Realizing Complete Solid-Solution Reaction to Achieve Temperature Independent LiFePO 4 for High Rate and Low Temperature Li-Ion Batteries

Open AccessCCS ChemistryRESEARCH ARTICLE3 Mar 2022Realizing Complete Solid-Solution Reaction to Achieve Temperature Independent LiFePO4 for High Rate and Low Temperature Li-Ion Batteries Bingqiu Liu, Qi Zhang, Yiqian Li, Yuehan Hao, Usman Ali, Lu Li, Lingyu Zhang, Chungang Wang and Zhongmin Su Bingqiu Liu Northeast Normal University, Changchun, 130024 Jilin , Qi Zhang Northeast Normal University, Changchun, 130024 Jilin , Yiqian Li Northeast Normal University, Changchun, 130024 Jilin , Yuehan Hao Northeast Normal University, Changchun, 130024 Jilin , Usman Ali Northeast Normal University, Changchun, 130024 Jilin , Lu Li *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Northeast Normal University, Changchun, 130024 Jilin , Lingyu Zhang Northeast Normal University, Changchun, 130024 Jilin , Chungang Wang *Corresponding authors: E-mail Address: [email protected] E-mail Address: [email protected] Northeast Normal University, Changchun, 130024 Jilin and Zhongmin Su Northeast Normal University, Changchun, 130024 Jilin https://doi.org/10.31635/ccschem.022.202201776 SectionsSupplemental MaterialAboutAbstractPDF ToolsAdd to favoritesDownload CitationsTrack Citations ShareFacebookTwitterLinked InEmail The lithium iron phosphate battery (LiFePO4 or LFP) does not satisfactorily deliver the necessary high rates and low temperatures due to its low Li+ diffusivity, which greatly limits its applications. The solid-solution reaction, compared with the traditional two-phase transition, needs less energy, and the lithium ion diffusivity is also higher, which makes breaking the barrier of LFP possible. However, the solid-solution reaction in LFP can only occur at high rates due to the lattice stress caused by the bulk elastic modulus. Herein, pomegranate-like [email protected] nanoclusters with ultrafine [email protected] subunits (8 nm) (PNCsLFP) were synthesized. Using in situ X-ray diffraction, we confirmed that PNCsLFP can achieve complete solid-solution reaction at the relatively low rate of 0.1C which breaks the limitation of low lithium ion diffusivity of the traditional LFP and frees the lithium ion diffusivity from temperature constraints, leading to almost the same lithium ion diffusivities at room temperature, 0, −20, and −40 °C. The complete solid-solution reaction at all rates breaks the shackles of limited lithium ion diffusivity on LFP and offers a promising solution for next-generation lithium ion batteries with high rate and low temperature applications. Download figure Download PowerPoint Introduction Lithium ion batteries (LIBs) have been extensively used in electric vehicles (EVs) and hybrid electric vehicles (HEVs).1–6 Olivine-structured LiFePO4 (LFP) has become a promising candidate for cathode materials due to its excellent thermal stability, low cost, and environmental safety.7–12 However, the major problems of LFP are low electronic conductivity and lithium ion diffusivity. Li+ is incorporated strictly at the LFP/FePO4 interface due to its normal two-phase mechanism, leading to an extremely slow rate of Li+ diffusion. Moreover, a two-phase reaction needs high overpotential to drive it during lithiation/delithiation, which results in slow kinetics. Thus, the structure of LFP seriously restricts the Li+ diffusion rate in LIBs, especially at low temperatures, thus limiting its high reaction rate and application at low temperatures in EVs and HEVs.13–15 The discovery of a solid-solution reaction gives us hope that we can solve these problems. The solid-solution reaction, compared with a traditional two-phase transition, is a single-phase reaction pathway (reaction via the formation of solid solutions with intermediate lithium contents varying between the equilibrium lithium-poor and lithium-rich phases), which exempts the formation of a two-phase interface, as well as the coherency strain.16–19 In 2011, Ceder et al.20 reported that, according