Enhancement of the dimethyl ether carbonylation activation via regulating acid sites distribution in FER zeolite framework

Summary

The carbonylation of dimethyl ether (DME) with CO is a key step for ethanol synthesis from syngas, but traditional mordenite (MOR) zeolite shows low catalytic stability. Herein, various FER zeolite nanosheets were prepared with four types of organic templates. The catalytic performance of FER in DME carbonylation is strongly dependent on the location of strong acid site in framework, which can be effectively regulated by altering organic template. FER-MORP sample synthesized with morpholine shows the highest DME conversion of 53%, thus, giving a methyl acetate space-time yield (STY_MA) of 0.889 mmol g^–1 h^–1. DFT calculation, NH₃-IR, ¹H/²⁷Al/²⁹Si MAS NMR, and in situ DRIFTS results indicate that morpholine directs more Al species, or strong Brønsted acid sites (BAS), to locate in 8-membered ring (8-MR) channels, which not only enhances carbonylation activity but also suppresses formation of coke species. The catalytic performance is well maintained within 4 repeated recycles (∼460 h).

Graphical abstract

Highlights

• •Series of nanosheets FER zeolites with high crystallinity were synthesized
• •Use of morpholine as template agent directs more Al siting in 8-MR channel of FER
• •Elevation of BAS in 8-MR promotes DME carbonylation by forming more acyl species
• •DME conversion, MA selectivity, and lifetime reach 53%, 99.5%, and ∼460 h on FER-MORP

Physical inorganic chemistry; Interface science; Energy materials

Introduction

Ethanol as a clean gasoline additive shows higher energy density and octane number than traditional methyl tert-butyl ether.^1,2,3 It is generally produced from microbial fermentation and ethylene hydration, but these processes are costly or consume large amounts of cereal crops. Recently, a new route has been proposed for ethanol synthesis, which involves in dimethyl ether (DME) carbonylation to methyl acetate (MA) and subsequent hydrogenation.^4,5 This process makes it possible of synthesis of ethanol from coal, natural gas, and biomass via syngas. In this process, DME carbonylation is the rate-limiting step.

In 2006, Iglesia and co-workers⁶ found that acidic mordenite (H-MOR) is a potential catalyst for DME carbonylation under mild reaction conditions because of its high catalytic activity and MA selectivity. This catalyst made a breakthrough for carbonylation as it avoids using Rh or Ir organometallic complexes as catalysts and iodide compounds as promoters. Therefore, it attracts considerable research interest.^7,8,9,10 However, the severe coke deposition in the 12-membered ring (12-MR) channel causes it suffer from a rapid deactivation.¹¹ A consensus has been reached that the Brønsted acid sites (BAS) in the 8-MR channel of H-MOR are active sites for DME carbonylation.⁸ Thus, more attentions have been paid to increase of BAS in 8-MR channels and removal of BAS from 12-MR channels. Simple methods are pre-adsorption of pyridine¹² and selective dealumination,¹³ as they can effectively poison or decrease BAS in 12-MR channels. Ion-exchange of H⁺ with Cu²⁺,¹⁴ Zn²⁺,¹⁵ or Co²⁺ ions¹⁶ can not only play this role but also enhance catalytic activity. Therefore, demo DME carbonylation catalyst is Cu-MOR.

Another method is controllable synthesis of MOR zeolite with more Al species or BAS located in 8-MR and less in 12-MR channels, as the BAS distribution in zeolite framework is closely related to the Al locations.¹⁷ It has been shown that the Al location in lattice sites can be regulated by altering synthesis materials, such as silicon source, aluminum source, and counter-cations, and adding appropriate amounts of heteroatoms.^18,19,20 Silicon sol as silica source leads to location of more Al atoms in 10-MR straight and sinusoidal channels of ZSM-5, whereas tetraethoxysilane gave larger number of Al atoms in intersection cavities.¹⁸ Na⁺ or (Na⁺ + Li⁺) as counter-cation shifts considerable numbers of Al atoms from 10-MR channels to intersection cavities in the synthesis of ZSM-11.²¹ Recently, some researchers found that an appropriate organic templating molecules could direct framework Al atoms to desirable lattice sites in zeolite.^17,22 For example, due to the space confinement effect, the organic template agents prefer to locate in the 10-MR channel of FER.²² This allows more Na⁺ ions to be distributed in the 8-MR channel to balance the negative charge of [AlO₄]^-, and thus, increasing the content of acid sites in the 8-MR channel of FER zeolite.

