Surface and Electrochemical Studies on Silicon ... - ACS Publications

Jan 31, 2018 - Present-day discussions on the improvement of a lithium battery system mainly center on the increase of energy density and enhanced ove...
17 downloads 6 Views 5MB Size
Research Article Cite This: ACS Appl. Mater. Interfaces 2018, 10, 7096−7106

www.acsami.org

Surface and Electrochemical Studies on Silicon Diphosphide as Easyto-Handle Anode Material for Lithium-Based Batteriesthe Phosphorus Path Romy Reinhold,*,†,‡ Ulrich Stoeck,† Hans-Joachim Grafe,† Daria Mikhailova,† Tony Jaumann,† Steffen Oswald,† Stefan Kaskel,‡ and Lars Giebeler† †

Leibniz Institute for Solid State and Materials Research (IFW) Dresden e.V., Helmholtzstraße 20, D-01069 Dresden, Germany Department of Inorganic Chemistry, Technische Universität Dresden, Bergstraße 66, D-01069 Dresden, Germany



S Supporting Information *

ABSTRACT: The electrochemical characteristics of silicon diphosphide (SiP2) as a new anode material for future lithiumion batteries (LIBs) are evaluated. The high theoretical capacity of about 3900 mA h g−1 (fully lithiated state: Li15Si4 + Li3P) renders silicon diphosphide as a highly promising candidate to replace graphite (372 mA h g−1) as the standard anode to significantly increase the specific energy density of LIBs. The proposed mechanism of SiP2 is divided into a conversion reaction of phosphorus species, followed by an alloying reaction forming lithium silicide phases. In this study, we focus on the conversion mechanism during cycling and report on the phase transitions of SiP2 during lithiation and delithiation. By using ex situ analysis techniques such as X-ray powder diffraction, formed reaction products are identified. Magic angle spinning nuclear magnetic resonance spectroscopy is applied for the characterization of long-range ordered compounds, whereas X-ray photoelectron spectroscopy gives information of the surface-layer species at the interface of active material and electrolyte. Our SiP2 anode material shows a high initial capacity of about 2700 mA h g−1, whereas a fast capacity fading during the first few cycles occurs which is not necessarily expected. On the basis of our results, we conclude that besides other degradation effects, such as electrolyte decomposition and electrical contact loss, the rapid capacity fading originates from the formation of a low ion-conductive layer of LiP. This insulating layer hinders lithium-ion diffusion during lithiation and thereby mainly contributes to fast capacity fading. KEYWORDS: lithium batteries, anode materials, silicon phosphide, deactivation, XPS, NMR, XRD

1. INTRODUCTION Present-day discussions on the improvement of a lithium battery system mainly center on the increase of energy density and enhanced overall performance parameters. As graphite, with a specific capacity of 372 mA h g−1, seems to have reached almost the theoretical limit in advanced Li-based batteries, silicon anode (composite) materials have garnered valuable attention, and there has already been fast and optimistic progress to enhance the capacity further. The ability of silicon to form intermetallic compounds with lithium enhances the specific capacity up to 3600 mA h g−1.1−5 As demonstrated by the compound Li3.75Si compared to graphite, alloy formation allows the insertion of almost four lithium ions instead of one lithium ion per C6 formula unit in graphite,6 which explains the superior specific capacities of Li− Si compounds. However, with the high amount of inserted lithium, a 300% volume expansion of the Si lattice occurs during lithiation and delithiation, which needs to be managed to promote silicon becoming a high-performance electrode.7 In principle, three main events are responsible for the fast degradation of Si anodes: (I) Huge volume changes cause © 2018 American Chemical Society

crack formation and pulverization because of the alloying− dealloying mechanism followed by electrical contact loss. (II) The solid electrolyte interface (SEI) is permanently formed and grows continuously because of the pulverization, always generating fresh, SEI-uncovered surfaces. This action increases the loss of electrical contact and leads to a high electrolyte consumption and, thereby, rapid capacity fading. (III) Displacement of a part of the anode composite during cycling additionally increases the contact loss problem.6,8−11 One possibility to overcome this problem is to nanostructure the sample.3,8,11,12 Therefore, silicon nanoarchitectures such as silicon nanowires,13−15 nanopillars,16,17 thin films,18 and nanoparticles11,12,19 have been tested. Another promising approach to facilitate high-performance electrode materials is the in situ formation of the electrochemically active material. Therefore, binary nonmetal compounds, for example, phosphides, have attracted noticeable Received: December 8, 2017 Accepted: January 26, 2018 Published: January 31, 2018 7096

DOI: 10.1021/acsami.7b18697 ACS Appl. Mater. Interfaces 2018, 10, 7096−7106

Research Article

ACS Applied Materials & Interfaces

at room temperature where HF was added until a clear solution without any residue was obtained. Three measurements for each sample after standardization were conducted, and the average concentration was taken to calculate the quantitative element content. This procedure was repeated for all produced batches used here. 2.2. Characterization of Structure and Morphology. XRD patterns were recorded on a STADI P diffractometer (STOE) with a curved Ge(111) crystal as the monochromator in Debye−Scherrer mode. Cu Kα1 radiation (λ = 1.54056 Å) was used as the X-ray source with a step width of Δ2θ = 0.02°. Samples were filled in a glass capillary (diameter 0.5 mm) and fire-sealed. Rietveld analyses were performed using the software FullProf implemented into the software package WinPLOTR.31,32 On the basis of literature data, the XRD patterns were analyzed for SiP233 and Li3P.34 Scanning electron microscopy (SEM) images were taken on a Gemini LEO 1530 (Zeiss) with a Schottky emitter and an acceleration voltage of 10 kV. For energy-dispersive X-ray spectroscopy (EDXS) measurements, a Bruker XFlash Detector 4010 was applied. The exposure to air was minimized. Transmission electron microscopy (TEM) experiments were carried out on a FEI Tecnai F30 with 300 kV acceleration voltage equipped with a field emission gun. The material was dispersed in ethanol through sonication and drop-coated on a copper grid with lacey carbon layer as the sample holder. Selected area electron diffraction (SAED) and high-resolution TEM (HRTEM) images were taken. The Gatan DigitalMicrograph software was used for a detailed analysis. The X-ray photoelectron spectra were recorded on a Physical Electronics PHI 5600 CI equipped with a hemispherical analyzer allowing high-sensitivity and high-resolution experiments. Energy scale and binding energy of the spectrometer were calibrated with Cu and Au foils at the binding energies of Cu 2p3/2 (932.67 eV) and Au 4f7/2 (84.00 eV). All measurements were conducted at a base pressure of 10−10 mbar. Monochromatic Al Kα radiation (350 W, 1486.7 eV) with a pass energy of 29 eV and a step width of 0.1 eV in high-resolution mode was used as an X-ray source. For ablating the upper layers, Ar+ ions with an energy of 3.5 keV were utilized. The sputtering speed of 3.3 nm/min was referenced to SiO2. For all XPS measurements, cycled electrodes were washed once with dimethyl carbonate (DMC) to remove some electrolyte decomposition residues and dried in a vacuum oven at 80 °C overnight. Recorded spectra were normalized to C 1s (284.4 eV). Peak fitting was performed with a MagicPlot Student edition using a Gaussian-A function. To prevent contamination with air, the PHI 04-110 transfer vessel (physical electronics) was used to transport the sample from the glovebox to the XP spectrometer.35 Solid-state magic angle spinning (MAS) NMR experiments were conducted with a wide bore 7 T magnet using a Tecmag Apollo NMR spectrometer as well as an NMR Service Eagle H-X 4 mm CP/MAS probe. NMR spectra are recorded for 7Li, 29Si, and 31P nuclei, which were internally referenced to 1 M LiCl in D2O, TMS, and 85 wt % H3PO4, respectively. For all NMR measurements, cycled electrodes were washed three times with DMC to remove the organic and inorganic substances from the electrolyte decomposition. The washed electrodes were dried in a vacuum oven at 80 °C overnight. ZrO2 rotors (4 mm) were packed with powder samples and sealed with tight-fitting Kel-F caps in an argon-filled glovebox. MAS was performed using dry nitrogen gas at 10 kHz. Peak fitting was performed with the MagicPlot Student edition using a Lorentzian-A function, and the intensities were normalized to 1. 2.3. Electrochemical Characterization. For electrode preparation, 50 wt % SiP2 were mixed with 40 wt % Super P (Timcal) and 10 wt % PVDF 1013 (Solvay). The mixture was pressed onto a copper net (Ø 12 mm) and dried at 80 °C overnight under vacuum. The Swagelok cells were assembled in an argon-filled glovebox (H2O < 1 ppm, O2 < 0.1 ppm) using a two-electrode configuration with lithium metal discs (ChemPur, 250 μm thickness) as the counter electrode, two glass fiber separators (Whatman), and 250 μL of electrolyte. As the electrolyte, 1 M LiPF6 in dimethyl carbonate/ethylene carbonate (1:1 v/v) (Selectilyte LP30, BASF) was used. Electrochemical tests were conducted at a constant temperature of 25 °C in a climate chamber. A multichannel VMP3 potentiostat (Bio-Logic) was applied

