4-1 All-Solution-Processed Inverted Organic Solar Cell with a Stacked Hole-Transporting Layer
4-1.1 Introduction
OSCs have attracted increasing attention as next-generation devices for energy conversion from solar light to electricity [36, 75-77]. They can be manufactured costeffectively and are compatible with flexible substrates. The conventional forward bulk heterojunction architecture has a notable drawback of a low-work-function cathode, which allows diffusion of oxygen into the organic active layer and hence degrades the device performance [78, 79]. This can be avoided by employing an inverted device structure and replacing the low-work-function metal Al with less airsensitive high-work-function metals [80-84].
ZnO is a promising hole-blocking layer (HBL) candidate owing to its relatively high electron mobility and high transparency [85]. ZnO can be prepared by various methods such as atomic layer deposition (ALD) and the sol–gel process [86, 87]. The sol–gel process is attractive because it can fabricate a large-area transparent conductor oxide (TCO) using simple and inexpensive equipment in comparison to other methods.
The most traditionally and frequently used buffer layer in polymer OSCs is PEDOT:PSS. It has many advantages such as high transparency, a high-work-function, smooth morphology, and good conductivity [88-90]. Devices based on copper phthalocyanine (CuPc) have been studied extensively because of the low-cost, chemical and physical stability, and outstanding photophysical and electrochemical properties of CuPc [91]. The CuPc higher HOMO/LUMO levels are approximately 5.0/3.6 eV, and excitons can be dissociated at the corresponding donor/acceptor interface. Thereafter, free electrons can reach the cathode and holes can be transported
43
and collected by the anode. Because of complementary absorption spectra, more electron–hole pairs can be generated, and more holes can be extracted to the transparent conductive electrode, and greater Jsc can be obtained. Hence, all solution processes with low investment costs for fabricating OSCs will be feasible for commercial applications. Spray-coated PEDOT:PSS has been used as a top electrode in solution-processed inverted OSCs [92]. A solution-processed silver nanowire has also been demonstrated to be a transparent electrode with a mesh shape or as a thin film [93, 94].
In this work, we have prepared an inverted OSC with the structure ITO /ZnO/
P3HT:PCBM /PEDOT:PSS/CuPc/Ag. ZnO, P3HT:PCBM, PEDOT:PSS, CuPc, and Ag are all formed as films by a solution process. The device performance is characterized and discussed.
4-1.2 Experimental Procedure
The device structure and energy band diagrams of an inverted OSC with a stacked HTL are illustrated in Fig. 4-1. The ZnO film is formed by the following sol–gel technique. The precursor solution consists of 0.1M zinc acetate dihydrate and 0.1M monoethanolamine in 2-methoxyethanol [95]. It was spincoated onto glass/ITO substrates at 2000 rpm for 40 s. To obtain the ZnO planar film, the substrate was immediately placed onto a hot plate that was preheated at 250 ℃[91, 96] and annealed for 10 min. P3HT and PCBM blend films were spincoated onto the ZnO film with a 1 : 0:8 weight ratio solution in 1,2-dichlorobenzene at 700 rpm using the slow-growth method, followed by annealing on a hot plate at 120 ℃ for 10 min. The HTLs, including PEDOT:PSS, CuPc, CuPc/PEDOT:PSS, and PEDOT:PSS/CuPc, were all spin-coated onto the active layer. CuPc powder was dissolved in C3H3O (IPA) to a concentration of 5wt%. The 5wt% CuPc solution was spin-coated onto a P3HT:PCBM film and then baked at 120 ℃ for 10 min in N2 atmosphere. The anode
44
is fabricated from silver nanoparticles. A silver solution consisting of silver nanoparticles dissolved in IPA and alpha-terpineol, is added dropwise onto the HTL by the titration technique using a Pasteur burette and then baked at 120 ℃ for 5 min in N2 atmosphere. To confirm the formation of stacked layers of P3HT:PCBM/CuPc, a sample with a stacked structure of glass/ITO/P3HT:PCBM/CuPc was prepared.
Figure 4-2 shows a SEM cross-sectional image and demonstrates the formation of P3HT:PCBM/CuPc stacked layers. The film thicknesses of P3HT:PCBM and CuPc are 124 and 58 nm, respectively.
