Enhancing the performance of organic phototransistors using a sandwich-heterostructure†

Chi Yan‡ ^a, Qi Wang‡ ^a, Weijie Gong ^a, Jie Lu ^a, Yao Yin ^a, Chuan Xiang ^a, Di Xue *^a, Zi Wang ^b, Lizhen Huang *^a and Lifeng Chi *^a
aInstitute of Functional Nano & Soft Materials (FUNSOM), Jiangsu Key Laboratory for Carbon-Based Functional Materials & Devices, Soochow University, 199 Ren’ai Road, Suzhou, 215123, Jiangsu, P. R. China. E-mail: chilf@suda.edu.cn; lzhuang@suda.edu.cn; dixue0130@suda.edu.cn
^bSuzhou Laboratory, Suzhou 215123, China

First published on 15th August 2024

---

---

Organic phototransistors (OPTs), combining the advantages of both light susceptibility and electronic signal amplification functions, are a class of three-terminal optoelectronic devices that hold significant potential for high-performance photodetection equipment, flexible and wearable electronics, and artificial vision systems.^1–3 Generally, a high-performance phototransistor with the corresponding light absorption ability usually requires efficient exciton dissociation to separate photoexcited carriers, good charge transport to generate a high photocurrent, and a low dark current to deliver high detectivity. In the design of high-performance organic thin-film phototransistors, introducing an organic heterostructure configuration is a promising strategy, which not only facilitates efficient exciton dissociation efficiency but also possibly prolongs the minority carrier lifetime via trapping to generate a high photogain.^4,5 However, the performance of organic heterostructure phototransistors is still far from satisfactory, partially due to the complex impacts on charge transport, the dark current and contact resistance issues originating from the organic–organic (O–O) interface.^6,7 The rough interface and high defect density of traditional planar polycrystalline film-based heterojunctions will lead to unfavourable transport of charge carriers, scattering of photogenerated excitons, and low charge-transfer efficiency. These limitations lead to a low channel current and poor photoresponse in the currently reported thin-film heterostructure OPTs. Furthermore, the existence of the interface charge transfer effect or charge trapping ability at the heterostructure often leads to a high dark current, greatly limiting the detectivity and photosensitivity of the OPTs. Therefore, it is highly desirable to create high-performance thin-film heterostructure OPTs that can simultaneously achieve low dark current and superior photoresponse under illumination.

Herein, we report a facile strategy to modulate both the dark-current and photoresponse of thin-film heterostructure OPTs. By introducing a sandwich heterostructure where an opposite polarity organic semiconductor is deposited between the bottom transport layers and the source–drain electrodes, a significant increment in mobility and photoresponse, as well as a decrease in dark current can be realized. In addition, compared to the traditional planar heterostructure OPTs, the devices based on a sandwich heterostructure not only retain the efficient processes of charge transfer, but also greatly reduce the adverse effects including the internal noise and dark current. This facile and general construction approach offers a new route for developing high-performance photoresponse devices.

High mobility implies efficient charge carrier transport and collection properties, which typically contribute to the high photogain and excellent performance of the OPTs.⁸ Molecular structures of the organic material used are illustrated in Fig. 1a. 2,7-Dioctyl[1]-benzothieno[3,2-b][1]benzothiophene (C8-BTBT), as a typical high-mobility semiconductor, is adopted as the transport layer. The dicyanovinyl terthiophene derivative (DCV3T) is a highly absorbing terthiophene derivative and bears the strong electron withdrawing character of the dicyanovinyl end groups, thus increasing both the ionization energy and electron affinity.^9,10Fig. 1b depicts the configuration of the OPTs, where highly ordered C8-BTBT thin films are deposited on the SiO₂/Si substrate as the donor layers, and DCV3T films were positioned between the source–drain electrodes and C8-BTBT films. Most importantly, the DCV3T films partially covered the channel and confined solely to the region beneath the source–drain electrode, thus forming a sandwich-like heterojunction thin film structure C8-BTBT/DCV3T-sandwich.¹¹ The morphology and molecular stacking of the single-component C8-BTBT thin film and the heterojunction thin film were studied using atomic force microscopy (AFM) and X-ray diffraction (XRD). AFM height images (Fig. 1c and d) show that the C8-BTBT thin film grown on the SiO₂ surface exhibits terraced-like crystalline morphology, with good continuity and large grain sizes. As shown in Fig. 1d, DCV3T films preferentially grow around the grain boundaries of C8-BTBT thin films. Additionally, the XRD patterns in Fig. 1e confirm the high crystallinity of both thin films, in accordance with the results of AFM.⁸

