J. MATER. CHEM., 1994, 4( l), 133-137 Chemical Lithium Insertion into Sol-Gel Lamellar Manganese Dioxide MnO,.,,mH,O Philippe Le Goff," Noel Baffier," Stephane Bach,*b and Jean-Pierre Pereira-Ramosb a Laboratoire de Chimie Appliquee de /'€tat Solide, C.N.R.S. URA 302 ENSCP 7 7, rue Pierre et Marie Curie, 75237 Paris Cedex 05, France Laboratoire d'Electrochimie, Catalyse et Synthese Organique, C. N. R.S. UM 28 2, rue Henri Dunant, 94320 Thiais, France Chemical lithium insertion in the sol-gel manganese oxide MnO,.,,.nH,O has been performed using n-butyllithium as a reducing agent. Open-circuit voltage (OCV) experiments, IR, XRD and thermal analysis on lithiated samples Li,MnO,~,,~nH,O 0 <x 60.45show that lithium insertion into the host lattice induces a slight structural change: from x =0.05, a relative gliding of MnO, layers occurs, leading to the transformation of the initial hexagonal Structure into a closely related monoclinic phase.The higher the lithium content, the stronger the Li-hydrogen bond interactions which explains the contraction of the lattice observed along the c direction. The low magnitude of the structural changes is consistent with the high reversible behaviour found for this compound used as rechargeable cathodic material for lithium batteries. During the last several years, extensive research has been developed on manganese dioxide in order to obtain rechargeable Li/MnO, cells. From the wide variety of manga- nese dioxides, only the birnessite group exhibits a layered structure' which would theoretically provide attractive proper- ties as a rechargeable cathodic material for secondary Li batteries.Reversible insertion of ca. 1 Li ion per mole of oxide would be effectively possible without destruction of the host lattice thus leading to a theoretical Faradaic capacity of ca. 300 A h kg-'. The birnessite is usually synthesized by dehydration of buserite.2 This latter compound is prepared from an oxidation of an aqueous Mn(OH), suspension by oxygen or chlor-ine leading to sodium birnessite with the formula Na,,2sMnOl .,,-0.64H20. After removal of sodium ions from the structure by an acid treatment performed at room tempera- ture,, the Mn0,.84.0.71H,0 oxide is obtained. Regarding the corresponding sol-gel compound, a layered manganese di- oxide is obtained from mixed oxides Na,.,MnO,, KMnO, or K,.,,MnO, synthesized via a sol-gel pro~ess.~ These com- pounds are then transformed into a lamellar sol-gel oxide MnO,~,,.nH,O (0.60G n G0.76) by an acid treatment at room ternperat~re.~Contrary to what happens for the classical compound Mn01~,,-0.7 1H20,the highly preferred orientation of these latter is maintained during acid treatment.The layered structure of the sol-gel MnO,,,,~nH,O com-bined with these textural properties have been proved to induce remarkable electrochemical behaviour. Indeed, ca. 0.85 Li ions enter reversibly into the host lattice with high kinetics for Li transport, which corresponds to a specific capacity of ca.250 A h kg-' available in the potential range 4.2/2 V. Its good cycling behaviour has been previously ascribed to texture and structural properties. However, its electrochemical behav- iour has only been correlated to preliminary X-ray diffraction experiments. In particular, no data about the role of structural water have been provided. In order to achieve further insight into the Li insertion process in sol-gel MnO,.,,.nH,O and to specify the structural changes occurring during Li insertion-extraction, we have undertaken the structural investigation of chemically lithiated samples. In this paper, we report additional IR, X-ray, DSC and TG experiments as well as new potentiometric measure- ments which are discussed in relation to the electrochemical properties of the cathodic material.Experimental X-Ray diffraction experiments were performed with a Philips diffractometer using Cu-Kcc radiation. Thermal