Theoretical Studies on the CO2 Reduction to CH3OH on Cu(211)
- 744 Downloads
CO2 reduction has been pursued for decades as an effective way to produce useful fuels and to mitigate global warming at the same time. Methanol synthesis from CO2 hydrogenation over Cu-based catalysts plays an important role in the chemical and energy industries. However, fundamental questions regarding the reaction mechanism and key reaction intermediates of this process are still unclear. To address these issues, we studied the CO2 hydrogenation process using density functional theory (DFT) combined with van der Waals (vdW) force corrections, finding that the most effective pathway proceeds along the reaction series CO* → CHO* → CH2O* → CH2OH* → CH3OH* with the reactive intermediate CH2O*, which is consistent with experimental findings. Additionally, we find that water molecules play an inhibiting role in the reaction, while H bonds and vdW forces have an essential effect on the reaction mechanisms. These findings shed light on the reaction mechanism of CH3OH formation from CO2 hydrogenation and reveal the essence of H2O in this reaction, providing a useful basis for preceding studies.
KeywordsDensity functional theory CO2 reduction Electrocatalysis Copper
With a growing energy demand and fast depletion of fossil fuels, it is urgent to find new energy sources. One potential approach is to fix atmospheric CO2 into value-added products using renewable power sources without additional CO2 generation, such as solar source, wind, and tidal energy [1, 2]. Formation of oxygenated compounds, such as methanol, ethanol, and formic acid from CO2 reduction, has been intensively investigated in view of their added value . Especially the formation of methanol has drawn much attention because of its minimum hydrogen requirement and its large demand as a bulk chemical [2, 4, 5, 6, 7].
There have been extensive investigations on copper-based catalysts, and it is well accepted that copper is the active center [15, 16, 21, 22, 23, 24]. However, recently, some studies have indicated that the Zn/ZnO substrate has an essential effect on the reaction, which can improve the performance of catalysts to some extent [5, 25, 26, 27, 28, 29]. Although the effective sites have been studied thoroughly, there is still some debate on their exact nature. As the focus of our studies is on the catalytic reaction as well as the influence of co-adsorbated water on the mechanism of methanol formation, here we concentrate on the pure Cu metal. It was found that the activity of Cu(211) for CO2 as well as for CO hydrogenation is more advanced than on Cu(111) [25, 30]. Moreover, carbon-labeling experiments have shown that CO2 is the major reactant under industrial conditions [12, 20, 31, 32]. In addition, 13C-labeling experiments indicated that methanol was generated exclusively from the adsorbed CH2O* intermediate and not from the adsorbed surface formate .
Great efforts have been spent to explore the exact reaction mechanisms of CO2 hydrogenation [24, 34, 35, 36, 37, 38, 39], but yet no complete consensus of the pathways has been achieved. Traditionally, two possible reaction mechanisms have been proposed. One is the direct hydrogenation of CO2 into HCOO (formate) [24, 25, 38]. Yang et al. have shown that it would be hydrogenated into H2COO (dioxomethylene), followed by the formation of H2CO (formaldehyde) . In contrast, Studt et al. proposed that HCOO would first form HCOOH, followed by the formation of H2COOH. Then, the H2COOH is converted to H2CO and OH . Along the alternative reaction mechanism, CO2 first reacts to CO through an OCOH intermediate, and the following CO hydrogenation can proceed through CHO or COH intermediates [37, 38, 39, 40, 41, 42]. In case CHO is formed, CH2O will also be present as an intermediate. In most studies, it is stated that once CH2O is formed, the next protonation product is CH3O [25, 38, 43, 44, 45]. In fact, this step can again follow different reaction pathways, e.g., the formation of CH2OH with an O–H bond formed. Ye et al. have studied the formation of CH2OH from CH2O on In2O3(110), but found it energetically less favorable than the formation of CH3O . Although many studies have already addressed the reaction mechanism, there still exists a controversy on the exact pathways, which has been the main motivation of the present work.
