Formation of cathode buffer layer by surface segregation of fluoroalkyl-modified ZnO for polymer solar cells

Abstract

Novel zinc oxide nanoparticles (ZnO NPs) modified by silanization using triethoxy-1H,1H,2H,2H-tridecafluoro-n-octylsilane (TTFO), referred to as ZnOF NPs, have been successfully synthesized. Driven by the surface segregation behavior of the fluoroalkyl chains ascribed to their low surface energy, ZnOF NPs can migrate from the blend system with poly(3-hexylthiophene):[6,6]-phenyl-C₆₁-butyric acid methyl ester (P3HT:PCBM) to the surface of the active layer during the annealing process and consequently self-assemble as a cathode buffer layer. The addition of functionalized ZnOF NPs assists in the stacking of the P3HT chains to form a favorable morphology of the active layer with remarkable phase separation, especially upon annealing optimization in o-DCB solvent. The best power conversion efficiency (PCE) of 2.4% is achieved with an open-circuit voltage (V_oc) of 0.49 V, short-circuit current density (J_sc) of 8.1 mA cm⁻² and fill factor (FF) of 61%, based on the self-assembled cathode buffer layer of ZnOF NPs (5 wt%) upon annealing in o-DCB. Therefore, this novel approach could realize the fabrication of both the active layer and cathode buffer layer through a single step, which not only simplifies the fabrication procedure and reduces the manufacturing cost of polymer solar cells, but also increases the PCE, by reduction of the Schottky barrier at the interface, and the stability of the devices.

---

Introduction

Organic photovoltaic devices (OPVs) based on conjugated polymer materials have been paid a tremendous amount of attention, partly because they feature many advantageous properties, including low weight, mechanical flexibility, and low-temperature and low-cost fabrication of large-area devices.^1–3 Recently, increasing research efforts have been devoted to enhancing the photovoltaic efficiency. Optimization of the morphology of the active layer by improving the processing methods or adding some additives in the active layer, as well as utilizing an effective buffer interface (including a hole and electron transporting layer), are valid for realizing high efficiency, since they greatly affect the flow of charge carriers across or along the material interfaces.^4–10

The buffer interface has been commonly employed to improve the contact between the active organic layer and electrodes. Lithium fluoride (LiF),¹¹ cesium carbonate (Cs₂CO₃),¹² titanium oxide (TiO_x),¹³ zinc oxide (ZnO)¹⁴ and self-assembled monolayers¹⁵ are used as n-type electron transporting layer materials in polymer solar cells. They create an effective metal–organic interface, which decreases the contact resistance between the active layer and cathode electrode, forms a favorable dipole moment across the junction and makes the Fermi levels of the layers match well to enhance the efficiency of charge collection.⁸ The buffer layer also prevents oxygen and humidity diffusing into the active layer, thereby improving the lifetime of the unpackaged devices.

Generally speaking, the formation of the conventional cathode buffer layer (CBL) needs an additional fabrication step. For instance, LiF is prepared through vacuum deposition,^11,15–17 while ZnO and TiO_x are dipped or spin coated.^{13,14,18–20} The additional processes have an adverse effect on the active layer, make the crafting complicated and increase the manufacturing cost to some extent. Furthermore, owing to the vertical phase separation with thermodynamic instability, the elements of the additional buffer layer may diffuse into the active layer,⁵ which could affect the efficiency and the stability of the solar cells. Hence, methods to fabricate both organic materials and cathode buffer layer in a single step and keep the materials stable over a long time are valuable. Zhang et al. reported the application of non-conjugated polymers (surfactants), such as poly(ethylene oxide) (PEO), as the buffer layer between the active layer and the metal cathode to improve the performance of the polymer devices.²¹ Cheun et al. studied the role of thermally-induced vertical phase segregation and crystallization on the photovoltaic performance of bulk heterojunction inverted polymer solar cells.²² Clark et al. found that there was a thin 6,6-phenyl-C₆₁-butyric acid methyl ester (PCBM)-rich layer between the active layer and the air interface during film casting. Furthermore, the phenomenon of spontaneous vertical phase separation in polymer photovoltaic devices, decided by the different free energies, had been verified by angle-resolved X-ray photoelectron spectroscopy (ARXPS).²³ Yao et al. successfully synthesized a novel fluoroalkyl side-chain diblock copolymer, poly(3-hexylthiophene)-block-poly[3-(4-(3,3,4,4,5,5,6,6,7,7,8,8,8-tridecafluorooctyloxy)phenyl-decyloxy)thiophene] (P3HT-b-P3FAT). The fluorinated P3FAT polymer chains could spontaneously aggregate on the surface of poly-(3-hexylthiophene) (P3HT) during spin coating processes, driven by the low surface energy of the fluoroalkyl side chains.²⁴ Wei et al. applied a fullerene derivative with a fluorocarbon chain in bulk-heterojunction polymer solar cells. It could migrate to the surface of the organic layer during spin-casting owing to the low surface energy of the fluorocarbon, and formed a very thin buffer layer between the polymer and the metal electrode in one step.²⁵ Nevertheless, so far there have been few reports on fabricating organic active layers and cathode buffer layers based on the self-assembly of inorganic nanoparticles (such as ZnO NPs) simultaneously through a single step.

