Fluoroethylene Carbonate as an Important Component in Electrolyte

Jun 2, 2014 - ABSTRACT: The effect of fluorinated ethylene carbonate (FEC) as a cosolvent in alkyl carbonates/LiPF6 on the cycling performance of high...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/Langmuir

Fluoroethylene Carbonate as an Important Component in Electrolyte Solutions for High-Voltage Lithium Batteries: Role of Surface Chemistry on the Cathode Elena Markevich,*,† Gregory Salitra,*,† Katia Fridman,† Ronit Sharabi,† Gregory Gershinsky,† Arnd Garsuch,‡ Guenter Semrau,‡ Michael A. Schmidt,‡ and Doron Aurbach† †

Department of Chemistry Bar-Ilan University, Ramat Gan 52900 Israel BASF SE, Ludwigshafen 67056, Germany



S Supporting Information *

ABSTRACT: The effect of fluorinated ethylene carbonate (FEC) as a cosolvent in alkyl carbonates/LiPF6 on the cycling performance of highvoltage (5 V) cathodes for Li-ion batteries was investigated using electrochemical tools, X-ray photoelectron spectroscopy (XPS), and highresolution scanning electron microscopy (HRSEM). An excellent cycling stability of LiCoPO4/Li, LiNi0.5Mn1.5O4/Si, and LiCoPO4/Si cells and a reasonable cycling of LiCoPO4/Si cells was achieved by replacing the commonly used cosolvent ethylene carbonate (EC) by FEC in electrolyte solutions for high-voltage Li-ion batteries. The roles of FEC in the improvement of the cycling performance of high-voltage Li-ion cells and of surface chemistry on the cathode are discussed.



INTRODUCTION Fluoroethylene carbonate (FEC) has been much investigated for many applications in Li-ion batteries.1−8 It was used both as a cosolvent and a solid electrolyte interphase (SEI) formation additive for graphite,1,3,9−11 silicon,12−19 and other anodes,18,20−22 and it has demonstrated a variety of benefits in terms of the cycling performance and effective passivation of Li battery anodes. Much less attention has been given to the investigation of the effect of FEC on the performance of cathodes in Li-ion batteries.3,23−25 Recently, we reported a drastic improvement in the cycling stability of LiCoPO4 cathodes in electrolyte solutions containing FEC, instead of the commonly used ethylene carbonate (EC), as a cosolvent with DMC as a mandatory component in solutions for Li-ion batteries.26 LiCoPO4, with an olivine structure, possesses a high operating voltage (redox potential of 4.8 V vs Li/Li+), a flat voltage profile, and a high theoretical capacity of about 170 mAh/g.27 However, as we showed previously,28 LiCoPO4 is a very unstable cathode material in LiPF6-containing electrolyte solutions. A delithiated LixCoPO4 phase (x ≪ 1) repeatedly formed during charge−discharge cycling, resulting in Li-ion cells with an extremely low chemical stability, which is due to the fact that Co3+, unlike other olivines, in CoPO4 exists in a high-spin configuration.29 This delithiated (charged) state is prone to a nucleophilic attack of F− anions on the P atoms, leading to the degradation of the material. The use of FEC as a cosolvent in electrolyte solutions leads to a marked decrease in capacity fading of LiCoPO4 cathodes.26 In addition, we also communicated6 on the excellent performance of full Li-ion cells composed of another high-voltage (5 V) © 2014 American Chemical Society

cathode material, LiNi0.5Mn1.5O4 spinel, and a columnar silicon film anode when using FEC-containing electrolyte solutions. The performance of these cells with FEC-containing solutions was superior to that of similar cells containing EC-based electrolyte solutions. In this article, we present the results of a comparative study of EC- and FEC-based electrolyte solutions for high-voltage Liion systems and discuss the role of FEC in improving the cycling stability of high-voltage cathodes.



EXPERIMENTAL SECTION

Carbon-coated nanopowders of LiCoPO4 were prepared by hydrothermal synthesis as described previously.30 The carbon content in the samples was determined by an Eager, Inc. model 200C H, N, S analyzer and comprised 1.53 wt %. The surface area of the sample, equal to 11.7m2/g, was calculated using the Brunauer−Emmett−Teller (BET) model from the adsorption isotherms of N2 gas at 77 K using an Autosorb-1-MP apparatus (Quantachrome Corporation). LiNi0.5Mn1.5O4 powder was obtained from BASF SE. Composite LiCoPO4 electrodes comprised 80 wt % active material, 15 wt % carbon black (SuperP, Superior graphite, USA), and 5 wt % PVDF (Aldrich). LiNi0.5Mn1.5O4 composite electrodes comprised 90 wt % active material, 5 wt % carbon black, and 5 wt % PVDF. The cathode sheets were fabricated by spreading a slurry (a suspension of active cathode material powder and carbon black in PVDF/N-methylpyrrolidon) on aluminum foil current collectors with a doctor blade device. Typically, LiCoPO4 electrodes contained 2 to 3 Received: April 9, 2014 Revised: May 25, 2014 Published: June 2, 2014 7414

