Fabrication of carbon nanotube hybrid films as transparent electrodes for small-molecule photovoltaic cells

Abstract

We demonstrate stable carbon nanotube (CNT) hybrid films as transparent electrodes for small-molecule photovoltaic cells. A photonic curing process is utilized to enhance the doping, and construct CNT hybrid films with solution-processed CuI and evaporated MoO₃, respectively. These novel CNT–CuI and CNT–MoO₃ hybrid films exhibit stable sheet resistances of 70 and 110 Ω per square at around 80% transmittance, respectively. OPV cells are fabricated by evaporating a tetraphenyldibenzoperiflanthene and fullerene bilayer heterojunction on a series of CNT hybrid films. The corresponding optimization of OPV cells are carried out on CNT–CuI and CNT–MoO₃ hybrid films. The correlations between the cell performance and the surface morphology, sheet resistance and transmittance of these CNT hybrid films are discussed in detail. The optimum cell on the CNT–CuI film with a nanostructured cascade-type architecture exhibits a power conversion efficiency of 2.97%.

---

Introduction

Organic photovoltaic (OPV) cells are regarded as a promising source of renewable energy, due to their advantages of being lightweight, low-cost and flexible.^1,2 In the last ten years, significant progresses in OPV cells have been achieved by the development of molecules and device fabrication processes. High power conversion efficiencies (PCE) of around 10% have been achieved with a bulk heterojunction (BHJ) using solution-processed polymers, or co-evaporated small molecules.^3–8 Many efforts have been made to optimize the structures of the BHJ, while isolated regions are usually formed, resulting in the formation of charge trapping sites.^9–12 Constructing nanostructured organic thin films to achieve an ideal phase-separation morphology with an interpenetrating network has also been widely investigated.^9–15 However, these works are mostly carried out on indium-tin-oxide (ITO) coated glass substrate. Future OPV cells should be flexible on plastic substrate in order to differentiate them from conventional crystalline Si PV cells.¹⁶ Therefore, establishing the fabrication process on non-traditional low-cost and flexible electrodes is of great importance for boosting OPV cells for daily use.^17–25

Thin films of carbon nanotube (CNT) exhibiting high electrical conductivity and high optical transparency, can be a transparent electrode instead of ITO.²⁶ Compared with other potential candidates of transparent electrodes, CNT films not only enable a lower cost process, but also provide a more stretchable and flexible platform with stronger mechanical strength.²⁷ However, doping with acids is usually required to improve the contact between CNTs, and thus achieve a high conductivity in CNT films. Most of the doped CNT films are not stable to air, temperature, or humidity, showing a rapid decrease in conductivity after doping.²⁸ The stability problem could be more severe during actual device process, especially in vacuum process. Thus, very few papers report the vacuum-evaporated OPV cells using CNT as transparent electrodes.

This work aims to develop a methodology for fabricating efficient OPV cells via vacuum evaporating small molecules on CNT transparent electrodes. Here, CuI and MoO₃, both of which are widely used as hole transport layers for OPV applications, are utilized to dope CNT transparent electrodes, respectively. Both of CuI and MoO₃ are typical wide-bandgap semiconductors with deep work function of above 5.2 eV, being close to the highest occupied molecular orbital (HOMO) of donor molecules. MoO₃ is also used as an electron acceptor for doping organic semiconductor and CNTs,^29,30 while CuI can act a structural template to control the molecular growth to improve the OPV cells.³¹ In addition to the promising electrical properties, CuI and MoO₃ have moderate melting point (bellow 800 °C) for structural modification. Firstly, we develop a full solution process to fabricate CNT–CuI hybrid films by utilizing a photonic curing process. The CNT–CuI films exhibit similar sheet resistances to the our previous report using vacuum-evaporated CuI film.³² On the other hand, a novel CNT–MoO₃ hybrid films are also fabricated with stable conductivity in vacuum. Secondly, OPV cells are fabricated with using tetraphenyldibenzoperiflanthene (DBP) as a donor and fullerene (C60) as an acceptor.^33,34 Corresponding optimizations of OPV cells are carried out on CNT–CuI and CNT–MoO₃ hybrid films. Crystalline diindenoperylene (DIP) is introduced to construct a cascade-type cell architecture on CNT–CuI hybrid film smoothed with poly(3,4-ethylenedioxythiophene) polystyrene sulfonate (PEDOT:PSS) film. The PCE of optimized OPV cell on CNT–CuI reaches 2.97%, which is comparable to the efficient small-molecule OPV cells using metal nanowires and graphene as transparent electrodes.^35–40

