Post‐Polymerization Strategy via Dual Site Clicking for Synthesizing Intrinsically Cross‐Linkable Semiconducting Polymers

Abstract

ABSTRACT

Crosslinked organic semiconductors have opened the way for various fabrication techniques in the field of organic electronics owing to their three‐dimensional network structure with high solvent resistivity. However, recent efforts to synthesize cross‐linkable semiconducting polymers have been limited by their low molecular weights and yields. In this study, this limitation is overcome by a novel post‐polymerization strategy. A reagent with a cross‐linkable functional group, (3‐mercaptopropyl)trimethoxysilane, is attached to a diketopyrrolopyrrole‐based donor–acceptor copolymer (DPPTT) via thioesterification and para‐fluoro‐thiol reaction, modifying two sites simultaneously. This modification preserves the molecular weight and electrical properties of the original polymers. In addition, the use of click chemistry enables high yield (98%) without any purification. The modified DPPTT demonstrated high resistance to organic solvents (80% retention dipped in 1‐chlorobenzene for 1 h). Exploiting this benefit, an ultrathin flexible array of 100 organic field‐effect transistors fabricated using conventional photolithography showed high‐performance reliability. Thus, this study provides a universal strategy to synthesize versatile polymer semiconductors for practical organic electronics.

Introduction

With the growing demand for versatile electronic materials, organic semiconductors have been extensively studied over the past few decades. Their low cost, flexibility, stretchability, and diversity have filled the gaps that the inorganic counterparts would not be able to reach [1–10]. Among these unique properties, solution processability is the key advantage that enables the effective printing or coating of semiconductors over a large area [11–13]. A common strategy for endowing solution processability is to attach solubilizing side chains to the backbone of semiconducting polymers [14, 15]. However, their high solubility is double-edged; thin films fabricated from solution-processed polymers can be easily damaged when exposed to subsequent processing solvents [16]. This critically narrows the options available for subsequent fabrication processes for electronic devices with multiple layers.

To overcome the dark aspect of high solubility, the chemical resistance of the fabricated polymer film has been enhanced using chemical crosslinking strategies [17–20]. With chemical crosslinking, polymers covalently bond to form a network, thereby suppressing damage from subsequent processes. These methods are generally achieved by introducing cross-linkable agents or synthesizing intrinsically cross-linkable polymers. The former approach involves fabricating films using a blended solution of a semiconducting polymer and crosslinking agents, followed by external stimuli that crosslink the resulting film [21–23]. However, crosslinking agents are commonly insulating and crosslinkers often induce phase separation between the polymer and agent, degrading the electrical characteristics [24]. The latter approach effectively avoids this problem by directly attaching cross-linkable functional groups to polymers [25–27]. However, its potential has been largely limited by the poor yield of cross-linkable monomer synthesis resulting from the functional-group attachment process (e.g., bromination yield is 23% for siloxane-attached monomer) [28]. Moreover, the existence of cross-linkable sites along the reacting monomers reduces solubilizing power, sterically hinders the polymerization reaction, and causes unintended crosslinking reactions during the polymerization process, leading to the low molecular weight of the polymer (e.g., molecular weight of siloxane-incorporating polymer is 39 kDa) [29].

This study presents an effective strategy for synthesizing intrinsically cross-linkable donor–acceptor (DA) copolymers characterized by both high yields and high molecular weights. Our approach focuses on post-polymerization modification—a powerful tool for modification of polymer semiconductors after polymerization [30–32]. In detail, the attachment of a cross-linkable molecule (3-mercaptopropyl)trimethoxysilane (MPS) to both ends and along the backbone of the diketopyrrolopyrrole (DPP)-based DA copolymer (DPPTT) takes place after polymerization through click chemistry (Figure 1a). The attached functional groups then covalently crosslink the polymers via silanol condensation, resulting in a solvent-resistive film (Figure 1b). This study provides a detailed account of the synthetic route and mechanisms used for the synthesis of an intrinsically cross-linkable DPPTT series, referred to as the Cx-DPPTT series, through post-polymerization modification reactions. In addition, we present compelling evidence of enhanced solvent resistance, which was verified by fabricating patterned polymer thin films using conventional photolithography techniques. Furthermore, we illustrate the advantages of intrinsically cross-linkable semiconducting polymers by showcasing the development of ultrathin flexible organic field-effect transistor (OFET) arrays using these materials.

