A comparative XPS surface study of Li₂FeSiO₄/C cycled with LiTFSI- and LiPF₆-based electrolytes

Abstract

X-Ray photoelectron spectroscopy (XPS) has been used to characterise the surfaces of carbon-coated Li₂FeSiO₄ cathodes extracted from Li-ion batteries in both a charged and discharged state. 1 M lithium bis(trifluoromethylsulfonyl)imide (LiTFSI) and lithium hexafluorophosphate (LiPF₆) based electrolytes were used with ethylene carbonate (EC) and diethyl carbonate (DEC) as organic solvents. The LiTFSI-based electrolyte exhibited high salt stability and no significant formation of LiF. However, solvent reaction products from EC were found together with lithium carbonate. A LiPF₆-based electrolyte, on the other hand, showed inferior salt stability with Li_xPF_y, Li_xPO_yF_z and LiF species formed on the surface. Solvent reaction products together with lithium carbonate were also found. There are also indications that Li₂FeSiO₄ is degraded by the HF formed in the electrolyte by the hydrolysis of LiPF₆. A better understanding of the surface chemistry of carbon-coated Li₂FeSiO₄ after the first cycles in a Li-ion battery has thus been achieved, thereby facilitating the optimisation of Li-ion batteries based on this potentially cheap and electrochemically most promising cathode material giving excellent capacity retention: <3% drop over 120 cycles.

---

Introduction

Li₂FeSiO₄ is a recent addition to the group of interesting and potentially cheap cathode materials for large-scale lithium-ion battery applications: it shows great promise, especially in terms of its electrochemical stability. Li₂FeSiO₄ has a theoretical capacity of 166 mAh g⁻¹ based on the reaction:

Li₂FeSiO₄ → LiFeSiO₄ + Li⁺ + e⁻ (1)

In practice, a capacity of 120 mAh g⁻¹ has been achieved with only a ∼3% loss in capacity over more than 120 cycles.¹ Furthermore, the presence of strong Si–O bonds should promote the same lattice stabilization effect exploited in LiFePO₄.² The shift observed in the potential plateau from 3.10 to 2.85 V between the first and second cycles has been suggested to be related to structural rearrangement to a more stable structure.³ While details of this process are still not totally understood, it has been seen that occupation of the Li- and Fe-sites becomes more randomised between the first and subsequent cycles.

Surface-film formation at both the anode and cathode electrode–electrolyte interfaces is an undisputed source for irreversible capacity loss in the Li-ion battery. Extensive studies of graphite anodes in a Li-ion battery have shown that, during the first cycle, the so-called solid electrolyte interphase (SEI) is formed as a film completely covering the graphite particles. This film contains different quantities of polymeric degradation products, as well as inorganic species like LiF and Li₂CO₃. These inorganic species are normally incorporated into the polymeric matrix of the surface film. Ultimately, the composition of the SEI depends on the specific choice of lithium salt and organic solvent used in the electrolyte.^4–8

Surface-film formation on cathodes and its influence on battery performance has also been studied systematically. In particular, the transition-metal oxides have been shown to participate actively in the film formation,^9–13 resulting in films comprising a mixture of polycarbonates and various electrolyte-salt derived compounds. When using the LiPF₆ salt, these compounds are typically: LiF, Li_xPF_y and Li_xPO_yF_z. This surface-film formation is not only electrochemically driven, it can also have a chemical origin as a result of storage in the electrolyte.^10–12 There is no significant difference in the elemental composition of the surface-film after cycling and storage at elevated temperature compared to that at room temperature. However, the thickness and degree of coverage of the film are found to increase.¹¹ Indeed, the observed increase in ac-impedance for LiNi_0.8Co_0.2O₂ electrodes has been suggested to originate in the thicker surface layer formed at higher temperatures.^14,15

From a study of electrochemically cycled LiFePO₄ electrodes,¹⁶ it was concluded that no solvent-based reaction products were formed on the surface, indicating that the oxygen in the phosphate group does not take part in any such reactions, unlike the oxygen in the Co and Mn oxides. In LiFePO₄, the surface film consisted only of different salt-based products. Depth profiling could show that this film did not completely cover the electrodes or, at least, was not thicker than the detection limit of a few nanometres.

