Efficient Synthesis of Ethanol from CH4 and Syngas on a Cu-Co/TiO2 Catalyst Using a Stepwise Reactor.

Ethanol synthesis from CH4 and syngas on a Cu-Co/TiO2 catalyst is studied using experiments, density functional theory (DFT) and microkinetic modelling. The experimental results indicate that the active sites of ethanol synthesis from CH4 and syngas are Cu and CoO, over which the ethanol selectivity is approximately 98.30% in a continuous stepwise reactor. DFT and microkinetic modelling results show that *CH3 is the most abundant species and can be formed from *CH4 dehydrogenation or through the process of *CO hydrogenation. Next, the insertion of *CO into *CH3 forms *CH3CO. Finally, ethanol is formed through *CH3CO and *CH3COH hydrogenation. According to our results, small particles of metallic Cu and CoO as well as a strongly synergistic effect between metallic Cu and CoO are beneficial for ethanol synthesis from CH4 and syngas on a Cu-Co/TiO2 catalyst.

Owing to the diminishing supply of fossil fuels and rising crude oil prices, an alternative fuel source must be developed. Ethanol synthesis has recently attracted increasing attention because of its nontoxic nature and ability to be produced from renewable sources1. In general, there are two main methods of ethanol synthesis: one is fermentation derived from corn or sugar cane and hydration of petroleum-based ethylene, and the other is CO hydrogenation123456. Ethanol synthesis from syngas has recently received attention owing to food shortages. To the best of our knowledge, Rh-based catalysts are the best catalysts that show relatively high ethanol selectivity78910. However, the high cost of Rh limits its application in industry.

C2-oxygenate synthesis from CH4 and CO2 is thermodynamically unfavourable at low temperatures, but this can be overcome through a stepwise reaction technology that has been proposed by our group11. In this process, *CH4 is first adsorbed on the catalyst surface (M) and then dissociated to generate CHx-M; subsequently, the *CO2 species is inserted into the C-M bond to form *CHxCOO before finally forming acetic acid from *CHxCOO hydrogenation121314. It was found that the Pd-Co and Cu-Co bi-metal supported on TiO2 catalysts exhibited good activity for acetic acid from CH4 and CO215. Because CO2 has a relatively high reduction potential (1.9 V to CO2−), the conversion is difficult16. If CO2 is replaced by CO, then the conversion of CO is possibly better than that of CO2. Therefore, we propose a method of ethanol synthesis from CH4 and syngas in a stepwise reactor.

Although the activity of the Pd-Co/TiO2 catalysts is better than that of the Cu-Co/TiO2 catalysts for acetic acid synthesis from CH4-CO2 in the stepwise reactor15, we chose the Cu-Co/TiO2 catalysts for ethanol synthesis from CH4 and syngas, considering the price of Pd. Finally, the reaction mechanisms of ethanol from CH4 and syngas were studied on Cu-Co/TiO2 using density functional theory (DFT) and microkinetic modelling. The result may be useful for computational design and optimizations of Cu-Co/TiO2 catalysts.

Result and Discussion

Experimental result

Figure 1 shows the H2-TPR profile of the Cu-Co/TiO2 catalyst before reaction. The H2-TPR curves show four main peaks. The peak at 178 °C can be assigned as the reduction of CuO to Cu, and the peak at 238 °C is attributed to Cu–Co spinal phase (such as CuxCo3-xO4 oxides)417. The peaks at approximately 276 and 394 °C are assigned the reduction of Co3O4 → CoO and CoO → Co181920. Note that the reduction temperatures of Co3O4 → CoO and CoO → Co are approximately 450 and 550 °C18, which are higher than that of our catalyst. The reason for this observation is that the Cu species are first reduced at low temperatures to form metallic Cu nanoparticles, which subsequently catalyse the reduction of nearby Co species1819. As a result, the reduction temperature of the Co species in the Cu-Co/TiO2 catalyst is lower than that of the pure Co species. No reduction peak of TiO2 is detected, which is in accordance with our X-ray powder diffraction (XRD) (Fig. S1) results. XRD and high-resolution transmission electron microscopy (TEM) (Fig. S2) also show that Cu species and Co species are uniformly dispersed on the catalyst surface.

