- Email: [email protected]
Exotic solid state ion conductor from fluorinated titanium oxide and molten metallic lithium
Journal of Power Sources 400 (2018) 16–22
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Exotic solid state ion conductor from fluorinated titanium oxide and molten metallic lithium
T
Federico Bertasia, Gioele Pagota, Keti Vezzùa,∗∗, Enrico Negroa, Paul J. Siderisb, Steven G. Greenbaumc, Hiroyuki Ohnod, Bruno Scrosatie, Vito Di Notoa,∗ a
Section of Chemistry for Technology, Department of Industrial Engineering, University of Padova, Via Marzolo 9, 35131 Padova, Italy Chemistry Department, Queensborough Community College, CUNY, Bayside, NY, 11364, USA c Hunter College of the City University of New York, Physics Department, 695 Park Avenue, New York, NY, 10065, USA d Laboratory of Biotechnology and Life Science, Tokyo University of Agriculture and Technology, 2-24-16, Nagacho, Koganei, Tokyo, 184-8588, Japan e Istituto Italiano di Tecnologia, Genova, Italy b
H I GH L IG H T S
G R A P H I C A L A B S T R A C T
inorganic solid-state electrolyte for • An lithium batteries is prepared through an innovative one-step synthesis.
electrolyte shows a high con• The ductivity due to the hopping of Li ca-
•
tions at the interfaces between anatase -based NPs. Good prototype performance and full solid-state cyclic voltammetry are demonstrated.
A R T I C LE I N FO
A B S T R A C T
Keywords: Solid electrolyte Fluorinated titanium oxide Lithium battery Molten lithium Ion conduction
The potential for metallic lithium batteries that exhibit high specific capacities has stimulated a large interest within the energy research field. For safety reasons, the use of metallic lithium anodes requires electrochemically stable electrolytes. However, to date there has been limited success in this area. This work introduces a solid, lithium single-ion conductor thus providing new perspectives in the field of solid-state lithium-batteries. This new-concept material (LiFT), obtained by a direct reaction of nanometric fluorinated titanium oxide (FT) with molten metallic lithium, consists of nanoparticles (NPs) with anionic surface groups that are neutralized with lithium cations. The material displays fast lithium ion transport via an efficient migration mechanism occurring at the interfaces between different nanoparticles. The electrolyte comprises 1.34 mol kg−1 of Li and a conductivity of 2.8·10−4 S cm-1 at 25 °C is demonstrated. This level of performance, in conjunction with a native electrochemical stability towards lithium, extremely low cost starting materials (TiO2) and a facile one-pot synthesis, renders this electrolyte very attractive for applications in future full solid-state lithium batteries.
1. Introduction Lithium ion batteries, LIBs, are promising electrochemical devices
∗
that will likely play a crucial role in future energy technologies [1]. However, to assure their proper application in the emerging markets of road electrification and renewable energy storage, LIBs still require
Corresponding author. Corresponding author. E-mail addresses: [email protected] (K. Vezzù), [email protected] (V. Di Noto).
∗∗
https://doi.org/10.1016/j.jpowsour.2018.07.118 Received 26 January 2018; Received in revised form 13 July 2018; Accepted 29 July 2018 0378-7753/ © 2018 Published by Elsevier B.V.
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
2.3. Synthesis of the LIthiated fluorinated titanium oxide electrolyte (LiFT)
improvements in terms of safety, energy density and cost [2,3]. In this respect, the electrolyte, owing to its strong role on determining current density; time stability and safety; is arguably the most important component of the cell. Several classes of electrolytes including: polymer membranes [4]; ionic liquids [5] and inorganic solid-state materials [6] have been proposed in the last ten years, with the solid-state electrolytes, both polymeric and inorganic, exhibiting the best performance in terms of mechanical and chemical properties, especially in conjunction with metallic lithium or Li alloy-based anodes. New full solid-state electrolytes should present a: a) high ionic conductivity, which would enable high rate battery performance; and b) high chemical and electrochemical compatibility with the electrode materials. However, these two points are not achieved in state of the art solid-state electrolytes, which generally suffer of low conductivity at room temperature and poor interfacial properties with the electrodes. In addition, in order to be realistically applied to practical devices, they must also be of low cost and environmentally compatible. Several classes of solid-state electrolytes have been proposed in the last decades including: Lisicon, Nasicon, Perovskites, Antiperovskites, Garnets and glassy materials [7,8]. Lisicon -like materials typically show high conductivity values (up to 2.5 × 10−2 S cm−1 at 298 K) and suitable contact properties with the electrodes, nevertheless their stability against lithium anode is still an on-going issue [9,10]. Nasicon -based materials, which show conductivity values as high as 6 × 10−3 S cm−1 at 298 K [11] often present titanium reduction problems and generally show poor electrode wettability with high grain boundary resistance. Perovskites materials generally present high RT conductivity [12] (> 10−3 S cm−1 at RT) but poor electrochemical stability and high grain boundary resistance [13]. On the contrary, preliminary work on metastable antiperovskite -based materials, has shown promising features from both a conductivity and electrochemical stability window (ESW) point of view [14]. Finally garnet -base materials present both a high conductivity (up to 1 × 10−3 S cm−1 at 298 K) and good wettability and interfacial properties [15,16]. In this scenario, the herein presented solid-state lithium single-ion conducting material stands as a significant breakthrough in the field. The conductivity in this material arises from an effective lithium ion long range migration process occurring at the interfaces between functionalized nanoparticles. Due to the presence of a fluorinated shell on a TiO2 core, the proposed electrolyte (LiFT) does not undergo reduction upon contact with molten metallic Lithium during the synthesis process. This unusual behaviour makes this material extremely compatible with a metallic lithium anode of a “Lithium” battery.
LiFT electrolyte was prepared by suspending FT in a large excess of molten metallic lithium in an argon filled glovebox at 220 °C, with molten lithium acting as both a reagent and as a solvent. After cooling, the excess of metallic lithium was removed by reacting with ethanol thus obtaining lithium ethoxide, which is removed through several washing with EtOH. The resulting powder was then dried overnight under vacuum at 100 °C.
2.4. Materials characterization Elemental analyses were performed using a Spectro Arcos inductively-coupled plasma atomic emission spectroscopy (ICP-AES). The preparation of samples was conducted by reacting a known amount of material with hydrofluoric acid in a CEM MDS-2100 microwave digestion system. High-resolution transmission electron microscopy (HRTEM) measurements were performed using a 300 kV Jeol 3010 electron microscope equipped with a Gatan slowscan 794 CCD camera. A “GNR Analytical Instruments” eXplorer spectrometer, at a potential of 40 kV and a current of 30 mA applied to a Cu Kα1 source, was used to obtain the X-ray diffraction data. Measurements were carried out in a transmission mode between 5 and 60° every 0.1° sandwiching each sample between two Mylar windows. IR spectra were determined using a Nicolet Nexus spectrometer with a resolution of 2 cm−1 in the range 4000-50 cm−1. The MIR spectra were measured in the ATR mode while the FIR spectra measured in the transmission mode. For both cases, the samples were loaded inside an argon filled glove box. For the FIR measurements, two polyethylene windows were used and each spectrum was obtained after averaging 1000 scans. The ionic conductivity measurements were carried out by using a Novocontrol Alpha-A analyzer in the frequency range 10 mHz-10 MHz and in 25–155 °C temperature range, with a temperature control better than ± 0.2 °C, by exploiting a homemade heating-cooling system operating with liquid nitrogen. The sample, in the form of a pellet with known dimensions, was sandwiched between two circular platinum electrodes and placed into a hermetically sealed Teflon cell, assembled in an argon filled glove box and kept under argon for the duration of the test.
2.5. NMR measurements Both wideline 7Li and magic-angle spinning (MAS) 19F nuclear magnetic resonance (NMR) spectroscopy were used in this study in order to obtain additional information on the lithium, and the fluorine environments. Measurements were performed on a Varian-S Direct Drive spectrometer operating at 117.1 MHz and 283.5 MHz for 7Li and 19 F respectively. Powdered samples were packed into hermetically sealed 3.2 mm zirconia rotors inside an argon glovebox and 19F spectra were recorded at a spinning rate of 19–25 kHz. Free-induction decays (FIDs) were obtained using a phase cycled π/2 pulse – acquire – recycle delay sequence, and ECHOs were acquired using a typical phase cycled spin echo pulse sequence (π/2 pulse – τ – π pulse – τ – acquire). Spectra were gathered by Fourier transformation of the FIDs or trailing halfECHO signal. π/2 pulse widths of 4 and 3 μs were used for 19F and 7Li, respectively, with recycle delays of 2–10 s. Typically, several thousand transients were signal averaged before processing. Static (non-spinning) 7 Li spectra were obtained to assess the effect of temperature on Li+ motion through linewidth changes. The spectral frequency scales in the corresponding figures, as given in the normalized units of ppm, are relative to the 7Li and 19F chemical shifts of an aqueous solution of LiCl and CFCl3, respectively. A cross polarization (CP) experiment between 19 F and 7Li was performed to probe internuclear distances, with Hartmann-Hahn matching conditions optimized for solid LiF.
