supplementary information for - nature research€¦ · i o ed. supplementary information for...

In the format provided by the authors and unedited.

Supplementary Information for

Ultrathin metal-organic framework nanosheets for electrocatalytic oxygen evolution

Shenlong Zhao 1, 2, 4, Yun Wang 3, Juncai Dong 5, Chun-Ting He 6, Huajie Yin 1, 3, Pengfei An

5, Kun Zhao 1, Xiaofei Zhang 4, Chao Gao 1, Lijuan Zhang 1, 2, Jiawei Lv 1, Jinxin Wang 1, 2,

Jianqi Zhang 1, Abdul Muqsit Khattak 1, Niaz Ali Khan 1, Zhixiang Wei 1, Jing Zhang 5,

Shaoqin Liu 2*, Huijun Zhao 3* & Zhiyong Tang 1*

*Corresponding author E-mail: [email protected], [email protected], [email protected] (Z.T)

Contents

1. Supplementary Methods

2. Supplementary Figures 1-41

3. Supplementary Tables 1-5

4. Supplementary References 1-27

© 2016 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

SUPPLEMENTARY INFORMATIONARTICLE NUMBER: 16184 | DOI: 10.1038/NENERGY.2016.184

NATURE ENERGY | www.nature.com/natureenergy 1

http://dx.doi.org/10.1038/nenergy.2016.184

Supplementary Methods

EXAFS curve fitting details

Curve fitting is performed with Artemis software (Ref. 42) based on the EXAFS equation,

which is expressed in terms of single and multiple-scattering expansion:

2 2 20 2

( , , )( ) exp( 2 / ( ) 2 )sin(2 ( ) 2 ( ))

effc

N F k Rk S R k k kR k k

kR

(1)

where 02 ( )ek m E E represents a scale conversion from the photo energy (E, eV) to the

wave number (k, Å-1) of the excited photoelectron as measured from absorption threshold E0.

The sums is over a series of equivalent scattering paths, Γ, which originate at the central

absorption atoms, travelling to one or more of the neighboring atoms, and then back to the

original central atoms. The equivalent scattering paths, with a degeneracy of NΓ, are grouped

according to the atomic number of the passed atoms and the total path length RΓ of the

photoelectron. The dependence of the EXAFS oscillatory structure on path length and energy

is reflected by the sin(2 ( ) 2 ( ))ckR k k term, where ( )k is the effective scattering

phase shift for path Γ. ( , , )effF k R denotes the effective scattering amplitude for path Γ.

The amplitude decay due to inelastic scattering is captured by the exponential term

exp( 2 / ( ))R k , where λ(k) is the photoelectron mean free path. Additional broadening effect

due to thermal and structural disorder in absorber-scatterer(s) path lengths is accounted by the

Debye-Waller term 2 2exp( 2 )k

. S02 is a many-body amplitude-reduction factor due to

excitation in response to the creation of the core hole. In this work, while the scattering

amplitudes and phase shifts for all paths as well as the photoelectron mean free path are

theoretically calculated by ab initio code FEFF 9.05 (Ref. 41), the variable parametres that are

determined by using the EXAFS equation to fit the experimental data are NΓ, RΓ, and σΓ2. The

S02 parametre is determined in the fit of NiO and CoO standards, and used as fixed value in

the rest of the EXAFS models. All fits are performed in the R space with k-weight of 3. The

EXAFS R-factor (Rf) that measures the percentage misfit of the theory to the data is used to

evaluate the goodness of the fit.

For the EXAFS curve fitting of Ni (Co) K-edge data, below three cases are considered:



SUPPLEMENTARY INFORMATIONDOI: 10.1038/NENERGY.2016.184


bulk NiCo-MOFs, NiCo-UMOFNs and Ni-UMOFNs (Co-UMOFNs). All the data are

modeled with a radius of 4.0 Å cluster. Specifically, the first six most important paths,

including five back-scattering paths and one double-scattering path, are used, as schematically

demonstrated in Supplementary Fig. 25. It should be noted that while there are two (four) sets

of non-equivalent sites of 3 and 4 (1, 2, 3, and 4) of the metal atoms in bulk NiCo-MOFs

(NiCo-UMOFNs), they can be uniformly described by the six paths. In another word, the

EXAFS measures the average local radical distribution of atoms around the central absorption

metal atoms. The best-fit results are shown in Fig. 4a and Supplementary Figs 26 and 27. The

fitting parametres are listed in Supplementary Tables 2 and 3. The results of NiO and CoO

standards and the DFT optimized bond distances for bulk NiCo-MOFs are added for

comparison. It is obvious that the apparent bond distances and Debye-Waller factors for the

Ni-O/Co-O and Ni-Ni/Co-Co paths in NiO/CoO reference are consistent with the reported

values1, validating our multiple-shell fitting method. Also, the apparent bond distances for the

first six paths around Ni and Co atoms in bulk NiCo-MOFs agree well with the predicted

values by DFT optimization of the crystal structure determined from PXRD refinement

(Supplementary Tables 2 and 3). We also notice that the Debye-Waller factors for the Ni-O1

(0.009 Å-2) and Ni-Ni (0.014 Å-2) paths in bulk NiCo-MOFs are considerably larger than

those of NiO standard (0.005 Å-2 for Ni-O path and 0.006 Å-2 for Ni-Ni path). The

Debye-Waller factor is highly correlated with the coordination number, so the variability of

the Debye-Waller factor should be considered carefully during the fitting process. It is well

established that the Debye-Waller factor results from the thermal disorder and static disorder.

Through close examination on the Ni-O bond distribution, we can find that the NiO crystal

adopts ideal NiO6 octahedra with six identical Ni-O bond distances (2.08 Å), indicating that

the Debye-Waller factor value for the Ni-O path mainly originates from the thermal vibration.