to canonical Monte Carlo simulations, the solid-solution reaction in LFP needs very small overpotential (∼30 mV) and can significantly improve the kinetics of Li+ diffusion with a fast 1D channel. In 2014, Zhang et al.21 observed evidence for the solid-solution reaction in LFP during charging at 5C and 10C. The single-phase process of the solid-solution reaction needs less energy, and the lithium ion diffusivity is also higher, which is very important for high rate and low temperature applications.22 Thus, it is of great importance to realize the solid-solution reaction. However, the complete solid-solution reaction in LFP can only occur at relatively high rates which has been proved by many researchers, resulting in unsatisfactory practical applications.23 The main reason for this phenomenon is that the lattice stress caused by bulk elastic modulus will play a dominant role due to the large size of primary LFP and only because the high rate can break through this limit.24,25 Thus, achieving the complete solid-solution reaction at all rates is of great value to both practical applications and basic research. And coating carbon materials is also essential to enhance the electrical properties and stability of LFP while reducing the particle size due to its poor electronic conductivity. Therefore, the solid-solution reaction at all rates will become the main strategy to fundamentally solve the problem of low lithium ion diffusivity and achieve high rate and low temperature performance. Inspired by the subunit structure of pomegranate fruits, we developed a facile and unique method for the first time to synthesize pomegranate-like [email protected] nanoclusters with ultrafine [email protected] subunits (∼8 nm) (PNCsLFP) by designing the peel and core of the kernels in the pomegranate-like structure as carbon and ultrafine LFP subunits. The ultrafine primary LFP nanoparticles make the surface energy large enough to cause the room-temperature miscibility gap to taper further inward to achieve complete solid-solution reaction at all rates.26 This unique nanostructure leads to complete solid-solution reaction at all rates to achieve almost the same lithium ion diffusivities at room temperature, 0, −20, and −40 °C. Due to the advantages of the ultrafine [email protected] subunits, the PNCsLFP exhibit excellent rate performance (125.2 and 111.2 mA h g−1 at 50C and 100C) at room temperature. Even at low temperatures (−20 and −40 °C), the PNCsLFP still display remarkable rate performance (79.2 and 40.1 mA h g−1 at 20C under −20 and −40 °C, respectively) and outstanding cycling stability (122.1 and 86.8 mA h g−1 at 0.5C under −20 and −40 °C after 300 cycles with 97.6% and 95.2% capacity retention, respectively). Furthermore, the PNCsLFP//graphite full cell also displays splendid electrochemical performance at both room temperature and −20 °C. Experimental Methods Synthesis of pomegranate-like LiFePO4@C nanoclusters (PNCsLFP) In a 20 L glass jar, a polyacrylic acid (PAA) aqueous solution (0.2 g mL−1, 40 mL) and 2.47 g Li2CO3 were added in deionized water (4 L) and ultrasonically dispersed for 30 min. After that, isopropyl alcohol (IPA) (16 L) was dripped into the conical flask under magnetic stirring to form a suspension. Subsequently, 3.07 g NH4H2PO4 and 5.34 g FeCl2.4H2O were added into the suspension under magnetic stirring for 6 h at room temperature to obtain the FePO4/PAA-Li NSs. The obtained FePO4/PAA-Li NSs were centrifuged and washed several times by anhydrous ethanol and finally dried at 50 °C for 12 h for further experimentation. The highly dispersed FePO4/PAA-Li NSs were ground and annealed at 400 °C for 5 h, followed by sintering at 700 °C for 10 h under a high-purity argon atmosphere with a heating rate of 2 °C min−1 to obtain the PNCsLFP. (Further details about these procedures can be found in the Supporting Information.) Results and Discussion Synthesis and characterization Scheme 1 illustrates the relatively simple preparation of this unique PNCsLFP. First, PAA, Li2CO3, and