ALKALI METAL ELECTROCHEMICAL CELL WITH COATED SOLID ELECTROLYTE
The present invention is concerned with alkali metal electrochemical cells employing a solid electrolyte separating two electrodes of which at least one comprises an alkali metal.
An example of such an alkali metal electrochemical cell is the sodium sulphur cell in which sodium forms one electrode and sulphur, together with sodium polysulphides, forms the other electrode. Cells are also known employing sodium as both electrodes. In operation, the electrodes are normally liquid which implies elevated operating temperatures for the cell. The solid electrolyte has to be electronically insulating but conductive to cations of sodium. Beta alumina is the preferred material for the solid electrolyte in sodium/sulphur and sodium/sodium cells, although ionically-conductive glass electrolytes can also be employed.
A long standing problem with cells of this kind employing a solid electrolyte, especially beta alumina, is the maintenance of the integrity of the electrolyte. It is important that the electrolyte does not permit intermixing of the electrode materials, or provide any electronic conduction paths between the electrodes. In spite of great care in production of the electrolyte bodies, these still often have limited life time before failure.
It is thought that many of the mechanisms which limit the life of the solid electrolyte body used particularly in sodium-sulphur cells relate to degradation of the electrolyte at the sodium/electrolyte interface.
Ulti ately such degradation can lead to electronic conduction through the solid electrolyte separating membrane with resulting cell failure. It is believed that surface defects of the electrolyte body, such as sub-critical cracks and blemishes, can act as initiation sites for the development of filaments or dendrites of sodium which can then propagate through the electrolyte.
Howevermuch care is taken in preparing the electrolyte body to ensure the minimum of surface defects, even minute such defects appear eventually to be potential causes of cell failure.
The present invention aims to tackle this problem.
In one aspect the invention provides an alkali metal electrochemical cell comprising a solid electrolyte which is electronically insulating and conductive to cations of the alkali metal, said electrolyte having surface pores and/or cracks, wherein said surface pores and/or cracks, over at least a part of a surface of the electrolyte which is exposed to the alkali metal, are at least partially filled with a material which is, at the operating temperature and during operation of the cell, a liquid which is electronically insulating or semiconductive, is conductive to cations of the alkali metal, wets the electrolyte and forms an interface with the alkali metal. The term "semiconductive" is used herein to mean having an electronic resistance higher than that of a metal.
The provision of such a liquid at least partially filling minute pores and/or cracks in the surface of the solid electrolyte exposed to the alkali metal can substantially reduce the incidence of cell failure resulting from failure of the electrolyte.
It is believed that in many instances, electrolyte failure results from propagation of minute
cracks and blemishes on the electrolyte surface. In the absence of the aforementioned liquid, such cracks may become filled with sodium and thereby present a relatively lower resistance path from the inner tip of the crack (in the body of the electrolyte) and the sodium metal at the outer face of the electrolyte. This lower resistance path produces in turn a sodium ion current concentration at the crack tip during the charge cycle of the cell. Propagation of cracks is caused through the development of a Poiseuille pressure gradient within the cracks caused by the localised high flux density of sodium metal. Although much development work has resulted in improved cell life times, the complete elimination of surface flaws is extremely difficult.
The liquid at least partially filling the pores and/or cracks in the surface of the electrolyte, as described in this invention, is believed to reduce the risk of propagation of cracks in the electrolyte by the above described mechanism. Because such cracks and flaws are at least partially filled with the liquid, rather than sodium metal, and the liquid is electronically insulating or at least semiconducting, there is no longer a lower resistance path completely along any cracks or flaws in the electrolyte, thereby reducing the tendency for a sodium ion current concentration at the inner tip of any crack. This in turn reduces the tendency for increased sodium metal flux along the cracks.
Preferably, said liquid coats said electrolyte surface or surface part, thereby completely filling pores and/or cracks and forming a liquid/liquid interface between the liquid coating and the molten alkali metal.
