Naphthalocyanine-Based NIR Organic Photodiode: Understanding the Role of Different Types of Fullerenes

In this work, we presented experimental observation on solution-processed bulk heterojunction organic photodiode using vanadyl 2,11,20,29-tetra tert-butyl 2,3 naphthalocyanine (VTTBNc) as a p-type material. VTTBNc is blended with two different acceptors, which are PC61BM and PC71BM, to offer further understanding in evaluating the performance in organic photodiode (OPD). The blend film of VTTBNc:PC71BM with a volumetric ratio of 1:1 exhibits optimized performance in the VTTBNc blend structure with 2.31 × 109 Jones detectivity and 26.11 mA/W responsivity at a −1 V bias. The response and recovery time of VTTBNc:PC71BM were recorded as 241 ms and 310 ms, respectively. The light absorption measurement demonstrated that VTTBNc could extend the light absorption to the near-infrared (NIR) region. The detail of the enhancement of the performance by adding VTTBNc to the blend was further explained in the discussion section.


Introduction
Organic photodiodes (OPDs) are growing tremendously as an alternative to inorganic photodiodes due to their unique advantages of large-area fabrication, lightweight, low cost, detection tunability and mechanical flexibility [1][2][3][4][5]. For the last few years, the solutionprocessed bulk heterojunction (BHJ) was the prevalent design used in OPDs and organic photovoltaics (OPVs) [6][7][8][9]. It appears to be the most effective approach for an organic semiconductor with respect to the processing costs. This design mixes the interfacial region between donor and acceptor on a few nanometres scales to enhance charge separation. The progress in OPDs development contributes to the production of many applications such as image sensors, cameras, remote sensing, camcorders, automation and optical transmission system [6,10]. Therefore, the key parameters of an OPD such as detectivity, responsivity and external quantum efficiency (EQE) should be enhanced to achieve a high OPD performance.
In the fabrication of BHJ OPDs, the commonly used electron acceptor material is based on fullerene derivatives. Among others, [6,6]phenyl-C61-butyric acid methyl ester (PC 61 BM) and [6,6]phenyl-C71-butyric acid methyl ester (PC 71 BM) are the most widely used since they possess high electron affinities and high solubility in various

Experimental
VTTBNc and PC 61 BM were purchased from Sigma Aldrich (St. Louis, MO, USA), while PC 71 BM was obtained from Luminescence Technology Corp. (Hsinchu, Taiwan). The molecular structure of VTTBNc, PC 61 BM and PC 71 BM are illustrated in Figure 1a-c. In the preparation of the photoactive solution, 20 mg of each material were dissolved separately in 1 mL chloroform and stirred overnight in the nitrogen-filled glove box. Then, the blend solution was prepared by mixing VTTBNc with PC 61 BM and PC 71 BM in three different volume ratios (1:0.5, 1:1, 1:1.5) and filtered using a 0.20 µm nylon filter. Meanwhile, the pre-patterned indium tin oxide (ITO) coated glass substrate (Ossila, Sheffield, UK) was sequentially cleaned in an ultrasonic bath with soap water, deionized (DI) water, acetone, isopropanol and deionised (DI) water (Sigma Aldrich) for 15 min each and dried with nitrogen purge before being treated with UV-ozone for another 5 min. Next, a poly (3,4ethylenedioxythiophene) polystyrene sulfonate solution (PEDOT:PSS, PH1000 from H.C Stack, Newton, MA, USA) was filtered using a 0.45 µm nylon filter (LabServ, Derbyshire, UK) and spin-coated (Laurell, North Wales, PA, USA) at room temperature at 3000 rpm for 60 s on top of cleaned ITO substrate to form a 40 nm thin layer and annealed for 30 min at 130 • C as the hole transport layer. The photoactive layer was formed by spin-coating the blended solutions at 2000 rpm for 30 s on top of the PEDOT:PSS surface and annealed at 110 • C for 20 min. The sample was transferred to the thermal evaporator for the deposition of aluminum top contact under a base pressure of 2 × 10 −7 torr through a shadow mask that eventually formed a 0.045 cm 2 photoactive area. Finally, the samples were encapsulated with the glass using epoxy glue (Ossila) and cured under UV light. The schematic diagram of OPD is shown in Figure 1d. drich). In order to support this, the ultraviolet photoelectron spectroscopy (UPS) measurement of the BL3.2U was also conducted at Synchrotron Light Research Institute (SLRI, Nakhon Ratchasima, Thailand). Next, the current-voltage characteristics were obtained using Keithley 236 Source Measure Unit equipped with an Oriel (Xenon arc lamp) (Tektronix, Beaverton, OR, USA) solar simulator with 100 mW/cm 2 input power illumination, and the external quantum efficiency (EQE) was acquired using PVE300 Photovoltaic EQE (IPCE) system (Bentham, Berkshire, UK). The absorption spectra of the photoactive layers were observed with ultraviolet-visible (UV-Vis) spectroscopy (LAMBDA 900, Perkin Elmer, Waltham, MA, USA). Furthermore, characterizations of photoactive layers were observed using Raman and photoluminescence (PL) spectroscopy under 325nm excitation wavelength. Lastly, the Atomic Force Microscopy (AFM) images of the photoactive layers were also obtained using Hitachi AFM5100 using S1-DF3 cantilever (Hitachi, Tokyo, Japan).

