M.A.Santa Anaa, E.Benaventeb, J.Páeza, and G.Gonzáleza*
a Department of Chemistry, Faculty of Sciences, Universidad de Chile,
Casilla 653, Santiago de Chile
b Department of Chemistry, Universidad Tecnológica Metropolitana,
Av. José Pedro Alessandri 1242, Santiago de Chile ]]>
In memorian of Dr. Guido S. Canessa C.
ABSTRACT
Kinetic and thermodynamic properties of the products of the intercalation of lithium in layered transition metal oxides and sulfides are strongly determined by lithium coordinative environment. Thus, the electrode potentials as well as the activation energies for lithium diffusion are higher in MoO3 than in MoS2. A similar effect is appreciated for the products of the co-intercalation of lithium and poly(ethylene oxide) into MoS2. The effect of lithium coordinative environment on these properties may be understood by analyzing a theoretical model considering host-guest back-donation charge transfer achieved by the local polarization of the lithium ligands in the interlaminar spaces.
KEYWORDS: Lithium intercalation compounds, molybdenum trioxide, molybdenum disulfide, nanocomposites, lithium diffusivity.
]]> RESUMENLas propiedades termodinámicas y cinéticas de los productos de la intercalación de litio en óxidos y sulfuros de metales de transición están fuertemente determinadas por la coordinación del ion litio por sus alrededores. Así, tanto los potenciales de electrodo como las energías de activación para la difusión de litio son mayores para el MoO3 que para el MoS2. Un efecto similar se aprecia en los productos de la co-intercalación de litio y poli(oxietileno) en MoS2. El efecto de los alrededores coordinativos del litio en esas propiedades puede ser comprendido analizando un modelo teórico que considera una retrodonación de carga anfitrión-huésped producida por la polarización local de los ligantes del litio en los espacios interlaminares.
PALABRAS CLAVES: Compuestos de Intercalación de litio, trióxido de molibdeno, disulfuro de molibdeno, nanocompositos, difusividad de litio.
INTRODUCTION
Inclusion compounds may be in general seen as two interacting phase systems. A relatively rigid structure defining low dimensional spaces and, a second one, formed by the atoms, molecules, or ions which inserted in these spaces have a relatively high mobility[1]. In the case of the intercalation compounds, the host is a layered solid defining two dimensional interlayer spaces in which the guest species may be intercalated[2,3].
]]> Of special interest is the intercalation of alkali metals, which often leads to products with a wide stoichiometry range, often reaching high metal concentration so they may be useful as metal reservoirs. Among these, the intercalation of lithium in layered solids is specially important because of its applicability in the construction of electrode materials for secondary rechargeable lithium batteries[4]. Indeed, a modern concept of such a kind of batteries is the lithium-ion battery commonly known as "rocking-chair" battery[5] in which at least one of the electrodes is an intercalation compound. Graphite and carbonaceous compounds are frequently used as the negative electrode and transition metal sulfides and oxides as both cathode and anode[5-7].Lithium intercalation compounds meet well most of the characteristics needed for good electrode materials. Namely, because of the high potential of the couple Li/Li+ and the relatively high lithium intercalation degree they lead to products with high energy density. Moreover, the relatively high mobility of lithium ion in the interlaminar spaces permits rapid migration and diffusion of lithium from the surface to the bulk of the electrode, thus not only leading to devices with relatively high power, but also avoiding the accumulation of high reactive lithium in the surface, thus improving materials safety.
However, all these properties are determined by the coordination of the lithium ion in the interlaminar spaces. Indeed, the redox potential of the couple Li/Li+ is directly determined by the stabilization of the lithium cation which also determines its actual positive charge. A well known example of this feature is the influence of the donicity of the solvent (DN) on the reduction potential of lithium [8]. On the other side, the diffusion of lithium species will also depend on the activation energies for lithium jumping between two neighboring (generally octahedral) sites in the interlaminar spaces (vide infra). That is also certainly influenced by the stabilization of these species by coordination in the ground as well as in the activated state. So, both thermodynamic and kinetic properties of the system may be modulated by selecting an appropriated lithium environment, i.e. by an adequate selection or design of the host.
In this paper a contribution to understanding the effect of the coordinative environment of lithium in the interlaminar spaces of MoO3 and MoS2 is attempted by analyzing both the experimental results obtained from the intercalation of lithium in these matrices and a theoretical model for guest-host charge transfer[9]. Further, the analysis is extended to some nanocomposites prepared by the co-intercalation of lithium and electron pair donors such as poly(ethylene oxide) into MoS2.
EXPERIMENTAL
Molybdenum disulfide (Fluka, purum, size 1-3 µm), Molybdenum trioxide(Merck p.a.) and poly(ethylene oxide) (PEO) (Aldrich, molecular weight 100.000) were used as received. Butyl lithium was freshly prepared according to standard procedure [10]. Water was distilled twice and carefully degassed. The pH of aqueous solutions was continuously controlled and adjusted by adding 0.5 M nitric acid. n-Hexane (Aldrich p.a.) was dried and distilled under argon. Reactions were performed in Schlenk flasks under a dry argon atmosfere.
Typical synthesis procedure for poly(ethylene oxide)-MoS2 nanocomposites: 1.2 g (7.18 mmol) LixMoS2 (xª1) prepared by reacction of MoS2 with butyllithium in n-hexane [11] were treated with PEO following two different procedures: (i) LixMoS2 was treated directly with an aqueous solution of 0.62 g (1.42 mmol) PEO in pure water under stirring during 24 h at room temperature, (ii) LxMoS2 was hydrolyzed in pure water and the product flocculated by neutralizing the solution with nitric acid, washed with pure water, and then treated with the aqueous polymer solution under stirring for 24 h at room temperature. The products were washed with water and n-hexane and dried under vacuum. The elemental analysis for products obtained from procedures (i) and (ii) were respectively, Li0.1MoS2(PEO)0.5 (calc.): Li 0.63% (0.66%), C 5.75% (5.64%), H 1.02% (0.94%); and Li0.1MoS2(PEO)1.0: Li 0.63% (0.74%), C 13.2% (12.4%), H 2.22% (2.06%).
]]> Intercalation degree was tested by X-ray powder diffraction analysis performed in a Siemens D-5000 diffractometer with Cu Ka radiation.Compounds with different lithium stoichiometry such as Lixhost were achieved by galvanostatic reduction of the hosts at current density of 150 µA cm-1 using a Potentiostat-Galvanostat PARC model 175 in the cell Li / 1M LiClO4, PC / Host, measuring the open circuit voltage under quasi-equilibrium conditions of the cathode after a relaxation step lapse of 10 h. The variation of x was 0.025 in each step.
The cathodes were 6 mm diam. pellets of 20-25 mg of pure active material for MoS2 and Li0.1MoS2(PEO)n. In the case of MoO3 a mixture of MoO3 and graphite powder in the ratio 85:15 (w/o) containing 20-25 mg of MoO3 was used. Used cells were similar to those previously described [12,13].
Lithium diffusion coefficients were determined by the galvanostatic pulse relaxation technique[14] at different temperatures and at different lithium concentrations.
RESULTS AND DISCUSSION
Molybdenum disulfide may be described as a lamellar solid formed by layers of MoS2 units bonded covalently by S-bridges. These layers are stacked defining interlaminar bidimensional van der Waals spaces flanked by sulfur atoms. In the pristine solid each molybdenum atom is coordinated by six S-atoms in a prismatic arrangement. After intercalation however, an octahedral coordination of molybdenum is stabilized (vide infra) [15]. In the interlaminar spaces are thus defined two kind of sites, octahedral and tetrahedral sites, in which lithium could be located. Considering the well known difficulties for intercalating neutral species into MoS2, solvent effects have been neglected.
Orthorhombic molybdenum trioxide may be described as a layered structure in which the layers, held together by weak van der Waals forces, are stacked in an staggered arrangement. The interlaminar van der Waals spaces are in this case flanked by oxygen atoms [16].
In studies related with intercalation of alkali-metal cations into MoO3, two types of sites have been described. Assuming a displacement of each layer respect to its neighbors, cubic and tetragonal sites are defined. The occupation of these sites appears to depend on the size of the alkali cations as well as on the presence of electron pair donors as water or other solvents. Thus lithium, for instance, actually has an octahedral coordination sphere because of water molecules occupying the neighboring cubic sites [16]. Considering the ability of the oxygen sheets to display different kind of rearrangements, a distorted octahedral coordination of lithium cannot be excluded, specially in the case of intercalation reactions in which water or other donor solvent are absent.
]]> In both transition metal sulfides and oxides, the intercalation process is mainly governed by guest-host charge transfer [2,3]. The cations are normally naked without solvating shell and negative charge is distributed leading to polyanionic host. In the case of the intercalation of lithium in transition metal sulfides it has been observed that the charge transfer is partial; i.e. only a part of the charge of the alkali metal atom is transferred to the host. Thus, according to a model developed by Mendizábal et al.[17] -- which can be seen as a molecular approach using a finite-size cluster to represent the solid-- the charge transfer in the intercalation of lithium in TiS2 is only about 80%. Such calculations are in accord with the experimental evidences obtained from NMR measurements [18,19]. Moreover, studying the X-ray photoelectron spectra of a series of lithium compounds, we have observed that in the case of molybdenum disulfide with a relatively high intercalated lithium content, the transference appears to be rather low. Thus, for the compound Li0.8MoS2, the 1s electron in the intercalated lithium has a binding energy Eb of 55.6 eV which is more similar to that of the metal, 55.5 eV, than to those of ionic lithium compounds, e.g. 59.9 eV for LiBF4 [20]. The strength of the host-guest interactions, and thus the stability of the intercalation products, depend on such a transfer.According to the theoretical model mentioned above, the magnitude of the charge transfer is determined by the electronegativity of the host which, in turn, is a function of the charge previously acquired by the system. However for understanding the partial charge retained by lithium, the electronic polarization of the sulfur ligands induced by its coordination with lithium should be considered. Through back donation, the electronic charge returns to lithium. From this model it is clear that the charge transfer and, specially, the charge retained by the lithium atom depends on the nature of lithium ligands. The ligand polarization, which is rather appreciable for the sulfide, should be considerable lower in the case of oxygen which is certainly a much harder donor [21] than sulfur.
The concept charge capacity, i.e. the ability of one atom or group of atoms for absorbing electronic charge, has been often used for understanding the effect of lithium intercalation on the chemical potential of the system[22,23]. Theoretically, the charge capacity K can be written as the ratio between the charge (Q) transferred to the atom or group of atoms and the corresponding change of the electronic chemical potential[9]:
K=Q/(µelº- µel)
where µelº is the chemical potential before electron transfer.
The charge capacity can be experimentally represented by the differential capacity, i.e. the inverse derivative of the quasi equilibrium voltage-composition curve vs. intercalation degree, usually named incremental capacity. As observed in Fig 1, Dx/DE vs. x for MoO3 shows a rather normal behavior at low lithium content but increases again at about x=0.8 indicating a severe change in the lithium potential energy. That could be associated to a change in the geometry of the sites occupied by lithium. As analyzed before, that is quite possible for this host which has shown to have more than one site for accommodating alkali metal ions.
| Fig. 1. Incremental capacity for LixMoO3. |
 | Fig. 2. Variation of the of the reduction potential with lithium content along the intercalation of lithium in pure MoS2, and MoS2-PEO intercalates. |
An other way of changing the lithium-ion coordinative environment in the interlaminar spaces and thus the potential and the ionic conductivity of the intercalation products is the co-intercalation of both, lithium and some organic electron pair donors. (See curves for LixMoS2(PEO)y nanocomposites in Fig. 2).
Our approach to this problem was the co-intercalation of some well known ion-conducting polymers as are poly(ethylene oxide) and poly(acrylonitrile)[12,13,24-26]. Both polymers have been widely used as composites with lithium salts and a plasticizer, commonly an organic solvent, as solid or semi-solid conducting electrolytes[27-29].
The intercalation of these polymers into MoS2 following carefully specific synthesis methods yields a series of phases. Characterization of the products by elemental analysis, X-ray diffraction analysis, and thermal analysis, among others, show that they are pure laminar phases with different polymer content in the interlaminar spaces[26].
In addition to the determination of thermodynamic parameters -- as the electron chemical potential, the density of states, and bond orders --the theoretical model outlined above [9] may be also applied for modeling the lithium diffusion mechanism in the interlaminar spaces of MoS2. Lithium diffusion activation energies correspond indeed to the energy needed for moving the lithium from an octahedral site into the next tetrahedral one.