Experimental

CNT transparent electrode

Single-walled CNT with a mean diameter of 1.5 nm was synthesized by enhanced direct injection pyrolytic synthesis (eDIPS) method.⁴¹ The detailed description about CNT dispersion and deposition is given elsewhere.³² CNT films with different optical transmittance were prepared by a doctor blade method. Here, a photonic curing process (NovaCentrix PulseForge1300) and solution immersion process in 2-propanol were used to remove the polymer dispersant in CNT films. CuI and MoO₃ were used for CNT doping, respectively. CuI solution (0.05 M) was prepared by dissolving the CuI (Sigma-Aldrich) powder in acetonitrile (Wako) under vigorous stirring. The CuI solution was then spin-coated on CNT with a spin speed of 1500 rpm. As-prepared CuI film was dried at 100 °C for 10 min on a hotplate. 10 nm-thick MoO₃ was vacuum evaporated on CNT film at 2 × 10⁻⁵ Pa. Here, photonic curing process was further used to improve the doping effects, and correspondingly decrease the sheet resistance. The details about photonic curing is shown in Fig. S1.†

OPV cell fabrication

OPV cells were fabricated on patterned CNT electrodes with optical transmittance ranging from 70% to 90%. All of the small molecule materials were commercially available, C60 (Frontier carbon) was used as bought, while DBP (Lumtec) and DIP (Lumtec) were further purified three times prior to use. The thickness and growth rate were monitored with a quartz crystal oscillator. All of films were evaporated with a deposition pressure of about 2 × 10⁻⁵ Pa. 10 nm-thick DIP (or not), 15 nm-thick DBP, 50 nm-thick C60 films and 10 nm-thick (BCP) films were deposited as donor, acceptor and exciton blocker, respectively. The growth rates were kept at 0.1–0.2 Å s⁻¹ for organic materials, and 1 Å s⁻¹ for Al electrodes. The active device area was 6 mm². For some devices, PEDOT:PSS (Clevios P VP AI 4083) was spin coated with a thickness of 40 nm. Here, PEDOT:PSS aqueous dispersion was mixed with 5 vol% 2-propanol (Wako) to improve its wettability on hydrophobic surface of CNT. For comparison, OPV cells with same device structures were also prepared on commercial ITO coated glass substrate, which has a sheet resistance of 10 Ω per square with 88% transmittance.

Characterization

The cell preparation process as well as the electrical measurements were performed at room temperature without air exposure (vacuum or N₂ gas atmosphere). The current density–voltage (J–V) characteristics of the cells were measured under simulated AM 1.5G solar illumination with a Keithley 2401 Digital Source Meter. Incident power was calibrated by using a standard silicon photovoltaic to match 1 sun intensity (100 mW cm⁻²). External quantum efficiency (EQE) spectra were collected using a xenon lamp, integrated with a computer controlled monochromator. To calculate the internal quantum efficiency (IQE), the absolute reflectance of actual cell was measured with an integrating sphere system. Atomic force microscopy (AFM, SII SPA300) analyses were performed with the dynamic mode. The optical absorption was taken on a V-780 (JUSCO) spectrometer. The sheet resistance was collected by Loresta MCP-T370 (Mitsubishi Chem.) with a standard four-point probe.

Results and discussion

Fig. 1(a) schematically illustrates the solution-based process for CNT–CuI hybrid film, and Fig. 1(b)–(d) give their corresponding AFM images showing the structural evolution. The randomly distributed CNT network in Fig. 1(b) indicates that the CNT was well dispersed. Spin coating of a CuI solution resulted in the formation of continuous CuI films with a few of nanoparticles, as shown in Fig. 1(c). Fig. S2† indicates that CuI film just covered the CNT surface to form a CNT/CuI bilayer, while apparently, CuI could not penetrate into the porous CNT networks. It is possibly due to the hydrophobic characteristics of CNT. We have shown that the photonic curing enables the structural modification of CNT–CuI film, while leads to slight structural damages in CNT networks.⁴² Fig. 1(d) shows that after photonic curing, CuI film evolved into the randomly distributed nanoparticles with diameters of 40–100 nm. Interestingly, most of the nanoparticles were individually located at the cross points of two or more CNTs. We also note that some nanoparticles with smaller size penetrated deep into the CNT films. Moreover, Fig. S3† gives the AFM image of CNT–CuI after mild curing. The continuous CuI film was divided to several small agglomerations, accompanied with the formation of smaller nanoparticles. It indicates that the rapid thermal process induced by photonic curing reached the melting point (606 °C) of CuI. Compared to evaporated CuI film, spin-coated CuI film exhibited very different initial surface morphology, while photonic curing caused the formation of similar CNT–CuI hybrid structure.³² These interconnecting nodes are highly desired for ideal CNT–CNT contacts to achieve efficient charge transport.

Fig. 1 (a) Schematics of fabrication process of the CNT–CuI hybrid film. The AFM images of (b) CNT film, (c) CNT/CuI bilayer film and (d) CNT–CuI hybrid film after photonic curing. The scale bars are 200 nm. The heights are 40, 15 and 60 nm for (b), (c) and (d), respectively.

