Design of Efficient Catalysts with Double Transition Metal Atoms on ...
Supporting Information
Design of Efficient Catalysts with Double Transition Metal Atoms on C2N Layer
Xiyu Li,1,† Wenhui Zhong,2,† Peng Cui,1 Jun Li,1 Jun Jiang1,*
1
Hefei National Laboratory for Physical Sciences? at the Microscale, iChEM (Collaborative Innovation Center of Chemistry for Energy
Materials), School of Chemistry and Materials Science, University of Science and Technology of China, Hefei, Anhui 230026, P. R. China 2
Guizhou Provincial Key Laboratory of Computational Nano-Material Science, Institute of Applied Physics, Guizhou Synerget-ic Innovation
Center of Scientific Big Data for Advanced Manufacturing Technology, Guizhou Normal College, Gaoxin Road 115, Guiyang, Guizhou 550018, P. R. China
*Corresponding author. E-mail: [email protected]
†These authors contributed equally.
S1
The GGA+U method was applied to describe partially filled d-orbitals by considering coulomb and exchange corrections.1 Referring to reported theoretical studies of system containing 3d transition metal, we used 4 eV as correlation energy (U) and 1 eV as the exchange energy (J).22 The empirical correction method (DFT-D2) was played to describe the long-range van der Waals (vdW) interactions.1 The periodic boundary condition was set with a 15 Å vacuum region above the plane of one C2N unit cell. All geometric structures were fully relaxed until energy and forces were converged to 10-5 eV and 0.01 eV Å-1, respectively. The Brillouin zone was sampled with 5×5×1 Monkhorst-Pack k-meshes for the geometry optimization, and the kinetic energy cutoff is set to be 550 eV in the plane-wave expansion. The HSE06 functional2,3 was utilized to examine and validate the simulated band structure of monolayer C2N. The method of climbing image nudged elastic band (CI-NEB) was used for transition state search.13 Three images were inserted into initial and final states. The spring force between adjacent images was 5.0 eV Å-1, and images were optimized until the forces on each atom are less than 0.03 eV Å-1. All calculations were performed with the PBE+vdW+U method. An implicit solvation model (VASPsol)4 was employed to examined the water-solvent effect on O2 dissociations for TM2@C2N and ORR pathways for Co2@C2N.
S2
Figure S1. (a) Top and side view of atomic structure of monolayer C2N-h2D in the unit cell. Energy band structures of monolayer C2N-h2D computed with the functional of PBE (b) and HSE06 (c).
S3
Figure S2. Top view of optimized geometries of Pt-N in Pt@C2N (a), Co-N in Co@C2N (b), Ni-N in Ni@C2N (c), Cu-N in Cu@C2N (d). With two TM-N bonds formed in TM@C2N, there are only negligible changes of lattice and planarity of C2N induced by anchoring one TM into the N-hole of C2N. The bond lengths are given in Å.
S4
Figure S3. Top view of optimized geometries of Pt-N/Pt-Pt in Pt2@C2N (a), Co-N/Co-Co in Co2@C2N (b), Ni-N/Ni-Ni in Ni2@C2N (c), Cu-N/Cu-Cu in Cu2@C2N (d). With four TM-N bonds formed in TM@C2N, there are slight distortion of C2N plane induced by anchoring two TM into the N-hole of C2N. The bond lengths are given in Å.
S5
Table S1. The computed binding energies (Eb) of TM clusters on C2N, TM polarization positive charges extracted from the C2N plane in TM@C2N and TM2@C2N, and the bulk cohesive energy (eV/atom) of four TMs. TM charge (e+)
Eb (eV) TM@C2N TM2@C2N TM to TM@C2N
TM@C2N TM2@C2N
Bulk cohesive energy(eV/atom)
Pt
4.69
7.02
2.33
0.47
0.64
5.99
Co
6.21
11.30
5.09
0.96
1.50
5.36
Ni
4.85
7.48
2.63
0.74
1.14
4.38
Cu
3.85
5.76
1.91
0.68
1.10
3.38
All of the polarization charges were obtained from Bader charge analysis. For TM1-2@C2N, it is calculated by subtracting the Bader charges of the individual C2N monolayer and TM atoms from that of the hybrid system TM1-2@C2N. The same protocol applies to calculate the polarization charges between O2 and TM1-2@C2N.
