Patent Application
20100239916
The present invention relates to a non-aqueous electrolyte and has particular reference to a non-aqueous electrolyte which can be used in a primary or secondary battery, such as a lithium battery, in a supercapacitor, in an electrochromic device or in a solar energy cell. Furthermore, the present invention relates to a non-aqueous electrolyte when used in any of the foregoing devices.
Lithium batteries are known in non-rechargeable and in rechargeable form. Such batteries comprise positive and negative electrodes with a non-aqueous electrolyte disposed between them.
In a rechargeable lithium ion battery (secondary battery) the positive electrode of the battery can for example be LiCoO2 (referred to as the “cathode” in Li-battery community) and the negative electrode can for example be carbon (referred to as the “anode” in Li-battery community). In a non-rechargeable battery (primary battery) the positive electrode can for example be MnO2 and the negative electrode can be lithium metal. Various different types of electrolyte are known. For example there is the class of liquid electrolytes comprising at least one ionically conducting salt such as Li(TFSI), i.e. lithium bis(trifluorosulphonyl)imide, LiPF6, i.e. lithium hexafluorophosphate or LiClO4, i.e. lithium perchlorate which are present, with a low degree of dissociation within a non-aqueous solvent such as a mixture of DME (dimethylethane) and EC (ethylene carbonate), a mixture of DEC (diethylene carbonate) and EC, or a mixture of DMC (dimethyl carbonate) and EC or PC (propylene carbonate) or combinations thereof.
In addition there are so-called dry polymer electrolytes. In these electrolytes the salt is selected as before (i.e. for example from Li(TFSI), LiPF6 or LiClO4) and is dispersed in a polymer or mixture of polymers. Suitable polymers comprise PEO (polyethylene oxide), PVDF (polyvinylene difluoride), PAN (polyacrylonitrile), and PMMA (polymethyl methyl acrylate).
Furthermore, there are so called polymer gel electrolytes. These have the same basic composition as the dry polymer electrolytes recited above but include a solvent, for example a solvent of the kind recited in connection with the liquid electrolytes given above.
The known liquid electrolytes described have the advantage that they have a high ionic conductivity up to a transference number of 6 and a high conductivity of 10−2 S/cm. In addition the liquid properties ensure good wetting of the electrode surface. They are however dangerous because leakage can occur, so that safety considerations arise. In addition they can lead to passivation effects which are undesirable.
The dry polymer electrolytes do not result in good wetting of the electrodes, the conductivities which can be achieved are quite low and there is also not much scope for modifying the chemical composition of the ingredients. However, the electrolytes are good safety-wise and no leakage occurs.
With the polymer gel electrolytes the change in liquid content results in reductions in the conductivity and there is also the danger of leakage.
The object of the present invention is to provide an electrolyte comprising a lithium salt and a solvent as before but with an improved conductivity.
In order to satisfy this object there is provided, in accordance with the present invention, a non-aqueous electrolyte including
The applicants have namely found that the addition of fine oxide particles, e.g. in powder or elongate particle form, leads to a substantial increase in conductivity but with no disadvantages.
The electrolyte preferably has a low degree of dissociation, preferably with association constant in the range from 1×10−1 to 108/l−1·mol−1.
When used in a primary or secondary lithium battery having positive and negative electrodes, the oxide should be selected such that it does not react with the material of either of said positive and negative electrodes.
The non-aqueous electrolyte of the present invention is not restricted to use in a battery, it can for example be used in a supercapacitor, in electrochromic devices such as electro-chromic displays or in a solar energy cell.
In the non-aqueous electrolyte of the invention the ionically conductive salt is selected from the group comprising Li(TFSI), LiPF6 and LiClO4.
Moreover, the non-aqueous, anhydrous solvent is preferably selected from the group comprising DEC/EC, DMC/EC, PC, carbonate based solvents related to any of the foregoing, DMSO, organic sulphur compounds, THF, AN and mixtures of any of the foregoing.
The oxide used for the invention is preferably selected from the group comprising oxides exhibiting acidic properties, for example SiO2, TiO2 and oxides exhibiting basic properties, for example Al2O3, MgO and any mixtures thereof.
The average particle size of the oxide for particles of approximately spherical shape, is selected to be less than 5 μm and preferably less than 2 μm, with no lower limit other than that set by manufacturing techniques used to produce said oxide. For elongate particles, such as nano-wires or nano-tubes, the average diameter is selected to be less than 1 μm, preferably less than 100 nm, there being no limit on the length of such elongate particles.
The amount of oxide present in the electrolyte is preferably such as to give the electrolyte a consistency between that of a liquid and a solid, preferably a consistency similar to that of a soggy sand, i.e. a liquid and sand mixture having a consistency such that sedimentation effects do not occur.
