Trichlorosilanes as Anchoring Groups for Phenylene‐Thiophene Molecular Monolayer Field Effect Transistors

Self assembled monolayer field effect transistors (SAMFETs) are reported using a phenylene‐thiophene containing semiconducting mesogen attached through a trichlorosilane anchoring group. Monolayer films, covalently attached to silicon dioxide substrates, form in less than 10 h from solution, thanks to the accelerated reaction of the trichlorosilane anchor. Devices exhibit mobilities as high as 1.7 × 10−2 cm2 V−1 s−1, currents of up to 15 μA (on/off current ratio of 106) with device yields close to unity over large areas for channel lengths up to 100 μm.


Molecular Design Considerations
The choice of the semiconducting core of the SAM molecule is crucial to the realization of high-performance SAMFETs with a dense and ordered semiconducting monolayer. It has been recognized that conjugated mesogens can be self-organized into large area monodomains without defects caused by grain boundaries by processing in the mesophase. [ 18,19 ] Conjugated mesogens based on a phenylene-bithiophene core (PTTP) are promising semiconductors employed in OFETs due to the simple synthesis and favorable liquid crystalline phase behavior. [19][20][21] Further, the introduction of the phenylene units lowers the HOMO level leading to improved device stability in

Introduction
Self assembled monolayers (SAMs) are dense molecular layers deposited at an interface spontaneously from solution. They have become a critical technology in organic electronics for the deposition of functional nanoscale fi lms from the bottom up. SAMs have been extensively used to modify the surface of dielectric layers and metal electrodes but have recently shown great promise as single molecular layer dielectrics or semiconductors in organic fi eld effect transistors (OFETs). [1][2][3][4][5][6][7][8][9][10][11][12][13][14][15][16] Self assembled monolayer fi eld effect transistors (SAMFETs) have been demonstrated consisting of a single molecular layer of organic semiconductor, self assembled from solution. [8][9][10][11][12][13][14]16 ] The principal building blocks of these devices are molecules with a π-conjugated semiconducting core that is coupled to a surface anchoring group capable of covalently binding to a surface. These molecules create a self-limiting monolayer between source and drain contacts that functions as a two dimensional charge transport layer, comparable in thickness to the charge ambient, humid conditions relative to that of the analogous oligothiophenes. The sequence of liquid crystalline phases (I-SmA-G), in a PTTP mesogen, allowed melt processing of homeotropically aligned fi lms and strong intermolecular π-π interaction between adjacent molecules. [ 19 ] Hence, PTTP was chosen as the semiconducting core for the molecular monolayer, n-hexyl (C 6 ) chain as the end-capping group and a fl exible dodecane (C 12 ) spacer linked the semiconducting core to the trichlorosilane anchoring group.

Synthesis
Compounds 7a and 7b were prepared as depicted in Scheme 1 . In brief, 1 was lithiated with n -BuLi, followed by workup with 1,12-dibromododecane to afford 2 , which in turn underwent dehydrohalogenation upon treatment with t BuOK. Compound 4 was prepared according to the procedure of Hu et al. [ 22 ] Treatment with n -BuLi and tributyltin chloride afforded 5 . Stille cross coupling between 3 and 5 with Pd(PPh 3 ) 4 in DMF proceeded smoothly to yield compound 6 in excellent yield. Hydrosilylation of 6 with dimethylchlorosilane or trichlorosilane afforded compounds 7a and 7b respectively in near quantitative conversion under optimum conditions determined by Ponomarenko et al. (33 ppm Pt, 80 eq silane, 90 °C). [ 23 ] Compound 6 exhibited a narrow smectic mesophase between 214 °C and 218 °C indicated by a typical fan texture when viewed between crossed polarizers (see Supporting Information).

