Interlayer gap widened α-phase molybdenum trioxide as high-rate anodes for dual-ion-intercalation energy storage devices

Employing high-rate ion-intercalation electrodes represents a feasible way to mitigate the inherent trade-off between energy density and power density for electrochemical energy storage devices, but efficient approaches to boost the charge-storage kinetics of electrodes are still needed. Here, we demonstrate a water-incorporation strategy to expand the interlayer gap of α-MoO3, in which water molecules take the place of lattice oxygen of α-MoO3. Accordingly, the modified α-MoO3 electrode exhibits theoretical-value-close specific capacity (963 C g−1 at 0.1 mV s−1), greatly improved rate capability (from 4.4% to 40.2% at 100 mV s−1) and boosted cycling stability (from 21 to 71% over 600 cycles). A fast-kinetics dual-ion-intercalation energy storage device is further assembled by combining the modified α-MoO3 anode with an anion-intercalation graphite cathode, operating well over a wide discharge rate range. Our study sheds light on a promising design strategy of layered materials for high-kinetics charge storage.


Supplementary Tables
Supplementary Table 1 Raman peak assignment for MoO3 and e-MoO3. The results are also compared with the reported MoO3 and MoO3-x.

Supplementary Note 3
As observed in Supplementary Fig. 7, the rectangle frame indicates the peaks corresponding to Li compounds in Li-MoO3, Air-LiMoO3, and N2-LiMoO3. However, these peaks are not found in e-MoO3 based samples, which implies Li was removed from the crystal lattice during sonication and wash in water.

Supplementary Note 4
Supplementary Fig As shown in Supplementary Fig. 12, the valence band margin of α-MoO3 is 1.0 eV below the Fermi level, and the second electron cutoff is located at 166.5 eV. Thus, the work function is determined to be 3.4 eV for α-MoO3. By contrast, the work function of e-MoO3 is calculated to be 4.0 eV, which is slightly higher than that of α-MoO3. Moreover, the increment in band density near the Fermi level was reported to be assigned to the oxygen defects ( Supplementary Fig. 13). 26

Supplementary Note 6
Supplementary Fig. 14 compares the O K-edge NEXAFs spectra of α-MoO3 and e-MoO3. The high-intensity peak within 528-537 eV originates from the hybridization of O 2p with Mo 4d states, which is split into the t2g and eg (dz 2 , dx 2 −y 2 ) like peaks. In comparison with the eg peak of α-MoO3, the intensity increment in the eg peak of e-MoO3 is because of the increased density of conduction band by oxygen vacancies. 27,28 Supplementary Note 7 Compared with α-MoO3, α-MoO3-x·xH2O shows an additional valance band near valance band maximum, which shows a typical Mo-d characteristic (Supplementary Fig. 16). This newly formed band bridges the gap between valance band and conduction band, thus explains for the greatly improved conductivity of our e-MoO3 sample. Of note is that this newly formed valance band is in good agreement with the increment in valance band observed in Supplementary Fig. 13. The mixture was then stirred for 24 hours at room temperature under the protection of inert argon atmosphere.

Supplementary
The lithiated compound was separated by vacuum filtration and washed several times with hexane.
Subsequently, the sample was placed in 100 mL of distilled water and stirred for half an hour. Finally, the sample was centrifuged and dried in a vacuum oven at 80 °C.
SEM images show almost no difference between the morphologies of initial δ-MnO2 (Supplementary Fig.   24a and b) and treated δ-MnO2 (denoted as e-MnO2, Supplementary Fig. 24c and d). The electrochemical performance of δ-MnO2 and e-MnO2 as cathodes (loading mass of 1.6 mg cm −2 ) for Znion aqueous batteries was assessed in a two-electrode system with Zn foil as anode (Supplementary Fig. 25).
To prepare the electrodes, active materials (δ-MnO2 and e-MnO2), binder (polyvinylidene fluoride) and Super P carbon black were mixed with a weight ratio of 8:1:1. The mixture was dispersed in N-methyl pyrrolidone and coated on a piece of carbon cloth, followed by drying at 80 °C for 12 hours under vacuum. As shown in Supplementary Fig. 25a, the specific capacity of δ-MnO2 decays fast from 286 to 224 mAh g −1 during the first five charge/discharge cycles at 100 mA g −1 . In contrast, e-MnO2 depicts gradually increasing capacity from 250 to 270 mAh g −1 during the first five cycles at 100 mA g −1 . When the current density is increased to 1 A g −1 , the specific capacities of δ-MnO2 and e-MnO2 are 82 and 102 mAh g −1 , indicating the better rate capability of e-MnO2 than that of δ-MnO2. Moreover, e-MnO2 shows better cycling performance than δ-MnO2 ( Supplementary Fig. 25b). After 500 cycles at 1 A g −1 , the capacity of e-MnO2 first increases from 120 to 150 mAh g −1 and then keeps almost stable, while the capacity of δ-MnO2 gradually decreases from 113 to 93 mAh g −1 . All these results imply the feasibility of our H2O-incorporation strategy in improving the charge storage ability of birnessite δ-MnO2.

