The synthesis process of v-MoO3 PNB is schematically illustrated (Fig. S1, supporting information), which start a modified hydrothermal method reported previously for the synthesis of MoO3 nanobelt (MoO3 NB),18 which was then treated by FeCl3 solution.36 The v-MoO3 PNB was finally obtained by partially reduction upon thermally annealing with NaH2PO2·H2O as reductant at 250 ºC for 1 h according to the following reactions:
2 NaH2PO2·H2O (s) → PH3 (g) + Na2HPO4 (s) + 2 HO2 (g)
MoO3 + PH3 → MoO3-x + x /4 H3PO4
The as-prepared v-MoO3 PNB, similar to that of MoO3 NB, shows typical x-ray diffraction (XRD) patterns of orthorhombic MoO3 phase (JCPDS#35-0609) with three dominant peaks that can be well indexed to (020), (040), and (060) crystalline facets. (Fig. 1a). Notably, one can obviously observe negative shifts of peak (020), (040), and (060) for v-MoO3 PNB relative to those of MoO3 NB (Fig. S2, supporting information), implying the increase of lattice spacing in v-MoO3 PNB arising from the introduction of oxygen vacancies.37, 38 Fig. 1b-c show the typical scanning electron microscopy (SEM) images of MoO3 NB, displaying a smooth nanobelt morphology with an average width of 240 nm and lengths about 5-10 μm (Fig. S3, supporting information), while the v-MoO3 PNB show an eminently rough surface due to the etching treatment of FeCl3 solution (Fig. 1d-e) with well maintaining nanobelt structure (Fig. S4, supporting information). It is noted that the color changes from the grey-white of MoO3 NB to the blue of the v-MoO3 PNB (Fig. S5, supporting information), further confirming inducing the surface defects of oxygen vacancies. The transmission electron microscopy (TEM) image convinces the existence of the porous structure in v-MoO3 PNB (Fig. 1f). High-resolution TEM image (Fig. 1g) implies that the well-defined lattice fringe with an interplanar distance of 3.44 Å, corresponding the (020) plane of MoO3. Moreover, some atomic-leveled distortion could be clearly observed (remarked by red dots, also see Fig. S6, supporting information), which is due to the presence of oxygen vacancies. The selected area electron diffraction (SAED) pattern was shown in Fig.1h, indicating the single-crystalline nature of the v-MoO3 PNB, which is attributed to the (010) zone axis diffraction.
Raman spectroscopy and electron paramagnetic resonance (EPR) spectroscopy were carried out to study the samples. Both the MoO3 NB and the v-MoO3 PNB show three sharp Raman bands located at 663 (B2g/B3g, vas, O-Mo-O stretch), 816 (Ag, vas, Mo=O stretch) and 993 cm-1 (Ag, vas, Mo=O stretch, Fig. 2a),38 confirming they are orthorhombic structure of α-MoO3. The v-MoO3 PNB exhibits a remarkable reduction in Raman peak intensity compared to the MoO3 NB, implying the successful introduction of oxygen vacancies. Fig. 2b shows the Electron Paramagnetic Resonance (EPR) spectroscopy of MoO3 NB and v-MoO3 PNB. The MoO3 NB almost shows no EPR signal while the v-MoO3 PNB manifests a remarkable EPR signal in fields of 3250-3750 Gs. The g1 can be ascribed to the O- as the paramagnetic center and the g2 is related to the Mo5+ as the paramagnetic center.39, 40 The presence of Mo5+ further indicates the presence of oxygen vacancy in v-MoO3 PNB.
X-ray photoelectron spectroscopy (XPS) was conducted to further investigate the surface properties and chemical composition of the samples. The survey spectroscopy shows the sign of Mo, C, and O elements in both MoO3 NB and v-MoO3 PNB (Fig. S7, supporting information). Fig. 2c presents the high-resolution XPS spectra of Mo 3d, which can be divided into four peaks, 236.1 eV for Mo6+ 3d3/2, 232.9 eV for Mo6+ 3d5/2, 235.4 eV for Mo5+ 3d3/2, and 232.2 eV for Mo5+ 3d5/2, respectively.41, 42 The atomic ratio of Mo5+/Mo6+ can be employed to character the oxygen vacancies of MoO3. For MoO3 NB, the atomic ratio of Mo5+/Mo6+ is estimated to be 0.13 (corresponding to MoO2.93), while the corresponding value for v-MoO3 PNB increases to 0.38 (corresponding to MoO2.82), suggesting that the partial reduction of MoO3 and the oxygen vacancies have been increased after the reduction treatment. The peak of 530.6 eV is observed for both MoO3 and v-MoO3 PNB from the O1s core-level XPS spectra (Fig. 2d), which is ascribed to the Mo-O bond. Moreover, the O 1s XPS peak of 531.4 eV for MoO3 NB indicates the absorption of OH- on the surface of MoO3 NB, while for the v-MoO3 PNB, the peak corresponding to OH- is missing, and there appear two peaks located at 532.4 and 531.2 eV that can be assigned to the sign of PO3- and H2PO4-,43 further confirming the partial reduction of MoO3 by phosphate ions during the annealing process.