S6
Table S2. The computed diffusion barriers (eV) of TM atoms on the perfect monolayer of C2N, graphene and h-BN. C2N (eV)
graphene (eV)
h-BN (eV)
Pt
2.97
0.69
0.53
Co
3.91
0.11
0.03
Ni
3.04
0.14
0.02
Cu
3.33
0.08
0.02
S7
Figure S4. Computed energy band structures of TM@C2N (a) and TM2@C2N (b) (From left to right: TM=Pt, Co, Ni, Cu), showing the metallic nature. Here Cu@C2N, Pt2@C2N, Ni2@C2N, Cu2@C2N contain no magnetic moment (see Table S2 below), while the others use black and red curves to represent spin up and down electrons.
S8
Table S3. Computed magnetic moment per TM atom (MTM), total magnetic moment per unit cell (M ) of TM@C2N and TM2@C2N, and the exchange energies (Eex) for U
2×2 supercell of TM@C2N and one unit cell of TM2@C2N. TM@C2N
TM2@C2N
Pt
Co
Ni
Cu
Pt
Co
Ni
Cu
MTM (µB)
1
2
1
0
1
2
1
0
MU (µB)
1
3
1
0
0
0
0
0
Eex (meV)
-1
36
-4
0
0
109
45
0
A 2×2 supercell of TM@C2N was employed to investigate magnetic states with the exchange energy (Eex). The exchange energy is defined as Eex = EFM − EAFM, in which EFM and EAFM represent the energy of ferromagnetic state and antiferromagnetic state, respectively. Negative values of Eex indicate that the ferromagnetic state is the ground state, otherwise antiferromagnetic state is the ground one. The Eex and magnetic moment per unit cell are listed in Table S3. Here Pt/Ni atoms interact ferromagnetically with each other, while Co atoms interact antiferromagneticly. Cu@C2N is a nonmagnetic system. For the case of TM2@C2N, we investigated the magnetic coupling between two TM atoms in a unit cell. The ground state of Pt2/Cu2@C2N is the nonmagnetic state, while Co2/Ni2@C2N prefers antiferromagnetic state. The total magnetic moment per unit cell (MU) of TM2@C2N is zero for all double TM systems. Besides, the adsorption and dissociation of O2 were simulated on TM2@C2N with ground state. It is noted that the magnetic coupling of TMs in TM/TM1-2@C2N can only slightly affect the adsorption and dissociation of O2.
S9
Figure S5. The atomic structure of O2 physically adsorbed on pure C2N monolayer. The bond length is in Å.
S10
Figure S6. From left to right: the optimized geometries of initial state (IS), transition state (TS), final state (FS), and the potential energy profile of one adsorbed O2 molecule dissociated on Pt@C2N (a), Co@C2N (b), Ni@C2N (c), Cu@C2N (d). The bond lengths are in Å. The O2 adsorbed on Pt@C2N is side-on configuration; the end-on configuration of O2 is preferred for system of Co/Ni/Cu@C2N.
S11
Figure S7. From left to right: the optimized geometries of initial state (IS), transition state (TS), final state (FS), and the potential energy profile of one adsorbed O2 molecule dissociated on Pt2@C2N (a), Co2@C2N (b), Ni2@C2N (c), Cu2@C2N (d). The bond lengths are in Å. The O2 adsorption are all side-on configuration.
S12
Table S4. The computed charges trapped by the adsorbed O2 after adding one extra electron to TM@C2N and TM2@C2N to model the external field effect (normally induced by electrochemistry or photocatalysis process). O2
charge
after
one
additional electron (e-) TM@C2N TM2@C2N Pt
-0.24
-0.16
Co
-0.22
-0.25
Ni
-0.21
-0.25
Cu
-0.24
-0.24
S13
Figure S8. The dependence of reaction barrier on O2 polarization negative charge induced by external field of one additional electron in TM@C2N and TM2@C2N.
S14
Table S5. The computed charges donated from the C2N plane to O2 induced by its adsorption to TM@C2N and TM2@C2N. charge donated from C2N to O2 (e-) TM@C2N TM2@C2N Pt
-0.22
-0.16
Co
-0.33
-0.42
Ni
-0.14
-0.29
Cu
-0.17
-0.26
S15
Figure S9. Atomic structures of relaxed geometries for various ORR chemical species adsorbed on the Co2@C2N. (a−d) H, O, OH, and OOH adsorption on the central Co-Co of Co2@C2N, respectively.
S16
Figure S10. The reaction pathway from O2 dissociation, and the two subsequent hydrogenation of the atomic O to generate H2O molecular. Here, the transition states (TS) are marked by the red rectangular box, and the corresponding reaction and activation energy are presented above the TS in the form of (∆E, Ea).
