Bulk Structure and Electrochemical Performance of . The bulk structure of the CFx materials was characterized using XRD, TEM, SAED, HAADF-STEM imaging, and EDS mapping. As shown in Figure S1, F atoms bond with C atoms in an sp3 hybrid structure; the C atoms have a cyclohexane chair-type structure with AA stacking in the CFx structure.13,23 Figure 1b shows a typical XRD pattern of the CFx sample with the broad diffraction peaks representing the disordered phase. The peak at 13° is assigned to the (002) reflection of the CFx.9,10,24 The peak at 41° corresponds to (100) and (011) reflections, which is related to the C-C bond length and CF interlayer distance. Figure 1c shows a typical TEM bright field image of the CFx material with large sheet-like topography, and the corresponding high resolution TEM image (Figure 1d) shows a disordered amorphous structure. As shown in Figure 1e, there are two diffraction rings representing the (100), (011) and (110) planes of CFx in the SAED pattern, which is consistent with the XRD results. As shown in the HAADF-STEM image and corresponding color mixed EDS mapping of the CFx sample (Figure S2), C and F distribution are uniform in the sample.
Next, we tested CFx (x = 0.88) as a cathode material for K-ion batteries (KIBs), Na-ion batteries (SIBs), and LIBs, and typical discharge curves are shown in Figure 1f. Under the same CFx discharge current, the KIBs, NIBs, and LIBs discharge specific capacities are 796.1 mAh g-1, 770.6 mAh g-1, and 741.6 mAh g-1, respectively, and the discharge plateaus are ~2.7 V, ~2.6 V, and ~2.4 V, respectively, which is in agreement with our previous work.13 As the ionic radius of K, Na, and Li is 138 pm, 102 pm, and 76 pm, respectively, the discharge rate specific capacity of Li-ion electrode materials is normally larger than that of Na- and K-ion materials because of the higher mobility of Li+.25,26 However, the inverse phenomenon of this electrochemical performance is observed, when using CFx as the cathode in KIBs, NIBs, and LIBs, probably due to the Strokes’ radius and electrolyte-assisted reaction mechanism.
Ion Migration Kinetics of K/Na/Li Insertion. The real time K/Na/Li ion insertion mobility and reaction mechanism of the CFx cathode was studied using the in situ all-solid-state nanobattery TEM technique (Figure 1a), which has been applied to study the reaction mechanism of electrode materials for batteries at high resolution.26,27 We used M/M2O (M = K, Na, and Li) on a W tip as the counter electrode/solid electrolyte and CFx on a lacey carbon Cu TEM grid as the active electrode for in situ experiments. Figure 2, Figure 3, Figure S36, and Movie S1-3 show the real time in situ time series images and analysis results during K/Na/Li ion transport in a nano-sized CFx sample. As shown in Figure 2a and Movie S1, K uniformly intercalates into the CFx during the discharge process and the projected area of the entire sample is almost unchanged (Figure 2d), suggesting no volume change during potassiation. Upon K intercalation, a clear phase boundary can be seen between pristine CFx and the reacted domain in Figure 2a. The two phase boundaries are more obvious in Figure S3, which shows the STEM-EDS elemental mapping of C, F and K. The two-phase reaction mode is further confirmed by electron diffraction patterns. As shown in Figure 3b, comparing the SAED pattern of unreacted region (I) with the SAED pattern of region II, the additional rings that arise due to diffraction from KF can be seen. The SAED pattern after potassiation suggests that CFx converted to KF and C, which is the same as previous work.13 Figure 3c shows the high resolution TEM (HRTEM) images of CFx after potassiation. The C product maintains amorphous features, and KF nanoparticles (~4 nm) are formed and uniformly distributed in the amorphous carbon matrix. Furthermore, considering the HAADF image contrast is roughly proportional to Z1.7 (Z represents the atomic number), it was possible to confirm the two-phase boundary using a video taken with in situ HAADF-STEM imaging. Figure S6a–d