Electrochemical Analysis of the Carbon-Encapsulated Lithium Iron

This is an open access article published under an ACS AuthorChoice License, which permits copying and redistribution of the article or any adaptations for non-commercial purposes.

Article Cite This: ACS Omega 2018, 3, 6446−6455

Electrochemical Analysis of the Carbon-Encapsulated Lithium Iron Phosphate Nanochains and Their High-Temperature Conductivity Profiles K. P. Abhilash,†,‡,§ P. Christopher Selvin,∥ B. Nalini,⊥ Hui Xia,‡,§ Stefan Adams,# and M. V. Reddy*,†,#

Downloaded via 95.85.70.94 on July 6, 2018 at 07:59:56 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.

†

Department of Physics and #Department of Materials Science and Engineering, National University of Singapore, 117542, Singapore ‡ School of Materials Science and Engineering and §Herbert Gleiter Institute of Nanoscience, Nanjing University of Science and Technology, Nanjing 210094, China ∥ Department of Physics, Bharathiar University, Coimbatore 642046, Tamilnadu, India ⊥ Department of Physics, Avinashilingam University for Women, Coimbatore 642043, Tamilnadu, India S Supporting Information *

ABSTRACT: Carbon-encapsulated LiFePO4 (LFP) nanochains were prepared as a cathode material for lithium batteries by sol− gel method using citric acid as the carbon source. The prepared LFP/C material is characterized by structural, morphological, and electrochemical characterization. LFP/C shows an orthorhombic olivine structure with “Pnma” space group having an average particle size of 50 nm. The uniform distribution of LFP particles coated by the carbon matrix as a nanochain array has been analyzed by scanning electron microscopy and transmission electron microscopy analysis of the sample. The electrochemical performance of the LFP/ C nanochain has been analyzed using galvanostatic cycling, cyclic voltammetry, and impedance analysis of the assembled batteries. The sol−gel-derived LFP/C nanochain exhibits better capacity and electrochemical reversibility in line with the literature results. The high-temperature conductivity profile of the sample has been recorded from room temperature to 473 K using impedance analysis of the sample. The transport dynamics have been analyzed using the dielectric and modulus spectra of the sample. A maximum conductivity up to 6.74 × 10−4 S cm−1 has been obtained for the samples at higher temperature (448 K). The nucleation and growth at higher temperature act as factors to facilitate the intermediate phase existence in the LiFePO4 sample in which the phase change that occurs above 400 K gives irreversible electrochemical changes in the LFP/C samples.

■

particles4 and particle size reduction.5 The preparation of LFP and exploring its better electrochemical performance are quite tedious because the preparation conditions seriously affect the cycling performance of the same in a large extent. Here, citric acid-assisted sol−gel process has been used for the preparation of carbon-encapsulated LFP nanochain interlocks. Here, citric acid acts as a carbon source which can also prevent the oxidation of ferrous ions and offers a network chainlike array of LFP with better electronic conductivity.6 The present method shows a simple strategy for the effective carbon encapsulation to improve the capacity of the sample when compared with the other costly approaches such as the surface modification using graphene, nitrogen incorporation, and so forth. Even though LiFePO4 is an active field of current research, the dielectric and

INTRODUCTION

Lithium metal phosphates LiMPO4 (M = Fe, Mn, Co, and Ni) with olivine-type structure have attracted considerable attention as perspective cathode materials for lithium-ion batteries (LIBs).1,2 Among various alternative cathode materials, lithium iron phosphate (LFP), discovered by Goodenough in 1990s, is gaining significant attention because of its low cost, high discharge potential (very flat voltage curve around 3.4 V), large theoretical capacity (169 mA h/g), good thermal stability, excellent cycling performance, and high abundance with environmentally benign and safe nature.1−3 In LFP, the hexagonal close-packed lattice of oxygen has a two-dimensional channel network creating potential fast diffusion paths for lithium ions. The low electronic conductivity on the order of 10−6 to 10−8 S cm−1 and the slow lithium ion diffusion are the major obstacles preventing the application of an olivine cathode in a practical battery. An appreciable method to improve the conductivity is carbon coating around the © 2018 American Chemical Society

Received: March 20, 2018 Accepted: June 4, 2018 Published: June 15, 2018 6446

DOI: 10.1021/acsomega.8b00527 ACS Omega 2018, 3, 6446−6455

Article

ACS Omega modulus analyses of LiFePO4 have been less explored, especially at higher temperatures, which would give an indepth knowledge toward the mechanism of ionic transport and the relaxation mechanism of the material. It is a key factor to optimize the ionic conductivity and diffusion mechanism of LiFePO4 because they dictate the cell capacity and cycle life. Also, the cathode material should be stable in its structure and should give a better conductivity profile even at higher temperature for its application potential especially when it is used in all-solid-state assembly.7 In one of the previous attempt, Xu et al. reported an increase in ionic conductivity of LiFePO4 on high-temperature sintering using the Arrhenius plot.8 The study shows that the electronic structure of LFP mainly supports to the conductivity of the sample by first-principle calculations. The effect of annealing on the transport properties of LiFePO4 using a defect chemical model had been reported by Amin and Maier9 The annealed sample improves the conductivity of the sample around 4 orders of magnitude (∼3.1 × 10−6 S cm−1) when compared with the nonannealed (∼6 × 10−10 S cm−1) samples because of the transition from a defect regime to a region of ionic deficiency by the compensation of ionic disorder with the electronic defect. Rosaiah et al.10 have investigated the electrochemical properties of nanocrystalline LiFePO4 prepared by hydrothermal method from room temperature (RT) to 373 K in which the sample exhibits low ionic conductivity on the order of 10−7 S cm−1. In LFP material, the lithium insertion or extraction proceeds at 3.45 V versus lithium metal through a two-phase reaction between LiFePO4 and FePO4. Other than this, the existence of intermediate LixFePO4 solid-state solutions has been reported by Delacourt et al.11 The electrochemical behavior of this sample due to the predicted intermediate phase existence has scarcely been reported in the literature. The effect of these intermediate phases and the nature of lithium insertion and extraction at different temperatures are still challenging issues to be addressed. The present work aims to synthesize an LFP nanochain array by citric acid-assisted sol−gel method. The effective carbon coating using citric acid as the carbon source has been revealed in the present study. The major concern of the present approach is electrochemical characterization and the identification of hopping dynamics and relaxation processes of the sample at higher temperatures (from RT to 473 K). The role of temperature on the intermediate phase existence of LFP/C sample has been addressed in the present study.

