Poly(2‐vinylpyridine) as an Additive for Enhancing N‐Type Organic Thin‐Film Transistor Stability

N‐type organic semiconductors are particularly susceptible to degradation by ambient air. One such solution to this issue is to include additives in the inks these semiconductors would be cast from that would enhance device stability after film deposition. This method would reduce the number of processing steps needed to fabricate devices compared to other stabilization methods, such as encapsulation. In this study, the stabilization of n‐type performance of the semiconductor poly{[N,N′‐bis(2‐octyldodecyl)‐naphthalene‐1,4,5,8‐bis(dicarboximide)‐2,6‐diyl]‐alt‐5,5′‐(2,29‐bisthiophene)} (P(NDI2OD‐T2)) when it is blended with an increasing proportion by weight of poly(2‐vinylpyridine) (P2VP) is reported. The simple synthesis of P2VP also makes it an ideal candidate material for large‐scale applications. Concentrations as low as 0.1% P2VP incorporated into the P(NDI2OD‐T2) blends provided an immediate stabilization effect, and at 10% and 50%, longer‐term stability after one week is observed.


Introduction
[3] While this technology is promising, widespread adoption requires further improvements on the stability and performance of these devices. [4,5]Electron-transporting materials (n-type) are particularly susceptible to unstable performance due to their typically lower-lying LUMO (lowest unoccupied molecular orbital) level where electron transport occurs being close to the redox potentials of oxygen and water, present in ambient air. [6]While developing higher-performing and more stable n-type materials has been the focus of significant research in recent years, [7][8][9] materials that are currently commercially available and viable for production at scale are still in need of improvement for device applications.While the design and scale-up of new materials is an important pursuit, fabricating air-stable devices with existing materials through simple doping/blending can accelerate commercialization of OTFT-based applications.
Device encapsulation can enable air stable device operation, [10,11] but can be challenging due to the choice of solvent for effective orthogonal processing.As the field moves toward designing semiconductors with solubility in more environmentally-neutral solvents, fewer choices become available to enable orthogonal processing without dissolving or damaging the semiconductor during deposition of an encapsulant.Additionally, encapsulation is not perfect, and air or moisture diffusion can occur through the encapsulation layer over time.Therefore, it is essential to develop new strategies for improved air stability either to replace or to complement encapsulation.
Dopants and additives have been studied for enhancing both stability and performance of OTFTs. [12,13]DMBI (4-(1,3-dimethyl-2,3-dihydro-1H-benzoimidazol-2-yl)phenyl)dimethyl-amine), a popular n-type additive, does not hinder device stability while enhancing performance. [14,15]However, its synthesis is complex which can lead to increased costs.Recently, aromatic amines have shown to lend improvements to both stability and performance of n-type OTFTs. [16]We demonstrated that the addition of amine solvents such as pyridine can lead to enhanced stability of n-type OTFTs. [17]However, this improved stability was determined to be short lived, likely due to the volatility of the solvent escaping the semiconductor layer over time.Therefore, in this study we aim to maximize the stability through the use of non-volatile pyridine functional polymers.Poly(2-vinylpyridine) (P2VP) is an ideal candidate as a non-volatile pyridine-based additive as it contains only the pyridine functional group attached to the polymer backbone, to ensure minimize any impact on performance caused by other functional groups, and to avoid bulky side chains that may interrupt packing of the semiconductor.We synthesized P2VP and blended it with n-type polymer poly{[N,N′-bis(2-octyldodecyl)-naphthalene-1,4,5,8-bis(dicarboximide)−2,6-diyl]-alt-5,5′-(2,29-bisthiophene)} (P(NDI2OD-T2)) in various ratios.We found that the air stability of the OTFTs could be improved without P2VP disrupting the stacking of the P(NDI2OD-T2) chains resulting in minimal drop in device performance.

