Atomic-level energy storage mechanism of cobalt hydroxide electrode for pseudocapacitors

Abstract

Cobalt hydroxide is a promising electrode material for supercapacitors due to the high capacitance and long cyclability. However, the energy storage/conversion mechanism of cobalt hydroxide is still vague at the atomic level. Here we shed light on how cobalt hydroxide functions as a supercapacitor electrode at operando conditions. We find that the high specific capacitance and long cycling life of cobalt hydroxide involve a complete modification of the electrode morphology, which is usually believed to be unfavourable but in fact has little influence on the performance. The conversion during the charge/discharge process is free of any massive structural evolution, but with some tiny shuffling or adjustments of atom/ion species. The results not only unravel that the potential of supercapacitors could heavily rely on the underlying structural similarities of switching phases but also pave the way for future material design for supercapacitors, batteries and hybrid devices.

ARTICLE

Received 27 Jul 2016 | Accepted 9 Mar 2017 | Published 8 May 2017

Ting Deng1,2,3, Wei Zhang1,2,3,4,5, Oier Arcelus4, Jin-Gyu Kim6, Javier Carrasco4, Seung Jo Yoo6,

Weitao Zheng1,2,3, Jiafu Wang1,2,3, Hongwei Tian1,2,3, Hengbin Zhang1,2,3, Xiaoqiang Cui1,2,3 & Telo Rojo4,7

Cobalt hydroxide is a promising electrode material for supercapacitors due to the high capacitance and long cyclability. However, the energy storage/conversion mechanism of cobalt hydroxide is still vague at the atomic level. Here we shed light on how cobalt hydroxide functions as a supercapacitor electrode at operando conditions. We nd that the high specic capacitance and long cycling life of cobalt hydroxide involve a complete modication of the electrode morphology, which is usually believed to be unfavourable but in fact has little inuence on the performance. The conversion during the charge/discharge process is free of any massive structural evolution, but with some tiny shufing or adjustments of atom/ion species. The results not only unravel that the potential of supercapacitors could heavily rely on the underlying structural similarities of switching phases but also pave the way for future material design for supercapacitors, batteries and hybrid devices.

DOI: 10.1038/ncomms15194 OPEN

Atomic-level energy storage mechanism of cobalt hydroxide electrode for pseudocapacitors

1 Department of Materials Science, Jilin University, Changchun 130012, China. 2 Key Laboratory of Mobile Materials MOE, Jilin University, Changchun 130012, China. 3 State Key Laboratory of Automotive Simulation and Control, Jilin University, Changchun 130025, China. 4 CIC Energigune, 01510 Minano, Spain.

5 Ikerbasque, Basque Foundation for Science, 48013 Bilbao, Spain. 6 Department of Electron Microscopy Research, Korea Basic Science Institute, Daejeon 34133, South Korea. 7 Departamento de Qumica Inorgnica, Universidad del Pas Vasco, UPV/EHU, 48080 Bilbao, Spain. Correspondence and requests for materials should be addressed to W.Z. (email: mailto:[email protected]

Web End [email protected] ) or to W.T.Z. (email: mailto:[email protected]

Web End [email protected] ).

NATURE COMMUNICATIONS | 8:15194 | DOI: 10.1038/ncomms15194 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 1

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15194

Pseudocapacitance is the Faradaic charge transfer between the electrolyte and the (sub)surface of a suitable metal oxide/hydroxide electrode, involving reversible redox

reactions, electroadsorption or intercalation processes1. Redox reaction is the main property in the functionality of supercapacitors (SCs)2,3, but it also applies to a variety of other energy storage devices such as fuel cells4,5 and rechargeable batteries6,7. Irreversible morphological changes during electrochemical cycles cause capacitance and cyclability degradations. So controlling morphological stability, therefore, is the common motif to unlock the potential of novel energy storage and conversion SC systems. Theoretically, in nickelmetal hydride batteries, the charge/discharge mechanism is based on the motion of H species, the oxidation and reduction reactions can be interpreted as hydrogen ions switching through:

Ni(OH)2 OH -NiOOH H2O e (refs 6,8,9). For a

hydroxide-based supercapacitive electrode (for example, Co(OH)2), with high theoretical capacitance and long cycling life, it is well accepted that the charge/discharge mechanism is Co(OH)2 OH -CoOOH H2O e (refs 10,11). Both

Co(OH)2 and CoOOH have the same layered cobalt structure and the large interlayer spacing that facilitates the intercalation of H species. Ni(OH)2 exploited in nickelmetal hydride batteries shows a pair of well-separated Faradaic redox peaks, with a large voltage separation (40.10.2 V) between oxidation and reduction due to a phase transition12. This indicates a ne battery behaviour according to the perspective of Simons et al.13. Co(OH)2 shows a pair of redox peaks as well, however, the separation of these two peaks is vaguer and smaller, which indicates a so-called battery-mimic mechanism. Provided with such a mechanism in a metal hydroxide electrode, the high specic pseudocapacitance and long cycling life of a SC and relevant structural transformations should be explained by a more fundamental motif rather than only morphological stability.

Here we show the energy storage/conversion mechanism of Co(OH)2 electrode, which can retain 95.7% of its initial capacitance after 8,000 cycles. Furthermore, along with in situ experimental examinations, the density functional theory (DFT) calculations provide more specic details of this mechanism, showing that the conversion during the charge/discharge procedure is free of any massive structural evolution, with only some tiny shufing or adjustments of atom/ion species instead. According to our results, a battery-mimic mechanism at the atomic level can be applicable to pseudocapacitance, which may open an avenue in the design of new SC electrodes.