The invention will now be described in further detail with reference to basic designs of Lithium batteries known in the prior art as shown in
Referring first of all to
It has already been established by others that transition metal oxides, more specifically the oxides Fe, Ni, Co, Cu, show promising performance as anode materials for rechargeable lithium batteries. The reversible Li-storage mechanism is due to the formation and decomposition of Li2O upon the electrochemical insertion/extraction of Li. The key point for the reversibility seems to be the formation of a special microstructure in which metal clusters are dispersed uniformly with Li2O at a nanoscale after electrochemical reaction of metal oxide with Li.
German patent application 102 42 694.5 assigned to the present applicants recognizes that LiF, which is electrochemically inactive, shows electrochemical activity when it is dispersed with a transition metal at atomic or nanoscale level. A series of transition metal fluorides (M═Ti, V, Mn, Fe, Co, Ni, Cu) were investigated and led to favorable results.
The experimental setup was as follows:
The working electrodes comprised either TiF3A:TiF3:PVDF (9:1) on Ti foil; or TiF3B:TiF3:CB:GP:PVDF (10:0.5:0.5:1) on Ti foil. CB denotes carbon black, GP denotes graphite and PVDF denotes polyvinylene di-fluoride. The pasting procedures for the electrode film on the Ti-foil are similar to those reported in H. Li. L. H. Shi, W. Lu, X. J. Huang, L. Q. Chen, J. Electrochem. Soc., 148, A915 (2001)
The electrochemical cells tested were 2-electrode cells for electrochemical testing similar in layout to the known cell of
The materials can be the same as those described with reference to
Two examples will now be given for the preparation of non-aqueous, anhydrous electrolytes:
The composite electrolyte was prepared in the form of a mixture of fine particles of ceramic oxides (SiO2, TiO2 and Al2O3, radius: r ˜0.15 μm) with a non-aqueous solution of 0.1M LiClO4 in Methanol (MeOH). Although MeOH was selected for this test it is not a preferred solvent for lithium batteries. However, because the invention is effective using MeOH it appears certain that it will work better in the preferred solvents, i.e. in a non-aqueous, anhydrous solvent selected from the group comprising DME/EC, DEC/EC, DMC/EC, PC, carbonate based solvents related to any of the foregoing, DMSO, organic sulphur compounds, THF, AN and mixtures of any of the foregoing, and indeed practical tests have confirmed this.
In this example the oxides were dried in vacuum at 250° C. for 24 hours prior to composite preparation and all samples were prepared under an Argon atmosphere in a glove box (closed container with a window and gloves for handling the materials involved). Room temperature conductivity was measured for various volume fractions using impedance spectroscopy with the samples placed between two parallel stainless steel electrodes of a home-built cell (samples loaded under Argon). The impurity effects were excluded by pre-washing the oxides in the liquid solvents. The conductivity was better than 2×10−3 S/cm.
This is interpreted as suggesting that the conductivity enhancement in the liquid-solid composite is attributable to the formation of a highly conducting layer around the oxide filler with a higher Li ion conductivity compared to that in the solution caused by adsorption of the anion and breaking up of the ion-pair. This is also supported by ζ-potential measurements (φoxide≈5×10−5, [LiClO4]≈10−3M) arising out of the effective surface charge showing a more negative potential measuring from Al2O3 (−18.3 mV) to SiO2 (−36.3 mV). In the regime of lower oxide fractions, i.e. in the colloidal regime, the double layer repulsion works against percolation and leads to the very poor effect in the case of SiO2. The earlier but smoother increase (lower percolation threshold) in that regime of the Al2O3 composite is in agreement with a lower charge density at the Al2O3 surface. As soon as the particles are forced to be sufficiently close (Debye length, λ˜100 nm) the interfacial conductivity percolates (in a cubic arrangement the minimal particle-particle distance (δ) is assessed to be 2δ at φ≈6.5%). At this volume fraction range the situation resembles the Nafion-type electrolytes in which the counter ion is attached to the polymer backbone and the proton is mobile within tiny water filled channels. Since the freed counter ion needs solvent to be mobile the conductivity breaks down at higher φ, (in cubic arrangement δ≈0 for φ≈52%). The maximum conductivity enhancement was observed for SiO2-composite with σm=1.2×10−2 S/cm which is of the order of a high conducting electrolyte used for Li-battery. The interfacial picture is also supported by the fact that the SiO2 with larger surface area per unit volume, i.e. with an average particle size of 0.3 μm, (inset
The role of the oxides in dissociation of ion-pairs in non-aqueous solution was further investigated by preparing composite electrolytes comprising a solution of an Li-salt in tetrahydrofuran (THF) with added SiO2. The THF exhibits a significantly lower dielectric constant (∈THF=7.4) compared to MeOH (∈MeOH=32.6). Indeed, as shown in
The effective composite conductivity (σm) can be written as follows:
σm=βsφsσs+β1φ1Δσm,1 1)
where φs and φ1 are the volume fractions for solution and high conducting layer respectively whereas σs and Δσm,1 are respectively the ionic conductivities for the solution and excess conductivity of the high conducting layer. The β-factors measure the degree of percolation and are of the order of unity for a parallel switching (only percolating pathways). They are distinctly less and depend on φ if this is not the case. Equation (1) leads to the estimate of lower limit of the maximum interfacial conductivity according to
which is plotted in
(where u is the mobility of the Li ions, F is the Faraday constant, ∈ is the dielectric constant, T is the absolute temperature and co is the concentration in the first layer adjacent to the adsorption layer) for semi-infinite boundary conditions and large effects while it tends to uFco for a vanishing solution channel width (given a sufficient thickness to maintain mobility).
