Modified natural kaolin clay as an active, selective, and stable catalyst for methanol dehydration to dimethyl ether

In this work, the production of dimethyl ether (DME) from methanol over natural kaolin clay modified through impregnation with various percentages of H2SO4, WO3, or ZrO2 catalysts was investigated. The prepared catalysts were characterized via X-ray fluorescence, X-ray diffraction, Fourier transform infrared spectroscopy, scanning electron microscopy, and N2-sorption analysis. The acidity of these catalysts was determined through the dehydration of isopropyl alcohol and the chemisorption of pyridine. The catalytic activity performance revealed that the addition of modifiers into kaolin enhanced the latter’s activity toward DME production. In addition, the kaolin clay modified with 10 wt% ZrO2 exhibited excellent activity of 98% conversion with 100% selectivity at 275 °C. Moreover, this catalyst could proceed the reaction for a long time (6 days) without any noticeable deactivation. The remarkable improvement in the catalytic performance achievement was well correlated with the acidity and the structure of the catalysts.

The structural changes in the kaolin material due to impregnation with different percentages of H 2 SO 4 , WO 3 , and ZrO 2 were studied using the XRD technique. The XRD profiles of the K400 sample and of those treated with H 2 SO 4 , WO 3 , and ZrO 2 are shown in Fig. 1b. The K400 sample shows well-defined reflections at 2θ values of 12.3°, 20.3°, 23.1°, 24.9°, 38.3° and 39.3° which well-matched with the data bank of kaolinite as a major phase (COD card No: 9009230). Diffraction peaks of quartz (COD card No: 1011176) could be found at 2θ = 50.2° and 60.1°. Illite (COD card No: 9013732) was detected at 2θ = 26.6°. Muscovite (COD card No: 1000042) was also observed at 2θ = 35°. In addition, Calcite (COD card No: 1010928) was detected at 2θ of 29.5°, 43.2°, 47.6° and 48.6°. The XRD diffraction patterns of the kaolin modified by 5-15 wt% of H 2 SO 4 showed that the peak intensity of kaolin clay gradually decreased with increasing H 2 SO 4 percentage. This behavior indicates that acid treatment leads to structural disorders, which affects the crystallinity of the kaolin 37 . In addition, the XRD peak that assigned to calcite phase at 2θ = 28.9° is completely disappeared due to the reaction occurred between CaCO 3 and sulfuric acid. The intercalation of WO 3 also slightly lowered the intensities of the characteristic's peaks of the kaolin clay. Moreover, new reflections appeared at 2θ = 23.3°, 23.8°, 38.1° and 38.9°, which typically characterize triclinic WO 3 (COD card No: 1010618). The intensities of these reflections gradually increased with increasing WO 3 loading on the kaolin. On the other hand, the XRD patterns of the kaolin samples incorporated with different loading percentages of ZrO 2 were similar to that of K400, indicating that the intercalation with ZrO 2 did not damage the kaolin crystal structure. This means that all the diffraction peaks of kaolin still appeared after the intercalation; nevertheless, the peak intensities slightly decreased. This result reflects that ZrO 2 was successfully incorporated into the kaolin clay support. New peaks came out at 2θ = 30.2°, 50.7°, and 60.2°, and their intensities increased with increasing ZrO 2 weight percentage were observed. These new peaks corresponded to the tetragonal phase of ZrO 2 (COD card No: 1525706), which may have a role in enhancing the catalytic activity of these samples.
