Tris(oxalato)phosphates, method for their preparation and their use

This application is a 371 of PCT/EP/00/04301 filed May 12, 2000 now WO 01/07450.

The invention relates to tris(oxalato)phosphates, M[P(C 2 O 4 ) 3 ], a method for their preparation and the use of tris(oxalato)phosphates, inter alia, as conducting salts in electrochemical storage systems.

Electrochemical storage systems are e.g. batteries and so-called supercapacitors. Electrolyte solutions comprising a conducting salt and an aprotic solvent are used in these systems. Modern systems, such as e.g. lithium ion batteries, have a high power density and output voltage (often≧3 V). Aprotic electrolyte systems are required for these cells.

Lithium hexafluorophosphate (LiPF 6 ) is currently used as the conducting salt in all commercial lithium ion batteries. This salt has the necessary prerequisites for use in high-energy cells, i.e. it is readily soluble in aprotic solvents, it leads to electrolytes with high conductivities, and it has a high degree of electrochemical stability. Oxidative decomposition occurs only at potentials>approx 4.5 V.

However, LiPF 6 also has serious disadvantages which are chiefly attributed to its lack of thermal stability. A dissociation, although slight, into LiF and PF, which can lead to cationic polymerization of the solvent caused by the Lewis acid PF., takes place in solution.

On contact with moisture, corrosive hydrogen fluoride is liberated, which on the one hand makes handling difficult because of its toxicity and corrosiveness, and on the other hand can lead to (partial) dissolution of the transition metal oxides employed as the cathode material (e.g. LiMn 2 O 4 ). The cycle stability of the electrochemical energy store concerned is affected in this manner.

Against this background, there have been intensive efforts with the aim of developing alternative conductive salts. Above all, lithium salts with perfluorinated organic radicals have been tested as such salts. In particular lithium trifluoromethanesulfonate, lithium bis(trifluoromethanesulfonyl)-imide and the lithium methides, the simplest parent substance of which is lithium tris(trifluoromethanesulfonyl)methide, have been mentioned. These salts also have disadvantages which have so far impeded their use in commercial lithium batteries. The first mentioned salt does not impart a sufficiently high conductivity to the electrolytes prepared with it. The salts mentioned last indeed have a conductivity equivalent to that of LiPF 6 , but they are of no commercial interest because of the expensive preparation process. Furthermore, the imide is corrosive towards aluminium sheets, which are employed as current diverters in many battery systems. Because of the high fluorine content of the compounds, exothermic reactions with the lithium of the electrode are, moreover, to be feared under adverse conditions.

Lithium hexafluorophosphate and all the abovementioned conductive salt alternatives have the common feature of a rather high fluorine content. On the basis of this fact, the preparation costs are comparatively high and certain safety precautions have to be taken in the disposal or recycling of spent batteries in order to avoid emission of fluorine-containing substances (e.g. toxic and corrosive hydrogen fluoride HF).

The lithium borate complex salts [(R′O) 2 B(OR″) 2 ]Li described in DE 19633027 A1 represent a considerable advance. In these, R′ and R″ are the same or different, R′ and R″ are optionally bonded to one another by a single or double bond, and R′ and R″ in each case individually or together denote an aromatic ring from the group consisting of phenyl, naphthyl, anthracenyl or phenanthrenyl, which can be unsubstituted or mono- to tetrasubstituted by A or Hal, wherein Hal represents fluorine or chlorine and A is an alkyl residue having 1 to 8 C atoms, which can in turn be mono- to tetrahalogenated.

The stabilities of the non-fluorinated derivatives, which are indeed improved, but are in no way adequate for the 3 V systems required, are a disadvantage of these compounds. Thus, e.g. unsubstituted lithium bis[1,2-benzenediolato(2-)-O,O′]borate(1-) (which is the 2:1 complex of pyrocatechol) already decomposes when an anodic potential of 3.6 V is exceeded. This value is significantly below that of the standard conductive salt LiPF 6 (approx. 4.5 V). Even in the case of these chelatoborates, too, only fluorine-substituted derivatives are sufficiently stable to oxidation.

A chelatophosphate, namely lithium tris[1,2-benzenediolato(2)-O,O′]phosphate, has been tested as another alternative (M. Handa, M. Suzuki, J. Suzuki, H. Kanematsu, Y. Sasaki, Electrochemical and Solid-State Letters, 2 (2) 60-62 (1999)). This salt has a wider electrochemical stability range than the corresponding boron compound (start of decomposition from about 3.7 V), but the maximum conductivities which can be achieved for electrolyte solutions prepared with this are below 4 mS/cm; i.e. significantly below the standard given by LiPF 6 .

Salts with large cations (e.g. N(R 1 R 2 R 3 R 4 ) + , wherein R 1 , R 2 , R 3 and R 4 are independently of one another H or an alkyl group having 1 to 8 C atoms, are commonly used for supercapacitors since these are largely inert towards the electrode materials.