attention toward enhancing the overall electrochemical performance of silicon-based anodes. On some nonsilicon phosphides, first studies have been performed to identify suitable electrode materials.20−24 The main advantage addressed by this material is the formation of a stable matrix, where the silicon active material is embedded. From this point of view, the matrix-supported nanostructuring of the silicon active material potentially buffers the silicon lattice expansion during de-/lithiation as well as the material displacement and minimizes or suppresses continuous SEI growth or other parasitic reactions.25 As reported for Sn4P3,20 metallic tin was formed in situ during lithiation. On the basis of NMR studies,20 a general three-step conversion mechanism was developed with (i) the decomposition of Sn4P3 to form nanoparticulate Sn and LiP followed by (ii) the transformation of lithium phosphide (LiP) into trilithium phosphide (Li3P) and (iii) the Li−Sn alloy formation. Another promising anode material with a similar conversion mechanism was found with Cu3P which forms copper and Li3P after the first discharge cycle.26 On the basis of this idea, Reinhold and co-workers27,28 and Duveau et al.29 presented pioneering studies about the electrochemical behavior of SiP2 in lithium- and sodium-ion batteries achieving capacities of 2750 mA h g−1 (vs Li/Li+),28 2380 mA h g−1 (vs Li/Li+),29 and 1160 mA h g−1 (vs Na/ Na+)29 after the first discharge. Kwon et al.30 proposed a reaction mechanism of SiP2 suggesting (1) a topotactic transition of SiP2 to LixSiP2 (x ≤ 1.8), followed by (2) an amorphization and a (3) final conversion to Li13Si4 and Li3P. Additionally, they demonstrated a SiP2/C nanocomposite with enhanced electrochemical performance, especially with regard to higher coulombic efficiencies (CEs) of about 99.9% over cycling. We propose, however, that the mechanism of the transformation of SiP2 into an electrochemically active electrode composite is much more complex with respect to mechanistic intermediate steps and phase formation, as previously suggested. Our study therefore focuses on fundamental investigations about the synthesis of SiP2 via ball milling and the reaction during cycling. Furthermore, we reveal first the results of ex situ X-ray diffraction (XRD), solid-state NMR spectroscopy, and Xray photoelectron spectroscopy (XPS) giving deeper insights into the chemical and structural states at different reaction steps taking place in the SiP2 nanoparticle as well as on their surface. We hereby focus on the fate of phosphorus species undergoing a conversion reaction during cycling. These results are related to electrochemical measurements to allow new conclusions on the underlying reaction mechanism.

2. EXPERIMENTAL SECTION 2.1. Synthesis. The planetary ball mill Pulverisette 7 premium line (Fritsch) was used to synthesize SiP2 from elemental powders. Silicon (Aldrich, −325 mesh, 99% purity) and red phosphorus (Alfa Aesar, −100 mesh, 98.9% purity) powders were stoichiometrically (1:2) mixed in a silicon nitride beaker with 20 silicon nitride balls (Ø 10 mm). Milling was performed under Ar atmosphere at a rotational velocity of 800 rpm for 45 min. This milling procedure was performed four times, where every second cycle was operated in reverse direction. After eight milling cycles, the final reaction state of the synthesis was reached. The precision of the synthesis was quantitatively evaluated with chemical analysis by inductively coupled plasma optical emission spectrometry (ICP-OES) with an iCAP 6500 Duo View of ThermoFisher Scientific. For the analysis preparation, the as-prepared SiP2 after ball milling was weighed inside a glovebox under Ar inert gas atmosphere, and the sample was dissolved in half-concentrated HNO3 7097

DOI: 10.1021/acsami.7b18697 ACS Appl. Mater. Interfaces 2018, 10, 7096−7106

Research Article

ACS Applied Materials & Interfaces for the electrochemical measurements. Galvanostatic cycling with potential limitation (GCPL) was performed between 0.01 and 1.2 V versus Li/Li+ at a current density of 50 mA g−1. The current density and the specific capacity were calculated based on the mass of SiP2.

3. RESULTS AND DISCUSSION 3.1. Characterization of the As-Prepared Silicon Diphosphide. As known from the binary phase diagram of silicon and phosphorus,36 SiP is the only thermodynamically stable phase at 25 °C. Consequently, there are only a few publications regarding the synthesis of SiP2, for example, by vapor transport techniques33 or by solution synthesis in molten salts.37 The high-energy ball milling (HEBM) technique represents a scalable and environment-friendly alternative to obtain metastable SiP2, as shown in recent publications.27−30 SiP2 crystallizes in a pyrite-type cubic system, space group Pa3̅, with the lattice parameter a = 5.7045 Å.33 Silicon is distorted octahedrally coordinated by P−P dumbbells, whereas phosphorus is surrounded by three silicon and one phosphorus atom in a tetrahedral configuration (Figure S1 in the Supporting Information). The diffraction pattern of the asprepared SiP2 is presented in Figure 1. According to the

Figure 2. SEM secondary electron image of the as-prepared SiP2 (a), a secondary electron image (b), and the corresponding P (c) and Si (d) EDXS mapping images.

TEM gives information about the arrangement of the crystallites in the material (Figure 3). The SAED pattern (Figure 3a) of SiP2 confirms the presence of crystalline silicon

Figure 1. XRD pattern of the as-prepared SiP2 (red dots) with the calculated profile (black solid line) according to the Rietveld analysis of the structure model from SiP2.33 Difference curve resulted from the subtraction of the observed to calculated data (blue solid line). Some Miller indices are omitted for clarity.