Figure 1

H2-TPR profile before reaction.

Figure 2 displays the Co 2p, Cu 2p, Cu LMM, and O 1s XPS spectra of the Co-Cu/TiO2 catalyst before and after reaction. As shown in Fig. 2a, the binding energies of Co 2p3 before and after are similar to each other, being located at approximately 780.4 eV. The intensity of the shakeup satellite of Co 2p3 before the reaction is obviously lower than that after the reaction. Therefore, the Co species before the reaction is mainly Co3O4 and Cu–Co spinal phase, a similar shape of the Co 2p3/2 core level spectra is also observed for mixed CuxCo3-xO4 oxides212223. After reaction, CoO is the main phase2124. This result is similar to our previous result, in which CoO is the main phase in the Co-Pd/TiO2 catalysts under 400 °C for a 2 h reduction using in-situ XPS21. Note that some CoO is reduced to metallic Co at 400 °C according to the TPR result, but the metallic Co is not detected by XPS. The reason for this observation is that a small amount of CoO is reduced at approximately 400 °C, i.e., much CoO is not reduced, according to the TPR result. Thus, the Co 2p3 peak of CoO overlaps with that of metallic Co, and the intensity of CoO is larger than that of metallic Co; as a result, the metallic Co is not detected by XPS.

Figure 2

Co 2p (a), Cu 2p (b), Cu LMM (c) and O 1s (d) XPS spectra before and after reaction.

For Cu 2p (Fig. 2b), a shakeup satellite is observed at approximately 942 eV before the reaction, and the binding energy of Cu 2p3 before the reaction is approximately 933.5 eV, which can be assigned to Cu2+ (CuO, 933.7 eV)2526. The result shows that the surface is covered by CuO before the reaction. After the reaction, the shakeup satellite disappears, indicating that the CuO is reduced. The metallic Cu and Cu2O cannot be distinguished using Cu 2p3, whereas they could be distinguished from the Cu LMM Auger spectra (Fig. 2c). As shown in Fig. 2c, the kinetic energy of Cu LMM after the reaction is approximately 917.9 eV. The kinetic energy is slightly smaller than the kinetic energy of metallic Cu (918.4 eV) but is obviously larger than that of Cu2O (916.2 eV)2526.This result indicates that the surface is covered by metallic Cu after the reaction. The kinetic energy of Cu LMM is approximately 918.3 eV, which is similar to the kinetic energy of CuO (918.1 eV)2526, further verifying the presence of CuO on the surface before the reaction.

In the case of O 1s (Fig. 2d), the peaks at 529.7 and 531.3 eV are assigned as lattice oxygen and O(H) species respectively before reaction27. After reaction, a new peak appears at 532.7 eV. The peak can be attributed to C=O or O-C-O28, because the productions adsorb on the surface. For the Ti species, the binding energies are approximately 458.5 eV before and after the reaction (Fig. S3), which can be assigned to the TiO229. The result shows that the TiO2 could not be reduced during the reaction. In general, the surface is mainly covered by CuO and Co3O4 before the reaction, whereas the surface is mainly covered by Cu and CoO after the reaction, in agreement with the H2-TPR result. In other words, the metallic Cu and CoO are the active sites for the CH4-syngas conversion.

