Facile synthesis of C–FeF₂ nanocomposites from CFx: influence of carbon precursor on reversible lithium storage

Abstract

Transition metal fluorides are an important class of cathode materials for lithium batteries owing to their high specific energy and safety. However, metal fluorides are electrical insulators, exhibiting slow reaction kinetics with Li. Consequently, metal fluorides can show poor electrochemical performance. Instead, carbon–metal fluoride nanocomposites (CMNFCs) were suggested to enhance electrochemical activity. Chemical synthesis of CMNFCs poses particular challenges due to the poor chemical stability of metal fluorides. Recently, we reported a facile one-step method to synthesize carbon–FeF₂ nanocomposites by reacting fluorinated carbon (CFx) with iron pentacarbonyl (Fe(CO)₅) at 250 °C. The method resulted in C–FeF₂ nanocomposites with improved electrochemical properties. Here, we have synthesized four different C–FeF₂ nanocomposites by reacting four different CFx precursors made of petro-coke, carbon black, graphite, and carbon-fibers with Fe(CO)₅. Electrochemical performance of all four C–FeF₂ nanocomposites was evaluated at 25 °C and 40 °C. It is shown that the nature of CFx has a critical impact on the electrochemical performance of the corresponding C–FeF₂ nanocomposites. The C–FeF₂ nanocomposites were characterized by using various experimental techniques such as X-ray diffraction, scanning electron microscopy, transmission electron microscopy, resistivity measurement, and ⁵⁷Fe Mössbauer spectroscopy to shed light on the differences in electrochemical behaviour of different C–FeF₂ nanocomposites.

---

Introduction

Conversion electrode materials based on metal fluorides are interesting as cathode materials for lithium batteries due to their high energy density compared to the insertion based electrode materials.^1–6 Metal fluorides react with lithium at a relatively high voltage, and more than one electron per metal can be transferred, which results in high energy density. For example, FeF₂ can react with 2.0 Li at a potential of 2.66 V, with a specific capacity of 571 mA h g⁻¹, which leads to a theoretical specific energy of 1518 W h kg⁻¹. On the other hand, metal fluorides pose certain challenges as electrode materials. Metal fluorides are electrical insulators, show slow reaction kinetics with lithium, and consequently, they exhibit large voltage hysteresis between discharge and charge processes. In addition, high volume changes associated with metal fluorides pose further challenges.⁴ Therefore, stable anchoring of metal fluoride nanocrystallites in a conductive carbon matrix is necessary to provide the electronic path, to improve the reaction kinetics and to buffer the volume changes. To overcome these issues, carbon metal fluoride nanocomposites (CMFNCs) were suggested in earlier work.^7,8

A general approach to synthesize CMFNCs is mechanical milling of a conductive carbon with a desired metal fluoride.^7,8 While mechanical milling effectively reduces the particle size of metal fluorides, the high energy applied in mechanical milling process leads to the destruction of the original carbon structure and produces disordered carbons with less conductive interfaces. Consequently, CMNFCs obtained by mechanical milling show limited cycling stability in lithium half cells.^7,8 Alternatively, chemical methods have been reported to synthesize CMNFCs, which showed much higher cycling stability.^9–24 Recently, we have reported a facile method for the synthesis of carbon–metal fluoride nanocomposites by reacting graphite fluoride (CFx) with iron pentacarbonyl (Fe(CO)₅) at 250 °C.^25–27 The C–FeF₂ nanocomposites obtained by this method delivered high reversible capacity in lithium half cells.^25,27 Further, pretreatment of the CFx precursor by mechanical milling has a significant impact on the reversible capacity, which we attributed to the reduced particle size of CFx.²⁵ We also showed that optimum fluorine to carbon ratio is necessary to achieve high reversibility.²⁶

CFx has a great advantage as a precursor material for the synthesis of CMFNCs, as it is a source of both conductive carbon and fluoride ions. Defluorination of CFx restores the carbon to its original state and the desired carbon structure can be predesigned. More importantly, the reaction occurs in one step and is quick (could be finished in 1 h).²⁷ Further, addition of conductive carbon is not required to prepare the electrodes. In this study, we used various fluorinated carbon compounds namely, petro-coke, carbon black, graphite, and carbon-fibers to synthesize C–FeF₂ nanocomposites. The unique feature of such nanocomposites is that, while the average size of the FeF₂ crystallites is almost the same, the carbon matrix in which FeF₂ nanocrystallites are embedded is different. This provides a unique opportunity to study and better understand the requirements for the design of carbon-nanocomposites for high reversible lithium storage.

