High Voltage Magnesium-ion Battery Enabled by Nanocluster Mg3Bi2

This study demonstrates a cost-effective route to fabricate stable and high voltage rechargeable Mg-ion battery potentially for grid-scale ... are man...
2 downloads 3 Views 1MB Size
Download PDF
Subscriber access provided by AUSTRALIAN NATIONAL UNIV

Article

High Voltage Magnesium-ion Battery Enabled by Nanocluster Mg3Bi2 Alloy Anode in Noncorrosive Electrolyte Yi-Hong Tan, Wei-Tang Yao, Tianwen Zhang, Tao Ma, Lei-Lei Lu, Fei Zhou, Hong-Bin Yao, and Shu-Hong Yu ACS Nano, Just Accepted Manuscript • DOI: 10.1021/acsnano.8b01847 • Publication Date (Web): 27 Apr 2018 Downloaded from http://pubs.acs.org on April 28, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

High Voltage Magnesium-ion Battery Enabled by Nanocluster Mg3Bi2 Alloy Anode in Noncorrosive Electrolyte Yi-Hong Tan,†,‡ Wei-Tang Yao,*,‡ Tianwen Zhang,† Tao Ma,† Lei-Lei Lu,† Fei Zhou,† Hong-Bin Yao,*,† Shu-Hong Yu† †

Division of Nanomaterials & Chemistry, Hefei National Research Center for Physical Sciences at the Microscale, CAS Center for Excellence in Nanoscience, Hefei Science Center of CAS, Collaborative Innovation Center of Suzhou Nano Science and Technology, Department of Chemistry, University of Science and Technology of China, Hefei 230026, China. ‡

Key Subject Laboratory of National Defense for Radioactive Waste and Environmental Security, Southwest University of Science and Technology, Mianyang 621010, China. *Corresponding author: [email protected] and [email protected] ABSTRACT: Currently, developing high voltage (beyond 2 V) rechargeable Mg-ion batteries still remains a great challenge owing to the limit of corrosive electrolyte and low compatibility of anode material. Here we report a facile one step solid state alloying route to synthesize nano-clustered Mg3Bi2 alloy as a high-performance anode to build up 2 V Mg-ion battery using noncorrosive electrolyte. The fabricated nano-clustered Mg3Bi2 anode delivers a high reversible specific capacity (360 mAh g-1) with excellent stability (90.7% capacity retention over 200 cycles) and high Coulombic efficiency (average 98%) at 0.1 A g -1. The good performance is attributed to the stable nanostructures, which effectively accommodate the reversible Mg 2+ ions insertion/de-insertion without losing electric contact among clusters. Significantly, the nano-clustered Mg3Bi2 anode can be coupled with high voltage cathode Prussian Blue to assemble a full cell using noncorrosive electrolyte, showing a stable cycling (88% capacity retention over 200 cycles at 0.2 A g-1) and good rate capability (103 mAh g-1 at 0.1 A g-1 and 58 mAh g-1 at 2 A g-1). The energy and power density of as-fabricate full cell can reach up to 81 Wh Kg-1 and 2850 W Kg-1, respectively, which are both the highest values among the reported Mg-ion batteries using noncorrosive electrolytes. This study demonstrates a cost-effective route to fabricate stable and high voltage rechargeable Mg-ion battery potentially for grid-scale energy storage. KEYWORDS: Mg3Bi2 alloy, Nano clusters, Noncorrosive electrolyte, High voltage, Magnesium-ion battery Rechargeable Mg-ion battery is intriguing for future large scale stationary energy storage due to its low cost, no dendritic hazards, and high theoretical energy density.1-4 However, there are many challenges remained to be addressed for high performance rechargeable Mg-ion batteries, including incompatibility of anode-electrolyte-cathode system, formation of passivated solid-electrolyte interphase (SEI) layer, and sluggish-kinetics of Mg2+ ions transport in electrodes.5-7 In past few decades, the conventional electrolytes with similar compositions to that used in Li-ion batteries, such as Mg(ClO4)2, Mg(PF6)2 or Mg(TFSI)2 salts in carbonate or ether solvents have been tried in Mg-ion batteries and displayed very low compatibilities with electrodes.8-10 The formation of insulating SEI layer on the surface of Mg metal anode would block the diffusion of Mg2+ ions into Mg anode limiting reversible plating/stripping of Mg2+ ions. Presently, only several electrolytes could enable reversible plating/stripping of Mg2+ ions at the Mg metal anode, such as Grignard reagents (RMgX in ethers, R = organic alkyl or aryl group and X= halide like Cl or Br).11-12 Furthermore, to address the kinetics sluggish issue of Mg2+ ions in the cathode, the dual-salt electrolyte systems have been proposed.13-18 For example, Nazar et al.