In this work, ZnO NPs were modified with triethoxy-1H,1H,2H,2H-tridecafluoro-n-octylsilane (TTFO), resulting in functionalized ZnOF NPs with the fluoroalkyl chains giving a surface segregation function. During the solvent annealing and thermal annealing process of the polymer solar cells, the low surface energy of the fluorocarbon could drive ZnOF NPs to migrate to the surface of the organic layer to form a cathode buffer layer (CBL). Thus, both the active layer and CBL can be fabricated in a single step, which not only simplifies the fabrication process and reduces the manufacturing cost of the solar cells, but also increases the stability of the device, because the vertical phase separation in polymer photovoltaic devices is spontaneous and thermodynamically stable. On the other hand, the movement of ZnOF NPs may actuate the polymer chains to stack in a more orderly manner, consequently forming a favorable heterojunction morphology of the active layer with remarkable phase separation, which is beneficial for boosting the power conversion efficiency (PCE) of polymer solar cells. The amount of added ZnOF NPs and the degree of ZnOF NP migration, regulated by the film-formation process upon annealing in different solvents, produces obvious effects on the stacking of P3HT chains, the bulk heterojunction (BHJ) morphology and the photovoltaic performance. After optimizing the self-assembly of CBL with 5 wt% ZnOF NPs and annealing in the good solvent o-DCB, the PCE of the devices based on an P3HT:PCBM active layer is enhanced to 2.4% with an open-circuit voltage (V_oc) of 0.49 V, short-circuit current density (J_sc) of 8.1 mA cm⁻² and fill factor (FF) of 61%, in comparison to the PCE of 1.2% for the device without a CBL.

Experimental section

Materials

The poly(3-hexylthiophene) (P3HT, M_w = 48 300 g mol⁻¹, head-to-tail, regioregularity > 90%, Rieke Metals, Inc.) and [6,6]-phenyl-C₆₁-butyric acid methyl ester (PCBM, 99.5%; Nano-C) were used as received. ZnO NPs were synthesized following the process reported by Beek et al.²⁶ The general procedure for the preparation of ZnO NPs is as follows: 1.23 g of zinc acetate dihydrate (Zn(CH₃COO)₂·2H₂O, 98%, AR, Aldrich) was dissolved in methanol (55 mL) at 60 °C under vigorous stirring. 25 mL potassium hydroxide (KOH) (90%, AR, Aldrich, 0.34 mmol mL⁻¹) solution in methanol was dropped into the Zn(CH₃COO)₂·2H₂O solution over 20 min under vigorous stirring. The reaction was held at 60 °C for an additional 2 h to yield a homogeneous, clear and transparent solution, which contained ZnO NPs. After the ZnO NP solution fell to room temperature, triethoxy-1H,1H,2H,2H-tridecafluoro-n-octylsilane (TTFO) was added under vigorous stirring for 3 h to graft the fluoroalkyl functional groups onto the ZnO NPs. Finally, the solution was left to precipitate for another 2 h. In this report, the ZnO NPs grafted with functional fluoroalkyl chains were referred to as ZnOF NPs. The precipitate was separated by centrifugation and was washed three times with methanol. O-Dichlorobenzene (o-DCB) was used to disperse the ZnOF NPs, with a content of 20 mg mL⁻¹.

Fabrication of devices

Poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate) (PEDOT:PSS) was prepared on indium tin oxide (ITO) substrates by spin coating. The conductive ITO substrates were cleaned in acetone, soap water, deionized water and isopropanol successively by ultrasonication and dried, then treated with UV ozone for 20 min before spin coating PEDOT:PSS. Then, the substrates were annealed at 140 °C on a hot plate for 20 min. After that, P3HT : PCBM (1 : 1, w/w) blend or NPs (ZnO or ZnOF) dissolved in o-DCB were spin-coated on the ITO/PEDOT:PSS substrate layer at 800 rpm for 40 s to form the active layer. Upon solvent annealing, the films were placed respectively in o-DCB and acetone vapor for 3 h. The additional cathode buffer layer was fabricated by spin-coating of ZnO or ZnOF NPs (20 mg mL⁻¹ dissolved in methanol) above the active layer at 4000 rpm for 60 s, followed by thermal annealing on a hot plate in a glovebox at 150 °C for 10 min. Finally, a 90 nm Al electrode was thermally deposited under vacuum (3 × 10⁻⁶ Torr) and the film was slowly cooled down to room temperature.

Characterization

Hydrogen nuclear magnetic resonance (¹H NMR) and Fourier transform infrared spectroscopy (FTIR) were used to confirm the chemical structure of the ZnOF NPs. ¹H NMR spectra were collected on a Bruker AV 400 NMR spectrometer with deuterated CDCl₃ as the solvent and with tetramethylsilane (δ = 0) as the internal standard. FTIR spectra were recorded by KBr tablet on a Shimadzu IRPrestige-21 FT-IR spectrophotometer. All FTIR spectra were collected at room temperature over a scanning range of 400–4000 cm⁻¹. Field emission transmission electron microscopy (TEM) using a JEOL, JEM-2100F was used to observe the dispersibility of the ZnO and ZnOF NPs. Atomic force microscopy (AFM) images were measured on a Nanoscope III A (Digital Instruments) scanning probe microscope using the tapping mode. To gain insight into the photophysics and electronic properties of the pure ZnO and ZnOF NPs, UV-vis-NIR spectroscopy (performed using a Perkin-Elmer Lambda 750 with integrating sphere) was utilized. Survey X-ray photoelectron spectroscopy (XPS) was applied as an effective measure to certify the content of C 1s, F 1s, Zn 2p and O 1s on the surface of P3HT:PCBM:ZnOF (5%) on the ITO substrate annealed in different solvent vapors. Ultraviolet photoelectron spectroscopy (UPS) was used for determining the energy-level diagram of the related materials. The current–voltage (J–V) characteristics of all devices were processed using a Keithley 2400 Source Meter to test the mixed layer applied to the photovoltaic devices under unencapsulated conditions and were measured in air under 100 mW cm⁻² simulated AM 1.5 G irradiation (Abet Solar Simulator Sun 2000). A photo mask with an area of 0.04 cm² was used to define the active area of the device irradiated by light.