2. Experimental section 2.1. Materials NH4HF2 (99.999%), FeTiO3 (99.9%), metallic Lithium (99.9%) and Ethanol (> 99.8%) were Sigma-Aldrich products. Prior to use, ethanol was further dried by distillation over CaO. All handling, transfer and storage of reagents and products were carried out in Ar inert atmosphere dry boxes.
2.2. Synthesis of the fluorinated titanium oxide (FT) precursor FT was prepared by reacting FeTiO3 with an aqueous solution (30% wt) of NH4HF2 at 105 °C in a closed vessel for 1 h. The weight ratio between FeTiO3 and NH4HF2(aq) was equal to 1:1.75. The resulting mixture was then filtered and the pH raised to 9 by adding NH4OH(aq). The increase in pH led to the precipitation of a white powder that was first dried and then hydropyrolyzed at 450 °C in presence of steam (20 g min−1) for 2 h. The resulting powder was then dried overnight under vacuum at 100 °C. 17
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
electrochemical testing (see later). The key step during the LiFT synthesis is the formation of the fluorinated titanium oxide (FT) that is stable in molten lithium. This surface fluorination of the TiO2 NPs results in the formation of a passivating surface “shell” that prevents the molten lithium from coming into direct contact with the under laying TiO2 “core”. It is expected that the presence of fluoride on the surface of FT NPs raises the Fermi level of TiO2 thus protecting the NPs from potential reduction reactions.
2.6. Electrochemical measurements Li deposition/stripping, transference number, EIS and conductivity measurements are performed on LiFT self-standing pellets with an average thickness in the order of ≈150 μm. The lithium deposition/ stripping test was performed at 25 °C and at 0.10 and 0.05 mV s−1 scan rate using a VSP Bio-Logic 5-channel potentiostat-galvanostat. The full solid-state CV is performed using a three-electrode configuration, with a copper working electrode (nominal surface area of 6 mm2), while lithium served as both the counter and the reference electrode. A layered LiFT-copper pellet was coupled with a lithium metal electrode and heated to ensure an optimal contact. The cell was then housed in a Teflon holder and the measurements were collected three days after its preparation. Electrochemical Impedance Spectroscopy, EIS, was used to evaluate the lithium transference number in LiFT as well as its stability towards the metallic lithium electrode. The EIS measurements were carried out in the 1 Hz-10 MHz frequency and with an applied potential of 100 mV at 25 °C.
3.2. Structure and morphology The structures of the pristine FT and that of LiFT are probed by a series of complementary techniques, including X-ray diffraction (XRD), High-Resolution Transmission Electronic Microscopy (HR-TEM), Magic Angle Spinning (MAS) Nuclear Magnetic Resonance spectroscopy (NMR), Cross-Polarization MAS experiment (CPMAS) and variable temperature static 7Li NMR spectroscopy (see methods section). The diffraction profile of FT and LiFT (Fig. 2b), are both coincident with that of anatase titania [24], thus demonstrating that the synthetic method employed in this investigation does not result in any kind of intercalation compound. Fig. 2 also shows the HR-TEM images obtained for FT and LiFT. No significant changes to the morphology or size of the core NPs is observed following the reaction of FT with molten lithium. Indeed, both FT and LiFT consist of a distribution of NPs with average grain sizes less than 500 nm. In addition, the inter-planar distance, measured by HR-TEM, of 3.5 Å for FT and 2.4 Å for LiFT, is consistent with that of the (101) and (004) reflections of anatase titania [24]. The coincidence of the HR-TEM images and XRD patterns suggests that no lithium intercalation occurs in LiFT and that lithium ions are therefore located exclusively on the surface of the anionic NPs. XPS measurements on FT (not shown) have revealed: a) a F/Ti molar ratio equal to 0.07; and b) the presence of both bridging and terminal fluorine atoms. Infrared spectra of LiFT, FT and anatase titania (Fig. S1) are measured using Attenuated Total Reflectance methods (FT-IR ATR) owing to the higher degree of sensitivity of this technique to the surface structure of materials respect to a conventional transmission configuration. The reaction of ilmenite (FeTiO3) with ammonium hydrogen difluoride is thought to result in a fluorine functionalized anatase titania surface consisting of fluoride ions that are neutralized by ammonium cations in place of the hydroxide functionalities found on the surface of typical titania NPs [25]. Indeed, this is observed by comparing the FT-IR ATR spectra of FT with that of anatase titania. The broad band observed between 2500 and 3600 cm−1 and the sharper ones at 3666 and 1624 cm−1, attributed to the OH stretching and bending modes of anatase titania [26], are not present in the spectra of FT (Fig. S1, a), which rather displays peaks at 3261 and 1420 cm−1, attributed to the NH stretching and bending modes, respectively. Following the lithiation process, no IR active bands are observed above 2500 cm−1, and there is a significant reduction in the intensity of the NH bending mode observed at 1420 cm−1. This suggests that most of the ammonium groups of FT have been replaced by lithium ions during the formation of LiFT. A strong, broad Ti-O stretching band for anatase titania is detected just below 900 cm−1. For FT, this band is of diminished intensity and is present alongside another strong band occurring at 943 cm−1, which is attributed to the Ti-F stretching mode of a TiOF3like species [27]. It should be noted that this second band exhibits a substantial intensity only in FT, implying that fluorination promotes the formation of a “shell” of concatenated TiF2 moieties. Upon subsequent lithiation, surface group of Ti-F-Li+ are formed. Bands associated with Ti-O bonds are also observed in the FIR region for all three compounds (Fig. S1, b). Two strong ones are revealed at 453 and 318 cm−1, which we assume to be associated with the E1u and A2u phonon modes of anatase titania [28]. In the spectrum of LiFT there are three additional bands that are not observed in those of FT nor of TiO2. Broad bands at 388 and 357 cm−1, as well as another one occurring at 296 cm−1, are associated with ν(LiF) Ag and B1u modes of lithium ions coordinated by
2.7. Battery prototypes The batteries were prepared using lithium metal as anodes and lithium cobalt oxide, LiCoO2 (LCO) (89.1%wt LiCoO2, 8.9%wt graphite SK6, 2.0%wt PVDF), as cathode, separated by the LiFT electrolyte. In the case of prototype testing, one drop (15 μL) of lithium-salt-free ethylene carbonate-dimethyl carbonate (EC/DMC) solution was added between the electrolyte and the cathode in order to improve the charge injection and the surface properties of the cathodic side. A careful optimization of the cathode composition (active material/conductive additive/ binder ratio) and morphology (multilayered structure) are the targets of ongoing studies and will allow to avoid the use of any trace of solvents. 3. Results and discussion 3.1. Synthesis The electrolyte reported in this work is based on cheap fluorinated anatase TiO2 nanoparticle, directly surface functionalized with Li cations through a novel one-step reaction with molten metallic lithium. The electrolyte is hereafter referred by the acronym LiFT (LIthiated Fluorinated Titanium oxide). Several materials based on TiO2 and TiOF2, obtained predominantly by hydrothermal methods, have been proposed as electrodes for lithium batteries. In these materials it has been demonstrated that: a) Li cation intercalation occurs within the TiO2 structure [17-22]; or b) conversion reaction takes place with formation of LiF [23]. In the case of the proposed electrolyte (LiFT), XRD measurements show no evidence of lithium ion intercalation nor conversion behaviour (see Paragraph 3.2). On the contrary, Li cations, whose presence is measured by ICP-AES are present only on the surface of TiO2 -based nanoparticles, which behave like macroanions. The first step of the LiFT synthesis involves the preparation of FT (Fluorinated Titanium oxide) by reacting FeTiO3 with NH4HF2 (Fig. 1, Step I). A suspension of FT nanopowder in a large excess of molten Li is then kept at 220 °C for 2 h under constant stirring (Fig. 1, step II). During the course of the reaction it is observed that the FT changes in colour from light yellow to blue as it becomes LiFT. ICP-AES and elemental analysis results indicate that the obtained LiFT electrolyte has the formula Li0.128 [(NH4)0.007TiO1.98F0.07]. It should be noted that the reaction between molten lithium and non-fluorinated TiO2, as well as with other oxides (including SiO2 and Fe2O3), does not produce a lithiated material. This is owing to the associated explosive oxides reduction reaction. LiFT, on the contrary, is stable when in direct contact with molten lithium at 220 °C, this clearly demonstrating that LiFT is fully compatible with metallic lithium anodes, as indeed verified by 18
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
Fig. 1. Graphical illustration of the two step synthesis of LiFT. Step I: Synthesis of FT precursor; Step II: lithiation of FT to obtain the LiFT electrolyte.
