development of heterogeneous catalysts for

Катализ в химической и нефтехимической промышленности UDC 542.934.8’7:547.261:547.271 DOI 10.18412/1816-0387-2018-4-6-30

DEVELOPMENT OF HETEROGENEOUS CATALYSTS FOR DEHYDRATION OF METHANOL TO DIMETHYL ETHER: A REVIEW © 2018 y.

Hamed Bateni1*, Chad Able2

1

Department of Chemical and Biological Engineering, Iowa State University, United States of America 2 Department of Chemical and Biomolecular Engineering, Ohio University, United States of America

Dimethyl ether (DME) is a promising multisource and multipurpose clean fuel and value-added chemical synthesized from syngas. This process can be either performed in a single stage (direct process) using a dual catalysis system or a two stage (indirect process) where syngas is first converted into methanol and then dehydrated to produce DME. While the dehydration reaction has been studied extensively over multiple decades, to date no review has been conducted on the catalysts involved in the methanol dehydration reaction. This work demonstrates the state of the art in catalyst preparation and analysis for this application. The dominant catalysts are studied extensively in this work, including γ-Al2O3 and various zeolites, such as ZSM-5, Y, beta and mordenite as well as their relevant modifications. Additionally, silicaalumina, mesoporous silicates, aluminum phosphate, silicoaluminophosphates, heteropoly acids (HPAs), metal oxides, ion exchange resins and quasicrystals are discussed in this work, owing to the wide variety of catalysts available and studied for the purposes of methanol dehydration to DME. Keywords: dimethyl ether, methanol dehydration catalysts, DME synthesis catalysts.

1. Introduction Increase in global energy demand along with the environmental issues of regular fossil fuels have prompted a worldwide pivot toward alternative fuels [1—3]. However, the research has confirmed that fossil based fuels will still dominate for several decades [4]. Therefore, clean burning fuels are considered key in alleviating the environmental challenges normally associated with fossil fuels [5, 6]. Dimethyl ether (DME) is one of the most promising clean alternative fuels [5]. DME does not form any explosive peroxides unlike other ethers, facilitating handling and storage [7]. The carbon monoxide and unburned hydrocarbon emissions of DME are significantly lower than those of natural gas due to the absence of C—C bonds, presence of C—H and C—O bonds, and relatively high oxygen content (about 35 %) in its structure [7, 8]. DME has a higher cetane number than conventional petroleum

Hamed Bateni – Research Assistant, Department of Chemical and Biological Engineering, Iowa State University, Ames, Iowa 50011, USA. Phone: +1-740-590-7588. E-mail: [email protected] Chad Able – Research Assistant, Department of Chemical and Biomolecular Engineering, Ohio University, Athens, Ohio 45701, USA. E-mail: [email protected]

6

diesel with a similar auto-ignition temperature [8, 9]. Moreover, burning DME does not produce any emission of particulate matter and toxic gases such as NOx [10—13]. Owning similar properties as propane and butane, similar vapor pressure to that of LPG, and ease of liquefaction (at pressures above 0.5 MPa), DME can take advantage of available infrastructure for LPG, making it a promising candidate as an alternative fuel [14]. DME is also an important intermediate to produce other value-added chemicals such as olefins [15]. DME is produced through a direct or indirect production process [7, 16]. The indirect process includes the conversion of synthesis gas (or syngas) to methanol followed by the dehydration of methanol to DME (MTD process) [9]. The direct process converts syngas to methanol and dehydrates methanol to DME in the same reactor, known in literature as the syngas-to-DME (STD) process [15]. Regardless of the synthesis method, methanol dehydration to DME needs to be catalyzed with an acidic catalyst, such as γ-Al2O3 [17]. In addition, methanol dehydration is an exothermic reaction; therefore, DME production is thermodynamically favorable at lower temperatures. The reaction at higher temperatures may lead to the formation of other hydrocarbons and coke [14].

Катализ в промышленности, т. 18, № 4, 2018

Катализ в химической и нефтехимической промышленности Although there is a virtual consensus regarding the surface reaction as the rate-controlling step, the methanol adsorption mechanism is still controversial [18—22]. The dissociative adsorption and molecular (associative) adsorption are the most popular mechanisms that have been proposed for alcohol dehydration (Fig. 1) [18, 23, 24]. In the associative (concerted) pathway, two molecules of methanol are adsorbed and directly react to form water and DME. In the case of the dissociative (stepwise) pathway, methanol adsorption, followed by water elimination, leads to an adsorbed methyl group and water. The methyl group then reacts with another methanol molecule to form DME [18]. Several studies have been performed to recognize the preferred route for methanol dehydration [19—22, 25, 26]. Blaszkowski and Santen [21] used a 3T sites cluster approach in their density functional theory (DFT) study to identify the mechanism of methanol to DME dehydration over H-ZSM-5. They concluded that the associative pathway was the favorable route for the MTD reaction [21]; however, the periodic nature of zeolite structure undermines the reliability of a simple cluster model [20]. A later work by Moses and Norskov [20] using a periodic DFT model to study the reaction mechanism over ZSM-22 showed that the dissociative pathway is predominant. A van der Waals corrected periodic DFT study of the MTD reaction over H-ZSM-5 showed that the associative pathway is preferred for a typical set of conditions; however, a transition in the mechanism from the associative to dissociative route was observed at elevated temperatures. Therefore, at a range of temperatures both mechanisms may be active at different catalytic sites [26]. The periodic gradient-corrected DFT calculated MTD reaction over silica-supported Keffin tungsten HPA clusters demonstrated that the associative mechanism was the main

route [19]. Alharbi et al. [18] showed that the MTD reaction using both the HPA and H-ZSM-5 catalysts follows a similar mechanism. Therefore, the intrinsic mechanisms of the surface reaction depends on the nature of the catalyst and the type of active sites, which could also shift at elevated operating conditions resulting in development of different kinetic expressions [19—21, 25—28]. Different catalysts have been utilized for methanol dehydration including bulk and modified γ-Al2O3 [7, 27, 29—34], bulk and modified H-ZSM-5 [35—49], heteropoly acids (HPAs) [18, 50—54], silicoaluminophosphates (SAPOs) [55—59], aluminosilicates [60—64], and ion exchange resins [35, 65—68]. However, the research continues for finding a proper catalyst with promising properties including satisfying stability, proper acidity, high activity, hydrophobic surface, high selectivity toward DME, and low production (and recovery) cost. The article published by Dr. James J. Spivey is perhaps the very first review on dehydration catalysts for methanol to DME conversion [69]. Sun et al. [70] has published a great review on the catalysis chemistry of DME synthesis and briefly discussed the chemistry of the reaction over alumina and zeolite and how the modification affects the chemistry of MTD. Sarvanan et al. [71] has recently published a review paper focusing on the progress of the heterogeneous bifunctional hybrid catalysts for single-stage DME synthesis from syngas (STD process) in which alumina and zeolite were also discussed as the common representatives of methanol dehydration catalysts. However, to the best of our knowledge, no review has been published over the past 25 years focused on the development of the catalysts for the methanol to DME reaction. This study presents a comprehensive overview on the development of different solid catalysts for methanol dehydration to provide an insight for further investigation in the field according to the identified gaps. Despite the frequent use of alumina and zeolite for methanol dehydration, they do not meet all the required characteristics of an ideal catalyst for this reaction. It is our opinion that a supported catalyst possessing adequate acidity (both Lewis and Brønsted sites), stability (hydrothermal and mechanical), and textural and structural properties can be more appropriately tuned for the MTD reaction. Meanwhile, the synthesis method, i. e. vapor or liquid phase, and their characteristics should be taken into account for catalyst selection and synthesis.

2. Aluminum oxide Fig. 1. Methanol to DME conversion through associative and dissociative pathway (Reprinted with permission from reference [18]. Copyright © 2015, American Chemical Society)

Kataliz v promyshlennоsti, vol. 18, № 4, 2018

Aluminum oxide or alumina is a low cost chemical compound with a wide range of applications due to reasonable chemical and thermal stability [72—74]. In general, as a hydrous alumina precursor is heated, hydroxyl groups are

7

Катализ в химической и нефтехимической промышленности

Fig. 2. Decomposition sequence of hydrous alumina during heating up to 1200 °C (Reprinted with permission from reference [75]. Copyright © 2003, John Wiley & Sons) driven off leaving a porous solid structure of activated alumina [75]. Fig. 2 shows the decomposition sequence of several hydrous precursors with the approximate temperature of phase transitions. The presence of corundum and transition structures provide different important properties in the alumina [76]. The α-Al2O3 (corundum form) possesses closely packed oxygen ions in a hexagonal structure, resulting in promising mechanical, electrical, and thermal properties [77]. On the other hand, γ-Al2O3 owns a transition structure with a cubic formation of closely packed oxygen ions resulting in a mesoporous structure with high surface area and surface acidity [77, 78]. Therefore, γ-Al2O3 has gained a great deal of attention as an adsorbent or a solid acid catalyst [76]. γ-Al2O3 can be produced by dehydration of a precursor (boehmite or gibbsite) or crystallization of amorphous aluminum oxide [78, 79]. It is noteworthy that the properties of the precursor can significantly affect the properties of the final catalyst [75, 80]. Therefore, Abbattista et al. [78] developed a method for preparation of high surface area amorphous aluminum oxide, in which thermal decomposition of aluminum nitrate in the presence of tartaric acid at 180 °C was followed by further thermal and hydrothermal treatments. Although applying a higher temperature decreased the surface area of the final catalyst, this method secured the formation of gamma-phase alumina [78]. The alumina catalyst prepared from sodium aluminate possesses only weak acid sites while the alumina prepared

