Abstract: Introduction Rechargeable batteries with high energy density have been required for recent applications such as hybrid electric vehicles and large-scale energy storage. Rechargeable batteries based on fluoride shuttle (FSBs) have attracted much attention as a candidate above, owing to their high theoretical energy density by selecting appropriate metal/metal fluoride couples as active materials of negative- and positive-electrode[1-3]. One of the important things of the battery is adoption of the good conductive media for fluoride species, that is, an electrolyte[4] . Another is reversible fluorination/defluorination of active materials at the interface with electrolyte. Previously, we have demonstrated archetypal reversible FSB using a liquid electrolyte containing organic fluoride at room temperature[3]. However, this battery needed to improve the utilization and cyclability of its active material. Since the common metal fluoride shows low electronic- and ionic-conductivity, the active materials should be minified for FSBs. In this study, metal fluoride nanoparticle, e.g. BiF 3 , was prepared via sol–gel method using a trifluoroacetate precursor for improvement of the charge/discharge reactions in FSB. Bismuth trifluoride having 302 mAh/g in theoretical capacity is considered as a candidate for positive active material[1-6]. Then, we studied reversible defluorination/fluorination reaction of thus prepared BiF 3 nanoparticles in a liquid electrolyte of FSB at room temperature, and valence change of bismuth species in the consequent electrode was evaluated using synchrotron ex-situ XANES measurement. Experimental Bismuth trifluoride nanoparticle was prepared from bismuth trifluoroacetate precursor via decarboxylation at 453 K in an oleylamine solution under argon atmosphere. At that time, specific amount of trifluoroacetic acid and/or trifluoroacetic anhydride were added as a supplementary fluorine source. After the nanoparticles were precipitated via addition of toluene and ethanol sequentially, the resulting precipitation was centrifuged and then re-dispersed in toluene containing oleic acid as a dispersant. Finally, nanoparticles were obtained after stripping of ligand using a trimethyloxonium tetrafluoroborate. Size of the resulting nanoparticles can be controlled by precursor concentration, decarboxylation reaction time and its temperature. Electrochemical measurement was carried out using a three-electrode cell with a VSP-300 BioLogic potentiostat/galvanostat in an argon-filled glove box at room temperature. The cell was composed of a carbon cloth as a counter electrode, a silver wire as a quasi-reference electrode, and an ionic liquid containing an organic fluoride as a liquid electrolyte. The liquid electrolyte was prepared by dissolving N -methyl- N -propylpiperidinium fluoride (MPPF) or N,N,N -trimethyl- N -neopentylammonium fluoride (NpF) in N,N,N -trimethyl- N -propylammonium bis(trifluoromethanesulfonyl)amide (TMPA-TFSA) with molar ratio of 1:10. Nanoparticles-supported carbon cloth electrode as a working electrode was prepared by soaking the cloth in BiF 3 -nanoparticles/acetonitrile dispersion. The prepared nanoparticles via the sol–gel method and the defluorination/fluorination products in electrochemical measurements were evaluated using XRD, XPS, SEM-EDS, and so on in our laboratory. Furthermore, valence change of bismuth species was evaluated using synchrotron ex-situ XANES measurement at SPring-8, Japan. Results and Discussion Thus obtained nanoparticles via above-mentioned procedure had a few tens of nanometers in diameter and showed low crystallinity. Moreover, it was suggested that small amount of oxygen would be doped into the nanoparticles according to elemental analyses. Then, non-doped BiF 3 nanoparticles having hexagonal crystalline structure had been acquired after further fluorination-treatment. Consequent BiF 3 nanoparticles-supported carbon cloth showed ca. 240 mAh/g (80%) as an initial discharge capacity, and retained ca. 100 mAh/g for more than five cycles at room temperature. Furthermore, in ex-situ bismuth L 3 -edge XANES spectra, valence of bismuth species changed from trivalent to zero (metal) during the defluorination reaction and similar reverse change was elucidated in fluorination reaction. In this way, the size of the active material was minified by the preparation method, thereby improving the utilization and cyclability in FSB. This method will be also applied for other positive- and negative-active material having high gravitational theoretical capacity. Acknowledgements This work was supported by the “Research and Development Initiative for Scientific Innovation of New Generation Batteries 2 (RISING2)” of the New Energy and Industrial Technology Development Organization (NEDO), Japan. References [1] M.A. Reddy and M. Fichtner, J. Mater. Chem. , 21 , 17059 (2011). [2] F. Gschwind, G. Rodriguez-Garcia, D.J.S. Sandbeck, A. Gross, M. Weil, M. Fichtner, N. Hörmann, J. Fluorine Chem. , 182 , 76 (2016). [3] K. Okazaki, Y. Uchimoto, T. Abe, Z. Ogumi, ACS Energy Lett. , 2 , 1460 (2017). [4] V.K. Davis et al., Science , 362 , 1144 (2018). [5] T. Yamanaka, K. Okazaki, T. Abe, K. Nishio, Z. Ogumi, ChemSusChem , 12 , 527 (2019). [6] T. Yamanaka, K. Okazaki, T. Abe, Z. Ogumi, ACS Appl. Energy Mater. , DOI:10.1021/ acsaem.9b01803.

