Synthesis of life1 - Xmxpo4 / graphene nanocomposite as cathode material to improve electrochemical properties for lithium - ion batteries

0.5740 0.22015 1.1596 STM2 1.2588 2.0851 0.4305 0.22015 1.1596 Mn2+, Ni2+ STM3 1.2588 2.2241 0.4305, 0.1454 0.22015 1.1596 Y3+ STY1 1.2588 2.6970 0.1149 0.22015 1.1596 STY2 1.2588 2.7245 0.0766 0.22015 1.1596 STY3 1.2588 2.7523 0.0383 0.22015 1.1596 Ni(NO3)2.6H2O = 290,8; Mn(NO3)2.6H2O = 287,04; Y(NO3)3.6H2O = 383 Bảng 2.5. The mass/volume of precursor using for synthesis of LiFe1-xMxPO4/Gr Samples LiOH.H2O (g) FeSO4.7H2O (g) M(NO3)n.6H2O (g) C6H8O6 (g) H3PO4 (g) Graphene (g) M 41.96 278.01 (*) 176.12 97.99 12 STN1-G1 1.2588 2.6411 0.1454 0.22015 1.1596 0.08605 STN2-G2 1.2588 2.6411 0.1454 0.22015 1.1596 0.1721 STM1-G1 1.2588 2.2241 0.5740 0.22015 1.1596 0.0859 STM2-G2 1.2588 2.2241 0.5740 0.22015 1.1596 0.1718 STY1-G1 1.2588 2.7245 0.0766 0.22015 1.1596 0.08633 STY2-G2 1.2588 2.7245 0.0766 0.22015 1.1596 0.17266 STM3-G1 1.2588 2.2241 0.4305, 0.1454 0.22015 1.1596 0.08453 6 2.2. The Analysis Method Iron contents in LFP and nickel doped LFP were analyzed by the volumetric titration method using KMnO4 in concentrated H2SO4 medium to fully oxidized-Fe2+ to Fe3+. The LFP and LFNP samples were stirred with 50 mL H2SO4 until the formation of a clear green solution was created. The solution was then titrated with 0.0125N KMnO4 The LFP, LiFe1-xMxPO4, LiFe1-xMxPO4/Gr crystalline structure, phase purity and the particles size were characterized using a Rigaku/max 2500Pc and D8 Brucker X-ray diffractometer (XRD) with Cu-Kα radiation (λ=1.5418 Å, 2: 0° to 90° at a scan rate 0.25 -1.00 °/s). Raman measurements were performed using Horiba Jobin Yvon LabRAM HR300 system with 514.5 nm laser radiation, and the resolution of the measurement system was 2 cm-1. Raman spectra ranges from 100- 3000 cm-1. With 1 µm penetration, the vibration of LFNP bonds was determined and its structure and thin nanographene was identified. The system uses two excitation lasers and 4 magnetic ways (600 lines / mm to 2400 lines/mm). The controled heat speed is completed by software with an error of ± 0.1oC, measuring materials in powder form. Thermogravimetric analysis (TGA) technique with the type of Seratam LABSYS Evo TG-DSC at a heating rate of 10 oK/min in argon environment was implemented to determine the impurities phase contents in the samples. The mass in a sample of 100-150 mg its decreases it’s following when heating from 100 -1200 oC, rate scan 5-10 oC/minutes, time for scan reaching 50 minutes was performed at HCM City University Department of Education. This method was used as a predication of thermal stability with the calculation of crystallization content with the sample. Flame Atomic Absorption Spectrometry (FAAS) on AA-6800 machine (Shimadzu, Japan) ionte in air-acetylene flame at about 2700 °C. This technique is typically used for determinations in the mg/ L range, and may be extended down to a few μg /L with radiation wavelength: 253.7 nm, sensitivity: 0.1 ppm, operation mode: CV-AAS at HCM city University of Science. Energy dispersive X-rays (EDX - Hitachi SU6600, pressure < 10 pA, time lines: 30- 40 s, resolution: 127 eV) analysis was conducted to identify the elements occurring in structure. And using EDS Edax Team software analyzed data. The composite morphology and particle size were characterized by Scanning Electron Microscopy (SEM– type of 4800 machine: 5 kV, 8.5 mm x 20.0 k, the SEM images was studied at National Institute of Hygiene and Epidemiology), Field Emission Scanning Electron Microscopy (FESEM-Hitachi SU6600 equipment with machine parameters: 10 kV, 9.8 mm x 20 k at Institute for Nanotechnology in HCM city) and Transmission Electron Microscope (TEM-JEM-1400 with resolution of 0.2-0.38 nm, capacity of 100 kW, magnification of 2-3 Å), High-Resolution Transmission Electron Microscopy (HRTEM-FEI T20 at Singapore). Those methods was deeply investigated determination of particle morphology as well as analyzing the graphene film coating structure on material particles. In addition, with high-resolution HR-TEM technique, it’s possible to identify graphene film image and thickness and number of layers (sheet). X-ray Photoelectron spectroscopy (XPS) is well-method was used to determine the percentage of elements on the thin layer of materials (<10 nm) including Li metal, especially, determination of the percentage of Fe2+ into structural materials was studied at AXIS-NOVA (XPS) X-ray photoelectric spectrometer (Kratos) at DGIST University (Korea) with AlK filter (hυ = 1486.6 eV, 150 W, 2.6 × 10-9 Torr). 