One‐Step Sixfold Cyanation of Benzothiadiazole Acceptor Units for Air‐Stable High‐Performance n‐Type Organic Field‐Effect Transistors

Abstract Reported here is a new high electron affinity acceptor end group for organic semiconductors, 2,1,3‐benzothiadiazole‐4,5,6‐tricarbonitrile (TCNBT). An n‐type organic semiconductor with an indacenodithiophene (IDT) core and TCNBT end groups was synthesized by a sixfold nucleophilic substitution with cyanide on a fluorinated precursor, itself prepared by a direct arylation approach. This one‐step chemical modification significantly impacted the molecular properties: the fluorinated precursor, TFBT IDT, a poor ambipolar semiconductor, was converted into TCNBT IDT, a good n‐type semiconductor. The electron‐deficient end group TCNBT dramatically decreased the energy of the highest occupied and lowest unoccupied molecular orbitals (HOMO/LUMO) compared to the fluorinated analogue and improved the molecular orientation when utilized in n‐type organic field‐effect transistors (OFETs). Solution‐processed OFETs based on TCNBT IDT exhibited a charge‐carrier mobility of up to μ e≈0.15 cm2 V−1 s−1 with excellent ambient stability for 100 hours, highlighting the benefits of the cyanated end group and the synthetic approach.


Material characterizations
Cyclic and square-wave voltammograms were recorded using a Metrohm Autolab PGStat101 potentiostat/galvanostat. The experimental setup consisted of an Ag/Ag + reference electrode, a platinum wire counter electrode and an FTO working electrode, and all measurements were carried out under nitrogen at room temperature. Measurements were performed in anhydrous, degassed solutions of CH2Cl2 with tetrabutylammonium hexafluorophosphate (0.1 M) electrolyte. After each measurement, an arbitrary amount of ferrocene was added to the solution as an internal reference. Square-wave voltammetry (SWV) measurements were conducted in a 0.1 M solution of [n-Bu4N]PF6 in CH2Cl2 at 25 Hz frequency, 20 mV step size, and 50 mV pulse height. The SWV potentials were referenced to those of ferrocene when a ferrocene/ferrocenium reference redox system of 4.8 eV below the vacuum level was used as an internal standard, and the conversion from electrochemical potentials to electron volts was done using the formula E(eV) = -Eredox -4.8 eV. 2,3 Any solvent effects were neglected.

Raman spectroscopy
A Renishaw inVia Raman microscope with a 50× objective in a backscattering configuration was used to collect both PL and Raman spectra. Films for the measurements were prepared by spin-coating from 10 mg mL -1 solutions in chloroform onto glass substrates. To avoid photooxidation all samples were measured in a nitrogen purged Linkam sample chamber. All measurements were taken with a defocused laser spot with a radius of ≈ 10 µm, laser powers and acquisition times were optimised and kept consistent between samples. Both Raman and PL spectra were recorded using 514 nm laser excitation. Raman spectra were baselined by subtraction of a polynomial function; PL spectra are shown with no baselining or instrument response corrections.

Ambient Photoemission Spectroscopy (APS)
Ionisation potentials were measured by ambient photoemission spectroscopy (APS) measurements using an APS04 air photoemission system (APS04, by KP Technology) and a 2 mm gold tip under atmospheric conditions. Films for the measurements were prepared by spincoating from 10 mg mL -1 solutions in chloroform onto ITO substrates. Films were annealed at 80°C, 150°C and 200°C for three minutes in an inert atmosphere. All samples were grounded via the ITO substrate. Measurements were taken at multiple positions on the films to ensure reproducibility. The APS data were processed using the protocol described by Baikie et al. 4 This involved taking the cube root of the measured photoemission, fitting the resultant linear region, and extrapolating to zero photoemission to find the HOMO level of the semiconductor.

Density functional theory (DFT) calculation
DFT calculations were conducted using Gaussian 09 software on the Imperial College High-Performance Computing Service. 5 All simulations were carried out on single molecules in the gas phase at the B3LYP level of theory with the basis set 6-31G(d,p). [6][7][8] Alkyl side chains were replaced by methyl groups to reduce the computation time. Structures were optimized to a local minimum energy conformation, and frozen dihedral angles were used to simulate molecular conformational changes. Frequency calculations were carried out to simulate the Raman spectra; an empirical scaling factor of 0.97 was applied to the calculated wavenumber. 9 Visualization of the simulated vibrational modes using GaussView 6.0.16 software was used to aid Raman peak assignment alongside consultation with the literature.