CO2 reduction experiments demonstrated that under variable reaction conditions, water does play different roles. With Cu as an effective catalyst, the main products for the electrochemical reduction of CO2 in an aqueous electrolyte are CH4, C2H4, and HCOOH, while under gaseous conditions, CH3OH will form as the main product [5, 25, 46]. Montoya et al. found that without the charged water layer at the metal surface, the formation of CO dimers is prohibitively endergonic . In other words, water in the electrochemical interface may accelerate the reaction to some extent. In contrast, it was found that water molecules that are formed during the reaction might suppress the formation of CH3OH from syngas [17, 32]. Here, one explanation is that these water molecules block active copper sites and thus inhibit the adsorption of CO2 for the follow-up reactions . On the other hand, it was argued that these water molecules accelerate the crystallization of Cu and ZnO, resulting in the catalyst’s deactivation . Notably, the latter explanation implies that crystallization of Cu and ZnO leads to a decrease in surface area, being consistent with the former explanation that water molecules can block active copper sites. Further, Zhao et al. have studied the effect of H2O on this reaction, finding that H2O plays an inhibiting role in the hydrogenation of HCOO into methanol in the formate pathway . However, in their studies, only a single water molecule was introduced, which is insufficient to mimic the catalyst surface under aqueous conditions. Thus, further studies are required to understand the effect of co-adsorbed water on the reaction process.
All theoretical data were obtained using the CASTEP code with ultrasoft pseudopotentials and the generalized gradient approximation (GGA) exchange-correlation functional proposed by Perdew, Burke, and Ernzerhof (PBE) combined with the TSsurf dispersion corrections [49, 50, 51, 52]. This is motivated by the fact that while PBE alone is known to overestimate binding energies for adsorption processes on transition-metal surfaces [52, 53, 54], it had been demonstrated that the inclusion of entropy and thermochemical contributions as well as van der Waals (vdW) forces yields an improved description of the CO2 hydrogenation [40, 43, 55]. We adopted a plane wave basis set with an energy cutoff of 400 eV, and used a 2 × 2 × 1 Monkhorst-Pack k-point scheme for the Brillouin zone sampling. All calculations were performed on the Cu(211) surface orientation with five layer slabs, where the bottom three layers were fixed at the bulk-truncated structure, leaving the top two surface layers and the adsorbents as well as co-adsorbed water to be fully relaxed. The periodic images of the slab were separated by a vacuum gap being at least 15 Å wide. To investigate the reaction pathways, the transition state (TS) search procedure as implemented in the CASTEP code was used, which employs a combination of linear and quadratic synchronous transit (LST/QST) algorithms with subsequent conjugate gradient methods .
Results and Discussion
Cu(211) With and Without H2O Chain
Adsorption energies, E ad, in electron volts, of different molecules on bare Cu(211) and Cu(211) with a H2O chain using the PBE and PBE + TSsurf approximations
E ad (eV)
PBE + TSsurf
Cu(211) with H2O chain
PBE + TSsurf
Overall, vdW interactions seem to have a significant effect on the adsorption energies of small molecules, especially when being only weakly adsorbed. This should additionally have an impact on reaction mechanisms, an aspect that will be elucidated in the next sections.
CHO*, CH2O*, and CH3O* Pathways
For the CHO* pathway, CO* hydrogenation proceeds via the hydrogenation of the carbon atom of CO*, following the stepwise hydrogenation with the formation of CHO*, CHOH*, and CH2OH* intermediates. Lastly, CH2OH* is hydrogenated into CH3OH. The formation of CHO* from CO* is an endothermic process (reaction energy E r = 0.70 eV) with an energy barrier of E a = 1.05 eV, which is consistent with the results of Asthagiri et al. (E r = 0.71 eV and E a = 1.03 eV) . The hydroxyl formation (CHOH*) is the determining step for this pathway, which is endothermic by 0.18 eV with a barrier of 1.38 eV. Then, CHOH* can be further hydrogenated into CH2OH* with a relatively small barrier of 0.61 eV and an overall energy gain of E r = −0.71 eV. The final hydrogenation step, the formation of CH3OH, releases 1.09 eV and has a barrier of 0.71 eV, which is higher than the barrier reported by Mavrikakis et al. (E a = 0.51 eV) .
The difference between the CH2O* and the CH3O* pathways is that the stepwise hydrogenation of CH2O* to CH3OH* is through CH3O* (CH3O* pathway) or its isomer CH2OH* (CH2O* pathway). In both mechanisms, the formation of CH3OH* is from CH2O* where one C–H and one O–H bonds need to be formed. Along the CH2O* pathway, the O–H bond is formed first, followed by the formation a C–H bond, while the order is the opposite for the CH3O* pathway. Here, the process of CH2O* hydrogenation into CH3O* is very facile with a barrier of only 0.38 eV and an exothermicity of 0.97 eV, which coincides well with the results of Li et al. (E a = 0.35 eV and E r = −0.92 eV) . Compared to the process of CH2OH* formation (E a = 0.95 eV and E r = −0.35 eV), the formation of CH3O* from CH2O* seems more favorable. However, the CH3OH* formation from CH3O* (being the determining step) is kinetically prohibitive, which must overcome a barrier of 1.32 eV, which is 0.15 eV higher than the value obtained by Mavrikakis et al. . Thus, considering all the elementary steps, the CH2O* pathway seems more promising than the CH3O* pathway.