The presence of the liquid coating has other advantages. The liquid of the coating is one which wets the electrolyte. A problem with alkali metal
electrochemical cells, particularly sodium cells, is that the alkali metal does not always readily wet the surface of the electrolyte. As a result, even when an electrolyte is immersed in liquid sodium, regions of the surface of the electrolyte may not be wetted, with the result that during operation of the cell, there may be concentrations of ionic current flow particularly at the boundaries of non-wetted electrolyte. Apart from increasing the apparent resistance of the cell during operation, this current concentration, or any degree of non-homogenous conduction at the electrolyte surface, can have deleterious effects in terms of cell life. It is usually necessary with prior art cells to repeatedly charge and discharge new cells a few times, to promote complete wetting of the electrolyte, before the cell resistance drops to its expected operational value. In this example of the present invention, the electrolyte surface is wetted by the liquid coating, which in turn will form a continuous interface with the liquid alkali metal. In this way, the above mentioned problems of wetting of the electrolyte by the alkali metal are substantially obviated.
The liquid of the coating may be an ionically conducting salt which is molten at the operating temperature of the cell. For a cell in which the alkali metal is sodium and the solid electrolyte is beta alumina, the coating is conveniently sodium amide.
Sodium amide is electronically insulating and its ionic conductance to sodium ions is substantially less than the electronic conductance of molten sodium. It is, therefore, important to ensure that the liquid coating covering the electrolyte is as thin as possible.
A convenient and successful method of providing the liquid coating of sodium amide in a sodium electrochemical cell with a beta alumina electrolyte
includes the step of exposing the electrolyte and/or the sodium of the cell to ammonia gas, so that the ammonia reacts with sodium to form sodium amide which coats the electrolyte at least where exposed to sodium in operation or at least partially fills pores and/or cracks. In one procedure, the electrolyte is exposed to the ammonia so that ammonia is absorbed on at least a portion of the electrolyte surface to be exposed to the sodium.
The ammonia may react with loosely bonded sodium in the surface regions of the electrolyte itself or directly with liquid sodium when this is added. The two reactions may be described as follows:
2Na + 2NH3 > 2NaNH2 + H2 (reaction with Na liquid)
+ - + - + - + - (NH4 + NH2 ) + (Na + Beta ) > NH4 Beta + Na NH2
(reaction with "beta" i.e. beta alumina)
Another method according to the invention of making a sodium electrochemical cell with a beta alumina electrolyte includes the step of adding to the or each sodium electrode sodium amide or a reagent which reacts with sodium to form sodium amide. Normally between O.lg and lOg of sodium amide will be provided for each lOOcc of sodium, and preferably between 0.5g and 5g. Ammonium chloride may be used as a reagent or alternatively the additive may be one or more of sodium amide, ammonium chloride, ammonium azide or an amine.
In one method of making a cell, the electrolyte and/or sodium may be exposed to ammonium chloride vapour.
Figures l to 9 of the drawings are graphical representations of the resistance performance of various
cells to illustrate the improved electrolyte wetting resulting from examples of the invention.
In examples of the invention, electrolyte cups made from beta alumina as used in manufacturing sodium sulphur cells were placed in a stainless steel vessel which was evacuated. Ammonia gas was introduced into the vessel from a gas bottle. The vessel was then sealed and passed into a glove box under dry argon where the sample electrolyte cups were extracted without any atmospheric contamination.
Within the glove box, the electrolyte cups were placed on a heating plate and sodium at 150 C was introduced into the cups. As the cups and sodium were heated, at approximately 210°C a transparent film of sodium amide became visible wetting the entire surface of the electrolyte sample. At this stage, the sodium within the cup formed a visually flat 0° contact angle with the electrolyte wall. Although this indicated wetting of the surface of the electrolyte by sodium, movement of the sample did not result in any sodium adhering to the electrolyte surface. Thus, it appears there is a contiguous coating of the liquid sodium amide between the electrolyte and the sodium.
Figure 1 shows the measured resistance of a sample obtained by partially immersing the outer surface of the cup in sodium and contacting the two sodium bodies to a source of electric current. The total currents indicated in the Figure correspond to current densities through the beta alumina electrolyte of between 30 and 133
—2 mA cm . As can be seen, the resistance of the electrolyte cup varies between 22 and 28 m.ohms during the test in which the current was reversed every fifteen minutes. The theoretical resistance of the electrolyte cups used is 25 m.ohms ± 3m.ohms, however the exact
sodium contact area was difficult to establish in this test. It was apparent, however, that the sodium amide layer produced did not increase the anode resistance to any significant extent.