dx.doi.org/10.1021/la501368y | Langmuir 2014, 30, 7414−7424

Langmuir

Article

Figure 1. Curves of discharge capacity (a) and irreversible capacity (b) vs cycle number obtained upon galvanostatic cycling between 3.5 and 5.2 V at a rate of C/5 h for LiCoPO4 electrodes in EC-based (hollow dots) and FEC-based (full dots) electrolyte solutions (30 °C). mg/cm2 of active mass, and LiNi0.5Mn1.5O4 electrodes contained 3 to 4 mg/cm2 of active mass. Silicon thin-film electrodes were prepared by dc magnetron sputtering.19,31 The surface density of the a-Si film was around 1.3 mg/cm2 (∼6 μm thick). The electrolyte solution was 1 M LiPF6 in EC/DMC 1:1 (ECbased) or FEC/DMC 1:4 (FEC-based) (both Li-battery grade from Merck, KGaA; solvent ratios are given by weight). Two-electrode cells comprising a LiCoPO4 or LiNi0.5Mn1.5O4 cathode, a PE separator (Setela Tonen, Japan), electrolyte solution, and a Li foil or a-Si film negative electrode were assembled in a glovebox filled with pure argon and sealed in coin-cells (2523, NRC, Canada). Galvanostatic cycling of these cells was carried out using a computerized multichannel analyzer from Arbin Instruments (USA), model BT2000. LiCoPO4/Li and LiCoPO4/Si cells were cycled with a current rate of C/5, and LiNi0.5Mn1.5O4/Li and LiNi0.5Mn1.5O4/Si cells were cycled with a current rate of C/8. The potential range was 3.5−5.2 V in all cases. The constant current−constant voltage (CCCV) procedure included charge step performed at a current rate of C/5 to 5.2 V followed by a potentiostatic step at 5.2 V for 1 or 5 h, as indicated, and discharge to 3.5 V (C/5). All cycling tests were performed at 30 °C. Before the use of the Si film electrodes as anodes in full cells, they were galvanostatically prepassivated and prelithiated in two-electrode coin-type cells containing Li counter electrodes. The cells comprising silicon electrodes and Li counter electrodes were cycled galvanostatically with voltage cutoff limits of 10 mV and 1.2 V and a current density of 120 mA/g in the first cycle and 600 mA/g in four subsequent cycles. Finally, the Si electrodes were discharged galvanostatically to 10 mV versus Li/Li+ and then potentiostatically at 10 mV versus Li/Li+ for 24 h, and then they were withdrawn from

the Si/Li cells and used for the preparation of the complete cells with a LiNi0.5Mn1.5O4 or LiCoPO4 cathode. XPS measurements were carried out with an AXIS-HS system (Kratos Analytical, Inc., England) using monochromatic Al Kα radiation. For these measurements, cycled electrodes taken from dismantled cells were thoroughly washed with DMC, vacuum-dried, and then transferred to the spectrometer using a homemade transfer system containing a magnetic manipulator and a gate valve. All binding energies (BE) were corrected with respect to the BE value of C 1s at 285 eV. HRSEM images were obtained using a FEI xHR-SEM Magellan 400L microscope.



RESULTS AND DISCUSSION Figure 1 shows the cycling results of LiCoPO4/Li cells with two electrolyte solutions (EC- and FEC-based). For the cells cycled with the FEC-based electrolyte solution, significantly better capacity retention and lower irreversible capacity were observed compared to those of cells cycled with the EC-based solution. The anodic stability limits do not differ markedly for two electrolyte solutions. (see Supporting Information, Figure S1). The FEC-based solution possesses slightly higher anodic stability compared to that of the EC-based electrolyte, suggesting that cyclic carbonates govern the anodic limit of the electrolyte solutions, in accordance with ref 25. As we have previously reported,28,32 the main reason for the capacity fading of LiCoPO4 cathodes is their instability toward the attack of HF on the phosphate group of the delithiated (charged) form of the material. 7415

dx.doi.org/10.1021/la501368y | Langmuir 2014, 30, 7414−7424

Langmuir

Article

Figure 2. SEM images of the pristine LiCoPO4 electrode (a, b) and electrodes cycled in the FEC-based (c, d) and EC-based (e, f) electrolyte solutions.

We showed that a nucleophilic attack of F− anions on the P atoms of the olivine compound in the delithiated state leads to the breakdown of the P−O bonds of the phosphate anions with the formation of LiPO2F2, which is soluble in the electrolyte solution:

PO4 3 − + HF + H+ ↔ PO3F2 − + H 2O PO3F2 − + HF + H+ ↔ PO2 F−2 + H 2O PO2 F−2 + HF + H+ ↔ POF3 + H 2O 7416

(1)

dx.doi.org/10.1021/la501368y | Langmuir 2014, 30, 7414−7424

Langmuir

Article

Figure 3. 19F NMR spectra of the 1 M LiPF6 in EC/DMC 1:1 (a) and 1 M LiPF6 in FEC/DMC 1:4 (b) electrolyte solutions after the addition of 2000 ppm of water. Kinetics of the accumulation of HF and LiPO2F2 (c). Decrease in the content of LiPF6 (d) after the addition of 2000 ppm of water in the two electrolyte solutions (1 M LiPF6 in EC/DMC 1:1 and 1 M LiPF6 in FEC/DMC 1:4), as indicated.

This process leads to the degradation of the active material near its surface, which is accompanied by dissolution of the product in the electrolyte solution and, as we showed previously,31 by structural degradation of olivine and Li depletion. HRSEM images of LiCoPO4 electrodes cycled in EC- and FEC-based electrolyte solutions provide clear visual evidence for the mechanism of their capacity fading. In Figure 2a,b, one can see the images of a pristine LiCoPO4 composite electrode. Figure 2c,d shows images of electrodes cycled in an FEC-based electrolyte solution. Both sets of images reflect a similar morphology of the electrodes’ particles, which indicates that the particles keep their shape and structure when the electrodes are cycled in the FEC solutions. To the contrary, the electrodes cycled in the EC-based solution exhibit severe structural degradation (Figure 2e,f). The particles are strongly eaten away by the electrolyte solution. Moreover, from the general view of the electrode’s surface (Figure 2f), one cannot observe any intact particles. Thus, the cathode active material dissolves in the EC-based electrolyte solution. It is important to note that the destructive processes (1) involve a chain reaction. Water is formed in these processes and reacts further with new portions of PF6− anions to give HF:

Thus, one of the main factors that controls the rate of the degradation of the active material is the rate of reaction of water with PF6−, which produces HF. The second decisive factor that determines the stability of LiCoPO4 cathode in the given electrolyte solution is the ability of the components of the solution to form effective protective films on the LiCoPO4 particles’ surface, which can prevent direct contact between the active mass and nucleophilic F− ions in solution. In order to distinguish between these two possible reasons for the marked improvement in the cycling behavior of LiCoPO4 cathodes in the FEC-based electrolyte solutions, we launched a series of experiments. In our previous work,15 we reported the results of direct 19F NMR measurements of the amount of HF that is formed in both electrolyte solutions after the addition of water on the basis of the well-known reaction of PF6− anions with water:28,33 LiPF6 + 2H 2O → LiPO2 F2 + 4HF

We added 2000 ppm of water to both EC- and FEC-based electrolyte solutions under an argon atmosphere, transferred them immediately to Teflon NMR probe tubes, sealed the tubes, and measured 19F NMR spectra after 30 min or after 1, 3, and 7 days (Figure 3a,b). Figure 3c,d presents kinetics of the accumulation of HF and LiPO2F2 and the decrease in the content of LiPF6. The concentrations of all F-containing species were calculated by the integration of the area of the corresponding peaks in the 19F NMR spectra by taking into consideration a starting concentration of LiPF6 equal to 1 M. It can be seen that water reacts faster with LiPF6 and produces

PF−6 + H 2O ↔ F− + POF3 + 2HF POF3 + H 2O ↔ PO2 F−2 + HF + H+ PO2 F−2 + H 2O ↔ PO3F2 − + HF + H+

(3)

(2) 7417

dx.doi.org/10.1021/la501368y | Langmuir 2014, 30, 7414−7424

Langmuir

Article

Figure 4. Galvanostatic cycling of LiCoPO4/Li cells in a 1 M LiPF6 in EC/DMC 1:1 electrolyte solution with a potentiostatic step at different potentials performed for 24 h during the second charge (a) and with a potentiostatic step at 5.2 V (CCCV cycling procedure) for 1 or 5 h after every galvanostatic charge (b), as indicated (C/5; 30 °C).

more HF in the FEC-based electrolyte solution. Thus, according to the chain decomposition process (1 and 2), which includes the step of HF formation by the interaction of water and LiPF6, one can conclude that the marked improvement of the cycling performance of LiCoPO4/Li cells in the FEC-based electrolyte does not result from the lower content of HF in this electrolyte solution. It is likely that FEC participates in the formation of protective surface films on these cathodes. In order to understand why a drastic difference in the performance of LiCoPO4/Li cells was observed in the two electrolyte solutions that differ from each other only by the nature of the cyclic organic carbonate, namely, EC or FEC, we studied the effect of the cycling procedure on the capacity retention of LiCoPO4/Li cells in both electrolyte solutions. We found that in the EC-based electrolyte solution there is also a passivation process that may be able to protect LiCoPO4 from fast degradation. Such a process can be observed at high potentials and results from electrochemical reactions of the components of the electrolyte solution. This phenomenon is demonstrated in Figure 4. This figure compares the cycling results obtained in the standard cycling procedure (charge and discharge with a current rate of C/5) with results obtained for the same cycling procedure in which the second charge potentiostatic steps of different potentials were added during 24

h. It can be seen that potentiostatic steps in the potential range 4−4.4 V do not affect the results. At higher potentials (4.8−5.0 V), a slight improvement in the capacity fading was observed, and applying a 5.1−5.2 V potentiostatic step leads to a pronounced improvement in the cycling stability of LiCoPO4 electrodes in the EC-based solution. This experiment proves that in the standard EC-based electrolyte solution the formation of a protective surface film on the surface of the cathode is possible and starts at about 5.1 V versus Li/Li+. Figure 4b exhibits another cycling procedure in which the potentiostatic steps were performed in every cycle after every galvanostatic charge at 5.2 V for 1 or 5 h (CCCV). The longer the potentiostatic steps, the better the capacity retention that we observed. It is remarkable that one potentiostatic step of 24 h performed only in the second cycle was more effective than potentiostatic steps in every cycle for 5 h. This means that the protective surface films formed in the EC-based solutions at high potentials are reasonably stable (once formed, they protect the electrodes for prolonged cycling), but they are formed very slowly. These observations are in line with the results of the analysis of the electrolyte solutions rinsed from the cycled LiCoPO4/Li cells and after prolonged storage tests performed using 19F NMR spectroscopy. The highest content of LiPO2F2 in the electrolyte solution was observed after a standard cycling 7418

dx.doi.org/10.1021/la501368y | Langmuir 2014, 30, 7414−7424

Langmuir

Article

procedure (about 17% of P atoms were in the form of LiPO2F2). The prolonged storage tests (2 weeks) of the LiCoPO4/Li cells at OCV did not reveal LiPO2F2, suggesting that the lithiated (discharged) form of LiCoPO4 is reasonably stable in the HF-containing electrolyte solution. Finally, the storage tests at 5 V for the same period resulted in the formation of a smaller amount of difluorophosphate species than that measured after repeated galvanostating cycling (about 8% of P atoms were in the form of LiPO2F2). These results show that LiCoPO4 is stable in the lithiated form and degrades in the dilithiated state and that prolonged polarization at high voltage leads to the formation of the protective surface films. In turn, the application of both cycling protocols (as described above) to Li/LiCoPO4 cells containing the FECbased solution has a detrimental affect on the performance of the cells (Figure 5). It is clear that at high potentials two

Figure 6. C 1s XPS spectra of LiCoPO4 electrodes: (a) pristine LiCoPO4 electrodes (magenta curve) and LiCoPO4 after 100 cycles in the FEC-based (red curve) and EC-based (black curve) electrolyte solutions. (b) Discharged LiCoPO4 electrodes after potentiostatic step at different potentials, as indicated, performed during 24 h.