In addition, the effect of zeolite structure, e.g., FER,²³ ETL,²⁴ SZR,²⁵ and CHA,²⁶ on the catalytic performance was investigated. It is shown that FER zeolite with a two-dimensional (2D) lamellar structure consisting of 10- and 8-MR channels in [001] and [010] directions, respectively, exhibits comparable catalytic activity as MOR.^27,28 Similar to MOR zeolite, the BAS in the 8-MR channels are also assumed to be the main active sites for catalyzing DME carbonylation.^29,30,31 As its 10-MR channel has stronger space confinement effect on coke formation than the 12-MR channel of MOR,³² it shows higher catalytic stability. With the inspiration, the BAS distribution was regulated by some researchers.^29,30,33 Shen and co-workers proved that elevation of the proportion of Al atoms in the 8-MR pore of FER can greatly improve the activity and durability for DME carbonylation.³³

Since the Al siting is always related to the organic structure-directing agent (OSDA), it might be possible to controllably regulate the location of Al in FER framework by changing the structure of OSDA. Herein, various FER zeolites were prepared using ethylenediamine (EDA), 1,6-diaminohexane (DAH), piperidine (PI), and morpholine (MORP) during synthesis. Interestingly, the unused morpholine as a template molecule results in 50.2% of Al atoms located on the T2+T4 sites and a considerable increase in strong BAS in 8-MR channels of FER zeolite, and hence, a great enhancement in catalytic activity and stability.

Results and discussions

Physicochemical properties of synthesized FER-x zeolites

Figure 1A shows that all the samples have pure FER structure with high crystallinity (81.5%–100%). The crystals of these samples all exhibit a nanosheet habit with the thickness of 50–100 nm (Figures 1C–1F). N₂ sorption results show that all the synthesized FER zeolites possess similar specific surface area and micropore volume, being in the range of 281–296 m² g^–1 and 0.16–0.22 cm³ g^–1 (Figure 1B, Table S1). These characterization results show that organic template has insignificant effect on the crystal structure, morphology, and texture properties of samples.

Figure 1.

Characterizations of structure and morphology of FER-x zeolites

(A) XRD patterns.

(B) N₂ sorption isotherms.

(C–F) FE-SEM images of FER-MORP, FER-EDA, FER-DAH, and FER-PI zeolites.

The acid properties of various FER-x zeolites were investigated by NH₃-TPD, NH₃-IR, and Py-IR. Figure 2A shows that two NH₃ desorption peaks are present in the ranges of 150°C–300°C and 350°C–550°C, which are due to desorption of NH₃ from weak and strong acid sites, respectively.³⁴ These two peaks are located at higher temperature in the profile of FER-MORP than in those of FER-DAH, FER-EDA, and FER-PI, indicating a stronger acid strength. The acid strength of BAS was further evaluated by NH₃-IR with the spectra collected at 150°C, 250ºC, and 350ºC (Figure S1). The intense peak at 1450 cm⁻¹ and the weak one at 1620 cm⁻¹ are characteristic of BAS and Lewis acid sites (LAS), respectively.³⁵ The peak areas (1450 cm⁻¹) collected at 150°C, 250ºC, and 350ºC are considered to be correspondent to the amounts of total, medium, and strong BAS, respectively.³⁶ Interestingly, FER-MORP shows larger I₃₅₀/I₁₅₀ ratio than the other three samples, indicating that it has more strong BAS. As the strength of BAS in the 8-MR channel of FER is much stronger than those in the 10-MR channel,¹⁰ it can be deduced that FER-MORP may have more BAS in the 8-MR channels.

Figure 2.

Acidic properties of various FER-x zeolites

(A) NH₃-TPD.

(B) Py-IR.

(C) ¹H MAS NMR.

(D) ²⁹Si MAS NMR spectra.

The total BAS (B_total) content in zeolite can be obtained from ¹H MAS NMR,^37,38,39 where the signals at 4.2 and 1.8 ppm are assigned to bridging hydroxyl (Si–OH–Al) groups (BAS) and nonacidic surface silanol (Si–OH) species (Figure 2C), while the other two at 2.8 and 0.95 ppm are due to extra-framework aluminum (Al–OH) species (LAS).³⁷ Therefore, the total Brønsted acid content of zeolites was quantificationally calculated according to the intensity of peak at 4.2 ppm.³⁹ On the contrary, the results tested by pyridine-adsorbed IR is mainly the BAS in 10-MR (B_10-MR) (Figure 2B), as pyridine as a basic probe molecule usually characterizes the acidity in void space with window opening larger than 10-MR due to the steric restriction.⁴⁰ Therefore, the amount of BAS in 8-MR (B_8-350) was calculated by deducting B_10-MR from B_total. Table 1 displays the total BAS (B_total), BAS in 10-MR channel (B_10-MR), and BAS in 8-MR channel (B_8-MR) of four synthesized FER-x zeolites. It can be seen that morpholine as template effectively elevates the content of BAS in 8-MR channels of FER framework.

