Chemically oxidized γ-MnO2

Chemically oxidized g-MnO2 for lithiu and intercalation/deintercalation proper Won Il Jung,a Miki Nagao,a Ce´dric Pitteloud,a Keiji Itoh Received 24th July 2008, Accepted 25th September 2008 First published as an Advance Article on the web 6th January 2009 DOI: 10.1039/b812683f al utro hic iate spe pla others and form inner channels through the structure, which have two types of intergrowth, pyrolusite (1 � 1 channels) and 8,10–14 playing important roles when lithium ions are inserted and extracted reversibly. is is a powerful tool to obtain ring intercalation and dein- especially for local structure f protons, long range crystal- uctuations.18,25,26 Moreover, it length scales in disordered hase transition or structural samples without long-range PAPER www.rsc.org/materials | Journal of Materials Chemistry D ow nl oa de d by Z he jia ng U niv ers ity on 08 O cto be r 2 01 2 Pu bl ish ed o n 06 Ja nu ar y 20 09 o n ht tp :// pu bs .rs c.o rg | d oi: 10. 103 9/B 812 683 F View Online / Journal Homepage / Table of Contents for this issue g-MnO2 in Backus-Naur form. See DOI: 10.1039/b812683f crystallographic order could be analyzed using real-space func- tions, or radial distribution function RDF(r), derived from the Fourier transformation of the structural factor, S(Q). In this study, various states of charge and discharge during the first two cycles were identified by ex-situ X-ray diffraction analysis. Moreover, information on the possible occupation of lithium ions was obtained for chemically lithiated g-MnO2. The neutron total scattering analysis was sufficient to investigate structural changes of g-MnO2 during the chemical oxidation or lithiation processes. aDepartment of Electronic Chemistry, Interdisciplinary Graduate School of Science and Engineering, Tokyo Institute of Technology, 4259 Nagatsuta, Midori-ku, Yokohama, 226-8502, Japan. E-mail: kanno@echem.titech.ac. jp; Fax: +81 45 924 5401; Tel: +81 45 924 5409 bResearch Reactor Institute, Kyoto University, Kumatori-cho, Sennan-gun, 590-0494, Japan † Electronic supplementary information (ESI) available: Background of the structure for g-MnO2, oxidation mechanism, cyclability data for g-MnO2, comparison of ex-situ XRD patterns during the first two cycles (sample RAM), the data file format of DIFFaX simulation for ramsdellite (2 � 1 channels). Another notable feature of g-MnO2 is the existence of cation vacancies and protons in the structure as suggested by Ruetschi.15,16 Ruetschi proposed that there are vacancies in g-MnO2 and each empty Mn 4+ site is coordinated to four protons in OH� form to compensate the charge of manganese ions. These protons are called ‘‘Ruetschi’’ protons that are localized in the vacant octahedra and perma- nently occupied.15,16 On the other hand, the protons located in Neutron total scattering analys direct structural information du tercalation of lithium ions,24,25 information including positions o lographic order, and short-range fl can show the behavior over the crystalline materials during a p transformation. In addition, the Structural changes of chemically lithiated g-MnO2 during chemic X-ray diffraction combined with a stacking fault analysis and ne structure analysis. Lithium ions intercalate into the crystallograp prior to those around ‘‘Coleman’’ protons for the chemically lith ‘‘Coleman’’ protons have O–H bond lengths of 1.2 A˚ and 1.9 A˚, re chemical oxidation. The chemically intercalated lithium ions are octahedra with an O–Li length of 1.9 A˚. Introduction Lithium ion secondary batteries made a notable turning point in the field of personal communications such as mobile phones and personal digital assistants (PDAs). In the last decade, their specific capacity has been increased more than 2 times compared to the first commercialized battery owing to the development of anode materials and their cell designs.1 LiCoO2 currently in extensive use as a commercial cathode material possesses a high discharge voltage of 3.8–4.2 V vs. Li/Li+ and good cyclability. However, new cathode materials have been investigated in order to obtain a higher capacity, lower cost, and higher thermal stability.2–4 Manganese dioxides are promising cathode materials due to their high capacity and low toxicity. In particular, g-MnO2 has attracted great attention because it has relatively higher discharge voltage compared to the other types of manganese dioxides.5–9 Regarding the structural features of g-MnO2, the chains of edge-sharing MnO6 octahedra share corners with 800 | J. Mater. Chem., 2009, 19, 800–806 m secondary batteries: structure ties† ,b Atsuo Yamadaa and Ryoji Kanno*a oxidation were characterized by n total scattering with a local sites around ‘‘Ruetschi’’ protons d g-MnO2. The ‘‘Ruetschi’’ and ctively, which decreased after the ced at the center of oxygen conjunction with Mn3+ ions are called ‘‘Coleman’’ protons,17 which could be easily removed from the structure by heat-treatment at about 150 �C.18 For this reason, relevant heat- treatment processes have been utilized to prepare proton-free manganese dioxides as cathode materials for primary lithium batteries,19 because complete de-protonation from MnO2 would be difficult at ambient temperature. Previously, we developed a new method to remove the protons by chemical oxidation using NO2BF4 as an oxidant without any structural change. The achieved specific discharge capacity of g-MnO2 was estimated to be 275 mAh/g in the first cycle and its reversible capacity could reached 250 mAh/g with two discharge plateaux in subsequent cycles.20 On the other hand, inelastic neutron scattering studies experimentally confirmed the existence of both types of protons in g-MnO2, but their geometrical information is limited to the dynamics of these protons.21–23 Therefore, further inspection of a geometrical correlation between the protons and adjacent spaces is highly necessary, because they are This journal is ª The Royal Society of Chemistry 2009 materials were mixed with acetylene black and Teflon powder Neutron total scattering measurement Neutron scattering was measured with a high-intensity total scattering spectrometer, HIT-II, installed at the pulsed neutron source at the Neutron Science Laboratory (KENS, Tsukuba, Japan). The structural factors, S(Q), where Q ¼ 4psinql, 2q is the diffraction angle, and l is the neutron wavelength, were obtained from the scattering intensity. The Q length of HIT-II around 50 A˚�1 was used, and the corrections on the background, empty container, attenuation, and multiple scattering as well as normalization to the scattering length from a vanadium rod were carefully considered. The radial distribution function, RDF(r) was derived from the Fourier transformation of S(Q). Results and discussion X-Ray diffraction analysis Fig. 1 shows the charge–discharge profiles of chemically oxidized g-MnO2 (product name TKV-B) obtained after 24 hours � D ow nl oa de d by Z he jia ng U niv ers ity on 08 O cto be r 2 01 2 Pu bl ish ed o n 06 Ja nu ar y 20 09 o n ht tp :// pu bs .rs c.o rg | d oi: 10. 103 9/B 812 683 F View Online with a weight ratio of 50:10:2.5, and pelletized with a size of 1 cm2. Cells were composed by stacking a cathode pellet between two pieces of Al mesh as current collectors and lithium foil as a reference electrode separated by a polypropylene separator soaked in electrolyte (1 M LiPF6 in EC:DEC ¼ 3:7 vol.%). The cells were assembled in a glove box with an Ar atmosphere. The charge–discharge experiments were examined over a voltage range of 2.0–4.5 V vs. Li/Li+ with a current density of 0.1 mA/cm2 at 25 �C. The partial charge–discharge experiments for X-ray diffrac- tion analysis were carried out with a current density of 0.1 mA/cm2 (C/100). The CC/CV mode was adopted at the end of cut-off voltages to reduce the polarization of cells. The compo- sition was fixed at an interval of 20% of state-of-charge/discharge during the first two charge and discharge cycles. Partially charged or discharged coin-type cells were carefully dissembled and the cathode pellets were soaked and rinsed with a small amount of dimethylcarbonate (DMC) to remove residual elec- trolyte under Ar atmosphere for several days. All dissembling and sampling processes for ex-situ X-ray diffraction analysis were conducted under Ar atmosphere. X-Ray diffraction analysis (XRD, CuKa radiation, l¼ 1.54A˚, 50kV, 150mA, Rigaku RU200B) was used for the structure characterization of g-MnO2 and partially charged/discharged samples. Patterns were recorded over the 2q range 10–80� with a step size of 0.03� and a count time of 3.0 s. Experimental Preparation of chemically oxidized and lithiated g-MnO2 Chemically oxidized g-MnO2 was synthesized as follows. Elec- trolytic manganese dioxides (Mitsui Mining and Smelting Co., Ltd., product name TKV-B) were firstly heated under vacuum at 100 �C for 12 hours in order to remove water. After heat treatment, the powder was mixed with an oxidant (nitronium tetrafluoroborate, NO2BF4) in acetonitrile solution and stirred for 24 hours under Ar gas atmosphere. After filtering and vacuum drying process at 120 �C for 12 hours, the final product was kept under Ar atmosphere for several days. Chemically lithiated g-MnO2 was obtained from the treatment with lithium iodide (LiI) and n-butyllithium (n-BuLi) as reducing agents. For the chemical lithiation using LiI, chemically oxidized g-MnO2 powder was mixed with an equal amount by moles of LiI in acetonitrile solution and stirred for 24 hours at room temperature under Ar atmosphere. After filtering and vacuum drying at 120 �C for 12 hours, the product was kept under Ar atmosphere for several days. For the chemical lithiation using n-BuLi, chemically oxidized g-MnO2 powder was mixed