Figure 5. Galvanostatic cycling of LiCoPO4/Li cells in a 1 M LiPF6 in FEC/DMC 1:4 electrolyte solution with a potentiostatic step at different potentials performed for 24 h during the second charge and with a potentiostatic step at 5.2 V (CCCV cycling procedure) for 5 h after every galvanostatic charge, as indicated (C/5; 30 °C).

pristine electrodes. In the case of the EC-based electrolyte, this observation is related to the process of the removal of the surface conducting carbon layer, which ensures the electronic conductivity of the electrodes. This phenomenon is likely attributed to the exfoliation of the carbon layer because of the dissolution of the underlying LiCoPO4. Indeed, carbon particles were identified (using micro-Raman spectroscopy) on the PE separators of Li/LiCoPO4 cells that were cycled in EC-based solutions.34 The exfoliation of the conducting carbon coating layer is by itself an additional factor leading to the failure of LiCoPO4 electrodes. A comparison between the XPS spectra of LiCoPO4 electrodes cycled in the EC-based (black curves) and FEC-based (red curves) solutions shows, in the case of the former, that along with a decrease in the intensity of the peak at 285 eV in the C 1s spectra, an increase in the intensity of the Co 2p, O 1s, and P 2p peaks is observed (Figure 7). Thus, the process of degradation of the carbon surface layer is also reflected by these variations in the relative intensities of the XPS signals related to the amorphous carbon surface layer and the elements that belong to the underlying active material. This result is in line with the HRSEM observations (Figure 2e,f), which demonstrates that the process of the dissolution of the active material occurs from inside the LiCoPO4 rods and leads to their transformation to fragments of tube-like particles with a free carbon layer newly formed at the inner surface. At the same time, the higher signal of the F atoms in the F 1s spectrum of the electrodes cycled in the FEC-based solution suggests a higher content of F-containing species in the surface

concurrent processes occur, namely, a negative process of degradation and a positive process of protective surface film formation. The formation of the protective surface films in the FEC-based solution is very fast, so at a current rate of C/5, the effective protective layer has a chance to be formed even in the first cycle. Then, applying the potentiostatic steps at high potentials has a detrimental affect, as it only increases the degradation process of LiCoPO4. In the case of the FEC-based electrolyte solution, the positive process is fast and the electrodes have time for passivation during the standard cycling procedure without the need for potentiostatic steps. In the EC-based electrolyte solution, the formation of the protective surface films occurs at a much slower rate than that in the FEC-based solution; thus, the delithiated LixCoPO4 moieties, which are sensitive to attack by F− anions in solution, have no time to become effectively passivated. Surface characterization of LiCoPO4 electrodes (containing PVDF and carbon) cycled for equal periods of time in the ECand FEC-based electrolyte solutions was carried out by XPS. The XPS spectra of cycled and pristine LiCoPO4 electrodes are shown in Figures 6a (C 1s) and 7. The intensity of the peak of the amorphous carbon at 285 eV in the C 1s spectra (Figure 6a) decreases substantially for the electrodes cycled in the EC-based electrolyte solution compared to those cycled in the FEC-based solution and 7419

dx.doi.org/10.1021/la501368y | Langmuir 2014, 30, 7414−7424

Langmuir

Article

Figure 7. Co 2p, F 1s, O 1s, and P 2p XPS spectra of pristine LiCoPO4 electrodes (magenta curves) and LiCoPO4 electrodes cycled in the EC-based (black curves) and FEC-based (red curves) electrolyte solutions.

film in this case (Figure 7b). These components relate both to the decomposition products of LiPF6 salt, namely, PFx and POxFy species, and C−F containing organic products of FEC transformation (a shoulder at about 686 eV; red curve35,36).

Analysis of the C 1s spectra enables one to estimate relative content of carbonates and other products of decomposition of organic carbonate components in electrolyte solutions. It can be seen that the ratio of 7420

the the the the

dx.doi.org/10.1021/la501368y | Langmuir 2014, 30, 7414−7424

Langmuir

Article

intensities for the two main peaks observed in the C 1s spectra of the cycled electrodes at about 290 and 287 eV is different. The first peak is composed of the signals of carbon atoms in a three-oxygen environment (CO3-like), which is related to inorganic carbonates MCO3, ROCO2M, ROCO2R (dialkylcarbonates), or −(OCO2R)n− (polycarbonates), as well as of carbon atoms in −CF2− chains of PVDF (Me = Li, Co). The second one consists of the signals of CO-like carbon atoms of the organic carbonate compounds at 287.5 eV as well as the signals related to the poly(ethylene oxide) (PEO)-like polymers. As was shown previously,37 the estimation of the relative intensities of these two peaks is indicative of the relative content of carbonate species in the surface films. Equal intensities of these peaks are a characteristic of a ROCO2M spectrum or of a mixture of equal amounts of MCO3 and dialkylcarbonates (or an equivalent amount of polycarbonates). At the same time, the predominance of CO3-like carbon atoms points to the higher content of MCO3, which does not contain CO-like carbon atoms. To the contrary, the marked predominance of the second peak (at about 287.5 eV) is indicative of the presence of dialkyl- or polycarbonates that contain 2 CO-like carbon atoms per each CO3-like one. In our case, it can be clearly seen that electrodes cycled in the FECbased electrolyte solutions (red curve) exhibit a predominant signal related to CO3-like carbon atoms compared to the signal associated with CO-like carbon atoms. For the electrodes cycled in the EC-based electrolyte solution (black curve), the opposite relationship between the intensities of these two peaks is observed. In addition, for the FEC-based solutions, the maximum of the peak observed in the vicinity of 287.5 eV is shifted to a lower BE of about 286.7 eV, suggesting that PEOlike polymer species are present in a higher content in surface films formed on the cathodes in the FEC solutions.38−40 Thus, it may be inferred that the surface films that are formed in the FEC-based electrolyte solution contain more PEO-like polymer species and MCO3 compared to those formed in the EC-based electrolyte solution. In turn, the surface films formed in the ECbased electrolyte solution seem to have a higher content of polycarbonate species. The O 1s spectra (Figure 6c) of these electrodes correlate well with the above suggestion. It can be seen that the relative content of OCO oxygen atoms at 534.3 eV41 is markedly higher in the case of the EC-based electrolyte solution. In the P 2p spectra of pristine electrodes and electrodes cycled in the FEC-based electrolyte solution, the peak of the phosphate groups (134.4 eV) related to the unchanged olivine structure42−44 dominates the spectra. For the electrodes cycled in the EC-based solution, POxFy components prevail (135.3 eV),44 suggesting the almost total surface degradation of the olivine phosphate structure as a result of the nucleophilic attack of the F− anions on the P atoms. We also analyzed the surface films that were formed slowly on the surface of LiCoPO4 cathodes during potentiostatic steps and that protected the electrodes from very fast degradation in the EC-based electrolyte solution, as shown in Figure 4a. The XPS spectra were collected from LiCoPO4 electrodes that were polarized for 24 h at three different potentials, namely, 4.4, 4.9, and 5.2 V, as indicated, and then fully discharged at 3.5 V in LiCoPO4/Li cells (Figure 6b). It can be seen that for the electrodes polarized up to 4.4 V the XPS spectra are very similar to that of pristine, in good agreement with the results of galvanostatic cycling of such electrodes. This observation is in full agreement with the results reported by Edstrom et al.41