shows a series of real time HAADF-STEM images, the raw video is shown in Movie S4. As shown in Figure S6b, the brighter domain represents the reaction area with K ion intercalation, and the darker domain represents the pristine area. As shown in Movie S4, movement of the phase boundary was observed during K ion insertion, and this clear phase boundary is characterized using EDS mapping and EDS spectra (Figure S6e–h). Due to an increasing projected reaction area during the phase boundary motion, the K ion diffusivity in CFx is estimated to be ~232.9 nm2/s in Figure 2d, based on the equation D = d2/t, where D is ion diffusivity, d2 is the projection area, and t is the diffusion time.28
Sodiation of the CFx sample exhibits a similar two-phase response, as shown in Figure 2b, 3d–f, S4 and S6i–p, and Movie S3, S5. Observing the TEM and HAADF-STEM images, we conclude that the volume of CFx remains unchanged after sodiation (Figure 2e), and the SAED pattern (Figure 3e) shows a two-phase reaction representing the transformation of CFx to NaF and C, which is in agreement with previous studies.29,30 After Na intercalation, 6–7 nm NaF nanoparticles are evenly distributed in the amorphous carbon matrix (Figure 3f). A two-phase boundary is observed in the elemental mapping of C, F, and Na (Figure S6m–o), and the Na ion diffusivity is ~479.7 nm2/s, as calculated using the projected reaction area (Movie S2, Figure 2e). As a parallel comparison, the reaction of CFx with Li+ was also examined through in situ TEM, and the raw data are shown in Movie S3, Figure 2c, 2d and Figure S5. Upon Li ion intercalation, the projected area of CFx remained the same and become smaller (because of sample rotational movement), indicating the volume remains unchanged. According to STEM-EDS mapping (Figure S4), C and F are remained uniformly distributed, suggesting no product aggregation phenomena. The two-phase boundary moves fast but is still clear, and the Li ion diffusivity is ~2133.0 nm2/s, as calculated using the projected reaction area (Movie S3). As shown in Figure S5f–g, the EELS Li K edge and F K edge spectra are in agreement with that reported for LiF,31 suggesting that LiF is formed after Li ion intercalation. However, in the SAED pattern of pristine CFx (region V) and after lithiation (region VI) (Figure 3h), the rings of region VI could be assigned to (111), (220), and (311) of Li2O, indicating that Li2O is formed after lithiation. As shown in Figure 3i, Li2O nanoparticles with a size of ~3 nm are uniformly distributed in the amorphous carbon matrix after lithiation. According to previous work, LiF crystals have been formed and detected using XRD analysis and NMR spectra in liquid Li ion primary batteries during discharge process.8,9,13,16,17 Considering the all-solid-state environment in in situ TEM, amorphous LiF is formed in the solid-state Li-CFx system, which is consistent with that reported by Rangasamy.32 As for the formed Li2O, Li will diffuse along the surface of the materials during in situ TEM, and the active Li will react with trace O2 in the TEM instrument, resulting in Li2O formation in CFx.
Upon formation of a solid-state M-CFx in situ TEM system, all alkali ions uniformly diffuse across the materials, with diffusivities of 232.9 nm2/s, 479.7 nm2/s, and 2133.0 nm2/s for K, Na, and Li ion insertion in CFx, respectively. However, in the liquid electrolyte, the discharge specific capacities of CFx for KIBs, NIBs and LIBs are 796.1 mAh g-1, 770.6 mAh g-1, and 741.6 mAh g-1, respectively, indicating that the ion mobility of K is higher than that of Na and Li. In the liquid-state M-CFx system, large fluoride crystals are formed and a higher K+ diffusivity than Na+ and Li+ was observed (10-14-10-10 cm2s-1 for K+/Na+ diffusion and 10-15–10-11 cm2s-1 for Li+ diffusion). 13 However, if organic solvent molecules are absent from a solid-state battery system, the ion diffusivity is determined by alkali ion insertion ability, as demonstrated by the high ion diffusivities observed for the Li ion in our study using in-situ TEM.
