English

• Interfaces in organic electronics play a pivotal role in modulating the device performance, which has been fully verified in the past three decades [1-6]. Thereinto, the introduction of polymer dielectrics into organic devices offers more possibility of tuning the device interface and then optimizing the device performance due to designable monomeric structures and adjustable surface characteristics [7]. Despite remarkable progress in achieving efficient charge transport has been made on polymer dielectric, the reported studies mainly discussed the performance of a certain organic semiconductor (OSC) on either dielectric layer with variable interfaces [8-11], or a fixed dielectric layer [12-14], resulting in the unclear relationship between the dielectrics and chemical structures of semiconductors. Therefore, to further investigate the selectivity of specific dielectric layer to semiconducting molecules is necessary, which would help more effectively screen the dielectric and semiconductor materials with better compatibility for high-performance organic devices. As one of the earliest polymer dielectrics used in organic field-effect transistors, polyimide (PI) with satisfactory advantages of high/low temperature stability, solvent resistance and excellent stability [15-18], has exhibited promising applications in today's microelectronics [6,7]. However, it is worth noting that most of the devices based on bare PI dielectrics show lower mobility than thin-film amorphous silicon devices of 0.5–1 cm² V⁻¹ s⁻¹ [17,19-21], only a few cases of modified PI could obtain the mobility exceeding 1 cm² V⁻¹ s⁻¹ (Fig. S1 in Supporting information) [22-25]. On this basis, how to correctly evaluate the compatibility of bare PI with semiconductors and improve the device performance is becoming the key point.

  Here, three groups of organic small molecules with and without alkyl chains are used as active layers to vapor-deposit onto bare PI for constructing thin-film devices. It was found that bare PI had a molecule-selective interface, which preferred to select the semiconducting molecules with alkyl chains forming efficient conducting channel for high charge transport possibly due to the strong intermolecular interactions at the interface. As a result, the mobility of devices with bare PI and alkyl-chains-based molecular structure could reach 1.2 cm² V⁻¹ s⁻¹, which is one of the highest mobilities of thin film transistors based on bare PI dielectric reported in the existing literature (Fig. S1). Simultaneously, a low activation energy (E_A) of 3.01 meV was achieved, demonstrating a low degree of disorder present in the films of molecules with alkyl side chains on the surface of the PI dielectric. In addition, the better compatibility of bare PI with semiconductors could also promote the device with efficient charge transport at low temperature (e.g., 100 K) and obtain a relatively stable device at low temperature. The selectivity relationship between PI and semiconductors provides an instructive strategy for the application of PI dielectric in high-performance organic thin-film transistors at room and low temperatures.

  As shown in Fig. 1a, the OFET device with bottom-gate top-contact (BGTC) structure was used to study the molecule-screening interface of bare PI dielectric layer. Here, three groups of representative organic small molecules with and without alkyl chains were selected (Fig. 1a). Vacuum-deposited method is an effective approach to investigate the interface effect of the device due to the sensitivity of molecular growth to the device interface [5,13]. First of all, the surface morphology of the semiconductor layer on the PI dielectric layer was analyzed by atomic force microscopy (AFM). As shown in Fig. S2 (Supporting information), the roughness of interface based on organic semiconductors without alkyl side chains (e.g., DPA, DNTT and 2,6-DNA) was larger than those with alkyl side chains (e.g., C6DPA, C10DNTT, C10-2,6-DNA). Obviously, more ordered morphology of thin films was observed from deposited molecules with alkyl side chains on PI (Fig. S3 in Supporting information). As a result, it is clearly shown in Fig. 1b that the devices using semiconductors without alkyl side chains exhibited suppressed mobility of ~10⁻³ cm² V⁻¹ s⁻¹ due to the weak interaction between the PI strands and the semiconductors [26], but those semiconductors with alkyl side chains could construct efficient conducting channel on bare PI for up to three orders of magnitude higher performance possibly due to the strong interaction between the PI strands and the semiconductors. The representative transfer and output curves based on different semiconductors were shown in Figs. 1c–e and Fig. S4 (Supporting information), respectively. What is more, from the analysis of the transfer curve (Figs. 1c–e), not only the output current but also the on/off ratio has been improved by several orders of magnitude using organic semiconductors with alkyl side chains. In order to verify the uniqueness of molecule-screening PI interface, bi-polymer dielectrics on the basis of polystyrene (PS) and polyphenylene ether (PPO) as the buffer layers to modify the PI were used to prepare the OFET device (Fig. 2a). It was obvious that the introduction of PPO and PS buffer layers could enhance the device mobility based on some small molecules, but the existence of PPO and PS would change the screening characteristics of molecules at the interface (Figs. 2b and c, Figs. S5 and S6 in Supporting information). These results could further suggest that the apparent difference in the device mobility based on bare PI was attributable to the selective interaction between PI and semiconductor molecules.