They showed that no solvent reaction or decomposition products (e.g., polycarbonates, semicarbonates, and Li2CO3) were detected on the carbon-coated LiFePO4 surface after cycling. This finding is expected, as LiFePO4 cathodes work at relatively low potentials, with a maximal high cutoff potential below 4 V. Electrodes polarized to 4.9 V undergo only slow degradation, and their spectra are supposed to reflect the formation of surface films comprising carbonate and PEO-like species. The intensity of the relevant XPS peaks related to these surface moieties are indeed maximal in the spectra related to electrodes polarized in the EC-based solutions at 5.2 V. The intensity of the peaks related to the surface films formed on the electrodes polarized to 5.2 V correlates with their better cycling stability in EC-based solutions. The main difference between the surface films that were formed on the electrodes cycled in a standard galvanostatic procedure without a potentiostatic polarization (Figure 6a, black curve) and those whose working protocol included a potentiostatic step at high potentials (Figure 6b) is a higher content of PEO-like polymers compared to that of organic carbonate/polycarbonate components in the surface films of the latter ones, as reflected by the XPS data. Obviously, a high content of organic carbonate/polycarbonate species is a detrimental factor for the effective passivation of the high-voltage cathodes. If our conclusion about the better passivation of the surface of the high-voltage LiCoPO4 cathode in the FEC-based electrolyte solution is correct, then this phenomenon would be reflected by a better performance of other high-voltage cathodes in the FEC-based solutions compared to that in ECbased ones. For verification, we compared the performance of one more 5 V cathode material, LiNi0.5Mn1.5O4 spinel, in both electrolyte solutions. Figure 8a shows the results of the galvanostatic tests of LiNi0.5Mn1.5O4/Li cells performed with the EC- and FEC-based electrolyte solutions. The full capacity of Si anodes and LiNi0.5Mn1.5O4 cathodes comprised 6 mAh and 0.66 mAh, respectively. It can be seen that in both cases the cells demonstrate a very stable cycling behavior and that advantages of one solution composition over the other are not evident. This finding is not surprising, as LiNi0.5Mn1.5O4 spinel is a high-voltage cathode material that is much more stable in a LiPF6-containing electrolyte solution compared to its stability in LiCoPO4. However, for full LiNi0.5Mn1.5O4/Si cells, as described in ref 6, a marked improvement in the performance in the FEC-based electrolyte solution over that in the EC-based one was observed. Typical results for these two systems are shown in Figure 8b. In the EC-based solution, these full cells can withstand only 20−30 charge−discharge cycles. After that, a sudden increase in the irreversible capacity coupled with a drastic growth of the charge capacity value and a decrease in the discharge capacity are always observed. This behavior is typical for the so-called “shuttle” mechanism when the species that are formed on the anode as a result of side parasitic reduction processes are transferred to the cathode side and are oxidized on the cathode during charging.45−48 Obviously, after 20−30 cycles of stable cycling in the EC-based electrolyte, the rate of side parasitic reactions on the electrodes becomes comparable or even higher than that of the Faradaic process related to Liion intercalation, and, finally, this situation leads to the failure of the cells. As we noted in the Experimental Section, Si anodes were galvanostatically prepassivated and partially prelithiated in Si/Li half cells containing Li counter electrodes before their use as anodes for the full cells. This pretreatment procedure ensures a 7421

dx.doi.org/10.1021/la501368y | Langmuir 2014, 30, 7414−7424

Langmuir

Article

Figure 9. Discharge capacity vs cycle number obtained upon galvanostatic cycling of LiCoPO4/Si cells in a 1 M LiPF6 FEC/ DMC 1:4 electrolyte solution (30 °C).

