Solution-Processed PEDOT:PSS/MoS2 Nanocomposites as Efficient Hole-Transporting Layers for Organic Solar Cells

An efficient hole-transporting layer (HTL) based on functionalized two-dimensional (2D) MoS2-poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) composites has been developed for use in organic solar cells (OSCs). Few-layer, oleylamine-functionalized MoS2 (FMoS2) nanosheets were prepared via a simple and cost-effective solution-phase exfoliation method; then, they were blended into PEDOT:PSS, a conducting conjugated polymer, and the resulting hybrid film (PEDOT:PSS/FMoS2) was tested as an HTL for poly(3-hexylthiophene):[6,6]-phenyl-C61-butyric acid methyl ester (P3HT:PCBM) OSCs. The devices using this hybrid film HTL showed power conversion efficiencies up to 3.74%, which is 15.08% higher than that of the reference ones having PEDOT:PSS as HTL. Atomic force microscopy and contact angle measurements confirmed the compatibility of the PEDOT:PSS/FMoS2 surface for active layer deposition on it. The electrical impedance spectroscopy analysis revealed that their use minimized the charge-transfer resistance of the OSCs, consequently improving their performance compared with the reference cells. Thus, the proposed fabrication of such HTLs incorporating 2D nanomaterials could be further expanded as a universal protocol for various high-performance optoelectronic devices.


Introduction
Organic solar cells (OSCs) have many striking properties such as flexibility, solution processability, light weight, and simple manufacturing, especially if compared with their inorganic counterparts. To enhance their performance, numerous strategies have been proposed, including novel photoactive materials, morphology control, interfacial engineering, plasmonic nanoparticles incorporation, and alternative buffer layers and electrodes [1][2][3][4][5][6]. Their power conversion efficiency (PCE) has been recently improved up to >13% with rapid advances in new photovoltaic materials [7]. In the typical bulk heterojunction (BHJ) OSCs configuration, a photoactive blend layer consisting of acceptor/donor pairs is sandwiched between a bottom transparent anode and a top low-work-function cathode, combined with the corresponding interlayers. Such interlayers are crucial for determining the overall PCE and stability of OSCs because they reduce the potential energy barrier between photoactive layer and electrodes, enhancing the extraction of holes and electrons at the anode and cathode, respectively.

Synthesis of FMoS 2 Nanosheets and PEDOT:PSS/FMoS 2 Hybrids
FMoS 2 nanosheets were synthesized according to the liquid-phase exfoliation method reported in literature [45], with small modifications. Briefly, bulk MoS 2 powder (200 mg) was bath-sonicated in oleylamine (2 mL) by using a Branson ultrasonic bath for 20 min and successively stirred at 60 • C for 12 h in an N 2 -filled glove box. Then, 1,2-dichlorobenzene (DCB) (18 mL) was added, and the dispersion was further bath-sonicated for 5 h. The resulting suspension was centrifuged at 4000 rpm, and the top 80% dark-green color supernatant, which contains excess oleylamine, DCB, and FMOS 2 was collected. Then, the FMoS 2 nanosheets were separated by adding excess acetone, followed by sonication for 2 min and high-speed centrifugation (10000 rpm). The separated FMoS 2 nanosheets were settled at the bottom of the centrifuge tube, which was re-dispersed in a small amount of IPA by mild sonication, and different concentrations (5,20, and 50 µL) of the resulting dispersion were added into PEDOT:PSS:methanol (1:1 V%) aqueous solutions, which were successively ultrasonicated for 30 min to obtain PEDOT:PSS/FMoS 2 hybrid solutions.