Results

Structural characterization of Co(OH)2. Figure 1a shows the

X-ray diffraction pattern of the as-prepared a-Co(OH)2, and

scanning electron microscopy (SEM) and transmission electron microscopy (TEM) analysis. The peak at 26 belongs to the carbon bre paper (CFP) substrate whose intensity is much weaker than the features of a-Co(OH)2, indicating that a-Co(OH)2 is well spread on the substrate. For more accurate information, we conducted synchrotron X-ray diffraction characterization (Supplementary Fig. 1), which conrms the peaks for (100) and (110), but also reveals the existence of a peak at 9.47 (l 1.24

nm) indexed as (003). The (003) peak is more intense and sharper than other peaks, indicating that a-Co(OH)2 grows preferentially along the direction o0034. In addition, Supplementary Fig. 1 also reveals the presence of nitrates into a-Co(OH)2, which are incorporated during the synthesis. As shown in Fig. 1b,f, the active material forms an interlaced petal-like morphology, with each petal composed of several a-Co(OH)2 sheets. A variety of pore sizes inside those petals is typically

believed to increase the number of adsorbed charge species. In a magnied high-resolution TEM image, the a-Co(OH)2 (100)

planes can clearly be identied (Fig. 1g, where P stands for porous regions). Such unique morphology facilitates redox reactions by shortening ion diffusion distance and, therefore, enhancing ionic conductivity, which is expected to yield a high specic pseudocapacitance. To be noted, we chose CFP as the substrate, not only because its chemical stability and the ability to shorten the distance of electron transport14 but also CFP can provide a robust bonding with Co(OH)2. We examined the surface state of the interface before and after the electrodeposition of Co(OH)2 (Supplementary Fig. 2), and we can see that a strong interaction was yielded between the substrate and the active material. The Co(OH)2 synthesized on Ni foam (Supplementary

Fig. 3) with a higher capacitance compared to Co(OH)2 on CFP, however, shows a very poor cyclability (only 71.4% capacitance retention after 1,000 cycles). So another advantage of CFP is the strong interaction between CFP and Co(OH)2, which prevents the active material from being peeling off, thus leading to a stable platform to study the intrinsic properties of Co(OH)2.

However, it is important to notice that the synthesis of a-Co(OH)2 generally results in poor crystalline and disordered structures1517, which can rapidly transform into the more stable b-phase in alkaline media18. Therefore, before conducting electrochemical experiments, we settled the electrode in the electrolyte (KOH solution) for 15 min to make OH inltrate into the pores to fully impregnate Co(OH)2 (allow ions permeate into electrodes and make full use of active materials). After bathed in the electrolyte, we observed a concomitant, most spectacular morphological change (Fig. 1c), where the loose-stacked interlaced petals transformed into regular hexagonal platelets (the same phenomenon was also found in the residuals). This suggests that a-Co(OH)2 interacts with hydroxide like a scalpel that removes the interlayer nitrates. According to Fourier transform infrared spectra (Supplementary Fig. 4a), an intense absorption band at 1,384 cm 1 is due to NO stretching vibration of NO 3 in the case of the as-prepared Co(OH)2. The relative intensity and the integral area of the peak reduced (Supplementary Fig. 4b) after bathing in KOH for 15 min, indicating the removal of most NO 3 species. However, the

NO 3 ions in inner layers of Co(OH)2 cannot be removed immediately, resulting in the remain of the peak. The wide band at 3,473 cm 1 is assigned to the OH stretching modes of interlayer water and of H-bound OH group. After bathing, a sharp and intense band appeared at 3,629 cm 1, which is characteristic of free OH groups in b-type Co(OH)2. And other bands are due to CoO stretching and CoOH bending vibrations. As a result, the layers become more compact, with an ultimate hexagonal shape and presenting various thicknesses. Such drastic change is also reected by the X-ray diffraction pattern shown in Fig. 1a, which indicates that the actual reactant in the alkaline environment was b-Co(OH)2. In addition, our

DFT calculations indicate that the low-index hexagonal (0001) surface (Supplementary Fig. 5) is among the three lowest-energy surfaces of b-Co(OH)2 (Supplementary Table 1), which is consistent with the observed regular hexagonal platelets.

The change of specic surface area after this morphology modication was examined via N2 adsorption (N2 isotherms in

Supplementary Fig. 6). The BrunauerEmmettTeller (BET) surface area of the as-prepared Co(OH)2 was found to be56.220 m2 g 1. After the modication, BET surface area increased to 110.280 m2 g 1. The appearance of hexagonal

Co(OH)2 platelets and the corresponding activation of some pores are responsible for the specic surface area increase, which is favourable for the redox reaction. However, the limited increase of BET surface area is still too low to play a signicant role in

2 NATURE COMMUNICATIONS | 8:15194 | DOI: 10.1038/ncomms15194 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15194 ARTICLE

enhanced electrochemical performance. Indeed, we note that an increase of the BET surface area does not necessarily involve an increase of the specic capacitance, as shown by Ghodbane et al.19 These authors prepared a series of MnO2 allotropic phases with different BET surface areas, and the specic capacitance measured

for each MnO2 phase did not follow the trend of the BET surface area. The contact of electrode and electrolyte is inevitable, and the coincidence is that the species that make Co(OH)2 ready for electrochemical tests is just the electrolyte (KOH solution) itself. The management of military transformation instead of its civilian

a b c d e

CFP

CFP

2 3

5

Intensity (a.u.)

1

4

6

1 2

20 40 60

2[afii9835] (degree)

f g h i j

P

(100)

(100)C

P

(010) (010)C

(100)C

P

60o

(100)

(010)

(110)

(100)

Figure 1 | Characterization of Co(OH)2 before and after in situ XAS experiment. (a) X-ray diffraction patterns of as-prepared Co(OH)2, a1 and a2 are indexed to a-Co(OH)2 (100) and (110) (PDF, card No. 46-0605), and Co(OH)2 after bathing in 1 M KOH for 15 min, b1b6 are indexed to (001), (100), (101), (102), (110) and (111) (PDF card No. 30-0443), respectively. (be) SEM images of as-prepared Co(OH)2, Co(OH)2 after bathing in KOH and after in situ XAS. Scale bars, 2 mm, 2 mm, 2 mm and 400 nm, respectively. (fj) TEM images of Co(OH)2 and Co(OH)2 after in situ XAS. (f,h) High-resolution

TEM (HRTEM) images of the areas marked by the red circles in the insets. Scale bars, 10 nm and the inset scale bars, 200 nm. (g,i) HRTEM images that are highlighted by a white square in f,h, respectively. Scale bars, 3 nm. (j) A local selected area electron diffraction pattern of Co(OH)2 after in situ XAS. The inset scale bar, 0.5 mm.