Abstract: Introduction Fluoride shuttle batteries (FSBs) are attracting interest as candidates for a next generation battery with high energy density.[1-3]FSBs utilize the following redox reactions. MF x + x e - → M + x F - (at the positive electrode, M being metal) (1) M’ + y F - → M’F y + y e - (at the negative electrode, M’ being metal) (2) To realize high battery performance, it is necessary to elucidate and control reactivity and mechanisms in FSB reactions. BiF 3 is a strong candidate for a cathode material in FSBs. In the present work, structural transformation, reactivities and mechanisms in FSB reactions of BiF 3 single microparticles were studied by insitu Raman microscopy. BiF 3 microparticles were partly embedded in a gold plating film (BiF 3 /gold, orthorhombic BiF 3 (o-BiF 3 ) and cubic BiF 3 (c-BiF 3 )). By using a Raman cell consisting a BiF 3 /gold cathode, a Pb wire anode and an ionic liquid-based electrolyte,[2] reactivities and mechanisms in defluorination were found to be different for o-BiF 3 and c-BiF 3 . Experimental BiF 3 /gold was prepared by embedding o-BiF 3 microparticles (Fluorochem Ltd.) in a gold plating film on a gold foil. [4,5] Briefly, a gold foil and gold plating solution for deposition of 24 K gold were put into a small vessel, and then a small amount of o-BiF 3 powder was put in the solution. Then a platinum wire was placed in the solution and a current was applied between the wire and the gold foil. The amount of deposited gold corresponded to that of the gold plating film with a thickness of 1 μm without BiF 3 . An organic fluoride (1-methyl-1-propylpiperidinium fluoride: MPPF) was dissolved in an ionic liquid (N,N,N-trimethyl-N-propylammonium bis(trifluoromethanesulfonyl)amide: TMPATFSA) at a molar ratio of 1:10 and the resultant solution was used as an electrolyte.[2] The electrochemical cell had a quartz window and a Pb wire as a counter electrode. The distance between BiF 3 /gold and the window was 30 μm to 4 μm, and the space between them was filled with the electrolyte. The cell was sealed with Kalrez and placed in flowing Ar during Raman microscopy measurements. Raman microscopy was conducted with an NRS-4500 Raman spectrometer (JASCO Corporation) and 532 nm laser light. Results At OCV (0.7 V), o-BiF 3 gradually transformed into cubic c-BiF 3 . When the voltage of the BiF 3 /gold cathode vs the Pb anode was decreased from OCV to 0.05 V step by step, direct defluorination (eq. (1)) of the surfaces of only o-BiF 3 started from their contours at 0.45 V (Figs. 1a and 1b) and then extended to their center parts and was mostly completed at 0.2 V. The results suggest that the rate-limiting process of direct defluorination of o-BiF 3 is electronic conduction at the surface of o-BiF 3 . Then defluorination of c-BiF 3 started at a voltage below 0.2 V by both direct defluorination (eq. (1)) and dissolution-deposition mechanisms (BiF 3 → Bi 3 + + 3F − , Bi 3 + + 3e − → Bi). In the direct mechanism, the nucleus of Bi first appeared near the edge of c-BiF 3 microparticles, and the nucleus grew for defluorination to proceed over the whole surface, suggesting that the rate limiting process is formation of the nucleus of Bi. The direct defluorination of c-BiF 3 was much slower than that of o-BiF 3 . Defluorination of c-BiF 3 by the dissolution-deposition mechanism was fast and dominant at a high power of the excitation beam, probably due to a thermal effect. The results of the present work provide important implication for the development of electrodes and electrolytes with proper solubility of BiF 3 and for better utilization of reactions by the two mechanisms in order to realize FSBs with high performance. ACKNOWLEDGMENT This work was supported by the New Energy and Industrial Technology Development Organization (NEDO) under contract from the Research & Development Initiative for Scientific Innovation of New Generation Batteries 2 (RISING2). REFERENCES [1] F. Gschwind, G. Rodriguez-Garcia, D. Sandbeck, A. Gross, M. Weil, M. Fichtner, N. Hörmann, J. Fluorine Chem. 182 , 76 (2016). [2] K.-I. Okazaki, Y. Uchimoto, T. Abe, Z. Ogumi, ACS Energy Lett. 2 , 1460 (2017). [3] V. K. Davis et al., Science 362 , 1144 (2018). [4] T. Yamanaka, K. Okazaki, T. Abe, K. Nishio, Z. Ogumi, ChemSusChem , 12 , 527 – 534 (2019). [5] T. Yamanaka, T. Abe, K. Nishio, Z. Ogumi, J. Electrochem. Soc., 166 , A635-A640 (2019). Figure 1. Results of Raman mapping of an area on a BiF 3 /gold sample in which both o-BiF 3 and c-BiF 3 particles were distributed. Only defluorination of o-BiF 3 occurred at 0.45 V. The arrows in (b) indicate that defluorination started at the contours of o-BiF 3 particles. Figure 1