[IMAGE OMITTED. SEE PDF]

Materials and Methods

Instrumentation and Materials

3,6-Bis(5-bromothiophen-2-yl)-2,5-bis(2-octyldodecyl)pyrrolo[3,4-c]pyrrole-1,4(2H,5H)-dione, and 2,5-bis(trimethylstannyl)-thieno[3,2-b]thiophene were purchased from Ossila. Tri(o-tolyl)phosphine was purchased from Tokyo Chemical Industry Co. Ltd. All other chemicals were purchased from Sigma-Aldrich and used without further purification. All reactions were carried out under a nitrogen atmosphere using standard Schlenk line techniques with dry solvents unless stated otherwise. ¹H NMR data were obtained using a Varian Oxford 300 spectrometer at 300 MHz, with chemical shifts referenced to tetramethylsilane. UV–Vis absorption and visible PL spectra were recorded on UV1900 (Shimadzu) and Fluoromax-4 (HORIBA), respectively. Fourier-transform infrared (FT-IR) spectroscopy was performed using a Shimadzu IRTracer-100 instrument. The number-average (M_n) and weight average (M_w) molecular weights, and the polydispersity index (PDI) of the polymer products were determined by high-temperature gel-permeation chromatography (HT-GPC) with Agilent 1200 HPLC and miniDAWN TREOS using polystyrene as the standard in trichlorobenzene at 80°C (HPLC grade). Thermal gravimetric analysis and differential scanning calorimetry were performed on a SDT 650 (TA Instrument) at a heating scan rate of 10°C/min under nitrogen atmosphere. X-ray photoelectron spectroscopy (XPS) and ultraviolet photoelectron spectroscopy (UPS) were performed for thin films on a Si substrate using an ESCALAB 250xi instrument (Thermo Fisher Scientific). Two-dimensional grazing incidence X-ray diffraction (2D GIXD) measurements were performed at the 3C and 9A beamlines at the Pohang Accelerator Laboratory, Korea. The topology and thickness of the films were analyzed using an atomic force microscope (NX10; Park System).

Synthesis of DPPTT-H

3,6-Bis(5-bromothiophen-2-yl)-2,5-bis(2-octyldodecyl)pyrrolo[3,4-c]pyrrole-1,4(2H,5H)-dione (0.234 g, 0.23 mmol), 2,5-bis(trimethylstannyl)-thieno[3,2-b]thiophene (0.107 g, 0.23 mmol), tris(dibenzylideneacetone)dipalladium (3.6 mg), tri(o-tolyl)phosphine (4.9 mg), and anhydrous 1-chlorobenzene (20 mL) were charged in a 25 mL flask. The mixture was heated to 130°C and stirred for 48 h, followed by the addition of 2-bromothiophene (0.5 mL) to react with the unreacted trimethylstannyl end groups for 8 h under stirring. Then, 2-(tributylstannyl)thiophene (0.38 mL) was added, and the mixture was stirred for 8 h under stirring. The reaction mixture was then cooled to ambient temperature and poured into 200 mL of methanol. The resulting solid was filtered, washed with methanol, and dried. The solid was further purified by Soxhlet extraction using acetone, methanol, dichloromethane, and n-hexane, and then dissolved in chloroform. After the extraction, CHCl₃ was collected and dried under reduced pressure. A blue solid was obtained after removing the solvent (0.219 g, 95.2%).

Synthesis of DPPTT-1F

The procedure before end-capping was the same as that used for DPPTT-H. After polymerization at 130°C for 48 h, bromopentafluorobenzene (0.15 mL) was added to react with unreacted trimethylstannyl end groups for 8 h under stirring. Then, 2-(tributylstannyl)thiophene (0.38 mL) was added, and the mixture was stirred for 8 h under stirring. The subsequent steps were the same as those used for DPPTT-H. After Soxhlet extraction and residual solvent removal, a blue solid was obtained (0.218 g, 94.8%).

Synthesis of DPPTT-2F

The procedure before end-capping was the same as that used for DPPTT-H. After polymerization at 130°C for 48 h, bromopentafluorobenzene (0.15 mL) was added and allowed to react with the unreacted trimethylstannyl end group for 8 h under stirring. Subsequently, tributyl(pentafluorophenyl)stannane (0.42 mL) was injected, and the mixture was stirred for 8 h under stirring. The subsequent steps were the same as those used for DPPTT-H. After Soxhlet extraction and residual solvent removal, a blue solid was obtained (0.220 g, 95.4%).

Synthesis of Low Molecular Weight DPPTT-2F

3,6-Bis(5-bromothiophen-2-yl)-2,5-bis(2-octyldodecyl)pyrrolo[3,4-c]pyrrole-1,4(2H,5H)-dione (0.234 g, 0.23 mmol), 2,5-bis(trimethylstannyl)-thieno[3,2-b]thiophene (0.107 g, 0.23 mmol), tris(dibenzylideneacetone)dipalladium (3.6 mg), tri(o-tolyl)phosphine (4.9 mg), and anhydrous 1-chlorobenzene (30 mL) were charged in a 50 mL flask. After polymerization at 130°C for 24 h, bromopentafluorobenzene (0.15 mL) was added and allowed to react with the unreacted trimethylstannyl end group for 8 h under stirring. Subsequently, tributyl(pentafluorophenyl)stannane (0.42 mL) was injected, and the mixture was stirred for 8 h under stirring. The subsequent steps were the same as those used for DPPTT-H. After Soxhlet extraction and residual solvent removal, a blue solid was obtained (0.166 g, 72.0%).

