# Revealing molecular-level surface redox sites of controllably oxidized black phosphorus nanosheets

## Abstract

Bulk and two-dimensional black phosphorus are considered to be promising battery materials due to their high theoretical capacities of 2,600 mAh g−1. However, their rate and cycling capabilities are limited by the intrinsic (de-)alloying mechanism. Here, we demonstrate a unique surface redox molecular-level mechanism of P sites on oxidized black phosphorus nanosheets that are strongly coupled with graphene via strong interlayer bonding. These redox-active sites of the oxidized black phosphorus are confined at the amorphorized heterointerface, revealing truly reversible pseudocapacitance (99% of total stored charge at 2,000 mV s−1). Moreover, oxidized black-phosphorus-based electrodes exhibit a capacitance of 478 F g–1 (four times greater than black phosphorus) with a rate capability of ~72% (compared to 21.2% for black phosphorus) and retention of ~91% over 50,000 cycles. In situ spectroelectrochemical and theoretical analyses reveal a reversible change in the surface electronic structure and chemical environment of the surface-exposed P redox sites.

## Main

Exfoliated black phosphorus (BP) nanosheets have emerged as a two-dimensional (2D) analogue of graphene1,2,3,4,5,6,7. Bulk and 2D BP can achieve theoretical capacities of 2,596 and 2,600 mAh g–1 for lithium-ion and sodium-ion storage, respectively, through alloying (MP → M2P → M3P, where M = Li or Na)8,9,10. 2D BP nanosheets have advantages over bulk BP due to their shortened ion diffusion pathways, larger surface-to-volume ratios and high storage site utilization11,12,13. However, BP electrodes suffer from low rate and cyclic capabilities due to their sluggish kinetics and large volume change (~307% for Li storage) arising from (de-)alloying8.

Here, we demonstrate that by coupling nanosheets of oxidized black phosphorus (oBP) and reduced graphene oxide (rGO), the surface redox pseudocapacitance exceeds the kinetic and stability limitations of previously explored BP. The pseudocapacitance of 2D oBP has two features: (1) molecular-level redox-active P=O sites obtained from controlled ozone-driven oxidation, revealing a surface redox charge storage capability with high reversibility (coulombic efficiency of 99.6%, even at a high rate of 50 A g−1) and high utilization efficiency (92% of theoretical capacitance); (2) strong interlayer coupling of redox-active oBP nanosheets with electronically conductive rGO via covalent bonding, resulting in high electronic conductivity, increased stability and decreased volume expansion of the BP8,9. These factors result in ultrahigh rate and cyclic performance. We correlated the surface chemistry and interface structure of our oBP-based hybrids with the fast and reversible pseudocapacitive behaviour of surface-exposed P redox sites. Using in situ spectroelectrochemical methods and density functional theory (DFT) calculations we probed the voltage-dependent chemical shift of the P=O sites, the variation of the P 2p peak, the V shape of the Ag2 vibration and strong proton binding.

## Synthesis and characterization of oBP-based hybrids

Free-standing oBP/rGO (foBG) hybrid films were obtained through controlled ozonation, vacuum filtration and subsequent thermal treatment (Supplementary Fig. 1 and Supplementary Sections 1 and 2). The exfoliation of BP into fewer than five layers, corresponding to 5–10 nm of height, was confirmed (Supplementary Figs. 2 and 3). After ozone treatment, the BP nanosheets were partially oxidized into oBP and well hybridized with GO. For assembly of self-standing films, it was crucial to disperse the oBP nanosheets homogeneously in aqueous solution. Oxidation of oBP was limited to an O/P ratio of 0.89 to prevent degradation14. The oxygen-containing groups in oBP and GO interact via hydrogen bonding to form a robust self-standing oBP/GO (oBGO) film, where the nanosheets were stacked alternately layer by layer (Supplementary Section 4). After thermal annealing at 800 °C, oBGO films were converted to foBG films (Supplementary Fig. 5). As shown in the proposed structure of foBG in Fig. 1a, the covalent bonds of C–P(=O) and C–O–P(=O) at the oBP/rGO heterointerface (denoted Int and Int, respectively) were formed through thermal decarboxylation and dehydrolysis reactions (Supplementary Section 5) and are critical for constructing the hybrid architecture by stabilizing the linking of the oBP and rGO nanosheets as well as providing fast charge transport. Moreover, some oxidized P not associated with interfacial covalent bonding remains intact at the surface (denoted Sur). The electrical conductivity of the foBG film was 1,194.15 S cm–1, while that of oBGO was too low to measure. Such a high electrical conductivity of foBG was attributed to the restored conjugation of rGO (Supplementary Table 1). Under atmospheric exposure (Supplementary Fig. 6), the chemical structure of foBG did not change significantly for one month due to stable bonds and passivation by rGO (Supplementary Section 7)8.

