A study on highly active Cu-Zn-Al-K catalyst for CO₂ hydrogenation to methanol

2.1 Catalyst preparation

CuZnAl samples were synthesized by using traditional coprecipitation method with Cu/Zn/Al mole ratio of 1:1:1. In a typical procedure, 0.1074 mol of Cu(NO₃)₂·3H₂O (Panreac, 98%), 0.1074 mol Zn(NO₃)₂·6H₂O (Aldrich, 99%) and 0.1074 mol Al(NO₃)₃·9H₂O (Panreac 98%) were dissolved in deionized water for preparation of a 2 M solution (Solution 1). Solution 2 was prepared by dissolving K₂CO₃ (0.203 mol) (CDH, 98%) and KOH (0.7406 mol) (CDH, 98%) in deionized water for 4 M solution. Solution 2 was added into solution 1 drop wise under constant stirring condition. The precipitation continued until pH of the solution reached 11. Then the mixture was stirred for 30 min after the base addition and the precipitate was aged at 70 °C for 18 h. Further, the suspension was filtered and the solid residue was seperated. The wet solid was dried at 110 °C for overnight to obtain the CZA-K catalyst. Identical procedure was followed for the synthesis of CZA-Na and CZA catalysts by using 4 M solutions of Na₂CO₃/NaOH and (NH₄)₂CO₃/NH₄OH respectively. The resultant dried solids were calcined at 400 °C for 5 h under static air.

2.2 Characterization of samples

All the calcined and spent CZA samples were characterized by using the following techniques.

BET surface area-poresize analysis was done by using Quantachrome Nova Station (USA). The samples were pretreated under vacuum for 2 h at 150 °C. XRD analysis was done by using EQUINOX 1000 Inel XRD machine at Co Kα = 1.7902 Å with acquisition of 2θ from 10° to 110°. X-ray Fluorescence (XRF) analysis was carried out using AMPTEK XRF Kit in Equinox 1000 (France) system and was operated at 100 kV and 30 mA. XPS analysis was done by using SPECS GmbH analysis system containing Mg Kα 1253.6 eV X-ray source and the samples binding energy (BE) was attained by using C1s 284.8 eV correction.

CO₂-TPD analysis was done by using micromeritics AutoChem HP 2950 V3.02 instrument. The mass spectral data was obtained by using ThermoStarTM GSD 320 quad core mass spectrometer. The m/z values followed were: m/z = 18 (H₂O), m/z = 28 (CO) and m/z = 44(CO₂). In a typical experiment, calcined samples of 150 mg each was placed in a quartz tube and pretreated at 200 °C in He flow (50 cm³ min⁻¹) for 1 h. Subsequently, the temperature of the sample was brought to 40 °C and saturated with 10% CO₂-He for 1 h. Then, the sample was purged with He flow (50 cm³ min⁻¹) for 1 h. Desorption of CO₂ was performed over the temperature range of 75–400 °C at a ramping rate of 10 °C min⁻¹. Desorption stream was analyzed simultaneously by using a TCD and mass detector through an automated split valve. HAADF-STEM-EDX images were collected on Tecnai 200 kV D1234 SuperTwin microscope with camera length of 97 cm. TG analysis was carried out on STA-449 F3, NETZSCH instrument by using 20 mg of spent sample loading.

Insitu DRIFTs analysis was carried out using BrukerTensor-II FTIR spectrometer equipped with DLATGS detector, Harrick praying mantis with high temperature and high pressure reaction chamber (4 mm thick ZnS windows (≥6 MPa)). Approximately 0.1 g of catalyst was loaded and it was reduced at 300 °C for 2 h under the flow of H₂ (20 cm³ min⁻). Subsequently, the chamber flushed with argon gas for 1 h. Then the reaction feed-gas (CO₂/H₂ = 1:4 mol ratio) was introduced into the chamber with an applied pressure of 3.0 MPa and collected the sample data.

High pressure temperature programmed reduction (HP-TPR) of the calcined CZA samples (0.150 g) were studied using 10% H₂–Ar mixture gas (50 cm³ min⁻) at 3.0 MPa applied pressure by using micromeritics AutoChem HP 2950 V3.02 instrument at a ramping rate of 10 °C min⁻.