Fig. 2. Structural investigations. (a) HR-TEM images and corresponding interplanar distances. (b) XRD patters for the LiFT and FT materials. (c) 19F MAS (25 kHz spin rate) NMR spectra of FT and LiFT (top and middle); 7Li – 19F CPMAS (500 μs contact time) spectrum of LiFT (bottom). (d) Static wide-line 7Li NMR spectra of LiFT at various temperatures.
fluoride functionalities present in the “shell” of LiFT [29,30]. Fig. 2c (top) displays the 19F MAS-NMR spectra for FT and LiFT at 25 °C. There are common features present in both spectra, including a strong resonance at −18 ppm, attributed to the bridging fluorine atoms. This signal is accompanied by significant spinning sidebands. A smaller signal at −87 ppm is assigned to terminal fluoride, while another (minor) component at +35 ppm remains unassigned at this stage. The sharp resonance centred at about −120 ppm is assigned to PTFE,
presents as background from the probe and rotor assembly. The significant difference between the two spectra is the occurrence of a strong resonance at −201 ppm only in the spectrum for LiFT. This likely results from the lithiation of the terminal fluorine atoms. In order to confirm this hypothesis, a 7Li – 19F cross polarization measurement was carried out; only those 19F nuclei that at are in close proximity to the 7Li ones are observed (within 1–2 Å) (see Fig. 2c, bottom spectrum). Indeed, the strong resonance that occurs at −201 ppm in the 19F MAS 19
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
between the lithium metal electrodes and the LiFT electrolytes. By fitting the Nyquist plot, Re(Z) vs. frequency, Im(Z) vs. frequency and tanδ profiles with the equation associated with the related equivalent circuit [31] (see Fig. S2 in the Supplementary Information), it is possible to demonstrate that in the Li|LiFT|Li cell, both the charge transfer (Rct) and the bulk electrolyte (Rb) resistances decrease with time, this in turn signifying that the overall electrochemical kinetics, as well as the solid electrolyte conductivity, improve over time, see Fig. 3b. A possible explanation can be ascribed to the conversion from bridging to terminal fluoride ions, on the surface of LiFT particles. Following this, the enhanced concentration of coordination sites for Li migration leads to an increased number of percolation pathways, thus increasing the electrolyte conductivity. Fig. 3c compares the conductivity (σ) Arrhenius plots of FT and LiFT. One striking difference is that the ionic conductivity of the latter exceeds that of the former by several orders of magnitude, reaching 2.8·10−4 S cm-1 at 25 °C with an activation energy of 50.4 ± 1.5 kJ mol−1. Conductivity values of this order of magnitude approach those of polymer electrolytes commonly used in lithium-ion batteries [32]. This unique behaviour is also confirmed by analyzing the real component of the conductivity (σ′) measured over a frequency range from 10 mHz to 10 MHz and at variable temperature from 25 °C to 155 °C with 10 °C intervals, see inset of Fig. 3c. Although it is reasonable to assume that in both FT and LiFT the long range charge migration events take place owing to hopping processes between neighbouring anionic sites present on the surface of NPs, the energy barrier for this process in FT is double than that of LiFT, suggesting a higher density of mobile charge species and a lower average charge migration distance between neighbouring anionic sites. Therefore, in accordance with the solid state NMR results, we may conclude that the long-range charge migration events in LiFT largely take place owing to Li+ cation exchange phenomena occurring between neighbouring terminal fluoride anions present in the “shell” of LiFT NPs. The measured unitary lithium transference number [33] (tLi+), see Fig. S3 in the Supplementary Information, demonstrates that LiFT is a single-ion conductor and confirms that the mobility of the lithium-neutralizing fluoride ions
NMR spectrum is the only one observed under CP conditions with the short contact time used. The observed chemical shift and CP parameters are similar to those for LiF, although the presence of crystalline LiF is not observed by XRD. Taking all together, the occurrence of this strong lithium fluorine interaction, in conjunction with ICP-AES, XRD and IR data, supports the assumption that: a) lithium is present only on the surface of the NP, acting as to neutralize the terminal fluoride ions; b) the bulk of material consists of anatase; and c) no fluorine is present in bulk LiFT material. Fig. 2d shows the static wide-line 7Li NMR spectra recorded as a function of temperature. Changes in the lineshapes with increasing temperature are attributed to motional narrowing, which is partially masked by structural heterogeneity. The primary line broadening contribution is the nuclear quadrupole interaction, with additional influence from the 7Li - 19F dipole-dipole interactions with the terminal fluorides discussed above. Unfortunately, high-resolution MAS measurements (not shown) shed no additional light on the structural heterogeneity because both the quadrupole and dipolar broadening mechanisms are effectively averaged out by MAS, and the chemical shift range for Li in diamagnetic compounds is small. After cooling the sample from elevated temperature (the spectra were recorded with increasing temperature), there is a significant change in the spectral lineshape, this being attributed to an increased structural heterogeneity. A possible mechanism for this behaviour is the thermal conversion of bridging F− to terminal F− and a resulting rearrangement of Li+ sites on the nanoparticle surface. 3.3. Electrochemical characterization Fig. 3a shows that the lithium deposition-stripping process for LiFT takes place at low overvoltages with a peak current higher than 200 mA cm−2. In addition we can then affirm that LiFT is highly compatible with the lithium metal anode, as also confirmed by the impedance spectroscopy response of a symmetric Li|LiFT|Li cell, see inset of Fig. 3b. No diffusive spike is observed at low frequency, suggesting that there is no formation of blocking layers at the interfaces
Fig. 3. Electrochemical characterizations of LiFT. (a): Cyclic voltammetry of the lithium plating-stripping process on a copper working electrode (at 25 °C). (b): Impedance measurements Nyquist plots for Li|LiFT|Li cell, fitted with an equivalent circuit based on two elements in series (at 25 °C). Each element consists of a resistance in parallel with a constant phase capacity (See also Fig. S2). (c) Arrhenius conductivity plots of LiFT and FT. The inset shows the real component of the conductivity vs. frequency and temperature. The fitting of conductivity vs. T−1 data provides a correlation coefficient of 0.997 and 0.992 for LiFT and FT, respectively.
20
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
conducting material, suitable for the development of next generation, all solid-state lithium batteries. Acknowledgements V.D.N., F.B., G.P., K.V. and E.N. thank the strategic project “MAESTRA” and the project SID 2016 of the University of Padua for funding. The NMR Program at Hunter College is supported by a single investigator grant from the Basic Energy Sciences Division of the U.S. Department of Energy (No. DE-SC0005029). V.D.N. thanks the University Carlos III of Madrid, Spain, for the “Càtedras de Excelencia UC3MSantander” (Chair of Excellence UC3M-Santander). The authors thank Dr Stefano Zeggio and Dr Fabio Bassetto from Breton S.p.a. for their contribution in the development of the pristine FT material and for the fruitful discussions. Appendix A. Supplementary data Fig. 4. Voltage profiles of the charge-discharge initial cycling of a Li|LiFT|LiCoO2 cell at room temperature and at various rates and in the 3.50–4.25 V range.
Supplementary data related to this article can be found at https:// doi.org/10.1016/j.jpowsour.2018.07.118. References
is indeed negligible.
