Organic photodetectors (OPDs) in the image sensing, communications, ultraviolet (UV)–visible–near infrared (NIR) region, and chemical/biological sensing show a wide range of commercial and scientific applications.^[1–5] Conventionally, most of the OPDs are photodiodes with photoresponse originating from the collection of photogenerated charges.^[6,7] Due to the limitation of photon harvesting efficiency, exciton dissociation efficiency, and photogenerated charge carrier collection efficiency, the external quantum efficiency (EQE) of photodiode-type OPDs is lower than 100%.^[8,9] In spite of the photodiode-type OPDs possessing relatively low working voltage and noise current, the responsivity (R) is not too high due to the EQE values of <100%. In order to achieve highly sensitive optical detection, environmental noise should be generally reduced by complex equipment.^[10,11] Therefore, the OPDs with facile fabrication, high responsivity R, and normalized detectivity (D*) are in urgent need.^[12–15] The critical steps for improving normalized detectivity of OPDs are to enhance responsivity.^[16] In recent years, photomultiplication (PM) structure has been widely applied to the OPDs so as to further improve the optoelectronic performance. The efforts on trap-assisted PM have always been attempted in OPDs and become a hot topic recently.^[12–19]

Wang et al. proposed a solution processed PM-type OPD based on P3HT:PC₆₁BM blend film with PEIE modified ITO electrode, which exhibits a R of 14.25 A·W⁻¹ under 550 nm (3.09 μW/cm²) at −1.0 V bias.^[16] However, the D* is not improved correspondingly. Luo et al. reported PM-type OPDs with a 5 nm C₆₀ as hole block layer between ITO and active layer which show a R of 1.33 A·W⁻¹ under 850 nm (0.782 mW/cm²) at −3.0 V bias.^[17] However, a higher evaporation temperature of C₆₀ will be complex to fabricate. Nie et al. utilized a photon-controllable “valve" formed by PBDTT-DPP: PC₇₁BM/MoO₃ to accumulate and release external circuit carriers, realizing EQE>100%.^[18] However, this structure needs donor material with its electron trap levels lower than the lowest unoccupied molecular orbital (LUMO) of PC₇₁BM, so the materials selection will be limited. Recently, Wang et al. reported that an EQE of 53500% was achieved by trap-assisted PM via a small weight of PC₇₁BM doped in P3HT.^[19] Whereas this large EQE value brought defect of too high operating voltage (−60 V). From the perspective of fabrication process, trap doping is an easy method to realize PM.

In order to solve the shortcomings of the above-mentioned methods, such as special materials and too high working voltage, the bulk heterojunction (BHJ) structure of high exciton dissociation efficiency combined with the method of trap-assisted PM is applied to the OPDs. In this paper, we investigate the PM-type OPDs prepared with P3HT:PC₆₁BM (and PBDT-TT-F:PC₆₁BM) blends doped with low LUMO level of C₆₀ (4.5 eV, lower than PC₇₁BM LUMO level) using the one-step solution process. The PM mechanism of this OPD is investigated systematically, especially for the varying relation between PM efficiency and C₆₀ concentration. Additionally, the device performance of the two donor–acceptor combinations is also compared. Therefore, our works may provide an easy method for obtaining highly sensitive OPDs with low working voltage, high R and D* simultaneously.