analysis measurements were performed in air at heating rates of 10 "C min-' using a Netszch STA 409 analyser with the simultaneous recording of weight losses (TG) and temperature variations (DSC). The mean oxidation state 'Z' of manganese was determined by a chemical titration using iron@) sulfate6 with an accuracy of f0.02. Chemical analysis of the compounds was made by atomic absorption measurements with a Varian 2150 appar- atus. Two successive sulfuric acid treatments of sol-gel Nao.,MnO, prepared as mentioned in ref. 5 ensure a complete removal of sodium ions from the compound.This was checked by atomic absorption experiments. Infrared spectra were recorded on a Perkin-Elmer 783 spectrophotometer by grinding the powder into KBr pellets. The electrochemical measurements were performed in pro- pylene carbonate (PC), twice distilled, obtained from 'Fluka' and used as received. Anhydrous lithium perchlorate was dried under vacuum at 200°C for 12 h. The electrolytes were prepared under a purified argon atmosphere. Results and Discussion According to the literature,' the structural model for the manganese(1v) oxides with layered structure is chalcophanite ZnMn,O7-3H,O. This structure consists of single sheets of water molecules between layers of edge-sharing MnOb octahedra, with Zn atoms located between the water layer and oxygens of the MnO, layer.In the case of classical birnessite, it is necessary to substitute Mn2+ ions for Zn2+ ions. The stacking sequence along the c-axis would thus be: O-Mn'V-O-Mn"-H,O-Mn"-O-Mnlv-O-and the perpendicultr distance between two consecutive MnO, layers is ca. 7.24 A. In this type of structure, vacancies exist in the layer of linked MnO, octahedra and the total water content is variable. Chemical Synthesis of Lithiated Compounds From thermodynamic data, butyllithium is well known to be particularly suitable to perform the Mn4+ reduction into Mn3+ ions., Since the lithiation reaction occurs at room temperature the chemically reduced compound should be similar to the electrochemical one.The reduction reaction, which has already been described in the literature,' was carried out as follows. The sol-gel compound (1g) was mixed with a hexane solution in a test-tube first placed in an inert atmosphere. An appropriate ratio of n-butyllithium in hexane was added and the mixture was stirred for a period of 1 or 2 days at room temperature. The reaction took place according to: MnO,.,,.nH,O +xC,H,Li~Li,MnO,,,,.nH,O +x/2C,H,, (1) The final product was obtained after filtering, washing with hexane and drying at room temperature. The amount of lithium in the compound was determined by atomic absorp- tion spectroscopy. The following x values were obtained: 0.05; 0.1; 0.15; 0.20; 0.25; 0.30; 0.35; 0.40; 0.45. Electrochemical Data Electrochemical Li Insertion Li insertion into sol-gel MnOl,8s.nH,0 has been shown to occur reversibly in two steps in the potential range 4.25/2 V us.Li/Lif with a maximum Faradaic yield equal to 0.85 F per mol of oxide [Fig. l(a)]. From these results it was suggested that only Mn" ions were involved in the charge transfer so that the most convenient formula for the sol-gel compound is Mn'V,~,5Mn",~,s01,8,.nH,0.9 Previous thermodynamic and kinetic results as well as the XRD experiments performed on electrochemically lithiated electrodes indicated a two-step process for the Li insertion reaction: for the Li content 0 <x <0.4, Li insertion occurs in the potential range 4.25-2.85 V with high kinetics for Li transport, while for a higher Li concentration range 0.4 <x <0.9 Li insertion is more difficult and a voltage quasi- plateau at ca.2.8 V is observed. In terms of cycling capacity, the specific capacity slightly decreases from an initial value of 200 A h kg-' (0.7 F mol-') to ca. 150 A h kg-' by the 50th cycle. In a few words, the major advantage of the sol-gel lamellar manganese dioxide consists of an unusually high depth of discharge which makes this compound a very promis- ing rechargeable MnO, in comparison with various MnO, forms studied for secondary Li batteries.'