Understanding trends in C–H bond activation … · Understanding Trends in C-H Bond Activation in Heterogeneous Catalysis Allegra A. Latimer,1,

Supporting Information

for

Understanding Trends in C-H Bond Activation in

Heterogeneous Catalysis

Allegra A. Latimer,1,‡ Ambarish R. Kulkarni,1,‡ Hassan Aljama,1 Joseph H. Montoya,3 Jong Suk Yoo,1 Charlie Tsai,1

Frank Abild-Pedersen,1,2 Felix Studt,1,2 Jens K. Nørskov1,2,*

1 SUNCAT Center for Interface Science and Catalysis, Department of Chemical Engineering, Stanford

University, 450 Serra Mall Stanford, California 94305, USA

2 SUNCAT Center for Interface Science and Catalysis, SLAC National Accelerator Laboratory, 2575 Sand Hill

Road, Menlo Park, California 94025, United States

3 Environmental Energy Technologies Division, Lawrence Berkeley National Laboratory, 1 Cyclotron Road,

Berkeley, California 94720, USA

* corresponding author: [email protected]

‡ equal contribution

Understanding trends in C–H bond activation in heterogeneous catalysis

SUPPLEMENTARY INFORMATIONDOI: 10.1038/NMAT4760

NATURE MATERIALS | www.nature.com/naturematerials 1

http://dx.doi.org/10.1038/nmat4760

2

Table of Contents 1. Formation Energy Calculations and Kinetic Analysis Details ......................................... 3

2. Supplementary Tables .......................................................................................................... 4 3. Supplementary Figures ........................................................................................................ 9

3

1. Formation Energy Calculations and Kinetic Analysis Details

Formation Energies for a given site were calculated relative to gas-phase !"# and #" as: $%('(#)!*) = $('(#)!*) −

*"$ !"# − )

"− *

.$ #" − $('() (1)

All transition state energies are calculated relative to the clean active site and the relevant gas-phase

reactant. By assuming that the active site coverage is thermodynamically determined, we implicitly assume that the activation barrier corresponding to formation of the active site motifs is small. To test this assumption, we calculate the O2 dissociation barrier for IrO2 surface and bis-µ-oxo/MOR motif as 0.2 eV and 0.64 eV respectively. These low O2 dissociation barriers suggests that our initial assumption is plausible for similar systems.

Active site formation was assumed to be thermodynamically limited. The active site coverages for O2 and N2O activation of M-O active sites and O2/H2O activation of M-OH active sites were calculated as:

/00"2345624578 =9:;<

=>?/ABC

DE 9:;<=>?/ABC

(2)

/0F"02345624578 =GH;:GH;

<=>?/ABC

DEGH;:GH;<=>?/ABC

(3)

/0I =9J;:9:;<

=>?ABC

KL

DE 9J;:9:;<=>?ABC

KL (4)

where:

M% 1 = M '# − D"M #" − M(') (5)

M%(2) = M '# − M P"# − M ' + M(P") (6)

M%(3) = M '#! − D"M !"# − M ' − D

.M(#") (7)

4

2. Supplementary Tables

Table S1. Description of material classes shown in universal scaling line. Active site motif Description CHA A 36 atom primitive rhombohedral unit cell for chabazite was used.

The BEEF-vdW optimized lattice constants are calculated as a = 9.334 Å and α = 94.54°. As all 12 T atoms for CHA are equivalent, one Si atom was replaced by Al and Fe, Co, Ni or Cu cations were added to balance the charge. Only cation positions and adsorption in the 8MR was considered.

CHA-2Al A CHA structure with two Al atoms separated by one Si atom in the 8MR was considered.

CaO

A bridging oxygen adsorbed on the 110 surface was the active oxygen.

Doped (Ca/Mg)O

A Ca/Mg atom adjacent to an active lattice oxygen on a 110 surface was replaced with Li or Na.

Graphene Initial structure was obtained from published literature.2 Only the N decorated version of the structure was used.

MOF-74

Initial structure was obtained from Verma et al1 and the lattice constantes were optimized using DFT for the Fe variant (a=13.8 Å, b=15.2 Å, c=15.2 Å, α=117.8°, β=98.7°, γ=98.7°) and used for other cations. A 108-atom unitcell was used with 12 metal atoms ( 11 Mg and 1 transition metal cation).