The surface layer formed at 55 °C on fresh uncycled Li₂FeSiO₄ electrodes compared with that formed on electrodes cycled with a 1 M LiTFSI EC:PC (1 : 1) electrolyte has previously been studied by photoelectron spectroscopy using both synchrotron and monochromatized AlKα radiation.¹ At ambient temperature, fresh electrodes exposed to air showed larger amounts of carbonate-based compounds, such as Li₂CO₃ and LiHCO₃, on their surfaces than electrodes stored under an inert atmosphere. This indicates that lithium is withdrawn from the original structure on exposure to air. After electrochemical cycling, the main component of the surface film is LiTFSI salt in its original form, together with small amounts of solvent reaction products, possibly lithium carboxylates. No LiF or carbonate-based compounds are found on the surface.

Here, the Li₂FeSiO₄ electrode surface layer is studied in detail by photoelectron spectroscopy in the light of earlier results for LiTFSI- and LiPF₆-salt based electrolyte systems.

Experimental

Li₂FeSiO₄ was prepared by the solid-state reaction of Li₂SiO₃ with FeC₂O₄·2H₂O. These starting materials were dispersed in acetone, mixed thoroughly and then ground together with 10 wt% polyethylene glycol. After evaporating the acetone, the mixture was heated to 700 °C for 20 h in a flow of CO–CO₂ (50 : 50) gas to suppress the oxidation of Fe²⁺. The crystal structure was readily confirmed by X-ray powder diffraction (not shown here). The carbon content of the synthesised material was determined to be 4.68(5) wt% using the carbon combustion method for elemental analysis.

The material was cycled electrochemically using “coffee-bag”-type cells. Electrodes were prepared by spreading a mixture of 90 wt% active material, 5 wt% carbon black and 5 wt% EPDM binder dissolved in cyclohexane onto an aluminium foil; the thickness of the active layer was ca. 25 µm. Circular electrodes (3.14 cm²) with a typical loading of 3–5 mg of active material were dried under vacuum at 120 °C, prior to assembly of the electrochemical cells. Half-cells of configuration <Li₂FeSiO₄|glass-wool separator soaked in electrolyte|Li foil> were assembled in an Ar-filled glove-box (<5 ppm H₂O and O₂). Two different electrolytes were used for the study: 1 M LiTFSI (lithium bis(trifluoromethylsulfonyl)imide) and 1 M LiPF₆ (lithium hexafluorophosphate). Each lithium salt was dissolved in a mixture of ethylene carbonate (EC) and diethyl carbonate (DEC) in a 2 : 1 volumetric ratio. The LiTFSI salt was dried under vacuum at 120 °C prior to electrolyte mixing; the LiPF₆ salt was dried at 80 °C. The solvents were used as received.

Galvanostatic cycling was performed at 55 °C using a Digatron BTS 600 testing system. Cells were precycled three times at a C/20 rate (current density: 4.7 mA g⁻¹) between 2.0 and 3.7 V. Cycling was stopped either in the discharged or the charged state. The cells were then dismantled in the glove-box and small pieces of the electrodes cut out, mounted on a sample-holder, and transported to the XPS equipment, using a specially designed transport chamber to avoid contamination by air or moisture. All measurements were performed on unwashed electrodes to preserve any surface species formed during the cycling. XPS measurements using monochromatized Al Kα at 1486.6 eV were performed on a PHI 5500 system. The spectrometer energy scale was calibrated using Cu2p_3/2 (932.7 eV), Ag3d_5/2 (386.2 eV) and Au4f_7/2 (84.0 eV) emissions of 3keV sputter-cleaned metal foils.

Results and discussion

Electrochemical characterisation

Fig. 1 shows the first charge and discharge cycle for the 1 M lithium bis(trifluoromethylsulfonyl)imide (LiTFSI) EC:DEC (2 : 1) cell (a) and the 1 M hexafluorophosphate (LiPF₆) EC:DEC (2 : 1) cell (b). Both cells were cycled at 55 °C and at a C/20 rate, and both exhibited two distinct plateaus: one at 3.10 V during charge and one at 2.76 V during discharge. The following charge (not shown) had a plateau at 2.85 V; this shift in potential from the first to subsequent charge-cycles has been discussed in a previous paper and will not be elaborated upon further here.³ The initial capacity for the LiTFSI-based cell is 100 mAh g⁻¹, while it is a little lower for the LiPF₆-based cell: 93 mAh g⁻¹. As reported earlier, Li₂FeSiO₄ exhibits a stable capacity of 120 mAh g⁻¹ with an electrolyte based on the LiTFSI-salt together with EC and propylene carbonate (PC) in a volumetric ratio of 1 : 1.³

Fig. 1 First-cycle voltammograms for carbon-coated Li₂FeSiO₄ electrodes at 55 °C and C/20 cycle rate for an electrode with 1 M LiTFSI EC:DEC (2 : 1) (a) and 1 M LiPF₆ EC:DEC (2 : 1) (b).