Absorption Characteristics
In order to determine the energy bandgap of the VTTBNc, the cyclic voltammetry (CV) technique was used. The CV was set up using platinum as the counter electrode with Ag/AgCl as the reference electrode and ITO as the working electrode. The supporting electrolyte was prepared by using 0.01 M hydroquinone in HCl with pH 3.0 (Sigma Aldrich). In order to support this, the ultraviolet photoelectron spectroscopy (UPS) measurement of the BL3.2U was also conducted at Synchrotron Light Research Institute (SLRI, Nakhon Ratchasima, Thailand). Next, the current-voltage characteristics were obtained using Keithley 236 Source Measure Unit equipped with an Oriel (Xenon arc lamp) (Tektronix, Beaverton, OR, United States) solar simulator with 100 mW/cm 2 input power illumination, and the external quantum efficiency (EQE) was acquired using PVE300 Photovoltaic EQE (IPCE) system (Bentham, Berkshire, UK). The absorption spectra of the photoactive layers were observed with ultraviolet-visible (UV-Vis) spectroscopy (LAMBDA 900, Perkin Elmer, Waltham, MA, USA). Furthermore, characterizations of photoactive layers were observed using Raman and photoluminescence (PL) spectroscopy under 325nm excitation wavelength. Lastly, the Atomic Force Microscopy (AFM) images of the photoactive layers were also obtained using Hitachi AFM5100 using S1-DF3 cantilever (Hitachi, Tokyo, Japan). Figure 2 shows the absorption spectra of the single-component material and the blended films from the 300 nm to 900 nm wavelength region. Figure 2a illustrates the absorption spectra of the VTTBNc, PC 61 BM and PC 71 BM. It was observed that VTTBNc exhibits two broad peaks at B-band (333 nm) and Q-band region (761 nm and 810 nm). These absorption peaks are similar to the absorption spectra reported by Dhanya and Menon (2012) [20]. The absorption peak at the B-band correlates to the excitation of localized π-π* transition that occurred between the bonding and anti-bonding molecules. Meanwhile, the absorption peak at the Q-band is related to the charge transfer within the core unit [20,21]. Next, PC 61 BM reveals a maximum absorption peak located at 335 nm, and PC 71 BM has several shoulder peaks observed at 373 nm and 473 nm. Figure 2a also shows that PC 71 BM has a stronger and broader absorption range in the visible spectrum compared to PC 61 BM. The less symmetrical structure of PC 71 BM helps to relax the forbidden transitions in longer wavelengths and increases light absorption compared to PC 61 BM [22,23].

Absorption Characteristics
can be inferred that blends with PC71BM have a stronger absorption in the visible range compared to the blend with PC61BM. Additionally, both blends with a blend ratio of 1:0.50 exhibit high absorption at the Q-band region due to high VTTBNc content in the blend. VTTBNc:PC61BM shows almost the same peak intensity at the shorter wavelength for all ratios. However, the absorption peak at 328 nm in VTTBNc:PC71BM reduces with increasing PC71BM content. It is also evident from the two graphs that the absorption peak at 700 nm to 900 nm decreased with the increase in PC61BM and PC71BM content.

Determination of VTTBNc Energy Bandgap
The optical band gap is commonly determined using the UV-Vis absorption spectrum and the Tauc plot model [24]. The optical band gap energy can be determined by simply extrapolating a line to the linear region of the Tauc plot graph plotted from the absorption spectrum. From the Tauc plot shown in Figure 3a, the Q-band gap energy (Eg1) and B-band gap (Eg2) of VTTBNc are measured as 1.36 eV and 2.57 eV, respectively. This result is slightly low compared to the band gap reported by Dhanya and Menon (2012), which is between 1.7 eV and 1.8 eV of the Q-band gap and 3.1 eV and 3.4 eV of the B-band gap [20]. The difference in the band gap value is due to the red-shifted property in the absorption range observed in this experiment, resulting in lower band gap values. This property is good in enhancing the OPDs detection at the longer wavelength. Q-band gap energy (Eg1) was chosen due to half of the solar photon flux occurring at the low band gap energy and also essential in the development of NIR OPD [25].  VTTBNc:PC 61 BM shows almost the same peak intensity at the shorter wavelength for all ratios. However, the absorption peak at 328 nm in VTTBNc:PC 71 BM reduces with increasing PC 71 BM content. It is also evident from the two graphs that the absorption peak at 700 nm to 900 nm decreased with the increase in PC 61 BM and PC 71 BM content.

