Using Molecular Design to Increase Hole Transport: Backbone Fluorination in the Benchmark Material Poly(2,5‐bis(3‐alkylthiophen‐2‐yl)thieno[3,2‐b]‐thiophene (pBTTT)

The synthesis of a novel 3,3′‐difluoro‐4,4′‐dihexadecyl‐2,2′‐bithiophene monomer and its copolymerization with thieno[3,2‐b]thiophene to afford the fluorinated analogue of the well‐known poly(2,5‐bis(3‐alkylthiophen‐2‐yl)thieno[3,2‐b]‐thiophene) (PBTTT) polymer is reported. Fluorination is found to have a significant influence on the physical properties of the polymer, enhancing aggregation in solution and increasing melting point by over 100 °C compared to nonfluorinated polymer. On the basis of DFT calculations these observations are attributed to inter and intramolecular S…F interactions. As a consequence, the fluorinated polymer PFBTTT exhibits a fourfold increase in charge carrier mobility compared to the nonfluorinated polymer and excellent ambient stability for a nonencapsulated transistor device.


Introduction
Polythiophenes and thiophene containing derivatives have been the subject of much research in the fi eld of organic semiconductors, owing to their relative stability, ease of synthesis, tuneable energy levels, and propensity for self-assembly. The popularity of poly(3-hexylthiophene) ( P3HT ) as an active material in both organic-fi eld-effect transistors (OFETs) and organic photovoltaics has led to many derivatives being investigated over the past decade. [1][2][3] In many semicrystalline polymers such as polythiophene derivatives, increased aggregation strongly correlates with increased transistor performance, providing the polymer aggregates and crystallites are appropriately interconnected with so-called 'tie-molecules'. [4][5][6][7] Processing techniques such as aromatics and can be overcome by chemically fusing the units together. [ 12,18 ] Another approach is to promote nonbonding interactions between adjacent aromatic monomers to assist backbone planarization. [19][20][21][22] We were particularly interested to investigate the effect of thiophene fl uorination in PBTTT because nonbonding S … F interactions have previously been shown to have a planarization effect in conjugated polymers. [23][24][25][26][27] In addition, fl uorination could increase the IP of PBTTT as it does for F-P3ATs , and hence the oxidative stability of a polymer often considered as on the cusp of air stability. [ 13,24,25,[28][29][30] In considering PBTTT , there are two possible sites of fl uorination: on the fused thieno [3,2b ]thiophene or on the alkylated bithiophene co-monomer. We chose to explore the infl uence of fl uorination on the bithiophene monomer because in this arrangement the fl uorine substituents are 'head-to-head' with respect to each other which should maximize any possible S … F interactions.
In this manuscript, we present a comparative study of hexadecyl derivatives of PBTTT and poly(2,5-bis(4-fl uoro-3-hexadecyl-thiophen-2-yl)thieno [3,2- We explore the effect of fl uorination on backbone planarity though density functional theory (DFT), present the synthesis and characterization of these polymers, and probe the infl uence of fl uorination on OFET performance and air stability, as well as thin fi lm morphology through X-ray diffraction analysis and atomic force microscopy (AFM).

DFT Calculations
DFT calculations are a useful tool to predict and assess the potential planarity of conjugated polymers. [ 19 ] Trimeric units (for A-B copolymer systems) are often used as analogues to the polymer as they strike an appropriate balance between predicting the basic properties of interest, while allowing these calculations to be completed within reasonable computational time. For this reason, we opted to run the geometry optimization calculations on units where the hexadecyl side chains were replaced with propyl groups in order to emulate the steric bulk near the polymer backbone while keeping computational time low. The minimum energy conformations of the monomeric species were optimized using the B3LYP functional and a 6.31g(d) basis-set, and from these optimized geometries the minimum energy conformations of the corresponding trimers ( BTTT and FBTTT ) were calculated using the same level of theory.