MOR A orthorhombic unit cell was used for DFT lattice constant optimization (a=18.3 Å, b=20.6 Å, c=7.6 Å). The M-O-M, bis-µ-oxo and M3O3 are located in the main zeolite channel as shown in Figure 2.

MgO

Either a bridging oxygen adsorbed on the 110 surface or a stoichiometric surface oxygen on the 211 surface was the active oxygen.

Nanoparticles

A 13 atom cluster of Au or Ag was used.

5

Ofcc- or Oot- Promoted Constrained Metals

On a 111 surface, oxygen was adsorbed at an fcc or on-top site. Instead of allowing CH3 to be stabilized by the metal surface in the TS, methane was constrained to go through a radical intermediate. While this system in not physically meaningful, it supports the idea that the scaling shown in main text Figure 1 is indeed universal if methane activation proceeds through a radical intermediate, irrespective of catalyst material.

Ofcc-Promoted Unconstrained Metals

On a 111 surface, oxygen was adsorbed at an fcc or on-top site. On noble metals, such as Au and Ag, the CH3-surface interaction is so weak that the most stable methane activation pathway is through a radical intermediate. Therefore, no geometric constraints were employed in these transition state calculations.

O-Promoted Rutile Oxides

As shown in Figure S1(b), oxygen adsorbed at 110 cus sites activates methane through a radical intermediate. This is significant because the surfaces of active catalyst materials such as RuO2 and IrO2 have oxygen coverages close to unity at standard temperatures and pressures. Therefore, calculations of non-oxygen promoted methane activation for these catalysts may be less relevant.

Table S2. Hydrogen affinity, formation, and transition state energies during C-H bond activation process. All energies were referenced to H2O, O2, and CH4 in the gas phase. If energies are not provided, they were not calculated for the given catalyst.

Class Catalyst Ef (active site)

EH ETS

(CH4) ETS

(C2H6) ETS

(CH3OH) AEI-M-O-Co Co -1.80 -0.40 0.8

AEI-M-O-Cu Cu 0.01 -1.11 0.22

AEI-M-O-Fe Fe -2.16 -0.64 0.61

AEI-M-O-Ni Ni -1.01 -0.71 0.55

AFX-M-O-Co Co -1.80 -0.31 0.78

AFX-M-O-Cu Cu 0.07 -0.97 0.21

AFX-M-O-Fe Fe -2.16 -0.56 0.68

AFX-M-O-Ni Ni -0.98 -0.66 0.54

Alkali Doped CaO Li-doped CaO -1.68 0.08

Alkali Doped MgO Li-doped MgO -2.59 -1.51 0.15

Alkali Doped MgO Na-doped MgO -1.51 0.44 BO2 Terminated Perovskite

(100) SrIrO3 -0.66 -0.46 0.69

BO2 Terminated Perovskite (100)