Results and discussion

The synthetic route of the modified ZnO NPs (henceforth referred to as ZnOF NPs) and the fabrication process of the polymer solar cells with self-assembled ZnOF NPs as the cathode buffer layer has been shown in Scheme 1. The functional fluoroalkyl chains were successfully grafted onto the surface of the ZnO NPs by the reaction between the silane coupling agent of triethoxy-1H,1H,2H,2H-tridecafluoro-n-octylsilane (TTFO) and the hydroxyl groups of the ZnO NPs, which could be verified by Fourier transform infrared spectroscopy (FTIR) and hydrogen nuclear magnetic resonance (¹H NMR) characterization. As shown in the FTIR spectra of the ZnO NPs, TTFO and functionalized ZnOF NPs (Fig. 1), a broad band from 3000 cm⁻¹ to 3500 cm⁻¹ is ascribed to the dangling hydroxyl groups on the surface of the ZnO and ZnOF NPs. For the ZnOF NPs, the bands at 1246 cm⁻¹, 1200 cm⁻¹ and 745 cm⁻¹ are attributable to the asymmetrical C–F stretching mode of the –CF₂– and –CF₃ groups. The bands around 1122 cm⁻¹, 1140 cm⁻¹ and 1020 cm⁻¹ originate from the stretching mode of the Si–O groups. The band around 1170 cm⁻¹ was the result of the stretching mode of the C–O groups. The appearance of new bands proves that the fluorinated substitution of the ZnO NPs has been realized.

Scheme 1 Synthetic route of the functionalized ZnOF NPs and fabrication process of the polymer solar cells with self-assembled ZnOF NPs as the cathode buffer layer.

Fig. 1 FTIR spectra of TTFO, bare ZnO and ZnOF NPs.

Fig. 2 shows the ¹H NMR spectra of the ZnO NPs, TTFO and ZnOF NPs. The peak at 1.55 ppm (proton a) is assigned to protons from the hydroxyl groups on the surface of the ZnO NPs. The peaks at 1.22 ppm (9H, proton b), 3.83 ppm (6H, proton c), 0.88 ppm (2H, proton d) and 2.15 ppm (2H, proton e) are ascribed to the different CH₂ and CH₃ groups of TTFO. For the ZnOF NPs, the appearance of a new single-peak at 1.28 ppm (proton f), assigned to methyl groups, has been observed. Meanwhile, the peaks at 1.22 ppm (b) and 3.83 ppm (e) belonging to TTFO disappear, which demonstrates that the functional ZnOF NPs have been synthesized successfully.

Fig. 2 ¹H NMR (400 MHz, CDCl₃) of ZnO NPs, TTFO and ZnOF NPs.

Grafting the fluoroalkyl chains onto the surface of the ZnO NPs effectively increased their dispersion, which could be proven by field emission transmission electron microscopy (TEM). The TEM images of the ZnO NPs and ZnOF NPs (4 mg mL⁻¹ in o-DCB) are shown in Fig. 3. It can be seen that the dispersion of the ZnOF NPs is much better than that of the ZnO NPs. Bringing in fluoroalkyl chains can prevent ZnO NPs from aggregating. The high resolution transmission electron microscopy (HRTEM) images reveal clear lattice fringes of the ZnO and ZnOF NPs. The marked spacing for the lattice fringes of the ZnO NPs is about 0.28 nm, agreeing well with the expected separation of the (002) planes in wurtzite ZnO. For ZnOF NPs, the spacing of the lattice fringes is about 0.28 nm, which displays almost no variation after grafting with fluoroalkyl chains.

Fig. 3 TEM images of (a) ZnO NPs and (b) ZnOF NPs. The insets show the corresponding high-resolution TEM images.

The modified ZnOF NPs have been investigated as additives for mixing with poly(3-hexylthiophene):[6,6]-phenyl-C₆₁-butyric acid methyl ester (P3HT:PCBM) blend for application in polymer solar cells. Driven by the surface segregation of the fluoroalkyl chains due to their low surface energy, ZnOF NPs can migrate from the active layer based on the P3HT:PCBM blend to the surface during the annealing process and self-assemble as a cathode buffer layer (CBL). The process flow of the polymer solar cell fabrication is illustrated in Scheme 1, which shows the fabrication of both the active layer and CBL through a single step. The upward migration of the ZnOF NPs would induce the stacking of the P3HT polymer chains, consequently promoting the formation of a tailored heterojunction morphology of the active layer. In addition, the amount of added ZnOF NPs is a critical factor in the formation of a continuous and dense buffer layer by self-assembly, which creates an effective contact interface between the active layer and cathode electrode and helps to resist current leakage.

To investigate the self-organization ability of the polymer chains in the active layer after incorporation of the ZnOF NPs, the UV-vis absorption spectra of P3HT:PCBM blended with different amounts of ZnOF NPs have been measured and are shown in Fig. 4a, which show a red shift of the P3HT:PCBM absorption. The distinct shoulder peaks corresponding to vibronic features at about 610 nm, 560 nm and 510 nm are related to the 0–0, 0–1 and 0–2 transitions to the vibronic modes of the excited electronic state.^27,28 The height of the peak strongly depends on the intensity of the interchain π–π stacking of the P3HT polymer.²⁹ From the enlarged view of the partial spectrum in Fig. 4b, the red shift and the raised intensity are more pronounced, which indicates that the local ordering and crystallinity of P3HT improve after adding ZnOF NPs. This is because the ZnOF NPs possessing many fluorine atoms with low surface energy tend to migrate to the surface of the active layer in the film-formation process, which influences the crystallization thermodynamic and kinetic parameters of P3HT.³⁰ When the weight percent of ZnOF NPs is below 5 wt%, the self-assembly of the ZnOF NPs is favourable for the stacking of the P3HT chains, consequently improving the crystallinity. The positive effect further increases when the amount of ZnOF NPs added increases. When the weight percent of the ZnOF NPs reaches 10%, the stacking of P3HT can be suppressed by the aggregation of residual ZnOF NPs within the active layer. More remarkably, there is also a red shift for the 10 wt% amount, compared to that of the pure P3HT:PCBM blend. The addition of ZnOF NPs as “dilutants” can increase the extent of phase separation between donor and acceptor, as well as the crystallization of P3HT.