The major chemical bonds of C–P(=O) (Int) and C–O–P(=O) (Int) observed in foBG demonstrate that the dangling oxidized P species of oBP are covalently bound to oxygen-containing groups of rGO and stabilized at the interface (Fig. 1a). Solid-state 31P-NMR spectroscopy (Fig. 1b and Supplementary Fig. 7) for foBG revealed peaks located at around 48, 28, 19, 9, −3 and −20 ppm, which are assigned to C–P, P–P, P–O, O–P=O, C–O–P(=O) and C–P(=O) bondings, respectively. The oxidized P species can be divided into C–O–P(=O) and C–P(=O) interfacial covalent bonds and P–O and O–P=O dangling surface bonds15,16 (Supplementary Section 8). The Raman spectra of foBG showed characteristic 2D BP peaks at 359, 435 and 461 cm−1 assigned to $${{\mathrm{A}}_{\mathrm{g}}^1}$$, B2g and $${{\mathrm{A}}_{\mathrm{g}}^2}$$ modes, respectively (Fig. 1c)17. The exfoliation and orthorhombic phase of BP in foBG were verified by the change in peak position, curve area and full-width at half-maximum (Supplementary Figs. 810 and Supplementary Section 9). In the region of 300–500 cm–1, the Raman peaks of oBGO were dramatically weakened and broadened by the formation of oxidative defects10, and then restored by the formation of covalent C–P(=O) and C–O–P(=O) bonds. Moreover, a new Raman peak at ~750 cm–1 and downshift of the G band were associated with these two bonds. IR peaks of oBGO and foBG at 1,027, 1,132 and 1,639 cm−1 were assigned to P–O bonds for the former and symmetric stretching modes of P=O bonds for others (Fig. 1d)18. These peaks were distinguished from C–O and C=O peaks at 1,084 and 1,726 cm−1 (refs 19,20). New foBG peaks in the range of 500–800 cm–1 also supported the formation of C–P(=O) and C–O–P(=O) bonds (Fig. 1e and Supplementary Section 10). XPS spectra in Fig. 1f–h (Supplementary Fig. 12) reveal atom percentages of C 1s, P 2p and O 1s species to be 77.0, 12.2 and 10.8%, respectively. In the P 2p scan of 2D BP, two peaks at 130.2 and 131.0 eV were assigned to P–P bonds of P 2p3/2 and P 2p1/2, respectively. By contrast, additional broad peaks appeared at 131.9, 133.1 and 134.8 eV due to slight oxidation (Fig. 1f)21. The first two peaks indicate the formation of dangling (O–P=O) and bridging (P–O–P) phosphorus–oxygen bonds, while the last peak corresponds to phosphorus pentoxide (P2O5)22. All oxidative P peaks were intensified when 2D BP was hybridized to oBGO (Fig. 1g). The striking contrast between the P 2p scans of oBGO and foBG suggests the formation of C–P(=O) and C–O–P(=O) bonds between oBP and rGO (Fig. 1h and Supplementary Section 11). The covalent C–P(=O) and C–O–P(=O) fraction obtained from XPS was found to be 43% at the interface, consistent with the NMR data. The new peaks of C 1s and O 1s scans at 283.8 and 530.9 eV also support the formation of C–P(=O) and C–O–P(=O) bonds (Supplementary Fig. 13).