8

from aluminum isopropoxide provides strong acid sites, resulting in an extensive isomerization during alcohol dehydration which decreases the DME production yield [80]. Depending on the solution used for precipitation of alumina, i. e. ammonia or sodium hydroxide, the final catalyst may exhibit different activities. In fact, it was reported that NH4+–TiO2/Al2O3 revealed a higher activity than Na+–TiO2/Al 2O3 catalyst for methanol dehydration [37]. It is worth mentioning that, regardless of the calcination temperature, γ-alumina only possesses Lewis acid sites [81]. Increasing the calcination temperature may lead to increases in particle size [81] resulting in lower acidity [15, 81—83]. Alumina possesses a hydrophilic surface which makes it more susceptible to water due to its stronger affinity with the alumina surface. This results in the occupancy of the active sites with water and a lower catalyst activity for the methanol dehydration reaction [82—84]. The catalytic properties of alumina also depend on its crystalline structure and texture affected by the method of synthesis [82, 85]. Industrial scale production of γ-alumina is typically done through a sequential process in which oxy-hydroxide is subjected to precipitation, drying, and calcination [85]. However, other methods include hydrothermal treatment of an alkoxide derived-alcogel, sol-gel processing, and spray-pyrolyzed powder [86]. There are several reports in the literature on the effect of synthesis methods and their operating conditions on the final characteristics of γ-alumina [33, 72, 77, 85]. Appropriately chosen hydro-

Катализ в промышленности, т. 18, № 4, 2018

Катализ в химической и нефтехимической промышленности thermal conditions can result in successful control of pore size distribution in the final alumina [72]. Below, we discuss some of the most important synthesis methods in the literature for the MTD reaction. A great deal of effort has been devoted to synthesizing nanostructured alumina in which hypercritical drying of xerogels and surfactant are among the most important considerations in controlling the textural properties [15, 82, 83, 85]. Sanchez-Valente et al. [85] synthesized nanostructured alumina using sol-gel and investigated the effect of aging on the properties of the final catalyst. This concept was integrated in preparation of an alumina-based mesoporous catalyst via a three-step sol-gel method using nitric acid and a cationic surfactant (hexadecyltrimethylammonium bromide) as a template to prepare mesoporous nanocrystalline γ-Al2O3 [33]. The catalyst possessed a mesoporous structure with a high specific surface area and a broad pore size distribution with uniform size and shape. The use of surfactant increased the concentration of weak to medium strength sites while limiting the strong sites, resulting in a superior performance for the MTD reaction at 300 °C and atmospheric pressure [33]. The mesoporous nanostructured alumina can be also synthesized via an acid free sol-gel method in which the final catalyst still possesses a high surface area and uniform porous structure [34]. The precipitating agent used in co-precipitation of a γ-Al2O3 nanocatalyst also has a salient effect on the performance of the catalyst. Hosseini and Nikou [15] reported ammonium salts resulting in higher agglomeration, surface area, and acidity compared to sodium salts (carbonate or bicarbonate). The relatively poor performance of alumina prepared by co-precipitation of NaAlO2 suspension with HCl for the MTD reaction strengthened such a conclusion [36]. Simultaneous hydrolysis and condensation reactions employed through the sol-gel method offer a promising opportunity for controlling the physical, chemical, and textural properties of the final catalyst. This method usually employs an aluminum precursor (e. g. aluminum iso-propoxide, AIP), a hydrolysis rate controller (e. g. acetic acid, AA), and an organic solvent (e. g. 2-propanol, IPA) to synthesize the intermediate gel for catalyst preparation. Water has important effects on the degree of hydrolysis, the species initially formed, and the subsequent condensation reaction. Acetic acid effectively controls the morphology of the structure due to selective adsorption on high-energy sites of the boehmite particles nucleated during aging. Therefore, it is possible to adjust the crystallinity of the boehmite (and thus γ-Al2O3 produced via thermal decomposition of boehmite) to acquire high surface area nanosized particles. This also improves the concentration of the acid

Kataliz v promyshlennоsti, vol. 18, № 4, 2018

sites in the final catalyst, leading to a promising catalytic performance at lower operating temperatures with respect to commercial γ-Al2O3 [86]. Liu et al. [87, 88] introduced a novel method for in-situ preparation of a slurry catalyst for methanol dehydration in the liquid phase. They prepared the aged gel similarly to the aforementioned method and directly dispersed it into liquid paraffin with treatment under N2 for 12 h (without prior calcination). AlOOH was the main phase of the catalyst despite its transformation to γ-Al2O3 during the reaction. The fresh catalyst possessed a mesoporous structure with an average pore size of 1.9 nm and was rich with moderate strength acid sites [87]. There have been great efforts in modifying the alumina with different agents to improve its performance for the MTD reaction. Although phosphoric acid modification was traditionally a common method of improving the performance of γ-Al2O3, modification with titanium oxide, titanium sulfate, niobium oxide, boric acid, and halogenating agents showed superior results [32, 89—91]. Table 1 shows a list of modified alumina used in methanol-to-DME conversion. Besides the modifying agent listed in Table 1, C2H2NbO4 [89], (NH4)2SO4 [97], Na2CO3 [42], and CH2O [42] were also used to modify alumina for dehydration reaction in the single-stage DME synthesis process. The modifying agent can incorporate into the structure and increase the Brønsted acidity of the catalyst (e. g. in the case of sulfate solution) [94] or replace the external hydroxyl group (e. g. in the case of fluorine solution) and destroy oxy-bridges in the alumina framework [32]. The concentration of solution needs to be optimized in order to prevent an extensive erosion of the surface resulting in demolition of small pores, widening the pores, and potentially weakening the structure of the catalyst [32]. Silica based materials were also used to modify the alumina and improve the stability of the catalyst [31, 95, 98, 99]. Mollavali et al. [98] showed that an increase in the Si content resulted in an increase in the BET surface area, pore volume, strength of acid sites, and total acidity of the catalyst. However, increasing the Si content over 6 % decreased the activity of the catalyst for DME formation (lower methanol conversion despite similar DME selectivity). Alumina-based mesoporous materials have also attracted attention as catalysts or catalyst supports [100—102]. Among them, boria-alumina catalysts have been used as an active catalyst in different reactions [103—108]. Xiu et al. [96] prepared a boria-alumina composite using a combination of surfactant templating, sol-gels, and evaporation-induced self-assembly (EISA). Therefore, the composite itself had acidity as expected from a non-siliceous oxide material [109—112]. The catalyst showed thermal stability in the

9

Катализ в химической и нефтехимической промышленности Table 1 The performance of modified alumina in methanol-to-DME reaction Feed composition

Feed rate

MeOH conversion, %

DME selectivity, %

1.5

MeOH

5.6 gcat·h·(molMeOH)–1

~30.0

98.7

[36]

290

0.1

MeOH

(LHSV) 0.9 h–1

86.1

100

[92]

–

300

–

11 % MeOH in N2

(WHSV) 1 h–1

76.2

100

[93]

5 % Ti(SO4)2/γ-Al2O3

Ti(SO4)2

240

NA

21 % MeOH in N2

(GHSV) 3.4 L·g–1cat·h–1

~85

99.99

[94]

5 % B2O3/γ-Al2O3

H3BO3

250

0.1

13.3 % MeOH in N2

(GHSV) 3.6 L·g–1cat·h–1 N2b

49.1

NA

[95]

Modified catalyst

Modifying solution

γ-Al2O3

–

275

γ-Al2O3

–

γ-Al2O3

T, °C P, MPa

Boria-γ-Al2O3

H3BO3

350

NA

0.1 % MeOH in N2

85

100

[96]

5 % ZrO2/γ-Al2O3

ZrO(NO3)2·6H2O

250

0.1

13.3 % MeOH in N2a

(GHSV) 3.6 L·g–1cat·h–1 N2b

60.3

NA

[95]

1 % SiO2/γ-Al2O3

SiO2

250

0.1

13.3 % MeOH in N2a

(GHSV) 3.6 L·g–1cat·h–1 N2b

74.4

NA

[95]

0.1

a

(GHSV) 3.6 L·g–1cat·h–1 N2b

72

NA

[95]

b

71.7

NA

[95]

(GHSV) 3.6 L·g–1cat·h–1 N2b

37.5

NA

[95]

b

29.5

NA

[95]

27.1

NA

[95]

5 % SiO2/γ-Al2O3

SiO2

250

(WHSV) 4 h

–1

Ref.

13.3 % MeOH in N2

–1

a

10 % SiO2/γ-Al2O3

SiO2

250

0.1

13.3 % MeOH in N2

1 % SiO2/γ-Al2O3

SiO2

250

0.1

8.9 % MeOH, 8.6 % Water in N2a

(GHSV) 3.6 L·g a

–1

cat·h

–1

–1

5 % SiO2/γ-Al2O3

SiO2

250

0.1

8.9 % MeOH, 8.6 % Water in N2

(GHSV) 3.6 L·g

10 % SiO2/γ-Al2O3

SiO2

250

0.1

8.9 % MeOH, 8.6 % Water in N2a

(GHSV) 3.6 L·g–1cat·h–1 N2b

с

Fluorinated-γ-Al2O3

NH4F

250

0.1

MeOH in N2

Fluorinated-γ-Al2O3

NH4F

400

NA

MeOH in Arс

NA

с

Chlorinated-γ-Al2O3

NH4Cl

400

MeOH in Ar

cat·h

N2

(LHSV) 9.27 h

–1

N2

~81

NA

[32]

4.98·10–2 mol·h–1

~82

100

[90]

–2

~90

100

[90]

4.98·10

mol·h

–1

a

Value was calculated based on the data reported in the reference. Calculated based on the data available on the original manuscript (the density of methanol at room temperature was used if required). c There was not enough data available on the original manuscript to calculate the percentage of methanol in nitrogen. NA – not available. b

mesoporous structure in the temperature range of 225— 375 °C where the methanol conversion reached about 85 % with 100 % DME selectivity [96]. Although γ-alumina was the most common alumina to be used as an MTD catalyst, other crystalline phases, i. e. η-Al 2O3, θ-Al 2O3, (χ+γ)-Al 2O3, δ-Al 2O3, α-Al 2O3, and κ-Al 2O3, have also been evaluated for dehydration of methanol to DME [81, 113]. The catalytic activities in the MTD reaction decreased in the order of η-Al 2O3, γ-Al 2O3, θ-Al 2O3, (χ+γ)-Al 2O3, δ-Al 2O3, α-Al 2O3, and κ-Al 2O3 wherein the latter two did not show any catalytic properties [113].

3. Zeolites Zeolites are porous aluminosilicates, natural or synthetic, that have achieved prolific use as catalysts and adsorbents. They are comprised of silicates (SiO4) and aluminates (ions AlO4–) linked in a tetrahedral fashion via oxygen atoms. The discovery and historical uses of zeolites, while interesting, fall outside of the scope of this paper; however, the use of zeolite materials to synthesize hydrocarbons through dehydration of organic compounds has been studied for almost a century [114].