XPS characterisation

Surface characterisation by XPS has been conducted for both the 1 M LiTFSI EC:DEC (2 : 1) system and the 1 M LiPF₆ EC:DEC (2 : 1) systems. The XPS spectra for electrodes in an intercalated state are compared with those for deintercalated electrodes for both LiTFSI- and LiPF₆-based electrolytes in Fig. 2. The respective high-resolution spectra are compared in Fig. 3. The data has been background-corrected by subtracting Tougaard-type background functions. The spectra have been curve-fitted using a combination of Gaussian and Lorenztian curves (Voigt-type). These curves, together with the summation curve, are drawn as thin lines while the raw data appears as points. An overview of peak assignments and binding energies is given in Table 1.

Table 1 Assignment of XPS peaks from carbon-coated Li₂FeSiO₄ and surface/electrolyte components, based on ref. 1, 10 and 21–24. The values given for the absolute binding energies can vary for different samples, depending on the relative positions of the Fermi levels (E_F) with respect to the vacuum level (E_vac)

Assignments	Measured binding energy/eV
	C1s	F1s	Li1s	O1s	P2p3/2	Si2p3/2	S2p3/2	N1s
C/Li2FeSiO4	284.5	—	55.2	531	—	101.8	—	—
Carbon black	284.5	—	—	—	—	—	—	—
Hydrocarbon	285.3	—	—	—	—	—	—	—
LiN(SO2CF3)2	293.3	689.1	56–57	532.5	—	—	169.4	399.7
LiF	—	685.5	56–57	—	—	—	—	—
Li2CO3	290.2	—	56–57	532	—	—	—	—
R(CH2CH2O)n–R	286.7	—	—	533	—	—	—	—
LiPF6	—	688	56–57	—	137.8	—	—	—
LixPFy	—	687–688	56–57	—	136.5	—	—	—
LixPOyFz	—	—	56–57	534	134.5–135	—	—	—
–Si–F	—	—	—	—	—	102.7	—	—
–SOx	—	—	—	—	—	—	167.5	—

---

Fig. 2 A survey of the XPS AlKα spectra for lithiated and delithiated carbon-coated Li₂FeSiO₄ electrodes cycled with 1M LiTFSI EC:DEC (2 : 1) and 1M LiPF₆ EC:DEC (2 : 1).

Fig. 3 XPS spectra for lithiated (top) and delithiated (bottom) carbon-coated Li₂FeSiO₄ electrodes cycled with 1M LiTFSI EC:DEC (2 : 1) and 1M LiPF₆ EC:DEC (2 : 1). The C1s (a, b), F1s (c, d), O1s (e, f), Li1s/(Fe3p) (g, h), S2p (i), N1s (j), P2p (k) and Si2p (l, m) core-level emissions are presented. In the case of the Li1s/Fe3p emissions (g, h), the Fe3p peak for Fe(III)O is superimposed (thick solid line) to facilitate the separation of the overlapping Li1s and Fe3p levels.