Figure 1. Refined XRD pattern of the LFP/C sample.

parameters with a unit cell volume of 301 Å3 taking into account the electron correlation during the phase change which are in line with these values. Citric acid used during the preparation not only acts as a solvent for the formation of sol− gel but also acts as a carbon source for the formation of LFP/C interface which suppresses the growth of LFP crystals resulting in a well-arranged nanochain array.13 The absence of any other signals in the XRD pattern indicates that there are no unwanted impurities, such as Fe2O3 and Li3Fe2(PO4)3, or any other Ferelated compounds in the prepared LFP/C samples.14 No carbon-related peak was obtained, which shows that carbon is present in the amorphous phase.15,16 The scanning electron microscopy (SEM) micrograph of LFP/C (Figure 2A,B) shows the uniform formation of olivine nanoparticles attached end to end, indicating the formation of a nanochain array. The LFP/C particles have a spherical morphology dispersed in carbon matrix that inhibits the growth of the particles. The spherical aggregates formed from the network structure of uniformly carbon-coated nanospheres can increase the electronic conductivity of LFP/C and provide good electronic contact between the crystallites.17,18 The SEM micrograph shows the network of aggregated nanochains. The residual carbon imprints are not much visible in the SEM micrograph, which ensures the uniform coating. More details about the surface carbon distribution and the real texture of the particles have been portrayed from the transmission electron microscopy (TEM) micrograph of the LFP/C sample.19 Figure 2C−F shows the typical TEM micrograph of the LFP/C sample. The TEM micrograph gives a clear indication of the uniform carbon coating, preventing the aggregate growth of LFP nanospheres to form a nanochain interlock array. The nanospheres forming the chain embedded in the carbon matrix can be visible from the TEM micrographs. The particles with average size of 50 nm have been uniformly carbon-coated around the particles and also embedded in outer carbon layers, which are clearly visible from the TEM micrographs. The Raman spectra analysis and the energy-dispersive X-ray spectrometry (EDXS) composition analysis has been included as the electronic supplementary information (Supporting Information-Figures S1 and S2A,B) to further verify the structure and carbon content present in the sample. Temperature-Dependent Conductivity Profile of LFP/ C. Figure 3 shows the Z′−Z″ plot of LFP/C at RT and at different elevated temperatures with Ag|LiFePO4|Ag configuration. The obtained plots were fitted using Z-View software and the fitting parameters are listed in Table 1. In all figures, the solid line depicts the original data and the round symbol line represents the fitted data. The inset picture shows the most

■

RESULTS AND DISCUSSION Structural and Morphological Characterization. The X-ray diffraction (XRD) pattern of the sol−gel-derived LFP/C nanopowder is shown in Figure 1. LFP/C crystallizes to form an orthorhombic olivine structure in which the FeO6 octahedra form a layer in the “bc” plane. The adjacent layers are linked by PO4 units leaving some space for the one-dimensional array of Li atoms along the “a” axis.12 The XRD pattern of LFP/C is assigned as orthorhombic phase with “Pnma” space group using the Topas software, when indexed below a maximum deviation of 0.030° for all indexed lines, which is consistent with the PCPDF no. 81-1173. The high-intensity background signals originated because of the Fe content in the sample under Cu Kα radiation. The unit cell parameters are a = 10.120 Å, b = 6.07 Å, and c = 4.60 Å with a unit cell volume of 297 Å3. This value is slightly higher than the values reported in the literature. Zhou et al. predicted more suitable and self-consistent lattice 6447

DOI: 10.1021/acsomega.8b00527 ACS Omega 2018, 3, 6446−6455

Article

ACS Omega

Figure 2. Typical SEM (A,B) and TEM (C−F) micrographs of an LFP/C sample.

Figure 3. Nyquist plot of LFP/C at different temperatures [(a) rt, (b) 323 K, (c) 348 K, (d) 373 K, (e) 398 K, (f) 423 K, and (g) 448 K] and (h) the equivalent circuit model and related parameters used for fitting.

fitted equivalent circuits. The room-temperature (301 K) Nyquist plot (Figure 3a) shows distinct grain and grain boundary regions in the sample, which have been resolved by fitting the plot using a suitable equivalent model.20,21 In the equivalent circuit (Figure 3h), the block containing resistance (R) and the constant phase element (CPE) represents one semicircle. The grain and grain boundary effect has been

resolved from the capacitance value provided by the CPE obtained after fitting. The grain component usually bears a capacitance value in the range of 10−12 to 10−14 F and the grain boundary in the range of 10−11 to 10−10 F. From the value of the capacitance, the grain and grain boundaries have been differentiated. The conductivity of the samples (grain and grain boundary conductivities) has been calculated by 6448

DOI: 10.1021/acsomega.8b00527 ACS Omega 2018, 3, 6446−6455

Article

ACS Omega

cathode material for LIBs by Goodenough et al.,23 in 2001, Prosini et al. reported that the addition of fine particles of carbon black during the synthesis of LiFePO4 can greatly improve the electrochemical performance. The composite material so obtained is conductive and no additional carbon black needs to be added during the electrode preparation. The capacity seems to be increased with an increase in temperature.24 The pioneering work on LFP carried out by Armand et al. showed that LiFePO4 with carbon coating can achieve better capacity. The study explains the method of heating to deposit the conductive carbon coating on the LFP particles which improves the rate and capacity of the material. After this work, the LFP material as a green cathode material for LIBs became an important candidate for the scientific world.25 After that, many of the researchers proved the importance of carbon coating on LFP materials. The conductivity of the noncarboncoated sample is only at the range of 10−7 to 10−9 S cm−1.26 The present sample has been obtained with conductivity in the range of 10−5 to 10−4 S cm−1, which clearly shows the effect of uniform carbon encapsulation on the LFP particles. Even though an appreciable increase in the ionic conductivity occurs with the rise in annealing temperature, an inductive loop appeared in the low-frequency region, which may be related to some phase transformation that occurs at higher temperatures.27 The inductive loop at the lower-frequency end occurs because of the grain boundary nucleation and growth at higher temperatures.28 The existence of LiFePO4 intermediates has been thoroughly investigated by different authors.29−32 Delacourt et al. have reported the temperature-assisted XRD pattern of LiFePO4 which clearly suggests a phase change in the range of temperatures between 423 and 473 K.11 This phase change gives no signature in its differential scanning calorimetry curve but originates from the coexistence of different stable LixFePO4 phases as suggested by Zhou et al.29,30 This can be correlated with the higher-temperature inductive loop formation that occurs in the LFP sample. This inductive loop suggests the possibility of accumulation of the charge carriers near the contact electrode, which may lead to the formation of some resistive interface when coupled with the electrolytes during battery formation. The present study explores that the nucleation and growth at higher temperature act as a factor which facilitate the intermediate phase existence in LiFePO4 as suggested by the first-principle calculations of Zhou et al., in which the phase change that occurs above 400 K gives irreversible electrochemical changes in the LFP/C samples. After 448 K, the conductivity plot does not show the regular

