Published on 30 January 2014. Downloaded by MEDICAL RESEARCH COUNCIL LABORATORY OF MOLECULAR BIOLOGY on 27/10/2014 10:28:59.
PCCP View Article Online
PAPER
Cite this: Phys. Chem. Chem. Phys., 2014, 16, 6027
View Journal | View Issue
Progress of the conversion reaction of Mn3O4 particles as a function of the depth of discharge Daisuke Yonekura,ab Etsuro Iwama,a Naoki Ota,a Masanori Muramatsu,a Morihiro Saito,a Yuki Orikasa,c Wako Naoid and Katsuhiko Naoi*abd A comprehensive investigation of the morphological and interfacial changes of Mn3O4 particles at different lithiation stages was performed in order to improve our understanding of the mechanism of the irreversible conversion reaction of Mn3O4. The micronization of Mn3O4 into a Mn–Li2O nanocomposite microstructure and the formation of a solid electrolyte interphase (SEI) on the Mn3O4 surface were carefully observed and characterized by combining high-resolution transmission electron microscopy (HRTEM), X-ray photoelectron spectroscopy (XPS), and in situ X-ray absorption fine structure (XAFS) measurements. Accumulation of a thin SEI film of 2–5 nm thickness on the surfaces of the Mn3O4 particles
Received 22nd January 2014, Accepted 30th January 2014 DOI: 10.1039/c4cp00334a
due to their catalytic decomposition was observed at a depth of discharge (DOD) of 0%. As the DOD increases from 25% to 75%, the SEI layer composed of Li2CO3 and LiF continues to grow to 20–30 nm, and Li2O nanoparticles are clearly observed. At 100% DOD, the Mn–Li2O particles with diameters of 2–5 nm become totally encapsulated within a huge organic–inorganic coating structure, while the
www.rsc.org/pccp
overall starting shape of the particles remains.
Introduction Lithium-ion batteries (LiBs) are the most widely used rechargeable devices for storing electrochemical energy because of their high volumetric energy density and excellent service life. However, improved energy and power capability continue to be in demand for consumer electronics and (more importantly) electronic automotive applications.1 Metal oxides such as VO2,2 MnxOy,3 Fe2O3,4 Fe3O4,5 CoO,6 Co3O4,7 NiO,8 CuO,9 ZnO,10 MoO3,11 SnO2,12 and SiO2,13 which can act as anode materials because of their various conversion reactions, have attracted intense interest because of their high theoretical capacity (700–1500 mA h g 1) as alternative materials replacing the conventional graphite anodes (372 mA h g 1).5 Among them, Mn3O4 has a large theoretical capacity of 937 mA h g 1.14–24 However, Mn3O4 materials exhibit poor cyclability, because the conversion reaction of Mn3O4 (Mn3O4 + 8Li+ + 8e # 3Mn + 4Li2O) is largely irreversible.18,24 There is another issue of the inappropriate voltage range for practical applications of LiBs (0–3 V vs. Li/Li+).14–24 a
Department of Applied Chemistry, Tokyo University of Agriculture & Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8558, Japan. E-mail: [email protected] b Advanced Capacitor Research Center, Tokyo University of Agriculture & Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8558, Japan c Graduate School of Human and Environmental Studies, Kyoto University, Yoshida-Nihonmatsucho, Sakyo-ku, Kyoto, Japan d Division of Art and Innovative Technologies, K & W Inc, 1-3-16-901 Higashi, Kunitachi, Tokyo 186-0002, Japan
This journal is © the Owner Societies 2014
While the number of reports on the electrochemical performance of Mn3O4 as an anode material has increased, few reports have clarified the complicated mechanism of its conversion reaction.22–24 Among these reports, the common lithiation–delithiation process is as follows: Discharge, (i) Mn3O4 + Li+ + e - LiMn3O4 [1.5–0.5 V vs. Li/Li+] (ii) LiMn3O4 + Li+ + e - Li2O + 3MnO [1.5–0.5 V vs. Li/Li+] (iii) MnO + 2Li+ + 2e - Li2O + Mn [0.5–0.0 V vs. Li/Li+] Charge, (iv) Li2O + Mn - MnO + 2Li+ + 2e [0.5–3.0 V vs. Li/Li+]. During the conversion reaction, Mn3O4 is micronized and disintegrated into fragments while it is transformed into MnO and subsequently Mn. At the same time, Li2O, which is thermodynamically stable, is inevitably produced as a side product of the Mn3O4 conversion. It is understood that a solid electrolyte interphase (SEI) forms at the interface between the active material and the electrolyte at voltages lower than 1.0 V. X. Fang et al. reported the high-resolution transmission electron microscopy (HRTEM) observation of Mn3O4 nanorods with diameters of 100–150 nm and lengths of 1–2 mm in the full lithiation and delithiation states, along with the ex situ X-ray diffraction (XRD) measurements.22 From their HRTEM images, the existence of an a-Mn crystalline phase with particle diameters of 5–10 nm was suggested as one of the end products of the Mn3O4 conversion reaction, but the XRD patterns of a-Mn
Phys. Chem. Chem. Phys., 2014, 16, 6027--6032 | 6027
Published on 30 January 2014. Downloaded by MEDICAL RESEARCH COUNCIL LABORATORY OF MOLECULAR BIOLOGY on 27/10/2014 10:28:59.
View Article Online
Paper
unfortunately overlapped with a Cu substrate peak. The existence of MnO after full delithiation was also suggested by their HRTEM images, indicating that after the delithiation process, Mn normally returned to the 2.0+ valence state. Interestingly, broad peaks attributed to the diffraction of the (111) and (220) planes of Li2O were observed in the XRD patterns of fully lithiated Mn3O4. Furthermore, H. Kim et al. reported on Mn3O4–graphene composites.23 Their Mn3O4 nanoparticles (approximately 20–30 nm) exhibited a high capacity of 1050 mA h g 1 at 0.1 C (100 mA g 1) owing to the good dispersion on the graphene. Their samples were characterized by ex situ XRD and Raman spectroscopy at different stages of lithiation and delithiation (0.01, 0.5, and 1.0 V). In the Raman spectra, peaks attributed to Li2CO3 were observed at 0.01 V, while the Mn3O4 peaks disappeared. These Li2CO3 peaks were apparently a result of the transformation of Li2O into Li2CO3 after the exposure of the samples to air and the subsequent reaction with CO2. Lowe et al. reported an excellent study by in situ XRD and X-ray absorption spectroscopy at detailed different potentials. In their report, it is indicated that during the first lithiation, lithium diffuses into Mn3O4 to form LiMn3O4, then diffusion of Li+ and O2 out of LiMn3O4 leads to the formation of MnO, and finally either manganese or oxygen ions diffuse out of MnO to create Mn clusters.24 Unlike other reports,22,23 the existence of Li2O was not observed in their XRD spectra. The coexistence of other reactions such as reversible film formation, irreversible solvent decomposition, or capacitance at the nanoparticle surfaces was also suggested. As mentioned above, there are few, but excellent, studies on the understanding of the Mn3O4 conversion reaction mechanism. However, these reports still lack detailed description of (1) the Mn micronization process, (2) change of the evaluated valence state of Mn, (3) visible localisation of conversion-derived Li2O and electrolyte-derived SEI (composed of Li2CO3 and LiF). Especially, if one is to realize a reversible conversion reaction, a better understanding of the Li2O generation is inevitable, including its location, the voltage at which it occurs, and whether or not a SEI forms. However, because of the intrinsic low electronic conductivity of Mn3O4 (1.8 10 5 S m 1),14 the addition of conductive carbon is necessary in order to evaluate Mn3O4 electrodes. As a consequence, detailed HRTEM observations of isolated Mn3O4 have not been reported. Therefore, to tackle these issues and to help have a better understanding of the lithium reaction procedure of metal oxides, we chose Mn3O4 as a model material. In this study, comprehensive and accurate observation of the morphological and interfacial changes of lone Mn3O4 particles at different charge–discharge stages was achieved by combining fine-scale HRTEM observations with in situ X-ray absorption fine structure (XAFS) and X-ray photoelectron spectroscopy (XPS) measurements.