IPA were added into deionized water to form the PAA-Li nanospheres (NSs).27,28 Subsequently, FePO4 was generated on the network of the PAA-Li NS after adding NH4H2PO4 and FeCl2·4H2O. The product was annealed at 350 °C for 5 h and then 700 °C for 10 h under an argon atmosphere to carbonize the PAA shell and afford PNCsLFP. Figure 1a displays the transmission electron microscopy (TEM) image of PAA-Li NSs with a uniform size of 110 nm. Subsequently, FePO4/PAA-Li NSs were obtained by adding NH4H2PO4 and FeCl2·4H2O (Figure 1b). Finally, the PNCsLFP were obtained by calcinations (Figure 1c). The high-resolution TEM (HRTEM) image in Figure 1d and Supporting Information Figure S1c displays the single PNCLFP composed of numbers of ultrafine [email protected] subunits. The inset of Figure 1d suggests that the LFP nanoparticle subunit (8 nm) is coated with a carbon layer (2 nm) and the visible lattice fringes with a spacing of ∼0.43 nm corresponding to the (101) plane. Figure 1e shows scanning electron microscopy (SEM) images of the PNCsLFP with uniform spherical morphology (110 nm). Particularly, the NC is organized by many ultrafine subunits (8 nm), which can shorten the Li+ diffusion path and accelerate the Li+ transport throughout the electrode (Figure 1f). The elemental mapping images shown in Figures 1g–1k further confirm that the PNCsLFP contain elements of Fe, P, O, and C. In the current case (1.3 g of FeCl2·4H2O), about 4 g of PNCsLFP can be successfully fabricated at a time. Scheme 1 | Schematic illustration of the fabrication process of PNCsLFP. Download figure Download PowerPoint Figure 1 | TEM images of (a) PAA-Li NSs, (b) FePO4/PAA-Li NSs, SEM images of (c) PNCsLFP, (d) a single [email protected] NC. (e) PNCsLFP, (f) large magnified image of the PNCsLFP, inset: HRTEM image, the elemental mapping images of a single PNCLFP: (g) TEM image and corresponding (h) C, (i) O, (j) Fe, and (k) P. Download figure Download PowerPoint Figure 2a displays the X-ray diffraction (XRD) patterns of PNCsLFP, nanosized [email protected] and microsized LFP/C, and all diffraction peaks are indexed to olivine LFP (JCPDS No. 83-2092). PNCsLFP show weak and widened diffraction peaks, revealing an ultrafine particle size in PNCsLFP. The average particle size of LFP nanocrystals in the PNCsLFP and nanosized [email protected] can be calculated to be 8 and 62 nm by using Scherrer's formula (D = 0.89λ/βcosθ), which matches well with the results of TEM and SEM. The calculated crystal lattice parameters are a = 10.338 Å, b = 6.007 Å, c = 4.705 Å, V = 292.2 Å3 for PNCsLFP, and a = 10.347 Å, b = 6.011 Å, c = 4.704 Å, V = 292.6 Å3 for nanosized [email protected] Both of the parameters are very close to the values of standard LFP (a = 10.334 Å, b = 6.010 Å, c = 4.693 Å, V = 291.5 Å3), indicating a high crystallinity of both the samples.29 The amount of carbon in the PNCsLFP and nanosized [email protected] was tested by thermogravimetric analysis measurements, as illustrated in Supporting Information Figures S2 and S4. By taking the theoretical weight gain (5.07 wt %) of pure LFP, as shown in Supporting Information, the contents of carbon in PNCsLFP and nanosized [email protected] are calculated to be 8.6 and 9.0 wt %, respectively, revealing that the carbon contents of PNCsLFP and nanosized [email protected] are comparable with each other. Raman spectra of PNCsLFP was further determined to probe the structural information of carbon (Figure 2b). The peaks at 1338 and 1591 cm−1 are attributed to the D and G bands, respectively. The intensity ratio of the D band and G band (ID/IG) stands for the degree of surface disorder of carbon materials, and low ID/IG stands for high graphitization. The ID/IG value of PNCsLFP is calculated to be 0.94, revealing a higher graphitization degree, and thus, it can improve the electronic conductivity. The X-ray photoelectron spectroscopy (XPS) was tested to examine the composition of PNCsLFP. The wide XPS spectrum shows that Li 1s, P 2p, and O 1s are located at 55.2, 133.6, and 531.9 