Keywords

Introduction

In terms of a specific power traditional electrochemical system of a lithium-ion battery, manufactured since 1991 (lithium cobaltate–graphite), approaches its theoretical limit [1, p. 100]. One of the new electrochemical systems of a lithium-ion battery, such as lithium iron phosphate–lithium titanate, has ultimately higher power. It is conditioned by specific features of current-producing processes in two-phase systems, as well as the essential necessity to use functional electrode materials in the nanosized form [10, pp. 74, 203]. It is obvious that in terms of specific power the lithium iron phosphate–lithium titanate system will lose out to the lithium cobaltate–graphite system due to reduced voltage [10, p. 590]. Simultaneously, a number of applications, such as fixed energy storage units or load leveling systems, require batteries tolerant to high charging/discharging currents, while specific power becomes unimportant for them.

That is why a new electrochemical system for lithium-ion battery with a positive electrode based on doped lithium iron phosphate and a negative electrode based on doped lithium titanate was developed.

Relatively low electronic and lithium conductivity (10−13 S cm−1) [25, p. 589; 26, p. 1237; 2, p. A103; 21, p. 1241], as well as low discharge capacity, which is less than that of graphite, can be qualified as disadvantages of Li4Ti5O12. However, these disadvantages are largely compensated by high cycling stability, especially at high charge/discharge currents, while silicon and tin quickly lose their initial capacity due to degradation resulting from significant volume change at lithium intercalation. To increase electronic and lithium conductivity, there were attempts of heterovalent doping of Li4Ti5O12 with divalent (Cu2+), trivalent (Cr3+, Sc3+, Al3+, Tb3+), and quintavalent (Ta5+) cations [25, p. 590; 26, p. 1238; 2, p. A103]. In a number of cases, increased conductivity of obtained materials was observed but electrochemical properties of the doped Li4Ti5O12 were not studied. A number of authors have reported results of electrochemical performances of the doped lithium titanates, but the range of cycling was from 1 to 3 V [22, p. 13198]; [4, p. 396]; [23, p. 1443]; [8, p. 375]; [11, p. 128]; [12, p. 1036]; [9, p. 2250]; [7, p. 748].

Different ways for further improvement of this active material have been extensively studied for the last few years: development of advanced nanostructured LFP-carbon composites; replacement of carbon by conductive, electrochemically active polymers; doping of LFP by the ions of transition metals; and so on. In particular, LFP doping with vanadium has been suggested as a way for the increase in mobility and diffusion coefficient of Li+ ions due to lattice expansion and Li–O interaction weakening [19, p. 2956]. The authors of [20, p. 207] have studied the structure and properties of LiFe0.9V0.1PO4 and found that the cathode properties of the doped counterpart, including reversible capacity, cyclability, and rate capability are better than those of LiFePO4. Later the lengthening and weakening of Li–O bond and improvement of the electrochemical performance especially under the high C rate were confirmed by the examples of LiFe0.95V0.05PO4 [24, p. A730], LiFe0.97V0.03PO4 [18, p. 842], and LiFe0.99V0.01PO4 [13, p. 1019]. Other ions of transition metals, such as Mn [15, p. 446], V [3, p. 280], Mg [16, p. 340], Ni [17, p. 830], Co [6, p. 145], Mo [5, p. 9963] can act as dopants.

Experimental

Lithium iron phosphate of the Li0.99Fe0.98Y0.01Ni0.01PO4 composition was synthesized using the sol–gel method. At the first stage of synthesis, initial reagents were dissolved in stoichiometric ratios in deionized water. Fe(NO3)3·9H2O (Sigma-Aldrich, >98%), Li2CO3 (Sigma-Aldrich, >98%), (NH4)2HPO4 (Sigma-Aldrich, >99%), Ni(NO3)2·6H2O (Sigma-Aldrich, >98.5%), Y2O3 were used as reagents. When heated, the Y2O3 oxide was preliminarily dissolved in the concentrated HNO3. The solutions obtained after mixing of initial reagents were vaporized with post-heat treatment in an inert atmosphere at (100–1000 ℃) in order to obtain the crystallized product. Composite materials with carbon were obtained to increase the electrical conductivity of Li0.99Fe0.98Y0.01Ni0.01PO4. The carbonaceous coat is usually applied using pyrolysis of organic compounds at high temperature (600–900 ℃) in the inert atmosphere [14, p. 538]. In this paper, glucose was chosen as the source of carbon. Samples of doped lithium iron phosphate were ground with glucose samples with different weight and were annealed at 800 ℃ in an inert atmosphere. In these conditions, carbonization is observed. The carbon content in the composites was determined thermogravimetrically and was 6–12%.