Fabrication of OSCs
The OSCs having device architectures of ITO/PEDOT:PSS/P3HT:PCBM/LiF/Al and ITO/(PEDOT:PSS/FMoS 2 )/P3HT:PCBM/LiF/Al were fabricated as follows. ITO-coated glass substrates were cleaned via sequential ultrasonication in acetone, IPA, and distilled water, followed by oxygen plasma treatment for 10 min; then, they were spin-coated with a PEDOT:PSS (Clevios P VP Al 4083) or PEDOT:PSS/FMoS 2 solution at 4000 rpm for 40 s and dried at 130 • C for 30 min to complete the HTL deposition. Next, an active layer consisting of a P3HT:PCBM (1:0.6 wt%) binary blend solution was spin-coated on the resulting HTL layer at 2500 rpm for 40 s inside an N 2 -filled glove box and annealed at 150 • C. Finally, LiF and Al layers were deposited by thermal evaporation. The active area of the fabricated OSCs was 0.06 cm 2 .

Characterization
The absorption properties of the samples were analyzed using a UV-Vis absorption spectrometer (Cary 5000, Varian, Inc.). Raman spectra were recorded on a Horiba Jobin-Yvon spectrometer. The emission properties were investigated with a luminescence spectrometer (LS55 Perkin Elmer). The TEM measurements were carried out on a JEOL JSM-2100-F system. The surface morphologies were investigated using a tapping-mode atomic force microscope (Veeco D3100). The water contact angles of the samples were measured with a KSV CAM 101 instrument. The GIWAXS analysis was conducted at the PLS-II 9A U-SAXs beamline of the Pohang Accelerator Laboratory (Korea) at the following operating conditions: incidence angle of~0.12 • , wavelength of 1.12 Å, and sample-to-detector distance of 224 nm. The GIWAXS patterns were recorded using a 2D charge-coupled device camera (Rayonix, SX-165, USA) with an exposure time of 10-30 s. The JV properties of the solar cells were measured with a Keithley 2400 solar cell IV measurement system under AM 1.5 G illumination at 100 mW cm −2 .
Raman spectroscopy is a powerful nondestructive technique for monitoring structural changes in 2D materials [46]. The Raman spectrum of bulk MoS 2 showed two characteristic peaks at 374.83 and 402.05 cm −1 corresponding, respectively, to the E 1 2g and A 1g vibrational modes ( Figure 2a); the first arose from the in-plane vibration of Mo and S atoms, while the second resulted from the out-of-plane vibrations of sulfur [47,48]. As regards FMoS 2 , the peaks for both the E 1 2g and A 1g vibrational modes were blue-shifted toward higher wavenumbers (respectively, 382.66 and 405.66 cm −1 ), suggesting interactions between oleylamine and MoS 2 . Moreover, the wavenumber difference between these two vibrational modes is closely related to the layer number present in the MoS 2 nanosheets [49], and in detector distance of 224 nm. The GIWAXS patterns were recorded using a 2D charge-coupled device camera (Rayonix, SX-165, USA) with an exposure time of 10-30 s. The JV properties of the solar cells were measured with a Keithley 2400 solar cell IV measurement system under AM 1.5 G illumination at 100 mW cm -2 .
Bulk MoS 2 is an indirect bandgap semiconductor that does not exhibit any photoluminescence; however, upon exfoliation, its luminescence increases with decreasing its layer thickness, so that single-layer MoS 2 shows the highest photoluminescence due to its transition into a direct bandgap semiconductor [52,53]. As expected, FMoS 2 exhibited significant photoluminescence (see the PL spectra in Figure S1, Electronic Supporting Information (ESI)), which clearly proves the successful layer thinning of MoS 2 during the functionalization process.  Raman spectroscopy is a powerful nondestructive technique for monitoring structural changes in 2D materials [46]. The Raman spectrum of bulk MoS2 showed two characteristic peaks at 374.83 and 402.05 cm -1 corresponding, respectively, to the E ¹ 2g and A1g vibrational modes ( Figure 2a); the first arose from the in-plane vibration of Mo and S atoms, while the second resulted from the out-of-plane vibrations of sulfur [47,48]. As regards FMoS2, the peaks for both the E ¹ 2g and A1g vibrational modes were blue-shifted toward higher wavenumbers (respectively, 382.66 and 405.66 cm −1 ), suggesting interactions between oleylamine and MoS2. Moreover, the wavenumber difference between these two vibrational modes is closely related to the layer number present in the MoS2 nanosheets [49], and in our case, this difference decreased from 27.2 cm -1 for bulk MoS2 to 23 cm -1 for FMoS2 nanosheets, demonstrating the successful exfoliation of MoS2 nanosheets during the oleylamine treatment.