The FTIR spectra of the unmodified K400 and of the modified kaolin calcined at 400 °C ( Fig. 2) were taken to analyze the vibrational bands and the interface interaction potentially responsible for promoting catalytic activity. The K400 sample showed all the characteristic vibration bands of kaolinite. The bands at 3694, 3670, and 3652 cm −1 were assigned to an inner-surface OH-stretching modes of kaolinite 38 , whereas the band at 3620 cm −1 was associated with the stretching mode of the inner hydroxyl group of kaolinite 39 . The small and broad band at 3450 cm −1 may be attributed to water physisorbed on the surface of kaolin 40 . The band at 1430 cm −1 was ascribed to a vibrational mode of CO 3 2− group which is due to the presence of calcite 41 . The bands at 1632, and 915 cm −1 were attributed to water functional group (H-O-H) 41 32 , respectively. After the kaolin was modified by low ratios of H 2 SO 4 (1-3 wt%), a little variation in the band patterns was observed. However, with a further increase in the acid loading (5-15 wt%), some bands appeared, disappeared, or shifted, and their intensities also changed. In this issue, the structural hydroxyl vibrations bands are progressively decreased due to the dehydroxylation process which caused by the penetration of acid protons into the kaolin layers and attack the structural OH group 47 . Moreover, an increase in the wavenumber from 1632 cm −1 in the K400 sample to 1667 cm −1 in the modified kaolin 48 , and the latter band was more intense. Furthermore, the stretching vibration of -OH from water adsorbed was appeared broader, more intense and at much lower wavenumber 3100 cm −149 . These results confirmed that the modification of kaolin by H 2 SO 4 influences the OH bending and stretching vibrations of water. The change in the position and the intensities of the latter two bands in the modified kaolin may have a role in the catalytic dehydration of methanol to DME. Meanwhile, two doublet bands also appeared at 610 and 594 cm −1 , which may be due to the bending modes of SO 4 −250,51 . The asymmetric and symmetric stretching modes of O=S=O also detected at 1120-1230 cm −1 and 1010-1080 cm −1 , respectively, and these bands are overlapped with that of the Si-O-Si band 52 . The enhancement in the band's intensities in 960-1545 cm −1 region with increasing the loading of H 2 SO 4 on kaolin was observed. This enhancement may be due to the increase of the concentrations of the stretching vibration of the S=O bond, the symmetric vibrations of Si-O-S bridges 53 , and the increase in the amorphous silica percentages 54 . It is also noted that the band corresponded to the presence of calcite (1430 cm −1 ) disappeared due to the reaction of CaCO 3 with H 2 SO 4 . This observation agrees with that obtained from the XRD peak analysis; the diffraction peaks of calcite disappeared in the H 2 SO 4 modified kaolin samples. The FTIR spectra of the kaolin treated with different percentages of WO 3 are presented in Fig. 2. No apparent change in the intensities and the positions of the characteristic bands of kaolin was observed. This may be due to the weak intensity of the characteristic bands of WO 3 or the band corresponding to W-O-W that appeared at a position of 800 cm −1 which like that of the quartz band. Figure 2 shows negligible change in the intensities or the positions of the characteristics bands of kaolin when it was impregnated with ZrO 2 . No characteristic bands of ZrO 2 appeared in these samples because of the very low intensities of the bands related to ZrO 2 itself or the bands corresponding to Zr-O-Zr or Zr-O-Si have positions similar to those of the band's characteristics of the K400 structure. Accordingly, the absence of the FTIR bands due to the presence of WO 3 and ZrO 2 in the modified samples indicates that both oxides are highly dispersed or incorporated into the kaolin framework.
