Ferrum fluorides as nanostructured conversion cathodes: in situ XAFS and XRD study using synchrotron radiation.
Transitional metals fluorides have been known from 1960s in the field of rechargeable batteries. They belongs to the conversion materials, and by reaction with lithium can be reduced to the mixture of zero valent metal and lithium fluoride, resulting the so-called more-than-one electron reaction. However, possibly due to the lack of nanostructuration of commercial compounds, fluorides gave the way for more immediately interesting and promising materials.
Since 1960s various technologies for producing of nanocompounds have been developed, so fluorides became one of the leading research directions for Li-ion industry. Most efforts are being put into iron trifluoride for several reasons: it’s more environmentally friendly than others, cheaper, requires reasonable synthesis conditions, has relatively low polarization, etc.
In our study we pursue the goal to determine structural changes which take place inside the full cell and to look at those processes in situ during cycle. Experiment was carried on the B station of BM01 (SNBL) beamline at ESRF, Grenoble, as a mixed XAFS/XRD experiment. Samples were prepared with synthesis of iron fluoride nanoparticles inside reduced graphene oxide sheets which increases conductivity. The material offers a stable discharge energy of 600-700 Wh/kg over 100 cycles, which is higher than the widely applied cathode materials (300-500 Wh/kg). Each sample we cycled with 20 mAh/g in the 1,2-4,2 V range, while measuring Fe K-edge XAFS spectra in transmission mode and XRD patterns with 15 minutes interval. For measurements we used self-made test cells with glassy carbon windows connected to Gamry potentiostats responsible for cycling and data acquisition.
The Li intercalation in the first discharge is different from subsequent cycles. Up to 1.8 V, a maximum of 0.66 Li will be inserted into the channels of the framework structure of initial FeF3 0.33H2O (from synthesis conditions). Then peaks in the XRD patterns disappear and nanocrystalline LiF and Fe phases will form below 1.8 V. On charge, the ReO3-type FeF3 phase with higher density will form instead of the open-framework structure. On second and subsequent discharge reactions, we will form LiFe2F6 instead, which then also converts to LiF/Fe. This is of theoretical nature, because of the nanocrystalline structure of the involved phases, which is too small to detect in XRD, and also similar Fe-F6 environments, which make the FeF3 phases very hard to distinguish in XAS.
Results of the x-ray studies were associated with cycling data to obtain structure-charge state dependency. HTB structure of the as-prepared material has open intercalation channels as a result, full electrochemical reaction can be separated into initial intercalation of one Li- anion per formula unit and following conversion reaction involving two more Li-, which gives us 3LiF/Fe mixture and a complete three electron transition. To prove this we performed principal component analysis (PCA) on the series of XAFS experimental spectra. We have used FitIt software to mathematically decompose the series of the Fe K-edge spectra at different voltages into independent sub-spectra. It was found that all spectra for discharge process can be reproduced as a combination of three components. First component corresponds to HTB structure, second to the intercalated structure with Fe2+ charge state and the third one corresponds to metallic Fe. We have observed that pure Fe forms after HTB conversion to intercalated phase. We also performed a set of ab initio calculations and DFT modeling for different concentrations of Li in cathode material. Theoretical simulations for the Fe K-edge XANES are in progress now to figure out if we can distinguish intercalated HTB structure from the LiFe2F6 phase in the XANES data.