The morphology and structure of the foBG hybrid were characterized by scanning electron microscopy (SEM), high-resolution transmission electron microscopy (HR-TEM) and scanning TEM (STEM) (Fig. 2). The foBG hybrid film with thickness of 2–3 µm exhibited a layer-by-layer porous structure consisting of the oBP and rGO hybrid nanosheets which were randomly distributed and alternated, but not self-restacked (Fig. 2a,b). The pore size, pore volume and surface area of the foBG film were measured as 7.55 nm, 0.02 cm3 g–1 and 10.02 m2 g–1, respectively, using N2 adsorption/desorption isotherms (Supplementary Fig. 14). The cross-section of the foBG film was smoother than that of GO film due to uniform loading of puckered oBP and the alternately stacked hybrid architecture. This oBP was deposited onto the rGO surface in a face-to-face manner (Fig. 2c). A top view of foBG reveals that oBP has an orthorhombic crystalline structure (Fig. 2d,e), similar to pristine BP (Supplementary Fig. 15). In contrast, the cross-sectional HR-TEM images of foBG show a disordered structure of oBP and rGO at the amorphorized heterointerface (Fig. 2f). The lattice fringes of oBP and rGO were observed at the interface, as marked by red and yellow lines, respectively. The spectroscopic data reveal the formation of interlayer C–P(=O) or C–O–P(=O) bonds associated with an amorphorized interface. Direct identification of the atomic-level lattice distortion within the heterointerface was challenging due to distortion of the structure as a result of the electron beam23,24. Nonetheless, the structure of the interface is more disordered than the crystalline structure of the oBP nanosheets due to the presence of strain at the oBP/rGO interface and incommensurate stacking of the layers25. Energy-dispersive spectroscopy (EDS) elemental mapping using high-angle annular dark field scanning TEM (HAADF-STEM) of foBG revealed C, O and P elements (Fig. 2g), and their overlay images confirm a uniform distribution of each element (Fig. 2h and Supplementary Fig. 16). Furthermore, self-restacking inhibition of the oBP and rGO nanosheets was verified by the broadened X-ray diffraction (XRD) peaks of foBG compared to pristine samples (Fig. 2i and Supplementary Section 15). Accordingly, lattice fringe spacings of 0.54 and 0.34 nm are observed, corresponding to the interlayer distances of the (004) and (002) planes in oBP and rGO, respectively (Fig. 2f).

## Electrochemical behaviour of oxidized BP-based hybrids

The electrochemical behaviour of foBG was analysed using cyclic voltammetry (CV) in a three-electrode configuration in 1 M H2SO4 electrolyte. The current response in the CV curves of foBG increased towards negative potentials (Fig. 3a), indicating potential-dependent pseudocapacitive behaviour26. The CV curve of foBG showed the typical rectangular shape of ideal capacitive behaviour, while oBGO showed a similar pseudocapacitive feature, but with a much lower capacitance than foBG (Fig. 3b). Considering the low surface area (10.02 m2 g−1) of foBG and there being no gas evolution, the pseudocapacitive feature of the foBG is not related to reversible hydrogen storage in the designated potential window27. Moreover, the CV curves of 2D BP were highly distorted and resistive (Fig. 3b and Supplementary Fig. 18d). The ultrafast pseudocapacitive response of foBG becomes more prominent for scan rates from 5 to 2,000 mV s–1, as indicated by the rectangular shapes at high rates (Supplementary Fig. 18a). In contrast, oBGO yield a highly resistive CV shape. The fast pseudocapacitive behaviour of foBG was further confirmed by galvanostatic charge/discharge (GCD) curves in a two-electrode configuration, as shown in Fig. 3c (Supplementary Figs. 19 and 20). The discharge curve of foBG at 1 A g–1 showed a nearly triangular feature with a slight tilted end (insets of Fig. 3c), consistent with the CV feature. The specific capacitance of foBG at 1 A g–1 was evaluated as 478 F g–1, which was 1.8, 4.2 and 4.5 times greater than the 271, 115 and 107 F g–1 for oBGO, rGO and 2D BP, respectively. The capacitance retention of foBG was 72.4% even when the current density increased by a factor 50—much greater than the 48.0, 37.5 and 21.2% for oBGO, rGO and 2D BP, respectively, implying ultrahigh rate capability (Fig. 3d). The coulombic efficiency (99.6%) of foBG at the high rate of 50 A g1 also suggests exceptional reversibility. The fast capacitive dynamics of foBG associated with high capacitance and rate capability were investigated using electrochemical impedance spectroscopy (EIS) over frequencies ranging from 10–2 to 105 Hz (Fig. 3e). The equivalent series resistance of foBG was found to be 0.40 Ω, which is substantially lower than the 8.85 and 2.65 Ω obtained for 2D BP and oBGO, respectively. In the high-to-mid frequency range, foBG exhibited a lower charge transfer resistance of 19.75 Ω (compared to 48.11 and 32.8 Ω for 2D BP and oBGO, respectively).