The performance of a series of OPTs under dark conditions was evaluated in advance, where for comparison, we prepared planar heterostructure C8-BTBT/DCV3T OPTs with full coverage of DCV3T thin films as shown in Fig. S1 (ESI†). The transport characteristics are summarized in Fig. 2a and c and Fig. S2 (ESI†), and all devices exhibit typical p-channel field-effect behavior, but their threshold voltage (V_th) and source–drain current (I_DS) show significant differences. From the transfer curves, it is observed that after the addition of DCV3T, both types of heterojunction transistors exhibit a significant increase in the I_DS, but the changes in their off-state currents are considerably dissimilar. The C8-BTBT/DCV3T planar heterostructure OPTs show a significant increase in the off-state current from 10⁻¹² A to 10⁻¹⁰ A. This distinction apparently points to the strong electron-withdrawing ability of the dicyanovinyl end-group-based DCV3T. This property promotes electron transfer to the DCV3T layer, while holes accumulate in the C8-BTBT layer, resulting in a notable rise in channel conductivity and elevated off-state current levels of the devices.⁴ However, for C8-BTBT/DCV3T-sandwich heterostructure OPTs, the off-state current display a value similar to that of C8-BTBT OPTs, indicating that the introduction of a sandwich-like structure can effectively reduce the off-state current of heterojunction-based transistors. To quantitatively compare the performance differences, transistor parameters were extracted based on the saturation region equations, and the distribution of hole mobilities of the series of OPTs is shown in Fig. 2c and Fig. S2 (ESI†). Both types of OPTs significantly increased, where the mobilities of C8-BTBT/DCV3T-sandwich heterostructure OPTs are up to 4.5 cm² V⁻¹ s⁻¹, superior to those of C8-BTBT/DCV3T planar heterostructure OPTs.

The 1/f noise of the device is primarily present at low frequencies, stemming from the fluctuations in electronic states induced by defects and disorder. And the internal noise in the three OTFTs devices was evaluated at V_G = 0 V and f = 1 Hz frequency, as shown in Fig. 2d and Fig. S3 (ESI†).¹² The C8-BTBT/DCV3T planar heterostructure OPTs exhibit a significantly higher noise, which should be attributed to the interface charge effect and traps at the traditional bilayer planar heterojunction structure, increasing the probability of carrier scattering.¹³ On the other hand, the introduction of a sandwich heterostructure strategy significantly ameliorates this deficiency. Additionally, all devices exhibit small hysteresis behavior, as shown in Fig. S4 (ESI†), with C8-BTBT/DCV3T planar heterostructure OPTs exhibiting the largest hysteresis window (ΔW) of approximately 2.93 V. The hysteresis of devices based on sandwich-like structures is suppressed, in agreement with noise results. To evaluate the transition rate between the on and off states of devices and the interface trap density, the subthreshold swing (SS) and the trap density (N_trap) are calculated, as shown in Table S1 (ESI†).¹⁴ C8-BTBT/DCV3T planar heterostructure OPTs show a significant increase in SS and N_trap, which is attributed to the increased contact areas for C8-BTBT/DCV3T planar heterostructure OPTs, leading to more interface defects or traps. In addition, a larger SS implies a large number of trap states in the carrier transport channel, thereby slowing down the device's on–off transition rate.¹⁵ In contrast, the change in SS and N_trap of C8-BTBT/DCV3T-sandwich heterostructure OTFTs show a slight change and even can be ignored, indicating that the sandwich-like structure effectively reduces the introduction of interface defects or traps by reducing the effective contact area between C8-BTBT and DCV3T.