That is, Ni-O bond distances in NiO crystal just can be rationally considered as a lower limit

for distribution of all the Ni-O bond distances. As comparison, according to DFT/XRD

prediction on our system (Supplementary Table 2), the NiO6 octahedra in bulk NiCo-MOFs is

featured by a rather distorted NiO6 octahedra adopting 2+2+2 Ni-O bond-distance distribution,

suggesting that the static disorder of bulk NiCo-MOFs is rather strong and thus leads to a

larger Debye-Waller factor of the Ni-O bond distance than that of NiO. Moreover, the





distorted NiO6 octahedra can usually make the Debye-Waller factor value larger2. For

example, the distorted NiO6 octahedra in NaNiO2 are so severe that σ2Ni-O value is as high as

0.015 Å-2. For our system, because MOFs are more flexible than the NiO standard, NiO6

octahedra distortion in our system occurs. Therefore, based on the above analysis the larger

Debye-Waller factors for both Ni-O and Co-O in bulk NiCo-MOFs are reasonable and within

allowance range, enabling us to obtain accurate structural features of NiCo-UMOFNs,

especially for the analysis on whether the Ni/Co atoms are coordinatively unsaturated.

As compared to bulk NiCo-MOFs, the apparent bond distances and the Debye-Waller

factors for all the first six paths around Ni atoms in NiCo-UMOFNs only show slight

variation, confirming consistency between the crystal structures of NiCo-UMOFNs and bulk

NiCo-MOFs. While the coordination number for the Ni-Ni path is fixed to 6.0, the

coordination numbers for the Ni-O and Ni-C paths show a systematic and obvious decreases

within the statistical uncertainty, e.g. from 6.0 to 5.7 (1) for the nearest-neighbor Ni-O1 path

and from 3.0 to 2.3 (1) for the Ni-C path, respectively. As shown in Supplementary Fig. 25,

the coordination numbers of Ni-Ni are constant, because the Ni-Ni path arises from the

intra-layer’s metal atom scattering in the (100) lattice plane and has no relationship with the

layer number of NiCo-UMOFNs. As comparison, the Ni-C path and Ni-O path are ascribed to

the C and O atoms along the [100] lattice direction, respectively, so that they strongly depend

on the layer number of NiCo-UMOFNs and the coordination status of O atoms around the Ni

atoms. It is worth noting that with similar scattering amplitudes, the coordination number of

the multiple-scattering Ni-O1-C (N2) path is closely correlated with that of the Ni-C path (N1)

(N2 = 2*N1) (Supplementary Table 2). This fact is very crucial for accurately extracting the

coordination number (N1) of the Ni-C path. To further quantitatively understand the reason

causing the decreased coordination number in NiCo-UMOFNs, a series of theoretically

structural models of NiCo-UMOFNs are constructed including different layer numbers with

completely unsaturated and completely saturated O atoms attached to the outmost two layers

(Supplementary Fig. 28), which can be considered as an appropriate approximation for the

lower and upper limits of O coordination status in NiCo-UMOFNs, respectively. The

corresponding result about the various coordination numbers is summarized in Supplementary

Fig. 28. Impressively, the coordination number of the Ni-C path, 2.3 (1), matches well with





the corresponding ideal value of 2.25 for NiCo-UMOFNs model with four metal coordination

layers within the statistical uncertainty, which is also in good agreement with the result

obtained by AFM measurement (Fig. 1d). More interestingly, one can see that the

coordination number for the Ni-O1 path is 5.7 (1), which is within the range of the ideal

values of completely unsaturated O atoms (the coordination number for the Ni-O1 path is 5.5)

and completely saturated O atoms (the coordination number for the Ni-O1 path is 6.0).

Similar phenomenon occurs for the higher nearest-neighbor Ni-O2 and Ni-O3 paths. All these

results suggest that the Ni atoms are partially unsaturated on the surface of NiCo-UMOFNs.

Analogously, Co atoms in NiCo-UMOFNs and Co-UMOFNs and Ni atoms in Ni-UMOFNs

are also analysed, showing the similar qualitative conclusion.





Supplementary Figures

Supplementary Figure 1. SEM image of NiCo-UMOFNs. The 2D morphology of

NiCo-UMOFNs can be also discerned by SEM survey (Supplementary Fig. 1). Note that due

to the resolution limit, the quality of SEM image is not as good as that of TEM image (Fig.

1a).





Supplementary Figure 2. PXRD patterns of the samples obtained at different

preparation conditions as well as the simulated diffraction pattern based on the crystal

structure. As shown in Supplementary Fig. 2, NiCo-UMOFNs (blue curve) and bulk

NiCo-MOFs (red curve) possess the same crystal phase, evidenced by comparing the PXRD

patterns of the samples.





Supplementary Figure 3. SEM image of bulk NiCo-MOFs prepared by hydrothermal

method at 140oC. Supplementary Fig. 3 shows that NiCo-UMOFNs and bulk NiCo-MOFs

not only possess the same crystal phase, but also have the similar morphology.





Supplementary Figure 4. Pawley refinement of PXRD pattern of bulk NiCo-MOFs. The

Pawley refinement result is further obtained based on the PXRD pattern of bulk NiCo-MOFs,

revealing that our product is isostructural to the previously reported Ni-based MOFs [no.

985792, Cambridge Crystallographic Data Centre (CCDC)] (Ref. 22).





Supplementary Figure 5. Crystal structure of NiCo-UMOFNs. a, Magnified crystal

structure of NiCo-UMOFNs. b, Overall crystal structure of NiCo-UMOFNs. Color scheme

for chemical representation: green for Ni, purple for Co, red for O, grey for C and white for H.

As for the ideal bulk crystal, the Co and Ni atoms are octahedrally coordinated by six O atoms

in two ways: (1) four O atoms from carboxylates and the other two from hydroxyls (right

green shadowed octahedron in Supplementary Fig. 5a); or (2) two O atoms from carboxylates

and the other four form hydroxyls (left green shadowed octahedron in Supplementary Fig. 5a).

It needs to be stressed that on the surfaces of the nanosheets, Ni or Co centres are partially

five coordinated due to edge growth restriction, and these coordinatively unsaturated metal

sites would bind solvent/reactant molecules reversibly3-5. Supplementary Fig. 5b further

demonstrates the overall crystal structure of NiCo-UMOFNs with three coordination

structural layers. Obviously, ultrathinning of the NiCo-MOFs is expected to generate rich

coordinatively unsaturated metal sites on their surfaces. Note that these unsaturated tetragonal

pyramidal metal centres exposed on the surfaces would serve as the active sites in the

electrocatalytic reactions4, 6.