LiTFSI

The C1s spectrum for the intercalated electrode with LiTFSI (Fig. 3a) displays a sharp dominant peak at 284.5 eV (FWHM: 0.8 eV) due to the presence of different (aliphatic) carbon species, i.e., the carbon black, the carbon coating on the Li₂FeSiO₄ particles and the EPDM binder. A smaller peak is observed at 285.3 eV within the main emission peak, which can be attributed to various hydrocarbons, including the EPDM binder. The –CF₃ groups of the LiTFSI salt can be found at a fairly high binding energy of 293.3 eV. Since the line width of this peak is quite narrow (FWHM: 1.05 eV) with a symmetric peak shape, the LiTFSI salt seems to be present on the surface and has not reacted significantly or become degraded. This is confirmed for the F1s spectra (Fig. 3c), where the large peak at 689.0 eV is assigned to the –CF₃ group in LiTFSI, and the much smaller peak at 685.5 eV is attributed to a minimal amount of LiF. A similar situation has been observed for the previously reported LiFeSiO₄/LiTFSI/EC:PC (1 : 1) electrode system in the intercalated state.¹ However, the peak at 290.3 eV corresponding to lithium carbonate and that at 286.8 eV were not reported in the previous study. The latter peak originates most probably from an oxygen-containing polymeric compound of polyethylene oxide (PEO) type, formed by a salt-anion initiated polymerisation of EC on the electrode.^6,17–20 These two findings are confirmed in the O1s spectrum (Fig. 3e), where the large peak at 533.4 eV contains both the LiTFSI oxygen and PEO-type polymer contributions. The smaller peaks at 531.9 eV and 531.1 eV can probably be attributed to Li₂CO₃ and the silicate oxygen in Li₂FeSiO₄, respectively. The Li1s/Fe3p emissions are shown in Fig. 3g. Due to the overlap of those regions, the Fe3p emission of a Fe(III)O (99%, Merck) powder sample is superimposed as a reference. The intensity has been normalized to that found for the Si2p peak of Li₂FeSiO₄ (discussed later), and an empirical correction constant has been applied for all measurements to fit the slope of the Li1s/Fe3p signal of the electrodes. To a first approximation the lithium contributions have thus been separated using this procedure. The emission found at 55.2 eV represents the Li⁺ in Li₂FeSiO₄, whereas that centred at 56.6 eV can be attributed to a number of components in the SEI (see Table 1).

The S2p emission of the –SO₂– group of the LiTFSI salt is shown in Fig. 3i. In the following, the binding energies of spin–orbit split core-level emissions are always referred to the main emission. The doublet separation (DS) of the S2p_3/2 and S2p_1/2 components is 1.185 eV. The peak is composed of a main emission at 169.4 eV and a weak shoulder component at 167.5 eV. This group must therefore be slightly reduced, leading to a shift to lower binding energies. The N1s spectrum in Fig. 3j is composed of only one component located at 399.7 eV. The peak reveals a FWHM of 1.7 eV, corresponding to the imide group of LiTFSI. The Fe2p core-level, however, cannot be addressed, since it overlaps with strong satellite peaks of the intense F1s emission.

The deintercalated electrode (Fig. 3a) shows more or less the same binding energies for the features in the C1s emission structure as the intercalated electrode, but with grossly differing relative intensities. The largest peak is no longer the carbon-coating and carbon-black/EPDM binder peak (at 284.6 eV), but the –CF₃ emission of the LiTFSI salt at 293.3 eV, while the intensity of hydrocarbon peak at 285.4 eV is slightly decreased. The decrease of the emissions at lower binding energies is accompanied by a strongly enhanced component (at 286.7 eV) corresponding to the PEO-type polymer. The products seem to contain mainly carboxylic groups, but also carbonates, since the Li₂CO₃ peak in the 290.6 eV region has increased slightly. This observation is supported by the normalized peak intensities given in Table 2. This is also reflected in the O1s spectra (Fig. 3e), which shows an intense LiTFSI and PEO peak at 533.3 eV and a Li₂CO₃ peak at 532.0 eV. This latter peak only indicates the presence of Li₂CO₃, contrary to the intercalated electrode, where an additional peak was observed attributed to Li₂FeSiO₄. The F1s spectrum of the deintercalated electrode (Fig. 3c) remains more or less the same compared to that for the intercalated electrode. The feature at 685.3 eV reveals a slightly increased though still minimal amount of LiF. As expected for a charged electrode, the Li1s signal at 55.2 eV (Fig. 3g) corresponding to Li⁺ in the Li_2−xFeSiO₄ structure is notably reduced compared to that for the intercalated electrode. On the other hand, the Li emission at 56.7 eV is increased as a result of the thicker surface layer.