Fig. 2 shows the doping-induced variations in optical transmittance and sheet resistance. Here, we fabricated a series of CNT films with well-controlled transmittances. In Fig. 2(a), intrinsic CNT films exhibited high sheet resistances of above 1000 Ω per square, which were slightly decreased after spin-coating of CuI. Obviously, the continuous CuI film with a high sheet resistance of above 10⁵ Ω per square contributed limitedly to the conductivity of CNT films. However, the photonic curing greatly decreased the sheet resistance by a factor of 10. As a result, the CNT–CuI films with transmittance of around 75%, 80%, 85% and 90% exhibited sheet resistances of 50 ± 10, 70 ± 10, 100 ± 10, and 180 ± 20 Ω per square, respectively. The detailed effects of CuI doping have been discussed with spectroscopies.³⁹ Moreover, Fig. 2(b) shows that the photonic curing process led to a total removal of the S22 peak at around 1350 nm, a well-known phenomenon for doping semiconducting CNT with acid.^43,44 It indicates that photonic curing encouraged CuI doping, and induced the transition from semiconducting to metallic CNTs, where carriers (holes) were efficiently injected into the band structure of CNT. After photonic curing, CuI nanoparticles modified the contact between the CNTs, and thus dramatically improved the network conductivity, while the transmittance remained constantly high. On the other hand, the sheet resistances are comparable to those using vacuum evaporated CuI.³¹ Their differences in morphology suggest that optimization of CuI thickness and photonic curing process may further improve the conductivity. Comparing with the transmittance spectra of ITO, with similar transmittance at 550 nm, CNT–CuI electrode absorbs less in the wavelength of above 600 nm. Therefore, CNT–CuI is promising for further improving OPV cells using narrow-bandgap materials or tandem structures, which absorb near infrared light.

Fig. 2 (a) Optical transmittance (λ = 550 nm) versus sheet resistance, (b) CuI doping-, and (c) thickness-dependent transmittance spectra (excluding glass) of the CNT–CuI films. The transmittance spectra of ITO are provided for reference in (c).

It has been reported that MoO₃ can be an efficient and stable dopant to improve the conductivity of CNT film, while it required thermal annealing at above 450 °C for 3 h in inert gas atmosphere.²⁹ Here, we utilized photonic curing to modify CNT–MoO₃ films. Fig. 3(a) gives the surface morphology of CNT–MoO₃ film before curing. No clear structural variation induced by MoO₃ can be observed. Unlike CuI, 10 nm thick MoO₃ film by vacuum evaporation exhibited amorphous state (or crystallites) with small grain size of 10 nm, as shown in Fig. S4.† After photonic curing with same parameters used for CNT–CuI, MoO₃ nanoparticles with diameters of 50–100 nm appeared in Fig. 3(b). These nanoparticles fully covered CNT film. Further curing process with stronger power finally resulted in the formation of a novel CNT–MoO₃ hybrid film, as shown in Fig. 3(c). Note that these nanoparticles are also individually located at the cross points of CNT bundles as interconnecting nodes. MoO₃ has a melting point of 795 °C, being much higher than the softening point of glass substrate (550–700 °C) used in the present work. Thus, photonic curing process realized a quick high temperature process on low-temperature substrate. It also indicates the universality of our method to fabricate CNT hybrid networks with different nanoparticles for desired applications. In Fig. 3(d), S22 peak disappears, suggesting that photonic curing simultaneously enhanced charge-transfer doping of MoO₃. However, we note 3–4% drop in transmittance ranging from 500 to 1200 nm after photonic curing. Similar phenomena occurred for MoO₃ pure layer after photonic curing, while the reason remains unclear. As shown in Fig. 3(e), photonic curing also increased the conductivity of CNT–MoO₃ by a factor of around 10. As a result, the CNT–MoO₃ films with transmittance of around 72%, 76%, 82% and 87% exhibited sheet resistances of 90 ± 10, 110 ± 10, 220 ± 30, and 350 ± 50 Ω per square, respectively. The results are comparable to the results using high-temperature annealing.²⁹

Fig. 3 AFM images of the CNT–MoO₃ films (a) before, after (b) mild and (c) final photonic curing. The scale bars are 200 nm. The heights are 25, 40 and 50 nm for (a), (b), and (c), respectively. (d) Transmittance spectra and (e) optical transmittance versus sheet resistance of the CNT–MoO₃ films. The transmittance spectra of ITO are provided for reference in (d).