Supplementary Figure 6. EDS spectrum of NiCo-UMOFNs, and the inserted table

summarizes the weight and atomic ratio of elements inside NiCo-UMOFNs.





Supplementary Figure 7. XPS spectrum of NiCo-UMOFNs, and the inserted table

summarizes the weight ratio and atomic ratio of elements inside NiCo-UMOFNs.





Supplementary Figure 8. Tridimensional BFDH morphology of NiCo-UMOFNs drew out

by Mercury software. As shown in Supplementary Fig. 8, the strongest SAXS peak in Fig. 2b

should be assigned to the largest exposed (200) planes, corresponding to the metal close

packing layers (Fig. 2b).





Supplementary Figure 9. N2 adsorption-desorption isotherm of NiCo-UMOFNs at 77K.

According to classification by IUPAC7, the N2 isotherm of NiCo-UMOFNs at 77 K

(Supplementary Fig. 9) is identified as type II with H4-type hysteresis loop. This result is also

in good agreement with the close packing crystallographic nature (without microporous

sorption) of NiCo-UMOFNs. The measured Brunauer-Emmett-Teller (BET) surface area of

NiCo-UMOFNs is 209.1 m2 g-1, which should be associated with the slit-like mesopores

formed by aggregation of nanosheets (Ref. 15).





Supplementary Figure 10. TEM images of Ni-UMOFNs and Co-UMOFNs. a,

Ni-UMOFNs. b, Co-UMOFNs. Supplementary Figs 10a and 10b are typical TEM images of

Ni-UMOFNs and Co-UMOFNs, respectively, clearly exhibiting the ultrathin layered

morphology.





Supplementary Figure 11. SAXS patterns of Co-UMOFNs, Ni-UMOFNs and

NiCo-UMOFNs. The SAXS patterns of Co-UMOFNs and Ni-UMOFNs are same as that of

NiCo-UMOFNs.





Supplementary Figure 12. Crystal structure schematic diagrams of Ni-UMOFNs and

NiCo-UMOFNs. a, Ni-UMOFNs. b, NiCo-UMOFNs. Color scheme for chemical

representation: green for Ni, purple for Co, red for O, grey for C and white for H. The

corresponding crystal structure of both Ni-UMOFNs and Co-UMOFNs is displayed in

Supplementary Fig. 12, suggesting three types of UMOFNs (Ni-UMOFNs, Co-UMOFNs and

NiCo-UMOFNs) share the same crystal structure.





Supplementary Figure 13. Molecular structures of Ni-UMOFNs and Co-UMOFNs. a,

Ni-UMOFNs. b, Co-UMOFNs. Color scheme for chemical representation: green for Ni,

purple for Co, red for O, grey for C and white for H. Supplementary Fig. 13 and Fig. 2a

indicate that UMOFNs have the well-defined structure and accurate atom arrangement, which

are crucial for mechanism study in many fields.





Supplementary Figure 14. AFM images of Ni-UMOFNs and Co-UMOFNs. a,

Ni-UMOFNs. b, Co-UMOFNs. Ni-UMOFNs and Co-UMOFNs possess uniform thickness of

3.1 nm, suggesting that all UMOFNs studied in our work contain four metal coordination

layers.





Supplementary Figure 15. N2 adsorption-desorption isotherms of Ni-UMOFNs and

Co-UMOFNs at 77K. The close packing crystallographic nature (without microporous

sorption) is also found for both Ni-UMOFNs and Co-UMOFNs. The slit-like mesopores are

also formed by aggregation of Ni-UMOFNs or Co-UMOFNs, resulting in the BET values of

180.2 m2 g-1 or 176.1 m2 g-1.





Supplementary Figure 16. CV curves of NiCo-UMOFNs in O2-saturated 1 M KOH

solution at scan rate of 5 mV s-1. To evaluate the OER activity, all the glassy carbon

electrodes modified by the samples (diametre: 5 mm, loading amount: 0.2 mg cm-2) are

measured in an O2-saturated 1 M KOH solution using a conventional three-electrode system

at a slow scan rate of 5 mV s-1 to minimize capacitive current. The black and red curves in

Supplementary Fig. 16 represent the measured cyclic voltammetry (CV) curves before and

after iR compensation, respectively.





Supplementary Figure 17. RRDE measurement of NiCo-UMOFNs under different disk

potentials in N2-saturated 1 M KOH solution. In order to confirm the bubbles are O2, the

rotating ring-disk electrode (RRDE) measurement is performed in a N2-saturated 1 M KOH

solution at a rotating speed of 1600 rpm. The glass carbon disk and Pt ring are used to load

the samples and reduce O-based intermediates, respectively. There is negligible ring current

when 0 V is applied to the disk electrode (black curve in Supplementary Fig. 17); however, a

typical linear sweep voltammetry (LSV) curve corresponding to O2 reduction appears at the

ring electrode after 1.5 V is applied to the disk electrode (red curve in Supplementary Fig. 17),

indicating generation of O2.





Supplementary Figure 18. Gas chromatography curves of NiCo-UMOFNs before and

after OER at the given overpotential of 0.3 V. The gas chromatography survey clearly

confirms that the produced bubbles are O2 (red curve in Supplementary Fig. 18).





Supplementary Figure 19. Magnified polarization curves of Ni-UMOFNs,

NiCo-UMOFNs, Co-UMOFNs and bulk NiCo-MOFs. a, Comparative potential at 0.1 mA

cm-2 current density. b, At 10 mA cm-2 current density. As displayed by the magnified

polarization curves in Supplementary Fig. 19 (the original polarization curves are shown in

Fig. 3a), the anodic current recorded on the NiCo-UMOFNs modified electrode presents a

sharp onset potential (Eonset) at 1.42 V (defined as a potential required for reaching an ORR

current density of 0.1 mA cm-2), much better than the electrodes equipped with Ni-UMOFNs,

Co-UMOFNs and bulk NiCo-MOFs of the Eonset at 1.46 V, 1.53 V and 1.48 V, respectively

(Supplementary Fig. 19a). Supplementary Fig. 19b displays that the overpotential at 10 mA

cm-2 for the NiCo-UMOFNs, Ni-UMOFNs, Co-UMOFNs, bulk NiCo-MOFs and commercial

RuO2 electrode is 250 mV, 321 mV, 371 mV, 317 mV and 279 mV, respectively. Note that

the overpotential is calculated by the measured potential at 10 mA cm-2 minus the

thermodynamic OER potential (E0H2O/O2 = 1.229 V).