Fabrication of SAMs and Devices
Highly doped (n++) silicon substrates with 300 nm of thermally grown oxide were used as the gate electrode and dielectric respectively. To enhance the anchoring of the silanes to the silicon dioxide the substrates were rinsed and sonicated in acetone, propan-2-ol and methanol followed by activation in UV-Ozone. The substrates were transferred into a nitrogen glove box and placed in a sealed vessel containing a solution of compound 7a or 7b in chloroform at a concentration of 0.5 mg mL −1 (3 m M ) that had been fi ltered through a 0.2 μm PTFE syringe fi lter. The substrates were incubated in this solution for between 1 h and 7 days, transferred to a vial of clean solvent and sonicated for 30 minutes to remove physisorbed material and impurities, followed by drying under a stream of nitrogen. The substrates were annealed on a hot plate for 2 h at 120 °C to remove residual solvent. Transistors were completed by the evaporation of gold source and drain electrodes through a shadow mask to a nominal thickness of 50 nm.
The monochlorosilane ( 7a ) showed a very low reactivity at a silicon dioxide surface and it was not possible to form complete monolayers from this compound, in contrast to the previously reported quinquethiophene derivative. [ 12 ] The trichlorosilane derivative ( 7b ) was more reactive and working monolayer fi eld effect transistors were fabricated in less than 10 h over large areas (4 cm 2 ) for long channel lengths (up to 100 µm).

Atomic Force Microscopy
The growth and morphology of the SAMs formed by deposition of 7a and 7b were investigated by atomic force microscopy. The monochlorosilane derivative ( 7a ) exhibited a slow growth rate and incomplete monolayer formation was observed even after several days in solution. After 1 day ( Figure S14, Supporting Information) the islands are small (<100 nm) and cover approximately 25% of the surface. A line scan shows that the islands are approximately 3 nm in height, which is slightly lower than the predicted height of 7a (3.5 nm), suggesting that the molecules are tilted toward the substrate surface. The presence of larger islands, with a larger height profi le (4.5 nm) suggests that these molecules stand in a more upright position. After immersion for 3 days the coverage increases to approximately 35% with the formation of some larger islands and aggregates that could not be removed by sonication. It appears that the condensation reaction between the surface and the chlorosilane is so slow that the solution reaches equilibrium with the substrate. The interaction between the conjugated semiconducting core and the polar silicon dioxide inhibits the growth of full monolayer fi lms. Attempts to fabricate OFET devices led to insulating properties due to limited percolation paths for charge carriers between the source and drain electrodes. NMR spectroscopy of aliquots taken during the deposition process confi rmed that a large portion of the solution was still in the form of free monochlorosilane (identifi ed by the presence of the two methyl groups bonded to silicon, δ = 0.41 ppm) confi rming that the lack of full coverage was not due to consumption of 7a by self-condensation in solution.
The increased reactivity of the trichlorosilane derivative ( 7b ) was aided by the higher solubility of 7a over 7b , reducing the amount of solution aggregation possible during monolayer formation. Figure 1 shows the growth of 7b on silicon dioxide, after 1 hour the surface coverage is already greater than 60%, after 3 hours the coverage increases to 90% and complete coverage is observed after 9 hours. As with the monochlorosilane, compound 7b grows in the form of small islands (<100 nm) that are distributed homogeneously across the surface and ultimately lead to the formation of crystalline monolayers with little or no grain boundaries.
It has been seen that monochlorosilanes have different self assembly kinetics than the corresponding di-and tri-chloro derivatives. [ 27 ] In order to gain some insight into our system a simple, fi rst order, Langmuir kinetics model was used to estimate the rate constants for the two systems. [ 28 ] Fitting the model to the coverage measured from the AFM images showed that the rate constant for the trichlorosilane compound (0.23 s −1 ), is fi ve times greater than that of the monochlorosilane derivative (0.04 s −1 ) (see Supporting Information for details of the calculation).
Complete monolayers of 7b have a surface roughness of ≈2 nm. Working devices were successfully fabricated from these fi lms, even at large channel lengths (100 µm), confi rming the long range order of the molecular monolayer. The choice of trichlorosilane anchoring group allows for a crosslinking of the molecules within the fi lm, improving fi lm stability and potentially benefi tting the molecular packing and overall device performance. Previous work has shown that trichlorosilane SAMs generally lead to higher grafting densities and smaller tilt angles than mono-or di-chlorosilane derivatives, [ 29 ] In SAMFET devices the improved molecular packing of these SAMs should translate to higher device performance.