Device fabrication and characterizations
Transistor characterization was carried out under nitrogen using a Keithley 4200 parameter analyser. All films were prepared and characterized under inert atmosphere. Bottom gate/top contact (BG/TC) devices were fabricated on heavily doped n + -Si(100) wafers with 300 nm thick thermally grown SiO2. The Si/SiO2 substrates were treated with trichloro(octadecyl)silane (ODTS) to form a self-assembled monolayer. TCNBT IDT was dissolved in chloroform (20 mg mL -1 ) and spun cast at 2000 rpm for 60 seconds from a room temperature solution and annealed at 150°C for 3 minutes. Al (50 nm) or Au (30/40 nm) source and drain electrodes were deposited under vacuum through shadow masks. The channel width and length of the transistors were 1500 μm and 50 μm, respectively. The transfer and output characteristics were determined in a N2-filled glove box using a Keithley 4200 source meter.
The saturation-regime mobility of the transistor was determined using the equation Ids = (WCi /2L)µsat(Vg-VT) 2 , where Ids is the source-drain current, Ci (10 nF cm -2 ) is the capacitance per unit area, L is the channel length, W is the channel width, and Vg and VT are the gate and threshold voltages, respectively.
Top gate/bottom contact (TG/BC) devices were fabricated on glass substrates using Au (40 nm) source-drain electrodes and a Cytop dielectric. The channel width and length of the transistors were 1000 μm and 30/40 μm, respectively. Both TFBT IDT and TCNBT IDT were dissolved in chloroform (10 mg mL -1 ) and spin coated at 2000 rpm from a room temperature solution for 60 seconds before being annealed at 120°C and 150°C for 3 minutes, respectively.
The saturation-regime mobility of the transistor was determined using the equation Ids = (WCi /2L)µsat(Vg-VT) 2 , where Ids is the source-drain current, Ci (2.1 nF cm -2 ) is the capacitance per unit area, L is the channel length, W is the channel width, and Vg and VT are the gate and threshold voltages, respectively. For the air stability TFT measurements, the saturation mobility was calculated from the slope of the second derivative of the drain current versus gate voltage.
The reaction mixture was heated to 80 to 85°C for 6 hours. After cooling down to room temperature, the solution was decanted in 500 mL of ice water and the solution was neutralised with 50% NaOH to reach a pH of 7-8. After filtration, the filtrate was extracted with ethyl acetate (3 x 500 mL) and dried over MgSO4. Evaporation of ethyl acetate under reduced pressure afforded the product as brown solid (7.
When no starting material remained, the reaction was left to cool to room temperature. The excess bromine was reduced with sat. Na2SO3 and a flaky white solid crashed out, which was filtered and washed with deionised water (~1 L) to afford the brominated product TFBT-Br as       APS values indicate deepening of HOMO energy levels (measurement error ±0.05 eV) and increased crystallinity. The Fermi level also deepens following the same trend. The LUMO energy level was calculated by the addition of the optical band gap (obtained from the intersection wavelength of the absorption and photoluminescence spectra (Figure S5  a)).  ! !!" # = (a * ) 2 +(b * ) 2 + (c * ) 2 +2a*b*cos(γ*) = (q 100 ) 2 +(q 010 ) 2 + (q 001 ) 2 +2q 100 q 010 cos(γ*)

k= cos(γ-ϕ)*sin(γ)/sin(ϕ)
q pi-pi cos(ϕ) = ka*, q pi-pi =2π/d pi-pi =2π/b·sin(ϕ) a*= 2π/a·sin(γ) (2π/b·sin(ϕ)) x cos(ϕ) =h 2π/a·sin(γ) h= cos(ϕ)*a*sin(γ)/b·sin(ϕ) 9. Raman-structure changes upon film annealing Figure S18. a) (Top) Normalised simulated Raman spectra of TCNBT IDT at its minimum energy structure and with an induced dihedral torsion of 60°, (bottom) experimental normalised and baselined Raman spectra taken at 514 nm excitation of TCNBT IDT films at different annealing temperatures. Arrows indicate matching peak changes that occur upon moving towards a more planar structure in the simulated spectra and upon annealing in the experimental spectra. The similarity between the changes suggests that a more planar structure is adopted by the molecules upon annealing. b) Normalised and baselined Raman spectra of TCNBT IDT with peaks numbered for assignment (see Table S17 below) for simulated spectra (top) and experimental (bottom) spectra taken on an 80°C annealed film at 514 nm excitation.

b) a)
TC/BG Device Optimization -Active layer annealing temperature effect Figure S23. Active layer film annealing temperature dependence of device performance for TCNBT IDT. Films a) as-cast b) 80°C c) 150°C and d) 200°C annealing temperature output curves. The film annealed at 150°C was the best performing film with the highest drain current, minimum contact resistance and highest saturation mobility value. For the output measurements gate voltage (Vg) was applied from 0 to 60 V (10 steps). Figure S24. In both figures, the TCNBT IDT film was annealed at 150°C. a) OFET performance without device annealing and b) OFET performance after the device was annealed at 100°C for 3 minutes.   Figure S25. The air stability of TCNBT IDT in the solid-state was tested using a top contact/bottom gate architecture. A silicon substrate was treated with octadecyltrichlorosilane (100 μL in 10 mL, 10 minutes at 100°C) and TCNBT IDT was spin coated from a chloroform solution (20 mg/mL, 2000 rpm) on top. The films were annealed at different temperatures and left under ambient conditions for 10 weeks [as cast films (w/o-red), 80°C annealed film (blue), 150°C annealed film (yellow). Every 3 weeks, fresh Al contacts were evaporated on top, and the device was measured. As the graph above indicates, TCNBT IDT film mobility remained stable for the annealed films but decreased for the as-cast film.