Figure 3 shows the NH3-TPD spectra of the Cu/TiO2 catalyst. A larger peak and four main NH3 desorption peaks are detected. The first peak at approximately 118 °C is attributed to the weak acid, the second and third peaks are assigned to the mediate strong acid, and the peak at 591 °C is assigned to the strong acid. Our group has been studying the activation and conversion of CO2 and CH4 over Cu-Co catalysts supported on different solid acid supports, such as γ-Al2O3, ZrO2/SO42−, and HZSM530. The activation ability of CH4 on the Cu-Co catalyst increases with increasing acid intensity, but the too strong acid is not beneficial for the formation of active species. In other words, the appropriate acid intensity of the Co-Cu/TiO2 catalyst is favours for the conversion of CH4 and syngas.

Figure 3

NH3-TPD before reaction.

Table 1 shows the formation rate and the selectivity of the products on the Cu-Co/TiO2 catalyst. As shown in Table 1, the formation rates of CH3OH, C2H5OH and CH3COOH are 1.90, 139.37 and 0.51 mg · gcat.−1·h−1, respectively, and the corresponding selectivities of CH3OH, C2H5OH and CH3COOH are 1.34%, 98.30% and 0.36%, respectively. The result shows that the formation rate and selectivity of C2H5OH are far greater than those of CH3OH and CH3COOH, indicating that the Cu-Co/TiO2 catalyst is beneficial for the formation of C2H5OH. In addition, only C2H5OH, CH3OH, and CH3COOH are produced; these species are easily separated.

Table 1

The STY (mg·gcat.−1·h−1) and selectivity (%) of products on Cu-Co/TiO2 catalyst.

	CH3OH	C2H5OH	CH3COOH	C2H6	H2O
STY	1.90	139.37	0.51	—c	—
selectivitya	1.34	98.30	0.36	—	—
sb	11.23	88.77	0	0	0

aExperiment result.

bMicrokinetic modeling.

cNot detected by experiment.

DFT results

Ethanol synthesis from CH4 and syngas on CoCu(111) surface

The adsorption configurations of possible intermediates involved ethanol synthesis from CH4 and syngas on the CoCu(111) surface are shown in Fig. S4, and the corresponding adsorption parameters are listed in Table 2. Figures S5–S10 show the energy barriers, the reaction energies and the TS structures of ethanol synthesis from CH4 and syngas on the CoCu(111) surface.

Table 2

The adsorption energies (Eads, eV) and adsorption configurations (d, Å) of possible intermediates at their preferable adsorption sites.

Species	Eads	dCu-X (Å)a	dCo-X(Å)	Adsorption site
CH4	−0.11			—
CH3	−1.69		2.214	briCo
CH2	−4.23	2.153	2.058	fccCo
CH	−5.81	2.065	1.958	fccCo
C	−6.43		1.924	briCo
H	−2.67	1.874	1.764	fccCo
CO	−1.27		2.030	briCo
CO2	−0.53		−1.924	topCo
H2O	−0.08			—
CHO	−1.76		2.203	briCo
COH	−2.72	2.153	1.971	fccCo
O	−3.94	2.012		fccCu
CH2O	−0.26			—
CHOH	−2.21		2.083	briCo
CH3O	−1.92		2.125	briCo
CH2OH	−1.43		2.118	topCo
CH3OH	−0.51		2.467	topCo
C2H6	−0.24			—
CH3CO	−1.88		2.052	topCo
	−1.82	2.039		topCu
CH2CO	−1.67	2.256	2.124	fccCo/C(-H)-briCo,C(-O)-topCu
CHCO	−3.12	2.384	2.037	fccCo/C(-H)-briCo,C(-O)-topCu
CCO	−4.94	2.147	1.979	fccCo
CH3COH	−2.34		1.987	topCo
CH3CHO	−0.59		2.040	topCo
CH3CHOH	−1.75		2.171	topCo
C2H5OH	−0.43		2.289	topCo
CH3COO	−0.92		2.073	briCo/O-topCo, O-topCo
CH3COOH	−0.27			—

aThe nearest bond length, X stands for H, C or O.