fabricated a 2 V Mg metal-based battery using all-phenyl complex with lithium chloride (LiCl-APC) dual-salt electrolyte.19 Yao et al. reported a 2.2 V Mg-Na hybrid battery based on dual-salt electrolyte of [Mg2(µCl)2][AlCl4]2/dimethoxythane (DME) with NaAlCl4.20 However, these electrolytes are very air-sensitive, difficult to be synthesized and handled.21-22 Even worse, the existence of free halogen ions in the electrolytes are highly corrosive with commonly used current collectors (such as stainless steel and nickel), hindering their broad applications in low cost Mg-ion batteries beyond 2 V.14-15, 23-28 To enable the compatibility of conventional noncorrosive electrolytes in Mg ion batteries, alloying typed anode materials have been proposed to replace Mg metal anodes.29 Recent studies demonstrated that some p-block metals such as Bi,30-31 Sb,30 Sn,32 In33 and Pb34 are able to conduct reversible magesiation/de-magesiation at a low potentials. Bi is one of the most attractive anodes due to the low potential (~0.25 V vs. Mg2+/Mg) and high theoretical specific capacity (385 mAh g-1, surpassing the graphite anode in Li ion battery).35-36 The pioneering work on electrochemical alloying/de-alloying of Mg

ACS Paragon Plus Environment

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 11

Figure 1. The phase and electrochemistry characterizations of the directly alloying Mg 3Bi2. (a) Binary Mg-Bi phase diagram for choosing alloying temperatures. (b) PXRD patterns of products obtained at different alloying temperatures. (c) Typical charge/discharge voltage profiles of obtained products in half cells coupling with Mg metal anode and LiCl-APC electrolyte. (d) Specific discharge capacities and corresponding CEs of obtained products versus cycling number at current density of 0.1 A g-1.

in Bi anode was done by Arthur et al., in which the compatibility of Bi anode with conventional electrolyte has been demonstrated.30 To improve the rate performance and cycling stability of Bi anode, Shao et al. proposed Bi nanotubes to enhance the kinetics of Mg2+ ion diffusion and accommodate the large volume change during the insertion/de-insertion of Mg2+ ions.31 However, these proposed nanostructured Bi anodes have to be electrochemically magnesiated in advance to couple with conventional cathodes and electrolytes, which is not applicable for large scale grid scale energy storage system. Herein, we propose a directly alloyed nano-clustered Mg3Bi2 anode to enable a stable 2 V rechargeable Mg-ion battery using noncorrosive electrolyte and low cost current collectors. The alloying synthetic strategy for the nano-clustered premagnesiated Bi anode is firstly optimized according to the BiMg binary phase-diagram. The obtained nano-clustered Mg3Bi2 anode shows high reversibility for Mg2+ ion deinsertion/insertion without pulverization due to its nanoscale clustered structures. A variety of potential cathode materials such as the Prussian Blue (PB, Na0.61Fe[Fe(CN)6]0.94•1.08H2O)

nanocubes, hydrated vanadium pentoxide (V2O5) nanosheets37 and birnessite-manganese dioxide (δ-MnO2)38-39 are employed to verify the proof-of-concept Mg-ion full cell with noncorrosive magnesium(II) bis(trifluoromethanesulfonyl) imide (Mg(TFSI)2)-lithium bis(trifluoromethanesulfonyl)imide (LiTFSI)/acetonitrile (AN) dual-salt electrolyte. Attractively, the fabricated Mg3Bi2-PB full cell exhibits high discharge voltage at ~2 V and excellent reversible specific capacity of 92 mAh g-1 over 200 cycles at high current density of 0.2 A g-1.