To evaluate the H+ storage capability, cyclic voltammetry (CV) tests were carried out in 1.0 M H2SO4 with a standard three-electrode system with Ag/AgCl as reference electrode, Pt wire as counter electrode, and the MoO3 modified carbon paper (1*1 cm2) as working electrode, respectively. Fig. 3a shows the first three CV curves of v-MoO3 PNB at a scan rate of 1 mV s-1 at the potential range from -0.5 to 0.3 V. One can observe three pairs of redox peaks of 0.074 V/0.01 V, 0.032 V/-0.04 V, and -0.371 V/-0.384 V that are related to the insertion and deinsertion of H+ into the v-MoO3 PNB electrode. Notably, the peak current of v-MoO3 PNB is larger than that of MoO3 NB (Fig. 3b), implying a higher H+ storage capacity for v-MoO3 PNB, which can be rationally attributed to the higher conductivity (Fig. S8, supporting information) and more activity sites derived from large specific surface area (Fig. S9, supporting information). Moreover, the reaction kinetics of H+ insertion and deinsertion were further studied by recording the CV curves at different scan rates from 0.5 to 10 mV s-1 (Fig. 3c). The peak currents and the peak separations increase along with the scan rate increases, which is due to the enhanced polarization at high scan rates. The Randles-Sevcik equation was employed to determine the relationship between the peak currents (ip) and scan rate (v). The ip linearly increases with the square root of the scan rate (v1/2) over the scan rate range, with the excellent linearity close to 1 (Fig. 3d), indicating a diffusion process control of the H+ insertion and deinsertion on the v-MoO3 PNB electrode.
The charge and discharge profiles from 1 A g-1 to 20 A g-1 of the v-MoO3 PNB are shown in Fig. 4a. Even at an ultra-high current of 20 A g-1, three obvious pairs of charge and discharge plateaus can be observed, which match well with the three-step redox behavior in the CV curves. While the charge and discharge plateaus of MoO3 NB can’t be recognized at large current (Fig. S10, supporting information), implying the high rate capability for the v-MoO3 PNB thanks to its improved conductivity. Fig. 4b shows the capacity at different current density of the v-MoO3 PNB, which delivers discharge capacities of approximately 248.2, 245.4, 210.8, and 198.7 mAh g-1 at currents of 1, 5, 10, and 20 A g-1, respectively. The Coulombic efficiency is in the range of 97 - 99% can be obtained when the current is reduced to 1 A g-1. However, the MoO3 NB exhibits a poor rate performance with low specific capacity and inferior capacity recovery (Fig. S11, supporting information). Fig. 4c shows the stability of the v-MoO3 PNB for H+ insertion and deinsertion. At a current of 10 A g-1, above 92% of its initial capacity can remain with an average Columbia efficiency of 96% after 200 cycles, which is superior to those of MoO3 NB (Fig. S12, supporting information).
Given its quite good H+ storage capability, we developed a alkali-acid Zn-MoO3 hybrid battery that OH- are involved on alkaline Zn conversion anode while H+ insertion/deinsertion takes place in acidic v-MoO3 PNB cathode (Fig. 5a).