As shown in Figs S5 and S6, *CH4 dehydrogenation are as follows: *CH4 → *CH3 → *CH2→ *CH → *C, which is in accordance with the previous studies of CH4 dehydrogenation on different metals and alloys using DFT313233343536373839. Table S1 shows the adsorption parameters of *CH4, *CH3, *CH2, *CH and *C on a Cu site. Comparing Table 2 with Table S1, the binding strengths of *CH3, *CH2, *CH and *C on a Co site (−1.69, −4.23, −5.81 and −6.43 eV) are found to be obviously larger than those on a Cu site (−1.23, −3.82, −5.21 and −5.46 eV); the binding strength of *H on a Co site (−2.67 eV) is slightly larger than that on a Cu site (−2.45 eV); and the binding strength of *CH4 on a Co site (−0.11 eV) is similar to that on a Cu site (−0.10 eV). The observed trend is in agreement with the result of Liu et al.37. We also studied *CH3 and *CH2 formation on fccCu (Fig. S7); the energy barriers are 1.92 and 1.18 eV, which are far larger than that on a Co site. The result indicates that CH4 dehydrogenation prefers to occur on Co sites versus Cu sites.

As shown in Fig. S8, *CHO and *CH2O formation are likely from *CO hydrogenation. Again, *CH2O hydrogenation is superior to dissociation. Because the energy barrier of *CH3O formation (0.89 eV) is similar to that of *CH2OH formation (0.82 eV), *CH2OH and *CH3O further reactions are considered. In the case of *CH2OH further reaction, *CH2 formation occurs slightly easier than *CH3OH formation. Similarly, *CH3O prefers to be dissociated into *CH3 and *O. A previous study showed that the energy barrier of *CH3O+ *H → *CH3OH + * (0.76 eV) is lower than that of *CH3O+ * → *CH3 + *O (1.05 eV) on a Rh(111) surface40. The energy barrier of *CH3O+ *H → *CH3OH +* (1.07 eV) is obviously smaller than that of *CH3O+ * → *CH3 + *O (2.22 eV) on a Cu(211) surface; however, the energy barrier of *CH3O+ *H → *CH3OH +* (1.41 eV) is slightly lower than that of *CH3O+ * → *CH3 + *O (1.67 eV) on a Rh doped Cu(211) surface. Thus, Zhang et al. considered that C-O scission is difficult to perform on Cu-based catalysts and the promoter Rh facilitates *CH3 formation. The results show that the promoter Rh increases the productivity and selectivity of ethanol synthesis from syngas on Cu-based catalysts41. The energy barrier and reaction energy of *CH3OH+ * → *CH3 + *OH are 0.81 and 0.10 eV, and the energy barrier of *CH3OH+ * → *CH3 + *OH is higher than that of the desorption energy of CH3OH (0.51 eV). This result shows that CH3OH desorption occurs on the surface.

According to the above results, *CH3, *CH2, *CH and *C are the possible intermediates during the process of *CH4 dehydrogenation, and *CH3 and *CH2 are the possible intermediates from C-O scission during the process of methanol from syngas. Therefore, *CH3, *CH2, *CH and *C reactions with *CO are considered in this section. As shown in Fig. S9, the energy barriers of *CH3CO, *CH2CO, *CHCO and *CCO are in the following order: *CH3CO (0.49 eV) > *CH2CO (1.55 eV) > *CHCO (1.71 eV) > *CCO (2.07 eV). The result shows that the insertion ability of *CO decreases with decreasing H number of *CHx(x = 0, 1, 2, 3). Finally, *C2H5OH is synthesized through *CH3COH and *CH3CHOH from *CH3CO further hydrogenation (Fig. S10).

It is well known that the Cu-based catalysts also have applications in the water-gas shift (WGS) reaction or the reverse WGS reaction424344. Because of the complication of both reactions, we only consider *CO2 and *H2O formation. The energy barriers of *CO + *O → *CO2 + * and *OH + *H → *H2O + * are 0.81 and 1.43 eV, respectively (Fig. S10). The energy barrier of *CH3+ *CO2 → *CH3COO + * are 1.13 eV, which is higher than those of *CH3CO (0.49 eV) and *C2H6 (0.89 eV) formation, indicating that CH3COO formation is difficult. There is only one product of *CH3COO hydrogenation, *CH3COOH, for which the energy barrier and reaction energy are 0.93 and 0.08 eV, respectively.