2.3 Activity tests

Catalytic activity tests were performed using PID Eng & Tech micro reactor equipped with gas flowmeters and temperature controllers. These tests were carried out in the temperature range of 220–250 °C at 2400 h⁻ GHSV of CO₂/H₂ mole ratio equals to 1:4 and 3.0 MPa applied pressure. For example, 0.5 g of calcined catalyst was loaded into a stainless steel (SS) tube reactor and subsequently catalyst was treated with hydrogen gas (20 cm³ min⁻) at 300 °C for 2 h. Further, the reactor temperature was set back to 220 °C and purged with nitrogen gas. After attaining the steady state at set temperature, CO₂/H₂ feed gas was introduced into the reactor to attain the required pressure. The reaction stream at each temperature directed to Agilent 7890A GC equipped with TCD detector with Restek Plot Q column for CO₂ and CO evaluation and FID detector with HP Plot Q column for methane and methanol estimation. The CO₂ conversion (%X CO₂), methanol selectivity (%S MeOH) and methanol space time yield (STY, mmol. g_cat⁻. h⁻) calculated using following equations. The calibrated results were reproducible with in the sample standard deviation of ± 0.1%. $$\%XCO2=\frac{(InitialmolesofCO2-FinalmolesofCO2)}{\mathit{Initial}molesofCO2}\times 100$$ $$\%SMeOH=\frac{\mathit{Methanol}moles}{(InitialmolesofCO2-FinalmolesofCO2)}\times 100$$ $$\mathit{STY}MeOH=\frac{\left(\frac{\%XCO2}{100}\right)\times \left(\frac{\%SMeOH}{100}\right)\times (GHSV)}{22.4}$$

3.1 Influence of CO₂/H₂ mole ratio on methanol space time yield

The CO₂ to H₂ mole ratio in the present study was varied between 1:3 and 1:5 and the obtained results on CZA-K at 2400 h⁻ GHSV and 240 °C are presented as Fig. 1. Methanol yield was found low at 1:3 CO₂/H₂ ratio (12.2 mmol·g_cat⁻·h⁻), on the other side, maximum methanol yield was obtained at 1:5 CO₂/H₂ ratio (14.8 mmol·g_cat⁻·h⁻). It is obvious from the results that sluggish improvement in the methanol yield was observed from 1:4 CO₂/H₂ mole ratio to 1:5. The results suggest that at 1:3 ratio the partial pressure of H₂ was low while at 1:5 ratio the CO₂ partial pressure was low to improve the methanol formation at 2400 h⁻ GHSV. Hence, the optimum mole ratio of CO₂/H₂ was found to be around 1:4 on CZA-K catalyst and it was used for further studies.

Close

Fig. 1

Effect of CO₂/H₂ mole ratio on methanol yield.