Experimental section

Synthesis

Different CFx precursors were kindly provided by Advanced Research Chemicals (ARC). Fe(CO)₅ was purchased from Aldrich. C–FeF₂ nanocomposites were synthesized in sealed Swagelok® type stainless steel (SS) reactors. In a typical synthesis, required amount of Fe(CO)₅ was added to 0.25 g of CFx powder in the SS reactor and closed with VCR fittings inside an argon-filled glove box. The SS reactor was placed inside a tube furnace, and the temperature was raised from room temperature to 250 °C with a heating rate of 5 °C min⁻¹. The reaction was carried out at 250 °C for 24 hours; then the reactor was allowed to cool down naturally. Pressure developed due to the formation of gaseous side product was released carefully, and the reactor was opened in the Ar-filled glove box. The resulting black powder was collected carefully.

Characterization

Powder X-ray diffraction (PXRD) patterns were recorded in Bragg–Brentano geometry in the 2θ range 10–70° using a Philips X'pert diffractometer equipped with Mo Kα radiation. In the Debye–Scherrer mode, patterns were collected using a STOE Stadi P diffractometer equipped with a Dectris Mythen 1K linear silicon strip detector and Ge(111) double crystal monochromator (Mo Kα1 radiation, λ = 0.7093 Å). The samples were loaded into 0.7 mm glass capillaries (Hilgenberg borosilicate glass no 50) in an argon-filled glove box. For the refinement of the XRD pattern, we used MAUD software.²⁸ Scanning electron microscopy (SEM) was performed with a LEO 1530 at 15 kV using carbon tape as substrate. Transmission electron microscopy was carried out on an aberration (image) corrected Titan 80-300 (FEI Company) operated at 80 kV equipped with a Gatan imaging filter Tridiem 863. The material for TEM studies consisted of powder sample free from solvents. Since the samples were sensitive to the electron beam at 300 kV, resulting in the amorphization of the graphitic carbon around the FeF₂ nanoparticles, the TEM studies were carried out at 80 kV. The Mössbauer spectra were collected using a standard transmission Mössbauer setup with a ⁵⁷Co in Rh-Matrix source operated in constant acceleration mode. All Isomer shifts (IS) are given with respect to bcc-Fe at room temperature. The spectra were fitted using the WinNormos software by R. A. Brand. For electrical resistivity measurements, the powders were pressed into a 13 mm diameter pellets with a pressure of 10 tons per m². The resistivity of the nanocomposites was measured by the Van der Pauw method.

Electrochemical studies

Electrochemical measurements were performed in Swagelok® type cells. The electrode fabrication and electrochemical cells were assembled in an argon-filled glove box. Electrodes were fabricated by mixing the as-synthesized material and polyvinylidene fluoride (PVDF) in the weight ratios of 90 : 10. A slurry containing the above mixture was prepared by using N-methyl-2-pyrrolidinone and was spread on stainless steel (SS) foil (area: 1.13 cm²) and dried on the hot plate at 160 °C for 12 h. Typically, each electrode contained 4–6 mg of the C–FeF₂ nanocomposite. Lithium foil (Goodfellow) was used as the negative electrode, and a borosilicate glass fiber sheet (separator) saturated with 1 M LiPF₆ in 1 : 1 ethylene carbonate (EC)/dimethyl carbonate (DMC) (LP30, Merck) was used as the electrolyte. The cells were placed in an incubator (Binder) to maintain a constant temperature of 25 °C or 40 °C. Electrochemical studies were carried out using an Arbin battery cycling unit. For electrochemical impedance measurements (EIS), cells were assembled and equilibrated at open circuit voltage (OCV) for 24 h. Impedance measurements were made using Zahner IM6 electrochemical workstation.