Ethanol synthesis from CH4 and syngas on Cu(111) and Co(111) surfaces

According to the above result, there are two key factors for ethanol synthesis: one is *CHx formation; the other is C-C bond formation. For *CHx formation, there are two methods: one is from CH4 decomposition; the other is from C-O bond scission during the process of methanol synthesis. Liu et al. studied CH4 decomposition on Co(111) and Cu(111) surfaces37. They found that the energy barrier of *CH3 formation from *CH4 dehydrogenation on a Cu(111) surface is obviously higher than that on a Co(111) surface (1.14 vs. 1.88 eV) using the same calculation parameters. The results show that CH4 decomposition preferably occurs on a Co site.

Regarding *CHx formation during the process of methanol synthesis from syngas, our previous results showed that the energy barriers of *CH3OH and *CH3 formation from *CH3O, *CH3O and *CH2 formation from *CH2O, and *CH2O and *CH formation from *CHO on a Cu(111) surface are 0.63 and 2.18 eV, 1.00 and 2.05 eV, and 0.93 and 2.05 eV, respectively45. Zhang et al. and Mehmood et al. studied ethanol synthesis from a Cu(211) surface and methanol decomposition on Cu4 nanoparticles; they also found that the ability of hydrogenation is greater than that of C-O scission4146. The result indicates that *CHx is not formed on a Cu(hkl) surface during methanol synthesis. In the case of a Co surface, Mehmood et al. proposed that the energy barriers of *CH3OH and *CH3 formation from *CH3O, *CH3O and *CH2 formation from *CH2O, and *CH2O and *CH formation from *CHO on Co4 nanoparticle are 1.48 and 1.66 eV, 0.86 and 1.11 eV, and 1.43 and 2.13 eV, respectively46. The energy barriers of *CH3OH and *CH3 formation from *CH3O and *CH3O and *CH2 formation from *CH2O are similar to each other, but the energy barrier of *CH formation is higher than that of *CH2O formation. The result shows that *CH2 and *CH3 species formation are likely on a Co surface.

According to the above results, it was found that the formation of *CH, *CH2 and *CH3 during the process of *CH4 dehydrogenation and *CH3OH formation on a single Co active site are possible; however, it is impossible on a single Cu active site. Therefore, we only consider C-C formation on the Co(111) surface. The energy barriers, reaction energies and TSs are shown in Fig. S11.

Fig. S11 shows that the energy barriers of *C2H6, *CH3CO, *CH2CO and *CHCO formation are 0.76, 1.04, 1.40 and 2.02 eV, respectively. Comparing the energy barriers of C-C formation on the CoCu(111) surface, it was found that the energy barriers of *C2H6 (0.76 vs. 0.89 eV) and *CH2CO (1.40 vs. 1.55 eV) formation on the Co(111) surface are slightly lower than those on the CoCu(111) surface, whereas the energy barriers of *CH3CO (1.04 vs. 0.49 eV) and *CHCO (2.02 vs. 1.71 eV) formation on the Co(111) surface are higher than those on the CoCu(111) surface. The result also shows that Co-Cu based catalysts change the reaction path. In addition, the energy barrier of *C2H6 formation is lower than those of *CH3CO, *CH2CO and *CHCO formation. The result shows that *C2H6 formation is preferable, for which hydrocarbon formation is preferred versus C2 oxygenate. The result is in agreement with the experiment results in which a Co-based catalyst is one of catalysts for the F-T reaction474849. Therefore, ethanol synthesis from CH4 and syngas requires two active sites: CoO and metallic Cu. In addition, because ethanol synthesis from CH4 and syngas requires a synergistic effect between metallic Cu and CoO, small particles of CoO and metallic Cu are required.