Experimental Electrochemical characterization The electrochemical characterization was performed using Mn3O4 electrodes coated on Cu substrates. Mn3O4 electrodes were prepared by coating a thoroughly mixed N-methyl pyrrolidone
6028 | Phys. Chem. Chem. Phys., 2014, 16, 6027--6032
PCCP
(NMP) solution containing Mn3O4 particles as an active material (95 wt%) and polyvinylidene difluoride (PVdF) as a binder (5 wt%) onto a Cu foil (10 mm). The electrode was then dried at 80 1C in vacuo. The thickness of the Mn3O4 electrodes was adjusted to ca. 30 mm. The total loading mass on Cu foil was ca. 0.9 mg. CR2032-type half-cells were assembled with a Li metal electrode and the Mn3O4 electrode. The electrolyte was a mixture of ethylene carbonate (EC) and diethyl carbonate (DEC) containing 1.0 M lithium hexafluorophosphate (LiPF6) as an electrolyte salt. Charge– discharge tests were performed on assembled half-cells at 20 mA g 1 (0.02 C) in the voltage range of 0–3.3 V vs. Li/Li+. Microscopy and spectroscopy The detailed crystal nanostructure of Mn3O4 was observed by HRTEM (HITACHI H-9500) at different voltages from 0 to 3.3 V. Prior to the HRTEM observation, the Mn3O4 electrode was removed from the Cu foil, ultrasonically treated in DEC, and finally dropped onto a microgrid (Okenshoji, type B Cu150 mm). The composition of the compounds deposited on the Mn3O4 electrodes was analyzed by XPS (ESCA-3400, Shimadzu Corp.) using the MgKa line as the X-ray source. In situ XAFS measurements at the Mn K-edges were performed in a transmission mode for the composite samples at the beam line BL01B1 of the synchrotron radiation facility, Spring-8 (Hyogo, JAPAN). In order to obtain in situ XAFS data with a reasonable quality, the thickness of the electrodes was adjusted by adding super-growth carbon nanotubes (SGCNTs) and Ketjen black (KB) to the Mn3O4 electrode (total loading mass of ca. 1.0 mg). Using this Mn3O4–KB–SGCNT (60/36/4) mixed electrode, laminate-type two-electrode cells were assembled with lithium metal foil. A discharge test was performed on the assembled laminate cell in the voltage range of 0–3.3 V during the first discharge at a rate of 0.02 C. The XAFS spectra were recorded at an equilibrium state, i.e., at a stabilized voltage after a rest period for each measuring point. The obtained XAFS spectra were analyzed using the spectral fitting software REX2000 (Rigaku Corp.) to evaluate the ratios of Mn species with different valance states such as Mn0, Mn2+, and Mn3+.
Results and discussion In order to avoid the influence of conductive carbon, we deliberately prepared electrodes composed of only Mn3O4 and a PVdF binder. A discharge curve for the Mn3O4 electrode is shown in Fig. 1. From the open-circuit voltage (OCV; point A), the voltage rapidly dropped to 0.3 V, and then a small bump was observed before the voltage reached a plateau. The large voltage drop and the small bump are considered to be due to the overpotential of the initial reaction of Mn3O4 without the presence of conductive carbon (Mn3O4 + Li+ + e - LiMn3O4 and LiMn3O4 + 1.6Li+ + 1.6e - 1.3Li2O + 3MnO). In the plateau region, from point B at a depth of discharge (DOD) of 25% to point D at a DOD of 75%, the conversion reaction from MnO into Mn and Li2O proceeds. After the plateau, at DOD = 75%, the voltage continuously decreased down to 0 V (point E).
This journal is © the Owner Societies 2014
Published on 30 January 2014. Downloaded by MEDICAL RESEARCH COUNCIL LABORATORY OF MOLECULAR BIOLOGY on 27/10/2014 10:28:59.
View Article Online
PCCP
Paper
Fig. 1 First discharge curve of Mn3O4/PVdF on a half-cell with a Li/1 M LiPF6(EC + DEC)/Mn3O4 structure in the voltage range of 0–3.3 V at a current density of 20 mA g 1 (0.02 C). Points A (0% DOD), B (25% DOD), C (50% DOD), D (75% DOD), and E (100% DOD) denote the stages at which XAFS, XPS, and HRTEM measurements were carried out.
DOD = 0–25% At DOD = 0%, the XANES and EXAFS spectra correspond well to the Mn3O4 reference spectra, showing that Mn did not change immediately after the contact with the electrolyte (Fig. 2(a)–(c)). The valence number of Mn before cycling is 2.66+, which is equal to that of Mn3O4. However, from the O1s XPS spectra at a 0% DOD, the peak attributed to –OH at 535 eV was observed in addition to the Mn3O4 peak at 532.5 eV (Fig. 3). This –OH peak was not observed from the surface of pristine Mn3O4 particles, suggesting that a protective film may have formed on Mn3O4 through decomposition of the electrolytes due to the catalytic effect of Mn3O4. Comparison of the HRTEM images of the pristine particles and the particles at 0% DOD supports the above XPS results (Fig. 4(a)–(d)). A thin film of 2–5 nm thickness accumulated on the surface of Mn3O4 particles at 0% DOD, and is considered to be a SEI. Meanwhile, the particle morphology and high crystallinity did not change after the material was soaked in the electrolyte. Beyond a DOD of 25%, the XANES spectra drastically changed from those observed at 0% DOD. The pre-edge and main peak shifted to lower energies as the electron density on Mn increased during reduction of Mn3O4, which corresponds well to previously reported results.24 These XANES spectra and EXAFS spectra correspond well to the MnO reference spectra, suggesting that Mn changed from a valence of 2.66+ to 2.0+. This indicates that the transformation of Mn3O4 into MnO was complete at a DOD of 25%. In the C1s XPS spectra at 25% DOD, a peak appeared at around 290.5 eV that was attributed to Li2CO3, which is possibly a product of the reduction of EC and DEC (Fig. 3). This result corresponds well with the appearance of the peak at around 56 eV in the Li1s spectra, which is attributed to Li2CO3 and Li2O. These results and a decrease in the intensity of the peaks in the Mn2p spectra suggest that the surface of MnO particles started to be covered by a new SEI layer mainly composed of Li2CO3 with some coexisting Li2O. At this stage, the possibility of LiF formation, which is another main SEI compound formed by decomposition of LiPF6, can be excluded because no obvious peak was observed in the F1s spectra. A magnified HRTEM image (inset of Fig. 5(e)) supports the
This journal is © the Owner Societies 2014
Fig. 2 Evaluated changes in the valence state of Mn vs. DOD (top) and normalized Mn K-edge XANES and EXAFS spectra of the Mn3O4–KB composite during the first discharge (bottom). The XANES spectra of Mn3O4, MnO, and Mn foil are included as a reference for valence states of 2.66+, 2+, and 0, respectively. At DOD = 0%, before cycling, the spectra were identical to the Mn3O4 reference spectra, suggesting that Mn itself did not change just after contact with the electrolyte. At 25% DOD, the obtained spectra corresponded well to the MnO reference spectra. At 75% DOD, the XANES spectra have more-or-less the same shape as the Mn reference spectra, and the average valence number of Mn was 0. In the EXAFS spectra, before cycling, the spectrum was the same as that of Mn3O4, suggesting that no crystallographic structure change occurred. At 25% DOD, the positions of the first peak (corresponding to Mn–O) and the second peak (Mn–Mn) remained unchanged, while the peak intensities diminished. The decrease in the intensity without any shift in the peak position suggests a decrease in the coordination number, reflecting distortion of the Mn3O4 crystal structure or a disordered structure. At 75% DOD, there are no clear peaks in the obtained spectra. This corresponds well to the Mn reference spectra.