In general, the reaction energies should be negative (exothermic) along the reaction pathways. While, the second OH hydrogenation reaction is endothermic with a reaction energy of 0.35 eV along the O2 dissociation pathway and OOH dissociation pathway. It should be noted that ORR often involves many reaction steps, in which some endothermic elementary steps could occur to achieve overall exothermic ORR. For instance, the reaction energies of OH hydrogenation reaction of ORR on layered SiC sheet is also endothermic with of 0.46~0.54 eV.5 In our work, the overall ORR reaction energy along the O2 dissociation pathway and OOH dissociation pathway on Co2@C2N is -13.82 eV and -6.62 eV, respectively (Figure 4c). In addition, the product H2O of OH hydrogenation reaction can be easily removed due to the weak adsorption energy and O2 competitive adsorption. These suggests that the ORR could proceed on Co2@C2N.
S17
Figure S11. The hydrogenation process of the adsorbed O2 and the formation and subsequent dissociation of the OOH species: O atom and OH molecular. After the hydrogenation of the atomic O, the configuration with two OH molecular transform into that result from the hydrogenation of two dissociated O atom on the pathway of O2 dissociation. And the remaining steps are the same as those along the O2 dissociation pathway. Here, the transition states (TS) are marked by the red rectangular box, and the corresponding reaction and activation energy are presented above the TS in the form of (∆E, Ea).
The adsorbed structures of reaction species on Co2@C2N, such as H, O, OH, OOH, were displayed in Figure S9. There are two nitrogen atoms with lone pair electrons in one hole of the configuration of Co2@C2N, with which the hydrogen atoms tend to form H-N bonds in the plane (Figure S9a). For the O2 dissociation pathway, following the O2 dissociation with no barrier (exothermic with reaction energy of 1.25 eV), the formed oxygen atom can easily take the near H from N due to its higher electronegativity to form OH, shown in Figure S10. Subsequently, the two lone-pair-electron N atoms can capture another two H atoms, respectively (Figure S18
S10). And the two following hydrogenation reactions will proceed to two H2O molecules of the final ORR product with activation barriers of 0.11 eV (endothermic with reaction energy of 0.07 eV ) and 0.39 eV (endothermic with reaction energy of 0.35 eV), respectively. In addition, the pathway of OOH dissociation (Figure S11), initiated with a hydrogenation reaction to form an adsorbed OOH with an activation barrier of 0.28 eV (exothermic with reaction energy of 0.02 eV). Subsequently, the OOH is decomposed to an O and an OH by the break of O-O bond with activation energy of 0.09 eV (endothermic reaction energy of 0.14 eV). The formed O atom will undergo hydrogenation reaction to generate two OH. Overcoming a barrier of 0.31 eV, this configuration transforms into the 2OH configuration in O2 dissociation pathway. And the remaining steps are the same as those in the O2 dissociation pathway. Considering the extremely low dissociation barrier of OOH (0.09 eV), it is not necessary to study the hydrogenation reaction of OOH and the reaction pathway of HOOH dissociation. Hence, we investigated the reaction pathways for ORR on Co2@C2N using first-principles DFT calculations. Our DFT results indicate that the O2 would be chemisorbed to the two central Co atoms, and a four-electron O2 dissociation pathway would be kinetically favorable, in which the H2O formation from OH hydrogenation is the rate-determining step with an activation energy of 0.39 eV.
S19
Figure S12. Calculated free energy diagrams for ORR on Co2@C2N along O2 dissociation pathway. U is the applied electrode potential and the limiting potential for ORR is as high as 0.30 eV, which indicate that the O2 dissociation ORR pathway is thermodynamically viable for some critical electrode potentials.
The change of Gibbs free energy (∆G) for all ORR step was calculated by35 ∆G = ∆E + ∆Ezpe - T∆S + ∆GpH -1/2GH2 + neU The reaction energy (∆E) can be obtained by analyzing the DFT total energies. The harmonic vibrational frequency calculations were performed to determine the zero point energy ∆Ezpe. ∆S is the entropy difference between the adsorbed state ant the gas phase, and T is of room temperature 298.15 K. ∆GpH = 2.303kBT pH is the free energy contribution depending on the variations of H concentration, and the value of pH was assumed to be zero for acidic medium in this work. The contribution of potential was computed assuming (H+ + e−) =1/2GH2 – neU (pH = 0), where GH2 is the free energy of H2 molecule, n is the number of transferred electrons, e represents the transferred electron, and U is the operating electrochemical potential relative to the reversible hydrogen electrode (RHE). The entropies of the free molecules (such as O2) can be taken from the NIST database6 and the energy contribution from the configuration entropy in the adsorbed state was not included.