Figure 1(c) depicts the energy level diagram of materials used here and we can know that the LUMO level of C₆₀ is much lower than that of P3HT, PBDT-TT-F and PC₆₁BM. So electron traps could be formed in the BHJ active layer after doping C₆₀, as shown in Fig. 1(c), which will be confirmed by a series of tests later. In OPDs, the dark current density (J_dark) is one of the most important electrical parameters, which determines the detection ability of weak signals.^[20] A main contribution to the J_dark is the charge carrier injection through the respective counter electrode, e.g., electrons being injected through the anode.^[6] Due to the rather low electron mobility of PEDOT:PSS, the electrons are directly injected into the LUMO level of PC₆₁BM from ITO anode.^[8] Therefore, there exists a relatively high electron injection barrier of ∼ 1.1 eV and the hole injection barrier of ∼ 0.9 eV (1.0 eV) that prevents hole injection from Al cathode into the highest occupied molecular orbital (HOMO) level of P3HT (PBDT-TT-F) under reverse bias in darkness. Figures 2(a) and 2(b) show the J–V characteristics of C₆₀-doped P3HT:PC₆₁BM and C₆₀-doped PBDT-TT-F:PC₆₁BM devices in dark condition and under 460 nm (P_in = 0.21 mW·cm⁻²) illumination, respectively. First of all, from the J_dark in the reverse bias regime, it is obvious that both of the two structure devices exhibit relatively low J_dark due to the high injection barrier. Additionally, J_dark is further decreased after doping C₆₀, which may be attributed to the reduced electron injection from ITO into the LUMO level of PC₆₁BM because some injected electrons are trapped in C₆₀ near ITO. The decreased J_dark indicates that the incorporation of C₆₀ has a significant influence on the charge injection and the transportation of charge carriers. In general, charge transport in organic semiconductor based diode follows a trap-limited space-charge-limited current (SCLC) model: J ∝ V^k.^[17] To directly investigate the difference of charge migration after doping C₆₀, we further do SCLC analysis based on the dark J–V curves of P3HT:PC₆₁BM and P3HT:PC₆₁BM:1.6 wt.% C₆₀ devices under reverse bias, as shown in Fig. 2(c). In order to find the exponent k (Fig. 2(c)), we piecewise fit the measured data with the relation J ∝ V^k. It can be seen that the slope k is close to 1 either for device A without C₆₀ (k = 1.27) or device C doped with 1.6 wt.% C₆₀ (k = 0.83) in the low voltage region, signifying a near Ohmic regime (k = 1) described by J_Ω = qμNV/L.^[21] Here q is the elementary charge, μ is the carrier mobility, N is the volume density of charge, V is the applied voltage, and L is the thickness of the active layer. It is generally accepted that Ohmic conduction originates from the low intrinsic conductivity of the organic semiconductor and the mobility is mainly influenced by the trapping of charge carrier.^[22,23] Therefore, the slope k for device C is slightly lower than device A, which is caused by the decreased carrier mobility because that some injected electrons from ITO are trapped in C₆₀. At moderate voltage, the injection carrier increases and space charge emerges inside the materials. Hence, J–V relationship becomes nonlinear. When the trap states are not filled or do not affect the charge transport, referring to trap-free SCLC regime (TF-SCLC, k = 2), the J–V relation is described by Mott–Gurney law^[17]

Fig. 2.

	Figure Option ViewDownloadNew Window
	Fig. 2. (a), (b) J–V curves in darkness and and under 460 nm (0.21 mW·cm−2) illumination for the C60-doped P3HT:PC61BM and C60-doped PBDT-TT-F:PC61BM devices. (c) Experimental measured dark J–V curves (symbols) and the simulated plots (red line, J ∝ Vk) of the devices under reverse bias. (d) EQE–V curves of the two kinds of the active layer devices under 460 nm (0.21 mW·cm−2) illumination.

It can be clearly seen that the light current density (J_light) of C₆₀-doped devices exhibits an obvious improvement compared with their respective undoped devices. For example, device A shows only J_light = 5.01 × 10⁻⁵ A·cm⁻² under −1.0 V, however device B, C, and D exhibit larger J_light (1.99 × 10⁻⁴, 2.55 × 10⁻⁴, and 1.68 × 10⁻⁴ A·cm⁻² correspondingly). To give a more intuitive comparison, we calculate the EQE values based on light J–V curves according to Eq. (3), and the EQE–V curves are shown in Fig. 2(d). It is apparent that all the EQE values of C₆₀-doped devices have improved significantly. Especially, the EQE values of device B, C, and D are all greater than 100% at −1.0 V bias (255.2%, 327.5%, 215.9%, respectively). Whereas the EQE value of device A is no more than 100% even at −3.0 V bias. Furthermore, these EQE values increase first and then decrease as the increase of C₆₀ concentration. It means that the electron traps are mainly induced by adding C₆₀ rather than intrinsic defect in active layer films.^[8] Noted that these EQE values of C₆₀-doped devices increases with the applied voltage, but the EQE values of undoped device A are nearly unchanged as the applied bias increases. A similar phenomenon happens to C₆₀-doped PBDT-TT-F:PC₆₁BM devices. Thus, we can conclude that the improvement of device optoelectronic performances is attributed to the external carrier injection after doping C₆₀. The detailed explanation about the EQE value of >100% will be discussed later.