[1] M. Armand, J.-M. Tarascon, Building better batteries, Nature 451 (2008) 652–657. [2] E. Quartarone, P. Mustarelli, Electrolytes for solid-state lithium rechargeable batteries: recent advances and perspectives, Chem. Soc. Rev. 40 (2011) 2525–2540. [3] B. Scrosati, J. Hassoun, Y.-K. Sun, Lithium-ion batteries. A look into the future, Energy Environ. Sci. 4 (2011) 3287–3295. [4] V. Di Noto, S. Lavina, G.A. Giffin, E. Negro, B. Scrosati, Polymer electrolytes: present, past and future, Electrochim. Acta 57 (2011) 4–13. [5] M. Armand, F. Endres, D.R. MacFarlane, H. Ohno, B. Scrosati, Ionic-liquid materials for the electrochemical challenges of the future, Nat. Mater. 8 (2009) 621–629. [6] P. Knauth, Inorganic solid Li ion conductors: an overview, Solid State Ionics 180 (2009) 911–916. [7] B. Zhang, R. Tan, L. Yang, J. Zheng, K. Zhang, S. Mo, et al., Mechanisms and properties of ion-transport in inorganic solid electrolytes, Energy Storage Mater 10 (2018) 139–159. [8] M.H. Braga, N.S. Grundish, A.J. Murchison, J.B. Goodenough, Alternative strategy for a safe rechargeable battery, Energy Environ. Sci. 10 (2017) 331–336. [9] Y. Kato, S. Hori, T. Saito, K. Suzuki, M. Hirayama, A. Mitsui, et al., High-power allsolid-state batteries using sulfide superionic conductors, Nat. Energy 1 (2016) 16030. [10] N. Kamaya, K. Homma, Y. Yamakawa, M. Hirayama, R. Kanno, M. Yonemura, et al., A lithium superionic conductor, Nat. Mater. 10 (2011) 682–686. [11] M. Pérez-Estébanez, J. Isasi-Marín, D.M. Többens, A. Rivera-Calzada, C. León, A systematic study of Nasicon-type Li1+xMxTi2−x(PO4)3 (M: Cr, Al, Fe) by neutron diffraction and impedance spectroscopy, Solid State Ionics 266 (2014) 1–8. [12] Y. Inaguma, M. Nakashima, A rechargeable lithium-air battery using a lithium ionconducting lanthanum lithium titanate ceramics as an electrolyte separator, J. Power Sources 228 (2013) 250–255. [13] C. Hua, X. Fang, Z. Wang, L. Chen, Lithium storage in perovskite lithium lanthanum titanate, Electrochem. Commun. 32 (2013) 5–8. [14] Y. Li, W. Zhou, S. Xin, S. Li, J. Zhu, L. Xujie, et al., Fluorine-doped antiperovskite electrolyte for all-solid-state lithium-ion batteries, Angew. Chem. Int. Ed. 55 (2016) 9965–9968. [15] H.Y.P. Hong, Crystal structure and ionic conductivity of Li14Zn(GeO4)4 and other new Li+ superionic conductors, Mater. Res. Bull. 13 (1978) 117–124. [16] B. Liu, Y. Gong, K. Fu, X. Han, Y. Yao, G. Pastel, et al., Garnet solid electrolyte protected Li-Metal batteries, ACS Appl. Mater. Interfaces 9 (2017) 18809–18815. [17] F. Bonino, L. Busani, M. Lazzari, M. Manstretta, B. Rivolta, B. Scrosati, Anatase as a cathode material in lithium—organic electrolyte rechargeable batteries, J. Power Sources 6 (1981) 261–270. [18] G. Sudant, E. Baudrin, D. Larcher, J.-M. Tarascon, Electrochemical lithium reactivity with nanotextured anatase-type TiO2, J. Mater. Chem. 15 (2005) 1263–1269. [19] G. Armstrong, A.R. Armstrong, P.G. Bruce, P. Reale, B. Scrosati, TiO2(B) nanowires as an improved anode material for lithium-ion batteries containing LiFePO4 or LiNi0.5Mn1.5O4 cathodes and a polymer electrolyte, Adv. Mater. 18 (2006) 2597–2600. [20] M.V. Reddy, S. Madhavi, G.V. Subba Rao, B.V.R. Chowdari, Metal oxyfluorides TiOF2 and NbO2F as anodes for Li-ion batteries, J. Power Sources 162 (2006) 1312–1321. [21] Y.-G. Guo, Y.-S. Hu, W. Sigle, J. Maier, Superior electrode performance of nanostructured mesoporous TiO2 (anatase) through efficient hierarchical mixed conducting networks, Adv. Mater. 19 (2007) 2087–2091. [22] L. Chen, L. Shen, P. Nie, X. Zhang, H. Li, Facile hydrothermal synthesis of single crystalline TiOF2 nanocubes and their phase transitions to TiO2 hollow nanocages as anode materials for lithium-ion battery, Electrochim. Acta 62 (2012) 408–415. [23] S.-T. Myung, M. Kikuchi, C.S. Yoon, H. Yashiro, Y.-K. Sun, A new synthetic method
3.4. Prototype testing After elucidating the favourable electrochemical properties of LiFT, we proceeded with the evaluation of its relevance as an advanced medium for the development of advanced batteries. Fig. 4 shows the voltage profiles of the initial 40 cycles of the Li|LiFT|LiCoO2 battery run at 25 °C and at various rates within the 3.50–4.25 V range. In this battery one drop (15 μL) of a salt-free EC/ DCM solution are added on the cathodic side prior to couple with the Li/LiFT anodic pellet. Notably, the battery may operate at room temperature delivering a capacity of the order of 135 mAh·g−1 (LiCoO2), namely a value that approaches that expected for the lithium cobalt oxide cathode-based cells. We believe that this behaviour provides further convincing evidence of the high conductivity of our LiFT electrolyte. Some minor decay in capacity is observed at increasing rates; however the initial value at C/20 is fully recovered at the end of cycling, this in turn clearly showing that the battery is not affected by irreversible polarizations. In addition, the capacity delivered in charge is almost totally recovered in discharge with no evidence of irregularities in the charge profiles, offering good confirmation of the high compatibility of our electrolyte with the lithium metal electrode, another characteristic rarely found in previous literature work. Also remarkable is the fact that the battery operates around 3.8 V, thus giving a theoretical energy density of the order of 570 W h kg−1. 4. Conclusions Although still preliminary, the results above discussed clearly demonstrate the significant advantages of LiFT in terms of ionic conductivity, compatibility with the lithium metal anode, and cost, over other existing/competing ion conducting materials. The synthesis upon reaction with molten metallic lithium renders LiFT intrinsically inert upon further contact with Lithium (e.g. Li metal anode). In addition, a unique conductivity mechanism trough Li cations exchange at the interfaces between different LiFT nanoparticles is demonstrated. This peculiarity makes LiFT significantly different from conventional stateof-the-art solid-state electrolytes, in which the conductivity typically takes place at both the grain boundary and trough the bulk of the material. These unusual and rare electrochemical properties, combined with the cheap synthesis, the high stability with the lithium metal anode, the long life time and, especially, the high room temperature conductivity, render LiFT a very promising new solid-state lithium21
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
[24]
[25]
[26] [27] [28]
Phys. D Appl. Phys. 38 (2005) 1415–1420. [29] R.L. Redington, Infrared vibrational spectra of matrix-isolated cyclic Li2F2, Li3F3, and Li4F4 isotopomers, J. Chem. Phys. 102 (1995) 7325–7331. [30] D. Oschetzki, G. Rauhut, Pushing the limits in accurate vibrational structure calculations: anharmonic frequencies of lithium fluoride clusters (LiF)n, n = 2-10, Phys. Chem. Chem. Phys. 16 (2014) 16426–16435. [31] E. Barsoukov, J.R. Macdonald, Impedance Spectroscopy, John Wiley & Sons, Inc, Hoboken, NJ, USA, 2005. [32] B. Scrosati, K.M. Abraham, W. Van Schalkwijk, J. Hassoun, Lithium Batteries: Advanced Technologies and Applications, John Wiley & Sons, Inc, Hoboken, NJ, USA, 2013. [33] J. Evans, C.A. Vincent, P.G. Bruce, Electrochemical measurement of transference numbers in polymer electrolytes, Polymer 28 (1987) 2324–2328.
of titanium oxyfluoride and its application as an anode material for rechargeable lithium batteries, J. Power Sources 288 (2015) 376–383. K. Burdett, Structural-electronic relationships in inorganic solids: powder neutron diffraction studies of the rutile and anatase polymorphs of titanium dioxide at 15 and 295 K, J. Am. Chem. Soc. (1987) 3639–3646. N.M. Laptash, I. Maslennikova, L. Kurilenko, N. Mishchenko, Ilmenite fluorination by ammonium hydrodifluoride: a new ammonium oxofluorotitanate, russ, J. Inorg. Chem. 46 (2001) 28–34. K. Morishige, S. Ogawara, S. Sasaki, Hydrated surfaces of particulate titanium dioxide prepared by pyrolysis, J. Phys. Chem. 20 (1985) 4404–4408. N.M. Laptash, I.G. Maslennikova, T.A. Kaidalova, Ammonium oxofuorotitanates, J. Fluorine Chem. 99 (1999) 133–137. M. Grujić-Brojčin, M.J. Šćepanović, Z.D. Dohčević-Mitrovi, I. Hinić, B. Matović, G. Stanišić, et al., Infrared study of laser synthesized anatase TiO2 nanopowders, J.