Fig. 4 (a) Normalized absorbance spectra of P3HT:PCBM blend with different contents of ZnOF NPs, without solvent annealing, and (b) the enlarged view of the partial spectrum.

The atomic force microscopy (AFM) images of P3HT:PCBM blended with different amounts of ZnOF NPs without solvent annealing have been displayed in Fig. 5. The blank P3HT:PCBM without ZnOF NPs exhibits a smooth surface morphology and a root mean square (RMS) of 2.64 nm. After adding ZnOF NPs (2.5 wt%) into the P3HT:PCBM blend, the RMS increases to 3.75 nm. From the UV-vis absorption spectra (Fig. 4), we have presumed that the addition of ZnOF NPs can increase the phase separation between donor and acceptor and the crystallization of P3HT, so the RMS growing is reasonable.³¹ On the other hand, this may result from the small quantity of ZnOF NPs could not self-assemble as a continuous and dense layer on the surface, but rather were only isolated nanoclusters. When the weight percent of ZnOF NPs rises from 2.5% to 5%, the RMS decreases to the minimum value (1.05 nm), owing to the formation of a favorable morphology after self-assembly of an appropriate amount of ZnOF NPs. However, the excess ZnOF NPs (10 wt%) not only self-assemble as a complete buffer layer, but also remain partly within the active layer, which impairs the crystallinity of P3HT. Compared to P3HT:PCBM blended with ZnOF NPs (5 wt%), the addition of ZnOF NPs (10 wt%) results in a more rough morphology (RMS of 1.86 nm) and a relative blue shift in the UV-vis absorption related to P3HT:PCBM (Fig. 4). It is noteworthy that the addition of ZnOF NPs is in favor of the stacking of P3HT chains, which agrees well with the obvious red shift in the UV-vis absorption. Furthermore, the amount of ZnOF NPs has an influence on the degree of crystallinity of P3HT to some extent. However, excess ZnOF NPs remain partly within the active layer, which provide bridges for the cross-linking of PCBM rather than a nucleating agent, attributed to the hydrophilic characteristics,³² resulting in apparent aggregation of PCBM. On the other hand, it may hinder the segmental chain motion of P3HT, ultimately leading to a relatively poor ordering.³³

Fig. 5 AFM height images of P3HT:PCBM blend with different contents of ZnOF NPs, without solvent annealing. The image sizes are all 3 μm × 3 μm.

The ZnOF NPs with fluoroalkyl chains have undergone a gradual evolution from blending with P3HT:PCBM to self-assembling as a surface segregated monolayer (SSM), during which the migration degree of the ZnOF NPs may have a direct relationship with the stacking of P3HT chains. Therefore, adjusting the dynamic self-assembly process of the ZnOF NPs by regulating the film-formation procedure could be an effective approach. During the film formation, the optimized thermodynamic and kinetic parameters, such as polymer state in solution, solvent vapor pressure, and solubility of solutes,^34,35 contribute to the migration of the ZnOF NPs, consequently enhancing the stacking of P3HT and the aggregation of the nanoparticles. Solvent annealing treatment can prolong the film-forming process and provide adequate time for the polymer chains and ZnOF NPs to self-organize,³⁶ which is attributed to the solvent vapor penetrating into the film and increasing the space between the polymer chains. At the same time, the migration of the ZnOF NPs gives P3HT chains extra stress, which is beneficial to forming oriented structures and enhancing the crystallinity.³⁷

In order to explore the relationship between the film-formation process and migration extent of the ZnOF NPs, the self-assembly of P3HT:PCBM blend containing ZnOF NPs (5 wt%), without annealing or with annealing in acetone or o-DCB vapor, were investigated. The corresponding surface coverage of C 1s, F 1s, Zn 2p and O 1s elements has been measured using X-ray photoelectron spectroscopy (XPS), as shown in Fig. 6. The coverage of F 1s has a positive correlation with the content of ZnOF NPs. The highest and the lowest coverage of F 1s have been observed for the blend system upon post-annealing in o-DCB and acetone, respectively. This may result from the difference between the solubility of P3HT in the solvents. Upon post-annealing, o-DCB is a good solvent and assists in the stretching of the P3HT chains, consequently providing plenty of space for the floating movement of the ZnOF NPs to the surface.³⁸ As a consequence, the surface segregated monolayer displays the highest coverage of F 1s. Meanwhile, self-organization of P3HT chains into an ordered structure driven by the unobstructed migration of ZnOF NPs agrees well the strongest intensity and largest red shift of the absorption peak associated with P3HT:PCBM (Fig. 7), which will be described in detail later. However, acetone is a poor solvent and upon post-annealing causes polymer chains to accumulate and form crystalline aggregates, which reduces the channels for ZnOF NPs migrating to the surface, consequently resulting in the lowest surface coverage of F 1s.

Fig. 6 (a) Survey X-ray photoelectron spectra and (b) the coverage of C 1s, F 1s, Zn 2p and O 1s on the surface of P3HT:PCBM:ZnOF (5 wt%) on ITO substrate.

Fig. 7 UV-vis normalized absorbance spectra of P3HT:PCBM blend film: (a) without or with ZnO NPs (5 wt%) and (c) without or with ZnOF NPs (5 wt%); (b) and (d) are the corresponding partial enlargements.