Results and discussion

We used four different CFx precursors synthesized by fluorinating petro-coke (FPC), carbon-black (FCB), graphite (FG), and carbon-fibers (FCF) (Advanced Research Chemicals). These CFx precursors largely differ in terms of morphology, particle size and surface area (Source: Advanced Research Chemicals). Selected physical properties of these materials are given in Table S1 (see ESI†). In the case of FPC, FCB and FCF fluorine to carbon ratio is between 1.05–1.12, whereas in the case of FG it is 0.95. Fig. S1† shows the XRD patterns of the CFx precursors (see ESI†). All samples show two broad peaks: one at 5° and another at 18° which are typical for highly fluorinated carbon materials. Fig. S2† shows SEM images of the CFx precursors (see ESI†). In the case of FPC, the particle size is in the range of 1–30 μm, with an average particle size of 8.0 μm. In the case of FCB big agglomerates can be seen. However, these agglomerates consist of smaller particles which are less than 200 nm in diameter. In the case of FG, the particle size is in the range of 1–10 μm with an average particle size of 2.0 μm. In the case of FCF, the diameter of the fibers is in the range of 10–30 μm while the length of the fibers is up to 100 μm. Despite the bigger particle size of the FCF, the surface area of 344 m² g⁻¹ suggests that the material is porous. The high surface area of the highly fluorinated carbons FCB (1.12) and FCF (1.1) suggests that during high-temperature fluorination some carbon was converted into a CF₄ gas which leads to pores.

Four types of C–FeF₂ nanocomposites were synthesized by reacting different CFx samples with Fe(CO)₅. PC–FeF₂ was synthesized by reacting fluorinated petro coke, FPC with Fe(CO)₅ for 24 h at 250 °C. Similarly, CB–FeF₂, G–FeF₂, and CF–FeF₂ were synthesized by reacting Fe(CO)₅ with FCB, FG, and FCF respectively at 250 °C for 24 h. Fig. 1 shows the XRD patterns of the as-synthesized carbon–FeF₂ nanocomposites recorded in capillary mode. In the case of PC–FeF₂, CB–FeF₂ and G–FeF₂ all the peaks could be indexed to rutile-type FeF₂. However, few extra peaks were observed in the case of CF–FeF₂ (indicated with an asterisk * in Fig. 1). We also noticed that the relative intensity of (210) plane is growing from CB–FeF₂ < PC–FeF₂ < G–FeF₂ < CF–FeF₂. The origin of the additional peaks could be due to the formation of iron carbide (which was confirmed by ⁵⁷Fe Mössbauer spectroscopy). Rietveld refinement was performed including iron carbide as a secondary phase. The carbon was not taken into account. The refinement parameters are given in Table S2 (see ESI†). The hump around 10° is due to the capillary and was fitted in the background using a Gaussian peak in addition to a polynomial function for the baseline. Further, refinement results showed that in the case of CB–FeF₂ there was no iron carbide whereas, in the case of PC–FeF₂, G–FeF₂ and CF–FeF₂ iron carbide was present as a secondary phase with a maximum of 22.5 wt% in CF–FeF₂. The average crystallite size was growing from PC–FeF₂ (11 nm) < G–FeF₂ (12 nm) < CF–FeF₂ (16 nm) < CB–FeF₂ (17 nm).

Fig. 1 XRD patterns of carbon–FeF₂ nanocomposites synthesized from various CFx precursors (recorded in capillary mode).

In order to further elucidate the origin of the additional reflections found in the XRD patterns, we recorded ⁵⁷Fe Mössbauer spectra of all the nanocomposites as Mössbauer spectroscopy is a sensitive local probe for the ⁵⁷Fe environments. Fig. 2a shows the measured Mössbauer spectra of the four C–FeF₂ nanocomposites together with their respective fit model. The obtained Mössbauer hyperfine parameters are summarized in Table S3 (see ESI†). All four samples show two quadrupole doublets as major components. The first one, and most important one is characteristic for FeF₂ with an IS of 1.33 mm s⁻¹, and a quadrupole splitting (QS) of about 2.77 mm s⁻¹. These values are in agreement with data in the literature.²⁹ The second doublet has an IS of about 0.47 mm s⁻¹, which is characteristic for the presence of Fe³⁺ component. However, no evidence was found for the existence of any crystalline phases of FeF₃ from the XRD measurements. This signal could be possibly due to aerial oxidation of the sample during the sample transfer, although care has been taken to avoid oxidation. This hypothesis has been checked by investigating a purposely air exposed sample (CB–FeF₂) (Fig. 2b). After exposure to air, the sample consists of pure Fe³⁺ components, which is very similar to the Fe³⁺ components found in the fresh samples. This finding makes the oxidation as an explanation for the occurrence of the Fe³⁺ phase very likely. The Mössbauer spectrum of the CB–FeF₂ sample was fully reproduced by these two doublets. All other samples exhibit three additional sub-spectra, which are magnetically split with a magnetic hyperfine field BHF of about 22, ∼19 and ∼11 T respectively. They can be attributed to the three Fe-sites present in Fe_1−xC_x alloys (x ∼ 0.2–0.3) with triangular, prismatic structure.³⁰ The relative ratios given in Table S3† describe the spectral area fraction of the sub-spectra and may be used to compare the four samples among each other. However, it is not possible to give numbers for the phase fractions of the components as the Debye–Waller factors for these different phases may differ, which in turn leads to a different resonant area of the respective Fe environment.