Rietveld analysis, the lattice parameter is determined to be a = 5.7055(2) Å with an averaged crystallite size of 13 nm. Refinement of the P occupancy with the fixed Si content led to the composition SiP2. SEM images (Figure 2a) reveal particles with a size ranging from several nanometers up to a few micrometers indicating agglomeration with a broad particle size distribution. Besides particle size determination, SEM combined with EDXS offers elemental analysis and the elemental distribution of the compound. Figure 2b shows a section of the sample and the corresponding EDXS mappings of phosphorus (Figure 2c) and silicon (Figure 2d). As expected, both elements are uniformly distributed throughout the sample suggesting the formation of a homogenous compound. All particles consist of both elements as expected from the XRD measurement. The elemental composition, as determined by ICP-OES, supported these results (Table S1). The molar elemental content of silicon and phosphorus is in accordance with a composition of 1 to 2, as expected.

Figure 3. SAED (a) and HRTEM image (b) of the as-prepared SiP2. Amorphous SiOx layer is indicated by red arrows. 7098

DOI: 10.1021/acsami.7b18697 ACS Appl. Mater. Interfaces 2018, 10, 7096−7106

Research Article

ACS Applied Materials & Interfaces

Figure 4. O 1s (a), P 2p (b), and Si 2p (c) X-ray photoelectron spectra of the as-prepared SiP2 before and at different ablation depths.

diffraction methods, and thereby the two Si 2p signals at 100.0 eV (p3/2) and 100.8 eV (p1/2) are assigned to SiP2. In the P 2p spectra, the signals at 130.0 and 129.1 eV correspond to the binding energies of p1/2 and p3/2, respectively. The signal at 129.1 eV indicates the formation of a phosphide. The measured binding energy position of 129.1 eV is in good agreement with literature values of NiP2,38,39 which is isostructural to SiP2.40 Before cycling, SiP2 is analyzed after being in contact with LP30 electrolyte for 2 h by the XPS analysis. Figure 5 shows the

diphosphide because of the ring diffraction pattern, which is superimposed with the calculated one. The circular shape of the SAED pattern indicates polycrystallinity with variously oriented crystallites. In Figure 3b, a representative HRTEM image of the sample presents three nanocrystallites oriented in several directions with the typical d values of SiP2 of the indexed planes. Furthermore, HRTEM allows the determination of crystallite size which ranges from approximately 10 to 20 nm in diameter. Therefore, the crystallite size obtained by HRTEM image is in the same size range and confirms XRD analysis. Additionally, a thin amorphous layer of about 2 nm is observed, which wraps the SiP2 nanoparticles. To obtain information on the nature and origin of the thin surface layer and to find out the intrinsic composition, XPS experiments were conducted. Figure 4 exhibits the recorded spectra for the O 1s (a), P 2p (b), and Si 2p (c) binding energies (BE). From these measurements in combination with the results of SEM−EDXS and TEM, a thin SiO2 layer with a typical O 1s signal at 532.4 eV (Figure 4a) and two Si 2p signals at 103.2 eV (p1/2) and 102.3 eV (p3/2) (Figure 4c) are detected on the surface of the SiP2 particles. The composition of the sample provided by XPS is summarized in Table S2 and Figure S3 (Supporting Information). Because of the sensitivity of the chosen method, occurrence of carbon (2−6 at %) is explained by contaminations of the sample. The C content reduces during sputtering. By evaluating the fit of the signal areas of the Si 2p binding energy (without sputtering), the proportion of silicon is distributed in 20% of SiO2 and 80% of SiP2. Therefore, a fraction of 6.2 at % silicon is assigned to SiO2, which perfectly fits with the oxygen content of 12.4 at %. Regarding the SEM− EDXS analysis and the TEM images, the thin amorphous layer around the particles mainly consists of silicon dioxide. This observation may be explained by the natural oxide layer, which is almost unavoidable on silicon particles. Even after using HEBM, the natural oxide layer remains during SiP2 synthesis. Accordingly, 25 at % silicon binds to 50 at % phosphorus which is in good agreement with the calculated and expected values. After the removal of several atomic layers by Ar+ ion sputtering for 5 min, the signals of SiO2 shift to lower binding energies in the O 1s and the Si 2p spectra corresponding to lower electron density at the silicon atoms. This observation indicates the formation of SiOx compounds with 1 ≤ x ≤ 2. Assuming that 6 at % oxygen forms SiO species, an excess of 8 at % silicon is observed. This observation might be explained by amorphous silicon, which cannot be detected by the XRD analysis or by the formation of substoichiometric Si−P compounds. To the best of our knowledge, no reference XPS data can be found for SiP2. However, silicon diphosphide is identified via

Figure 5. P 2p X-ray photoelectron spectra of SiP2 after contact with LP30 electrolyte before and at different ablation depths.

P 2p X-ray photoelectron spectra. A thick phosphate-containing (LixPOy) (2p3/2 132.8 eV, 2p1/2 133.8 eV) interfacial layer as well as some LiP (2p3/2 128.7 eV, 2p1/2 129.6 eV) are already formed when SiP2 is exposed to the LP30 electrolyte substantiating the high reactivity of nanoparticulate SiP2. The signals at 137.8 eV (P 2p1/2) and 136.7 eV (P 2p3/2) are assigned to residues of the LiPF6 conducting salt. After sputtering for 10 min, no signals related to SiP2 are observed. Most likely, an already-formed thick interfacial layer between electrolyte and electrode material exists, as implied by the high SiP2 reactivity. 3.2. Mechanistic Studies on Silicon Diphosphide Conversion. To investigate the mechanism of SiP2 transformation depending on the cycling state, ex situ XRD measurements were performed. XRD presupposes a longrange order as found in crystals, but some intermediates formed during cycling may not correspond to this requirement. Solidstate MAS NMR experiments on the 31P nucleus were carried 7099