Microkinetic modelling

To date, most possible reactions during the reaction of CH4 and syngas have been studied using DFT. Table 3 summarises the optimal reaction pathways for ethanol synthesis on the CoCu(111) surface together with the corresponding activation barriers. In this section, to estimate the results from DFT, the selectivity of the possible products involved in ethanol synthesis from CH4 and syngas under our experimental condition was studied using a microkinetic model50. Similar kinetic modelling has been successfully applied for various reactions405152. As shown in Table 3, the adsorption processes (R1, R2 and R3) are assumed to be in equilibrium. The other surface species involved in the R4-R22 reaction can be described according to the pseudo-steady-state approximation50. The relative selectivity (s) values are defined as si = ri/i, where r is the relative rate for each product, and i denotes CH3OH, C2H5OH, C2H6, CH3COOH and H2O. The detailed description of the microkinetic model is shown in the supplement.

Table 3

The optimal reaction pathways for ethanol synthesis on CoCu(111) surface together with the corresponding activation barriers(Ea, eV).

No.	Elementary reactions	Ea	No.	Elementary reactions	Ea
1	CH4(g) + * → CH4*		12	CH2* + H* → CH3* + *	0.61
2	CO(g) + * → CO*		13	CH4* + * → CH3* + H*	1.28
3	H2(g) + 2* → 2H*		14	CH3* + CO* → CH3CO* + H*	0.49
4	CO* + H* → CHO* + *	1.09	15	CH3CO* + H*→ CH3COH * + *	0.86
5	CHO* + H* → CH2O* + *	0.72	16	CH3COH* + H*→ CH3CHOH* + *	0.62
6	CH2O* + H* → CH3O* + *	0.89	17	CH3CHOH* + H*→ C2H5OH(g) + *	0.28
7	CH2O* + H* → CH2OH* + *	0.82	18	CO* + O* → CO2* + *	0.81
8	CH3O* + * → CH3* + O*	0.62	19	CH3* + CO2* → CH3COO* + *	1.13
9	CH3O* + H* → CH3OH(g) + 2*	1.17	20	CH3COO* + H* → CH3COOH* + *	0.93
10	CH2OH* + * → CH2* + OH *	1.09	21	H* + OH* → H2O (g) + *	1.43
11	CH2OH* + H* → CH3OH(g) + 2*	1.38	22	CH3* + CH3* → C2H6(g) + *	0.89

According to our DFT results and the microkinetic model, the relative selectivity of CH3OH, C2H5OH, C2H6, CH3COOH and H2O are determined under our experimental conditions (PCH4 = 0.95 atm, PCO = 0.5 atm, =0.5 atm and T = 300 °C). As shown in Table 1, the relative selectivities of CH3OH and C2H5OH are 11.23 and 88.77%; the relative selectivities of C2H6, CH3COOH and H2O are very small and can be ignored. Compared with the experiment result, it is found that the selectivity of CH3OH using the microkinetic model is higher than that the experiment result, whereas the selectivities of C2H5OH and CH3COOH using the microkinetic model are lower than the experiment results. The differences in selectivity between our theoretical and experimental results could be caused by many effects. The first possible reason is that the Cu-Co alloy is not formed during 400 °C calcinations5354, and the experiment results show that the Cu-CoO interface is the best model. The second possible explanation for the selectivity differences between our theoretical and experimental results is the presence of defect sites. To our best knowledge, defects can have a major role in catalysis by affecting the energy barriers55565758. Nonetheless, ethanol synthesis from CH4 and syngas on CoCu(111) provides useful insights into the experiment to a certain degree. In the future, we plan to investigate the Cu-CoO interface and defects for ethanol synthesis from CH4 and syngas.