The absorption properties of the FMoS2 nanosheets were further investigated via UV-Vis absorption spectroscopy; their spectrum ( Figure 2b) clearly showed two characteristic absorption peaks of MoS2 at 618 and 677 cm -1 corresponding, respectively, to the A1 and B1 direct excitonic transitions with the energy split from valence band spin-orbital coupling [50]. Furthermore, unlike bulk MoS2, FMoS2 yielded dark-greenish dispersion in 1,2dichlorobenzene. These results clearly indicate some alteration in the surface properties of MoS2 due to the oleylamine treatment [51].
Bulk MoS2 is an indirect bandgap semiconductor that does not exhibit any photoluminescence; however, upon exfoliation, its luminescence increases with decreasing its layer thickness, so that single-layer MoS2 shows the highest photoluminescence due to its transition into a direct bandgap semiconductor [52,53]. As expected, FMoS2 exhibited significant photoluminescence (see the PL spectra in Figure S1, Electronic Supporting Information (ESI)), which clearly proves the successful layer thinning of MoS2 during the functionalization process. The TEM images of the FMoS 2 nanosheets are displayed in Figure 2c,d, showing a thin nanosheet morphology with sizes of several hundred nanometers. A careful observation of the nanosheet edges reveals the presence of few-layer nanosheets, confirming the effectiveness of the liquid-based exfoliation with oleylamine. AFM measurements were carried out ( Figure S2, ESI) to further evaluate the layer thickness; that of FMoS 2 was~6.7 nm, suggesting the existence of few-layer nanosheets, while the reported thickness of monolayer MoS 2 ranges between 0.9 and 1.2 nm [54].
To improve the performance of conventional PEDOT:PSS-based HTL for OSCs, we incorporated it with FMoS 2 via a simple solution-blending method because we believed that the introduction of 2D sheet-like MoS 2 functionalized with a long-chain primary alkyl amine (oleylamine) would have made the PEDOT:PSS surface more hydrophobic, facilitating the following deposition of the hydrophobic active layer. In addition, the amine group of oleylamine tends to be located near Mo atoms in MoS 2 due to metal-amine interactions, while its long alkyl chain with -CH 3 groups is oriented toward the active layer, and this kind of configuration should enforce the active layer with a desirable molecular orientation for efficient charge transport in OSCs; P3HT thin films deposited on insulator substrates modified with -CH 3 groups formed face-on orientation because of π-H interactions [55,56].
The contact angles of ITO with PEDOT:PSS and PEDOT:PSS/FMoS 2 containing 5, 20, and 50 µL of FMoS 2 were 30 • , 47 • , 54 • , and 56 • , respectively ( Figure S3, ESI), which indicates that the hydrophobicity of PEDOT:PSS was slightly increased by the FMoS 2 addition and, hence, the hydrophobic active layer solution was more compatible on hybrid HTL than that of the hydrophilic PEDOT:PSS one.
The surface morphology of the various samples was compared via tapping-mode AFM analysis ( Figure S4, ESI); the root-mean-square (rms) roughness value of PEDOT:PSS was 1 nm and decreased down to 0.69 nm for PEDOT:PSS/FMoS 2 (5), suggesting a smooth surface morphology in the hybrid HTL. However, PEDOT:PSS/FMoS 2 (50) exhibited an rms roughness value of 0.97 nm, indicating that the addition of higher FMoS 2 concentrations would decrease the film smoothness.