X-Ray Photon Electron Spectroscopy
X-ray photon electron spectroscopy was performed on both compounds 7b (trichlorosilane) and 7a (monochlorosilane) deposited on to silicon dioxide substrates (see Supporting Information for further details). All spectra confi rm the presence of carbon and sulphur. The calculated sulphur to carbon ratio for compound 7b is 2/38 = 0.053, the measured value from XPS of 0.051 is confi rmation of surface coverage by the semiconducting molecular monolayer. Films derived from compound 7a (monochlorosilane), which formed incomplete monolayers, showed the presence of chlorine in the XPS spectra, indicating unreacted species within the fi lm. In contrast fi lms deposited from the trichlorosilane derivative 7b contained no chlorine and it can be postulated that the monolayer has completely reacted with the surface and that a signifi cant portion of the fi lm has crosslinked.

Variable Angle, Attenuated Total Refl ectance -FTIR
The position of C-H stretching frequencies for the alkyl chains of a SAM observed in the IR spectrum gives information about the degree of crystallinity within a monolayer. [ 1,24 ] To investigate the degree of crystallinity of fi lms of compound 7b , SAMs were deposited on the native oxide of bare silicon substrates and investigated by variable angle ATR infra red spectroscopy (VATR-FTIR). Spectra were recorded at a resolution of 2 cm −1 and a total of 1024 scans were averaged. Figure 2 shows the C-H stretching region of the IR spectrum. The peaks are assigned as follows, symmetric and asymmetric CH 2 absorptions (ν s (CH 2 ), ν a (CH 2 )) are observed at 2851.6 cm −1 and 2919.2 cm −1 respectively, the symmetric CH 3 (ν s (CH 3 )) absorption appears at 2879.7 cm −1 and the in-plane CH 3 (ν a (CH 3 ,ip)) at 2964.7 cm −1 . The out-of-plane CH 3 and Fermi resonance splitting of the symmetric CH 3 were obscured by the other peaks. Table 1 compares the peak positions of the CH stretches observed for the SAMFET ( 7b ) fi lm with those assigned for crystalline and liquid alkanes and a crystalline octadecyltrichlorosilane (OTS) monolayer fi lm (Supporting Information Figure S20). [ 24 ] The SAMFET fi lms are consistent with a highly crystalline molecular monolayer demonstrating the potential of this molecule for large area deposition of well-ordered monolayers.

Transistor Characterization
Typical transfer and output characteristics for a device with full monolayer coverage (channel length of 90 µm, channel width of 2000 µm) are presented in Figure 3 . The saturation and linear mobility of the transistor was found to be 1.6 × 10 −2 cm 2 V −1 s −1 and 1.7 × 10 −2 cm 2 V −1 s −1 respectively, with an on/off current ratio of greater than 10 6 and an onset voltage close to zero. Functioning devices were observed for all channel lengths (20-100 µm) and a device yield of 100% was obtained (out of 36 devices tested), confi rming the large area applicability of the fabrication procedure. Good electrical charge injection from the contacts is confi rmed by the linearity in the output characteristics passing through the origin.
The dependence of the saturation and linear mobility on channel length (20-100 µm) are presented in Figure 4 for 9 devices fabricated on the same substrate. In general the linear mobility is slightly higher than the saturation mobility but both parameters show the same dependence on the channel length. Below 50 μm the mobility reduces as a consequence of the increased contribution from contact resistance at these shorter channel lengths. For long channel lengths (>40 μm) the mobility plateaus and stays the same up to the maximum channel length of 100 μm. The consistent mobility at channel lengths greater than 40 µm is confi rmation of long-range charge transport through a densely packed monolayer with minimal conduction barriers arising from grain boundaries or incomplete coverage. Similar behavior was reported by Mathijssen et al., who reported that the exponential decay of mobility with channel length seen in earlier attempts at creating SAM-FETs was a result of incomplete coverage. [ 30 ] Other extracted parameters, threshold voltage, subthreshold swing and hysteresis, remain constant for all channel lengths and are presented in Table 2 along with the saturation and linear mobility. The averages are taken over all devices with channel lengths greater than 40 µm (total of 24 devices).