cathodes, which are much more effective and are formed much faster than the surface films formed in EC-based electrolyte solutions. It is remarkable that these conclusions were made on the basis of the results of the electrochemical performance of LiCoPO4, a high-voltage cathode material that exhibits severe degradation in LiPF6-containing electrolyte solutions; thus, this approach offers a simple model system for the fast estimation of the effectiveness of different electrolyte solutions, additives, and cycling protocols for the operation of 5 V cathodes. The results of XPS analysis of LiCoPO4 cathodes suggest that the surface films that are formed in the FEC-based electrolyte solutions contained more PEO-like polymer and inorganic carbonate species compared to those formed in the EC-based electrolyte solution, where a higher content of polycarbonate species is observed. The cycling procedure highly influences the capacity retention of the LiCoPO4 cathodes. With the EC-based electrolyte solutions, a potentiostatic step at 5.1−5.2 V performed in the initial period of cycling leads to a pronounced improvement in the cycling stability of LiCoPO4 electrodes. Obviously, in this electrolyte solution at high potentials, the process of passivating the surface of high-voltage cathodes occurs because of the formation of protective surface films as the result of electrochemical transformations of the components of the electrolyte solution. This protective surface film is reasonably stable, but the process of its formation in the ECbased electrolyte solution occurs much more slowly than that in the FEC-based electrolyte. The superiority of FEC-based electrolyte solutions over conventional EC-based solutions for the performance of highvoltage cathodes is evident from the cycling results obtained with full cells. Excellent performance of full cells comprising amorphous silicon film anodes and high-voltage LiNi0.5Mn1.5O4 spinel cathodes was achieved by replacing of the EC-based electrolyte solution by the FEC-based one. Additionally, a reasonable cycling performance of LiCoPO4/Si full cells was demonstrated for the first time with the FEC-based electrolyte solution.

Figure 8. Curves of charge (full dots) and discharge (hollow dots) capacity vs cycle number obtained upon galvanostatic cycling of LiNi0.5Mn1.5O4/Li (a) and LiNi0.5Mn1.5O4/Si (b) cells at a current rate of C/8 in the EC-based and FEC-based electrolyte solutions, as indicated (30 °C).

reasonable passivation of Si anodes in both electrolyte solutions, and Si anodes with a surface density of 1.3 mg/cm demonstrated at least several hundred stable cycles of Si/Li half cells even in the nonoptimal EC-based electrolyte solution.15 Thus, the cathode side is obviously responsible for the fast failure of the full LiNi0.5Mn1.5O4/Si cells in the EC-based electrolyte solution. It is clear that good passivation of both the cathode and anode surfaces would prevent this failure scenario. Indeed, when the cells were cycled in the FEC-based electrolyte, they demonstrated very stable cycling, with a charge−discharge efficiency approaching 100%. Moreover, we succeeded in running LiCoPO4/Si full cells (the capacity of LiCoPO4 cathodes comprised 0.33 mAh) that demonstrated a reasonable cycling performance for more than a hundred cycles in the FEC-based electrolyte solution (Figure 9). To the best of our knowledge, this is the first example of the stable cycling of full high-voltage cells containing undoped LiCoPO4 cathodes, which are usually very unstable even during cycling, in half cells with Li counter electrodes in conventional Li-ion battery (EC-based) electrolyte solutions.



CONCLUSIONS FEC-based electrolyte solutions result in dramatically improved cycling behavior of LiCoPO4/Li cells compared to that with standard EC-based electrolyte solutions. The remarkable improvement in the cycling of LiCoPO4/Li cells with the FEC-based solutions does not result from the lower content of HF in these electrolyte solutions. FEC participates in the formation of protective surface films on these high-voltage



ASSOCIATED CONTENT

S Supporting Information *

Electrochemical windows of EC- and FEC-based electrolyte solutions. This material is available free of charge via the Internet at http://pubs.acs.org. 7422