7 The materials was mixed with acetylene black and copolymer binder (PVdF–HFP) (weight ratio 80:10:10) in N–methyl pyrrolidone (NMP) to make cell electrode. The ink solution was pasted on the aluminum foil with 0.1 mm thickness and dried in vacuum for 24 h. Electrochemical properties were specifically determined by different routes. Cyclic voltammetry method is the best choice to investigate the process of kinetic ion Li+ de-intercalation/intercalation on the face of surface, this measurement is performed on an electrochemical meter MGP2 (Bio-logic, France) with a scanning speed of 0.1-100 mV.s-1 with a voltage range of 2.5 to 4.0 V. Using A charge/discharge cycling method was studied structural stabilities and capacity performance. A charge/discharge cycling test for Swagelok–type battery was carried out in liquid electrolyte LiPF6/EC–DMC (1:1) at room temperature. Cells were assembled in a glove box under argon atmosphere with < 2 ppm H2O. Electrochemical studies were carried out using a MPG2 Galvano/Potentiostat (Bio–Logic, France; Applied Physical Chemistry Laboratory, University of Science, VNUHCM) in the potential window of 3.0-4.2 V versus Li/Li+ in the galvanostatic mode at the C/10 regime. The Electrochemical Impedance Spectra (EIS) supports the determination of conductivity and diffusion coefficient. In addition, the method supports determine the effect of diffusion or charge transfer when the battery performs charging. In this method, the battery system is assembled as in the charging and discharging procedure and measured on VSP devices (Bio-Logic, France) between 2.7 and 4.2 V versus Li+/Li. CHAPTER 3. RESULTS AND DISCUSSION 3.1. LFP materials electrode for Li-ion batteries 3.1.1. Characterization of crystalline structure and phase composition LiFePO4 was prepared to synthesize by solvothermal with variety of the different condition (table 2.3). Among them, ST01 sample using the ratio of precursors of Li: Fe: P = 3: 1: 1 was dissolved in ethylene glycol solvents (volume ratio EG/W = 4: 1). Ascorbic acid was added into the solution as an agent prevents oxidation Fe2+ to Fe3+. Structure of material having orthorhombic olivine-type shape with Pnma space group (JPCDS 96-101- 1112) without impurity phase and calculation of lattice parameters (Å) shows a =10.334 Å, b = 6.010 Å, c = 4.693 Å. Specifically, the sharp peaks in the patterns of LFP indicate that the powders are well crystallized and intensity its characteristic main peaks at the diffraction angles 2θ: 36o, 30o, 26o, 15o correspondent with the crystal planes of {311};{211},{202};{111}{200}. In addition, the amount of diffraction peaks includes over 20 peaks with low baseline and without spotting (Fig.3.1) Figure 3.1. XRD pattern of synthesized LFP comparative with standard LFP 8 The elements impact the rate of chemical reaction, type of solvents and the temperature heating on LFP phase structure. For the effect of firing temperature, LFP materials after drying have formed olivine phase but can still be mixed with organic or complex water-based impurities (XRD results do not detect impurities with small amount of compound less than 5.wt% such as Fe3(PO4)2 which can be pyrolysis at 200-300 oC . To solve this problem that is elevate these impurities need to heat the material in gas atmosphere and perform thermal analysis to determine the amount of contaminant decomposing the decomposition temperature. Furthermore, the results of TGA analysis of ST01 (L1) and ST01 (L2) specimen also support determination