Chapter 3 Experimental Section
3.2 Material fabrication
3.2.5 F-doped SnO 2 deposition on ZnO nanorods
To deposit the FTO films on ZnO nanorods by ultrasonic spray pyrolysis, the initial solution was prepared from 0.5 moles of stannous chloride in 1.0 L of deionized water.
In order to promote solubility, 5% HCl was added into the stannous chloride precursor solution. Upon stirring, the solution immediately became transparent, indicating its solubilization. Finally, a 50% ammonium fluoride precursor was mixed at room temperature and stirred for 5 min. The deposition temperature was set at 400 °C for all of the depositions, and the deposition time was 0.5-3 min. The carrier gas flow rate was maintained at 20 L/min in air.
3.2.6 Fabrication of solar cells
Bulk heterojunction: The indium tin oxide (ITO) coated glass was purchased from Merck and resistivity is 10 ohms/sq. The ITO-coated glass was pre-cleaned (DI water,
34
acetone, ethanol and isopropyl alcohol) and treated with oxygen plasma prior to use.
PEDOT:PSS layer (50 nm) was spin-coated at 2000 rpm and annealed at 120 ºC for 30 min. The CdS/P3HT layers (200 nm) were spin-coated from their corresponding dichlorobenzene solutions (30 mg/mL) at 1500 rpm and annealed at 160 °C for 60 min, followed by thermal evaporation of an aluminum electrode. Al top electrodes (100 nm) were deposited by thermal evaporation through a shadow mask, resulting in individual devices with 0.1 cm2nominal area.
Ordered heterojunction: The indium tin oxide (ITO) coated glass was purchased from Merck and resistivity is 10 ohms/sq. The ITO-coated glass was pre-cleaned (DI water, acetone, ethanol and isopropyl alcohol) and treated with oxygen plasma prior to use.
The ZnO nanorods were grown on the substrate, as mentioned above (Chapter 3.2.1).
The CdS/P3HT layers (200 nm) were spin-coated from their corresponding dichlorobenzene solutions (30 mg/mL) at 1500 rpm and annealed at 160 °C for 60 min. The PEDOT:PSS layer (50 nm) was spin-coated at 2000 rpm, followed by thermal evaporation of an Au electrode.
35
Chapter 4
Fabrication and characterization of MgO-doped ZnO nanorod arrays
4.1 Introduction
Research on ZnO has generated great interests for its potential applications in photonics, especially on short wavelength light-emitting, UV lasing and transparent conducting materials due to its wide direct gap of 3.37 eV, large exciton binding energy of 60 meV at room temperature and its promising versatile applications [122].
Recently, doped ZnO are of technological importance because of their great potential for applications, such as transparent conducting electrodes [123].
Alloying ZnO with MgO makes ZnMgO a potential candidate for future optoelectronic devices because of its wide band gap, less lattice mismatch with ZnO as the ionic radius of Mg2+ and Zn2+ are similar [124, 125]. The band gap of the MgZnO alloys could be expanded from 3.3 to 4.20 eV and the microstructure and PL property of MgO–ZnO mixture have also been studied by several groups [126, 127].
In these works, this alloy produced a stronger UV luminescence at room temperature that was found to be excitonic in nature. These recent developments open up enormous interest in this materials system.
Recently, single-crystal ZnO nanorods have drawn considerable interests as building blocks or basic units for the fabrication of nanosized devices due to their wide band gap and a large exciton binding energy with high crystalline quality. In
36
order to realize ZnMgO-based electronic and optoelectronic devices [128, 129], band gap engineering of the nanorods would be prerequisite and many researchers have reported the technologies by incorporating MgO into ZnO. However, few studies have been investigated into ZnO-MgO core-shell nanorods grown by coating MgO nanoparticles on the surface of ZnO nanorods using wet-chemical solution process.
Therefore, in this work, a simple method was used to fabricate MgO-coated ZnO nanorods and a rapid post-annealing was performed to investigate the effect of thermal treatment on the surface microstructure and luminescent properties of MgO-coated ZnO heterostructured nanorods.
4.2 Surface morphologies after annealing treatments
Figure 4.1 shows the scanning electron microscopy (SEM) images taken from several samples with highly uniform and densely packed arrays of ZnO nanorods and MgO-coated ZnO nanorods grown on Si substrate. Figure 4.1(a) and 4.1(b) illustrate as-grown ZnO nanorods and MgO-coated ZnO nanorods. It can be observed that MgO nanoparticles have been successfully deposited on the surface of the ZnO nanorods.