Export to PPT

3.2 Catalytic activity studies

Catalytic activity results of CZA catalysts are presented as Fig. 2. Activity studies were carried out in the reaction temperature range of 220–250 °C under 3 MPa applied pressure of CO₂/H₂ feed-gas (1:4 mol ratio) with 2400 h⁻ GHSV. Among the studied catalysts, the maximum CO₂ conversion of 16% was obtained on CZA-K at 250 °C. On the other hand, only 13 and 12% of CO₂ conversion was obtained on CZA-Na and CZA catalysts respectively at this temperature. Further, maximum methanol selectivity of 98% was obtained on CZA-K with 10% CO₂ conversion at 220 °C. Methanol selectivity of studied catalysts was decreased about 5 to 7% with increase in the reaction temperature from 220 to 250 °C. Besides, selectivity of CO and CH₄ was slightly increased with increase in the reaction temperature. At 240 °C, 96% methanol selectivity with 14% of CO₂ conversion was observed on CZA-K catalyst. Whereas, only 11 and 94% followed by 9 and 92% CO₂ conversion and methanol selectivity was observed on CZA-Na and CZA catalysts respectively at this temperature. The greater catalytic activity of CZA-K was attributed to its greater reduced surface copper content (0.42%) followed by CZA-Na (0.30%). For comparison purpose, CZA-K washed catalyst activity was also studied under identical reaction conditions. However, the CO₂ conversion and methanol selectivity was decreased to 11.5 and 92% respectively at 240 °C. The results are in agreement with FTIR-DRIFTs data wherein intense CO formation was observed on CZA-K washed catalyst over unwashed one. Moreover, FTIR-DRIFTs spectra suggests the interaction of CO₂ with the potassium existed in the CZA framework led to K—O—(CO)—O species formation. To some extent, CO₂ dissociation to CO and subsequent CH₄ formation was limited by this species in presence of H₂. Residual sodium effect on the catalytic activity of Cu-Zn-Al in methanol synthesis using CO₂ and H₂ was reported by Jun et al. (1998). Lower CO₂ hydrogenation activity on this catalyst was attributed to low Cu dispersion and non-facile reduction of CuO phase in the presence of residual sodium. In contrast, we have observed very good CO₂ hydrogenation activity on our lab synthesized CZA-Na catalyst studied under similar reaction conditions. The major differences between Jun et al., and the present study as follows: (i) Jun et al., studied the lower mole ratio of Zn and Al in Cu-Zn-Al catalyst (1:0.8:0.15) whereas, Cu-Zn-Al mole ratio in this study was 1:1:1 (ii) The precipitation of CZA-Na was done at pH = 7, on the other hand, we precipitated at pH = 11 (iii) CZA-Na catalyst was air calcined at 350 °C for 12 h, whereas, we calcined our catalyst at 400 °C for 5 h in static air (iv) the overall reducibility of our CZA-Na catalyst was about 95% and copper metal area was about 18.0 m²/g versus 1.0 m²/g.

Close

Fig. 2

Catalytic activity results of CZA catalysts at CO₂/H₂ (1:4) GHSV 2400  h⁻ and 3.0 MPa.

Export to PPT

Fig. 3 demonstrates time on stream results on CZA catalysts at 240 °C with CO₂/H₂ (1:4 mol ratio) feed-gas pressure of 3.0 MPa and 2400 h⁻ GHSV. Steady catalytic activity was observed for 100 h of continuous operation on CZA-K catalyst with 14% of CO₂ conversion and 96% of methanol selectivity. On the other hand, CZA-Na and CZA catalysts showed steady CO₂ hydrogenation activity for 54 and 36 h respectively. Further, a slight decrease in the activity results was associated with carbon deposits formation on the active sites of CZA-Na and CZA catalysts observed through TGA studies (Table 1, Fig. S4). In addition, the particle size of spent CZA (60–120 nm) and CZA-Na (50–100 nm) catalysts found high over CZA-K (20–40 nm) suggests that sintering of active metal could be another reason for their low catalytic activity. Hence, CZA-K is a promising catalyst for mass production of methanol from CO₂ and H₂. Further, detailed catalyst characterization was reported and possible structure–activity correlation was discussed.

Close

Fig. 3

CZA catalysts stability test at 240 °C with CO₂/H₂ (1:4) GHSV 2400  h⁻ and 3.0 MPa.

Export to PPT

3.3 XRD analysis

Powder X-ray diffraction patterns of calcined and spent CZA catalysts were obtained by using Co (Kα = 1.7902 Å) source are presented as Fig. 4. All the calcined samples showed a broad X-ray pattern in the 2θ range between 40 and 45° with X-ray reflection maxima at 2θ = 42.7° which suggests the existence of Cu₂O and CuO phases. However, CuO phase was prominent in CZA-Na calcined sample over other two CZA calcined samples. On the other hand, spent CZA catalysts principally demonstrated metallic copper and Cu₂O phases. Furthermore, γ-Al₂O₃ phase was observed at 2θ = 65.7, 67.2 and 82.1° for CZA-Na and CZA spent samples. The results suggest that small sized metallic copper and Cu₂O crystals were obtained for CZA-K catalyst through K₂CO₃/KOH precipitation.

