Si–O–AL–Zn thin-film negative electrode for lithium-ion batteries
The stability of the silicon-containing thin films is significantly higher than that of amorphous silicon, not to mention crystalline silicon. In amorphous and crystalline silicon the considerable (up to three times) increase in the specific volume by the introduction of lithium leads to rapid destruction of material and loss of capacitive characteristics of the films. To achieve the specific capacity over 1000 mA·h/g is possible, as a rule, for films whose thickness does not exceed 300 nm [1–6]. To achieve the negative electrode specific capacity at a rate of 0.8 mA·h/cm2 and above, the thickness of the silicon-containing film needs to reach several micrometers. As experience shows and, in particular, the results of the present project, with increase in the thickness of the film, its stability is drastically reduced.
Of these options, the highest specific capacity with good cyclability was found in Si-Al-Zn composite . It should be noted that high values of capacity and of cyclability of this material were obtained for films with a thickness close to 100 nm. In  a composite SixZnyAlz films and the effect of mass percentage of elements on specific capacity and cyclability of electrode material were studied. In particular, it was found that in the interval of 26 ≤ x ≤ 47 high material capacity (1072 mA·h/g or higher) is combined with high cyclability (95% of initial capacity was retained at the 50th cycle). The content of Zn for ensuring high capacity and stability must be in the interval of 16 < y < 69, as the phase transition of Si-Li compound from the amorphous state to the crystalline is effectively suppressed, and the discharge capacity at the 50th cycle is up to 90% of initial capacity. For the same reasons, the z value must be in the interval of 22 ≤ z ≤ 46.
In  the three-component magnetron sputtering at a constant current with independent control is considered as one of methods of obtaining SixZnyAlz thin films. An example of film deposition using three targets is presented, when the power of the first magnetron (Si) was 185 W, and of the second and third magnetrons (Zn and Al) – 50 W each.
The object of the present work is experimental verification of the results of ; study of the possibility of obtaining of Si–O–Al–Zn composite films, including with low oxygen content; obtaining high values of surface specific capacity and cyclability at the film thickness of 2–4 µm.
OF EXPERIMENTAL SAMPLES
Films of silicon composite with a thickness of 2–4 µm were fabricated by magnetron sputtering on the titanium foil at a constant current in an argon and oxygen plasma from two targets Si0,9Al0,1 and Zn with use of the installation MVU TM MAGNA 10 (NIITM, Zelenograd), shown in Fig.1. Before spraying the titanium foil substrate with a thickness of 12–18 µm were cleaned in argon plasma and heated to 200°C for 60–120 s. The pressure of argon or argon and oxygen during deposition was maintained equal to 1.5 Pa. For samples 1 and 2 the actuation gas mixture consisted of argon and oxygen (flow rate of 0.05 liters/hour). For samples 3–9 only argon was used. The power of the Si-Al magnetron was maintained equal to 600 W, the power of the Zn magnetron was varied from 50 W for samples 1, 2, 3, 8, 9 to 75 W for samples 4, 6, 7, and to 100 W for sample 5. The deposition time was 39–40 minutes for all samples. To measure thickness of films and study their structure a polished silicon wafer for deposition of a witness sample was placed in the same installation. The second witness sample of the glassceramics with size of 60 × 48 mm2 was used to measure the resistivity of the films.
The thickness of Si–O–Al–Zn film, the morphology of the chip and of the film surface on the silicon witness sample was investigated using Quanta 3D 200i scanning electron microscope. The elemental composition of the films was investigated using EDAX energy dispersive spectroscopy detector for Quanta 3D 200i electronic microscope. The study of samples by the method of x-ray phase analysis was performed using ARL X'tra (Thermo Scientific, Switzerland) powder diffractometer with Kα radiation of copper with a wave length of λ=1.5418 Å, while tube voltage was 30 kV with current of 30 mA. The registration was conducted with Bragg-Brentano focusing. The x-ray diffraction database PDF-2/Release 2009 was used for interpretation of data. Studies of Raman spectra were carried out using EnSpectr R532 Raman spectrometer, which is equipped with semiconductor laser with a wavelength of 532 nm, power of 20 mW, and has a spectral resolution of 6 cm-1 and spectral range of 140–6030 cm-1. Registration settings: exposure of 2 s, 100 passes. The resistivity of films was measured by four-probe method on glassceramic second witness sample. Using VLR 200 scales the weight of the Si–O–Al–Zn film was estimated as the difference of the masses of glassceramic witness sample before and after film deposition.
STUDY OF SURFACE
AND CHEAP MORPHOLOGY
The main trend in the morphology changes of the chip and surface of the films is the development of surfaces at increase of power of sputtering of a zinc target. Example of the dependence of surface and cheap morphology on the magnetron power is shown in Fig.2.