For OPV cells using evaporated small molecules, controlling the molecular growth is of great importance to achieve high performances. We deposited a few of buffer layers to modify the CNT transparent electrodes, and fabricate series of OPV cells with a DBP (15 nm)/C60 (50 nm) bilayer heterojunction architecture. Firstly, a 40 nm-thick PEDOT:PSS film, a well-known hole extraction layer was spin coated to smoothen the surface of CNT films. PEDOT:PSS film has been widely used for fabricating ITO-free OPV cells.^35–40 Fig. 4 shows the cell structure and performances, and Table 1 summaries the detailed parameters. For the bare CNT–CuI film, large leakage currents can be observed in both of dark and light J–V curves in Fig. 4(b) and (c), suggesting that the large surface roughness of CNT led to severe damages (shorting the electrodes) in the cell. Although its short-circuit current (J_SC) and EQE was very comparable to that for reference cell prepared on ITO, the poor rectifying characteristics resulted in a very low PCE of 0.19%. However, depositing a PEDOT:PSS film on CNT–CuI significantly improved the rectifying characteristics. Compared to the reference cell on ITO, the cell on CNT–CuI exhibited large decreases in open-circuit voltage (V_OC) and fill factor (FF), possibly because of the much larger sheet resistance of electrode and the large leakage current.

Fig. 4 (a) Device structure, current density versus voltage (J–V) characteristics under (b) dark and (c) 1 sun illumination, and (d) external quantum efficiencies (EQE) of the OPV cells with PEDOT:PSS/DBP/C60 structure.

Table 1 Device performances of DBP/C60 OPV cells using CNT hybrid transparent electrodes

Transparent electrode	Transmittance (%)	Sheet resistance (Ω per square)	Buffer layer	PCE (%)	JSC (mA cm^−2)	VOC (V)	FF	Series resistance (Ω cm^2)	Shunt resistance (Ω cm^2)
a Measured before coating with PEDOT:PSS (PP).
ITO	88	10	w/o	2.28	4.1	0.90	0.62	6	1.8 × 10^3
CNT–CuI	85	100 ± 10	w/o	0.19	3.5	0.17	0.32	14	1.3 × 10^2
CNT–CuI	85	100 ± 10	PP	1.95	5.8	0.76	0.44	38	7.2 × 10^2
CNT(HNO3)	85	110 ± 10^a	PP	1.40	5.7	0.80	0.30	111	4.2 × 10^2
ITO	88	10	PP/DIP	3.43	6.3	0.91	0.60	8	3.5 × 10^3
CNT–CuI	75	50 ± 10	PP/DIP	2.33	6.6	0.67	0.55	13	1.6 × 10^3
CNT–CuI	80	70 ± 10	PP/DIP	2.98	7.3	0.77	0.53	20	2.0 × 10^3
CNT–CuI	85	100 ± 10	PP/DIP	2.70	7.7	0.79	0.45	22	6.2 × 10^2
CNT–CuI	90	180 ± 20	PP/DIP	2.75	8.2	0.84	0.40	55	5.5 × 10^2
ITO	88	10	MoO3	2.42	4.2	0.90	0.64	5	2.1 × 10^3
CNT–MoO3	72	90 ± 10	MoO3	1.61	5.2	0.58	0.53	12	6.0 × 10^2
CNT–MoO3	76	110 ± 10	MoO3	1.87	6.6	0.61	0.46	27	7.2 × 10^2
CNT–MoO3	82	220 ± 30	MoO3	1.54	6.3	0.61	0.40	39	8.1 × 10^2
CNT–MoO3	87	350 ± 50	MoO3	1.62	6.8	0.66	0.35	50	1.0 × 10^3

---

On the other hand, 40% increase in J_SC can be observed. The enhancement in J_SC is identified by the EQE spectra in Fig. 4(d), where the peaks at wavelengths of λ = 560 nm and λ = 605 nm correspond to the DBP absorption region, while the peak at λ = 440 nm corresponds to the C60 absorption region. As a result, the PCE of cell on CNT–CuI reached 1.95%, being close to the cell on bare ITO (2.28%). Moreover, we fabricated similar cell on CNT doped with HNO₃. Its series resistance of 111 Ω cm², being 3-fold larger than that for CNT–CuI, implies that device process in vacuum greatly deteriorated the film conductivity. The poor FF of 0.30 demonstrates that such unstable doping method is not applicable. On the other hand, it has been reported that CuI nanoparticles can be an efficient structural template to improve OPV cells.³¹ Note that the cells on HNO₃-doped CNT and CNT–CuI hybrids exhibited very similar J_SC and EQE spectra. Apparently, here, the variations in optical transmittance and surface morphology due to the existence of CuI nanoparticle can be negligible. On the other hand, PEDOT:PSS film on ITO showed a very smooth surface with a root mean square (rms) roughness of 1.1 nm in Fig. 4(b), being slightly smaller (around 1.9 nm) than that of bare ITO, while it showed a hill-like morphology (see Fig. 4(c)) with a high rms roughness of 4.6 nm on CNT–CuI. It indicates that the formation of PEDOT:PSS film strongly depended on the surface morphology of CNT. The rough PEDOT:PSS surface may lead to a rough organic film, thus larger DBP/C60 interfacial area was obtained for more efficient exciton dissociation. We will discuss the improvement in J_SC in following section.