Supplementary Figure 20. OER activity polarization curves and EIS of NiCo-UMOFNs.

a, Polarization curves of NiCo-UMOFNs/Cu foam and of bare Cu foam in 1 M KOH solution.

Inset is the magnified polarization curve of NiCo-UMOFNs/Cu foam. b, EIS of bare Cu foam

and glassy carbon at the open circuit voltage under same experimental condition. To exclude

the catalytic activity effect of the substrate, the bare copper foam is also tested under the same

condition (Supplementary Fig. 20a). Compared with UMOFNs/Cu foam (blue curve), the bare

copper foam shows no OER activity in the testing range (black curve), demonstrating that the

OER activity originates from the NiCo-UMOFNs sample. Impressively, Supplementary Fig.

20a reveals that the catalytic performance is further improved after NiCo-UMOFNs are

loaded on the high conductivity substrate like copper foam. The Eonset (less than 1.39 V) and

Ej=10 (less than 1.42 V) can be clearly distinguished in the inset of Supplementary Fig. 20a. To

compare the conductivity of galssy carbon (GC) and copper foam (Cu foam) substrates,

electrochemical impedance spectroscopy (EIS) measurement is performed. In Supplementary

Fig. 20b, the diametre of the semicircle with the real impedance axis (') present the charge

transfer resistance (Rct) at the electrode/electrolyte interface. The Rct of Cu foam (~ 2.5 Ω) is

much smaller than that of GC (~ 17.2 Ω) under the same experimental condition, suggesting

the favourable conductivity and charge transfer ability of Cu foam compared with GC.





Supplementary Figure 21. OER polarization curves of Ni-UMOFNs/Cu foam,

NiCo-UMOFNs/Cu foam and Co-UMOFNs/Cu foam in 1 M KOH solution. a, Normal

polarization curves. b, Magnified polarization curves. Ni-UMOFNs and Co-UMOFNs are

also loaded on the Cu foam, and tested under the same experimental condition

(Supplementary Fig. 21). Analogously to the NiCo-UMOFNs/Cu foam electrode,

Ni-UMOFNs/Cu foam and Co-UMOFNs/Cu foam electrodes (the potential at 10 mA cm-2 is

1.467 V and 1.532 V for Ni-UMOFNs and Co-UMOFNs, respectively) exhibit obviously

increased OER performance, compared with that of Ni-UMOFNs/GC and Co-UMOFNs/GC

(the potential at 10 mA cm-2 is 1.550 V and 1.600 V for Ni-UMOFNs and Co-UMOFNs,

respectively). It should be noted that the activity order of the three samples loaded on Cu

foam is consistent with that of the three samples loaded on GC (NiCo-UMOFNs

Ni-UMOFNs Co-UMOFNs), which again suggests that the improved OER activity of

UMOFNs/Cu foam should be attributed to the enhanced conductivity of the substrate.





Supplementary Figure 22. Koutecky-Levich plots from NiCo-UMOFNs, Ni-UMOFNs,

Co-UMOFNs and bulk NiCo-MOFs electrocatalyzed OER.





Supplementary Figure 23. Configuration and working mechanism of rotating ring-disk

electrode (PINE). Supplementary Fig. 23 is a photo of the PINE measurement device. The

inset is the enlarged drawing of the RRDE. The RRDE contains the ring and disk parts: the

ring part is made of platium for collecting the intermediates or products, and the disk part is

made of glass carbon for supporting catalysts. Moreover, the ring part and disk part can be

controlled independently by a software with the blue and red connectors (the pink dotted

circle), respectively.





Supplementary Figure 24. XAS spectra record to monitor X-ray radiation damage on

NiCo-UMOFNs over course of measurement. a, Ex situ Ni K-edge XAS. b, In situ Ni

K-edge XAS. c, Ex situ Co K-edge XAS. d, In situ Co K-edge XAS. The insets show the

magnified XANES region with no obvious change. Possible X-ray radiation damage on the

NiCo-UMOFNs samples is carried out by recording a series of XAS spectra under continuous

X-ray irradiation either in air (Supplementary Figs 24a and 24c) or in 1 M KOH solution

(Supplementary Figs 24b and 24d). It is evident that in both cases, the third XAS spectrum

(blue curves after 45 min irradiation) of both Ni and Co K-edge in NiCo-UMOFNs shows no

change compared with the first (black curve after 15 min irradiation) and second (red curve

after 30 min irradiation) scan spectra, indicating that the obtained XAS data are original and

not altered with X-ray irradiation.





Supplementary Figure 25. Schematic model in UMOFNs system. Schematic model of the

first six most important paths for the four sets of non-equivalent metal sites (Site 1, Site 2,

Site 3 and Site 4 are denoted as S1, S2, S3, and S4, respectively) in NiCo-UMOFNs. The M

represents the metal Ni/Co atoms; O1, O2 and O3 represent the first, second, third nearest

neighbor coordination O atoms. The small dashed circles indicate the missed C and bonded O

atoms on the surface metal atoms, and the black arrows denote X-ray scattering paths. All

sites can be uniformly described by the six paths within allowance range. Bulk NiCo-MOFs

have only two sets of non-equivalent sites (Site 3 and Site 4) for the metal atoms. Unlike bulk

NiCo-MOFs with definite coordination environment for Ni/Co atoms, the average

coordination parametres for NiCo-UMOFNs depend strongly on the layer number of metal

atom and the oxygen unsaturation attached to the surface metal atoms.





Supplementary Figure 26. Fourier-transformed magnitude in R space for NiO, CoO,

Ni-UMOFNs and Co-UMOFNs. a, NiO. b, CoO. c, Ni-UMOFNs. d, Co-UMOFNs.

Measured and calculated spectra are matched very well for all samples. The best-fit

parameters are shown in Supplementary Tables 2 and 3.