DOI: 10.1021/acsami.7b18697 ACS Appl. Mater. Interfaces 2018, 10, 7096−7106

Research Article

ACS Applied Materials & Interfaces

of a passivation layer appearing only in the first cycle. However, reactions occurring in this potential range are associated with the decomposition of the electrolyte on fresh nonpassivated electrode active material surfaces. These electrolyte decomposition products form an SEI on the SiP2 nanoparticles and may transform the upper layer of a SiP2 particle. This principle is well-documented for carbonate-based electrolytes in contact with the anode materials, as used for lithium-ion battery, and is also indicated by the XPS analysis, as discussed above.11,41−44 Figure 7 shows the XRD patterns as well as the corresponding NMR spectra of SiP2 at different lithiation states for certain cycling steps. It should be noted that the points A2* and A4* are only introduced for better discussion of the electrochemical behavior. The plot of the first GCPL cycle is added as Figure 7a. Starting at 3 V versus Li/Li+ and going to 0.01 V (46th h) describes the lithiation, whereas the potential increase to 1.2 V versus Li/Li+ (82nd h) is related to the delithiation of the active material. The corresponding XRD patterns and the NMR spectra are shown in Figure 7b,c, respectively. The initial material (gray curve), as detected by XRD, fits well with the X-ray data for SiP2. Therefore, the narrow signal in the 31P NMR at 150 ppm is assigned to pure SiP2. This result is in good agreement with other phosphides such as NiP2 (185 ppm)21 showing the same distorted octahedral coordination of the central atom by P atoms. In the cathodic direction, after the first discharging stepdown to 0.7 V versus Li/Li+ (A1, green), a decrease in intensity of the diffraction signal is observed as a result of a reduced content of the crystalline phase. The broad signal at 20°−25° 2θ indicates the amorphous carbon additive. A partial conversion of SiP2 to amorphous Li3P represented by the 31P NMR signal at −280 ppm occurs. The broadening of the SiP2 signal indicates an ongoing amorphization of SiP2. Additional signals between −50 and 50 ppm are assigned to binary lithium phosphorus phases such as LiP7 (0 ppm) and LiP5 (−20 ppm)45 formed as intermediates, which were also observed during the cycling of black phosphorus.46 Moreover, the signal at −190 ppm is attributed to LiP formation,45 which was often found during the cycling of phosphides, such as Sn4P3 or CoP3.22,23 Discharging to a lower voltage of about 0.5 V (A2, red) results in further amorphization of SiP2 without a crystallization of Li3P nanoparticles, as shown by XRD. The reduction peak at 350 mV is observed according to the incipient formation of crystalline Li3P, which is completed around 0.1 V versus Li/Li+ (Figure 6, A2*, gray) and correspond to the 31P NMR signal at −280 ppm, as reported in the literature.20,45 The small drop in the range of 0.01 V ≤ E ≤ 0.1 V, more pronounced and better visible in Figure 6, might be interpreted as the formation of a lithium silicide phase, which can typically occur below 0.1 V versus Li/Li+ 11 and is taken as an indicator for the successful formation of amorphous LixSi.9 Crystalline Li15Si4, the highest lithiated phase of silicon, cannot be observed by XRD (Figure 7b) even at 10 mV, only confirming the formation of amorphous LixSi. Finally, at 10 mV versus Li/Li+ (A3, blue), crystalline Li3P is visible as evidenced by NMR and crystallizes with an average crystallite size of about 80 nm in a structure type with the hexagonal space group P63/ mmc and the lattice parameters a = 4.3483(1) Å and c = 7.7168(2) Å which exhibits a layered structure of alternating Li2P and Li layers.47,48 In many cases, this conversion-like reaction is observed in electrochemically active phosphides at low potentials versus Li/Li+.22,24 It should be noted that besides

out to overcome the limitation of XRD and to allow for the short-range order information acquisition. For this kind of ex situ experiments, cells were cycled in the potential window of 0.01−1.2 V versus Li/Li+ until a defined potential was reached. Discharging is hereby defined as the insertion of lithium atoms. To evaluate the characteristic discharging and charging potential versus Li/Li+, the differential capacity plot of SiP2 cycled in the above-mentioned potential window is shown in Figure 6. After each redox reaction indicated by a peak

Figure 6. Typical differential capacity plot of a SiP2 electrode during the first cycle in a potential window between 0.01 and 1.2 V at a current density of 50 mA g−1. Points A1 to A5 indicate the samples displayed in 31P NMR and XRD plots in Figure 7.

occurring in the differential capacity plot, a sample was taken and analyzed with the above-mentioned techniques. Although all other events are discussed in combination with Figure 7, the small peak at around 0.6 V versus Li/Li+ (Figure 6) is particularly paid attention here. It is attributed to the formation

Figure 7. GCPL curves of SiP2 discharged to 0.01 V (fully lithiated) and charged to 1.2 V (delithiated) (a) and their corresponding ex situ XRD patterns (λ = 1.54056 Å) (b), and ex situ 31P NMR MAS spectra (c) at different potential steps. 7100

DOI: 10.1021/acsami.7b18697 ACS Appl. Mater. Interfaces 2018, 10, 7096−7106

Research Article

ACS Applied Materials & Interfaces

Figure 8. P 2p spectra of cycled SiP2 samples after first, second, and third half-cycles without (a) sputtering and after sputtering for (b) 5 min and (c) 10 min.

Figure 9. P 2p spectra of cycled samples after first (a), second (b), and third (c) half-cycle without sputtering (“surface”) and after sputtering 5 min (∼16 nm), 10 min (∼33 nm), 15 min (∼50 nm), and 30 min (∼100 nm).

LiP completes at 0.9 V versus Li/Li+ (Figure 6, A4*, gray). No reformation of SiP2 is observed after full delithiation to 1.2 V versus Li/Li+ (A7, Figure 7, turquoise). Similar behavior was observed for the lithiation of crystalline silicon causing amorphization, without reformation of crystalline silicon after full delithiation.50 The SiP2 reformation is additionally thermodynamically unfavorable, which implicates an insufficient ability to reform this phase at the available conditions.36,51 To prove the suggested mechanism along with gaining information on the particle composition, XPS experiments were conducted. Figure 8 shows the phosphorus 2p spectra of SiP2 after discharging to 0.01 V (first half-cycle), charging to 1.2 V (second half-cycle), and discharging to 0.01 V versus Li/Li+ (third half-cycle). Figure 8 represents all samples without depth-profile sputtering. A close examination of the phosphorus region exhibits two overlapping signals. The binding energy signals at 137.8 eV (p1/2) and 136.7 eV (p3/2) are caused by PFy species originating from incomplete electrolyte removal, whereas the signals at 133.2 eV (p3/2) and 134.2 eV (p1/2) are allocated to P−O species such as LixPOy-like compounds.41,52 It is confirmed that PFy species remain after electrolyte removal at the surface of the particles, as indicated by an observed signal at 687 eV in the F 1s spectra (Supporting Information, Figure S6). Comparing all samples without sputtering, no significant differences are identified. After 5

SiP2, Li3P, and LiP, other phosphorus-containing species from SEI formation show signals in 31P NMR which creates difficulties for quantitative assignment, but the main intensity is related to Li3P and LiP. During delithiation, a first oxidation signal is detected at 300 mV11,49 resulting from the progressive degradation of the final Li3P and LixSi phases. At a potential of 0.4 V versus Li/Li+ (A4, orange), the XRD pattern shows an increase of the Li3P diffraction line-broadening indicating smaller Li3P crystallites with an average size of 18 nm. This behavior indicates the beginning of the transformation from a crystalline to an amorphous state. 31P NMR exhibits two different signals at −280 and −190 ppm corresponding to Li3P and LiP, respectively. As there is only a change in intensity ratios of these signals, no SiP2 forms during delithiation. Because of the change in intensity ratios, a partial transformation of Li3P to LiP is found. In contrast to Li3P showing a high ionic conductivity of 10−4 (Ω cm)−1, LiP is an insulator.47 In the case of an incomplete reverse reaction, the formation of new compounds stops at LiP and the high loss in capacity mainly results from the low ionic conductivity of LiP. The Li3P particles are lithium-depleted at the outer surface, and a protective but insulating amorphous LiP layer forms, hindering further delithiation. With this barrier, phosphorus is unable to transfer into a new reaction path to reform SiP2. After further forced lithium-ion removal, the reaction of Li3P to amorphous 7101

DOI: 10.1021/acsami.7b18697 ACS Appl. Mater. Interfaces 2018, 10, 7096−7106

Research Article

ACS Applied Materials & Interfaces

Figure 10. Assumption of a core−shell model for the interfaces between the electrochemically active SiP2 particle and the electrolyte.

