Improved Performance of Ternary Solar Cells by Using BODIPY Triads

Two boron dipyrromethene (BODIPY) triads, namely BODIPY-1 and BODIPY-2, were synthesized and incorporated with poly-3-hexyl thiophene: (6,6)-phenyl-C61-butyric acid methyl ester (PCBM) P3HT:PCBM. The photovoltaic performance of BODIPY:P3HT:PCBM ternary solar cells was increased, as compared to the control binary solar cells (P3HT:PCBM). The optimized power conversion efficiency (PCE) of BODIPY-1:P3HT:PCBM was improved from 2.22% to 3.43%. The enhancement of PCE was attributed to cascade charge transfer, an improved external quantum efficiency (EQE) with increased short circuit current (Jsc), and more homogeneous morphology in the ternary blend.


Introduction
A power conversion efficiency (PCE) of bulk heterojunction solar cells of over 13% has been achieved through remarkable progress in the development of narrow gap polymers and non-fullerene acceptors [1][2][3][4][5][6][7][8][9][10][11]. Nevertheless, a single polymer cannot absorb the entire electromagnetic range of the sun spectrum, thus, limiting the amount of photon harvesting capabilities of those solar cells. To solve this issue, tandem solar cells, composed of stacking layers of different chromophores that display complementary absorption to each other, have been demonstrated [1,12,13]. Although tandem cells can improve efficiency up to 17% [14,15], their fabrication process is still complicated and comparatively expensive [12,[16][17][18].
Another promising solar cell that can retain the advantages from of bulk heterojunction solar cells and tandem solar cells is the ternary solar cells [19][20][21]. A ternary system offers several advantages including: (a) Light harvesting can be improved by the complementary absorption of the third component. (b) Photovoltaic parameters, short circuit current (J sc ), open circuit voltage (V oc ), and fill factor (FF), can be optimized by varying the ratio of donor or acceptor moieties. (c) The morphology and phase separation of the active layer, which play a key role in determining the efficiency, can be adjusted and controlled. (d) The fabrication method is easier than those of multi-layer or tandem solar cells [17,[21][22][23][24]. To date, ternary solar cells provide a PCE of up to 12% [21]. Typically, an active layer of ternary solar cells contains a donor (D), an acceptor (A), and a third component. The third component can be either a polymer or a small organic compound that serves as a donor or acceptor. Thereby, ternary solar cells are classified as two donors/one acceptor (D1/D2/A) or and one donor/two acceptors (D/A1/A2) [17,22,25,26]. It has been reported that ternary solar cells based on polymer/small (BODIPY-1) and CBZ-BODIPY-CBZ (BODIPY-2) were mixed with P3HT with different weight ratios. BODIPY:P3HT donors were blended with fullerene derivative (6,6)-phenyl-C61-butyric acid methyl ester (PC 61 BM), an acceptor, to fabricate ternary solar cells. The photophysical properties of the blended active material were studied. The solar cell efficiency was investigated and the influence of blending BODIPY on the photovoltaic parameter, the external quantum efficiency (EQE) and the surface morphology were discussed.

Solar Cell Fabrication
The patterned ITO substrates (25 mm × 20 mm) were cleaned with commercial detergent, Alconox solution (10 wt.%), isopropanol, and DI water under sonication for 30 min, respectively. Then, the ITO substrates were treated with UV light (UV-LED, 365 nm, Larson Electronics, Kemp, TX, USA) for 30 min before use. The 80 µL of filtered PEDOT: PSS was deposited onto the ITO substrates by using the spin coating technique at 3000 rpm for 40 s. After annealing at 140 • C for 30 min in air, the substrates were quickly transferred to a glove box filled with nitrogen gas (RH~10%). The 50 µL of P3HT (and/or BODIPY-1, BODIPY-2): PCBM (1:1 by wt.) (20 mg/mL in chloroform) solution was spin coated onto the PEDOT: PSS layer at 2500 rpm for 30 s. Then, the 60 µL of TiO 2 solution was coated on active layer at 4000 rpm for 30 s. The coated substrates were annealed at 110 • C for 10 min and then transferred to thermal evaporator to deposit the aluminum cathode (film thickness~90-110 nm). The active area of the device was 4 mm 2 .