RESULTS AND DISCUSSION The nano-clustered Mg3Bi2 alloy anode was synthesized via a facile one step solid state alloying reaction using commercially available Bi (≥200 mesh) and Mg (100-200 mesh) powders as raw materials, which is suitable for large scale production (Figure S1). We set the stoichiometric ratio of MgBi as 3.3:2 (with 10 wt.% excess of Mg to compensate the evaporation loss) and selected different alloying temperatures (300 oC, 500 oC, 650 oC and 750 oC, denoted as MB-300, MB500, MB-650 and MB-750, respectively)

2 Environment ACS Paragon Plus

Page 3 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

Figure 2. Nanoscale morphologies characterization of products obtained at different alloying temperatures. (a-d) SEM image of MB-300, MB-500, MB-650 and MB-750, respectively. (e-h) TEM image of MB-300, MB-500, MB-650 and MB-750, respectively.

to investigate the alloying process of Mg and Bi powders. These alloying temperatures were chosen according to the specific points such as the melting point of Bi or Mg in the Mg-Bi binary phase diagram (Figure 1a). The obtained phases were firstly characterized by the powder X-ray diffraction (PXRD, Figure 1b) to reveal the phase purity. At 300 oC, most Mg and Bi powders were at solid state and the alloying reaction was processed to a limited extent resulting in separated Bi and Mg phase, and a little very low crystalline Mg3Bi2 phase in the product. As the reaction temperature raised up to 500 oC, the Bi powder would melt and alloy with Mg at the surface of Mg particles to form the hexagonal structure Mg3Bi2 (JCPDS no.80-1098), while, the core of Mg particles would still maintain single Mg phase. When the temperature further increased to 650 oC, the obtained product was pure Mg3Bi2 alloy without impurity. However, with the reaction temperature increasing to 750 oC, the phase of Bi gradually appeared in the obtained product, which can be ascribed to too much evaporation loss of Mg at such high temperature.40 The Mg2+ ions storage behaviors of directly alloying synthesized products were evaluated in half-cells by adopting Mg foils as the counter electrodes and the LiCl-APC as the electrolytes. The cells were charged to 1.6 V (de-magnesiation process) and then discharged to 0.05 V (magnesiation process) at the current density of 0.1 A g-1. The typical charge/discharge voltage profiles are shown in Figure 1c. It can be seen that the MB-300 electrode shows largest overpotential (230 mV) and delivers the lowest capacity (21 mAh g-1) in comparison to other electrodes, which is consistent with the PXRD result that the Mg particles were not fully alloyed with Bi at 300 oC. With the alloying reaction temperature increasing, the alloyed products were generated and thus, accordingly, the electrochemical behavior of MB-500 electrode exhibits normal demagnesiation/magnesiation processes with low overpotential and enhanced specific capacity (189 mAh g-1). With the completely alloying at 650 oC, it could be seen that the MB-650 electrode exhibits highest charge/discharge specific capacity reaching up to 350 mAh g-1. However, with further increasing the reaction temperature to 750 oC, the specific capacity of asobtained MB-750 electrode decreases to 260 mAh g-1, which can be ascribed to the inadequate alloying of Bi due to the evaporation loss of Mg at 750 oC. It is worth noting that the

initial charge voltages of alloyed electrodes are a little higher than that in following cycles indicating that an activation process is necessary for the highly crystalline electrodes obtained by the directly alloying reactions (Figure S2). In particular, the initial charge voltage plateau of MB-300 is lower than other alloys because of the low crystallinity of as-formed Mg3Bi2. And the existence of low content of low crystalline Mg3Bi2 in MB-300 was confirmed by the high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) image and corresponding elementary mappings (Figure S3) as well. Furthermore, the long-term cycling stabilities of as-obtained electrodes were tested as well (Figure 1d). The results reveal that all electrodes except the MB-300 behaved an initial activation process with the increase of specific capacities. Among all electrodes, the MB-650 electrode delivered the highest specific capacity of 360 mAh g1 after initial activation cycles with a following slow capacity decay to 263 mAh g-1 over 200 cycles, corresponding to 90.7% capacity retention with respect to the initial capacity (290 mAh g-1), which is better than that of MB-750 (90.5%) and MB-500 (84.9%). In addition, the MB-300 electrode displayed much low Coulombic Efficiency (CE) in initial tens of cycles comparing to other electrodes because of very low alloying extent of Mg in this electrode. Notably, the average CE of MB-650 can reach up to 98.4% in 200 cycles indicating highly reversible insertion/deinsertion of Mg2+ ions in this electrode. The electrochemical impedance analysis of as-fabricated electrodes (Figure S4) indicated the lowest ions transfer resistance (875Ω) at the interfaces of MB-650 electrode, which is also consistent with the best electrochemical performance of MB-650 electrode implying that the Mg2+ ions transport in this electrode was improved mostly. Beside the alloy phase purities of as-synthesized electrode materials, the nanostructures of Mg3Bi2 alloy particles are believed to be different and highly related to their electrochemical performances. As shown in Figure 2a, the MB300 electrode material presents out original micro-size particle morphologies of Mg and Bi powders because very little alloying reaction occurred at 300 oC. The transmission electron microscopy (TEM) image and the selected area electron diffraction (SAED) pattern (Figure S5a)