J–V curves were determined using a Keithley 2400 source meter. The sample
was illuminated using an AM1.5G simulated solar spectrum from a filtered Xe arc lamp source. The light intensity of the solar simulator was calibrated using a Si photodetector. A crosssectional image of the stacked layers was obtained by a SEM.
AFM was used in an attempt to correlate the characteristics of the surface morphology.
4-1.3 Results and Discussion
As the light enters from the ITO side, transmittance from ZnO is important to realize high current density. The transmittance depends on the crystal size of ZnO, which in turn depends on the concentration of precursors in the sol–gel. Using optimized parameters to fabricate ZnO films on ITO substrates, we measured the optical transmission spectrum of ITO/ZnO. Figure 4-3 shows that ITO/ZnO absorbs light only in the wavelength region of < 400 nm, and thus is very transparent throughout the visible light region. The transmittance of ITO/ZnO is approximately 90%, which coincides with the absorbance region of the active layer. The change in current density is consistent with the variation in transmittance from ZnO films.
Therefore, ZnO has a very good electron-transporting capability. It also serves as an exciton-blocking layer because the valence band of ZnO (7.5 eV) is much higher than
45
the HOMO of the acceptor/donor (6.1 eV)/(5.2 eV). The fabricated inverted OSC device structures are summarized as follows. All the structures used ZnO as the HBL but different materials as the HTL.
Figure 4-4 shows the J–V curves of inverted OSCs with various HTL materials under 100mW/cm2 white light illumination in air. The device characteristics of the inverted OSCs are summarized in Table 4-1. C-1 with an HTL of PEDOT:PSS shows FF, PCE, Jsc, and Voc values of 36.53%, 1.13%, 5.82mA/cm2, and 0.53V, respectively.
An inverted OSC fabricated from ITO/ZnO/P3HT:PCBM/Ag without an HTL exhibits FF, PCE, Jsc, and Voc values of 32.38%, 0.61%, 3.75 mA/cm2, and 0.50 V, respectively. It is clear that the PEDOT:PSS can act as an HTL to efficiently enhance holes to be transported from the active layer to the anode. However, the interface between P3HT:PCBM and PEDOT:PSS is not stable owing to the chemical reaction.
Instead of PEDOT:PSS, CuPc is introduced as a buffer layer in C-2. CuPc has reasonably high conductivity and can be selected as a hole-conducting layer for solar cells [97]. Its visible absorption range is wide and it has good chemical, heat, and light stability [98]. Therefore, we use it as an HTL layer. Comparison of C-1 and C-2 reveals that in the latter Jsc decreased from 5.82 to 4.28 mA/cm2 and PCE decreased from 1.13 to 0.75%. Notably, Rs is attributed to the ohmic loss in the whole device, which relates to the overall quality of the films. The increase in Rs from 44.65 to 61.38 Ω cm2 is due to the formation of a degraded ohmic contact. This is attributed to the solubility problem of CuPc, which is almost insoluble in common solvents.
C-3 with a buffer layer of CuPc/PEDOT:PSS shows an FF of 36.39% and a PCE of 0.95%. Jsc and Voc are 4.96 mA/cm2 and 0.53 V, respectively. Comparison of C-3 and C-4 reveals that in the latter PCE increased from 0.95 to 1.24%. The PCE is strongly governed by the number of carriers, which were generated in the active layer, arriving at the electrodes. Examination of the HOMO of PEDOT:PSS (5.1 eV) and
46
CuPc (5.0 eV) shows that a stacked PEDOT:PSS/CuPc HTL forms a stepwise energy level configuration, as shown in Fig. 4-1(b), which enhances the hole transportation to the Ag anode. Therefore, the Jsc of C-4 with a PEDOT:PSS/CuPc HTL is superior to that of C-3. However, according to Fig. 4-1(c), the energy level of C-3 is discontinuous at the CuPc/PEDOT:PSS interface, subsequently degrading the electrical performance markedly. Because the Ag nanoparticles have superior film properties, the wet-coated Ag layer was employed as the top anode in the bottom-illuminated OSCs.