All the synthesized PEDOT:PSS and PEDOT:PSS/FMoS 2 hybrid films exhibited similar UV-Vis transmittance values ( Figure S5a, ESI), showing that the FMoS 2 addition did not affect any absorption property of the PEDOT:PSS matrix. As regards the P3HT:PCBM (active layer) films spin-coated on glass substrates predeposited with PEDOT:PSS or PEDOT:PSS/FMoS 2 HTLs ( Figure S5b), for all the samples, their absorbance ranged from 400 to 650 nm, with a maximum at 512 nm, and two shoulders around 550 and 600 nm. The existence of vibronic feature at 600 nm suggests that the P3HT film existed in a high degree of ordered crystalline lamella due to strong interchain interactions [57].
The current-voltage (JV) characteristics of the fabricated P3HT:PCBM OSCs having PEDOT:PSS/FMoS 2 as HTL are shown in Figure 3a. Their performance is compared with that of reference devices having PEDOT:PSS as HTL in Table 1. The reference cells showed PCE = 3.25%, J sc = 7.92 mA cm −2 , V oc = 0.671 V, and FF = 0.61. The FMoS 2 incorporation led to significant PCE and J sc improvements; in particular, the device based on PEDOT:PSS/FMoS 2 (5) exhibited the highest PCE, J sc , and FF.
The external quantum efficiency (EQE) measurements (Figure 3b) showed improved EQE for the hybrid HTL-based OSCs compared with the reference cells and confirmed also their increased J sc , demonstrating the enhanced charge extraction at the HTL/active layer interface and the charge collection at the electrodes [58,59]. The photovoltaic parameters such as PCE, J sc , FF and V oc as a function of FMoS 2 in PEDOT:PSS HTLs are plotted in Figure 3c, d, e and f respectively. Nanomaterials 2019, 9, x FOR PEER REVIEW 6 of 12  To understand the charge transport, we analyzed the microstructure (chain-orientation and crystallinity) of the active layer (P3HT:PCBM) on both the PEDOT:PSS and PEDOT:PSS/FMoS2 samples by GIWAXS (Figure 4 and Figure 5). Charge transport in conjugated polymers occurs either in the π-π staking direction or the chain backbone one, which is the fastest but its vertical alignment of chains backbones along the z direction is rarely observed [60,61]. In general, P3HT crystallizes into two main configurations, namely, edge-on and face-on orientations; in the former, both chain backbone and π-π staking directions lie parallel to the substrate; in the latter, π-π staking occurs perpendicular to the substrate, which is a desirable orientation in OSCs for vertical charge transport [62]. Figure 4 shows the GIWAXS diffraction patterns of P3HT:PCBM thin films deposited on PEDOT:PSS and PEDOT:PSS/FMoS2 HTLs. In both cases, the thin films exhibited strong (100), (200), and (300) diffractions along the z axis, confirming the existence of the strong edge-on lamellae      To understand the charge transport, we analyzed the microstructure (chain-orientation and crystallinity) of the active layer (P3HT:PCBM) on both the PEDOT:PSS and PEDOT:PSS/FMoS 2 samples by GIWAXS (Figures 4 and 5). Charge transport in conjugated polymers occurs either in the π-π staking direction or the chain backbone one, which is the fastest but its vertical alignment of chains backbones along the z direction is rarely observed [60,61]. In general, P3HT crystallizes into two main configurations, namely, edge-on and face-on orientations; in the former, both chain backbone and π-π staking directions lie parallel to the substrate; in the latter, π-π staking occurs perpendicular to the substrate, which is a desirable orientation in OSCs for vertical charge transport [62]. Figure 4 shows the GIWAXS diffraction patterns of P3HT:PCBM thin films deposited on PEDOT:PSS and PEDOT:PSS/FMoS 2 HTLs. In both cases, the thin films exhibited strong (100), (200), and (300) diffractions along the z axis, confirming the existence of the strong edge-on lamellae configuration of P3HT [63]. In addition, the absence of π-π staking peak (010), corresponding to the face-on orientation near the z axis, indicates that P3HT preferentially adopted the edge-on configuration in both PEDOT:PSS and PEDOT:PSS/FMoS 2 HTLs. Since the use of -CH 3 group-functionalized substrates tends to