  Figure 1

  

  Figure 1.  (a) Schematic diagram of organic thin-film FETs with BGTC and different semiconductor molecular structures. (b) Mobility distribution of transistors based on different semiconductors and bare PI. The typical transfer curves of transistors based on bare PI dielectrics using (c) DPA and C6DPA, (d) DNTT and C10DNTT, (e) 2,6-DNA and C10–2,6-DNA as active layers.

  DownLoad: Full-Size Img PowerPoint

  Figure 2

  

  Figure 2.  (a) Schematic diagram of the device with buffer layers. Mobility distribution of transistors based on different semiconductors and (b) PPO/PI and (c) PS/PI. (d) Schematic diagram of organic single crystal transistor devices constructed by transferring pre-grown organic single crystals and gold electrodes, respectively. Transfer curves of devices with pre-grown (e) DPA, (f) C6DPA and (g) DNTT single crystals transferred onto PI dielectric, respectively. Inset: optical images of the fabricated single crystal OFETs. Scale bar: 90 µm.

  DownLoad: Full-Size Img PowerPoint

  Subsequently, we transferred organic single crystals with order molecular stacking on bare PI surface to further investigate its interface effect. The details about the fabrication of single crystal are shown in the experimental section in Supporting information [27]. Here, DPA, C6DPA and DNTT were chosen as examples to prepare the single-crystal OFETs (Fig. 2d). In this kind of device, the source/drain gold electrodes were constructed by transferring the gold film by mechanical probes [28]. The typical transfer and output curves of single-crystal devices were shown in Figs. 2e and f, and Figs. S7a and b (Supporting information), respectively, compared to C6DPA device (μ = 1.06 cm² V⁻¹ s⁻¹), DPA device (μ = 4.08 cm² V⁻¹ s⁻¹) exhibited even better charge transport performance. In addition, DNTT single crystal devices also had higher mobility (μ = 1.71 cm² V⁻¹ s⁻¹) compared with thin-film devices (Fig. 2g and Fig. S7c in Supporting information). These results showed that the PI interface could also produce efficient charge transport based on organic semiconductors with order molecular stacking (e.g., single crystal), but this interface had an influence on the directly vacuum-deposited growth of molecules with different chemical structures.