Close

Fig. 4

(a) CZA-cal (b) CZA-spent (c) CZA-K cal (d) CZA-K spent (e) CZA-Na cal and (f) CZA-Na spent.

Export to PPT

3.4 High-pressure TPR studies

Reducibility of CZA calcined samples were studied under high pressure (3 MPa) of 10% H₂-Ar probe gas and the obtained profiles were presented as Fig. 5. A wide range of TCD signal was observed in the reduction temperature range between 175 and 340 °C for all the three CZA samples. Further, reduction peak temperature maximum (Tmax) at 240 or 260 °C was associated with the reduction of dispersed and or isolated CuO to Cu₂O in CZA ternary oxide.

Close

Fig. 5

HP-TPR profiles of CZA samples.

Export to PPT

The sharp reduction profile of CZA-Na sample suggests the existence of relatively large sized CuO crystals over other two samples. On the other hand, a broad reduction profile in the temperature range between 260 and 350 °C for CZA-K sample suggests the continuous process of Cu₂O reduction into metallic copper. TPR results are in agreement with XRD results wherein, broad X-ray reflections were obtained for CZA-K sample. Furthermore, higher reducibility (95%) was observed for CZA-Na sample compared to CZA-K (70%) and CZA (63%) samples. No ZnO reduction was observed in these samples upto 350 °C of HP-TPR. We have also performed ambient pressure TPR experiments for these samples and observed a similar reduction pattern, however, the TPR profiles was shifted to higher side of the temperature scale by 10–20 °C.

3.5 CO₂-TPD-mass analysis

CO₂-TPD-mass analysis results of CZA samples are presented as Fig. 6. It is obvious that a broad desorption of CO₂ in the temperature range between 100 and 275 °C was observed with peak Tmax at 145 °C for CZA-Na and CZA-NH₃ samples. On the other hand, a broad CO₂ desorption peak in the temperature range between 75 and 300 °C was observed for CZA-K sample with Tmax at 110 and 145 °C. According to Kai et al. (2018) CO₂ can adsorb on potassium in Mg-Al-K hydrotalcites at lower adsorption temperatures. Further, Pasupulety et al. (2015a, 2015b) reported a CO₂ desorption peak at 130 °C for C₂ZA sample and attributed to linear mode of CO₂ sorption on metallic copper. In the present study, the CO₂ desorption observed at Tmax equals to 110 °C was attributed to K interacted CO₂ in CZA-K sample. Further, CO₂ interacted with metallic copper was desorbed at Tmax equals to 145 °C in all the three samples. The decreasing order of CO₂ desorption capacity (Table 1) as follows: CZA-K (22.0 µmol g_cat⁻) > CZA (19.0 µmol g_cat ⁻) > CZA-Na (18.0 µmol g_cat⁻). Hence, the results suggest that the potassium presence in the CZA framework improved the sorption capacity of CO₂ for CZA-K sample and there by influenced the CO₂ hydrogenation activity.

Close

Fig. 6

CO₂-TPD-mass analysis results of CZA samples.