min of Ar+ ion-sputtering to remove the upper layer (Figure 8b), only small discrepancies are found. P−O species (P 2p3/2 at 132.8 eV and P 2p1/2 at 133.8 eV), originating from electrolyte decomposition of carbonate-based solvents, are still observed in all samples. As observed by XRD measurements, Li3P is formed at the fully lithiated state showing a signal at very low P 2p binding energies of 126.2 eV (p3/2) and 127.2 eV (p1/2), which nicely fits the energy range of the few binding energies reported for phosphides, and the reference values for Li3P, which vary between 126127 eV.39,53 However, after delithiation to 1.2 V versus Li/Li+, Li3P is completely consumed. Interestingly, a minor signal at 128.7 eV (p3/2) and 129.6 eV (p1/2) indicates the formation of a thin LiP layer. To the best of our knowledge, no reference XPS data for LiP is available. Because of its structure consisting of P−P chains, we would expect a signal around 129−130 eV, which is close to red phosphorus with a P 2p binding energy at 130.5 eV.54 Investigations regarding metallic monophosphides52 (M = Cr, Mn, Fe, and Co) exhibit signals in the low energy range of the P 2p spectra at around 129 eV (2p3/2) and 130 eV (2p1/2) applicable to the observed signals for LiP. Comparing the signal intensities of LiP after 5 min (Figure 8b) and 10 min (Figure 8c) of sputtering, an increase in intensity corresponding to an increase in LiP concentration is recognized suggesting a higher content of insulating LiP in the lower layer than in the upper layer. After cycling of the SiP2 anodes, depth-profile sputtering in combination with XPS analyses was conducted (Figure 9). After the first discharge and before sputtering (Figure 9a, “surface”), LixPOy and PFy are both detected as the two main constituents of the upper SEI-related layer. The first sputtering of 5 min ablate the electrolyte-dependent PFy part of the SEIlayer, and three different compounds are observed, where LixPOy seems to spread into deeper parts of the layer but LiP and Li3P also become visible. By longer sputtering until 30 min (almost 100 nm in depth) into the layer, the proportion of phosphate-like species (P 2p3/2 132.8 eV, P 2p1/2 133.8 eV) decreases, whereas the amount of Li3P (P 2p3/2 126.2 eV, P 2p1/2 127.2 eV) increases. As mentioned above, LiP (P 2p3/2 128.7 eV, P 2p1/2 129.6 eV) is observed most prominently after 5 and 10 min sputtering or about 16 and 33 nm, respectively, in depth. Later or in larger depths, the LiP signal becomes less intense and the lower limit of layer is reached by the sputtering process. After removing lithium ions (Figure 9b), an interface consisting of phosphate species and an increasing content of lithium phosphides (LiP and Li3P) depending on the sputtering depth are observed. Taking into account that 31P NMR measurements can indicate the existence of amorphous Li3P, an amorphization of Li3P and an increasing thickness of the LiP layer are shown by changes of line-broadening and by

comparison of the relative intensities of Li3P and LiP visualized by an increase of the LiP peak at −190 ppm, respectively. This thicker layer results in enhanced passivation properties and in fast capacity fading after a few cycles. In contrast to the first discharging cycle, after the second discharge (Figure 9c), the composition of the phosphorus species changed significantly. On the surface, more phosphate species are detected, and the total amounts of both Li3P and LiP increases when sputtered into larger depths. As an additional feature, an incomplete reaction of LiP is a most likely situation, which is demonstrated by the large LiP content in the first charging reaction (second half-cycle). With increasing LiP content, more Li is withdrawn from the underlying storage reaction, and this behavior may be another reason for fast capacity fading. To take this situation to extremes, a core−shell model seems to be possible where from the first-formed Li3P, a LiP layer remains, which is unable to fully react during charging. On the basis of our results, we developed a model of the passivation mechanism of the Li3P particles (Figure 10). Two major effects are observed: (1) growing of the interface between Li3P particle and electrolyte (green) and (2) growing of the insulating LiP layer (blue fit). The former effect leads to the assumption of the continuous degradation of electrolyte components during cycling. The latter effect contributes to faster capacity fading hindering lithium-ion diffusion. As the fully lithiated state of SiP2 at 0.01 V versus Li/Li+ is always denoted as LixSi and Li3P in previous sections, 7Li NMR experiments were carried out to address the stoichiometry and structure of the formed compounds. As expected from 31P NMR, shown in Figure 7, Li3P is formed at low potentials resulting in two signals at −1 and 5 ppm which are assigned to planar and tetrahedrally coordinated Li atoms, respectively.20,24 Within the Li2P layers, the Li atoms are located in one coordination sphere with the phosphorus atoms resulting in a threefold planar bonding (Supporting Information, Figure S7), whereas the Li atoms between these layers are tetrahedrally bound. Concerning the occupancy of the lithium atoms, two out of three are linked planar resulting in a doubled intensity compared to tetrahedrally coordinated Li atoms. This finding is in excellent agreement with the experimental data depicted in Figure 11 (gray). The most significant signal appears at 16 ppm and most likely originates from the formation of a lithium silicide phase as well as other lithium species such as Li2CO3, Li2O, LiF, and semiorganic lithium salts emerging from the decomposition of the electrolyte. Both reactions may contribute significantly to the signal as the CE within the first cycle is very low (60−80%) indicating severe side reactions.55 The formation of Li2O is often described in compounds consisting of natural silicon oxide layers. As mentioned above, after the ball milling 7102

DOI: 10.1021/acsami.7b18697 ACS Appl. Mater. Interfaces 2018, 10, 7096−7106

Research Article

ACS Applied Materials & Interfaces

Figure 11. 7Li MAS NMR spectrum of SiP2 at the fully lithiated state, 0.01 V vs Li/Li+ referenced to LiCl where Li3P and LixSi (1.7 ≤ x ≤ 3.25) are present.

synthesis, SiP2 particles are surrounded by a thin silicon oxide layer. During lithiation, Li2O forms (Figure S8) and consumes additional Li+ ions. However, as described in literature, Li2O might be beneficial for further cycling explained by improved transport properties.43,44 Key et al.56 reported 7Li NMR data for all known intermetallic LixSi compounds, which are stable at the room temperature. Because of the influence of the local environment of silicon atoms on the NMR signal, individual signal shift directions are observed for Si-rich and Si-poor compounds. For example, isolated Si−Si dumbbells in the Sirich phase Li7Si3 lead to a deshielding of the Li nucleus, which results in NMR signal shifts to higher ppm values compared to LixSi phases with a high Li content.56 Therefore, intermetallic Li−Si phases with a high Li/Si ratio exhibit shifts to lower ppm values. On the basis of their investigations, Key et al.56 proposed the signal located at 12−22 ppm consisting of Si−Si dumbbells. This description leads to the assumption that the fully lithiated SiP2 phase may form Li12Si7, Li7Si3, or Li13Si4. Further investigations providing additional information are in preparation. The results of the mechanistic studies of silicon diphosphide suggest (1) a formation of intermediate LiPx phases such as LiP5 and LiP7, (2) their subsequent conversion to LiP, followed by (3) Li3P formation and simultaneously (3′) lithiation of silicon nanoparticles. During discharging, the following reactions are suggested

Figure 12. Voltage profiles of the SiP2 negative electrode cycled between 0.01 and 1.2 V vs Li/Li+ at a current density of 50 mA g−1 (a) and calculated specific discharging capacity for the first 50 cycles (b).