We further deposited a DIP film on PEDOT:PSS to construct a DIP/DBP/C60 cascade-type architecture to improve the cell performances. It has been demonstrated that such device structure not only provides larger interface area over conventional planar cell, but also enables more efficient exciton transfer in DBP film.⁴⁵ As shown in Fig. 5, OPV cells on a series of CNT–CuI films were fabricated to understand the effects of transmittance and the resistance on performances. Table 1 shows that the series resistance is approximately proportional to the sheet resistance of CNT–CuI. Correspondingly, the FF decreased from 0.55 to 0.40 with increasing the sheet resistance from 50 to 180 Ω per square. However, J_SC increased from 6.6 to 8.2 mA cm⁻² with increasing transmittance from 75% to 90%. Fig. 5(f) identifies that the variations in EQE spectra are related to the transmittance spectra in Fig. 2(c). Especially, the cell on CNT–CuI with a transmittance of 75% exhibited a large decrease in EQE spectra from 400 to 500 nm. It suggests that the CNT–CuI film with low transmittance absorbed too much light to efficiently harvest light for C60 film.

Fig. 5 (a) Device structure, AFM images of the 40 nm-thick PEDOT:PSS films on (b) ITO and (c) CNT–CuI, and current density versus voltage (J–V) characteristics under (d) dark and (e) 1 sun illumination, and (f) external quantum efficiencies (EQE) of the OPV cells on CNT–CuI with PEDOT:PSS/DIP/DBP/C60 structure.

Moreover, all the OPV cells prepared on CNT–CuI exhibited smaller V_OC than ITO cell (0.90 V). Generally, V_OC is determined by the difference between the HOMO of the donor and the lowest unoccupied molecular orbital (LUMO) of the acceptor, while it depends on a number of parameters including illumination intensity, temperature, mobility of charge carrier, and electrode/organic interfaces.⁴⁶ Large leakage current (rectifying properties) and high sheet resistance due to CNT–CuI electrode are the possible reasons for the degradations in V_OC by comparing with ITO cell. The detrimental effect of the leakage current to V_OC was demonstrated, where the V_OC was 50% decreased by increasing the leakage current by two orders of magnitude in small-molecular solar cells.⁴⁷ On the other hand, for the OPV cells on CNT–CuI electrodes, we note that V_OC increased from 0.67 to 0.84 V with transmission of the CNT–CuI electrodes. The shunt resistance (in Table 1) seems to be independent on V_OC, suggesting that the variations in V_OC can be attributed to the actual illumination intensities due to the different transmittances of CNT–CuI. Trap-assisted recombination, which degrades the V_OC becomes more important when the light-generated charge-carrier density is smaller at lower illumination intensity.⁴⁸ The results indicate that both of high transmittance and low sheet resistance of CNT transparent are important to achieve efficient OPV cells. In addition, comparing to the reference cell fabricated on ITO, all of the cells on CNT–CuI exhibited smaller V_OC, and FF, but larger J_SC. The increase in the EQE spectra from 400 to 650 nm, identifies that cascade structure was successfully established on CNT–CuI,⁴⁴ and the DBP/C60 contact area was larger on CNT–CuI film. As a result, the best cell fabricated on the CNT–CuI film with 80% transmittance exhibited J_SC of 7.3 mA cm⁻², V_OC of 0.77 V, FF of 0.53, leading to a PCE of 2.97%, being close to the cell (3.43%) on conventional ITO. According to our knowledge, it is the first efficient OPV cell based evaporating small molecules on CNT transparent electrode.

Furthermore, to understand the origins of the improvement in J_SC by using CNT–CuI electrode, we measured the total reflectance (R) of the OPV cells and the transmittance (T) of the CNT–CuI films (Fig. 2(c)), and calculated the IQE by further excluding the absorption of CNT–CuI:

IQE(λ) = EQE(λ)/[1 − R(λ) − (1 − T(λ))] (1)

In Fig. 6, the OPV cells on series of CNT–CuI films exhibited very similar IQE spectra in whole wavelength, indicating that the sheet resistance of CNT–CuI played a negligible role in the efficiencies in exciton dissociation, charge transfer and charge collection. Thus, those variations in J_SC should be attributed to the absorption of CNT–CuI electrodes. On the other hand, compared to the reference cell on ITO, IQE spectra increased from 60% to 70%, and from 70% to 85% for C60 and DBP absorption region, respectively. It implies that larger DBP/C60 interface due to the porous CNT–CuI film led to increase in exciton dissociation efficiency in whole absorption region, resulting in 20% increase in J_SC. On the other hand, the 40% increase in J_SC is observed in OPV cell without DIP layer, compared with ITO cell, as shown in Fig. 4. Clearly, increasing the interfacial area could be more efficient for exciton dissociation without DIP layer, because more excitons would be annihilated on DBP/ITO interface. Overall, constructing nanostructured template is an efficient strategy for improving the OPV cells no mater ITO is used or not.^31,45 Especially, capitalizing the structural characteristics (1-dimensional nanostructures) of non-traditional electrodes including CNT, metal nanowires may provide a new strategy for efficient OPV cells.