eV, respectively (Figure 2c). The Fe 2p spectrum in Figure 2d splits into two peaks, and the peaks locate at 710.7 and 724.2 eV, corresponding to Fe 2p3/2 and Fe 2p1/2 with an energy separation of 13.5 eV, which is similar to the recently reported spectra of Fe2+ in LFP.30 To investigate the pore structure of the PNCsLFP, we carried out Brunauer–Emmett–Teller (BET) analysis of the nitrogen adsorption desorption isotherms (Figures 2e and 2f). The PNCsLFP sample exhibits a typical isotherm of type IV with a hysteresis loop extending from P/P0 = 0.4 to 1.0, indicating a mesoporous characteristic,31 which can be easily understood as a result of the stacking of nanoparticles. The measured BET specific surface area of PNCsLFP is 134.4 m2 g−1 with an average pore diameter of 3.627 nm, which benefits rapid Li+ transport and electrolyte accessibility. Figure 2 | (a) XRD pattern and (b) Raman spectrum of the three samples. XPS for the PNCsLFP (c) the survey spectrum, (d) subspectra for Fe 2p. (e) Nitrogen adsorption–desorption isotherms and (f) pore size distributions of PNCsLFP. Download figure Download PowerPoint Electrochemical performance in half cell To evaluate the electrochemical properties of the PNCsLFP, an electrochemical study compared with nanosized [email protected] and microsized LFP/C product was performed by galvanostatic charge/discharge techniques, and 1C equals 170 mA g−1. Detailed synthesis steps and characterization of nanosized [email protected] and microsized LFP/C are shown in the experimental section in the Supporting Information ( Supporting Information Figures S3 and S5). Figure 3a shows their typical charge/discharge curves at 1C. The discharge capacity of PNCsLFP is 159.5 mA h g−1, much higher than that of nanosized [email protected] (142.9 mA h g−1) and microsized LFP/C (136.0 mA h g−1). Moreover, the polarization between the charge and discharge plateaus of PNCsLFP was smaller than nanosized [email protected] and microsized LFP/C, and the potential-plateau of PNCsLFP was more stable than that of other samples (the inset of Figure 3a), indicating that the kinetics were further improved by ultrafine [email protected] subunits. Figure 3b displays the charge/discharge curves of PNCsLFP at various rates. The high and stable discharge potential-plateau of PNCsLFP at every current rate reveals both the higher power density and energy density. The electrode discharges to 166.1 mA h g−1 at 0.2C rate, which is close to the theoretical capacity of LFP (170 mA h g−1). The material also delivers a discharge capacity of 146.1 mA h g−1 at a high rate of 10C, and the capacities stay as high as 136.7 and 125.0 mA h g−1 at 20C and 50C rates. More excitingly, even when the rate is up to 100C, the capacity can still remain at 110.6 mA h g−1, which is significantly better than most of the published LFP/C nanomaterials ( Supporting Information Table S1). Figure 3c shows the rate capabilities of PNCsLFP, nanosized [email protected] and microsized LFP/C evaluated from 0.2C to 100C. The PNCsLFP deliver stable capacities of 165.2, 161.7, 159.5, 156.4, 151.9, 146.1, 136.2, 125.2, and 111.2 mA h g−1, respectively, as the current densities rise from 0.2C to 0.5C, 1C, 2C, 5C, 10C, 20C, 50C, and 100C. The reversible capacities of PNCsLFP almost return to the initial values of 163.9 mA h g−1 as the discharging rate changes back to 0.2C. As the rate rises, the difference in capacities among these three electrodes become more and more obvious, indicating the rate performance of PNCsLFP is much better than the other samples. PNCsLFP deliver 110.6 mA h g−1 even at 100C, compared with nanosized [email protected] and microsized LFP/C with only 59.7 and 30.5 mA h g−1 at such high rate. The ultrafine primary particles of pomegranate-like [email protected] subunits with a thin conductive carbon lead to the high power rate performance. The details that further prove the excellent rate performance are exhibited in the Supporting Information (Figures 3e and 3f and Supporting Information Figure S6). Figure 