Experimental investigations on metal-insulator transition and structure in amorphous carbon films

Abstract

Conductivity of carbon films prepared from TCPA at pyrolysis temperatures of 700°C, 800°C, and 900°C was measured by varying the temperature between 1.4 K to 300 K. To study the effect of substrates, films were deposited on quartz and alumina substrates. The concept of reduced activation energy helps in clearly distinguishing the metallic, critical, and insulating behavior in the samples. Analysis done using the various models developed for disordered electronic systems confirms this classification. Conductivity exhibits a transition from metallic to insulating behavior through the critical region when the pyrolysis temperature is varied from 900°C to 700°C. Films in the critical regime show metallic behavior at high temperature, and the onset of critical behavior is seen only at low temperature (< 5 K). A crossover from Mott to ES variable range hopping is seen in the insulating samples. Carbon films deposited on quartz substrate show a metal-to-insulator transition as a function of pyrolysis temperature. Carbon films deposited on alumina substrate show a saturation of conductivity (below 8 K), irrespective of the pyrolysis temperature. However, it is intriguing that even the samples in the insulating regime show this kind of saturation. Remarkably, as the measurement temperature is lowered, the films exhibit possibly a metal-to-critical/insulator transition. Above 10 K, the carbon films deposited on quartz and alumina substrates show almost identical transport properties at various pyrolysis temperatures. The effect of substrate on the behavior of dc conductivity is remarkable below ~10 K. A plausible reason for the observed variation in the conductivity behavior with substrates might be due to the thermal stress introduced by the thermal expansion mismatch between the substrate and the carbon film. The effect of magnetic field on the conductivity of the carbon films might probably shed more light on the issue as regards the observed saturation in conductivity, which shall be discussed in the next chapter.

In order to perform the stress corrections, the temperature-dependent resistance change due to thermally induced strain needs to be subtracted from the experimental temperature dependence of resistance. To obtain this, one needs the data for the coefficient of piezoresistivity as a function of temperature (G(T)) and also the coefficient of linear thermal expansion as a function of temperature (?(T)), in the x, y, and z directions. These data are not available for the material under study and so the corrections could not be done. However, the method to perform them is given as follows. If R(T) is the resistance of the sample as a function of temperature, and Ro is the resistance at an arbitrary reference temperature, the corrected resistance temperature dependence characteristics, Rc(T), is given by: Rc(T) = R(T) - ?Re(T) where ?Re(T) is the change in resistance due to temperature-induced strain. ?Re(T) / Ro = G(T) [?x + ?y + ?z] + Sx - Sy - Sz where ?x, ?y, ?z represent the thermally induced strain in the x, y, and z directions respectively, and Sx(T) = ?x(T) ?T.