TiO2 micro-flowers composed of nanotubes and their application to dye-sensitized solar cells

TiO2 micro-flowers were made to bloom on Ti foil by the anodic oxidation of Ti-protruding dots with a cylindrical shape. Arrays of the Ti-protruding dots were prepared by photolithography, which consisted of coating the photoresists, attaching a patterned mask, illuminating with UV light, etching the Ti surface by reactive ion etching (RIE), and stripping the photoresist on the Ti foil. The procedure for the blooming of the TiO2 micro-flowers was analyzed by field emission scanning electron microscopy (FESEM) as the anodizing time was increased. Photoelectrodes of dye-sensitized solar cells (DSCs) were fabricated using TiO2 micro-flowers. Bare TiO2 nanotube arrays were used for reference samples. The short-circuit current (Jsc) and the power conversion efficiency of the DSCs based on the TiO2 micro-flowers were 4.340 mA/cm2 and 1.517%, respectively. These values of DSCs based on TiO2 micro-flowers were higher than those of bare samples. The TiO2 micro-flowers had a larger surface area for dye adsorption compared to bare TiO2 nanotube arrays, resulting in improved Jsc characteristics. The structure of the TiO2 micro-flowers allowed it to adsorb dyes very effectively, also demonstrating the potential to achieve higher power conversion efficiency levels for DSCs compared to a bare TiO2 nanotube array structure and the conventional TiO2 nanoparticle structure.


Background
Dye-sensitized solar cells (DSCs) have received much attention since Grätzel and O'Regan achieved a remarkable level of efficiency through their use of mesoporous TiO 2 films as a photoanode for DSCs in 1991 [1]. DSCs have several advantages compared to Si or copper indium gallium selenide (CIGS) solar cells as follows: (a) DSCs can be fabricated with non-vacuum processes, as opposed to Si or CIGS solar cells. The use of nonvacuum equipment offers the possibility to reduce costs. (b) Wet etching processes such as saw damage etching and texturing, which are widely used in Si solar cells, are not required during the fabrication of DSCs. The fabrication of DSCs is thus simplified without a wet etching process. (c) Colorful DSCs can be easily fabricated because dyes have various colors according to their light absorption characteristics. Although DSCs have these merits, the relatively low power conversion efficiency has become the main cause which limits the commercialization of DSCs.
The effort to find alternative TiO 2 nanostructures has been an important issue to researchers who attempt to increase the power conversion efficiency of DSCs. Various types of nanotechnologies have been applied to alternative TiO 2 nanostructures such as nanorods [13], nanowires [14,15], nanotubes [16,18,19,22,23,25,[27][28][29][30]42,43], nanohemispheres [21,24], and nanoforests [17,20]. These structures were used to increase the surface area for dye adsorption and to facilitate charge transport through TiO 2 films. Of these nanostructures, the TiO 2 nanotube structure has the best potential to overcome the limitations of the TiO 2 nanoparticle structure. A previous report showed that the electronic lifetimes of TiO 2 nanotube-based DSCs were longer than those of TiO 2 nanoparticle-based DSCs [30]. Due to the one-dimensional structure of TiO 2 nanotube arrays, charge percolation in TiO 2 nanotube-based films is easier than it is in the TiO 2 nanoparticle structure [27][28][29].
In this study, TiO 2 micro-flowers composed of nanotubes were fabricated by means of dot patterning, Ti etching, and anodizing methods. The dot patterning and etching of Ti substrates increased the anodizing area to form TiO 2 nanotubes. By controlling the anodizing time, beautiful TiO 2 micro-flowers were successfully made to bloom on Ti substrates and were applied to the photoelectrodes of DSCs. To the best of our knowledge, this is the first study to report the fabrication of TiO 2 micro-flowers and their application to DSCs. The TiO 2 micro-flower structure is strongly expected to enhance the possibility to overcome the limitations of the TiO 2 nanoparticle structure.

Methods
To fabricate the protruding dot patterns on a 0.5-mmthick Ti foil (99%, Alfa Aesar Co., Ward Hill, MA, USA), 5-μm-thick negative photoresists (PR; L-300, Dongjin Co., Hwaseong-Si, South Korea) were coated on a flat layer of Ti foil using a spin coater (Mark-8 Track, TEL Co., Tokyo, Japan). The coated photoresists were softly baked at 120°C for 120 s and hardly baked at 110°C for 5 min. A dot-patterned photomask was used for PR, the patterning process via UV light exposure. UV light having an energy of 14.5 mJ/s was used for illumination for 5 s, and the PR were developed. The PR at areas not exposed to UV light were removed.
The PR-patterned Ti foil was dry-etched at 20°C for 30 min using reactive ion etching (RIE) equipment (ICP380, Oxford Co., Abingdon, Oxfordshire, UK). BCl 3 and Cl 2 were used as the etchant gas in the RIE process with a top power of 800 W and a bottom power of 150 W. The photoresists on the UV-exposed area served to protect the flat Ti surface during the RIE process. Only the Ti surface at the area not exposed to UV was etched out. The remaining photoresist after the RIE process was stripped at 250°C for 20 min using a photoresist stripper (TS-200, PSK Co., Hwaseong-si, South Korea). O 2 and N 2 gases were used to remove the photoresist at a power of 2,500 W.