1M KOH

a b

120 100

80

Capacitance retention (%)

CE

RE

KOH

CFP

Co(OH)2

Co(OH)2

KOH

60 40 20

Membrane

Co(OH)2+OH

CoOOH+H2O+e

0 0 2,000 4,000 6,000 8,000

Cycle number

WE

c d

0.0080.006

0.004 B

A

H

G F E

Cobalt oxide Cobalt hydroxide

7,760 7,770

Energy (eV)

Current (A)

C

D

0.002

0.002 0.004

0.2 0.1 0.1 0.2 0.3 0.4 7,710 7,720 7,730 7,740 7,750

Normalized intensity (a.u.)

0

0

Potential (V)

Figure 2 | Cycling life of Co(OH)2 and design of the XAS experiment. (a) The synthesized Co(OH)2 shows excellent cyclic stability, and the pseudocapacitance retention reaches 95.7% of its initial value after 8,000 cycles. (b) The design of the reaction cell for in situ XAS experiment. Within the cell, a three-electrode system is utilized, WE, CE and RE stand for working, counter and reference electrodes, respectively. (c) The in situ cyclic voltammetry curve of Co(OH)2 in the reaction cell and the scan rate is 0.4 mVs 1. AH labels correspond to the sampling points of XAS. (d) Comparison of XANES data collected on as-prepared Co(OH)2 electrode and CoO standard sample.

NATURE COMMUNICATIONS | 8:15194 | DOI: 10.1038/ncomms15194 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 3

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15194

counterpart at the level of the Co(OH)2 unit cell shows higher crystallinity according to the X-ray diffraction patterns, which indicates better conductivity of Co(OH)2.

Electrochemical performance evaluation. Supplementary Fig. 7 show cyclic voltammetry curves of Co(OH)2 and galvanostatic charge/discharge curve in 1 M KOH solution, respectively. The existence of a pair of redox peaks indicates a one-electron-transfer process and conrms that the pseudocapacitance behaviour is rooted in a redox reaction. During charge, Co(OH)2 is oxidized to CoOOH, whereas in discharge the reaction is reversed. To be noted, there is a second redox process that has been discussed in many works: CoOOH OH -CoO2

H2O e . The second process is related to many factors, such as

specic surface area10,20. The interlayer distance between adjacent Co(OH)2 single sheets can also inuence the redox reaction, which occurs on the interface between electrolyte and active materials21. In our work, in addition to the low specic surface area of Co(OH)2 we synthesized, the decreased interlayer spacing due to the loss of NO 3 species hinders the complete penetration of the electrolyte. Thus, the oxidation process to

CoOOH is expected to dominate. According to X-ray absorption spectroscopy (XAS) sampling that is statistically averaging, and the following results of in situ XAS experiments, the second process was incomplete and can be neglected when compared to the rst reaction. However, under these circumstances, Co(OH)2 still exhibited great performance, which pales the role of interlayer spacing, nominal BET surface and some other factors in activating this performance. Surprisingly, there is no signicant capacity loss after 8,000 cycles (95.7% capacity retention) as shown in Fig. 2a. The pathways for ion transport and diffusion to inner layer of Co(OH)2 could be forged and opened, which is

responsible for capacitance increase after 5,000 and 7,000 cycles. The difference of peak currents is caused by semiconductor nature and aggregation of large mass. When the mass gets lower, the same phenomenon is also presented (Supplementary Fig. 8) with better electrochemical properties (high specic capacitance, excellent rate capability and long cycling life) (Supplementary Fig. 9). A unit of 800 F g 1 (at the current density of 2 A g 1)

was achieved, which can be calculated according to the equation22: Cm C/m (I Dt)/(DV m), where Cm (F g 1) is

the specic capacitance, I (mA) the discharge current, Dt (s) the discharging time, DV (V) represents the potential drop during discharge process and m (mg) the mass of the Co(OH)2.

Accordingly, 756, 654 and 640 F g 1 was also yielded at the current density of 4, 8 and 16 A g 1, respectively. Even at the current density of 32 A g 1, a capacitance of 523 F g 1 was obtained, which proves the good rate capability of Co(OH)2. And

an asymmetric Co(OH)2-active carbon capacitor was fabricated and showed excellent performance (the capacitance reaches71 F g 1 at the current density of 1 A g 1 and the energy density of the asymmetric capacitor can be calculated as20.74 Wh kg 1 at a power density of 1,450 W kg 1; Supplementary Fig. 10). Nevertheless, a large mass is a prerequisite for acquiring good XAS signals in KOH aqueous solution, where X-ray intensities decay. For comparison, XAS measurements were also performed on the as-prepared electrode, the electrode after in situ experiment, as well as CoO as reference. Operated in a self-designed electrolytic cell (Fig. 2b, see the photographs of the device in Supplementary Fig. 11), the cyclic voltammetry curve shows no obvious difference (Fig. 2c).

In situ atomic level mechanism analysis. We chose in situ XAS as a powerful tool to qualitatively and quantitatively analyse how

a b

H2O

OH

Normalized intensity (a.u.)

Charge

P2

A

B

C

D

Co(OH)2

E

F

G

H

Discharge

H+

CoOOH

7,770

7,760

7,750

7,740

7,730

7,720

7,710

c

Energy (eV)

a

b

c d

A

B

C

D

E

FG

H

P1

P1

Normalized intensity (a.u.)

Normalized intensity (a.u.)

P2

Positive shift Negative shift

7,710 7,720 7,730 7,740 7,750 7,760 7,770

Energy (eV)

7,710 7,720 7,730 7,740 7,750 7,760 7,770

Energy (eV)

Figure 3 | Comparison of in situ XANES collected on the electrode and the reaction model of Co(OH)2 and CoOOH transformation. (a) In situ XANES spectra of a whole charge/discharge cycle in a three-dimensional mode. AH spectra correspond to the positions marked in Fig. 2c. (b) The reaction model of Co(OH)2 and CoOOH phase transformation. Co(OH)2 loses a hydrogen to becomes CoOOH during charge, and vice versa, a hydrogen is incorporated into CoOOH during discharge. (c) In situ XANES spectra of the charge process. (d) In situ XANES spectra of the discharge process.