In order to prove this hypothesis further, comparative experiments based on P3HT:PCBM blend mixed with ZnO and ZnOF NPs, as well as the blank system (without adding any nanoparticles), were carried out, and the corresponding UV-vis absorption spectra (in Fig. 7) without annealing or upon annealing in different solvents were measured. Compared to the blank system of P3HT:PCBM blend without solvent annealing, the red shift and raised intensity of the absorption peak related to P3HT:PCBM blend upon annealing in o-DCB is pronounced, which indicates that good solvent annealing optimization improves the local ordering and crystallinity of P3HT. For P3HT:PCBM blend upon annealing in acetone, the absorption peak displays an apparent blue shift and decreased intensity, due to acetone as a poor solvent for post-annealing causing the P3HT chains to accumulate and form crystalline aggregates which is not beneficial for the formation of an ordered structure. Furthermore, the absorption peaks of P3HT:PCBM blend mixed with ZnO NPs (5 wt%) without annealing or annealing in o-DCB and acetone all exhibit a consistent blue shift compared to the pure P3HT:PCBM, owing to the existence of ZnO NP aggregation which may hinder the ordered stacking of P3HT. After adding self-assembling ZnOF NPs with excellent dispersibility, upon the same annealing process there are apparent red shifts in the absorption peaks in comparison to the blank system. This is because the migration of ZnOF NPs arising from the surface segregation of the fluoroalkyl chains gives P3HT chains extra stress to form an oriented structure and enhance the crystallization capacity.³⁷ However, with the incorporation of pristine ZnO NPs, the extra stress does not exist. Therefore, it can be concluded that the migration of ZnOF NPs attributed to the surface segregation of the fluoroalkyl chains has obvious effects on the ordered stacking of P3HT chains.

The morphologies of pure P3HT:PCBM and P3HT:PCBM blend mixed with ZnOF NPs (5 wt%), without annealing or upon annealing in different solvent vapors, were measured by AFM, and the corresponding height images are shown in Fig. 8. For pure P3HT:PCBM, acetone as a poor solvent for post-annealing causes the polymer chains to aggregate, resulting in a more rough morphology with an RMS of 4.02 nm compared to that without annealing (RMS of 2.64 nm). Upon annealing in the good solvent o-DCB, there is a more remarkable phase separation between donor and acceptor, because o-DCB helps the P3HT chains to stack and PCBM to aggregate to some extent. Thus the RMS increases to 4.28 nm. With incorporation of ZnOF NPs (5 wt%) into the P3HT:PCBM blend, the RMS values change to 1.05 nm, 2.77 nm and 3.04 nm without annealing, and with annealing in acetone and o-DCB, respectively. It can be illustrated that the self-assembly of the ZnOF NPs as a surface segregated CBL gives a more smooth morphology than that of pure P3HT:PCBM without any buffer layer. In combination with the former UV-vis results, it can be concluded that the existence of functionalized ZnOF NPs assists in the stacking of donor polymer chains to form a favorable morphology of the active layer with remarkable phase separation, especially upon annealing optimization in o-DCB solvent.

Fig. 8 AFM height images of pure P3HT:PCBM and P3HT:PCBM blend mixed with ZnOF NPs (5 wt%), without annealing or upon annealing in different solvent vapors. The image sizes are all 3 μm × 3 μm.

The self-assembled ZnOF NPs as a CBL applied in polymer solar cells has been carried out, and the pristine and fluoroalkyl modified ZnO NPs spin-coated above the active layer as an additional CBL have also been studied for comparison. The corresponding ultraviolet photoelectron spectroscopy (UPS) has been carried out and the bandgaps of the ZnO and ZnOF NPs have been measured, as shown in Fig. 9. From the calculated energy level data in Table 1, there is little change in the lowest unoccupied molecular orbital (LUMO) levels for the ZnO and ZnOF NPs. The small variation of the LUMO from −4.4 eV to −4.2 eV may not have a direct effect on the electron transport, but it matches well with the energy level of PCBM and the Ag electrode, which is in favor of the electron extraction. The J–V characteristics of devices based on different cathode buffer layers formed by surface segregation of fluoroalkyl modified ZnO or additional spin-coating of pristine and fluoroalkyl modified ZnO NPs above the active layer (in Fig. 9), and the corresponding photovoltaic parameters, including open circuit voltage (V_oc), short-circuit current density (J_sc), fill factor (FF) and power conversion efficiency (PCE) (in Table 2), have been described. For the devices based on the additional CBL of ZnO and ZnOF NPs, the PCEs are 2.0% and 2.2%, respectively. For the pure P3HT:PCBM active layer without annealing, the absence of the CBL results in low J_sc and FF values, and consequently an inferior PCE (only 1.2%). Incorporation of appropriate ZnOF NPs self-assembling as a CBL optimizes the heterojunction morphology of the active layer, which is better for the enhancement of J_sc and FF. Upon annealing in the good solvent o-DCB, the stacking of the donor P3HT chains helps to form the most favorable morphology of the active layer with remarkable phase separation, achieving the best PCE of 2.4% with V_oc of 0.49 V, J_sc of 8.1 mA cm⁻² and FF of 61.1%. It is noticed that the photovoltaic performance of the devices with self-assembled CBLs all display relatively low voltage and high photocurrent, in comparison to the devices with additional CBLs. The low voltage results from the residual ZnOF additive in the active layer (P3HT:PCBM blend), which can change the built-in potential. This conclusion agrees well with the study by Ning et al.³⁹ The conduction band of the ZnO NPs is −4.4 eV and that of the ZnOF NPs is −4.2 eV (Table 1), which indicates that ZnOF should provide a higher V_oc and built in potential than the non-functionalized ZnO NPs. However, the opposite tendency is observed, from which it appears that residual ZnOF in the active layer is acting as a recombination site affecting the transport properties of the system and reducing the V_oc of the device. A comprehensive effect of the built in voltage, work-function and active layer composition on V_oc has been described previously by Garcia-Belmonte.^40,41 The high photocurrent comes from the ordered stacking of donor P3HT chains and the formation of a self-assembled thin and complete CBL. A thick CBL may decrease the photocurrent, which has been reported by Mbule et al.⁴² Therefore, this novel approach could realize the fabrication of both active layer and CBL through a single step, which not only simplifies the fabrication procedure and reduces the manufacturing cost of the polymer solar cells, but also increases the PCE by reduction of the Schottky barrier at the interface.