Table 2 Normalized intensities of the peaks representing PEO-type surface species (C1s), the respective salt species (P2p, S2p), the C-coating (C1s) and the Li₂FeSiO₄ cathode material (Si2p); cf.Table 1. The integrated intensities were normalized to atomic sensitivity factors (ASF) of the PHI 5500 spectrometer.²⁵ The total intensity for each sample has been normalized to unity

Core-level	Normalized intensity (cps/ASF)
	C1s	P2p	S2p	C1s	Si2p3/2
Binding energy of components/eV	286.7	all	all	284.5	101.8
Assignments	PEO	LiPF6	LiTFSI	C-coating	Li2FeSiO4
ASF	0.314	0.525	0.717	0.314	0.368
Sample
LiTFSI lithiated	0.16	—	0.25	0.55	0.03
LiTFSI delithiated	0.39	—	0.47	0.14	0.01
LiPF6 lithiated	0.18	0.03	—	0.74	0.05
LiPF6 delithiated	0.22	0.06	—	0.69	0.04

---

The S2p and N1s features of the electrolyte (shown in Fig. 3i and 3j) are broadened and the intensities of the main lines are increased. The increased FWHM for both N1s and S2p suggests an inhomogeneous LiTFSI salt distribution, possible degradation products or a slight partial charging of the surface species.

These results clearly indicate an increased thickness in the surface layer, containing reaction products as well as unreacted electrolyte species on the deintercalated electrode. The intensities from features in the underlying electrode are therefore reduced. The amount of oxidized/reacted species, mainly decomposition products of the solvents, increases significantly compared to the intercalated electrode (see Table 2). The overall increase in electrolyte and decomposition products, i.e., LiTFSI, PEO compounds and Li₂CO₃, suggests a dynamic and possibly reversible SEI coverage of the electrode, dependant on the electrode potential.

LiPF₆

The situation using an LiPF₆-based electrolyte is somewhat different for both intercalated and deintercalated electrodes compared to that for LiTFSI. The C1s spectra of the intercalated and deintercalated electrodes (Fig. 3b) are basically the same, with insignificant changes in relative intensities. The major contribution to these spectra is the peak at 284.6 eV, assigned to the carbon-coating, carbon-black and EPDM binder, together with the hydrocarbon and EPDM binder peak at 285.2–285.3 eV. Similar peaks are found at 286.4–286.6 eV and 290.1–290.2 eV to those found for the LiTFSI electrodes; these originate from PEO-type polymer compounds as well as Li₂CO₃. However, the amounts of these compounds appear less here than on the LiTFSI electrodes. Meanwhile, the overall carbon intensity is about three times higher than for LiTFSI, indicating a larger carbon content at the surface and hence a lower SEI coverage. No increase is observed on deintercalation (Table 2).

The F1s spectra (Fig. 3d) reveal large amounts of LiF at 685.5–685.6 eV both for the intercalated and the deintercalated electrodes. The relative intensity of the LiF peak for the deintercalated electrode is roughly double that for the intercalated. The LiPF₆-salt contribution is found at binding energies of 687.7–687.9 eV; it is ∼1.2 eV lower than for the –CF₃ groups of LiTFSI. However, the quite broad peak widths (FWHM: 1.8 and 2.0 eV, respectively) are a further indication that the LiPF₆ salt is unstable during cycling, and has degraded to Li_xPF_y-type compounds. The same situation has been found in a previous study of LiFePO₄ with a 1M LiPF₆ EC:DEC (2 : 1) electrolyte.¹⁶ It can be noted that the amount of fluorine observed for LiPF₆ is 3–4 times lower than for LiTFSI. However, the amount of reacted LiF found for the LiTFSI system is only ∼10% of that observed for LiPF₆. This amount may be higher at the interface, but the signal is reduced due to the higher SEI coverage.

The peak contributing most to the O1s spectra both for the intercalated (531.2 eV) and deintercalated (531.6 eV) electrodes is assigned to Li₂CO₃, with probable contributions also from the silicate oxygen in Li₂FeSiO₄ (Fig. 3f). The peaks at higher binding energies are generally attributed to PEO-type polymer compounds and, to a lesser extent, possible oxygen-containing decomposition products of LiPF₆, e.g., Li_xPO_yF_z. The Li1s/Fe3p region (shown in Fig. 3h) has been treated in the same way as the LiTFSI sample. Two lithium emissions are observed (at 55.0 and 56.5 eV), corresponding to the Li₂FeSiO₄ electrode and surface phases, respectively. On deintercalation, the relative intensity of the peak at lower binding energy is decreased, as seen earlier for the LiTFSI salt.