inhomogeneous distribution of lithiated silicon nanoparticles, as well as probably different SEI thicknesses, and a higher potential may be needed to fully delithiate the silicon nanoparticles. As a result, the phase-formation mechanism is described as follows LixSi(1.7 ≤ x ≤ 3.75) → Si + x Li+ +

Li3P → LiP + 2Li + 2e

A similar mechanism was proposed by Jang et al. and León et al.20 for Sn4P3, suggesting LiP as the intermediate phase resulting in fast capacity fading of the anode material because of the low ionic conductivity and most probably leading to the incomplete reaction and Sn encapsulation, as described above. Because of the irreversibility of the delithiation mechanism, a fast capacity fading should occur during cycling. In Figure 12a, the charging and discharging curves are displayed as specific capacity versus voltage plot. After discharging to 0.01 V versus Li/Li+, a capacity of 2750 mA h g−1 (Figure 12b) is achieved. Forming the final phase of Li3P and probably LixSi, a theoretical capacity of 2300 mA h g−1 (Li12Si7) to 2700 mA h g−1 (Li13Si4) can be reached. Experimental and theoretical capacities for the first cycle are in good agreement, whereby differences are explained by electrolyte decomposition during the first cycle and the irreversible reaction back to SiP2, as indicated by the low CE of 65% (Figure 12b). Regarding the first charging cycle (Figure S9), about three Li+ ions were removed until 0.75 V versus Li/ Li+. By further charging, about three Li+ ions were extracted, being in good agreement with the incomplete reverse reaction to form LiP. In the second discharging half-cycle, most of the

(1)

2LiPx (5 ≤ x ≤ 7) + 1.6−1.7Li+ + 1.6−1.7e− → 2LiP (2) +



2LiP + 4Li + 4e → 2Li3P

(3)

Si + x Li+ + x e− → LixSi(1.7 ≤ x ≤ 3.75)

(3′)

(5) 58

SiP2 + 0.3−0.4 Li+ + 0.3−0.4 e− → 2LiPx (5 ≤ x ≤ 7) + Si

(4)



During charging, the reformation of silicon nanoparticles and delithiation of Li3P are expected. As described in literature, the delithiation of lithium silicide phases proceeds at around 300− 500 mV,57 whereas Li3P is delithiated above 0.8 V versus Li/ Li+.46 The observed charging behavior (Figure 12a) showing no distinct voltage plateaus might be described by the kinetic inhibition of some silicon nanoparticles, caused by the 7103

DOI: 10.1021/acsami.7b18697 ACS Appl. Mater. Interfaces 2018, 10, 7096−7106

Research Article

ACS Applied Materials & Interfaces

react quantitatively without the formation of a passivation layer, this electrode material will become a suitable alternative for the conventional nanostructured Si anodes. One possibility to reach this goal may be achieved by the use of electrolyte additives with better complexing properties for phosphorus species.

capacity is recovered leading to the assumption that the lithium-ion diffusion is hindered during delithiation (charging), whereas diffusion during lithiation (discharge) is more favored. This finding corresponds to the results of the XPS experiments suggesting the continuous formation of a LiP layer during charging, which is neither degraded during charging nor during discharging. The accessibility of Li3P is blocked leading to a high loss of capacity after the first charging cycle. Comparing the discharging characteristics of the first and the second cycle, especially the plateau between 0.5 and 0.3 V versus Li/Li+, decreases rapidly leading to a low reversibility of the reaction LiP to Li3P. Comparing the charging characteristics of the first and the second cycle, only a slight difference in the length of the plateau is observed. Besides these effects, common fatigue phenomena, such as loss of contact arising from stress during (de)lithiation within the particles and the locally restricted crystallization of Li3 P particles, are expected to have contributed to the low CE. After 10 cycles, a specific capacity of only 370 mA h g−1 (Figure 12b) remains the abovementioned critical factors. The plateaus almost completely disappeared at this stage, leading to the assumption that some reactions only partially occur.



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acsami.7b18697. Crystal structure of silicon diphosphide, SEM−EDXS analysis of the as-prepared silicon diphosphide, elemental composition determined by ICP-OES of the as-prepared silicon diphosphide, elemental composition provided by XPS of the as-prepared silicon diphosphide during sputtering, elemental composition provided by XPS of the as-prepared silicon diphosphide, XRD pattern of trilithium phosphide with corresponding Rietveld analysis, voltage versus time plot of SiP2, F 1s spectra of cycled samples, crystal structure of trilithium phosphide, O 1s spectra of cycled samples, and lithiation followed by delithiation depending on potentials (PDF)



4. CONCLUSIONS SiP2 was synthesized by a mechanochemical technique. This anode material exhibits a high first-discharge capacity of about 2700 mA h g−1 undergoing a conversion reaction from SiP2 → LiPx → LiP + Si, finally forming crystalline Li3P and amorphous LixSi. Because of the formation of a thin LiP layer and its low ionic conductivity, only 65% (1800 mA h g−1) of the initial discharging capacity is recovered after the first charging cycle. Furthermore, the reformation of the thermodynamically metastable SiP2 phase upon delithiation was not confirmed leading to the assumption that the capacity fading is partly attributed to this irreversible reaction. After the first conversion reaction of SiP2, a typical alloying-like lithiation mechanism, as expected for Si, occurs. On the basis of XPS measurements, we concluded the formation of a layered core−shell structure around electrochemically active Li3P, consisting of low ionically conductive LiP, responsible for fast capacity fading after a few cycles. Moreover, decomposition reactions of the carbonatebased electrolyte lead to the formation of a passivating layer resulting in low CE and finally also low capacity retention. In summary, four main factors contribute to fast capacity fading of silicon diphosphide: (1) No reformation of SiP2 (2) Formation of an insulating, continuously growing layer of LiP during lithium removal (3) Strong decomposition of carbonate-based electrolyte in the first cycles (4) Volume effects by the (repeated) formation of large crystals, similar to that for Li3P, and the transition to an amorphous state with less volume and higher flexibility to organize the particles in the electrode composite during de-/lithiation resulting in a loss of contact Nevertheless, our results shed light on the complex electrochemical reaction pathway and show that SiP2 is, in principle, an interesting anode material for future lithium technology similar to Li−S, but significant further performance improvements are necessary that mainly concern LiP. In case the formation of LiP can be controlled, for example, by suppression of the LiP reaction pathway or by allowing LiP to

AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. ORCID

Romy Reinhold: 0000-0001-7911-4630 Daria Mikhailova: 0000-0002-8197-1807 Stefan Kaskel: 0000-0003-4572-0303 Lars Giebeler: 0000-0002-6703-8447 Author Contributions

The manuscript was written through contributions from all authors. All authors have given approval to the final version of the manuscript. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS We gratefully acknowledge Dr. Ahmad Omar for helpful discussions and the German Federal Ministry of Education and Research (BMBF) for the financial support through the Excellent BatteryWING center BamoSa “BatteriesMobility in Saxony” (grant no. 03X4637).



