Introduction

To avoid the interference of surrounding signals in practical applications such as image sensing, fiber optic communication, and bioinformatics monitoring, it is necessary to prepare the organic photodetectors (OPDs) with specific spectral selectivity1,2,3. The function of spectrally selective or narrow-band photo-detection in practical applications is usually realized by combining a broad-band photodetectors with a dedicated optical filter which will increase the complexity and cost of device design, and the filters are generally rigid and cannot be applied to flexible OPD devices4,5,6,7. Therefore, developing narrow-band OPDs without filters is one of the current research hotspots. Based on material synthesis and device structure design, filter-free narrow-band OPDs are mainly obtained in the following four different ways8,9,10,11,12,13,14: (1) design and synthesis of spectrally selective organic photosensitive materials; (2) exciton generation narrowing; (3) charge collection narrowing; and (4) exciton dissociation narrowing.

In 2009, Fukuda et al. achieved a blue light spectral response with a bandwidth of about 120 nm using F8BT monolith as an active layer, but the external quantum efficiency (EQE) of the device was only 0.12% due to the low exciton dissociation rate of the monolithic node15. In 2013, Park et al. prepared a green light narrow-band OPD based on a DMQA: DCV3T photoactive layer16. Apart from tuning the absorption spectra, narrow-band OPDs with specific spectral response can also be realized through reasonable device structure design and optimization. In 2017, Mischok et al. prepared distributed bragg reflection (DBR) electrodes with high reflectivity through 21 alternating layers of SiO2 and TiO2 which obtained a narrow-band OPD device with the full width at half maximum (FWHM) of 21 nm and a 17% EQE at 880 nm17. In the same year, Siegmund et al. achieved the dynamic EQE peaks of OPDs from 875 nm to 1550 nm by adjusting the thickness of the function layers18. However, the microcavity structure requires precise wavelength matching conditions and the disadvantages of complicated preparation processes, high cost and angular dependence of the response spectra also restricts its development. In 2015, Ardalan et al. proposed the concept of charge collection narrowing (CCN)19. When the thickness of the bulk heterojunction (BHJ) active layer is large enough (1.5–3 μm), the stronger absorbing bands are absorbed by the active layer near the incident indium tin oxide (ITO) electrode, and the photogenerated carriers generated in this band are too far away from the top electrode (Al electrode), which makes it difficult to drift to the Al electrode and be collected. On the other hand, the light energy in the weakly absorbing band in the active layer can be absorbed by the entire active layer, and this band of photogenerated carriers exists in the entire active layer, thus can eventually be absorbed by the electrodes to generate the photoresponse. However, due to the large thickness of the active layer, the narrow-band OPDs based on CCN generally have the drawbacks of low photoresponse and slow response speed20,21. In addition, in 2020, Fei Huang et al. proposed the concept of exciton dissociation narrowing (EDN) and prepared a new narrow-band OPD device with a response crest of 860 nm, an EQE response of 65%, a FWHM of 50 nm, and a specific detectivity (D*) of 1013 Jones22. EDN further broadens the research direction of low-cost and high-performance narrow-band OPDs. However, in the preparation of double-layer planar heterojunction, to prevent the destruction of the donor layer by solvents during the preparation of the acceptor layer, the corresponding cross-linking agent is added to the preparation of the donor layer, which limits the universality of the method23,24.