Fig. 9 (a) Ultraviolet photoelectron spectroscopy (UPS) of ZnO and ZnOF NPs, (b) the bandgaps of ZnO and ZnOF NPs, (c) energy-level diagram of the related materials used for device fabrication and (d) J–V characteristics of the devices: ITO/PEDOT:PSS/P3HT:PCBM:(with or without addition of ZnOF NPs)/Al and ITO/PEDOT:PSS/P3HT:PCBM/additional ZnOF or ZnO layer/Al.

Table 1 Energy levels of ZnO and ZnOF NPs

Samples	Ecutoff^a (eV)	Eonset^b (eV)	HOMO^c (eV)	LUMO^d (eV)
a The high binding-energy cutoff.b The onset relative to the Fermi level (EF) of Au (at 0 eV), where the EF was determined from the Au substrate.c Calculated according to HOMO = hν − (Ecutoff − Eonset), where hν is the incident photon energy, hν = 21.22 eV.d Calculated from the HOMO level and optical band gap obtained from UV-vis absorption spectra.
ZnO NPs	15.56	0.56	−7.6	−4.4
ZnOF NPs	14.58	0.36	−7.0	−4.2

---

Table 2 Photovoltaic performance of polymer solar cells based on different cathode buffer layers formed by surface segregation of fluoroalkyl modified ZnO or additional spin coating above active layer^a

Different cathode buffer layers	Jsc (mA cm^−2)	Voc (V)	FF (%)	PCE (%)
a All values represent averages from fifteen 0.04 cm^2 devices on a single chip. The structure of the devices: ITO/PEDOT:PSS/P3HT:PCBM:(with or without added ZnOF NPs)/Al and ITO/PEDOT:PSS/P3HT:PCBM/additional ZnOF or ZnO layer/Al.
Blank (w/o NPs, w/o annealing)	6.4 ± 0.7	0.46 ± 0.03	38.5 ± 1.9	1.2 ± 0.0
ZnOF (2.5%, w/o annealing)	6.5 ± 0.2	0.46 ± 0.02	42.7 ± 0.7	1.3 ± 0.4
ZnOF (5%, w/o annealing)	7.9 ± 0.0	0.47 ± 0.01	48.0 ± 3.3	1.8 ± 0.1
ZnOF (10%, w/o annealing)	7.1 ± 0.2	0.48 ± 0.01	40.0 ± 3.1	1.3 ± 0.1
ZnOF (5%, acetone annealing)	7.5 ± 0.2	0.49 ± 0.01	43.7 ± 1.1	1.6 ± 0.1
ZnOF (5%, o-DCB annealing)	8.1 ± 0.1	0.49 ± 0.00	61.1 ± 1.4	2.4 ± 0.1
ZnOF (additional layer)	6.8 ± 0.3	0.53 ± 0.01	56.2 ± 2.5	2.0 ± 0.1
ZnO (additional layer)	6.7 ± 0.3	0.55 ± 0.01	59.2 ± 1.9	2.2 ± 0.1

---

Conclusions

The spontaneous self-assembly of ZnOF NPs as a surface-segregated cathode buffer layer driven by the low surface energy of fluoroalkyl chains was present in the polymer solar cells based on P3HT:PCBM as the active layer. The amount of ZnOF NPs added and the degree of migration, regulated by the film-formation process upon annealing in different solvents, produce obvious effects on the stacking of P3HT chains, the heterojunction morphology and the photovoltaic performance. A small quantity of ZnOF NPs could not self-assemble as a continuous and dense layer on the surface, but rather appear only as isolated nanoclusters. Excess ZnOF NPs (up to 10 wt%) might suppress the stacking of P3HT due to the aggregation of residual ZnOF NPs within the active layer. The addition of appropriately functionalized ZnOF NPs assists the stacking of P3HT chains to form a favorable heterojunction morphology of the active layer with remarkable phase separation, which has been confirmed by the obvious red shift and raised intensity of the absorption peak related to P3HT:PCBM, as well as the RMS variation from AFM characterization. Upon post-annealing, o-DCB as a good solvent helps P3HT chains to stretch, consequently providing plenty of space for the segregation movement of ZnOF NPs to the surface. Meanwhile, P3HT chains could self-organize into an ordered structure driven by the unobstructed migration of the ZnOF NPs. However, acetone as a poor solvent for post-annealing causes polymer chains to accumulate and form crystalline aggregates, which reduces the channels for ZnOF NPs migrating to the surface. After optimization of the addition amount of ZnOF NPs and annealing in the good solvent o-DCB, the PCE of the devices based on a P3HT:PCBM active layer is enhanced to 2.4% with an open-circuit voltage (V_oc) of 0.49 V, short-circuit current density (J_sc) of 8.1 mA cm⁻² and fill factor (FF) of 61%, in comparison to the PCE of 1.2% for the device without any CBL.