Export to PPT

3.6 FTIR-DRIFTs studies

FTIR-DRIFTs studies on CZA-K catalyst was carried out at different chamber temperatures under CO₂/H₂ atmosphere (1:4 mol ratio) at 3.0 MPa pressure and the results are presented as Fig. 7A. All the spectra were collected for 2 h. No methoxy (—OCH₃) species was detected on this catalyst upto 180 °C. However, at 180 °C, symmetric and asymmetric vibrations of bidentate carbonate species were observed at 1360 and 1600 cm⁻¹ respectively. These carbonate vibrations were intensified with increase in the chamber temperature from 180 to 240 °C. Further, at 220 and 240 °C weakly intensified new bands were appeared at 2100 and 3000 cm⁻¹ respectively corresponds to CO and hydrocarbons. It is well established that CO formation was due to the reverse water gas shift reaction under the studied conditions. These type of surface carbonate intermediates were reported on K promoted hydrotalcites by Kai et al. (2018). After 2 h of CO₂ hydrogenation reaction on CZA-K at 3 MPa, the gas flow was switched to argon (20 cm³ min⁻) under ambient pressure at 240 °C to see the existing surface intermediates on this catalyst (brown line, Fig. 7A(g)). Essentially four signals were detected at 1360, 1420, 1580 and 2380 cm⁻¹ respectively corresponds to symmetric bidentate carbonates, K interacted carbonates (K—O—(CO)—O), asymmetric bidentate carbonates (slightly shifted to lower frequency in the absence of water vapor) and CO₂—H₂ vibrations in argon gas flow. The carbonates formation on K is in agreement with CO₂-TPD-mass analysis data wherein low temperature (Tmax = 110 °C) CO₂ sorption was observed for CZA-K sample. Under different gas atmospheres and chamber temperatures negligible variation in the formation of symmetric bidentate carbonates over asymmetric bidentate carbonates was observed on CZA-K. Hence, assymmetric bidentate carbonates were active and converting into methoxy species to methanol. According to Wang et al. (2020) bicarbonates were more active in the presence of water vapor and transformed into methanol via formate intermediate on Cu-ZnO-ZrO₂ catalyst. For comparison purposes, we have carried out the DRIFTs analysis on CZA-K washed catalyst to find out the role of K in the formation of methanol under identical reaction conditions (dark yellow line, Fig. 7A(f)). It is noteworthy that, K washed catalyst demonstrated relatively higher CO intermediate formation with considerable methoxy species. The results suggest that K presence in the CZA framework playing a vital role in methanol formation on CZA-K catalyst via K—O—(CO)—O species.

Close

Fig. 7

(A) DRIFTs spectra collected on CZA-K at different chamber temperatures (a) 23 °C, 3.0 MPa (b) 100 °C, 3.0 MPa (c) 180 °C, 3 MPa (d) 220 °C, 3 MPa (e) 240 °C, 3 MPa (f) spectra collected on CZA-K washed catalyst at 240 °C, 3 MPa and (g) spectra collected on CZA-K at 240 °C in argon gas flow, 0.1 MPa. (B) Time resolved transient DRIFTs spectra on CZA-K catalyst at 240 °C under CO₂-H₂ feed-gas pressure of 3.0 MPa.

Export to PPT

The time resolved transient DRIFTs spectra on CZA-K catalyst at 240 °C under CO₂-H₂ feed-gas pressure of 3.0 MPa is presented as Fig. 7B. Apart from CO₂-H₂ vibrations, new bands corresponds to —OCH₃, mono, bi-dentate carbonates, —CO and -C-H vibrations were detected on CZA-K catalyst. It is noteworthy that, a slight improvement in the relative intensity of —CO and —OCH₃ species was observed on this catalyst at different time intervals. The results indicate the steady formation of —OCH₃ and mono, bi-dentate carbonate intermediates on CZA-K catalyst for12h duration. In addition, DRIFTs spectra obtained on CZA-Na and CZA catalysts were presented as supplementary Fig. S1. Relatively higher CO formation was observed for these catalysts compared to CZA-K at 240 °C under 3.0 MPa CO₂/H₂ (1:4) feed gas pressure.

3.7 BET surface area and pore size distribution

BET surface area and poresize distribution results of spent CZA samples are presented as Fig. 8(a), (b) and Table 1. All the three samples exhibited type-IV isotherm with H4 hysteresis loop (Fig. 8a). Loop H4 often associated with narrow slit like pores (Sing et al., 1985) in studied samples. However, CZA-Na and CZA samples illustrated hysteresis loop starting from P₀ = 0.4 to 1, whereas, hysteresis loop was observed from P₀ = 0.8 to 1 in CZA-K sample. Among the studied catalysts, greater BET surface area was observed for CZA-Na (56.0 m² g⁻) followed by CZA-K (50.0 m² g⁻) and CZA sample (49.0 m² g⁻).

Close

Fig. 8

BET surface area and poresize distribution results of spent CZA samples.