4 NATURE COMMUNICATIONS | 8:15194 | DOI: 10.1038/ncomms15194 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15194 ARTICLE

the structural changes occur at atomic level. Thus, we can nd out what actually controls the morphology to be stable and thus conclude how to minimize the inuence of morphological change in actual operations. Figure 2d shows the X-ray absorption near-edge structure (XANES), the absorption K-edge for as-prepared Co(OH)2 is quite close to that of CoO, suggesting that the valence of cobalt is 2. (The CoO, purchased from Alfa Aesar, may

contain up to 10% Co3O4, which is noted in the label. The existence of Co3O4 with higher valence state makes some positive shifting of XANES for CoO when compared to pure CoO.) However, the spectrum of the electrode after 30 cycles shows a shift in the edge position to a higher energy (Supplementary Fig. 12). As mentioned above, thick layers (large mass) of Co(OH)2 is key to acquire good XAS signals. But this large mass resulted in conductivity decrease, which certainly would reduce the reversibility of the transformation of Co(OH)2 and CoOOH.

So some CoOOH remained and led to positive shift in Supplementary Fig. 12. In Fig. 1h, no obvious pores can be observed. b-Co(OH)2 can be well identied, as indicated by the (100) and (010) crystal planes with an acute angle 60 (Fig. 1i). According to the comparison of the spectra, Co element presents two valences indicating the co-existence of Co(OH)2 and

CoOOH, which is also in accordance with the cyclic voltammetry curve and proved by selected area electron diffraction pattern (Fig. 1j), where b-Co(OH)2 and CoOOH (denoted as C) exist coherently with an orientation relationship of (100)b//(100)C,(010)b//(010)C, o0014b//o0014C. The XANES spectra collected in Fig. 3a summarizes the overall charge/discharge process, which is schematically represented in Fig. 3b. In addition, Fig. 3c,d separately compares the in situ XANES spectra for the charge and discharge process, respectively. The absorption K-edges in Fig. 3c show a gradual positive shift while the potential increases from 0.3 to 0.4 V (Supplementary Table 2). This

proves that the Co valence increases during the charge process23. The P1 peak gains intensity during the charge process due to a decrease in the local disorder of the nearest neighbours; the P2 peak also gains intensity, which indicates a stronger CoCo shell. Consistently, we observe a negative shift and intensity decrease of P1 and P2 peaks during discharge (Fig. 3d). This evidences the gradual valence decrease of Co atoms. Interestingly, spectra A and H show almost the same absorption K-edge positions, demonstrating the valence circulation. However, no other features can be observed in the oscillation pattern, which suggests the absence of a major structural change.

To gain more insight into the proposed structural models shown in Fig. 3b, we carried out a series of DFT calculations. Figure 4a shows the computed energy prole for the Co(OH)2-

CoOOH phase transformation. We rst considered the removal of one H atom per formula unit from Co(OH)2 to form CoOOH, using the DFT-optimized lattice constants of Co(OH)2 (a b 3.176 ; c 9.358 ). We found that a structure with

one H atom per CoO layer (state A) is 571 meV more stable than a system that alternates fully hydrogenated CoO layers with empty ones (state B). We then examined the evolution of structures A and B into the CoOOH ground state (state C*), which involves the concerted migration of O and/or OH species. In addition, we considered two different scenarios regarding the imposed lattice constants of the cell during the phase transformation. In the left side of the energy prole in Fig. 4a, we xed the lattice constants to those of the DFT-optimized Co(OH)2,

whereas in the right side we considered the lattice constants of the CoOOH ground state (a b 3.036 ; c 8.862 ). In both

cases we found that the lowest transition state structures (T1 and T1*, respectively) correspond to the simultaneous diffusion of one O atom and one OH group within each CoO layer. T1* transition state is only 44 meV lower in energy than the T1, which indicates

that Co(OH)22CoOOH lattice relaxation plays a minor role in the stabilization of the transition state structures. Overall, our DFT results suggest that the structural transformation towards CoOOH on Co(OH)2 deprotonation involves the rearrangement of both O atoms and OH groups, with lattice relaxation affecting mainly to the stabilization of the CoOOH ground state.

In situ extended X-ray absorption ne structure (EXAFS) spectra showing similar patterns were recorded. This implies no much difference of the coordination environment (Fig. 4b). P3 and P4 in the EXAFS curves correspond to the CoO and CoCo shells24. When the charge process is proceeding, both peaks have a gradual negative shift (Supplementary Table 6), indicative of the decreasing distance of CoO and CoCo. The increasing intensities of both peaks show that structural disorder is reduced while the general structure is maintained (Fig. 4c). The entirely opposite situation occurs during the discharge process. Since the phase transformation is continually changing during the redox reaction, we thereof t the spectra as three spectra (A, D and H). The tted curves are shown in Fig. 4eg, and the tting results are listed in Supplementary Table 7. The coordination numbers (N) of CoO and CoCo are lower than theoretical values, which indicate phase distortions, the presence of non-diffraction, small particle domain and also the presence of a minority of phase. The variation trend of CoO and CoCo bond lengths (R) is consistent with in situ EXAFS spectra. R values of CoO and CoCo decrease for the charge process while increase on discharging. CoO/CoCo bond lengths vary within 0.06/0.22 during each charge/discharge cycle, which indicate the feasibility of this transformation. In addition, our DFT results indicate that the energy difference between the directly deprotonated Co(OH)2 and ground-state CoOOH structures (states A and C* in Fig. 4a, respectively) are separated by only 0.59 eV, with an activated energy barrier as low as 0.76 eV (1.34 eV) for the forward (backward) process (transition state T1* in Fig. 4a). These results lead us to correlate excellent pseudocapacitance properties with the structural similarity of Co(OH)2 and CoOOH, in which massive atom rearrangements are unnecessary for this mutual redox transformation, and thus save much energy from phase transition and enhance columbic efciency. Co(OH)2 and

CoOOH delicately adjust themselves to transform mutually into each other. The merit offers this reaction with a rapid switching rate, thus contributes to high power density, excellent rate capability and long cycling life. So to completely understand the reaction process, we chose in situ XAS to qualitatively and quantitatively study the structure transformation. To be complementary, we showed the SEM images before and after in situ XAS experiments. The hexagonal platelets in Fig. 1d,e show more quantity, more regularity, better homogeneousness and more uniform sizes than that of the Co(OH)2, in Fig. 1c, which spontaneously transformed from the as-prepared Co(OH)2. This is attributed to the electrochemical process. The bathed Co(OH)2, which may not be so regularly transformed from a-type, were gradually trimmed to be more regular under potential loading after an entire electrochemical cycle. Another key observation is that the platelets were uniformly distributed on the substrate with a minority of small scratches on the surface. This is the evidence of volume expansion and contraction. In contrast with the pristine porous structure, a more compacted, layered structure was formed consisting of a majority of b-Co(OH)2 (Fig. 1h,i).