In practical sodium sulphur cells it is normally required to provide some wicking means on the sodium side of the electrolyte to ensure that the maximum surface area of the electrolyte is exposed to sodium at all stages of charge and discharge of the cell. In the next example of the invention, a cell sub-assembly comprising the electrolyte cup fitted with wicking means was similarly exposed to ammonia in the evacuated steel vessel before being transferred to the glove compartment for sodium/sodium cell tests in dry argon. A control sample of sub-assembly which had not been exposed to ammonia was also tested. Both the control sample and that which had been exposed to ammonia were from the same batch of sub-assemblies so that there intrinsic properties should have been similar.
The resistance characteristics of the control and ammonia exposed samples were then measured and the results are shown in Figure 2. lThese results indicate that the ammonia treated sample have substantially lower resistance at lower temperatures, presumably arising from improved "wetting" of the electrolyte by the sodium. It should be understood that in the case of the ammonia treated samples, there is in fact no direct wetting contact between the sodium and the electrolyte in view of the intervening coating of molten sodium amide.
In summary, therefore, the tests have established that treatment of electrolyte with ammonia gas prior to exposure to sodium produces a thin coating or film of sodium amide over the electrolyte (at temperatures above
the melting point of sodium amide, 208 C) , and has no apparent deleterious effect on the electrical performance of the cell. In fact, improved wetting characteristics are additionally noted which indicate that the treatment may enable the cells to be operated at lower temperatures.
In a further example of the present invention, a series of twelve sodium/sodium cells were tested. Of these cells, six were exposed to ammonia gas to produce a sodium amide coating, and six cells were untreated and used as control samples. The cells were all cycled at 4amps for a period 0.5h, which is equivalent to lAh cyclic discharge in each direction. Figure 3 illustrates the measured cell resistance for the first five cycles of the untreated cells, and shows some resistance variability, with some two to three cycles being required before the cells attain their equilibrium resistance. The five cells which had been treated with ammonia gas were then electrically tested in an identical manner and the results are shown in Figure 4. Not only do the treated cells show equilibrium resistance from the beginning of the first cycle, but resistance variation between the different cells is smaller.
In a further example, sodium-sulphur cells were constructed in which the sodium electrodes had been treated with ammonia gas during manufacture. These cells were repeatedly electrically cycled and the cells quickly attained their expected operating resistance at the usual operating temperature. The cells were then cooled down and the resistance characteristics plotted against temperature as shown in Figure 5. A corresponding plot for an untreated cell is shown in Figure 6. This shows at least that the operating resistance of a treated cell is no worse than that of an untreated cell.
In still further examples of the invention, instead of exposing the electrolyte and/or sodium to ammonia gas, the cells were treated by the addition of sodium amide (NaNH_) to the sodium. Sodium amide was added in some samples to provide O.lg of sodium amide per lOOcc sodium, and in other samples to provide 5g per lOOcc. The cells were then tested to monitor the cell resistance during initial warm up after manufacture, using a 50mA current which commenced after the sodium melting point of 98°C. The resistance characteristics of five test cells with this sodium amide addition to the sodium are shown in Figure 7, and it can be seen that all the cells have very similar characteristics whether "doped" with O.lg or 5g sodium amide per lOOcc.
Comparing these resistance values against the plot for an untreated but well cycled cell shown in Figure 6, it can be seen that the treated cell has the expected resistance characteristics without the need for any charge cycling to achieve these. Thus these results demonstrate that the addition of sodium amide assures good wetting of the beta alumina on initial war up of the treated cells, without the need to provide a long temperature soak at 350°C or for cycling at these temperatures which is normally required in prior art cells to promote wetting. Subsequent further resistance plots on cooling down the cells and reheating them are shown in Figures 8 and 9 and indicate that the electrolyte does not de-wet on cool down.
All the above examples describe providing a sodium amide coating or addition either by directly doping the sodium with sodium amide, or exposing the electrolyte or the sodium to ammonia gas. Other ammonium compounds may also be used which will react with sodium to produce sodium amide. An important criteria is that the by-product of this reaction must be benign in the
environment of the cell.
Ammonium chloride has been used successfully. This material sublimes at 340 C and sample cells were treated by exposing them to the vapour produced during sublimation. Subsequent filling with sodium again showed that wetting of the electrolyte surface was promoted.
Apart from ammonium chloride, which produces sodium chloride as a by-product, other materials may be considered including ammonium azide (NH.N. . Further organic derivatives of ammonia may be considered including amines.