3d shows the cyclic performance of the PNCsLFP for 500 cycles at 1C. The reversible capacity of PNCsLFP remained at 157.6 mA h g−1 at 1C after 500 cycles. The corresponding average coulombic efficiency approached 99.5%, revealing the highly reversible Li+ insertion/extraction kinetics. And capacity retention of 98.5% is obtained for PNCsLFP after 500 cycles compared with the lower retentions of 93.8% and 90.5% for nanosized [email protected] and microsized LFP/C, demonstrating the improved cycle stability and reversibility of PNCsLFP. Figure 3 | (a) Charge/discharge curves of PNCsLFP, nanosized [email protected] and microsized LFP/C at 1C. (b) Typical charge/discharge curves of PNCsLFP from 0.2C up to 100C. (c) Rate performance of PNCsLFP from 0.2C up to 100C, and then back to 0.2C. (d) Cycling performance of PNCsLFP during 500 cycles at 1C. (e) Log(i)–log(v) curves for the anodic and cathodic scan. (f) Contribution of the pseudocapacitive fraction at 2 mV s−1. (g) Long cycling performance of the PNCsLFP, nanosized [email protected] and microsized LFP/C at 10C. (h) Long cycling performance of the PNCsLFP at 20C and 50C. Download figure Download PowerPoint In situ XRD To further understand the structural evolution of PNCsLFP during Li ion extraction and insertion, we performed in situ XRD at 0.1C between 2.0 and 4.2 V (Figure 4b). Corresponding intensity contour maps are shown in Figure 4a. Note that there is no new phase formation or phase transition during the whole charge/discharge process, demonstrating a complete solid-solution behavior in PNCsLFP. Normally for LFP, phase transition via two-phase kinetics occurs at low rates, whereas a solid-solution reaction occurs at relatively high rates.23 However, in our case, even at a low rate of 0.1C, the PNCsLFP still achieve solid-solution reaction (Figure 4c). The complete solid-solution reaction at all rates can greatly improve the utilization rate of LFP, increase the diffusion rate of lithium ions, and obtain high specific capacity and excellent rate performance. To investigate the free energy and voltage during the solid-solution reaction, canonical Monte Carlo simulations of the LFP cluster expansion have been performed by Ceder et al.20 Small simulation cells (2 × 3 × 3 unit cells) were used in nonequilibrium, but low-energy solid-solution states were captured. The overall shape that the solid-solution free energy exhibits is almost flat regardless of the Li content except near XLi ≈ 0 and XLi ≈ 1. And the Li states display short-range ordering in the solid-solution state. In summary, the Li+ is distributed equally in the 1D diffusion channels, and the solid-solution reaction defines a continuous lithiation path because the Li insertion reaction is a topotactic process, in which all 1D diffusion channels are active in either lithiation or delithiation. Hence, the Li+ diffusion along the 1D diffusion channels in solid-solution reaction is faster than bulk diffusion in the two-phase transition. And some concepts describe the expressions for the critical radius (r*) and critical nucleation barrier (ΔGr*), r* = 2γ·v/(|φ| − Δgs) and ΔGr* = 16π·γ3·v2/3(|φ| − Δe)2, where γ is the interfacial energy of LFP/FePO4,32v is the molar volume of LFP, φ is the overpotential, and Δgs is the coherency strain energy.33 For normal LFP, an overpotential of more than 50 mV must be applied, and ΔGr* is very large. The large energy required to form the critical nucleus makes nucleation very difficult. It is more meaningful to consider the overpotential φ by the solid-solution pathway from the following equation 20: Δ φ = − Δ μ Li = − ( ∂ Δ G ∂ x Li ) T (1)After being determined from first-principles calculations, the overpotential of the solid-solution mechanism needs only ∼30 mV. Compared with the two-phase transition, this single-phase transformation path can be available at very low overpotential, and the availability obviates the need for nucleation and growth.20 The free energy profile in the literature also explains why (de)lithiation of