Energy level consideration of source/channel/drain for performance enhancements of N- and P-channel organic FETs

Both n and p-channel operation of pentacene TFTs have been recently reported from the viewpoints of the channel substrate [1] and the appropriate source(S)/drain(D) material selection [2]. This paper discusses a possible way to achieve better FET performances for both channels as well as a determination mechanism of the channel type. We investigated perfluoropentace (C ₂₂F₁₄) (PF-pentacene) for n-channel and pentacene (C ₂₂H₁₄) for p-channel FETs. On the basis of the energy level consideration for both channel material and S/D metals, we show a systematic guideline for achieving a better OFET performance. First, the crystalline structures of PF-pentacene and pentacene are shown in Fig. 1. Although the XRD pattern of two materials is similar to each other, there is a big difference of the SEM picture. This fact may be related to the effective mobility in the channel due to a difference of the grain-grain interaction between two materials. But, this paper focuses another aspect determining the OFET performance. The energy level structures of PF-pentacene and pentacene were determined by using the photoemission apparatus, AC-3 (Rikenkeiki), and the spectroscopic ellipsometry (SE). 50nm thick PF-pentacene and pentacene films were grown by the vacuum evaporation on 30 nm-thick SiO₂ thermally grown on p- and n-silicon substrates, respectively. From the AC-3 results, HOMO levels of PF-pentacene and pentacene were estimated to be 6.7eV and S.0eV, as shown in Fig.2. On the other hand, LUMO-HOMO energy gaps of them were estimated to be 1.7eV and 1.8eV from the absorption edge of SE as shown in Fig. 3. Thus, LUMO levels were calculated to be S.0eV for PF-pentacene and 3.2eV for pentacene, respectively. The energy level diagrams of them are depicted in Fig. 4, together with the work-function (WF) levels of Au and Al. It is noted that the energy barrier is very small between the LUMO level of PF-pentacene and the WF of Au, while the HOMO level of pentacene is quite similar to the WF of Au. Those results guide us how to achieve higher FET performance in terms of both substrate material and S/D material selections. Fig. 5 shows Ids-Yds characteristics for both n- and p-channel FETs by changing the substrate and S/D materials. FET size was W/L=1000μm/300μm and the measurement was conducted in a vacuum. The mobility of pentacene and PF-pentacene FET with Au source drain electrode was 0.2 cm²/Vs and 0.02 cm²/Vs, respectively. Note that Al(S)/PF-pentacene/Au(D) structure is better than Au(S)/PF-pentacene/Au(D) for n-channel FET, while Au/pentacene/Au one is better than any Al electrode FETs for p-channel one. Since Al source for PF-pentacene and Au source for pentacene have no carrier injection barrier (a very small if any) expected from the results in Fig. 4, while Al drain for PF-pentacene and Al source for pentacene lead to high barriers. Yasuda et al. reported an n-channel pentacene FET operation with calcium S/D electrodes on the basis of the energy level consideration [2]. However, electron mobility in pentacene was much smaller than our PF-pentacene FET mobility. Thus, it can be concluded that appropriate combination of the channel materials with S and D metals will be a key to achieve higher performance dual channel FETs. Furthermore, it is interesting to note that a finite offset of Vds in Fig. 5 enables to realize a switched diode, which means a diode on-state and no current off-state, in addition to the simple FET characteristics. This function may lead to a new circuit operation mode of OFETs without any dopant control employed in the conventional semiconductor technology. Finally, it is worthy of mention that it will be interesting to apply these structures to the bilayer ambipolar devices. The bilayer FET of pentacene/PF-pentacene/SoO₂/Si structure with Au S/D electrodes clearly shows ambipolar Ids-Vds characteristics as shown in Fig. 6. In case of bilayer device with different S/D electrodes, it is expected to show novel functional characteristics. In conclusion, we investigated the channel material selection associated with the S/D electrodes on the basis of the energy level alignment of the source/channel/drain structure. We also discussed a switched diode using a simple OFET in terms of a functional device. These results are useful for designing high performance organic device family.