Supplementary Figure 27. K3-weighted spectra in K space for NiCo-UMOFNs,

Ni-UMOFNs, Co-UMOFNs, bulk NiCo-MOFs, NiO and CoO. a and e, NiCo-UMOFNs. b,

Ni-UMOFNs. f, Co-UMOFNs. c and g, bulk NiCo-MOFs. d, NiO. h, CoO. Measured and

calculated spectra are matched very well. The best-fit parametres are shown in Supplementary

Tables 2 and 3.





Supplementary Figure 28. Dependence between metal (Ni/Co) layer number and

coordination numbers of the six paths in NiCo-UMOFNs. a, M-O1. b, M-C. c, M-O1-C. d,

M-O2. e, M-M. f, M-O3. The values are made on a series of theoretical structural models of

NiCo-UMOFNs with increasing metal atom layer number. The complete unsaturation (dotted

red curves) and saturation (dotted black curves) of the surface metals can be considered as an

appropriate approximation for the lower and upper limit of O coordination status in

NiCo-UMOFNs, respectively. As listed in Supplementary Table 2, the coordination number of

the Ni-C path is estimated to be 2.3 (1), which matches well with the corresponding ideal

value of 2.25 for NiCo-UMOFNs model with four metal coordination layers within the

statistical uncertainty (dashed line in Supplementary Fig. 28b). Such a model of four metal

coordination layers is also in good agreement with the result obtained by AFM measurement

(Fig. 1d). Interestingly, the measured coordination number for the Ni-O1 path is 5.7 (1)

(Supplementary Table 2), which is within the range of the ideal values of completely

unsaturated O atoms (the coordination number for the Ni-O1 path is 5.5) and completely

saturated O atoms (the coordination number for the Ni-O1 path is 6.0) (Supplementary Fig.

28a). Similar phenomenon occurs for the higher nearest-neighbor Ni-O2 and Ni-O3 paths

(Supplementary Figs 28d and 28f). These results suggest that the Ni atoms on the surface of

NiCo-UMOFNs are partially unsaturated. Moreover, Co atoms in NiCo-UMOFNs and

Co-UMOFNs and the Ni atoms in Ni-UMOFNs are also analysed, showing the similar

qualitative result.





Supplementary Figure 29. Structural models of the surface metals on NiCo-UMOFNs

with different oxygen saturation status. a, Fully unsaturated (denoted as 4L-S1-5-S2-5,

where 4L, S1-5, and S2-5 stands for 4 metal coordination layers (4L), site 1 with 5

coordination number (S1-5), and site 2 with 5 coordination number (S2-5), respectively). b,

Partially unsaturated oxygen on the sites of S1 (denoted as 4L-S1-5-S2-6). c, Partially

unsaturated oxygen on the sites of S2 (denoted as 4L-S1-6-S2-5). d, fully saturated (denoted

as 4L-S1-6-S2-6) oxygen on the two sets of the non-equivalent metal sites (S1 and S2) for the

surface metal atoms.





Supplementary Figure 30. Experimental and theoretical analysis of Co K-edge XANES

spectra. Comparison of Co K-edge XANES experimental and theoretical spectra including

NiCo-UMOFNs, bulk NiCo-MOFs, and theoretical structural models shown in

Supplementary Fig. 29.





Supplementary Figure 31. In situ XAS spectra of NiCo-UMOFNs and bulk NiCo-MOFs.

a, c and e, Comparative XANES data collected on as-prepared bulk NiCo-MOFs and

NiCo-UMOFNs as well as on bulk NiCo-MOFs and NiCo-UMOFNs. b, d and f, Comparative

EXAFS data collected on as-prepared bulk NiCo-MOFs and NiCo-UMOFNs as well as on

bulk NiCo-MOFs and NiCo-UMOFNs during potentiostatic OER experiments.





Supplementary Figure 32︱Partial DOS calculations of 3d eg states in NiCo-UMOFNs

and bulk NiCo-MOFs. a and b, Ni5c and Co5c in NiCo-UMOFNs. c and d, Ni6c and Co6c

in bulk NiCo-MOFs. Notably, the unfilled eg state of Ni atom (Supplementary Fig. 32a) is less

than that of Co (Supplementary Fig. 32b). Furthermore, the unfilled eg states of

coordinately-unsaturated metals are much less than those of fully-coordinated metals

(Supplementary Figs 32a and 32b vs Supplementary Figs 32c and 32d).





Supplementary Figure 33. Schematic diagrams for atomic structure change in

NiCo-UMOFNs. a, without. b, with OOH*. The atoms within the black square represent the

adsorbed OOH on the surface. Color scheme for chemical representation: green for Ni, purple

for Co, red for O, grey for C, and white for H.





Supplementary Figure 34. Illustration of different magnetic structures of NiCo-MOFs.

The green and blue arrows indicate direction of metal cations’ major spin components. Color

scheme for chemical representation: lilac for metal, red for O, grey for C, and white for H.

Supplementary Fig. 34 summarizes possible magnetic structures in NiCo-MOFs. It is found

that the antiferromagnetic (AFM) structure is more stable than the ferromagnetic (FM)

structure. And, AFM I is the most stable configuration, showing the coupling between metal

cations through aromatic rings (Supplementary Table 4). As indicated in Supplementary

Table 5, the magnetic moments for Co and Ni ions are about 2.7 B and 1.7 B. This means

that Co2+ has the high spin states with 3 unpaired electrons. As comparison, Ni2+ has the

valence electronic configuration as 3d8, and its high spin and low spin states are same with a

theoretical value of 2 B. After formation of the intermediates on the surfaces, the AFM

structures are still kept with almost no change of magnetic moments.





Supplementary Figure 35. XPS characterization of NiCo-UMOFNs, Ni-UMOFNs and

Co-UMOFNs. a, Ni 2p3/2. b, Co 2p3/2. Supplementary Figs 35a and 35b present the

high-resolution XPS spectra of Co 2p3/2 and Ni 2p3/2 of NiCo-UMOFNs, Ni-UMOFNs and

Co-UMOFNs. The XPS peak at 855.7 eV is assigned to Ni 2p3/2 of Ni-UMOFNs (black curve

in Supplementary Fig. 35a), while the XPS peak centered at 781.7 eV is attributed to Co 2p3/2

of Co-UMOFNs (black curve in Supplementary Fig. 35b), As comparison, the XPS peaks

corresponding to Ni 2p3/2 and Co 2p3/2 of NiCo-UMOFNs are shifted to 856.5 eV and 781.0

eV, respectively, further confirming the strong interaction between Co2+ and Ni2+ in

NiCo-UMOFNs8.