3 Environment ACS Paragon Plus

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 11

Figure 3. Ex-situ PXRD and STEM characterizations of MB-650 electrode at different electrochemical states. (a) PXRD patterns of MB650 electrodes recorded along the first charge/discharge profiles. (b-e) STEM images and the corresponding Mg and Bi elementary mappings at (b) open-circuit voltage, charged to (c) 0.75 V, (d) 1.6 V, and finally discharged to (e) 0.05 V. (f) A schematic illustration of deinsertion/insertion Mg2+ ions in Mg3Bi2 alloy particles (inset high-resolution TEM of images, left: Mg3Bi2 nanoparticle in (e) and right: Bi nanoparticle in (d).

further confirmed the solid single crystal microparticles in the MB-300 electrode (Figure 2e). The Mg2+ ions conduction is highly limited in the micro-size particles due to the low conductivity of Mg2+ ions in bulk materials. Thus, the specific capacity delivered by MB-300 is the lowest one among assynthesized electrode materials. The scanning electron microscopy (SEM) image and TEM image of MB-500 electrode material is shown in Figure 2b and f, respectively. The synthesized electrode material are still solid microparticles (Figure 2b) and the inhomogeneous contrast inside the particles shown in Figure 2f indicates that the particles are alloyed polycrystalline, which is also confirmed by SAED analysis (Figure S5b). But the microscale size of particles still limited the conduction of Mg2+ ions in the electrode resulting in much lower specific capacity of MB-500 in comparison to MB-650 and MB-750. As shown in Figure 2c, the morphologies of MB650 electrode materials are totally different from that of MB300 and MB-500, which present out the porous microparticles with secondary nanocluster structures inside. The TEM image (Figure 2g) further indicates that the nanoparticles in the clustered structures are in the size of 100~200 nm. This nanocluster structure would facilitate the infusion of electrolyte into electrode particles and increase the contact area between electrolyte and electrode, meanwhile, the nanoparticles inside the clusters could also sustain the volume change and strain during the magesiation/de-magesiation processes. Thus, the MB-650 electrode delivered the highest specific capacity and excellent cycling stability among the MB-series electrode materials. Although MB-750 electrode (Figure 2d, h) has the similar nanostructures with that of MB-650, the delivered specific capacity is not the highest one due to the insufficient alloying of Bi particle as confirmed by the PXRD result (Figure 1b). In all, the excellent electrochemical performance of MB-

650 electrode can be ascribed to its well alloying crystalline and nano-cluster structures. Due to its best performance, the electrochemical processes of MB-650 electrode were further analyzed in combination with ex-situ PXRD and TEM. Firstly, the cyclic voltammetry (CV) scan at 0.1 mV s-1 (Figure S6a) shows that an anodic peak at 0.85 V appeared in the initial oxidation process which is corresponding to the de-insertion of Mg2+ ions from the alloyed type Mg3Bi2 anode. In the following reverse reduction process, the cathodic peak at ~0.07 V is related to the insertion of Mg2+ ions back into Bi clusters. In the second cycle, the anodic peak moved to a lower value of ~0.5 V and cathodic peak rose to ~0.1 V, which are mainly caused by the electrochemical modification of the grains of alloyed Mg3Bi2 particles after initial redox process (known as the electrochemical grinding)41. In subsequent redox cycles, the peaks corresponding to Mg deinsertion/insertion slowly evolved, along with gradually increased peak current density (the CV curve stabilized after ~10 cycles), indicating an activation process with the increase of specific capacity in the cycling test as shown in Figure 1d. Typical charge/discharge voltage profiles (Figure S6b) show that the charge voltage plateau of first cycle is at 0.8 V, and the charge plateaus of following cycles is at 0.4 V, which is consistent with the activation process as confirmed by CV curves. The MB-650 electrode also exhibits excellent rate capability (Figure S6c). The discharge capacity is 365 mAh g−1 at 0.038 A g-1. When the current density gradually increased to 1.9 A g-1, the MB-650 could still remain a specific capacity of 251 mAh g-1. Figure S6d shows the long-term galvanostatic charge/discharge cycling performance of MB-650 electrode at 0.38 A g-1, which indicates a good stability with a specific capacity of ~312 mAh g-1 at initial cycles and retained the capacity of 209 mAh g-1 over 300 cycles.