  The E_A calculated from temperature-variable experiments is an important factor to assess the degree of disorder presented at the interface [29,30]. Based on the low-temperature stability of PI dielectric layer [17,29], the devices using different semiconductors were carried out temperature-variable electrical tests (Fig. S8 in Supporting information). The temperature-dependent mobility curves showed the typical hopping mode in all kinds of devices because of the decreasing values of mobility with the decrease of the temperatures (Fig. S9 in Supporting information). According to the μ = μ₀·exp(-E_A/k_BT) [29], where μ is carrier mobility at different temperatures, E_A is activation energy, k_B is Boltzmann constant, and T is thermodynamic temperature. The calculated E_A consistently indicated that the values of devices using organic semiconductors with alkyl side chains (C6DPA, 11.20 meV, Fig. 3a; C10DNTT, 3.01 meV, Fig. 3b), were significantly lower than that using organic semiconductors without alkyl side chains (DPA, 24.15 meV, Fig. S10 in Supporting information; DNTT, 40.4 meV [31]). The calculated value of E_A could clearly indicate that the organic semiconductors with alkyl chains preferred to form high-quality packing of semiconductor molecules and a lower degree of interface energy disorder at the interface of PI. It was worth noting that our fabricated devices exhibited one of the lowest values of E_A (3.01 meV) compared with the previous reports [11,30,32-39] as shown in Table S1 (Supporting information), which demonstrated that our devices could still maintain efficient charge transfer, even at low temperatures. For example, the device based on C10DNTT on bare PI dielectric showed the mobility of 1.2 cm² V⁻¹ s⁻¹ at 300 K, and then, the decrease of temperatures only reduced the mobility by 25%, maintaining 0.9 cm² V⁻¹ s⁻¹ at 100 K (Fig. S9b in Supporting information). More excitingly, the device using C10DNTT and PS/PI dielectric could still maintain the mobility about 1.0 cm² V⁻¹ s⁻¹ at 100 K (Fig. S11 in Supporting information). This efficient charge transport performance at low temperature indicated the excellent compatibility of the low temperature resistance of PI and high-quality stacking of C10DNTT for further low-temperature applications. In addition, we also found that PI-based devices all showed atmospheric stability after one year (Fig. 4a). Moreover, the efficient charge transport would contribute to the high-performance optical figures of merit in organic phototransistors (OPTs) [35,40]. We fabricated the thin film OPTs using bare PI dielectric and the semiconductors with alkyl side chains and high mobility (e.g., C6DPA, C10DNTT). The best value of calculated optical figures of merit, including photosensitivity (P), photoresponsivity (R) and detectivity (D^*), were plotted in Figs. 4b and c. It was clearly seen that the OPTs using bare PI dielectric and the semiconductors with alkyl side chains showed higher performance than that with bare PI dielectric and alkyl-side-chains-free semiconductors [31]. Furthermore, the switching performance based on semiconductors with alkyl side chains was also higher than that devices using semiconductors without alkyl side chains, as shown in Figs. S12-S14 (Supporting information), respectively.

  Figure 3

  

  Figure 3.  Fitting results of low temperature mobility and temperature according to Arrhenius formula. (a) C6DPA/bare PI and C6DPA/PS/PI, (b) C10DNTT/bare PI and C10DNTT/PS/PI.

  DownLoad: Full-Size Img PowerPoint

  Figure 4

  

  Figure 4.  (a) The stability performance of device mobility based on bare PI and different semiconductors. The best value of photosensitivity (P), photoresponsivity (R) and detectivity (D^*) using (b) DPA and C6DPA; (c) DNTT and C10DNTT.

  DownLoad: Full-Size Img PowerPoint

  Although the potential application of polyimide dielectric in organic field-effect transistors has been verified, the device performance based on polyimide dielectric still lagged behind other dielectric materials. In this work, we have found there is interfacial molecular screening of polyimide dielectric on semiconductors in organic field-effect transistors. On this basis, the molecules without alkyl side chains are difficult to form an ordered layered stack during vacuum deposition, leading to suppressed electrical properties. Comparatively, the polyimide dielectric layer could induce the semiconductor molecules with alkyl side chains with order packing, resulting in excellent electrical properties. The compatibility between the polyimide and semiconductor with alkyl side chains promotes the combined devices maintaining excellent performance even at low temperature. Our research showed that the compatibility of the gate dielectrics with semiconductors is vital for constructing efficient conducting channel for high charge transport, and is also the basis for exploiting more complex device applications.

  Acknowledgments

  The authors are grateful to acknowledge financial support from National Key Research and Development Program (Nos. 2021YFA0717900, 2022YFE0124200), National Natural Science Foundation of China (Nos. 62004138, 52273190, 61905121, U2241221), and Haihe Laboratory of Sustainable Chemical Transformations.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2023.108796.

  
  
• 1. [1]

     M.Y. Sun, C.C. Zhang, D. Chen, et al., SmartMat 2 (2021) 213. doi: 10.1002/smm2.1030

  2. [2]

     A. Tsumura, H. Koezuka, T. Ando, Appl. Phys. Lett. 49 (1986) 1210–1212. doi: 10.1063/1.97417

  3. [3]

     Y. Yao, Y. Chen, H. Wang, P. Samorì, SmartMat 1 (2020) e1009.

  4. [4]

     C. Wang, X.T. Zhang, W.P. Hu, Chem. Soc. Rev. 49 (2020) 653–670. doi: 10.1039/C9CS00431A

  5. [5]