Table 1. Conductivity Data of LFP/C Extracted from the Nyquist Plots temperature (K)

grain resistance (Ω)

grain boundary resistance (Ω)

303 323 348 373 398 423 448

940 970 1227 845 464 227 79

2550 1900

total conductivity (S cm−1) 5.02 8.46 1.04 1.52 2.76 5.02 6.74

× × × × × × ×

10−5 10−5 10−4 10−4 10−4 10−4 10−4

σg = d /R gS

(1)

σgb = d /R gbS

(2)

where, d is thickness of the pellet, S is the area of the pellet, σg is the grain conductivity, σgb is the grain boundary conductivity, Rg is the grain resistance, and Rgb is the grain boundary resistance. Here, uniform conditions are followed for the sample to measure the conductivity and the high-temperature features were measured from the same sample. The conductivity of the pellet has been measured by considering it as a disc and the thickness of the pellet as the length of the disc. The bulk conductivity of the sample is 1.36 × 10−4 S cm−1 at RT. The total conductivity of the sample is found to be 5.02 × 10−5 S cm−1 which is an order of magnitude lesser when compared with the bulk conductivity. At 323 K (Figure 3b), the distinct grain and grain boundary regions are not visible and the equivalent circuit differentiates the grain and grain boundary resistance. The conductivity at the grain region is decreased at 323 K, which may be assumed due to the increase in unit cell volume at elevated temperatures. However, the total conductivity (8.46 × 10−5 S cm−1) of the sample is found to be improved. At higher temperatures, the grain and grain boundary regions are not differentiated, which makes the difference in the equivalent circuit model of the samples at elevated temperatures. As temperature increases, the intercept point on the real axis shifts toward the origin, indicating the decrease in the total resistance of the material. The resistance decreases up to 79 Ω, increasing the conductivity up to 6.74 × 10−4 S cm−1 at 448 K (175 °C). The increase in conductivity at high temperature may be attributed to the hopping mechanism of thermally activated charge carriers.22 After the identification of LFP as a better

Figure 4. (a,b) Arrhenius plot (a) and conductivity spectra (b) of the LFP/C sample. 6449

DOI: 10.1021/acsomega.8b00527 ACS Omega 2018, 3, 6446−6455

Article

ACS Omega behavior, which clearly establishes that the intermediate phase transitions has a key role in the anomalous high-temperature electrochemical behavior of the olivine-type LiFePO4 sample. Arrhenius Plot. The Arrhenius plot (Figure 4a) [represented as: log(σ) at different temperatures vs 1000/T] exhibits two portions: low-temperature region with low activation energy (0.44 eV) and high-temperature region with high activation energy (1.57 eV). The low-temperature linear portion suggests that the sample adheres to the Arrhenius behavior up to 373 K. The exponential growth above 373 K reveals the requirement of higher activation for the hightemperature electrochemical processes. The activation energy of the sample has been calculated using the Arrhenius equation ⎛ −E ⎞ σ σ = 0 exp⎜ a ⎟ ⎝ kT ⎠ T

Table 2. dc Conductivity (σdc) Calculated from Conductivity Spectra temperature (K) 303 323 348 373 398 423 448

⎛ −E ⎞ (N (ze)2 d 2) ϑ0 exp⎜ a ⎟f ⎝ kT ⎠ kT

3.5990 4.2305 7.1866 1.5077 2.0832 3.6274 7.4077

× × × × × × ×

10−6 10−6 10−6 10−5 10−5 10−5 10−5

frequency and gets saturated at high frequencies. The fractional increase that occurs in the sample results in an increase in the value of dielectric constant. The accumulation of charge carriers near the electrode results in an initial hike in the dielectric spectra at the lower-frequency region. The increase in dielectric constant results from the increase in the charge carrier concentration which in turn increases the conductivity of the sample. At higher frequencies, a decrease in the value of the dielectric constant occurs due to the periodic reversal of the applied field which is so high that there is no further ion diffusion in the field direction. Hence, the polarization-proliferated charge accumulation decreases, leading to the decrease in the value of dielectric constant at the higher frequency region. At RT, the polarization effect is more when compared with the same at elevated temperature, which may be the reason for the increase in the dielectric value of 303 and 323 K when compared with other high temperatures. Also, the dielectric constant increases with temperature as the orientation is facilitated due to the increase in temperature, which in turn increases the effective dipole. The dielectric loss spectra are shown in Figure 5b. Dielectric loss is the energy dissipated and includes the comprehensive contribution from ionic transport and the polarization of charges or a dipole. The increase in the dielectric loss at lower frequencies arises due to free charge motion that occurs within the material. These values, which arise from the free charge buildup in the interface between the material and electrode, do not contribute to the bulk dielectric processes but lead to the conductivity relaxation process. The relationship between the conductivity σ and dielectric loss factor ε″ can be given by σ ε″ = (5) ωε′ σ and ε″ are strongly temperature-dependent and increase with increase in temperature. The conductivity loss plots generally divulge into two prominent relaxation processes: the lower-frequency αrelaxation peak and high-frequency β-relaxation peak. The αrelaxation peak pronounced at higher temperature may be caused by the movement of main-chain dipole segments and the β-relaxation peak at higher frequencies may be caused by the side group dipoles and the nearest part of the backbone.38,39 In the present study, the absence of multiple dielectric loss peaks suggests the occurrence of the single relaxation process. The increase in ε′ at the lower-frequency window as temperature increases is caused by the polarizability that arises due to multiple components such as electronic, ionic, and interfacial orientation of ions. Electrical Modulus Analysis. The complex electric modulus spectrum represents the measure of the distribution

