Field-effect transistors are at the heart of every digital electronic device, from laptops to digital alarm clocks. Typically, the active material in these transistors is an inorganic semiconductor, such as silicon; however, there is a strong drive to replace these heavy, expensive materials with cheap, lightweight organic semiconductors. During operation, a voltage applied to the gate electrode of an organic field-effect transistor (oFET) (see Figure 1) induces the accumulation of charge carriers (electrons or holes) in a thin volume at the interface of the organic and dielectric materials, called the "accumulation layer". At some threshold charge density, electrical current flows between the source and drain electrodes (i.e. the transistor is switched "on"). The mobility of accumulated charge carriers is defined by the molecular structures in this thin interfacial slab of the semiconductor.
Vibrational sum frequency generation (VSFG) spectroscopy is an invaluable tool for the characterization of interfaces.4 Briefly, beams of visible and IR light pulses are focused at a sample interface and a wave-mixed pulse is generated at the sum of their frequencies (wSFG = wIR + wvis) (see Figure 1). The output signal is generated almost exclusively by the material interfaces. This signal is significantly enhanced when the IR light frequency is resonant with a molecular vibrational at that interface, thereby enabling it to extract the IR absorption spectrum of only those interfacial molecules producing the necessary break in symmetry. Furthermore, by collecting the VSFG spectra with several polarization combinations for the visible and IR pump beams, the average orientations of the resonant transition moments can be determined. Preliminary VSFG measurements have recently been reported on oFETs, demonstrating the feasibility of applying this technique to this system. In the following pages, three experiments are proposed using polarization selective (PS-) VSFG to characterize the static and time-dependent interfacial molecular structures in oFETs prior to and during device operation.
A. Molecular Structure in oFETs Switched "Off"
The first PS-VSFG experiments that we propose on oFETs will study the accumulation layer structure in the absence of an applied gate or drain voltage. The first target molecule will be N-N'-dioctyl-3,4,9,10-perylene tetracarboxylic diimide (PTCDI-c8). We will use the orientation of the CH2 and CH3 stretching modes to determine the configurations of the alkyl chains on evaporated PTCDI-c8 films. Gauche defects in the alkyl chains will result in increased CH2 absorption bands in the VSFG spectrum. We will also measure the orientations of the C=O vibrations on PTCDI to determine the orientation of the aromatic system of the molecule. We have measured preliminary VSFG spectra from PTCDI-c8 on a Si/SiO2 gate electrode surface in the alkyl (not shown) and C=O spectral regions (Figure 2). One might assume that a vibrational mode such as the symmetric C=O stretch (1696 cm-1) on the aromatic core of PTCDI-c8 would not exhibit VSFG signal due to molecular symmetry, however these modes are indeed VSFG-active; the functional groups at the interface are not equivalent to those away from the interface. Furthermore, we can also observe the asymmetric C=O (1670 cm-1) and the aromatic C=C ring stretching bands in this region (1583 cm-1). These data were collected with a single polarization combination for the visible and IR pump beams (ppp). This precludes detailed orientational calculations of these transition moments, but it is clear that there are several vibrational modes to monitor on the body and periphery of PTCDI-c8 that can be independently used to track structure and orientation. We will use this approach to understand the influence of processing conditions (substrate temperature, functionalization, and post-deposition annealing) on the interfacial structure
B. Molecular Structure in oFETs Switched "On"
Once the orientations of the molecules at the organic/dielectric interface have been established over a range of processing conditions, we will apply a voltage to the gate electrode to accumulate charges at the organic/dielectric interface. Our experimental design will allow us to monitor changes in the IR spectra and average orientation of the interfacial molecules as the accumulation layer is charged. Presumably, if a spectral shift characterizes the increase in charge density in the accumulation layer, we would expect this shift trend to continue as conduction turns on; however, if surface-localized traps states become depleted (or filled) due to conduction, the spectral shifting could slow, stop, or reverse its trend due to mobilization of carriers in the accumulation layer.
Preliminary data from our lab for an experiment of this type are shown in Figure 2, in which VSFG from the C=O and C=C stretching region was measured at two gate voltages. A spectral shift of the C=O symmetric stretching frequency is observed in response to the 25 V gate bias, as well as a change in its intensity using the ppp polarization combination. If the symmetric stretching intensity is found to increase in the orthogonal polarization combinations at the expense of the ppp output, this would be indicative of a reorientation of the molecule under applied potential. It is also interesting to note that the asymmetric C=O stretch at 1670 cm-1 shows no change in band shape or position. The origin of this effect could be determined by performing these experiments with additional polarization combinations.
C. Time Resolving Molecular Changes During Switching
Once the initial (oFET "off") and final (oFET "on") molecular environments have been characterized using PS-VSFG, we will synchronize the application of the gate voltage with the arrival of the VSFG pulse pair at the sample to determine the time dependence of the structural changes. By varying the time delay between when the gate voltage is applied and when the VSFG signal pump pulse pair arrive at the sample, we will be able to follow a particular VSFG signature for configurational change in real time. The time resolution in this case will be limited by how quickly the gate voltage can be applied, setting nanosecond switching as the upper limit. The structural measurements (for all experiments in this proposal) will be compared to the electrical response time of the oFET to determine their relevance to the device performance. For example, it is not at all clear that the reorientation of alkyl functional groups should play any role in the conductance switching event in these devices. This experiment will provide a direct correlation between interfacial structure and device function. We ultimately plan to test many oFET molecular variants (beyond PTCDI-c8) to determine the structure-function relationship that causes some organic materials to have sluggish and irreproducible switching times. We will fit these kinetic data to functional forms and develop a model that describes the switching process from a microscopic perspective.
435 Amundson Hall, 421 Washington Ave. SE, Minneapolis, MN, 55455
P: 612-626-0713 | F: 612-626-7805