Determination of VTTBNc Energy Bandgap
The optical band gap is commonly determined using the UV-Vis absorption spectrum and the Tauc plot model [24]. The optical band gap energy can be determined by simply extrapolating a line to the linear region of the Tauc plot graph plotted from the absorption spectrum. From the Tauc plot shown in Figure 3a, the Q-band gap energy (E g1 ) and B-band gap (E g2 ) of VTTBNc are measured as 1.36 eV and 2.57 eV, respectively. This result is slightly low compared to the band gap reported by Dhanya and Menon (2012), which is between 1.7 eV and 1.8 eV of the Q-band gap and 3.1 eV and 3.4 eV of the B-band gap [20]. The difference in the band gap value is due to the red-shifted property in the absorption range observed in this experiment, resulting in lower band gap values. This property is good in enhancing the OPDs detection at the longer wavelength. Q-band gap energy (E g1 ) was chosen due to half of the solar photon flux occurring at the low band gap energy and also essential in the development of NIR OPD [25].  Next, the HOMO and LUMO levels of VTTBNc are determined by measuring the oxidation and reduction onset potential using CV measurement [26,27]. A proper alignment of HOMO-LUMO energy levels between donor and acceptor is essential for the electron transfer in the device. The HOMO and LUMO energy levels of VTTBNc is calculated using the equations below: where EOX and ERED refer to the oxidation and reduction onset potentials, respectively, and the 4.4 eV refers to the absolute potential of Ag/AgCl electrode relative to the vacuum level [28,29]. From the CV plot in Figure 3b, it is observed that VTTBNc altered the oxidation and reduction peaks when compared with the bare ITO. The oxidation and reduction onset potential of VTTBNc are observed at 0.742 V and −0.662 V, respectively. Thus, the HOMO and LUMO levels of VTTBNc are determined to be at −5.14 eV and −3.74 eV, respectively, with a band gap energy of 1.4 eV. The value of band gap energy is almost the same as the value recorded from UV-Vis (1.36 eV). This suggests a small interface barrier between the electrode and VTTBNc thin film in the CV system that gives rise to the equilibrium state occurring quickly [30,31]. Besides CV, the determination of energy levels can also be obtained using the ultraviolet photoelectron spectroscopy (UPS) technique by determining their HOMO level. From UPS spectra (Figure 4), the HOMOonset and cut-off energy of VTTBNc are observed at 0.70 eV and 35.06 eV, respectively. The HOMO level can be determined by utilizing the following equations: Work function ℎ Cuttoff energy (3)

HOMO Work function HOMO
where hv (39.5 eV) is the energy of an incident photon used in this experiment. By using the cut-off energy and the HOMOonset observed in the UPS spectra, the HOMO level is determined to be 5.14 eV. This result supports the HOMO level determined in CV as both methods obtained the same HOMO level value, which is 5.14 eV. Meanwhile, the HOMO and LUMO energy levels of PC61BM and PC71BM are obtained from the literature [32]. The work function and energy levels alignment for all materials are illustrated in Figure 5. Next, the HOMO and LUMO levels of VTTBNc are determined by measuring the oxidation and reduction onset potential using CV measurement [26,27]. A proper alignment of HOMO-LUMO energy levels between donor and acceptor is essential for the electron transfer in the device. The HOMO and LUMO energy levels of VTTBNc is calculated using the equations below: where E OX and E RED refer to the oxidation and reduction onset potentials, respectively, and the 4.4 eV refers to the absolute potential of Ag/AgCl electrode relative to the vacuum level [28,29]. From the CV plot in Figure 3b, it is observed that VTTBNc altered the oxidation and reduction peaks when compared with the bare ITO. The oxidation and reduction onset potential of VTTBNc are observed at 0.742 V and −0.662 V, respectively. Thus, the HOMO and LUMO levels of VTTBNc are determined to be at −5.14 eV and −3.74 eV, respectively, with a band gap energy of 1.4 eV. The value of band gap energy is almost the same as the value recorded from UV-Vis (1.36 eV). This suggests a small interface barrier between the electrode and VTTBNc thin film in the CV system that gives rise to the equilibrium state occurring quickly [30,31]. Besides CV, the determination of energy levels can also be obtained using the ultraviolet photoelectron spectroscopy (UPS) technique by determining their HOMO level. From UPS spectra (Figure 4), the HOMO onset and cut-off energy of VTTBNc are observed at 0.70 eV and 35.06 eV, respectively. The HOMO level can be determined by utilizing the following equations: where hv (39.5 eV) is the energy of an incident photon used in this experiment. By using the cut-off energy and the HOMO onset observed in the UPS spectra, the HOMO level is determined to be 5.14 eV. This result supports the HOMO level determined in CV as both methods obtained the same HOMO level value, which is 5.14 eV. Meanwhile, the HOMO and LUMO energy levels of PC 61 BM and PC 71 BM are obtained from the literature [32]. The work function and energy levels alignment for all materials are illustrated in Figure 5.  The low band gap could result in smaller energy offset between HOMO and LUMO of donor and acceptor material. This allows the reduction in open-circuit voltage loss by minimizing the loss due to the charge transfer state between donor and acceptor [17]. Therefore, the utilization of light absorption in the NIR region was expected to enhance the photocurrent generation of the device. It is also suitable to make transparent OPD with high transmittance in the visible regions by utilizing such high NIR absorbing material.   The low band gap could result in smaller energy offset between HOMO and LUMO of donor and acceptor material. This allows the reduction in open-circuit voltage loss by minimizing the loss due to the charge transfer state between donor and acceptor [17]. Therefore, the utilization of light absorption in the NIR region was expected to enhance the photocurrent generation of the device. It is also suitable to make transparent OPD with high transmittance in the visible regions by utilizing such high NIR absorbing material.

