Reinhard Knödler

Thermal properties of sodium-sulphur cells

The heat capacity and the rate of heat generation of Na/S cells during discharge and charge were determined. The measurements were carried out in a furnace with very low heat loss and low heat capacity (quasi-adiabatic arrangement). A linear relationship between (1/I)(dT/dt) andI, whereI is the discharge or charge current and dT/dt the temperature gradient, was obtained. From these plots the heat capacity of the cell and the entropy term could be determined. It turned out that, due to a steep entropy increase beyond about 80% state of discharge, the heat generation rate increased strongly in this region. During charging, this effect causes a cooling effect at low currents. The data presented here are important for the design of the thermal management system of an electric vehicle battery.

Kinetics of thermal cycling of sodium polysulphides in sodium/sulphur cells

In the sodium/sulphur battery the anode material is liquid sodium and the cathode material molten Na2Sx, soaked in graphite felt. The value of x decreases from about 20 to 3 during the discharge process. This battery is a promising candidate for electric-vehicle traction purposes [1]. It is necessary in this application to cool down the battery from its operating temperature of 350°C to room temperature from time to time. This sometimes caused cell failures owing to the action of the sodium polysulphide which serves as cathode material [2]. Therefore, investigations have to be carried out in order to understand and eventually overcome these difficulties. There are data available on the heat of fusion, the heat capacity and the undercooling behaviour [3] as well as data on thermal expansion coefficients of various polysulphides [4]. In the following, results are reported on the kinetics of the freezing and thawing process, as determined from variations of the cooling and heating rate. For the thermal-cycling experiments standard sodium/sulphur cells were used with a capacity of about 44A h. The cylindrically shaped cells have a diameter of 3.6 cm and a height of 23 cm. They were discharged until the composition Na2S 3 in the cathode was reached which could be detected by the e.m.f, of 1.78V at 350 ° C. There are about 130 g Na2

Heat loss measurements on an enclosure for high temperature batteries

3, corresponding to about 0.9 tool, in the cathode compartment. The cells were thermally cycled between 80 and 330 ° C with rates between 1.3 and 35 ° C h 1. The exothermic heat during freezing and the endothermic heat during thawing were determined by a specially designed calorimeter in which the cells were operated. Details of this equipment which allows the measurement of the thermal power with an accuracy of 0.1 W are given elsewhere [5]. In principle, the electric power of the furnace required to maintain a linear rate of ternperature change is recorded. If an exothermic reaction takes place, the power has to be decreased, while during an endothermic reaction the power has to be increased. Figs. 1 and 2 show the thermal power during thawing and freezing, respectively. For the thawing process, this behaviour is the same for every cycle at a given heating rate. However, freezing starts each cycle at a different temperature because undercooling is governed by random nucleation. Therefore, the temperature range in Fig. 2 can, within certain limits, vary from cycle to cycle. The peak power during thawing is directly proportional to the heating rate, whereas peak power values during freezing scatter considerably below the 10 ° C h -~ cooling rate, being nearly constant at faster rates. This behaviour, which is shown in Fig. 3 suggests that if high values of heat evolved during freezing are to be avoided, because of possible cell failure, low cooling rates will probably be favourable. The freezing curves themselves can in principle be explained by nucleation processes [6]. However, contrary to theory no nucleation will occur when the melt is held at constant temperature, e.g. 220 ° C, even after several days. Therefore, and because of the lack of more data, no attempt was made at this time to extract activation energies from the measurements. The thawing curves in Fig. 1 may be attributed to an activated process, as a plot of In P against I /T indicates, where P is the thermal power and T is the temperature. However, the melting of a crystallized material itself needs no activation. An explanation for the observed behaviour could possibly be the melting of different phases, shown by the phase diagram to be Na2S 4 and Na2S2 [7]. In addition other phases may be present [8]. This idea is supported by the very early beginning of the thawing process,

e.m.f. and impedance of Na-S cells and sodium poly sulphides during freeze-thaw cycles

Abstract Heat loss measurements are reported for an enclosure which is suitable for high temperature batteries operating between 300 and 470 °C. The enclosure is rectangular and has a load-bearing vacuum insulation. The heat loss measurements are used to determine the thermal conductivity of the insulation. At 450 and 300 °C, a thermal conductivity of 3.2 and 2.0 mW / (m K), respectively, was determined. These values are smaller by a factor of about 20 than can be obtained with conventional (non-evacuated) thermal insulations. This is an important step towards the realization of electric vehicles driven by high temperature batteries.

Individual rechargeable electric cell

The dependence of the e.m.f. and the resistance of Na-S cells on temperature was determined in order to understand the processes occuring during freezing and thawing of the positive electrode. It turned out that there is a strong tendency for the sodium polysulphides even in graphite felt to undercool and to crystallize only at a very slow rate in the frozen state. At the freezing point e.m.f. and impedance show discontinuities which could be interpreted using the phase diagram. The freeze-thaw behaviour of Na2S3 is different from that of the other compositions concerning its conductivity, diffusion properties and adhesion to the wall of the casing. These findings are useful for the analysis of failures due to thermal cycling.