22
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Exotic solid state ion conductor from fluorinated titanium oxide and molten metallic lithium
T
Federico Bertasia, Gioele Pagota, Keti Vezzùa,∗∗, Enrico Negroa, Paul J. Siderisb, Steven G. Greenbaumc, Hiroyuki Ohnod, Bruno Scrosatie, Vito Di Notoa,∗ a
Section of Chemistry for Technology, Department of Industrial Engineering, University of Padova, Via Marzolo 9, 35131 Padova, Italy Chemistry Department, Queensborough Community College, CUNY, Bayside, NY, 11364, USA c Hunter College of the City University of New York, Physics Department, 695 Park Avenue, New York, NY, 10065, USA d Laboratory of Biotechnology and Life Science, Tokyo University of Agriculture and Technology, 2-24-16, Nagacho, Koganei, Tokyo, 184-8588, Japan e Istituto Italiano di Tecnologia, Genova, Italy b
H I GH L IG H T S
G R A P H I C A L A B S T R A C T
inorganic solid-state electrolyte for • An lithium batteries is prepared through an innovative one-step synthesis.
electrolyte shows a high con• The ductivity due to the hopping of Li ca-
•
tions at the interfaces between anatase -based NPs. Good prototype performance and full solid-state cyclic voltammetry are demonstrated.
A R T I C LE I N FO
A B S T R A C T
Keywords: Solid electrolyte Fluorinated titanium oxide Lithium battery Molten lithium Ion conduction
The potential for metallic lithium batteries that exhibit high specific capacities has stimulated a large interest within the energy research field. For safety reasons, the use of metallic lithium anodes requires electrochemically stable electrolytes. However, to date there has been limited success in this area. This work introduces a solid, lithium single-ion conductor thus providing new perspectives in the field of solid-state lithium-batteries. This new-concept material (LiFT), obtained by a direct reaction of nanometric fluorinated titanium oxide (FT) with molten metallic lithium, consists of nanoparticles (NPs) with anionic surface groups that are neutralized with lithium cations. The material displays fast lithium ion transport via an efficient migration mechanism occurring at the interfaces between different nanoparticles. The electrolyte comprises 1.34 mol kg−1 of Li and a conductivity of 2.8·10−4 S cm-1 at 25 °C is demonstrated. This level of performance, in conjunction with a native electrochemical stability towards lithium, extremely low cost starting materials (TiO2) and a facile one-pot synthesis, renders this electrolyte very attractive for applications in future full solid-state lithium batteries.
1. Introduction Lithium ion batteries, LIBs, are promising electrochemical devices
∗
that will likely play a crucial role in future energy technologies [1]. However, to assure their proper application in the emerging markets of road electrification and renewable energy storage, LIBs still require
Corresponding author. Corresponding author. E-mail addresses: [email protected] (K. Vezzù), [email protected] (V. Di Noto).
∗∗
https://doi.org/10.1016/j.jpowsour.2018.07.118 Received 26 January 2018; Received in revised form 13 July 2018; Accepted 29 July 2018 0378-7753/ © 2018 Published by Elsevier B.V.
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
2.3. Synthesis of the LIthiated fluorinated titanium oxide electrolyte (LiFT)
improvements in terms of safety, energy density and cost [2,3]. In this respect, the electrolyte, owing to its strong role on determining current density; time stability and safety; is arguably the most important component of the cell. Several classes of electrolytes including: polymer membranes [4]; ionic liquids [5] and inorganic solid-state materials [6] have been proposed in the last ten years, with the solid-state electrolytes, both polymeric and inorganic, exhibiting the best performance in terms of mechanical and chemical properties, especially in conjunction with metallic lithium or Li alloy-based anodes. New full solid-state electrolytes should present a: a) high ionic conductivity, which would enable high rate battery performance; and b) high chemical and electrochemical compatibility with the electrode materials. However, these two points are not achieved in state of the art solid-state electrolytes, which generally suffer of low conductivity at room temperature and poor interfacial properties with the electrodes. In addition, in order to be realistically applied to practical devices, they must also be of low cost and environmentally compatible. Several classes of solid-state electrolytes have been proposed in the last decades including: Lisicon, Nasicon, Perovskites, Antiperovskites, Garnets and glassy materials [7,8]. Lisicon -like materials typically show high conductivity values (up to 2.5 × 10−2 S cm−1 at 298 K) and suitable contact properties with the electrodes, nevertheless their stability against lithium anode is still an on-going issue [9,10]. Nasicon -based materials, which show conductivity values as high as 6 × 10−3 S cm−1 at 298 K [11] often present titanium reduction problems and generally show poor electrode wettability with high grain boundary resistance. Perovskites materials generally present high RT conductivity [12] (> 10−3 S cm−1 at RT) but poor electrochemical stability and high grain boundary resistance [13]. On the contrary, preliminary work on metastable antiperovskite -based materials, has shown promising features from both a conductivity and electrochemical stability window (ESW) point of view [14]. Finally garnet -base materials present both a high conductivity (up to 1 × 10−3 S cm−1 at 298 K) and good wettability and interfacial properties [15,16]. In this scenario, the herein presented solid-state lithium single-ion conducting material stands as a significant breakthrough in the field. The conductivity in this material arises from an effective lithium ion long range migration process occurring at the interfaces between functionalized nanoparticles. Due to the presence of a fluorinated shell on a TiO2 core, the proposed electrolyte (LiFT) does not undergo reduction upon contact with molten metallic Lithium during the synthesis process. This unusual behaviour makes this material extremely compatible with a metallic lithium anode of a “Lithium” battery.
LiFT electrolyte was prepared by suspending FT in a large excess of molten metallic lithium in an argon filled glovebox at 220 °C, with molten lithium acting as both a reagent and as a solvent. After cooling, the excess of metallic lithium was removed by reacting with ethanol thus obtaining lithium ethoxide, which is removed through several washing with EtOH. The resulting powder was then dried overnight under vacuum at 100 °C.
2.4. Materials characterization Elemental analyses were performed using a Spectro Arcos inductively-coupled plasma atomic emission spectroscopy (ICP-AES). The preparation of samples was conducted by reacting a known amount of material with hydrofluoric acid in a CEM MDS-2100 microwave digestion system. High-resolution transmission electron microscopy (HRTEM) measurements were performed using a 300 kV Jeol 3010 electron microscope equipped with a Gatan slowscan 794 CCD camera. A “GNR Analytical Instruments” eXplorer spectrometer, at a potential of 40 kV and a current of 30 mA applied to a Cu Kα1 source, was used to obtain the X-ray diffraction data. Measurements were carried out in a transmission mode between 5 and 60° every 0.1° sandwiching each sample between two Mylar windows. IR spectra were determined using a Nicolet Nexus spectrometer with a resolution of 2 cm−1 in the range 4000-50 cm−1. The MIR spectra were measured in the ATR mode while the FIR spectra measured in the transmission mode. For both cases, the samples were loaded inside an argon filled glove box. For the FIR measurements, two polyethylene windows were used and each spectrum was obtained after averaging 1000 scans. The ionic conductivity measurements were carried out by using a Novocontrol Alpha-A analyzer in the frequency range 10 mHz-10 MHz and in 25–155 °C temperature range, with a temperature control better than ± 0.2 °C, by exploiting a homemade heating-cooling system operating with liquid nitrogen. The sample, in the form of a pellet with known dimensions, was sandwiched between two circular platinum electrodes and placed into a hermetically sealed Teflon cell, assembled in an argon filled glove box and kept under argon for the duration of the test.
2.5. NMR measurements Both wideline 7Li and magic-angle spinning (MAS) 19F nuclear magnetic resonance (NMR) spectroscopy were used in this study in order to obtain additional information on the lithium, and the fluorine environments. Measurements were performed on a Varian-S Direct Drive spectrometer operating at 117.1 MHz and 283.5 MHz for 7Li and 19 F respectively. Powdered samples were packed into hermetically sealed 3.2 mm zirconia rotors inside an argon glovebox and 19F spectra were recorded at a spinning rate of 19–25 kHz. Free-induction decays (FIDs) were obtained using a phase cycled π/2 pulse – acquire – recycle delay sequence, and ECHOs were acquired using a typical phase cycled spin echo pulse sequence (π/2 pulse – τ – π pulse – τ – acquire). Spectra were gathered by Fourier transformation of the FIDs or trailing halfECHO signal. π/2 pulse widths of 4 and 3 μs were used for 19F and 7Li, respectively, with recycle delays of 2–10 s. Typically, several thousand transients were signal averaged before processing. Static (non-spinning) 7 Li spectra were obtained to assess the effect of temperature on Li+ motion through linewidth changes. The spectral frequency scales in the corresponding figures, as given in the normalized units of ppm, are relative to the 7Li and 19F chemical shifts of an aqueous solution of LiCl and CFCl3, respectively. A cross polarization (CP) experiment between 19 F and 7Li was performed to probe internuclear distances, with Hartmann-Hahn matching conditions optimized for solid LiF.