REFERENCES

(1) Obrovac, M. N.; Chevrier, V. L. Alloy negative electrodes for Liion batteries. Chem. Rev. 2014, 114, 11444−11502. (2) Park, C.-M.; Kim, J.-H.; Kim, H.; Sohn, H.-J. Li-alloy based anode materials for Li secondary batteries. Chem. Soc. Rev. 2010, 39, 3115− 3141. (3) Szczech, J. R.; Jin, S. Nanostructured silicon for high capacity lithium battery anodes. Energy Environ. Sci. 2011, 4, 56−72. (4) Su, X.; Wu, Q.; Li, J.; Xiao, X.; Lott, A.; Lu, W.; Sheldon, B. W.; Wu, J. Silicon-based nanomaterials for lithium-ion batteries: a review. Adv. Energy Mater. 2014, 4, 1300882. (5) McDowell, M. T.; Lee, S. W.; Nix, W. D.; Cui, Y. 25th Anniversary Article: Understanding the lithiation of silicon and other alloying anodes for lithium-ion batteries. Adv. Mater. 2013, 25, 4966− 4985. 7104

DOI: 10.1021/acsami.7b18697 ACS Appl. Mater. Interfaces 2018, 10, 7096−7106

Research Article

ACS Applied Materials & Interfaces (6) Goriparti, S.; Miele, E.; De Angelis, F.; Di Fabrizio, E.; Zaccaria, R. P.; Capiglia, C. Review on recent progress of nanostructured anode materials for Li-ion batteries. J. Power Sources 2014, 257, 421−443. (7) Wu, H.; Chan, G.; Choi, J. W.; Ryu, I.; Yao, Y.; McDowell, M. T.; Lee, S. W.; Jackson, A.; Yang, Y.; Hu, L.; Cui, Y. Stable cycling of double-walled silicon nanotube battery anodes through solid-electrolyte interphase control. Nat. Nanotechnol. 2012, 7, 310−315. (8) Wu, H.; Cui, Y. Designing nanostructured Si anodes for high energy lithium ion batteries. Nano Today 2012, 7, 414−429. (9) Shi, L.; Kunz, U. Mechanical behavior during electrochemical and mechanical deactivation of an aged electrode in a lithium-ion pouch cell. Energy Technol. 2016, 4, 1520−1530. (10) Mauger, A.; Julien, C. Nanoscience supporting the research on the negative electrodes of Li-ion batteries. Nanomaterials 2015, 5, 2279−2301. (11) Jaumann, T.; Herklotz, M.; Klose, M.; Pinkert, K.; Oswald, S.; Eckert, J.; Giebeler, L. Tailoring hollow silicon-carbon nanocomposites as high-performance anodes in secondary lithium-based batteries through economical chemistry. Chem. Mater. 2015, 27, 37−43. (12) Liu, X. H.; Zhong, L.; Huang, S.; Mao, S. X.; Zhu, T.; Huang, J. Y. Size-dependent fracture of silicon nanoparticles during lithiation. ACS Nano 2012, 6, 1522−1531. (13) Cui, L.-F.; Ruffo, R.; Chan, C. K.; Peng, H.; Cui, Y. Crystallineamorphous core-shell silicon nanowires for high capacity and high current battery electrodes. Nano Lett. 2009, 9, 491−495. (14) Laïk, B.; Eude, L.; Pereira-Ramos, J.-P.; Cojocaru, C. S.; Pribat, D.; Rouvière, E. Silicon nanowires as negative electrode for lithium-ion microbatteries. Electrochim. Acta 2008, 53, 5528−5532. (15) Krause, A.; Dörfler, S.; Piwko, M.; Wisser, F. M.; Jaumann, T.; Ahrens, E.; Giebeler, L.; Althues, H.; Schädlich, S.; Grothe, J.; Jeffery, A.; Grube, M.; Brückner, J.; Martin, J.; Eckert, J.; Kaskel, S.; Mikolajick, T.; Weber, W. M. High area capacity lithium-sulfur full-cell battery with prelithiated silicon nanowire-carbon anodes for long cycling stability. Sci. Rep. 2016, 6, 27982. (16) Lee, S. W.; McDowell, M. T.; Choi, J. W.; Cui, Y. Anomalous shape changes of silicon nanopillars by electrochemical lithiation. Nano Lett. 2011, 11, 3034−3039. (17) Green, M.; Fielder, E.; Scrosati, B.; Wachtler, M.; Moreno, J. S. Structured silicon anodes for lithium battery applications. Electrochem. Solid-State Lett. 2003, 6, A75−A79. (18) Maranchi, J. P.; Hepp, A. F.; Kumta, P. N. High capacity, reversible silicon thin-film anodes for lithium-ion batteries. Electrochem. Solid-State Lett. 2003, 6, A198−A201. (19) McDowell, M. T.; Ryu, I.; Lee, S. W.; Wang, C.; Nix, W. D.; Cui, Y. Studying the kinetics of crystalline silicon nanoparticle lithiation with in situ transmission electron microscopy. Adv. Mater. 2012, 24, 6034−6041. (20) León, B.; Corredor, J. I.; Tirado, J. L.; Pérez-Vicente, C. On the mechanism of the electrochemical reaction of tin phosphide with lithium. J. Electrochem. Soc. 2006, 153, A1829−A1834. (21) Boyanov, S.; Bernardi, J.; Bekaert, E.; Ménétrier, M.; Doublet, M.-L.; Monconduit, L. P-Redox mechanism at the origin of the high lithium storage in NiP2-based batteries. Chem. Mater. 2009, 21, 298− 308. (22) Kim, Y.-U.; Lee, C. K.; Sohn, H.-J.; Kang, T. Reaction mechanism of tin phosphide anode by mechanochemical method for lithium secondary batteries. J. Electrochem. Soc. 2004, 151, A933− A937. (23) Pralong, V.; Souza, D. C. S.; Leung, K. T.; Nazar, L. F. Reversible lithium uptake by CoP3 at low potential: role of the anion. Electrochem. Commun. 2002, 4, 516−520. (24) Poli, F.; Wong, A.; Kshetrimayum, J. S.; Monconduit, L.; Letellier, M. In situ NMR insights into the electrochemical reaction of Cu3P electrodes in lithium batteries. Chem. Mater. 2016, 28, 1787− 1793. (25) Gao, H.; Xiao, L.; Plümel, I.; Xu, G.-L.; Ren, Y.; Zuo, X.; Liu, Y.; Schulz, C.; Wiggers, H.; Amine, K.; Chen, Z. Parasitic reactions in nanosized silicon anodes for lithium-ion batteries. Nano Lett. 2017, 17, 1512−1519.