dx.doi.org/10.1021/la501368y | Langmuir 2014, 30, 7414−7424

Langmuir



Article

Roué, L.; Lestriez, B. New insights into the silicon-based electrode’s irreversibility along cycle life through simple gravimetric method. J. Power Sources 2012, 220, 180−184. (17) Nakai, H.; Kubota, T.; Kita, A.; Kawashima, A. Investigation of the solid electrolyte interphase formed by fluoroethylene carbonate on Si electrodes. J. Electrochem. Soc. 2011, 158, A798−A801. (18) Chockla, A. M.; Klavetter, K. C.; Mullins, C. B.; Korgel, B. A. Tin-seeded silicon nanowires for high capacity Li-ion batteries. Chem. Mater. 2012, 24, 3738−3745. (19) Elazari, R.; Salitra, G.; Gershinsky, G.; Garsuch, A.; Panchenko, A.; Aurbach, D. Li ion cells comprising lithiated columnar silicon film anodes, TiS2 cathodes and fluoroethyene carbonate (FEC) as a critically important component. J. Electrochem. Soc. 2012, 159, A1440− A1445. (20) Chockla, A.; Klavetter, K.; Mullins, C.; Korgel, B. Solutiongrown germanium nanowire anodes for lithium-ion batteries. ACS Appl. Mater. Interfaces 2012, 4, 4658−4664. (21) Wilhelm, H.; Marino, C.; Darwiche, A.; Monconduit, L.; Lestriez, B. Significant electrochemical performance improvement of TiSnSb as anode material for Li-ion batteries with composite electrode formulation and the use of VC and FEC electrolyte additives. Electrochem. Commun. 2012, 24, 89−92. (22) Klavetter, K.; Wood, S.; Lin, Y.; Snider, J.; Davy, N.; Chockla, A.; Romanovicz, D.; Korgel, B.; Lee, J.; Heller, A.; Mullins, C. A highrate germanium-particle slurry cast Li-ion anode with high Coulombic efficiency and long cycle life. J. Power Sources 2013, 238, 123−136. (23) Wu, B.; Ren, Y.; Mu, D.; Liu, X.; Zhao, J.; Wu, F. Enhanced electrochemical performance of LiFePO4 cathode with the addition of fluoroethylene carbonate in electrolyte. J. Solid State Electrochem. 2013, 17, 811−816. (24) Liao, L.; Cheng, X.; Ma, Y.; Zuo, P.; Fang, W.; Yin, G.; Gao, Y. Fluoroethylene carbonate as electrolyte additive to improve low temperature performance of LiFePO4 electrode. Electrochim. Acta 2013, 87, 466−472. (25) Hu, L.; Zhang, Z.; Amine, K. Fluorinated electrolytes for Li-ion battery: an FEC-based electrolyte for high voltage LiNi0.5Mn1.5O4/ graphite couple. Electrochem. Commun. 2013, 35, 76−79. (26) Sharabi, R.; Markevich, E.; Fridman, K.; Gershinsky, G.; Salitra, G.; Aurbach, D.; Semrau, G.; Schmidt, M.A.; Schall, N.; Bruenig, C. Electrolyte solution for the improved cycling performance of LiCoPO4/C composite cathodes. Electrochem. Commun. 2013, 28, 20−23. (27) Padhi, K.; Nanjundaswamy, K.S.; Goodenough, J.B. Phosphoolivines as positive-electrode materials for rechargeable lithium batteries. J. Electrochem. Soc. 1997, 144, 1188−1194. (28) Markevich, E.; Sharabi, R.; Gottlieb, H.; Borgel, V.; Fridman, K.; Salitra, G.; Aurbach, D.; Semrau, G.; Schmidt, M.A.; Schall, N.; Bruenig, C. Reasons for capacity fading of LiCoPO4 cathodes in LiPF6 containing electrolyte solutions. Electrochem. Commun. 2012, 15, 22− 25. (29) Ehrenberg, H.; Bramnik, N.N.; Senyshyn, A.; Fuess, H. Crystal and magnetic structures of electrochemically delithiated Li1‑xCoPO4 phases. Solid State Sci. 2009, 11, 18−23. (30) Nuspl, G.; Wimmer, L.; Eisgruber, M. U.S. Patent 2007/ 0054187 A1. (31) Elazari, R.; Salitra, G.; Gershinsky, G.; Garsuch, A.; Panchenko, A.; Aurbach, D. Rechargeable lithiated silicon-sulfur (SLS) battery prototypes. Electrochem. Commun. 2012, 14, 21−24. (32) Sharabi, R.; Markevich, E.; Borgel, V.; Salitra, G.; Aurbach, D.; Semrau, G.; Schmidt, M.A.; Schall, N.; Stinner, C. Significantly improved cycling performance of LiCoPO4 cathodes. Electrochem. Commun. 2011, 13, 800−802. (33) Plakhotnyk, A.V.; Ernst, L.; Schmutzler, R. Hydrolysis in the system LiPF6-propylene carbonate-dimethyl carbonate-H2O. J. Fluorine Chem. 2005, 126, 27−31. (34) Sharabi, R.; Markevich, E.; Borgel, V.; Salitra, G.; Gershinsky, G.; Aurbach, D.; Semrau, G.; Schmidt, M.A.; Schall, N.; Stinner, C. Raman study of structural stability of LiCoPO4 cathodes in LiPF6 containing electrolytes. J. Power Sources 2012, 203, 109−114.

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



REFERENCES

(1) Jeong, S.; Inaba, M.; Mogi, R.; Iriyama, Y.; Abe, T.; Ogumi, Z. Surface film formation on a graphite negative electrode in lithium-ion batteries: atomic force microscopy study on the effects of film-forming additives in propylene carbonate solutions. Langmuir 2001, 17, 8281− 8286. (2) Choi, N.-S.; Yew, K. H.; Lee, K. Y.; Sung, M.; Kim, H.; Kim, S.-S. Effect of fluoroethylene carbonate additive on interfacial properties of silicon thin-film electrode. J. Power Sources 2006, 161, 1254−1259. (3) Ryou, M.; Han, G.; Lee, Y.; Lee, J.; Lee, D.; Yoon, Y.; Park, J. Effect of fluoroethylene carbonate on high temperature capacity retention of LiMn2O4/graphite Li-ion cells. Electrochim. Acta 2010, 55, 2073−2077. (4) Liu, B.; Li, B.; Guan, S. Effect of fluoroethylene carbonate additive on low temperature performance of Li-ion batteries. Electrochem. Solid-State Lett. 2012, 15, A77−A79. (5) Hu, L.; Zhang, Z.; Amine, K. Fluorinated electrolytes for Li-ion battery: an FEC-based electrolyte for high voltage LiNi0.5Mn1.5O4/ graphite couple. Electrochem. Commun. 2013, 35, 76−79. (6) Fridman, K.; Sharabi, R.; Elazari, R.; Gershinsky, G.; Markevich, E.; Salitra, G.; Aurbach, D.; Garsuch, A.; Lampert, J. A new advanced lithium ion battery: combination of high performance amorphous columnar silicon thin film anode, 5 V LiNi0.5Mn1.5O4 spinel cathode and fluoroethylene carbonate-based electrolyte solution. Electrochem. Commun. 2013, 33, 31−34. (7) Seng, K.; Li, L.; Chen, D.; Chen, Z.; Wang, X.; Liu, H.; Guo, Z. The effects of FEC (fluoroethylene carbonate) electrolyte additive on the lithium storage properties of NiO (nickel oxide) nanocuboids. Energy 2013, 58, 707−713. (8) Eom, K.; Joshi, T.; Bordes, A.; Do, I.; Fuller, T. The design of a Li-ion full cell battery using a nano silicon and nano multi-layer graphene composite anode. J. Power Sources 2014, 249, 118−124. (9) Profatilova, I.; Kim, S.; Choi, N. Enhanced thermal properties of the solid electrolyte interphase formed on graphite in an electrolyte with fluoroethylene carbonate. Electrochim. Acta 2009, 54, 4445−4450. (10) Tsubouchi, S.; Domi, Y.; Doi, T.; Ochida, M.; Nakagawa, H.; Yamanaka, T.; Abe, T.; Ogumi, Z. Spectroscopic characterization of surface films formed on edge plane graphite in ethylene carbonatebased electrolytes containing film-forming additives. J. Electrochem. Soc. 2012, 159, A1786−A1790. (11) Liao, L.; Zuo, P.; Ma, Y.; An, Y.; Yin, G.; Gao, Y. Effects of fluoroethylene carbonate on low temperature performance of mesocarbon microbeads anode. Electrochim. Acta 2012, 74, 260−266. (12) Lin, Y.; Klavetter, K.; Abel, P.; Davy, N.; Snider, J.; Heller, A.; Mullins, C. High performance silicon nanoparticle anode in fluoroethylene carbonate-based electrolyte for Li-ion batteries. Chem. Commun. 2012, 48, 7268−7270. (13) Chockla, A.; Bogart, T.; Hessel, C.; Klavetter, K.; Mullins, C.; Korgel, B. Influences of gold, binder and electrolyte on silicon nanowire performance in Li-ion batteries. J. Phys. Chem. C 2012, 116, 18079−18086. (14) Profatilova, I.; Stock, C.; Schmitz, A.; Passerini, S.; Winter, M. Enhanced thermal stability of a lithiated nano-silicon electrode by fluoroethylene carbonate and vinylene carbonate. J. Power Sources 2013, 222, 140−149. (15) Markevich, E.; Fridman, K.; Sharabi, R.; Elazari, R.; Salitra, G.; Gottlieb, H. E.; Gershinsky, G.; Garsuch, A.; Semrau, G.; Schmidt, M. A.; Aurbach, D. Amorphous columnar silicon anodes for advanced high voltage lithium ion full cells: dominant factors governing cycling performance. J. Electrochem. Soc. 2013, 160, A1824−A1833. (16) Mazouzi, D.; Delpuech, N.; Oumellal, Y.; Gauthier, M.; Cerbelaud, M.; Gaubicher, J.; Dupré, N.; Moreau, P.; Guyomard, D.; 7423

