Hybrid inverted organic photovoltaic cells using electrospun zinc oxide nanofibers as electron transport

80

XRD measurement. Characteristic peaks for ZnO are rather weak and obscure, which indicates only few crystalline ZnO structures form under this calcination condition. After further calcination at 500 for 2 hours, five diffraction peaks at 31.76º, 34.34º, 36.20º, 56.50º and 62.84º are observed, corresponding to (100), (002), (101), (110), and (103) of wurtzite crystal structure, respectively. All of these diffraction peaks are consistent with the reported data for ZnO of a wurtzite hexagonal phase. No characteristic peaks for other impurities, except for the substrate, were found, which means the phase of the fibers obtained after calcination at 500 for 2 hours are rather pure. All of these demonstrations imply that calcination condition played an important role in removing the PVP component form the composite fibers and improving the crystallinity of ZnO nanofibers.

Figure 5.14 Thermogravimetric analysis curves of pure (a) PVP and (b) PVP-ZnO fibers.

5.2.6 Conclusions

I have demonstrated the controllability of the diameter of electrospun ZnO-PVP composite nanofibers in the range from micrometer down to several tens of nanometer. Two key factors, molar concentration of zinc acetate and PVP concentration in the precursor solution, were found to determine the morphology and diameter of the electrospun fibers. The work that applies the ultrathin ZnO nanofibers on the solar cells to increase the surface area is currently underway in our laboratory. I believe that the control method described here could extend the application of ZnO nanofibers to more geometry-dependent devices.

5.3 Hybrid inverted organic photovoltaic cells using electrospun zinc oxide nanofibers as electron

81

and inexpensive to fabricate, flexible, and customizable on the molecular level, and they have lower potential for negative environmental impact. The poly(3-hexylthiophene) (P3HT) and (6,6)-phenyl C61 butyric acid methyl ester (PCBM) blend is one of the most promising materials for PSC based on composite blends of a π-conjugated polymer. Excitons in the blend film are dissociated through the photoinduced charge transfer from the electron donor in P3HT to the acceptor in PCBM [146].

However, it is difficult to achieve the ideal device architecture by simply blending the two materials together because the constituents determine the domain size during the phase separation process. Moreover, operational stability of the devices is poor under ambient conditions for practical applications because of many complex degradation processes in the photoactive as well as other functional layers. To deal with this issue, many groups use blends of oxide semiconductor nanoparticles embedded in a conjugated polymer matrix for a hybrid polymer/inorganic device [147]. They found that photoinduced charge transfer can also occur between a conjugated polymer and a metal oxide semiconductor such as ZnO, TiO2, and SnO2 [148-150]. These devices presently suffer from incomplete exciton dissociation due to imperfect dispersion of the nanoparticles in the blend and low carrier mobility through the nanoparticle pathways in the blend.

Recent studies show that electron conducting nanowires on electron transport layer (ETL) with their length reaching to most of the photoactive layer (P3HT:PCBM) thickness improves charge collection efficiency [151, 152]. Some researchers plant ZnO nanofibers in the ETL of PSC to improve charge collection efficiency and reduce recombination rate [153]. In this work, network of ZnO nanofibers were electrospun directly onto ETL and infiltrated by subsequently spin coated P3HT/PCBM blend films to increase the interface area between ETL and active layer. A wide variety of surface treatments for the ZnO nanofibers were employed to remove the residual organic component in the nanofibers.

5.3.2 Experimental details

A zinc acetate precursor gel using for spin coating ZnO thin layer was prepared using zinc acetate (0.15 g), 2-methoxyethanol (3.0 ml), ethanolamine (0.05 ml), and isopropanol (1.5 ml). And three kinds of poly(vinylpyrrolidone) (PVP) (0.3 g), with molecular weight of 40,000, 360,000, and 1,300,000, were added into the precursor to fabricate ZnO nanofibers. Poly(3-hexylthiophene) (P3HT) and [6,6]-phenyl C61 butyric acid methyl ester (PCBM) solution A home-built electrospinning system was used to spray the gel through a glass capillary with a 100 μm inner diameter at the blunt tip. A stable high voltage was generated by a regulated power supply (HJPQ-30P1, Matsusada Precision) and applied to the gel through a copper wire in a glass capillary.

82

Figure 5.15 Architecture of the devices in this study

The OPV devices in this study were shown in Figure 5.15. Patterned indium tin oxide (ITO)-coated glass substrates were pre-cleaned for 10 min consequently in deionized water, acetone, and isopropanol by ultrasonic agitation, and treated in an ultraviolet cleaner for 20 min. Firstly, a nucleation layer of ZnO was spin coated onto the ITO at 1000 rpm for 1 min and thermal annealed at 300 °C for 15 min in air. Then ZnO nanofibers were electrospun onto the thin layer of ZnO at 6 kV. After electrospinning, the samples were annealed at 300 °C for 30 min in air. Poly(3-hexylthiophene) (P3HT) and [6,6]-phenyl C61 butyric acid methyl ester (PCBM) blend films were spun-coated onto the network of ZnO nanofibers from a 1:1wt-ratio solution in 1,2-chlorobenzene (20 mg of P3HT /ml of solvent) at 600 rpm for 1 min. Finally, a MoO3 layer of 10 nm and a gold electrode of 80 nm were evaporated onto the P3HT:PCBM layer by vacuum deposition.