BaMoO3 -3.24 1.03

BO2 Terminated Perovskite (100) BaNbO3 -4.35 1.44

6


BaVO3 -2.64 0.90


KMoO3 -2.59 0.76


KWO3 -4.01 1.21

BO2 Terminated Perovskite (100) NaWO3 -4.18 1.25


SrNbO3 -4.57 1.53


SrVO3 -2.62 0.88


NaSiO3 -2.42 -1.42


ZrAlO3 -3.00 -2.70


KNbO3 -0.68 -0.76 0.86


NaNbO3 -0.74 -0.67 0.93

BO2 Terminated Perovskite (100) LaTiO3 -4.41 1.38


NaTaO3 -1.17 -0.94


KTaO3 -1.18 -0.89


CaTaO2N -0.41 -0.87 0.63

BO2 Terminated Perovskite (100) NaNbO3 -3.02 -0.81 0.58


KNbO3 -2.79 -0.92 0.48

CaO CaO 1.13 -2.11 -0.37

CHA-M-O Fe -2.29 -0.43 0.84 0.51 0.31

CHA-M-O Co -1.74 -0.30 0.79 0.46 0.26

CHA-M-O Ni -0.95 -0.71 0.52 0.18 -0.01

CHA-M-O Cu 0.09 -1.07 0.20 -0.13 -0.32

CHA-M-O-2Al Fe -0.83 -0.54 0.57

CHA-M-O-2Al Co -0.60 -0.60 0.54

CHA-M-O-2Al Ni -0.07 -0.83 0.27

CHA-M-O-2Al Cu 0.94 -1.29 0.02

CHA-M-OH Fe -2.73 1.40 2.25

CHA-M-OH Co -2.04 0.75 1.65 1.33 1.12

CHA-M-OH Ni -1.66 0.45 1.27 0.92 0.64

CHA-M-OH Cu -0.98 -0.55 0.79 0.44 0.19

CHA-M-OH-2Al Fe -1.36 -0.36 0.61

CHA-M-OH-2Al Co -1.21 -0.49 0.56

7

CHA-M-OH-2Al Ni -0.89 -0.75 0.27

CHA-M-OH-2Al Cu -0.35 -1.37 0.02

GN_N4-O Fe -1.13 0.18 1.23

GN_N4-O Co 0.11 -0.69 0.59

GN_N4-O Ni 1.47 -1.00 0.38

GN_N4-O Cu 1.49 -1.17 0.24

GN_N4-OH Fe -0.95 0.72 1.65 1.50 1.25

GN_N4-OH Co -0.58 0.40 1.37 1.22 0.98

GN_N4-OH Ni 0.47 -0.56 0.80 0.66 0.44

GN_N4-OH Cu 0.32 -0.44 0.82 0.71 0.48

M-O-SAPO-34 Fe -2.35 -0.52 0.58 0.23 0.02

M-O-SAPO-34 Co -1.98 -0.33 0.73 0.41 0.22

M-O-SAPO-34 Ni -1.17 -0.66 0.46 0.11 -0.09

M-O-SAPO-34 Cu -0.19 -1.05 0.17 -0.16 -0.32

MAZ-M-O-Co Co -1.54 -0.27 0.66

MAZ-M-O-Cu Cu 0.43 -1.05 0.06

MAZ-M-O-Fe Fe -2.13 -0.36 0.7

MAZ-M-O-Ni Ni -0.77 -0.63 0.34

MgO MgO(211 step) -2.09 -0.94 0.60

MgO O-promoted MgO(110)

1.26 -2.04 -0.43

MOF-74 Fe -1.06 -0.15 1.30

MOF-74 Co -0.43 -0.22 1.00

MOF-74 Ni 0.55 -0.94 0.50

MOF-74 Cu 1.33 -1.16 0.23

MOR-bis-S-oxo Fe -5.34 0.16 1.24

MOR-bis-S-oxo Co -4.42 0.00 1.06

MOR-bis-S-oxo Ni -2.79 -0.52 0.52

MOR-bis-S-oxo Cu -0.17 -0.84 0.18

MOR-bis-S-oxo CuNi -1.92 -0.55 0.29

MOR-M-O-Co Co -1.63 -0.31 0.74

MOR-M-O-Cu Cu 0.18 -1.01 0.17

MOR-M-O-Fe Fe -2.09 -0.62 0.57

MOR-M-O-M Fe -4.46 0.91 1.69 1.39

MOR-M-O-M Co -4.00 0.50 1.39 1.03 0.86

MOR-M-O-M Ni -2.80 0.11 1.15 0.74 0.58

MOR-M-O-M Cu -1.53 -0.42 0.71 0.21 -0.03

MOR-M-O-Ni Ni -0.94 -0.66 0.51

MOR-M3O3 Ni -1.46 -0.12 0.85

8

MOR-M3O3 Cu -0.26 -0.80 0.37

Nanoparticles Ag -1.21 -0.31 0.93

Nanoparticles Au -0.68 -0.25 0.93 Ofcc-Promoted Constrained

Metals Cu -1.50 0.37 1.35

Ofcc-Promoted Constrained Metals

Pt -0.87 0.60 1.57

Ofcc-Promoted Constrained Metals Rh -1.83 0.83 2.24

Ofcc-Promoted Constrained Metals

Ir -1.55 1.01 2.04


Ag -0.28 -0.41 0.85


Au 0.25 -0.19 0.77

Oot-promoted Constrained Metals

Ag 1.09 -1.23 0.34


Cu 0.21 -0.73 0.76


Pd 0.39 -0.72 0.80


Pt 0.40 -0.73 1.04


Rh -0.50 -0.21 1.35


Ir -0.62 -0.07 1.82

Oxygen-promoted rutile oxides (110)