formation of a new layer as the thickness of the surface layer increases from 2–5 to 5–10 nm. This magnified HRTEM image also suggests lattice distortion due to the change in structure from a mixture of tetrahedral and octahedral Mn3O4 to octahedral MnO. DOD = 25–50% As shown in Fig. 2(a), a linear decrease in the Mn valence number continued from DOD = 25% to 50% as further lithiation of MnO and transformation into metallic Mn proceeded. The progress of the lithiation is supported by the decrease in the EXAFS peak intensity from DOD = 25% to 50%, which indicates an increase in the lattice distortion and a decrease in the coordination number of the Mn atoms. The peak intensity in the Mn2p XPS spectra further decreased beyond 25% DOD, suggesting that the thickness of the SEI layer around MnO and Mn increased (Fig. 3). In the C1s and F1s spectra, no obvious change was observed between DOD = 25% and 50%, but the peak intensity at 56 eV in the Li1s spectra increased while the Mn3O4 peaks at 644 and 656 eV decreased. These results indicate that the SEI layer growth continues without any composition change, and
Phys. Chem. Chem. Phys., 2014, 16, 6027--6032 | 6029
Published on 30 January 2014. Downloaded by MEDICAL RESEARCH COUNCIL LABORATORY OF MOLECULAR BIOLOGY on 27/10/2014 10:28:59.
View Article Online
Paper
PCCP
Fig. 4 High- and low-magnification HRTEM images of pristine Mn3O4 ((a) and (b)) and of the electrode material at 0% DOD ((c) and (d)). The difference between pristine Mn3O4 and that at 0% DOD is that the latter had been dipped in the electrolyte (1 M LiPF6/EC + DEC). HRTEM images of pristine Mn3O4 show highly crystalline particles with diameters of 20–300 nm. The particle morphology and high crystallinity did not change after the particles were soaked in the electrolyte, while a thin film of 2–5 nm thickness accumulated. This film is considered an SEI (4(d)). Fig. 3 XPS spectra (C1s, F1s, Li1s, Mn2p, and O1s) with possible peak assignments of the Mn3O4 electrode surface at different DOD stages. In the O1s spectrum obtained before cycling, a peak attributed to –OH at 535 eV was observed. This result suggests that the –OH film was formed on Mn3O4 just after contact with the electrolyte, which corresponds well to the HRTEM observations (Fig. 4(d)). As the DOD increased, the –OH peak gradually decreased in intensity and disappeared at a 75% DOD. In the Mn2p spectra, the Mn peaks at 644 and 656 eV also decreased in intensity with increasing DOD and finally disappeared at a DOD of 75%. This supports the idea that another film forms on the Mn3O4. At the same DOD of 75%, the C1s peak for Li2CO3 at 290.5 eV, the F1s peak for LiF at 686 eV, and the Li1s peak for Li2CO3, LiF, and Li2O at 56 eV are clearly observed. These peaks suggest that the new material is composed of a mixture of these Li2CO3, LiF, and Li2O compounds.
Li2O particles clearly appear. In the HRTEM images in Fig. 5(g) and (h), a thickened surface film (10–20 nm) is observed around the particles. Furthermore, disintegration within the particles is clearly observed, resulting in nanograins of 2–5 nm, although the global shape of the initial particle is preserved. The above results suggest that the metallic Mn nanoparticles and unreacted MnO particles are dispersed uniformly with Li2O on a nanoscale level and are covered by a thick SEI layer mainly composed of Li2CO3. DOD = 50–75% At the end of the plateau at DOD = 75%, the lithiation of MnO is completed as the Mn valence state drops to 0.0 (Fig. 2(a)). The XANES spectra correspond well to that of the metallic Mn. In the C1s spectra, the intensity of the peak attributed to Li2CO3 further increases, while at the same time the peak at 686 eV in the F1s spectra becomes pronounced, indicating the formation of LiF.
6030 | Phys. Chem. Chem. Phys., 2014, 16, 6027--6032
Fig. 5 High- and low-magnification HRTEM images of Mn3O4 after discharging to DODs of 25% ((e) and (f)) and 50% ((g) and (h)). The image of Mn3O4 at DOD = 25% shows distortion of the Mn3O4 lattice (f), although the overall shape of the starting particles is preserved (e). In addition, the thickness of the SEI film apparently increased to 10 nm (e). At a DOD of 50%, the thickness of the SEI obviously increased (g), and black dots of less than 5 nm diameter began to appear. Meanwhile the lattice stripes of crystalline Mn3O4 became unclear, probably because of the thicker SEI layer (h).
This journal is © the Owner Societies 2014
Published on 30 January 2014. Downloaded by MEDICAL RESEARCH COUNCIL LABORATORY OF MOLECULAR BIOLOGY on 27/10/2014 10:28:59.
View Article Online
PCCP
The disappearance of the Mn peaks in the Mn2p spectra suggests that the Mn nanoparticles are completely surrounded by the thickened SEI layer. HRTEM images taken at this stage revealed a nanocomposite microstructure consisting of Mn–Li2O completely covered by a further thickened SEI layer (20–30 nm thickness). The unclear nanoparticles in the images compared to those observed at DOD = 50% also support the complete covering of the nanoparticles by the thick SEI layer. These results suggest that Mn may have accelerated the decomposition kinetics of the electrolytes. DOD = 75–100% From DOD = 75% to 100%, the XANES spectra remained unchanged and almost overlapped that of metallic Mn, showing that the valence number remained constant, close to 0.0 (Fig. 2). In the EXAFS spectra, the peak from Mn metal at around 2.2 Å (Mn–Mn) was observed, while overall the peaks are ambiguous, indicating a disordered structure containing Mn nanograins. This result suggests that this electrochemical reaction, which consumed the last 25% of the discharge, was not related to the Mn species but instead to the further decomposition of the electrolyte. In the XPS spectra, however, no obvious change was observed between DOD = 75% and 100%, suggesting that the composition of SEI did not change as its thickness increased (Fig. 3). This idea is supported by the HRTEM images, which show thickened and merged SEI layers that look like polymeric networks completely surrounding the Mn–Li2O microstructures, which can no longer be observed (Fig. 6(k)).
Fig. 6 High- and low-magnification HRTEM images of Mn3O4 after discharging to DOD = 75% ((i) and (j)) and 100% ((k) and (l)). At a DOD of 75% (i), the surfaces of the Mn3O4 particles were further covered by the SEI layer, while the composition of the SEI layer mostly did not change beyond DOD = 50%. At DOD = 100%, the thickness of the SEI remained unchanged (k), while the boundary between the Mn3O4 particles and the SEI layer became unclear because of the complete mix of the two compounds (l). The starting Mn3O4 particle shapes remained despite the total decomposition of the starting 100 nm particles into particles with diameters of 5 nm or smaller.
This journal is © the Owner Societies 2014
Paper
These results are consistent with reports of Tarascon’s group and other groups that discharged metal oxides such as CoO and RuO2 form a polymer/gel-like coating.25–28 We believe that such encapsulation of dispersed Mn–Li2O nanoparticles within the merged SEI layer isolates Mn and Li2O from the electronic path/ contact. The isolation of active materials hinders the reversible reaction of Mn into Mn3O4, limiting the recovery of the Mn valence state to 2.0+, as reported by Lowe et al. and others.22,24
Conclusions The Mn3O4 conversion reaction was investigated at different DODs by a combination of in situ XAFS, XPS, and HRTEM observations to clarify how the reaction proceeds from the crystallographic and interfacial viewpoints. The results are summarized in separate illustrations of the crystallographic structure and the surface morphology in Fig. 7. Crystallographic structure changes Starting from a highly crystalline Mn3O4 particle with a diameter of 20–300 nm, as seen in the HRTEM images taken at DOD = 0% after the particle was soaked in the electrolyte (Fig. 7, point A), the
Fig. 7 Illustrations of the material transformation of Mn3O4 during the first discharge. Crystallographic changes of Mn3O4 (left) and morphological changes of Mn3O4 particles (right) are shown for different DOD stages from DOD = 0% (point A) to DOD = 100% (point E).