Conclusions

In the paper, ethanol synthesis from CH4 and syngas on a CoCu-based catalyst was studied using experiments, DFT and microkinetic modelling. The experimental results indicated that ethanol can be synthesised at high efficiency from CH4 and syngas on the Cu-Co/TiO2 catalyst, over which the selectivity of ethanol is approximately 98.30%. It was found that the active sites of ethanol synthesis are metallic Cu and CoO, with metallic Cu and CoO uniformly dispersed on the catalyst surface.

Most possible ethanol formation pathways from methanol and syngas were systematically investigated on CoCu(111) surface. The DFT result showed that ethanol synthesis from CH4 and syngas requires the synergistic effect between metallic Cu and CoO, and ethanol is not synthesised on single metallic Cu and CoO. On the CoCu(111) surface, *CH3 is the primary CHx species. *CH3 forms via three pathways: the first is *CH4 dehydrogenation, the second is C-O scission of *CH3O, and the third is CH2 hydrogenation from C-O scission of *CH2OH. Subsequently, *CH3CO forms from the *CO and *CH3 reaction. Finally, ethanol is synthesised through *CH3COH hydrogenation. The microkinetic modelling result showed that there is only CH3OH and C2H5OH, for which the selectivity of ethanol is lower than that of the experiment result. We think that the difference between the theoretical and experimental results could be mainly caused by issues with the model and the presence of defect sites. Future work will focus on the Cu-CoO interface and defects for ethanol synthesis from CH4 and syngas.

Experimental and theoretical methods

Catalyst preparation

The preparation of the support TiO2 was as follows: 24 g of NaOH was introduced into 60 mL of distilled water (10 M NaOH solution), and then, 1.0 g of commercial TiO2 powder (P25, Degussa) was dispersed into the 10 M NaOH solution with continuous stirring for 2 h. The mixture was transferred into a Teflon-lined autoclave, and then, the mixture was heated to 150 °C for 24 h under sealed conditions. Subsequently, the mixture was allowed to cool to room temperature. The powder was washed using distilled water until the pH of the powder was approximately 7. The neutral powder was washed using 0.1 mol/L HNO3 and then washed again using distilled water until the pH of the powder was approximately 7. After drying for 10 h at 75 °C, the obtained precipitate was calcined in air at 400 °C for 10 h, and the heating rate was 1 °C /min. Finally, the support TiO2 was obtained5960.

The preparation method of the Cu-Co/TiO2 catalyst was the equal volume impregnation method. TiO2, Co(NO3)2·6H2O and Cu(NO3)2·3H2O were dissolved into ethylene glycol solution. After stirring for 12 h, the resulting slurry was dried for 12 h at 150 °C. Subsequently, the catalyst was calcined in air at 400 °C for 4 h at the heating rate of 2 °C/min. Finally, the Cu-Co/TiO2 catalyst was obtained21. The Cu and Co loading on TiO2 were 12 and 6 wt.%.

Catalyst characterization

XPS was performed using a V.G. Scientific ESCALAB250 with focused monochromated Al Kα. The residual pressure inside the analysis chamber was set to <2.0 × 10−9 mbar. For H2 temperature-programmed reduction (TPR) experiment, 50 mg catalyst was loaded into a fixed-bed reactor. The heating rate was 10 °C/min until the temperature is 600 °C using a temperature controller. The reduction gas was H2 and N2 which the ratio was 5:95 with a flow rate of 30 mL/min. NH3-TPD experiment was used on a TP-5000 instrument. 100 mg catalyst adsorbed NH3 at 50 °C until saturation, then purged the physisorbed NH3 using He for 30 min. Finally, the NH3-TPD data were collected in flow He from 50 to 800 °C which the heating rate was 10 °C/min.