Acknowledgements

The financial support for this work is provided by the National Natural Science Foundation of China (51273088 and 51302130), National Science Fund for Distinguished Young Scholars (51425304), National Basic Research Program of China (973 Program 2014CB260409), and Doctoral Programs Foundation of Ministry of Education of China (Grants 20133601120006). Zhijuan He and Licheng Tan contributed equally to this work.

References

1. Z. G. Zhang and J. Wang, Structures and properties of conjugated donor–acceptor copolymers for solar cell applications, J. Mater. Chem., 2012, 22, 4178–4187.
2. F. Zhang, M. Ceder and O. Inganäs, Enhancing the photovoltage of polymer solar cells by using a modified cathode, Adv. Mater., 2007, 19, 1835–1838.
3. N. S. Kang, B. K. Ju, T. W. Lee, D. H. Choi, J. M. Hong and J. W. Yu, Organic photovoltaic devices with a crosslinkable polymer interlayer, Sol. Energy Mater. Sol. Cells, 2011, 95, 2831–2836.
4. H. L. Yip, S. K. Hau, N. S. Baek, H. J. Ma and K. Y. Alex, Polymer solar cells that use self-assembled-monolayer-modified ZnO/metals as cathodes, Adv. Mater., 2008, 20, 2376–2382.
5. Y. Zhang, K. Tajima and K. Hashimoto, Nanostructure formation in poly(3-hexylthiophene-block-3-(2-ethylhexyl)thiophene)s, Macromolecules, 2009, 42, 7008–7015.
6. Y. E. Ha, M. Y. Jo, J. Park, Y. C. Kang, S. J. Moon and J. H. Kim, Effect of self-assembled monolayer treated ZnO as an electron transporting layer on the photovoltaic properties of inverted type polymer solar cells, Synth. Met., 2014, 187, 113–117.
7. A. F. Eftaiha, J. P. Sun, I. G. Hill and G. C. Welch, Recent advances of non-fullerene, small molecular acceptors for solution processed bulk heterojunction solar cells, J. Mater. Chem. A, 2014, 2, 1201–1213.
8. G. Li, V. Shrotriya, J. S. Huang, Y. Yao, T. Moriarty, K. Emery and Y. Yang, High-efficiency solution processable polymer photovoltaic cells by self-organization of polymer blends, Nat. Mater., 2005, 4, 864–868.
9. M. T. Dang, L. Hirsch, G. Wantz and J. D. Wuest, Controlling the morphology and performance of bulk heterojunctions in solar cells. Lessons learned from the benchmark poly(3-hexylthiophene):[6,6]-phenyl-C₆₁-butyric acid methyl ester system, Chem. Rev., 2013, 113, 3734–3765.
10. Y. Geng, Q. Wei, K. Hashimoto and K. Tajima, Dipole layer formation by surface segregation of regioregular poly(3-alkylthiophene) with alternating alkyl/semifluoroalkyl side chains, Chem. Mater., 2011, 23, 4257–4263.
11. Y. F. Zhang, R. Y. Liu, S. T. Lee and B. Q. Sun, The role of a LiF layer on the performance of poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate)/Si organic–inorganic hybrid solar cells, Appl. Phys. Lett., 2014, 8, 083514.
12. M. F. Xu, Y. J. Liao, F. S. Zu, J. Liang, D. X. Yuan, Z. K. Wang and L. S. Liao, Work-function tuneable and aqueous solution-processed Cs₂CO₃ for high-performance polymer solar cells, J. Mater. Chem. A, 2014, 2, 9400–9404.
13. Y. C. Yeh, S. S. Li, C. C. Wu, T. W. Shao, P. C. Kuo and C. W. Chen, Stoichiometric dependence of TiO_x as a cathode modifier on band alignment of polymer solar cells, Sol. Energy Mater. Sol. Cells, 2014, 125, 233–238.
14. P. S. Mbule, T. H. Kim, B. S. Kim, H. C. Swart and O. M. Ntwaeaborwa, Effects of particle morphology of ZnO buffer layer on the performance of organic solar cell devices, Sol. Energy Mater. Sol. Cells, 2013, 112, 6–12.
15. D. Kim, J. Yoon and H. Kim, Selective area atomic layer deposited ZnO nanodot on self-assembled monolayer pattern using a diblock copolymer nano-template, J. Nanosci. Nanotechnol., 2012, 12, 1483–1486.
16. M. G. Helander, Z. B. Wang, L. Mordoukhovski and Z. H. Lu, Comparison of Alq3/alkali-metal fluoride/Al cathodes for organic electroluminescent devices, J. Appl. Phys., 2008, 104, 094510.
17. C. Y. Neo and J. Y. Ouyang, LiF doped mesoporous TiO₂ as the photoanode of highly efficient dye-sensitized solar cells, J. Power Sources, 2013, 241, 647–653.
18. Y. F. Zhang, R. Y. Liu, S. T. Lee and B. Q. Sun, The role of a LiF layer on the performance of poly(3,4-ethylenedioxythiophene):poly(styrenesulfonate)/Si organic–inorganic hybrid solar cells, Appl. Phys. Lett., 2014, 104, 083514.
19. X. C. Bao, L. Sun, W. F. Shen, C. P. Yang, W. C. Chen, R. Q. Yang and P. M. Amassian, Facile preparation of TiO_X film as an interface material for efficient inverted polymer solar cells, J. Mater. Chem. A, 2014, 2, 1732–1737.
20. S. W. Lee, Stoichiometric dependence of TiO_x as a cathode modifier on band alignment of polymer solar cells, Sol. Energy Mater. Sol. Cells, 2014, 125, 233–238.