Instrumentations
1 H-NMR spectra were recorded on a 400 MHz NMR spectrometer (Ascend 400, Bruker BioSpin GmBH, Rheinstetten, Germany) and were reported in ppm using the residual proton resonance of the solvent as the internal standard (CDCl 3 at 7.26 ppm). 13 C-NMR spectra were proton decoupled and recorded on a 100 MHz Bruker spectrometer using the carbon signal of the deuterated solvent as the internal standard. Thermogravimetric analysis (TGA) of BODIPY-1 and BODIPY-2 were measured using a thermogravimetric analyzer (Diamond TA/Q50, TA instruments, New Castle, DE, USA). The absorption spectra of BODIPY and P3HT solution were recorded on a Thermo Scientific UV-Genesys 10 s spectrophotometer. The solid state absorption spectra of BODIPY and P3HT and blended films were recorded on a UV-Vis-NIR spectrophotometer (SolidSpec-3700, Shimadzu, Kyoto, Japan). Fluorescence spectra were recorded on a PerkinElmer LS 55 fluorescence spectrometer. Electrochemical Impedance spectroscopy (EIS) (PGSTAT302, Metrohm Autolab B. V., Utrecht, the Natherlands) was conducted on an AUTOLAB. The surface morphology of thin films were characterized using an atomic force microscope (AFM) (NaioAFM, Nanosurf, Liestal, Switzerland). Contact angles were measured by a static optical contact angle meter (KINO SL150E). The current density-voltage (J-V) curves were measured using a Keithley 2400 source-measurement unit and a Newport 94011A solar simulator. The external quantum efficiency (EQE) measurement was performed on a Newport Oriel QEPVSI-b under the irradiation of a Xenon lamp equipped with a monochromator controlled via USB connection through TracQ software (version 6.7, Newport Corporation, Irvine, CA, USA).
Materials 2020, 13, x FOR PEER REVIEW 5 of 15 were confirmed using 1 H-NMR and 13 C-NMR, respectively (see Supplementary Materials, Figure S1-S6). Thermal gravimetric analysis was carried out to investigate the thermal degradation. The 2.5% decomposition temperatures (Td) of BODIPY-1 and BODIPY-2 were 200 °C and 225 °C ( Figure S7), respectively, indicating that they are stable for use as the active layers of solar cells [52].

Optical Characterizations
The absorption spectra of BODIPY-1 and BODIPY-2 exhibited narrow absorption with absorption maxima at 537 and 532 nm, and high extinction coefficients of ∼3 × 10 4 and 8 × 10 4 M −1 cm −1 , respectively ( Figure 2a). P3HT showed a broader spectrum, with absorption maxima at 453 nm, while the solid-state absorption of the P3HT was red-shifted by of ∼50 nm with respect to that of the P3HT in solution (Figure 2b). The absorption spectra of BODIPY blended P3HT films are a combination of the individual spectra BODIPY and P3HT, indicating cooperative of light harvesting.  Table 1. The control device fabricated from

Optical Characterizations
The absorption spectra of BODIPY-1 and BODIPY-2 exhibited narrow absorption with absorption maxima at 537 and 532 nm, and high extinction coefficients of~3 × 10 4 and 8 × 10 4 M −1 cm −1 , respectively ( Figure 2a). P3HT showed a broader spectrum, with absorption maxima at 453 nm, while the solid-state absorption of the P3HT was red-shifted by of~50 nm with respect to that of the P3HT in solution ( Figure 2b). The absorption spectra of BODIPY blended P3HT films are a combination of the individual spectra BODIPY and P3HT, indicating cooperative of light harvesting. were confirmed using 1 H-NMR and 13 C-NMR, respectively (see Supplementary Materials, Figure S1-S6). Thermal gravimetric analysis was carried out to investigate the thermal degradation. The 2.5% decomposition temperatures (Td) of BODIPY-1 and BODIPY-2 were 200 °C and 225 °C ( Figure S7), respectively, indicating that they are stable for use as the active layers of solar cells [52].