In the case of LixMoS2, for which we have determined the activation energy by measuring the lithium diffusion coefficients at different temperatures, it can be seen that the diffusion mechanism does not change with the lithium content as indicated by the unchanged DHþ observed in the studied stoichiometric range. That may be graphically appreciated in Fig. 3 in which the behavior of LixMoS2 is compared with those of other PEO– based nanocomposites (vide infra). However, the diffusion coefficients even decrease with increasing lithium content. That should be due to the variables contained in the pre-exponential factor of the Arrhenius relationship and may be interpreted in a rough approach as proportional to the corresponding activation entropy changes[30].
| ]]>
Fig. 3. Influence of lithium concentration in the activation energy for the diffusion of lithium |
A different behavior is observed for MoO3. Thus, as observed in Fig. 4, in the same lithium concentration range studied for the MoS2 (x=0.1-0.6) the activation energy decreases with increasing lithium content. However, at high lithium content, at about x=0.8, the activation energy again shows a relatively high value. The influence of lithium intercalation degree on lithium diffusion activation energy should be related with the effect of the charge transferred to the host. As shown schematically in Fig.5 because of lacking effective back-donation mechanisms -- probably due to a deficient overlap of the lithium-ion empty orbitals with those of its oxygen local environment in the host -- an accumulation of the negative charge in the matrix oxygen layers is produced, affecting thus the chemical potential of lithium in both the ground and the activated state in a different degree.
 | Fig. 4 Influence of lithium concentration in the activation energy for the diffusion of lithium in LixMoO3. |
 | Fig. 5 Comparative charge hostguest charge exchange and lithium diffusion activation energies in molybdenum trioxide and molybdenum disulfide. X(Q), host electronegativity /ref. 17). |
Environment effects on lithium diffusion may be appreciated in the case of the PEO-MoS2 nanocomposites. Lithium diffusion coefficients in LixMoS2 intercalated with one or one half mol PEO per mol molybdenum disulfide differ each other not only in the magnitude but also in its behavior against the variation of lithium content. This feature are clearly observed in Fig. 3 in which the activation energies of these nanocomposites with those of the LixMoS2 may be compared. In all the cases, a relatively linear behavior of DH# with lithium content is observed. However, this parameter is not constant as in MoS2. Indeed, it increases slightly for the compound with one mol PEO and rather notoriously for the complex with 0.5 mol PEO. The activation energy values observed for the PEO intercalation complexes result to be quite higher than that for pure MoS2 and of the same magnitude order than that observed in MoO3. That confirms our hypothesis that the activation enthalpy should be determined by the energy of the ground state which, in turn, is determined by the coordinative environment of the lithium ion. The nearly constant value of DH# for the complex LixMoS2(PEO)1.0, whose behavior is similar to that of pure MoS2, indicates that practically the same mechanism is valid for the whole x-range. Contrastingly, for the complexes LixMoS2(PEO)0.5 the situation is rather different and a constant change of the activation energy is apparent.
]]> Interlaminar distances observed for the phases commented above permit in a first approach to assume for the compound with one mol PEO per mol MoS2 a PEO-bilayer structure. The analysis of the IR spectra of the products as well as their comparison with those informed for other PEO intercalates, as those in MoO3[31] and CuFeS2[32], and free PEO[33] confirm such an assumption. Indeed, the vibrational bands observed for the polymer in the complexes Li0.1MoS2(PEO)1 would correspond to a bilayer constituted by two PEO strands with zigzag conformations, leading to a configuration similar to that existent in the free polymer. In the complex Li0.1MoS2(PEO)0.5, in turn, the polymer would have a simple zigzag configuration.The structural view of the discussed PEO-MoS2 phases illustrated schematically in Fig. 6, which can be obtained from powder X-ray diffraction as well as IR analysis, agrees with both the thermodynamic and kinetic features discussed above, and corresponds, moreover, to the behavior expected for the coordination of lithium ion in a sulfur or an oxygen environment.
 | Fig. 6 Schematic description of the intercalates LixMoS2(PEO)0.5 and LixMoS2(PEO)1.0 |
CONCLUSIONS
The comparison of the behavior of lithium in environments with different Lewis-base properties discussed above lead to following conclusions: i. The lithium chemical potential is strongly affected by its environment in the intercalated phase. Thus, the activity of lithium ion increases with the hardness of lithium coordination sphere, so higher electrode potentials are observed in oxygen hosts as in MoO3 or in a PEO environment. ii. The activation energy for lithium diffusion results to be higher for oxygen than for sulfur lithium ligands, thus reflecting a higher stabilization of the ion ground state by hard donors. iii. According to theoretical considerations the main cause of the dependence of thermodynamic and dynamic properties of lithium ion in the intercalated state on its environment appears to be the polarization of the latter. Thus high polarizable sulfur ligands favor electron back donation mechanisms leading to a net host-guest charge transfer lower than in hard oxygen medium.