with an equal molar ratio of n-BuLi in n-hexane solution and stirred for 24 hours at room temperature under Ar atmosphere. After vacuum drying at 60 �C for 6 hours to remove impurities and n-hexane, it was kept under Ar atmosphere for several days. Preparation of samples for electrochemical characterization and X-ray diffraction To characterize the electrochemical and structural properties, 2032 coin-type cells were assembled as follows. The cathode This journal is ª The Royal Society of Chemistry 2009 Fig. 1 Charge–discharge profiles of chemically oxidized g-MnO2 (TKV- B) at optimized conditions. exchange at 25 C followed by filtering under Ar atmosphere and vacuum drying at 120 �C. The cell showed a discharge capacity of about 275 mAh/g in the first cycle, and a reversible capacity of about 250 mAh/g with two discharge plateaux in subsequent cycles. The discharge capacity at the sixth cycle reached 244 mAh/g, indicating a relatively good cyclic retention. The X-ray diffraction was measured at five different state-of- charge (SOC) and depth-of-discharge (DOD) stages. Fig. 2 shows X-ray diffraction patterns of the optimized TKV-B at different partial charge/discharge states. The strong reflections around 18� and 42� to 45� come from the sample holder. A broadening of the distance between the (221) and (240) at DOD 60% and new peak appearances around 46�, 56� and 63� at DOD 100% (marked with a square in Fig. 2(a)) corresponds to a decrease in microtwinning defects (Tw). 11 The decrease in microtwinning defects combined with heat treatment of g-MnO2 was observed in previous work.11 However, we found that chemically oxidized g-MnO2 exhibited the usual property during the discharge process, and the peaks remained during charge– discharge process in this work. J. Mater. Chem., 2009, 19, 800–806 | 801 D ow nl oa de d by Z he jia ng U niv ers ity on 08 O cto be r 2 01 2 Pu bl ish ed o n 06 Ja nu ar y 20 09 o n ht tp :// pu bs .rs c.o rg | d oi: 10. 103 9/B 812 683 F View Online We found overall peak shifting to low scattering angle during the discharge in Fig. 2(a) and (c), and shifting again to higher angles reversibly during the charge in Fig. 2(b). These observa- tions are related to the insertion and extraction of lithium ions, and the reversible peak shifting indicates that g-MnO2 has a stable structure during the charge–discharge process. For further inspection of the new peak around 25� which shifted reversibly during the charge–discharge (marked with dotted lines and arrows in Fig. 2), the DIFFaX program was employed to simulate these structural variations induced by their stacking sequence; the structural transformation parameters are given in Fig. 3.18,20,26 Due to the high level of defects in g-MnO2, it is impossible to fit the data by conventional X-ray methods and to get absolutely quantitative results. However, DIFFaX simu- lation can provide information about the microtwinned structure in g-MnO2. The patterns of stacking sequence are defined as two types in the ramsdellite structure, +R and –R. The ratio of +R and –R is a relative value for the left and right twinned structure on the (021) plane. For pyrolusite, the patterns of +P and –P are also determined in the same way. The changes in the stacking and we identified that the main direction vector was changed Fig. 2 Ex-situ X-ray diffraction patterns for g-MnO2 (TKV-B) at different partial charging/discharging states: (a) the first discharging state, (b) the first charging state and (c) the second discharging state (asterisks are due to sample holder). 802 | J. Mater. Chem., 2009, 19, 800–806 from +R to �P and finally to +P.20 The parameters (+R:�R:+P:�P ¼ 76.5:12.5:2.0:8.0) at ambient temperature changed into (+R:�R: + P: � P ¼ 4.2:11.3:67.6:16.9) after heat- treatment at 400 �C. The changes were irreversible because they are caused by collapse of inner channels due to the insertion and extraction of lithium ions. Similar behavior was also observed in the charge–discharge process. The main direction vector was changed from +R to �P, and finally to +P. The parameters of as-prepared material (+R: �R: +P: �P ¼ 76.5:13.5:2.0:8.0) changed into (+R: �R: +P: �P ¼ 57.5:13.5:11.5:17.5) after it was fully charged. More interestingly, the parameter changes during the charge and discharge were less than those of the heat-treated sample, and its structural transformation was reversible. Chemically oxidized and lithiated g-MnO2 sequence of g-MnO2 during heat-treatment from ambient temperature to above 400 �C were reported in our previous work, Fig. 3 The simulated results for g-MnO2 during the first discharging state using DIFFaX and their transformation parameters. X-Ray diffraction patterns of chemically oxidized g-MnO2 and chemically lithiated g-MnO2 by reducing agents, LiI and n-BuLi, are shown in Fig. 4. As the lithiation level increase by chemical