Optical Characterizations
The absorption spectra of BODIPY-1 and BODIPY-2 exhibited narrow absorption with absorption maxima at 537 and 532 nm, and high extinction coefficients of ∼3 × 10 4 and 8 × 10 4 M −1 cm −1 , respectively ( Figure 2a). P3HT showed a broader spectrum, with absorption maxima at 453 nm, while the solid-state absorption of the P3HT was red-shifted by of ∼50 nm with respect to that of the P3HT in solution (Figure 2b). The absorption spectra of BODIPY blended P3HT films are a combination of the individual spectra BODIPY and P3HT, indicating cooperative of light harvesting.  Table 1. The control device fabricated from

Photovoltaic Performance
Binary and ternary solar cells were fabricated in a regular architecture (ITO/PEDOT:PSS/ BODIPY:P3HT:PC 61 BM/TiO 2 /Al) under the same conditions. The current-voltage (J-V) characteristic curves of solar cells are presented in Figure 3c,d and the photovoltaic performance parameters including short circuit current density (J sc ), open circuit voltage (V oc ), fill factor (FF) and power conversion efficiency (PCE) are summarized in Table 1. The control device fabricated from P3HT:PCBM provided a PCE of 2.22%, with a J sc of 6.77 mA/cm 2 , a V oc of 0.63 V, and an FF of 0.52. Binary solar devices fabricated from BODIPY-1:PCBM and BODIPY-2:PCBM have PCEs of 1.85% and 1.92%, respectively. The lower performance of BODIPY:PCBM solar cells is attributed to the narrow absorption properties of BODIPY, resulting in lower J sc values. The photovoltaic performance of ternary solar cells were optimized by varying weight ratio of P3HT and BOIDPY. We found that incorporating a lower quantity of BODIPY in P3HT:PCBM led to a reduction in the performance of devices (see Table S1). Interestingly, increasing BODIPY concentration in P3HT:PCBM significantly enhanced J sc values, leading to the improvement of device efficiency as compared to those of P3HT:PCBM or BODIPY:PCBM binary solar cells (Table 1 and Figure 3e). Ternary solar cells containing BODIPY-1:P3HT (0.6:0.4) exhibited the highest PCE (3.43%) due to a large enhancement in J sc (9.45 mA/cm 2 ) and an increase in FF (0.54). The efficiency was raised by 55% as compared to the control P3HT device. Incorporating BODIPY-2 with P3HT also improved the solar cell performance. The device fabricated with BODIPY-2:P3HT (0.7:0.3) provided a PCE of 3.20%, with the highest J sc value of 9.96 mA/cm 2 . However, the V oc values of the champion devices are slightly increased as compared to the P3HT device. In general, the V oc of the P3HT:PCBM solar cells is limited by a high voltage loss (∆E loss ), in a range of 1.10-1.35 eV [11,[53][54][55][56]. This loss is caused by three factors: (i) the radiative recombination from the absorption above the bandgap (~0.30 eV), (ii) an additional radiative recombination from the absorption below the bandgap (>0.6 eV), and (iii) the non-radiative recombination (>0.3 eV) [11,[53][54][55][56]. In this work, P3HT:PCBM binary solar cell has a voltage loss of 1.26 eV. Whereas, the voltages loss values of BODIPY-1:P3HT (0.6:0.4) and BODIPY-2:P3HT (0.7:0.3) ternary solar cells are 1.12 eV and 1.10 eV, respectively. It should be noted that PCE of champion cells are higher than those of reported small BODIPY molecule used in ternary solar cells [48,57]. This suggested that our BODIPY triads and P3HT are a suitable combination to improve the efficiency of ternary solar cells.