2. Experimental section 2.1. Materials NH4HF2 (99.999%), FeTiO3 (99.9%), metallic Lithium (99.9%) and Ethanol (> 99.8%) were Sigma-Aldrich products. Prior to use, ethanol was further dried by distillation over CaO. All handling, transfer and storage of reagents and products were carried out in Ar inert atmosphere dry boxes.
2.2. Synthesis of the fluorinated titanium oxide (FT) precursor FT was prepared by reacting FeTiO3 with an aqueous solution (30% wt) of NH4HF2 at 105 °C in a closed vessel for 1 h. The weight ratio between FeTiO3 and NH4HF2(aq) was equal to 1:1.75. The resulting mixture was then filtered and the pH raised to 9 by adding NH4OH(aq). The increase in pH led to the precipitation of a white powder that was first dried and then hydropyrolyzed at 450 °C in presence of steam (20 g min−1) for 2 h. The resulting powder was then dried overnight under vacuum at 100 °C. 17
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
electrochemical testing (see later). The key step during the LiFT synthesis is the formation of the fluorinated titanium oxide (FT) that is stable in molten lithium. This surface fluorination of the TiO2 NPs results in the formation of a passivating surface “shell” that prevents the molten lithium from coming into direct contact with the under laying TiO2 “core”. It is expected that the presence of fluoride on the surface of FT NPs raises the Fermi level of TiO2 thus protecting the NPs from potential reduction reactions.
2.6. Electrochemical measurements Li deposition/stripping, transference number, EIS and conductivity measurements are performed on LiFT self-standing pellets with an average thickness in the order of ≈150 μm. The lithium deposition/ stripping test was performed at 25 °C and at 0.10 and 0.05 mV s−1 scan rate using a VSP Bio-Logic 5-channel potentiostat-galvanostat. The full solid-state CV is performed using a three-electrode configuration, with a copper working electrode (nominal surface area of 6 mm2), while lithium served as both the counter and the reference electrode. A layered LiFT-copper pellet was coupled with a lithium metal electrode and heated to ensure an optimal contact. The cell was then housed in a Teflon holder and the measurements were collected three days after its preparation. Electrochemical Impedance Spectroscopy, EIS, was used to evaluate the lithium transference number in LiFT as well as its stability towards the metallic lithium electrode. The EIS measurements were carried out in the 1 Hz-10 MHz frequency and with an applied potential of 100 mV at 25 °C.
3.2. Structure and morphology The structures of the pristine FT and that of LiFT are probed by a series of complementary techniques, including X-ray diffraction (XRD), High-Resolution Transmission Electronic Microscopy (HR-TEM), Magic Angle Spinning (MAS) Nuclear Magnetic Resonance spectroscopy (NMR), Cross-Polarization MAS experiment (CPMAS) and variable temperature static 7Li NMR spectroscopy (see methods section). The diffraction profile of FT and LiFT (Fig. 2b), are both coincident with that of anatase titania [24], thus demonstrating that the synthetic method employed in this investigation does not result in any kind of intercalation compound. Fig. 2 also shows the HR-TEM images obtained for FT and LiFT. No significant changes to the morphology or size of the core NPs is observed following the reaction of FT with molten lithium. Indeed, both FT and LiFT consist of a distribution of NPs with average grain sizes less than 500 nm. In addition, the inter-planar distance, measured by HR-TEM, of 3.5 Å for FT and 2.4 Å for LiFT, is consistent with that of the (101) and (004) reflections of anatase titania [24]. The coincidence of the HR-TEM images and XRD patterns suggests that no lithium intercalation occurs in LiFT and that lithium ions are therefore located exclusively on the surface of the anionic NPs. XPS measurements on FT (not shown) have revealed: a) a F/Ti molar ratio equal to 0.07; and b) the presence of both bridging and terminal fluorine atoms. Infrared spectra of LiFT, FT and anatase titania (Fig. S1) are measured using Attenuated Total Reflectance methods (FT-IR ATR) owing to the higher degree of sensitivity of this technique to the surface structure of materials respect to a conventional transmission configuration. The reaction of ilmenite (FeTiO3) with ammonium hydrogen difluoride is thought to result in a fluorine functionalized anatase titania surface consisting of fluoride ions that are neutralized by ammonium cations in place of the hydroxide functionalities found on the surface of typical titania NPs [25]. Indeed, this is observed by comparing the FT-IR ATR spectra of FT with that of anatase titania. The broad band observed between 2500 and 3600 cm−1 and the sharper ones at 3666 and 1624 cm−1, attributed to the OH stretching and bending modes of anatase titania [26], are not present in the spectra of FT (Fig. S1, a), which rather displays peaks at 3261 and 1420 cm−1, attributed to the NH stretching and bending modes, respectively. Following the lithiation process, no IR active bands are observed above 2500 cm−1, and there is a significant reduction in the intensity of the NH bending mode observed at 1420 cm−1. This suggests that most of the ammonium groups of FT have been replaced by lithium ions during the formation of LiFT. A strong, broad Ti-O stretching band for anatase titania is detected just below 900 cm−1. For FT, this band is of diminished intensity and is present alongside another strong band occurring at 943 cm−1, which is attributed to the Ti-F stretching mode of a TiOF3like species [27]. It should be noted that this second band exhibits a substantial intensity only in FT, implying that fluorination promotes the formation of a “shell” of concatenated TiF2 moieties. Upon subsequent lithiation, surface group of Ti-F-Li+ are formed. Bands associated with Ti-O bonds are also observed in the FIR region for all three compounds (Fig. S1, b). Two strong ones are revealed at 453 and 318 cm−1, which we assume to be associated with the E1u and A2u phonon modes of anatase titania [28]. In the spectrum of LiFT there are three additional bands that are not observed in those of FT nor of TiO2. Broad bands at 388 and 357 cm−1, as well as another one occurring at 296 cm−1, are associated with ν(LiF) Ag and B1u modes of lithium ions coordinated by
2.7. Battery prototypes The batteries were prepared using lithium metal as anodes and lithium cobalt oxide, LiCoO2 (LCO) (89.1%wt LiCoO2, 8.9%wt graphite SK6, 2.0%wt PVDF), as cathode, separated by the LiFT electrolyte. In the case of prototype testing, one drop (15 μL) of lithium-salt-free ethylene carbonate-dimethyl carbonate (EC/DMC) solution was added between the electrolyte and the cathode in order to improve the charge injection and the surface properties of the cathodic side. A careful optimization of the cathode composition (active material/conductive additive/ binder ratio) and morphology (multilayered structure) are the targets of ongoing studies and will allow to avoid the use of any trace of solvents. 3. Results and discussion 3.1. Synthesis The electrolyte reported in this work is based on cheap fluorinated anatase TiO2 nanoparticle, directly surface functionalized with Li cations through a novel one-step reaction with molten metallic lithium. The electrolyte is hereafter referred by the acronym LiFT (LIthiated Fluorinated Titanium oxide). Several materials based on TiO2 and TiOF2, obtained predominantly by hydrothermal methods, have been proposed as electrodes for lithium batteries. In these materials it has been demonstrated that: a) Li cation intercalation occurs within the TiO2 structure [17-22]; or b) conversion reaction takes place with formation of LiF [23]. In the case of the proposed electrolyte (LiFT), XRD measurements show no evidence of lithium ion intercalation nor conversion behaviour (see Paragraph 3.2). On the contrary, Li cations, whose presence is measured by ICP-AES are present only on the surface of TiO2 -based nanoparticles, which behave like macroanions. The first step of the LiFT synthesis involves the preparation of FT (Fluorinated Titanium oxide) by reacting FeTiO3 with NH4HF2 (Fig. 1, Step I). A suspension of FT nanopowder in a large excess of molten Li is then kept at 220 °C for 2 h under constant stirring (Fig. 1, step II). During the course of the reaction it is observed that the FT changes in colour from light yellow to blue as it becomes LiFT. ICP-AES and elemental analysis results indicate that the obtained LiFT electrolyte has the formula Li0.128 [(NH4)0.007TiO1.98F0.07]. It should be noted that the reaction between molten lithium and non-fluorinated TiO2, as well as with other oxides (including SiO2 and Fe2O3), does not produce a lithiated material. This is owing to the associated explosive oxides reduction reaction. LiFT, on the contrary, is stable when in direct contact with molten lithium at 220 °C, this clearly demonstrating that LiFT is fully compatible with metallic lithium anodes, as indeed verified by 18
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
Fig. 1. Graphical illustration of the two step synthesis of LiFT. Step I: Synthesis of FT precursor; Step II: lithiation of FT to obtain the LiFT electrolyte.