To assess the change in the composition of Si–O–Al–Zn nanocomposite on the film thickness, energy dispersive analysis of elemental composition was carried out at accelerating voltages of 10 and 25 kV. The results of the analysis are presented in tables 1 and 2.
The initial ratio of the elements in the Si-Al target is: Al (wt.%) / Si (wt.%) = 0.107. According to table.2, after deposition the ratio of the concentrations of Al and Si in films is somewhat lower. Its average value calculated using the data of table.2 is: Al (wt.%) / Si (wt.%) = 0.096. However, the dispersion of ratios of the experimental concentration values from sample to sample is small, which indirectly confirms the validity of the data of energy dispersive analysis.
The concentrations of the components of all SixZnyAlz samples are in the intervals of 29.51 ≤ x ≤ 68.94, 12.48 < y < 61.27, 4.30 ≤ z ≤ 7.70 near the surface, and in the intervals of 31.56 ≤ x ≤ 77.47, 10.78 < y < 57.94, 3.55 ≤ z ≤ 7.48 in volume of the film. Although it is not three-component system (as in , if magnetron sputtering was used), for easy comparison of table.1 data with data of  they can be represented using Roseboom concentration triangles (Fig.3, 4).
A feature of the films is the enrichment of surface with oxygen compared to the more deeply lying layer of the film for samples 1–5 and almost the same film composition by oxygen for samples 6–9. The low rate of oxygen supply of 0.05 l/h for samples 1–2 and no oxygen supply for samples 3–9 have practically no effect on the oxygen content in the film. The Al content in each sample varies slightly, at the same time his change from sample to sample is more noticeable in a near-surface layer of a film. A change in the Zn content in the film does not always corresponds to the change in power of Zn-magnetron and hence to the rate of growth of the film. This fact requires additional checks for the next series of samples. The presence of carbon in small quantities is caused, most likely, by his adsorption during contact of the film with the atmosphere.
X-RAY PHASE ANALYSIS OF FILMS
The results of x-ray phase analysis are based on the following ICDD PDF2 database cards: Ti – 00-044-1294; Si – 00-027-1402; Zn – 01-073-6858. It should be noted that when using the titanium substrate the signal from the small quantity of aluminum is essentially not noticeable. X-ray diffraction patterns of the samples 1–9 of the titanium foil are shown in Fig.5.
Samples 1 and 2 do not contain the crystallized silicon, but contain a small fraction of crystalline zinc. In samples 3–5 the crystalline phase of Zn increases with increasing power of Zn-magnetron: at 50 W sizes of blocks of coherent scattering for Zn are 12.1 nm, at 75 W they reach 27.8 nm, and at 100 watts – 52.0 nm. In turn, the sizes of blocks of coherent scattering for Si fall with increasing of the power of the Zn-magnetron: 12.7 nm, 7.9 nm and 2.4 nm at powers of 50 W, 75 W and 100 W, respectively. X-ray diffraction analysis of samples 6, 7 and 8, 9 confirmed that the crystalline phase of Zn increased with increasing of power of the Zn-magnetron. As for the crystallized phase of silicon, it is observed only in samples 7 and 9.
Fig.6 presents the Raman spectra of the samples 3–9 in the range of 150–1150 cm–1. Table 2 contains data of relative content of phases of amorphous and crystalline silicon in the samples that are obtained by qualitative assessment of the processed spectra.
Sharp peaks in the area of 505–511 cm–1 correspond to crystalline phase of silicon. Peaks are shifted from the well-known line of 519 cm–1 for monocrystalline silicon, which may indicate a violation of the periodicity of the lattice due to embedding of aluminium atoms . As shown in , the shift of the peak of monocrystalline silicon on such a value may be caused by the crystallization of Si in the form of grains with a size of 2–3 nm that for a sample 5 will be consistent with the results of x-ray diffractometry.
Peaks of more flat form in the area of 450–490 cm–1 indicate the presence of silicon in the amorphous state. Feature around 930 cm–1 is typical for crystalline silicon. In  the spectra of crystalline, amorphous, polycrystalline and microcrystalline silicon are presented. General view of the spectra presented in Fig.6 is close to microcrystalline silicon. Thus, all samples contain amorphous and crystalline silicon in various proportions.
DISCUSSION RESULTS OF STUDY
OF MORPHOLOGY, ELEMENTAL
AND PHASE COMPOSITION OF FILMS
Technological parameters of the manufacture of experimental samples and their physical characteristics are presented in table.3. The table shows that the results of x-ray phase and Raman analyses have some differences. This can be explained by the fact that the laser in Raman spectrometer allows to analyze the surface layer of the film not thicker than 100 nm. Approximate calculations of the extinction ratio for the wavelength of 532 nm for amorphous silicon give the value of the radiation penetration depth of 85 nm. For metals this value is an order of magnitude smaller, whereas in the single-crystal silicon it is about 2 µm . These estimates help to explain the differences in the data obtained by x-ray phase and Raman analyses, considering that they belong to different volumes of the material. Thus, x-ray phase and Raman analyses complement each other in determining the phase composition.