To further understand the alkali ion intercalation behavior in CFx, DFT is applied to calculate the structure and physical-chemical properties after Li/Na/K ion intercalation (Figures S7–10). As shown in Figure S8, the binding energies for Li, Na, and K adsorption in C4F4 at high concentrations of M2C4F4 are 4.80 eV, 2.49 eV, and 0.45 eV, respectively, which indicates Eb(Li) > Eb(Na) > Eb(K). This relationship is the same as results reported by Yoon33 and our previous work13. A high bind energy leads to a high ionic conductivity, which is consistent with our experimental results (2133.0 nm2/s for Li, 479.7 nm2/s for Na, and 232.9 nm2/s for K). The partial density of states of pristine CF and Li/Na/K adsorbed CF (M2C4F4) is also acquired through DFT calculations and shown in Figure S9. Herein, and based on the theoretical calculations, the CF structure is a semiconductor with a bandgap of 2.4 eV, which is consistent with previous work. 33 For a successful application as battery cathodes, good electrical conductivity is important. After adsorption of Li/Na/K in CF, the Fermi level contains various electronic states, indicating improved electric conductivity. As shown in Figure 1c, the initial discharge plateau is low, and then increases to the original state as the discharge process progresses, this is related to the increased conductivity resulting from Li/Na/K adsorption. Furthermore, the ion diffusion pathways and energy profiles for the diffusion of Li/Na/K ions in CF are also calculated and shown in Figure 4. Diffusion of Li/Na/K ions at the interlayer of CF proceeds via a zigzag pathway (Figure 4a–b); the energy diffusion barrier for Li/Na/K ions is 0.137 eV, 0.568 eV, and 1.591 eV, respectively (Figure 4c–e). A low energy diffusion barrier indicates a high ionic mobility, which further confirms our experimental results.
Phase Transformation and Structural Evolution. As discussed above, different alkali ions resulted in an unchanged volume during the solid-state reaction; alkali fluoride nanoparticles evenly distributed in the carbon matrix were the resultant products. To probe the phase evolution of CFx during alkali ion insertion, in situ electron diffraction is applied to study the samples during the entire discharge process and the results are shown in Figure 5, Figure S11 and Movie S6-S8. Figure 5a presents the radial intensity profiles as a function of reaction time retrieved from in situ SAED (Movie S6) upon K ion intercalation. The original radial intensity profiles are acquired by integrating the intensity of a series of SAED patterns along the r direction over the full 2π range, and the final radial intensity profiles are obtained after background subtraction using the power-law model (Figure 5a). Before potassiation, characteristic diffraction rings for CFx corresponding to (100), (011) and (110) peaks were observed (Figure S10a). As K ion intercalation progresses, several new diffraction rings appeared, initially faint, then becoming stronger with time. In the corresponding radial intensity profiles (Figure 5a), new diffraction peaks that become stronger with time appeared, and these new diffraction peaks are assigned to the (200) and (220) planes of KF. As shown in Figure 5b, the diffraction peaks for pristine CFx remain and shift left over lower distances, indicating the spacing of monolayer CF increased and the C-C bond length increased after K ion intercalation. Moreover, Figure 5c shows the full width at half maxima (FWHM) for the KF peak (200), and the corresponding calculated crystal size of KF-based on the Scherrer equation.34 After potassiation, ~3.3-nm KF nanoparticles are immediately formed, and the particle sizes remain the same with increasing reaction time, which is in agreement with HRTEM results. The formation of KF nanoparticles also demonstrates the migration of F ions during the reaction, implying a conversion reaction occurs during potassiation. As a parallel comparison, the reaction of CFx with Na ions was examined with in situ electron diffraction, the raw data from a typical experiment are shown in Movie S7. The radial intensity profile and corresponding diffraction patterns with false colors are shown in Figure 5d–f and Figure S10b. As the Na ion intercalation progresses, several new diffraction rings which could be indexed to the cubic NaF (space group Fm-3m), appeared, faint initially and becoming stronger with time. As shown in Figure 5e, and different to Na ion insertion, the CFx