Before the anodizing process, Ti foil samples patterned with protruding dots were successively sonicated with acetone, ethanol, and deionized (DI) water to remove any residue on their surfaces. TiO 2 micro-flowers, consisting of TiO 2 nanotubes, were fabricated by the anodization of the Ti foil sheets which had been patterned with protruding dots in an ethylene glycol solution containing 0.5 wt% NH 4 F. A constant potential of 60 V with a ramping speed of 1 V/s was applied between the anode and the cathode. Pt metal was used as a counter cathode. The anodizing time was controlled for the successful blooming of the TiO 2 micro-flowers. The as-anodized TiO 2 nanotubes were rinsed with DI water and annealed at 500°C for 1 h. The morphologies of the TiO 2 nanotubes and the micro-flowers were studied by field emission scanning electron microscopy (FESEM, Hitachi SU-70, Tokyo, Japan). The as-anodized and annealed TiO 2 nanotubes were analyzed by X-ray diffraction (XRD; Rigaku D/ MAX-RC, Cu Kα radiation, Rigaku Corporation, Tokyo, Japan) to confirm the crystallization characteristics.
Results and discussion Figure 1 shows FESEM images of Ti-protruding dots which have a cylindrical shape. The Ti surface at the UV-exposed area was flat because the cross-linked photoresist blocked the etching by reactive ions. However, the surface at the area not exposed to UV was very rough due to the RIE in the vertical direction. The diameter and height of the protruding dots were approximately 4 and 5 μm, respectively.
The microstructures while increasing the anodization time from 1 to 7 min are shown in Figures 2, 3, 4, 5, and 6. Figure 2 shows FESEM images of a Ti surface which was patterned with protruding dots and anodized for 1 min at 60 V in an ethylene glycol solution containing 0.5 wt% NH 4 F. Anodized Ti dot arrays are shown in        Figure 2d shows that the TiO 2 nanotubes grew vertically from the wall of the protruding dots. When the anodization time was increased to 2 min, small cleavages formed between the top areas and side walls of the protruding dots, as shown in Figure 3. Figure 3b,c shows approximately 700-nm-thick TiO 2 nanotube arrays.
When the anodization time was increased to 4 min, beautiful TiO 2 micro-flowers started to bloom. The arrays of TiO 2 micro-flowers are shown in Figure 4a. The thickness of each TiO 2 nanotube is linearly correlated with the extent to which the TiO 2 micro-flowers bloom. The blooming of the TiO 2 micro-flowers is due to the severe cleavages of the TiO 2 nanotubes between the top areas and the side walls of the protruding dots. As the anodization time was increased to 5 min, core bundles of nanotubes in TiO 2 micro-flowers were slightly bent in random directions, as shown in Figure 5a,b,c,d. This occurred due to the difference in the growing speed of each TiO 2 nanotube in the core bundles. The measured thickness of the TiO 2 nanotubes in Figure 5d was 2 μm. As the anodization time was increased to 7 min, the center area of the core nanotube bundles in the TiO 2 micro-flowers was removed, as shown in Figure 6a,b,c. Figure 6d shows the cleavage areas of the TiO 2 microflowers. The structure of the TiO 2 nanotubes in that area collapsed due to the additional etching by the fluorine ions in the anodizing solution. Figure 7 shows the schematic mechanism involved in the blooming of the TiO 2 micro-flowers. One of the Tiprotruding dots from the photolithography and RIE process shows a cylindrical shape in Figure 7a. Figure 7b shows that the TiO 2 nanotubes grew in a vertical direction from the Ti surface due to the anodizing process. As the thicknesses of the TiO 2 nanotubes at the cylindrical upper side (area A) and at the cylinder side (area C) increased, the Ti-supporting metal at the cylinder corner (area B) was completely converted into TiO 2 nanotubes. The TiO 2 nanotubes without Ti-supporting metal in area B finally fell onto the TiO 2 nanotubes which had grown in area C, as shown in Figure 7c. Several horizontal cleavages in area B formed due to the collapse of the TiO 2 nanotubes in area B. Several vertical cleavages in areas B and C were also observed, resulting from the volume expansion when the Ti was converted into TiO 2 nanotubes. Volume expansion in an organic anodizing solution was reported previously [44]. Figure 7d shows that the growing TiO 2 nanotubes in area C pushed and pushed TiO 2 nanotubes between areas A and B to area C. More horizontal cleavages in area B were created due to the pushing of the TiO 2 nanotubes, and these cleavages formed the multi-layered petals in the TiO 2 micro-flowers. Figure 7c,d shows the blooming of beautiful TiO 2 micro-flowers. This is a first blooming of TiO 2 micro-flowers. The thickness of the TiO 2 nanotubes in areas A and C gradually increased with the anodization time. Finally, all Ti metal was converted into TiO 2 nanotubes, leaving no additional Ti metal to support the TiO 2 nanotubes in area A. Figure 7e shows that the TiO 2 nanotubes without Ti-supporting metal in area A were detached from the center of the nanotube bundles. This removal of the TiO 2 nanotubes in area A left an empty core in the TiO 2 micro-flowers. These TiO 2 micro-flowers with empty cores are different from those shown in Figure 7c,d. This result represents a second blooming of the TiO 2 micro-flowers. Figure 8 shows the results of an XRD analysis of the as-anodized TiO 2 micro-flowers and the annealed TiO 2 micro-flowers. Figure 8a shows only the Ti peaks, revealing that the as-anodized TiO 2 nanotubes in the microflowers have an amorphous crystal structure. However, if  the as-anodized TiO 2 nanotubes are annealed at 500°C for 1 h, the crystal structure of the TiO 2 nanotubes is converted into the anatase phase. Anatase peaks and Ti peaks were found, as shown in Figure 8b. From the XRD results, it can be confirmed that the annealed TiO 2 micro-flowers exist in the anatase phase.
As shown in Figure 9, bare TiO 2 nanotubes and TiO 2 micro-flowers were applied for use in DSC photoelectrodes. DSCs based on bare TiO 2 nanotube arrays were used as reference samples to compare the J-V characteristics with DSCs based on TiO 2 micro-flowers. Photoelectrodes based on bare TiO 2 nanotubes were prepared by an anodizing process of flat Ti foil. On the other hand, photoelectrodes based on TiO 2 micro-flowers were fabricated by an anodizing process of Ti foil patterned and shaped such that they approximated cylindrical protruding dots. Figure 10 shows the J-V characteristics of DSCs based on the bare TiO 2 nanotubes and TiO 2 micro-flowers when the thicknesses of the TiO 2 nanotubes are 1.5 and 2.0 μm, respectively. When the thickness of the TiO 2 nanotubes was 1.5 μm, the short-circuit current (J sc ), open-circuit voltage (V oc ), and power conversion efficiency of the DSCs based on the TiO 2 micro-flowers were slightly higher than those of the bare TiO 2 nanotubes, as shown in Figure 10 and Table 1. However, the fill factor of the samples based on the TiO 2 microflowers showed a decrease compared to that of the bare samples. When the thickness of the TiO 2 nanotubes was increased from 1.5 to 2.0 μm, the J sc of the DSCs based on the TiO 2 micro-flowers increased from 3.838 to 4.340 mA/cm 2 . This appears that the improvement of J sc in the TiO 2 micro-flower samples is due to the increased surface area for dye adsorption. The efficiency of DSCs based on TiO 2 micro-flowers reached 1.517%. The obtained efficiency levels were relatively low, as the thicknesses of the TiO 2 nanotubes were very thin at 1.5 and 2.0 μm. The thickness of the TiO 2 nanoparticle layer in the conventional DSCs was approximately 20 μm. If the thickness of the TiO 2 micro-flowers is increased, its efficiency will also increase. The performance levels of DSCs based on these TiO 2 micro-flowers will also improve if the morphologies of the protruding dots, such as the dot diameter, the distance between adjacent dots, and the height of the cylindrical protrusions, are tailored. Our future work will concentrate on all of these factors to attain the maximum efficiency level from DSCs based on TiO 2 micro-flowers. The conclusion of this report is that DSCs based on TiO 2 micro-flowers have the potential to achieve higher efficiency levels compared to DSCs based on normal TiO 2 nanotubes and TiO 2 nanoparticles.

Conclusion
Ti-protruding dots with a cylindrical shape were fabricated by coating with a photoresist, illuminating with UV light, etching the Ti surface via the RIE method, and stripping the photoresist on Ti foil. When the Tiprotruding dots were anodized for over 3 min, beautiful arrays of TiO 2 micro-flowers successfully bloomed on the Ti foil sheets. The blooming TiO 2 micro-flowers were applied as the photoelectrodes of DSCs. The J-V characteristics of the DSCs based on the TiO 2 microflowers were compared with those based on bare TiO 2 nanotubes. The J sc and power conversion efficiency values of DSCs based on TiO 2 micro-flowers were higher than those of bare samples. TiO 2 micro-flowers facilitated better dye adsorption, resulting in higher J sc values. The TiO 2 micro-flowers had a larger surface area  for dye adsorption compared to that of bare TiO 2 nanotubes. The efficiency of the DSCs based on the TiO 2 micro-flowers was found to reach 1.517%. The efficiency levels of the DSCs based on the TiO 2 micro-flowers were relatively low compared to those of conventional DSCs based on TiO 2 nanoparticle structures, as the thickness of the TiO 2 nanotubes in the micro-flowers was very small. To improve the efficiency of DSCs based on TiO 2 micro-flowers, our future work will concentrate on controlling the characteristics of the dot patterns such as the dot diameter, the distance between adjacent dots, and the height of the protruding dots.