10

There are many different zeolites used in the methanol dehydration reaction, including ZSM-5 [35—49], Y [35, 45, 48], clinoptilolite [115, 116], and mordenite [35, 48, 117]. Many of these catalysts are used in their H-form, i. e., H-ZSM-5, H-Y, and H-mordenite. To produce this H-form of ZSM-5, for example, the typical strategy begins with a commercial NH4 -ZSM-5 catalyst with a given ratio of silica to alumina; this catalyst is calcined to decompose the ammonium groups and protonate the zeolite [18, 39, 40, 45, 46]. The process is roughly identical for producing H-Y zeolite [45, 94]. At times, the catalyst is first obtained in its Na-form (Na-ZSM-5); this is ion exchanged with ammonium nitrate prior to calcination [47]. Unlike the γ-alumina counterpart, zeolite catalysts possess Brønsted acidity in addition to Lewis acidity; this is due to the presence of the O atoms in the zeolite structure which weaken the proton’s bond to the zeolite via partial electron transfer [118]. This leads to a preference for methanol over water, meaning that the zeolite variants do not experience dramatic deactivation when water is present in the feedstock, nor during the methanol dehydration reaction [36, 40]. This is believed to be due to the water complexes that form on the acid sites of zeolite catalysts, acting as mild donors [119]. This is a strong advantage considering the

Катализ в промышленности, т. 18, № 4, 2018

Катализ в химической и нефтехимической промышленности production of water in methanol synthesis from syngas as well as the dehydration to DME [37]. On the other hand, the selectivity to DME is a fair bit lower in general, as the zeolite activity will tend to make longer chain hydrocarbons from the DME unless the reaction parameters are carefully tuned to prevent this [36, 40, 42, 119]. This is additionally dependent on the morphological characteristics of the zeolite structure [115]. Lopez [119] notes that the microporous nature of mordenite results in a rapid deactivation of the catalyst due to coking, whereas ZSM-5 possesses a much lower decoking temperature. This phenomenon starts to occur around 250 °C [18]. In addition, the formation of aforesaid hydrocarbons may lead to coking on the surface of the zeolite catalyst [119], although this can partially passivate the catalyst, lowering the activity and increasing DME selectivity in longer time-on-stream tests [40]. The effect of micro- and mesoporosity plays a role in both DME selectivity and CO conversion, rising to paramount importance in catalyst modification which will be elucidated further. Since these zeolite catalysts are comprised primarily of silicate and aluminate covalent bonds, the ratio of silica to alumina (or Si/Al ratio hereafter) warrants further discussion. Decreasing the Si/Al ratio increases the amount of active alumina within the zeolite structure, increasing the acid strength; this peaks around a Si/Al ratio of 30 for H-ZSM-5 as per Alharbi et al.’s findings [18]. This leads to a direct increase in the activity of the methanol dehydration reaction; however, as with other methods of increasing activity with zeolite catalysts, this usually decreases the dimethyl ether selectivity in both ZSM-5 [40] and ferrierite structures [120]. In addition, zeolites which possess high alumina content will experience partial deactivation in the presence of high partial pressures of water in the feed, due to the presence of Lewis acid sites; this is not as dramatic as the deactivation of gamma-alumina [37]. As a result of the issue of selectivity, altering the acidity and the pore size distribution of the zeolite catalysts becomes paramount in ensuring optimum dimethyl ether synthesis. Sodium modification of zeolite catalysts is a common strategy [36, 39—41]. This is believed to partially block the Brønsted acidity that is believed to contribute to the synthesis of hydrocarbons. Kim et al. [39] examined FTIR spectra for the catalyst variants and found that the 1540 cm–1 band associated with Brønsted acidity disappeared with the addition of Na. Another strategy in doping the ZSM-5 catalyst is reducing micropores in favor of mesopores [39, 44]. Since micropores are believed to be diffusion limited in the methanol dehydration reaction, and thus lead to unwanted products such as hydrocarbons and coke formation, a method of either a) widening the pore diameter or b) supplanting micropores with mesopores via a molecular

Kataliz v promyshlennоsti, vol. 18, № 4, 2018

sieve is sought after [44]. This ties into the above concept of reducing strong acid sites; this method of increasing mesoporosity generally decreases overall acidity of the catalyst. This strategy of metallic modification of the existing zeolite structure is a major focus for current research. Khandan et al. [121] carried out the methanol dehydration reaction over a variety of catalysts including ZSM-5, Y, mordenite, ferrierite, and beta zeolite as well as silica and alumina. The reactions were carried out at 523 K and 3 MPa in a batch reactor with a starting methanol concentration of 0.962 mol/L. Under these conditions, Na- and H-mordenite had the highest DME selectivity. Interestingly, increasing the Si/Al ratio decreased both conversion and selectivity, i. e. methanol dehydration was hindered as well as producing more unwanted hydrocarbons. Following the success of the H-mordenite, various improvements were attempted with Cu, Mg, Ni, Zn, Al, Zr and Na-H-mordenite being synthesized via wet impregnation. All of these modifications caused a decrease both in BET surface area and total pore volume. The NH3-TPD analysis produced qualitative data on the relative strength of the weak, medium and strong acid sites corresponding to peaks between 100—300 °C, 300— 500 °C, and 500—700 °C, respectively. While the weak and medium acid sites increased depending on the metal used, all of the modifications experienced a sharp decrease in the strength of strong acid sites. The selectivity toward DME is correlated with the total of weak and medium acid sites and inversely correlated with the strong acidity. In addition, the Al, Zr, Ni and Na-modifications experience much higher stability than the other modifications and in stark contrast to the other zeolites. Hassanpour et al. [122] compared the effectiveness of ZSM-5 and Mordenite zeolite structures in the methanol dehydration reaction. The ZSM-5 structure (here denoted MDHC-1) outperformed the others but experienced partial deactivation at temperatures past 300 °C. Na-modified MDHC-1 possessed lower surface areas and pore volumes. The Na modifications generally outperformed the unmodified MDHC-1 except for 120 mol.% Na; this is in spite of loss of acidity. The modifications also improved selectivity to as high as 99.9 %. Further work by this group [123] tested a series of mordenite zeolites. Here, the H-mordenite exhibits the highest stability after 180 hours as well; however, it experiences partial deactivation in 20 wt.% water. Because of the coke formation in the H-mordenite supercage, the catalyst cannot be regenerated using water (i. e. the carbon is trapped once deposited). With zeolites beside ZSM-5, other interesting modifications can be made. Zheng et al. [124] synthesized a composite zeolite with Beta and Mordenite zeolite structures (denoted as BMZ). This catalyst was synthesized via a mixture of pre-synthesized Beta zeolite with sodium aluminate and

11

Катализ в химической и нефтехимической промышленности varying masses of sodium hydroxide; this step controlled the alkalinity of the catalyst which tuned the performance. Nitrogen adsorption/desorption data provides insight to the degree of micro- and mesoporosity within the different catalysts; while microporosity remains the main contributor to surface area (micropore volume at a level of 0.16—0.19 cm3/g), the enlargement of micropores results in a degree of mesoporosity between 0.04 and 0.06 cm3/g. Increasing amounts of sodium hydroxide in the BMZ formation led to a decreasing BET surface area and pore volume as expected. As compared to a mechanically mixed composite of Beta and Mordenite, as well as compared to a ZSM-5 catalyst with a Si/Al ratio of 50, the BMZ catalyst attained equilibrium conversion at lower temperatures and maintained selectivity to DME at higher temperatures than its counterparts. Later work within the same research group [125] involved a porous composite zeolite BFZ with Beta zeolite cores and Y zeolite polycrystalline shells as a method of improving on the typical H-Y zeolite. An H-Y catalyst was compared to this new catalyst as H-BFZ-1 and H-BFZ-2 (denoting crystallization times of 24 and 28 h) and physically mixed with a methanol synthesis catalyst for use in a synthesis gas to dimethyl ether reaction. H-Y has a much higher surface area and 3 times the number of micropores, roughly; however, the composites boast a much higher mesopore volume (0.19 and 0.15 vs. 0.02). When compared to these BFZ composites, H-Y possesses the most acid sites and the strongest acid strength. However, the composite zeolites have a much higher CO conversion (94.2 and 89.8 vs. 79.9 %) and a slightly higher selectivity toward DME. This lends to the idea that: 1) closely calibrating acid strength is necessary for DME formation; 2) mesoporosity is more beneficial than microporosity when considering the synthesis of dimethyl ether, due perhaps to diffusion limitations. Fei et al. [126] compared standard H-Y zeolite to various modifications including Fe, Co, Ni, Cr, and Zr-modified H-Y zeolite in the methanol dehydration reaction. This was also used in the STD reaction using an H-Y supported Cu/Mn/Zn catalyst. The strong acid peak increased with the Fe and Cr modifications and slightly increased with the Co and Zr modifications. The middle strength peak is also more pronounced in the Zr modification. In time-onstream tests of up to 15 h on the various catalysts in methanol dehydration (carried out at H2/CO — 1.5, temperature — 245 °C, pressure — 2 MPa, space velocity — 1500 h–1), only the Ni and Zr modifications maintained a high methanol conversion of ~86 %; this correlates with the amount of coking found on each catalyst. More work from the same group [127] used rare earth metals, namely, La, Ce, Pr,

12

Nd, Sm, and Eu modifications on the Y zeolite. By way of XRD data analysis, it is suggested that the modification via rare earth metals successfully penetrates the Y supercage. When looking at ammonia TPD profiles (Fig. 3), the middle strength acid peak (between 340 and 490 °C) was more pronounced in the case of La, Ce and Pr modifications. In addition, the strong acid sites increased in Nd and Sm modified Y. Eu-Y experienced little change in acid site strength. With regards to conversion, H-Y experienced a drastic decrease in conversion of methanol after only 10 hours on stream; this was improved in the exchange of all the rare metals, especially La, Ce, Pr and Nd-Y with La-Y maintaining the highest methanol conversion after 10 hours with only a 3 % loss. In addition, all of these modified zeolites had a higher initial conversion, owing to the higher acid strength of the modifications; 87.5 % for H-Y zeolite vs. 92.0—94.6 % for all of the rare earth modifications. As well as impregnation with rare earth metals, research has been conducted regarding impregnation with halogen ions, hereafter referred to as halogenation. Conducted at the Egyptian Petroleum Research Institute, Aboul-Fotouh et al. [128] treated mordenite with ammonium chloride and fluoride, as well as hydrochloric and hydrofluoric acids, in