Fig. 6 Internal quantum efficiencies (IQE) of the OPV cells on CNT–CuI with a PEDOT:PSS/DIP/DBP/C60 structure.

MoO₃ is a wide bandgap semiconductor, which has been widely used in OPV cells as a hole extraction layer.³⁰ We further deposited a 10 nm-thick MoO₃ film on CNT–MoO₃ to modify the interface between the CNT and organic layer. Thus, the unique morphology (Fig. 3(c)) can remain to construct nanostructured OPV cell as a template without PEDOT:PSS coating. The results of OPV cells fabricated on CNT–MoO₃ are given in Fig. 7. As shown in Fig. 3(d) in spite of the fact that very similar CNT films were fabricated for CuI and MoO₃ doping, photonic curing caused 3–4% drop in transmittance of CNT–MoO₃. We note that the cells on CNT–MoO₃ films exhibited very similar transmittance/resistance dependence with the cells on CNT–CuI. Table 1 shows that series resistance increased from 12 to 50 Ω per square with increasing the sheet resistance from 90 to 350 Ω per square, resulting in a reduction in FF from 0.53 to 0.35. The J_SC increased from 5.2 to 6.8 mA cm⁻², while the V_OC increased from 0.58 to 0.66 V with increasing the transmittance from 72 to 87%. Moreover, compared to the reference cell fabricated on ITO, the cells on CNT–MoO₃ exhibited larger J_SC. Apparently, the MoO₃ nanoparticles in CNT–MoO₃ film led to rougher DBP film, and correspondingly larger DBP/C60 surface area for more efficient exciton dissociation. The EQE spectra in Fig. 7(d) indicate that CNT–MoO₃ with similar 87% transmittance led to about 100% and 50% increases in C60 and DBP absorption region, respectively. More efficient light harvesting in C60 absorption region is possibly attributed to two reasons: (1) more excitons in C60 layer can be dissociated at rough DBP/C60 interface because C60 (20–40 nm) exhibits higher diffusion length than DBP (10 nm);⁴⁵ (2) excitons in C60 layer can be dissociated at C60/MoO₃ interface⁴⁹ since DBP film with a thickness of 15 nm could not fully cover the rough CNT–MoO₃ surface. Moreover, all the OPV cells prepared on CNT–MoO₃ exhibited smaller V_OC and FF, comparing with ITO cell. It indicates that the large leakage current and sheet resistance should be further improved. The best cell fabricated on the CNT–MoO₃ film with 80% transmittance exhibited a PCE of 1.87%, which was 2.42% for the reference cell on ITO.

Fig. 7 (a) Device structure, current density versus voltage (J–V) characteristics under (b) dark and (c) 1 sun illumination, and (d) external quantum efficiencies (EQE) of the OPV cells on CNT–MoO₃ with a MoO₃/DBP/C60 structure.

Conclusions

We demonstrated a series of CNT–CuI and CNT–MoO₃ hybrid films as transparent electrodes for small-molecule OPV cells. Photonic curing process was utilized to encourage the doping effects and construct novel CNT hybrids with spin-coated CuI and vacuum-evaporated MoO₃. The CNT–CuI and CNT–MoO₃ hybrid films exhibited stable sheet resistances of 70 and 110 Ω per square at around 80% transmittance, respectively. We showed the importance of PEDOT:PSS film, which can smoothen the rough surface of CNT–CuI, and correspondingly suppress the leakage current of the OPV cell. Although the large sheet resistance of CNT–CuI led to poor FF and V_OC, its rough surface resulted in high J_SC. The DBP/C60 bilayer cell on CNT–CuI exhibited a PCE of 1.95%, which is much higher than that (1.40%) on CNT with unstable HNO₃ doping. Further construction of DIP/DBP/C60 cascade-type device architecture on CNT–CuI greatly improved the cell performance, especially for photon-current generation efficiency. The best cell fabricated on the CNT–CuI film with 80% transmittance exhibited J_SC of 7.3 mA cm⁻², V_OC of 0.77 V, FF of 0.53, leading to a PCE of 2.97%, being close to that for the reference cell (3.43%) on ITO. On the other hand, DBP/C60 bilayer cells were directly fabricated on CNT–MoO₃ hybrids without further PEDOT:PSS coating. The nanostructured CNT–MoO₃ also greatly improved photon-current generation efficiency, especially in C60 absorption region, while the poor FF and V_OC resulted in the maximum PCE of 1.87%. We believe that in spite of the fact that the sheet resistance of CNT hybrids should be further reduced compared with ITO, the present methodology constructing nanostructured architecture is of great importance for developing high performance OPV cells no matter whether ITO is used or not.