Table 1.

Physicochemical properties of various FER-x samples

Catalysts	Si/Al ratio	Cryst. (%)	Brønsted acid content (μmol g–1)				NH3-IR (%)
			Btotala	B10-MRb	B8-MRc	B8-350d	I350/I150e
FER-MORP	8.79	100	1054.7	222.2	832.5	363.8	43.7
FER-EDA	8.85	88.8	1030.8	226.2	804.6	333.9	41.5
FER-DAH	8.99	85.2	949.0	218.3	730.7	273.4	37.4
FER-PI	8.58	81.5	936.3	209.3	727.0	195.6	26.9

^aB_total is tested by ¹H MAS NMR.

^bB_10-MR is tested by Py-IR.

^cB_8-MR is the amount of Brønsted acid in 8-MR obtained by the difference of B_total and B_10-MR.

^dB_8-350 is the amount of strong Brønsted acid in 8-MR obtained by the B_8-MR and I₃₅₀/I₁₅₀.

^eI₃₅₀/I₁₅₀ represents the peak areas (1450 cm⁻¹) collected at 350°C to that at 150°C in NH₃-IR.

Effect of organic template on the Al locations in the framework of FER zeolite

FER zeolite has four crystallographically distinct T sites. T1 and T3 sites are located in the 10-MR channels, while T2 and T4 sites are positioned in the 8-MR channels.²⁰ Recent investigations confirm that the BAS located in T2 and T4 sites are the main catalytic centers for DME carbonylation, whereas those in T1 and T3 sites show lower carbonylation activity.²³ The thermodynamic distribution probability of Al and relative substitution energies of Al for Si in different T sites were evaluated by periodic density functional theory (DFT) calculation (Figures S2–S5 and Table S2). It shows that morpholine directs more Al atoms to occupy T2 sites with the probability as high as 97%. With respect to ethyldiamine and piperidine, although T2 is still the favorable site for locating Al, its occupation probability is decreased to 72.9% and 79.1%, along with an increase in T1 site to 27% and 20.8%, respectively. Concerning 1,6-diaminohexane, more Al atoms are inclined to occupy T1 site (71.7%), followed by T4 site (26.7%). These calculation results theoretically evidence that the thermodynamic stability of Al atoms in different FER lattice sites is highly related to organic template structure.

The ²⁷Al MAS NMR spectra of the four FER-x zeolites show an intense peak centered at about 54 ppm and a small peak at –2 ppm (Figure S6), which are characteristic of tetrahedral framework Al and octahedral extra-framework Al species, respectively.^41,42,43 A deconvolution of the spectra in the range of 30–65 ppm indicates that several types of Al species are present in FER zeolite (Figure 3). The peaks at around 30–40 and 40–45 ppm are ascribed to pentacoordinated and/or distorted tetrahedral Al species, respectively.^44,45 The formation of these extra-framework Al species usually leads to the decrease of carbonylation activity,^46,47 despite that their exact structure and configuration are still not clear. It is found that FER-MORPP has smaller extra-framework Al and distorted tetrahedral Al species than that of FER-EDA, FER-DAH, and FER-PI. This results in the decrease of total BAS content in the order of FER-MORPP > FER-EDA > FER-DAH > FER-PI, although these four samples have similar Si/Al ratio. This is also verified by the ²⁹Si MAS NMR result (Figure 2D). The resonance peaks at –100 and –107 ppm are corresponded to the Si atoms bonded to two Al atoms (Si(2Al)) and one Al atom (Si(1Al)), respectively, while that at –112 and –116 ppm can be attributed to the Si atoms bonding without vicinal Al atoms (Si(0Al)).^33,48 The structure of Si(2Al), e.g., Al–O–Si–O–Al configuration, is generally unstable and generates less active and weaker BAS due to the strong repulsive forces.⁴⁹ The weaker signal at –100 ppm indicates formation of less numbers of Al–O–Si–O–Al species on FER-MORPP than on the other samples (Table S3). Thus, more isolated bridging Si–OH–Al species are formed on FER-MORP, consequently, a higher B_total content was obtained on FER-MORP.