Herein, we utilized the thin film transfer printing technology (TFTPT) in preparing active layers with BHJ on the organic photo-filtering layers and successfully realized the narrow-band red and near-infrared (NIR) OPD devices without any external filters. By adding appropriate amounts of poly[4,8-bis(5-(2-ethylhexyl)thiophen-2-yl)benzo[1,2-b;4,5-b’]dithiophene-2,6-diyl -alt-(4-(2-ethylhexyl)-3-fluorothieno[3,4-b]thiophene-)-2-carboxylate-2-6-diyl] (PTB7-Th) and poly[(2,6-(4,8-bis(5-(2-ethylhexyl)thiophen-2-yl)-benzo[1,2-b:4,5-b’]dithiophene))-alt-(5,5-(1’,3’-di-2-thienyl-5’,7’-bis(2-ethylhexyl)benzo[1’,2’-c:4’,5’-c’]dithiophene-4,8-dione)] (PBDB-T) into the photo-filtering layer, the absorption spectra of the photo-filtering layer in the wavelength of 580–720 nm and 300–450 nm were effectively complemented. The NIR OPD devices with the response peak of 790 nm and the corresponding spectra rejection ratios (SRRs) of 20, 36, and 256 at 378 nm, 691 nm, and 555 nm were prepared. By optimizing the thickness of the photo-filtering layer, the narrow-band NIR device with a response peak at 790 nm and a FWHM of 62 nm was prepared with the responsivity (R) and D* at 790 nm of 0.24 A/W and 1.39 × 1012 Jones, respectively, under the bias voltage of -5 V. In this device, the incident high-energy short-wavelength photons are firstly absorbed by the photo-filtering layer and then generate excitons, but they cannot be dissociated and finally annihilated because of the lack of a donor-acceptor interface. Only low-energy long-wavelength photons can penetrate the donor layer to reach the donor-acceptor interface and generate photogenerated charges, which are finally collected by the corresponding electrodes to form the photoresponse. The organic photo-filtering layers have both optical properties of light filtration and electrical properties of hole transport and electron blocking. The responding wavelengths of the device and the corresponding FWHMs can be flexibly controlled by adjusting the absorption spectra of the BHJ active layers and the photo-filtering layers.

Results and discussion

Fabrication process of OPDs based on organic photo-filtering layers using TFTPT

Molybdenum oxide (MoO3), PTB7-Th, poly(3-hexylthiophene-2,5-diyl) (P3HT), PBDB-T, 3,9-bis(2-methylene-(3-(1,1-dicyanomethylene)-indanone))-5,5,11,11-tetrakis(4-hexylphenyl)-dithieno[2,3-d:2’,3’-d’]-s-indaceno[1,2-b:5,6-b’]dithiophene (ITIC), 3,9-bis(2-methylene-((3-(1,1- dicyanomethylene)-6,7-difluoro)-indanone))-5,5,11,11-tetrakis(4-hexylphenyl)-dithieno[2,3-d:2’,3’-d’]-s-indaceno[1,2-b:5,6-b’]dithiophene (IT4F), poly[(2,6-(4,8-bis(5-(2-ethylhexyl-3-fluoro) thiophen-2-yl)-benzo[1,2-b:4,5-b’]dithiophene))-alt-(5,5-(1’,3’-di-2-thienyl-5’,7’-bis(2-ethylhexyl)benzo[1’,2’-c:4’,5’-c’]dithiophene-4,8-dione)] (PM6) and 2,2’-((2Z,2’Z)-((12,13-bis(2-ethylhexyl) -3,9-diundecyl-12,13-dihydro-[1,2,5]thiadiazolo[3,4-e]thieno[2”,3’’:4’,5’]thieno[2’,3’:4,5]pyrrolo[3,2-g]thieno[2’,3’:4,5]thieno[3,2-b]indole-2,10-diyl)bis(methanylylidene))bis(5,6-difluoro-3-oxo-2,3-dihydro-1H-indene-2,1-diylidene))dimalononitrile (Y6) were bought from Shanghai Daeyeon Chemical Co., Ltd. Polydimethylsiloxane (PDMS) stamps was ordered from Solarmer Materials Inc. ITO anodes coated on glass substrates were purchased from the CSG Holding Co., LTD. The chemical structures of the relevant materials and their energy bands are shown in Fig. S1 (Electronic supporting information, ESI).

We combined EDN and TFTPT to prepare narrow-band OPD devices with tunable response spectra with the basic structure is ITO/electron blocking layer/photo-filtering layer/NIR BHJ active layer/cathode modification layer/aluminum (Al). The NIR BHJ active layer was prepared by TFTPT. MoO3 was selected as the electron-blocking material and P3HT: PTB7-Th: PBDB-T as the photo-filtering layer. MoO3 precursor solutions are based on literature25.