Before delving into the investigation of its optoelectronic transistor performance, we compared the absorption spectra of this series of thin films, as shown in Fig. S5 (ESI†). The C8-BTBT film shows distinct absorption in the range of 320–400 nm, with a strong absorption peak around 360 nm. The absorption spectra of the bilayer heterojunction film are close to the superposition of the C8-BTBT and DCV3T films.¹⁶ Therefore, we chose light at a wavelength of 365 nm as the incident light, and studied the photoresponse characteristics of the series of OPTs under different light intensities as shown in Fig. 3a and b and Fig. S6a (ESI†). With light intensity increasing, the I_DS significantly increased, and the V_th of all OPTs shifted towards the positive direction, indicating the generation of photo-generated charge carriers in the channel.^17,18 To accurately evaluate the photoresponse performance of the three types of OPTs, we extracted photosensitivity (P), responsivity (R), and detectivity (D*) as key performance parameters based on the respective equations (see the ESI†). For C8-BTBT OPTs, the maximum P is 3.90 × 10⁸, while after the introduction of DCV3T, there is a significant difference in photoresponse properties as shown in Fig. S6b (ESI†). The P value of C8-BTBT/DCV3T-sandwich heterostructure OPTs experienced a slight increase, 5.23 × 10⁸, whereas for C8-BTBT/DCV3T planar heterostructure OPTs, a significant decline was observed, dropping to 1.27 × 10⁶. The difference in D* between the two heterostructures, as observed from Fig. 3c, is consistent with that in P. D* is represented by responsivity and noise, including the grain noise caused by the dark current, dielectric layer noise, Johnson noise, etc. To accurately calculate D*, the relationship D* = RA^1/2(i_n)⁻¹ is used, and the noise power density is measured as shown in Fig. S5 (ESI†). The higher noise current in the C8-BTBT/DCV3T planar heterostructures results in lower D*, only achieving 10¹⁶ jones, whereas the C8-BTBT/DCV3T-sandwich heterostructures, with their lower noise current, exhibit a significant increase in D*, reaching 10¹⁷ jones. Meanwhile, the sandwich heterostructure OPTs exhibit higher R under different illumination power densities compared to the traditional planar heterostructure OPTs as shown in Fig. 3d. Based on the previous discussion, we speculate that the use of a sandwich structure has reduced the contact area between DCV3T and C8-BTBT, reducing the impact of charge transfer on lateral transport and the possibility of exciton scattering, while enhancing the charge injection effect. Moreover, the optical figures of merit for C8-BTBT/DCV3T-sandwich heterostructure OPTs are better than those of most previously reported organic thin film phototransistors, as shown in Table S2 (ESI†).

To further investigate the reasons for the performance improvement achieved by the sandwich heterostructures, the photocurrent response curves of the heterostructure OPTs were tested, and the decay time constant (τ) was calculated as shown in Fig. 3e and f.¹⁹ The I_DS of heterostructure OPTs does not return to the initial state of low dark current within a certain time after light removal, but rather, after a slight decrease, it remains relatively constant at a higher current level, exhibiting the typical persistent photocurrent (PPC) phenomenon.²⁰ By fitting the response curves after light removal with a single exponential decay, it is found that the τ for C8-BTBT/DCV3T planner heterostructure OPTs is 182.93, while for C8-BTBT/DCV3T-sandwich heterostructure OPTs, τ reaches 394.50. The τ is associated with the recombination of photo-generated carriers, and C8-BTBT/DCV3T planar heterostructure OPTs have a small decay time constant. More trap states have been introduced for planar heterostructures, which might capture both electrons and holes, acting as recombination centers and resulting in a rapid decay of photogenerated carriers. After using the sandwich heterostructure, the contact area between C8-BTBT and DCV3T decreased, greatly improving this situation.