For a better understanding of the relation between structural similarity and the redox process, it is interesting to consider the iron-series oxides and hydroxides. These three transition metals share similar electronic congurations (3d64s2, 3d74s2 and 3d84s2), but exhibit quite different pseudocapacitance properties. For example, Ni(OH)2 has a large pseudocapacitance9,25, holding

NATURE COMMUNICATIONS | 8:15194 | DOI: 10.1038/ncomms15194 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 5

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15194

a

2

1

0

Lattice constants of Co(OH)2 Lattice constants of CoOOH Lattice constants

of Co(OH)2

C*

T3

T2

T1

T3* T2*

C

T1* A*

B

Energy (eV)

B

A

B*

A

c

b A/A* B/B* C/C* T1/T1* T2/T2* T3/T3*

FT ()

b c d

A

B

C

D

E

FG

H Positive shift

P3

P4

Negative shift

P3

FT (r)

FT (r)

P4

0 2 4 6

A

B

C

D

E

F

G

H

R ()

0 1 2 3 4 5 6

R ()

0 1 2 3 4 5 6

R ()

e f g

H

Fitting

FT (r)

A

Fitting

FT (r)

D Fitting

FT (r)

0 1 2 3 4 5 6

0 1 2 4 5 6

3

R ()

Figure 4 | Phase transformation energy prole using DFT and comparison of in situ EXAFS collected on the electrode together with their corresponding tted curves. (a) All energies are referenced to the total energy of the CoOOH ground state (state C*). States AC stand for minima and transition structures are indicated by T1T3; their structures are schematically shown in the insets. Asterisks denote the use of the lattice constants of CoOOH ground state instead of those of Co(OH)2. Notice that structures with and without asterisks both correspond to CoOOH and, therefore, contain the same amount of H atoms. Red, white and blue balls stand for O, H and Co atoms, respectively. Total energy values for all the minima and transition states, and their corresponding optimized atomic structures are given in Supplementary Tables 35. (b) In situ EXAFS spectra of a whole charge/discharge cycle in a three-dimensional mode. AH labels correspond to the positions marked in Fig. 2c. (c,d) In situ EXAFS spectra of the charge and discharge process, respectively. (eg) The tting results of spectra A, D and H. The black solid lines represent the experimental curves while the red dot lines are the tted curves.

0 1 2 3 4 5 6

R ()

R ()

a similar structure (Supplementary Fig. 13) to NiOOH that facilitates phase transformation during charge. This motif is in perfect accordance with the case of Co(OH)2/CoOOH. When

Co(OH)2 acts as a SC electrode, a similar battery-mimic mechanism is more favourable, as Co(OH)2 and CoOOH have similar lattice structures just like the case of Ni(OH)2/NiOOH to facilitate the redox process. It is Faradaic in origin, involving the passage of charge across the double layer within the change of potential. Within a certain working window, a material with more distinct oxidation states is expected to achieve higher capacitance1. In contrast, Fe oxides show poor pseudocapacitance

properties3,2628. The dissimilarities of structural types (Supplementary Fig. 14) between Fe oxides and hydroxides pull down SC, which demands a fast phase transformation, especially for pseudocapacitance materials. The lattice structures change between different Fe phases is expected to cost too much energy due to the required massively rearrangement of atoms. In the end, it is the similarity between pristine and charged phases that originates excellent performance of Co(OH)2 and Ni(OH)2, which Fe oxides and hydroxides do not possess. The novelty and originality of our work lies right in that we successfully interpreted why Co(OH)2 shows distinct peaks in cyclic

6 NATURE COMMUNICATIONS | 8:15194 | DOI: 10.1038/ncomms15194 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15194 ARTICLE

voltammetry curves and plateaus in the galvanostatic charge/ discharge, similar to the behaviour of batteries, but still enabling excellent power densities and long cycling life with considerable energy densities. The foundation of the Co(OH)2, synthesized through one-step method without binders, having excellent capacity and cyclability, is the intrinsic structural similarity and the battery-mimic mechanism we demonstrated, which spares unnecessary and unfavourable massive atom rearrangements, and offers practical potentials of excellent electrochemical properties. SC electrode materials use their active surface to storage energy, while battery electrode materials are bulk materials for ion intercalation/de-intercalation. From the perspective of SCs, the structural similarity between Co(OH)2 and CoOOH spares massive structural changes for this phase transformation, resulting in excellent power densities and long cycling life. From the perspective of batteries, H intercalation/deintercalation for energy store/release (a mechanism similar to, for example, Li intercalation/de-intercalation) is the key factor to achieve the observed high energy densities. This can benet future supercapacitor materials designs to yield high energy densities. In particular, the transformation between Co(OH)2 and

CoOOH is a good model to pursue enhanced properties and minimize drawbacks, providing a novel concept for the hybridization of SCs and batteries. For the hybrid material design, we should design materials with structural similarity between their origin and charged phases so that we can preserve the excellent properties of SCs, such as quick charge/discharge capability and high power density. On the other hand, based on this battery-mimic mechanism, we can perform some theoretical calculations to predict the maximum thickness of active materials, which can be utilized for ion intercalation/de-intercalation to enhance energy density. And of course, other factors like conductivity should be also under consideration. Co(OH)2 is a beacon that shows us the way to pursue the hybrid device of SCs and batteries for further developments, but there are many work to be done and new materials needs to be delicately designed for hybridization of SCs and batteries.

DiscussionThe structural similarity between Co(OH)2 and CoOOH enables a battery-mimic mechanism, resulting in high specic pseudocapacitance and long cycling life. The different behaviours of iron-series elements prove this structure-to-property afnity. Co(OH)2 electrodes with a battery-mimic mechanism blurs therefore the distinction between SCs and batteries. At molecular level, the similarity between pristine and charged phases is ultimately responsible for the observed enhanced properties and performance. Our results may radically alter the design route of new electrodes for SC, battery and hybrid device materials. Lithiumion batteries can deliver a high energy density, but are limited by power density and cycling stability29,30. On the other hand, SCs still suffer from low energy density. Our ndings pave the way for new pseudocapacitance materials with high energy density and high power density for a wide range of applications, including high-power-density batteries with considerable energy densities. Our work also sheds light on the promising future for the hybridization of SCs and batteries, with great potential for future energy storage applications.

