Computational Design of Single Molybdenum Catalysts for Nitrogen

Subscriber access provided by Iowa State University | Library

C: Surfaces, Interfaces, Porous Materials, and Catalysis

Computational Design of Single Molybdenum Catalysts for Nitrogen Reduction Reaction Qinye Li, Siyao Qiu, Chuangwei Liu, Ming-Guo Liu, Lizhong He, Xiwang Zhang, and Chenghua Sun J. Phys. Chem. C, Just Accepted Manuscript • DOI: 10.1021/acs.jpcc.8b11509 • Publication Date (Web): 04 Jan 2019 Downloaded from http://pubs.acs.org on January 8, 2019

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

The Journal of Physical Chemistry

Computational Design of Single Molybdenum Catalysts for Nitrogen Reduction Reaction Qinye Li,a Siyao Qiu,b Chuangwei Liu,c Mingguo Liu,d Lizhong He,a Xiwang Zhang,a,* Chenghua Sunb,e,* a

b

School of Chemical Engineering, Monash University, Clayton, VIC 3800, Australia Science & Technology Innovation Institute, Faculty of Science, Dongguan University of

Technology, Dongguan 523808, China c

School of Chemistry, Monash University, Clayton, VIC 3800, Australia

d

Hubei Key Laboratory of Natural products Research and Development, College of Biological

and Pharmaceutical Sciences, China Three Gorges University, Yichang 443002, China e

Department of Chemistry and Biotechnology, Faculty of Science, Engineering & Technology,

Swinburne University of Technology, Hawthorn, VIC 3122, Australia

*[email protected] (X.Z.) *[email protected] (C.S.)

ACS Paragon Plus Environment

1

The Journal of Physical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 18

ABSTRACT

Starting from molybdenum (Mo) embedded in black phosphorus (BP), 17 single Mo catalysts with various combinations of ligands, including phosphorous (P), boron (B), nitrogen (N), sulphur (S) and carbon (C), have been computationally examined as catalysts for nitrogen reduction reaction. Among them, Mo-PC2, Mo-PB2 and Mo-BC2 have been identified as the most promising catalysts, offering an overall overpotential less than 0.60 V. Mo-BC2 is particularly attractive as it also shows a high NRR selectivity over hydrogen evolution reaction. Such high performance essentially is originated from the mediation of the ligands, which effectively shift the d-band center of Mo-atom towards the Fermi energy.


2

Page 3 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction Molybdenum (Mo) offers a strong capacity to adsorb nitrogen (N2), but Mo-N2 interaction is too strong for full nitrogen reduction reaction (NRR),1,2 a key step for ammonia (NH3) synthesis.3-5 Therefore, additional mediation is needed. Schrock and co-workers demonstrated N2 reduction to ammonia using Mo-N complex bearing tetradentate triamidoamine ligands,6 achieving 8 equiv. of ammonia based on Mo-atom of the catalyst. Later, Mo complex bearing PNP (2,6-bis(di-tert-butyl phosphinomethyl) pyridine) ligands has been developed, in which phosphorous (P) ligands have been introduced to optimize the performance, achieving 23 equiv. of ammonia with the catalyst, of which one molybdenum atom producing 12 equiv. of ammonia.7 When more P-ligands have been introduced, like a mer-triphosphine, up to 63 equiv. of ammonia can be produced, based on the molybdenum atom.8 Last year, Mo mediated by PNP-type pincer ligands even sets a new record of 415 equiv. NH3 per Mo.9,10 These achievements vividly demonstrated the importance of ligands optimization in the design of NRR catalysts. Following these successes in design of molecular complex, the searching of Mo-based NRR catalysts has been extended to inorganic materials and single-Mo catalysts in recent years.11-17 Luis et al. examined M3C2 MXene (M=d2-d4 transition metals), including Mo3C2 (Mo is threecoordinated as Mo-C3), which shows a strong capacity to adsorb N2.11 Mo-N bonding has also been considered for NRR catalysis.12-14 As reported by Li et al., MoN2 nanoshseet may generate N-vacancy under electrochemical conditions, which is active to fill N2 via the Mo-N3 bonding and offer excellent NRR performance, especially after Fe-doping.14 In addition, single-Mo catalysts have been incorporated onto boron nitride (BN) nanosheet,15 and N-doped graphene,16 for which Mo-N bonding has been identified as an effective way to optimize the performance of Mo-centre. In contrast, P- and boron (B)-ligands are rarely studied in the design of inorganic NRR catalysts,


