290 THE BELL SYSTEM TECHNICAL JOURNAL, APRIL 1951 



5,3 Thermal Conductivity 



A comparison method was employed to determine the thermal con- 

 ductivity of the carbon films^^ The conductivity of a carbon film deposited 

 on a flat silica plate was compared to that of an identical silica plate to the 

 surface of which a thin foil of lead or other metal was fastened with glycerine. 

 One end of each of these plates was securely clamped to a heavy copper 

 base and to the opposite ends were clamped identical copper blocks supphed 

 with heater windings. Differential thermocouples permitted determination 

 of temperature differences between the two heated blocks and of the temp- 

 erature drops along the specimens. The temperature drop along the speci- 

 mens which were about 3 cm in length, never exceeded 12 deg C, and the 

 entire apparatus was contained in a heavy copper cyHnder immersed in a 

 constant-temperature oil bath. Calibration of the apparatus with foils of 

 different metals and of various thicknesses showed that the relative thermal 

 conductivities of the two specimens were accurately proportional to the 

 powers dissipated in the two heaters when the temperatures and temperature 

 drops were the same. 



By comparison with pure lead, iron, copper, nickel, and aluminum, the 

 thermal conductivity of carbon films about 1 X 10~^ cm thick was found 

 to be 0.08 watt cm-^ deg C~^ This value is in good agreement with the con- 

 ductivity of black carbon determined by other methods^^, and it was in- 

 dependent, within the limits of accuracy of the method, of the conditions 

 under which the carbon was deposited. 



Specimens cut from samples of crystal graphite with their base planes 

 parallel to their lengths were found by the same method to have thermal 

 conductivities greater than that of pure copper, 4.0 watt cm-^ deg C~^ 

 with a temperature coefficient of about — 0.0054 deg C~^ According to one 

 series of measurements, the conductivity was greater than that of pure 

 silver, this abnormally high conductivity of graphite along the base plane 

 being in agreement with other measurements^^ 



Specimens suitable for the determination of thermal conductivity along 

 the c-axis of graphite by this method could not be procured. However, 

 specimens of like size cut from crystal graphite with their axes along the 

 c-axis and the base plane, respectively, and oxidized to destroy any ori- 

 entation produced at their edges by cutting, were clamped to the surface 

 of a heated copper block with a small crystal of orthonitrophenol placed 

 on the upper surface of each. The temperature of the block required to melt 

 these crystals was noted and in this way it was found that the necessary 

 temperature gradient along the c-axis was at least five times as great as 

 that along the base plane, thus providing an approximate value of 0.8 watt 

 cm~* deg C"^ for the thermal conductivity of graphite along the c-axis. The 