4 Environment ACS Paragon Plus

Page 5 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

The crystalline structural variation of MB-650 electrode involved in the electrochemical alloying/de-alloying process was further investigated by ex-situ PXRD and TEM. As shown in Figure 3a, in the initial charge progress, the PXRD peaks correspond to Mg3Bi2 gradually disappeared while the peaks of Bi (JCPDS no. 44-1246) appeared and kept almost unchanged when the cell voltage raised up to 1.6 V. In the subsequent discharge process to 0.05 V, PXRD peaks of Mg3Bi2 were recovered while the peaks of Bi disappeared, indicating the well crystalline structure recovery with the insertion of Mg2+ ions into Bi hosts. To further understand the advantage of Mg3Bi2 electrode with nano-clustered structure, we used ex-situ highangle annular dark field scanning transmission electron microscope (HAADF-STEM) and energy-dispersive X-ray (EDX) elementary mapping to track its morphology and composition variations during the de-insertion/insertion processes of Mg2+ ions (Figure 3, b-e). In the pristine state, the coexistence of Mg and Bi elements were confirmed by the elementary mappings, in which the Mg and Bi mapping signals are well overlapped with the nanoparticle morphology (Figure 3b). When the MB-650 electrode was charged to 0.75 V, the mapping signal of Mg became weaker (Figure 3c), which is agreement with the de-insertion of Mg2+ ions at this voltage. After further charging the MB-650 electrode to 1.6 V, the Mg mapping signal became the weakest while the Bi nanoparticle

morphology and corresponding mapping signal remained unchanged, indicating the good stability of Bi host at nanoscale (Figure 3d). Upon discharging the MB-650 electrode to 0.05 V, the intensity of Mg mapping signal was almost recovered to the same level of MB-650 at the initial state (Figure 3e), which is also confirmed by X-ray photoelectron spectroscopy (XPS, Figure S7) that a majority of Mg and a little of Li re-inserted into the Bi host. During the whole charge/discharge progresses, the Bi mapping signal intensity kept the same indicating the stability of Bi host for insertion and de-insertion of Mg2+ ions. The variation of Bi/Mg ratio is also revealed in the corresponding EDX spectra (Figure S8), which is consistent with the signal intensity variation of elementary mappings. Combining the PXRD, STEM image and corresponding elemental mappings, we can speculate that nano-clustered Mg3Bi2 electrode is good to store/release Mg2+ ions reversibly without a noticeable nonreversible deterioration. The phase transformation of MB-650 electrode during the charge/discharge process is schematically illustrated in Figure 3f. With charging to 1.6 V, the phase of MB-650 alloy changed to the rhombohedral Bi (R-3m) and the lattice spacing of 3.2 Å shown in the high resolution TEM image is corresponding to that of (012) planes42. In the subsequent discharging process, with the re-insertion of Mg2+ ions into the Bi nanoparticles, the crystal structure of the electrode was re-magnesiated to the

Figure 4. Electrochemical characterizations of PB-MB-650 full cell. Discharge specific capacities and CEs versus cycling numbers of PBMB-650 full cells with (a) 0.5 M/ 1 M/ 2 M Mg(TFSI)2-1 M LiTFSI/AN electrolytes and (b) 0.1 M/ 0.5 M/ 1 M LiTFSI-2 M Mg(TFSI)2/AN electrolytes, respectively. (c) CV curves (Scan rate: 0.1 mV s-1) and (d) galvanostatic discharge/charge voltage profiles of PB-MB-650 full cell with 1 M LiTFSI-2 M Mg(TFSI)2/AN electrolyte, respectively. (e) Cycling performance at 0.2 A g-1 and corresponding CE, and (f) rate performance of PB-MB-650 full cell using 1 M LiTFSI-2 M Mg(TFSI)2/AN electrolyte, respectively.