diffraction peak for does not shift during Na ion intercalation, suggesting the spacing of the carbon monolayer product is unchanged. As shown in Figure 5f, ~6.3-nm NaF nanoparticles are formed based on the (200) peak of NaF; their size is retained with an increasing reaction time. Lithiation of CFx exhibits a similar phase evolution phenomenon, and is shown in Movie S8, Figure 5g-i and Figure S10c. Upon Li ion intercalation, several new diffraction rings appear, initially faint, and becoming stronger over time (Figure S10c). As shown in Figure 5h, as the reaction time increases, new, higher, diffraction peaks appear; these new peaks can all be indexed to the cubic Li2O phase (space group Fm-3m). Furthermore, the (100), (011), and (110) peak position for the resulting carbon monolayer was maintained with increasing reaction time (Figure 5h), implying that the amorphous feature of CFx remained unaltered after lithiation. As shown in Figure 5i, the Li2O nanoparticle size slowly increases from 3 nm to 5 nm, which may account for Li aggregation on the surface. As we know, if amorphous LiF is formed in a solid-state Li-CFx system, no LiF crystal diffraction rings are detected during in situ electron diffraction. During K/Na/Li intercalation, the original diffraction peaks for the (100) and (011) planes change, which is connected to the embedded ion structure. Figure S7a shows the optimized structure of M2C4F4 which is calculated using DFT. As shown in Figure S7a, the distance of the CF monolayer for pristine C4F4 is 5.574 Å, and becomes larger, 6.288 Å, 6.592 Å, and 7.285 Å, after Li/Na/K ion insertion, respectively. The corresponding diffraction profile is calculated using CrystalDiffract software based on the optimized structure and is shown in Figure S7b. The peak position for the (100) and (011) plane shifts to the right with increased distance after K ion insertion, while the peak position remains unchanged for Na ion insertion, which is consistent with experimental results. Remarkably, the peak position shift left due to the transformation from chair configuration to a flat graphite-like structure during Li intercalation, which is in accordance with in-situ experiments and as reported in previous studies. 13,35
As mentioned above, the electrochemical solid-state reaction of CFx with alkali metal ions is expected to behave according to the following equation:
CFx + xK+ + xe- = C + xKF (crystal) (1)
CFx + xNa+ + xe- = C + xNaF (crystal) (2)
CFx + xLi+ + xe- = C + xLiF(amorphous) (3)
This reaction formula is similar to that of liquid-state batteries, the only difference being that crystalline LiF is produced in the liquid-state mode. Upon K/Na/Li ion insertion, the resulting carbon monolayer maintains its amorphous features. However, the CFx monolayer distance becomes larger with ion insertion, resulting in (100) and (011) peak shifts during K ion intercalation. In liquid-state batteries, the alkali fluoride will dissolve and then aggregate on the CFx surface, resulting in large crystals (~200–300 nm) and theoretical volume expansion of the electrode (56.9% for K insertion, 32.9% for Na insertion, 22.9% for Li insertion).8,11,13 However, in the solid-state system, alkali fluoride nanoparticles are formed immediately and the size of the alkali fluoride nanoparticles remains unchanged with time. As the alkali fluoride nanoparticles are uniformly distributed in the amorphous carbon matrix and the size remains unchanged during the reaction, volume expansion does not occur, which is beneficial for solid-state battery systems. Furthermore, the diffusivity is 232.9 nm2/s (2.329 × 10-12 cm2/s), 479.7 nm2/s (4.797 × 10-12 cm2/s), and 2133.0 nm2/s (2.133 × 10-11 cm2/s) for K, Na, and Li ion insertion, respectively, in solid CFx cells, which is comparable to the diffusivity of Li/Na/K ions in liquid-state alkaline cells (10-14–10-10 cm2 s-1 for K+/Na+ diffusion and 10-15–10-11 cm2 s-1 for Li+ diffusion) 13, this also indicates that CFx is a suitable cathode for solid-state cells. In order to test the electrochemical performance as a cathode in a solid-state battery, the solid-state Li/PEO-LiTFSI/CFx cell is fabricated using PEO-LiTFSI as the solid electrolyte. As shown in Figure S11, the discharge capacity of the solid-state Li/PEO-LiTFSI/CFx cell is 507.4 mAh g-1 (1215.4 Wh kg-1) at room temperature with a discharge current of 0.02C, indicating that CFx can be used as a cathode material for high energy density solid-state primary batteries.