Methods

Synthesis of Co(OH)2. Our previous work demonstrated that the electrodeposition is an ideal synthesis method to prepare binder-free electrode for better performance17. In this work, a CFP was used as the substrate. Before the preparation, the CFP was ultrasonically cleaned with acetone, ethanol and distilled water for 30 min, respectively. Co(OH)2 thin lm was electrodeposited in an aqueous solution containing 1.2 M Co(NO3)2 (analytical reagent (AR), Z99.0%, Xilong

Chemical Co., Ltd.). The deposition process was performed at 45C in a

conventional three-electrode electrolytic cell consisting of the CFP (2 cm2 in area), a platinum counter electrode (2 cm2 in area) and a saturated calomel electrode reference. The deposition potential was controlled at 0.9 V for 60 min. The

electrodeposition process of the Co(OH)2 lm can be expressed as follows31:

NO 3 7H2O 8e ! NH4 10OH 1

Co2 2OH ! Co OH

2 2

After the deposition, the prepared electrode was rinsed consequentially with double-distilled water and dried in vacuum oven at 60C for 12 h. The mass of the deposited Co(OH)2 was measured for the weight difference before and after the electrochemical deposition by means of a micro-balance (Sartorius BT125D) with an accuracy of 0.01 mg.

Physical characterization. Surface morphology of the prepared electrode, after bathing in 1 M KOH for 15 min, and after in situ XAS measurements, was studied by using SEM (an Hitachi SU8000 scanning electron microscope operated at 2 kV). The detailed microstructural characterization was performed by TEM(JEM ARM 1300S for high-resolution TEM images and Libra 200 HT Mc Cs for selected area electron diffraction pattern and bright eld (BF) images). The specic area was examined by N2 adsorption measurements using an auto JW-BK132F instrument from JWGB SCI. & TECH instruments. Fourier transform infrared spectra were obtained from Perkin Elmer Spectrum One B instrument. X-ray photoelectron spectroscopy (XPS) spectra were acquired from ESCALAB-250 instrument with a monochromatic Al Ka radiation source and a hemisphere detector with an energy resolution of 0.1 eV. The Co(OH)2 phase was investigated using X-ray diffraction (RIGAKU D/MAX2500, l 1.54 nm). For more detailed

information, X-ray diffraction spectra of the prepared Co(OH)2 electrode was collected at Shanghai Synchrotron Radiation Facility, on BL14B1 beamline with an electron energy of 2 GeV.

Electrochemical characterization. Electrochemical measurements were carried out on a computer-controlled potentiostat (CHI660E, CH instrument, Shanghai) with a three-electrode electrochemical cell contain 1 M KOH aqueous solution as electrolyte. The working electrode was the Co(OH)2. A platinum plate and an saturated calomel electrode were used as counter electrode and reference electrode, respectively. The electrolyte (1 M KOH) was prepared from high-purity KOH pellets (AR, Z85.0%, Sinopharm Chemical Reagent Co., Ltd.) by adding 56 g of pellets to 1 l distilled water. Before the electrochemical activity was measured, the Co(OH)2 was bathed in the electrolyte for 15 min to make OH inltrated in the interlayer space. The active carbon (AC) electrode was fabricated using method as follows: a mixture of AC, acetylene black, naon (wt%: 85:10:5) and a small amount of ethanol was prepared by stirring 6 h to produce a homogeneous paste. This paste was pressed onto Ni foam to produce AC electrode. A two-electrode cell conguration was used to measure the performance of an asymmetric SC in 1 M KOH aqueous electrolyte solution. The working electrodes were the prepared Co(OH)2 and AC electrodes mentioned above, and they were placed together and separated by a porous non-woven cloth separator. Then cyclic voltammetry was used to investigate its activity. Its charge/discharge ability was measured by a galvanostatic test. The energy and power densities of the asymmetric capacitor were calculated as follow: E 0.5 C V2, P I V/m, where E (Wh kg 1) is the

energy density and P (W kg 1) is the power density.

In situ Co K-edge XAS. X-ray absorption spectra were collected at Beijing Synchrotron Facility (BSRF) on beamline 1W1B. The BSRF storage ring is operated at the electron energy of 2.2 GeV with beam current of 250 mA. A Si (111) double-crystal monochromator was applied. The beam size used at the sample position was about 900 300 mm2. For electrochemical experiments, no transmission data could

be collected, and then uorescence mode was used. All the data were collected at ambient temperature. Curve tting was performed with Artemis and IFEFFIT software24,32, EXAFS curve tting of these three spectra was carried out using both Co(OH)2 and CoOOH data as the starting parameters. Before collecting the XAS spectra, the electrode was cycled 30 times in a three-electrode cell to reach a steady state. In situ electrochemistry was performed in the self-made device (Fig. 1b) and cyclic voltammetry was used from 0.3 to 0.4 V. To meet the condition of XAS

sampling, we set the scan rate at 0.4 mV s 1. Spectra were acquired in Quick-EXAFS (QXAS) mode. Cobalt oxide (99.995%, Alfa Aesar, may contain up to 10%

Co3O4) was purchased to be the reference.

DFTcalculations. Spin-polarized DFT calculations were performed using supercell approach and the PerdewBurkeErnzerhof (PBE) functional33 as implemented in the Vienna ab initio simulation package (VASP, version 5.4.1)34. We treated explicitly the H (1s), O (2s, 2p) and Co (3s, 3p, 3d, 4s) electrons as valence states expanded in plane waves with a cut-off energy of 520 eV, whereas the remaining electrons were replaced by PBE-based projector-augmented wave potentials35. Total energies and electron densities were computed using the DFT U approach

of Dudarev et al.36, in which a Hubbard U-like term describing the onsite Coulomb interactions (Ueff U J, that is, the difference between the Coulomb U and

NATURE COMMUNICATIONS | 8:15194 | DOI: 10.1038/ncomms15194 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 7

ARTICLE NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15194

exchange J parameters, hereinafter referred to as simply U) is added to the PBE

functional. We used a value of U 3.3 eV for Co atoms, which was obtained by

Ceder et al.37 on the basis of oxide formation energies and agrees well with other U values recently proposed to describe different cobalt oxide and hydroxide phases38, as well as CoOOH (ref. 39) and Co-doped NiOOH (ref. 40) surfaces. We used a unit cell containing two formula units (Co2O4H2). Equilibrium lattice constants were computed allowing the atomic positions, lattice vectors and cell shape to relax with a residual force threshold of 0.02 eV 1 and using a

Monkhorst Pack grid with 8 8 2 k-point sampling. These computational

settings guarantee a tight convergence in equilibrium distances (better than0.001 ). Transition structures for the considered phase transformations were located by using the nudged elastic band algorithm41 using ve images along each pathway. The optimized lattice constants were used for all of the nudged elastic band calculations, which were performed at constant volume.