Discussion
Bottom gate top contact (BGTC) OTFTs were fabricated using films of P(NDI2OD-T2) blended with varying ratios of poly(2vinylpyridine) (P2VP) on silicon wafers with SiO 2 as the gate dielectric (Figure 1).Blends of 0.1, 1, 5, 10, and 50 wt.% of P2VP in P(NDI2OD-T2) were compared to a baseline of pure P(NDI2OD-T2).The devices were fabricated and annealed in glovebox prior to electrical characterization.These OTFTs were first characterized in an inert environment, then in air after 30 min of air exposure.The devices were then left in air for 7 days and characterized again to assess the effect of P2VP on the device stability.Figure 2 reports the electron mobility (μ e ) and threshold voltage (V T ) for the corresponding OTFTs, and the characteristic transfer curves can be found in the supporting information (Figure S1, Supporting Information).As shown in Figure 2, when the lowest concentration studied, 0.1 wt.% P2VP, is added to P(NDI2OD-T2), an initial drop in μ e was observed compared to pure P(NDI2OD-T2) when characterized under inert atmosphere.Increasing the amount of P2VP resulted in further drop in μ e in an inert environment, but the V T also decreased with increase in P2VP loading, dropping to ≈0 V with the addition of ≥ 5 wt.%P2VP in the blends.The drop in mobility is likely due to incorporation of a non-conducting element in the semiconductor layer, while the drop in V T could be a result of a doping effect from the P2VP. [13,18,19]When characterizing in air, the presence of as little as 0.1 wt.% of P2VP resulted in a greater μ e and reduced V T in air compared to the baseline devices characterized in inert atmosphere.OTFT devices with 10 and 50 wt.%P2VP were the only conditions which displayed any electrical performance in air after 7 days, suggesting the addition of P2VP improved the device stability.
To correlate the device performance with the morphology of films of blended P(NDI2OD-T2) and P2VP, we characterized the films by scanning transmission X-ray microscopy (STXM) and grazing-incidence wide-angle X-ray scattering (GIWAXS).[22] Unlike AFM, this technique provides domain composition through the entire film, not just the surface, providing a more representative understanding of the blend.STXM maps (Figure 3) of our blends show the distribution of P(NDI2OD-T2) and P2VP, with white pixels indicating pure P(NDI2OD-T2) domains, black pixels indicating pure P2VP domains and gray pixels representing a blend.While P2VP and P(NDI2OD-T2) are well dispersed throughout the whole volume of each film, some degree of phase separation occurred between P(NDI2OD-T2) and P2VP, as evidenced by the variation in composition through the film.Hildebrand solubility parameters for all three polymers are relatively close in value, being 18.7, 21.3, 20.5 MPa 1/2 for PS, [23] P2VP, [24] and P(NDI2OD-T2) [25] respectively.This suggests that the polymers should all have some miscibility, leading to the observed mixing of the polymer phases. [26]ith the addition of 0.1% P2VP (Figure 3a), the majority of the film is pure P(NDI2OD-T2).Films with 10% P2VP (Figure 3b) showed a greater degree of phase separation, with some regions having 80% P(NDI2OD-T) (20% P2VP), further suggesting the P2VP had become more concentrated in those regions.Both films with addition of 0.1 and 10 wt.% P2VP were characterized by having pure P(NDI2OD-T2) domains intermixed with blended domains of P2VP and P(NDI2OD-T2).However, for the sample of 50 wt.%P2VP (Figure 3c) there were no pure P(NDI2OD-T2) domains, as the P2VP permeated the entire film.With 50 wt.%P2VP, some regions showed as little as 28% P(NDI2OD-T2), with regions having a maximum of 89% P(NDI2OD-T2), again indicating significant phase separation from P2VP.An additional version of this figure with an adjusted scale is included for enhanced visual fidelity in the Supporting Information (Figure S2, Supporting Information).Increasing P2VP concentration in the blends also increased the observed sizes of P2VP-rich domains.These results indicate that at low P2VP content, there are still pure P(NDI2OD-T2) domains which participate in the charge transport similar to a pure P(NDI2OD-T2) film, which is why only a small drop in μ e was observed (Figure 2).As the P2VP content is increased to 50 wt.%there are no pure P(NDI2OD-T2) domains that span the entire film thickness, and therefore the charge transport through the film is significantly affected, as evidenced by the drop in μ e compared to the pure P(NDI2OD-T2) film in inert atmosphere (Figure 2).Among all the samples, no regions of pure P2VP were observed, suggesting P2VP is partially miscible in the P(NDI2OD-T2) domains which is likely favourable for increasing the interactions between the two polymers and favourable for increased device stability.
GIWAXS can provide information related to the crystalline regions of the film, which are directly related to the charge transport characteristics of the film.The results from the GIWAXS analysis are shown in Figure 4.At all P2VP loadings we observe the characteristic semi-crystalline P(NDI2OD-T2) peaks. [27]This indicates that even at 50 wt.%P2VP, the regions of film with a  high concentration of P(NDI2OD-T2) will likely still exhibit similar packing as the neat films.We do observe a drop in P(NDI2OD-T2) GIWAXS peak intensity with the addition of 50 wt.%P2VP which suggests that while the P(NDI2OD-T2) packing is similar, there is less of it, which is consistent with a film that has half the P(NDI2OD-T2) content.Therefore, the addition of P2VP does not appear to interrupt the packing of P(NDI2OD-T2), suggesting charge transport is similar through the P(NDI2OD-T2) domains and that the drop in OTFT performance is likely due to the reduction in P(NDI2OD-T2) domain size (or quantities) with the addition of P2VP.Overall, STXM and GIWAXS results indicate that P2VP appears to confer better long-term stability to the device when it is more dispersed into the film and does not impact the packing of P(NDI2OD-T2) to achieve this.