Fig. 2 Mössbauer spectra of C–FeF₂ nanocomposites measured at room temperature.

From the XRD and Mössbauer studies, the presence of iron carbide is evident. Formation of iron carbide indirectly suggests the intermediate formation of iron nanoparticles by the decomposition of Fe(CO)₅; these iron nanoparticles further react with CFx and forms C–FeF₂. The absence of iron carbide in CB–FeF₂ could be explained by its smaller particle size. Due to the smaller particle size, the reaction between iron nanoparticles and FCB is fast, and the formation of iron carbide is mitigated. The smaller the particle size of the fluorinated carbon, the faster is the reaction with iron nanoparticles and hence no formation of iron carbide is observed in the case of FCB. However, the iron carbide content is large in the case of G–FeF₂ compared to PC–FeF₂ (the opposite is expected because of the small particle size of FG compared to FPC); this could be due to the availability of free carbon in the precursor (F/C ratio 0.95) which could readily react with iron nanoparticles and form iron carbide.

To investigate the microstructure of the nanocomposites, scanning electron microscopy (SEM) and transmission electron microscopy (TEM) analysis was performed on these nanocomposites. Fig. 3 shows the SEM images of the nanocomposites. Additional low magnification SEM images were shown in Fig. S3 (see ESI†). The bulk morphologies were similar to unreacted precursors. In the case of PC–FeF₂ (Fig. 3a), FeF₂ crystallites are embedded in carbon layers. No FeF₂ particles are seen on the surface of the sample. However, in the case of CB–FeF₂ (Fig. 3b), some of the FeF₂ crystallites are protruding on the surface, and few crystallites (up to 50 nm) are also seen on the surface (shown with arrows). This could be due to the high surface to volume ratio of the sample. The G–FeF₂ (Fig. 3c) looks similar to PC–FeF₂, and no FeF₂ particles are seen on the surface. All the particles are embedded in the carbon layers. The CF–FeF₂ (Fig. 3d) sample appears quite different from all the nanocomposites. Apart from the FeF₂ crystallites embedded in the carbon layers, a large number of FeF₂ crystallites (up to 300 nm) can be observed on the surface of the fibers. The reason for such behavior is not understood at present.

Fig. 3 High resolution SEM images of (a) PC–FeF₂ (b) CB–FeF₂ (c) G–FeF₂ and (d) CF–FeF₂.

Fig. 4 shows the bright-field TEM images of the C–FeF₂ nanocomposites. All the nanocomposites show a similar morphology. The high resolution TEM image of G–FeF₂ is shown in Fig. S4 (see ESI†). The FeF₂ crystallites are embedded in the carbon matrix. The crystallite sizes are in the range of 5–12 nm. Fig. 5 shows the SEAD patterns corresponding to the TEM images shown in Fig. 4. SAED of PC–FeF₂ (Fig. 5a), CB–FeF₂ (Fig. 5b) and CF–FeF₂ (Fig. 5d) exhibit a similar pattern and demonstrate the nanocrystalline nature of the FeF₂. However, the SAED pattern of CF–FeF₂ shows additional rings in agreement with the lattice distances seen in iron carbide. In the case of G–FeF₂ (Fig. 5c) in addition to the nanocrystalline FeF₂ few rings with bright spots were observed, which is attributed to the graphite as the d-values of the spots match with the d-values from graphite.

Fig. 4 Bright-field TEM images of (a) PC–FeF₂ (b) CB–FeF₂ (c) G–FeF₂ and (d) CF–FeF₂.

Fig. 5 Selective area electron diffraction (SAED) of TEM images shown in (a) PC–FeF₂ (b) CB–FeF₂ (c) G–FeF₂ and (d) CF–FeF₂.