Figure 3.

Al siting in different lattice T sites of various FER-x zeolites

(A–D) ²⁷Al MAS NMR spectra of FER-MORP (A), FER-EDA (B), FER-DAH (C), and FER-PI (D).

(E–H) Calculated occupation probability of Al in FER-MORP (E), FER-EDA (F), FER-DAH (G), and FER-PI (H).

Four components can be deconvolved from asymmetric signal in the range of 50–65 ppm in ²⁷Al MAS NMR because of the presence of four crystallographically distinct T sites. The chemical shifts (δ_cs) of these four components in ²⁷Al NMR estimated by DFT calculation^50,51 and the proportion of various peaks obtained by curve fitting of the ²⁷Al MAS NMR spectra are listed in Tables S4 and S5, respectively. The components at 57.5–60.0 and 56.5–57.5 ppm are related to the Al located at the T1 and T3 sites, while the other two at 54.5–55.0 and 51.1–52.5 ppm correspond to Al distributed in T2 and T4 sites, respectively. FER-MORP zeolite obviously shows higher occupation probability at T2+T4 sites (50.2%) than FER-EDA (46.1%), FER-DAH (41.7%), and FER-PI (39.0%). This gives another evidence for morpholine to direct more Al atoms or acid sites in the 8-MR than in the 10-MR channel and intersection. Notably, DFT calculation indicates that piperidine as a template agent is thermodynamically conducive to promoting Al siting in the 8-MR channel, being in contrast to the observation of ²⁷Al MAS NMR that more Al atoms are located at the 10-MR channel of FER. This implies that piperidine has a low template-directing ability during synthesis and the Al distribution is also affected by other crystallization conditions. One evidence is that the FER zeolite prepared by piperidine has lower crystallinity than other samples (Figure 1A, Table 1), while the result of crystallization curve also proves that piperidine needs higher synthesis temperature to achieve the same crystallinity as that of morpholine (Figures S7−S9).

Catalytic performances of FER-x zeolites for carbonylation of DME

Figure 4 displays the catalytic results for carbonylation of DME with CO over the four FER-x zeolites. Although all the samples show the MA selectivity ≥99.5%, their catalytic activities are largely different. FER-MORP shows a higher DME conversion (53%) than FER-EDA (48%), FER-DAH (39%), and FER-PI (34%). This is due to its larger amount of strong BAS in the 8-MR channels, which are considered to be primary active sites for DME carbonylation.^7,52 Figure 4B reveals that a linear relationship exists between the MA space-time yield (STY_MA) and amount of strong BAS in 8-MR channels, confirming that the strong BAS in 8-MR play a vital role in DME carbonylation.⁵² As shown in Table S6, FER-MORP also exhibits higher STY_MA (0.889 mmol g^–1 h^–1) and DME conversion (53%) than those of conventional FER zeolites (0.17–0.78 mmol g^–1 h^–1 and 11%–31% for STY_MA and DME conversion, respectively). This confirms the potential of FER-MORP for realistic applications.

Figure 4.

Catalytic performance for DME carbonylation reaction over FER-x zeolites

(A) Catalytic results of FER-x zeolites as a function of reaction time. Reaction conditions: 210°C, 3 MPa, DME/CO = 1/30, GHSV = 1544 mL g⁻¹ h⁻¹.

(B) Relationship between STY_MA and strong BAS content in 8-MR channels.

(C) DME conversion and MA selectivity of FER-MORP within four recycles in DME carbonylation.

In addition, smaller numbers of BAS in 10-MR channels endow FER-MORP with lower coking rate (0.048% h^–1) than FER-EDA (0.084% h^–1), FER-DAH (0.089% h^–1), and FER-PI (0.094% h^–1) (Figure S10). Thus, a higher catalytic stability is observed on FER-MORP, as corroborated by the DME conversion and MA selectivity well maintained at 38% and 99.5% after 132 h. Figure 4C shows the catalytic results obtained with four recycles with regeneration (∼460 h). Interestingly, the activity and selectivity of FER-MORP in DME carbonylation were fully recovered after each run, substantiating that FER-MORP can be well reused. This makes the lifetime of FER-MORP surpass most of the reported H-MOR in DME carbonylation (Table S7), although H-MOR shows higher catalytic activity because of its stronger acidic strength.⁵³

As shown in Figure S11, the crystal structure and morphology of FER-MORP are well maintained after four repeated tests, despite that the surface area is slightly decreased due to formation of small amounts of coke species.