Figure 4-5(a) shows the contact angle of the P3HT:PCBM film. It appears to be 90° and indicates that the active layer of P3HT:PCBM is a hydrophobic film. Figure 4-5(b) shows the contact angle of the PEDOT:PSS film. It also appears to be nearly 90°. Therefore, it is difficult to form PEDOT:PSS on P3HT:PCBM. We dissolve PEDOT:PSS in IPA to decrease the contact angle, as shown in Fig. 4-5(c). IPA dilutes PEDOT:PSS and lets the active layer be removed from the chemical reaction with PEDOT:PSS, leading to the stability of the P3HT:PCBM/PEDOT:PSS interface.
Hence, PEDOT:PSS can successfully form a film on the active layer.
Figures 4-6(a) and 4-6(b) show the surface morphologies of the CuPc/PEDOT:PSS and PEDOT:PSS/CuPc films, respectively, observed by AFM.
The RMS roughnesses are 3.63 nm for the film in Fig. 4-6(a) and 3.20 nm for the film in Fig. 4-6(b). The RMS roughness of PEDOT:PSS/CuPc is lower than that of CuPc/PEDOT:PSS, which indicates that a smooth PEDOT:PSS/CuPc film increases charge carrier transfer and improves the PCE of inverted OSCs [58, 59].
4-1.4 Summary
An inverted OSC is successfully fabricated by using all-solution processes. We prepared a ZnO film by a low-cost sol–gel spin-coating technique. An Ag anode was fabricated from Ag nanoparticles by a drop titration technique using a Pasteur burette.
47
Experimental results show that the PEDOT:PSS/CuPc stacked HTL provides a stepwise hole-transporting energy diagram configuration, which subsequently increases the charge carrier transporting capability and extracts holes from the active layer to the anode. Owing to effective hole blocking by ZnO and hole transporting by the PEDOT:PSS/CuPc layer, the inverted OSC can lead to a significant improvement in Jsc. An inverted OSC with the optimized structure of ITO/ZnO/P3HT:PCBM/PEDOT:PSS/CuPc/Ag exhibits Voc of 0.53V, Jsc of 6.13 mA/cm2, FF of 37.53%, and PCE of 1.24% under simulated AM 1.5G illumination of 100mW/cm2. Hence, a solution process is feasible for fabricating low-cost and large-area solar energy devices.
48
Table 4-1 Device characteristics of inverted OSCs with various HTLs studied in this work.
49
Fig. 4-1 (a) Device structure used in this study. Energy band diagrams of inverted OSC with a stacked HTL of (b) CuPc/PEDOT:PSS and (c) PEDOT:PSS/CuPc.
50
Fig. 4-2 SEM cross-sectional image of glass/ITO/P3HT:PCBM/CuPc stacked layers.
51
Fig. 4-3 Transmission spectra of ITO and ITO/ZnO films.
52
Fig. 4-4 J-V characteristics of the inverted OSCs with various HTLs under 100 mW/cm2 white light illumination in air.
53
(a) (b)
(c)
Fig. 4-5 Images of a drop of diiodomethane on surface of (a) P3HT:PCBM, (b) PEDOT:PSS, and (c) PEDOT:PSS:IPA films.
54
(a)
(b)
Fig. 4-6 AFM 2D and 3D images of (a) CuPc/PEDOT:PSS and (b) PEDOT:PSS/CuPc.
55
4-2 Efficiency Enhancement of Solution-Processed Inverted Organic Solar Cells with a Carbon Nanotubes-doped Active Layer
4-2.1 Introduction
OSCs based on blends of conjugated polymers and fullerenes offer a promising alternative for the development of low-cost, and flexible photovoltaic technology [36, 79, 99, 100]. In a conventional structure of OSCs, PEDOT:PSS is usually applied as a HTL, which is coated onto the anode of ITO. In an inverted structure, the interface of ITO/PEDOT:PSS can be avoided, and low-work-function aluminum (Al) can be replaced with less air-sensitive high-work-function metals such as Au and Ag [82-85, 80, 101, 102,]. However, the interface between ITO and PEDOT:PSS is not stable and the chemical reaction between ITO and PEDOT:PSS degrades device performance.