Acknowledgements

This work was supported by JSPS Grant-in-Aid for Research Activity Start-up. The authors acknowledge the help from Ms Y. Yokota.

Notes and references

1. S. R. Forrest, Nature, 2004, 428, 911.
2. G. Li, R. Zhu and Y. Yang, Nat. Photonics, 2012, 6, 153.
3. Z. He, B. Xiao, F. Liu, H. Wu, Y. Yang, S. Xiao, C. Wang, T. P. Russell and Y. Cao, Nat. Photonics, 2015, 9, 174.
4. S. Sista, Z. Hong, L. M. Chen and Y. Yang, Energy Environ. Sci., 2011, 4, 1606.
5. L. Lu, T. Zheng, Q. Wu, A. M. Schneider, D. Zhao and L. Yu, Chem. Rev., 2015, 115, 12666.
6. B. Kan, M. Li, Q. Zhang, F. Liu, X. Wan, Y. Wang, W. Ni, G. Long, X. Yang, H. Feng, Y. Zuo, M. Zhang, F. Huang, Y. Cao, T. P. Russell and Y. Chen, J. Am. Chem. Soc., 2015, 137, 3886.
7. K. Cnops, B. P. Rand, D. Cheyns, B. Verreet, M. A. Empl and P. Heremans, Nat. Commun., 2014, 5, 3406.
8. M. E. Ragoussi and T. Torres, Chem. Commun., 2015, 51, 3957.
9. F. Yang, M. Shtein and S. R. Forrest, Nat. Mater., 2005, 4, 37.
10. J. G. V. Dijken, M. D. Fleischauer and M. J. Brett, J. Mater. Chem., 2011, 21, 1013.
11. Y. Zheng, R. Bekele, J. Ouyang and J. Xue, Org. Electron., 2009, 10, 1621.
12. N. Li and S. R. Forrest, Appl. Phys. Lett., 2009, 95, 123309.
13. Y. Zhou, T. Taima, T. Miyadera, T. Yamanari, M. Kitamura, K. Nakatsu and Y. Yoshida, Appl. Phys. Lett., 2012, 100, 233302.
14. X. He, F. Gao, G. Tu, D. Hasko, S. Huttner, U. Steiner, N. C. Greenham, R. H. Friend and W. T. S. Huck, Nano Lett., 2010, 10, 1302.
15. Z. P. Wang, T. Miyadera, T. Yamanari and Y. Yoshida, ACS Appl. Mater. Interfaces, 2014, 6, 6369.
16. F. C. Krebs, N. Espinosa, M. Hösel, R. R. Søndergaard and M. Jørgensen, Adv. Mater., 2014, 26, 29.
17. W. J. Silva, H. P. Kim, A. Yusoff and J. Jang, Nanoscale, 2013, 5, 9324.
18. C. Roldán-Carmona, O. Malinkiewicz, A. Soriano, G. M. Espallargas, A. Garcia, P. Reinecke, T. Kroyer, M. I. Dar, M. K. Nazeeruddin and H. J. Bolink, Energy Environ. Sci., 2014, 7, 994.
19. W. Wang, T. S. Bae, Y. H. Park, D. H. Kim, S. Lee, G. Min, G. H. Lee, M. Song and J. Yun, Nanoscale, 2014, 6, 6911.
20. T. R. Andersen, H. F. Dam, M. Hosel, M. Helgesen, J. E. Carle, T. T. Larsen-Olsen, S. A. Gevorgyan, J. W. Andreasen, J. Adams, N. Li, F. Machui, G. D. Spyropoulos, T. Ameri, N. Lemaitre, M. Legros, A. Scheel, D. Gaiser, K. Kreul, S. Berny, O. R. Lozman, S. Nordman, M. Valimaki, M. Vilkman, R. R. Sondergaard, M. Jorgensen, C. J. Brabec and F. C. Krebs, Energy Environ. Sci., 2014, 7, 2925.
21. J. S. Yu, I. Kim, J. S. Kim, J. Jo, T. T. Larsen-Olsen, R. R. Søndergaard, M. Hösel, D. Angmo, M. Jørgensen and F. C. Krebs, Nanoscale, 2012, 4, 6032.
22. G. A. dos Reis Benatto, B. Roth, M. V. Madsen, M. Hösel, R. R. Søndergaard, M. Jørgensen and F. C. Krebs, Adv. Energy Mater., 2014, 3, 1400732.
23. M. Kaltenbrunner, M. S. White, E. D. Głowacki, T. Sekitani, T. Someya, N. S. Sariciftci and S. Bauer, Nat. Commun., 2012, 3, 770.