In order to deeply understand the correlation between the electronic structure and photoresponse behavior of the entire interface between C8-BTBT and DCV3T at the thin film level, we investigated the photoluminescence (PL) of single-component C8-BTBT films and C8-BTBT/DCV3T heterojunction films. As shown in Fig. 4a, the C8-BTBT film exhibits high-intensity fluorescence in the 350–400 nm range, while the fluorescence intensity of the heterojunction film with DCV3T added to the upper layer significantly decreases, showing an obvious fluorescence quenching phenomenon, indicating charge transfer or exciton dissociation occurring in the heterojunction film after illumination, which benefits the improvement of the device's electrical performance.¹⁸ This is also confirmed by the significant decrement in the exciton lifetime observed on the C8-BTBT/DCV3T heterostructure films (Fig. S7, ESI†).

Furthermore, ultraviolet photoelectron spectroscopy (UPS) and X-ray photoelectron spectroscopy (XPS) studies were conducted using highly ordered pyrolytic graphite (HOPG) as the substrate to investigate the electronic structure evolution during the deposition of different amounts of DCV3T on C8-BTBT. The HOMO region and secondary electron cutoff edge given by the UPS spectrum of C8-BTBT/DCV3T heterostructures with different thicknesses of DCV3T relative to the Fermi level were used to determine the vacuum level (VL).²¹ As shown in Fig. 4b, as the thickness of DCV3T increases from 0 to 100 Å, VL gradually decreases from 4.14 eV to 3.88 eV. Additionally, during the deposition of DCV3T, sulfur (S) signals derived from DCV3T appear, as shown in Fig. 4c. The S 2p signal derived from C8-BTBT (green curve) shows a noticeable shift towards higher binding energy (∼0.3 eV), while the signal from DCV3T molecules shifts slightly towards higher binding energy (∼0.1 eV) (red curve). Considering the relatively weak van der Waals forces at the molecular level, electron transfer occurs from the N atom in DCV3T to the S atom in C8-BTBT. This clearly confirms the occurrence of charge transfer between C8-BTBT and DCV3T. Therefore, the sandwich-structure with a discontinuous DCV3T layer can reduce the negative impact from the strong charge transfer in the dark while maintain the efficient charge injection and transport in transistors, leading to high photoresponse.

In conclusion, we have proposed a straightforward strategy by introducing sandwich heterostructures, thereby constructing high-performance UV OPTs. Compared to the traditional C8-BTBT/DCV3T planar heterostructure OPTs, the sandwich heterostructures have significantly improved the mobility and photoresponsivity due to the reduction in carrier scattering probability and the decrease in interfacial defects. It also effectively mitigates issues such as increased internal noise and excessive dark current in planar heterojunction transistors. This approach offers a new perspective for the design of high-performance optoelectronic devices.

The authors acknowledge financial support from the National Natural Science Foundation of China (Grant No. 22222205, 52173176) and the Suzhou Key Laboratory of Surface and Interface Intelligent Matter (Grant SZS2022011). This work is also supported by Collaborative Innovation Center of Suzhou Nano Science & Technology.

Data availability

The data supporting this article have been included as part of the ESI.†

Conflicts of interest

There are no conflicts to declare.

