A Comprehensive Physics-Based Compact Model for CNT Thin Film Transistors—Part II: Ohmic Contact

In continuation to the preceding article, this part presents a study of electrical transport in Ohmic contact carbon nanotube thin film transistors (CNT TFTs). Here, the transistor model is implemented as a special case of the previously presented generalized model. Based on comparisons between simulated and experimental data, it is observed that optical phonon scattering in Ohmic contact CNT TFTs is fairly independent of gate and drain biases as opposed to the Schottky contact counterparts. Moreover, it is observed that transport is significantly affected by the channel length and alignment between CNTs as optical phonon scattering is found to be more pronounced for smaller channel lengths and greater alignments. The success of the model is supported by good agreement between simulations and a wide range of experimental data. This is verified by calculating the relative root mean square error (RMSE) which shows an average deviation in the order of ~15%. In comparison to previous CNT TFT models, this work introduces a generalized model, considering both Schottky and Ohmic contact-induced ambipolar and unipolar modes of transport, closely following the essential transport physics.

I. Introduction

In recent times, high-performance carbon nanotube thin film transistors (CNT TFTs) comparable to the current state of the art have been reported [1], [2], [3], [4], [5], [6], [7]. These reports indicate toward the large potential of CNT TFTs in future electronics. It is observed that in almost all the presented cases, the device had a superior quality channel of uniform, a high-density network of straight, pristine CNT segments invariably contacted by metals Pd or Au, for source-drain contacts. These metals have been reported to form Ohmic contacts with CNTs [8], [9]. Therefore, in addition to the quality of the CNT film, the choice of contact metal is also expected to play an important role in delimiting the performance of the transistor. As mentioned in the preceding article, junctions formed between CNTs and metals are expected to be relatively free from the effects of metal-induced gap states (MIGSs), with energy bands closely following the Schottky-Mott rule; and there are several instances of both theoretical and experimental reports in support of this claim [10], [11], [12]. The contact of the device is dependent on the relative positioning of the metal Fermi level () with respect to the intrinsic Fermi level () of the CNT, and an Ohmic junction is expected when lies close to the CNT conduction/valence bands. Ideally, for metal and CNT work functions , CNT bandgap , and electron affinity , Ohmic contacts are expected for ; and , corresponding to the n- and p-type polarities, respectively. This is confirmed by previous studies where Ohmic contacts with p-type nature for high work function metals such as Pd, Au, and so on, () and n-type for low-work function metals such as rare earth metals Sc, Y, and Hf (), and so on, have been reported [8], [9], [10], [13], [14], [15]. From the above, the ideal Ohmic contact can be viewed to be equivalent to the Schottky contact without the barrier at the metal/CNT junction. Probably this is the reason why Ohmic contact CNT FETs have been reported to be relatively transparent, affording ballistic transport of carriers across them. However, such high-quality transport is not observed for all metals satisfying the metal-CNT work function relation, as high contact resistances have also been reported for metals such as Ni, Rh, and so on, suggesting that there may also be other factors delimiting the quality of the contacts [9], [16]. Nonetheless, the above picture indicates the importance of an accurate theoretical model of electronic transport in Ohmic contact CNT TFTs. As such, an analysis would probably enable to identification of the critical parameters affecting the device performance and provide insights toward attaining and exceeding performance limits.