     H.L. Chen, W.N. Zhang, M.L. Li, G. He, X.F. Guo, Chem. Rev. 120 (2020) 2879–2949. doi: 10.1021/acs.chemrev.9b00532

  6. [6]

     D.Y. Ji, L.Q. Li, H. Fuchs, W.P. Hu, Acc. Mater. Res. 2 (2021) 159–169. doi: 10.1021/accountsmr.0c00112

  7. [7]

     D.Y. Ji, T. Li, W.P. Hu, H. Fuchs, Adv. Mater. 31 (2019) 1806070. doi: 10.1002/adma.201806070

  8. [8]

     X.Q. Jiang, C.F. Sun, C.C. Zhang, S.S. Cong, Chin. Chem. Lett. 32 (2021) 3364–3367. doi: 10.1016/j.cclet.2021.04.005

  9. [9]

     P.C. Gu, M.H. Hu, S. Ding, et al., Chin. Chem. Lett. 29 (2018) 1675–1680. doi: 10.1016/j.cclet.2018.03.034

  10. [10]

      B.H. Lee, B.B.Y. Hsu, S.N. Patel, et al., Nano Lett. 16 (2016) 314–319. doi: 10.1021/acs.nanolett.5b03868

  11. [11]

      B. Kang, N. Park, J. Lee, et al., Chem. Mater. 27 (2015) 4669–4676. doi: 10.1021/acs.chemmater.5b01207

  12. [12]

      X. Zhang, T. Li, Chin. Chem. Lett. 28 (2017) 2058–2064. doi: 10.1016/j.cclet.2017.09.008

  13. [13]

      T. Yokota, T. Kajitani, R. Shidachi, et al., Nat. Nanotech. 13 (2018) 139–144. doi: 10.1038/s41565-017-0018-6

  14. [14]

      N. Seiki, Y. Shoji, T. Kajitani, et al., Science 348 (2015) 1122–1126. doi: 10.1126/science.aab1391

  15. [15]

      T. Sekitani, Y. Kato, S. Iba, et al., Appl. Phys. Lett. 86 (2005) 073511. doi: 10.1063/1.1868868

  16. [16]

      S. Pyo, H. Son, K.Y. Choi, M.H. Yi, S.K. Hong, Appl. Phys. Lett. 86 (2005) 133508. doi: 10.1063/1.1894587

  17. [17]

      D.Y. Ji, L. Jiang, X.Z. Cai, et al., Ore. Electron. 14 (2013) 2528–2533. doi: 10.1016/j.orgel.2013.06.028

  18. [18]

      D.Y. Ji, Y.D. Wang, L.F. Chi, H. Fuchs, Adv. Funct. Mater. 25 (2015) 3855–3859. doi: 10.1002/adfm.201500771

  19. [19]

      Y.Z. Li, D.Y. Ji, H.L. Dong, J.Z. Li, W.P. Hu, RSC Adv. 5 (2015) 88059–88062. doi: 10.1039/C5RA17382E

  20. [20]

      T. Someya, T. Sekltani, S. Iba, et al., Proc. Natl. Acad. Sci. U. S. A. 101 (2004) 9966. doi: 10.1073/pnas.0401918101

  21. [21]

      L.H. Chen, P. Lin, J.C. Ho, et al., Synth. Met. 161 (2011) 1527–1531. doi: 10.1016/j.synthmet.2011.05.002

  22. [22]

      D.Y. Ji, L. Jiang, H.L. Dong, et al., J. Mater. Chem. C 2 (2014) 4142–4146. doi: 10.1039/c4tc00119b

  23. [23]

      X.Z. Cai, D.Y. Ji, L. Jiang, et al., Appl. Phys. Lett. 104 (2014) 063305. doi: 10.1063/1.4865239

  24. [24]

      B. Yonghwa, L. Sooman, Y.E. Joo, et al., ACS Appl. Mater. Interfaces 6 (2014) 15209. doi: 10.1021/am5035076

  25. [25]

      Y. Kato, S. Iba, R. Teramoto, et al., Appl. Phys. Lett. 84 (2004) 3789–3791. doi: 10.1063/1.1739508

  26. [26]