Notes and references

1. K. J. Baeg, M. Binda, D. Natali, M. Caironi and Y. Y. Noh, Adv. Mater., 2013, 25, 4267–4295 CrossRef CAS PubMed .
2. J. Liu, W. Hu and L. Jiang, Sci. Bull., 2023, 68, 1474–1477 CrossRef CAS PubMed .
3. X. Huang, D. Ji, H. Fuchs, W. Hu and T. Li, ChemPhotoChem, 2019, 4, 9–38 CrossRef .
4. D. Xue, W. Gong, C. Yan, Y. Zhang, J. Lu, Y. Yin, J. Zhang, Z. Wang, L. Huang and L. Chi, Adv. Funct. Mater., 2024, 2402884 CrossRef .
5. S. Wang, X. Shi, J. Gong, W. Liu, C. Jin, J. Sun, Y. Peng and J. Yang, Nano Lett., 2024, 24, 3204–3212 CrossRef CAS PubMed .
6. D. Xue, Q. Wang, M. Xie, W. Gong, Y. Zhang, Y. Yin, Y. Wei, J. Lu, J. Zhang, S. Duhm, Z. Wang, L. Chi and L. Huang, Adv. Opt. Mater., 2023, 12, 2302091 CrossRef .
7. E. Ercan, C. C. Hung, G. S. Li, Y. F. Yang, Y. C. Lin and W. C. Chen, Nanoscale Horiz., 2023, 8, 632–640 RSC .
8. J. Huang, J. Du, Z. Cevher, Y. Ren, X. Wu and Y. Chu, Adv. Funct. Mater., 2017, 27, 1604163 CrossRef .
9. C. Uhrich, R. Schueppel, A. Petrich, M. Pfeiffer, K. Leo, E. Brier, P. Kilickiran and P. Baeuerle, Adv. Funct. Mater., 2007, 17, 2991–2999 CrossRef CAS .
10. R. Schueppel, C. Uhrich, M. Pfeiffer, K. Leo, E. Brier, E. Reinold and P. Baeuerle, Chem. Phys. Chem., 2007, 8, 1497–1503 CrossRef CAS PubMed .
11. D. Xue, Y. Zhang, W. Gong, Y. Yin, Z. Wang, L. Huang and L. Chi, Sci. Chin. Chem., 2022, 65, 2567–2575 CrossRef CAS .
12. F. Ling, Q. Du, Y. Zhang, X. Zheng, A. Wang, C. Zhu, W. Wang, F. Wang and S. Qin, Org. Electron., 2024, 124, 160942 CrossRef .
13. Y. T. Lee, Y. T. Huang, S. P. Chiu, R. T. Wang, T. Taniguchi, K. Watanabe, R. Sankar, C. T. Liang, W. H. Wang, S. S. Yeh and J. J. Lin, ACS Appl. Mater. Interfaces, 2024, 16, 1066–1073 CrossRef CAS .
14. H. Li, T. Jiang, Y. Zheng, Y. Zou, S. Qi, G. Tian, D. Ji, L. Li and W. Hu, Adv. Opt. Mater., 2022, 10, 2200614 CrossRef CAS .
15. W. Deng, Y. Lv, X. Ruan, X. Zhang, R. Jia, Y. Yu, Z. Liu, D. Wu, X. Zhang and J. Jie, Laser Photonics Rev., 2022, 16, 2200283 CrossRef CAS .
16. M. Pei, J. Guo, B. Zhang, H. Wang, Y. Zhu, Q. Wang, K. Tsukagoshi, Y. Shi and Y. Li, Mater. Today Energy, 2021, 21, 100711 CrossRef CAS .
17. H. Xu, J. Li, B. H. Leung, C. C. Poon, B. S. Ong, Y. Zhang and N. Zhao, Nanoscale, 2013, 5, 11850–11855 RSC .
18. F. Huang, X. Wang, K. Xu, Y. Liang, Y. Peng and G. Liu, J. Mater. Chem. C, 2018, 6, 8804–8811 RSC .
19. H. Shao, Y. Li, W. Yang, X. He, L. Wang, J. Fu, M. Fu, H. Ling, P. Gkoupidenis, F. Yan, L. Xie and W. Huang, Adv. Mater., 2023, 35, e2208497 CrossRef PubMed .
20. R. Jia, X. Wu, W. Deng, X. Zhang, L. Huang, K. Niu, L. Chi and J. Jie, Adv. Funct. Mater., 2019, 29, 044010 Search PubMed .
21. Q. Wang, J. Yang, A. Gerlach, F. Schreiber and S. Duhm, J. Phys. Mater., 2022, 5, 044010 CrossRef CAS .

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4cc03227f

‡ These authors contributed equally to this work.

This journal is © The Royal Society of Chemistry 2024