promote the face-on orientation of P3HT [55,56], we aimed to improve such configuration of the active layer by incorporating the described oleylamine (having -CH 3 groups)-functionalized MoS 2 into PEDOT:PSS, but we did not observe any significant difference in its molecular orientation, maybe because the low FMoS 2 concentrations used were not sufficient for such change. Thus, we can conclude that the PCE and J sc enhancement in the OCSs having PEDOT:PSS/FMoS 2 as HTL may be due to its surface compatibility for the active layer deposition, as observed in the AFM and contact angle measurements. Nanomaterials 2019, 9, x FOR PEER REVIEW 7 of 12 configuration of P3HT [63]. In addition, the absence of π-π staking peak (010), corresponding to the face-on orientation near the z axis, indicates that P3HT preferentially adopted the edge-on configuration in both PEDOT:PSS and PEDOT:PSS/FMoS2 HTLs. Since the use of -CH3 groupfunctionalized substrates tends to promote the face-on orientation of P3HT [55,56], we aimed to improve such configuration of the active layer by incorporating the described oleylamine (having -CH3 groups)-functionalized MoS2 into PEDOT:PSS, but we did not observe any significant difference in its molecular orientation, maybe because the low FMoS2 concentrations used were not sufficient for such change. Thus, we can conclude that the PCE and Jsc enhancement in the OCSs having PEDOT:PSS/FMoS2 as HTL may be due to its surface compatibility for the active layer deposition, as observed in the AFM and contact angle measurements.  Electrical impedance spectroscopy (EIS) was performed to investigate the charge transport dynamics of the OSCs fabricated with PEDOT:PSS and PEDOT:PSS/FMoS2(5) as HTL ( Figure 6). This analysis allowed us to observe the current response by applying alternating current voltage as a function of frequency; the OSCs with PEDOT:PSS/FMoS2(5) demonstrated slightly lower charge transfer resistance, revealing that the holes were configuration of P3HT [63]. In addition, the absence of π-π staking peak (010), corresponding to the face-on orientation near the z axis, indicates that P3HT preferentially adopted the edge-on configuration in both PEDOT:PSS and PEDOT:PSS/FMoS2 HTLs. Since the use of -CH3 groupfunctionalized substrates tends to promote the face-on orientation of P3HT [55,56], we aimed to improve such configuration of the active layer by incorporating the described oleylamine (having -CH3 groups)-functionalized MoS2 into PEDOT:PSS, but we did not observe any significant difference in its molecular orientation, maybe because the low FMoS2 concentrations used were not sufficient for such change. Thus, we can conclude that the PCE and Jsc enhancement in the OCSs having PEDOT:PSS/FMoS2 as HTL may be due to its surface compatibility for the active layer deposition, as observed in the AFM and contact angle measurements.  Electrical impedance spectroscopy (EIS) was performed to investigate the charge transport dynamics of the OSCs fabricated with PEDOT:PSS and PEDOT:PSS/FMoS2(5) as HTL ( Figure 6). This analysis allowed us to observe the current response by applying alternating current voltage as a function of frequency; the OSCs with PEDOT:PSS/FMoS2(5) demonstrated slightly lower charge transfer resistance, revealing that the holes were Electrical impedance spectroscopy (EIS) was performed to investigate the charge transport dynamics of the OSCs fabricated with PEDOT:PSS and PEDOT:PSS/FMoS 2 (5) as HTL ( Figure 6). This analysis allowed us to observe the current response by applying alternating current voltage as a function of frequency; the OSCs with PEDOT:PSS/FMoS 2 (5) demonstrated slightly lower charge transfer resistance, revealing that the holes were effectively transported from the active layer to the anode (ITO). In order to elucidate the origin of the improvement in the photovoltaic performance, especially both FF and J sc for PEDOT:PSS/FMoS 2 (5), we further calculated the resistance of the devices. In general, it is well known that lower series resistance (R S ) and higher shunt resistance (R SH ) are required to achieve higher FF in