Synthesis of Cx-DPPTT Series

In a 6 mL pressure-resistant vial with a stirrer bar, the DPPTT series (0.010 mmol, 10 mg) and solvent (2 mL) were added, and the solution was gently heated and stirred until all the polymers had visibly dissolved, after which (3-mercaptopropyl)trimethoxysilane (1324 eq.), and K₂CO₃ (1780 eq.) were added. The mixture was further stirred at 80°C for 12 h before cooling down to room temperature and poured into 200 mL of stirring methanol. The solid was washed with acetone and hexane to remove the excess nucleophiles. A blue solid was obtained after removing the solvent (9.8 mg, 98.0%).

Synthesis of Modified DPP

In a 2-mL pressure-resistant vial with a stirrer bar, 3,6-bis(5-bromothiophen-2-yl)-2,5-bis(2-octyldodecyl)pyrrolo[3,4-c]pyrrole-1,4(2H,5H)-dione (1) (10.19 mg, 0.01 mmol) and tetrahydrofuran (3 mL) were added, and a solution was gently heated and stirred until all the solutes had visibly dissolved, followed by the addition of 1-octanethiol (10 eq, 0.1 mmol) and base (15 eq, 0.15 mmol). The reaction mixture was then sealed and stirred at 85°C for 12 h. After cooling to room temperature, the solvent was evaporated under reduced pressure. The crude product was purified by chromatography on silica with 0%–50% dichloromethane in hexane as eluent. Isolated dark purple solid was obtained with 24% yield. ¹H-NMR (300 MHz, CDCl₃) δ 8.86 (d, J = 3.9 Hz, 1H), 8.60 (d, J = 4.2 Hz, 1H), 7.21 (d, J = 4.2 Hz, 1H), 7.14 (d, J = 4.1 Hz, 1H), 3.94 (dd, J = 9.5, 8.1 Hz, 4H), 2.96 (t, J = 7.3x (2) Hz, 2H), 1.88 (m, 2H), 1.69 (q, J = 7.3x (4) Hz, 2H), 1.42 (m, 2H), 1.32–1.19 (m, 73H), 0.93–0.80 (m, 15H).

Preparation of Photopatterned Polymer Films

Glass substrates (2.5 cm × 2.5 cm) were prepared by sequential ultrasonication with acetone, isopropyl alcohol, and deionized water. The substrates were dried under a nitrogen stream and cleaned with ultraviolet ozone plasma for 30 min. Cx-DPPTT-H, Cx-DPPTT-1F, and Cx-DPPTT-2F were dissolved in 1-chlorobenzene at a concentration of 7 mg mL⁻¹ in a N₂ environment. The solutions were then transferred to ambient air for spin coating (1500 rpm for 60 s). The samples were then transferred back into a N₂ environment and annealed at 150°C for 20 min. Then, a positive photoresist (AZ GXR 601, AZ Electronic Materials) was spin-coated (3000 rpm for 30 s) onto the polymer film and annealed at 120°C for 60 s. UV light was illuminated on the photoresist with a photomask and developed using a developer (AZ MIF 300, AZ Electronic Materials). The area without the photoresist was etched using O₂ plasma (60 W for 50 s). The lift-off process was performed in acetone, leaving only a patterned polymer film on the substrate.

OFET Fabrication Based on SiO₂/p++ Si Wafer

A heavily p-doped Si wafer with 300 nm thick thermally grown SiO₂ (capacitance = 10.8 nF cm⁻²) was used as the substrate for the OFET of the bottom gate/top contact. The substrate surface was modified with an octadecyltrimethoxysilane (OTMS, Gelest Inc.) self-assembled monolayer according to the following steps. The Si wafers were cleaned with a piranha solution (7:3, H₂SO₄: H₂O₂) to functionalize the surface with hydroxyl groups and then rinsed with deionized water. The functionalized substrates were spin-coated (3000 rpm for 30 s) with OTMS solution in trichloroethylene (Sigma Aldrich, ≥ 99%) and stored under NH₃ atmosphere for 12 h at room temperature (~300 K). All the substrates were cleaned with toluene, acetone, isopropyl alcohol, and deionized water. The substrates were then transferred to an N₂ atmosphere for further processing. DPPTT-H, DPPTT-1F, DPPTT-2F, Cx-DPPTT-H, Cx-DPPTT-1F, and Cx-DPPTT-2F were dissolved in anhydrous chloroform (Sigma Aldrich, ≥ 99%) at a concentration of 7 mg mL⁻¹ and spin-coated (2000 rpm for 60 s) on the modified substrates. The devices were annealed at three optimal temperatures (as fabricated, 150°C, 200°C) for 20 min. Au (50 nm) was thermally evaporated (rate: 1 Å/s) on the polymer film through a shadow mask as source/drain electrodes (channel length: 100 μm, channel width: 1000 μm).