reduction, the diffraction peaks were shifted to lower scattering angle due to insertion of lithium ions into g-MnO2 as shown in Fig. 4(b) and (c). From the XRD patterns of lithiated g-MnO2 by n-BuLi (Fig. 4(c)), we found a new peak clearly observed around 25� (marked with a black square), which is the same phenomenon as occurred in chemically oxidized g-MnO2 during the charge– discharge as mentioned above (Fig. 2). Regarding the DIFFaX simulation results for a layered structure model containing stacking faults (Fig. 3), it is believed that changes of the stacking sequence are able to facilitate insertion of lithium ions into layered structures. On the other hand, the new peak appearing around 25� from a 20% depth-of-discharge state might be due to the change of stacking sequence from +R to –P in g-MnO2 during electro- chemical reactions. In contrast to the electrochemically lithiated g-MnO2, there is no detectable evidence around 25 � in the diffraction patterns of chemically lithiated g-MnO2 by LiI This journal is ª The Royal Society of Chemistry 2009 Fig. 4 X-Ray diffraction patterns of g-MnO2: (a) chemically oxidized g-MnO2, (b) chemically lithiated g-MnO2 by LiI and (c) chemically lithiated g-MnO2 by n-BuLi (asterisks are due to sample holder). D ow nl oa de d by Z he jia ng U niv ers ity on 08 O cto be r 2 01 2 Pu bl ish ed o n 06 Ja nu ar y 20 09 o n ht tp :// pu bs .rs c.o rg | d oi: 10. 103 9/B 812 683 F View Online (Fig. 4(b)), which means that the stacking sequence was main- tained during the chemical lithiation process. It can be clearly seen that the mechanism of chemical lithiation is different from that of electrochemical lithiation as described in Fig. 2(a). Fig. 5(a), (b) and (c) show the first charge profiles of untreated g-MnO2, vacuum-dried g-MnO2, and chemically oxidized g-MnO2 with an upper cut-off voltage of 4.5 V, respectively. Even though there is no lithium in the untreated g-MnO2 elec- trode, the cell was charged with a capacity of 12 mAh/g. This unidentified initial charge capacity decreased to 9 mAh/g after a vacuum drying process. According to the literature,20 this initial charge capacity is consistent with the electrochemical results of removing some of the ‘‘Coleman’’ protons from the host structure by heat-treatment. The chemically oxidized Fig. 5 The first charge profiles of g-MnO2 and chemically lithiated g-MnO2; (a) untreated g-MnO2, (b) vacuum dried g-MnO2 at 120 �C, (c) chemically oxidized g-MnO2, (d) chemically lithiated g-MnO2 by LiI and (e) chemically lithiated g-MnO2 by n-BuLi. This journal is ª The Royal Society of Chemistry 2009 g-MnO2 showed a first charge capacity of 3 mAh/g, which indicates that most of the protons governing electrochemical (de)intercalation could be removed by chemical oxidation. Fig. 5(d) and (e) show the charge profiles of chemically lithi- ated g-MnO2 by reducing agents LiI and n-BuLi, respectively. The charge capacity of lithiated g-MnO2 by LiI was 94.5 mAh/g, which gave the composition of x¼ 0.35 in LixMnO2. The lithium content is much smaller than the value of x¼ 1.0 in LixMnO2 for chemically lithiated g-MnO2 by n-BuLi, which is calculated from the capacity value of 291.3 mAh/g shown in Fig. 1. This differ- ence of depth-of-lithation between the different chemically lithiated g-MnO2 samples seems to be caused by different redox potentials of the reducing agents. The redox potential of LiI is 2.8 V vs. Li/Li+, with one plateau at around 3.5 V in Fig. 5(d), whereas the redox potential of n-BuLi is 1.0 V vs. Li/Li+ with two plateaux at around 3.1 V and 3.6 V in Fig. 5(e).28,29 The plateau at higher voltage (3.5–3.6 V) is associated with the sites around the Mn4+ vacancies and ‘‘Ruetschi’’ protons where the vacancy clustering is formed, and that at lower voltage (3.0 V) was associated with either octahedral site or around Mn3+ inner channels of pyrolusite and ramsdellite.16,30,31 When the chemical lithiation proceeded up to the depth-of-discharge of 33.5% (x ¼ 0.35 in LixMnO2), lithium ions were inserted into the interstitial space formed by Mn4+ vacancies along the inner channels. Moreover, with continuing chemical lithiation up to the depth-of-discharge of 100% (x ¼ 1.0 in LixMnO2), lithium ions were inserted directly into the octahedral sites of inner channels in pyrolusite and ramsdellite. Neutron total scattering analysis The intercalated protons into the g-MnO2 structure were placed at the center of a regular octahedron of oxygen atoms and then formed the inner channels. All O–H bonding distances were found to be set equal to about 2 A˚ by inelastic neutron scattering analysis.21,22 The degree of covalent bonding between the protons and the oxygen atoms in the channel walls was detected by NMR in previous reports.32,33 They also suggested that the protons moved off-center toward one of the oxygen atoms with a O–H bonding distance of