(3)

where σ0 is the pre-exponential factor, T is the absolute temperature, Ea is the activation energy of conduction, and k is the Boltzman constant. According to the Nernst−Einstein relationship, the ionic conductivity σ=

σdc (S cm−1)

(4)

where, ϑ0′ is the attempt frequency, Ze is the charge of a mobile ion (in the case of Li+, Z = 1), “d” is the hopping distance of the ion, N is the number density of mobile ions, and “f ” is the correlation factor (considered as 1 in most cases). Comparing eqs 3 and 4, the pre-exponential factor is connected with the hopping distance, number density, and attempt frequency of mobile ions,33 which in turn are influenced by the annealing process. Hence, the increase of pre-exponential factor should be ascribed to the increase of the ionic hopping distance (d), variations in the number density, and the attempt frequency with the rise of annealing temperature. The sharp increase in activation energy after 423 K is due to the inductive loop formation as verified from the Nyquist plots. Conductivity Spectra. The nature of conduction and the mechanism of hopping of the charge carriers/ions at different temperatures have been divulged by the conductivity spectra. The conductance spectra at different temperatures have been depicted in Figure 4b. The conductivity spectra consist of two regions: intermediate plateau and high-temperature dispersion regions.34,35 In the present sample, the frequency-independent plateau region is more prominent. The width of the plateau region increases with the increase in temperature. The ionic relaxation process initiates at a frequency called characteristic frequency (ωs), after which the transition of long-range to short-range hopping mechanism occurs. With an increase in temperature, the characteristic frequency ωs also increases with a shift to the higher-frequency region. The mechanism of conductivity relaxation is in line with the Jonscher’s power law.36 The increase in the width of the plateau region with the increase in temperature points that the long-range hopping mechanism inputs more share to the total conductivity of the sample. The dc conductivity (σdc) (Table 2) has been calculated by extrapolating the conductivity spectra to the Y axis. The hopping rate (ωp = 2σdc) also increases with increase in temperature of the sample. Dielectric Analysis. The real part of dielectric permittivity (ε′)37 with respect to the frequency has been depicted in Figure 5a. Generally, the dielectric constant decreases with increase in 6450

DOI: 10.1021/acsomega.8b00527 ACS Omega 2018, 3, 6446−6455

Article

ACS Omega

Figure 5. Real (a) and imaginary (b) part of dielectric spectra for the LFP/C sample at different temperatures.

Figure 6. Real (a) and imaginary (b) part of frequency-dependant modulus spectra of the LFP/C sample at different temperatures (the inset in Figure 6b shows the magnified image to indicate the shift in the peak position).

of ion energies or configurations in the structure describing the electrical relaxation and macroscopic properties of the materials.19,40,41 Figure 6a,b represents the frequency dependence of M′(ω) and M″(ω) at different temperatures. In the M′ versus log(ω) plot, M′ at the lower-frequency region reaches zero, suggesting the negligible contribution from electrode polarization.42−44 The higher value of M′ suggests the relaxation process that occurs at the higher-frequency region. The pattern at the higher-frequency side did not show a uniform order of increase or decrease with temperature, pointing that the relaxation is not fully temperature-dependant but the local dispersion-based relaxation also contributes to the relaxation process. The effect of temperature in the relaxation process will be obtained from the analysis of the imaginary part of modulus value at different frequency windows. The M″ versus log(ω) plot is shown in Figure 6b and the inset shows the magnified image of the lower-frequency region of the sample at different temperatures. The M″ versus log(ω) plot gives the peak relaxation frequency (f max) from which the relaxation time (τ) can be intended. The broad nature of the peak suggests that the relaxation time is not single-valued but disseminated continuously around the mean value.45 The peak relaxation frequency (f max) and the relaxation time (τ) are tabulated in Table 3. The increase in temperature results in a shift in the plot toward the higher-frequency region, but at higher temperatures (423 and 448 K) where the inductive effect is more prominent (as shown from the Nyquist plot), does not show any shift in the peak position. This can be correlated with the occurrence of intermediate phase which explains the deviation of electrochemical behaviors. The M″/Mmax″ versus log(ω) (Figure 7) plot furnishes the effect of temperature in the relaxation phenomena. The plot gives an overlapping peak

Table 3. Peak Relaxation Frequency (f max) and the Relaxation Time (τ) Calculated from the M″ vs log(ω) Plot f max

temperature 303 323 348 373 398 423 448

1 2 2 2 2 2 1

986 197 258 508 545 023 998

380 222 961 244 520 656 029

τ 8.01638 7.24714 7.04907 6.34849 6.25553 7.86871 7.96964

× × × × × × ×

10−8 10−8 10−8 10−8 10−8 10−8 10−8

Figure 7. M″/Mmax″ vs log(ω/ωmax) plot of the LFP/C sample at different temperatures.

at higher-frequency windows suggesting that the conductivity relaxation mechanism is more or less temperature independent, suggesting the prominent dispersion-related conductivity 6451

DOI: 10.1021/acsomega.8b00527 ACS Omega 2018, 3, 6446−6455

Article

ACS Omega

Figure 8. (a) Specific capacity vs voltage plot of an LFP/C nanochain for selected cycles at a current density of 30 mA g−1 (inset shows the magnified image of the end part of the discharge cycles). (b) Capacity vs cycle number plot of an LFP/C nanochain at different current densities [30, 60 mA g−1, 170 mA g−1 (1 C), and 540 mA g−1 (3 C)]. (c) CV plot of an LFP/C nanochain at different scan speeds. (d) Peak current vs square root of scan rate for an LFP/C nanochain.