Phys. Chem. Chem. Phys., 2014, 16, 6027--6032 | 6031
Published on 30 January 2014. Downloaded by MEDICAL RESEARCH COUNCIL LABORATORY OF MOLECULAR BIOLOGY on 27/10/2014 10:28:59.
View Article Online
Paper
material is reduced to MnO during the transition from DOD = 0% (point A) to 25% (B). During this transition, the MnO particles, whose valence number was confirmed to be 2.0+ by in situ XAFS measurements, retain their whole-particle morphologies but begin to exhibit partial distortion due to the structural change from a mixture of tetrahedral and octahedral Mn3O4 to octahedral MnO. At DOD = 50%, the lithiation of MnO, transformation into Mn metal, and production of Li2O proceed. The crystallographic structure of this material should be the most complicated among those at the various DOD stages, owing to the existence of three different phase compositions, Mn, MnO, and Li2O. Thus, the lattice distortion and a decrease in the coordination number of Mn occur, which is seen as a decrease in the EXAFS peak intensity. At DOD = 75% (point D), the transformation from MnO into Mn metal and Li2O is complete. At DOD = 100% (point E), the crystallographic state remains unchanged from that observed at DOD = 75% (D), with the same mixture of Mn metal and Li2O. Morphological changes At the initial stage, with DOD = 0% (point A), after contact with the 1 M LiPF6/EC + DEC electrolyte, a very thin protective hydroxidebased film of 2–5 nm thickness is formed by decomposition due to the catalytic effect of Mn3O4 particles. The thickness of the film increases from 2–5 to 5–10 nm up to DOD = 25% (point B), while the main component of the film changes from hydroxide to Li2CO3, which is a product of the reduction of the carbonate solvents (EC and DEC). At this stage, the particles start to show wrinkles on portions of their surfaces due to the crystallographic transformation. The particle morphology clearly changes at DOD = 50% (point C), where 2–5 nm Mn nanograins appear along with amorphous Li2O from the reduced MnO. Meanwhile, the starting particle begins to crack because of the distortion due to the crystallographic transformation. The thickness of the SEI film further increases from 5–10 to 10–20 nm, without an accompanying composition change. The SEI composition begins to change between DOD = 50% and 75% (D), where a new peak attributed to LiF is observed in the XPS spectra along with a Li2CO3 peak. The mixed Li2CO3–LiF SEI continues to grow to 20–30 nm thickness and eventually covers the entire structure consisting of Mn–Li2O nanoparticles. At this stage, the average valence state of Mn is almost 0, suggesting the termination of the Mn-related reaction. At 0 V vs. Li/Li+ (DOD = 100%, point E), the Li2CO3–LiF mixed SEI grows without changing in composition. Its precise thickness, however, is difficult to determine from the HRTEM images, because it has merged over the particle surfaces. This merged SEI covers the microstructure consisting of Mn/Li2O particles of a few nm size, which are now totally encapsulated within a huge organic–inorganic coating structure. However, the overall shape of the starting particles is retained.
Acknowledgements The synchrotron radiation experiments were performed at Spring-8 with the approval of the Japan Synchrotron Radiation Research Institute (JASRI). This work was supported by JSPS KAKENHI Grant Number 25249140.
6032 | Phys. Chem. Chem. Phys., 2014, 16, 6027--6032
PCCP
References 1 M. Armand and J. M. Tarascon, Nature, 2008, 451, 652–657. 2 K.-C. Zhou, D.-M. Cao and Z.-Y. Li, Trans. Nonferrous Met. Soc. China, 2006, 16, 517–521. 3 Z.-H. Wang, L.-X. Yuan, Q.-G. Shao, F. Huang and Y.-H. Huang, Mater. Lett., 2012, 80, 110–113. 4 M. F. Hassan, Z. Guo, Z. Chen and H. Liu, Mater. Res. Bull., 2011, 46, 858–864. 5 P. Lian, X. Zhu, H. Xiang, Z. Li, W. Yang and H. Wang, Electrochim. Acta, 2010, 56, 834–840. 6 H. Qiao, L. Xiao, Z. Zheng, H. Liu, F. Jia and L. Zhang, J. Power Sources, 2008, 185, 486–491. 7 H. Kim, D.-H. Seo, S.-W. Kim, J. Kim and K. Kang, Carbon, 2011, 49, 326–332. 8 X. H. Huang, J. P. Tu, C. Q. Zhang, X. T. Chen, Y. F. Yuan and H. M. Wu, Electrochim. Acta, 2007, 52, 4177–4181. 9 H. Wang, Q. Pan, J. Zhao and W. Chen, J. Alloys Compd., 2009, 476, 408–413. 10 X. H. Huang, X. H. Xia, Y. F. Yuan and F. Zhou, Electrochim. Acta, 2011, 56, 4960–4965. 11 M. F. Hassan, Z. P. Guo, Z. Chen and H. K. Liu, J. Power Sources, 2010, 195, 2372–2376. 12 M. Winter and J. O. Besenhard, Electrochim. Acta, 1999, 45, 31–50. 13 B. Guo, J. Shu, Z. Wang, H. Yang, L. Shi, Y. Liu and L. Chen, Electrochem. Commun., 2008, 10, 1876–1878. 14 H. Dhaouadi, A. Mandani and F. Touati, Mater. Lett., 2010, 64, 2395–2398. 15 S.-Y. Liu, J. Xie, Y.-X. Zheng, G.-S. Cao, T.-J. Zhu and X.-B. Zhao, Electrochim. Acta, 2012, 66, 271–278. 16 H. Wang, L.-F. Cui, Y. Yang, H. S. Casalongue, J. T. Robinson, Y. Liang, Y. Cui and H. Dai, J. Am. Chem. Soc., 2010, 132, 13978–13980. 17 N. Lavoie, P. R. L. Malenfant, F. M. Courtel, Y. Abu-Lebdeh and I. J. Davidson, J. Power Sources, 2012, 213, 249–254. 18 G. Yang, Y. Li, H. Ji, H. Wang, P. Gao, L. Wang, H. Liu, J. Pinto and X. Jiang, J. Power Sources, 2012, 216, 353–362. ˜ a, Chem. Mater., 2011, 23, 19 J. Gao, M. A. Lowe and H. D. Abrun 3223–3227. 20 Z. H. Wang, L.-X. Yuan, Q.-G. Shao, F. Huang and Y.-H. Huang, Mater. Lett., 2012, 80, 110–113. 21 D. Pasero, N. Reeves and A. R. West, J. Power Sources, 2005, 141, 156–158. 22 X. Fang, X. Lu, X. Guo, Y. Mao, Y.-S. Hu, J. Wang, Z. Wang, F. Wu, H. Liu and L. Chen, Electrochem. Commun., 2010, 12, 1520–1523. 23 H. Kim, S.-W. Kim, J. Hong, Y.-U. Park and K. Kang, J. Mater. Res., 2011, 26, 2665–2671. 24 M. A. Lowe, J. Gao and H. D. Abruna, J. Mater. Chem. A, 2013, 1, 2094–2103. 25 P. Poizot, S. Laruelle, S. Grugeon, L. Dupont and J. M. Tarascon, Nature, 2000, 407, 496–499. 26 J. M. Tarascon, S. Grugeon, M. Morcrette, S. Laruelle, P. Rozier and P. Poizot, C. R. Chim, 2005, 8, 9–15. 27 S. Gurgeon, S. Laruelle, L. Dupont and J.-M. Tarascon, Solid State Sci., 2003, 5, 895–904. 28 P. Balaya, H. Li, L. Kienle and J. Maier, Adv. Funct. Mater., 2003, 13, 621–625.