After annealed at 900 °C in O2 atmosphere, Figure 4.1(c) shows the MgO-coated ZnO nanorods still kept hexagonal shape and the morphology of ZnO nanorods remained almost unchanged as compared to that of the as-grown ZnO nanorods. However, it was noted that the surface of the MgO-coated ZnO nanorods annealed at 900 ºC in O2
atmosphere was flattened and became rather smooth. In contrast, as annealed at 800
°C in H2/N2, it was observed in Figure 4.1(d) that some of the MgO-coated ZnO nanorods are slightly collapsed, probably because H2/N2 reduction atmosphere caused surface etching [130].
37
Figure 4.1 SEM images of (a) ZnO nanorods, (b) as-made MgO-coated ZnO nanorods, and annealed MgO-coated ZnO nanorods at (c) 900 °C in O2, and (d) 800
°C in H2/N2.
Figure 4.2(a) shows the X-ray diffraction (XRD) patterns of as-made and annealed MgO-coated ZnO nanorods. A weak (200) peak characteristic of the cubic MgO phase was detected for the as-made sample as confirmed to the XRD pattern of the MgO precursor in the inset of Figure 4.2(a) which is obtained at 350 ºC. In contrast, as the MgO-coated ZnO nanorods were annealed at 900ºC in O2 or 800 °C in N2-H2
atmosheres, no MgO phase can be detected from the XRD patterns, indicating an interaction occurred in between ZnO and MgO. It is known that ZnO and MgO react easily to form MgZnO alloys at annealing temperatures higher than 700 °C [131]. In addition, it was also found that the ZnO (0 0 2) diffraction peak of the as-made and the annealed samples (in O2) is located at 34.46° and 34.52°, respectively. It is thought that Zn2+ ions in the ZnO lattice were replaced partly by Mg2+ ions with
300 nm
(b) (a)
300 nm
300 nm
(d)
300 nm
(c)
38
smaller radium, resulting in the decrease of the lattice constant along the c-plane [132].
According to the phase diagram [133], the thermodynamic solid solubility limit of MgO in ZnO was less than 4 mol%. Within the range of MgO 4 mol%, the MgZnO had a hexagonal phase similar to that of ZnO. Therefore, it can be deduced that MgZnO was probably formed.
Figure 4.2 XRD patterns of the MgO-coated ZnO nanorods annealed in various atmospheres. The inset is the XRD pattern of MgO precursor annealed at 350 °C.
It was postulated that during the process of hydrolysis, OH- could react with Mg2+, leading to MgO nanoparticles adhering better on the surface of ZnO nanorods [134], as shown in Figure 4.3(a). After thermal treatment at 800 ºC in O2 atmosphere, hetrostructured core (ZnO)-shell (Mg-compound) nanorods were observed as evidenced from transmission electron microscopy (TEM) image in Figure 4.3(b). The diameter of the core (ZnO nanorods) and the shell (MgOx) thin layer is about 70 nm and 7 nm, respectively. High resolution transmission electron microscopy (HRTEM) image of the MgO-coated ZnO nanorods in Figure 4.3(c) shows the (002) lattice fringe of the ZnO nanorod core and an amorphous layer can be identified in shell-MgOx thin layer, implying the diffusion of Mg into the ZnO nanorods during MgZnO alloy formation. Furthermore, the HRTEM image in Figure 4.3(d) obtained
39
from 900 °C annealing sample demonstrated that no significant amorphous MgOx was observable and the lattice fringes do notshow obvious distortions due to the small difference between the ionic radius of Mg2+ and Zn2+.
Figure 4.3 TEM image of (a) as-made MgO nanoparticles on the surface of ZnO nanorods. (b) Low-magnification TEM and (c) HRTEM of annealed MgO-coated ZnO nanorod at 800 °C in O2 atmosphere and (d) HRTEM image of 900 °C -annealed MgZnO alloy nanorod
4.3 Annealing effect on photoluminescence
Figure 4.4 illustrates the room-temperature photoluminescence (PL) properties of ZnO and MgO-doped ZnO nanorods. The ultraviolet (UV) emission peak of ZnO is generally attributed to the exciton-related activity, and the deep level emission generally results from structural defects, single ionized vacancies, and impurities
(c)
MgOX
ZnO
MgO Nanoparticles
ZnO
(a)
20 nm
(b)
10 nm MgOx
ZnO
MgZnO alloy
(d)
2 nm
40
[135].When the MgO-coated ZnO nanorods were annealed above 800 °C in O2
atmosphere, a blue shift of the near-band-edge emission was observed in Figure 4.4(a) due to the formation of MgZnO alloy. These results are in reasonable agreement with the thermal diffusion of Mg atoms at the MgO/ZnO heterointerface with the formation of MgZnO alloys [136].
The UV emission peaks at 371 and 367 nm for the samples annealed at 800 °C and 900 °C, respectively, are attributed to the increased substitution ratio of Mg on Zn sites and the band gap energy. Furthermore, an enhancement of the near band edge emission is achieved by annealing MgO–ZnO mixture in oxygen due to the reduction of structure defects.6 In addition, it was also found that the blue shift of the near-band-edge emission for the MgO-coated ZnO nanorods annealed in H2/N2
atmosphere above 700°C, as shown in Figure 4.4(b). Moreover, with an increase of the temperature of annealing up to 800 °C, a blue band peaked at 454 nm was observed. It indicates that MgO diffusion process may produce a new luminescent center to emit the blue emission in H2/N2 reduction atmosphere. However, the explanation on the blue emission produced by annealing the sample at 800 °C in H2/N2 is still not fully understood at present.
Figure 4.4 Room-temperature PL spectra of ZnO nanorods annealed at various temperatures in (a) O2, (b) H2/N2 atmospheres.
(a)
O
2350 400 450 500 550 600 650 As-made
350 400 450 500 550 600 650 Wavelength (nm)
41
4.4 Summary
The microstructure and photoluminescent properties of MgO-coated ZnO nanorods prepared by solution techniques are investigated. The samples annealed at 700– 900
°C show well crystalline wurtzite structure of the ZnO nanorods. Annealing at high temperatures (>700 °C) in O2 atmospheres leads to Mg diffusion in ZnO and MgZnO alloy formation. The room-temperature UV emission of the ZnO nanorods enhances with the increase of annealing temperature due to the reduction of structure defects.
Furthermore, a blue shift in the near-band-edge emission was observed as a result of the alloy band gap widening. Moreover, it is noted that the visible emission from the MgO-doped ZnO nanorods annealed at 800 °C in H2/N2 is centered at about 454 nm.
In contrast, the visible emission from the MgZnO nanorods annealed O2 is at 595 nm.
42
Chapter 5
Improvement of charge injection in nanostructured ZnO/P3HT hybrid solar cells
5.1 Introduction
Recently, organic-inorganic hybrid solar cells have attracted a great deal of interest recently due to their potential application in developing low-cost, large-area, mechanically flexible photovoltaic devices [50]. In organic-inorganic hybrid solar cells with planar junctions, the power conversion efficiency (PCE) is limited because the exciton diffusion length of the donor material is typically significantly shorter than its absorption length, resulting in recombining easily. In order to overcome this problem, a bulk heterojunction structure has been developed for organic-inorganic hybrid solar cells. Inorganic nanocrystalline materials can be used as alternative electron acceptors TiO2 and ZnO nanocrystals [87, 102], or photo-absorption such as CdSe quantum dots [87, 137], which may be dispersed in solution together with the organic donor to form QDs-decorated ZnO-based hybrid solar cell. Several theoretical studies conclude that the ideal polymer-inorganic device topology is a perfect vertical array of single-crystal nanorods of the appropriate dimensions and pitch, encased in a film of the polymer [138].A key advantage of using a vertical array of single-crystal nanorods as electron acceptors is that it provides direct channels for electron and hole transport to the electrodes. Besides, Electron transport in crystalline wires is expected to be several orders of magnitude faster than percolation through a random
43
polycrystalline network. In addition, QDs can provide the ability to match the solar spectrum better by tuning their particle size.
Organic-inorganic hybrid solar cells convert sunlight to electrical power by splitting photogenerated excitons across an interface between an electron donor and an electron acceptor material [139].Hence,the study on the interface engineering is essential toward improving excitonic solar cells performance. Recently there are some reports on improvement in the photocurrent by the chemical modification with ruthenium dyes [140],carboxylic, and phosphoric acid groups [120],which have the potential to improve polymer wetting and charge transfer dynamics. Generally, interface modifiers are known to passivate inorganic surface states by chemically interacting with surface dangling bonds [141]. However, the mechanism for the improvement by the chemical modification is poorly understood. Therefore, in this study, the ZnO nanorods were directly decorated with CdS quantum dots and the surface energetic was changed by using different-sized quantum dots (QDs) to investigate the effect of energy level of QDs on the charge injection efficiency. In this study, ZnO nanorods were directly decorated with CdS quantum dots and the energy level by controlling the size of CdS quantum dots. It is demonstrated that the photovoltaic efficiency of the ZnO nanorods/P3HT solar cells can be improved 4-fold by decorating the nanorod arrays with CdS QDs.
5.2 Optical properties of CdS QDs
Figure 5.1 shows the absorption spectra of the three different-sized CdS QDs employed in the present study. These particles exhibit absorption in the UV or visible with an onset corresponding to particle size. The shift of the onset absorption to lower wavelengths with decreasing particle size represents size quantization effects in these particles. By comparing the excitonic transition (361, 404, 435 nm) to the absorption
44
curve reported by Yu and co-workers [121], the particle mean diameter of these samples was estimated as 2.2, 3.2 and 4.6 nm respectively. These different-sized QDs would be then deposited on ZnO nanorods for spectroscopic and photovoltaic investigation.
350 400 450 500 550 600
Absorbance
Wavelength (nm) 2.2 nm 3.2 nm 4.6 nm
Figure 5.1 Absorption spectra of 2.2, 3.2, and 4.6 nm diameter CdS quantum dots in toluene.
5.3 Decoration of CdS QDs on ZnO nanorods
Figure 5.2(a) shows the cross-sectional scanning electron micrograph (SEM) of typical single-crystal ZnO nanorods on the ITO glass substrate. The nanorods have lengths and diameters ca. 320 nm and 90 nm, respectively. Figure 5.2(b) shows dark-field TEM images of a ZnO nanorod decorated with CdS QDs (4.6 nm) and demonstrates CdS QD coverage on the nanorods was so high to cover the nanorods.
The attached QDs appear as randomly oriented crossed-fringe patterns on the nanorods surface, and energy dispersive spectrometer (EDS) confirms that they are CdS (Figure 5.2(c)). These figures prove we achieve relatively high coverage of CdS QDs on ZnO nanorods and a uniform coverage throughout the film. Moreover, because prolonging time of immersion in CdS solution does not further increase CdS absorption, we assume the coverage of CdS particles on ZnO nanords surface to be a
45
monolayer. Such monolayer coverage of the CdS particles is analogous to TiO2 films modified with sensitizing dyes [142].
Figure 5.2 (a) Cross-sectional scanning electron micrograph of ZnO nanorods. (b) Dark-field transmission electron micrograph of a ZnO nanorod decorated with CdS
(b)
50 nm
1μm
(a)
(c)
46
QDs. (c) Element composition analysis of an individual ZnO nanorod decorated with CdS QDs.
5.4 Annealing effect on P3HT
The P3HT film annealed at 200 °C, around the melting point of P3HT [143], suggest that the polymer completely fills the nanorod array, as shown in Figure 5.3(a).
Moreover, X-ray diffraction (XRD) patterns show that annealed the composite film substantially improves the P3HT crystallinity, which should enhance its hole mobility [144]. Intensifies and sharpens of the P3HT (100) diffraction peak are improved significantly, which originates from crystals with their alkyl spacing direction (100) normal to the substrate (Figure 5.3(a)). Diffraction peaks of CdS QDs don’t appear due to a fewer amount of CdS QDs. Figure 5.3(b) illustrates the UV-vis spectra of ZnO/P3HT annealed at 220 °C and without annealing. The UV-vis spectra of composite films show the development of a well-defined shoulder at ~610 nm upon annealing. This feature is connected with an increase in molecular order and enhanced hole mobility in the P3HT film [145].
Figure 5.3 (a) XRD patterns of ZnO/CdS/P3HT with annealing at 220 °C and without annealing. (b) Absorption spectra of ZnO/CdS/P3HT with annealing at 220 °C and without annealing.
(a) (b)
47
5.5 Effect of CdS QD size
Figure 5.4 shows the current density-voltage (J-V) characteristics of hybrid solar cells based on P3HT and ZnO nanords with and without CdS QDs under simulated AM 1.5G irraadiation at an intensity of 100 mW cm−2 in the air. Modifying the ZnO nannorods with CdS QDs, short-circuit current density JSC of the hybrid solar cell reached 1.14 mA cm−2, which was 2.9 times larger than that without CdS QDs. (Table 2 and Figure 4). It seems plausible that the increased photocurrent is attributed to the efficient photosensitizing effect of the CdS QD for metal oxide owing to the absorption bands in the visible region. To determine the origin of the increase in JSC, we measured the J-V curves of solar cells based on P3HT and ZnO nanords modified with three different-sized CdS QDs, as can be seen in Figure 5.4. It shows that when the absorption bands of CdS QD shift to UV region (2.2 nm and 3.2 nm), the JSC are still improved. The JSC varies with particle sizes and the maximum photocurrent is seen with 3.2 nm diameter CdS particles. It suggests that other mechanisms should be involved in the increase in JSC except photocurrent from QDs. For electron transfer from donor to acceptor to occur, the lowest unoccupied molecular orbital (LUMO) of donor needs to be ~0.5 eV higher than the LUMO of the acceptor [146]. In this case of P3HT, however, this energy difference is much higher, namely 1.2 eV so we suspected that the ZnO/P3HT interface is not a suitable one for charge separation.
Therefore, we replaced ZnO/P3HT interface with CdS QDs/P3HT interface which is known to readily split excitons generated in P3HT. Expected energy levels of ZnO, CdS, and P3HT, determined from literature values [147-149], are presented in Figure 5.5(a). The conduction band of bulk CdS is lower in energy than LUMO of P3HT, enabling an electron transfer cascade from P3HT to the CdS interface layer to ZnO and, thereby, potentially enhancing charge separation at the ZnO interface.
48
Additionally, due to the valence band level of bulk CdS relative to P3HT, the CdS QD is expected to obstruct hole transfer between P3HT and ZnO, and thus to localize hole in the P3HT away from the metal oxide surface. Thus, the CdS QDs serve as an electronic mediator that enhances the electron transfer efficiency from P3HT to ZnO.
Moreover, because of the small electron effective mass (me =0.2 mo) versus the significantly larger hole mass (mh= 0.8 mo), the shift in the conduction band energy is significantly greater than the shift in valence band energy for quantized particles resulting in increasing the driving force for charge injection. Thus, we can optimize the interface energetics by adjusting the size of CdS QDs.
0.0 0.1 0.2 0.3 0.4 0.5
-1.5 -1.0 -0.5 0.0
Current density (mA/cm2 )
Potential (V)
ZnO nanorod 4.6 nm 3.2 nm 2.2 nm
Figure 5.4 Photovoltaic performances of the devices with different diameter CdS quantum dots.
49
Table 2 Summary of the device performance of different diameter CdS quantum dots.
JSC /mA cm-2 VOC / V FF (%) PCE (%)
Only ZnO 0.39 0.30 40 0.05
2.2 nm 1.05 0.36 39 0.15
3.2 nm 1.38 0.38 41 0.21
4.6 nm 1.17 0.35 40 0.16
The energy of the lowest exciton state (ECdS*), a bound pair of 1S electron in the conduction band (CB) and 1S hole in the valence band (VB), is given by [150,151]
R the effective masses of electrons and holes, respectively. The second and third terms are the confinement energy of the 1S electron and 1S hole, respectively. The fourth term results from electron hole coulomb attraction. Following the Brus’ theory [150,151], we estimated the shift in the conduction band energy for QDs with an average diameter 2.2 nm, 3.2 nm, and 4.6 nm is about 0.9 eV, 0.5 eV, and 0.2 eV respectively. Therefore, when we employ 3.2 nm CdS instead of 4.6 nm quantum dots, the energy difference between the CB of CdS and the LUMO of P3HT is reduced to 0.5 eV which is optimal value for electron transfer.19 Because of the optimal offset, easy electron transfer from P3HT to CdS causer the JSC to increase 1.3 times. In contrast with 2.2 nm CdS QDs, lower JSC should be mainly due to less overlap between 2.2 nm CdS QDs absorption and the solar spectrum.
50
550 600 650 700 750 800
P3HT/glass
P3HT/glass