OFET Flexible Array Fabrication

2.5 cm × 2.5 cm glass plates were prepared as supporting plates during device fabrication and cleaned routinely by ultrasonication with acetone, isopropyl alcohol, and deionized water. Five-micrometer-thick parylene-c was deposited by the parylene coater (OBTPC 500, Obang Technology) and annealed in a vacuum oven at 120°C overnight to enhance the adhesion with glass plates. Photopatterned source/drain electrodes (Ti [3 nm] and Au [30 nm]) were then deposited. The substrate was cleaned using an O₂ plasma treatment before spin-coating the OSC. The Cx-DPPTT-2F film was coated and patterned as described for the photopatterned polymer films. Cytop (CTL-809 M; Asahi Glass Co.) diluted in Cytop solvent (CT-Solv.180; Asahi Glass Co.) (1:4 = CTL-809 M: CT-Solv.180, volume ratio) was spin-coated as the first layer of gate dielectric and dried in a vacuum oven at 50°C for 12 h. Parylene-C (300 nm) (the second layer of gated dielectric) was vacuum deposited by the parylene coater. Al (50 nm) was thermally evaporated onto the gate electrodes using a shadow mask. The complete device was then mechanically peeled off.

Photothermal Deflection Spectroscopy

A photothermal deflection spectroscopy (PDS) set-up was used to measure sub-bandgap absorptions. This technique is based on the heat energy released from the surface of the sample when monochromatic light is absorbed. An inert liquid surrounding the sample dissipates this thermal energy, changes its refractive index, and consequently deflects a laser beam that is sent at grazing incidence along the surface of the substrate. Using a quadrant detector connected to a lock-in amplifier, the deflection of the laser beam is recorded as a function of the monochromatic pump wavelength, resulting in a reading of absorbance.

Electrical Characterization

The electrical performances of the OTFTs were characterized using Keithley 4200-SCS under vacuum (≈ 10⁻² Torr). Field-effect mobility was calculated for the saturation region: I_DS = (W/2 L) μ_FET C_i (V_GS—V_TH)², where W and L are the channel width and length, respectively; and C_i is the capacitance per unit area of the dielectric layer.

Results and Discussion

The structures and synthetic scheme of the DPPTT series end-capped with pentafluorobenzene (PFB) and/or thiophene are shown in Figure 1a, and the overall synthetic routes are shown in Figure S1 in the Supplementary Information (SI). We depicted the pristine DPP monomer conformation as CH···O interacting structure since, in the thermodynamic aspect, it is slightly more stable than S···O interacting one [33]. However, when it comes to DPPTT thin-film fabricated by spin-coating, those two conformations can be mixed due to the abrupt solidification of the polymers, in which kinetic factors are more dominant than thermodynamics. In this article, DPPTT polymers are drawn as S···O interacting conformation, as common in other studies [34–36]. All polymers were synthesized via the Stille coupling reaction in 1-chlorobenzene. The polymers were then treated with MPS at 80°C for 12 h in the presence of K₂CO₃ as the base, followed by vacuum filtration. This procedure deprotonates MPS by K₂CO₃, leading to a substitution reaction of MPS with para position fluorine—para-fluoro-thiol reaction that is a regioselective click reaction to substitute thiol molecule for para-sited fluorine under base—and nucleophilic addition of MPS to amide of DPP moieties, resulting in Cx-DPPTT series with a yield above 95% (Figures S2, S3) [37–41]. The significance lies in the fact that yield calculations are confined to this specific modification reaction, not polymerization, because of the independence of the polymerization and crosslinking modification processes. The Cx-DPPTT series comprises three variations: Cx-DPPTT-H, which features an unaltered chain end; Cx-DPPTT-1F, in which one side of the chain end has been modified; and Cx-DPPTT-2F, with modifications at both ends of the chain. The detailed synthetic procedures are described in the Materials and Methods section. The synthesized Cx-DPPTT series exhibited good solubility in chlorinated organic solvents such as chloroform, 1-chlorobenzene, and o-dichlorobenzene (Figure S4). The number-average molecular weight ($$$ \overline{M_{\mathrm{n}}} $$$), weight average molecular weight ($$$ \overline{M_{\mathrm{w}}} $$$), and PDI of the polymer products were determined by HT-GPC using polystyrene as the standard in trichlorobenzene at 80°C, exhibiting high molecular weight over 160 kDa (Table 1). The minor decrease in the molecular weight after modification was attributed to the filtration of thermally crosslinked aggregates from the high temperature of the GPC. In comparison to cross-linkable polymers that have been previously studied, the modified Cx-DPPTT demonstrated a notably higher molecular weight, reaching 110.8 kDa for Cx-DPPTT-2F and attaining 99% yield, while maintaining a low PDI of 1.2. (Figure S5 and Table 1). Thermogravimetric analysis showed good thermal stability of all the modified polymers with 5% weight loss (T_d) near 370°C, which is the same as those of pristine DPPTT series (Figure S6). Differential scanning calorimetry measurements revealed no endothermic or exothermic behavior between 50°C and 300°C (Figure S7). UV–Vis spectra were measured to observe the optical properties of the synthesized polymers as thin films on a glass substrate. All synthesized polymers showed similar spectral characteristics, with broad absorption bands extending from 600 to 1000 nm, indicating intramolecular charge transfer by strong DA interactions (Figure S8) [42].