This journal is © the Owner Societies 2014
PCCP View Article Online
PAPER
Cite this: Phys. Chem. Chem. Phys., 2014, 16, 6027
View Journal | View Issue
Progress of the conversion reaction of Mn3O4 particles as a function of the depth of discharge Daisuke Yonekura,ab Etsuro Iwama,a Naoki Ota,a Masanori Muramatsu,a Morihiro Saito,a Yuki Orikasa,c Wako Naoid and Katsuhiko Naoi*abd A comprehensive investigation of the morphological and interfacial changes of Mn3O4 particles at different lithiation stages was performed in order to improve our understanding of the mechanism of the irreversible conversion reaction of Mn3O4. The micronization of Mn3O4 into a Mn–Li2O nanocomposite microstructure and the formation of a solid electrolyte interphase (SEI) on the Mn3O4 surface were carefully observed and characterized by combining high-resolution transmission electron microscopy (HRTEM), X-ray photoelectron spectroscopy (XPS), and in situ X-ray absorption fine structure (XAFS) measurements. Accumulation of a thin SEI film of 2–5 nm thickness on the surfaces of the Mn3O4 particles
Received 22nd January 2014, Accepted 30th January 2014 DOI: 10.1039/c4cp00334a
due to their catalytic decomposition was observed at a depth of discharge (DOD) of 0%. As the DOD increases from 25% to 75%, the SEI layer composed of Li2CO3 and LiF continues to grow to 20–30 nm, and Li2O nanoparticles are clearly observed. At 100% DOD, the Mn–Li2O particles with diameters of 2–5 nm become totally encapsulated within a huge organic–inorganic coating structure, while the
www.rsc.org/pccp
overall starting shape of the particles remains.
Introduction Lithium-ion batteries (LiBs) are the most widely used rechargeable devices for storing electrochemical energy because of their high volumetric energy density and excellent service life. However, improved energy and power capability continue to be in demand for consumer electronics and (more importantly) electronic automotive applications.1 Metal oxides such as VO2,2 MnxOy,3 Fe2O3,4 Fe3O4,5 CoO,6 Co3O4,7 NiO,8 CuO,9 ZnO,10 MoO3,11 SnO2,12 and SiO2,13 which can act as anode materials because of their various conversion reactions, have attracted intense interest because of their high theoretical capacity (700–1500 mA h g 1) as alternative materials replacing the conventional graphite anodes (372 mA h g 1).5 Among them, Mn3O4 has a large theoretical capacity of 937 mA h g 1.14–24 However, Mn3O4 materials exhibit poor cyclability, because the conversion reaction of Mn3O4 (Mn3O4 + 8Li+ + 8e # 3Mn + 4Li2O) is largely irreversible.18,24 There is another issue of the inappropriate voltage range for practical applications of LiBs (0–3 V vs. Li/Li+).14–24 a
Department of Applied Chemistry, Tokyo University of Agriculture & Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8558, Japan. E-mail: [email protected] b Advanced Capacitor Research Center, Tokyo University of Agriculture & Technology, 2-24-16 Naka-cho, Koganei, Tokyo 184-8558, Japan c Graduate School of Human and Environmental Studies, Kyoto University, Yoshida-Nihonmatsucho, Sakyo-ku, Kyoto, Japan d Division of Art and Innovative Technologies, K & W Inc, 1-3-16-901 Higashi, Kunitachi, Tokyo 186-0002, Japan
This journal is © the Owner Societies 2014
While the number of reports on the electrochemical performance of Mn3O4 as an anode material has increased, few reports have clarified the complicated mechanism of its conversion reaction.22–24 Among these reports, the common lithiation–delithiation process is as follows: Discharge, (i) Mn3O4 + Li+ + e - LiMn3O4 [1.5–0.5 V vs. Li/Li+] (ii) LiMn3O4 + Li+ + e - Li2O + 3MnO [1.5–0.5 V vs. Li/Li+] (iii) MnO + 2Li+ + 2e - Li2O + Mn [0.5–0.0 V vs. Li/Li+] Charge, (iv) Li2O + Mn - MnO + 2Li+ + 2e [0.5–3.0 V vs. Li/Li+]. During the conversion reaction, Mn3O4 is micronized and disintegrated into fragments while it is transformed into MnO and subsequently Mn. At the same time, Li2O, which is thermodynamically stable, is inevitably produced as a side product of the Mn3O4 conversion. It is understood that a solid electrolyte interphase (SEI) forms at the interface between the active material and the electrolyte at voltages lower than 1.0 V. X. Fang et al. reported the high-resolution transmission electron microscopy (HRTEM) observation of Mn3O4 nanorods with diameters of 100–150 nm and lengths of 1–2 mm in the full lithiation and delithiation states, along with the ex situ X-ray diffraction (XRD) measurements.22 From their HRTEM images, the existence of an a-Mn crystalline phase with particle diameters of 5–10 nm was suggested as one of the end products of the Mn3O4 conversion reaction, but the XRD patterns of a-Mn
Phys. Chem. Chem. Phys., 2014, 16, 6027--6032 | 6027
Published on 30 January 2014. Downloaded by MEDICAL RESEARCH COUNCIL LABORATORY OF MOLECULAR BIOLOGY on 27/10/2014 10:28:59.
View Article Online
Paper
unfortunately overlapped with a Cu substrate peak. The existence of MnO after full delithiation was also suggested by their HRTEM images, indicating that after the delithiation process, Mn normally returned to the 2.0+ valence state. Interestingly, broad peaks attributed to the diffraction of the (111) and (220) planes of Li2O were observed in the XRD patterns of fully lithiated Mn3O4. Furthermore, H. Kim et al. reported on Mn3O4–graphene composites.23 Their Mn3O4 nanoparticles (approximately 20–30 nm) exhibited a high capacity of 1050 mA h g 1 at 0.1 C (100 mA g 1) owing to the good dispersion on the graphene. Their samples were characterized by ex situ XRD and Raman spectroscopy at different stages of lithiation and delithiation (0.01, 0.5, and 1.0 V). In the Raman spectra, peaks attributed to Li2CO3 were observed at 0.01 V, while the Mn3O4 peaks disappeared. These Li2CO3 peaks were apparently a result of the transformation of Li2O into Li2CO3 after the exposure of the samples to air and the subsequent reaction with CO2. Lowe et al. reported an excellent study by in situ XRD and X-ray absorption spectroscopy at detailed different potentials. In their report, it is indicated that during the first lithiation, lithium diffuses into Mn3O4 to form LiMn3O4, then diffusion of Li+ and O2 out of LiMn3O4 leads to the formation of MnO, and finally either manganese or oxygen ions diffuse out of MnO to create Mn clusters.24 Unlike other reports,22,23 the existence of Li2O was not observed in their XRD spectra. The coexistence of other reactions such as reversible film formation, irreversible solvent decomposition, or capacitance at the nanoparticle surfaces was also suggested. As mentioned above, there are few, but excellent, studies on the understanding of the Mn3O4 conversion reaction mechanism. However, these reports still lack detailed description of (1) the Mn micronization process, (2) change of the evaluated valence state of Mn, (3) visible localisation of conversion-derived Li2O and electrolyte-derived SEI (composed of Li2CO3 and LiF). Especially, if one is to realize a reversible conversion reaction, a better understanding of the Li2O generation is inevitable, including its location, the voltage at which it occurs, and whether or not a SEI forms. However, because of the intrinsic low electronic conductivity of Mn3O4 (1.8 10 5 S m 1),14 the addition of conductive carbon is necessary in order to evaluate Mn3O4 electrodes. As a consequence, detailed HRTEM observations of isolated Mn3O4 have not been reported. Therefore, to tackle these issues and to help have a better understanding of the lithium reaction procedure of metal oxides, we chose Mn3O4 as a model material. In this study, comprehensive and accurate observation of the morphological and interfacial changes of lone Mn3O4 particles at different charge–discharge stages was achieved by combining fine-scale HRTEM observations with in situ X-ray absorption fine structure (XAFS) and X-ray photoelectron spectroscopy (XPS) measurements.
