Measurements are reported on the temperature dependence of the d.c. conductivity in the c-direction of kish graphite. These results show that both the magnitude and temperature dependence of the c-axis conduction are similar to results reported in single crystal graphite, but qualitatively different from observations made in pyrolvtic graphite. Because of differences in the impurity and defect density between kish graphite and single crystal graphite. the present results on kish graphite provide support for a band conduction model for c-axis conduction in graphite.
I. I~TRODUC~ON
The mechanism and magnitude of the c-axis electrical conductivity in graphite has been a source of recent controversy. Representative values for the room temperature c-axis conductivity in natural single crystal graphite samples [ 1,2] are in the range 150-230 (a cm) '. This can be compared with the electrical conductjvity in the basal planeIll (r, = 2.7 x 104 (fi cm))' at the same temperature to yield an anisotropy ratio of a,, iuV in the range I IO < uo/uc < 170. In Table 1 a summary is given of experimentally derived values for the c-axis electrical conductivity in natural single crystal graphite samples [ l-61. In highly ordered pyrolytic graphite, on the other hand, representative values for the room temperature c-axis conductivity are in the range 6 (a cm))' < a; < 10 (a cm)-', though the basal plane conductivity u" is of comparable magnitude to that in natural single crystal samples [7,8]. A summary of c-axis conductivity data[7-101 for pyrolytic graphite is given in Table 2. From these tables we see that the anisotropy ratio in the electrical conductivity at room temperatme is more than one order of magnitude greater in highly ordered pyrolytic graphite than in single crystal graphite.
As the temperature is lowered to 4"K, these tables show that this discrepancy is further increased by at least another order of magnitude. In natural single crystal graphite, both a, and crc. increase by about a factor of 15 between room temperature and 4"K, so that the anisotropy ratio is roughly temperature independent[l]. In contrast, cr, for highly ordered pyrolytic graphite increases by about a factor of 10 between 300 and 4Β°K while Us decreases by about a factor of 2. Therefore, the anisotropy ratio (a,/~,-) at 4Β°K is -300 times greater in highly ordered pyrolytic graphite samples than in natural single crystal graphite. This discrepancy must be consi-