Anode: Zn + 2OH- ↔ ZnO + H2O
Cathode: MoO3-x + H+ ↔ HyMoO3-x
In this as-developed hybrid device, the redox reactions of anode and cathode can proceed in their optimal conditions with potential of harvesting the so-called electrochemical neutralization energy (ENE),44, 45, 46, 47, 48 which can significantly enhance the voltage and energy density of energy devices. Fig. 5b shows the CV curves of the Zn anode in alkali and the v-MoO3 PNB cathode acid at a scan rate of 1 mV s-1. Both electrodes show redox activity at their individual potential windows, which indicates the availability and feasibility for fabricating alkali-acid Zn-MoO3 hybrid battery. The as-constructed alkali-acid Zn-MoO3 hybrid battery (Fig. 5b) shows one prominent pair of redox peaks in the potential range of 1.2 V to 1.6 V. As a result, the battery shows a high open-circuit voltage (OCV) of about 1.85 V (Fig. 5c), higher than most of aqueous battery.49, 50, 51, 52, 53 And the value is constant for 60 mins, implying the decent stability of the hybrid battery. Moreover, a single battery can power a red light-emitting diode (LED, 1.8-2.2 V, Fig. S13, supporting information), further demonstrating the high voltage of the as-built battery. It is noted that the OCV of the battery varies when the electrolyte is different (Fig. S14, supporting information) because the H+ insertion and deinsertion are sensitive to the pH value of electrolytes, including the current and potential (Fig. S15, supporting information).
Fig. 5d shows the charge and discharge profiles from 1 A g-1 to 10 A g-1 of the alkali-acid Zn-MoO3 hybrid battery. The distinct charge and discharge plateaus can be observed in all curves at various current density. The alkali-acid Zn-MoO3 hybrid battery can deliver a discharge capacities of 198.2, 170.2, 157.8, and 139.2 mAh g-1 at currents of 1, 2, 5, and 10 A g-1, respectively (Fig. 5e), which implies the battery can afford an excellent rate performance. Moreover, when the current density was back to 1 A g-1, the reversible specific capacity could be recovered, indicating excellent reversibility of the battery even after high rate current. Meanwhile, high Coulombic efficiency in the range of 94 - 98% can be achieved. Fig. 5f exhibits the charge and discharge profiles (GCD) of the alkali-acid Zn-MoO3 hybrid battery in different cycle number at a current density of 5 A g-1, in which one can clearly observe two pairs of GCD voltage plateaus. The dominant GCD voltage plateaus are located at about 1.45 V and 1.85 V for the first charge-discharge process, and they are positioned at a stable level of about 1.91 V and 1.51 V in the following cycles, indicating that the alkali-acid Zn-MoO3 hybrid battery shows a good reversibility and cycling stability. Furthermore, long-term cycling performance of the alkali-acid Zn-MoO3 hybrid battery confirms the excellent stability. As shown in Fig. 5g, above 90% of its initial capacity can be achieved with an average Columbia efficiency of 95% after 200 cycles at a current of 5 A g-1. The phase of the v-MoO3 PNB has not been changed after long-term stability test, except for peaks from Nafion at 18º and 26.4º, and crystalline carbon paper at about 25º (Fig. S16, Supporting Information). In addition, the morphology and microstructure of the v-MoO3 PNB also keep almost unchanged, as demonstrated by the SEM images of the material (Fig. S17, Supporting Information). Moreover, the CV curve of Zn anode shifts for high potential, while that of v-MoO3 PNB cathode shifts to low potential (Fig. S18, Supporting Information), which may be due to the crossover of H+ and OH-. For this purpose, we monitor the pH value variation of the catholyte and anolyte during cycling. The concentration of H+ and OH- decrease along with cycling (Fig. S19, Supporting Information), which induces the capacity decay of the alkali-acid Zn-MoO3 hybrid battery.
To investigate the mechanism of the alkali-acid Zn-MoO3 hybrid battery, the ex-situ XRD measurements of v-MoO3 PNB electrode were performed during the charge and discharge process (Fig. 6). As shown in Fig. 6b, all the XRD patterns during the charge-discharge processes show three predominant peaks of (020), (040) and (060), indicating that no phase transition exists in the processes, which suggests a solid solution reaction during H+ ions insertion and deinsertion. Moreover, the peak located at about 18º is assigned to the Nafion. When discharged to 1.0 V, the peak at 12.7º for the (020) planes of the v-MoO3 PNB shifts to 12.63 V, associated with the (002) planes of H0.88MoO3,16 implying the H+ ions have been inserted into the v-MoO3 PNB. When charged to 1.7 V and 2.2 V, the peak derived from the (020) planes returns to 12.67º and 12.7º. During the subsequent discharge process, the (020) peak shifts from 12.7º to 12.63º. These observations reveal that the insertion and deinsertion of H+ are reversible on the v-MoO3 PNB.