Gallium-doped lithium titanate was synthesized using the citrate method. Titanium tetrabutylate (99%, Alfa Aesar) and lithium carbonate (99%, Fluka) were dissolved in the ethanol-nitric acid mixture (volume ratio 5:1), and gallium solutions (99.99%, Aldrich) in nitric acid and citric acid (98%, Sigma) were added in the minimum quantity of water. Lithium carbonate was taken with a 5% surplus to prevent possible losses of lithium during subsequent annealing at high temperatures. The obtained mixture was heated sequentially at 95 ℃ during 24 h and at 250 ℃ during 5 h. The so formed precursor was ground in an agate mortar to a smooth paste that was subjected to final annealing at 800 ℃ during 5 h in air.

The X-ray phase analysis (XPA) of synthesized samples was conducted with a Rigaku D/MAX 2200 diffractometer with the CuKα radiation. The Rigaku Application Data Processing software package was used for spectra processing. X-ray patterns were processed in FullProf Suite program (WinPlotr), lattice parameters were updated using the Checkcell debugging tool.

The microstructure of samples and study of the composition of elements of materials were analyzed with Carl Zeiss NVision 40 scanning electron microscope at accelerating voltage of 1 kV.

Cathode and anode masses for electrodes were prepared using the ratios as follows: 85 mass% of doped lithium iron phosphate (or doped lithium titanate), 10 mass% of carbon black, 5 mass% of PVDF. The latter was dissolved in N-methylpyrrolidone. Active masses based on doped lithium iron phosphate and doped lithium titanate were heated and applied to aluminum foil substrate with MSK-AFA-II-Automatic Thick Film Coater. After drying, electrode plates were placed in roller press. The semi-finished product was pressed at 2t. Rolled electrode sheets were cut into ready electrodes sized 55 x 55 mm2 with MiniMarker A2 laser marker, which were subsequently used for assembling batteries. To determine the specific electrochemical capacity of cathode and anode material, small electrodes sized 1.5 x 1.5 cm2 were cut out, and galvanostatic studies were carried out in three-electrode electrochemical cells. The thickness of the positive electrode’s active layer was 90 µm. The thickness of the negative electrode’s active layer was about 50 µm. The difference in thickness was conditioned by the difference in specific capacity of doped lithium iron phosphate and doped lithium titanate. Batteries and electrochemical cells were assembled in a glove box in a dry argon atmosphere. The 1 M LiPF6 in a mixture of ethylene carbonate-diethyl carbonate-dimethyl carbonate (1:1:1) prepared in the laboratory of the Frumkin Institute of Physical Chemistry and Electrochemistry of the Russian Academy of Sciences from Battery Grade commercial reagents was used as an electrolyte.

Results and Discussion

X-ray diffraction analysis method showed that the X-ray pattern of the synthesized sample of Li0.99Fe0.98Y0.01Ni0.01PO4 contains reflexes of LiFePO4 (triphylite, orthorhombic modification, Pnma space group). Reflexes of other phases are not detected. X-ray patterns of Li0.99Fe0.98Y0.01Ni0.01PO4 were compared with the Card No. 81-1173 of the powder database of diffraction standards PDF2. Based on the obtained data, the inference should be drawn that the obtained materials represent lithium iron phosphate with the structure of olivine.

When using the sol–gel synthesis method, the chemical composition of obtained materials must be determined by the ratio of initial reagents. However, lithium compounds can be volatile at high temperatures of final annealing, which can result in stoichiometric impurity in the end product. That is why the content of cations and phosphate included as compounds of the material was determined using inductively coupled plasma mass-spectrometry (ICP-MS). For this purpose, the samples were preliminarily dissolved and the solution composition was analyzed. It was demonstrated that the composition of obtained samples corresponds to the initial load (Table 1).

Table 1 Elements content according to inductively coupled plasma mass-spectrometry
Full size table

Following the results of scanning electron microscopy, the size of particles of Li0.99Fe0.98Y0.01Ni0.01PO4 is about 50 nm. The particles form agglomerates from 100 nm to 300 nm (Fig. 1).

Fig. 1
figure 1

Micrographs of Li0.99Fe0.98Y0.01Ni0.01PO4/C composite

Full size image

X-ray patterns of doped lithium titanate samples contain reflexes of Li4Ti5O12 only (Card No. 72-0426 PDF-2 database), which evidences that the obtained material is single-phase. Radii of cations of Ga3+, Ti4+, and Li+ are similar in size, that is why gallium ions can get inserted into both, positions of titanium, and positions of lithium. Ga3+ ion attracts oxygen ions more intensively than Li+ ion considering Coulomb interaction, which results in a greater lattice contraction. To verify this assumption, the structure of doped lithium titanate was updated using the Rietveld method. According to the data obtained as a result of updating the structure using the Rietveld method of the sample of Li4+xTi5-xGaxO12 composition at x = 0.1 (Tables 2, 3), gallium ions occupy both, positions of lithium (8a), and positions of titanium (16d).