We recently reported the screening of different solvent additives and their effect on increasing the air stability of P(NDI2OD-T2) based OTFTs. [17]We found that pyridine was effective at increasing the n-type stability by increasing the resistance of the polymer to oxygen degradation.In this previous study we demonstrated that the improved stability was not a result of a change in morphology but rather a result of the presence of the appropriate pyridine functional group.In the present study, we wanted to confirm that the increased stability comes from the pyridine functional group and not change in morphology, aligning with the results of the previous study.Therefore, we fabricated a series of OTFT devices using blends of P(NDI2OD-T2) with poly(styrene) (PS), which were prepared and characterized under the same conditions.As PS has a similar structure to P2VP without the nitrogen atom present, this makes it possible to determine whether the amine is contributing to the observed sta-bilizing effect.Figure 5 presents the results obtained with the P(NDI2OD-T2) / PS blends.The presence of PS does not have the same initial impact on performance as P2VP.Only a small initial decrease in μ e was observed.In fact, the presence of 10 and 50 wt.%PS resulted in an increase in μ e while the devices were in air.However, this does not confer a long-term stabilizing effect.After 7 days, none of the devices displayed functionality in air regardless of PS content.The initial increase in μ e with the addition of 10-50 wt.% of PS is temporary.This indicates that the amine indeed has a stabilizing effect on these devices.This initial increase in performance in air observed from 10 and 50 wt.%PS blends may be due to an initial reduction of air and/or moisture diffusion into the devices due to the PS on the device surface.Similar to the addition of P2VP, the presence of PS was also found not to impact the packing of P(NDI2OD-T2) in the thin films by GIWAXS analysis (Figure S3, Supporting Information).Furthermore, no significant change in V T was observed for the OTFTs with PS blending suggesting the effects are not influencing the charge injection and contact resistance of the OTFTs, and do not interact with P(NDI2OD-T2) in the same way as P2VP which resulted in a decreased V T .
While the high percentages of PS did provide some stabilizing effect on the short-term, the presence of P2VP provided greater stabilization of performance over time.Comparing the performance difference between devices characterized in the glovebox to the initial air exposure, the P(NDI2OD-T2) + 50 wt.%P2VP devices retained 49.5% of their μ e while the P(NDI2OD-T2) + 50 wt.%PS devices only retained 16.5% of their μ e (Figure 6).Again, no device performance was observed when using PS after 7 days.Comparing the results from P2VP in the present study to our previous results from the screening that identified pyridine as a potential stabilizing moiety (Table 1) shows that μ e remains in the same order of magnitude, and V T is comparable as well.
Cyclic voltammetry (CV) was performed on blended films of P(NDI2OD-T2) and P2VP (Figure 7).With the inclusion of P2VP, the oxidation peaks disappear when 0.1 wt.% of P2VP is introduced into the film.As the percent of P2VP increases, the reduction peaks also disappear, until there are no visible redox events occurring.At 5 wt.%P2VP and above, there are no visible redox peaks remaining.In addition to this, when the voltammograms are run for multiple cycles, increasing the amount of P2VP decreases peak spreading and shifting in the voltammograms in subsequent cycles (Figure S4, Supporting Information).This decrease in the magnitude of changes observed in CV measurements may also indicate stabilization toward these redox reactions, as there is greater reversibility.In our previous work, [17] it was found that exposing P(NDI2OD-T2) films to pyridine also resulted in a loss of visible redox peaks in this range.Therefore, these results are consistent with the previous observations with pyridine suggesting the P2VP is improving device stability through the reduction of P(NDI2OD-T2) oxygen oxidation.Table 1.Comparison of stabilization by pyridine vapour exposure [17] and P2VP blending.
Pyridine [ 17] P2VP [50%] μ e in air [cm 2 V −1 s −1 ] 5 .6 × 10 −3 Taken with the results from the STXM in Figure 3, it appears that the morphologies that enhance contact with P2VP and P(NDI2OD-T2) led to greater long-term stability.If P2VP is acting as a stabilizing element by preventing these redox events, it may need to be in contact with P(NDI2OD-T2) to provide longer term stability.Future studies that focus on enhancing the contact between P2VP and P(NDI2OD-T2) would be valuable to determine the extent of this effect, and to optimize the stabilization of the devices.

Conclusion
Incorporating as little as 0.1% P2VP blended with P(NDI2OD-T2) provided stabilization when exposed to air.However, at least 10% P2VP was required for long-term stability over 7 days.In order to confirm the impact was due to the pyridine moiety of P2VP, devices were also fabricated with PS and these devices did not exhibit improved stability.Cyclic voltammetry indicates P2VP may prevent damaging redox reactions with air or water from occurring.The morphology of the blends may explain why longterm stability is observed in blends with higher ratios of P2VP.In these higher-ratio blends, there is sufficient P2VP present to have few to no regions with only P(NDI2OD-T2), thus ensuring contact with the stabilizing effects from P2VP.

Experimental Section
Polymer Synthesis: Poly(2-vinylpyridine) (P2VP, 97%, Sigma-Aldrich) was synthesized by Nitroxide Mediated Polymerization (NMP) employing a process similar to literature procedures. [28]First, the initiator NHS-BlocBuilder (NHS-BB) was synthesized by coupling of BlocBuilder-MA (Arkema) and N-hydroxysuccinimide (Oakwood, 98%) following literature procedures. [29]The homopolymerization of 2VP was performed in a 50-mL three-neck round bottom glass flask fitted with a condenser and a magnetic stir bar.Two of the necks in the flask were sealed with rubber septa while the centre neck was connected to the condenser and was capped with a septum pierced with a needle to serve as a vent for the system.The NHS-BlocBuilder (0.030 g, 0.065 mmol) was dissolved in 2VP monomer (3.2 g, M n, Target = 49.3 kg mol −1 ).The solution was then bubbled with nitrogen gas for 30 min at room temperature to remove oxygen and then heated to 120 °C using a heating mantle with a heat controller while maintaining the nitrogen purge.The start of the reaction was taken when the target temperature was reached (t = 0 min).The solution was allowed to react for 60 min before the reaction was stopped by cooling the solution to room temperature.The polymer product was precipitated in hexanes, filtered using filter paper and allowed to dry under vacuum overnight.The purified polymer has number-average molecular weight (M n = 25.000kg mol −1 ) and dispersity (Ð) = 1.3.M n and Ð were determined by size exclusion chromatography performed on a Malvern Omnisec equipped with Omnisec Resolve pump and autosampler (CHR7100) with two T6000M columns and Omnisec Reveal differential refractive index (CHR6000), diode-arraybased UV-vis spectrometer, and Viscotek SEC-MALS 20 multi-angle light scattering detectors.THF (HPLC, Caledon Laboratory Chemicals) was used as the eluent at 30 °C with a flow rate of 1 mL min −1 .
Device Preparation: Si/SiO 2 substrates (Ossila, SiO 2 thickness 300 nm) were cleaned by successive 5-min sonication steps while submerged in water, acetone, and methanol.Substrates were dried with a flow of nitrogen, then plasma treated for 10 min.Immediately following sonication, substrates were rinsed with water, then isopropyl alcohol, then dried again with a nitrogen flow.The substrates were then immersed in a solution of 1% (v/v) octyltricholorosilane (TCI Chemicals, OTS) in toluene (Sigma-Aldrich) and heated at 70 °C for 16 h.Following OTS treatment, the substrates were rinsed with toluene to remove excess solution, then dried in a vacuum oven for 1 h at 70 °C.
Semiconductor and Electrode Deposition: Solutions of P(NDIOD-T2) (1-Material) blended with P2VP or PS were prepared as follows.P(NDI2OD-T2) was dissolved in 1,2-dichlorobenzene (Sigma-Aldrich) at 10 mg mL −1 by stirring on a hotplate at 50 °C for 1 h.P2VP (or PS) solutions were separately prepared at 10 mg mL −1 in 1,2-dichlorobenzene by the same method.These solutions were then mixed together to create the appropriate percentages of P2VP (or PS) by mass: 0% (no P2VP or PS), 0.1%, 1%, 5%, and 50%, such that the total concentration of all polymer in solution was 10 mg mL −1 .For example, the 50% sample contained a concentration of 5 mg mL −1 P2VP (or PS) and 5 mg mL −1 P(NDI2OD-T2).These solutions were cast by spin coating with 80 μL of solution at 2000 rpm.The films were annealed at 150 °C for 1 h under vacuum.