TABLE 1 Material properties of polymers. ^aE_HOMO was measured by UPS. ^bE_LUMO was obtained by the addition of the optical bandgap to HOMO level, where the optical bandgap was calculated from the UV–Vis spectra using the Tauc plot method.

	$$$ \overline{{\mathrm{M}}_{\mathrm{w}}}/\overline{M_{\mathrm{n}}} $$$[kDa]	Td [°C]	λ solmax [nm]	λ filmmax [nm]	Eg [eV]	EHOMOa [eV]	ELUMOb [eV]
DPPTT-H	164.4/123.9	398	807	822	1.32	−5.18	−3.86
Cx-DPPTT-H	160.9/121.3	372	817	822	1.34	−5.05	−3.71
DPPTT-1F	146.9/114.8	396	819	828	1.34	−5.18	−3.84
Cx-DPPTT-1F	140.4/95.4	386	813	816	1.36	−5.06	−3.70
DPPTT-2F	136.8/108.5	398	811	821	1.34	−5.18	−3.84
Cx-DPPTT-2F	110.8/92.1	371	810	822	1.36	−5.07	−3.71

To analyze the exact modification mechanism of the DPP moieties along the polymer chain, we modeled a monomer-scale reaction of DPP with a selected thiol molecule (Figure 2a). We used octanethiol instead of MPS as the modifying agent to avoid inevitable condensation with the silica gel in the chromatography column. Other modification conditions were identical to those used for polymer modification, resulting in modified DPP (mDPP). ¹H-NMR measurement was carried out to reveal how the octanethiol reacts with DPP. A distinct triplet peak at 2.98 ppm, quintet peak at 1.66 ppm, and septet peak at 1.44 ppm with all identical integral values in the chemical shift appeared, indicating that one equivalent octanethiol was successfully attached to DPP by means of nucleophilic addition (Figures 2b and S9). This result indicates that octanethiol was successfully attached to DPP. The chemical shifts of protons from the two thiophene moieties, which showed two doublet peaks, were split into four doublet peaks with half of integral value; this is caused by the conformational change to asymmetric structure. In particular, doublet peak at 3.94 ppm turned to doublet of doublets by the appearance of the doublet peak at 3.97 ppm, indicating a structural change near the alkyl chain. To verify the exact site and mechanism of attachment, we observed attenuated total reflectance-FT-IR spectroscopy (ATR-FTIR) spectra for pristine DPP and mDPP (Figure 2c). Pristine DPP exhibited a peak centered at 1671 cm⁻¹ that is associated with the CO stretching mode of the two amides in DPP. The modeled reaction, however, deconvolutes the peak into two peaks centered at 1656 cm⁻¹ and 1732 cm⁻¹. This deconvolution occurs because one of the two amides is modified into a different type of carbonyl group; thiolate addition to the carbonyl group deforms one of the amide structures, forming a thioester (Figure 2a). The FTIR spectra predicted from density functional theory (DFT) calculations supported the modified structure, indicating that DPP underwent thioesterification with a thiol (Figure S10). In addition, DFT-calculated electron density maps show that thioesterification occurs exclusively at one of the two amides (Figure S11). This selective modification arises because the DPP modification alters the charge distribution of the amide carbon atom, shifting it from positive to negative and thus preventing subsequent thioesterification of the second amide.

[IMAGE OMITTED. SEE PDF]