Control of charge carrier transport at heterojunctions in multilayer structures is one of the most important issues that require resolution for the improvement of OSCs [103, 104]. Hence, metal oxides, such as nickel oxide (NiO), V2O5, WO3, and MoO3, have been used in OSCs as a buffer layer [50, 105, 107]. However, MOx films are mostly prepared by thermal vacuum evaporation, which is hardly used for the fabrication of large-area PSCs and leads to high energy consumption. Compared with thermal-vacuum-deposited [108, 109] layers, low-temperature-coating solution-processed V2O5 interlayers are more facile and favorable for the low-cost mass production of devices such as ‘‘roll-to-roll’’ printed thin-film solar cells.
At present, a number of ZnO films doped with various metallic ions have been widely studied for the manipulation of their optical and electrical properties, such as Al, V, W, In2O3, and titanium [110-115]. However, there is limited discussion of the structural optical properties of titanium-doped ZnO (TZO). Among the various types of doped ZnO thin films, TZO films have been investigated recently for their unique electrical, magnetic, and sensing properties [116-118]. It has been demonstrated that
56
larger preferentially c-axis-oriented ZnO films doped with titanium have better optical properties than pure ZnO films [119,120]. To fulfill the deficiencies of PCBM as a source of electron acceptors and transporters, carbon nanotubes (CNTs) and nanorods may be employed because they are better electronic conductors; moreover, because of their high aspect ratio, the quantity required for fabricating efficient bulk heterojunctions is very low compared with that for PCBM, thereby reducing the overall cost of OSCs devices. It is expected that the introduction of CNTs in the active layer of OSCs will lead to optimized carrier transport through the ballistic pathway provided by CNTs.
In this work, a CNT near-infrared-sensitive material is doped into the active layer of P3HT:PCBM and solution-processed TZO is used as an HBL to fabricate inverted OSCs. Its device structure is ITO/TZO/P3HT:PCBM:CNTs/V2O5/Ag and its related performance is determined and discussed.
4-2.2 Experimental Procedure
The inverted OSCs in this work comprised a CNTs, P3HT and PCBM blend thin film sandwiched between ITO and a metal cathode. Figure 4-7(a) schematically shows the device structure, whereas Figs. 4-7(b)–4-7(d) show the energy level diagrams of various materials used to fabricate the device. The TZO film forms by following a sol-gel technique as illustrated. A precursor solution consists of 0.1 M zinc acetate dihydrate and 0.1 M monoethanolamine in 2-methoxyethanol [121].
Acetic acid and titanium is added as a dopant into the ZnO solution, while keeping the molar ratio of Ti dopant at 0.003 g/ml. As the light enters from ITO side, the transmittance of TZO is important in realizing high current density. The transmittance depends on the crystal size of TZO, and similarly relies on the concentration of precursors in the sol. It was spin-coated onto glass/ITO substrates at 1000 rpm for 40 s. For the TZO planar film, the substrate was immediately placed onto a hot plate that
57
was preheated at 200℃ and annealed for 90 min. CNT, P3HT and PCBM blend films were spin coated onto TZO film at a 0.02:1:0.8 wt-ratio solution in 1,2-dichlorobenzene at 700 rpm using the slow-growth method, followed by annealing on a hot plate at 150 ℃ for 10 min. The V2O5 layer was obtained by spin-coating an aqueous solution onto the active layer at a coating speed of 1300 rpm for 30 s, followed by baking at 120℃ for 5 min. Moreover, the anode, 100 nm of Ag, was thermally deposited on top of V2O5 in a deposition chamber under a vacuum pressure of 10-6 Torr.
J–V curves were determined using a Keithley 2400 source meter. The sample
was illuminated using an AM1.5G simulated solar spectrum from a filtered Xe arc lamp source. The light intensity of the solar simulator was calibrated with a Si photodetector.
4-2.3 Results and Discussion
Fig. 4-8 the optical transmittance compares ITO with ITO/TZO over the wavelength range from 200 nm to 900 nm. The transmittance depends on the crystal size of TZO, which simultaneously depends on the concentration of precursors in the sol-gel. All the films are found to be highly transparent in the visible wavelength region with an average transmittance over 90%. The observed high transparency in the visible spectral region makes these films suitable for HBL applications. From the inset of Fig. 4-8, we can also see that the UV absorption edge shifts to short wavelength when titanium doped into ZnO films.