5 Environment ACS Paragon Plus

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Mg3Bi2 and the lattice spacing of 2.3 Å is corresponding to the d-spacing of (110) planes of hexagonal Mg3Bi2 (P-3ml). Moreover, TEM images in Figure S9 further show that there is no obvious morphological change at different voltages, which indicates that such nano-clustered structure can withstand the volume shrinkage/expansion of Mg3Bi2 alloy with deinsertion/insertion of Mg2+ ions, leading to substantially good cycling stability and high CE. The crystalline structure and composition stability of MB-650 alloy anode after different numbers of cycles were also characterized by PXRD.43 As shown in Figure S10, the Mg3Bi2 alloy anode maintains a good crystalline structure. However, after numbers of cycles the PXRD peaks of Mg3Bi2 alloy became weak and the peaks of Bi appeared, which indicates that Mg2+ cannot be fully re-inserted into Bi hosts after multiple cycles possibly due to the deactivation of Bi particles. This is also consistent with the capacity decay of the Mg3Bi2 alloy anode The applicability of Mg3Bi2 alloy anodes with conventional noncorrosive electrolytes and high voltage cathodes is very attractive to fabricate high energy density Mg-ion batteries. It is well known that the transportation pathways of Mg2+ ions into Mg metal anode are blocked due to the formation of poor Mg2+ ion conductive layer on the surface of Mg metal anode in conventional electrolytes24, 31. In our synthesized MB-650 electrode, the insertion voltage (196 mV vs. Mg2+/Mg) of Mg2+ ions in Bi host is a little higher than that of Mg2+ deposition voltage (-200 mV vs. Mg2+/Mg)9 on the Mg metal anode, which can prevent the formation of passivation layer, implying the compatibility of MB-650 electrode with conventional electrolyte44. To evaluate the performance of the proof-of-concept Mg-ion battery with conventional noncorrosive electrolyte, the full Mgion cell using MB-650 anode to couple with high voltage Prussian Blue (PB) cathode was tested in a stainless-steel coin cell (denoted as PB-MB-650). First of all, the high quality PB nanocrystals were synthesized via a previously reported method45. The obtained PB nanocrystals present high crystalline cubic morphologies (Figure S11a, b) and high pure phase with a low content of crystalline water (Figure S11c, d). In general, the active mass ratio of PB cathode to Mg3Bi2 anode was set as 1:1.3 in the full cell because the specific capacity of the full cell decayed very fast if the mass ratio of anode to

Page 6 of 11

cathode was 1:1 (Figure S12). Noting that the specific capacities of full cells are calculated based on the mass of cathode active materials. Since Mg2+ ions are hard to intercalate into the PB hosts at first due to their strong solvation effect with the electrolyte solvent, we introduced a monovalent cation, Li+, into the electrolyte by adding the salt of LiTFSI 19. To explore the impact of concentrations of salts in electrolytes on the electrochemical performance of as-fabricated full cells and find out an appropriate electrolyte, the cycling performance of PBMB-650 full cell was evaluated under different concentrations of Mg(TFSI)2 and LiTFSI in acetonitrile. As shown in Figure 4a, as the concentration of LiTFSI salt increases at a constant concentration of Mg(TFSI)2 at 2 M, the PB-MB-650 full cell exhibits improved CE and cycling stability. When the concentration of LiTFSI in the electrolyte reached to 1 M, the CE of the cell reached up to nearly 99% and the specific capacity of PB maintained at 97 mAh g-1 after 100 cycles. The concentration of Mg(TFSI)2 in the electrolyte is also important for the efficiency and stability of as-fabricated full cell. As shown in Figure 4b, as the concentration of Mg(TFSI)2 increasing to 2 M, the full cell behaved the highest CE and best cycling stability. In the 2 M Mg(TFSI)2-1M LiTFSI/AN electrolyte, the full cell displayed outstanding cycling stability at 0.1 A g-1, with only 4.5% of capacity decay after 100 cycles. Thus, the 2 M Mg(TFSI)2-1M LiTFSI/AN electrolyte was chosen as standard electrolyte for the following tests. It has been known that Mg3Bi2 anode presents a charge/discharge voltage plateau between 0.2-0.4 V (vs. Mg2+/Mg) in the half-cell (Figure 1c) and PB cathode displays one step charge/discharge voltage plateau between 1.6-2.9 V (vs. Mg2+/Mg)19. Thus, the working voltage of our fabricated PB-MB-650 full cell lies in the range of 1.0-2.5 V, which is also confirmed by the CV curves. Figure 4c presents the CV curves of PB-MB-650 full cell for initial three cycles at a scan rate of 0.1 mV s-1, from which we can see that the cathodic peak of initial cycle is centered at approximately 1.6 V and, subsequently, highly symmetric peak profiles are observed, showing good reversibility of both anodic (2.0 V) and cathodic (1.8 V) processes of the full cell. The galvanostatic charge/discharge voltage profiles of PB-MB-650 full cell of the initial three cycles at 0.1 A g-1 are shown in Figure 4d. At the first cycle, the discharge and charge specific capacity of full cell