The optimized ground-state geometries of the fl uorinated and nonfl uorinated trimers are shown in Figure 1 a. The bithiophene link exhibits a dihedral angle ( θ 1 ) of 12.7° in the nonfl uorinated trimer, but upon fl uorination this angle becomes 0.1°, indicating that the two fl uorinated thiophenes are essentially coplanar. We believe this planarization is Coulombic in nature, as is the case in the aforementioned F-P3AT systems. [ 24 ] Mulliken charges show that both the hydrogen atom in the 4-position of the thiophene unit and the proximate sulfur atom in the neighboring thiophene ring have a slight positive charge in the BTTT trimer, tending to electrostatic repulsion and therefore a nonnegligible torsional angle. In contrast, in FBTTT the 4-position is occupied by a fl uorine atom that exhibits a slight negative charge, leading to an attractive interaction, consequently reducing the torsional angle to near planarity. While we acknowledge that Mulliken charges must be treated with caution in conjugated systems, they nevertheless provide an indication of the sign of the charge on atoms, which in this case proves crucial to rationalizing the seemingly contradicting planarization incurred by substituting a hydrogen atom for a fl uorine atom that has a larger covalent radius. Also of interest is the very slight planarization of the thiophene-thienothiophene link ( θ 2 ), from 41.8° to 39.1° upon fl uorination of the bithiophene unit.
The planarizing effect of the fl uorine substitution is particularly evident when observing the change in potential energy as a function of θ 1 for the trimers, as shown in Figure 1 b. We note that the slight asymmetry in the potential energy scan is present regardless of the rotation direction calculated. Aside from the minimum energy conformation being closer to planarity in FBTTT , the barrier to rotation is also much larger in FBTTT than BTTT , and the syn conformation of the thiophene−thiophene link is strongly disfavored in the former case. This conformation places both the fl uorine atoms in close proximity, with the steric and electrostatic repulsion leading to the observed higher potential energy. This is less apparent in BTTT , presumably due to the smaller Van der Waals radius of hydrogen compared to fl uorine, and lower partial charge (+0.15 and −0.28 respectively). A Boltzmann analysis of the relative populations at room temperature ( Figure 1 c) clearly illustrates this. Indeed, while BTTT shows a non-negligible population of syn thiophenethiophene links ( θ 1 = ±180°), FBTTT exhibits a narrow distribution of conformations around the trans coplanar conformer ( θ 1 = 0°) and a very small proportion of the competing syn analogue. Together these results suggest that fl uorination will result in a signifi cant increase in the rigidity of the backbone and a preference for the trans coplanar conformation.
Visual representations of the HOMO and LUMO of BTTT and FBTTT trimers are shown in Figure S1 (Supporting Information). Both show extended delocalization of the two frontier molecular orbitals, typical of all-donor conjugated units. [ 17,44 ] The substitution of hydrogen atoms by fl uorine in the bithiophene unit seems to have only a minimal effect on the orbital distribution, in the form of a minor contribution to both the HOMO and the LUMO.

Synthesis
Building upon our recent synthesis of short chain 3-alkyl-4-fl uorothiophene derivatives, we decided to utilize the fl uorinated building block 1 and introduce the hexadecyl side chain via a cross-coupling methodology ( Scheme 1 ). [ 24 ] Since backbone fl uorination can result in a reduction in polymer solubility, the hexadecyl side chain was chosen as it is the longest solubilizing group which has been demonstrated to still afford good FET performance in PBTTT . [ 24,28,31 ] Alkylation of 1 was best achieved via Negishi coupling using commercially available hexadecylzinc bromide in the presence of catalytic Pd(dppf) Cl 2 . Superheating of the tetrahydrofuran (THF) solvent to 100 °C in a microwave reactor was required to afford good yields of 2 . This was purifi ed either by reverse-phase chromatography on C18-functionalized silica using a mixed acetonitrile/THF eluent, or the crude mixture was deprotected to give 3 , which was then purifi ed by normal-phase fl ash chromatography on silica. We opted for the latter for larger scale reactions, since unfunctionalized silica can support a higher loading capacity.