We have identified unique surface redox features of foBG that are completely different from the diffusion-controlled alloying/dealloying mechanisms of BP. To confirm the surface pseudocapacitance of foBG, the contributions of capacitive current to the overall stored charge were quantitatively evaluated by Dunn’s method28 and designated as the shaded areas of the corresponding CV curves at various scan rates (Supplementary Fig. 21). As shown in the plot of capacitive and diffusive contributions to total stored charge in Fig. 3f, the capacitive contribution of foBG at 5 mV s–1 is ~86%, which is much higher than the 37.0 and 65.2% values for bulk and 2D BP, respectively. This capacitive contribution becomes more significant and increases up to 99.0% of total stored charge at 2,000 mV s−1.

The electrochemical stability of foBG was investigated at 10 A g–1 (Fig. 3g). The capacitance retention of the foBG over 50,000 cycles was ~91% of initial capacitance—higher than the 64.6% for 2D BP (~25,000 cycles). This value was also higher than the 77.8 and 72.6% for rGO and oBGO, respectively. As shown by the oxidative stability (Supplementary Section 7), the formation of stable covalent bonds also improved the chemical stability. Therefore, the electrochemical degradation of redox-active P=O sites was inhibited by the covalent bonds of foBG, enabling ultralong cyclability.

## Surface redox pseudocapacitance of 2D oxidized BPs

To verify the unique charge storage feature of foBG, in situ and ex situ spectroelectrochemical methods were used. Because the external magnetic fields of P atoms depend on their chemical environment16, the molecular-level redox-active P=O sites were tracked during the charging/discharging process using in situ 31P NMR spectroscopy (Fig. 4a and Supplementary Section 17). NMR measurements revealed that when the proton is inserted into the electrodes (charging process, 0.0 → 0.8 V), the chemical shift of the P=O groups at 10 ppm is broadened and split due to strong proton binding. The redox transition of oxidized P sites associated with proton binding was further confirmed by broadening and shifting of the P=O peak and changes in the P 2p peak in in situ FTIR and ex situ XPS results (Supplementary Fig. 22 and Supplementary Section 18). The NMR results are consistent with a reduction of electron density around the nucleus of the P bound to the O atom due to the formation of P–O–H bonding, resulting in deshielding of the local magnetic field. This is analogous to bonding of a lone pair electron with a highly electronegative ion29,30. The P=O peak shifts to 13 ppm on charging and recovers to 10 ppm and narrows on extraction (discharging process, 0.8 → 0.0 V) due to shielding of the local magnetic field. These results indicate that the chemical environment of surface P-containing groups in foBG dramatically change during charging/discharging due to a redox transition of surface-exposed P=O sites binding with protons.

We also tracked the surface electronic structure of the foBG using in situ Raman spectra (Fig. 4b and Supplementary Fig. 23). The in-plane $${{\mathrm{A}}_{\mathrm{g}}^2}$$ mode of the in situ Raman spectra was upshifted to a higher wavenumber (from 460 to 465 cm–1) during charging and shifted back to a lower wavenumber on discharging, in a reversible manner (Fig. 4c). This trend in the $${\rm{A}}_{\rm{g}}^2$$ mode arises from the introduction of compressive strain from proton binding on surface active oxygen atoms of the P=O groups and release of the strain on successive charging/discharging. In contrast, negligible changes in the $${{\mathrm{A}}_{\mathrm{g}}^1}$$ vibration were observed. Furthermore, the G band of rGO was more significantly changed than the D band due to electronic changes arising from charge carrier density and compressive strain (Supplementary Fig. 24)31,32. As the surface atoms are different from interior atoms because of missing neighbour atoms affording weaker bonding, the change in the Raman vibrations in the $${{\mathrm{A}}_{\mathrm{g}}^2}$$ peak signifies coupling of the electronic structure with surface strain.

To identify the energetically favourable hydrogen binding sites of foBG, we carried out first-principles DFT calculations within the generalized gradient approximation (GGA)33,34,35,36. We used the opt-type van der Waals density functional to account for dispersion interactions37,38 to more accurately predict the binding energies and interlayer spacings of the layered structures39,40,41 (Supplementary Section 21). The calculated DFT binding energy (Eb) between oBP and rGO was calculated to be –53.0 meV per atom in rGO, with an interlayer spacing of 3.380 Å. The unique hydrogen adsorption sites and interaction with oBP were investigated in our simplified 2D oBP/rGO interface (Supplementary Figs. 2527). The representative oxygen dangling bond (P=O) of oBP and the carbonyl functional group (C=O) of rGO were the most energetically favourable reaction centres for hydrogen adsorption, related to P–H bond formation in the pure 2D BP (Supplementary Table 5 and Supplementary Section 21). Figure 4d presents the electron density difference isosurface after hydrogen adsorption on the sites (Supplementary Figs. 25 and 26), in which red and green regions correspond to electron accumulation and depletion regions, respectively (±0.005 eV Å3). In general, when the chemical bonds are saturated (namely the other adsorption sites tested in Supplementary Fig. 27), hydrogen adsorption was found to be less favourable (Supplementary Table 6). Thus, the DFT calculations support the idea that molecular-level P=O configurations of foBG could act as redox-active sites for proton accumulation by forming favourable chemical bonds with hydrogen.