The invention is therefore based on the object of eliminating the disadvantages of the prior art and of providing halogen-free electrochemically stable and thermally stable compounds which are readily soluble in aprotic solvents and are suitable for the preparation of electrolyte solutions with a good conductivity. Furthermore, catalysts should be found, such as those employed, for example, for the hydroamination of amines. The invention is furthermore based on the object of providing a method for the preparation of these compounds.

The object is achieved with tris(oxalato)phosphates of the general formula M[P(C 2 O 4 ) 3 ], where N=H, a metal or N(R 1 R 2 R 3 R 4 ), wherein R 1 , R 2 , R 3 and R 4 independently of one another are H or an alkyl group having 1 to 8 C atoms. Metal tris (oxalato) phosphates are as preferred compounds and hydrogen tris(oxalato)phosphate, lithium tris (oxalato) phosphate and sodium tris(oxalato)phosphate are as particularly preferred compounds.

It has been found, surprisingly, that these tris(oxalato)phosphates have the required property profiles and furthermore are easy to prepare. The compounds are readily to very readily soluble in polar-aprotic solvents.

FIG. 1 is a cyclic voltammogram of a solution of lithium-tris(oxalato phosphate), (Li[P(C 2 O 4 ) 3 ]) in EC/DMC (1:1) against an Ni electrode.

FIG. 2 is a graph showing the conductivity of a solution of lithium-tris(oxalato phosphate), (Li[P(C 2 O 4 ) 3 ]) in EC/DMC (1:1) at room temperature.

FIG. 3 is a graph showing the conductivity of a solution of lithium-tris(oxalato phosphate), (Li[P(C 2 O 4 ) 3 ]) in 1,2-DME/PC/DMC (1:1) at room temperature.

Lithium tris(oxalato)phosphate has been investigated in more detail. The solubilities in various polar-aprotic solvents are shown in Table 1.

THF=tetrahydrofuran; 1,2-DME=1,2-dimethoxyethane;

EC=ethylene carbonate; DMC=dimethyl carbonate; PC=propylene carbonate;

Et 2 O=diethyl ether

According to thermogravimetric findings, decomposition starts only above 150° C. The specific conductivities are significantly above those of the chelatophosphate compound described by Handa et al.: In binary carbonate mixtures up to approx. 7 mS/cm is recorded. If ether-functionalized co-solvents, such as e.g. tetrahydrofuran (THF), ethylene glycol ether, polyethers or 1,3-dioxolane, are added, even significantly higher conductivities are measured. A 20% solution in propylene carbonate/1,2-dimethoxyethane (1:1) thus has a conductivity of 9.7 mS/cm. On inert electrodes (e.g. platinum or nickel), electrochemical decomposition only starts significantly at above 4 V (see FIG. 1 ).

To prepare tris(oxalato)phosphates, in a first reaction step phosphorus pentachloride is reacted with anhydrous oxalic acid in the presence of an aprotic solvent according to the following equation:

PCl 5 +3C 2 O 4 H 2 →H[P(C 2 O 4 ) 3 ]+5HCl

For this, the oxalic acid is expediently initially introduced into the solvent and PCl 5 is measured in (on a laboratory scale with e.g. a solids measuring bulb). However, it is also possible initially to introduce PCl 5 into the solvent and then to add the oxalic acid. An ether (e.g. diethyl ether, tetrahydrofuran, 1,2-dimethoxyethane) or a carbonate (e.g. ethylene carbonate, dimethyl carbonate, diethyl carbonate, propylene carbonate) or a hydrocarbon (e.g. alkane having 5 to 12 C atoms or aromatics, such as e.g. benzene or toluene) or a halogenated hydrocarbon or a partly halogenated hydrocarbon or a mixture of these substances can be employed as the aprotic solvent. Depending on the dissolving power of the solvent used, the two reactants are at least partly dissolved (e.g. with ethers as the solvent) or merely suspended (e.g. with hydrocarbons as the solvent). The reaction temperature is −20 to 120° C., preferably 0 to 100° C. It has been found that the reaction was usually complete within a few minutes to hours (depending on the batch size, rate of addition of the reactants, dissolving power of the solvent, temperature).

The hydrogen chloride (HCl) forming as a by-product already largely escapes during the synthesis via the gas phase, depending on the solvent chosen and the reaction temperature. For use of the later product as a conductive salt in electrochemical systems, it is necessary for this to be largely freed from chloride, e.g. the chloride content in lithium tris(oxalato)phosphate should be<20 ppm.

There are several possibilities for the complete removal of HCl: When the reaction is complete, by boiling the reaction mixture under reflux, stripping by passing an inert gas stream (e.g. nitrogen or argon) through the reaction container, carrying out the reaction under reduced pressure or distilling off some or all of the solvent. When the solvent is removed completely, the chelatophosphoric acid is obtained as a solid, which can be liberated completely from volatile acid impurities under reduced pressure at temperatures of preferably 20 to 50° C.