performance of the assembled cell, the same cell after sweeping 50 cycles has been cycled at different current rates. The capacity versus cycle plots at different current densities of 30 and 60 mA g−1, 1 C (170 mA g−1), and 3 C (510 mA g−1) have been portrayed in Figure 8b. The figure shows a stable charge discharge profile with minimum capacity fading.48−50 The sample exhibits a capacity fading less than 1% between 5th and 50th cycles in all current rates, clearly establishing the better stability of the sample due to the uniform carbon encapsulation around the nano-LFP/C by simple sol−gel route. The stable charge−discharge cycles obtained even at higher current rate result from the uniform carbon-encapsulated nanochain morphology of the sample. Cyclic Voltammetry Studies. Cyclic voltammetry (CV) studies have been carried out at different scan rates for better understanding toward the intercalation−deintercalation of Li ions in the sol−gel-derived LFP/C nanochains. Figure 8c shows the CV plots of the sample at different scan rates (0.1, 1, 5, 10, and 20 mV/s). The inset picture shows the individual plot at 0.1 mV/s. The CV plot at 0.1 mV/s scan rate shows the oxidation peak at 3.51 V and the corresponding reduction peak at 3.41 V. At higher scan speed, the oxidation potential shifts to the higher value, whereas the reduction potential shifts to the lower value with an increase in peak current, which is in accordance with the Randles−Sevcik relation.49 In a slow voltage scan rate, the diffusion layer will grow much farther from the electrode in comparison to a fast scan which reduces the current value. The CV parameters are tabulated in Table 4. For a perfect reversible system, the anodic and cathodic peak currents should have the same magnitude.51 The lower value for the present LFP (i.e., ic/ia not equal to 1) may be due to the fact that the Li ion undergoes a two-phase process during charge and discharge in which Li ions are most likely transported through regions with FePO4 or LiFePO4.23,52

relaxation at the higher-frequency window. The dispersive conductivity can be explained with respect to the predicted diffusion-controlled relaxation model.46 The M″/Mmax″ versus log(ω) plot shows that the broadness of the curve is independent of the temperature.47 Rosaiah et al.10 studied the electrochemical properties of LFP materials in the frequency range 1 Hz to 1 MHz in which the imaginary part of the modulus spectra exhibits multiple energy loss at higher temperatures. However, in the present study, the LFP/C sample prepared by citric acid-assisted sol−gel method exhibits single relaxation at the higher-frequency window. The almost constant full width half-maximum value (“β”) would also imply that the relaxation process is temperature independent. The frequency at which the intersection occurs with the real axis, the frequency at which the relaxation effects begin, and the peak frequency of M″ plot shift toward the higher-frequency side with an increase in temperature. The M″/Mmax″ versus log(ω) plot at 423 and 448 K does not exhibit an overlapping peak at a high-frequency window, which also substantiates the intermediate phase change-related accumulation of charge carriers. Electrochemical Characterization of the Sample. Galvanostatic Studies. Figure 8a shows the galvanostatic cycling results of the sol−gel-derived LFP/C nanochain. The sample has been cycled within the potential range of 2.5−4.5 V with a current density of 30 mA g−1 after the battery fabrication. Charge discharge profiles of the selected cycles (at 30 mA g−1) have been included for clarity. The inset picture shows the magnified image of the end portion of the discharge profile. The first cycle sweeps a maximum charge capacity of 161 mA h/g, which is followed by the discharge capacity of 154 mA h/g at 30 mA g−1. After the initial cycles, the stable capacity around 144 mA h/g has been retained by the sample in the proceeding cycles. The LFP/C nanochain shows a discharge capacity of 143 mA h/g after 50 cycles. To elucidate the stable 6452

DOI: 10.1021/acsomega.8b00527 ACS Omega 2018, 3, 6446−6455

ACS Omega

■

Table 4. CV Parameters of the LFP/C Nanochain

Vo

0.1 1.0 5.0 10.0 20.0

3.51 3.59 3.71 3.79 3.93

−VR

potential difference (V)

Io

IR

peak current ratio

3.31 3.26 3.14 3.05 2.93

0.20 0.33 0.57 0.74 1.00

0.0012 0.0037 0.0079 0.0107 0.0139

0.0009 0.0029 0.0063 0.0088 0.0123

0.75 0.78 0.79 0.82 0.88

CONCLUSIONS

The sol−gel-derived LFP/C nanochains show orthorhombic structure with an average particle size of 50 nm. The surface carbon-dispersed spherical particles interconnected as a nanochain array have been established by SEM and TEM characterization of the sample. The high-temperature conductivity analysis of the LFP/C nanochain array shows an increase in the value of conductivity with an increase in temperature. The nucleation and growth at higher temperature act as a factor which facilitate the intermediate phase existence in the LiFePO4 sample in which the phase change that occurs above 400 K gives irreversible electrochemical changes in the LFP/C samples. The difference in the activation energy calculated from the Arrhenius plot reveals that the energy barrier for the charge carriers for an activated process is higher at higher temperatures above 373 K. The conductivity spectra exhibit a long-range plateau region, suggesting that a long-range ionic movement contributes to the major share in conductivity. The dielectric constant increases with increase in temperature because of the effective orientation of the dipoles at elevated temperatures. The absence of multiple dielectric loss peaks suggests the occurrence of single relaxation process in the LFP/ C sample. The conductivity relaxation mechanism is more or less temperature independent, suggesting the prominent dispersion-related conductivity relaxation at the higherfrequency window. The frequency at which the intersection occurs with real axis, the frequency at which the relaxation effects begin, and the peak frequency of M″ plot shift toward the higher-frequency side with an increase in temperature. The M″/Mmax″ versus log(ω) plot at 423 and 448 K does not exhibit an overlapping peak at the high-frequency window, which also substantiates the intermediate phase change-related accumulation of charge carriers at higher temperatures. The better electrochemical performance of the sample has been derived from the uniform carbon encapsulation around the nanoparticles to form a nanochain array by the facile sol−gel method.

peak potential (V) scan rate (mV/s)

Article

This difference in electrochemical environments10 for Li movement may account for the difference in the magnitude of the peak current. The intercalation−deintercalation peak current is linearly related to the square root of scan rate (Figure 8b), establishing the reversible nature of the prepared electrodes. The CV plots and the cycling results verify the better electrochemical property of the sol−gel-derived carbonencapsulated LFP nanochains, which is in line with different literature reports.48,49,53,54 Impedance Analysis. Figure 9 shows the impedance plot of the assembled cell before and after cycling. The impedance

■

EXPERIMENTAL SECTION Preparation. The LFP/C cathode material has prepared by modified sol−gel method. Stoichiometric amounts of ferrous oxalate dihydrate (AR-99%, Aldrich) (9 g) and lithium nitrate (AR-99.5%, Aldrich) (3.45 g) were dissolved in 1 M nitric acid solution (30 mL) (AR-99.5%, Himedia), into which 10 mL of citric acid solution (AR-99.5%, Himedia) (66 wt %) was added dropwise with continuous stirring. A saturated solution of ammonium dihydrogen phosphate (AR-99.5%, Aldrich) (5 mL, 53 wt %) was then added to the mixture. The mixture was heated gently with continuous stirring for 2 h. The resulting gel precursor was dried in a tubular furnace for 5 days at 100 °C. The 5 days have been used for the complete removal of NH3 from the intermediate polymer gel matrix produced after the gel formation. This step has been performed at low temperature to reduce the particle size to a larger extent by the slow removal of NH3 from the precursor, which is necessary to maintain the preferred morphology of the sample. Then, the precursor was further calcined at 700 °C in the argon atmosphere for 2 h. The final product was brownish-black in color. The calcinated powder was dried and pelletized using the pelletizer by applying a uniform pressure of 16 kN/m2. Characterization. The prepared LFP/C nanopowder is structurally characterized by the powder XRD technique (XRD-

Figure 9. Nyquist plot of an LFP/C nanochain for the fresh cell (a) and after 50 cycles (b).

data of the full cell have been fitted using Z-View software. The circle line shows the original data and solid line shows the most fitted curve. The fresh cell impedance plot exhibits a single semicircle with combined surface film and charge-transfer resistance. The fresh cell exhibits an electrolyte dissociation resistance of 2.45 Ω, which is found to be vanished in the impedance plot recorded after 50 cycles. The fresh cell exhibits the combined surface film and charge-transfer resistance of 39.93 Ω. The impedance plot recorded after cycling shows two semicircles, which is evident from the equivalent circuit model used for fitting (inset Figure 9). After cycling, the surface film resistance has been resolved from the charge-transfer component.48,50,55 The sample exhibits the surface film resistance of 3.78 Ω and a charge-transfer resistance of 13.42 Ω. The lower impedance value substantiates the better performance of carbon-encapsulated nanochains prepared by sol−gel method. 6453

DOI: 10.1021/acsomega.8b00527 ACS Omega 2018, 3, 6446−6455

Article

ACS Omega

change of lithium iron phosphate during carbon coating. Nat. Commun. 2014, 5, 3415. (3) Molenda, J. Electronic Structure in Relation to the Reactivity of Nonstoichiometric AxMX2 Layered Compounds. Phys. Status Solidi B 1991, 165, 419−433. (4) Zaghib, K.; Mauger, A.; Gendron, F.; Julien, C. M. Magnetic studies of phospho-olivine electrodes in relation with their electrochemical performance in Li-ion batteries. Solid State Ionics 2008, 179, 16−23. (5) Yamada, A.; Chung, S. C.; Hinokuma, K. Optimized LiFePO[sub 4] for Lithium Battery Cathodes. J. Electrochem. Soc. 2001, 148, A224− A229. (6) Kope¢, M.; Yamada, A.; Kobayashi, G.; Nishimura, S. Structural and magnetic properties of Lix(MnyFe1−y)PO4electrode materials for Li-ion batteries. J. Power Sources 2009, 189, 1154−1163. (7) Miara, L.; Windmüller, A.; Tsai, C.-L.; Richards, W. D.; Ma, Q.; Uhlenbruck, S.; Guillon, O.; Ceder, G. About the compatibility between high voltage spinel cathode materials and solid oxide electrolytes as a function of temperature. ACS Appl. Mater. Interfaces 2016, 8, 26842−26850. (8) Xu, Y.-N.; Chung, S.-Y.; Bloking, J. T.; Chiang, Y.-M.; Ching, W. Y. Electronic Structure and Electrical Conductivity of Undoped LiFePO4. Electrochem. Solid-State Lett. 2004, 7, A131−A134. (9) Amin, R.; Maier, J. Effect of annealing on transport properties of LiFePO4: Towards a defect chemical model. Solid State Ionics 2008, 178, 1831−1836. (10) Rosaiah, P.; Jeevan Kumar, P.; Jayanth Babu, K.; Hussain, O. M. Electrical and electrochemical properties of nanocrystalline LiFePO4 cathode. Appl. Phys. A 2013, 113, 603−611. (11) Delacourt, C.; Poizot, P.; Tarascon, J.-M.; Masquelier, C. The existence of a temperature-driven solid solution in Li x FePO4 for 0 ≤ x ≤ 1. Nat. Mater. 2005, 4, 254−260. (12) Naik, A.; Zhou, J.; Gao, C.; Wang, L. Microwave Assisted Solid State Synthesis of LiFePO4/C Using Two Different Carbon Sources. Int. J. Electrochem. Sci. 2014, 9, 6124−6133. (13) Shiraishi, K.; Dokko, K.; Kanamura, K. Formation of impurities on phospho-olivine LiFePO4 during hydrothermal synthesis. J. Power Sources 2005, 146, 555−558. (14) Chang, Z.-R.; Lv, H.-J.; Tang, H.-W.; Li, H.-J.; Yuan, X.-Z.; Wang, H. Synthesis and characterization of high-density LiFePO4/C composites as cathode materials for lithium-ion batteries. Electrochim. Acta 2009, 54, 4595−4599. (15) Cui, Y.; Zhao, X.; Guo, R. Enhanced electrochemical properties of LiFePO4 cathode material by CuO and carbon co-coating. J. Alloys Compd. 2010, 490, 236−240. (16) López-Romero, S.; Chávez-Ramírez, J. Synthesis of TiC thin films by CVD from toluene and titanium tetrachloride with nickel as catalyst. Rev. Mater. 2007, 12, 487−493. (17) Zhang, Q.; Wang, S.; Zhou, Z.; Ma, G.; Jiang, W.; Guo, X.; Zhao, S. Structural and electrochemical properties of Nd-doped LiFePO4/C prepared without using inert gas. Solid State Ionics 2011, 191, 40−44. (18) Hsu, K.-F.; Tsay, S.-Y.; Hwang, B.-J. Synthesis and characterization of nano-sized LiFePO4cathode materials prepared by a citric acid-based sol-gel route. J. Mater. Chem. 2004, 14, 2690−2695. (19) Cao, Y. L.; Yu, L. H.; Li, T.; Ai, X. P.; Yang, H. X. Synthesis and electrochemical characterization of carbon-coated nanocrystalline LiFePO4 prepared by polyacrylates-pyrolysis route. J. Power Sources 2007, 172, 913−918. (20) Abhilash, K. P.; Sivaraj, P.; Selvin, P. C.; Nalini, B.; Somasundaram, K. Investigation on spin coated LLTO thin film nano-electrolytes for rechargeable lithium ion batteries. Ceram. Int. 2015, 41, 13823−13829. (21) Abhilash, K. P.; Selvin, P. C.; Nalini, B.; Nithyadharseni, P.; Pillai, B. C. Investigations on pure and Ag doped lithium lanthanum titanate (LLTO) nanocrystalline ceramic electrolytes for rechargeable lithium-ion batteries. Ceram. Int. 2013, 39, 947−952. (22) Rhouma, F. I. H.; Dhahri, A.; Dhahri, J.; Valente, M. A. Dielectric, modulus and impedance analysis of lead-free ceramics