dx.doi.org/10.1021/la501368y | Langmuir 2014, 30, 7414−7424

Langmuir

Article

(35) Ferraria, A. M.; Lopes da Silva, J. D.; Botelho do Rego, A.M. XPS studies of directly fluorinated HDPE: problems and solutions. Polymer 2003, 44, 7241−7249. (36) Xu, W.; Vegunta, S. S. S.; Flake, J. C. Surface-modified silicon nanowire anodes for lithium-ion batteries. J. Power Sources 2011, 196, 8583−8589. (37) Dedryvere, R.; Gireaud, L.; Grugeon, S.; Laruelle, S.; Tarascon, J.-M.; Gonbeau, D. Characterization of lithium alkyl carbonates by Xray photoelectron spectroscopy: experimental and theoretical study. J. Phys. Chem. B 2005, 109, 15868−15875. (38) Andersson, A.; Edstrom, K. Chemical composition and morphology of the elevated temperature SEI on graphite. J. Electrochem. Soc. 2001, 148, A1100−A1109. (39) Leroy, S.; Martinez, H.; Dedryvere, R.; Lemordant, D.; Gonbeau, D. Influence of the lithium salt nature over the surface film formation on a graphite electrode in Li-ion batteries: an XPS study. Appl. Surf. Sci. 2007, 253, 4895−4905. (40) Ensling, D.; Stjerndahl, M.; Nyten, A.; Gustafsson, T.; Thomas, J. O. A comparative XPS surface study of Li2FeSiO4/C cycled with LiTFSI- and LiPF6-based electrolytes. J. Mater. Chem. 2009, 19, 82− 88. (41) Edstrom, K.; Gustafsson, T.; Thomas, J. O. The cathodeelectrolyte interface in the Li-ion battery. Electrochim. Acta 2004, 50, 397−403. (42) Naumkin, A. V.; Kraut-Vass, A.; Gaarenstroom, S. W.; Powell, C. J. NIST X-ray Photoelectron Spectroscopy Database, NIST Standard Reference Database 20, Version 4.1. http://srdata.nist.gov/xps/ (accessed April 9, 2014). (43) Castro, L.; Dedryvere, R.; Ledeuil, J.-B.; Breger, J.; Tessier, C.; Gonbeau, D. Aging mechanisms of LiFePO4 // graphite cells studied by XPS: redox reaction and electrode/electrolyte interfaces. J. Electrochem. Soc. 2012, 159, A357−A363. (44) Verdier, S.; El Ouatani, L.; Dedryvère, R.; Bonhomme, F.; Biensan, P.; Gonbeau, D. XPS study on Al2O3- and AlPO4-coated LiCoO2 cathode material for high-capacity Li ion batteries. J. Electrochem. Soc. 2007, 154, A1088−A1099. (45) Lu, D.; Xu, M.; Zhou, L.; Garsuch, A.; Lucht, B. L. Failure mechanism of graphite/LiNi0.5Mn1.5O4 cells at high voltage and elevated temperature. J. Electrochem. Soc. 2013, 160, A3138−A3143. (46) Borgel, V.; Markevich, E.; Aurbach, D.; Semrau, G.; Schmidt, M. On the application of ionic liquids for rechargeable Li batteries: high voltage systems. J. Power Sources 2009, 189, 331−336. (47) Li, S. R.; Chen, C. H.; Xia, X.; Dahn, J. R. The impact of electrolyte oxidation products in LiNi0.5Mn1.5O4/Li4Ti5O12 cells. J. Electrochem. Soc. 2013, 160, A1524−A1528. (48) Dedryvere, R.; Foix, D.; Franger, S.; Patoux, S.; Daniel, L.; Gonbeau, D. Electrode/electrolyte interface reactivity in high-voltage spinel LiMn1.6Ni0.4O4/Li4Ti5O12 lithium-ion battery. J. Phys. Chem. C 2010, 114, 10999−11008.

7424

dx.doi.org/10.1021/la501368y | Langmuir 2014, 30, 7414−7424