Performance of VTTBNc:PC61BM and VTTBNc:PC71BM Organic Photodiode
Figures 6 and 7 show the current density-voltage (J-V) graphs of single materials and the BHJ OPDs examined in dark and light conditions. The J-V characteristics are investigated in reversed biased conditions to study the dark current and photocurrent generation, which later determine the responsivity and detectivity of the OPDs. The responsivity  [17]. Therefore, the utilization of light absorption in the NIR region was expected to enhance the photocurrent generation of the device. It is also suitable to make transparent OPD with high transmittance in the visible regions by utilizing such high NIR absorbing material.  7 show the current density-voltage (J-V) graphs of single materials and the BHJ OPDs examined in dark and light conditions. The J-V characteristics are investigated in reversed biased conditions to study the dark current and photocurrent generation, which later determine the responsivity and detectivity of the OPDs. The responsivity of the device refers to the electrical output produced per incident power input and can be calculated using the following equation: where R is the responsivity, J ph is the photocurrent density, J D is the dark current density and P is the power of incident light. Meanwhile, detectivity (D*) is defined as the ability of the OPD to detect a small photon signal which is calculated by the equation below [10,33]: where q is the elementary charge (1.602 × 10 −19 C). From the equation above, the shot noise (2q.J D ) from dark current is often assumed to be the dominant contribution that limits the detectivity, while flicker noise and thermal noise are ignored [10,33]. [10,33]: * 2q.
where q is the elementary charge (1.602 × 10 −19 C). From the equation above, the sh (2q.JD) from dark current is often assumed to be the dominant contribution that li detectivity, while flicker noise and thermal noise are ignored [10,33]. As seen in Figure 6, it is observed that PC71BM generated a high photocurr leads to higher responsivity compared with PC61BM. The responsivity of PC71BM ured as 16.40 mA/W, 20.66 mA/W, 21.20 mA/W at 0 V, −0.5 V, −1 V bias, resp While the responsivity of PC61BM is measured as 10.50 mA/W, 14.32 mA/W, 16.2 at 0 V, −0.5 V, −1 V bias, respectively, as listed in Table 1. The higher photocurren ation in PC71BM devices is mainly due to the stronger absorption in the visible re observed in Figure 2. However, a high dark current was also observed, causing th to have lower detectivity compared to PC61BM. This may be attributed to the high generation of charge carriers at the contacts when the voltage is applied [33]. As obtained stronger absorption and higher photocurrent generation, it also contains charge carrier generation compared to PC61BM, and in turn, tends to have a high th generation of charge carriers from the two electrodes contact.   As seen in Figure 6, it is observed that PC 71 BM generated a high photocurrent that leads to higher responsivity compared with PC 61 BM. The responsivity of PC 71 BM is measured as 16.40 mA/W, 20.66 mA/W, 21.20 mA/W at 0 V, −0.5 V, −1 V bias, respectively. While the responsivity of PC 61 BM is measured as 10.50 mA/W, 14.32 mA/W, 16.21 mA/W at 0 V, −0.5 V, −1 V bias, respectively, as listed in Table 1. The higher photocurrent generation in PC 71 BM devices is mainly due to the stronger absorption in the visible region, as observed in Figure 2. However, a high dark current was also observed, causing the device to have lower detectivity compared to PC 61 BM. This may be attributed to the high thermal generation of charge carriers at the contacts when the voltage is applied [33]. As PC 71 BM obtained stronger absorption and higher photocurrent generation, it also contains a higher charge carrier generation compared to PC 61 BM, and in turn, tends to have a high thermally generation of charge carriers from the two electrodes contact. 10 /4.41 × 10 ) Jones, (2.37 × 10 /1.53 × 10 ) Jones and (2.31 × 10 /1.21 × 10 ) Jones at 0 V, −0.5 V and −1 V, respectively. Figure 8 depicts the trend in detectivity and responsivity of the optimized ratio of both blends with respect to bias. According to detectivity values obtained in Tables 1 and 2, it is evident that the addition of VTTBNc has greatly improved the detectivity limit in VTTBNc:PC71BM device to about 186% with respect to PC71BMonly device compared to VTTBNc:PC61BM device, which increases only around 19% with respect to the PC61BM-only device.  (JD = Dark current density, Jph = Photocurrent density, R = Responsivity, D*= Detectivity and VOC = Open-circuit voltage).   Whereas for the VTTBNc:PC 71 BM blend, the ratio of 1:1.0 showed the highest detectivity of 3.85 × 10 11 Jones and 2.31 × 10 9 Jones at 0 V and −1 V, respectively. On the other hand, ratio 1:1.5 exhibits lower detectivity of 8.48 × 10 10 Jones and 2.05 × 10 9 Jones despite having higher responsivity of 9.55 mA/W and 28.48 mA/W at 0 V and −1 V, respectively. The high detectivity of ratio 1:1.0 may be due to the low dark current density obtained. A low dark current is favorable in photodetection applications as it indicates low noise and high detectivity in the OPD device as high dark current will lead to higher device power consumption and may accommodate the photodiode to hardly detect weak signal for signal readout [10]. Thus, the device with a ratio of 1:1.0 was considered an optimized device in VTTBNc:PC 71 BM as it has the highest detectivity values.
Between the two blended devices, VTTBNc:PC 71 BM has the highest responsivity and detectivity values compared to VTTBNc:PC 61 BM. As seen in Table 2 respectively. In addition to that, the detectivity for both optimized devices are (3.85 × 10 11 / 4.41 × 10 9 ) Jones, (2.37 × 10 9 /1.53 × 10 9 ) Jones and (2.31 × 10 9 /1.21 × 10 9 ) Jones at 0 V, −0.5 V and −1 V, respectively. Figure 8 depicts the trend in detectivity and responsivity of the optimized ratio of both blends with respect to bias. According to detectivity values obtained in Tables 1 and 2, it is evident that the addition of VTTBNc has greatly improved the detectivity limit in VTTBNc:PC 71 BM device to about 186% with respect to PC 71 BMonly device compared to VTTBNc:PC 61 BM device, which increases only around 19% with respect to the PC 61 BM-only device.    Meanwhile, high detectivity in VTTBNc:PC71BM caused the photo-response to have a stable detection in dark/light state. The main contribution of high dark current in small molecule OPD is due to the carrier injection phenomenon from electrodes that are caused by the low band gap characteristic in small molecules [34].   Figure 9a, there are spikes or noise observed in VTTBNc:PC 61 BM (1:1 and 1:1.5) at the dark states while no photo-response is detected in 1:0.5 due to the weak signal and high noise detected from the experiment. These spikes might be due to the contribution of high dark current and low photocurrent generation. Meanwhile, high detectivity in VTTBNc:PC 71 BM caused the photo-response to have a stable detection in dark/light state. The main contribution of high dark current in small molecule OPD is due to the carrier injection phenomenon from electrodes that are caused by the low band gap characteristic in small molecules [34].  Meanwhile, high detectivity in VTTBNc:PC71BM caused the photo-response to have a stable detection in dark/light state. The main contribution of high dark current in small molecule OPD is due to the carrier injection phenomenon from electrodes that are caused by the low band gap characteristic in small molecules [34].