OF NEGATIVE ELECTRODES
Samples 1–9 were tested in the models of half-elements to determine the specific characteristics for the introduction of lithium. Previously the samples were dried under vacuum at 120°C for 4 hours to remove adsorbed water. Next, in the box with atmosphere of dry argon models of batteries with the additional reference electrode were assembled. A main electrode was made of Si–O–Al–Zn, and a counter and reference electrodes – of lithium metal. The last one was rolled with a layer with thickness of about 100 microns on a substrate of nickel mesh with welded current output of nickel foil. The electrodes were separated by a nonwoven polypropylene. Models of batteries were filled with electrolyte LP 71. The area of the main electrode was 2.25 cm2. The models were tested using the automated charge-discharge measuring and computing complex АZRIVK – 0.05 A-5V of the NTC Booster company. For the sample 5 charging and discharge current was 700 mA, for the other samples – 250–310 mA. The potential difference in charge-discharge mode was in the range from 0.01 to 2.00 V.
The dependence of the discharge capacity on the number of charge/discharge cycles for Si–O–Al–Zn film samples is shown in Fig.7. The growth of the specific capacity of samples 1 and 2 in the first cycles, obviously, is caused by the partial recovery of the silicon oxide.
Specific capacity in the first cycle of the electrochemical tests for all the samples is in the range of 0.25–0.50 mA·h/cm2, which is below the values of specific capacity for the Si-O-Al films of similar thicknesses [12, 13]. By the nature of the change of discharge capacity during cycling all of the samples can be divided into three groups. The first group includes samples 2, 3, 6 and 8, whose specific capacity decreases dramatically during the first 10–15 cycles. The second group includes samples 1, 4, 5 and 9, which have the speed of decrease of capacity during cycling lower than the first group. The third group includes sample 7, which has no noticeable change of the discharge capacity from the tenth to the twentieth cycle of charge-discharge.
It is interesting to note the considerable difference in the cyclability for samples 6 and 7. These samples manufactured under the same technological parameters, have different thickness. Better cyclability have the sample 7, which has a greater thickness, although thick films are typically worse for cyclability. On the concentration triangle, which is built on the basis of elemental analysis (table.1), the points of these samples are virtually identical. At the same time the sample 7 differs from 6 and other samples (except a sample 5) by high content of crystal zinc. Significant differences were also observed in resistivity, which for sample 7 is 0.019 Ohm·cm, and for the sample 6 is 0.58 Ohm·cm. It is possible that a good cyclability of sample 7 and the relative stability of the sample 4 are caused by the low resistivity, which ensures the reversibility of lithiation and delithiation. With some reservations this can be attributed to the sample 5, which has sufficiently low resistivity (0.068 Ohm cm) and the highest content of crystalline zinc.
Overall, the comparison of the data of elemental analysis of the samples and their electrochemical testing does not give a definite answer to the question about the optimal ratio of elements that provides high capacity and good cyclability of Si–O–Al–Zn films. It is obvious that the electrochemical properties correlate well with the ratio of crystalline and amorphous phases of Si–O–Al–Zn composite film. In sample 7 this ratio of phases was the most successful.
It should be noted that the data of  are relevant to the films with a thickness of 70 nm, whereas the tables 1 and 2 show the results of averaging with a thickness of 2–3 microns. Thereafter, the concentration ratio shown in Fig.3 and 4, in a layer with thickness of 70 nm can significantly deviate from average values. Thus, when comparing the results of the present project with data of , it is necessary to keep in mind that it is legitimate only under the condition of high homogeneity of film thickness.
The composite Si–O–Al–Zn films are obtained by magnetron sputtering at a constant current and the studies of their elemental and phase composition are executed. It is shown that the combination of scanning electron microscopy, x-ray diffractometry and Raman spectrometry allows better study the morphology and estimate the phase composition of composite films.
The electrochemical tests showed significant differences in the discharge capacity and cyclability of films obtained at different modes of sputtering, which is caused by, in the first place, their elemental and phase composition. At the thickness of 3.5 µm and a specific ratio of amorphous and fine-crystalline phases of Si and Zn in the Si–O–Al–Zn composite film it is possible to achieve the specific capacity of the thin film negative electrode of 0.75 mA·h/cm2 with good cyclability charge-discharge. This result is comparable with capacity of Si-O-Al multilayer films, whose manufacturing technology is much simpler. Thus, significant advantages in capacity and stability of Si–O–Al–Zn films compared to the Si-O-Al  at a thickness of more than 2–4 µm are not revealed. ■
The project is executed at financial support of the Ministry of education and science of the Russian Federation. Agreement No. 14.576.21.0021 of 30 June 2014. Unique identifier of applied research (project) RFMEFI57614X0021.