Limits of Conductivity in ZnO Thin Films: Experiment and Theory

Transparent conductive oxides (TCOs) have major (multi-B) roles in applications such as flat-panel displays, solar cells, and architectural glass. The present workhorse TCO is indium-tin-oxide (ITO), but the recent huge demand for ITO has made In very expensive; moreover, it is toxic. The most commonly suggested replacement for ITO is ZnO, doped with Al, Ga, or In, and indeed the ISI lists 628 papers on Group-III-doped ZnO in 2009. However, to our knowledge, none of these papers has included calculations of donor ND and acceptor NA concentrations, the fundamental components of conductivity in semiconductors. We have developed a simple model for the calculation of ND and NA from temperature-dependent measurements of carrier concentration n, mobility μ, and film thickness d. With the inclusion of phonon scattering in the model, excellent fits of n and μ are obtained from 15 -- 300 K. Experimentally, we have shown that highly conductive ZnO films can be grown by pulsed laser deposition in a pure Ar ambient, rather than the usual O2 . In a 278- μ m-thick film, we have achieved a room-temperature resistivity ρ = 1.96 x10-4 Ω -cm, carrier concentration n = 1.14 x1021cm-3 , and mobility μ = 28.0 cm 2 V-s. From our model, we calculate N D = 1.60 x 10 21 andN A = 4.95 x 10 20 cm -3 however, the model also predicts that a significant reduction of NA would give μ = 42.5 cm 2 V-s and ρ = 7.01 x 10 -5 Ω -cm, a world record. Such a reduction in NA may be possible by in-diffusion of Zn after growth, since there is evidence that one of the major acceptor species in these films is the Zn-vacancy/GaZn complex. We can also decrease the resistivity by annealing in forming gas, and have recently attained ρ = 1.46 x 10 -4 Ω - cm , n = 1.01 x 10 21 cm -3 , and μ = 42.2 cm 2 V-s, giving ND = 1.13 x 10 21 and NA = 1.09 x 10 20 cm -3 . In very thin films, quantum effects must be considered.