In the three considered reaction sequences, the most promising one is the CH2O* pathway, of which the determining step is the formation of CHO*. The process of this pathway proceeds as CO* → CHO* → CH2O* → CH2OH* → CH3OH* and the reactive intermediate is CH2O*. The isotope labeling study for methanol synthesis on Cu/ZrO2 suggested that CH3OH is exclusively formed from CH2O* (not HCOO*), which agrees well with our results .
As stated above, in the present work, we concentrated on the role of water on the reaction mechanism and did not explicitly take zero-point energy and entropy corrections into account. Although this approach seems to be sufficient for such a comparative study, our absolute reaction energies (∆E) for the overall CH3OH formation are all about −2.0 eV, which is higher than the experimental value . This difference is mainly associated with the separated (not adsorbed) molecules as the following estimation shows. While the zero point energy contributions are 0.14 eV for CO and 1.35 eV for CH3OH , taking these contributions into account, we would obtain the overall reaction free energy to be 0.80 eV, which is relatively close to the experimental value of 0.95 eV . The remaining differences might certainly be attributed to the omitting of ∆pV contributions to the enthalpy.
Influence of a Co-adsorbed H2O Chain
To verify the effect of co-adsorbed water, we have calculated the same reaction pathways on Cu(211) in the presence of a co-adsorbed H2O chain (Fig. S2 and Table S3). The results reveal that due to water the barriers for CHO* and CH3OH* formation along the CHO* pathway increase by 0.08 and 0.14 eV, respectively. For the CH2O* pathway, the reaction barrier of each step is also increased by up to 0.34 eV (for the CH2OH* formation). Along the CH3O* pathway, the barriers of the former two steps increase by the same values as in the CH2O* pathway. However, the CH3O* and CH3OH* formation steps now have barriers that are increased by 0.41 and 0.03 eV in the presence of water. Meanwhile, the reaction CH3O* + H* → CH3OH* qualitatively changes from an endothermic process (E r = 0.10 eV) to an exothermic process (E r = −0.18 eV), but the larger barrier (1.35 eV) indicates that this step is prohibited dynamically. All the results indicate that the co-adsorbed H2O chain apparently suppresses this reaction, which is consistent with the experimental results that water molecules have an inhibiting effect [17, 32, 48].
Now, we discuss how water influences the overall reaction mechanisms. The energy diagram of Fig. S2 shows that with water, the CH2O* pathway is still the most effective reaction mechanism. Without water, the CO* molecule adsorbs at a step site of the (211) surface vertically through its C atom (Fig. 3 1a). In the presence of H2O, the CO moves to the (111) terrace (Fig. 3 1a) as the H2O chain occupies and blocks the step. Without water, CHO* adsorbed through their C and O atoms “lying” on the step, while the hydrogen atom pointed into the vacuum. Under the influence of the H2O chain, the adsorption configuration of CHO* did not change but was pushed to the terrace. CH2O* and CH2OH* intermediates are also shifted to the (111) terrace in the presence of water while again their configurations (all the atoms parallel to the surface) did not alter. Due to the presence of the H2O chain in the system, the adsorption sites of the intermediates have been pushed from the active step to the less active terrace. In addition, there are interactions between the intermediates (CH2O*, CH2OH*) with the co-adsorbed H2O molecules through H bonds, which in addition affect the molecular orientations. On one hand, these interactions lead to an increase of the adsorption energies of the intermediates and, on the other hand, result in increased activation barriers and thus lower activities. Therefore, our studies clearly show that the presence of water will certainly influence the adsorption sites and configurations of the intermediates.
Role of the vdW Forces
To explain the influence of vdW interactions on the reaction mechanisms, we analyzed the reaction barriers and energies of main steps for the CHO*/CH2O*/CH3O* pathways (Table S4) both on bare Cu(211) and Cu(211) with the H2O chain. For the CHO* pathway on bare Cu(211), the contribution of vdW forces to the reaction energy is up to 0.11 eV (18.6%; the percentage is the ratio of the vdW contribution to the overall energies) for the CO* + H* → CHO* step, where the contribution to the barrier is 0.11 eV (11.7%). The corresponding contribution in the CH2O* mechanism increases to 20.7% for the CH2OH* formation process. For the CH3O*, formation step along the CH3O* pathway, vdW contributions are even 26.7% on Cu(211). Under the influence of water, the vdW contributions further increase to 20, 30.8, and 52.6% for CHO*, CH2O*, and CH3O* pathways, respectively. It can be seen that due to the presence of water in the system, the contribution of the vdW forces has been increased significantly.
Overall, we find a preference for the CH2O* pathway, a result that differs from the pathway suggested by Nørskov et al. , who found a preference for a CO* → CHO* → CH2O* → CH3O* → CH3OH* sequence, which corresponds to our CH3O* pathway. However, in their work, the BEEF-vdW approach was employed to describe the vdW force, while the TSsurf method has been used in our work. The previous studies have found that the TSsurf method is applicable to metal surfaces where more precise results can be obtained compared to the other vdW schemes . The significant contribution of vdW interactions to the reaction pathways can effectively change the selectivity of different pathways.
COH* and HCOO* Pathways
As stated before, CH3OH* can also be formed directly from CO2*. Here, two different pathways have been proposed; one proceeds via the sequence CO2* → HCOO* → H2COO* → H2CO* and the other CO2* → HCOO* → HCOOH* → H2CO* [24, 25, 36, 38, 40]. The remaining steps to CH3OH* are equivalent to the mechanism discussed previously for the CO* hydrogenation. As for both proposed pathways the first step is the formation of HCOO* from CO2*, we calculated this process on Cu(211) with and without water. On bare Cu(211), this step is exothermic by −0.83 eV but requires overcoming a high barrier of 1.64 eV. In the presence of water, the barrier further increases even up to 2.98 eV and the exothermicity decreases to −0.33 eV. Although thermodynamically slightly preferred, again it seems kinetically hindered due to this high barrier. Therefore, it seems appropriate to disregard the direct CO2* pathways for the present reaction process.
In conclusion, we have performed detailed studies on the formation of CH3OH* on Cu(211), summarized in Scheme 1, and found that the most possible pathway proceeds via the reaction sequence CO* → CHO* → CH2O* → CH2OH* → CH3OH*, in which the active intermediate of CH2O* is in accord with experiments. We also found that the presence of co-adsorbed water at the preferred edge sites (represented as a H2O chain) has an inhibiting effect on the CO2 hydrogenation, which is again consistent with the experimental results. This suppression mainly stems from the effect of H bonds and vdW interactions. The H bonds induced by the H2O chain change the orientation of the intermediates and increase their stabilities as well as the reaction barriers. The vdW contributions also lead to an increase of the adsorption energies of intermediates and the correlated energy barriers for further reaction steps. Our results provide a comprehensive understanding of the mechanism and conceptually explain the experimental phenomenon (the inhibiting effect of H2O molecules).
The authors acknowledge support from the Deutsche Forschungsgemeinschaft (DFG) through the grant proposal (JA1072/9-1 and 9-2), which was part of the research unit DFG-FOR1376. Further, support by the Program for Thousand Young Talents Plan and the National Natural Science Foundation of China (No. 21673095, 51631004), the computing resources of High Performance Computing Center of Jilin University, and National Supercomputing Center in Jinan and in Shenzhen China are acknowledged. Finally, the authors also acknowledge the computer time supported by the state of Baden-Württemberg through the bwHPC project and the DFG through grant number INST40/467-1 FUGG.
- 25.M. Behrens, F. Studt, I. Kasatkin, S. Kühl, M. Hävecker, F. Abild-Pedersen, S. Zander, F. Girgsdies, P. Kurr, B.-L. Kniep, M. Tovar, R.W. Fischer, J.K. Nørskov, R. Schlögl, The active site of methanol synthesis over Cu/ZnO/Al2O3 industrial catalysts. Science 336, 893–897 (2012)CrossRefGoogle Scholar
- 27.S.A. Kondrat, P.J. Smith, P.P. Wells, P.A. Chater, J.H. Carter, D.J. Morgan, E.M. Fiordaliso, J.B. Wagner, T.E. Davies, L. Lu, J.K. Bartley, S.H. Taylor, M.S. Spencer, C.J. Kiely, G.J. Kelly, C.W. Park, M.J. Rosseinsky, G.J. Hutchings, Stable amorphous georgeite as a precursor to a high-activity catalyst. Nature 531, 83–87 (2016)CrossRefGoogle Scholar
- 60.J.B. Hansen, P.E. Højlund Nielsen, Methanol synthesis, in: handbook of heterogeneous catalysis, (Wiley-VCH Verlag GmbH & Co. KGaA, 2008)Google Scholar