External Quantum Efficiency (EQE)
The EQE spectra of binary solar cell and ternary solar cell are depicted in Figure 4. The EQE of P3HT device was 35% at 500 nm, while the EQE of BODIPY-1 and BOIDPY-2 devices were 20% at wavelengths of 400 and 550 nm. For the ternary blend system, the EQE of the ternary solar cells were higher than those of BODIPY and P3HT binary solar cells. The EQE spectra for BODIPY-1:P3HT (0.6:0.4) and BODIPY-2:P3HT (0.7:0.3) illustrated that EQE in BODIPY and P3HT absorption were increased up to 40% and covered in a broader range from 350 to 700 nm. In addition, the trend in EQE spectra and increase in the photocurrent generation are in good agreement.

External Quantum Efficiency (EQE)
The EQE spectra of binary solar cell and ternary solar cell are depicted in Figure 4. The EQE of P3HT device was 35% at 500 nm, while the EQE of BODIPY-1 and BOIDPY-2 devices were 20% at wavelengths of 400 and 550 nm. For the ternary blend system, the EQE of the ternary solar cells were higher than those of BODIPY and P3HT binary solar cells. The EQE spectra for BODIPY-1:P3HT (0.6:0.4) and BODIPY-2:P3HT (0.7:0.3) illustrated that EQE in BODIPY and P3HT absorption were increased up to 40% and covered in a broader range from 350 to 700 nm. In addition, the trend in EQE spectra and increase in the photocurrent generation are in good agreement.

Impedance Analysis
It is well known that the suppression of dark saturation current density (J0) can increase the Voc value of organic photovoltaics (OPVs) as a result of the reduction in charge recombination at the interface between the active layer and the anode/cathode layer [58,59]. Electrochemical impedance spectroscopy (EIS) was used to investigate the charge recombination process, which is directly related to the Voc in OPV devices. EIS was performed over a frequency range from 0.01 Hz to 1 MHz at an amplitude of 10 mV. The EIS spectra were fitted by using an equivalent circuit model, as shown in the inset of Figure 5. Rs, Rrec, and CPE represented the series resistance, the charge recombination resistance, and the constant phased element representing the chemical capacitance, respectively [60,61]. Figure 5 showed Nyquist plots of binary solar cells and ternary solar cells under 1 sun simulation. The semicircle represented the charge recombination resistance (Rrec) of the solar cell. It was found that the radius of the semicircle of the Nyquist plot was increased when BODIPY triads were applied in the ternary system. This suggests that charge recombination was significantly suppressed, resulting in the increase in Voc that was observed in champion devices. In addition, we found that the P value of CPE, simulated by an EIS software analyzer (Table S2, increased from 0.70 (P3HT device) to 0.90 (BODIPY-1:P3HT (0.6:0.4) device). This P value of the champion device is close to the P-value of an ideal capacitor (P = 1) without grain boundary and/or defect [59]. Since voltage loss values of BODIPY-1:P3HT and BODIPY-2:P3HT devices were similar (1.1 eV), a higher PCE of BODIPY-1:P3HT could be due to efficient charge transport in the active layer. This suggests that the active layer film formed by the ternary blended system has low charge trapping sites, originating from defects such as pin-holes or the grain boundaries of the film, and corresponds to the atomic force microscope (AFM) results), showing that the quality of the active film was higher and had greater homogeneity when BODIPY triads were applied in the active layer.