The PBDB-T: ITIC (IT4F) precursor solution was prepared by 10 mg of PBDB-T and ITIC (IT4F) dissolved in 995 µL of chlorobenzene and 5 µL of 1,8-diiodooctane. The PM6:Y6 precursor solution was prepared by 7.2 mg of PM6 and 8.6 mg of Y6 dissolved in 995 µL of chloroform and 5 µL of chloronaphthalene. The photo-filtering layer precursor solution was prepared by 20 mg of P3HT, 20 mg of PTB7-Th and 5 mg of PBDB-T dissolved in 1000 µL of chlorobenzene.

The pre-ordered ITO substrates were cleaned successively in de-ionized water, acetone and isopropanol, and then exposed to ultraviolet ozone. Then MoO3 solution was spined coat onto the surface of ITO at a speed of 2000 rpm. Afterwards, ITO substrates were transferred into the glove box and photo-filtering layers were spined coat with the thickness of 450 nm, respectively. Immediately after the active layers were spined coat onto the silicon substrates treated with ultraviolet ozone into the glove box and then the ITO substrates with photo-filtering layers and the silicon substrates with active layers were removed from the glove box. Then the pre prepared PDMS films were attached onto the active layer films and place them together with the silicon substrates in the deionized water. Afterwards, the PDMS films with an attached active layer film were taken out from deionized water and attached to the photo-filtering layers on the ITO substrates. Finally, the entire samples were quickly transferred to the glove box and annealed at 100℃ for 30 s. At this time, the PDMS films could be peeled off and the transfer process of the active layers was finished. After completing the above steps, the samples were transferred to a vacuum evaporation chamber in the vacuum evaporation equipment at < 5 × 10− 5 mbar. 1.5 nm cathode modification layer 8-quinolinolato lithium (Liq) and 100 nm Al electrodes are evaporated with the rate of Liq at 0.05 nm/s and the rate of Al electrodes at 0.5 ~ 1 nm/s. The step-by-step schematic of the device fabrication is shown in Fig. S2.

EQE spectra of OPDs with different photo-filtering layers

The transmission spectrum of the P3HT film, and the normalized absorption spectra of the films of P3HT and PM6:Y6 are showed in Fig. 1a. It can be seen that PM6:Y6 has a good photoresponse in the wavelength range of 300 ~ 950 nm with the peaks located at 628 nm and 812 nm, respectively. Whereas P3HT has a strong absorption in the wavelengths of 300 ~ 620 nm, with the absorption peak at 521 nm. As shown in Fig. 1b, we choose P3HT as the photo-filtering layer and non-fullerene PM6:Y6 as the active layer to prepare narrow-band NIR OPDs with the structure of ITO/MoO3/P3HT/PM6:Y6/Liq/Al, where photons of 300 ~ 620 nm wavelength are absorbed by the P3HT layer, and the photons greater than 620 nm penetrate through the P3HT layer, then be absorbed and dissociated into free carriers by the PM6:Y6 layer which are finally collected by the electrodes on both sides to form a narrow-band photoresponse with a FWHM of about 280 nm in the NIR wavelengths. From Fig. 1b, it can be seen that the increase of the thickness of the P3HT photo-filtering layers significantly enhances its visible light absorption, but it also causes a decrease in the EQE response of the device in the red-light region. To further reduce the FWHM of the device in the NIR region as well as ensuring the optical response, it is necessary to select suitable materials with good hole transporting ability and absorption spectra. Figure 1c shows the normalized absorption spectra of P3HT, poly[2-methoxy-5-(2-ethylhexyloxy)-1,4-phenylenevinylene] (MEH-PPV), PBDB-T and PTB7-Th. It can be seen that PTB7-Th has the maximum cut-off absorption wavelength (760 nm), and its absorption peaks are well complementary to that of P3HT in the visible range. We prepared the narrow-band OPD device with P3HT: PTB7-Th hybrid photo-filtering layer and PBDB-T: IT4F as the active layer with the device structure of ITO/MoO3/P3HT: PTB7-Th/PBDB-T: IT4F/Liq/Al. The EQE response is shown in Fig. 1d. It can be clearly found that the addition of P3HT: PTB7-Th significantly suppresses the EQE response of the devices in the wavelengths of 300 ~ 720 nm compared to the basic devices without photo-filtering layer, and the peak is located at 786 nm with a FWHM of 61 nm. However, the sub-peaks are clearly observed in the wavelengths of 580 ~ 720 nm and 300 ~ 450 nm, which will cause the signal crosstalk. Compared to the absorption spectra of P3HT and PTB7-Th, it can be found that the main reason for the formation of the sub-peaks is the weak light absorption ability of the P3HT: PTB7-Th photo-filtering layer in the corresponding range, and it is difficult for the nanometer-thick P3HT: PTB7-Th photo-filtering layer to completely absorb all the incident light in those wavelengths, and finally the photons are absorbed by the active layer and dissociated into free carriers to be collected by the electrodes, resulting in the photoresponse.

Fig. 1
figure 1

(a) Normalized absorption spectra of PM6:Y6 and P3HT films and transmission spectrum of P3HT film, (b) the EQE response spectra with P3HT photo-filtering layers of different thicknesses; (c) Normalized absorption spectra of the photo-filtering layer materials, (d) the EQE spectra of the device without/with P3HT: PTB7-Th hybrid photo-filtering layer (the inset in (d) shows the EQE response spectrum of the device with P3HT: PTB7-Th hybrid photo-filtering layer).

Full size image

To minimize the sub-peaks in the wavelength ranges of 580 ~ 720 nm and 300 ~ 450 nm for the narrow-band OPD devices, we further added PBDB-T into the photo-filtering layer to supplement the absorption of incident light in the wavelength ranges of 580 ~ 720 nm and 300 ~ 450 nm. Figure 2a shows the normalized EQE spectra of the devices corresponding to different optical filtering layers. It can be seen that with the addition of PBDB-T, the EQE of the device in the visible wavelengths decreases significantly. SRR is used to measure the rejection of visible light by the device. The equation is as follows26 :

$${\text{SR}}{{\text{R}}_\lambda }=\frac{{{\text{EQ}}{{\text{E}}_{{\text{target}}}}}}{{{\text{EQ}}{{\text{E}}_\lambda }}}$$
(1)

where EQEtarget represents the corresponding EQE of the device at the target wavelength, and EQEλ represents the EQE of the device at a wavelength of λ. As can be seen from Fig. 2a, when the concentration of P3HT: PTB7-Th: PBDB-T is 20:25:5 mg/mL, the device has the smallest sub-peaks in the visible wavelengths, and the corresponding SRRs at 378 nm, 691 nm, and 555 nm are 20, 36, and 256, respectively, which present a good light response suppression in the visible wavelengths. Figure 2b shows the mechanism of the narrow-band OPDs based on EDN, in which high-energy photons of short wavelengths are first absorbed by the photo-filtering layer and form excitons. However, due to the lack of donor/acceptor interfaces, the excitons cannot dissociate into free charges and ultimately undergo recombination annihilation. In contrast, the photons of long wavelengths cannot be absorbed by the photo-filtering layer so that they effectively reach to the active layer, where they are absorbed and dissociated into free electrons and holes to form the narrow-band photoresponse.

Fig. 2
figure 2

(a) Normalized EQE spectra of OPDs with different photo-filtering layers, and (b) the mechanism diagram of EDN-type narrow-band OPDs.

Full size image

Photoelectric performance of narrow-band OPDs based on photo-filtering layers

We optimized the mixing ratios and thickness of the photo-filtering layers to prepare a narrow-band NIR Device A with the structure of ITO/MoO3 (8 nm) /P3HT: PTB7-Th: PBDB-T (20:25:5 mg/mL; 450 nm)/PBDB-T: IT4F (150 nm) /Liq (1 nm) /Al (100 nm). A reference Device B without the photo-filtering layer was also prepared with the structure of ITO/MoO3/PBDB-T: IT4F/Liq/Al. Fig. S3 shows the current density-voltage (J-V) curves of the two devices in the ambient conditions of illumination (AM 1.5G) and darkness. Due to the filtering and blocking effect of the photo-filtering layer on the photons of visible wavelengths, only the photons of over 740 nm (NIR) can reach the active layer and be absorbed, and finally dissociate into free carriers, so the photoresponse current of Device A is much smaller than that of Device B. Under − 5 V bias, the dark current density of Device A is 9.21 × 10− 8 A/cm2, which is much lower than that of the reference Device B (1.30 × 10− 4 A/cm2). Therefore, in addition to the function of photo filtering, the reverse noise current suppression of the photo-filtering layers also contributes to enhance the performance of narrow-band OPDs.