24. Y. Liu, N. Qi, T. Song, M. Jia, Z. Xia, Z. Yuan, W. Yuan, K. Q. Zhang and B. Sun, ACS Appl. Mater. Interfaces, 2014, 6, 20670.
25. M. Kaltenbrunner, G. Adam, E. D. Głowacki, M. l Drack, R. Schwödiauer, L. Leonat, D. H. Apaydin, H. Groiss, M. C. Scharber, M. S. White, N. S. Sariciftci and S. Bauer, Nat. Mater., 2015, 14, 1032.
26. J. Du, S. Pei, L. Ma and H. Cheng, Adv. Mater., 2014, 26, 1958.
27. D. S. Hecht, L. Hu and G. Irvin, Adv. Mater., 2011, 23, 1482.
28. R. Jackson, B. Domercq, R. Jain, B. Kippelen and S. Graham, Adv. Funct. Mater., 2008, 18, 2548.
29. S. L. Hellstrom, M. Vosgueritchian, R. M. Stoltenberg, I. Irfan, M. Hammock, Y. B. Wang, Y. Gao and Z. Bao, Nano Lett., 2012, 12, 3574.
30. S. Murase and Y. Yang, Adv. Mater., 2012, 24, 2459.
31. Y. Zhou, T. Taima, T. Miyadera, T. Yamanari, M. Kitamura, K. Nakatsu and Y. Yoshida, Nano Lett., 2012, 12, 4146.
32. Y. Zhou, S. Shimada, T. Saito and R. Azumi, Carbon, 2015, 87, 61.
33. Y. Zhou, T. Taima, Y. Shibata, T. Miyadera, T. Yamanari and Y. Yoshida, Sol. Energy Mater. Sol. Cells, 2011, 95, 2861.
34. D. Fujishima, H. Kanno, T. Kinoshita, E. Maruyama, M. Tanaka, M. Shirakawa and K. Shibata, Sol. Energy Mater. Sol. Cells, 2009, 93, 1029.
35. F. Selzer, N. Weiß, D. Kneppe, L. Bormann, C. Sachse, N. Gaponik, A. Eychmüller, K. Leo and L. Müller-Meskamp, Nanoscale, 2015, 7, 2777.
36. C. Sachse, N. Weiß, N. Gaponik, L. Müller-Meskamp, A. Eychmueller and K. Leo, Adv. Energy Mater., 2014, 4, 1300737.
37. J. Y. Lee, S. T. Connor, Y. Cui and P. Peumans, Nano Lett., 2010, 10, 1276.
38. H. Park, Y. M. Shi and J. Kong, Nanoscale, 2013, 5, 8934.
39. H. Park, S. Chang, M. Smith, S. Gradecak and J. Kong, Sci. Rep., 2013, 3, 1581.
40. H. Park, R. M. Howden, M. C. Barr, V. Bulović, K. Gleason and J. Kong, ACS Nano, 2012, 6, 6370.
41. T. Saito, S. Ohshima, T. Okazaki, S. Ohmori, M. Yumura and S. Iijima, J. Nanosci. Nanotechnol., 2008, 8, 6153.
42. Y. Zhou, S. Shimada, T. Saito and R. Azumi, J. Appl. Phys., 2015, 118, 215305.
43. J. L. Blackburn, T. M. Barnes, M. C. Beard, Y. H. Kim, R. C. Tenent, T. J. McDonald, B. To, T. J. Coutts and M. J. Heben, ACS Nano, 2008, 2, 1266.
44. J. W. Jo, J. W. Jung, J. U. Lee and W. H. Jo, ACS Nano, 2010, 4, 5382.
45. Y. Zhou, T. Taima, T. Kuwabara and K. Takahashi, Adv. Mater., 2013, 25, 6069.
46. T. C. Monson, M. T. Lloyd, D. C. Olson, Y.-J. Lee and J. W. P. Hsu, Adv. Mater., 2008, 20, 4755.
47. D. B. Staple, P. A. K. Oliver and I. G. Hill, Phys. Rev. B: Condens. Matter Mater. Phys., 2014, 89, 205313.
48. N. K. Elumalai and A. Uddin, Energy Environ. Sci., 2016, 9, 391.
49. L. Chen, Q. Zhang, Y. Lei, F. Zhu, B. Wu, T. Zhang, G. Niu, Z. Xiong and Q. Song, Phys. Chem. Chem. Phys., 2013, 15, 16891.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ra01674j

---