Supplementary Figure 36. DFT calculations of NiCo-UMOFNs. a, Partial DOS of Ni 3d eg

states in Ni-UMOFNs (black curve) and CoNi-UMOFNs (red curve). b, Co 3d eg states in

Co-UMOFNs (black curve) and CoNi-UMOFNs (red curve). c, Electronic coupling between

Ni and Co and corresponding charge transfer process. d, Relationship between overpotential

of OER and occupancy of the eg-symmetry electron of metals. The change of PDOS of

unfilled metal 3d eg both in single metal- and bimetal-UMOFNs suggests that density of the

unoccupied eg state increases or decreases for Ni or Co after the coupling, respectively (pink

ovals in Supplementary Figs 36a and 36b). It has been reported that such coupling could

induce change of eg filling (Supplementary Fig. 36c) and further optimize their OER

performance (Supplementary Fig. 36d) (Ref. 6).





Supplementary Figure 37. Polarization curves of NiCo-UMOFNs and physical blending

samples under the same condition. The contrast experiment has been designed to

demonstrate that the electrocatalytic activity of the NiCo-UMOFNs is distinct from that of

simple blending of Ni-UMOFNs and Co-UMOFNs. As shown in Supplementary Fig. 37, the

electrocatalytic activity of the electrode equipped with the blending catalysts is much lower

than that of the NiCo-UMOFNs loaded electrode under the same testing condition.





Supplementary Figure 38. Cycling stability test of NiCo-UMOFNs electrode for 10,000

seconds. Supplementary Fig. 38 displays the current density measured by potentiostatically

cycling the voltage between 1.50 V and 1.52 V for a total of 10,000 seconds. Obviously, the

current density keeps almost unchanged, suggesting the good recoverability of the

NiCo-UMOFNs loaded electrode.





Supplementary Figure 39. Long term stability of NiCo-UMOFNs at constant

overpotential of 0.25 V and the stability comparison (inset) between NiCo-UMOFNs (red

curve) and commercial RuO2 (black curve) under the same condition. The ultralong-term

OER durability of NiCo-UMOFNs is also examined by the chronoamperometric method in an

O2-saturated 1 M KOH solution at 1600 rpm (Supplementary Fig. 39). The applied voltage is

set to 1.48 V. It is evident that the anodic current decay is ~ only 2.6 % after 200 h continuous

electrolysis. As comparison, the corresponding current loss at the RuO2 electrode under the

same condition is as high as 60.3 % after only 40,000 s electrolysis (inset of Supplementary

Fig. 39). These data highlight the good robustness and stability of the noble metal-free

NiCo-UMOFNs catalysts, which holds the great potential for practical application.





Supplementary Figure 40. PXRD patterns of NiCo-UMOFNs before and after OER

process. As shown in Supplementary Fig. 40, the crystal structure of NiCo-UMOFNs after

40,000 s electrocatalysis under a given potential of 1.48 V keeps unchanged.





Supplementary Figure 41. TEM image of NiCo-UMOFNs after OER process. The

ultrathin layered morphology is well maintained after 40,000 s electrocatalytic OER process

under a given potential of 1.48 V.





Supplementary Tables

Supplementary Table 1. Comparison of OER activity of UMOFNs (red words) and

recently reported active catalysts (blue words) in 1 M KOH solutions.

Sample Eonset (V)

Overpotential (at 2 mA cm-2)

Overpotential (at 10 mA cm-2)

Substrate Reference

NiCo-UMOFNs 1.42 235 mV 250 mV GCE This work

NiCo-UMOFNs/Cu foam

1.39 —— 189 mV Cu foam This work

LaCaCoOx ~1.46 370 mV —— GCE 9

FeCoNiOx ~1.42 ~260 mV —— GCE 10

Ni3N sheet 1.53 >520 mV —— GCE 11

FeNiOOH ~1.49 ~280 mV —— GCE 12

NiFe-LDH-CNTs 1.43 —— ~250 mV GCE 13

LiNiCo-OH ~1.52 —— 340 mV GCE 14

NiFe-LDH NS ~1.47 ~270 mV 300 mV GCE 15





Ultrathin CoMn-LDH

~1.50 ~290 mV 350 mV GCE 16

Co3O4@CoO SC ~1.52 ~290 mV 430 mV GCE 17

Echo-MWCNTs ~1.50 ~315 mV 360 mV GCE 18

CoFe(OOH) 1.53 ~420 mV —— GCE 19

NiCo2.7

OH 1.48 ~310 mV 350 mV GCE 20

Gelled-FeCoW —— —— 315 mV Au (111) 21

Gelled-FeCoW —— —— 191 mV Au foam 21

NiFe-LDH/NF 1.44 —— —— Nickel foam

22

CoxP/Cu foil 1.53 —— 345 mV Cu foil 23

NiCo LDH/CP 1.53 —— 367 mV Carbon paper

24

NixCo3-xO4 nanowire array/Ti

foil —— —— 370 mV Ti foil 25





Co3O4-C nanowires/Cu foil

1.41 —— 220 mV Cu foil 26

NiCo2O4 nanowires/Ti mesh

1.56 —— 370 mV Ti mesh 27

Note: All potentials were versus to reversible hydrogen electrode (RHE). All these works were tested under similar condition

in 1 M KOH solution. As for GCE, the catalysts were coated on the glassy carbon electrode (GCE) with Nafion binder; with

respect to other conductive substrates, the catalysts were loaded on the substrates directly.





Supplementary Table 2. Fitting parametres of Ni K-edge EXAFS curve for bulk

NiCo-MOFs, UMOFNs and NiOa.