Figure 5. The comparison of our cells with previously reported cells. (a) The Ragone plot (green area represents corrosive electrolyte systems and blue area represents noncorrosive electrolyte systems). (b) Charge/discharge cycles and capacity retention comparison bet ween our works and previously reported studies. The data summary for the comparison please see supplementary Table S1.

6 Environment ACS Paragon Plus

Page 7 of 11 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Nano

reached to 101 and 103 mAh g-1, respectively, with an initial CE of 98%. After the first full cycle, the PB-MB-650 cell was disassembled and the PB cathode was characterization by PXRD, STEM and corresponding EDX elementary mappings, respectively. As shown in Figure S13, PXRD patterns of PB cathode at the discharge and charge states are the same indicating the phase stability of PB during the insertion and deinsertion processes of Li+ and Mg2+ ions. The elementary mappings of PB (Figure S14) indicate that Mg2+ ions were inserted into the PB nanocubes at initial fully discharge state of 1 V. More attractively, the PB-MB-650 full cell exhibits a superior stable cycling performance at high current densities of 0.2 A g-1 for 200 cycles with a capacity retention of 88% and high average CE of 98% (Figure 4e). The PB-MB-650 full cell also exhibits excellent rate performance. As shown in Figure 4f, the discharge specific capacity of PB-MB-650 full cell at 0.1 A g-1 is ~99 mAh g-1 and gradually decreases to ~94 mAh g-1, ~86 mAh g-1 ~76 mAh g-1 and ~58 mAh g-1 when the current density increases from 0.1 A g-1 to 0.2 A g-1, 0.5 A g-1, 1 A g-1 and finally to 2 A g-1, respectively. When the current density changes back to 0.1 A g-1, the discharge specific capacity of PB-MB-650 full cell immediately recovers to ~97 mAh g-1 indicating the good rate capability of the PB-MB-650 full cell. To further demonstrate the versatility of MB-650 anode to couple with other high voltage cathodes for high energy density Mg-ion batteries using conventional noncorrosive electrolyte. V2O5 (VO) nanosheets and birnessite-MnO2 (MO) cathode materials were synthesized and used to couple with MB-650 anode. As shown in Figure S15 (a-c), the pure phase VO nanosheets have been successfully prepared. When they were coupled with MB-650, the full cell delivered high specific capacity of 268 mAh g-1 (based on the mass of VO) in the working voltage range of 0.8-2.7 V and displayed relatively stable cycling in 50 cycles at 0.2 A g-1 (Figure S15d-f). The high pure phase flower-like MO cathode materials were also prepared (Figure S16a-c). Comparably, the MO-MB-650 full cell delivered an initial specific capacity of 107 mAh g-1 in the working voltage range of 1.0-2.7 V and gradually faded to 55 mAh g-1 (based on the mass of MO) after 50 cycles at 0.2 A g-1 (Figure S16d-f). The EIS were tested on the PB-MB-650, VOPB-650, and MO-PB-650 full cells at the initial open circuit voltage to reveal their difference in charge transfer resistances (Figure S17). It was found that the interfacial resistance of PBMB-650 full cell is much smaller than that of VO-MB-650 and MO-MB-650 proving the best electrochemical performance of PB-MB-650 full cell. Notably, the energy density and span-life of our fabricated Mg-ion batteries are comparable or even better than previously reported Mg-ion batteries. On the basis of total weight of both active electrode materials, the specific energy densities of PBMB-650, VO-MB-650, and MO-MB-650 full cells are calculated as 81 Wh kg-1 (0.1 A g-1), 174 Wh kg-1 (0.2 A g-1), and 72 Wh kg-1 (0.2 A g-1), respectively (calculation details please see methods). Although several previously reported Mgion batteries (denoted in the green area in Figure 5a) displayed higher energy density, these battery systems employed highly air sensitive and corrosive electrolytes, which will heavily increase the cost and hazards of Mg-ion batteries. Only considering the noncorrosive electrolyte systems (denoted in the blue area in Figure 5a), our fabricated PB-MB-650 full cell has the highest energy density. More attractively, the power density of PB-MB-650 is also the highest one reaching up to 2850 W kg-1. The life-spans of our fabricated Mg-ion batteries are compared with previously reports in Figure 5b as well. The

PB-MB-650 full cell shows a good specific capacity retention (88.5%) up to 200 cycles.