Reaction mechanism for DME carbonylation on FER-x zeolites

The reaction mechanism for DME carbonylation on FER-x zeolites was investigated by in situ DRIFTS and DFT calculation. Figure 5 shows an obvious negative band at ∼3600 cm⁻¹ when DME enters into the in situ IR cell. This band is caused by the DME adsorbed on BAS.^54,55 Its intensity gradually decreases with the temperature, indicative of a desorption of DME at high temperature. The lower desorption rate of DME from BAS of FER-MORP than other three FER-x samples indicates a stronger adsorption of DME on FER-MORP, which benefits the subsequent carbonylation of DME with CO to form surface acetyl groups.

Figure 5.

In situ DRIFT spectra of DME adsorption on FER-x zeolites at different temperatures

(A, C, E, and G) OH regions of DME adsorption on FER-MORP, FER-EDA, FER-DAH, and FER-PI, respectively.

(B, D, F, and H) C-H vibration regions of DME adsorption on FER-MORP, FER-EDA, FER-DAH, and FER-PI, respectively.

In the C-H vibration region, the bands at 2945, 2897, and 2832 cm^–1, assigned to the C-H symmetric and asymmetric stretching vibrations of CH₃ and CH₂ groups, gradually decrease in intensity with the reaction temperature. When the reaction temperature is higher than 120°C, a new shoulder band at 2966 cm⁻¹, ascribed to the surface methoxy species,^29,54 appears along with a shift of the C-H asymmetric stretching band of -CH₂- groups from 2832 to 2839 cm⁻¹. In agreement with the OH-region vibrations, FER-MORP shows more intense bands at 2966 and 2839 cm⁻¹, implying that it has larger number of BAS in 8-MR channels, which favors DME activation and promotes the generation of more methoxy intermediates. This is supported by the time-dependent in situ DRIFTS result at 240°C (Figure 6), with the increase in the sorption time, more methoxy intermediates are detected on FER-MORP than on the other three samples.

Figure 6.

Time-dependent in situ DRIFT spectra of DME adsorption on FER-x zeolites at 240°C

(A) FER-MORP.

(B) FER-EDA.

(C) FER-DAH.

(D) FER-PI. The spectra were collected up to 60 min every 10 min.

After adsorption of DME on FER-x zeolites for 30 min and subsequent evacuation of 30 min at 210°C, 1.2 MPa CO was introduced into the reaction cell. When the reaction was carried out for 10 min, a new band at 1652 cm⁻¹ was observed, and its intensity enhanced with the reaction time (Figure 7). This band is ascribed to acetyl (CH₃CO∗) species.⁵⁵ Clearly, it is the most intense for FER-MORP, further confirming the higher activity of FER-MORP for the reaction between methoxy and CO due to the presence of more strong BAS in the 8-MR channels, being in line with the results of DFT calculation and ²⁷Al MAS NMR (Figure 3). It is further consolidated by the DFT calculation that formation of acetate species through carbonylation of methoxy species with CO shows lower free energy barrier and higher rate constant (122 kJ mol^–1 and 3.08 × 10^–1 s^–1) on T2 site in 8-MR channels than on T1 site in 10-MR channels (129 kJ mol^–1 and 6.09 × 10^–2 s^–1) (Table S8). Therefore, abundant strong BAS in 8-MR channels of FER-MORP not only enhances DME adsorption and activation but also promotes carbonylation of generated methoxy species with CO.

Figure 7.

In situ DRIFT spectra for carbonylation of methoxy with CO on FER-x zeolites at 210°C and 1.2 MPa

(A) FER-MORP.

(B) FER-EDA.

(C) FER-DAH.

(D) FER-PI.

(E) Dependence of 1652 cm⁻¹ peak intensity on FER-x zeolites as a function of reaction time.

Conclusions

A series of nanosheets FER zeolites with high crystallinity were synthesized using a two-step hydrothermal method. It was found that the distribution of Al atoms or acid sites in zeolite framework is closely related to the type and structure of the organic template agent. The results of DFT calculation, NH₃-IR, ¹H MAS NMR, ²⁷Al MAS NMR, and in situ DRIFTS suggest that use of morpholine as the organic template agent directs more Al atoms or strong acid sites to locate in the 8-MR channel of FER zeolite, thereby promoting the carbonylation reaction through forming more acetyl intermediates. Therefore, FER-MORP exhibits higher catalytic activity and stability in DME carbonylation. The DME conversion, MA selectivity, and MA space-time yield reach 53%, 99.5%, and 0.889 mmol g^–1 h^–1, respectively. Such a catalytic performance is well maintained even after four continuous cycles (∼460 h). This work not only designs an excellent catalyst for DME carbonylation but also provides an efficient method to controllably regulate the Al siting in FER zeolite framework.