Fig. 3. NH3–TPD data for H-Y zeolite (a) and its rare earth modifications: b – La-Y, c – Ce-Y, d – Pr-Y, e – Nd-Y, f – Sm-Y, and g – Eu-Y (Reprinted with permission from reference [127]. Copyright © 2007, Elsevier)

Катализ в промышленности, т. 18, № 4, 2018

Катализ в химической и нефтехимической промышленности Table 2 The performance of zeolite catalysts in methanol dehydration Modified catalyst

Modifying solution T, °C P, MPa

Feed composition

Feed rate

Methanol DME conver- selec- Ref. sion, % tivity, %

Na-modified H-ZSM-5

NaNO3

270

0.1

Methanol (crude) in N2

(LHSV) 10 h–1

~83

100

[38]

NaHZSM-5/γ-Al2O3

NaNO3 + γ-Al2O3

250

1.01

Methanol (crude)

(LHSV) 10 h–1

82

100

[39]

–1

NaH-ZSM-5

NaNO3

250

NA

Methanol in N2

(WHSV) 4 h

~60

100

[40]

Na-modified H-ZSM-5

NaNO3

300

1.6

Methanol in N2

(LHSV) 3.8 h–1

98

100

[122]

NaH-ZSM-5

NH4Cl

260

1.5

Methanol

(ST) 5.6 gcat·h·mol–1

~90

99

[36]

HCHO-modified H-ZSM-5

Na2CO3 + HCHO

260

5.51

54.5 % H2, 45.5 % CO

NaOH-treated NH4-ZSM-5

NaOH

260

1.5

Methanol

HFeZSM-5 HFeAlZSM-5 H-ZSM-5/MCM-41 MgO-modified H-ZSM-5

Fe(NO3)3

4500 mL·h

(ST) 5.6 gcat·h·mol

260

4

66.6 % H2, 33.3 % CO

(GHSV) 1500

Fe(NO3)3 + Al(NO3)3 260

4

66.6 % H2, 33.3 % CO

(GHSV) 1500

0.1

Methanol in N2

NaOH + CTAB Mg(NO3)2

220 260

4

66 % H2, 30 % CO, 4 % CO2

Sb2O3-modified H-ZSM-5

Sb2O3

260

4

61.4 % H2, 28.5 % CO, 2.8 % CO2, 7.3 % N2

Fluorinated H-ZSM-5

NH4F

275

NA

Methanol in Ar

·gcat–1 –1

–1

cm ·h ·gcat–1 cm3·h–1·gcat–1 3

–1

0.1 mL/min ·gcat–1

1500 mL·h

–1

1500 mL·h

–1

·gcat–1

4.98 · 10–2 mol/h –2

4.98 · 10

mol/h

67.0

*

~92 94.09

*

*

79.5

[42]

99

[36] *

54.9

[43]

95.49*

67.06*

[43]

~90

100

[44]

96.3

*

*

95.0

*

[47]

*

69.0

[131]

~90ab

NA

[129]

~80a

NA

[129]

64.5

Chlorinated H-ZSM-5

NH4Cl

275

NA

Methanol in Ar

Silylated H-ZSM-5

TEOS

260

2

66.6 % H2, 33.3 % CO

(WHSV) 1.7 h–1

~72*

63.5*

[46]

H-ZSM-5/H-Y (BFZ)

AlNaO2

250

5

60.8 % H2, 27.2 % CO, 4.8 % CO2, 3.2 % Ar

(SV) 1500 h–1

94.2*

67.9*

[125]

Fe-modified H-Y

Fe(NO3)3

245

2

60 % H2, 40 % CO

(SV) 1500 h–1

22.6*

–1

65.2*

[126]

*

37.5*

[126]

13.3*

[126]

*

[126]

Co-modified H-Y

Co(NO3)2

245

2

60 % H2, 40 % CO

(SV) 1500 h

8.7

Ni-modified H-Y

Ni(NO3)2

245

2

60 % H2, 40 % CO

(SV) 1500 h–1

7.5*

–1

*

Cr-modified H-Y

Cr(NO3)3

245

2

60 % H2, 40 % CO

(SV) 1500 h

70.9

Zr-modified H-Y

Zr(NO3)4

245

2

60 % H2, 40 % CO

(SV) 1500 h–1

71.4*

66.7

67.5*

[126]

Fluorinated H-Y

NH4F

275

NA

Methanol in Ar

4.98·10–2 mol/h

~70a

NA

[129]

Chlorinated H-Y

NH4Cl

275

NA

Methanol in Ar

4.98·10–2 mol/h

~70a

NA

[129]

La-modified H-Y

3+

La

245

NA

20 % MeOH in Ar

50 mL/min

92.20

97.40

[127]

Ce-modified H-Y

Ce3+

245

NA

20 % MeOH in Ar

50 mL/min

94.50

94.70

[127]

Pr-modified H-Y

2+

Pr

245

NA

20 % MeOH in Ar

50 mL/min

92.00

96.70

[127]

Nd-modified H-Y

Nd3+

245

NA

20 % MeOH in Ar

50 mL/min

94.60

92.70

[127]

Sm-modified H-Y

Sm3+

245

NA

20 % MeOH in Ar

50 mL/min

93.40

89.00

[127]

Eu-modified H-Y

3+

Eu

245

NA

20 % MeOH in Ar

50 mL/min

92.00

90.50

[127]

Cu-H-MOR

Cu(NO3)2

250

3

Methanol in isooctane

0.962 mol/L

95.10

87.10

[121]

Mg-H-MOR

NA

250

3

Methanol in isooctane

0.962 mol/L

92.90

85.20

[121]

Ni-H-MOR

NA

250

3

Methanol in isooctane

0.962 mol/L

89.50

92.90

[121]

Zn-H-MOR

NA

250

3

Methanol in isooctane

0.962 mol/L

97.40

90.80

[121]

Al-H-MOR

NA

250

3

Methanol in isooctane

0.962 mol/L

99.40

95.20

[121] [121]

Zr-H-MOR

NA

250

3

Methanol in isooctane

0.962 mol/L

98.80

94.10

Na-H-MOR

NA

250

3

Methanol in isooctane

0.962 mol/L

96.20

93.50

[121]

Fluorinated H-MOR

NH4F

300

NA

Methanol in Ar

4.98·10–2 mol/h

~90a

~98

[129]

Chlorinated H-MOR

NH4Cl

300

NA

Methanol in Ar

4.98·10–2 mol/h

~93a

~100

[129]

90

100

[124]

~27*

~77*

[116]

H-MOR/H-B (BMZ)

NaOH/AlNaO2

200

0.1

Methanol in N2

HNO3-modified Clinoptilolite

HNO3

275

4

66.6 % H2, 33.3 % CO

(WHSV) 2.26 g·h

–1

(GHSV) 600 cm3·h–1·gcat–1

*

This reaction was a single-stage synthesis gas to DME reaction – the conversions listed are for CO. Only a DME yield is reported which is shown here. b The numbers shown here are said to also be for Cl – this is assumed to be a typo. a

Kataliz v promyshlennоsti, vol. 18, № 4, 2018

13

Катализ в химической и нефтехимической промышленности order to evaluate the effect of halogenation and hydrohalogenation on the methanol dehydration reaction. The initial H-mordenite possessed a Si/Al ratio of 6 : 1; the subsequent doping methods increased this ratio. In conjunction with a loss of crystallite size, this led the authors to conclude that alumina had been removed from the zeolite framework via halogenation, replacing Si–O–Al bonds with shorter Si–O–Si bonds. At most of the temperatures studied between 100 and 300 °C, halogenation led to a higher conversion and a higher DME percentage in the product stream. Fluoridation’s effect was more significant than chlorination, and the acid form of the treatment (hydrohalogenation) superseded the ammonium treatment. However, selectivity was lost at higher temperatures. This work was continued in 2016 [129] with ZSM-5, mordenite and Y zeolites, additionally considering the potential impact of sonication on catalyst preparation. The sonicated Cl-H-mordenite possessed a much larger surface area (0.44 vs. 0.22 cm3/g) than its unsonicated counterpart; this effect is far less pronounced in F-H-mordenite. Ultrasonication (the effects of sonication and ultrasound irradiation) increases the activity of the chlorinated mordenite and zeolite but decreases the activity of the fluoridated counterparts; this may correlate with the increased catalyst surface area. Both chlorination and fluoridation increase the activity of the H-ZSM-5 catalyst; similar behavior is seen in H-Y. The difference in activity may originate from the increased number of acid sites due to chlorination and the increased acid strength from fluoridation; this ultimately follows the effects of other methods of doping on zeolites. Seo et al. [130] used the synthetic W zeolite for methanol dehydration. W zeolite possesses a low Si/Al ratio. The zeolite structure could not be decomposed in either case up to 800 °C, indicating stability for the W zeolite. Two methods of preparation were used: a hydrothermal treatment and a microwave treatment; the HT treated W zeolite had conversions 3—43 % from 250 to 325 °C, and the microwave treated W had 0—28 % conversion from the same temperature range. Clinoptilolite is a common natural zeolite with a low Si/Al ratio and large intersecting open channels of 8- and 10-member tetrahedral rings. Royaee et al. [115] utilized a clinoptilolite zeolite catalyst with an Si/Al ratio of 5.78 for methanol dehydration. The reactions were conducted at 350 °C and atmospheric pressure, as well as a partial pressure for methanol of 50 kPa and a WHSV of 4.78 h–1. In this study, the calcination temperature, ion-exchange solution concentration, calcination time and solution type were varied in the catalyst’s preparation. The authors found that, among other factors, ion-exchange solution concentration and calcination times were the strongest factors in the catalyst yield.