Figure 4-8 also illustrates the effect of CNT dopant on the absorption spectra of active layers. CNTs and the conjugated polymer have absorption in the infrared and visible regions, respectively. As a result, the composite has slightly broader absorption.
The incorporation of CNTs in P3HT:PCBM composite modifies its absorption spectrum: increase in the absorbance intensity. As a tentative explanation, structuring
58
of the polymeric chains around the carbon nanotubes could explain this increase in absorbance intensity, as reported by Ikeda et al. [122]. Our results are in agreement with previous studies.
Figure 4-9 shows the J-V curves of inverted OSCs under 100 mW/cm2 white light illumination in air. Device characteristics of inverted OSCs are summarized in Table 4-2. The D-1 without an HTL shows a Jsc of 3.90 mA/cm2, a Voc of 0.40 V, and a FF of 35.61%, resulting in a PCE of only 0.56%. The poor device performance is attributed to direct contact of active layer and Ag. It is possible for PCBM to transfer electrons to the Ag electrode. Comparing to D-1, the D-2 with V2O5 reached a higher performance with a Jsc of 4.88 mA/cm2, a Voc of 0.56 V, and a FF of 38.56% and PCE of 0.86%. A comparison of D-1 with D-2 reveals that the latter has an increased Jsc
from 3.90 to 4.88 mA/cm2. Despite the increased Jsc, the Rs, defined as the slope of the J-V curve at J = 0 mA/cm2, was estimated to be approximately 27.60 Ω cm2 in D-2, which is less than that of the solar cell of D-1 (34.40 Ω cm2). Such stepwise energy level configuration indeed increases Jsc in an inverted OSC device. The improvement of Jsc is explained that the V2O5 layer can availably transport holes and block electrons. The V2O5 thin film can module the Schottky barrier and form an ohmic contact at the organic/Ag interface, which has potential to act as a superior HTL.
In efforts to reduce the cell Rs of HTL in our devices, CNTs were embedded into active layer. We expect to reduce cell Rs and enhance optical absorptions in active layer. D-3 with ITO/ZnO/P3HT:PCBM/PEDOT:PSS:Ag NPs/Ag showed FF, PCE, Jsc and Voc of 46.73%, 1.25 %, 5.70 mA/cm2 and 0.47 V, respectively. The PCE was dramatically improved to 1.25% by doping CNTs into the active layer. The results show that the PCE increase mainly originates from the increase of the short circuit current. The increase in the photocurrent caused by the incorporated CNTs was likely
59
due to the increase of absorbed photons and charge carrier mobility.
D-4 significantly increases its Jsc from 5.70 to 6.50 mA/cm2. Consequently, the PCE improves significantly, increasing from 1.25 to 2.20%. This increase of Jsc was due to the decrease in Rs from 22.69 Ω cm2 to 9.32Ω cm2 when we compared to the D-3. TZO has a high transmittance and higher HOMO. Therefore, it can be the best candidate because of its extraordinary hole blocking performance.
The EQEs of three inverted OSCs were depicted in Fig. 4-10. The spectra clearly represents an increase in EQE of the device with a CNTs-doped active layer. The photoelectrons generated in P3HT can rapidly transfer to PCBM and CNT, while the holes generated by PCBM can transfer to P3HT and CNT, resulting in a faster charge carriers transport than that in pristine devices. Furthermore, a maximum EQE of ITO/TZO/P3HT:PCBM:CNTs/V2O5/Ag around 300 nm to 650 nm results in the optimal PCE. The increase of surface roughness (RMS about 2.77 nm) by using titanium doped ZnO shows EQE about 55% because of light diffusion.
The EQEs of three inverted OSCs were depicted in Fig. 4-10. The spectra clearly represents an increase in EQE of the device with a CNTs-doped active layer. The photoelectrons generated in P3HT can rapidly transfer to PCBM and CNT, while the holes generated by PCBM can transfer to P3HT and CNT, resulting in a faster charge carriers transport than that in pristine devices. Furthermore, a maximum EQE of ITO/TZO/P3HT:PCBM:CNTs/V2O5/Ag around 300 nm to 650 nm results in the optimal PCE. The increase of surface roughness (RMS about 2.77 nm) by using titanium doped ZnO shows EQE about 55% because of light diffusion.