BET measurements were conducted to determine the effects of the H 2 SO 4 , WO 3 , and ZrO 2 additions on the texture properties of the K400 sample. The isotherms of the K400 and the modified kaolin clay catalysts (Fig. 3) belonged to typical type II isotherms and possessed an H3 hysteresis loop. According to IUPAC classification, this demonstrates that the K400 and modified K400 catalysts have micropores formed owing to the presence of aggregates of plate-like particles, giving rise to slit-shaped pores 55 . Table 1 shows the S BET and porosity characteristics of the samples. The calculated value of the S BET of the K400 clay is 25.2 m 2 /g. Upon the impregnation of the K400 sample with H 2 SO 4 , the specific surface area and the pore volume fell regularly because of the partial collapse of the clay structure due to the strong acid penetration of the crystal structure 56 . The modification of the K400 sample with WO 3 also reduced the pore volume and specific surface area. These findings indicate that the WO 3 blocked the pores of the K400 sample, which are responsible for such decrease. Conversely, the modification of the K400 sample with 1-3 wt% ZrO 2 decreased the specific surface area and pore volume. The S BET value increased upon a further increase in the ZrO 2 content but was still lower than that of the K400 sample. The morphologies of the K400, 10% H 2 SO 4 /K400, 10% WO 3 /K400, and 10% ZrO 2 /K400 catalysts were studied using SEM, and the obtained images are shown in Fig. 4. The K400 sample has aggregates of semispherical structure and agglomerated particles. The SEM image of 10% H 2 SO 4 /K400 shows flaky particles stacked together.    Table 2). The increase in the total acidity of 10% H 2 SO 4 /K400 is due to the substitution of exchangeable cations of Na + and Mg 2+ by H + ions and the formation of bridging hydroxyl groups between nearest neighbor Al atoms of the Si atom 61 . Conversely, the increase in the total acidity of 10% WO 3 /K400 and 10% ZrO 2 /K400 is due to the presence of triclinic WO 3 and tetragonal ZrO 2 in the modified kaolin, respectively, as confirmed previously from the XRD results. Finally, based on the total acidity, the samples take the following order: K400 < 10% H 2 SO 4 /K400 < 10% WO 3 / K400 < 10% ZrO 2 /K400.
Catalytic dehydration of methanol. The catalytic dehydration of methanol over the K400 and modified kaolin catalysts was conducted to measure the catalytic activity in terms of methanol conversion in the temperature range of 200-350 °C (Fig. 6). As seen in Fig. 6, all catalysts exhibited nearly similar behaviors, where the methanol conversion increased with increasing reaction temperatures from 200 to 350 °C with 100% selectivity to DME. The observed differences in catalytic activity could be related to the acidity, the hydrophobicity 22 , and the texture properties of the catalysts. The K400 sample is an active catalyst with a maximum methanol conversion and DME yield ≈ 98% obtained at 350 °C. However, it shows very low activity at 200 °C (only 7% methanol conversion). The kaolin clay was impregnated with different percentages of H 2 SO 4 to improve its catalytic activity at relatively low temperatures, and the results are shown in Fig. 6a. It shows that at 200 °C, acid impregnation significantly increases the catalytic activity from 7 to 30% over the kaolin modified by 10 wt% of H 2 SO 4 . However, when the acid percentage was further increased to 15 wt%, a sharp decrease in the catalytic activity was observed. Moreover, the conversion of methanol increased linearly by increasing the percentages of H 2 SO 4 to 10 wt%. However, the maximum methanol conversion with DME yield ≈ 98% was obtained over the catalysts modified with 5 and 10 wt% of H 2 SO 4 at a reaction temperature of 300 °C. At 300 °C, the methanol conversion rates for this series of catalysts follow the sequence 15% H 2 SO 4 /K400 < K400 < 1% H 2 SO 4 /K400 < 3% H 2 SO 4 / K400 < 5% H 2 SO 4 /K400 = 10% H 2 SO 4 /K400. Conversely, at 200 °C, the methanol conversion increased to 45% over the 10% WO 3 /K400 catalyst (Fig. 6b). The maximum methanol conversion ≈ 98% and DME yield ≈ 98% were achieved over the 10% WO 3 /K400 catalyst at 300 °C. Moreover, at this temperature, the activities of the catalysts follow the order K400 < 1% WO 3 /K400 < 3% WO 3 /K400 = 5% WO 3 /K400 < 10% WO 3 /K400 = 15% WO 3 / K400. Additionally, in the case of the ZrO 2 -modified kaolin series of catalysts (Fig. 6c), at 200 °C, the methanol conversion increased from 7 to 48% with the increased percentage loading of ZrO 2 to 10%. At 300 °C, the catalytic activity follows the order K400 < 1% ZrO 4 /K400 < 3% ZrO 2 /K400 < 5% ZrO 2 /K400 < 15% ZrO 2 /K400 < 10% ZrO 2 /K400. However, the maximum methanol conversion is nearly 98% with a 98% yield of DME was obtained over the 10% ZrO 4 /K400 catalyst at 275 °C. A plot of the temperature where the complete dehydration of methanol to DME (T 98 ) for the K400 and modified kaolin catalysts is presented in Fig. 7a , to compare the catalytic performance of the most active catalysts. The reaction temperature required to obtain the complete conversion of methanol to DME takes the following order: K400 > 10% H 2 SO 4 /K400 = 10% WO 3 /K400 > 10% ZrO 2 /K400. Therefore, the 10% ZrO 2 /K400 catalyst had the highest catalytic activity at relatively lower temperature (275 °C). This high catalytic activity may be attributed to the formation of tetragonal zirconia which accompanied by high acidity values, as mentioned previously in the XRD and acidity sections. Our group previously found that the formation of the tetragonal phase of zirconia caused the greatest enhancement in the acidity and the catalytic activity of sulfated zirconia 17 . A similar enhancement was also obtained by modifying FePO 4 with 10% of ZrO 2 62 . Kou et al. 63 found that Zr-pillared clay samples are quite effective in dehydrating methanol to DME and hydrocarbons because of their porosity and acidity. Methanol dehydration to DME was studied on Zr-loaded P-containing mesoporous activated carbon catalysts 64 . The catalyst with 5.25% Zr-loading showed the highest methanol conversion ≈ 69% with 95% selectivity to DME. The high catalytic activity of this catalyst may be attributed to the increased acidity due to the formation of zirconium phosphate active species. Chmielarz et al. 65 found that the intercalation of porous clay heterostructures with Zr increased the conversion of methanol to 73% with nearly 100% selectivity to DME at 325 °C. They assumed that the acid sites on this catalyst are in the clay mineral layers and are related to the incorporated Zr 4+ cations. www.nature.com/scientificreports/ Catalyst stability. Catalyst stability in terms of conversion and selectivity is the most important parameter reflecting catalyst quality. Hence, the stability of the most active catalyst (10% ZrO 2 /K400) for methanol dehydration to DME was tested for 6 days at 275 °C. The results are depicted in Fig. 7b. The maximum conversion of methanol (98%) to DME with 100% selectivity was observed to be nearly unchanged for a long duration. This means that the catalyst did not exhibit any sign of deactivation under the reaction conditions, which was con-  www.nature.com/scientificreports/ firmed by completely matching the XRD patterns of the fresh and the spent catalyst (Fig. S1). Thus, this catalyst has excellent stability to produce DME from methanol. Table 3 shows the comparison of the catalytic performance of the modified kaolin catalysts with that of other clay minerals. Despite the different reaction conditions used in this study and other studies, the modified kaolin presented herein showed a remarkable activity and selectivity than those of the clay catalysts described in the literature. Additionally, the modified kaolin with ZrO 2 achieved the complete conversion of methanol with nearly 100% selectivity to DME at relatively low temperature (≈ 275 °C). Moreover, this result is in comparable with the other catalysts published in the literature. This advantage makes modified kaolin with H 2 SO 4 , WO 3 and ZrO 2 are promising catalysts for dehydrating methanol to DME.

Conclusion
Natural kaolin clay impregnated with different percentages of H 2 SO 4 , WO 3 , or ZrO 2 exhibited high catalytic performance in the gas-phase dehydration of methanol to DME at relatively low temperatures. The results revealed that the kaolin structure and acidic properties were greatly influenced by the ratios of the modifiers used. The observed differences in the catalytic activity of the modified and unmodified kaolin could be attributed to the acidity, hydrophobicity, and the texture properties of the catalysts. The catalyst containing 10 wt% of ZrO 2 exhibited excellent activity of 98% conversion with 100% DME selectivity. The high catalytic performance of this catalyst was correlated to the formation of tetragonal zirconia, which accompanied with high acidity values. Additionally, this catalyst displayed long-term stability toward methanol dehydration to DME up to 6 days without deactivation. Thus, it is a potentially suitable catalyst for producing DME from methanol at a temperature of 275 °C. The obtained results in this study will open doors for the preparation of highly active, eco-friendly, and cost-effective catalysts based on natural kaolin and their application in acid-catalyzed reactions.