Experimental Electrochemical characterization The electrochemical characterization was performed using Mn3O4 electrodes coated on Cu substrates. Mn3O4 electrodes were prepared by coating a thoroughly mixed N-methyl pyrrolidone
6028 | Phys. Chem. Chem. Phys., 2014, 16, 6027--6032
PCCP
(NMP) solution containing Mn3O4 particles as an active material (95 wt%) and polyvinylidene difluoride (PVdF) as a binder (5 wt%) onto a Cu foil (10 mm). The electrode was then dried at 80 1C in vacuo. The thickness of the Mn3O4 electrodes was adjusted to ca. 30 mm. The total loading mass on Cu foil was ca. 0.9 mg. CR2032-type half-cells were assembled with a Li metal electrode and the Mn3O4 electrode. The electrolyte was a mixture of ethylene carbonate (EC) and diethyl carbonate (DEC) containing 1.0 M lithium hexafluorophosphate (LiPF6) as an electrolyte salt. Charge– discharge tests were performed on assembled half-cells at 20 mA g 1 (0.02 C) in the voltage range of 0–3.3 V vs. Li/Li+. Microscopy and spectroscopy The detailed crystal nanostructure of Mn3O4 was observed by HRTEM (HITACHI H-9500) at different voltages from 0 to 3.3 V. Prior to the HRTEM observation, the Mn3O4 electrode was removed from the Cu foil, ultrasonically treated in DEC, and finally dropped onto a microgrid (Okenshoji, type B Cu150 mm). The composition of the compounds deposited on the Mn3O4 electrodes was analyzed by XPS (ESCA-3400, Shimadzu Corp.) using the MgKa line as the X-ray source. In situ XAFS measurements at the Mn K-edges were performed in a transmission mode for the composite samples at the beam line BL01B1 of the synchrotron radiation facility, Spring-8 (Hyogo, JAPAN). In order to obtain in situ XAFS data with a reasonable quality, the thickness of the electrodes was adjusted by adding super-growth carbon nanotubes (SGCNTs) and Ketjen black (KB) to the Mn3O4 electrode (total loading mass of ca. 1.0 mg). Using this Mn3O4–KB–SGCNT (60/36/4) mixed electrode, laminate-type two-electrode cells were assembled with lithium metal foil. A discharge test was performed on the assembled laminate cell in the voltage range of 0–3.3 V during the first discharge at a rate of 0.02 C. The XAFS spectra were recorded at an equilibrium state, i.e., at a stabilized voltage after a rest period for each measuring point. The obtained XAFS spectra were analyzed using the spectral fitting software REX2000 (Rigaku Corp.) to evaluate the ratios of Mn species with different valance states such as Mn0, Mn2+, and Mn3+.
Results and discussion In order to avoid the influence of conductive carbon, we deliberately prepared electrodes composed of only Mn3O4 and a PVdF binder. A discharge curve for the Mn3O4 electrode is shown in Fig. 1. From the open-circuit voltage (OCV; point A), the voltage rapidly dropped to 0.3 V, and then a small bump was observed before the voltage reached a plateau. The large voltage drop and the small bump are considered to be due to the overpotential of the initial reaction of Mn3O4 without the presence of conductive carbon (Mn3O4 + Li+ + e - LiMn3O4 and LiMn3O4 + 1.6Li+ + 1.6e - 1.3Li2O + 3MnO). In the plateau region, from point B at a depth of discharge (DOD) of 25% to point D at a DOD of 75%, the conversion reaction from MnO into Mn and Li2O proceeds. After the plateau, at DOD = 75%, the voltage continuously decreased down to 0 V (point E).
This journal is © the Owner Societies 2014
Published on 30 January 2014. Downloaded by MEDICAL RESEARCH COUNCIL LABORATORY OF MOLECULAR BIOLOGY on 27/10/2014 10:28:59.
View Article Online
PCCP
Paper
Fig. 1 First discharge curve of Mn3O4/PVdF on a half-cell with a Li/1 M LiPF6(EC + DEC)/Mn3O4 structure in the voltage range of 0–3.3 V at a current density of 20 mA g 1 (0.02 C). Points A (0% DOD), B (25% DOD), C (50% DOD), D (75% DOD), and E (100% DOD) denote the stages at which XAFS, XPS, and HRTEM measurements were carried out.
DOD = 0–25% At DOD = 0%, the XANES and EXAFS spectra correspond well to the Mn3O4 reference spectra, showing that Mn did not change immediately after the contact with the electrolyte (Fig. 2(a)–(c)). The valence number of Mn before cycling is 2.66+, which is equal to that of Mn3O4. However, from the O1s XPS spectra at a 0% DOD, the peak attributed to –OH at 535 eV was observed in addition to the Mn3O4 peak at 532.5 eV (Fig. 3). This –OH peak was not observed from the surface of pristine Mn3O4 particles, suggesting that a protective film may have formed on Mn3O4 through decomposition of the electrolytes due to the catalytic effect of Mn3O4. Comparison of the HRTEM images of the pristine particles and the particles at 0% DOD supports the above XPS results (Fig. 4(a)–(d)). A thin film of 2–5 nm thickness accumulated on the surface of Mn3O4 particles at 0% DOD, and is considered to be a SEI. Meanwhile, the particle morphology and high crystallinity did not change after the material was soaked in the electrolyte. Beyond a DOD of 25%, the XANES spectra drastically changed from those observed at 0% DOD. The pre-edge and main peak shifted to lower energies as the electron density on Mn increased during reduction of Mn3O4, which corresponds well to previously reported results.24 These XANES spectra and EXAFS spectra correspond well to the MnO reference spectra, suggesting that Mn changed from a valence of 2.66+ to 2.0+. This indicates that the transformation of Mn3O4 into MnO was complete at a DOD of 25%. In the C1s XPS spectra at 25% DOD, a peak appeared at around 290.5 eV that was attributed to Li2CO3, which is possibly a product of the reduction of EC and DEC (Fig. 3). This result corresponds well with the appearance of the peak at around 56 eV in the Li1s spectra, which is attributed to Li2CO3 and Li2O. These results and a decrease in the intensity of the peaks in the Mn2p spectra suggest that the surface of MnO particles started to be covered by a new SEI layer mainly composed of Li2CO3 with some coexisting Li2O. At this stage, the possibility of LiF formation, which is another main SEI compound formed by decomposition of LiPF6, can be excluded because no obvious peak was observed in the F1s spectra. A magnified HRTEM image (inset of Fig. 5(e)) supports the
This journal is © the Owner Societies 2014
Fig. 2 Evaluated changes in the valence state of Mn vs. DOD (top) and normalized Mn K-edge XANES and EXAFS spectra of the Mn3O4–KB composite during the first discharge (bottom). The XANES spectra of Mn3O4, MnO, and Mn foil are included as a reference for valence states of 2.66+, 2+, and 0, respectively. At DOD = 0%, before cycling, the spectra were identical to the Mn3O4 reference spectra, suggesting that Mn itself did not change just after contact with the electrolyte. At 25% DOD, the obtained spectra corresponded well to the MnO reference spectra. At 75% DOD, the XANES spectra have more-or-less the same shape as the Mn reference spectra, and the average valence number of Mn was 0. In the EXAFS spectra, before cycling, the spectrum was the same as that of Mn3O4, suggesting that no crystallographic structure change occurred. At 25% DOD, the positions of the first peak (corresponding to Mn–O) and the second peak (Mn–Mn) remained unchanged, while the peak intensities diminished. The decrease in the intensity without any shift in the peak position suggests a decrease in the coordination number, reflecting distortion of the Mn3O4 crystal structure or a disordered structure. At 75% DOD, there are no clear peaks in the obtained spectra. This corresponds well to the Mn reference spectra.