Tail-to-tail dimerization of 3-alkylthiophenes is usually easily achieved through lithiation followed by oxidative coupling using copper(II) chloride. [ 32 ] Though the initial lithiation of 3 was achieved, as confi rmed by quenching with a solution of iodine in dry THF, the oxidative coupling step proved problematic. Evidently, the fl uorine substitution at the 4-position plays a role in hindering the oxidative coupling with copper(II) chloride. We therefore decided to regioselectively metallate 3 using the sterically hindered Knochel-Hauser base before adding 0.5 equivalents of Ni(dppp)Cl 2 . Reductive elimination from the resulting Ni(II) salt afforded the desired tail-to-tail bithiophene 4 in good yield (73%). Subsequent bromination led to monomer 5 , which was copolymerized with 2,5-bis(trimethylstannyl) thieno [3,2b ]thiophene under microwave polymerization conditions to yield PFBTTT . [ 33 ] For comparison purposes, PBTTT was synthesized under identical microwave polymerization conditions. Both polymers were purifi ed by precipitation and subsequent Soxhlet extraction with methanol, acetone, and hexane to remove catalyst impurities and low molecular weight oligomers. In the case of PFBTTT the crude polymer was also extracted with chloroform to remove low molecular weight poly meric material, before extraction into chlorobenzene. In both cases, a fi nal purifi cation by precipitation of chlorobenzene solutions into methanol was performed.

Physical Properties
In order to ensure a fair comparison and exclude molecular weight effects, which are known to affect the performance of PBTTT , [34][35][36] we compared batches of similar molecular weight distribution PBTTT and PFBTTT ( M n 42 kDa, Ð 1.5 and M n 44 kDa, Ð 2.0, respectively, as measured by high-temperature gel-permeation chromatography (HT GPC) against polystyrene standards). The basic physical properties of PFBTTT all tend to suggest a greater degree of aggregation when compared to its nonfl uorinated analogue PBTTT . First, PFBTTT is only soluble in near-boiling chlorobenzene, rendering processing from this traditional solvent diffi cult. We therefore adopted 1,2,4-trichlorobenzene (TCB) as a solvent for most processing techniques due to its good solubilizing properties and high boiling point. In TCB, PFBTTT is readily soluble at temperatures exceeding 135 °C, yet precipitates soon after cooling to 130 °C, while PBTTT remains soluble for several minutes at room temperature.
The solution UV-vis spectra of PBTTT , both hot and at room temperature, exhibit a single broad absorption band typical of fully solvated polythiophene derivatives ( Figure 2 and Table 1 ). [ 37 ] The slight blueshift of 12 nm of the absorption maximum upon increasing the solution temperature is attributed to the increased backbone torsion, which reduces the effective conjugation length. In contrast, the hot and room temperature solution spectra of PFBTTT are very different. The hot solution spectrum shows that the majority of the absorption arises from solvated polymer, similar to PBTTT . There is very little difference in λ max for the two polymers in hot solution. When the FBTTT solution is cooled, the absorption spectrum redshifts by 69 nm to give an absorption spectrum which strongly resembles that of the thin fi lm, suggesting the polymer has planarized and aggregated in solution, similar to the formation of polythiophene aggregates in poor solvents. [38][39][40][41] Even in the hot solution, the presence of a slight shoulder at higher wavelengths suggests some aggregate is still present. Upon thin fi lm formation, both polymers display a redshift in λ max of around 70 nm compared to the hot solutions as the polymers planarize in the solid state. In the case of PFBTTT , this is accompanied by the appearance of pronounced shoulders both at longer and shorter wavelengths. These shoulders are typical of the vibronic progression observed in many polythiophenes and ascribed to the formation of ordered aggregates in the fi lm. For PBTTT , the vibronic features are less obvious in the as-spun fi lm, but become sharper upon thermal annealing at 200 °C. Thermal annealing also results in changes to the PFBTTT spectra, with the longer wavelength shoulder increasing in intensity and the short wavelength diminishing. These changes are consistent with some structural reordering in the fi lms upon annealing. Both polymers exhibit very similar optical band gaps, as measured by the absorption edge of the as-spun fi lms ( Table 1 ). Thermal annealing results in fi lms with an identical band gap.
The effect of the fl uorination on the IP of thin fi lms was minimal, with no difference observed within the experimental error (±0.05 eV) of the measurement by photoelectron spectroscopy in air (PESA) ( Table 1 ). This is in contrast to the results for F-P3AT s, in which a clear (≈0.25-0.4 eV) increase in IP was observed in comparison to the analogous P3AT s, as well as a widening in the optical band gap. [ 24 ] The DFT calculations suggest that fl uorination would result in a modest stabilization of the HOMO of 0.16 eV and the LUMO of 0.18 eV over the nonfl uorinated trimer. The difference between the experimental and the theoretical results may therefore be related to error of the PESA measurement, since we note that the transistor results (vide infra) show higher contact resistance for the fl uorinated polymer over the nonfl uorinated polymer, which would support a slight increase in IP upon fl uorination.