The EQE spectra of Device A under different bias voltages are shown in Fig. 3a and d. Device B exhibits a good EQE response in the wavelengths of 300 ~ 850 nm. Due to the photo-filtering effect of the photo-filtering layer in the wavelengths of visible light, Device A has a significant EQE response only in the wavelength range of 740 ~ 850 nm, with the peak of 790 nm and a FWHM of 62 nm, which has a superior narrow-band characteristic. With the increase of reverse bias, the corresponding EQE spectra of Device A do not change significantly, but the response intensities are significantly improved. At the peak of 790 nm, the corresponding EQE values of Device A are 14.72% and 37.50% at the bias of 0 V and − 5 V, respectively. As shown in Fig. 3b, the normalized EQE spectra of Device A under different bias voltages do not show any significant changes, and the peaks are all located at 790 nm, with a FWHM of 62 nm. In addition, under 0 V bias, the corresponding SRRs of Device A at 378 nm, 691 nm, and 555 nm are 37, 42, and 917, respectively, and under − 5 V bias, the corresponding SRRs are 42, 48, and 960, respectively. The experimental results show that the applied bias only changes the charge transport and collection efficiency in the device and has no significant effect on the optical response of the device, which can help to improve the reliability of the device in practical use.

Fig. 3
figure 3

(a) EQE spectra, (b) normalized EQE spectra (c) R and (d) D*of Device A under different reverse biases (the insets are the enlarged view of EQE and D*).

Full size image

R and D* of the devices can be calculated by Equation S-1 and S-2, and the specific parameters of the two devices at wavelength of 790 nm with different bias voltages are summarized in Table 1. Significantly, it has been pointed out that this calculation can overestimate the D*27,28. However, it is widely used and thus enables a fair comparison between reported OPD performances22,29. Combining Fig. 3c and d, although the R of Device A is much smaller than that of Device B without the bias, the D* of Device A is higher than that of Device B due to the suppression of noise current by the photo-filtering layer. With the increase of the bias voltage, a significant increase in optical response and noise current is observed in the Device A. From − 1 V to -5 V, the photocurrent density (Jph) of the device increases from 2.02 × 10− 4 A/cm2 to 5.01 × 10− 4 A/cm2, while the reverse noise current density (Jd) increases from 6.34 × 10− 9 A/cm2 to 9.21 × 10− 8 A/cm2. Since Jd increases much more than Jph, the corresponding D* of the device decreases as the reverse bias voltage increases. In addition, the applied bias voltages also provide a significant effect on the response speed of the device, and the reversed applied electric field can promote the free carrier transport within the devices, thus achieving the improving the response speed of OPDs. As shown in Fig. 4, the rise time (τr) and fall time (τd) of Device A under − 5 V are 0.12 ms and 0.21 ms, respectively. Meanwhile τr and τd of Device A under 0 V are 0.20 ms and 0.27 ms, respectively, indicating that the applied reverse bias can effectively reduce the response time of the device.

Fig. 4
figure 4

The (a) rise time and (b) fall time of narrow-band NIR OPD at − 5 V and the (c) rise time and (d) fall time at 0 V bias.

Full size image
Table 1 Performance parameters of narrow-band NIR OPD at 790 nm wavelength under different bias voltages.
Full size table