Impedance Analysis
It is well known that the suppression of dark saturation current density (J 0 ) can increase the V oc value of organic photovoltaics (OPVs) as a result of the reduction in charge recombination at the interface between the active layer and the anode/cathode layer [58,59]. Electrochemical impedance spectroscopy (EIS) was used to investigate the charge recombination process, which is directly related to the V oc in OPV devices. EIS was performed over a frequency range from 0.01 Hz to 1 MHz at an amplitude of 10 mV. The EIS spectra were fitted by using an equivalent circuit model, as shown in the inset of Figure 5. R s , R rec , and CPE represented the series resistance, the charge recombination resistance, and the constant phased element representing the chemical capacitance, respectively [60,61]. Figure 5 showed Nyquist plots of binary solar cells and ternary solar cells under 1 sun simulation. The semicircle represented the charge recombination resistance (R rec ) of the solar cell. It was found that the radius of the semicircle of the Nyquist plot was increased when BODIPY triads were applied in the ternary system. This suggests that charge recombination was significantly suppressed, resulting in the increase in V oc that was observed in champion devices. In addition, we found that the P value of CPE, simulated by an EIS software analyzer (Table S2, increased from 0.70 (P3HT device) to 0.90 (BODIPY-1:P3HT (0.6:0.4) device). This P value of the champion device is close to the P-value of an ideal capacitor (P = 1) without grain boundary and/or defect [59]. Since voltage loss values of BODIPY-1:P3HT and BODIPY-2:P3HT devices were similar (1.1 eV), a higher PCE of BODIPY-1:P3HT could be due to efficient charge transport in the active layer. This suggests that the active layer film formed by the ternary blended system has low charge trapping sites, originating from defects such as pin-holes or the grain boundaries of the film, and corresponds to the atomic force microscope (AFM) results), showing that the quality of the active film was higher and had greater homogeneity when BODIPY triads were applied in the active layer.

Morphology Study
Morphology of the photoactive layer is known as a critical factor that optimizes photon/electron conversion efficiency of ternary solar cells [62][63][64][65]. To determine surface roughness and phase behavior of the blended active layer, AFM has been investigated in tapping mode (Figure 6a-e). For an accurate comparison, all films were prepared with the same conditions for solar cell device fabrication. It is clearly seen that BODIPY-1:PCBM and BODIPY-2:PCBM films showed some void (pin-holes) and microphase separation (Figure 6a,b). The surface roughness of BODIPY blended with P3HT:PCBM exhibited lower root mean square of average height (RMS) values compared to those of the binary blends (Figure 6d,e). This indicated that blending BODIPY and P3HT yields a flatter and smoother surface. The RMS values of BODIPY-1 blended P3HT:PCBM and BODIPY-2 blended P3HT:PCBM were 0.94 nm and 1.61 nm, respectively. It should be noted that the champion device, BODIPY-1:P3HT:PCBM (0.7:0.3:1) developed a more homogeneous film with an improved morphology over the binary counterpart device. This could result in more efficient charge transport at the interfacial of the donor/acceptor area and enhance the PCE [59,[66][67][68].
The ω value of BODIPY-2 was −1.96, which suggests that BODIPY-2 molecules are likely to be located in the PCBM domain.

Morphology Study
Morphology of the photoactive layer is known as a critical factor that optimizes photon/electron conversion efficiency of ternary solar cells [62][63][64][65]. To determine surface roughness and phase behavior of the blended active layer, AFM has been investigated in tapping mode (Figure 6a-e). For an accurate comparison, all films were prepared with the same conditions for solar cell device fabrication. It is clearly seen that BODIPY-1:PCBM and BODIPY-2:PCBM films showed some void (pin-holes) and microphase separation (Figure 6a,b). The surface roughness of BODIPY blended with P3HT:PCBM exhibited lower root mean square of average height (RMS) values compared to those of the binary blends (Figure 6d,e). This indicated that blending BODIPY and P3HT yields a flatter and smoother surface. The RMS values of BODIPY-1 blended P3HT:PCBM and BODIPY-2 blended P3HT:PCBM were 0.94 nm and 1.61 nm, respectively. It should be noted that the champion device, BODIPY-1:P3HT:PCBM (0.7:0.3:1) developed a more homogeneous film with an improved morphology over the binary counterpart device. This could result in more efficient charge transport at the interfacial of the donor/acceptor area and enhance the PCE [59,[66][67][68].
If ω < −1, BODIPY will be located in the PCBM domain. We found that the ω value of BODIPY-1 was 0.25, which indicates that BODIPY-1 molecules preferentially locate at the P3HT/PCBM interface.
The ω value of BODIPY-2 was −1.96, which suggests that BODIPY-2 molecules are likely to be located in the PCBM domain.