Table 2 Results of updating the structure of Li3.812Ti4.972Ga0.1O12 at 25 ℃
Full size table
Table 3 Coordinates of atoms and isotropic parameters of thermal bias (B) at 25 ℃
Full size table

Moreover, there are about 2.5 times less of gallium ions in octahedral sites than in tetrahedral ones. In accordance with these results, the formula of gallium-doped lithium titanate should be as follows: Li3.812Ti4.972Ga0.1O12.

According to the data from the scanning electron microscopy, synthesized samples of Li3.812Ti4.972Ga0.1O12 represent a rather homogeneous crystalline mass (Fig. 2). Growth steps are clearly seen in the micrographs. The particle size varies in the range of 450–550 nm.

Fig. 2
figure 2

Mircograph of Li3.812Ti4.972Ga0.1O12 composite

Full size image

The results of galvanostatic cycling in Fig. 3 revealed that the specific discharge capacity of lithium iron phosphate doped with yttrium and nickel at the current density of 20 mA/g which corresponds to the current C/8 was about 160 mAh/g. The increased current density logically resulted in the decreased discharge capacity. However, even at the current density of 30 С the discharge capacity was about 56 mAh/g. The discharge potential of lithium iron phosphate doped with yttrium and nickel at low current density (C/8) was about 3.4 V. At increased current densities (30 C), the discharge potential of Li0.99Fe0.98Y0.01Ni0.01PO4 lowered insignificantly and was about 3.2 V.

Fig. 3
figure 3

Charge–discharge curves (a) and dependence of the discharge capacity on the current density (b) Li0.99Fe0.98Y0.01Ni0.01PO4

Full size image

The ability of lithium iron phosphate to withstand high currents is explained by two factors: first, the high ion conductivity of this material, and second, the small size of particles of synthesized material. The results of galvanostatic cycling of negative electrodes from doped lithium titanate are represented in Fig. 4. Traditionally, lithium titanate is discharged to potential 1 V, which corresponds to the insertion of 3 lithium ions per formula unit. Simultaneously, it was shown in a number of papers that doped samples of lithium nanotitanate can be cycled in a wider potential range (up to 0.01 V), in this case, the discharge capacity is increased up to 275 mAh/g, which is 75% over the discharge capacity registered in a narrower potential range.

Fig. 4
figure 4

Charge-discharge curves (a) and dependence of the discharge capacity on the current density (b) Li3.812Ti4.972Ga0.1O12

Full size image

Analysis of Fig. 4 shows that extending the cycling range leads to the increase in the discharge capacity, in which case the ability of lithium titanate to operate at high current densities up to 40C (6400 mA/g) is preserved. It is important to emphasize that the extended cycling range does not result in the increased degradation during the cycling. The discharge current equal to 6400 mAh/g corresponds to the charge for 1.5 min (or 40C regime), in which case the discharge capacity remains at the level of about 50 mAh/g. It is important to note that such a high discharge rate corresponds to the current of about 180 mA/cm2, which ensures the high power of the battery. Seven (7) negative and seven (7) positive electrodes were used to manufacture a stack battery of 1 Ah rated capacity. The results of cycling are represented in Fig. 5. As Fig. 5a shows, the battery discharge capacity at the current of C/5 equals to the rated capacity, and the average discharge voltage is about 1.8 V, which evidences of the battery’s insignificant ohmic resistance. When the current density is increased, the battery’s discharge capacity and average discharge voltage are reduced, however, even at increased current densities up to 30 C  the discharge capacity is about 22% of the rated capacity. Degradation and cyclic life of the lithium iron phosphate–lithium nanotitanate system battery were determined at the charging-discharging current 1 A, which corresponded to the so called cycle service 1C. As the Figures show, the increase in charging-discharging current results in the decreased discharge capacity, as well as decreased average discharge voltage. Change in the discharge capacity during the cycling for 950 cycles was 16 mAh on average, which is about of 16% of the rated capacity. Thus, degradation during the cycling was 0.017% per cycle.

Fig. 5
figure 5

Charge-discharge curves a and change in the discharge capacity b of the doped lithium iron phosphate-doped lithium titanate system battery

Full size image

Conclusions

In order to develop a battery with increased power specifications, new materials for the lithium-ion battery were synthesized: cathode material based on lithium iron phosphate doped with nickel and yttrium (Li0.99Fe0.98Y0.01Ni0.01PO4) and anode material based on doped lithium titanate (Li3.812Ti4.972Ga0.1O12). Both electrode materials were characterized by their ability to operate at increased current densities (up to 30 C ). The lithium-ion battery of the doped lithium iron phosphate-doped lithium titanate system was developed based on these materials. The energy density of the battery was about 100 Wh kg−1, while its specific power was about 2 kW kg−1. The battery of this electrochemical system is intended for use in fixed energy storage units.