S20
Figure S13. (a) The schematic arrangement of two TM atom embedded into two adjacent hole vacancies (Hole1 and Hole2) in TM@C2N. Here we have tried six kinds of configurations as labeled with number 1-6 in the Hole2 vacancy. (b) The schematic arrangement of two double-TM embedded into two adjacent hole vacancies (Hole1 and Hole2) in TM2@C2N. Here we have tried three kinds of configurations as labeled with number 1-3 in the Hole2 vacancy. Here we built a model of 2×1 C2N supercell with one or two TM atoms deposited in each hole. The optimized geometries for these configurations exhibit nearly the same C-N/TM-N bond lengths in two adjacent holes.
S21
Table S6. The Bader charge (e+) of TM in different hole vacancy sites in TM@C2N (structure in Figure S12). Pt
Co
Ni
Cu
TM@C2N Hole1 Hole2 Hole1 Hole2 Hole1 Hole2 Hole1 Hole2 1
0.47
0.47
0.96
0.96
0.74
0.74
0.68
0.68
2
0.47
0.47
0.96
0.96
0.74
0.74
0.68
0.68
3
0.47
0.46
0.96
0.97
0.74
0.75
0.68
0.68
4
0.47
0.47
0.96
0.96
0.74
0.74
0.68
0.68
5
0.46
0.47
0.96
0.96
0.74
0.74
0.68
0.68
6
0.47
0.47
0.96
0.96
0.74
0.74
0.68
0.68
S22
Table S7. The Bader charge (e+) of TM in different hole vacancy sites in TM2@C2N (structure in Figure S12).
Pt
Co
Ni
Cu
TM2@C2N Hole1 Hole2 Hole1 Hole2 Hole1 Hole2 Hole1 Hole2 1
0.63
0.64
1.46
1.46
1.14
1.14
1.10
1.10
2
0.68
0.69
1.46
1.47
1.16
1.16
1.09
1.09
3
0.64
0.66
1.46
1.46
1.15
1.15
1.08
1.09
S23
Table S8. The dissociation barriers (Ea) without solvent effect, in comparing to those (Easolv) in the water-solvated phase for O2 adsorbed on TM2@C2N.
TM2@C2N Pt Ea (eV)
Co
Ni
Cu
0.63 0.00 0.11 0.56
Ea solv (eV) 0.68 0.00 0.10 0.64
S24
Figure S14. The O2 and OOH dissociation initiated pathways for ORR catalyzed by Co2@C2N in the water-solvated phase, the reaction energy (∆Esolv) and activation energy (Easolv) for all steps are given in parentheses with the form of (∆Esolv, Easolv).
We have investigated the reaction pathways for ORR on Co2@C2N in the water-solvated phase. The results indicate that the solvent effect on the Co2@C2N-catalyzed ORR along the favorable O2 dissociation pathway is insignificant, when the H2O formation from OH hydrogenation is the rate-determining step with an activation energy of 0.39 eV.
S25
REFERENCES (1) Grimme, S. Semiempirical GGA-type Density Functional Constructed with a Long-range Dispersion Correction. J. Comput. Chem. 2006, 27, 1787-1799. (2) Heyd, J.; Scuseria, G. E.; Ernzerhof, M. Hybrid Functionals Based on a Screened Coulomb Potential. J. Chem. Phys. 2003, 118, 8207. (3) Heyd, J.; Scuseria, G. E., G. Efficient Hybrid Density Functional Calculations in Solids: Assessment of the Heyd–Scuseria–Ernzerhof Screened Coulomb Hybrid Functional. J. Chem. Phys. 2004, 121, 1187. (4) Mathew, K.; Sundararaman, R.; Letchworth-Weaver, K.; Arias, T. A.; Hennig, R. G. Implicit Solvation Model for Density-functional Study of Nanocrystal Surfaces and Reaction Pathways. J. Chem. Phys. 2014, 140, 084106. (5) Zhang, P.; Xiao, B. B.; Hou, X. L.; Zhu, Y. F.; Jiang, Q. Layered SiC Sheets: A Potential Catalyst for Oxygen Reduction Reaction. Sci. Rep. 2014, 4, 3821. (6)
Computational
Chemistry
Comparison
http://cccbdb.nist.gov/.
