| Abstrakt: |
The spectroscopic method of measuring temperature in spark discharges is discussed, bringing out the significance of the cooler zones which lie between the hottest part of the discharge and the surrounding air. A calculation of the relationship between line intensities and temperatures, with the conventional assumptions, leads to the result that most of the spectrum lines measured have their peak emission at between 5000 and 10,000°K. A more precise calculation for a selected example shows that the spectroscopically estimated temperature is independent of a doubling of the peak temperature; and peak temperatures cannot be correlated with experimentally observed temperatures. Consequently no such measurements can lead to reliable conclusions as to the highest temperatures in spark discharges; these can only be estimated using lines of highly ionised atoms whose transition probabilities are calculable. |