of crystallization temperature and the ability stable phase. It was heated in the argon environment through 3 main steps. - The process heating range of 30-300 oC. In this stage, we can observe clearly the TG plot variations in temperature with less than 20 % of weight loss by release residual water and de-composite of organic compound to perform CO2 gas and H2O steam. - The second stage heating from 300-500oC, the weight has trend decreasing slightly with 2-3 wt% from SO42-, OH- (base) - Continuous increasingly heating between 550-900 oC, this is a stable structure of material with less none of negligible weight was lost. Thereby indicated the small amount of carbon free non- bonded to the LFP particle. Most of carbon modification is expected to be chemical linking to the surface of LFP particles. Hence, the content of crystallization of thermal stability materials 500-650 oC and crystallization content accounted for 85% of suitable reference [32, 109] (Fig.3.2) Figure3.2: TG plot ST01 (1) and ST01 (2) 9 3.1.2. Analysis of chemical elements composition LFP material In order to determine chemical components material using different methods to look for the accurately chemical elements into LFP material was supported by advancement of structural analysis techniques such as: Energy Dispersive X-rays (EDS), Atomic Absorption Spectrometry (AAS) and X-ray Photoelectron spectroscopy. Actually, the analytical results show that full presence of elements in the material with relative content in accordance with the theoretical value. For instance, the analytical results EDS of LFP: the elemental components in LFP (ST01) including: Fe, C, O, P. The ratio each element accounting for: O (38-42%); Fe (15-18%); P (15-18%); C (6-9%) comparative with reference [163] (Fig. 3.3). Hình 3.3. EDS chart of ST01 sample XPS measurements of powder samples was not only considerable determination of the oxidation states of iron Fe2+ and Fe3+ at the extreme surface of the electrode materials but also determinant the component of elements by binding energy peak position of element such as: C (1s) is 284.5 eV, indicating the presence of carbon on the surface of the material due to the reduction of ascorbic acid during LFP synthesis to create a surface-covering carbon. The connecting energy of Li element (1s) has a position of 55.0 eV, characteristic for the atomic Li+ ion in the LFP material structure. At peak 531.8 eV of element O (1s) proved the presence of O2- in the P-O bond of anion (PO4)3- in the structured olivine type LFP. The observation of only one P 2p doublet at this binding energy reveals the presence of only one environment for phosphorus, in good agreement with a (PO4)3- phosphate group. Element P (2p) peaked at 133.2 eV characterizing the P5+ ion of the orthogonal LFP (orthorhombic) and FePO4 hexagonal structure In addition, the spectra of Fe (2p) were split into two parts due to the coupling of spin - orbit corresponding to Fe (2p3/2) and Fe (2p1/2) with peak positions at 711.6 eV and 725.1 eV. This is the connecting energy of Fe (2p3/2) with the charge of Fe3+ and Fe (2p1/2) with the charge of Fe2+. The peak of Fe (2p3/2) can be divided into two positions 711.2 and 715.1 eV; Similar peak Fe (2p1/2) can be divided into two peaks at 724.1 and 726.9 eV, completely separated from the spectrum of Fe3+ in the form of Fe2p elemental absorption peak decay with the document [184] (Fig. 3.4). The amount of ferric iron was determined by XPS data analysis reached 15 % while the values up to 20% determined from chemical titration method. 10 Figure 3.4. XPS spectra of LFP Figure 3.5. The XPS of Fe2+ 2p3/2 and Fe3+ 2p1/2 of LFP 3.1.3. Morphology and particle size of LFP SEM images was shown the crystal shape morphology of LFP as nanorod. Almost crystal phase similarly size, agglomerate, however, more and more heating trend is the agglomerate size tend to reduce respectively 100-500 oC with particles size between 50-150 nm. Average size approximately 136 nm was measured ImageJ software. Figure 3.6. SEM images of LFP with different heating degree : 100 oC (a); 400 oC (b); 550 oC (c) 3.1.4. Electrochemical properties of LFP 3.1.4.1. Evaluation ability of Li intercalation/de-intercalation by CV method To evaluate ability of Li intercalation/de-intercalation in surface materials using Cyclic Voltametric (CV) with different scan rate of from 0-100 µV/s. The cell was cycled between 2.8 and 4.2V and charged at 0.1C. It is shown that indicating two-phase nature of lithium extraction/insertion reactions between LiFePO4 and FePO4 (Fe3+/Fe2+) with flat voltage plateaus at 3.4-3.5V range (versus Li+/Li). The peaks appearing on CV curves are of the reversible oxidation-reduction process of Fe3+/Fe2+ corresponding intercalation/de-intercalation of Li into the structure without interference of strange phase (Fig. 3.7) 11 Figure 3.7. Cyclic voltammograms of LFP (ST01) with different scan rate (a) and comparison CV curves between ST01and ST02 a scan rate of 80 µV/s (b) From CV curve, determine cathode current density parameters (Ipc), anode current density (Ipa), redox potential of materials at various rate. Based on Randles-Sevcik equation, it’s possible to build a linear relationship between the cathode current strength (Ipc) according to the square root of the potential scanning rate (v1/2) and thereby can define the diffusion coefficient of the ion Li+ (DLi+) in the structure of cathode electrode materials. Using the above equation, the Li+ diffusion coefficients of ST01 (7.5.10-13 cm2.s-1), ST00 (1.6.10-13 cm2.s-1). 𝑖𝑝 = (2,69 × 10 5)𝑛3 2⁄ 𝑣1 2⁄ 𝐴𝐷1 2⁄ 𝐶∗ (3.1) Where peak current 𝑖𝑝, the number of electrons involved in the redox process n, the concentration of lithium in the electroactive material C* (mol.cm-3) was calculated as 0.0228 mol.cm-3, the electrode area A(cm2), the potential scan rate 𝑣 (V s-1), diffusion coefficient D. 3.1.4.2. Characterization of charge-discharge To compare charge-discharge voltage profile between two samples ST01, ST02 and ST00 using the fixed current discharge method at scan rate of C/10 and charge-discharge voltage between 3.4 and 3.5V (Vs. Li+/Li). All the samples have similar charge-discharge curve with flat plateaus corresponding to the lithium intercalation/de-intercalationin/out the olivine structure. As a result, the significant increase of specific capacity was observed for ST01 samples. Specifically, ST01 capacity up to 120 mAh.g-1 is higher than the ST00 reaching 95 mAh.g-1. The discharge efficiency reaches 60% after 20 cycles (Fig.3.8). The reduced discharge efficiency may be due to the Fe2+ material being corroded by the electrode, leading to un-stable structure. (a) (b) 12 Figure 3.8. Charge-discharge curve after 20 cycles at C/10 of ST00 and ST01 samples. 3.1.4.3. Conductivity of materials In Fig.3.9 the Nyquist impedance plots are observed semicircle line of two LFP material samples (ST01 and ST00). LFP was investigated with two stages obtaining: Kinetic control process at high frequencies (notice: A, C) and Warburg diffusion control process at low frequencies (notice: B, D). Actually, the Warburg diffusion process was dominated in total process. By the EIS plot, the resulting impedance spectra obtained the resistances of the samples respectively ST01 (100 ) and sample ST00 (200 ). According to the Randles equation - Sevcik the resistances inversely proportional diffusion coefficient so the resistances becomes smaller the greater the diffusion coefficient Figure 3.9. Nyquist plots of EIS of LFP Successfully synthesized LFP by solvothermal method (EG/H2O = 4:1), ascorbic acid takes important role an oxidizing preventive agency as well as reducing agent Fe3+to Fe2+and also changes different ratio Li:Fe:P, results are as follows with the best ratio 3:1:1, X-ray has strong diffraction intensity with lower background. Using ascorbic acid and EG/H2O solvents can control nanoparticle nano size to improve the capacity from 95 mAh.g-1 ( ST00) growing to 120 mAh.g-1 (ST01), the charging discharge efficiency reaches 60% after 20 cycles. Lastly, the ion conductivity () sample ST01 (2.6.10-3 S.cm-1) and ST00 (1.2.10-3 S.cm-1). This is very consistent with our purpose to improve conductivity by using EG/H2O solvents and ascorbic agents to limit oxidation of Fe2+. Furthermore, organic solvents make increasing viscosity is a good way to reduce a particle agglomeration lead to Li+ ions more flexible. The conductivity of LFP material was synthesized by our research has a higher conductivity than the other publication similar condition with 10-5 S/cm [52]. To increase these properties may add carbon as a nano film covering material [54] published with LFP (5.9.10-9 S.cm-1) and LFP/C (8.3.10-2 S.cm-1). 13 Although the conductivity (2.9.10-3 S.cm-1) and diffusion coefficient (7.5.10-13cm2.s-1) were improved significant, the capacity is low. It need to improve increasing material durability. 3.2. Doped-LFP materials 3.2.1. Analysis of structural crystal and component phase Metal doping is the excellent way to directly affects the olivine structure of LFP. Analysis of X-ray diffraction of LiFe1-xMxPO4 figure out that the material has olivine structure, the lattice parameters tend to decrease and coincide with the standard spectral up to 20 diffraction peaks (JPCDS 96-400-1850) (Fig 3.18). Further, after doping, pattern of X-ray diffraction was analyzed to find out without impurities, peak intensity highly and evenly rank of 2 = 40-70o. Thus, the olivine structure of doped-LFP has not changed phase comparison with LFP, however, this shift is consistent with the rule of the theory of crystallography involved in the structure of the chemical element depending un-doped atomic radius which can make change. For Ni, Mn has a radius smaller than Fe, so the diffraction points are slightly shifted to the right (Fig.3.19b). Therefore, to determine accurate this issue consideration the scope of each diffraction. Through the X- ray diffraction data, most diffraction peaks clearly illustrate the olivine LFP structure, the lattice parameters after the doping are insignificant, indicating the limitation of network defects and good dispersion in the material according to Vergard's law: From other side, it is possible to explain the change in the volume of lattice cells after doping through two factors: firstly, the atomic radius of the doped metal element, the second is the content of the metal element doping. Thus, doped- nickel, the volume of crystal lattice is most reduced, which is perfectly suitable for theory because Ni's radius is the smallest compared to Mn, Y, Fe. Although yttrium has the smallest radius but a large electronic number (Z = 89) and the amount of doping is quite small, it has little impact on the network cell volume. In theory, the volume is inversely proportional to the electrical conductivity because it increases the ion density and shortens the diffusion distance. Figure 3.19. Figure out of shift of doped creating LiFe1-xMxPO4 (Ni, Mn, Y) 1 4 4 4LiFe M PO LiFePO LiMPO (1 ) x x a xa x a     14 Raman spectroscopy is an effective method to determine the characteristic vibrational molecular bonds or functional groups through characteristic vibration frequencies [154]. It should be mentioned that this separation is a guide to discussion only, because the vibrations may be coupled. The Fe-O bond corresponds two vibrational bands rank of 150-300 cm-1 and 1200-1300 cm-1, other bonds obtaining M–O Ni–O, Mn–O, Y–O with oscillation region 500-800 cm-1. The separation of region modes at 995 cm–1 and 1067 cm-1 was observed P–O bond. In Figure 3.21a, two modes intensity of bond Fe–O và P–O is the most strongly was clearly proved a huge number of Fe, O, P in sample [154, 155]. Moreover, from the characteristic vibrational regions of the material, it has been shown that the oscillation area Fe-O has a much more intense peak intensity than Ni-O, Mn-O and Y-O, this is consistent in accordance with the content of mass components of the materials. In addition, the presence of doped elements makes slight shift, increasing the density of vibrations compared to LFP materials, indicating that the charge density increases gradually after doping (Fig. 3.21b) in the frequency range of 0-1300 cm-1 without C-O bond area. Hình 3.21. Raman spectra of LiFe1-xMxPO4 material in vibrational bond 0-1000 cm-1. 