IrO2 -0.70 -0.79 0.33 0.08 -0.40

Oxygen-promoted rutile oxides (110) MoO2 -3.18 0.98 1.83 1.56 1.24


PtO2 0.96 -1.78 -0.52 -0.75 -1.11


RhO2 0.45 -1.05 0.14 -0.09 -0.49


RuO2 -0.80 -0.09 0.95

Oxygen-promoted rutile oxides (110) IrO2(bridge) -2.36 -0.27 0.94


OsO2 -2.39 0.36


IrO2 (strain=95%) -0.71 -0.81 0.39


IrO2 (strain=105%)

-0.68 -0.73 0.45

Oxygen-promoted rutile oxides (110) Mo in TiO2 -2.98 0.92


GeO2 2.33 -2.16


TiO2 1.54 -1.61

9

3. Supplementary Figures

Figure S1. Example transition state geometries for materials that proceed via (a) a surface-stabilized CH3 intermediate (Rh 111) and (b) a radical intermediate (oxygen-promoted IrO2)

Figure S2. Carbon p-projected DOS of the TS for (a) Cu-O/CHA, (b) Cu-OH/CHA and Cu-O/(Mg) MOF-74 active sites. The black arrows in the PDOS indicate the unoccupied bands above the Fermi level for the down-spin channel. The radical nature of CH3 is clearly seen by the unpaired electron density in

the up-spin channel.

(a)!! (b)!!

(a)$ (b)$ (c)$

Cu)O/CHA$ Cu)OH/CHA$ Cu)O/(Mg)MOF)74$

10

Figure S3. Carbon p-projected DOS for TS for (a) oxygen-promoted (b) and non-oxygen promoted IrO2 illustrating the differences in the electronic structure of the radical-like and metal-like transition states. The radical nature of CH3 is clearly seen by the unpaired electron density in the up-spin channel and empty bands above the Fermi-level (black arrows). For the metal-like TS in (b), no unpaired electron

density is observed.

(a)$ (b)$

Oxygen$promoted$IrO2$(Radical8like$transi;on$state)$

Non8oxygen$promoted$IrO2$(Metal8like$transi;on$state)$

11

Figure S4. Spin-density difference for TS corresponding to (a) Cu-O/CHA, (b) Cu-OH/CHA and (c) Cu-O/MOF-74 active site showing the radical nature of the TS. The unpaired electron density in all the cases is delocalized over the methyl carbon, active oxygen and the metal atom. The iso surfaces at drawn at 0.05

eV/Å3.

Figure S5. Fraction of available active sites as a function of active site formation energy (as given in SI Eq. 2). Here, molecular oxygen is being used to create an MO active site at 450°C. Gf corresponds to the

Gf given in SI Eq. 5.

(a)$ (b)$ (c)$

12

Figure S6: Methane activation volcano for M-OH active sites. H2O/O2 activation is assumed to

equilibrate at 450 °C and methane activation takes place at 150 °C.

13

Figure S7: Volcano Analysis for using O2 as an oxidant for temperature (a) 150 °C (b) 450 °C and (c) 600 °C with methane activation occurring at 150 °C. (d) N2O used as an oxidizing agent at 450 °C.

Changing activation temperature has the effect of broadening the volcano and shifting the peak to more strongly oxygen binding materials. Using N2O instead of O2 has the effect of increasing activity for the

materials that bind oxygen more weakly, effectively shifting the peak to more weakly binding materials.

14

Figure S8. (a) Scaling of $I vs $% for all materials (b) Same data separated by ability of the

catalyst to delocalize charge

At first glance, a relatively poor scaling between $I and $% is observed (Figure S7(a)). However,

when the materials are divided based on their ability to delocalize charge upon oxygen adsorption, we

find that these two groups exhibit better scaling, with MAEs on the order of DFT error (Figure S7(b)).