(26) Poli, F.; Kshetrimayum, J. S.; Monconduit, L.; Letellier, M. New cell design for in situ NMR studies of lithium-ion batteries. Electrochem. Commun. 2011, 13, 1293−1295. (27) Reinhold, R.; Stoeck, U.; Giebeler, L.; Eckert, J.; Kaskel, S., Silicon non-metal compounds as precursors for nanoparticle synthesis and their application in electrochemical energy storage devices. 3rd Dresden Conference “Energy in Future” and 4th Lithium-Sulfur-Workshop, Dresden, Germany, 10.-11.11.2015. (28) Reinhold, R.; Stoeck, U.; Ahrens, E.; Eckert, J.; Kaskel, S.; Giebeler, L., Silicon non-metal compounds as new promising materials for lithium ion batteries, Abstract #A01−0140, 229th ECS Meeting, San Diego, USA, 29.05.-02.06.2016.. (29) Duveau, D.; Sananes Israel, S.; Fullenwarth, J.; Cunin, F.; Monconduit, L. Pioneer study of SiP2 as negative electrode for Li- and Na-ion batteries. J. Mater. Chem. A 2016, 4, 3228−3232. (30) Kwon, H.-T.; Lee, C. K.; Jeon, K.-J.; Park, C.-M. Silicon diphosphide: a Si-based three-dimensional crystalline framework as a high-performance Li-ion battery anode. ACS Nano 2016, 10, 5701− 5709. (31) Rietveld, H. M. A profile refinement method for nuclear and magnetic structures. J. Appl. Crystallogr. 1969, 2, 65−71. (32) Roisnel, T.; Rodríquez-Carvajal, J. WinPLOTR.: A windows tool for powder diffraction pattern analysis. Mater. Sci. Forum 2001, 378− 381, 118−123. (33) Donohue, P. C.; Siemons, W. J.; Gillson, J. L. Preparation and properties of pyrite-type SiP2 and SiAs2. J. Phys. Chem. Solids 1968, 29, 807−813. (34) Brauer, G.; Zintl, E. Metals and alloys. XXIII. Constitution of phosphides, arsenides, antimonides and bismuthides of lithium, sodium and potassium. Z. Phys. Chem. 1937, 37, 323−352. (35) Vogel, U.; Brachmann, E.; Oswald, S.; Menzel, S.; Gemming, T.; Eckert, J. Evaluation of a mobile vacuum transfer system for in vacuo XPS analysis using as-deposited Ti thin-films. Vacuum 2015, 117, 81− 84. (36) Olesinski, R. W.; Kanani, N.; Abbaschian, G. J. The P-Si (Phosphorus-Silicon) System. Bull. Alloy Phase Diagrams 1985, 6, 130−133. (37) SpringThorpe, A. J. The preparation of single crystal orthorhombic SiP2. Mater. Res. Bull. 1969, 4, 125−128. (38) Liu, Q.; Gu, S.; Li, C. M. Electrodeposition of nickelphosphorus nanoparticles film as a Janus electrocatalyst for electrosplitting of water. J. Power Sources 2015, 299, 342−346. (39) Li, G.; Yang, H.; Li, F.; Du, J.; Shi, W.; Cheng, P. Facile formation of a nanostructured NiP2@C material for advanced lithiumion battery anode using adsorption property of metal-organic framework. J. Mater. Chem. A 2016, 4, 9593−9599. (40) Bachhuber, F.; Rothballer, J.; Pielnhofer, F.; Weihrich, R. First principles calculations on structure, bonding, and vibrational frequencies of SiP2. J. Chem. Phys. 2011, 135, 124508. (41) Schroder, K. J.; Celio, H.; Webb, L. J.; Stevenson, K. J. Examining solid electrolyte interphase on crystalline silicon electrodes: influence of electrochemical preparation and ambient exposure conditions. J. Phys. Chem. C 2012, 116, 19737−19747. (42) Philippe, B.; Dedryvère, R.; Allouche, J.; Lindgren, F.; Gorgoi, M.; Rensmo, H.; Gonbeau, D.; Edström, K. Nanosilicon electrodes for lithium-ion batteries: interfacial mechanisms studied by hard and soft X-ray photoelectron spectroscopy. Chem. Mater. 2012, 24, 1107− 1115. (43) Schroder, K.; Alvarado, J.; Yersak, T. A.; Li, J.; Dudney, N.; Webb, L. J.; Meng, Y. S.; Stevenson, K. J. The effect of fluoroethylene carbonate as an additive on the solid electrolyte interphase on silicon lithium-ion electrodes. Chem. Mater. 2015, 27, 5531−5542. (44) Jaumann, T.; Balach, J.; Klose, M.; Oswald, S.; Langklotz, U.; Michaelis, A.; Eckert, J.; Giebeler, L. SEI-component formation on sub 5 nm sized silicon nanoparticles in Li-ion batteries: the role of electrode preparation, FEC addition and binders. Phys. Chem. Chem. Phys. 2015, 17, 24956−24967. 7105

DOI: 10.1021/acsami.7b18697 ACS Appl. Mater. Interfaces 2018, 10, 7096−7106

Research Article

ACS Applied Materials & Interfaces (45) Mayo, M.; Griffith, K. J.; Pickard, C. J.; Morris, A. J. Ab initio study of phosphorus anodes for lithium and sodium-ion batteries. Chem. Mater. 2016, 28, 2011−2021. (46) Park, C.-M.; Sohn, H.-J. Black phosphorus and its composite for lithium rechargeable batteries. Adv. Mater. 2007, 19, 2465−2468. (47) Nazri, G. Preparation, structure and ionic conductivity of lithium phosphide. Solid State Ionics 1989, 34, 97−102. (48) Nazri, G.; Julien, C.; Mavi, H. Structure of Li3X (X = N, P, As) superionic conductors: X-ray diffraction and FTIR studies. Solid State Ionics 1994, 70−71, 137−143. (49) Wang, F.; Wu, L.; Key, B.; Yang, X.-Q.; Grey, C. P.; Zhu, Y.; Graetz, J. Electrochemical reaction of lithium with nanostructured silicon anodes: a study by in situ synchrotron X-ray diffraction and electron energy-loss spectroscopy. Adv. Energy Mater. 2013, 3, 1324− 1331. (50) Li, J.; Dahn, J. R. An in situ X-ray diffraction study of the reaction of Li with crystalline Si. J. Electrochem. Soc. 2007, 154, A156− A161. (51) Obrovac, M. N.; Krause, L. J. Reversible cycling of crystalline silicon powder. J. Electrochem. Soc. 2007, 154, A103−A108. (52) Malmgren, S.; Ciosek, K.; Hahlin, M.; Gustafsson, T.; Gorgoi, M.; Rensmo, H.; Edström, K. Comparing anode and cathode electrode/electrolyte interface composition and morphology using soft and hard X-ray photoelectron spectroscopy. Electrochim. Acta 2013, 97, 23−32. (53) Wenzel, S.; Weber, D. A.; Leichtweiss, T.; Busche, M. R.; Sann, J.; Janek, J. Interphase formation and degradation of charge transfer kinetics between a lithium metal anode and highly crystalline Li7P3S11. Solid State Ionics 2016, 286, 24−33. (54) Grosvenor, A. P.; Wik, S. D.; Cavell, R. G.; Mar, A. Examination of the bonding in binary transition-metal monophosphides MP (M = Cr, Mn, Fe, Co) by X-ray photoelectron spectroscopy. Inorg. Chem. 2005, 44, 8988−8998. (55) Aurbach, D. Review of selected electrode-solution interactions which determine the performance of Li and Li ion batteries. J. Power Sources 2000, 89, 206−218. (56) Key, B.; Bhattacharyya, R.; Morcrette, M.; Seznéc, V.; Tarascon, J.-M.; Grey, C. P. Real-time NMR investigations of structural changes in silicon electrodes for lithium-ion batteries. J. Am. Chem. Soc. 2009, 131, 9239−9249. (57) Liu, N.; Wu, H.; McDowell, M. T.; Yao, Y.; Wang, C.; Cui, Y. A Yolk-Shell design for stabilized and scalable Li-ion battery alloy anodes. Nano Lett. 2012, 12, 3315−3321. (58) Jang, J. Y.; Park, G.; Lee, S.-M.; Choi, N.-S. Functional electrolytes enhancing electrochemical performance of Sn-Fe-P alloy as anode for lithium-ion batteries. Electrochem. Commun. 2013, 35, 72−75.

7106

DOI: 10.1021/acsami.7b18697 ACS Appl. Mater. Interfaces 2018, 10, 7096−7106