The thermal properties of the two polymers were investigated by differential scanning calorimetry (DSC). Our initial experiments utilized conventional DSC at heating rates of 10°C min −1 ( Figure S13, Supporting Information). In these experiments, we were able to clearly observe the LC phase of PBTTT , with two well-defi ned endotherms on heating, at 126 and 238 °C, which have previously been attributed to a side chain and backbone melt. [ 42,43 ] However, despite heating to 390 °C, we could not observe a well-defi ned backbone melt for PFBTTT . We therefore moved to fl ash DSC in which heating rates of 500 K s −1 are obtainable. Such rapid heating and cooling rates increase the sensitivity of the measurement and also allow the investigation of temperatures regimes not readily accessible with conventional DSC due to competing decomposition processes.
The fl ash DSC thermograms of both polymers are shown in Figure 2 . The effect of the two fl uorine atoms on the thermal properties is remarkable, with an increase in the backbone melt of approximately 100°C, from 250°C for PBTTT to 350 °C for PFBTTT . Note that melting enthalpies are not readily accessible with fl ash DSC due to the small sample size. This increase is much more substantial than would be expected by a simple molecular weight argument (based upon the increased atomic mass of F over H), and is therefore supportive of the increased intra-and intermolecular interactions that would result from the more coplanar bithiophene unit suggested by DFT calculations. It is also interesting that the temperature of the low-temperature endotherm reduces upon fl uorination, from 66 to 55 °C. A similar reduction in side-chain melting point is observed for a PBTTT analogue in which the alkyl side chains are moved from the bithiophene unit, in which each thiophene can rotate independently, to a thieno[3,2-b]thiophene in which case each alkyl side chain must move co-operatively. [ 30 ] A similar co-operative movement of the side chains might be expected for the fl uorinated bithiophene if the S…F interaction was pronounced, as suggested by the DFT calculations. Finally, we note that thermogravimetric analysis of PFBTTT demonstrates that it exhibits excellent thermal stability, with 5% weight loss occurring only beyond 420 °C (Figure S14, Supporting Information).

Transistor Fabrication and Characterization
The electrical properties of the polymers were studied employing thin fi lm transistors. Films were prepared by spin-coating from TCB in both cases, followed by annealing at 200 °C for 30 min. We note that the transistor performance of PBTTT , to the best of our knowledge, has not been previously reported for fi lms cast from TCB. Bottom-gate top-contact confi guration devices for both polymers exhibit typical unipolar hole transporting behavior with low hysteresis between the forward and reverse gate voltage ( V G ) sweeps and moderate on/off channel current ratios of ≈10 3 -10 4 , as shown in Figure 3 . It is notable that both the peak and average charge carrier mobility values increase upon fl uorination by approximately a factor of 4, with the best devices exhibiting a saturated mobility of ≈0.32 cm 2 V −1 s −1 for the fl uorinated polymer, compared to ≈0.069 cm 2 V −1 s −1 for the nonfl uorinated analogue. The performance data of the two polymers is summarized in Measured in 1,2,4-trichlorobenzene (TCB) at 140 °C relative to polystyrene standards; b) Measured in room temperature TCB, or hot (>80 °C) TCB in parentheses; c) Films spin-coated from TCB; d) Optical band gap estimated from the absorption onset of as-spun fi lms; e) Ionization potential measured by photoelectron spectroscopy in air (PESA) error ± 0.05 eV; f) Lowest unoccupied molecular orbital estimated from the optical band gap and PESA measurements; g) Measured by fl ash DSC at heating and cooling rates of 500 K s −1 .