Post-annealing was performed at 140 °C for 20 min in glove box. The current density-voltage characteristics of the OPVs were measured using Keithley series 2400 digital source under an AM 1.5G illumination and corrected to 100 mW/cm². The top graphic images of sample were obtained using AFM (Seiko, SPA300HV).

Absorption spectra were acquired using a high resolution ultraviolet-visible (UV–Vis) spectrophotometer (V-650, JASCO Corporation). Atomic concentration was measured with an X-ray photoelectron spectroscopy (AXIS Nova, Shimadzu).

5.3.3 Application of ZnO network of various morphology to inverted solar cells

As described in Section 5.1, the morphology of ZnO network can be controlled by electrospinning from the precursor solutions containing PVP of different molecular weight. To get optimized ZnO network for solar cell applications, I fabricated a variety of ZnO networks by varying molecular weight of PVP and electrospinning time separately. The diameter of resultant fibers was controlled by selecting the molecular weight of PVP, and the density of the ZnO network was modified by changing the electrospinning time from 1 to 7 min.

83

Figure 5.16 Current density-voltage curves of the inverted solar cells containing ZnO network of different morphology. (a), (b), and (c) correspond to the PVP40000, PVP360000, and PVP1300000 devices,

respectively.

Current density-voltage characteristic of most inverted solar cells in Figure 5.16 seems similar to each other. The power conversion efficiency of the solar cells with ZnO nanostructures is as many as that of spin coated reference sample. Some of these solar cells even show lower efficiency that the reference sample. It

84

is worth note in the PVP36 group that the longer electrospinning time, the lower device performance obtained. This suggests that the effectiveness of the ZnO network in these inverted solar cells is rather weak for some common reasons.

I checked the morphology of the ZnO network with AFM, see Figure 5.17. It is clearly shown in the topography that ZnO nanofibers of hundreds of nanometer in diameter accumulate on the substrate and the surface of the substrate is not fully covered by the ZnO nanofibers. We also noticed that the surface of the ZnO nanofibers is rather smooth. The ZnO grains cannot be identified from the fibers. So, one possible reason for the low performance is insufficient crystallization of ZnO in the fibers. I investigated calcination condition of ZnO nanofibers for increasing the crystallinity of ZnO in the fibers in below.

Figure 5.17 AFM images of ZnO network fabricated with PVP36 for 3 min. From left to right are topography, error signal, and 3D image, respectively.

5.3.4 The effect of calcination condition on the morphology of ZnO nanofibers

As shown in the upper images of Figure 5.18, ZnO grains can be identified from the fibers, which is more distinct in the fibers electrospun on the silicon substrate. However, I also find that the ZnO grains are still embedded in the fibers. It indicates that the PVP component in the fibers inhibits the crystallization of ZnO during calcination. Comparing with the samples calcined at 300 ºC for 10 min, the samples calcined at 500 ºC for 2 hours show higher contrast, see lower images of Figure 5.18. The fibers are comprised of ZnO grains of about 50 nm in diameter. I also noticed that the beads in the ZnO network disappear after calcined at 500 ºC for 2 hours.

85

Figure 5.18 Topography of ZnO nanofibers on ITO coated substrate after calcination. Upper and lower samples are calcined at 300 ºC for 10 min and at 500 ºC for 2 hours, respectively.

The transmittance spectra of the samples are shown in Figure 5.19. The transmittance of the ZnO nanofibers calcined at 300 ºC for 10 min is about 95% from 400 to 800 nm, which can be attributed to the absorption of residual PVP component in the ZnO nanofibers. The transmittance of the ZnO nanofibers at this scale increase to almost 100% after being calcined at 500 ºC for 2 hours, which indicates that most of the PVP component has been removed during the calcination. It is worth note that an absorption band around 370 nm, corresponding to the energy gap of 3.2 eV for ZnO, is observed, which confirms that crystalline ZnO formed after calcined at 500 ºC for 2 hours.

86

Figure 5.19 Transmittance spectra of the ZnO nanofibers after calcination at 300 ºC for 10 min (black curve) and at 500 ºC for 2 hours (red curve).

5.3.5 UV treatment for ZnO-PVP nanocomposite fibers to remove residual PVP content

Previous work has shown that UV-ozone treatment can remove PVP on colloidal nanoparticle film surfaces [154]. The removal of PVP did not alter the size, shape or distribution of the nanoclusters in the films [154, 155]. I treated the ZnO-PVP nanocomposite fibers in with UV-ozone for different time to remove PVP in the fibers.