S26
and
Benchmark
Database.
Design of Efficient Catalysts with Double Transition Metal Atoms on C2N Layer
Xiyu Li,1,† Wenhui Zhong,2,† Peng Cui,1 Jun Li,1 Jun Jiang1,*
1
Hefei National Laboratory for Physical Sciences? at the Microscale, iChEM (Collaborative Innovation Center of Chemistry for Energy
Materials), School of Chemistry and Materials Science, University of Science and Technology of China, Hefei, Anhui 230026, P. R. China 2
Guizhou Provincial Key Laboratory of Computational Nano-Material Science, Institute of Applied Physics, Guizhou Synerget-ic Innovation
Center of Scientific Big Data for Advanced Manufacturing Technology, Guizhou Normal College, Gaoxin Road 115, Guiyang, Guizhou 550018, P. R. China
*Corresponding author. E-mail: [email protected]
†These authors contributed equally.
S1
The GGA+U method was applied to describe partially filled d-orbitals by considering coulomb and exchange corrections.1 Referring to reported theoretical studies of system containing 3d transition metal, we used 4 eV as correlation energy (U) and 1 eV as the exchange energy (J).22 The empirical correction method (DFT-D2) was played to describe the long-range van der Waals (vdW) interactions.1 The periodic boundary condition was set with a 15 Å vacuum region above the plane of one C2N unit cell. All geometric structures were fully relaxed until energy and forces were converged to 10-5 eV and 0.01 eV Å-1, respectively. The Brillouin zone was sampled with 5×5×1 Monkhorst-Pack k-meshes for the geometry optimization, and the kinetic energy cutoff is set to be 550 eV in the plane-wave expansion. The HSE06 functional2,3 was utilized to examine and validate the simulated band structure of monolayer C2N. The method of climbing image nudged elastic band (CI-NEB) was used for transition state search.13 Three images were inserted into initial and final states. The spring force between adjacent images was 5.0 eV Å-1, and images were optimized until the forces on each atom are less than 0.03 eV Å-1. All calculations were performed with the PBE+vdW+U method. An implicit solvation model (VASPsol)4 was employed to examined the water-solvent effect on O2 dissociations for TM2@C2N and ORR pathways for Co2@C2N.
S2
Figure S1. (a) Top and side view of atomic structure of monolayer C2N-h2D in the unit cell. Energy band structures of monolayer C2N-h2D computed with the functional of PBE (b) and HSE06 (c).
S3
Figure S2. Top view of optimized geometries of Pt-N in Pt@C2N (a), Co-N in Co@C2N (b), Ni-N in Ni@C2N (c), Cu-N in Cu@C2N (d). With two TM-N bonds formed in TM@C2N, there are only negligible changes of lattice and planarity of C2N induced by anchoring one TM into the N-hole of C2N. The bond lengths are given in Å.
S4
Figure S3. Top view of optimized geometries of Pt-N/Pt-Pt in Pt2@C2N (a), Co-N/Co-Co in Co2@C2N (b), Ni-N/Ni-Ni in Ni2@C2N (c), Cu-N/Cu-Cu in Cu2@C2N (d). With four TM-N bonds formed in TM@C2N, there are slight distortion of C2N plane induced by anchoring two TM into the N-hole of C2N. The bond lengths are given in Å.
S5
Table S1. The computed binding energies (Eb) of TM clusters on C2N, TM polarization positive charges extracted from the C2N plane in TM@C2N and TM2@C2N, and the bulk cohesive energy (eV/atom) of four TMs. TM charge (e+)
Eb (eV) TM@C2N TM2@C2N TM to TM@C2N
TM@C2N TM2@C2N
Bulk cohesive energy(eV/atom)
Pt
4.69
7.02
2.33
0.47
0.64
5.99
Co
6.21
11.30
5.09
0.96
1.50
5.36
Ni
4.85
7.48
2.63
0.74
1.14
4.38
Cu
3.85
5.76
1.91
0.68
1.10
3.38
All of the polarization charges were obtained from Bader charge analysis. For TM1-2@C2N, it is calculated by subtracting the Bader charges of the individual C2N monolayer and TM atoms from that of the hybrid system TM1-2@C2N. The same protocol applies to calculate the polarization charges between O2 and TM1-2@C2N.