14

Khoshbin, et al. [116] tested the synthesis of DME on admixed nanocatalysts of CuO–ZnO–Al 2O3 and HNO3-modified clinoptilolite. The natural clinoptilolite was treated with a 5 N HNO3 solution via ion exchange prior to rinsing and subsequent calcination treatment. The co-precipitated clinoptilolite/CZA catalyst was calcinated a second time before nanocatalyst formation. The treatment increased natural zeolite surface area from 13 to 96.9 m 2/g. Higher pressures correlate with higher CO conversions and higher DME selectivities. The CZA/clinoptilolite activity is overall low but is able to maintain a high time-on-stream with no deactivation after 15 hours. These manipulations of zeolite catalysts are summed up succinctly in Table 2, with their relevant conditions and DME yields/selectivities reported, if available. These modifications have effects on the acidity and morphological characteristics of the zeolite catalyst, and should serve as a baseline for future investigations into zeolite modification should new information arise.

4. Silica-alumina Both the amorphous and zeolitic silica-alumina (e. g. ZSM-5 and aluminosilicate) can be used for methanol conversion [132]. Silica-alumina catalysts own strong Brønsted and Lewis acid sites where the Brønsted sites can be reversibly converted to Lewis sites by dehydration at elevated temperatures [37, 69, 133]. However, due to the abundance of strong acid sites, silica-alumina catalysts are more suitable for conversion of methanol to larger hydrocarbons than DME [132, 134]. Therefore, the strength of the acid sites need to be controlled for effective contribution in the MTD reaction [69, 135, 136]. Thus, selective poisoning of the strong sites using weak bases was introduced for silica alumina [69, 133, 137, 138]. Xu et al. [37] evaluated the effect of Si content on the performance of commercial amorphous silica-alumina. Surprisingly, they observed that increasing the Si content decreased the methanol conversion; this could be due to either the reduction of basic sites (as acid-base pairs are assumed to be the active sites for the reaction) or blockage of the active sites by coke formation at early stages of the reaction [37]. The same result was concluded by Jun et al. [95] in the case of modified alumina with silica. On the contrary, Takeguchi et al. [139] observed the superior performance of silica-rich silica-alumina catalyst in the single stage syngas to DME conversion. They prepared a series of silica-alumina catalysts with different silica content via mechanochemical activation of the mixture of the silica gel and Al(OH)3 precursor followed by calcination at 550 °C for 3.5 h. Increase in Si content led to an increase in BET surface area

Катализ в промышленности, т. 18, № 4, 2018

Катализ в химической и нефтехимической промышленности and the ratio of Brønsted/Lewis acid sites; however, it did not affect the acid strength of the catalyst. The abundance of Lewis acid sites in the catalyst with low Si content made it more susceptible to high partial pressures of water in the STD reaction [139]. Aluminosilicate also demonstrated a promising potential for DME synthesis due to the abundance of Brønsted acid sites on the catalyst; the Al/Si content is an additionally important factor in the activity of the catalyst [140, 141]. Varisli et al. [141] synthesized a series of mesoporous aluminosilicates with an Al/Si ratio in the range of 0.03—0.18 via hydrothermal synthesis of sodium silicate and aluminum nitrate in the presence of a surfactant. The catalyst with an Al/Si ratio of 0.09 possessed superior performance without the formation of any coke during the reaction (up to 400 °C). The methanol conversion and DME selectivity increased with reaction temperature and reached 78 % and 99 %, respectively, at 400 °C [141]. The same results were reported by Tokay et al. [8] for the mesoporous aluminosilicate catalyst with an Al/Si ratio of 0.1 in which tetraethylorthosilicate (TEOS) was used as Si source.

5. Mesoporous silicates Mesoporous materials, especially mesoporous silicates, have received a great deal of attention due to their unique characteristics including high surface area, large pore size, highly ordered pore structure (hexagonal arrangement) and narrow pore size distribution [93, 142—144]. Meynen et al.

[145] provide a great overview on synthesis of mesoporous materials. The discovery of the M41S family, including MCM-41 and the SBA-n group, opened a new chapter in catalysis research [8, 142]. The low acidity of MCM-41 and SBA-15 limits their direct application as a dehydration catalyst; however, it shows great potential as a catalysis support [44, 51, 52, 65, 146—151]. On the other hand, the acidity of the mesoporous materials can be improved through the insertion of foreign metal ions such as aluminum into the structure [93, 150—152]. Table 3 highlights some of the characteristics of the mesoporous silica catalysts used for the MTD reaction. For instance, Naik et al. [152] synthesized mesoporous Al-MCM-41 catalyst with a SiO2/Al2O3 ratio of 30 : 1 using aluminum isopropoxide and tetraethylorthosilicate (TEOS) in the presence of a structure-directing agent (hexadecyltrimethylammonium bromide, CTAB) and a co-surfactant additive (tetrapropylammonium hydroxide, TPAOH). The addition of TPAOH during the synthesis of the catalysts facilitated the incorporation of Al ions in a mesoporous aluminosilicate framework, increasing the acidity, BET surface area, and mesopore volume of the catalyst while decreasing the pore diameters [152]. MCM-41 usually suffers from low thermal stability [8]; however, such a deficiency can be mitigated by increasing the pore wall thickness and the extent of silica condensation through either hydrothermal treatment or the addition of TPAOH. It is noteworthy that the use of a large quantity of TPAOH may result in the formation of disordered mesoporous zeolitic silica [152]. In addition, the

Table 3 The mesoporous silicate catalysts in methanol-to-DME reaction Modified catalyst

Special synthetic characteristics

T, °C P, MPa

Feed composition

Feed rate

Methanol DME converselecsion, % tivity, %

Ref.

MCM-41

no Al insertion, no TPAOH

400

NA

97.48 mmHg MeOH in N2

14 L/(g·h)

6

100

[152]

Al-MCM-41 (MC-2)

SiO2/Al2O3 = 30, no TPAOH

400

NA

97.48 mmHg MeOH in N2

14 L/(g·h)

45

98

[152]

Al-MCM-41 (MC-3)

SiO2/Al2O3 = 30, CTAB/TPAOH = 0.75

400

NA

97.48 mmHg MeOH in N2

14 L/(g·h)

72

98

[152]

Al-MCM-41 (MC-4)

SiO2/Al2O3 = 30, CTAB/TPAOH = 0.2

400

NA

97.48 mmHg MeOH in N2

14 L/(g·h)

73

97

[152]

–1

HMS

no Al insertion, DDA surfactant

300

0.1

11 % MeOH in N2

1h

10

NA

[93]

Al-HMS-5

Si/Al = 5, DDA surfactant

300

0.1

11 % MeOH in N2

1 h–1

91.2

94.1

[93]

Al-HMS-10

Si/Al = 10, DDA surfactant

300

0.1

11 % MeOH in N2

1 h–1

89

100

[93]

Al-HMS-20

Si/Al = 20, DDA surfactant

325

0.1

11 % MeOH in N2

1 h–1

20

100

[93]

–1

16

100

[93]

80

99.99

[8]

Al-HMS-35

Si/Al = 35, DDA surfactant

375

0.1

11 % MeOH in N2

Al-SBA-15

Si/Al = 10, PEO-PPO-PEO surfactant

350

NA

50 % MeOH in He

1h

SBA-15-SO3H-Al

Si/Al = 4.08

300

0.1

4 % MeOH in He

20 ml/min

~83

~100

[150]

MCF-SO3H-Al

Si/Al = 3.42

300

0.1

4 % MeOH in He

20 ml/min

~82

~100

[150]

MSU-S

SiO2/Al2O3 = 55, CTAB surfactant

380

0.1

Pure methanol

5 h–1

75

97

[153]

13.3 L g–1cat·h–1

TPAOH – tetrapropylammonium hydroxide; CTAB – hexadecyltrimethylammonium bromide; Al-HMS – aluminated hexagonal mesoporous material; DDA – dodecylamine; PEO-PPO-PEO – poly (ethylene glycol)-block-poly (propylene glycol)-block-poly (ethylene glycol).

Kataliz v promyshlennоsti, vol. 18, № 4, 2018

15

Катализ в химической и нефтехимической промышленности Si/Al content has an important effect on the performance of the aluminated mesoporous silica catalyst. Increases in Al content could increase the lattice disorder, catalyst acidity, Lewis/Brønsted acid site ratio, and methanol conversion while decreasing the surface area, pore volume, and DME selectivity of the catalyst. Therefore, the aluminated catalyst with higher Al content can be more susceptible to water poisoning due to the preferential adsorption of water on the Lewis acid sites rather than Brønsted sites [93]. The catalyst with a Si/Al ratio of 10 showed promising results for the MTD reaction (89 % methanol conversion and 100 % DME selectivity at 300 °C) and reasonable stability for long-term operation. Zhao et al. [154] synthesized SBA-15 with an ordered pore structure with much thicker pore walls than MCM-41, leading to better hydrothermal stability. A triblock co-polymer with good amphiphilic characteristics is needed as the organic structure-directing agent during the synthesis process [8, 154]. Azimov et al. [151] reported that impregnating SBA-15 with titanium isopropoxide (C12H28O4Ti) in isopropanol (Ti content of 5—10 %) improves its performance for the MTD reaction. Alumina was also used to impregnate SBA-15 synthesized via a single-stage hydrothermal process without any significant distortion on the support structure, where the methanol conversion and DME selectivity reached 80 and 99 %, respectively, at 350 °C [8]. On the other hand, SBA-15, which is formed using a surfactant and TEOS (as a silica source) in an acidic media, is usually subjected to further processing to secure the acid sites on the substrate [144, 150]. This process includes the anchoring of 3-(mercaptopropyl)trimethoxysilane (MPTMS) to form surface thiol groups (–SH) followed by oxidation to convert the thiol groups to sulfonic acid groups (–SO3H) using H2O2 [144, 149, 150]. These sulfonic acid groups provide Brønsted acidity in the final catalyst (SBA-15—SO3H) [144]. However, the substrate may have only a limited number of weak acid sites [150]. Similarly, the acidic properties of mesostructured cellular foam (MCF), synthesized by addition of a swelling agent such as mesitylene (1,3,5-trimethylbenzene) during the synthesis of SBA-15, can be improved by securing the sulfonic acid group on its structure as can be seen in the NH3–TPD profile, which is reported in Fig. 4 [150]. SBA-15—SO3H and MCF—SO3H were used as catalyst supports for deposition of aggregated alumina species to make an active solid acid catalyst for the MTD reaction [149, 150]. After formation of sulfonic acid groups on the surface, aluminum Keggin oligocations were deposited on the mesoporous support through ion-exchange to reach 10 mmol Al3+/g modified silica [150] or 1 mmol Al3+/g modified silica [149]. The resulting catalyst (modified