Photoluminescence and Raman Spectra
Furthermore, the photoluminescence behavior of the optimized ratio for both blends was observed under an excitation wavelength of 325 nm. Figure 10a shows the PL spectra of the optimized ratio of the blends; VTTBNc:PC61BM (1:1.5) and VTTBNc:PC71BM (1:1.0). Generally, PL measures the radiative illumination of a material. The intensity of PL spectra provides information on the charge carrier's recombination rate in order to obtain an effective charge transfer within the blended film [35]. It is evident that VTTBNc:PC71BM blend film has lower PL intensity than VTTBNc:PC61BM, especially in the range of 400-700 nm and 930-1150 nm. The PL spectra of VTTBNc:PC71BM are quenched around 22% at 600 nm compared to VTTBNc:PC61BM. The percentage of the quenching is calculated based on the difference in the PL intensity of both blends with respect to the VTT-BNc:PC61BM. This quenching generally suggests that VTTBNc:PC71BM has more effective photo-induced charge transfer in the blended film and demonstrates a lower recombination rate of the charge carriers [36].
Furthermore, Raman spectroscopy was used to observe the molecular vibration mode in different blended films. In this characterization, the poor and high-performance blend ratio from VTTBNc:PC71BM and VTTBNc:PC61BM OPDs was observed. Figure 10b shows the Raman spectra of VTTBNc:PC71BM (1:0.5 and 1:1.0) and VTTBNc:PC61BM (1:0.5 and 1:1.5). The assignments of the vibrational mode recorded in Table 4 were performed based on the literature [37]. It is observed that VTTBNc:PC71BM exhibits a sharp peak at 1566 cm −1 in both ratios, which refer to C=C stretch mode. Meanwhile, most peaks in VTT-BNc:PC61BM have lower intensity and broader peaks. The result also shows two sharp peaks in VTTBNc:PC71BM for both ratios that refer to C-H bending mode. From the literature, the sharp peak in the Raman spectra is often associated with the crystallinity and ordered structure of the film [35,38]. Therefore, VTTBNc:PC71BM films can be assumed to garner more crystalline and ordered structure compared to VTTBNc:PC61BM films. The