Electron energy loss spectroscopy (EELS) was performed to further understand the electronic structure of carbon and valence state of iron fluoride. Carbon K-edge spectra are shown in Fig. 6a and the F-K edge and Fe-L edge is shown in Fig. 6b. The carbon spectra of all the nanocomposites show a typical shape for graphitic or amorphous carbon where the “Energy loss near edge structure” (ELNES) indicates two peaks for the transition of the C K-shell electrons to π*- (284.5 eV) and σ*- (290.5 eV) antibonding states.³¹ The ELNES structure of G–FeF₂ shows the most defined peaks, which indicates a more ordered sp² hybridized carbon compared to the other three samples, where the structure is characteristic for amorphous carbon.³² The F-K edge, the Fe-L₃ edge, and the Fe-L₂ edge are shown in Fig. 6b. In case of significant amounts of FeF₃, a pre-peak would be expected in EELS and XAS measurements at 684.6 eV.^31,33 The absence of this pre-peak in the EELS spectra rules out the presence of significant amounts of FeF₃. This suggests that the observation of Fe³⁺ in the Mössbauer spectra is due to oxidation of the sample.

Fig. 6 Electron energy loss spectra (EELS) recorded for (a) carbon edge and (b) fluorine and iron edge.

The electrical resistivity of the samples was measured to understand the conducting nature of the composites. Fig. 7 shows the resistivities of C–FeF₂ nanocomposites. The resistivities of PC–FeF₂, CB–FeF₂, G–FeF₂ and CF–FeF₂ are 66, 1750, 7 and 314 ohm cm respectively. CB–FeF₂ composites showed high resistivity compared to PC–FeF₂, G–FeF₂ and CF–FeF₂. The large difference in the resistivity of CB–FeF₂ and other composites could be due to the lack of iron carbide in CB–FeF₂. Iron carbide shows semiconductivity or metallic conductivity depending on the composition.³⁴ Therefore the presence of iron carbide in PC–FeF₂, G–FeF₂ and CF–FeF₂ might result in the reduced resistivity. The very low resistivity of G–FeF₂ could be due to the large graphitic domains and the high carbon content.

Fig. 7 Resistivities of the C–FeF₂ nanocomposites.

Electrochemical studies

Even though the resistivity of CB–FeF₂ is high, initial electrochemical studies were performed without the addition of extra carbon as we aimed to understand the nature of the carbon precursor on the electrochemical properties of the nanocomposites. Fig. 8 shows the discharge/charge curves of C–FeF₂ nanocomposites for the first 20 cycles obtained at 25 °C. Capacities are calculated based on the total weight of the nanocomposite in the electrode (i.e., 90% of the total electrode weight). The total first discharge capacities of PC–FeF₂, CB–FeF₂, G–FeF₂ and CF–FeF₂ are 326, 540, 442, and 225 mA h g⁻¹ respectively. C–FeF₂ nanocomposites derived from FPC and FCF delivered much less capacity compared to the CB–FeF₂ and G–FeF₂ nanocomposites. The first discharge capacities correlated well with the particle size of the CFx precursors. Larger particle size resulted in less discharge capacity. It appears that, in the case of large particles, accessing the FeF₂ nanocrystallites embedded in the core of the particle is difficult. Consequently, some of the FeF₂ crystallites do not participate in the electrochemical reaction with lithium and hence less discharge capacity was observed. The first discharge curve of C–FeF₂ composite shows two voltage regions; sloping region between 3.0–1.8 V and by a plateau region between 1.8–1.3 V. Further, there is an additional voltage plateau observed in the case of CF–FeF₂, which could be due to inhomogeneity in particle size (evident from SEM, Fig. 3d). FeF₂ should react at a single voltage plateau with lithium.³⁵ Bigger particles of FeF₂ react at a lower voltage compared to nanocrystalline FeF₂ due to the sluggish kinetics, which might have resulted in two different voltage plateaus observed in CF–FeF₂. Indeed, two different sized FeF₂ crystallites were observed in CF–FeF₂ nanocomposites (Fig. 3d). In addition to the plateau at ∼1.8 V, all the C–FeF₂ nanocomposites delivered significant capacity in the 3.0–1.8 V region. The discharge capacity contribution in the 3.0 V to 1.8 V region is 37 mA h g⁻¹, 42 mA h g⁻¹, 75 mA h g⁻¹, 54 mA h g⁻¹ for PC–FeF₂, CB–FeF₂, G–FeF₂ and CF–FeF₂ respectively. This could be connected to the insertion of lithium into nanocrystalline FeF₂. Yamakawa et al. investigated the lithium insertion mechanism into nanocrystalline FeF₃ and FeF₂ by using solid-state NMR, XRD, and PDF analysis.³⁶ In the case of nanocrystalline FeF₂, they also observed significant capacity contribution in the 3.0 V to 1.8 V region, which was attributed to the insertion of Li in FeF₂. We can, therefore, attribute the capacity observed in the in the 3.0 V to 1.8 V region to the insertion of lithium into nanocrystalline FeF₂.