The dispersed C₆₀ in the active layer can be considered as electron traps, which has been discussed in the SCLC analysis above. Given that the electron traps will also influence the transportation of charge carriers generated by incident photons, we further do photoluminescence (PL) study and light-assisted capacitance measurements subsequently. PL is also a strong mean to prove the electron trapping ability of C₆₀ doped in active layer,^[13] which can trap electrons generated in P3HT and PC₆₁BM based on its energy level, shown in Fig. 1(c). Figure 3(a) depicts the PL spectra of P3HT, P3HT:C₆₀, P3HT:PC₆₁BM, and P3HT:PC₆₁BM:C₆₀ films respectively. Emission from P3HT is obviously quenched by C₆₀, demonstrating the electron transfer from P3HT to C₆₀. The P3HT emission from the P3HT:PC₆₁BM mixture is further quenched by adding C₆₀, which confirms the electron transfer from P3HT:PC₆₁BM mixture to C₆₀ and can be predicted from their energy levels as well.

Fig. 3.

	Figure Option ViewDownloadNew Window
	Fig. 3. (a) PL spectra of P3HT, P3HT:C60 (6:0.2 wt/wt), P3HT:PC61BM, and P3HT:PC61BM:C60 (6:6:0.2 wt/wt/wt) films. (b) C–V curves and (c) M–S plots of P3HT:PC61BM and P3HT:PC61BM:1.6 wt.% C60 devices in darkness and under 460 nm (5.54 mW·cm−2) illumination, respectively.

As is well known, light-assisted capacitance measurements have been used to study the device capacitance, which is determined by the charges at the electrode interfaces.^[16,24] Figure 3(b) shows the C–V curves of the P3HT:PC₆₁BM:C₆₀ devices in dark condition and under 460 nm (5.54 mW·cm⁻²) illumination, respectively. It can be clearly seen that the capacitance at zero bias increases from 41.7 nF to 45.9 nF (increments are 4.2 nF) and increases from 44.0 nF to 88.1 nF (increments are 44.1 nF) for undoped device A and 1.6 wt.% C₆₀-doped device C, respectively. It implies that the photogenerated electrons do accumulate at the interface of active layer/Al electrode for the C₆₀-doped devices under illumination. It is generally accepted that the capacitance of diode under small reverse and low forward voltage obeys Mott–Schottky equation^[25]

Fig. 4.

	Figure Option ViewDownloadNew Window
	Fig. 4. Schematic diagram of the operating mechanism of typical photovoltaic OPDs (a) and PM-type OPDs (b).

Incident photons absorbed in P3HT (PBDT-TT-F) and PC₆₁BM firstly generate electron–hole pairs, which separate at the interface of P3HT (PBDT-TT-F)/PC₆₁BM into free electrons and holes, for undoped devices, holes eventually end up on P3HT (PBDT-TT-F), and electrons end up on PC₆₁BM, finally, these free charges transport in their respective channels and are collected by anode/cathode, which is sketched in Fig. 4(a). However, for C₆₀-doped devices, photogenerated electrons are captured by C₆₀ traps due to the lowest LUMO energy level among all these components. The accumulated electrons near the cathode side will also induce hole accumulation of Al side (Fig. 4(b)), resulting in the band bending of P3HT (PBDT-TT-F) and the increase of Al work function.^[18,19,26] As a result, the barrier width (SCR of active layer/Al interface) will be reduced, which is sketched in Fig. 4(b). When the band bending reduces hole wavelength to less than or equal to the barrier width under enough illumination, the cathode holes can tunnel through the thin barrier region into the HOMO level of P3HT (PBDT-TT-F) under reverse bias. Now, these external hole injection current plus photogenerated current together contributes to the EQE value of >100%.^[12–19] It should be noted that the electron transfer from P3HT (PBDT-TT-F) to C₆₀, either direct transfer or indirect transfer through PC₆₁BM, is paramount for the observed improvement of device performance.

In order to exclude the other factors that may influence the device performance to some extent, such as the morphology and the light absorption of active layer film,^[27–29] next we investigate the difference of films after doping C₆₀ based on their AFM images and absorption spectra. Figures 5(a) and 5(b) show the AFM images of P3HT:PC₆₁BM:C₆₀ films with different C₆₀ concentrations. The neat P3HT:PC₆₁BM film exhibits a surface morphology with a root mean square roughness (RMS) of 1.26 nm, whereas the P3HT:PC₆₁BM:1.6 wt.% C₆₀ film exhibits an RMS of 1.01 nm. It is believed that there are not significant changes of morphology in these two films. Meanwhile, from the absorption spectra of the films shown in Fig. 5(c), we can observe that no apparent change of absorbance occurs between the undoped-films and C₆₀-doped films. Thus, the film morphology and light absorption cannot be the real reason for improved device performance.