3.2.2. The chemical component of LiFe1-xYxPO4 materials EDS and XPS methods support to determine the composition of elements and the relative percentage of the elements in the material. Chemical titration and XPS spectra can determine Fe2+ amount in structure phase. EDS –Mapping technology was studied the ability to distribute materials and elements present in the structure. The result of AAS analysis was approximately measured theoretical the ratio of doped-metal. However, AAS results do not clearly show the elements present in the structure so we also re-investigated EDS technique determination of the amount of chemical elements by EDS, especially, XPS. 15 Figure 3.22. The distribution of elements of LiFe1-xYxPO4 (STY2) using SEM-EDS –Mapping Figure 3.23. The EDS spectra of the grain of Ni, Y doped LiFePO4. For STN2 samples, the binding energy of Ni 2p3 (850 -900 eV) has appeared with a content of 1%, which confirms that Ni has participated in the material structure [45]. Similar to the STM3 sample that defines the binding energy area, there is element Ni 2p3 (850 -900 eV), Mn 2p3 (850 -900 eV). In addition, based on other binding energies, the presence of elemental components including Li, Ni, O, Fe, Mn can be accurately determined [86]. It’s also shown that the binding energy 710.1 eV. It indicates the oxidation state of Fe is +2. LiFe0.95Ni0.05PO4 (STN2) was used EDS analysis the result as Fig. 3.23. Comparison STN2 with STY2 (Fig. 3.23) through EDS spectrum analysis showed that the intensity of element C, O, P was almost unchanged. Fe content in two different samples is due to Fe in STY2 sample higher than STN2 sample. This is entirely consistent with the theory because the Y- element doped sample has lower Y content than Ni, so the amount of Fe in the STY2 sample is less replaced. 16 Figure 3.24. XPS spetra of LiFe1-xMxPO4 material Table 3.9b. The result analysis of elements mass (wt.%) 3.2.3. Morlopholy and size of LiFe1-xMxPO4 In three Ni doped samples, the particle size of the material is nanometer (70-100 nm), uniform, distinct like rod (nanorod) and the particles are more separate than the non-doped LFP material. This result is consistent with previous publications [52,142]. In particular, the STN2 sample has the best results. For Mn doping samples, the particle size is 20-150 nm and yttrium doping is 20-150 nm. SEM and FESEM images all showed a less uniform distribution of particle size in STN2 and STM2 samples compared to STN2 samples as shown in Fig .3.25. Figure 3.25. FeSEM of LiFe0.98Y0.02PO4 (STY2); LiFe0.95Ni0.05PO4 (STN2); LiFe0.8Mn0.2PO4 (STM2). 3.2.4. Electrochemical performances 3.2.4.1. The kinetic process of intercalation/de-intercalation Li+ ion Electrochemical properties of synthesized materials are essential role for application in lithium-ion batteries. Compared to pristine LFP, nickel doped LFP enhances the number of lithium ions intercalated into the structure, thus increases specific capacity of batteries. In fact, the cyclic Voltammetric (CV) with various rate from 20 to 200 V/s) at cycling test in the potential range 3.4 -3.5 V (vs Li+/Li) and the occurrence of oxidation peaks on the CV curve is the reversible oxidation reduction of Fe3+/ Fe2+ corresponding to the Sample Method C O P Fe Ni Y Mn STM2 EDS 4,96 41,08 15,10 19,04 - - 6,10 XPS 5,21 41,25 15,02 10,50 - - 6,25 STN2 EDS 6,27 40,99 14,01 20,03 3,10 - - XPS 5,49 42,05 16,20 21,00 3,10 STY2 EDS 5,69 40,78 18,80 20,22 - 0,01 - 17 intercalation/de-intercalation of Li+ ions into the olivine structure. The results of peak oxidation increasing respectively STN2, STY2, STM2. This can be explained that STM2 is the most currently density all of these. According to Randles-Sevcik equation to calculate the diffusion coefficient of the samples respectively in ascending order: ST01 <STY2 <STN2 <STM2. This result showed that the