Since the dielectric constant of CH3OH is comparable to that of typical Lithium battery electrolytes a marked improvement is expected when using improved “soggy sand” electrolytes of the invention based on electrolytes which are (meta)stable in Li-batteries. The experiments conducted to date give strong evidence for the possibility of enhancing ion transport in non-aqueous solutions by breaking ion-pairs at the interfaces to appropriate second phases. Beyond that they indicate the validity of heterogeneous doping mechanism proposed for inorganic crystals to be operative also for polymer electrolytes.
The experiments conducted provide evidence for the usefulness of composite electrolytes consisting of liquid non-aqueous salt solutions and solid insulating particles. At high second phase contents the investigated system show distinctly enhanced conductivities and are better described as “soggy sand” electrolytes, a viscous grain ensemble wetted by the liquid. Unlike solid-liquid composites described by Amita Chandra and Joachim Maier in Solid State Ionics 148, pages 153 to 158 (2002), the effects are not additive rather they are synergistic in the sense that the overall conductivity is distinctly higher than both of the constituent phases, an effect that is ascribed to interfacial interaction. The “soggy sand” electrolytes of the invention as described above combine enhanced conductivities with the favourable mechanical properties of the soft matter.
Polymer electrolytes under consideration consist of a salt dissolved in a covalent polymer matrix (the cases of polymer electrolytes in which one ion is strongly bound to the polymer (e.g. Nafion) are not expected to exhibit a heterogeneous doping effect as touched upon later). They are materials of prime importance in the context of electrochemical devices. Since the compositional flexibility however is not unlimited, strategies for optimising the conductivity properties of a given conductor are necessary.
In crystalline electrolytes, i.e. influencing the conductivity by adding second phase particles proved very helpful in different cases. Indeed it has been found [Scrosati, Wieczorek] that the dispersion of oxides leads to significant conductivity increases in crystalline electrolytes. Different explanations have been proposed, most of them considering mobility changes owing to segmental motion of the polymers, variation of the degree of crystallinity etc., whereas the effect in composite inorganic electrolytes, to which the present invention is directed, has been essentially attributed to a charge carrier concentration effect.
The heterogeneous doping effect consists of internally adsorbing one ion sort and hence effectively leading to dissociation. In the crystalline state e.g. AgCl:Al2O3 this may be the adsorption of Ag+ or F− ions connected with its removal from the energetically deep-lying regular positions resulting in the generation of vacancies as mobile carriers. In the covalent matrix the immobile ground state would be the undissociated ion pair with the conductivity effect consisting in the adsorption of one pair constituent resulting in breaking of the ion pair and generating a mobile counter ion as described by J. Maier, Prog. Solid St. Chem. 23 pages 171 to 263 (1995). Testing this hypothesis is rather difficult since a polymer matrix may be different in various ways. It was for this reason that a decision was made to test the effect of oxide additions to a salt in a liquid solvent of low polarity, i.e. MeOH. It was found that the conductivity is significantly enhanced with a percolation behaviour typical for interfacial conductivity and that the different impact of oxides with different basicity points towards a perchlorate adsorption and hence increased Li+ conductivity.
To investigate whether percolation type behaviour arises with a pronounced increase in conductivity for low volume fractions of SiO2, two low molecular solvents were prepared. The following liquid polymers were used:
The following properties were obtained for liquid polymers 1) and 2):
In a first step, the solvents were admixed with different concentrations of lithium salt (LiClO4). It was discovered that the highest values for the ionic conductivity were found at a concentration of 1 Mol. (see Table 1 below and
Table 1 shows the changes in ionic conductivity of electrolytes containing a PEG-150 solvent and for different concentrations of LiClO4, a reference sample of PEG-150 which contained no SiO2 was compared to samples of PEG-150 containing 1 vol. % SiO2 (7 nm-fumed SiO2) and a sample containing 1 vol. % SiO2 (10 nm grain size), at a steady state after 8 h for each respective lithium salt concentration.