21. F. C. Chen and S. C. Chien, Nanoscale functional interlayers formed through spontaneous vertical phase separation in polymer photovoltaic devices, J. Mater. Chem., 2009, 19, 6865–6869.
22. M. Fenoll, J. Shim, C. Fuentes-Hernandez, K. H. Sandhage and B. Kippelen, Roles of thermally-induced vertical phase segregation and crystallization on the photovoltaic performance of bulk heterojunction inverted polymer solar cells, Energy Environ. Sci., 2011, 4, 3456–3460.
23. M. D. Clark, M. L. Jespersen, R. J. Patel and B. J. Leever, Predicting vertical phase segregation in polymer–fullerene bulk heterojunction solar cells by free energy analysis, ACS Appl. Mater. Interfaces, 2013, 5, 4799–4807.
24. K. Yao, L. Chen, X. Chen and Y. W. Chen, Self-organized hole transport layers based on polythiophene diblock copolymers for inverted organic solar cells with high efficiency, Chem. Mater., 2013, 25, 897–904.
25. Q. S. Wei, T. Nishizawa, K. Tajima and K. Hashimoto, Self-organized buffer layers in organic solar cells, Adv. Mater., 2008, 20, 2211–2216.
26. J. E. Beek, M. M. Wienk, M. Kemerink, X. N. Yang and R. A. Janssen, Hybrid zinc oxide conjugated polymer bulk heterojunction solar cells, J. Phys. Chem. B, 2005, 109, 9505–9516.
27. S. Gartner, M. Christmann, S. Sankaran, H. Rohm, E. M. Prinz, F. Penth, A. Putz, A. E. Tureli and B. Penth, Eco-friendly fabrication of 4% efficient organic solar cells from surfactant-free P3HT: ICBA nanoparticle dispersions, Adv. Mater., 2014, 26, 6653–6657.
28. Y. Lin, Q. Wei, G. Qian, L. Yao and J. J. Watkins, Morphology control in TiO₂ nanorod/polythiophene composites for bulk heterojunction solar cells using hydrogen bonding, Macromolecules, 2012, 45, 8665–8673.
29. D. Chirvase, J. Parisi, J. C. Hummelen and V. Dyakonov, Influence of nanomorphology on the photovoltaic action of polymer–fullerene composites, Nanotechnology, 2004, 15, 1317–1323.
30. C. V. Hoven, X. D. Dang, R. C. Coffin, J. Peet, T. Q. Nguyen and G. C. Bazan, Improved performance of polymer bulk heterojunction solar cells through the reduction of phase separation via solvent additives, Adv. Mater., 2010, 22, E63–E66.
31. W. L. Wang, H. B. Wu, C. Y. Yang, C. Luo, Y. Zhang, J. W. Chen and Y. Cao, High-efficiency polymer photovoltaic devices from regioregular-poly(3-hexylthiophene-2,5-diyl) and [6,6]-phenyl-C-61-butyric acid methyl ester processed with oleic acid surfactant, Appl. Phys. Lett., 2007, 90, 183512.
32. W. Ma, X. Wang and J. Zhang, Effect of MMT, SiO₂, CaCO₃, and PTFE nanoparticles on the morphology and crystallization of poly(vinylidene fluoride), J. Polym. Sci., Part B: Polym. Phys., 2010, 48, 2154–2164.
33. B. Lim, J. Jo, S. I. Na, J. Kim, S. S. Kim and D. Y. Kim, A morphology controller for high-efficiency bulk-heterojunction polymer solar cells, J. Mater. Chem., 2010, 20, 10919–10923.
34. Y. Yao, J. H. Hou, Z. Xu, G. Li and Y. Yang, Effect of solvent mixture on the nanoscale phase separation in polymer solar cells, Adv. Funct. Mater., 2008, 18, 1783–1789.
35. X. Yang and J. Loos, Towardhigh-performance polymer solar cells: the importance of morphology control, Macromolecules, 2007, 40, 1353–1362.
36. C. Sinturel, M. Vayer, M. Morris and M. A. Hillmyer, Solvent vapor annealing of block polymer thin films, Macromolecules, 2013, 46, 5399–5415.
37. W. R. Hwang, G. W. M. Peters, M. A. Hulsen and H. E. H. Meijer, Modeling of flow-induced crystallization of particle-filled polymers, Macromolecules, 2006, 39, 8389–8398.
38. L. L. Chang, H. W. A. Lademann, J. B. Bonekamp, K. Meerholz and A. J. Moulé, Effect of trace solvent on the morphology of P3HT:PC₆₁BM bulk heterojunction solar cells, Adv. Funct. Mater., 2011, 21, 1779–1787.
39. Z. J. Ning, D. Zhitomirsky, V. Adinolfi, B. Sutherland, J. Xu, O. Voznyy, P. Maraghechi, X. Lan, S. Hoogland, Y. Ren and E. H. Sargent, Graded doping for enhanced colloidal quantum dot photovoltaics, Adv. Mater., 2013, 25, 1719–1723.
40. A. Guerrero, L. F. Marchesi, P. P. Boix, S. Ruiz-Raga, T. Ripolles-Sanchis, G. Garcia-Belmonte and J. Bisquert, How the charge-neutrality level of interface states controls energy level alignment in cathode contacts of organic bulk-heterojunction solar cells, ACS Nano, 2012, 6, 3453–3460.
41. A. Guerrero, N. F. Montcada, J. Ajuria, I. Etxebarria, R. Pacios, G. Garcia-Belmonte and E. Palomares, Charge carrier transport and contact selectivity limit the operation of PTB7-based organic solar cells of varying active layer thickness, J. Mater. Chem. A, 2013, 1, 12345–12354.
42. P. S. Mbule, T. H. Kim, B. S. Kim, H. C. Swart and O. M. Ntwaeaborwa, Effects of particle morphology of ZnO buffer layer on the performance of organic solar cell devices, Sol. Energy Mater. Sol. Cells, 2013, 112, 6–12.

---