S6
Table S2. The computed diffusion barriers (eV) of TM atoms on the perfect monolayer of C2N, graphene and h-BN. C2N (eV)
graphene (eV)
h-BN (eV)
Pt
2.97
0.69
0.53
Co
3.91
0.11
0.03
Ni
3.04
0.14
0.02
Cu
3.33
0.08
0.02
S7
Figure S4. Computed energy band structures of TM@C2N (a) and TM2@C2N (b) (From left to right: TM=Pt, Co, Ni, Cu), showing the metallic nature. Here Cu@C2N, Pt2@C2N, Ni2@C2N, Cu2@C2N contain no magnetic moment (see Table S2 below), while the others use black and red curves to represent spin up and down electrons.
S8
Table S3. Computed magnetic moment per TM atom (MTM), total magnetic moment per unit cell (M ) of TM@C2N and TM2@C2N, and the exchange energies (Eex) for U
2×2 supercell of TM@C2N and one unit cell of TM2@C2N. TM@C2N
TM2@C2N
Pt
Co
Ni
Cu
Pt
Co
Ni
Cu
MTM (µB)
1
2
1
0
1
2
1
0
MU (µB)
1
3
1
0
0
0
0
0
Eex (meV)
-1
36
-4
0
0
109
45
0
A 2×2 supercell of TM@C2N was employed to investigate magnetic states with the exchange energy (Eex). The exchange energy is defined as Eex = EFM − EAFM, in which EFM and EAFM represent the energy of ferromagnetic state and antiferromagnetic state, respectively. Negative values of Eex indicate that the ferromagnetic state is the ground state, otherwise antiferromagnetic state is the ground one. The Eex and magnetic moment per unit cell are listed in Table S3. Here Pt/Ni atoms interact ferromagnetically with each other, while Co atoms interact antiferromagneticly. Cu@C2N is a nonmagnetic system. For the case of TM2@C2N, we investigated the magnetic coupling between two TM atoms in a unit cell. The ground state of Pt2/Cu2@C2N is the nonmagnetic state, while Co2/Ni2@C2N prefers antiferromagnetic state. The total magnetic moment per unit cell (MU) of TM2@C2N is zero for all double TM systems. Besides, the adsorption and dissociation of O2 were simulated on TM2@C2N with ground state. It is noted that the magnetic coupling of TMs in TM/TM1-2@C2N can only slightly affect the adsorption and dissociation of O2.
S9
Figure S5. The atomic structure of O2 physically adsorbed on pure C2N monolayer. The bond length is in Å.
S10
Figure S6. From left to right: the optimized geometries of initial state (IS), transition state (TS), final state (FS), and the potential energy profile of one adsorbed O2 molecule dissociated on Pt@C2N (a), Co@C2N (b), Ni@C2N (c), Cu@C2N (d). The bond lengths are in Å. The O2 adsorbed on Pt@C2N is side-on configuration; the end-on configuration of O2 is preferred for system of Co/Ni/Cu@C2N.
S11
Figure S7. From left to right: the optimized geometries of initial state (IS), transition state (TS), final state (FS), and the potential energy profile of one adsorbed O2 molecule dissociated on Pt2@C2N (a), Co2@C2N (b), Ni2@C2N (c), Cu2@C2N (d). The bond lengths are in Å. The O2 adsorption are all side-on configuration.
S12
Table S4. The computed charges trapped by the adsorbed O2 after adding one extra electron to TM@C2N and TM2@C2N to model the external field effect (normally induced by electrochemistry or photocatalysis process). O2
charge
after
one
additional electron (e-) TM@C2N TM2@C2N Pt
-0.24
-0.16
Co
-0.22
-0.25
Ni
-0.21
-0.25
Cu
-0.24
-0.24
S13
Figure S8. The dependence of reaction barrier on O2 polarization negative charge induced by external field of one additional electron in TM@C2N and TM2@C2N.
S14
Table S5. The computed charges donated from the C2N plane to O2 induced by its adsorption to TM@C2N and TM2@C2N. charge donated from C2N to O2 (e-) TM@C2N TM2@C2N Pt
-0.22
-0.16
Co
-0.33
-0.42
Ni
-0.14
-0.29
Cu
-0.17
-0.26
S15
Figure S9. Atomic structures of relaxed geometries for various ORR chemical species adsorbed on the Co2@C2N. (a−d) H, O, OH, and OOH adsorption on the central Co-Co of Co2@C2N, respectively.
S16
Figure S10. The reaction pathway from O2 dissociation, and the two subsequent hydrogenation of the atomic O to generate H2O molecular. Here, the transition states (TS) are marked by the red rectangular box, and the corresponding reaction and activation energy are presented above the TS in the form of (∆E, Ea).