Fumed SiO2, also known as pyrogenic silica, is a non-crystalline, fine grain, low density and high surface area silica. It is not to be confused with silica fume, which is also known as microsilica. Fumed silica can be produced in a flame pyrolysis procedure, in which silicon tetrachloride or quartz sand is vaporized in an electric arc, at e.g. 3000° C., as is known to the person of ordinary skill in the art. The primary particle size of fumed silica is typically 5 to 50 nm. The fumed silica particles are non-porous and have a surface area of typically 50 to 600 m2·g−1 and a density of 2.2 g·cm−3. It has been found that fumed silica with properties in the ranges set out above can be exploited to advantage in a non-aqueous electrolyte in accordance with the present teaching. It should be noted that this does not mean that fumed silica outside the quoted ranges are not useful, they have simply not been investigated.
In a next step, the amount of SiO2 added to a solvent including 1 M of LiClO4 were investigated. In particular, different volume fractions of different grain size SiO2, 7 nm-fumed (water content 0.0%-9.8 ppm) or 10 nm non-fumed (water content 0.03%-281.3 ppm) respectively, in both 1M PEG-150-LiClO4 and 1M PEG-350-LiClO4 were investigated. For this the amount of oxide particles added to the solvent (fumed and non-fumed SiO2), were dried at ˜400° C. for 24 h, under vacuum and stored in a dry glove box under an argon atmosphere (water content approximately 0.1 ppm, before use). Following this the composite electrolytes were prepared by dispersing the SiO2 in the corresponding LiClO4/PEG solution using a Vortex shaking device to obtain well dispersed solution. The concentration of the SiO2 in the composite electrolytes were varied from 0.5 to 20% by volume fraction of SiO2.
Following the preparation of the sample ionic conductivity was measured under a controlled temperature and a controlled argon flow.
Table 2 below shows different volume fractions of SiO2 of two different grain sizes, 7 nm-fumed (water content 0.0%-9.8 ppm) or 10 nm (water content 0.03%-281.3 ppm) respectively, in 1M PEG-150-LiClO4.
Table 3 below shows the different volume fractions of SiO2 for different grain sizes, 7 nm-fumed (water content 0.0%-9.8 ppm) or 10 nm (water content 0.03%-281.3 ppm) respectively, in 1M PEG-350-LiClO4.
It was also observed that for solutions containing SiO2 (especially fumed; non-crystalline, fine-grain, low density and high surface area silica, with higher density of silanol (Si—OH) groups on the surface; where the particles are non-porous and have a surface area of 50-600 m2·g−1), more stable gel-like composite solutions are obtained and separation of the electrolyte into the solid oxide and liquid fractions is prevented. This is of particular significance because it means that an electrolyte can be used which contains only the oxide fraction, a Li salt and a non-aqueous solvent for the Li-salt. It is not necessary to use either a separator or polymeric components to stabilize the electrolyte. In all cases enhanced conductivity due to percolation is obtained and the electrolyte has self healing properties. Its gel-like consistency prevents the two electrodes of a cell from contacting each other.
The 7 nm fumed SiO2 (obtained from Sigma-Aldrich, Missouri USA) has a surface area BETarea=605.0 m2·g−1 and a density of 2.438 g·cm−3.
Similarly, the 10 nm SiO2 (non-fumed SiO2 obtained from Sigma-Aldrich, Missouri USA) has an approximate value for BETarea=537.9 m2·g−1.
Measurements with electrolytes including SiO2 particles having a particle size of 300 nm (BETarea=8.2 m2·g−1) were also conducted and were also seen to have a higher ionic conductivity than the pure electrolyte.
Although the above discussion has been given with respect to SiO2 as the oxide fraction, it is believed that other oxide fractions in a low volume range can also yield beneficial results. For example, for a 1M PEG-150 solution with an addition of a LiOTf (Lithium trifluoromethanesulfonate) salt and an Al2O3 oxide, which is preferably a fine grain powder with an approximate particle size which is less than 10 nm, typically in the range of 2 to 10 nm, a higher conductivity was found for volume fractions in the range from 0.5 vol. % to 6 vol. %. The maximum increase in ionic conductivity here was found to be approximately 1.3 times the ionic conductivity of the reference sample, however, these results were dependent on time.
| Number | Date | Country | Kind |
|---|---|---|---|
| 03018161.4 | Aug 2003 | EP | regional |
This application is a continuation-in-part of co-pending U.S. patent application Ser. No. 10/913,959, filed Aug. 6, 2004, which claims the benefit of European Patent 03 018 161.4, filed Aug. 8, 2003, the entirety of both are incorporated by reference herein.
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10913959 | Aug 2004 | US |
| Child | 12748649 | US |