Materials and methods
Materials. Synthesis of modified kaolin. The kaolin clay used in this study was collected from the Aswan area, Egypt, and sieved to an average particle size of 0.25 mm. The chemical analysis of the raw sample through X-ray fluorescence (Table 4) shows that the sample had a high percentage of SiO 2 and Al 2 O 3 , with a SiO 2 /Al 2 O 3 ratio equal to 1.5. The sample was thermally activated in a muffle furnace at 400 °C for 3 h to remove the physically adsorbed water and improve its surface properties before use. It was then labeled as K400. The modified kaolin was prepared by impregnating the K400 sample with an aqueous solution of known quantities of H 2 SO 4 or phosphotungstic acid or zirconyl nitrate. The impregnated samples were dried in an oven at 110 °C for 24 h and then calcined in a muffle furnace at 400 °C for 3 h under a static air atmosphere. The H 2 SO 4 , WO 3 , and ZrO 2 contents in the modified samples varied between 1 and 15 wt%. The modified samples were named x% modifier/K400, where x is the weight percentage of the modifier used.
Catalyst characterization. Thermal behavior (thermogravimetric (TG) and differential thermal analysis (DTA)) of the raw kaolin clay were studied using a Shimadzu thermal analyzer (Japan, 60H) using air as a heating atmosphere and a heating rate of 10 °C/min.
The structure of the catalysts was identified via X-ray diffraction (XRD) using a Philips diffractometer (model PW 2103/00) equipped with a Ni-filtered Cu Kα radiation (λ = 1.5408 Å). The samples were scanned over the 2ϴ range from 4° to 80°, at a scan rate of 2°/min.
Most functional groups of the catalysts were investigated via FTIR in the 4000-400 cm −1 region, which recorded on a Nicolet spectrophotometer (model 6700) by using KBr pellet technique. In this technique, the samples were prepared by mixing the catalyst powder with KBr (w/w 1:100) together and then compressed into thin pellets. The results were collected at a resolution of 2 cm −1 .
The catalyst microstructure and crystal morphology analysis were determined using a scanning electron microscope (JEOL Model JSM-5400 LV, Jeol, Tokyo, Japan).
The surface area and texture of the catalysts were determined via N 2 adsorption at − 196 °C using the gas adsorption apparatus Nova 3200 instrument (Quantachrom Instrument Corporation, USA). The Brunauer-Emmett-Teller (BET) model was used to calculate the surface area.
Isopropyl alcohol (IPA) was dehydrated as previously described by Said et al. 11 , and the pyridine FTIR (Py-FTIR) adsorption technique were used 59 . Those were done to measure the acidity of the catalysts. For pyridine FTIR adsorption experiments, about 30 mg of the samples were grounded and mixed with KBr and then pressed with a pressure in a self-supporting disc in air. The spectra of the samples without pyridine were recorded with a spectral resolution of 2 cm −1 . Afterward, the discs were heated to 200 °C in a drying oven for 3 h to remove any possible physisorbed species before saturated with pyridine for 7 h after evacuation in desiccator. Then, the www.nature.com/scientificreports/ pyridine excess was removed for 30 min under vacuum and the spectra were recorded. To determine the bands relevant to Lewis and Brønsted acidic sites, the spectra obtained after pyridine adsorption were subtracted from those obtained before pyridine adsorption (fresh samples) 70 .
Methanol dehydration. The methanol dehydration reaction was conducted in a conventional fixed-bed flow-type reactor using 500 mg of modified kaolin, 6000 ml/g/h gas hourly space velocity, and 4% methanol in the gas feed. The reaction was conducted in the temperature range of 200-350 °C under atmospheric pressure. The reactant and products were analyzed using a gas chromatograph (Pro-GC Unicam) with a thermal conductivity detector and a 2 m DNP glass column. At least three successive data points were taken for each reaction at equilibrium temperature (1 h). The average of these points was used in calculating the conversion and selectivity values. The methanol conversion and DME selectivity were calculated as previously described 62 .

Data availability
All data generated or analyzed during this study are included in this published article.