Photoluminescence and Raman Spectra
Furthermore, the photoluminescence behavior of the optimized ratio for both blends was observed under an excitation wavelength of 325 nm. Figure 10a shows the PL spectra of the optimized ratio of the blends; VTTBNc:PC 61 BM (1:1.5) and VTTBNc:PC 71 BM (1:1.0). Generally, PL measures the radiative illumination of a material. The intensity of PL spectra provides information on the charge carrier's recombination rate in order to obtain an effective charge transfer within the blended film [35]. It is evident that VTTBNc:PC 71 BM blend film has lower PL intensity than VTTBNc:PC 61 BM, especially in the range of 400-700 nm and 930-1150 nm. The PL spectra of VTTBNc:PC 71 BM are quenched around 22% at 600 nm compared to VTTBNc:PC 61 BM. The percentage of the quenching is calculated based on the difference in the PL intensity of both blends with respect to the VTTBNc:PC 61 BM. This quenching generally suggests that VTTBNc:PC 71 BM has more effective photo-induced charge transfer in the blended film and demonstrates a lower recombination rate of the charge carriers [36].
Furthermore, Raman spectroscopy was used to observe the molecular vibration mode in different blended films. In this characterization, the poor and high-performance blend ratio from VTTBNc:PC 71 BM and VTTBNc:PC 61 BM OPDs was observed. Figure 10b shows the Raman spectra of VTTBNc:PC 71 BM (1:0.5 and 1:1.0) and VTTBNc:PC 61 BM (1:0.5 and 1:1.5). The assignments of the vibrational mode recorded in Table 4 were performed based on the literature [37]. It is observed that VTTBNc:PC 71 BM exhibits a sharp peak at 1566 cm −1 in both ratios, which refer to C=C stretch mode. Meanwhile, most peaks in VTTBNc:PC 61 BM have lower intensity and broader peaks. The result also shows two sharp peaks in VTTBNc:PC 71 BM for both ratios that refer to C-H bending mode. From the literature, the sharp peak in the Raman spectra is often associated with the crystallinity and ordered structure of the film [35,38]. Therefore, VTTBNc:PC 71 BM films can be assumed to garner more crystalline and ordered structure compared to VTTBNc:PC 61 BM films. The high crystalline structure in VTTBNc:PC 71 BM supports its high OPD performance due to high charge carrier mobilities that are required for efficient charge extraction [39,40]. On the other side, Raman intensity at 1566 cm −1 for VTTBNc:PC 71 BM and 1467 cm −1 for VTTBNc:PC 61 BM for ratio 1:0.5 is low compared to other ratios (1:1.0 and 1:1.5) due to the lower content of PC 71 BM and PC 61 BM molecules.
Micromachines 2021, 12, x FOR PEER REVIEW 12 of 16 high crystalline structure in VTTBNc:PC71BM supports its high OPD performance due to high charge carrier mobilities that are required for efficient charge extraction [39,40]. On the other side, Raman intensity at 1566 cm −1 for VTTBNc:PC71BM and 1467 cm −1 for VTT-BNc:PC61BM for ratio 1:0.5 is low compared to other ratios (1:1.0 and 1:1.5) due to the lower content of PC71BM and PC61BM molecules.

Morphological and Surface Roughness Observations
The morphology and surface roughness of VTTBNc:PC61BM and VTTBNc:PC71BM thin films were investigated using AFM analysis with scan sizes of 5 × 5 μm and 0.5 × 0.5 μm. As shown in Table 5, four blend ratios with poor and optimized performance from VTTBNc:PC61BM (1:0.5 and 1:1.5) and VTTBNc:PC71BM (1:0.5 and 1:1.0) were chosen. From the AFM images with 0.5 × 0.5μm scan size, the ratio with higher n-type content exhibits higher surface roughness values which are 0.546 nm for VTTBNc:PC61BM (1:1.5) and 0.849 nm for VTTBNc:PC71BM (1:1.0). Meanwhile, the surface roughness for the low n-type content (1:0.5) is found to be 0.284 nm for VTTBNc:PC61BM and 0.461 nm for VTT-BNc:PC71BM. As seen in Table 5 with a 5 × 5 μm scan size, the clusters are more visible in a blend with higher n-type content. The presence of clusters might increase the surface roughness values. According to L. Benatto et al. (2020), high surface roughness can enhance the donor-acceptor contact area and assist the charge carrier diffusion in the interpenetrating layer [23]. Besides, the n-types used in the blended film also affect the surface roughness values. The key difference between PC61BM and PC71BM is the axial distance