Fig. 8 Electrochemical discharge/charge curves of C–FeF₂ nanocomposites for the first 20 cycles at 25 °C. The discharge and charge curves are obtained at a constant current of 20 mA g⁻¹ in the voltage window 1.3–4.3 V. The capacities are calculated with respect to total weight of the nanocomposite (i.e., 90% of the total electrode weight).

We also investigated the electrochemical performance of C–FeF₂ nanocomposites at 40 °C. Fig. 9 shows the discharge/charge curves of C–FeF₂ composites obtained at 40 °C. The first discharge capacities of PC–FeF₂, CB–FeF₂, G–FeF₂ and CF–FeF₂ are 390, 626, 507, and 318 mA h g⁻¹ respectively. The discharge capacity increased with increase in temperature in all the cases. Similarly, the capacities, in the 3.0 V to 1.8 V region also increased with an increase in cycling temperature. Fig. 10 shows the cycling behavior of C–FeF₂ nanocomposites obtained at 25 °C and 40 °C. The electrochemical performance of CB–FeF₂ and G–FeF₂ correlated well with the particle size and values of electrical resistivity. In the case of C–FeF₂ derived from FPC and FCF, capacity faded rapidly and the reversible capacity reduced to 20 mA h g⁻¹ within a few cycles. CB–FeF₂ shows high initial capacity, but capacity faded continuously with cycling. A reversible capacity of 145 mA h g⁻¹ was obtained after 70 cycles at 25 °C. In contrast, G–FeF₂ composites showed less reversible capacity and rapid capacity fading initially compared to CB–FeF₂, but the capacity stabilized after a few cycles. A reversible capacity of 163 mA h g⁻¹ was obtained after 100 cycles at 25 °C. At 40 °C, CB–FeF₂ samples showed much faster capacity fading, a reversible capacity of 184 mA h g⁻¹ was obtained after 30 cycles. On the other hand, in G–FeF₂ nanocomposites shows less capacity fading and high reversible capacity, a reversible capacity of 340 mA h g⁻¹ was obtained after 30 cycles, which is almost two times to that of CB–FeF₂ nanocomposites. The smaller particle size of FCB resulted in high initial capacity due to the reaction of all FeF₂ particles but resulted in continuous capacity fading as the carbon in which the FeF₂ nanocrystallites embedded is resistive. In contrast, in G–FeF₂, capacity faded initially, probably due to the larger particle size of CFx but showed much better cycling stability due to the high conducting nature of graphitic carbon backbone. Pereira et al. investigated the electrochemical properties of C–FeF₂ nanocomposites (at 60 °C) synthesized by a combination of chemical and mechanical milling.⁹ The C–FeF₂ nanocomposites showed high capacity of 420 mA h g⁻¹ in the first cycle, but capacity faded rapidly with cycling, similar to CB–FeF₂ nanocomposites investigated here.

Fig. 9 Electrochemical discharge/charge curves of C–FeF₂ nanocomposites for the first 20 cycles at 40 °C. The discharge and charge curves were obtained at a constant current rate of the 20 mA g⁻¹ in the voltage window 1.3–4.3 V. The capacities are calculated with respect to total weight of the nanocomposite (i.e., 90% of the total electrode weight).

Fig. 10 Electrochemical cycling of C–FeF₂ at a constant current of 20 mA g⁻¹ in the voltage window of 1.3–4.3 V (a) at 25 °C and (b) at 40 °C. The capacities are calculated with respect to the total weight of the nanocomposite (i.e., 90% of the total electrode weight).