the solar cell device [64]. Based on the J-V curves obtained from the Nanomaterials 2019, 9, 1328 8 of 12 devices, it is clearly revealed that the device with PEDOT:PSS/FMoS 2 (5) as HTL showed the lowest R S while maintaining higher R SH , leading to enhancement in charge extraction. The corresponding R S value of cells employing PEDOT:PSS/FMoS 2 (5) as HTL was 134.8 Ω·cm 2 , while the reference showed 180.0 Ω·cm 2 . Lower R S indicates that better interfacial contact and charge collection efficiency were obtained due to the addition of the conducting FMoS 2 layer. In the case of the R SH , no significant changes in the shunt resistance were observed for the devices. In the point of view of the identical R SH , barrier resistance at the interface and the leakage current level flowing across the photoactive layer is similar. Therefore, the addition of FMoS 2 might contribute to extract photoexcited charges efficiently by lowering the R S . Nanomaterials 2019, 9, x FOR PEER REVIEW 8 of 12 effectively transported from the active layer to the anode (ITO). In order to elucidate the origin of the improvement in the photovoltaic performance, especially both FF and Jsc for PEDOT:PSS/FMoS2(5), we further calculated the resistance of the devices. In general, it is well known that lower series resistance (RS) and higher shunt resistance (RSH) are required to achieve higher FF in the solar cell device [64]. Based on the J-V curves obtained from the devices, it is clearly revealed that the device with PEDOT:PSS/FMoS2(5) as HTL showed the lowest RS while maintaining higher RSH, leading to enhancement in charge extraction. The corresponding RS value of cells employing PEDOT:PSS/FMoS2(5) as HTL was 134.8 Ω•cm 2 , while the reference showed 180.0 Ω•cm 2 . Lower RS indicates that better interfacial contact and charge collection efficiency were obtained due to the addition of the conducting FMoS2 layer.
In the case of the RSH, no significant changes in the shunt resistance were observed for the devices. In the point of view of the identical RSH, barrier resistance at the interface and the leakage current level flowing across the photoactive layer is similar. Therefore, the addition of FMoS2 might contribute to extract photoexcited charges efficiently by lowering the RS.

Conclusions
The application of solution-processed PEDOT:PSS/FMoS2 hybrids as effective HTLs for OSCs has been successfully demonstrated. Raman, UV-Vis, PL, TEM, and AFM analyses confirmed the successful exfoliation of bulk MoS2 into few-layer nanosheets in the presence of oleylamine via a simple and cost-effective solution-based method. The OSCs fabricated with the synthesized PEDOT:PSS/FMoS2 hybrids as HTL exhibited PCE values up to 3.74%, which is 15.08% higher than that of the reference cells having simple PEDOT:PSS as HTL. The hybrid HTL films showed better surface properties for the deposition of the hydrophobic active layer, consequently, the charge-transfer resistance was minimized for OSCs fabricated with hybrid HTL compared with reference cells, improving the OSC performance. Due to their simple preparation method, 2D FMoS2-incorporated PEDOT:PSS-based HTL provides valuable alternative HTL for OSCs.

Conclusions
The application of solution-processed PEDOT:PSS/FMoS 2 hybrids as effective HTLs for OSCs has been successfully demonstrated. Raman, UV-Vis, PL, TEM, and AFM analyses confirmed the successful exfoliation of bulk MoS 2 into few-layer nanosheets in the presence of oleylamine via a simple and cost-effective solution-based method. The OSCs fabricated with the synthesized PEDOT:PSS/FMoS 2 hybrids as HTL exhibited PCE values up to 3.74%, which is 15.08% higher than that of the reference cells having simple PEDOT:PSS as HTL. The hybrid HTL films showed better surface properties for the deposition of the hydrophobic active layer, consequently, the charge-transfer resistance was minimized for OSCs fabricated with hybrid HTL compared with reference cells, improving the OSC performance. Due to their simple preparation method, 2D FMoS 2 -incorporated PEDOT:PSS-based HTL provides valuable alternative HTL for OSCs.