Morphological and Surface Roughness Observations
The morphology and surface roughness of VTTBNc:PC 61 BM and VTTBNc:PC 71 BM thin films were investigated using AFM analysis with scan sizes of 5 × 5 µm and 0.5 × 0.5 µm. As shown in Table 5, four blend ratios with poor and optimized performance from VTTBNc:PC 61 BM (1:0.5 and 1:1.5) and VTTBNc:PC 71 BM (1:0.5 and 1:1.0) were chosen. From the AFM images with 0.5 × 0.5µm scan size, the ratio with higher n-type content exhibits higher surface roughness values which are 0.546 nm for VTTBNc:PC 61 BM (1:1.5) and 0.849 nm for VTTBNc:PC 71 BM (1:1.0). Meanwhile, the surface roughness for the low n-type content (1:0.5) is found to be 0.284 nm for VTTBNc:PC 61 BM and 0.461 nm for VTTBNc:PC 71 BM. As seen in Table 5 with a 5 × 5 µm scan size, the clusters are more visible in a blend with higher n-type content. The presence of clusters might increase the surface roughness values. According to L. Benatto et al. (2020), high surface roughness can enhance the donor-acceptor contact area and assist the charge carrier diffusion in the interpenetrating layer [23]. Besides, the n-types used in the blended film also affect the surface roughness values. The key difference between PC 61 BM and PC 71 BM is the axial distance for fullerene moiety. The axial distance for PC 61 BM was reported to be 0.70 nm and for PC 71 BM was 0.80 nm [41]. The axial distance of fullerene moiety is the length of an allotrope of carbon, and this difference in the axial distance results in different surface roughness values. It can be assumed that the larger axial distance in fullerene moiety contributes to the larger cluster size on the surface. Some of the studies also emphasize the importance of the cluster in order to have charge delocalization and consequently reduce the recombination of charge carriers [42,43]. for fullerene moiety. The axial distance for PC61BM was reported to be 0.70 nm and for PC71BM was 0.80 nm [41]. The axial distance of fullerene moiety is the length of an allotrope of carbon, and this difference in the axial distance results in different surface roughness values. It can be assumed that the larger axial distance in fullerene moiety contributes to the larger cluster size on the surface. Some of the studies also emphasize the importance of the cluster in order to have charge delocalization and consequently reduce the recombination of charge carriers [42,43].  for fullerene moiety. The axial distance for PC61BM was reported to be 0.70 nm and for PC71BM was 0.80 nm [41]. The axial distance of fullerene moiety is the length of an allotrope of carbon, and this difference in the axial distance results in different surface roughness values. It can be assumed that the larger axial distance in fullerene moiety contributes to the larger cluster size on the surface. Some of the studies also emphasize the importance of the cluster in order to have charge delocalization and consequently reduce the recombination of charge carriers [42,43].  for fullerene moiety. The axial distance for PC61BM was reported to be 0.70 nm and for PC71BM was 0.80 nm [41]. The axial distance of fullerene moiety is the length of an allotrope of carbon, and this difference in the axial distance results in different surface roughness values. It can be assumed that the larger axial distance in fullerene moiety contributes to the larger cluster size on the surface. Some of the studies also emphasize the importance of the cluster in order to have charge delocalization and consequently reduce the recombination of charge carriers [42,43]. for fullerene moiety. The axial distance for PC61BM was reported to be 0.70 nm and for PC71BM was 0.80 nm [41]. The axial distance of fullerene moiety is the length of an allotrope of carbon, and this difference in the axial distance results in different surface roughness values. It can be assumed that the larger axial distance in fullerene moiety contributes to the larger cluster size on the surface. Some of the studies also emphasize the importance of the cluster in order to have charge delocalization and consequently reduce the recombination of charge carriers [42,43]. for fullerene moiety. The axial distance for PC61BM was reported to be 0.70 nm and for PC71BM was 0.80 nm [41]. The axial distance of fullerene moiety is the length of an allotrope of carbon, and this difference in the axial distance results in different surface roughness values. It can be assumed that the larger axial distance in fullerene moiety contributes to the larger cluster size on the surface. Some of the studies also emphasize the importance of the cluster in order to have charge delocalization and consequently reduce the recombination of charge carriers [42,43]. for fullerene moiety. The axial distance for PC61BM was reported to be 0.70 nm and for PC71BM was 0.80 nm [41]. The axial distance of fullerene moiety is the length of an allotrope of carbon, and this difference in the axial distance results in different surface roughness values. It can be assumed that the larger axial distance in fullerene moiety contributes to the larger cluster size on the surface. Some of the studies also emphasize the importance of the cluster in order to have charge delocalization and consequently reduce the recombination of charge carriers [42,43]. for fullerene moiety. The axial distance for PC61BM was reported to be 0.70 nm and for PC71BM was 0.80 nm [41]. The axial distance of fullerene moiety is the length of an allotrope of carbon, and this difference in the axial distance results in different surface roughness values. It can be assumed that the larger axial distance in fullerene moiety contributes to the larger cluster size on the surface. Some of the studies also emphasize the importance of the cluster in order to have charge delocalization and consequently reduce the recombination of charge carriers [42,43]. for fullerene moiety. The axial distance for PC61BM was reported to be 0.70 nm and for PC71BM was 0.80 nm [41]. The axial distance of fullerene moiety is the length of an allotrope of carbon, and this difference in the axial distance results in different surface roughness values. It can be assumed that the larger axial distance in fullerene moiety contributes to the larger cluster size on the surface. Some of the studies also emphasize the importance of the cluster in order to have charge delocalization and consequently reduce the recombination of charge carriers [42,43].  Figure 11 shows the EQE spectrum for the optimized OPDs for VTTBNc:PC 61 BM (1:1.5) and VTTBNc:PC 71 BM (1:1.0) as a function of wavelength. This characterization describes the photo-conversion capability of the OPD as the EQE is defined as the ratio between the numbers of charge carriers generated to the numbers of incident photons. It is observed that both OPDs exhibit the highest EQE peak at 400 nm, which is 2.74% for VTTBNc:PC 71 BM and 1.03% for VTTBNc:PC 61 BM. The low percentage of EQE may be due to the low charge carrier generated by the photoactive layer. The low charge carrier generated explained the low photocurrent density measured of both blended devices. The lower EQE values in the NIR region may be due to low light absorption, which lowers the charge carrier generation in the NIR region compared to the visible region.  Figure 11 shows the EQE spectrum for the optimized OPDs for VTTBNc:PC61BM (1:1.5) and VTTBNc:PC71BM (1:1.0) as a function of wavelength. This characterization describes the photo-conversion capability of the OPD as the EQE is defined as the ratio between the numbers of charge carriers generated to the numbers of incident photons. It is observed that both OPDs exhibit the highest EQE peak at 400 nm, which is 2.74% for VTT-BNc:PC71BM and 1.03% for VTTBNc:PC61BM. The low percentage of EQE may be due to the low charge carrier generated by the photoactive layer. The low charge carrier generated explained the low photocurrent density measured of both blended devices. The lower EQE values in the NIR region may be due to low light absorption, which lowers the charge carrier generation in the NIR region compared to the visible region.