CB–FeF₂ composites delivered high reversible capacity, but capacity faded rapidly with cycling due to the low electronic conductivity of the carbon backbone. Enhancing the electronic conductivity CB–FeF₂ might mitigate the capacity fading. Therefore, we deliberately added carbon nanofibers (CNF) to CB–FeF₂ nanocomposite and electrode were made similar to CNF free CB–FeF₂ electrodes. Fig. S5† shows the EIS spectra of PC–FeF₂, CB–FeF₂, G–FeF_2, CF–FeF₂ and PC–FeF₂ + CNF cells. The resistance of the cells decreased in the order of CB–FeF₂ > CF–FeF₂ > PC–FeF₂ > G–FeF₂ which is consistent with DC electronic conductivity measurements. After mixing of CNF, CB–FeF₂ cells showed lower resistance compared to all other C–FeF₂ cells, which confirms that the overall resistance of the CB–FeF₂ electrode reduced significantly after the addition of CNF. Fig. 11 shows the electrochemical performance of CB–FeF₂ + CNF cells cycled at 25 °C and 40 °C. The cyclic performance of the cells was similar to the cells of CNF free CB–FeF₂ cells. These results elucidate that the electronic conductivity of the carbon which is in direct contact with FeF₂ nanocrystallites plays a major role in determining the electrochemical performance of FeF₂ while the total amount of even highly conductive carbon that is present in the electrode plays an inferior role.

Fig. 11 Electrochemical cycling of CB–FeF₂ + CNF at a constant current of 20 mA g⁻¹ in the voltage window of 1.3–4.3 V (a) at 25 °C and (b) at 40 °C. The capacities are calculated with respect to total weight of the nanocomposite (i.e., 80% of the total electrode weight).

Conclusion

Four different C–FeF₂ nanocomposites were synthesized by reacting Fe(CO)₅ with four different CFx precursors at 250 °C. The four C–FeF₂ nanocomposites differed in the initial particle size of the carbon matrix, and electronic conductivity of the resulting C–FeF₂ nanocomposites. Both particle size and the electronic conductivity play a crucial role in determining the electrochemical performance of the C–FeF₂ nanocomposites. G–FeF₂ nanocomposites show less capacity fading and highly reversible capacity at 40 °C. Here, a reversible capacity of 340 mA h g⁻¹ was obtained after 30 cycles. The reversible capacity of G–FeF₂ nanocomposites can be improved further by reduction of the initial particle size. Further, the electronic conductivity of the carbon that is directly attached to the FeF₂ nanocrystallites plays a major role in determining the electrochemical performance of FeF₂ rather than the total amount of the carbon that is present in the electrode. The results obtained here provide an opportunity to study and understand the requirements for the design of conversion electrode materials for reversible lithium storage.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

MAR acknowledges Advanced Research Chemicals for kindly providing various CFx precursors used in this study.