      D.Y. Ji, X.M. Xu, L.F. Jiang, et al., J. Am. Chem. Soc. 139 (2017) 2734–2740. doi: 10.1021/jacs.6b12153

  27. [27]

      S.J. Zhou, Q.X. Tang, H.K. Tian, et al., ACS Appl. Mater. Interfaces 10 (2018) 15943–15951. doi: 10.1021/acsami.8b02304

  28. [28]

      P. Chen, T.L. Atallah, Z.Y. Lin, et al., Nature 599 (2021) 404–410. doi: 10.1038/s41586-021-03949-7

  29. [29]

      Y.S. Zheng, L. Yu, Z.W. Wang, et al., Adv. Mater. Interfaces 8 (2021) 2100217. doi: 10.1002/admi.202100217

  30. [30]

      Y.C. Mei, P.J. Diemer, M.R. Niazi, et al., Proc. Natl. Acad. Sci. U. S. A. 33 (2017) E6739–E6748.

  31. [31]

      H.C. Li, T. Jiang, Y.S. Zheng, et al., Adv. Opt. Mater. 10 (2022) 2200614. doi: 10.1002/adom.202200614

  32. [32]

      S. Wang, M.J. Ha, M. Manno, C.D. Frisbie, C. Leighton, Nat. Commun. 3 (2012) 1210. doi: 10.1038/ncomms2213

  33. [33]

      S. Schott, E. Gann, L. Thomsen, et al., Adv. Mater. 27 (2015) 7356–7364. doi: 10.1002/adma.201502437

  34. [34]

      T. Sakanoue, H. Sirringhaus, Nat. Mater. 9 (2010) 736–740. doi: 10.1038/nmat2825

  35. [35]

      D.Y. Ji, T. Li, J. Liu, et al., Nat. Commun. 10 (2019) 12. doi: 10.1038/s41467-018-07943-y

  36. [36]

      Y.S. Zheng, H.Z. Lin, T. Jiang, et al., Adv. Funct. Mater. 33 (2023) 2211742. doi: 10.1002/adfm.202211742

  37. [37]

      Y.J. Shi, L. Jiang, J. Liu, et al., Nat. Commun. 9 (2018) 2933. doi: 10.1038/s41467-018-05390-3

  38. [38]

      T. He, Y.F. Wu, G. Avino, et al., Nat. Commun. 9 (2018) 2141. doi: 10.1038/s41467-018-04479-z

  39. [39]

      I.N. Hulea, S. Fratini, H. Xie, et al., Nat. Mater. 5 (2006) 982–986. doi: 10.1038/nmat1774

  40. [40]

      K. Baeg, M. Binda, D. Natali, M. Caironi, Y. Noh, Adv. Mater. 25 (2013) 4267–4295. doi: 10.1002/adma.201204979

• 

  Figure 1  (a) Schematic diagram of organic thin-film FETs with BGTC and different semiconductor molecular structures. (b) Mobility distribution of transistors based on different semiconductors and bare PI. The typical transfer curves of transistors based on bare PI dielectrics using (c) DPA and C6DPA, (d) DNTT and C10DNTT, (e) 2,6-DNA and C10–2,6-DNA as active layers.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  (a) Schematic diagram of the device with buffer layers. Mobility distribution of transistors based on different semiconductors and (b) PPO/PI and (c) PS/PI. (d) Schematic diagram of organic single crystal transistor devices constructed by transferring pre-grown organic single crystals and gold electrodes, respectively. Transfer curves of devices with pre-grown (e) DPA, (f) C6DPA and (g) DNTT single crystals transferred onto PI dielectric, respectively. Inset: optical images of the fabricated single crystal OFETs. Scale bar: 90 µm.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Fitting results of low temperature mobility and temperature according to Arrhenius formula. (a) C6DPA/bare PI and C6DPA/PS/PI, (b) C10DNTT/bare PI and C10DNTT/PS/PI.

  下载: 全尺寸图片 幻灯片
  

  Figure 4  (a) The stability performance of device mobility based on bare PI and different semiconductors. The best value of photosensitivity (P), photoresponsivity (R) and detectivity (D^*) using (b) DPA and C6DPA; (c) DNTT and C10DNTT.

  下载: 全尺寸图片 幻灯片
•