Conclusions
In conclusion, the fabrication of OPDs using VTTBNc as a donor and PC61BM and PC71BM as acceptors were demonstrated. OPD performances were investigated along with the study on the optical, structural and morphological properties of VTTBNc:PC71BM and VTTBNc:PC61BM of different blend ratios. From the characterizations, we conclude that the optimized ratio obtained for VTTBNc:PC71BM and VTTBNc:PC61BM blend film is 1:1.0 and 1:1.5, respectively. VTTBNc:PC71BM and VTTBNc:PC61BM obtained detectivity of 2.31 × 10 9 Jones and 1.21 × 10 9 Jones with responsivity values of 26.11 mA/W and 10.89 mA/W, respectively. Both optimized blends obtained acceptable response and recovery time in milliseconds. The response and recovery time in optimized VTTBNc:PC71BM were found to be 241 ms and 310 ms, while in optimized VTTBNc:PC61BM is 516 ms and 481 ms, respectively. Comprehensively, the VTTBNc:PC71BM (1:1.0) has the highest performance among other OPD devices as it has considerably high photocurrent generation compared to other studied devices.

Conclusions
In conclusion, the fabrication of OPDs using VTTBNc as a donor and PC 61 BM and PC 71 BM as acceptors were demonstrated. OPD performances were investigated along with the study on the optical, structural and morphological properties of VTTBNc:PC 71 BM and VTTBNc:PC 61 BM of different blend ratios. From the characterizations, we conclude that the optimized ratio obtained for VTTBNc:PC 71 BM and VTTBNc:PC 61 BM blend film is 1:1.0 and 1:1.5, respectively. VTTBNc:PC 71 BM and VTTBNc:PC 61 BM obtained detectivity of 2.31 × 10 9 Jones and 1.21 × 10 9 Jones with responsivity values of 26.11 mA/W and 10.89 mA/W, respectively. Both optimized blends obtained acceptable response and recovery time in milliseconds. The response and recovery time in optimized VTTBNc:PC 71 BM were found to be 241 ms and 310 ms, while in optimized VTTBNc:PC 61 BM is 516 ms and 481 ms, respectively. Comprehensively, the VTTBNc:PC 71 BM (1:1.0) has the highest performance among other OPD devices as it has considerably high photocurrent generation compared to other studied devices.

Conflicts of Interest:
The authors declare no conflict of interest.