3


Page 4 of 18

although Mo-P3 complex shows great success.7,8 To further optimize its performance, systematic searching of optimum ligands is still needed. Herein, we envisaged the design of novel single Mo catalysts based on the Mo-P bonding through density functional theory (DFT) calculations. Inspired by the findings by Nishibayshi and co-workers,6-8 Mo-P3 bonding has been employed as a basic point through introducing single Mo on black phosphors (BP) to provide an ideal Mo-P3 bonding environment, followed by additional evaluation of various ligands combinations. As demonstrated below, Mo-BC2 has been successfully identified as new promising model catalyst for NRR. Methods The computational searching and calculations were carried out under the scheme of spinpolarized density functional theory (DFT) with the generalized gradient approximation (GGA) using the revised Perdew-Burke-Ernzerhof (PBE) functional.18,19 BP has been simulated with the (2×2) supercell, with single-Mo bonded by three P-ligands (Mo1/P16, dimension size: 8.98 Å × 6.62 Å), labelled as Mo-P3. The optimization of ligands is further carried out through replacing bonded P-atoms by new ligands, including C, N, S and B. NRR was studied through the calculation of free energy changes ΔG associated with all elementary steps, including N2 adsorption, reduction by six pairs of H+ and e-, and NH3 release, as described in the literature.20-22 The maximum ΔG, labelled as ΔGmax, was employed to identity the rate determining steps (RDS) and also was employed as an indicator to evaluate the NRR performance. Small ΔGmax means low energy request for full NRR. ΔGmax=1.08 eV for flat Ruthenium (Ru) (0001) was used as a reference for the discussion.23


4

Page 5 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Results Fig. 1(a) and (b) are the side and top views of optimized geometries of single Mo on BP, showing that Mo is three-coordinated and the planar shape of BP has been kept well. Specifically, the averaged distance between Mo and nearest phosphorus is 2.27 Å, slightly shorter than that of single Mo-P bond in typical Mo-based complexes (2.4-2.6 Å),24 indicating the effective Mo-P bonding. To further determine the oxidation state of Mo, Bader charge has been particularly calculated, as presented in Fig. 1(c). Mo shows an oxidation state of +0.48e, indicating net electron transferred from Mo to BP, which are mainly distributed in the first P-layer.

Figure 1. Mo-P3 Catalyst. (a) Top view and (b) Side view for single Mo adsorbed on BP; (c) Bader charges distribution. Mo and P are shown cyan and purple spheres. The charges shown in (c) are in unit of e. The N2 adsorption on single Mo adsorbed BP with three potential geometries was analyzed, including end-on, side-on and tilt end-on adsorptions25. The optimized geometries are shown in Fig. 2(a)-(c), together with the calculated adsorption energies in unit of eV, according to which


5


Page 6 of 18

end-on adsorption is more favourable by >0.2 eV than the other two, demonstrated by a calculated adsorption energy AE(N2*) of -1.15 eV. N-N distance increases from 1.10 Å (free N2) to 1.13 Å, indicating that N≡N bond is partially activated. Fig. 2(d) shows the detailed distribution of Bader charges after N2-adsorption. Clearly, the net charge on Mo has been slightly changed, from +0.48e to +0.71e, donating 0.41e to BP and 0.30e to N2. This result suggests a competition between N2 and the substrate for electrons. Interestingly, two N-atoms show quite different charge states, as 0.42e (bonded with Mo, NMo) and +0.12e (the outside one), as shown in Fig. 2(e). Another interesting change is that the charge for one of three P-ligands varies from -0.06e to -0.14e, suggesting that the ligands can strongly affect the charge state of Mo centre and consequently the charge transfer from Mo to N2, which is critical for N2 activation and reduction. Fig. 2(f) summarizes the basic information for charge transfer involved in N2-adsorption on such Mo-L3 catalysts. Specifically, the charge on Nout (the first hydrogenation site for NRR) is determined by NoutNMo and MoNMo electron transfer, highlighting the importance of ligands in the tuning of charge state of Mo-centre. For full NRR, three mechanisms have been investigated for Mo-P3, including Distal, Alternating and Enzymatic paths and their hybrid forms.26 In the case of Mo-P3, Distal path is preferred based on our tests (see Supporting Information, Fig. S1). The geometry of Mo-P3 catalyst and the calculated ΔG for each elementary step have been shown in Fig. 3(a) and (b), respectively. Accordingly, ΔGmax=1.04 eV has been derived for the rate-determining step (RDS), N2*+H++eN2H*, with all intermediate states shown in Fig. S2(a). To understand the N2-Mo interaction, we start from the molecular orbitals of free nitrogen whose bonding states of σ2p and π2p are fully occupied; therefore, electron injection to antibonding states or donation to catalysts can reduce the N-N bond order, activating N≡N bond. According to Fig. 2(e), electron injection 0.30e from Mo


6

Page 7 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


to N2 occurs, which is the basis for strong Mo-N interaction, leading to large AE. However, such charge transfer is highly localized to NMo. Nout, as the preferred site for the first H-adding, is positively charged (+0.12e); as a result, its σ2p is partially empty. Following this analysis, NRR performance of Mo centre may be improved through tuning its capacity of electron donation to N2. As P3 ligands are competing with N2 for electron transfer, it is likely to regulate the reactivity of Mo centre through introducing new non-metals to replace P-ligands.