Fig. 5.

	Figure Option ViewDownloadNew Window
	Fig. 5. AFM images of films for (a) P3HT:PC61BM and (b) P3HT:PC61BM:1.6 wt.% C60. (c) Absorption spectra of P3HT:PC61BM:C60 (0 and 1.6 wt.%) and PBDT-TT-F:PC61BM:C60 (0 and 1.3 wt.%), respectively.

It is known that the responsivity (R) and normalized detectivity (D*) of the OPDs strongly depend on the voltage and photon wavelength.^[16] Tables 1 summarizes the optoelectronic performances for P3HT:PC₆₁BM:C₆₀ and PBDT-TT-F:PC₆₁BM:C₆₀ OPDs under −1.0 V bias, in which the C₆₀ doping concentration is changed. It should be noted that the PM happening inside devices doped with C₆₀ is the key reason for the observed improvement of device performance. Specifically, the device performance improves first and then degrades complying with the increase of C₆₀ concentration. We attribute the main factor of the variation of device performance to the cathode hole tunneling injection. If C₆₀ concentration is lower, the trap numbers of the active layer near Al side will decrease as well. Hence, the trapped electrons in C₆₀ near the Al side will reduce, which induces weaker band bending and triggers less cathode hole tunneling injection. If C₆₀ concentration is higher, small molecule C₆₀ would like aggregating to form some isolated clusters or even continuous electron transport channels.^[8] A part of photogenerated electrons may be collected by cathode through these channels. Therefore the trapped electrons in the active layer near Al side will be also decreased, thereby resulting in the less cathode hole tunneling injection. Only when the C₆₀ concentration is moderate, the cathode hole tunneling injection can be effectively enhanced, which is supported by the optimal performance for device C and G in their respective active layer system.

Table 1.

Table 1.

Table 1. Optoelectronic performances for P3HT:PC61BM:C60 (A, B, C, and D) and PBDT-TT-F:PC61BM:C60 (E, F, G, and H) OPDs under −1.0 V and comparison with other reported PM-type OPDs. . Device structure λ/nm Pin/mW·cm−2 Vapp/V R/A·W−1 D*/Jones EQE/% Ref. A 0.24 5.22 × 1011 64.4 B with 0.4 wt.% C60 0.95 3.22 × 1012 255.2 C with 1.6 wt.% C60 1.21 4.22 × 1012 327.5 D with 4.8 wt.% C60 0.80 2.68 × 1012 215.9 this E 460 0.21 −1.0 0.23 1.86 × 1011 61.2 work F with 0.3 wt.% C60 0.48 7.50 × 1011 128.0 G with 1.3 wt.% C60 0.64 1.05 × 1012 173.4 H with 3.8 wt.% C60 0.43 4.90 × 1011 114.8 PEIE/P3HT:PC61BM 550 3.09 × 10−3 −1.0 14.25 1.04 × 1012 ∼ 2200 [16] C60/C60:PbPc/BCP 850 0.782 −3.0 1.33 – 192.9 [17] PBDTT-DPP:PCBM/MoO3 460 – 1.0 ∼ 20.0 7.81 × 1013 ∼ 1000 [18] P3HT:PC71BM(100:1 wt/wt) 650 – −20.0 – 1.4 × 1011 600 [19]

Table 1. Optoelectronic performances for P3HT:PC61BM:C60 (A, B, C, and D) and PBDT-TT-F:PC61BM:C60 (E, F, G, and H) OPDs under −1.0 V and comparison with other reported PM-type OPDs. .

