Subsurface Laboratory Methods 247 



The reproducibility has also been observed with the standard Georgia 

 kaolinite endothermic and exothermic peaks. Since this is consistent with 

 data in the literature and since the change in peak temperature of pure 

 hydrous minerals may easily vary 5° C, more precise calibration has 

 been considered unnecessary. Different sample blocks, thermocouples, and 

 furnace windings produce no change in peak temperature greater than 

 5° C. 



Although the thermocouples are made as similar as possible and ad- 

 justed to approximately the same heights in the sample holes, the sensi- 

 tivity varies slightly. Therefore, after replacement of one specimen holder 

 and the corresponding thermocouples by another, the first run is usually 

 made with standard Georgia kaolinite in all sample holes. This indicates 

 the relative sensitivity of the various thermocouples. It has been found 

 that these relative sensitivities remain essentially constant for the life of 

 the thermocouples unless the height of the thermocouple is changed as a 

 result of rough handling. 



All curves included in this description are based on the same sensi- 

 tivity for direct comparison. It has been found convenient to plot the 

 differential-thermal curve so that an exothermic peak is upward, while an 

 endothermic reaction is represented by a deviation downward from the 

 base-line curve. 



Theory 



The theory of differential thermal analysis has been presented by 

 Speil.^^ The following account, modified and corrected,^^ is included to 

 aid in introducing the present studies. 



Figure 103 compares two methods of dehydrating a clay mineral. The 

 static method produces the equilibrium-dehydration curve, while the dy- 

 namic method gives the differential-thermal curve. In the first instance, 

 the sample is held at each successively higher temperature until it has 

 reached equilibrium. In the second, the sample is heated at a constant rate, 

 thus extending the dehydration over a longer temperature range. Since 

 the thermal curve is a differential function, it depends only on those ef- 

 fects that do not occur simultaneously and equally in the specimen and 

 the inert material. Hence, there are only two thermal effects to consider, 

 the differential flow of heat from the block to or from the thermocouple in 

 the center of the sample and the heat of the thermal reaction. The differ- 

 ential-thermal curve of figure 103 represents an endothermic reaction. Be- 

 low temperature a the heat inflow to both thermocouples, sample and 

 inert material, is the same, and no difference in temperature is recorded. 

 At a the reaction starts absorbing heat from its surroundings, making the 

 sample couple cooler than the alundum couple. This effect increases until 

 at b the rate of heat absorption by the chemical reaction equals the rate 



^' Speil, Sidney, Berkelkamer, L. H., Pask, J. A., and Davies, Ben, op. cit. 



*' Error in Spiel's derivation pointed out by Dr. D. G. Brubaker, N. J. Zinc Co., Palmerton, P». 