The atomic concentrations of zinc, carbon, and oxygen in the UV-ozone treated ZnO–PVP nanocomposite fibers based on the Zn 2p, C 1s, and O 1s XPS spectra are summarized in Fig. 5.20. The atomic concentration of carbon from the PVP in the nanocomposite fibers is significantly reduced by UV-ozone treatment (from 29.6% to 22.9%). Conversely, the atomic concentrations of zinc and oxygen present in the nanocomposite fibers both increase from 25.3% and 45.1% to 28.3% and 48.8%, respectively, after UV-ozone treatment for 7.5 min. The relatively smaller increase in zinc atomic concentration compared to oxygen is due to the competition between the increases in oxygen content arising from UV-ozone treatment and the decrease in oxygen content coming from the removal of PVP. These results strongly support that UV-ozone treatment removes PVP content from the surface of the ZnO–PVP nanocomposite fibers.

The UV-ozone treated ZnO-PVP nanocomposite fibers were used as ETLs in inverted solar cells. Figure 5.21 shows the current density-voltage curves of the inverted solar cells. The performance of the inverted solar cell decrease after UV-ozone treatment for 5 min compared with the one treated for 2.5 min. While the performance dramatically increase after UV-ozone treatment for 7.5 min and exceed the performance of the solar cell treated for 2.5 min. All 3 solar cells have almost same Voc, but their Jsc vary significantly. This suggests that UV-ozone treatment for the ZnO-PVP nanocomposite fibers play a role in determining the solar cell performance. However, the power conversion efficiency of the UV-ozone treated solar cells is

87

lower than the reference sample without ZnO-PVP nanocomposite fibers. It indicates that some other reasons, instead of PVP content, limited the device performance.

Figure 5.20 Atomic concentrations of zinc, carbon and oxygen in ZnO-PVP nanocomposite fibers after UV-ozone treatment based on their XPS spectra.

Figure 5.21 Current density-voltage curves of the inverted solar cells with UV-ozone treated ZnO-PVP nanocomposite fibers as ETLs for various treatment time (2.5, 5, and 7.5 min) under 1.5G solar illumination

at 100 mW/cm².

88

5.3.6 Chemical surface treatment for ZnO-PVP nanocomposite fibers to remove residual PVP content Except UV-ozone treatment, I investigated chemical surface treatment for ZnO-PVP nanocomposite fibers to remove residual PVP content in the fibers. Three kinds of chemical surface treatments, immersing the ZnO-PVP nanocomposite fibers in ethanol solution, ultrasonic cleaning in ethanol for 10 min, and ultrasonic cleaning in acetone and IPA for 5 min respectively, were studied.

The atomic concentrations of zinc, carbon, and oxygen in the chemical surface treated ZnO–PVP nanocomposite fibers based on the Zn 2p, C 1s, and O 1s XPS spectra are summarized in Fig. 5.22. The atomic concentration of carbon from the PVP in the nanocomposite fibers is significantly increased from 25.8% to 51.1% by surface treatment, which can be attributed to the residual organic solvent. While, the atomic concentrations of zinc and oxygen present in the nanocomposite fibers increase obviously after ultrasonic cleaning, especially for the sample cleaned in ethanol for 10 min. These results indicate that these chemical surface treatments are effective in removal of the PVP content from the ZnO-PVP nanocomposite fibers.

Figure 5.22 Atomic concentrations of zinc, carbon and oxygen in ZnO-PVP nanocomposite fibers after chemical surface treatment based on their XPS spectra.

Inverted solar cells with ZnO-PVP nanocomposite fibers treated by various chemical surface treatments were fabricated, and their performance are shown in Figure 5.23. Comparing with the solar cell using ZnO thin layer as ETL, the performance of device cleaned in acetone and IPA for 5 min respectively is improved, which can be attributed to the removal of the PVP content from the ZnO-PVP nanocomposite fibers. As for

89

the solar cells treated in ethanol, however, negative effect on device performance is found. It can be explained that ZnO grains together with the PVP content are removed during the treatment.

Figure 5.23 Current density-voltage curves of the inverted solar cells with chemical surface treated ZnO-PVP nanocomposite fibers as ETLs under initial AM 1.5G solar illumination at 100 mW/cm².

The performance of the devices shown in Figure 5.23 are summarized in Table 5.1. When comparing device C with device D, PCE is improved. The increased interface area between the ITO and the active layer, which was reflected by the reductive series resistance and incremental shunt resistance, contributed to the improved performance of device C.

Table 5.1 Average devices performance of the inverted solar cells after various surface treatments for ETLs.

5.3.7 Conclusions

I investigated the application of ZnO-PVP nanocomposite fibers to inverted solar cells as ETL. I found that calcination at 500 ºC for 2 hour facilitates the formation of crystalline ZnO and the PVP content were removed during the calcination. UV-ozone treatment is effective for further removal of the PVP content in the ZnO-PVP nanocomposite fibers, however, it didn‘t contribute to the improvement of solar cell

90

performance directly. Chemical surface treatment, especially ultrasonic cleaning in acetone and IPA, can remove the PVP content in the ZnO-PVP nanocomposite fibers effectively. The inverted solar cell with ZnO-PVP nanocomposite fibers cleaned in PVP content in the ZnO-PVP nanocomposite fibers shows higher performance than the reference device with 20 nm ZnO thin layer as ETL, which is due to the increase of interface area between the ITO and the active layer.

91