CONCLUSIONS In summary, we report a facile directly alloying strategy to synthesize nano-clustered Mg3Bi2 alloy anode material for a low-cost and high voltage Mg-ion battery operating in the noncorrosive electrolyte. The MB-650 anode can be also coupled with various high voltage cathode materials including Prussian Blue, V2O5, and MnO2. Attractively, the fabricated PB-MB-650 full cell behaved high discharge voltage at ~2.0 V and delivered specific capacity of 92 mAh g-1 with a low capacity fade rate of less than 0.06% per cycle at 0.2 A g -1. Importantly, the energy density of as-fabricated PB-MB-650 full cell can reach up to 81 Wh kg-1, indicating its potential application for grid-scale energy storages

METHODS Synthesis of Mg3Bi2 alloy powders. Typically, the alloy compound Mg3Bi2 was prepared by mixing magnesium powder (99.5%, Macklin, 100-200 mesh) and bismuth powder (99.99% metals basis, Aladdin, ≥200 mesh) in a molar ratio of 3.3:2. A 10% weight excess of Mg is needed to balance the Mg evaporation loss during the alloying reaction process 40. The mixture was loaded in a tantalum crucible and heated at desired temperatures for 30 minutes under Ar atmosphere and then cooled down to room temperature naturally. The products were directly collected for the characterizations and electrochemical tests without further purifications. Synthesis of Prussian Blue nanoparticles. Prussian blue Na0.61Fe[Fe(CN)6]0.94•1.08H2O was synthesized according to a previously reported procedure45. Briefly, 2 mmol of Na4Fe(CN)6·10H2O (AR, Aladdin) and 1 mL of hydrochloric acid (37%, Aladdin) were added in 100 mL of deionized water and stirred vigorously to obtain a homogenous solution. The solution was then maintained at 60 °C for 4 hours under vigorous stirring to yield Na0.61Fe[Fe(CN)6]0.94•1.08H2O nanotubes. The product was collected by filtration, washed by deionized water and ethanol for three times and dried at 120 °C in a vacuum oven for 12 hours. Synthesis of V2O5 nanosheets. 0.5 g of V2O5 powder (99.8%, Alfa-AESAR) was added into 7.5 mL of deionized water and 2.5 mL of H2O2 (30 wt.% in H2O, Aladdin) to form a solution with a V2O5 concentration of ~ 0.3 M (n(H2O2): n(V) =8:1). The resulting solution was stirred for 15 min while kept in water bath at the room temperature and then sonicated for 15 min for the reactions. This solution was later diluted through adding 40 mL of deionized water and then sonicated for about 80 min until the solution turned into brownish red V2O5 gel. This gel was further dispersed and diluted through adding 50 mL of deionized water and stirred in de-ionized water until a homogenous red-colored, viscous solution was formed. This solution was freeze-dried for 2 days in a Labconco FreeZone 1 L freeze dryer. After drying, the V2O5 cryogel was annealed under ambient atmosphere at 400 °C for 1 hour to form V2O5 nanosheets. Synthesis of birnessite-MnO2. δ-MnO2 was synthesized using a typical hydrothermal method. Firstly, KMnO4 and KCl were dissolved in 40 mL of deionized water with a certain proportion (n(KMnO4): n(KCl) = 3:1). The mixed aqueous solution was stirring continuously for 10 min at room temperature. Then the resultant product was transferred into a Teflon-lined stainlesssteel autoclave (50 mL) and heated to 160 °C for 12 hours.

7 Environment ACS Paragon Plus

ACS Nano 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Subsequently, the autoclave was cooled down to room temperature naturally. The dark precipitate powders were collected and washed with distilled water and ethanol for several times and dried at 60 °C for 12 hours. Electrolyte preparations. All chemical preparations and electrochemical measurements were carried out under pure argon atmosphere in M. Braun, Inc., glove boxes (