James Lim Harvard College ‘16
Eugen Panaitescu1, Latika Menon1
1Department of Physics, Northeastern University
This study examined the effects of annealing parameters on the crystallization of anatase in order to optimize the operations of dye-sensitized solar cells (DSSCs). The nanotubes were synthesized using a two electrode Ti-Pt anodizing configuration in 99% ethylene glycol and 1% water. They were characterized to have lengths ranging from 60 to 80 μm and a high roughness factor of approximately 3,000. After nitrogen anneals under various combinations of the peak temperature and the duration of isothermal scanning, the crystalline nature of the nanotubes was studied. XRD spectra analysis revealed that anatase crystallization occurred at 250 ˚C, an observation previously unreported. In addition, a strong positive correlation was observed between the duration of isothermal scanning and the degree of crystallization, suggesting that it is possible to produce DSSC-applicable anatase at such low temperatures. SEM imaging also revealed two novel intermediary morphologies that enabled the modeling of surface collapsing and nucleation.
1.1 The Dye-Sensitized Solar Cell (DSSC)
Dye-sensitized solar cells (DSSCs) are a new generation of solar cells first developed in 1991 by Gratzel and O’Reagan . DSSCs operate between an anode, made of transparent conducting oxides (TCOs) deposited on the back of a glass slide, and a conducting cathode, which is a sheet of platinum or gold metal. Between the anode and the cathode are a substrate layer of titanium dioxide (TiO2), an electrolyte solution, and a mono-molecular layer of dye. The importance of the dye lies in its ability to absorb a wide range of the electromagnetic spectrum. Titanium dioxide has a wide band gap and thus can absorb only a small fraction of solar photons [2, 3], whereas the unique electronic structures of photosensitive dyes, such as the widely-used ruthenium-polypyridine dye (cis-bis(2,2-bipyridine)bis(isothiocyanato)ruthenium(II)), enable them to absorb photons with various energy levels .
When sunlight enters through the transparent anode and strikes the dye coated on the TiO2 layer, photons of sufficient energy are absorbed by the dye molecules, causing them to transition to an excited state. From this intermediary state, an exciton is produced and is immediately separated into its two components: the electron and the hole. The electron is directly injected into the TiO2 conduction band and travels to the anode via diffusion along the electron concentration gradient. Meanwhile, the hole is neutralized by a redox reaction chain in the electrolyte solution, often based on iodide/triiodide (I–/I3– ) . The electron at the anode flows along the external circuit to the counter electrode (CE), where it is incorporated into the redox chain. Figure 1 illustrates the operation mechanism of a standard DSSC.
Currently, DSSCs are at the center of extensive scientific research and show great promise as a novel method for generating solar energy. The replacement of high-cost silicon wafers with readily available TiO2 presents an enormous economic advantage over the conventional silicon-based solar cells. In addition, it has been reported that DSSCs have reached power production efficiency levels of up to 11% under laboratory conditions . Considering that commercially available multicrystalline silicon solar cells operate between efficiency levels of 14% and 19% , the coupled cost advantage and comparable efficiency of DSSCs make them viable as competitors in the solar cell market.
1.2 Titania (TiO2) Nanotubes
Enhancing surface area exposed to sunlight is critical in the development of more efficient DSSCs due to the importance of electron excitation in the operation mechanism. However, the volume of material used for DSSCs must be maintained below a certain level in order for DSSCs to remain economically advantageous. Thus, recent advancements in nanoscale technology have focused on developing highly ordered nanoporous and nanotubular structured materials. While various wide band gap oxides such as ZnO , SnO2 , and Nb2O5  have been investigated, currently the most common material used to create substrates for DSSCs is TiO2, also known as titania.
Titania has traditionally been applied to DSSCs in the form of nanocrystalline, mesoporous films. However, recent studies have reported that the structural disorder at the contact points between two crystalline nanoparticles can lead to enhanced scattering of free electrons, reducing electron mobility [11, 12]. Thus, an ordered and strongly interconnected nanoscale photoanode architecture offers most the potential for improved electron transport , and studies have already demonstrated that nanotube-based DSSCs outperform their nanocrystal-based counterparts . This finding has recently stimulated research regarding the synthesis of titania nanotubes as a substitute structure for the traditional nanocrystals. Titania nanotubes may also be synthesized through various previously tested means, such as hydrothermal synthesis [15, 16], sonochemical deposition , and the sol-gel process .
One of the simplest and most effective synthesis methods to date is electrochemical an odization, in which a thin titanium foil and a platinum-based cathode are immersed in a liquid electrolyte, typically hydrofluoric acid (HF) or potassium fluoride (KF) in ethylene glycol (EG). Once an anodizing voltage is passed through the setup, an oxide layer accumulates on the surface of the titanium foil via the following reaction mechanism:
The morphology of the titania nanotubes, specifically their lengths, diameters, and wall thicknesses can be controlled by adjusting the anodization voltage as well as the duration of current flow .