In general, the reaction energies should be negative (exothermic) along the reaction pathways. While, the second OH hydrogenation reaction is endothermic with a reaction energy of 0.35 eV along the O2 dissociation pathway and OOH dissociation pathway. It should be noted that ORR often involves many reaction steps, in which some endothermic elementary steps could occur to achieve overall exothermic ORR. For instance, the reaction energies of OH hydrogenation reaction of ORR on layered SiC sheet is also endothermic with of 0.46~0.54 eV.5 In our work, the overall ORR reaction energy along the O2 dissociation pathway and OOH dissociation pathway on Co2@C2N is -13.82 eV and -6.62 eV, respectively (Figure 4c). In addition, the product H2O of OH hydrogenation reaction can be easily removed due to the weak adsorption energy and O2 competitive adsorption. These suggests that the ORR could proceed on Co2@C2N.
S17
Figure S11. The hydrogenation process of the adsorbed O2 and the formation and subsequent dissociation of the OOH species: O atom and OH molecular. After the hydrogenation of the atomic O, the configuration with two OH molecular transform into that result from the hydrogenation of two dissociated O atom on the pathway of O2 dissociation. And the remaining steps are the same as those along the O2 dissociation pathway. Here, the transition states (TS) are marked by the red rectangular box, and the corresponding reaction and activation energy are presented above the TS in the form of (∆E, Ea).
The adsorbed structures of reaction species on Co2@C2N, such as H, O, OH, OOH, were displayed in Figure S9. There are two nitrogen atoms with lone pair electrons in one hole of the configuration of Co2@C2N, with which the hydrogen atoms tend to form H-N bonds in the plane (Figure S9a). For the O2 dissociation pathway, following the O2 dissociation with no barrier (exothermic with reaction energy of 1.25 eV), the formed oxygen atom can easily take the near H from N due to its higher electronegativity to form OH, shown in Figure S10. Subsequently, the two lone-pair-electron N atoms can capture another two H atoms, respectively (Figure S18
S10). And the two following hydrogenation reactions will proceed to two H2O molecules of the final ORR product with activation barriers of 0.11 eV (endothermic with reaction energy of 0.07 eV ) and 0.39 eV (endothermic with reaction energy of 0.35 eV), respectively. In addition, the pathway of OOH dissociation (Figure S11), initiated with a hydrogenation reaction to form an adsorbed OOH with an activation barrier of 0.28 eV (exothermic with reaction energy of 0.02 eV). Subsequently, the OOH is decomposed to an O and an OH by the break of O-O bond with activation energy of 0.09 eV (endothermic reaction energy of 0.14 eV). The formed O atom will undergo hydrogenation reaction to generate two OH. Overcoming a barrier of 0.31 eV, this configuration transforms into the 2OH configuration in O2 dissociation pathway. And the remaining steps are the same as those in the O2 dissociation pathway. Considering the extremely low dissociation barrier of OOH (0.09 eV), it is not necessary to study the hydrogenation reaction of OOH and the reaction pathway of HOOH dissociation. Hence, we investigated the reaction pathways for ORR on Co2@C2N using first-principles DFT calculations. Our DFT results indicate that the O2 would be chemisorbed to the two central Co atoms, and a four-electron O2 dissociation pathway would be kinetically favorable, in which the H2O formation from OH hydrogenation is the rate-determining step with an activation energy of 0.39 eV.
S19
Figure S12. Calculated free energy diagrams for ORR on Co2@C2N along O2 dissociation pathway. U is the applied electrode potential and the limiting potential for ORR is as high as 0.30 eV, which indicate that the O2 dissociation ORR pathway is thermodynamically viable for some critical electrode potentials.
The change of Gibbs free energy (∆G) for all ORR step was calculated by35 ∆G = ∆E + ∆Ezpe - T∆S + ∆GpH -1/2GH2 + neU The reaction energy (∆E) can be obtained by analyzing the DFT total energies. The harmonic vibrational frequency calculations were performed to determine the zero point energy ∆Ezpe. ∆S is the entropy difference between the adsorbed state ant the gas phase, and T is of room temperature 298.15 K. ∆GpH = 2.303kBT pH is the free energy contribution depending on the variations of H concentration, and the value of pH was assumed to be zero for acidic medium in this work. The contribution of potential was computed assuming (H+ + e−) =1/2GH2 – neU (pH = 0), where GH2 is the free energy of H2 molecule, n is the number of transferred electrons, e represents the transferred electron, and U is the operating electrochemical potential relative to the reversible hydrogen electrode (RHE). The entropies of the free molecules (such as O2) can be taken from the NIST database6 and the energy contribution from the configuration entropy in the adsorbed state was not included.