Fig. 2. Structural investigations. (a) HR-TEM images and corresponding interplanar distances. (b) XRD patters for the LiFT and FT materials. (c) 19F MAS (25 kHz spin rate) NMR spectra of FT and LiFT (top and middle); 7Li – 19F CPMAS (500 μs contact time) spectrum of LiFT (bottom). (d) Static wide-line 7Li NMR spectra of LiFT at various temperatures.
fluoride functionalities present in the “shell” of LiFT [29,30]. Fig. 2c (top) displays the 19F MAS-NMR spectra for FT and LiFT at 25 °C. There are common features present in both spectra, including a strong resonance at −18 ppm, attributed to the bridging fluorine atoms. This signal is accompanied by significant spinning sidebands. A smaller signal at −87 ppm is assigned to terminal fluoride, while another (minor) component at +35 ppm remains unassigned at this stage. The sharp resonance centred at about −120 ppm is assigned to PTFE,
presents as background from the probe and rotor assembly. The significant difference between the two spectra is the occurrence of a strong resonance at −201 ppm only in the spectrum for LiFT. This likely results from the lithiation of the terminal fluorine atoms. In order to confirm this hypothesis, a 7Li – 19F cross polarization measurement was carried out; only those 19F nuclei that at are in close proximity to the 7Li ones are observed (within 1–2 Å) (see Fig. 2c, bottom spectrum). Indeed, the strong resonance that occurs at −201 ppm in the 19F MAS 19
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
between the lithium metal electrodes and the LiFT electrolytes. By fitting the Nyquist plot, Re(Z) vs. frequency, Im(Z) vs. frequency and tanδ profiles with the equation associated with the related equivalent circuit [31] (see Fig. S2 in the Supplementary Information), it is possible to demonstrate that in the Li|LiFT|Li cell, both the charge transfer (Rct) and the bulk electrolyte (Rb) resistances decrease with time, this in turn signifying that the overall electrochemical kinetics, as well as the solid electrolyte conductivity, improve over time, see Fig. 3b. A possible explanation can be ascribed to the conversion from bridging to terminal fluoride ions, on the surface of LiFT particles. Following this, the enhanced concentration of coordination sites for Li migration leads to an increased number of percolation pathways, thus increasing the electrolyte conductivity. Fig. 3c compares the conductivity (σ) Arrhenius plots of FT and LiFT. One striking difference is that the ionic conductivity of the latter exceeds that of the former by several orders of magnitude, reaching 2.8·10−4 S cm-1 at 25 °C with an activation energy of 50.4 ± 1.5 kJ mol−1. Conductivity values of this order of magnitude approach those of polymer electrolytes commonly used in lithium-ion batteries [32]. This unique behaviour is also confirmed by analyzing the real component of the conductivity (σ′) measured over a frequency range from 10 mHz to 10 MHz and at variable temperature from 25 °C to 155 °C with 10 °C intervals, see inset of Fig. 3c. Although it is reasonable to assume that in both FT and LiFT the long range charge migration events take place owing to hopping processes between neighbouring anionic sites present on the surface of NPs, the energy barrier for this process in FT is double than that of LiFT, suggesting a higher density of mobile charge species and a lower average charge migration distance between neighbouring anionic sites. Therefore, in accordance with the solid state NMR results, we may conclude that the long-range charge migration events in LiFT largely take place owing to Li+ cation exchange phenomena occurring between neighbouring terminal fluoride anions present in the “shell” of LiFT NPs. The measured unitary lithium transference number [33] (tLi+), see Fig. S3 in the Supplementary Information, demonstrates that LiFT is a single-ion conductor and confirms that the mobility of the lithium-neutralizing fluoride ions
NMR spectrum is the only one observed under CP conditions with the short contact time used. The observed chemical shift and CP parameters are similar to those for LiF, although the presence of crystalline LiF is not observed by XRD. Taking all together, the occurrence of this strong lithium fluorine interaction, in conjunction with ICP-AES, XRD and IR data, supports the assumption that: a) lithium is present only on the surface of the NP, acting as to neutralize the terminal fluoride ions; b) the bulk of material consists of anatase; and c) no fluorine is present in bulk LiFT material. Fig. 2d shows the static wide-line 7Li NMR spectra recorded as a function of temperature. Changes in the lineshapes with increasing temperature are attributed to motional narrowing, which is partially masked by structural heterogeneity. The primary line broadening contribution is the nuclear quadrupole interaction, with additional influence from the 7Li - 19F dipole-dipole interactions with the terminal fluorides discussed above. Unfortunately, high-resolution MAS measurements (not shown) shed no additional light on the structural heterogeneity because both the quadrupole and dipolar broadening mechanisms are effectively averaged out by MAS, and the chemical shift range for Li in diamagnetic compounds is small. After cooling the sample from elevated temperature (the spectra were recorded with increasing temperature), there is a significant change in the spectral lineshape, this being attributed to an increased structural heterogeneity. A possible mechanism for this behaviour is the thermal conversion of bridging F− to terminal F− and a resulting rearrangement of Li+ sites on the nanoparticle surface. 3.3. Electrochemical characterization Fig. 3a shows that the lithium deposition-stripping process for LiFT takes place at low overvoltages with a peak current higher than 200 mA cm−2. In addition we can then affirm that LiFT is highly compatible with the lithium metal anode, as also confirmed by the impedance spectroscopy response of a symmetric Li|LiFT|Li cell, see inset of Fig. 3b. No diffusive spike is observed at low frequency, suggesting that there is no formation of blocking layers at the interfaces
Fig. 3. Electrochemical characterizations of LiFT. (a): Cyclic voltammetry of the lithium plating-stripping process on a copper working electrode (at 25 °C). (b): Impedance measurements Nyquist plots for Li|LiFT|Li cell, fitted with an equivalent circuit based on two elements in series (at 25 °C). Each element consists of a resistance in parallel with a constant phase capacity (See also Fig. S2). (c) Arrhenius conductivity plots of LiFT and FT. The inset shows the real component of the conductivity vs. frequency and temperature. The fitting of conductivity vs. T−1 data provides a correlation coefficient of 0.997 and 0.992 for LiFT and FT, respectively.
20
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
conducting material, suitable for the development of next generation, all solid-state lithium batteries. Acknowledgements V.D.N., F.B., G.P., K.V. and E.N. thank the strategic project “MAESTRA” and the project SID 2016 of the University of Padua for funding. The NMR Program at Hunter College is supported by a single investigator grant from the Basic Energy Sciences Division of the U.S. Department of Energy (No. DE-SC0005029). V.D.N. thanks the University Carlos III of Madrid, Spain, for the “Càtedras de Excelencia UC3MSantander” (Chair of Excellence UC3M-Santander). The authors thank Dr Stefano Zeggio and Dr Fabio Bassetto from Breton S.p.a. for their contribution in the development of the pristine FT material and for the fruitful discussions. Appendix A. Supplementary data Fig. 4. Voltage profiles of the charge-discharge initial cycling of a Li|LiFT|LiCoO2 cell at room temperature and at various rates and in the 3.50–4.25 V range.
Supplementary data related to this article can be found at https:// doi.org/10.1016/j.jpowsour.2018.07.118. References
is indeed negligible.