formation of a new layer as the thickness of the surface layer increases from 2–5 to 5–10 nm. This magnified HRTEM image also suggests lattice distortion due to the change in structure from a mixture of tetrahedral and octahedral Mn3O4 to octahedral MnO. DOD = 25–50% As shown in Fig. 2(a), a linear decrease in the Mn valence number continued from DOD = 25% to 50% as further lithiation of MnO and transformation into metallic Mn proceeded. The progress of the lithiation is supported by the decrease in the EXAFS peak intensity from DOD = 25% to 50%, which indicates an increase in the lattice distortion and a decrease in the coordination number of the Mn atoms. The peak intensity in the Mn2p XPS spectra further decreased beyond 25% DOD, suggesting that the thickness of the SEI layer around MnO and Mn increased (Fig. 3). In the C1s and F1s spectra, no obvious change was observed between DOD = 25% and 50%, but the peak intensity at 56 eV in the Li1s spectra increased while the Mn3O4 peaks at 644 and 656 eV decreased. These results indicate that the SEI layer growth continues without any composition change, and
Phys. Chem. Chem. Phys., 2014, 16, 6027--6032 | 6029
Published on 30 January 2014. Downloaded by MEDICAL RESEARCH COUNCIL LABORATORY OF MOLECULAR BIOLOGY on 27/10/2014 10:28:59.
View Article Online
Paper
PCCP
Fig. 4 High- and low-magnification HRTEM images of pristine Mn3O4 ((a) and (b)) and of the electrode material at 0% DOD ((c) and (d)). The difference between pristine Mn3O4 and that at 0% DOD is that the latter had been dipped in the electrolyte (1 M LiPF6/EC + DEC). HRTEM images of pristine Mn3O4 show highly crystalline particles with diameters of 20–300 nm. The particle morphology and high crystallinity did not change after the particles were soaked in the electrolyte, while a thin film of 2–5 nm thickness accumulated. This film is considered an SEI (4(d)). Fig. 3 XPS spectra (C1s, F1s, Li1s, Mn2p, and O1s) with possible peak assignments of the Mn3O4 electrode surface at different DOD stages. In the O1s spectrum obtained before cycling, a peak attributed to –OH at 535 eV was observed. This result suggests that the –OH film was formed on Mn3O4 just after contact with the electrolyte, which corresponds well to the HRTEM observations (Fig. 4(d)). As the DOD increased, the –OH peak gradually decreased in intensity and disappeared at a 75% DOD. In the Mn2p spectra, the Mn peaks at 644 and 656 eV also decreased in intensity with increasing DOD and finally disappeared at a DOD of 75%. This supports the idea that another film forms on the Mn3O4. At the same DOD of 75%, the C1s peak for Li2CO3 at 290.5 eV, the F1s peak for LiF at 686 eV, and the Li1s peak for Li2CO3, LiF, and Li2O at 56 eV are clearly observed. These peaks suggest that the new material is composed of a mixture of these Li2CO3, LiF, and Li2O compounds.
Li2O particles clearly appear. In the HRTEM images in Fig. 5(g) and (h), a thickened surface film (10–20 nm) is observed around the particles. Furthermore, disintegration within the particles is clearly observed, resulting in nanograins of 2–5 nm, although the global shape of the initial particle is preserved. The above results suggest that the metallic Mn nanoparticles and unreacted MnO particles are dispersed uniformly with Li2O on a nanoscale level and are covered by a thick SEI layer mainly composed of Li2CO3. DOD = 50–75% At the end of the plateau at DOD = 75%, the lithiation of MnO is completed as the Mn valence state drops to 0.0 (Fig. 2(a)). The XANES spectra correspond well to that of the metallic Mn. In the C1s spectra, the intensity of the peak attributed to Li2CO3 further increases, while at the same time the peak at 686 eV in the F1s spectra becomes pronounced, indicating the formation of LiF.
6030 | Phys. Chem. Chem. Phys., 2014, 16, 6027--6032
Fig. 5 High- and low-magnification HRTEM images of Mn3O4 after discharging to DODs of 25% ((e) and (f)) and 50% ((g) and (h)). The image of Mn3O4 at DOD = 25% shows distortion of the Mn3O4 lattice (f), although the overall shape of the starting particles is preserved (e). In addition, the thickness of the SEI film apparently increased to 10 nm (e). At a DOD of 50%, the thickness of the SEI obviously increased (g), and black dots of less than 5 nm diameter began to appear. Meanwhile the lattice stripes of crystalline Mn3O4 became unclear, probably because of the thicker SEI layer (h).
This journal is © the Owner Societies 2014
Published on 30 January 2014. Downloaded by MEDICAL RESEARCH COUNCIL LABORATORY OF MOLECULAR BIOLOGY on 27/10/2014 10:28:59.
View Article Online
PCCP
The disappearance of the Mn peaks in the Mn2p spectra suggests that the Mn nanoparticles are completely surrounded by the thickened SEI layer. HRTEM images taken at this stage revealed a nanocomposite microstructure consisting of Mn–Li2O completely covered by a further thickened SEI layer (20–30 nm thickness). The unclear nanoparticles in the images compared to those observed at DOD = 50% also support the complete covering of the nanoparticles by the thick SEI layer. These results suggest that Mn may have accelerated the decomposition kinetics of the electrolytes. DOD = 75–100% From DOD = 75% to 100%, the XANES spectra remained unchanged and almost overlapped that of metallic Mn, showing that the valence number remained constant, close to 0.0 (Fig. 2). In the EXAFS spectra, the peak from Mn metal at around 2.2 Å (Mn–Mn) was observed, while overall the peaks are ambiguous, indicating a disordered structure containing Mn nanograins. This result suggests that this electrochemical reaction, which consumed the last 25% of the discharge, was not related to the Mn species but instead to the further decomposition of the electrolyte. In the XPS spectra, however, no obvious change was observed between DOD = 75% and 100%, suggesting that the composition of SEI did not change as its thickness increased (Fig. 3). This idea is supported by the HRTEM images, which show thickened and merged SEI layers that look like polymeric networks completely surrounding the Mn–Li2O microstructures, which can no longer be observed (Fig. 6(k)).
Fig. 6 High- and low-magnification HRTEM images of Mn3O4 after discharging to DOD = 75% ((i) and (j)) and 100% ((k) and (l)). At a DOD of 75% (i), the surfaces of the Mn3O4 particles were further covered by the SEI layer, while the composition of the SEI layer mostly did not change beyond DOD = 50%. At DOD = 100%, the thickness of the SEI remained unchanged (k), while the boundary between the Mn3O4 particles and the SEI layer became unclear because of the complete mix of the two compounds (l). The starting Mn3O4 particle shapes remained despite the total decomposition of the starting 100 nm particles into particles with diameters of 5 nm or smaller.
This journal is © the Owner Societies 2014
Paper
These results are consistent with reports of Tarascon’s group and other groups that discharged metal oxides such as CoO and RuO2 form a polymer/gel-like coating.25–28 We believe that such encapsulation of dispersed Mn–Li2O nanoparticles within the merged SEI layer isolates Mn and Li2O from the electronic path/ contact. The isolation of active materials hinders the reversible reaction of Mn into Mn3O4, limiting the recovery of the Mn valence state to 2.0+, as reported by Lowe et al. and others.22,24
Conclusions The Mn3O4 conversion reaction was investigated at different DODs by a combination of in situ XAFS, XPS, and HRTEM observations to clarify how the reaction proceeds from the crystallographic and interfacial viewpoints. The results are summarized in separate illustrations of the crystallographic structure and the surface morphology in Fig. 7. Crystallographic structure changes Starting from a highly crystalline Mn3O4 particle with a diameter of 20–300 nm, as seen in the HRTEM images taken at DOD = 0% after the particle was soaked in the electrolyte (Fig. 7, point A), the
Fig. 7 Illustrations of the material transformation of Mn3O4 during the first discharge. Crystallographic changes of Mn3O4 (left) and morphological changes of Mn3O4 particles (right) are shown for different DOD stages from DOD = 0% (point A) to DOD = 100% (point E).