donor-acceptor polymers, these values are still high for an all-donor system. [ 17,44 ] While the mobilities reported here for PBTTT devices (max 0.069 cm 2 V −1 s −1 ) are lower than those often reported in the literature (0.1-1 cm 2 V −1 s −1 ), we attribute this mainly to the longer hexadecyl alkyl-chain length used in our case (as opposed to tetradecyl used previously in higher performing PBTTT -based transistors), as well as the different processing solvent (TCB compared to mixtures of chloroform and chlorobenzene). In fact, the few reports of hexadecyl-substituted PBTTT transistor properties involve either highly optimized and rigorous device fabrication process or a tailored 'spreadand-compress' fi lm formation on top of an ionic liquid. [ 29,45,46 ] The marginally higher threshold voltages ( V TH ) required for PFBTTT devices along with the slightly sigmoidal output curves and the increasing slope of the sd 1/2 I versus V g curve suggest increased contact resistance and/or localized high-energy trap states for PFBTTT over PBTTT . [ 47 ] Previously, the use of high work function electrodes like Pt has been shown to reduced contact resistance in PBTTT and may provide a possible solution to further improve device performance. [ 14 ] Previous reports have demonstrated that fl uorination of the conjugated side chain or backbone can result in improved device stability in the presence of ambient air, with the effect being ascribed to either an increase in the IP as a result of the electron withdrawing infl uence of fl uorine or to a kinetic effect reducing water ingress due to closer packing of the conjugated molecules in the solid state. [ 48,49 ] Therefore, we investigated the stability of the PBTTT and PFBTTT devices by removing them from the glove box and storing them in the dark under ambient conditions with an average temperature of 20 °C and relative humidity of 50%. The air-stability of PBTTT -based devices varies considerably in the literature. In the case of C 12 -and C 14 -PBTTT devices, stability in ambient conditions is known to be quite poor, with off current rising and charge mobility dropping to about 20% of the original value after 5 and 22 d, respectively. [ 13,30 ] Humidity was shown to have a signifi cant deleterious impact on the charge carrier mobility, with storage at low humidity shown to drastically improve the operational lifetime of these devices. [ 30 ] High humidity levels have been shown to result in the formation of charge traps in the fi lm in studies on related polythiophenes. [ 50 ] It is worth noting that the increase in off-current and shift in threshold voltage reported for many polythiophene derivatives may not be due solely to the effects of oxygen or water, but also to minor impurities in ambient air such as ozone, which can act as a reversible dopant Adv. Funct. Mater. 2015, 25, 7038-7048 www.afm-journal.de www.MaterialsViews.com  for the polymer. [ 51 ] To our knowledge, the only stability study performed on the C 16 -PBTTT was performed by Umeda et al. in which repeated stressing in ambient conditions over 2 d had little impact on the fi eld effect characteristics such as mobility and on/off ratio. [ 29 ] The transfer plots of PBTTT and PFBTTT after 3, 55, and 75 d storage in the dark in ambient air are shown in Figure 4 and the data are summarized in Table 3 . For both polymers the charge carrier mobility remained relatively constant over the test period, but we did observe changes in the threshold voltage and on/off channel current ratio over time. In particular, the off current rose rapidly after 3 d, with a large positive shift in the threshold voltage for both materials. Upon continued storage the threshold voltage moved back toward the original values, and the off currents dropped. The fl uorinated polymer consistently maintained a higher on/off ratio than PBTTT , with the device after 75 d exhibiting a value around 5 × 10 4 . The PFBTTT transistor also demonstrated a sharper turn-on with a narrower subthreshold swing than PBTTT . As discussed above, the shifts in threshold and increase in off-current are typical of oxidative doping making the device more conductive, and as such diffi cult to switch-off. The fl uctuations in threshold and off ratio over time suggest that this doping is reversible, and the lower currents for the fl uorinated polymer suggest it is less susceptible to such doping, in common with other fl uorinated materials. [ 22,52 ]

Thin Film Morphology