16

Fig. 4. NH3–TPD profile of pure silica supports and the modified support with alumina species: a – SBA-15 based catalyst, b – MCF based catalyst (Reprinted with permission from reference [150]. Copyright © 2016, Elsevier) SBA-15) possessed uniform, cylindrical, and double-sided open ended mesopores. It was observed that both Lewis and Brønsted acid sites are available on the supported catalysts; however, the Lewis acid sites are available in greater quantities [150]. Both supported catalysts showed promising results for methanol conversion (higher than commercial γ-alumina) in the temperature range of 225—300 °C, especially SBA-15—SO3H—Al due to higher content of acid sites compared to MCF—SO3H—Al [149, 150]. The DME selectivity remained about 100 % for the reaction at temperatures up to 300 °C [149]. It could be also concluded that higher Al3+ concentration results in higher catalyst activity [149, 150]. MSU-S is another type of mesoporous molecular sieve with promoted thermal stability and acidity [155]. To synthesize this catalyst, tetrapropylammonium hydroxide

Катализ в промышленности, т. 18, № 4, 2018

Катализ в химической и нефтехимической промышленности (TPAOH) and hexadecyltrimethylammonium bromide (CTAB) were used as the structure director and surfactant, respectively [156]. The direct assembly of nanoclustered zeolitic precursors results in the presence of a certain degree of zeolite-like order in the framework wall, enhancing the catalytic performance of MSU-S [155, 157]. As can be seen in the Table 3, the MSU-S mesoporous aluminosilicate with a SiO2/Al2O3 ratio of 55 has a promising potential for catalyzing the MTD reaction even though a slight reduction in the selectivity was observed at elevated temperatures [153].

6. Aluminum phosphate Aluminum phosphate (AlPO4) is a promising catalyst in the MTD reaction with high selectivity, low coke formation, and water resistance [62]. Yaripour et al. [60] prepared a series of aluminum phosphates with different Al/P molar ratios (1, 2, and 3) using co-precipitation in which ammonia was added to a solution of aluminum nitrate nonahydrate (ANN) and phosphoric acid. Increases in aluminum content increased the surface area and the pore volume while decreasing the acidity of the catalyst. In fact, Brønsted acid sites with moderate acid strength were more abundant on the surface of the catalyst, which resulted in higher methanol conversion and DME selectivity. The catalyst with an Al/P molar ratio of 2 had higher activity and less deactivation during the reaction in a fixed bed reactor at 300 °C and atmospheric pressure. Yaripour et al. [61] further studied the performance of aluminum phosphates with different Al/P molar ratios (0.5—3) in a fixed bed reactor at 1.6 MPa. The catalyst with an Al/P ratio of 1.5 provided the most promising and stable performance, with a DME yield of 81.69 % [61]. It was reported that the nature of phosphate precursors and the preparation method can affect the properties of the synthesized aluminum phosphate [63]. In a study, phosphoric acid (H3PO4) and diammonium hydrogen phosphate ((NH4)2HPO4) were used as phosphate precursors to prepare a series of aluminum phosphate catalysts with an Al/P ratio equal to 1. The catalysts were prepared using precipitation (AP-I and AP-III) and impregnation (AP-II and AP-IV) methods. In the case of precipitated catalysts, ammonia was added dropwise to a solution of ANN and a phosphate precursor followed by neutralization, drying (at 110 °C), and calcination (at 700 °C). In the case of impregnated catalysts, ammonium was added to ANN to get an aluminum hydroxide precipitate which was dried and calcined at the same conditions as the later method. The obtained aluminum oxide was then impregnated with a phosphate precursor. Although all the catalysts possessed moderate strength acid sites, the catalysts prepared via precipitation showed higher acidity than those prepared via impregna-

Kataliz v promyshlennоsti, vol. 18, № 4, 2018

tion. Moreover, the amorphous AlPO4, which is the active phase for the MTD reaction, was guaranteed in the case of precipitation. The catalyst prepared using impregnation method showed a very low activity. The impregnated catalyst with H3PO4 showed a slightly higher activity, however, it had a poor thermal stability compared to the catalyst prepared using (NH4)2HPO4. In the case of the catalyst prepared via precipitation, (NH4)2HPO4 seemed to show more promising results [63]. Although thermal pretreatment of aluminum phosphate in a neutral environment (using helium) did not improve the performance of the catalyst in the MTD reaction, hydrothermal pretreatment using 10 % water vapor in helium (at 100—300 °C) showed a positive impact on the catalytic activity of the AlPO4 catalyst [62]. It was reported that the presence of water molecules resulted in the formation of hydroxyls which were adsorbed and chemically bonded with the available phosphate groups in the aluminum phosphate catalyst, leading to the conversion of Lewis acid sites to Brønsted acid sites [62]. The hydroxyl groups on the phosphorous atoms (P–OH) are acidic and their strength is higher than Al–OH sites [63, 158]. Therefore, a higher amount of P–OH in the catalyst could correlate with a higher amount of available acid sites, which is in agreement with the reported results in the literature as seen in Table 4. It was proposed that the P–OH groups are involved in water condensation on the Brønsted acid sites leading to formation of the methoxy groups on the orthosphosphate surface, which can react with methanol to form DME. At reaction temperatures of 150—300 °C and atmospheric pressure in a fixed bed reactor, the pretreated catalyst performed optimally at 250 °C — this was likely due to the quantity and strength of the acid sites [62]. In addition to non-porous aluminum phosphate salts, aluminophosphate molecular sieves were also used to catalyze the MTD reaction. Aluminophosphate molecular sieves generally have weaker acid sites compared to their aluminosilicate counterparts, which makes them a better catalyst for DME production from methanol [59, 69]. It was also reported that isomorphous substitution of various polycharged cations into the aluminophosphate framework can improve the catalytic activity of the final catalyst [69]. Kikhtyanin et al. [159] synthesized aluminophosphates with an AlPO-5 structure using phosphoric acid (as a source of phosphorus), hydrated aluminum oxide (as a source of aluminum), and hydroxides or oxides of the respective elements (Be2+, B3+, Ga3+, Fe3+, Si4+, Ti4+, Mn4+, and V5+) for modified aluminophosphate catalysts. Galloaluminophosphate and ferroaluminophosphate frameworks containing B3+, Ga3+, Fe3+, Ti4+, Mn4+, and V5+ showed promising potential for DME synthesis at 380 °C while further increases in reaction

17

Катализ в химической и нефтехимической промышленности Table 4 The relative amount of P–OH group of hydrothermally treated AlPO4 samples from FTIR and Raman spectroscopy calculated by semi-quantitative analysis and the amount of acid sites by amine titration using methyl red indicator (Reproduced with permission from reference [62]. Copyright © 2010, Elsevier) Pretreatment temperature, °C

Relative amount of P–OHa

Amount of acid sitesb, mmol H+/gcat

FTIR at 2950 cm–1

Raman shift at 985 cm–1

100

0.49

0.00846

0.266

200

1.23

0.02431

0.276

250

1.35

0.03245

0.359

300

0.91

0.02871

0.340 –1

For the FTIR results: ratio of the area of de-convoluted hydroxyl peak at 2950 cm to that of Al–O–P peak at 833 cm–1. For Raman results: ratio of the area of P–OH at 985 cm–1 to that of PO21– at 1200 cm–1. b Determined by amine titration using methyl red indicator. a

Table 5 Methanol conversion and DME selectivity of different molecular sieves used in MTD reaction at different temperature (Reproduced with permission from reference [59]. Copyright © 2011, Elsevier) Sample

Temperature, °C

SAPO-5

SAPO-11

SAPO-41

AlPO-5

AlPO-11

AlPO-41

Methanol conversion, %

DME selectivity, %

250

85

68.1

300

98.6

12.1

350

100

1.1

400

100

0

200

43.7

100

250

84.1

100

300

80.1

83.2

350

84.2

41.7

400

95

10.5

250

81.3

82.1

300

92.1

51.4

350

100

3.7

400

100

0

250

34.6

100

300

79.6

99.9

350

81.0

99.6

400

79.2

91.7

250

43.8

100

300

79.3

99.9

350

81.4

99.9

400

76.1

98.8

250

47.1

100

300

80.3

99.9

350

82.4

99.9

400

75.9

98.8

WHSV – 1 h–1, time-on-stream – 5 min.

18

temperature to 450 and 500 °C significantly decreased the selectivity [159]. Dai et al. [59] compared the performance of three different aluminophosphates with similar one-dimension channels (AlPO-5, AlPO-11, and AlPO-41) in a fixed bed microreactor for the MTD reaction at 250— 400 °C and atmospheric pressure. As can be seen in Table 5, all three catalysts showed promising potentials for methanol dehydration to DME. The catalysts showed high DME selectivity (>99.9 %) at reaction temperatures as high as 250—350 °C resulting in a DME yield of over 80 % at 350 °C. The low activity of the catalyst at lower temperatures and negligible production of other hydrocarbons at high temperatures are attributed to the presence of only weak acid sites and the absence of moderate (and strong) acid sites on the aforementioned aluminophosphate molecular sieves [59].