Figure 2. N2 adsorption on Mo-P3 Catalyst. (a) End-on adsorption; (b) Side-on adsorption; and (c) Tilt end-on adsorption; (d) Bader charges distribution after N2 adsorption for Mo (cyan) and P (purple, only showing the first layer); (e) Bader charges for Mo and adsorbed N2 (blue); (f) Schematic configuration for the charge transfer involved with Mo-L3 catalysts after N2 adsorption. The charges are in unit of e.


7


Page 8 of 18

Following the above analysis, a series of model catalysts starting from Mo-P3 configuration were screened with p-block ligands, including B, C, N, S and P. The optimized geometries for these seventeen catalysts have been shown in Fig. S3 and Fig. S4. In most cases, Mo can be stably fixed by the ligands, except S-ligands which result in severe distortion, indicating low stability of these structures as listed separately in Fig. S4. Therefore, S-ligands are not recommended for the mediation of single-Mo catalyst, as confirmed later by full calculation of NRR. For NRR, the adsorption energy of single nitrogen, labelled as E(N*), has been widely employed as an indicator to search for potential catalysts and a volcano curve has often been obtained when the experimental values of turn-over frequency is plotted against E(N*).27 Promising NRR catalysts offer a value of E(N*) being close to that for Ru benchmark catalyst. Among the seventeen concept catalysts, Mo-P2B and Mo-PC2 shows E(N*)=-0.47 eV and -0.36 eV, being close to that of stepped Ru(0001),28 suggesting these two are promising NRR catalysts. To further examine this result, a calculation of full NRR over these catalysts has been carried out, with calculated ΔGmax plotted versus E(N*) in Fig. 4(a), in which the catalysts are indicated by the ligands for clarity. A reversed volcano curve between ΔGmax and E(N*)-E(N*)Ru has been indicated, confirming the validity of E(N*) indicator. It is worth to note that such validity origins from the linear relationship between the adsorption energies of intermediates; therefore, scattering or deviation from the linear is often seen, particularly when the RDS comes from N2 adsorption or NH3 release or the surface reconstruction during the reaction (e.g, for S-mediated single-Mo catalysts in this work, severe distortion has been found, see Fig. S4). In Fig. 3(a), several concept catalysts have been identified as ‘good candidates’ using E(N*) indicator, which have been confirmed to offer as excellent catalysts by the calculated ΔGmax (e.g., 0.43 eV for Mo-PC2 and ~0.60 eV for Mo-BC2 and Mo-PB2). Moreover, the RDS is similar with


8

Page 9 of 18 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Mo-P3 case, which comes from the *N2*N2H step (see dashed rectangle in Fig. 3(b)), but the overall energy request becomes smaller as shown by the full energy profile. Such performance is close or even lower than the half of that for flat Ru(0001) (0.54 eV)23, a benchmark NRR catalyst. Note N- and S-ligands do not bring notable improvement with respect to Mo-P3 system, being consistent with the observation for molecular catalysts.6,7 Using calculated ΔGmax as an indicator of the performance, a reversed volcano curve has been indicated by the dashed red line in Fig. 4(a), in which E(N)Ru is the adsorption energy of N-atom on stepped Ru(0001)28. It is also worth noting that (i) such volcano curve points out that an optimum value for E(N*) being close (but not exactly) to E(N)Ru, consistent with early study2; and (ii) for both P2B and PC2, E(N*) is very close to that of Ru benchmark catalyst, highlighting the value of these conceptual catalysts and importance of B- and C-ligands in the tuning of Mo-centre. For NRR, hydrogen evolution reaction (HER) is the key side reaction, leading to low Faraday efficiency (FE) when HER is more favourable than NRR, which has become a major challenge for most NRR catalysts.29-31 Therefore, additional calculations have been performed for these catalysts and presented with ΔGmax for HER versus that for NRR in Fig. 4(b), with the red dashed line showing ΔGmax(NRR) being equal to that for HER, in which case the points above the line indicate that NRR is favourable. In the case of Mo-P2S, sulphur diffuses into the BP lattice, resulting in heavy local strain energy associated with a notable deviation from the starting geometry. With Hadsorption, such strain has been partially released and the calculated ΔG(H*) is as low as -3.99 eV; as a result, releasing H2 becomes extremely difficult (>4.0 eV). For other catalysts without Sligands, the calculated ΔGmax for HER is in the range of 0.13-1.60 eV, but only those catalysts with ΔGmax(NRR)

Computational Design of Single Molybdenum Catalysts for Nitrogen

Recommend Documents