@I2 However, some unclear points remain; in particular, the unusual presence of water molecules in the structure of the material raises some questions about its effect on the Li insertion process.Chemical Li Insertion: OCVMeasurements For each composition, ca. 20 mg of pure chemically lithiated compound was pressed on a stainless-steel grid with a geo- 4400 40001. Fig. 1 OCV curves for (a) electrochemically and (b)chemically lithi- ated Li,MnO,,,,.nH,O samples J. MATER. CHEM., 1994, VOL. 4 metric area of 1 cm2. Under these conditions, equilibrium is considered to be reached when the open-circuit voltage remained stable (21 mV) for 20 h. Conversely, the composition change is ensured by coulo- metric titration for electrochemical experiments. As shown in Fig. 1, very similar OCV curves are found for the samples reduced according the two lithiation methods. This indicates the suitability of using data drawn from the present study on chemically lithiated samples to investigate the electrochemical behaviour of the sol-gel Mn0,,85.nH,0.X-Ray Diffraction Analysis of the Lithiated Compounds Fig. 2 shows the evolution of X-ray diffraction patterns for different amounts of Li in the compounds. For x =0, starting sol-gel lamellar manganese dioxide, two well defined diffrac- tion peaks, 001 and 002, correspond to the d-spacing between two consecutive Mn0, layers of the lamellar structure. This parameter is weakly dependent oq the water content, n: d= 7.24 A for n =0.64 H20up to 7.32 A for n =0.76 H20. Accord- ing to the position of the other diffraction peaks, th,e structure is of hexagonal type, with parameters a= 2.84 A and c= 14.64A. This sLructure is consistegt with the a-Na,.,MnO, form (c = 11.12 A, &spacing= 5.56 A), but with a higher inter- layer distance., Increasing of the interlayer distance is due to the extraction of Na' ions from the crystal framework, i.e.to an increasing MnIV content, during the acid treatment. From x =0.05, new diffraction peaks appear, the intensity of which grows with x,whereas intensities of peaks corre-sponding to the hexagonal structure decrease. From x =0.25, only the new diffraction peaks are present on the diagram. They correspond to .a monoclir$c phase with the following parameters: a =5.15 A; b =2.86 A; c = 14.29 A; p= 102.6". For higher Li contents, only the monoclinic phase exists, but ma- terials become less and less well crystallized. The hexagonal- monoclinic transformation is a well known phenomenon in Mn0,-based compound^.'^ The relationships between the parameters are (Fig.3): aMono ahexJ3 bMono ahex CMvlono sin P ZZ chex Such a structural change can be explained by a glide of the layers in the ab plane, one MnO, layer relative to another. The evolution of the spacing distance (c/2) uersus x is shown in Fig. 4. Two sets of data are achieved in the composition ranges 0 <x <0.02 and 0.05 <x <0.45 for which the interlayer spacing of the hexagonal and monoclinic phases decreases slightly (ca. 2%). A biphasic region is evidenced for 0.05<x<0.20 while a and b parameters of the monoclinic phase do not change significantly all along the lithium inser- tion process.Hence, Li insertion originates a hexagonal-monoclinic distortion corresponding to a maximum contrac- tion of the interlayer space (of the order of 5%) for 0 <x <0.9, with a volume change < 1%. In other respects, the lower interlayer space of the monoclinic phase combined with the higher lithium content makes the kinetics of lithium transport slower from ~~0.25.~ Nevertheless as reported,' no loss of reversibility is noted whatever the depth of discharge. IR Analysis of the Lithiated Compounds The evolution of infrared spectra uersus Li content x is shown in Fig. 5. The OH stretching region in the high-frequency spectrum between 4000 and 2500 cm-' corresponds, for x = 0 to a broad asymmetric absorption band centred near 3350 cm-', constituted of two components (3500 and J.MATER. CHEM., 1994, VOL. 4 * 0 I-. 