Contact resistance in organic transistors: Use it or remove it

Abstract

The contact resistance of organic field-effect transistors is revisited to address its fundamental origin, parametric interplays, and technological implications. In a time when flexible electronics powered by an organic circuit comes close to the market, the revelation of wide-spread carrier mobility overestimation has astonished the broad scientific community, as this may contradict some of the most significant developments made to date. Since the contact resistance was pointed out as the major reason behind the issue, the research into reducing or eliminating this resistance has become more intense and justified than ever. However, there have been other revelations that suggest the benefits of contact resistance in certain structures and applications. Therefore, it seems timely to fairly judge the true meaning and consequences of the contact resistance, and to provide a comprehensive view covering both its positive and negative aspects, which constitutes the main motivation of this article. To maximize the depth of discussion, several important backgrounds for contact effects will be recapitulated before tackling selected practical problems of contact resistance, and before clarifying when it should actually be minimized and when it could otherwise serve as a useful element.

Identifying Key Transport Mechanisms in Organic Antiambipolar Transistors

“…Therefore, the widely observed non-linearity of a square-root I D versus V G plot can be properly modelled by adjusting the γ, with γ = 2 reserved for the ideal (trap-free) case. [18,19] Both of our devices were fitted to Equation (1) yielding a good overall agreement between theory and experiment (Figure 1b). In addition to the monotonous decrease in I D , the non-ideality of the transfer curve became more and more pronounced upon cooling both OFETs (Figure S1, Supporting Information), implying a substantial increase in γ at low T. Figure 1c shows the extracted γ as a function of 1/T.…”

Bulk versus Contact Doping in Organic Semiconductors

“…The excellent agreement between the gate voltage ( V G ) versus drain current ( I D ) characteristics (or transfer characteristics) at a drain voltage ( V D ) of −12 V from the measurement and numerical simulation in Figure 1 a broadly supports the validity of the model and parameters used. We specifically took into account several unique properties of OFETs, including the effect of a charge injection barrier ( E b ) at the metal–semiconductor contact ( Figure 1 b) [ 30 , 31 , 32 ] and a trap- or disorder-limited transport mediated by an exponential density of states (DOS) ( Figure 1 c) [ 33 , 34 , 35 ]. The optimized parameters are listed in Table 1 .…”

“…Experimentally, bulk doping can be obtained by exposing a pristine organic semiconductor film to doping agents (usually in liquid or gas phase) or by forming a co-solution of a host-dopant mixture that is directly processed (e.g., by spin coating) to become a doped composite semiconductor layer [ 36 , 37 , 38 ]. Contact doping has also been widely adopted, as it is fundamentally associated with a generally high contact resistance of OFETs [ 32 ]. This type of doping can be experimentally achieved, for instance, by shadow masking upon thermal evaporation of dopant molecules, or by direct co-patterning of the blanket-deposited dopant and metal contact layers [ 39 , 40 , 41 , 42 ].…”

“…It was actually quite surprising that even the strongest p -type doping at a full penetration depth only had a relatively small effect. At this point, we hypothesized that if the entire OFET becomes more severely contact limited [ 32 ], contact doping could manifest more benefits. Therefore, we prepared another undoped base structure with E b = 0.5 eV.…”

“…According to our simulation, channel doping can provide a promising balance between these two techniques. It was shown to be effective in substantially increasing I D , while the V T shift remained in a realistic range owing to the contact depletion (therefore, the contact resistance provides a positive effect in this case [ 32 ]). Our simulation results were fully supported by an in-depth correlation between terminal characteristics and internal distribution, as well as extensive theoretical rationalization based on OFET device physics.…”