References

1. D. Andre, S. J. Kim, P. Lamp, S. F. Lux, F. Maglia, O. Paschosa and B. Stiaszny, J. Mater. Chem. A, 2015, 3, 6709–6732.
2. C. Li, K. Chen, X. Zhou and J. Maier, NPJ Comp. Mat., 2018, 4, 22.
3. H. Li, P. Balaya and J. Maier, J. Electrochem. Soc., 2004, 151, A1878–A1885.
4. G. G. Amatucci and N. Pereira, J. Fluorine Chem., 2007, 128, 243–262.
5. J. Cabana, L. Monconduit, D. Larcher and M. R. Palacin, Adv. Mater., 2010, 22, E170–E192.
6. M. Fichtner, Phys. Chem. Chem. Phys., 2011, 13, 21186–21195.
7. F. Badway, N. Pereira, F. Cosandey and G. G. Amatucci, J. Electrochem. Soc., 2003, 150, A1209–A1218.
8. I. Plitz, F. Badway, J. Al-Sharab, A. DuPasquier, F. Cosandey and G. G. Amatucci, J. Electrochem. Soc., 2005, 152, A307–A315.
9. N. Pereira, F. Badway, M. Wartelsky, S. Gunn and G. G. Amatucci, J. Electrochem. Soc., 2009, 156, A407–A416.
10. T. Li, L. Li, Y. L. Cao, X. P. Ai and H. X. Yang, J. Phys. Chem. C, 2010, 114, 3190–3195.
11. L. Li, F. Meng and S. Jin, Nano Lett., 2012, 12, 6030–6037.
12. S. W. Kim, D. H. Seo, H. Gwon, J. Kim and K. Kang, Adv. Mater., 2010, 22, 5260–5264.
13. C. Li, L. Gu, S. Tsukimoto, P. A. van Aken and J. Maier, Adv. Mater., 2010, 22, 3650–3654.
14. C. Li, L. Gu, J. Tong, T. Susumu and J. Maier, Adv. Funct. Mater., 2011, 21, 1391–1397.
15. C. Li, L. Gu, J. Tong and J. Maier, ACS Nano, 2011, 5, 2930–2938.
16. B. Li, D. W. Rooney, N. Zhang and K. Sun, ACS Appl. Mater. Interfaces, 2013, 5, 5057–5063.
17. Y. Ma, C. Zhang, G. Jia and J. Y. Lee, J. Mater. Chem., 2012, 22, 7845–7850.
18. K. Rui, Z. Wen, Y. Lu, J. Jin and C. Shen, Adv. Energy Mater., 2015, 5, 1401716.
19. Y. Lu, Z. Wen, J. Jin, X. Wu and K. Rui, Chem. Commun., 2014, 50, 6487–6490.
20. Y. Lu, Z. Wen, J. Jin, K. Rui and X. Wu, Phys. Chem. Chem. Phys., 2014, 16, 8556–8562.
21. Y. Lu, Z. Wen, K. Rui, X. Wu and Y. Cui, J. Power Sources, 2013, 244, 306–311.
22. R. Prakash, A. K. Mishra, A. Roth, Ch. Kübel, T. Scherer, M. Ghafari, H. Hahn and M. Fichtner, J. Mater. Chem., 2010, 20, 1871–1876.
23. R. Prakash, C. Wall, A. K. Mishra, C. Kübel, M. Ghafari, H. Hahn and M. Fichtner, J. Power Sources, 2011, 196, 5936–5944.
24. C. Wall, R. Prakash, C. Kübel, H. Hahn and M. Fichtner, J. Alloys Compd., 2012, 530, 121–126.
25. M. Anji Reddy, B. Breitung, V. S. K. Chakravadhanula, C. Wall, M. Engel, C. Kübel, A. K. Powell, H. Hahn and M. Fichtner, Adv. Energy Mater., 2013, 3, 308–313.
26. B. Breitung, M. Anji Reddy, V. S. K. Chakravadhanula, M. Engel, C. Kübel, A. K. Powell, H. Hahn and M. Fichtner, Beilstein J. Nanotechnol., 2013, 4, 705–713.
27. M. Anji Reddy, B. Breitung, C. Wall, S. Trivedi, V. S. K. Chakravadhanula, M. Helen and M. Fichtner, Energy Technol., 2016, 4, 201–211.
28. L. Lutterotti and P. Scardi, J. Appl. Crystallogr., 1990, 23, 246–252.
29. S. Ramasamy, J. Jiang, H. Gleiter, R. Birringer and U. Gonser, Solid State Commun., 1990, 74, 851–855.
30. G. LeCaer and E. Bauer-Grosse, Hyperfine Interact., 1989, 47, 55–67.
31. A. A. El-Barbary, S. Trasobares, C. P. Ewels, O. Stephan, A. V. Okotrub, L. G. Bulusheva, C. J. Fall and M. I. Heggie, J. Phys.: Conf. Ser., 2006, 26, 149–152.
32. K. N. Kushita and K. Hojou, Ultramicroscopy, 1991, 35, 289–293.
33. F. Cosandey, J. F. Al-Sharab, F. Badway, G. G. Amatucci and P. Stadelmann, Microsc. Microanal., 2007, 13, 87–95.
34. M.-C. Lee and G. Simkovich, Metal. Trans. A, 1987, 18, 485.
35. F. Wang, H.-C. Yu, M.-H. Chen, L. Wu, N. Pereira, K. Thornton, A. Van der Ven, Y. Zhu, G. G. Amatucci and J. Graetz, Nat. Commun., 2012, 3, 1201.
36. N. Yamakawa, M. Jiang, B. Key and C. P. Grey, J. Am. Chem. Soc., 2009, 131, 10525–10536.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c8ra07378c

‡ Presently at Materials Characterisation Division, Vikram Sarabhai Space Centre, Trivandrum, India.

---