To investigate the effect of the modification reaction on the energy distribution, we calculated and experimentally measured the energy levels of the highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) for all polymer series (Table 1). For the dimers, the DFT calculation results showed a significant increase in the bandgap and LUMO level with an increase in the number of mDPP moieties (Figure 3a). This blueshift was responsible for the decrease in backbone planarity. Thioesterification allows strong hydrogen bonding of the carbonyl group from DPP with amine groups, suppressing the S···O interaction with thiophene—the well-known interaction that minimizes rotation angle between donor and acceptor moieties [43]. However, the energy distribution of Cx-DPPTT-H, in which HOMO level was experimentally measured from the UPS (Figures S12, SI) and LUMO level was calculated by Tauc plot method from UV–Vis spectra, showed a minor increase in the HOMO and LUMO levels compared to those of DPPTT-H (Table 1). These are significantly deviated from the DFT-predicted energy distribution (especially the LUMO level) of the fully modified DPPTT polymer (Figures 3b and S13). These results imply that only a few DPP moieties were modified. We performed further investigations to estimate the proportion of modified DPPs along the DPPTT chain using XPS with respect to the annealed polymer thin films (Figure S14). XPS data of DPPTT-H and Cx-DPPTT-H thin films showed almost identical spectra before and after the modification. However, a distinct noisy shoulder at around 398 eV representing the secondary amine bond occurs after modification (Figure 3c). The ratio of the area is about 2.8% over the entire area, which convinces that only around 2.8% of DPP moieties participated in the thioesterification reaction, and thus a slight increase in bandgap occurred. This result is comparable with the bandgap difference ratio between DPPTT-H (1.32 eV) to Cx-DPPTT-H (1.34 eV) and DPPTT-H (1.32 eV) to the fully modified Cx-DPPTT-H (2.41 eV): that is 1.8%. The small participation in thioesterification would be attributed to steric hindrance which increases with the length of the polymer backbone in terms of polymer bending by intrachain interaction in an organic solvent [44].

[IMAGE OMITTED. SEE PDF]

PDS absorption spectra of DPPTT and Cx-DPPTT-2F provide further insight on the electronic structure of the polymers (Figure S15). PDS accurately measures the absorption tail below the bandgap, which is used to quantify the energetic disorders in the polymer films by extracting the Urbach energy [45]. The two polymers showed nearly identical bandgap as the results obtained from UPS and XPS measurements. The difference between the Urbach energy of the polymers was negligible: i.e., 37 meV for DPPTT-2F and 36 meV for Cx-DPPTT-2F. These values are comparable to the previously reported Urbach energy of the DPP derivatives, providing strong evidence that the modification minimally affects the conjugation of the original polymer [46, 47].

Bottom-gate top-contact (BGTC) OFETs were fabricated on SAM-treated SiO₂ substrates to compare the electronic properties of the synthesized polymers before and after modification. All semiconductor films were spin-coated and annealed in a moisture-controlled N₂ environment. Notably, the materials exhibited well-behaved p-type switching characteristics before and after the modification reaction, and the on/off current ratios were comparable to those of conventional conjugated polymer devices (~10⁴) (Figure S16a–f) [48]. The extracted field-effect mobilities (μ_FETs) demonstrated that the charge transport characteristics of the polymers were consistent regardless of the modification, with the lowest average μ_FET of 0.131 (±0.1) cm² V⁻¹ s⁻¹ for pristine Cx-DPPTT-H (Figure S16g, Table S1). Among the modified polymers, the Cx-DPPTT-2F film annealed at 200°C showed the best performance with the average μ_FET of 0.202 (±0.08) cm² V⁻¹ s⁻¹. The results confirm that the modification provides crosslinking capabilities without altering the electrical characteristics.

Further analysis on the electrical characteristics of the modified Cx-DPPTT-H, Cx-DPPTT-1F, and Cx-DPPTT-2F was conducted with top-gate bottom-contact (TGBC) OFETs. Polymethyl methacrylate (PMMA) was used as the gate dielectric. The transfer characteristics showed ambipolar charge transport characteristics with both electron and hole conduction (Figure S17) with low hysteresis. The reduced barrier of electron injection by charge crowding at the S/D electrode and semiconductor interface in TGBC configuration revealed the ambipolar nature of the synthesized polymer [49]. High μ_FET,electron = 0.690 cm² V⁻¹ s⁻¹ and μ_FET,hole = 1.07 cm² V⁻¹ s⁻¹ of Cx-DPPTT-2F show the excellent charge transport capabilities of the synthesized polymers. The increase in the charge mobilities compared with the BGTC transistors is attributed to a decrease in the energetic disorder near the transport level, due to the low polarizability of PMMA gate dielectrics compared to SiO₂/OTMS [50]. Field effect mobilities are anticipated to increase significantly with appropriate processing techniques such as solution shearing and engineering of the semiconductor-dielectric interface [51].

To confirm that crosslinking via silanol condensation from MPS moieties works, DPPTT-2F and Cx-DPPTT-2F thin films were fabricated and soaked in 1-chlorobenzene. The degree of crosslinking was then quantified using UV–Vis absorption spectra by comparing the film absorbance before and after soaking the film in 1-chlorobenzene [52, 53]. Notably, the Cx-DPPTT-2F thin film showed resistance over 1-chlorobenzene (Figure 4a,b). This solvent resistance was maintained for a week (Figure S18a, b). Moreover, the polymers with low molecular weight showed similar result, validating the crosslinking phenomenon (Figure S18c). To further reveal the existence of Si-O-Si bond, Si2p XPS spectra of thin films on a Si substrate were investigated (Figure S19). Here, the obvious peak at around 103.5 eV, which is responsible for the Si-O-Si bond when MPS was self-condensed, was shown [54]. These results demonstrate that the crosslinking phenomenon of the Cx-DPPTT series originated from the silanol condensation.