6 10 15 20 25 6 10 15 20 25 * 0 I. 6 10 15 20 25 , (f) 6 10 15 20 25 Hdegrees Fig. 2 Evolution of X-ray diffraction patterns for x=O (a); 0.05 (b); 0.1 (c); 0.15 (d); 0.25 (e); 0.45 (f) in chemically lithiated LiXMnO,,85.tIH,Osamples. For x =0, the diffraction peaks are charac-teristic of a hexagonal phase. From x=O.O5, new diffraction peaks (*) appear, corresponding to a monoclinic phase. For x=O.25, all the diffraction peaks are characteristic of the monoclinic phase only 3300cm-l) due to two kinds of hydroxyl group. As lithium accommodation proceeds, the stretching vibration band .assigned to hydroxyl groups situated near 3500 cm-' is pro-gressively shifted towards lower frequencies, leading for x= 0.45 to a more symmetric absorption band centred near 3400 cm-'.Such an evolution indicates the strengthening of amon I I .z I I 1sI I 13 I I II I igI I I II I I I I Fig. 3 The hexagonal-monoclinic transformation is a well known phenomenon in Mn0,-based compounds from ref. 13;(-) monoclinic; (---) hexagonal 6.80 0 0.10 0.20 0.30 0.40 0.50 lithium ratio Fig.4 Evolution of interlayer spacing with lithium content x for chemically lithiated Li,Mn0,,85.tIH,0 samples. The interlayer spacing is c/2 and c sinb/2 for the hexagonal (W) and monoclinic (0)phases, respectively HO-Li interactions during lithium accommodation.In fact, IR measurements show that the higher the degree of reduction, the stronger the interaction of Li' ions with hydrogen bonds between the water layer and oxygen atoms of the MnO, layers. Hence, it can be assumed that such a phenomenon explains the notable decrease of the interlayer spacing pointed out in XRD, especially in the Li composition range 0 <.Y d 0.2. However, since the samples have not been preserved from moisture, the basic character of interlayer water due to the presence of lithium ions results in a notable dissolution of atmospheric CO, to give Li,C03 as shown by the emergence of intense specific absorption bands at 1380 and 88Ocm-' for x>0.2. Thermal Analysis of the Lithiated Compounds Thermal analysis of lithiated compounds has been performed in order to study the dependence of the water content on the temperature. In the case of the starting sol-gel material MnO,,,,.nH,O, the simultaneous recorded thermal analyses (TG and DSC) were characterized by a total weight loss of ca.20% (Fig. 6).From room temperature to 250 "C,only one endothermic peak appeared at 120 "C corresponding to a weight loss of 11.5%, due to the departure of the interfoliar water (weakly bonded water). The weak weight loss (ca. 3%) observed up to 450 "C is probably due to the removal of more strongly bonded water or hydroxy groups. The last endo-thermic peaks, located at 540 and 950°C correspond to the transformation into a-Mn,03 and Mn304,respectivel\-.. J.MATER. CHEM., 1994, VOL. 4 n I I I I I I,,I I I I I 10 3200 2400 1600 1200 800 400 wavenumbedcm-' Fig. 5 IR spectra of Li,MnOl,,,~nHzO samples where x=(a) 0; (h)0.05;(c) 0.10; (d)0.15; (e)0.20; (f) 0.25; (g)0.30 (h)0.40 (i) 0.45 I-251 , , , 1-25 (u -30 0 200 400 600 800 1000 TI"C Fig.6 Simultaneous thermal analysis (TG and DSC) of the sol-gel lamellar manganese oxide for x =0 (20<T/"C<900). Thermal analy- sis measurements were performed in air at heating rates of 10 "Cmin-l The evolution of TG and DSC curves between 20 and 500 "C versus Li content x is shown in Fig. 7. From x =0.05, the main endothermic peak at 128°C splits into two well shaped peaks. The first one is located at ca. 1OO"C, while the other, more important, appears at 170 "C.Thus, insertion of Li' ions induces a new localisation of water leading to the formation of more strongly bonded water. In addition, the emergence of a broad exothermic peak located at ca. 220°C which increases with the lithium content, corresponding to the strengthening of the HO-Li interactions, is consistent with the IR analysis. Beyond 350 "C, X-ray diffraction analysis reveals the pres- 10--0---10--I I I -20 40 --20 -jj90 i 0--2010 100 .