Proposed Operating Mechanism
There are four fundamental operating models in ternary solar cells: the charge transfer, the energy transfer, the parallel-linkage, and the alloy model [17][18][19]23,26]. Here, we discussed a possible operating mechanism of BODIPY:P3HT:PCBM that enhance Jsc. Figure 7 presents the energy diagram of donor and acceptors. The HOMO and LUMO energy levels of BODIPY-1 and BODIPY-2 were estimated from our previous study [49]. The HOMO and LUMO offset (∆E) values between P3HT and BODIPY are about ∼0.1-0.2 eV, which are energetically feasible in term of driving the cascade charge transfer from P3HT to BODIPY and from BODIPY to PCBM [19,20]. Since P3HT can act as both a charge donor and an energy donor [28,75], we also investigate energy transfer process using a photoluminescence (PL) study. The PL spectra of blending P3HT and BODIPY revealed that P3HT emission can be quenched by a high BODIPY content (Figure 8). This implies that an energy transfer from P3HT to BODIPY was also presented. The energy transfer from P3HT to BODIPY molecules is likely favorable due to the substantial overlapping of the P3HT emission spectrum and the BODIPY absorption, as shown in the inset of Figure 8. Therefore, the boosted Jsc is attributed to cascade charge transfer and energy transfer and to the favorable morphology of the ternary blend.

Proposed Operating Mechanism
There are four fundamental operating models in ternary solar cells: the charge transfer, the energy transfer, the parallel-linkage, and the alloy model [17][18][19]23,26]. Here, we discussed a possible operating mechanism of BODIPY:P3HT:PCBM that enhance J sc . Figure 7 presents the energy diagram of donor and acceptors. The HOMO and LUMO energy levels of BODIPY-1 and BODIPY-2 were estimated from our previous study [49]. The HOMO and LUMO offset (∆E) values between P3HT and BODIPY are about~0.1-0.2 eV, which are energetically feasible in term of driving the cascade charge transfer from P3HT to BODIPY and from BODIPY to PCBM [19,20]. Since P3HT can act as both a charge donor and an energy donor [28,75], we also investigate energy transfer process using a photoluminescence (PL) study. The PL spectra of blending P3HT and BODIPY revealed that P3HT emission can be quenched by a high BODIPY content (Figure 8). This implies that an energy transfer from P3HT to BODIPY was also presented. The energy transfer from P3HT to BODIPY molecules is likely favorable due to the substantial overlapping of the P3HT emission spectrum and the BODIPY absorption, as shown in the inset of Figure 8. Therefore, the boosted J sc is attributed to cascade charge transfer and energy transfer and to the favorable morphology of the ternary blend.

Conclusions
In summary, we demonstrated a 55% enhancement in PCE in ternary solar cells (BODIPY-1:P3HT:PCBM), reaching 3.43%. The improvement in PCE is attributed to efficient charge transfer, and a smooth surface in the ternary blend that facilitates exciton dissociation and charge transportation. This finding suggests that BODIPY-1 and BODIPY-2 are suitable small molecules for improving the efficiency of the P3HT:PCBM system.

Conclusions
In summary, we demonstrated a 55% enhancement in PCE in ternary solar cells (BODIPY-1:P3HT:PCBM), reaching 3.43%. The improvement in PCE is attributed to efficient charge transfer, and a smooth surface in the ternary blend that facilitates exciton dissociation and charge transportation. This finding suggests that BODIPY-1 and BODIPY-2 are suitable small molecules for improving the efficiency of the P3HT:PCBM system.

Conclusions
In summary, we demonstrated a 55% enhancement in PCE in ternary solar cells (BODIPY-1:P3HT:PCBM), reaching 3.43%. The improvement in PCE is attributed to efficient charge transfer, and a smooth surface in the ternary blend that facilitates exciton dissociation and charge transportation. This finding suggests that BODIPY-1 and BODIPY-2 are suitable small molecules for improving the efficiency of the P3HT:PCBM system.
Funding: This research was funded by the Coordinating Center for Thai Government Science and Technology Scholarship Students, National Science and Technology Development Agency, Thailand, through the "Young Scientist Researcher Grant" This work was partially supported by King Mongkut's University of Technology Thonburi through the "KMUTT 55th Anniversary Commemorative Fund" and NANOTEC grant number P1952316.