7. Silicoaluminophosphate Silicoaluminophosphate (SAPO) is a class of crystalline microporous framework oxide molecular sieves with the properties of both zeolites and aluminophosphates as well as additional unique properties of its own. This class of materials shows structural diversity in its three-dimensional microporous framework structures. SAPO-40, SAPO-41, and SAPO-44 are some of the examples in this group. In terms of topology, the SAPO’s structure can be similar to zeolites (SAPO-34, SAPO-35, SAPO-37, and SAPO-42), aluminophosphates (SAPO-34, SAPO-35, SAPO-37, and SAPO-42), or both of them (SAPO-17 and SAPO-20) [160, 161]. SAPOs were also used as the dehydration catalyst for DME synthesis; however, there is a larger body of literature on methanol or DME conversion to other hydrocarbons over this group of catalysts [55, 57, 162—165]. Pop et al. [58] reported that H-SAPO-34 molecular sieves can be used in

Катализ в промышленности, т. 18, № 4, 2018

Катализ в химической и нефтехимической промышленности the same reactor for DME or olefin synthesis from methanol with only a temperature adjustment. A series of SAPOs including SAPO-34, -18, -5, and -11 were prepared via thermal treatment of an aqueous mixture of phosphoric acid, pseudo-boehmite, fumed silica, and tri-n-propylamin to be used in DME synthesis [56]. The SAPO-5 and -11 possessed a spherical morphology formed by agglomeration of small crystals, while SAPO-18 and -34 owned a corresponding platelet-like shape and well-shaped cubic pattern (Fig. 5). SAPO-5 consisted of unidimensional sheets of 12-membered ring channels with a pore size of about 0.74 nm connected to each other via 4- and 6-membered rings. By contrast, SAPO-11 possessed 10-membered rings in unidimensional sheets, resulting in elliptical pores with a 0.64 nm × 0.44 nm pore opening [56, 166]. The SAPO-34 and -18 provided a high surface area with comparable pore size (0.4 nm) and shape in the structure, only differing in crystallography [167]. SAPO-5 and -11 possessed weak and moderate strength acid sites, contrasting with the strong acid sites of SAPO-18 and -34; this led to a lower overall acidity for SAPO-5 and -11. The presence of P–OH and SiOHAl groups in the structure originated the weak and strong acid sites, respectively. The strength of the acid sites originated by SiOHAl groups is a function of the number of Si neighbors; thus, Si (1Al, 3Si) > Si (2Al, 2Si) > Si (3Al, 1Si) >

> Si (4Si) [168]. The evaluation of catalytic activity in a fixed bed reactor at 260 °C for the MTD reaction showed a rapid drop in the performance of SAPO-34 and -18 within a short reaction time, due to the formation of coke on the strong acid sites (coverage of the acid site) and pore blockage. In fact, inappropriate pore structure of these catalysts (small pore openings) and periodic expansion of the catalysts resulted in poor stability and a significant reduction in the activity of the catalyst, even in the presence of a small amount of hydrocarbon compounds. On the other hand, the relatively mild acid strength along with the appropriate pore structure of the SAPO-5 and -11 prevents the formation and buildup of carbonaceous compounds, resulting in a more stable catalytic performance in the MTD process. SAPO-11 showed a better performance for long-term methanol dehydration (more than 4 h). The performance of the catalyst was also evaluated in the STD process in which the dehydration catalyst was physically mixed with a methanol synthesis catalyst; the reaction was performed at 260 °C and 4.2 MPa with a H2/CO ratio of 1.5. As expected, the SAPO-34 and -18 showed a less stable performance while the SAPO-5 exhibited the highest overall CO conversion over the course of a 20 h reaction (long-term evaluation) [56]. On the contrary, Pop et al. [58] synthesized SAPO-34 through a different method and observed a promising per-

Fig. 5. SEM image of various SAPO catalysts: a – SAPO-5, b – SAPO-11, c – SAPO-18, d – SAPO-34 (Reprinted with permission from reference [56]. Copyright © 2007, Elsevier)

Kataliz v promyshlennоsti, vol. 18, № 4, 2018

19

Катализ в химической и нефтехимической промышленности formance in methanol dehydration (over 80 % conversion at temperatures above 200 °C). They prepared a gel oxide by addition of tetraethylammonium phosphate (template) to an amorphous phosphorus-alumina-silica gel while controlling the pH. The gel was then subjected to crystallization and drying to form a chabazite-CHA type structure in the resulting SAPO. Calcination of the SAPO at 600 °C for 4 h resulted in the complete removal of the template and formation of H-SAPO-34. The reaction at temperatures above 300 °C may significantly decrease the DME selectivity due to the formation of olefins [58]. Dai et al. [59] also evaluated the performance of one dimensional SAPO-5, SAPO-11, and SAPO-41 prepared via a hydrothermal method followed by calcination at 550 °C for 6 h in methanol dehydration. The catalysts possessed a BET surface area in the range of 200—250 m2/g with both weak and moderate acid sites; the total acidity decreased along the order of SAPO-41 > SAPO-5 > SAPO-11. In contrast with the results reported by Yoo et al. [56], SAPO-11 showed a more promising result among others at atmospheric conditions where the DME yield reached 84.1 % at 250 °C, as reported in Table 5. The presence of moderate acid sites could facilitate the MTD reaction at low temperatures; however, they can catalyze the methanol-to-olefin (MTO) reaction at higher temperatures. As the SAPO-11 has a very low concentration of moderate acid sites, the MTO reaction could not initiate at low temperatures. This catalyst showed a stable performance over the course of a 200 h reaction at 250 °C with no detectable coke formation on the surface [59]. Pinkaew et al. [169] prepared a core-shell like SAPO-46 encapsulated Cr/ZnO (methanol synthesis) capsule catalyst for the STD reaction using a physical coating method in which silica sol was used as an adhesive between the core and the shell. SAPO-46 possesses a three-dimensional structure of 12-member rings with weak Brønsted acidity. This catalyst showed great potential as a dehydration catalyst either in its encapsulated form or as a physical mixture; however, the DME selectivity of the former was twice that of the latter form.

8. Heteropoly acids Heteropoly acids (HPAs) have a very strong Brønsted acidity — even higher than conventional solid catalysts in some cases [18, 54]. HPAs own different molecular structures including Keggin and Dawson, which are more common structures for catalytic applications. The Keggin HPA is formulated as H8−n[X n+M12O40], where X is the central atom (P5+, Si4+, etc.), n is the oxidation state, and M is the metal ion (W6+, Mo6+, etc.); this structure possesses a central tetrahedron (XO4) surrounded by four octahedral MO6 units

20

stabilized by protons [53, 170]. The HPAs usually suffer from a low surface area which can be increased using a proper support [53, 141]. Keggin-type HPA catalysts have been widely used for dehydration of alcohol, especially ethanol [51, 52, 171—173]. Further studies have shown the positive effect of mesoporous silicate structures as the support for HPAs for the dehydration reaction [50, 52, 141, 148, 174]. Ciftci et al. [148] prepared a tungstophosphoric acid (TPA) incorporated silicate structured mesoporous catalyst using two different methods; namely, one-pot hydrothermal and impregnation of TPA into MCM-41. The Keggin structure was retained after impregnation [148]. The same result was reported by Herrera et al. [50] in which they used incipient wetness impregnation of TPA into MCM-41 and SBA-15. On the contrary, the hydrothermal method mostly destroyed the Keggin structure and replaced it by WOx clusters. It was observed that impregnation decreased the strength of the Brønsted acid sites [148]. The same result was reported by Varisli et al. [141]. The DME yield could reach to about 79 % at 220 °C via the impregnated catalyst, while the reaction yield over the hydrothermally synthesized catalyst was more limited [148]. Herrera et al. [50] quantified the amount of leaching of TPA in water at 50 °C from the supports. It was concluded that modification of MCM-41 and SBA-15 with aluminum ethoxide, titanium isopropoxide, or zirconium isopropoxide increased the strength of the binding between the support and TPA and consequently suppressed the leaching issue [50]. TiO2-supported heteropoly acid catalysts were also used for methanol dehydration [53, 54]. Two Keggin-type HPAs, H3PW12O40 (HPW) and H4SiW12O40 (HSiW), were deposited on TiO2 using incipient wetness impregnation with HPA loading in the range of 0.9—9.0 KU/nm 2 (KU — Keggin Units); the catalyst was then diluted using SiC prior to its use in the reactor to avoid hot spots. The supported catalysts provided a higher surface area compared to the bulk HPAs; however, an increase in the HPA content decreased the total surface area of the catalyst. It was found that an HPA loading of 2.3 KU/nm 2 is the optimum loading to achieve the highest DME yield. Lower HPA loadings may not provide monolayer coverage on the support or may result in a strong interaction between the acid protons of the HPAs and the support, hence preventing the participation of such protons in the methanol dehydration reaction. On the other hand, higher HPA loadings on the support resulted in the formation of larger HPA units, preventing the access of methanol to internal acid sites. It was also concluded that deposition of HPAs on TiO2 did not affect their Keggin structures regardless of HPA loading. Both the bulk and the supported HPA catalysts exhibit reasonably high methanol conversion rates and DME selectivity even at temperatures

Катализ в промышленности, т. 18, № 4, 2018

Катализ в химической и нефтехимической промышленности as low as 140 °C; however, the supported catalysts showed higher catalytic activities for the MTD reaction [54]. Alharbi et al. [18] investigated the relationship between the MTD reaction rate and the acid strength of both bulk and supported Keggin HPAs. The Keggin type HPAs, HPW and HSiW, were well-deposited on different catalyst supports, i. e., titania-based, silica-based, ZrO2, and Nb2O5, using the wet impregnation method. The cesium salts of Cs2.25H0.75PW12O40 and Cs2.5PW12O40 were also prepared. The dried calcined catalysts were ground and sieved to a 45—180 μm particle size. The supported HPAs provided a BET surface area of over 100 m 2/g with the exception of 15 % HPW/TiO2 which had a surface area of 45 m 2/g. It is noteworthy that the supported HPAs on SiO2 provided a higher BET surface area (over 200 m 2/g). As expected, the acid strength of the supported HPAs was less than the bulk HPA and the cesium salts [18]. However, Varisli et al. [141] reported that in the case of supported silicotungstic acid supported on aluminosilicate, the Brønsted acidity decreased while the Lewis acidity increased. It was revealed that the support affected the acid strength of the catalysts in the order of SiO2 > > TiO2 > Nb2O5 > ZrO2 simply due to the increase in the interaction between the HPAs and the support [18]. Alharbi et al. [18] showed that the 20 % HSiW/SiO2 catalyst offers stable performance at 150 °C and atmospheric pressure without deactivation for 4 h, where the methanol conversion and DME selectivity reached 89 % and 100 % for the MTD reaction, respectively. At temperatures above 200 °C, the DME selectivity dropped due to the formation of other light hydrocarbons, leading to coke formation and consequently catalyst deactivation [18]. The supported HSiW on aluminosilicate showed better thermal stability where no coke formation occurred up to 300 °C [141]. A comparison with the performance of the HZSM-5 for the MTD reaction revealed lower activities for the zeolite catalyst compared to the 20 % HSiW/SiO2 and Cs2.25H0.75PW12O40 (per unit catalyst weight). The authors [18] presented a fairly good linear relationship between the activity of HPA catalysts in the MTD reaction and their initial enthalpy of ammonia adsorption (as evidence of surface acidity). Carr et al. [19] have studied the effect of acid identity for methanol dehydration over silica-supported Keggin tungsten HPA clusters H8–nX n+W12O40 (X = P5+, Si4+, Al3+, and Co2+), where they provided an insight to the mechanism and kinetics of the reaction as well as the effect of acid site properties on the reaction. It was also concluded that the MTD reaction over HPA catalysts follow the same mechanism as the reaction over HZSM-5 [18]. Some of the HPAs are extremely sensitive to water which can result in extensive and continuous leaching over the course of the dehydration reaction [175]. To overcome such