R (Å) N

Sample Path XRD/D

FT EXAFS EXAFS σ2 (Å-2)

ΔE0 (eV)

Rf, %

NiOe Ni-O Ni-Nib

2.08 2.95

2.08 (1) 2.95 (1)

6.0c 12.0c

0.005 (1) 0.006 (1)

0.01 1.8 (3)

Bulk NiCo-MOFsf

Ni-O1 Ni-C Ni-O1-C Ni-O2 Ni-Nib Ni-O3

2.09 2.99 3.19 3.42 3.46 3.80

2.07 (3) 2.94 (1) 3.38 (3) 3.14 (2) 3.60 (2) 3.83 (4)

6.0c 3.0c 6.0c 3.0c 6.0c 5.0c

0.009 (2) 0.004 (1) 0.010 (2) 0.010 (2) 0.014 (2) 0.017 (3)

-4.7 (6) 2.8

NiCo-UMOFNsf Ni-O1 Ni-C Ni-O1-C Ni-O2 Ni-Nib Ni-O3

− − − − − −

2.06 (3) 2.94 (2) 3.37 (4) 3.11 (3) 3.60 (2) 3.82 (5)

5.7 (1) 2.3 (1)

4.6d 2.8 (1)

6.0c 4.8 (2)

0.009 (2) 0.003 (1) 0.010 (2) 0.012 (2) 0.015 (2) 0.017 (4)

-5.2 (5) 5.3

Ni-UMOFNsf Ni-O1 Ni-C Ni-O1-C Ni-O2 Ni-Ni Ni-O3

− − − − − −

2.06 (4) 2.99 (2) 3.39 (4) 3.13 (3) 3.59 (2) 3.77 (5)

5.8 (2) 2.2 (2)

4.4d 2.8 (2)

6.0c 4.8 (3)

0.010 (2) 0.003 (1) 0.009 (3) 0.004 (2) 0.019 (3) 0.021 (4)

-5.3 (9) 9.1

aO1, O2, and O3 represent the first, second, and third nearest neighbor coordination O atoms. The bond distances derived

from XRD measurement for NiO and from DFT calculation for bulk NiCo-MOFs are used for comparison. S02 is fixed to 1.0.

Values in parentheses indicate uncertainties. bBecause Ni and Co atoms are homogenously mixed in the bimetal-organic

framework and they are also indistinguishable by XAS due to their nearly identical scattering ability of photoelectron, only

Ni-Ni path is used for Ni-Ni and Ni-Co cases. cThe coordination number is fixed according to the crystal structure. dThe

coordination number of the multiple scattering path Ni-O1-C is set to be double that of the back scattering path Ni-C. eFitting

range: 3.7 ≤ k (/Å) ≤ 14.5 and 1.0 ≤ R (Å) ≤ 3.2. fFitting range: 2.3 ≤ k (/Å) ≤ 13.0 and 1.0 ≤ R (Å) ≤ 3.8.





Supplementary Table 3. Fitting parametres of Co K-edge EXAFS curve for bulk

NiCo-MOFs, UMOFNs and CoOa.

R (Å) N

Sample Path XRD/DFT

EXAFS EXAFS σ2 (Å-2) ΔE0 (eV) Rf, %

CoOe Co-O Co-Cob

2.12 3.00

2.12 (1) 3.01 (1)

6.0c 12.0c

0.007 (1) 0.008 (1)

-10.4 (8) 0.01

Bulk NiCo-MOFsf Co-O1 Co-C Co-O1-C Co-O2 Co-Cob Co-O3

2.09 2.99 3.19 3.42 3.46 3.80

2.07 (2) 2.97 (1) 3.38 (3) 3.14 (2) 3.62 (3) 3.83 (5)

6.0c 3.0c 6.0c 3.0c 6.0c 5.0c

0.012 (3) 0.004 (1) 0.010 (3) 0.006 (2) 0.015 (2) 0.019 (3)

-4.4 (6) 3.5

NiCo-UMOFNsf Co-O1 Co-C Co-O1-C Co-O2 Co-Cob Co-O3

− − − − − −

2.07 (3) 2.97 (2) 3.38 (4) 3.13 (2) 3.62 (2) 3.83 (4)

5.8 (2) 2.3 (2)

4.6d 2.8 (1)

6.0c 4.8 (2)

0.011 (2) 0.003 (1) 0.010 (3) 0.005 (2) 0.015 (3) 0.017 (3)

-4.1 (4) 7.3

Co-UMOFNsf Co-O1 Co-C Co-O1-C Co-O2 Co-Co Co-O3

− − − − − −

2.09 (3) 2.99 (1) 3.41 (4) 3.14 (2) 3.64 (2) 3.87 (4)

5.8 (2) 2.2 (3)

4.4d 2.8 (2)

6.0c 4.8 (3)

0.011 (2) 0.004 (2) 0.010 (3) 0.006 (1) 0.016 (3) 0.019 (4)

-4.6 (7) 12.1

aO1, O2, and O3 represent the first, second, third nearest neighbor coordination O atoms. The bond distances derived from

XRD measurement for CoO and from DFT calculation for bulk NiCo-MOFs are used for comparison. S02 is fixed to 1.0.

Values in parentheses indicate uncertainties. bBecause Co and Ni atoms are homogenously mixed in the bimetal-organic

framework and they are also indistinguishable by XAS due to their nearly identical scattering ability of photoelectron, only

Co-Co path is used for Co-Co and Co-Ni cases. cThe coordination number is fixed according to the crystal structure. dThe

coordination number of the multiple scattering path Co-O1-C is set to be double that of the back scattering path Co-C.

eFitting range: 3.3 ≤ k (/Å) ≤ 13.6 and 1.0 ≤ R (Å) ≤ 3.2. fFitting range: 2.3 ≤ k (/Å) ≤ 11.5 and 1.0 ≤ R (Å) ≤ 3.8.





Supplementary Table 4. Total energy differences (eV) for MOFs with various magnetic

structures. The total energy of AFM I is used as a reference.

AFM I AFM II AFM III FM

Ni-MOFs 0.000 0.002 0.001 0.003

Co-MOFs 0.000 0.039 0.083 0.190

NiCo-MOFs 0.000 0.106 0.072 0.107





Supplementary Table 5. Magnetic moments (B) of metal cations in the UMOFNs

without or with the adsorbed intermediates. Two values are given because there are two

types of metal cations with different chemical environments.