Shimazdu-6000) using a Cu Kα radiation source. The morphological characterization of the samples has been carried out using SEM (JEOL-JSM 6390) and TEM (JEOL JEM-3010; 300 kV) analysis. The electrical characterization of the pellets has been carried out by a complex impedance spectrometer (HIOKI LCR 3532-50; frequency range: 2 Hz to 5 MHz) using the silver paste as the contact electrode on either side with Ag| LiFePO4|Ag configuration from RT to 175 °C to analyze the high-temperature profile of the sample. Electrochemical studies have been carried out with a 2016 type of coin cell assembly. The electrodes were fabricated with the active material LFP/C, binder (Kynar 2801) and Super P conductive carbon black in the weight ratio of 70:15:15 using N-methylpyrrolidone as a solvent to the binder. Etched Al-foil (20 mm) has been used as the current collector. Lithium metal foil has been used as the negative electrode (anode) and 1 M LiPF6 in ethylene carbonate + dimethyl carbonate (1:1 volume ratio) (Merck, Selectipur LP40) has been used as the electrolyte. Charge−discharge cycling was carried out at an ambient temperature using a Bitrode battery tester (model, MCV16-0.5, USA). CV studies were carried out at different scan rates using the Solatron, 1470 battery test unit. Impedance spectroscopy was carried out with a Solartron impedance/gainphase analyzer (model SI 1255) coupled with a potentiostat (SI 1268). The frequency was varied from 0.18 MHz to 3 mHz with an alternating current signal amplitude of 10 mV.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsomega.8b00527. Raman spectra of LFP/C, EDXS spectra of the LFP/C, and atomic percentage of the elements present in LFP/C (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected], [email protected] (M.V. Reddy). ORCID

Hui Xia: 0000-0002-2517-2410 Stefan Adams: 0000-0003-0710-135X M. V. Reddy: 0000-0002-6979-5345 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS K.P.A. thanks NJUST, Nanjing, China for providing the Post Doctoral research fellowship to carry out the research work. M.V.R. and S.A. thank the National Research Foundation (NRF), Prime Minister’s Office, Singapore for support under its Competitive Research Programme (CRP Award NRF-CRP 10-2012-6).

■

REFERENCES

(1) Chang, C.-C.; Lee, K.-Y.; Lee, H.-Y.; Su, Y.-H.; Her, L.-J. Trimethyl borate andtriphenyl borate as electrolyte additives for LiFePO4 cathode with enhanced high temperature performance. J. Power Sources 2012, 217, 524−529. (2) Wang, J.; Yang, J.; Tang, Y.; Liu, J.; Zhang, Y.; Liang, G.; Gauthier, M.; Chen-Wiegart, Y.-C. K. C.; Norouzi Banis, M.; Li, X.; Li, R.; Wang, J.; Sham, T. K.; Sun, X. Size-dependent surface phase 6454