Phys. Chem. Chem. Phys., 2014, 16, 6027--6032 | 6031
Published on 30 January 2014. Downloaded by MEDICAL RESEARCH COUNCIL LABORATORY OF MOLECULAR BIOLOGY on 27/10/2014 10:28:59.
View Article Online
Paper
material is reduced to MnO during the transition from DOD = 0% (point A) to 25% (B). During this transition, the MnO particles, whose valence number was confirmed to be 2.0+ by in situ XAFS measurements, retain their whole-particle morphologies but begin to exhibit partial distortion due to the structural change from a mixture of tetrahedral and octahedral Mn3O4 to octahedral MnO. At DOD = 50%, the lithiation of MnO, transformation into Mn metal, and production of Li2O proceed. The crystallographic structure of this material should be the most complicated among those at the various DOD stages, owing to the existence of three different phase compositions, Mn, MnO, and Li2O. Thus, the lattice distortion and a decrease in the coordination number of Mn occur, which is seen as a decrease in the EXAFS peak intensity. At DOD = 75% (point D), the transformation from MnO into Mn metal and Li2O is complete. At DOD = 100% (point E), the crystallographic state remains unchanged from that observed at DOD = 75% (D), with the same mixture of Mn metal and Li2O. Morphological changes At the initial stage, with DOD = 0% (point A), after contact with the 1 M LiPF6/EC + DEC electrolyte, a very thin protective hydroxidebased film of 2–5 nm thickness is formed by decomposition due to the catalytic effect of Mn3O4 particles. The thickness of the film increases from 2–5 to 5–10 nm up to DOD = 25% (point B), while the main component of the film changes from hydroxide to Li2CO3, which is a product of the reduction of the carbonate solvents (EC and DEC). At this stage, the particles start to show wrinkles on portions of their surfaces due to the crystallographic transformation. The particle morphology clearly changes at DOD = 50% (point C), where 2–5 nm Mn nanograins appear along with amorphous Li2O from the reduced MnO. Meanwhile, the starting particle begins to crack because of the distortion due to the crystallographic transformation. The thickness of the SEI film further increases from 5–10 to 10–20 nm, without an accompanying composition change. The SEI composition begins to change between DOD = 50% and 75% (D), where a new peak attributed to LiF is observed in the XPS spectra along with a Li2CO3 peak. The mixed Li2CO3–LiF SEI continues to grow to 20–30 nm thickness and eventually covers the entire structure consisting of Mn–Li2O nanoparticles. At this stage, the average valence state of Mn is almost 0, suggesting the termination of the Mn-related reaction. At 0 V vs. Li/Li+ (DOD = 100%, point E), the Li2CO3–LiF mixed SEI grows without changing in composition. Its precise thickness, however, is difficult to determine from the HRTEM images, because it has merged over the particle surfaces. This merged SEI covers the microstructure consisting of Mn/Li2O particles of a few nm size, which are now totally encapsulated within a huge organic–inorganic coating structure. However, the overall shape of the starting particles is retained.
Acknowledgements The synchrotron radiation experiments were performed at Spring-8 with the approval of the Japan Synchrotron Radiation Research Institute (JASRI). This work was supported by JSPS KAKENHI Grant Number 25249140.
6032 | Phys. Chem. Chem. Phys., 2014, 16, 6027--6032
PCCP
References 1 M. Armand and J. M. Tarascon, Nature, 2008, 451, 652–657. 2 K.-C. Zhou, D.-M. Cao and Z.-Y. Li, Trans. Nonferrous Met. Soc. China, 2006, 16, 517–521. 3 Z.-H. Wang, L.-X. Yuan, Q.-G. Shao, F. Huang and Y.-H. Huang, Mater. Lett., 2012, 80, 110–113. 4 M. F. Hassan, Z. Guo, Z. Chen and H. Liu, Mater. Res. Bull., 2011, 46, 858–864. 5 P. Lian, X. Zhu, H. Xiang, Z. Li, W. Yang and H. Wang, Electrochim. Acta, 2010, 56, 834–840. 6 H. Qiao, L. Xiao, Z. Zheng, H. Liu, F. Jia and L. Zhang, J. Power Sources, 2008, 185, 486–491. 7 H. Kim, D.-H. Seo, S.-W. Kim, J. Kim and K. Kang, Carbon, 2011, 49, 326–332. 8 X. H. Huang, J. P. Tu, C. Q. Zhang, X. T. Chen, Y. F. Yuan and H. M. Wu, Electrochim. Acta, 2007, 52, 4177–4181. 9 H. Wang, Q. Pan, J. Zhao and W. Chen, J. Alloys Compd., 2009, 476, 408–413. 10 X. H. Huang, X. H. Xia, Y. F. Yuan and F. Zhou, Electrochim. Acta, 2011, 56, 4960–4965. 11 M. F. Hassan, Z. P. Guo, Z. Chen and H. K. Liu, J. Power Sources, 2010, 195, 2372–2376. 12 M. Winter and J. O. Besenhard, Electrochim. Acta, 1999, 45, 31–50. 13 B. Guo, J. Shu, Z. Wang, H. Yang, L. Shi, Y. Liu and L. Chen, Electrochem. Commun., 2008, 10, 1876–1878. 14 H. Dhaouadi, A. Mandani and F. Touati, Mater. Lett., 2010, 64, 2395–2398. 15 S.-Y. Liu, J. Xie, Y.-X. Zheng, G.-S. Cao, T.-J. Zhu and X.-B. Zhao, Electrochim. Acta, 2012, 66, 271–278. 16 H. Wang, L.-F. Cui, Y. Yang, H. S. Casalongue, J. T. Robinson, Y. Liang, Y. Cui and H. Dai, J. Am. Chem. Soc., 2010, 132, 13978–13980. 17 N. Lavoie, P. R. L. Malenfant, F. M. Courtel, Y. Abu-Lebdeh and I. J. Davidson, J. Power Sources, 2012, 213, 249–254. 18 G. Yang, Y. Li, H. Ji, H. Wang, P. Gao, L. Wang, H. Liu, J. Pinto and X. Jiang, J. Power Sources, 2012, 216, 353–362. ˜ a, Chem. Mater., 2011, 23, 19 J. Gao, M. A. Lowe and H. D. Abrun 3223–3227. 20 Z. H. Wang, L.-X. Yuan, Q.-G. Shao, F. Huang and Y.-H. Huang, Mater. Lett., 2012, 80, 110–113. 21 D. Pasero, N. Reeves and A. R. West, J. Power Sources, 2005, 141, 156–158. 22 X. Fang, X. Lu, X. Guo, Y. Mao, Y.-S. Hu, J. Wang, Z. Wang, F. Wu, H. Liu and L. Chen, Electrochem. Commun., 2010, 12, 1520–1523. 23 H. Kim, S.-W. Kim, J. Hong, Y.-U. Park and K. Kang, J. Mater. Res., 2011, 26, 2665–2671. 24 M. A. Lowe, J. Gao and H. D. Abruna, J. Mater. Chem. A, 2013, 1, 2094–2103. 25 P. Poizot, S. Laruelle, S. Grugeon, L. Dupont and J. M. Tarascon, Nature, 2000, 407, 496–499. 26 J. M. Tarascon, S. Grugeon, M. Morcrette, S. Laruelle, P. Rozier and P. Poizot, C. R. Chim, 2005, 8, 9–15. 27 S. Gurgeon, S. Laruelle, L. Dupont and J.-M. Tarascon, Solid State Sci., 2003, 5, 895–904. 28 P. Balaya, H. Li, L. Kienle and J. Maier, Adv. Funct. Mater., 2003, 13, 621–625.
This journal is © the Owner Societies 2014