The linear dynamic range (LDR) refers to a linear relationship between light intensity and device response. For a photovoltaic device, this follows the power law: J∝Pα, where J is the current, P is the incident light power, and α is typically close to 1 for an efficient organic solar cell. When α = 1, the current is linearly proportional to light power. However, when the bimolecular recombination rate of photogenerated carriers rises to compete with the charge collection rate, it will lead to a sub-linear relationship of the photocurrent and the optical power density of the device. Fig. S4 shows the linear dynamic range (LDR) of the device at different bias voltages. When there is no bias applied, the thick photo-filtering layer reduces the internal electric field, leading to a significant sub-linearity between the light intensity and device response30. With the gradual increase of the applied reverse bias voltage, the linearity of the device is gradually obvious, which is due to the fact that the applied reverse bias voltage facilitates the transport and collection of photogenerated holes. In addition, the reverse bias voltage increases the bulk noise current of OPDs, and the increase of the noise current limits the noise equivalent power and the detection ability of the device for weak light. As shown in Fig. S4, the minimum detected optical power densities of Device A at -1 and − 5 V are 1.27 × 10− 4 and 3.22 × 10− 3 mW/cm2, respectively. the LDRs of Device A calculated by Equation S-3 at -1 and − 5 V are 73 and 66 dB, respectively. The LDRs of the narrow-band NIR Device A at different bias voltages are summarized in Table 2. It is noting that the device shows the same photo current at incident optical power densities from 0.01 to 0.31 mW/cm2 from − 1 to -5 V bias. If the applied reverse bias of the device is adjusted based on the incident optical power, the LDR of the device can be extended with maximum and minimum optical power densities of 6.36 and 1.27 × 10− 4 mW/cm2, respectively, corresponding to a LDR of about 94 dB.

Table 2 LDRs of device A at different bias voltages.
Full size table

The universality of photo-filtering layers

The spectral response ranges and FWHMs of narrow-band OPD devices can be flexibly adjusted using TFTPT by adjusting the transmittance spectrum of the photo-filtering layer and the absorption spectrum of the active layer. Figure 5a and c show the narrow-band OPD devices based on PM6:Y6 and PBDB-T: ITIC active layers. Using P3HT: PTB7-Th: PBDB-T (20:25:5 mg/mL) as the photo-filtering layer, which can absorb the incident light within 720 nm and ensures more than 80% transmittance of the NIR light over 720 nm, and PM6:Y6 as the active layer, the narrow-band NIR OPD with the response peak located at 823 nm and the FWHM of 159 nm was prepared. Similarly, a narrow-band NIR OPD with a response peak at 712 nm and the FWHM of 103 nm was prepared by using PBDB-T: ITIC as the active layer and P3HT as the photo-filtering layer, demonstrating the universality of our strategy. Table 3 lists the performance comparison of some narrow-band NIR OPDs reported in the literature with that of our devices. The performance of OPD devices which are based on TFTPT and photo-filtering layers is also superior to that of inorganic PDs based on InGaAs and Germanium (Fig. 5d).

Fig. 5
figure 5

The (a) transmittance spectrum of P3HT: PTB7-Th: PBDB-T and the normalized absorption spectrum of PM6:Y6, (b) the transmittance spectrum of P3HT and the normalized absorption spectrum of PBDB-T: ITIC, (c) the EQE spectra of narrow-band OPDs based on PM6:Y6 and PBDB-T: ITIC, and (d) comparison of the performance between our devices and others reported in literatures.

Full size image
Table 3 Comparison of the performance parameters of narrow-band NIR OPDs.
Full size table

Conclusions

In conclusion, a strategy to prepare narrow-band OPD devices using TFTPT technology is proposed. We utilized TFTPT-based BHJ active layer on the photo-filtering layer to construct narrow-band OPDs. By using appropriate amounts of PTB7-Th and PBDB-T, a narrow-band NIR OPD device were prepared with a response peak at 790 nm and a FWHM of 62 nm. The R and D* at 790 nm are 0.24 A/W and 1.39 × 1012 Jones, respectively, under the bias voltage of -5 V. We also proved that by adjusting the absorption spectra of the organic photo-filtering layer and the active layer, our strategy can be used to prepare different narrow-band NIR OPDs with tunable response spectra.

Experimental sections

J-V curves of the OPDs were captured by a Keithley 2450, and the ON–OFF switching characteristics was measured under the driven of a RIGOL DG822 function generator. A Photonics Power Meter (OPHIR–NOVA II) was used to calibrate the light intensity before tests. The EQE of the OPDs was recorded by a solar cell measurements system (7-STAR Co.) The absorption and transmission spectra are tested by an UV-vis spectrophotometer.