It is noteworthy that the C₆₀-doped P3HT:PC₆₁BM devices show better optoelectronic performance comparing with C₆₀-doped PBDT-TT-F:PC₆₁BM devices according to Table 1. We consider there are two main factors contributing to the better performance of P3HT:PC₆₁BM:C₆₀ devices. The one factor is the higher exciton dissociation efficiency of P3HT:PC₆₁BM active layer due to larger energy offset of the LUMO energy levels of the donor and acceptor materials (P3HT/PC₆₁BM:0.7 eV, PBDT-TT-F/PC₆₁BM:0.3 eV).^[30] Hence, more photogenerated electrons are trapped in C₆₀-doped P3HT:PC₆₁BM active layer and higher efficiency of PM can be obtained. Another factor of performance improvement for P3HT:PC₆₁BM:C₆₀ devices is the increased hole injection from Al cathode into the HOMO level of donor. In order to compare the hole injection to different donor materials, we measured the J–V curves of the PM-type OPDs based on ITO/PEDOT:PSS/P3HT:C₆₀ (and PBDT-TT-F:C₆₀)/Al with all the weight ratio of 10:0.3 in darkness and under 460 nm (P_in = 5.54 mW·cm⁻²) illumination, respectively, as shown in Fig. 6(a). Because of the smaller hole injection barrier between the Al work function and the HOMO level of P3HT (0.9 eV, PBDT-TT-F/Al: 1.0 eV), the J_dark of P3HT:C₆₀ device is slightly larger than PBDT-TT-F:C₆₀ device. In fact, the higher J_light of P3HT:C₆₀ device at larger reverse bias regime (larger than −4.5 V) suggests more cathode hole injection into the HOMO level of P3HT. Figure 6(b) further depicts the EQE–V curves of P3HT:C₆₀ and PBDT-TT-F:C₆₀ devices based on their light J–V curves. It is clearly observed that when the EQE value is equal to 100%, the voltage of P3HT:C₆₀ device is about −5.1 V, which is lower than that of PBDT-TT-F:C₆₀ device (−5.9 V). The curve for P3HT:C₆₀ device increases much faster at large reverse bias, as shown in Fig. 6(b). Consequently, we can conclude that more cathode hole can be injected into the donor material with shallower HOMO level.

Fig. 6.

	Figure Option ViewDownloadNew Window
	Fig. 6. (a) J–V curves in darkness and under 460 nm (5.54 mW·cm−2) illumination for the P3HT:C60 (10:0.3, wt/wt) and PBDT-TT-F:C60 (10:0.3, wt/wt) devices. (b) EQE–V curves of the corresponding P3HT:C60 and PBDT-TT-F:C60 devices under 460 nm (5.54 mW·cm−2) illumination.

Table 1 also gives the comparison of optoelectronic performance of the OPDs fabricated in this paper with other PM-type OPDs. In short, for 1.6 wt.% C₆₀-doped P3HT:PC₆₁BM device C, the high R of 1.21 A·W⁻¹ is ∼5-fold higher than that of photodiode-type OPDs (λ = 500 nm, U = −1 V, R = 0.273 A·W⁻¹),^[6] the high D* of 4.22 × 10¹² Jones is ∼ 4-fold higher than that of PM-type OPDs under the same bias of −1.0 V (λ = 550 nm, P_in = 3.09 μW·cm⁻², D* = 1.23 × 10¹² Jones),^[16] which is comparable or even superior to some inorganic photodetectors.^[7,11,31] Moreover, under the premise of the same EQE, device C shows a much lower working voltage compared with the PM-type OPDs reported in Ref. [8] (−6 V). Although the EQE values are not greater than 1000%, the PM-type OPDs reported here with facile fabrication and relatively high device performance can solve the problem of too high working voltage and the limitation of special materials.^[18,19]

In summary, we have demonstrated a one-step solution processed PM-type OPD based on C₆₀-doped bulk heterojunction structure at low bias. Using the advantages of BHJ with high exciton dissociation efficiency as active layer, large PM under low bias can be achieved by doping a small weight of C₆₀ as traps. A schematic energy-level model combined with experimental measurements is capable of explaining the operating mechanism of the PM-type OPDs proposed here. The PM originates from cathode hole tunneling injection assisted by the trapped electrons in C₆₀ near the Al side. The efficiency of PM is related to the C₆₀ concentration in the active layer. In order to further improve the device performance, the donor material should have higher LUMO level and shallower HOMO level. On the basis, the 1.6 wt.% C₆₀-doped P3HT:PC₆₁BM device exhibits EQE = 327.5%, R = 1.21 A·W⁻¹, and D* = 4.22 × 10¹² Jones at −1 V under 460 nm (0.21 mW·cm⁻²) illumination. This paper provides a universal method of designing high PM OPDs with low operating voltage and facile fabrication.