Export to PPT

It is obvious from the poresize distribution graph in Fig. 8b that micro (<2 nm) and mesopores (2–7 nm) were dominant in CZA-Na and CZA samples. On the other hand, mespores in the pore width range of 2 to 7 nm were diminished for CZA-K sample. However, mesopores in the pore width range between 23 and 32.5 nm with pore width maximum of 28.5 nm were observed in potassium sample. The decreasing order of estimated pore volume for these samples as follows (Table 1): CZA-K (0.243 cc g⁻) > CZA (0.121 cc g⁻) > CZA-Na (0.105 cc g⁻). The results suggest that different precipitating agent used in the synthesis of CZA samples yielded different pore width and BET surface area.

3.8 HAADF-STEM-EDX studies

HAADF-STEM technique was used to identify the particle shape /size in spent CZA, CZA-K and CZA-Na catalysts and the images are depicted in Fig. 9. Irregular shaped particles with 60–120 nm size was observed for spent CZA sample (Fig. 9(a)). The corresponding EDX analysis revealed Cu, Zn and Al elements (Fig. 9(d)) in CZA sample. No nitrogen related EDX signals were detected in CZA sample. The result suggests that CZA precipitation with (NH4)₂CO₃/NH₄OH volatizes nitrogen residues during the calcination and or CO₂ hydrogenation process. On the other hand, two types of particles with platelet shape and tetragonal shape were obvious in Fig. 9(b). The focused beam EDX analysis of bulk tetragonal particles revealed Al as a main component. Whereas, the platelet particles were mainly associated with Cu and Zn elements with approximate length of 20–40 nm and the width of 2–5 nm. It is noteworthy that, potassium was found in the platelet particles of CZA upon careful EDX scanning (Fig. 9(e)). Further, The EDX analysis quantification of spent CZA-K sample as follows: Cu = 17.35 atm.%, Zn = 15.0 atm.%, Al = 21.0 atm.%, O = 46.0 atm.%, and K = 0.65 atm.%. Fig. 9(c) image revealed spherical particles about 50–100 nm size in Spent CZA-Na sample. The EDX analysis of CZA-Na revealed mainly Cu, Zn and Al elements. However, Na signal was masked by Cu and Zn EDX signal and it was not possible to quantify the Na content through this technique. Hence, Na content in spent CZA-Na sample was estimated through XRF technique as 0.60%.

Close

Fig. 9

HAADF-STEM-EDX analysis of spent CZA samples (a) CZA HAADF image and (d) CZA EDX; (b) CZA-K HAADF image and (e) CZA-K EDX; (c) CZA-Na HAADF image (f) CZA-Na EDX.

Export to PPT

3.9 XPS studies

XPS analysis of spent CZA samples are presented as Fig. 10. XPS results of Cu 2P3/2 orbital was depicted in Fig. 10a. Two peaks were observed at 935.0 and 933.0 eV after deconvolution of Cu 2P3/2 orbital corresponds to Cu²⁺ oxidation state related to CuO crystals and Cu¹⁺/Cu⁰ oxidation state was ascribed to Cu₂O/metallic copper (Pasupulety et al., 2015a, 2015b) in studied samples. However, the extent of the copper species on the surface was varied with respect to precipitating agent used in synthesis of these catalysts. The decreasing order of surface Cu¹⁺/Cu⁰ species as follows: CZA-K (0.42 atm.%) > CZA-Na (0.30 atm.%) > CZA (0.23 atm.%). Auger analysis of Zn was depicted in Fig. 10b. Principally Zn²⁺ species exhibited on the surface with minor portion of metallic zinc due to the partial reduction of ZnO at 987.4 and 991.0 eV respectively.

Close

Fig. 10

XPS and Auger analysis of spent CZA samples.

Export to PPT

In addition, CO₂ hydrogenation activity of CZA-K catalyst was compared to literature reports studied under similar conditions.

Table 2 shows the CO₂ conversion and methanol selectivity results reported on CZA hydrotalcite and Cu-Zn-Zr synthesized by different preparation methods. It is clear that excellant catalytic activity results were obtained on the present CZA-K catalyst due to the presence of K in the framework of CZA.