To highlight the importance of molecular-level design, we estimated the theoretical capacitance and fraction of redox-active P=O sites (Supplementary Section 22). The theoretical capacitance and area of the proposed basic unit for the redox-active oBP are 412 F g−1 and 0.1139 nm2, respectively42,43, indicating the existence of 8.78 P=O sites per nm2. Considering the actual capacitance (380 F g−1) of oBP in foBG, the utilization of oxidized P sites involved in pseudocapacitance is ~92%, indicating that 8.08 P=O sites per nm2 are active. This high utilization of P=O sites is attributed to the molecular-level control of oxidized P active sites44,45, which are stabilized by covalent bonding at the interface and in proximate contact with electronically conductive rGO with a low resistance.

We have demonstrated a new pseudocapacitive mechanism in oxidized P sites triggered by controlled oxidation and strong coupling of oBP with rGO at the amorphorized heterointerface. In particular, the oxidized P sites and interlayer bonds of foBG result in fast and reversible pseudocapacitive behaviour. The surface charge storage capability of the oBP was attributed to 92% utilization of the theoretical capacitance and reversibility of 99.6% coulombic efficiency at 50 A g−1. A pseudocapacitance of 478 F g–1 (99% of total stored charge) was obtained—exceeding the kinetic and stability limitations of existing BPs. Our contribution provides fundamental insight into the surface effects of 2D BP and offers a chemical strategy for molecular-level control to overcome the intrinsic limitations of existing energy storage materials for future applications.

## Methods

### Synthesis of foBG hybrid film

The GO solution was synthesized through the acidification, oxidation and exfoliation of natural graphite powder according to the modified Hummers method46. The quality of as-obtained GO has already been confirmed by our previous reports47. The homogeneous dispersion of GO and 2D BP in deionized (DI) water is crucial to form the uniform free-standing film. GO (20 mg) was dispersed in DI water as the solvent and ultrasonicated for 1 h to make a homogeneous GO dispersion. Subsequently, 30 ml of 2D BP solution was mixed with GO solution and tip-sonicated for 10 min to make a homogeneous mixture. The mixture was treated with ozone for 5 min. The 2D free-standing oBP/GO film was obtained by vacuum filtration through an Anodisc membrane filter, and the sample was noted as oBGO. To reduce the GO and form the stable state of oBP, the hybrid film was thermally treated in two steps. oBGO film was first heated at 250 °C for 1 h and the temperature then increased to 800 °C for 1 h to stabilize the oBP. During the heating process, the heating rate and flow rate of Ar gas were controlled at 10 °C min−1 and 100 ml min−1, respectively. After heating, the system was naturally cooled to room temperature with the same flow rate of Ar gas. Finally, the robust oBP/rGO hybrid film was obtained and noted as foBG.

### Characterization

Field-emission SEM (FE-SEM) images were obtained with a Philips SEM 535M equipped with a Schottky-based field-emission gun. HR-TEM images were collected on a JEM-3010 HRTEM (300 kV). HAADF-STEM images were obtained at 80 kV with a Titan G2 60-300 unit (FEI) with a low-background double tilt holder. EDS was used for elemental mapping of HAADF-STEM images. The XRD patterns were obtained on a Rigaku D/max IIIC (3 kW) with a θ/θ goniometer equipped with a CuKα radiation generator. The diffraction angle of the diffractograms was in the range 2θ = 2.5–70°. FTIR spectroscopy data were collected using a JASCO FT/IR-4700 spectrometer with attenuated total reflectance mode and standard DLaTGS detector. The Brunauer–Emmett–Teller surface areas and nitrogen adsorption–desorption isotherms were measured at 78 K BELSORP). Horvath–Kawazoe and Barrett–Joyner–Halenda analyses were used to calculate the average micro- and mesopore sizes. XPS data were obtained using a Thermo MultiLab 2000 system with an Al-Mg α X-ray source. There were no electron transfer processes during the XPS measurements of all samples. Raman and in situ Raman spectra were measured using a confocal micro-Raman spectrometer NRS-3100 (JASCO) system with a microscope with a ×100 lens and an excitation laser beam source (532 nm wavelength). Solid-state and in situ 31P magic angle spinning (MAS) NMR spectra were measured at 11.7 T on a Varian Unity Inova 500 MHz spectrometer equipped with a 1.2 mm Chemagnetics MAS probe head using a sample rotation rate of 20 kHz. The electrochemical characteristics of all samples were evaluated by measuring CV and GCD curves and using a BioLogic Science instrument VSP at room temperature. In addition, impedance spectroscopy data were obtained using a Solartron 1260 analyser.

## Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request. The data are not publicly available due to restrictions (for example, information that could compromise the privacy of research participants).

## References

1. 1.

Jiang, J.-W. & Park, H. S. Negative Poisson’s ratio in single-layer black phosphorus. Nat. Commun. 5, 4727 (2014).

2. 2.

Cai, Y., Zhang, G. & Zhang, Y.-W. Layer-dependent band alignment and work function of few-layer phosphorene. Sci. Rep. 4, 6677 (2014).

3. 3.

Ling, X. et al. Low-frequency interlayer breathing modes in few-layer black phosphorus. Nano Lett. 15, 4080–4088 (2015).

4. 4.

Reich, E. S. Phosphorene excites materials scientists. Nature 506, 19 (2014).

5. 5.

Chhowalla, M. et al. The chemistry of two-dimensional layered transition metal dichalcogenide nanosheets. Nat. Chem. 5, 263–275 (2013).

6. 6.

Petoukhoff, C. E. et al. Ultrafast charge transfer and enhanced absorption in MoS2–organic van der Waals heterojunctions using plasmonic metasurfaces. ACS Nano 10, 9899–9908 (2016).

7. 7.

Zhang, H. Ultrathin two-dimensional nanomaterials. ACS Nano 9, 9451–9469 (2015).

8. 8.

Sun, J. et al. Formation of stable phosphorus–carbon bond for enhanced performance in black phosphorus nanoparticle–graphite composite battery anodes. Nano Lett. 14, 4573–4580 (2014).

9. 9.

Sun, J. et al. A phosphorene–graphene hybrid material as a high-capacity anode for sodium-ion batteries. Nat. Nanotech. 10, 980–985 (2015).

10. 10.

Xu, G.-L. et al. Nanostructured black phosphorus/Ketjenblack–multiwalled carbon nanotubes composite as high performance anode material for sodium-ion batteries. Nano Lett. 16, 3955–3965 (2016).

11. 11.

Lukatskaya, M. R., Dunn, B. & Gogotsi, Y. Multidimensional materials and device architectures for future hybrid energy storage. Nat. Commun. 7, 12647 (2016).

12. 12.

Peng, L., Zhu, Y., Chen, D., Ruoff, R. S. & Yu, G. Two-dimensional materials for beyond lithium ion batteries. Adv. Energy Mater. 6, 1600025 (2016).

13. 13.

Wang, X., Weng, Q., Yang, Y., Bando, Y. & Golberg, D. Hybrid two-dimensional materials in rechargeable battery applications and their microscopic mechanisms. Chem. Soc. Rev. 45, 4042–4073 (2016).

14. 14.

Kwon, H. et al. Ultrathin and flat layer black phosphorus fabricated by reactive oxygen and water rinse. ACS Nano 10, 8723–8731 (2016).

15. 15.

Verkade, J. G. & Quin, L. D. Phosphorus-31 NMR Spectroscopy in Stereochemical Analysis (Wiley-VCH, Weinheim, 1987).

16. 16.

Smith, J. G. Organic Chemistry (McGraw-Hill, New York, 2013).

17. 17.

Liu, S. et al. Thickness-dependent Raman spectra, transport properties and infrared photoresponse of few-layer black phosphorus. J. Mater. Chem. C 3, 10974–10980 (2015).

18. 18.

Yu, X. et al. Elucidating surface redox charge storage of phosphorus-incorporated graphenes with hierarchical architectures. Nano Energy 15, 576–586 (2015).

19. 19.

Ma, G. et al. Phosphorus and oxygen dual-doped graphene as superior anode material for room-temperature potassium-ion batteries. J. Mater. Chem. A 5, 7854–7861 (2017).

20. 20.