LFP in the solid-solution reaction is so easy. Once very few Li+ ions are inserted in a nanoparticle, lithiation will move on if the potential is ∼20 mV (the slope of the free-energy profile) below the equilibrium potential. Because almost no potential changes between XLi ≈ 0.05 and XLi ≈ 0.9, lithiation will proceed rapidly as long as a nanoparticle has Li content (>≈0.05). Therefore, the solid-solution reaction can be an effective means to improve the low-temperature performance of LFP. The advantages of the solid-solution mechanism are significant: (1) the facile transformation; (2) higher diffusion coefficient of lithium ions; (3) lithiation of the nanoparticles is more homogeneous than that in the two-phase process and thus requires less energy. Figure 4 | (a) Contour map of in situ XRD patterns for PNCsLFP. (b) In situ XRD patterns for PNCsLFP during the first cycle at 0.1C between 2.0 and 4.2 V. (c) Schematic illustrating the difference between solid-solution reaction and two-phase transition. Download figure Download PowerPoint Li+ diffusion calculation Electrochemical impedance spectroscopy (EIS) was carried out to calculate the DLi quantitatively. The Nyquist plots of the PNCsLFP, nanosized [email protected], and microsized LFP/C at room temperature at an open-circuit potential are shown in Supporting Information Figure S7a. The Li+ diffusion coefficient (D) exhibits a noticeable synergistic effect. The values of D are calculated using the following equation: D = R 2 T 2 / 2 A 2 n 2 F 2 C 2 σ 2 (2)where T is the absolute temperature, A is the contact area between the active material and electrolyte, R is the gas constant, n is the number of electron per molecule during the redox reaction (n = 1 for the redox of Fe3+/Fe2+), C is the concentration of lithium-ions mol (7.69 × 10−3 cm3), F is the Faraday constant, and σ is the Warburg factor according to the following equation: − Z im = k + σ ɷ − 1 / 2 (3)The relationship between −Zim (Ω) and ɷ − 1 / 2 (frequency, Hz) in the low frequency region is shown in Supporting Information Figure S7b. From Eq. (2), the Warburg factor, σ, can be obtained by linearly fitting the plot between −Zim and ɷ − 1 / 2 . The lithium-ion diffusion coefficient of PNCsLFP is calculated to be 7.8 × 10−13, higher than 1.5 × 10−13 for nanosized [email protected] and 4.9 × 10−14 for microsized LFP/C according to Eq. (1), suggesting that the ultrafine LFP subunits with carbon coating benefit the improvement of Li+ diffusion processes leading to high ionic conductivity. (The EIS of PNCsLFP under 0, −20, and −40 °C are shown in Supporting Information Figures S7c, S7e, and S8a.) All Nyquist plots are similar to the Nyquist plot under room temperature. The plot between −Zim and ɷ − 1 / 2 is shown in Supporting Information Figures S7d, S7f, and S8b, and the lithium ion diffusion coefficients of PNCsLFP under 0, −20, and −40 °C are calculated to be 7.76 × 10−13, 7.58 × 10−13, and 7.30 × 10−13 cm2 s−1, which is very close to the lithium ion diffusion coefficient under room temperature (Table 1), suggesting that the ultrafine [email protected] subunits also accelerate lithium ion diffusion under low temperature. The long cycling performance of PNCsLFP at 10C is shown in Figure 3g. The PNCsLFP retain a discharge capacity of 144.3 mA h g−1 after 1000 cycles at 10C, corresponding to 97.5% capacity retention. The retention of nanosized [email protected] and microsized LFP/C are 90.5% and 76.5%, respectively. Even at 20C and 50C, the PNCsLFP still exhibit high discharge capacities of 132.8 and 120.1 mA h g−1 after 1000 cycles, corresponding to 96.9% and 94.6% capacity retention (Figure 3h). The excellent cyclability of PNCsLFP is caused by the fast capacity response of the carbon coating, which buffered the impact to the ultrafine LFP subunits at high rate.34 Meanwhile, the ultrafine carbon supports function as a skeleton in PNCsLFP, which enhances the whole structural robustness. Table 1 | Lithium Ion Diffusion Coefficient of PNCsLFP, Nanosized [email protected] and