ACKNOWLEDGMENTS
]]> Research partially financed by Fundación Andes (C12510), European Union (CI1-CT93-0330), DID Univ. de Chile and FONDECYT( 298 0040 and 198 1082).REFERENCES
1. J.L. Atwood, J.E.D. Davis, D.D. Macnicol, Eds. ,"Inclusion Compounds", Academic Press, New York, 1994. [ ]
2. M.S. Witthingan, Prog. Solid St. Chem. 12, 41 (1978). [ ]
3. H.Friend and A.D. Yoffe, Advances Physics 36,1 (1987). [ ]
4. C. Julien and G.A. Nazri, "Solid State Batteries. Materials Design and Optimization", Kluwer Academic Publishers, Boston, M A (1994). [ ]
5. B. Scrosati, J. Electrochem Soc. 139, 2276 (1992). [ ]
6. S. Megahed and B. Scrosati, Interface 4, 34 (1995). [ ]
7. E.J. Plichta and W.K. Behl, J. Electrochem Soc. 140, 46 (1993). [ ]
8. O. Duscheck and V. Gutmann, Mh. Chem. 104, 1259 (1973). [ ]
9. F. Mendizábal, R. Contreras and A. Aizman, Int. J. Quantum Chem. 56, 819 (1995). [ ]
10. N. Nudelman and A. Vitale, Org. Prep. Proc. 13, 44 (1981). [ ]
11. D. W. Murphy, F. J. Di Salvo, G. W. Hull and J. V. Waszczak, Inorg. Chem. 15, 17 (1976), [ ]
12. G. González, M.A. Santa Ana, E. Benavente, J.P.Donoso, T.M. Bonagamba, N.C. Mello, H. Panepucci, Solid State Ionics 85, 225 (1996). [ ]
13. G. González, M.A. Santa Ana, E. Benavente, J. Phys. Chem. Solids 58, 1457 (1997) [ ]
14. S. Basu and W.L. Worrel in "Fast Ion Transport in Solids", Edited by P. Vashista, J.N. Mundy and G.K. Sheny, North Holland, Amsterdam, 1979. [ ]
15. M. A. Py and R. R. Haering, Can. J. Phys. 61, 76 (1983). [ ]
16. D.M. Thomas, E.M. McCarron, Mat.Res.Bull., 21, 945 (1986). [ ]
17. F. Mendizábal, R. Contreras, A. Aizman, J. Phys. Condens. Matter 9, 3011 (1997). [ ]
18. B. Silbernagel and M. Whittingham, J. Chem. Phys. 64, 3690 (1976). [ ]
19. B. Silbernagel, Solid State Commun. 17, 361 (1975). [ ]
20. G. González, H. Binder, Bol.Soc. Chil. Quím. 41, 121 (1996). [ ]
21. a)R.G. Pearson, J.Am. Chem. Soc. 85, 3533 (1963) [ ]
(b) R.G. Pearson, Coord. Chem. Rev. 100, 403 (1990). [ ]
22. J.E. Huheey, J. Org. Chem. 36, 204 (1971). [ ]
23. J.E. Huheey, J. Phys, Chem. 69, 3284 (1965 [ ]
24. E. Benavente and G. González, Mat. Res. Bull. 32, 709 (1997). [ ]
25. G. González, M.A. Santa Ana, E. Benavente, Electrochim. Acta 43, 1327 (1998). [ ]
26. E. Benavente and G. González, Bol. Soc. Chil. Quím. 42, 555 (1997). [ ]
27. F. Croce, S. Passerini, A. Selvaggi, B. Scrosati, Solid State Ionics 40/41, 375 (1990). [ ]
28. D. Peramunage, D. Pasquariello, K.M. Abraham, J. Electrochem. Soc. 142, 1789 (1995). [ ]
29. K.M. Abraham and Z. Jiang, J. Electrochem. Soc. 143, 1 (1996). [ ]
30. M.A. Santa Ana, V. Sánchez and G. González., Electrochim. Acta, 11, 1773 (1995). [ ]
31. L.I. Nazar, H. Wu, W.P. Power, J. Mat. Chem. 5, 11 (1995). [ ]
32. C.Mujica, R.Durán, J.Llanos, R. Clavijo, Mat. Res. Bul. 5, 483 (1996). [ ]
33. R. Iwamoto, Y. Saito, H. Ishihada, H. Tadokoro, J. Polym. Sci. Part A-2, 6, 1509 (1983). [ ] ]]>