Na/Li Extraction Reaction Mechanism. As LiF cannot be electrochemically decomposed merely, it is widely accepted that the reaction of CFx with lithium is likely irreversible, meaning Li/CFx can only be used as a primary cell. However, recently, a reversible charge/discharge capacity of 786 mAh/g for a Na/CFx system was achieved, illustrating that CFx is a promising cathode material for SIBs.11 To investigate the Na ion extraction reaction mechanism of CFx, in situ TEM is applied to study the samples during the entire charge process, the raw data and analysis results are shown in Figure 6a–e, Figure S12 and Movie S9. As shown in Figure 6a, no phase boundaries occur during Na ion extraction. From 431 s to 435 s, the NaF product homogeneously and immediately decomposed under a 5-V charge, which is further confirmed by the SAED patterns. Before desodiation, the diffraction rings in Figure 6c can be index to NaF. After desodiation, the diffraction rings of NaF disappear, and the remaining diffraction rings in Figure 6e can be index to amorphous carbon, which has same features as that of pristine CFx. As shown in the HRTEM images (Figure 6b), NaF nanoparticles are evenly distributed in the amorphous carbon matrix before the charge process, during Na ion extraction NaF decomposes leaving amorphous carbon (Figure 6d). Figure S12 shows the HAADF-STEM image and corresponding EDS spectrum and elemental mapping of CFx after sodiation and desodiation. Comparing the disappearance of Na and F after the charging process indicates NaF decomposes into Na, which returns to the Na/NaxO probe, and F2, which evaporates into the vacuum system. Delithiation of CFx during the charging progress exhibits a similar decomposition phenomenon, as shown in Movie S10, Figure 6f–j and Figure S13. As shown in Figure 6f, there are also no phase boundaries during Li extraction. According to the SAED pattern in Figure 6h–j, the Li2O nanoparticles on the CFx surface all decompose following delithiation, which is further verified by the HRTEM images in Figure 6g–i. Interestingly, the F disappears from the STEM-EDS mapping and EDS spectrum, suggesting all the amorphous LiF decomposes into Li and F2 under a 5 V charge. The electrochemical solid-state reaction of CFx with Na/Li ions during the charging process is expected to behave according to the following equations:
C + xNaF (crystal) - xe- = C+ 1/2x F2 + x Na (4)
C + xLiF (amorphous) - xe- = C+ 1/2x F2 + x Li (5)
As discussed above, the in situ all-solid-state nanobattery TEM technique was employed to explore the reaction mechanism of CFx with K/Na/Li during ion intercalation/extraction. In the M/CFx liquid-state battery, the K ion diffusivity is higher than that of Li and Na ions, and the discharge products are amorphous carbon and large alkali fluoride crystals. In the solid-state battery system, the alkali ion intercalation reaction mechanism is significantly different from that of the liquid-state system, see Figure 7. In the M/CFx solid-state battery system, a two-phase reaction and phase boundary movement (Figure 2–3) is observed during alkali ion intercalation. Based on the two-phase boundary movement (Figure 2), the diffusivity of K/Na/Li ion intercalation in CFx is 232.9 nm2/s, 479.7 nm2/s, and 2133.0 nm2/s, respectively, which is further investigated through DFT calculations (Figure 4). According to the in situ electron diffraction results (Figure 5), crystalline KF, NaF nanoparticles, and amorphous LiF are formed and remain unchanged with increasing reaction time; furthermore, these alkali fluoride nanoparticles are uniformly distributed in the amorphous carbon matrix, resulting in no volume change. During the charging progress in the solid-state cell (Figure 6), the NaF product decomposes into Na and F2, leading to the possibility of a reversible Na/CFx battery, which is in agreement with previous studies.11, 12, 40, 45Converse to the understanding that the charging-induced decomposition of LiF is not possible, our observations show LiF simultaneously dissociates into F2 and Li following charging. According to the volume expansion limit and high ion diffusivity, the all-solid-state Li/PEO-LiTFSI/CFx cell possess a discharge capacity of 507.4 mAh g-1 with a high energy density of 1215.4 Wh kg-1 at room temperature (Figure S11).