DOI: 10.1021/acsomega.8b00527 ACS Omega 2018, 3, 6446−6455

Article

ACS Omega Ba0.8La0.133Ti1−x Sn x O3 (x=0.15 and 0.2). Appl. Phys. A 2012, 108, 593−600. (23) Padhi, A. K.; Nanjundaswamy, K. S.; Goodenough, J. B. Phospho-olivines as Positive-Electrode Materials for Rechargeable Lithium Batteries. J. Electrochem. Soc. 1997, 144, 1188−1194. (24) Prosini, P. P.; Zane, D.; Pasquali, M. Improved electrochemical performance of a LiFePO4-based composite cathode. Electrochim. Acta 2001, 46, 3517−3523. (25) Ravet, N.; Chouinard, Y.; Magnan, J. F.; Besner, S.; Gauthier, M.; Armand, M. Electroactivity of natural and synthetic triphylite. J. Power Sources 2001, 97−98, 503−507. (26) Yuan, L.-X.; Wang, Z.-H.; Zhang, W.-X.; Hu, X.-L.; Chen, J.-T.; Huang, Y.-H.; Goodenough, J. B. Development and challenges of LiFePO4cathode material for lithium-ion batteries. Energy Environ. Sci. 2011, 4, 269−284. (27) Hong, J.; Wang, C.; Kasavajjula, U. Kinetic behavior of LiFeMgPO4 cathode material for Li-ion batteries. J. Power Sources 2006, 162, 1289−1296. (28) Millet, P.; Dantzer, P. Impedance of metal hydride electrodes. Rate of phase transformation limited by nucleation and growth mechanisms. Electrochem. Commun. 1999, 1, 163−166. (29) Zhou, F.; Marianetti, C. A.; Cococcioni, M.; Morgan, D.; Ceder, G. Phase separation in LixFePO4 induced by correlation effects. Phys. Rev. B: Condens. Matter Mater. Phys. 2004, 69, 201101. (30) Zhou, F.; Cococcioni, M.; Marianetti, C. A.; Morgan, D.; Ceder, G. First-principles prediction of redox potentials in transition-metal compounds with LDA+U. Phys. Rev. B: Condens. Matter Mater. Phys. 2004, 70, 235121. (31) Yamada, A.; Koizumi, H.; Sonoyama, N.; Kanno, R. Phase Change in LixFePO4. Electrochem. Solid-State Lett. 2005, 8, A409− A413. (32) Morgan, D.; Van der Ven, A.; Ceder, G. Li Conductivity in Li[sub x]MPO[sub 4] (M = Mn, Fe, Co, Ni) Olivine Materials. Electrochem. Solid-State Lett. 2004, 7, A30−A32. (33) Xiong, Y.; Tao, H.; Zhao, J.; Cheng, H.; Zhao, X. Effects of annealing temperature on structure and opt-electric properties of ionconducting LLTO thin films prepared by RF magnetron sputtering. J. Alloys Compd. 2011, 509, 1910−1914. (34) Joncher, A. K. Physics of Thin Films, 2nd ed.; Academic Press: New York, 1980; pp 1−213. (35) Ingram, M. D. Ionic conductivity in glasses. Phys. Chem. Glasses: Eur. J. Glass Sci. Technol., Part B 1987, 28, 215−234. (36) Jonscher, A. K. Universal Relaxation Law, 1st ed.; Chelsea Dielectrics Press: London, 1996; pp 1−415. (37) Selvasekarapandian, S.; Vijayakumar, M. The ac impedance spectroscopy studies on LiDyO2. Mater. Chem. Phys. 2003, 80, 29−33. (38) Schmidt-Rohr, K.; Kulik, A. S.; Beckham, H. W.; Ohlemacher, A.; Pawelzik, U.; Boeffel, C.; Spiess, H. W. Molecular Nature of the.beta. Relaxation in Poly(methyl methacrylate) Investigated by Multidimensional NMR. Macromolecules 1994, 27, 4733−4745. (39) Cassu, S. N.; Felisberti, M. I. Poly(vinyl alcohol) and poly(vinylpyrrolidone) blends: 2. Study of relaxations by dynamic mechanical analysis. Polym. J. 1999, 40, 4845−4851. (40) Macedo, P. B.; Moynihan, C. T.; Bose, R. The long time aspects of this correlation function, which are obtainable by bridge techniques at temperatures approaching the glass transition. Phys. Chem. Glasses: Eur. J. Glass Sci. Technol., Part B 1972, 13, 171. (41) Provenzano, V.; Boesch, L. P.; Volterra, V.; Moynihan, C. T.; Macedo, P. B. Electrical Relaxation in Na2O.3SiO2 Glass. J. Am. Ceram. Soc. 1972, 55, 492−496. (42) Nobre, M. A. L.; Lanfredi, S. Dielectric loss and Phase transition of sodium potassium niobate ceramic investigated by impedance spectroscopy. Catal. Today 2003, 78, 529−538. (43) Barde, R. V.; Nemade, K. R.; Waghuley, S. A. AC conductivity and dielectric relaxation in V2O5-P2O5-B2O3 glasses. J. Asian Ceram. Soc. 2015, 3, 116−122. (44) Abhilash, K. P.; Christopher Selvin, P.; Nalini, B.; Somasundaram, K.; Sivaraj, P.; Chandra Bose, A. Study of the temperature dependent transport properties in nanocrystalline lithium

lanthanum titanate for lithium ion batteries. J. Phys. Chem. Solids 2016, 91, 114−121. (45) Armstrong, R. D.; Bell, M. F.; Metcalfe, A. A. In A Specialist Periodical Report-Electrochemistry; Thirsk, H. R., Ed.; The Chemical Society: Burlington House, London, 1978; pp 1−98. (46) Rao, K. J.; Baskaran, N.; Ramakrishnan, P. A.; Ravi, B. G.; Karthikeyan, A. Structural and Lithium Ion Transport Studies in Sol− Gel-Prepared Lithium Silicophosphate Glasses†. Chem. Mater. 1998, 10, 3109−3123. (47) Dyre, J. C. A simple model of ac hopping conductivity in disordered solids. Phys. Lett. A 1985, 108, 457−461. (48) Saravanan, K.; Balaya, P.; Reddy, M. V.; Chowdari, B. V. R.; Vittal, J. J. Morphology controlled synthesis of LiFePO4/C nanoplates for Li-ion batteries. Energy Environ. Sci. 2010, 3, 457−464. (49) Rao, R. P.; Reddy, M. V.; Adams, S.; Chowdari, B. V. R. Preparation, temperature dependent structural, molecular dynamics simulations studies and electrochemical properties of LiFePO4. Mater. Res. Bull. 2015, 66, 71−75. (50) Abhilash, K. P.; Christopher Selvin, P.; Nalini, B.; Jose, R.; Vijayaraghavan, R.; Chowdari, B. V. R.; Adams, S.; Reddy, M. V. Investigations on the influence of Sm 3+ ion on the nano TiO 2 matrix as the anode material for lithium ion batteries. J. Alloys Compd. 2017, 710, 205−215. (51) Gosser, D. K., Jr. Cyclic Voltammetry: Simulation and Analysis of Reaction Mechanisms, 1st ed.; Wiley VCH Publishers: New York, 1994; pp 1−156. (52) Srinivasan, V.; Newman, J. Design and Optimization of a Natural Graphite/Iron Phosphate Lithium-Ion Cell. J. Electrochem. Soc. 2004, 151, A1517. (53) Bai, N.; Chen, H.; Zhou, W.; Xiang, K.; Zhang, Y.; Li, C.; Lu, H. Preparation and electrochemical performance of LiFePO4/C microspheres by a facile and novel co-precipitation. Electrochim. Acta 2015, 167, 172−178. (54) Chen, C.; Liu, G. B.; Wang, Y.; Li, J. L.; Liu, H. Preparation and electrochemical properties of LiFePO4/C nanocomposite using FePO4·2H2O nanoparticles by introduction of Fe3(PO4)2·8H2O at low cost. Electrochim. Acta 2013, 113, 464−469. (55) Nithyadharseni, P.; Abhilash, K. P.; Petnikota, S.; Anilkumar, M. R.; Jose, R.; Ozoemena, K. I.; Vijayaraghavan, R.; Kulkarni, P.; Balakrishna, G.; Chowdari, B. V. R.; Adams, S.; Reddy, M. V. Synthesis and Lithium Storage Properties of Zn, Co and Mg doped SnO 2 Nano Materials. Electrochim. Acta 2017, 247, 358−370.

6455

DOI: 10.1021/acsomega.8b00527 ACS Omega 2018, 3, 6446−6455