[1] M. Armand, J.-M. Tarascon, Building better batteries, Nature 451 (2008) 652–657. [2] E. Quartarone, P. Mustarelli, Electrolytes for solid-state lithium rechargeable batteries: recent advances and perspectives, Chem. Soc. Rev. 40 (2011) 2525–2540. [3] B. Scrosati, J. Hassoun, Y.-K. Sun, Lithium-ion batteries. A look into the future, Energy Environ. Sci. 4 (2011) 3287–3295. [4] V. Di Noto, S. Lavina, G.A. Giffin, E. Negro, B. Scrosati, Polymer electrolytes: present, past and future, Electrochim. Acta 57 (2011) 4–13. [5] M. Armand, F. Endres, D.R. MacFarlane, H. Ohno, B. Scrosati, Ionic-liquid materials for the electrochemical challenges of the future, Nat. Mater. 8 (2009) 621–629. [6] P. Knauth, Inorganic solid Li ion conductors: an overview, Solid State Ionics 180 (2009) 911–916. [7] B. Zhang, R. Tan, L. Yang, J. Zheng, K. Zhang, S. Mo, et al., Mechanisms and properties of ion-transport in inorganic solid electrolytes, Energy Storage Mater 10 (2018) 139–159. [8] M.H. Braga, N.S. Grundish, A.J. Murchison, J.B. Goodenough, Alternative strategy for a safe rechargeable battery, Energy Environ. Sci. 10 (2017) 331–336. [9] Y. Kato, S. Hori, T. Saito, K. Suzuki, M. Hirayama, A. Mitsui, et al., High-power allsolid-state batteries using sulfide superionic conductors, Nat. Energy 1 (2016) 16030. [10] N. Kamaya, K. Homma, Y. Yamakawa, M. Hirayama, R. Kanno, M. Yonemura, et al., A lithium superionic conductor, Nat. Mater. 10 (2011) 682–686. [11] M. Pérez-Estébanez, J. Isasi-Marín, D.M. Többens, A. Rivera-Calzada, C. León, A systematic study of Nasicon-type Li1+xMxTi2−x(PO4)3 (M: Cr, Al, Fe) by neutron diffraction and impedance spectroscopy, Solid State Ionics 266 (2014) 1–8. [12] Y. Inaguma, M. Nakashima, A rechargeable lithium-air battery using a lithium ionconducting lanthanum lithium titanate ceramics as an electrolyte separator, J. Power Sources 228 (2013) 250–255. [13] C. Hua, X. Fang, Z. Wang, L. Chen, Lithium storage in perovskite lithium lanthanum titanate, Electrochem. Commun. 32 (2013) 5–8. [14] Y. Li, W. Zhou, S. Xin, S. Li, J. Zhu, L. Xujie, et al., Fluorine-doped antiperovskite electrolyte for all-solid-state lithium-ion batteries, Angew. Chem. Int. Ed. 55 (2016) 9965–9968. [15] H.Y.P. Hong, Crystal structure and ionic conductivity of Li14Zn(GeO4)4 and other new Li+ superionic conductors, Mater. Res. Bull. 13 (1978) 117–124. [16] B. Liu, Y. Gong, K. Fu, X. Han, Y. Yao, G. Pastel, et al., Garnet solid electrolyte protected Li-Metal batteries, ACS Appl. Mater. Interfaces 9 (2017) 18809–18815. [17] F. Bonino, L. Busani, M. Lazzari, M. Manstretta, B. Rivolta, B. Scrosati, Anatase as a cathode material in lithium—organic electrolyte rechargeable batteries, J. Power Sources 6 (1981) 261–270. [18] G. Sudant, E. Baudrin, D. Larcher, J.-M. Tarascon, Electrochemical lithium reactivity with nanotextured anatase-type TiO2, J. Mater. Chem. 15 (2005) 1263–1269. [19] G. Armstrong, A.R. Armstrong, P.G. Bruce, P. Reale, B. Scrosati, TiO2(B) nanowires as an improved anode material for lithium-ion batteries containing LiFePO4 or LiNi0.5Mn1.5O4 cathodes and a polymer electrolyte, Adv. Mater. 18 (2006) 2597–2600. [20] M.V. Reddy, S. Madhavi, G.V. Subba Rao, B.V.R. Chowdari, Metal oxyfluorides TiOF2 and NbO2F as anodes for Li-ion batteries, J. Power Sources 162 (2006) 1312–1321. [21] Y.-G. Guo, Y.-S. Hu, W. Sigle, J. Maier, Superior electrode performance of nanostructured mesoporous TiO2 (anatase) through efficient hierarchical mixed conducting networks, Adv. Mater. 19 (2007) 2087–2091. [22] L. Chen, L. Shen, P. Nie, X. Zhang, H. Li, Facile hydrothermal synthesis of single crystalline TiOF2 nanocubes and their phase transitions to TiO2 hollow nanocages as anode materials for lithium-ion battery, Electrochim. Acta 62 (2012) 408–415. [23] S.-T. Myung, M. Kikuchi, C.S. Yoon, H. Yashiro, Y.-K. Sun, A new synthetic method
3.4. Prototype testing After elucidating the favourable electrochemical properties of LiFT, we proceeded with the evaluation of its relevance as an advanced medium for the development of advanced batteries. Fig. 4 shows the voltage profiles of the initial 40 cycles of the Li|LiFT|LiCoO2 battery run at 25 °C and at various rates within the 3.50–4.25 V range. In this battery one drop (15 μL) of a salt-free EC/ DCM solution are added on the cathodic side prior to couple with the Li/LiFT anodic pellet. Notably, the battery may operate at room temperature delivering a capacity of the order of 135 mAh·g−1 (LiCoO2), namely a value that approaches that expected for the lithium cobalt oxide cathode-based cells. We believe that this behaviour provides further convincing evidence of the high conductivity of our LiFT electrolyte. Some minor decay in capacity is observed at increasing rates; however the initial value at C/20 is fully recovered at the end of cycling, this in turn clearly showing that the battery is not affected by irreversible polarizations. In addition, the capacity delivered in charge is almost totally recovered in discharge with no evidence of irregularities in the charge profiles, offering good confirmation of the high compatibility of our electrolyte with the lithium metal electrode, another characteristic rarely found in previous literature work. Also remarkable is the fact that the battery operates around 3.8 V, thus giving a theoretical energy density of the order of 570 W h kg−1. 4. Conclusions Although still preliminary, the results above discussed clearly demonstrate the significant advantages of LiFT in terms of ionic conductivity, compatibility with the lithium metal anode, and cost, over other existing/competing ion conducting materials. The synthesis upon reaction with molten metallic lithium renders LiFT intrinsically inert upon further contact with Lithium (e.g. Li metal anode). In addition, a unique conductivity mechanism trough Li cations exchange at the interfaces between different LiFT nanoparticles is demonstrated. This peculiarity makes LiFT significantly different from conventional stateof-the-art solid-state electrolytes, in which the conductivity typically takes place at both the grain boundary and trough the bulk of the material. These unusual and rare electrochemical properties, combined with the cheap synthesis, the high stability with the lithium metal anode, the long life time and, especially, the high room temperature conductivity, render LiFT a very promising new solid-state lithium21
Journal of Power Sources 400 (2018) 16–22
F. Bertasi et al.
[24]
[25]
[26] [27] [28]
Phys. D Appl. Phys. 38 (2005) 1415–1420. [29] R.L. Redington, Infrared vibrational spectra of matrix-isolated cyclic Li2F2, Li3F3, and Li4F4 isotopomers, J. Chem. Phys. 102 (1995) 7325–7331. [30] D. Oschetzki, G. Rauhut, Pushing the limits in accurate vibrational structure calculations: anharmonic frequencies of lithium fluoride clusters (LiF)n, n = 2-10, Phys. Chem. Chem. Phys. 16 (2014) 16426–16435. [31] E. Barsoukov, J.R. Macdonald, Impedance Spectroscopy, John Wiley & Sons, Inc, Hoboken, NJ, USA, 2005. [32] B. Scrosati, K.M. Abraham, W. Van Schalkwijk, J. Hassoun, Lithium Batteries: Advanced Technologies and Applications, John Wiley & Sons, Inc, Hoboken, NJ, USA, 2013. [33] J. Evans, C.A. Vincent, P.G. Bruce, Electrochemical measurement of transference numbers in polymer electrolytes, Polymer 28 (1987) 2324–2328.
of titanium oxyfluoride and its application as an anode material for rechargeable lithium batteries, J. Power Sources 288 (2015) 376–383. K. Burdett, Structural-electronic relationships in inorganic solids: powder neutron diffraction studies of the rutile and anatase polymorphs of titanium dioxide at 15 and 295 K, J. Am. Chem. Soc. (1987) 3639–3646. N.M. Laptash, I. Maslennikova, L. Kurilenko, N. Mishchenko, Ilmenite fluorination by ammonium hydrodifluoride: a new ammonium oxofluorotitanate, russ, J. Inorg. Chem. 46 (2001) 28–34. K. Morishige, S. Ogawara, S. Sasaki, Hydrated surfaces of particulate titanium dioxide prepared by pyrolysis, J. Phys. Chem. 20 (1985) 4404–4408. N.M. Laptash, I.G. Maslennikova, T.A. Kaidalova, Ammonium oxofuorotitanates, J. Fluorine Chem. 99 (1999) 133–137. M. Grujić-Brojčin, M.J. Šćepanović, Z.D. Dohčević-Mitrovi, I. Hinić, B. Matović, G. Stanišić, et al., Infrared study of laser synthesized anatase TiO2 nanopowders, J.
22