Kataliz v promyshlennоsti, vol. 18, № 4, 2018

a barrier, Ivanova et al. [175, 176] proposed the synthesis of an organic-inorganic hybrid composed of an ionic liquid protected by the HPA catalyst. 1-butyl-3-methyl imidazolium methanesulfonate (Bmim) was mixed with phosphomo3– 3– lybdic (PMo12O40 ) or phosphotungstic (PW12O40 ) ions to prepare a hybrid salt with a Bmim3PA molecular structure which protected the HPAs. Both catalysts showed a better performance compared to a water insoluble Cs2HPW12O40 acidic metal salt of HPAs, confirming the effectiveness of hybridization using an ionic liquid. As expected, the tungsten based structure showed higher thermal stability than a phosphomolybdic hybrid catalyst; however, its catalytic activity was lower due to stronger interactions between the imidazolium cation and the Keggin structure during the reaction. The partial decomposition of the hybrid catalyst resulted in the formation of more active and selective phases on the catalyst; however, a treatment at very high temperature resulted in MoO3 or WO3, facilitating the oxidation of methanol to CO2 [176]. Anwar et al. [177] reported that at a pretreatment temperature of 250 °C, the performance of supported HPAs (H3PW12O40 and H3PMo12O40) on α-alumina for the MTD reaction was improved at elevated temperatures (250— 500 °C) due to the conversion of Brønsted acid sites to Lewis acid sites by structural water losses at temperatures over 300 °C. Such conversion decreased the potential of further isomerization reactions or methane formation [177]. However, it was reported that the presence of water during the dehydration reaction over hybrid catalysts of HPAs and ionic liquids may limit the collapse of the Keggin structure [176]. Anwar et al. [177] claimed water can absorb on acid sites of the supported HPAs and decrease the efficiency of the catalysts.

9. Simple metal oxides Metal oxides have received a great deal of attention due to ease of alteration in their acid and base properties through composition modifications. Hesnel and Pines [178] reported ether formation on a non-supported (bulk) palladium oxide and also iridium oxide while the oxide of iron, cobalt, platinum, and copper did not catalyze the alcohol dehydration reaction. However, some supported transition metals, such as iron oxide supported on alumina, show potential for dehydration [69]. A limited amount of DME was formed on Pd/γ-Al2O3, Pd/TiO2, Pd/MgO, and Pd/ZrO2 during methanol synthesis due to the acidity of the support [179, 180]. Palladium oxide impregnated on fumed silica (Cab-O-Sil) could catalyze the alcohol dehydration [37, 181]. It was reported that an increase in the reaction temperature from 225 °C to 280 °C increased the methanol

21

Катализ в химической и нефтехимической промышленности conversion from 38 % to 77 %, while DME selectivity decreased from 78 % to 47 % [37]. The reduced nickel oxide supported on Cab-O-Sil was successfully utilized in the dehydration reaction, in which the metallic nickel acted as a Lewis acid and nickel oxide acted as a Lewis base (electron donor) [182—184]. It was also reported that the supported nickel oxide may possess Brønsted acid sites depending on the type of support. For example, although NiO/Al2O3— SiO2 shows proper Brønsted acid sites, NiO/Al 2O3 does not possess any Brønsted sites whatsoever [69]. Narasimhan and Swamy [185, 186] used MgM2O4 (M = Al, Fe, Cr) for alcohol decomposition where MgCr2O4 catalyzed both dehydration and dehydrogenation reactions while only the dehydration reaction occurred on MgAl2O4. MgFe2O4 did not show any potential for catalyzing the dehydration reaction. Although it was expected that addition of alkali metals as dopants may neutralize the acid sites and decrease the catalytic activity, Chakrabarty et al. [187] showed that doping the V2O5 with lithium actually improved its dehydration potential. However, doping with sodium decreased the catalyst activity as expected. Titanium oxide prepared by precipitation using titanium (III) chloride and ammonia was reduced at different temperatures in the range of 400—1000 °C. The TiO2 catalyst reduced at 600 °C showed the most promising potential for the MTD reaction at 400 °C while further increases in the reaction temperature resulted in DME decomposition to methane and carbon dioxide [188]. Vishwanathan et al. [189] prepared a series of TiO2—ZrO2 with different compositions using co-precipitation of TiCl4 and ZrCl4 with addition of NH4OH as a pH adjuster, followed by 24 h of aging and further calcination at 550 °C for 6 h. The calcined catalyst with a high TiO2 concentration owned a highly amorphous structure, attributed to a strong nitration between TiO2 and ZrO2. As can be seen in the FTIR spectra of pyridine adsorbed oxides (Fig. 6), the mixed oxides do not have Brønsted acidity due to the absence of peaks associated with the characteristic band at 1540 cm–1. Two intensive bands at 1446 and 1608 cm–1 are attributed to strong Lewis acid sites and the bands at 1489 and 1575 cm–1 refer to moderate and weak acid sites, respectively. The increase in the TiO2 content of the mixed catalyst resulted in more intensive IR absorbance peaks which are attributed to the increase in surface acid sites and their corresponding strength. Therefore, the mixed oxide catalyst with high TiO2 content could convert methanol to hydrocarbons. The TiO2—ZrO2 catalyst with Ti/Zr ratio in the range of 30/70 to 50/50 provided a higher surface area and DME production yield (over 80 %) at 340 °C and atmospheric pressure [189]. Titanium was also used in methanol dehydration in a sulfate form to enhance the performance of other acid cata-

22

Fig. 6. FTIR spectra of pyridine adsorbed on oxides: a – ZrO2, b – Ti/Zr = 1/9, c – Ti/Zr = 3/7, d – Ti/Zr = 1/1, e – Ti/Zr = 7/3, f – TiO2 (Reprinted with permission from reference [189]. Copyright © 2004, Springer)

Fig. 7. FTIR spectra of ammonia adsorption at room temperature on: a – Ti(SO4)2 modified γ-Al 2O3, b – γ-Al 2O3 (Reprinted with permission from reference [94]. Copyright © 2005, Elsevier) lysts [94]. As mentioned in section 2, Ti(SO4)2/γ-Al2O3 was achieved by impregnation, dried at 110 °C, and calcined at 500 °C for 1.5 h [94]. Fig. 7 shows the FTIR spectra of ammonia adsorption on Ti(SO4)2/γ-Al2O3 compared to the original γ-Al2O3. The bands at 1246 and 1612 cm–1 are attributed to the Lewis acid sites while the bands at 1475 and 1695 cm–1 originated from ammonium ions due to the interaction of Brønsted sites with ammonia. As can be seen, the modified catalyst owned proper Brønsted acid sites (higher than γ-Al2O3) resulting in conversion of methanol to DME without the formation of other hydrocarbons and coke [94]. Niobium oxides are the most common group of niobium based compounds in heterogeneous catalysis [190]. Al-

Катализ в промышленности, т. 18, № 4, 2018

Катализ в химической и нефтехимической промышленности though Nb2O5 has a very high melting point, the Tamman temperature (in which the surface atoms begin to diffuse) is much higher than typical catalytic reactions (200— 600 °C), limiting their applications [190]. However, the treated or modified Nb2O5 can be used as a promising dehydration catalyst [191]. Sun et al. [192] prepared a porous niobium phosphate (denoted as NbOPO4) through the hydrolysis of NbCl5 in phosphoric acid solution followed by further mixing with hexadecylamine, aging, and calcination. The NbOPO4 showed higher acidity compared to Nb2O5 due to the co-existence of P—OH and Nb—OH groups on the surface of the modified catalyst; the P—OH groups are stronger Brønsted acids. Both catalysts possessed Lewis acid sites due to the presence of coordinatively unsaturated Nb5+ cations [192, 193]. The NbOPO4 showed a higher surface area, stronger hydrophilicity, and many more adsorption sites (potential active sites) than Nb2O5. Although neither of the catalysts were as active as H-ZSM-5, they possessed great thermal stability and DME selectivity [192]. The NbOPO4 reached maximum activity at 320 °C with about 80 % methanol conversion and 100 % DME selectivity, while Nb2O5 was less active. By employing wetness impregnation methods via niobium (V) oxalate, Ladera et al. [194] took advantage of both TiO2 and Nb-based compounds by preparing a series of Nb-based/TiO2 catalyst with different Nb surface densities. All of these catalysts possessed medium and strong acid sites as per BET analysis; the NbO6 octahedral structures, specifically, have Lewis acid properties due to the Nb=O bonds which disappeared at higher Nb surface densities (around 4.5 at. Nb/nm 2). These higher surface density catalysts possessed increased Brønsted acidity and higher overall acid strength. The catalysts showed stable performances over the course of a 16 h methanol dehydration reaction in a fixed bed reactor at 300 °C and 6.1 kPa methanol (13 mmol·h–1·gcat–1). The increase in the Nb-loading improved the methanol conversion and the DME selectivity which could be due to the improvement of the acid property of the catalyst; however, methane and formaldehyde were also detected in these cases.

catalysts in different industrial processes including etherification of olefins with alcohols, dehydration of alcohols, and alkylation of phenols among others [195]. Ion exchange resins received attention for alcohol dehydration due to their ability to catalyze the reaction at relatively low temperatures (