While it is widely accepted that PBTTT possesses a high degree of order in the lamellar direction, the π-stacking direction exhibits a relatively high degree of paracrystallinity. [ 53,54 ] The factors that allow PBTTT to have high hole mobilities are therefore considered to be its relatively low-energy trap states and high edge-on orientation. [53][54][55] In order to assess the impact of PBTTT fl uorination on its orientational order, we performed grazing incidence wide-angle X-ray scattering (GIWAXS) on fi lms prepared in the same way as the OFET devices ( Figure 5 ). Both fi lms exhibit diffraction patterns consistent with lamellar ordering of the polymers with an edge-on orientation with respect to the substrate. In the as cast fi lms, the crystalline domains of PFBTTT possesses a higher degree of edgeon orientation, which is apparent from the more localized out-of-plane scattering patterns corresponding to diffraction in the lamellar direction. Indeed, arcing of these diffraction peaks is attributed to misalignment of the crystallites with respect to the surface. [ 15 ] The lamellar spacing is similar for both polymers, at 2.33 nm for PFBTTT and 2.35 nm for PBTTT . This is in agreement with the d -spacing previously observed for C 16 -PBTTT and suggests that the alkyl side chains of adjacent polymer backbones are interdigitated for both polymers. [ 56 ] Annealing the fi lms results in an increase in the intensity of the diffraction peaks for both poly mers, most clearly observed in the out-of-plane line profi les in the Supporting Information ( Figure S15). The d -spacing does not change for PFBTTT , while it slightly narrows for PBTTT to 2.315 nm, suggesting either enhanced crystallinity or a change in crystal orientation in the fi lm. That the in-plane lamellar diffraction peaks gradually disappear for PFBTTT with annealing would suggest the latter, that the misaligned polymer domains reorientate to become predominately edge-on. For PBTTT , the out-of-plane peaks also increase in intensity and the arcing reduces, consistent with a similar increase in edge-on alignment. However the in-plane scan suggests that the misaligned domains remain upon annealing ( Figure S15, Supporting Information), which was not the case for PFBTTT . High-resolution X-ray diffraction measurements also show an increase in intensity of the (100) and corresponding higher order diffraction peaks upon annealing for both polymers ( Figure S16, Supporting Information), thus confi rming the increased crystallinity suggested by GIWAXS. The greater orientational order in PFBTTT compared to PBTTT could therefore be one of the factors resulting in the increase in mobility upon fl uorination. We also note that PBTTT fi lms annealed at 150 °C exhibit a split in the (100) and higher order out-of-plane signals, possibly due to different degrees of interdigitation, and side-chain reorganization.
The surface morphology of the fi lms was also investigated by AFM in fi lms made under the same conditions as used for device fabrication. The fi rst point to note is that the fi lm morphology achieved for PBTTT from spin-coating from TCB Adv. Funct. Mater. 2015, 25, 7038-7048 www.afm-journal.de www.MaterialsViews.com  is very different from the terraced nanostructure observed when spin-coating from other solvents such as 1,2-dichlorobenzene and chloroform. [ 13,34,42 ] Indeed we observe that PBTTT appears to have a large quantity of pin holes distributed across the whole fi lm ( Figure S17, Supporting Information), which were likely formed by the solventvapor bubbles during drying. Although fi lms of PBTTT and PFBTTT have similar a similar RMS roughness (6.79 nm and 6.99 nm respectively), these holes probably have a detrimental effect on the continuity of the fi lm and introduce extra traps, potentially hindering the transport of charge carriers. However, in the case of PFBTTT , we observe a fi lm comprised of a more fi brillar morphology ( Figure S17, Supporting Information). Though highly entangled, the fi brillar structure could facilitate the release of solvent vapor, leaving a more continuous fi lm with fewer pinholes. In fact, when diluted to 2 mg mL −1 , PFBTTT forms interconnected nanofi brils, with heights of approximately 5 nm and widths of 200-500 nm ( Figure 6 ), unlike PBTTT under the same dilution ( Figure S18, Supporting Information). This network of intricately woven fi bers could be a morphological explanation for the increased mobility observed in PFBTTT . Indeed, recent studies have shown that interconnectivity of ordered domains is crucial to achieve high mobilities with polythiophene derivatives. [ 2,21,22 ]

Conclusions
In conclusion, we have described the synthesis of a novel 3,3′-difl uoro-4,4′-dihexadecyl-2,2′-bithiophene monomer and report its copolymerization with thieno[3,2b ]thiophene to afford the fl uorinated analog of the well-known PBTTT polymer. We fi nd that backbone fl uorination has a pronounced infl uence on the physical properties of the polymer, with a signifi cantly enhanced degree of aggregation compared to the nonfl uorinated analog. Remarkably, we fi nd that the incorporation of just two fl uorine atoms on the polymer backbone results in an increase in the polymer melting temperature of 100 °C. DFT calculations suggest this increased aggregation originates from a greater degree of backbone planarity and rigidity. A result of this greater rigidity is an enhancement of the edge-on orientational order, and consequently a significant increase in hole mobility in OFET devices, which is also aided by an appreciably interwoven fi brillar morphology. In addition, the fl uorinated polymer exhibits excellent ambient stability for a nonencapsulated transistor device. We believe Adv. Funct. Mater. 2015, 25, 7038-7048 www.afm-journal.de www.MaterialsViews.com  that these results demonstrate that backbone fl uorination is a useful tool in designing high-performance organic semiconductors.