UMOFNs OH* O* OOH*

Ni-UMOFNs 1.7 1.7, 1.7 1.6, 1.6 1.7, 1.7

Co-UMOFNs 2.7 2.7, 2.7 2.6, 2.8 2.7, 2.7

Ni in NiCo-UMOFNs 1.7 1.7, 1.7 1.7, 1.7 1.7, 1.7

Co in NiCo-UMOFNs 2.7 2.7, 2.7 2.6, 2.5 2.7, 2.7





Supplementary References

1. Rinaldi-Montes, N. et al. Interplay between microstructure and magnetism in NiO

nanoparticles: breakdown of the antiferromagnetic order. Nanoscale 6, 457-465 (2014).

2. Bediako, D. K. et al. Structure-activity correlations in a nickel-borate oxygen evolution

catalyst. J. Am. Chem. Soc. 134, 6801-6809 (2012).

3. Wu, H. et al. Unusual and highly tunable missing-linker defects in zirconium metal-organic

framework UiO-66 and their important effects on gas adsorption. J. Am. Chem. Soc. 135,

10525-10532 (2013).

4. Cliffe, M. J. et al. Correlated defect nanoregions in a metal-organic framework. Nat.

Commun. 5, 4176 (2014).

5. Vermoortele, F. et al. Synthesis modulation as a tool to increase the catalytic activity of

metal-organic frameworks: the unique case of UiO-66(Zr). J. Am. Chem. Soc. 135,

11465-11468 (2013).

6. Xie, J. F. et al. Defect-rich MoS2 ultrathin nanosheets with additional active edge sites for

enhanced electrocatalytic hydrogen evolution. Adv. Mater. 25, 5807-5813 (2013).

7. Sing, K. S. W. Ed. Reporting physisorption data for gas/solid systems with special

reference to the determination of surface area and porosity (provisional). Pure Appl. Chem. 54,

2201-2218 (1982).

8. Smith, R. D. L. Prevot, M. S. Fagan, R. D. Trudel, S. & Berlinguette, C. P. Water oxidation

catalysis: electrocatalytic response to metal stoichiometry in amorphous metal oxide films

containing iron, cobalt, and nickel. J. Am. Chem. Soc. 135, 11580-11586 (2013).

9. Suntivich, J. May, K. J. Gasteiger, H. A. Goodenough, J. B. & Shao-Horn, Y. A perovskite

oxide optimized for oxygen evolution catalysis from molecular orbital principles. Science 334,

1383-1385 (2011).

10. Smith, R. D. L. et al. Photochemical route for accessing amorphous metal oxide materials

for water oxidation catalysis. Science 340, 60-63 (2013).

11. Kun, K. et al. Metallic nickel nitride nanosheets realizing enhanced electrochemical water

oxidation. J. Am. Chem. Soc. 137, 4119-4125 (2015).

12. Trotochaud, L. Young, S. L. Ranney, J. K. & Boettcher, S. W. Nickel-iron oxyhydroxide

oxygen-evolution electrocatalysts: the role of intentional and incidental iron incorporation. J.





Am. Chem. Soc. 136, 6744-6753 (2014).

13. Gong, M. et al. An advanced Ni-Fe layered double hydroxide electrocatalyst for water

oxidation. J. Am. Chem. Soc. 135, 8452-8455 (2013).

14. Niu, K-Y. et al. Tuning complex transition metal hydroxide nanostructures as active

catalysts for water oxidation by a laser-chemical route. Nano Lett. 15, 2498-2503 (2015).

15. Song, F. & Hu, X. L. Exfoliation of layered double hydroxides for enhanced oxygen

evolution catalysis. Nat. Commun. 5, 4477 (2014).

16. Song, F. & Hu, X. L. Ultrathin cobalt-manganese layered double hydroxide is an efficient

oxygen evolution catalyst. J. Am. Chem. Soc. 136, 16481-16484 (2014).

17. Tung, C-W. et al. Reversible adapting layer produces robust single-crystal electrocatalyst

for oxygen evolution. Nat. Commun. 6, 8106 (2015).

18. Lu, X. Y. Yim, W-L. Suryanto, B. H. R. & Zhao, C. Electrocatalytic oxygen evolution at

surface-oxidized multiwall carbon nanotubes. J. Am. Chem. Soc. 137, 2901-2907 (2015).

19. Burke, M. S. Kast, M. G. Trotochaud, L. Smith, A. M. & Boettcher, S. W. Cobalt-iron

(oxy)hydroxide oxygen evolution electrocatalysts: the role of structure and composition on

activity, stability, and mechanism. J. Am. Chem. Soc. 137, 3638-3648 (2015).

20. Nai, J. W. et al. Efficient electrocatalytic water oxidation by using amorphous Ni-Co

double hydroxides nanocages. Adv. Energy Mater. 5, 1401880 (2015).

21. Zhang, B. et al. Homogeneously dispersed multimetal oxygen-evolving catalysts. Science

352, 333-337 (2016).

22. Lu, X. Y. & Zhao, C. Electrodeposition of hierarchically structured three-dimensional

nickel-iron electrodes for efficient oxygen evolution at high current densities. Nat. Commun. 6,

6616 (2015).

23. Jiang, N. You, B. Sheng, M. L. & Sun, Y. J. Electrodeposited cobalt-phosphorous-derived

films as competent bifunctional catalysts for overall water splitting. Angew. Chem. Int. Ed. 54,

6251-6254 (2015).

24. Liang, H. F. et al. Hydrothermal continuous flow synthesis and exfoliation of NiCo

layered double hydroxide nanosheets for enhanced oxygen evolution catalysis. Nano Lett. 15,

1421-1427 (2015).

25. Li, Y. G. Hasin, P. & Wu Y. Y. NixCo3-xO4 Nanowire arrays for electrocatalytic oxygen





evolution. Adv. Mater. 22, 1926-1929 (2010).

26. Ma, T. Y. Dai, S. Jaroniec, M. & Qiao, S. Z. Metal-organic framework derived hybrid

Co3O4-carbon porous nanowire arrays as reversible oxygen evolution electrodes. J. Am. Chem.

Soc. 136, 13925-13931 (2014).

27. Peng, Z. Jia, D. S. Al-Enizi, A. M. Elzatahry, A. A. & Zheng, G. F. From water oxidation

to reduction: homologous Ni-Co based nanowires as complementary water splitting

electrocatalysts. Adv. Energy Mater. 5, 1402031 (2015).





supplementary information for - nature research€¦ · i o ed. supplementary information for...

Documents