Limitations of the study

This study reports a series of FER-x zeolites synthesized with four different organic templates. The location of Al atoms and strong acid sites can be effectively regulated by altering organic template. As a result, the activity of FER was improved. Future research will focus on the resolution of the Al distribution in zeolite framework using more advanced characterization techniques.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
ethylenediamine, AR	Sinopharm Chemical Reagent Co., Ltd.	CAS: 107-15-3
1,6-diaminohexane, AR	Sinopharm Chemical Reagent Co., Ltd.	CAS 124-09-4
Piperidine, AR	Sinopharm Chemical Reagent Co., Ltd.	CAS: 45393-79-1
Morpholine, ≥99.0%	Sinopharm Chemical Reagent Co., Ltd.	CAS: 110-91-8
Al2(SO4)3 · 18H2O, ≥99%	Sinopharm Chemical Reagent Co., Ltd.	CAS: 7784-31-8
silica sol, SiO2:40wt %	Qingdao Ocean Chemical Co., Ltd.	CAS: 112926-00-8
NaOH, ≥96%	Sinopharm Chemical Reagent Co., Ltd.	CAS: 1310-73-2
NH4NO3, ≥96%	Sinopharm Chemical Reagent Co., Ltd.	CAS: 6484-52-2

Resource availability

Lead contact

Further information and requests for resources should be directed to and will be fulfilled by the lead contact, Weibin Fan (fanwb@sxicc.ac.cn).

Materials availability

This study did not generate new unique reagents, and all reagents were commercially available and used without purification.

Additional information

Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Experimental model and study participant details

This study does not use experimental models.

Method details

General reagent information

Commercially available reagents and chemicals were purchased and used without further purification.

CO (99.999 vol. %) and DME (3 vol. %/Ar) were purchased from Shanxi YiHong Gas Industry Co. LTD. DME (99.99 vol. %) were purchased from Dalian Special Gases LTD. Other reagents used in this article are listed in the key resources table.

General analytical information

X-ray diffraction (XRD) patterns were recorded in the 2θ range of 5–40° at a scanning rate of 1.0° min⁻¹ on a Bruker D8 ADANCE A25 X-ray diffractometer with Cu Kα radiation. The relative crystallinity of FER-x sample was calculated by comparing the integrated area of the diffraction peaks at 2θ of 9.3° and 25.2° with that of the most highly crystalline FER-MORP sample. The chemical compositions of samples were measured by an inductively coupled plasma-optical emission spectrometer (ICP-OES, iCAP 6300). The field-emission scanning electron microscopy (SEM) images were taken on a JEOL JSM-7001F instrument. N₂ sorption isotherms of samples were measured at −196°C on an automatic Micromeritics Tristar II 3020 instrument after being degassed at 300°C for 8 h in vacuum. The amount of deposited coke species on the used catalysts was estimated by a Rigaku Thermo Plus Evo 8120 thermogravimetric instrument. 10 mg Sample was heated from room temperature to 800 °C at a rate of 10 °C min⁻¹ in an air flow (30 mL min⁻¹).

Temperature-programmed desorption of NH₃ (NH₃-TPD) was carried out on a Micrometrics Autochem II 2920 equipped with a TCD detector. 0.1 g catalyst was loaded in a U-type fixed bed quartz reactor and activated at 550°C for 1 h in an Ar flow (30 mL min⁻¹). Then, it was cooled down to 120°C, and gaseous ammonia (5 vol % in Ar, 30 mL min⁻¹) was introduced and allowed to be adsorbed for 0.5 h, followed by flushing the sample with the Ar flow for 1 h. Finally, the temperature was heated from 120°C to 600°C at a rate of 10°C min⁻¹, and the desorbed NH₃ was monitored by a thermal conductivity detector (TCD).