Electrodes were deposited by physical vapour deposition through shadow masks.10 nm of chromium was deposited at a rate of 0.5 Å s −1 , then gold was deposited at a rate of 1 Å s −1 .Electrodes had the following dimensions: 30 μm channel length, 1000 μm channel width.20 individual transistor devices were prepared per substrate, and two substrates were prepared per condition.Devices were stored in a dry glovebox for a maximum of 1 day prior to electrical characterization.
Electrical Characterization: Devices were characterized first in a dry nitrogen glovebox.A Keithley 2614B sourcemeter was used to vary the gatesource voltage (V gs ) from 0-60 V and the source-drain voltage (V ds ) from 0-60 V to generate output curves, and V gs from 0-60 V while holding V ds at 50 V to generate transfer curves.Devices were then removed from the glovebox and characterized after 30 min of air exposure, with the characterization apparatus also exposed to air.After 7 days leaving the devices in ambient conditions, the devices were characterized again in air.
Scanning Transmission X-Ray Microscopy (STXM): Silicon nitride windows (SiN) with a 50 nm thickness on a Si frame were purchased from NORCADA.The SiN membranes were treated for 15 min with oxygen plasma to clean them.SiN membranes were spin-coated with films as previously described in Device Preparation.
Imaging and Analysis: STXM measurements were performed on the HERMES beamline at the SOLEIL synchrotron.A Fresnel Zone Plate with 25nm outer zone width was used to focus the monochromatized X-ray beam on the sample, and transmitted X-rays were collected using a PMT.The microscope chamber was maintained under a vacuum of 10 −4 mbar during all measurements.High resolution image stacks composed of 24 images (2 × 2 μm, 20 nm step size, 3 ms dwell time) at the carbon Kedge were obtained, in an energy range of [284.5-288] eV.The beam was maintained at linear horizontal polarization for all measurements.
The composition maps of P2VP:P(NDI2OD-T2) blends (Figure 3) were drawn from thickness values obtained from singular value decomposition (SVD) of energy stacks at the carbon K edge.Reference spectra were scaled down to an elemental thickness of 1 nm using simulated OD spectra (simulations parameters: P2VP monomer of chemical formula C 7 H 7 N with a density of 0.977 g cm − 3 and P(NDI2OD-T2) monomer of chemical formula C 62 H 88 N 2 O 4 S 2 with a density of 1 g cm −3 ).These 1 nm experimental OD spectra of pure materials were used to fit the energy stacks of the different blends following the SVD linear regression (blend spectrum = a * 1 nm P2VP spectrum + b * 1 nm P(NDI2OD-T2) spectrum) to yield thickness maps of both components in each blend (a and b in the above formula).The composition maps shown in Figure 3 and Figure S1 (Supporting Information) are obtained from the thickness maps as follows: % of P(NDI2OD-T2) = thickness of P(NDI2OD-T2) / (thickness of P2VP + thickness of P(NDI2OD-T2)).
The beamline energy was calibrated on a polystyrene film, and data was processed using aXis2000 (available at http://unicorn.chemistry.mcmaster.ca/aXis2000.html). [30]razing Incidence Wide-Angle X-Ray Scattering (GIWAXS): GIWAXS measurements were performed at the SIRIUS beamline at the SOLEIL Synchrotron in Saint-Aubin, France. [31]The samples were placed in a chamber that was flushed with helium gas to reduce air scattering.The sample-todetector distance was calculated to be 339 mm and the X-ray energy was 10 keV.The detector was placed at an angle of 8.9°.Grazing incidence patterns were taken at  = 0.1°with ten images taken at an exposure time of 10 s each.The final spectra were the sum of ten images.The GIWAXS data were calibrated against a silver behenate (AgBe) standard and analyzed using the GIXSGUI software package. [32]

Figure 2 .
Figure 2. Effect of blending P2VP with P(NDI2OD-T2) on a) mobility of the devices in a glovebox, b) relative retention of mobility of air-exposed devices compared to performance in the glovebox, c) threshold voltage of the devices in a glovebox, d) relative change of threshold voltage of air-exposed devices compared to performance in the glovebox.Statistical analysis in the box plots is derived from 35-40 individual OTFTs.

Figure 5 .
Figure 5.Effect of blending PS with P(NDI2OD-T2) on a) mobility of the devices in a glovebox, b) relative retention of mobility of air-exposed devices compared to performance in the glovebox, c) threshold voltage of the devices in a glovebox, d) relative change of threshold voltage of air-exposed devices compared to performance in the glovebox.Statistical analysis in the box plots is derived from 35-40 individual OTFTs.

Figure 6 .
Figure 6.Comparison of device stabilization by a) P2VP and b) PS.Error bars represent the standard deviation of 35-40 individual OTFTs.