Silberberg, M. S., Duran, R., Haas, C. G. & Norman, A. D. Chemistry: The Molecular Nature of Matter and Change (McGraw-Hill, New York, 2006)..

21. 21.

Kang, J. et al. Solvent exfoliation of electronic-grade, two-dimensional black phosphorus. ACS Nano 9, 3596–3604 (2015).

22. 22.

Edmonds, M. et al. Creating a stable oxide at the surface of black phosphorus. ACS Appl. Mater. Interfaces 7, 14557–14562 (2015).

23. 23.

Asensio, M. C. et al. Interfaces and heterostructures of van der Waals materials. J. Phys. Condens. Matter 28, 490301 (2016).

24. 24.

Novoselov, K., Mishchenko, A., Carvalho, A. & Neto, A. C. 2D materials and van der Waals heterostructures. Science 353, aac9439 (2016).

25. 25.

Gao, Y., Liu, Q. & Xu, B. Lattice mismatch dominant yet mechanically tunable thermal conductivity in bilayer heterostructures. ACS Nano 10, 5431–5439 (2016).

26. 26.

Acerce, M., Voiry, D. & Chhowalla, M. Metallic 1T phase MoS2 nanosheets as supercapacitor electrode materials. Nat. Nanotech. 10, 313–318 (2015).

27. 27.

Gao, Q., Demarconnay, L., Raymundo-Piñero, E. & Béguin, F. Exploring the large voltage range of carbon/carbon supercapacitors in aqueous lithium sulfate electrolyte. Energy Environ. Sci. 5, 9611–9617 (2012).

28. 28.

Wang, J., Polleux, J., Lim, J. & Dunn, B. Pseudocapacitive contributions to electrochemical energy storage in TiO2 (anatase) nanoparticles. J. Phys. Chem. C 111, 14925–14931 (2007).

29. 29.

Wen, Y. et al. Expanded graphite as superior anode for sodium-ion batteries. Nat. Commun. 5, 4033 (2014).

30. 30.

Griffin, J. M. et al. In situ NMR and electrochemical quartz crystal microbalance techniques reveal the structure of the electrical double layer in supercapacitors. Nat. Mater. 14, 812–819 (2015).

31. 31.

Frank, O., Dresselhaus, M. S. & Kalbac, M. Raman spectroscopy and in situ Raman spectroelectrochemistry of isotopically engineered graphene systems. Acc. Chem. Res. 48, 111–118 (2015).

32. 32.

van den Beld, W. T. et al. In-situ Raman spectroscopy to elucidate the influence of adsorption in graphene electrochemistry. Sci. Rep. 7, 45080 (2017).

33. 33.

Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865 (1996).

34. 34.

Kresse, G. & Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B 59, 1758 (1999).

35. 35.

Kresse, G. & Hafner, J. Ab initio molecular dynamics for liquid metals. Phys. Rev. B 47, 558 (1993).

36. 36.

Kresse, G. & Furthmüller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comp. Mater. Sci. 6, 15–50 (1996).

37. 37.

Klimeš, J., Bowler, D. R. & Michaelides, A. Chemical accuracy for the van der Waals density functional. J. Phys. Condens. Matter 22, 022201 (2009).

38. 38.

Klimeš, J., Bowler, D. R. & Michaelides, A. Van der Waals density functionals applied to solids. Phys. Rev. B 83, 195131 (2011).

39. 39.

Aykol, M., Kim, S. & Wolverton, C. Van der Waals interactions in layered lithium cobalt oxides. J. Phys. Chem. C 119, 19053–19058 (2015).

40. 40.

Kim, S. et al. Layered-layered-spinel cathode materials prepared by a high-energy ball-milling process for lithium-ion batteries. ACS Appl. Mater. Interfaces 8, 363–370 (2015).

41. 41.

Chen, K.-S. et al. Comprehensive enhancement of nanostructured lithium-ion battery cathode materials via conformal graphene dispersion. Nano Lett. 17, 2539–2546 (2017).

42. 42.

Zhi, M., Xiang, C., Li, J., Li, M. & Wu, N. Nanostructured carbon–metal oxide composite electrodes for supercapacitors: a review. Nanoscale 5, 72–88 (2013).

43. 43.

Du, H., Lin, X., Xu, Z. & Chu, D. Recent developments in black phosphorus transistors. J. Mater. Chem. C 3, 8760–8775 (2015).

44. 44.

Cao, J. et al. Supercapacitor electrodes from the in situ reaction between two-dimensional sheets of black phosphorus and graphene oxide. ACS Appl. Mater. Interfaces 10, 10330–10338 (2018).