Conclusion
To summarize, a novel self assembled monolayer fi eld effect transistor (SAMFET) has been produced, demonstrating an average saturation mobility of 1.5 × 10 −2 cm 2 V −1 s −1 . Large-area, high yield monolayer devices have been fabricated, in less than 10 hours on a silicon dioxide surface by utilizing a trichlorosilane anchoring group. These monolayer devices showed high mobilities, on the order of magnitude of the highest reported SAMFETs to date, confi rming the promising choice of this semiconducting core. The good performance demonstrates the long-range order, high coverage and good connectivity within the SAM fi lms. These thin fi lms have great promise for sensing applications as the ultrathin monolayer thickness could lead to improved sensitivity to analytes in these devices over the comparable bulk devices. Furthermore, the covalently bound monolayer fi lms offer stability against delamination in the presence of solvents.

Experimental Section
Further information on fi lm characterization, atomic force microscopy, X-ray photoelectron spectroscopy, IR and electrical measurements can be found in the Supporting Information. Monolayer Preparation : Solutions were prepared in an inert atmosphere to the desired concentration with the chosen solvent. The solutions were then heated to approximately 40 °C to facilitate the solvation of all of the material. They were then left to stand until reaching room temperature where they were then fi ltered through a 0.2 μm syringe fi lter into a sealed container with the freshly cleaned substrates. After the desired immersion time the substrates were transferred to sealed jars containing fresh solvent, removed from the inert atmosphere and sonicated for approximately 30 min. The clean substrates were fi nally dried on a hotplate at 120 °C for 2 h.
Atomic Force Microscopy (AFM) : AFM images were recorded on a Bruker Multimode 8 in Peak Force tapping mode at a resolution of 512 × 512 pixels. Cantilevers had a spring constant of approximately 0.350 Nm −1 with a resonant frequency of approximately 50-80 kHz. A modulation frequency of 2 kHz was used.
X-Ray Photoelectron Spectroscopy (XPS) : XPS measurements were performed using a Kratos Axis Nova with a monochromated aluminum K α X-ray source (1486.69 eV). Detailed spectra of individual peaks were taken at energy of 20 eV with a step size of 0.1 eV with a power of 225 W. The spectra were analyzed using CasaXPS.
Variable Angle Attenuated Total Refl ectance-Fourier Transform Infrared (VATR-FTIR) Spectroscopy : Monolayers were prepared on bare silicon substrates as this enhances the attenuation of the IR radiation. A Bruker Vertex 70 equipped with a Harrick Seagull ATR accessory and a germanium crystal is used in conjunction with a MCT detector. The sample is brought into contact with the crystal by applying a torque of 7.4 in.lbs to the sample mount. The angle of incidence of the source was 65 °. Both the background and the sample spectra were recorded at a resolution of 2 cm −1 for a total of 1024 scans.
Transistor Characterization : Transistors were tested in a top contact, bottom gate geometry. Thermally grown silicon dioxide (IDB Technologies) was used as the gate dielectric with a thickness of approximately 300 nm and a capacitance of 11.4 nFcm −1 . Transfer characteristics were recorded in the saturation and linear regimes with drain voltages of −60 V and −1 V respectively. Parameter extraction was performed using the standard fi eld effect transistor equations: An Agilent B1500a equipped with 3 SMUs was used in conjunction with a probe station. All transistors were tested in air at ambient temperature and pressure. Synthesis-General Considerations : Unless stated otherwise, all reagents and solvents were obtained from commercial suppliers and used without further purifi cation. Anhydrous THF was pre-dried over sodium wire prior to continuous distillation over sodium-benzophenone under an atmosphere of argon. Other anhydrous solvents were used as supplied (water content < 0.05%). 1 H and 13 C{ 1 H} NMR spectra were recorded on a Bruker 400 MHz or 500 MHz spectrometer. The progress of reactions was monitored by TLC on pre-coated silica or alumina plates visualized by ultraviolet light or staining with vanillin or potassium permanganate in the absence of a chromophore. Column chromatography employed Merck Kieselgel (60 Å) F 254 (230-400 mesh) silica or Aldrich neutral alumina and HPLC grade solvents. Distillation was carried out using standard Kugelrohr equipment with an oil diffusion pump. Purifi cation by recrystallization refers to the process of repeated recrystallization to a constant melting point. Melting points were recorded on a Linkam LTSE350 heating stage, quoted in degrees Celsius (°C) and are uncorrected.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.