Catalytic activity test

The diagram of the reaction apparatus is shown in Fig. 4 and was the same as the reaction apparatus of our previous paper on acetic acid synthesis from CH4 and CO212. There was 1.5 g of catalysts used in reactor A and B, respectively. Before the reaction, the catalyst in both reactor A and B was reduced with 30 vol % H2 and 70 vol % N2 at 400 °C for 2 h. Because H2 was found to inhibit the excessive dehydrogenation of methane during CH4 activation, CH4 and H2 were injected together12. The reaction was carried out at 300 °C at atmospheric pressure. The test procedure is as follows: first, 50 ml/min of CH4 and 5 ml/min of H2 were injected into reaction A; at the same time, syngas (50 ml/min of CO and 50 ml/min of H2) were injected into reaction B. After 300 s, the electromagnetic valve was changed over. Then, syngas (50 ml/min of CO and 50 ml/min of H2) were injected into reaction A, and at the same, 50 ml/min of CH4 and 5 ml/min of H2 were injected into reaction B. Subsequently, the cycle was repeated until the reaction was finished, and ethanol was produced from CH4 and syngas.

Figure 4

Schematic diagram of the experimental apparatus.

The obtained products from the reaction were analysed using a gas chromatograph (GC-950) equipped with a hydrogen flame detector and packed column. The contents of each component were studied using the external standard method. On-line gas phase analysis and off-line analysis of the liquid products were performed, with the off-line analysis assisting in the product identification. The detailed analysis procedure used is as follows. The products were cooled by a condensator. Next, the liquid products were obtained from the condensator every hour and were injected into the GC. The gas phase was collected every 70 s; the gas was not cooled and condensed by the condensator. Only the oxy-organics in the gas phase were considered. The space time yield (STY, S), the number of total moles of the hydrocarbon (n), and the selectivity of carbon atoms (x) were defined as S = K × A × V/m, n = and x = S × N/(n × M), where K, A, V, m, N and M are a constant using the external standard method, the area of products (i) using chromatography, gas flow, mass of catalyst, carbon number of the products and molar mass of products, respectively.

Computational methods

The geometries and transition state (TS) were calculated using the Dmol3 Materials Studio software6162. The calculation parameters were the same as those in our previous studies4563. The electronic structures were obtained by solving the Kohn−Sham equation self-consistently under spin-unrestricted conditions6465. DFT was also used for the core electrons by applying the PW91 generalised-gradient approximation to the exchange-correlation energy66. A double numeric quality basis set with polarisation functions was used. A self-consistent field procedure is carried out with a convergence criterion of 10−5 a.u. on energy and electron density, and the geometry is optimized under a symmetry constraint, with the convergence criteria of 10−3 a.u. on the gradient and 10−3 a.u. on the displacement. The TS was identified using the complete linear/quadratic synchronous transit method67.

The Cu(111) surface was cleaved from the face-centred cubic (fcc) crystal structure after optimisation; the theoretical equilibrium lattice constant of Cu was aCu = 3.685 Å, compared with the experimental value of aCu = 3.604 Å68. The surface was modelled using a six-layered mode p(3 × 3) super cell with nine atoms in each layer along with a 15 Å vacuum slab. The mass ratio of Cu:Co was 2 in the experiment, for which the molar ratio was approximately 1.8, and the molar ratio of Cu:Co of the CoCu(111) surface was 2 to simplify the model building. Next, three Cu atoms were replaced by Co atoms in each of the layers. The structure of the CoCu(111) surface after optimisation is shown in Fig. 5. During the calculation process, the bottom two layers were fixed, and other layers and adsorbates were allowed to relax. Meshes of 3 × 3 × 1 k-points were used for the CoCu(111) and Co(111) surfaces.

Figure 5

Side view (left) and top (right) view of the CoCu(111) surface after optimization.

Additional Information

How to cite this article: Zuo, Z.-J. et al. Efficient Synthesis of Ethanol from CH4 and Syngas on a Cu-Co/TiO2 Catalyst Using a Stepwise Reactor. Sci. Rep. 6, 34670; doi: 10.1038/srep34670 (2016).