45. 45.

Yu, X. et al. Emergent pseudocapacitance of 2D nanomaterials. Adv. Energy Mater. 8, 1702930 (2018).

46. 46.

Hummers, W. S. Jr & Offeman, R. E. Preparation of graphitic oxide. J. Am. Chem. Soc. 80, 1339–1339 (1958).

47. 47.

Choi, B. G. et al. Solution chemistry of self-assembled graphene nanohybrids for high-performance flexible biosensors. ACS Nano 4, 2910–2918 (2010).

## Acknowledgements

This work was supported by the R&D Convergence Program (CAP-15-02-KBSI) of the National Research Council of Science & Technology, the National Research Foundation funded by the Ministry of Science, ICT, and Future Planning (no. 2017M2A2A6A01021187) and the Energy Technology Development Project (ETDP) funded by the Ministry of Trade, Industry, and Energy (20172410100150), Republic of Korea. The authors thank the Korea Basic Science Institute for technical support and Y. Gogotsi for valuable discussions. S.K. and C.W. (DFT calculations) were supported by financial assistance award no. 70NANB14H012 from the US Department of Commerce, National Institute of Standards and Technology as part of the Center for Hierarchical Materials Design (CHiMaD). This research was supported in part through the computational resources and staff contributions of the Quest High Performance Computing Facility at Northwestern University, which is jointly supported by the Office of the Provost, the Office for Research, and Northwestern University Information Technology. S.K. is grateful for discussions with V. I. Hegde, M. Liu and S. Hao at Northwestern University.

## Author information

Authors

### Contributions

H.S.P., P.N. and X.Y. conceived and designed the experiments. P.N. and X.Y. performed sample fabrication, characterization and electrochemical measurements. P.N. and X.Y. conducted in situ Raman spectroscopy and analysed in situ Raman and in situ NMR data. S.K.P. and H.J.K. carried out in situ NMR spectroscopy. J.-Y.H., J.E.Y., J.K. and M.C. characterized bulk and 2D BP and interpreted data. W.L. and J.Y.H. carried out HAADF-STEM and EDS measurements. S.K. and C.W. performed DFT calculations and data analysis. P.N., X.Y., S.K., M.C. and H.S.P. co-wrote the paper. All authors discussed the results and commented on the manuscript.

### Corresponding author

Correspondence to Ho Seok Park.

## Ethics declarations

### Competing interests

The authors declare no competing interests.

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

## Rights and permissions

Reprints and Permissions

Nakhanivej, P., Yu, X., Park, S.K. et al. Revealing molecular-level surface redox sites of controllably oxidized black phosphorus nanosheets. Nature Mater 18, 156–162 (2019). https://doi.org/10.1038/s41563-018-0230-2

• Accepted:

• Published:

• Issue Date:

• ### Barrier function of graphene for suppressing the smoke toxicity of polymer/black phosphorous nanocomposites with mechanism change

• Wei Cai
• , Zhaoxin Li
• , Xiaowei Mu
• , Lingxin He
• , Xia Zhou
• , Wenwen Guo
• , Lei Song
•  & Yuan Hu

Journal of Hazardous Materials (2021)

• ### Highly flexible and conductive nanoporous Ag as good substrate for flexible hybrid supercapacitors

• Chaoyang Wang
• , Shengli Zhu
• , Yanqin Liang
• , Zhenduo Cui
• , Shuilin Wu
• , Chunling Qin
• , Shuiyuan Luo
•  & Akihisa Inoue

Journal of Alloys and Compounds (2021)

• ### Recent advances in off-grid electrochemical capacitors

• Buddha Deka Boruah

Energy Storage Materials (2021)

• ### Black Phosphorus Nanosheet with High Thermal Conversion Efficiency for Photodynamic/Photothermal/Immunotherapy

• Danyang Xu
• , Jun Liu
• , Yanxing Wang
• , Yingying Jian
• , Weiwei Wu
•  & Ruichan Lv

ACS Biomaterials Science & Engineering (2020)

• ### Super-tough MXene-functionalized graphene sheets

• Tianzhu Zhou
• , Chao Wu
• , Yanlei Wang
• , Antoni P. Tomsia
• , Mingzhu Li
• , Eduardo Saiz
• , Shaoli Fang
• , Ray H. Baughman
• , Lei Jiang
•  & Qunfeng Cheng

Nature Communications (2020)