In some cases, e.g. if diethyl ether is used as the solvent, the hydrogen chloride can also be removed by liquid/liquid separation. This is possible because two liquid phases are formed during the reaction: a heavy phase which contains the desired intermediate product in the form of an ether complex and little HCl, and a light phase floating on top in which the HCl is concentrated. Since the phosphorus compound is only very poorly soluble in ether, no noticeable amounts of product are to be found in the upper phase. The upper phase is separated off and the lower phase containing the intermediate product is extracted several times with pure ether, until no further acid is detectable in the upper phase.

Several of these process steps can also be combined for removal of HCl which is as complete as possible.

The intermediate product tris(oxalato)phosphoric acid (hydrogen tris(oxalato)phosphate), which is, as far as possible halide-free, is converted into the metal tris(oxalato)phosphate in a subsequent reaction step by reaction with the corresponding metal or metal derivative according to the following equations

H[P(C 2 O 4 ) 3 ]+M→M[P(C 2 O 4 ) 3 ]+½H 2 , or

H[P(C 2 O 4 ) 3 ]+M′B→M′[P(C 2 O 4 ) 3 ]+BH

in an aprotic solvent, described above, at temperatures of 0 to 80° C., preferably 10 to 50° C. All metals can be employed as the metals M but Li, Na, K, Rb and Cs are preferred. If a polyvalent metal (e.g. Mg, Zn or rare earths) is employed, the stoichiometry in the formulae given is to be adjusted accordingly. A preferred metal derivative M′B comprises a cation of the alkali metals mentioned or an ammonium or a substituted ammonium ion (N (R 1 R 2 R 3 R 4 ) + , wherein R 1 , R 2 , R 3 and R 4 independently of one another are H or an alkyl group having 1 to 8 C atoms, and a base B, wherein B=H or C(R 1 R 2′ R 3′ ) or N(R 1′ R 2′ ) or OR″, wherein R 1′ , R 2′ and R 3′ independently of one another are H or an alkyl group having 1 to 8 C atoms and R″ is an alkyl group having 1 to 8 C atoms.

Examples of alkali metal derivatives are hydrides (LiH, NaH, KH, RbH, CsH), organometallic compounds (e.g. methyllithium, butyllithium), amides (e.g. LiNH 2 , NaNH 2 , KNH 2 or lithium diisopropylamide) or alkoxides (e.g. Li tert-butoxide, Na tert-butoxide, Na tert-amoxide, K tert-butoxide). The hydrides and organometallic compounds are particularly preferred, since these are very reactive, easily removable, form acceptable by-products (hydrogen or the organic residues in the form of alkanes) and are commercially available.

To prepare the ammonium salt, dry ammonia NH 3 is passed in as the “metal derivative” M′B:

H[P(C 2 O 4 ) 3 ]+NH 3 →NH 4 [P(C 2 O 4 ) 3 ]

The quaternary ammonium salts can be prepared by double decomposition.

M″[P(C 2 O 4 ) 3 ]+X[N(R 1 R 2 R 3 R 4 )]→[N(R 1 R 2 R 3 R 4 )][P(C 2 O 4 ) 3 ]+M″X

where M″=H, Li, Na, K, Rb, Cs and where X=F, Cl, Br, I, NO 3 .

The intermediate compound where M″=H is preferred here, since in this case volatile acid, which can be removed via the gas phase, is formed in the reaction. If M″=Li, Na, K, Rb or Cs, a solvent in which M″X is insoluble is to be chosen, e.g. diethyl ether or hydrocarbons.

When the reaction has taken place, the product can remain in solution or can be liberated from the solvent by e.g. evaporation and drying. The product can be recrystallized for further purification.

The metal tris(oxalato)phosphates according to the invention of the general formula M[P(C 2 O 4 ) 3 ] are used as conductive salts in electrochemical storage systems, and in particular lithium tris(oxalato)phosphate Li[P(C 2 O 4 ) 3 ] can be used as a conductive salt in lithium ion batteries.

The tris(oxalato)phosphates according to the invention, in particular the free acid hydrogen tris(oxalato)phosphate, find use as catalysts and additives in organic synthesis, e.g. analogously to the compounds and reaction mechanisms described by Tokunaga, Eckert and Wakatsuki (“Ruthenium-catalysed intermolecular hydroamination of terminal alkines with anilines: a practicable synthesis of aromatic ketimines”, Angew. Chem. 1999, 111, no. 21, p. 3416-3419). It has been found there e.g. that HPF 6 and HBF 4 and their particular ammonium salts are effective additives in the hydroamination of alkines.

The subject matter of the invention is explained in more detail with the aid of the following examples: