Introduction
Monodisperse colloidal microspheres with controlled size are of key importance in colloidal chemistry and material chemistry because of the theoretical research on them in physical chemistry and their potential applications in optics, electronics, catalysis, sensors, and so forth [1-4]. Moreover, recent work on the crystallization of colloidal spheres has fully demonstrated the potential to obtain interesting and useful functionality not only from the constituent materials of the colloidal particles but also from the long-range order of the crystalline lattice [5-10]. As a result, much effort has been devoted to the synthesis of colloidal spheres of various materials, such as polystyrene [11], silica [12], ZnSe [13-17], ZnS [18-20], CdS [21], CdSe [22], Ag2Se [23], TiO2 [24-26], ZnO [27,28], PbS [29].
However, the reported colloidal spheres of semiconductors have mainly been synthesized by template-directed approach, in which the easy-fabrication colloidal spheres of materials are used as the template and are then coated or reacted with other materials. This is because semiconductors tend to grow anisotropically into non-spherical structures, including rods, wires, belts, tetrapods, cubes, and flowers. Hence, it is still a challenge to fabricate spherelike inorganic semiconductors in large quantities by a simple direct synthesis.
As an important II-VI semiconductor with a wide and direct band gap (3.37 eV) and a large exciton binding energy (60 eV), ZnO will attract much interest in the field of colloidal materials, owing to its specific electrical, catalytic, photochemical and optoelectronic properties [30-32]. Zhang’s group reported that ZnO colloid spheres synthesized by refluxing in diethylene glycol display good photovoltaic properties [33].
Here, ZnO quasi-spherelike colloids were synthesized by aqueous precipitate, using zinc salts and diethanolamine (DEA) as raw materials. The size of quasi-spherelike colloids can be modulated in the range of 80 nm-3 μm by adjusting starting concentration of zinc source, solvents, and electrolytes. The measurements of photovoltaic properties of different-size ZnO spheres demonstrate that quasi-spheres of 300-600 nm have the highest short-circuit current density (Jsc) and over light conversion efficiency (η). This result should be attributable to the resonant scattering existing in the quasi-spheres with the wavelength of the visible light, leading to the enhancement of photon absorption.
Experimental details
Synthetic procedures
All the chemical reagents used in our experiments were of analytical grade and used as received. Zn(NO3)2·6H2O (0.005 mol) was dissolved in deionized water (70 mL) under magnetic stirring and then DEA (30 mL, 98%) was added to form a transparent mixture. The same procedures were repeated to prepare the mixture with various starting concentration of zinc salts and solvents. Subsequently, such mixture was transferred to a two-necked flask and refluxed at 110°C for 2 h. A large quantity of white precipitates at the bottom of flask were collected and washed with alcohol and de-ionized water several times. The residues were then dried at 60°C in an oven for 3 h.
Characterization
The phase and crystal structure of the products were determined by X-ray diffraction (XRD, Siemens D-500 with Cu Kα radiation and normal 2θ scans). The morphology and microstructure of the sample were analyzed by scanning electron microscopy (SEM; JEOL JSM-820) and field emission scanning electron microscopy (FESEM; JEOL JSM-63357). The specific surface area was measured by a BET Surface Area Analyzer (Coulter products SA3100).
Preparation of ZnO film electrodes and measurement of dye-sensitized solar cells (DSSCs)
To fabricate ZnO films, the conductive glass substrates were first cleaned by sonicating in detergent, acetone and ethanol in order and then air-dried. Parallel edges of each substrate were covered with 10 μm-thick scotch tape to control the thickness of the film. A few drops of the resultant ZnO colloidal spheres were then placed onto the glass substrates and the films were formed by a doctor-blading process. The films were then immediately heated at a temperature of 350°C for 1 h, forming a layer of white film during the quick evaporation of the solvent. The thickness of the film is estimated to be 10 μm, which is the similar thickness as the spacers.
Before solar cell testing, the ZnO films were heated to 70°C and sensitized with standard ruthenium-based N3 red dye. The heated films were immersed in N3 dye with a concentration of 5×10-4 M in ethanol for 20 min. The samples were then rinsed with ethanol to remove excess dye on the surface and air-dried at room temperature. The counter electrode consisting of a platinum-coated silicon substrate was face to face placed on the ZnO film electrode. The two electrodes were then sandwiched together with two heavy duty clips.
The electrolyte in this study was a liquid admixture containing 0.5 M LiI, 50 mM I2, and 0.5 M 4-tertbutylpyridine in 3-methoxypropionitrile. The photovoltaic behavior was characterized when the cell devices were irradiated by simulated AM 1.5 sunlight with an output power of 100 mW·cm-2. An Ultraviolet Solar Simulator (Oriel 66902) with a 200 W Xenon Lamp Power Supply was used as the light source, and a Semiconductor Parameter Analyzer was used to measure current-voltage (I-V) curves.
Results and discussion
The morphology and size distribution of the as-prepared samples were investigated using the SEM technique. Figure 1 shows the SEM images of three typical samples prepared by refluxing within different parameters. The low-magnification SEM images in Figs. 1(a), 1(c) and 1(e) display a quasi-spherelike shape over a large area from sample 1, sample 2 and sample 3, respectively. The high-magnification SEM images in Figs 1(b), 1(d) and 1(f) show the sizes of the quasi-spheres in sample 1, sample 2 and sample 3, respectively. The results demonstrate that the size of quasi-spheres can be modulated by varying Zn2+ source and its concentration. The details of size distribution were analyzed by histograms in Fig. 2, which are depicted via measuring the diameters of the quasi-spheres in SEM images (Each sample was randomly selected 250 particles out). It is followed that the diameter of sample 1 is among 80-180 nm, sample 2 among 300-600 nm, sample 3 among 1.3-2.9 μm. This result shows that the size distribution of the as-prepared samples is broad.
Fig.1 SEM images of three typical ZnO quasi-spheres prepared by refluxing under different parameters. (a) and (b) 1 g of Zn(CH3COO)2·2H2O dissolved into the mixture of 110 mL ethanol and 2.5 mL DEA (sample 1); (c) and (d) 2.25 g of Zn(NO3)2·6H2O dissolved into the mixture of 80 mL H2O and 30 mL DEA (sample 2); (e) and (f) 1.5 g of Zn(NO3)2·6H2O dissolved into the mixture of 80 mL H2O and 30 mL DEA (sample 3) |
To analyze the detailed structure of the as-prepared quasi-spheres, high-magnification FESEM were used. Figures 3(a) and 3(b) show the FESEM images of samples prepared via a Zn(NO3)2·6H2O-H2O system and Zn(CH3COO)2·2H2O-C2H5OH system, respectively, both displaying hierarchical nanostructures. The typical quasi-spheres are the aggregates of smaller particles of about dozens of nanometer in diameter.
Figure 4 shows the XRD patterns acquired from the three samples. All of the diffraction peaks of each sample were indexed as a hexagonal phase of wurtize-type ZnO, consistent with JCPDS card No. 36-1451. No other impurities, such as Zn(NO3)2, Zn(OH)2, and other compounds were detected. The clearness and high intensity of diffraction peaks in the XRD pattern confirm that the products are well-crystallized. The width of diffraction peaks indicates that the as-prepared quasi-spheres are composed of small nanoparticles, which is in accord with the high-magnification FESEM observations. On the full width at half-maximum (FWHMs) of (100), (002) and (101) diffraction peaks, the average sizes of constructing subunits of the as-prepared ZnO quasi-spheres were calculated according to the Scherrer equation [34] and the results are shown in Table 1. The average sizes of the subunits of sample 1, sample 2 and sample 3 are 21, 19 and 22 nm, respectively.
Tab.1 Size of component nanoparticles of three samples calculated from XRD peaks |
| diffraction peak | sample 1 | sample 2 | sample 3 |
|---|---|---|---|
| (100) | 22.313±0.002 | 19.774±0.002 | 22.372±0.002 |
| (002) | 20.062±0.002 | 20.051±0.002 | 22.221±0.002 |
| (101) | 22.581±0.002 | 18.042±0.002 | 22.271±0.002 |
| average/nm | 21.322±0.002 | 19.292±0.002 | 22.291±0.002 |
The surface area and total pore volume of the ZnO quasi-spheres were measured by nitrogen sorption isotherms (the relative error of the used machine is smaller than 2% for the high speed gas sorption analyzer). The data were summarized in Table 2. Sample 2 has the largest BET surface area. With regard to the total pore volume of samples, it is getting smaller in the order of samples 1-3. On the basis of this result, it can be followed that the aggregation of subunits for the samples is more and more compact. Therefore, the compactness for samples 1-3 is getting better in sequence.
Tab.2 BET surface areas and total pore volumes of three ZnO samples with different sizes |
| sample | BET surface area/(m2·g-1) | total pore volume/(ccg-1) |
|---|---|---|
| 1 | 12.62 | 6.07×10-2 |
| 2 | 14.88 | 3.64×10-2 |
| 3 | 2.04 | 4.23×10-3 |
The three solar cells fabricated with different-size ZnO quasi-spheres (samples 1-3) were characterized by measuring I-V behavior when the cells were irradiated with a power density of 100 mW·cm-2 and in dark. The corresponding current density J-V curves were shown in Fig. 5. Table 3 summarizes the measured and calculated values of the cells’ parameters acquired from each J-V curve in Fig. 5(a), including Jsc, open-circuit voltage (Voc), maximum voltage (Vmax) out, maximum current (Jmax) out, filled factor (FF), η. It clearly shows that sample 2 achieves the highest Jsc and η, whereas sample 3 is the lowest. All the samples possess the similar Voc of approximately 600 mV and dark J before bias V of 600 mV.
Tab.3 Photovoltaic properties of DSSCs based on different-size ZnO spheres |
| ZnO film | Voc/mV | Jsc/(mA·cm-2) | Vmax/mV | Jmax/(mA·cm-2) | FF1)/% | η1)/% |
| sample 1 | 584 | 9.2 | 300 | 4.8 | 26.5 | 1.42 |
| sample 2 | 564 | 10.2 | 300 | 5.8 | 30.2 | 1.74 |
| sample 3 | 568 | 5.8 | 310 | 3.6 | 33.2 | 1.1 |
1) η and FF are calculated from and FF= Pout, max(VocIsc)-1, where Pout, max is the maximum output power density, Pin is the incident light power density, Voc is the photovoltage at open circuit |
In a DSSC, the Voc is dependent on the potential difference established by the V of the film and that of the Pt counter electrode, in which, the surface chemistry of the film’s material may influence the Voc. Since the three solar cells are fabricated with the similar materials and the same procedures. The similar Voc for them is expected. The Jsc is determined by the initial number of photogenerated carriers, the injection efficiency of electrons from dye molecules to semiconductor, and the recombination rate between the injected electrons and oxidized dye or redox species in the electrolyte. It is reasonable to assume the same injection efficiency and recombination rate for the given ZnO/N3/electrolyte systems. Accordingly, the initial number of photogenerated carriers may be significantly affected by the variation in the light-harvesting capability of photoelectrodes with different film structures. In the three fabricated DSSCs, the only difference between them is the diameter of ZnO quasi-spheres for the three samples, which may lead to the difference of the Jsc between them. The specific explanation is as follows.
Typically, light scattering [35] occurs with particle>100 nm in diameter. Further, Mie theory [36] and Anderson localization of light [37] provide the analytical description for the scattering of light by spherical particles and predict that resonant scattering may occur when the particle size is comparable to the wavelength of incident light. The secondary quasi-spheres for sample 2 are 300-600 nm in diameter and within the wavelength of visible light. They can therefore become efficient scatterers for visible light and even produce resonant scattering in the DSSC, which results in an increase in the generation of electron-hole pairs and light-harvesting capability of the photoelectrode. The Jsc and η are enhanced in the DSSC fabricated from sample 2. However, samples 1 and 3 are nano-sized and micrometer-sized, respectively, which are far from the wavelength of visible light. Hence, they can not produce resonant scattering and obtain a relatively low Jsc and η.
The three samples possess a similar FF in Table 3. The FF in a DSSC is affected by the capability of inhibition of the back electron transfer from ZnO to [38]. If ZnO film has a high connectivity between subunits, there is a rapid collection of photogenerated electrons and the degree of charge recombination get reduced so as to enhance the FF. The samples 1-3 get more and more compact in sequence as the above-mentioned discussion, leading to higher and higher FF value in the same sequence. However, the FF in this work is lower than the previous reports [33,39,40]. The polydisperse of the as-prepared ZnO quasi-spheres may have caused incomplete particle packing and cracking after sintering, which would result in low connectivity among quasi-spheres and influence the FF in a DSSC.
Conclusion
In this paper, ZnO quasi-spheres, aggregated nanoparticles with about 20 nm in diameter, were synthesized via a simple and low-temperature aqueous solution route. The size of ZnO quasi-spheres is easily tunable from dozens of nanometer to several micrometers by changing the type and concentration of zinc salts. The DSSCs constructed by ZnO quasi-spheres of 300-600 nm in diameter obtain the highest Jsc and η, compared to other samples with the diameter of 80-180 nm and 1.3-2.9 μm. The improvement of Jsc and η for sample 2 should be attributable to the resonant scattering produced in the photoelectrode that can enhance the light-harvesting capability. Although the property of DSSCs fabricated from the as-fabricated ZnO quasi-spheres is inferior to other reports, it indirectly confirms that the resonant scattering exists only in particles with the diameter of the wavelength of visible light.