[IMAGE OMITTED. SEE PDF]

The optimal crosslinking conditions for Cx-DPPTT were explored. We varied the experimental parameters for modification and film fabrication. The crosslinking of Cx-DPPTT films is a two-step process, with the initial step involving the hydrolysis of methoxy groups in MPS with the aid of water molecules to generate OH groups [55]. In the second step, these OH groups undergo condensation with other OH groups at the cross-linkable sites, forming a crosslinked polymer network (Figure 1b) [56]. Therefore, the density of cross-linkable sites and exposure to moisture in the film fabrication process are considered the most critical parameters determining the degree of crosslinking.

First, the density of the cross-linkable sites was controlled. As the MPS deprotonated by K₂CO₃ was attached to the end-capping PFB in the modification step, we could effectively control the density of the cross-linkable sites by varying the density of the end-capping PFB and the amount of base K₂CO₃ used in the modification [38]. With increasing PFB at the chain end, film retention increased from 39.8% for Cx-DPPTT-H to 80.9% for Cx-DPPTT-2F (Figure S20). Increasing the amount of K₂CO₃ during the modification process also contributed to a higher degree of crosslinking (Figure 4c). These results indicate that a high density of cross-linkable sites is crucial for obtaining a high degree of crosslinking.

Second, to study the relationship between the presence of water molecules and crosslinking, the level of exposure to moisture during the fabrication process was manipulated. As mentioned previously, moisture is crucial for crosslinking because it hydrolyzes the methoxy groups of the attached MPS to hydroxyl groups (Figure 4d). The film fabrication process was divided into three steps: (i) solution preparation, (ii) spin coating, and (iii) thermal annealing (Figure 4e). In each step, the exposure to moisture was precisely controlled. Both hydrous and anhydrous 1-chlorobenzene were used to regulate the solution preparation steps. In the spin-coating and thermal annealing steps, the surrounding atmosphere was varied by processing in either moisture-controlled N₂ or ambient atmosphere. Finally, we evaluated the films fabricated using the five different processing methods (Figure 4f). In processing method 2, the film was exposed to ambient conditions for the absorption of moisture and then returned to the N₂ atmosphere for thermal annealing. An increase in solvent resistance with an increase in exposure to moisture was evident; processing method 1 with no exposure to moisture showed 7.28% of film retention, whereas processing method 5 with all processes exposed to moisture showed 80.09%. To further investigate the effect of moisture on crosslinking, FT-IR spectra were measured to monitor the changes in the major functional groups (Si-OH, Si-OCH₃, and Si-O-Si) (Figure 4g). Increasing the exposure to moisture during the process decreased the intensities of the peaks corresponding to unreacted crosslinking sites, Si-OH stretching at 1040 cm⁻¹ and Si-OCH₃ at 1069 cm⁻¹. Especially, the Si-OCH₃ stretching peak disappeared for the films processed using methods 4 and 5. In addition, Si-O-Si stretching peak at 1108 cm⁻¹ increased with exposure to moisture, confirming that crosslinking could be effectively controlled.

The impact of moisture exposure on the electrical performance of the films fabricated using processing methods 1, 3, and 5 was evaluated. Devices fabricated with method 5 showed around 50% decrease in μ compared to those fabricated in method 1 for all Cx-DPPTT (Table S2). The reduction in μ_FET is attributed to the oxidation of polymers when exposed to moisture [57, 58]. On the other hand, method 3 exhibited no reduction in μ_FET, indicating that it is the optimal processing method for enhancing the degree of crosslinking without sacrificing electrical characteristics.

The impact of crosslinking on the crystalline nature and morphology was studied using grazing-incidence wide-angle X-ray scattering (GIWAXS) and atomic force microscopy (AFM) (Figure 5a). The (h00) peaks in the out-of-plane direction and the (010) peaks in the in-plane direction indicated that the edge-on orientation of the polymer crystals was consistent across all polymers (Figure S21). Polymers sharing the same end-capping groups exhibited analogous peaks in both the in-plane and out-of-plane directions (Figure 5b,c). The d-spacing of lamella packing and π-π stacking remained around the values of 20.5 Å and 3.8 Å, respectively, with negligible variations, for all the films. This confirms that the post-polymerization modification minimally affects the crystalline nature of the films (Table S3). We identified aggregated regions with sizes ranging from 10 to 20 nm in the crosslinked films (inset of Figure 5a). The GIWAXS patterns and AFM images before and after annealing indicated that these aggregates formed during the crosslinking process rather than during synthesis or film coating (Figure S22, Table S4). The density of aggregated regions was proportional to the density of crosslinked sites, with Cx-DPPTT-2F having the highest density of aggregated regions (Figure S23).