Surface calculations were performed using stoichiometric slab models cut along ve low-index surface directions of b-Co(OH)2 (space group P-3m1): (0001);

(1010); (1120); (0112); and (1014). Each slab model contained four to six atomic Co(OH)2 layers, with 15 of vacuum between slabs and centre layers xed to their bulk PBE U-optimal positions, leaving at least two layers relaxed on both

sides of the slab with a residual force threshold of 0.03 eV 1. Supplementary Fig. 3 shows the side and top views of the optimized geometries. Calculations were carried out using 800 eV plane-wave cut-off and k-point meshes of 12 12 1,

10 8 1, 14 12 1, 8 4 1 and 12 4 1 for (0001), (1010), (1120),

(0112) and (1014) surfaces, respectively. This set-up ensures a tight convergence in energies of at least 3 meV per atom. The surface energy of each slab of given area 2A and free energy Gslab was computed as g (Gslab Ngbulk)/2A, where N is the

number of Co(OH)2 formula units in the slab and gbulk is the free energy per formula unit of the ground state bulk Co(OH)2 (space group C2). We neglected the zero point energy and entropy corrections for Gslab and gbulk, which then become

equal to the DFT total energies, Eslab and Ebulk, respectively.

Data availability. The data that support the ndings of this study are available from the corresponding authors on request.

References

1. Conway, B. E. Electrochemical Supercapacitors: scientic Fundamentals and Technological Applications (Plenum, 1999).

2. Simon, P. & Gogotsi, Y. Materials for electrochemical capacitors. Nat. Mater. 7, 845854 (2008).

3. Wang, G., Zhang, L. & Zhang, J. A review of electrode materials for electrochemical supercapacitors. Chem. Soc. Rev. 41, 797828 (2012).

4. Arico, A. S., Bruce, P., Scrosati, B., Tarascon, J. M. & Van Schalkwijk, W. Nanostructured materials for advanced energy conversion and storage devices. Nat. Mater. 4, 366377 (2005).

5. Steele, B. C. H. & Heinzel, A. Materials for fuel-cell technologies. Nature 414, 345352 (2001).

6. Cheng, F. Y., Liang, J., Tao, Z. L. & Chen, J. Functional materials for rechargeable batteries. Adv. Mater. 23, 16951715 (2011).

7. Goodenough, J. B. & Park, K. S. The Li-ion rechargeable battery: a perspective.J. Am. Chem. Soc. 135, 11671176 (2013).8. Chen, J. & Cheng, F. Y. Combination of lightweight elements and nanostructured materials for batteries. Acc. Chem. Res. 42, 713723 (2009).

9. Li, H. B. et al. Amorphous nickel hydroxide nanospheres with ultrahigh capacitance and energy density as electrochemical pseudocapacitor materials. Nat. Commun. 4, 1894 (2013).

10. Cao, L., Xu, F., Liang, Y. Y. & Li, H. L. Preparation of the novel nanocomposite Co(OH)2/ultra-stable Y zeolite and its application as a supercapacitor with high energy density. Adv. Mater. 16, 18531857 (2004).

11. Gao, S. et al. Ultrahigh energy density realized by a single-layer beta-Co(OH)2 all-solid-state asymmetric supercapacitor. Angew. Chem. Int. Ed. 53, 1278912793 (2014).

12. Li, B. et al. Improved performances of beta-Ni(OH)2@reduced-graphene-oxide in Ni-MH and Li-ion batteries. Chem. Commun. 47, 31593161 (2011).

13. Simon, P., Gogotsi, Y. & Dunn, B. Where do batteries end and supercapacitors begin? Science 343, 12101211 (2014).

14. Huang, L. et al. Nickel-cobalt hydroxide nanosheets coated on NiCo2O4 nanowires grown on carbon ber paper for high-performance pseudocapacitors. Nano Lett. 13, 31353139 (2013).

15. Liu, Z., Ma, R., Osada, M., Takada, K. & Sasaki, T. Selective and controlled synthesis of alpha- and beta-cobalt hydroxides in highly developed hexagonal platelets. J. Am. Chem. Soc. 127, 1386913874 (2005).

16. Kong, L.B., Lang, J.W., Liu, M., Luo, Y.C. & Kang, L. Facile approach to prepare loose-packed cobalt hydroxide nano-akes materials for electrochemical capacitors. J. Power Sources 194, 11941201 (2009).

17. Zhao, C. et al. Synthesis of Co(OH)2/graphene/Ni foam nano-electrodes with excellent pseudocapacitive behavior and high cycling stability for supercapacitors. Int. J. Hydrogen Energy 37, 1184611852 (2012).

18. Gaunand, A. & Lim, W. L. From amorphous precipitates to sub-micronic crystalline platelets of Co(OH)(2): a kinetic study of the transformation process. Powder Technol. 128, 332337 (2002).

19. Ghodbane, O., Pascal, J. L. & Favier, F. Microstructural effects on charge-storage properties in MnO2-based electrochemical supercapacitors. ACS Appl. Mater. Interfaces 1, 11301139 (2009).

20. Chang, J.-K., Wu, C.-M. & Sun, I. W. Nano-architectured Co(OH)2 electrodes constructed using an easily-manipulated electrochemical protocol for high-performance energy storage applications. J. Mater. Chem. 20, 3729 (2010).21. Wang, L., Dong, Z. H., Wang, Z. G., Zhang, F. X. & Jin, J. Layered -Co(OH)2 nanocones as electrode materials for pseudocapacitors: understanding the effect of interlayer space on electrochemical activity. Adv. Func. Mater. 23, 27582764 (2013).