Two distinct peaks appear in the P2p spectrum (DS: 0.85 eV) at 134.8 and 137.9 eV for the intercalated and at 134.7 and 137.2 eV for the deintercalated electrodes (Fig. 3k). The LiPF₆ salt, with a binding energy of 137.9 eV for the intercalated electrode, moves to lower binding energy (137.2 eV) and thus represents not only the LiPF₆ salt but also its degradation product Li_xPF_y (at 136.5 eV). It has been shown earlier¹⁶ that LiPF₆ can decompose in the electrolyte to LiF and Li_xPF_y according to the reactions:

LiPF₆ → LiF + PF₅ (2)

PF₅ + 2xLi⁺ + 2xe⁻ → Li_xPF_5−x + xLiF (3)

LiPF₆ can also undergo hydrolysis according to:

LiPF₆ + H₂O → LiF + POF₃ + 2HF (4)

The oxygenated salt-derived species formed here are best summarized as Li_xPO_yF_z, and are responsible for the 134.8 and 134.7 eV peaks in Fig. 3k, respectively.

XPS Si2p

It has been debated as to whether the LiPF₆ electrolyte will react and subsequently degrade the Li₂FeSiO₄ framework. As seen in eqn (3), LiPF₆ can hydrolyse to form HF, which is thought to induce degradation of Li₂FeSiO₄.

Fig. 3l and 3m show the Si2p emission (DS: 0.61 eV) for intercalated and deintercalated electrodes from cells with LiTFSI- and LiPF₆-based electrolytes. A dominant Si emission is seen at 101.7–101.9 eV for both electrolytes. For LITFSI, only a single emission can be found at 101.8 eV for the intercalated and deintercalated electrode (Fig. 3l). The Si2p emission reveals an asymmetric peak shape for LiPF₆, suggesting a second emission at higher binding energies. This weak emission at 102.7 eV (Fig. 3m) indicates the presence of Si–F groups at the electrode surface, as expected when Li₂FeSiO₄ is degraded on reaction with the HF formed by hydrolysis of LiPF₆.

However, the Si2p emission is weakened due to the carbon-coating on the Li₂FeSiO₄ particles, and is even less visible for LiTFSI, where the SEI coverage is greater (see Table 2). There are some indications of degradation in the Li₂FeSiO₄/LiPF₆ system. Studies continue of the SEI phase on uncoated electrodes to help clarify this picture.

Conclusions

Here, the surface layer of Li₂FeSiO₄ has been investigated thoroughly for two electrolyte systems, and a better understanding of the overall surface chemistry has emerged. The experimental picture of the SEI layers formed is summarized schematically in Fig. 4.

Fig. 4 Schematic representation of the SEI formed on Li_2−xFeSiO₄ cathode surfaces for (a) LiTFSI and (b) LiPF₆ salt-containing electrolytes.

For the LiTFSI-based electrolyte, a polymeric surface layer incorporating LiTFSI and TFSI⁻ salt anions is formed from the polymerized solvent species, along with lithium carbonates. The amount of carboxylic/carboxalate bonding is strongly enhanced, especially for the delithiated LiTFSI/LiFeSiO₄ system. This has not been observed previously for the EC/PC electrolyte system, and may indicate the formation of a dynamic and even reversible SEI coverage, dependant on the state of charge/potential of the electrode. Further studies are needed to validate this observation.

The LiPF₆-based electrolyte forms a thin surface layer with only small amounts of carbon-based surface reaction products, both in the lithiated and delithiated states. The main contribution to the C1s spectra comes from the electrode itself. The commonly observed instability of the LiPF₆ salt, resulting in formation of Li_xPF_y, Li_xPO_yF_z and LiF species, was further confirmed in this work. Lithium carbonate is found on all electrodes studied, suggesting that the lithium carbonate initially present on the Li₂FeSiO₄ surface is insoluble in the EC/DEC electrolyte, as in the case of the EC/PC electrolyte. There are also indications that Li₂FeSiO₄ is degraded by the HF formed in the hydrolysis of LiPF₆.

Our study demonstrates well the difference in surface chemistry for two electrolyte systems. The most striking features are the spectral changes in the C1s region between the discharged and charged state for the LiTFSI system, while virtually no changes are observed for the LiPF₆-based samples. On the other hand, evidence has been found to suggest electrolyte-induced formation of LiF and the degradation of the Li₂FeSiO₄ particle surface by HF for the LiPF₆-containing electrolyte.