[IMAGE OMITTED. SEE PDF]

The achieved solvent resistance enabled the micropatterning of Cx-DPPTT films using conventional photolithography (Figure 6a). Without the issues of solvent-induced polymer dissolving or swelling during the patterning of polymer films, [59, 60] microscopic patterns with a critical dimension of 10 μm were successfully generated on the crosslinked Cx-DPPTT-2F film (Figure 6b) [33]. Furthermore, the AFM images verified that no topological damage was induced owing to its high solvent resistance (Figure 6c). These results demonstrate that our post-polymerization modification method significantly expands the range of available fabrication methods.

[IMAGE OMITTED. SEE PDF]

A 10 × 10 OFET flexible array on 5 μm thick parylene-c substrate was fabricated to demonstrate the feasibility of synthesized materials in the conventional photolithography process (Figure 7a). The Cx-DPPTT layer was fabricated using method 3 to minimize the degradation of the electrical characteristics while maintaining sufficient solvent resistance. The resulting transfer and output characteristics exhibited switching properties with an on/off current ratio > 10⁴ (Figure 7b,c). The statistic distribution of μ_FET, threshold voltage, and the distribution map of the calculated μ_FET are presented in Figure 7d,e. All devices fabricated operated well with the minimum μ_FET of 0.122 cm² V⁻¹ s⁻¹, indicating a 100% yield of 100 devices. A slight increase in average μ_FET (μ_{FET average} = 0.301 (±0.06) cm² V⁻¹ s⁻¹) was observed, compared to the devices fabricated on SiO₂/OTMS substrate with the same method (μ_{FET average} = 0.222 (±0.07) cm² V⁻¹ s⁻¹). This increase in μ_FET would originate from the low electron trap interface between the polymer semiconductor and CYTOP [61, 62].

[IMAGE OMITTED. SEE PDF]

Conclusion

We successfully introduced a post-polymerization modification strategy to synthesize intrinsically cross-linkable DPPTT. The para-fluoro-thiol reaction and thioesterification were used to modify both the backbone and chain end of the polymer. By attaching cross-linkable MPS to the polymer, high molecular weight could be achieved with a high yield of 95%. GIWAXS, AFM, and OFET analyses confirmed that the modification had a minor effect on the crystalline nature, film morphology, and electrical characteristics compared to the original polymer. Modified polymers can be efficiently crosslinked using moisture and heat. The degree of crosslinking increases under the following conditions: (i) increasing the density of the PFB end-capping reagent, (ii) increasing the amount of base used in the modification, and (iii) increasing exposure to moisture during the film fabrication process. The optimized polymer films showed high film retention of 80.09% after soaking in 1-chlorobenzene for 1 h. Finally, flexible OFET array was fabricated using conventional photolithography, with all devices operating with an average μ_FET of 0.301 cm² V⁻¹ s⁻¹, owing to the chemical resistance achieved. This result suggests that cross-linkable DPP-based polymers can be synthesized in high yields using click chemistry. We believe that our strategy can be extended to various DPP-based D-A copolymers with a simple prerequisite of PFB end-capping, eventually realizing various functionalities in high yields.

Author Contributions

Jaehoon Lee: conceptualization, data curation, formal analysis, writing – original draft. Seungju Kang: conceptualization, data curation, formal analysis, visualization. Eunsoo Lee: conceptualization, formal analysis, software, writing – original draft. Jiyun Lee: data curation, formal analysis, software. Tae Woong Yoon: data curation, formal analysis. Min-Jae Kim: methodology, validation. Yongjoon Cho: formal analysis. Mingfei Xiao: formal analysis. Yorrick Boeije: formal analysis. Wenjin Zhu: formal analysis. Changduk Yang: validation. Jin-Wook Lee: validation. Sungjoo Lee: methodology, validation. Guobing Zhang: supervision, validation. Henning Sirringhaus: supervision, validation. Boseok Kang: conceptualization, data curation, methodology, supervision, writing – review and editing.

Acknowledgments

Jaehoon Lee, Seungju Kang and Eunsoo Lee are contributed equally to this work. This work was supported by the International Research & Development Program of the National Research Foundation of Korea (NRF), funded by the Ministry of Science and ICT (2022K1A4A7A04094482), a National Research Foundation of Korea (NRF) grant funded by the Korean government (MSIT) (2023R1A2C1005015), and by the Nano & Material Technology Development Program through the National Research Foundation of Korea (NRF) funded by Ministry of Science and ICT (RS-2024-00403639). A portion of this research was conducted at the 3C SAXS-I and 9A U-SAXS beam.

Conflicts of Interest

The authors declare no conflicts of interest.

Data Availability Statement

Data will be made available on request.

References

M. Culebras, J. F. Serrano‐Claumarchirant, M. J. Sanchis, et al., “Conducting Pedot Nanoparticles: Controlling Colloidal Stability and Electrical Properties,” Journal of Physical Chemistry C 122, no. 33 (2018): 19197–19203, [DOI: https://dx.doi.org/10.1021/acs.jpcc.8b04981].