Supplementary Note 9
Identified by both CV and galvanostatic charge/discharge curves ( Supplementary Fig. 26), the anion intercalation into graphite displays high kinetic with a potential window of 3.5-5.3 V vs. Li + /Li. The capacity reaches 387 C g −1 at 0.1 mV s −1 , while 350 C g −1 is retained at 8 mV s −1 . Moreover, the capacity almost keeps the same during galvanostatic charge/discharge tests when the current density increased from 50 mA g −1 (capacity of 396 C g −1 ) to 500 mA g −1 (capacity of 392 C g −1 ). Graphite also demonstrates outstanding cycling stability with almost no capacity decay after 1000 charge/discharge cycles at 500 mA g −1 .

Supplementary Note 10
Supplementary

Supplementary Note 11
As shown in Supplementary Fig. 30, mass loading for e-MoO3 is 0.6 mg cm −2 , while the mass loading for graphite is 1.5 mg cm −2 . Since the loading mass of graphite is 2.5 times of that of e-MoO3, as well as the volume of graphite, is more than 5 times of that of e-MoO3 electrode, we conclude it would be further beneficiary to the full-cell energy density and power density if the cathode with comparable capacity could be developed.

Supplementary Note 12
The full e-MoO3//LiCoO2 Li-ion cell was assembled by using a two-electrode stainless steel Swagelok cell.
e-MoO3 coated on Cu foil and LiCoO2 coated on Al foil were used as anode and cathode, respectively, with a mass ratio of 1:2. Supplementary Fig. 33a presents the first three charge/discharge cycles of the e-MoO3//LiCoO2 device at 100 mA g −1 , revealing a reversible specific capacity of 309 C g −1 for the e-MoO3//LiCoO2 device within a voltage window of 0.2-3 V. CV curves were collected for the e-MoO3//LiCoO2 device at scan rates from 0.1 to 100 mV s −1 (Supplementary Fig. 33b). Supplementary Fig.   33c plots the specific capacity as a function of v −1/2 , which shows a close-to-linear shape. This means that the specific capacity of the device is significantly limited by the solid-state diffusion, and is dependent on the scan rate. This behavior is apparently different from the e-MoO3//graphite device, whose capacity is mostly independent on the scan rate. The Ragone plots of e-MoO3//LiCoO2 are calculated and compared with those of e-MoO3//graphite and other reported cation-ion batteries ( Supplementary Fig. 33d). Although the maximum energy density of e-MoO3//LiCoO2 device (146 Wh kg −1 ) is slightly higher than that of e-MoO3//graphite device (133 Wh kg −1 ), the power density of e-MoO3//LiCoO2 is far below that of e-MoO3//graphite device. Finally, the cycling stability of the e-MoO3//LiCoO2 device was assessed at 400 mA g −1 (Supplementary Fig. 33e). The device can be stably operated for about only 60 cycles. Afterwards, the capacity decays very fast at a capacity decay rate of 1.28 C g −1 cycle −1 . On the other hand, by employing commercial LiNi 1-x-y Co x Mn y O 2 (NCM) as cathode, the fabricated full Li-ion cell based on e-MoO 3 is hoped to achieve slightly higher energy density and better cycling stability. However, the power density is predicted to be still much lower than our e-MoO3//graphite device, as the rate performance of NCM is significantly lower than anion-intercalation graphite. Therefore, we conclude that the limitation of constructing high-rate Li-ion batteries based on e-MoO3 anode lies in the unsatisfying performance of cathode, especially the rate performance. Developing cation-intercalation cathode with high discharge potential, large rate capability and long-term cycling stability would be beneficial for further boosting the performance of Li-ion batteries based on e-MoO3 anode.