Pyridine-adsorption Fourier transform infrared spectra (Py-IR) were measured on a Bruker Tensor 27 FT-IR spectrometer equipped with an MCT detector cooled by liquid nitrogen. Self-supported zeolite wafer was firstly loaded in the cell and pre-treated at 400°C for 2 h in vacuum (10⁻² Pa). Then, pyridine vapor (200 Pa) was introduced into the cell for adsorption at 150°C for 0.5 h. The physically adsorbed pyridine was evacuated at 10⁻² Pa for 1.5 h. Subsequently, the IR spectra were collected at 150°C, and the amounts of BAS (1545 cm⁻¹) and Lewis (1450 cm⁻¹) acid sites (LAS) were calculated by the following equation: $C=\frac{A}{\epsilon }\times \frac{S}{m}\times 1000$ where C represents the acid site amount (μmol g⁻¹), A is the band area (cm⁻¹), S is the wafer surface area (1.327 cm²), ε is the molar extinction coefficient (cm μmol⁻¹), and m is the sample mass (mg). The extinction coefficients used for the calculation of BAS and LAS are 1.13 and 1.28 cm μmol⁻¹ respectively.⁵⁶

The BAS with different acid strengths were measured by NH₃-adsorption FTIR spectra. Self-supported zeolite wafer was firstly loaded in the cell and pre-treated at 400°C for 2 h in vacuum (10⁻² Pa). Then, NH₃ (200 Pa) was introduced into the cell for adsorption at 35°C for 0.5 h. The physically adsorbed NH₃ was then evacuated at 10⁻² Pa for 1.5 h. The IR spectra were recorded at 150°C, 250°C and 350°C, respectively.

The ²⁷Al, ²⁹Si and ¹H magic angle spinning (MAS) nuclear magnetic resonance (NMR) spectra were measured on a 600 MHz Bruker Avance III nuclear magnetic resonance spectrometer operating at a magnetic field of 14.1 T. The ²⁷Al MAS NMR spectra were recorded at a spinning rate of 12 kHz and a π/12 pulse width of 0.3 μs, while the ²⁹Si MAS NMR spectra were acquired at a spinning rate of 8 kHz, a π/2 pulse width of 4.2 μs and a recycle delay of 20 s. In the case of ¹H MAS NMR spectra, they were measured at a spinning rate of 12 kHz, a π/2 pulse width of 3.05 μs and a recycle delay of 5 s and scanned for 64 times.

The DME adsorption, activation and carbonylation behaviors on various samples were investigated by in situ diffuse reflectance infrared Fourier transform (DRIFT) spectroscopy that is performed on a Bruke Vertex 80 instrument equipped with an MCT detector and an in situ cell (with ZnSe windows) and set in a transmission pool. Before the adsorption of DME (3 vol. %/Ar), the sample was firstly pre-treated at 450°C for 2 h in vacuum (10⁻² Pa). Then, it was cooled to 40°C for adsorption of DME for 10 min. The gaseous and physically adsorbed DME was removed by evacuation at 10⁻¹ Pa for 0.5 h. The spectra of adsorbed DME on the samples at different temperature and evolved with time at 240°C were collected. For the DME carbonylation, after saturation adsorption of DME for 20 min, the CO was introduced and the reaction was performed at 1.2 MPa and 210°C. The in situ IR of formed acetyl groups with reaction time were then detected.

Synthesis of FER zeolite with different types of OSDAs

The FER sample was hydrothermally synthesized with NaOH, Al₂(SO₄)₃ · 18H₂O, silica sol, deionized water and EDA, DAH, PI or MORP. Typically, 2.38 g NaOH and 4.33 g Al₂(SO₄)₃ · 18H₂O were firstly dissolved in 35.1 g deionized water. Then, 19.50 g silica sol (40 wt % SiO₂) was added into the above solution drop by drop. This is followed by adding certain amounts of organic template. The resultant gel was stirred at room temperature for 2 h, and then, sealed into a Teflon-lined stainless steel autoclave, which was carried out at 120°C and 170°C for 48 and 24 h (15 rpm), respectively. The crystallization was stopped by quenching the autoclave in cold water. The solid product was centrifuged, washed with deionized water for several times, dried at 80°C overnight, and calcined at 570°C for 13 h. The Na-formed sample was ion-exchanged with 1 M NH₄NO₃ solution at 80°C for 4 h (Liquid/Solid = 40 mL g⁻¹), and repeated for 3 times. After that, it was dried at 80°C and calcined at 560°C for 8 h. The resultant solid was designated as FER-x (x represents the organic template). Notably, the yield of FER zeolites in synthesis process is above 90%.