We divide materials based on a Bader charge analysis in which charges per atom in the catalyst after

active site formation (oxygen adsorption) are plotted as a function of the charges before active site

formation. On materials that delocalize charge well, these points should form a straight line with a slope

of nearly one, as the additional charge introduced to the system has been shared among many catalyst

atoms. However, the adsorption site in materials that delocalize charge poorly deviates significantly from

the line. Bader charge analysis (Figures S8-S10) indicates that materials that exhibit higher conductivity

(metals, GN, MOF-74, metallic oxides and perovskites) tend to delocalize charge better than more

insulating substrates (zeolites, insulating oxides). We note that materials better able to delocalize charge

(MOF-74, decorated graphenes, and noble metals) are not likely to be the best candidates for methane

activation, as negative motif formation energies (higher number of active sites) and negative hydrogen

abstraction energies (lower C-H activation) are desirable.

15

Figure S9. Change in atomic charges before and after oxygen adsorption on Chabazite (a-d) and

(Mg)/MOF-74 (e-h) for Fe, Co, Ni and Cu variations. The atomic charges are calculated using the DDEC method3 as it has been previously shown to give chemically meaningful charges for zeolites.4 The dotted

red lines indicate the charge on the active O atom of the M-O active site motif. For M-O/CHA, adsorption of O atom induces significant changes in the charge of the underlying metal cation (black arrow). In

contrast, O adsorption does not change the cation charges significantly, and the positive charge is delocalized over the large framework. Differences in the behavior of the active oxygen charge (red dotted line) are also seen for CHA (changes from -0.61 to -0.29) and MOF-74 (remains essentially constant at ~

-0.5).

(a)$

(b)$

(c)$

(d)$

(e)$

(f)$

(g)$

(h)$

CHA$

CHA$

CHA$

CHA$

MOF(74$

MOF(74$

MOF(74$

MOF(74$$

16

Figure S10. Change in atomic Bader charge5 before and after oxygen adsorption on (a) IrO2 and (b)

MoO2 surfaces. The dotted red lines indicate the charge on the active O atom of the M-O active site motif. For IrO2, adsorption of O atom induces significant changes in the charge of the underlying metal cation (black arrow, ~ 0.45 e). In contrast, O adsorption does not change the cation charges significantly for

MoO2 ( ~ 0.1 e), and the positive charge is delocalized over the surface.

(a)$ (b)$

17

Figure S11. Change in DDEC atomic charges before and after oxygen adsorption on 3d, 4d and 5d transition metals for M-O active site in

Chabazite. For 3d metals, the positive charge induced by oxygen adsorption significantly changes the charge of the underlying cation. However, for 4d and 5d cations, the change in atomic charges is significantly smaller. This may be due to better charge delocalization and smaller charge

transfer for the larger cations.

18

References 1. Verma, P.; Vogiatzis, K. D.; Planas, N.; Borycz, J.; Xiao, D. J.; Long, J. R.; Gagliardi, L.; Truhlar, D. G. Mechanism of Oxidation of Ethane to Ethanol at Iron(IV)-Oxo Sites in Magnesium-Diluted Fe-2(dobdc). Journal of the American Chemical Society 2015, 137 (17), 5770-5781. 2. Deng, D.; Chen, X.; Yu, L.; Wu, X.; Liu, Q.; Liu, Y.; Yang, H.; Tian, H.; Hu, Y.; Du, P.; Si, R.; Wang, J.; Cui, X.; Li, H.; Xiao, J.; Xu, T.; Deng, J.; Yang, F.; Duchesne, P. N.; Zhang, P.; Zhou, J.; Sun, L.; Li, J.; Pan, X.; Bao, X. A single iron site confined in a graphene matrix for the catalytic oxidation of benzene at room temperature. Science Advances 2015, 1 (11). 3. Manz, T. A.; Sholl, D. S. Chemically Meaningful Atomic Charges That Reproduce the Electrostatic Potential in Periodic and Nonperiodic Materials. Journal of Chemical Theory and Computation 2010, 6 (8), 2455-2468. 4. Fang, H.; Demir, H.; Kamakoti, P.; Sholl, D. S. Recent developments in first-principles force fields for molecules in nanoporous materials. Journal of Materials Chemistry A 2014, 2 (2), 274-291. 5. Tang, W.; Sanville, E.; Henkelman, G. A grid-based Bader analysis algorithm without lattice bias. Journal of Physics: Condensed Matter 2009, 21 (8), 084204.

Documents

Understanding trends in C–H bond activation … · Understanding Trends in C-H Bond Activation in Heterogeneous Catalysis Allegra A. Latimer,1,