22. Zhao, C. M. et al. Ultrahigh capacitive performance from both Co(OH)(2)/ graphene electrode and K3Fe(CN)(6) electrolyte. Sci. Rep. 3, 2986 (2013).

23. Pantelouris, A., Modrow, H., Pantelouris, M., Hormes, J. & Reinen, D. The inuence of coordination geometry and valency on the K-edge absorption near edge spectra of selected chromium compounds. Chem. Phys. 300, 1322 (2004).

24. Rehr, J. J. & Albers, R. C. Theoretical approaches to X-ray absorption ne structure. Rev. Mod. Phys. 72, 621654 (2000).

25. Wang, H. L., Casalongue, H. S., Liang, Y. Y. & Dai, H. J. Ni(OH)(2) nanoplates grown on graphene as advanced electrochemical pseudocapacitor materials.J. Am. Chem. Soc. 132, 74727477 (2010).26. Wasiski, K., Walkowiak, M., Prolniczak, P. & Lota, G. Capacitance of Fe3O4/rGO nanocomposites in an aqueous hybrid electrochemical storage device. J. Power Sources 293, 4250 (2015).

27. Qu, Q., Yang, S. & Feng, X. 2D sandwich-like sheets of iron oxide grown on graphene as high energy anode material for supercapacitors. Adv. Mater. 23, 55745580 (2011).

28. Li, Y. et al. The impact of morphologies and electrolyte solutions on the supercapacitive behavior for Fe2O3 and the charge storage mechanism. Electrochim. Acta 178, 171178 (2015).

29. Tarascon, J. M. & Armand, M. Issues and challenges facing rechargeable lithium batteries. Nature 414, 359367 (2001).

30. Choi, N. S. et al. Challenges facing lithium batteries and electrical double-layer capacitors. Angew. Chem. Int. Ed. 51, 999410024 (2012).

31. Zhao, T., Jiang, H. & Ma, J. Surfactant-assisted electrochemical deposition of a-cobalt hydroxide for supercapacitors. J. Power Sources 196, 860864 (2011).

32. Newville, M. EXAFS analysis using FEFF and FEFFIT. J. Synchrotron Radiat. 8, 96100 (2001).

33. Perdew, J. P. et al. Erratum: restoring the density-gradient expansion for exchange in solids and surfaces. Phys. Rev. Lett. 102, 039902 (2009).

34. Kresse, G. & Furthmller, J. Efcient iterative schemes for ab initio total-energy calculations using a plane-wave basis set. Phys. Rev. B 54, 1116911186 (1996).

35. Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B 59, 17581775 (1999).

36. Dudarev, S. L., Botton, G. A., Savrasov, S. Y., Humphreys, C. J., A. & Sutton, A. P. Electron-energy-loss spectra and the structural stability of nickel oxide: an LSDA U study. Phys. Rev. B 57, 15051509 (1998).

37. Wang, L., Maxisch, T. & Ceder, G. Oxidation energies of transition metal oxides within the GGA U framework. Phys. Rev. B 73, 195107 (2006).

38. Chen, J. & Selloni, A. First principles study of cobalt (Hydr)oxides under electrochemical conditions. J. Phys. Chem. C 117, 2000220006 (2013).

39. Garcia-Mota, M. et al. Importance of correlation in determining electrocatalytic oxygen evolution activity on cobalt oxides. J. Phys. Chem. C 116, 2107721082 (2012).

40. Costanzo, F. Effect of doping b-NiOOH with Co on the catalytic oxidation of water: DFT U calculations. Phys. Chem. Chem. Phys. 18, 7490 (2016).

41. Jnsson, H., Mills, G. & Jacobsen, K. W. Classical and Quantum Dynamics in Condensed Phase Simulations (World Scientic, 1998).

Acknowledgements

This research is supported by the National Natural Science Foundation of China(No. 51372095), Special Funding for Academic Leaders (No. 419080500630) at Jilin University, and Korea Basic Science Institute grant (KBSI, T37210) to J.-G.K. W.Z. acknowledges the KBSI under the R&D programme (No. D36700) supervised by the Ministry of Science, ICT and Future Planning Korea. Beijing Synchrotron Radiation Facility (BSRF) is greatly acknowledged for providing us the beam time on beamline 1W1B for the XAS measurement, and Shanghai Synchrotron Radiation Facility (SSRF) on beamline BL14B1 for the X-ray diffraction measurement. J.C. acknowledges support by The Royal Society through the Newton Alumnus scheme. We are also grateful for computer resources to the Barcelona Supercomputer Center and the Arina cluster (Universidad del Pas Vasco/Euskal Herriko Unibertsitatea). We express our sincere thanks to Sibing Wang, Haifeng Zhao, Xiyang Wang, Haoxiang Wang and Dong Wang for their stimulating discussions and help.

8 NATURE COMMUNICATIONS | 8:15194 | DOI: 10.1038/ncomms15194 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications

NATURE COMMUNICATIONS | DOI: 10.1038/ncomms15194 ARTICLE

Author contributions

W.Z., W.T.Z. and T.D. conceived and coordinated the study and led writing of the manuscript; T.D. conducted the XAS and X-ray diffraction analysis; O.A. and J.C. contributed the theoretical calculations and analysis; J.-G.K., W.Z. and S.J.Y. contributed the microscopy and microanalysis; T.D. and J.W. and contributed the electrochemical test. All co-authors commented on the manuscript.

Additional information

Supplementary Information accompanies this paper at http://www.nature.com/naturecommunications

Web End =http://www.nature.com/ http://www.nature.com/naturecommunications

Web End =naturecommunications

Competing interests: The authors declare no competing nancial interests.

Reprints and permission information is available online at http://npg.nature.com/reprintsandpermissions/

Web End =http://npg.nature.com/ http://npg.nature.com/reprintsandpermissions/

Web End =reprintsandpermissions/

How to cite this article: Deng, T. et al. Atomic-level energy storage mechanism of cobalt hydroxide electrode for pseudocapacitors. Nat. Commun. 8, 15194doi: 10.1038/ncomms15194 (2017).

Publishers note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations.

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the articles Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/

Web End =http://creativecommons.org/licenses/by/4.0/

r The Author(s) 2017

NATURE COMMUNICATIONS | 8:15194 | DOI: 10.1038/ncomms15194 | http://www.nature.com/naturecommunications

Web End =www.nature.com/naturecommunications 9