Synthesis of 3-Fluoro-4-Hexadecylthiophene (3) : In a dry 20 mL Biotage microwave vial under argon, 1 (2.00g, 6.15 mmol), and [1,1′-bis(diphenylphosphino)ferrocene]dichloropalladium(II) dichloromethane complex (251 mg, 0.307 mmol) were added. The vial was capped, evacuated, and backfi lled with argon three times, before adding hexadecylzinc bromide solution (16.0 mL, 0.5 M in THF). After stirring at room temperature for 2 min, the vial was heated for 1 h in a microwave reactor at 100 °C. The resulting mixture (solid at room temperature) was heated to 40 °C, poured into acetone, and fi ltered. The solvent of the fi ltrate was removed in vacuo, and the residue passed through a pad of silica using hexane, and the solvent was again removed in vacuo. In a 100 mL round-bottomed fl ask, the resulting crude mixture of 2 was dissolved in dry THF (10 mL) and cooled to 0 °C before N -tetrabutylammonium fl uoride (21.5 mL, 1 M solution in THF) was added. The reaction was stirred for 2 h before quenching with water and extracting with diethyl ether. The organic extracts were dried over MgSO 4 and the solvent removed in vacuo. The crude mixture was purifi ed by column chromatography on silica, using hexane as eluent, to yield 3 as a white waxy solid (300 mg, 15% over two steps).  (4) : In a dry 20 mL Biotage microwave vial under argon, 3 (400 mg, 1.23 mmol) was dissolved in dry THF (3.5 mL), and (2,2,6,6-tetramethylpiperidinyl) magnesium chloride lithium chloride complex (1.59 mL, 1 M solution in THF/toluene) was added dropwise at room temperature. The reaction was stirred for 1 h before a dispersion of [1,3-bis(diphenylphosphino)propane] dichloronickel(II) (333 mg, 0.615 mmol) in dry THF (7 mL) was added. The solution turned from orange to dark brown/black and solidifi ed. After diluting with THF, the reaction mixture was poured into dilute HCl and extracted with chloroform. The solvent was removed in vacuo, and the residue passed through a small pad of silica using dichloromethane as eluent. The solvent was removed in vacuo, and recrystallized from acetone, to yield 4 as a pale yellow solid (290 mg, 73% Synthesis of 5,5 ′ -Dibromo-3,3 ′ -Difl uoro-4,4 ′ -Dihexadecyl-2,2 ′ -Bithiophene (5) : In a 100 mL round-bottomed fl ask wrapped in foil, 4 (265 mg, 0.408 mmol) was dissolved in a mixture of chloroform (20 mL) and acetic acid (3 mL), and to this solution was added N -bromosuccinimide (154 mg, 0.815 mmol). The solution was stirred overnight, quenched with saturated sodium sulfi te, and extracted with chloroform. The organic layer was washed with 1 M sodium hydroxide, water, and brine, and the solvent was removed in vacuo. The crude product was recrystallized from a mixture of acetone and ethyl acetate (1:1), to yield 5 as a pale yellow solid (245 mg, 75%  [3,2-b]-thiophene] (PFBTTT) : In a dry 0.5-2 mL Biotage microwave vial, 6 (202.9 mg, 0.2509 mmol), bis(2,5-trimethylstannyl)thieno[3:2b ] thiophene (116.9 mg, 0.2509 mmol), tris(dibenzylideneacetone) dipalladium(0) (4.1 mg, 2 mol%), and tris( o -tolyl)phosphine (6.1 mg, 8 mol%) were added, and the vial capped and evacuated for 10 min. After backfilling with argon, degassed chlorobenzene (1.2 mL) was added, and the solution was degassed for a further 10 min. The mixture was then heated in a microwave in steps as follows: 100, 120, 140, 160 °C for 2 min each, and finally 180 °C for 30 min. After cooling to room temperature, the dark purple gel was precipitated in methanol from chlorobenzene, and purified by Soxhlet extraction (glass thimble), washing with methanol, acetone, hexane, (each overnight), chloroform (3 h), and finally extracting the polymer with chlorobenzene. Most of the solvent was removed in vacuo, before precipitating the polymer into methanol and filtering (184 mg, 93% DFT Calculations : DFT calculations were carried out using the B3LYP hybrid functional and the 6-31g(d) basis set in the GAUSSIAN09 software package. [ 59 ] Alkyl chains were replaced with a propyl group to simplify calculations and reduce computational time. Structures were optimized, and a frequency analysis was performed. Potential energy scans were performed on the trimers using the redundant coordinate editor and scanning the indicated dihedral angle in 36 steps of 10° increments.