Acknowledgements

This work has been supported by the Global Climate and Energy Project (GCEP) of Stanford University (USA), the Swedish Energy Agency (STEM), the Swedish Science Council (VR) and the Swedish Governmental Agency for Innovation Systems (Vinnova).

References

1. A. Nytén, M. Stjerndahl, H. Rensmo, H. Siegbahn, M. Armand, T. Gustafsson, K. Edström and J. O. Thomas, J. Mater. Chem., 2006, 16, 3483.
2. A. K. Padhi, K. S. Nanjundaswamy and J. B. Goodenough, J. Electrochem. Soc., 1997, 144, 1188.
3. A. Nytén, S. Kamali, L. Häggström, T. Gustafsson and J. O. Thomas, J. Mater. Chem., 2006, 16, 2266.
4. E. Peled, J. Electrochem. Soc., 1979, 126, 2047.
5. A. M. Andersson, A. Henningsson, H. Siegbahn, U. Jansson and K. Edström, J. Power Sources, 2003, 119–121, 522.
6. A. M. Andersson and K. Edström, J. Electrochem. Soc., 2001, 148, A1100.
7. A. M. Andersson, M. Herstedt, A. Bishop and K. Edström, Electrochim. Acta, 2002, 47, 1885.
8. K. Edström, M. Herstedt and D. P. Abraham, J. Power Sources, 2006, 153, 380.
9. D. Aurbach, K. Gamolsky, B. Markovsky, G. Salitra, Y. Gofer, U. Heider, R. Oesten and M. Schmidt, J. Electrochem. Soc., 2000, 147, 1322.
10. T. Eriksson, A. M. Andersson, A. G. Bishop, C. Gejke, T. Gustafsson and J. O. Thomas, J. Electrochem. Soc., 2002, 149, A69.
11. T. Eriksson, A. M. Andersson, C. Gejke, T. Gustafsson and J. O. Thomas, Langmuir, 2002, 18, 3609.
12. G. Pistoia, A. Antonini, R. Rosati and D. Zane, Electrochim. Acta, 1996, 41, 2683.
13. A. M. Andersson, D. P. Abraham, R. Haasch, S. MacLaren, J. Liu and K. Amine, J. Electrochem. Soc., 2002, 149, A1358.
14. C. H. Chen, J. Liu and K. Amine, J. Power Sources, 2001, 96, 321.
15. D. P. Abraham, J. Liu, C. H. Chen, Y. E. Hyung, M. Stoll, N. Elsen, S. MacLaren, R. Twesten, R. Haasch, E. Sammann, I. Petrov, K. Amine and G. Henriksen, J. Power Sources, 2003, 119–121, 511.
16. M. Herstedt, M. Stjerndahl, A. Nytén, T. Gustafsson, H. Rensmo, H. Siegbahn, N. Ravet, M. Armand, J. O. Thomas and K. Edström, Electrochem. Solid-State Lett., 2003, 6, A202.
17. M. Herstedt, A. M. Andersson, H. Rensmo, H. Siegbahn and K. Edström, Electrochim. Acta, 2004, 49, 4939.
18. E. Peled, D. Bar Tow, A. Merson, A. Gladkich, L. Burstein and D. Golodnitsky, J. Power Sources, 2001, 97–98, 52.
19. L. Vogdanis and W. Heitz, Macromol. Chem. Rapid Commun., 1986, 7, 543.
20. J.-C. Lee and M. Litt, Macromolecules, 2000, 33, 1618.
21. M. Herstedt, D. P. Abraham, J. B. Kerr and K. Edström, Electrochim. Acta, 2004, 49, 5097.
22. H. Ota, Y. Sakata, X. Wang, J. Sasahara and E. Yasukawa, J. Electrochem. Soc., 2004, 151, A437.
23. N. M. Brown, J. A. Hewitt and B. J. Meenan, Surf. Interface Anal., 1992, 18, 187.
24. L. J. Rendek, G. S. Chottiner and D. A. Scherson, J. Electrochem. Soc., 2002, 149, E408.
25. J. F. Moulder, W. Stickle, P. E. Sobol and K. D. BombenHandbook of X-Ray Photoelectron Spectra, Perkin-Elmer, 1995.

---