-200 300 400 500 1°C Fig. 7 Simultaneous thermal analysis (TG and DSC) of chemically lithiated sol-gel lamellar manganese oxide for x=O (a); 0.05 (b); 0.10 (c); 0.45 (d) in Li,Mn0,.,,~nH,0 samples (20d T/"C<500). Thermal analysis measurements were performed in air at heating rates of 10°Cmin-l ence of the Li,,,Mn,O, spinel structure which could be correlated to the small additional exothermic peak observed in thermal analysis (Fig.7). TG data show that the quantity of water which is given off before 200 "C remains approximately constant (ca. 0.6 H,O/Mn) indicating that lithium insertion does not affect the concentration of interlayered water. However, beyond J. MATER. CHEM., 1994, VOL. 4 x=0.25, a slight and regular decrease of interlayered water is observed. This departure of water is correlated with the formation of Li,CO, evidenced on IR spectra. Conclusions The present work has clarified the lithium insertion process into the lamellar compound Mno1,8,*nH2O synthesized via the sol-gel process.IR, XRD and thermal analysis of chemi-cally reduced samples Li,MnOl,8,.nH,0 (0<x <0.45) have shown that lithium ions enter the interlayer space leading to strong HO-Li interactions. Interlayer hydrogen bonds are then significantly affected which could explain the contraction of the host lattice in the c Girection: the interlayer distance diminishes from 7.29 to 6.9 A for O<x <0.35. A similar trend was found for this parameter in the case of electrochemically li thia ted samples.’ The lithium insertion process into the hexagonal structure of the sol-gel MnO,,,,.nH,O results from x=O.O5 in a glide of MnO, layers, leading to the formation of a monoclinic phase from x=O.25 with a biphasic region for 0.05dxd0.2.However, the monoclinic phase is very close to that of the initial hexagonal structure lattice. From x =0.25, the shorter monoclinic interlayer distance induces a decrease in the kinetics of lithium transport in the lamellar compound Mn0,,8,.nH20. Finally, the structural rearrangements are then minimized as lithium insertion proceeds, which is in good agreement with the high electrochemical reversibility encountered for this MnO,,,,.nH,O material. The authors are grateful to Dr. J. P Labbe for valuable discussions and to H. Lepesant for IR spectra recording. Financial support by the Direction des Recherches, Etudes et Techniques (DRET) is gratefully acknowledged. References 1 R. G. Burns and V. M. Burns, in Proceedings of the Manganese Dioxide Symposium, I.C.Sample Office, Cleveland, OH 1975, p. 305. 2 R. Giovanoli, E. Stahll and W. Feitknecht, Helu. Chim. Acta, 1954, 31,2322. 3 R. Giovanoli, E. Stahll and W. Feitknecht, Helu. Chim. act^, 1970, 53,453. 4 S. Bach, M. Henry, N. Baffier and J. Livage, J. Solid State Chem., 1990,88,325. 5 N. Baffier, S. Bach and J. P. Pereira-Ramos, in Solid Statt lonics, ed. M. Balkanski, T. Takahashi and H. L. Tuller, Ihevier, Amsterdam, 1992, p. 55. 6 M. J. Katz, R. C. Clarke and W. F. Nye, Anal. Chem, 1956, 28, 507. 7 M. B. Dines, Muter. Res. Bull., 1975,10,287. 8 J. Rouxel, J. Chim. Phys., 1986,83, 850. 9 S. Bach, J. P. Pereira-Ramos, N. Baffier and R. hlessina, Electrochim. Acta, 1991,36, 1595. 10 T. Nohma, Y. Yamamoto, K. Nishio, I. Nakane and N. Furukawa, J. Power Sources, 1990,32,373. 11 J. M. Tarascon, E. Wang, F. K. Shokoohi, W. R. McKinnon and S. Colson, J. Electrochem. SOC.,1991, 138,2859. 12 J. M. Tarascon and D. Guyomard, J. Electrochem. Soc., 1991, 138,2864. 13 J. P. Parant, R. Olazcuaga, M. Devalette, C. Fouassier and P. Hagenmuller, J. Solid State Chem., 1971,3, 1. Paper 3/03381C; Received 1lth June, 1993