S20
Figure S13. (a) The schematic arrangement of two TM atom embedded into two adjacent hole vacancies (Hole1 and Hole2) in TM@C2N. Here we have tried six kinds of configurations as labeled with number 1-6 in the Hole2 vacancy. (b) The schematic arrangement of two double-TM embedded into two adjacent hole vacancies (Hole1 and Hole2) in TM2@C2N. Here we have tried three kinds of configurations as labeled with number 1-3 in the Hole2 vacancy. Here we built a model of 2×1 C2N supercell with one or two TM atoms deposited in each hole. The optimized geometries for these configurations exhibit nearly the same C-N/TM-N bond lengths in two adjacent holes.
S21
Table S6. The Bader charge (e+) of TM in different hole vacancy sites in TM@C2N (structure in Figure S12). Pt
Co
Ni
Cu
TM@C2N Hole1 Hole2 Hole1 Hole2 Hole1 Hole2 Hole1 Hole2 1
0.47
0.47
0.96
0.96
0.74
0.74
0.68
0.68
2
0.47
0.47
0.96
0.96
0.74
0.74
0.68
0.68
3
0.47
0.46
0.96
0.97
0.74
0.75
0.68
0.68
4
0.47
0.47
0.96
0.96
0.74
0.74
0.68
0.68
5
0.46
0.47
0.96
0.96
0.74
0.74
0.68
0.68
6
0.47
0.47
0.96
0.96
0.74
0.74
0.68
0.68
S22
Table S7. The Bader charge (e+) of TM in different hole vacancy sites in TM2@C2N (structure in Figure S12).
Pt
Co
Ni
Cu
TM2@C2N Hole1 Hole2 Hole1 Hole2 Hole1 Hole2 Hole1 Hole2 1
0.63
0.64
1.46
1.46
1.14
1.14
1.10
1.10
2
0.68
0.69
1.46
1.47
1.16
1.16
1.09
1.09
3
0.64
0.66
1.46
1.46
1.15
1.15
1.08
1.09
S23
Table S8. The dissociation barriers (Ea) without solvent effect, in comparing to those (Easolv) in the water-solvated phase for O2 adsorbed on TM2@C2N.
TM2@C2N Pt Ea (eV)
Co
Ni
Cu
0.63 0.00 0.11 0.56
Ea solv (eV) 0.68 0.00 0.10 0.64
S24
Figure S14. The O2 and OOH dissociation initiated pathways for ORR catalyzed by Co2@C2N in the water-solvated phase, the reaction energy (∆Esolv) and activation energy (Easolv) for all steps are given in parentheses with the form of (∆Esolv, Easolv).
We have investigated the reaction pathways for ORR on Co2@C2N in the water-solvated phase. The results indicate that the solvent effect on the Co2@C2N-catalyzed ORR along the favorable O2 dissociation pathway is insignificant, when the H2O formation from OH hydrogenation is the rate-determining step with an activation energy of 0.39 eV.
S25
REFERENCES (1) Grimme, S. Semiempirical GGA-type Density Functional Constructed with a Long-range Dispersion Correction. J. Comput. Chem. 2006, 27, 1787-1799. (2) Heyd, J.; Scuseria, G. E.; Ernzerhof, M. Hybrid Functionals Based on a Screened Coulomb Potential. J. Chem. Phys. 2003, 118, 8207. (3) Heyd, J.; Scuseria, G. E., G. Efficient Hybrid Density Functional Calculations in Solids: Assessment of the Heyd–Scuseria–Ernzerhof Screened Coulomb Hybrid Functional. J. Chem. Phys. 2004, 121, 1187. (4) Mathew, K.; Sundararaman, R.; Letchworth-Weaver, K.; Arias, T. A.; Hennig, R. G. Implicit Solvation Model for Density-functional Study of Nanocrystal Surfaces and Reaction Pathways. J. Chem. Phys. 2014, 140, 084106. (5) Zhang, P.; Xiao, B. B.; Hou, X. L.; Zhu, Y. F.; Jiang, Q. Layered SiC Sheets: A Potential Catalyst for Oxygen Reduction Reaction. Sci. Rep. 2014, 4, 3821. (6)
Computational
Chemistry
Comparison
http://cccbdb.nist.gov/.
S26
and
Benchmark
Database.
