Highly-ordered Triptycene Modifier Layer Based on Blade Coating for Ultraflexible Organic Transistors

We present a highly ordered surface modification layer for polymers based on ambient solution-processed triptycene (Trip) derivatives for high-mobility organic thin-film transistors (OTFTs). The nested packing of Trip molecules results in the formation of 2D hexagonal arrays, which stack one-dimensionally on the surface of polymer dielectrics without anchoring groups. The Trip surface was previously shown to be preferable for the growth of organic semiconductors (OSCs), and hence for enhancing the mobility of OTFTs. However, although the Trip modifier layer has been realized by thermal evaporation in a high-vacuum environment (TVE), it still has grain-boundary disorders that hinder the optimal growth of OSCs. To fabricate OTFTs with higher mobility, a disorder-free Trip layer is needed. We developed highly ordered Trip layers on polymer dielectrics via blade coating. In addition, we clarified that the highly ordered Trip modifier layer enhances the mobility of the OTFTs by more than 40%, relative to the disordered Trip layer prepared by TVE. Finally, we realized a ring oscillator composed of OTFTs with a highly ordered Trip layer.


Comparison of spin-coated and drop-casted TripOMe
We produced TripOMe using both the drop-casting and spin-coating processes. To fabricate the OTFTs, a 30-nm Au layer was thermally evaporated onto 1.0 µm parylene film supported on a glass film through a shadow mask to serve as the gate electrode. The substrate areas were 2 × 3 cm 2 . Next, before the deposition of parylene as the gate dielectric, the Au surface was treated with oxygen plasma at 100 W for 3 min. Then, approximately 30 nm of parylene was deposited by CVD. For the Trip formation based on the spin-coating and drop-casting processes, the TripOMe powder was dissolved in mesitylene (Wako chemicals) to give a 0.5 mM solution. For the spin coating, the first spin speed was 500 rpm which was maintained for 5 s. In the next step, the spin speed was increased to 3000 rpm and held for 20 s. In the dropcasting process, 100 µL of solution was cast to cover the entire parylene surface. After the Trip formation, all the substrates were annealed at 120 °C for 1 h under a vacuum at around 100 Pa. Finally, 30-nm DNTT and 50-nm Au layers were thermally deposited through a shadow mask to form the active layer and source and drain electrodes, respectively. The channel length and width of the OTFTs were 50 and 500 µm, respectively.
Supplementary Figure S1a-c shows the transfer characteristics of OTFTs with pristine parylene, spin-coated TripOMe and drop-casted TripOMe, respectively. The average mobility of the OTFTs with pristine parylene as a reference device were 0.46 ± 0.01 cm 2 V −1 s −1 ; whereas, OTFTs with spin-coated TripOMe did not exhibit any enhancement in their mobility, relative to those with pristine parylene, as shown in Supplementary Figure S1a The average mobility was 0.25 cm 2 /Vs in those OTFTs with the spin-coated TripOMe. We found that the spin-coated TripOMe did not exhibit increased capacitance relative to that with pristine parylene. This implies that spin coating cannot form a Trip film with enough thickness to improve OTFTs performance. In contrast, OTFTs with drop-cast TripOMe exhibited an enhanced level of mobility of 1.92 ± 0.47 cm 2 V −1 s −1 . However, the characteristics deviated considerably, as shown in Supplementary Figure S1c. This is because the quality of the drop-cast film depends on the random drying of the solution. This random drying prevents the formation of a uniform Trip film.

Mobility extraction and calculation of reliable factor (r)
In this study, we extracted mobility (μ) in saturation regime based on basic gradual channel approximation. In this approximation, the mobility can be expressed as: where L and W are the channel length and channel width, Ci is the gate channel capacitance per unit area, ID is the source-drain current, and VG is the voltage in the gate terminal. As described in the main text, Ci was estimated from the capacitance of the metal/dielectric/metal structures. The channel length and width are described in Fabrication and measurement of OTFTs. To extract the mobility and threshold voltage, we applied fitting curves to the √ID-VG transfer characteristics to determine the slope. With regard to the fitting range, some data points around small ranges were selected at the intermediated VG.
Based on the above-mentioned method, we estimated the reliability factor (r) for the extracted mobility of OTFTs with TVE-TripOMe and BC-TripOMe. In modified definition, r can be expressed as: [1] r = (2) The denominator expresses the square of the slope of the fitting curve to extract the mobility.
The numerator expresses the square of the slope of the transfer curve of the electrically equivalent TFT in ideal conditions. VGMAX is the maximum gate voltage to get the transfer characteristics. In this report, VGMAX is -2.5 V. |ID|VG =VGMAX  These values suggest the high reliability of the mobility extraction in the saturation regime. In the main manuscript, we showed the averaged r for 10 OTFTs, as used in Figure 2d.       Figure S10c. The FWHM on BC-TripH was found to be larger than that on BC-TripOMe. As discussed in the main text, this result indicates that crystallites of DNTT on BC-TripH lower crystallinity relative to BC-TripOMe. In other words, DNTT on BC-TripH exhibits more disorders than that on BC-TripOMe. In addition, an AFM image revealed that DNTT on BC-TripH was composed of smaller crystalline structures than either pristine parylene or BC-TripOMe as shown in Supplementary Figures S6 and S11. Thus, the smaller and more-disordered DNTT grain on TripH could be the main reason why OTFTs with TripH exhibit a lower mobility.
Second, to investigate the origin of the smaller and more-disordered DNTT grains on TripH, we observed the morphology of TripH films by using AFM. Supplementary   Figure S12a,b shows the morphology of TVE-and BC-TripH, respectively. The RMS were 5.09 and 1.94 nm, respectively. In the TVE process, many small terraces with a large step of around 10 nm were formed. On the other hand, the BC process dramatically improved the surface morphology of TripH. In addition, it should be emphasized that the surface of TripH has a lower surface energy than that of TripOMe (see Supplementary Table S2). However, the use of TripH led to smaller and more-disordered DNTT grains. As discussed in the main manuscript, the phase states (that is, the packing) of SAMs also influence the growth of organic semiconductors. These effects of TripH imply that TripH could present a greater degree of undesirable packing than TripOMe for the growth of organic semiconductors.
Surprisingly, the difference between TripH and TripOMe is only a methoxy group. This result is important not only for synthesizing new Trip derivatives but also for establishing a deeper understanding of the effect of the phase states in the self-assembled layers of organic devices.

Table S2
Surface energy of parylene with/without triptycene derivatives