Characterization : 1 H, 19 F, and 13 C NMR spectra were recorded on a Bruker AV-400 (400 MHz), using the residual solvent resonance of chloroformd or 1,1,2,2-tetrachloroethaned 2 and are given in ppm. Microwave experiments were performed in a Biotage initiator V 2.3. Polymer molecular weight and dispersity ( Ð ) analysis was completed via GPC in TCB at 140 °C using a Polymer Laboratories PL-220 HT GPC instrument calibrated against polystyrene standards. Electrospray mass spectrometry was performed with a Thermo Electron Corporation DSQII mass spectrometer. UV-vis spectra were recorded on a UV-1800 Shimadzu UV-vis spectrometer. Flash chromatography was performed on silica gel (Merck Kieselgel 60 230-400 mesh) or on reverse phase silica (Biotage SNAP KP-C18-HS cartridges). PESA measurements were recorded with a Riken Keiki AC-2 PESA spectrometer with a power setting of 5 nW and a power number of 0.5. Samples for PESA were prepared on glass substrates by spin-coating. DSC measurements, using ≈3 mg of material, were conducted under nitrogen at scan rate of 10 °C min −1 with a TA DSC-Q20 instrument. Flash DSC was performed on a Mettler Toledo Flash DSC 1 at a scan rate of 500 K s −1 . AFM images were obtained with a Picoscan PicoSPM LE scanning probe in tapping mode. GIWAXS measurements were performed at D-line, Cornell High Energy Synchrotron Source (CHESS) at Cornell University. A wide band-pass (1.47%) X-ray with a wavelength of 1.15 Å was shone on the samples with a grazing incidence angle of 0.15 o . A Pilatus 200k area detector was placed at a distance of 195 mm from the samples. A 1.5 mm wide tantalum rod was used to block the intense scattering in the smallangle area. The exposure time was 1 s. High-resolution X-ray diffraction measurements were carried out at G2, CHESS at Cornell University. The thin fi lms were aligned on a Kappa diffractometer to record the θ -2 θ www.afm-journal.de www.MaterialsViews.com scans. The wavelength of X-ray was 1.107 Å. An attenuator was used to allow ≈1/7 beam fl ux through and avoid saturation in the case of PBTTT sample at 200 °C annealing.
Device Fabrication and Characterization : All fi lm preparation and characterization steps were carried out under inert atmosphere. Bottomgate/top-contact devices were fabricated on heavily doped n + -Si (100) wafers with 400 nm thick thermally grown SiO 2 . The Si/SiO 2 substrates were treated with trichloro(octadecyl)silane to form a self-assembled monolayer. The polymers were dissolved in hot TCB (5 mg mL −1 ) and spin cast at 2000 rpm from a hot solution for 60 s before being annealed at 200 °C for 30 min. Au (30 nm) source and drain electrodes were deposited onto the polymer fi lm under vacuum through shadow masks. The channel width and length of the transistors are 1000 and 50 µm, respectively. Transistor characterization was carried out under nitrogen using a Keithley 4200 parameter analyzer. Mobility was extracted from the slope of I D 1/2 versus V G .

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author. Additional data relating to the paper can be found at doi.org/10.6084/m9.fi gshare.1539547