Microsized LFP/C at Room Temperature, 0, −20, and −40 °C Materials Lithium Ion Diffusion Coefficient (25 °C, ×10−13) Lithium Ion Diffusion Coefficient (0 °C, ×10−13) Lithium Ion Diffusion Coefficient (−20 °C, ×10−13) Lithium Ion Diffusion Coefficient (−40 °C, ×10−13) PNCsLFP 7.8 7.76 7.58 7.3 Nanosized [email protected] 1.5 1.03 0.745 0.021 Microsized [email protected] 0.49 0.336 0.224 0.014 Low temperature performance The high Li+ diffusion coefficients at low temperatures encouraged us to investigate the low-temperature capability of PNCsLFP. Figures 5a and 5b display the charge/discharge curves of the PNCsLFP electrodes with satisfactory charge/discharge potential plateaus at both −20 and −40 °C. The PNCsLFP electrode can deliver high discharge capacities of 130.5 and 96.1 mA h g−1 at 0.1C under −20 and −40 °C. The capacities of the PNCsLFP electrode decrease slightly to 127.4 and 124.5 mA h g−1 at 0.2C and 0.5C under −20 °C, and 93.0 and 88.2 mA h g−1 at 0.2C and 0.5C under −40 °C with the increase of rate, respectively. Even at 10C and 20C, the PNCsLFP still deliver 89.1 and 79.2 mA h g−1 under −20 °C, and 53.2 and 40.1 mA h g−1 under −40 °C. The high capacity under low temperature can be ascribed to the ultrafine [email protected] subunits, which reduce the ion diffusion distance and enlarge the electrode/electrolyte interface area. Figures 5c and 5d show the rate performance of the PNCsLFP. It can be seen that there is only a little reduction in capacity from 0.1C to 20C even under ultralow temperatures of −20 and −40 °C. The low-temperature performances at high rates of as-prepared PNCsLFP are much better than previously reported values ( Supporting Information Table S1). Moreover, we also tested the long-term cycling stability of PNCsLFP for 300 cycles at 0.5C. In Figure 5e, the PNCsLFP electrode displays capacities of 122.1 and 86.8 mA h g−1 under −20 and −40 °C after 300 cycles, corresponding to 97.6% and 95.2% capacity retention, indicating the outstanding cycle reversibility and stability of PNCsLFP under low temperatures due to the ultrafine [email protected] subunits. Figure 5 | Typical charge/discharge curves of the PNCsLFP electrode at (a) −20 and (b) −40 °C. (c) Rate capability of the PNCsLFP electrode at −20 and −40 °C. (d) Nyquist plots of PNCsLFP electrode at (a) −20 and −40 °C. (e) Cycling performance of PNCsLFP at 0.5C under −20 and −40 °C. Download figure Download PowerPoint Electrochemical performance in full cell We further explored the practical application of PNCsLFP at room temperature and low temperatures through lithium-ion full batteries with PNCsLFP and graphite ( Supporting Information Figure S9) as cathode and anode materials, respectively (Figure 6a). Figure 6b exhibits the representative charge-discharge curves cycled at 0.1C between 2.0 and 3.7 V at room temperature. The PNCsLFP//graphite full cell with an average working voltage of 3.3 V delivers an ultrahigh capacity of 154.3 mA h g−1. The rate performance at room temperature shown in Figure 6c is obtained in a wide current density range from 0.1C to 20C. The delivered capacity still remains 121.3 mA h g−1 even at 20C (whole charge-discharge within only 30 s), and the capacity retention is up to 78.6% compared with the value at 0.1C. To further confirm the cycling capability, the electrode was tested for 500 cycles at 1C. The full cell displays 137.7 mA h g−1 after 500 cycles with coulombic efficiency around 100%, corresponding to a capacity retention of 94.3%, which suggests a superior cycling stability (Figure 6g). The ultrahigh capacity, excellent rate performance, and outstanding cycling capability of PNCsLFP half cell at low temperatures encouraged us to further test its performance in a full cell. As expected, the electrode exhibits a high and stable working voltage of ≈3.2 V at −20 °C (Figure 6e). The full cell exhibits a remarkable rate performance shown in Figure 6f which delivers 100.0, 96.7, 94.2, 91.0, 86.6, 78.7, and 67.8 mA h g−1 at 0.1C, 0.2C, 0.5C, 1C, 2C, 5C, and 10C. The