DME carbonylation

DME carbonylation was carried out at 210°C and 3.0 MPa in a stainless steel fixed-bed reactor with an inner diameter of 8 mm. Typically, 2.4 g catalyst (20–40 mesh) was loaded and pre-treated at 500°C for 2 h in a N₂ flow (50 mL min⁻¹). Then, it was cooled to 210°C, and DME and CO gas mixture with a molar ratio of 1/30 was introduced at gas hourly space velocity (GHSV) of 1544 mL g⁻¹ h⁻¹. The product was kept at 160°C and analyzed by an online Shimadzu GC-2010 Plus gas chromatograph. DME conversion and methyl acetate (MA) selectivity were calculated by the following equations:

$${\mathrm{X}}_{\mathrm{D}\mathrm{M}\mathrm{E}}=\left(1-\frac{2\mathrm{D}\mathrm{M}{\mathrm{E}}_{\mathrm{o}\mathrm{u}\mathrm{t}\mathrm{l}\mathrm{e}\mathrm{t}}}{2\mathrm{M}{\mathrm{A}}_{\mathrm{o}\mathrm{u}\mathrm{t}\mathrm{l}\mathrm{e}\mathrm{t}}+\mathrm{M}\mathrm{e}\mathrm{O}{\mathrm{H}}_{\mathrm{o}\mathrm{u}\mathrm{t}\mathrm{l}\mathrm{e}\mathrm{t}}+\sum _1^{\mathrm{n}}\mathrm{n}{\mathrm{C}}_{\mathrm{n}}{\mathrm{H}}_{\mathrm{m}\mathrm{o}\mathrm{u}\mathrm{t}\mathrm{l}\mathrm{e}\mathrm{t}}+2\mathrm{D}\mathrm{M}{\mathrm{E}}_{\mathrm{o}\mathrm{u}\mathrm{t}\mathrm{l}\mathrm{e}\mathrm{t}}}\right)\times 100\%$$

$$Se{l}_{MA}=\left(\frac{2M{A}_{outlet}}{2M{A}_{outlet}+MeO{H}_{outlet}+\sum _1^{n}n{C}_n{H}_{moutlet}}\right)\times 100\%$$

The production rate of MA (space time yield, STY_MA) was defined as the average moles of MA produced at the reaction duration of 8–76 h [mmol g⁻¹·h⁻¹].

The used FER-MORP zeolite was regenerated in air flow (50 mL min⁻¹). The detailed regeneration procedure was described in Figure S12.

DFT calculations

The periodic density functional theory (DFT) calculations for the substitution energies and probability of Al for Si, and the reaction kinetics of DME carbonylation were carried out using Vienna Ab Initio Simulation Package (VASP 5.4.1).^57,58 The Perdew, Burke, and Ernzerhof (PBE) exchange–correlation functional was used and the projected augmented wave (PAW) method was employed to estimate the electron–ion interactions with the kinetic energy cutoff of 400 eV.^59,60 The Monkhorst Pack mesh k-points of (1 × 1 × 2) were applied by allowing convergence to 1 × 10⁻⁴ eV of the total electronic energy and below 0.05 eV/Å of the remaining total force. The D3 Becke-Jonson (BJ) correction was adopted to account for the dispersion interactions.⁶¹ The periodic structure models of FER zeolite containing 36 Si atoms sited at 4 distinct T sites were obtained from the zeolite database. The relative substitution energy (ΔEi) is the energy difference resulted from the substitution of Si with Al in T_i lattice site of FER zeolite. It can be calculated by the following equation:⁴⁰

$${\Delta E}_i=\left(E_{\left(Al-Z\right)}+E_{Si\left(OH\right)4}\right)-\left(E_{Si-Z}+E_{Al{\left(OH\right)}_3}\right)$$

The probabilities (P_i) of Al occupying at T_i site are estimated by the equation:

$$P_i=\frac{m_i\mathit{exp}\left(-\Delta E_i/Rt\right)}{\sum _i{m}_i\mathit{exp}\left(-\Delta E_i/Rt\right)}$$

where ΔE_i is the relative substitution energy of Al for Si at T_i site, Rt is the hydrothermal synthesis temperature (170°C) of FER zeolite, and m_i is the multiplicity of T_i site.

Transition state structures for carbonylation reaction were obtained by using the climbing image nudged elastic band method (CI-NEB).⁶² For each optimized stationary point, vibrational analysis was performed at the same level of theory to determine its character (minimum or saddle point).

Quantification and statistical analysis

This study does not include statistical analysis or quantification.

Additional resources

This study has not generated or contributed to a new website/forum and it is not part of a clinical trial.

Published: August 28, 2023

Data and code availability

• •Data reported in this paper will be shared by the lead contact upon reasonable request.
• •This paper does not report original code.
• •Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon reasonable request.