Titania exists naturally in several different crystalline forms, with varied applications. For example, rutile is widely used in dielectrics and high-temperature oxygen gas sensors, while anatase is preferred in chemical catalysis and, of particular relevance to this study, dyesensitized solar cells. This is due to the fact that anatase is more photoactive than rutile, has a larger specific surface area and is more prone to hydroxylation .
The crystalline or amorphous nature of titania nanotubes varies with the conditions under which they are synthesized and annealed. The conversion of amorphous titania nanotubes into their preferred anatase forms has been observed under annealing temperatures around 280 ˚C, while the anatase crystals are known to transform into rutile in the 650–900 ˚C range . During this annealing process, however, there is a high risk of the titania nanotubes losing their morphology in a process dubbed “collapsing”. Since collapsing damages the solar energy harvesting pathways inside titania nanotubes, a more efficient DSSC design requires a greater understanding of the conditions under which the risk of collapsing is minimized.
This study explored the effects of two annealing parameters on the crystalline nature of titania nanotubes. First, the effects of varying peak annealing temperatures on anatase crystallization were studied by comparing samples annealed under temperatures between 180 and 285 ˚C. Second, in order to explore the relationship between sustained temperatures and crystallinity, the duration of isothermal scanning at each peak annealing temperature was varied between 5 and 180 minutes. Furthermore, special attention was given to the morphologies of the nanotubes at various intermediary stages.
Numerous trials, each with a unique combination of these two parameters, were performed on a differential scanning calorimeter (DSC). Afterwards, the heat flux data from each trial was analyzed, while the crystalline structure of the annealed nanotubes was studied using X-ray diffraction (XRD) and scanning electron microscopy (SEM).
2 Materials and Methods
In preparation for anodization, titanium foils (25 × 35 × 0.127 mm3, 99.7 % purity, Sigma Aldrich) were cleaned and sonicated for 15 minutes with isopronanol. After being washed with deionized (DI) water, the foils were placed at the anode. Using a two electrode configuration with the as-prepared titanium foils as the anode and platinum as the cathode, the titania nanotubes were synthesized via anodization, as shown schematically in Figure 2.
The anodization medium was a 0.05 M potassium fluoride (KF) solution of 1 % H2O and 99 % ethylene glycol by volume, and the anodization voltage was 60 V. The tubes of sample code IO59 were anodized for 360 minutes, while the tubes of sample code Dalmau were anodized for 480 minutes. While the duration of anodization was not among the target synthesis parameters, it was adjusted to yield nanotubes of varying morphologies.
After anodization, the samples were cleaned again first with DI water and then with isopropanol. Finally, they were sonicated for 30 minutes for a thorough surface cleansing of titania debris and solution residue, and dried in air flow. This entire procedure was conducted at a surrounding room temperature of approximately 20 ˚C. At the conclusion of the preparation process, the specimens were geometrically characterized by the roughness factor and porosity of the tubes.
Before the annealing took place, the differential scanning calorimeter (DSC, TAC 7/DX, Perkin Elmer) was calibrated using tin and indium reference samples. In both cases, the observed melting point and latent heat came within ± 5 ˚C and ± 10 % of the reported values, respectively. After confirming the accuracy of the DSC, the as-prepared amorphous titania nanotubes were annealed at varying temperatures. The samples were heated at either 0.5 ˚C/min or 10 ˚C/min to peak temperatures between 180˚C and 350 ˚C and were placed under isothermal scanning at the peak temperature for periods varying from 10 minutes to 90 minutes. The annealing conditions of each sample pan are provided in Tables 1 and 2.
|IO59 (0.5 ˚C/min, 90 min)|
|Peak Temperature (˚C)||275||280||250||270||285|
Table 1. The annealing conditions of IO59 samples The IO59 samples were heated to varying peak temperatures between 250˚C and 285˚C. To maintain consistency, both the heating rate (0.5˚C/min) and the duration of isothermal scanning (90 minutes) were kept constant.
|Peak Temperature (˚C)||200||220||240||180||285||270||260||250|
|Scanning Period (min)||90||5||180||120||30|
Table 2. The annealing conditions of Dalmau samples. The first half of the Dalmau samples were heated to varying peak temperatures in order to determine the lower crystallization temperature bound. Then, the second half was heated to 250˚C and isothermally scanned for periods ranging from 5 minutes to 180 minutes. For all samples, the heating rate was kept constant at 10˚C/min.
After annealing, the superficial morphology of the annealed nanotubes was examined with a scanning electron microscope (SEM, Hitachi S4800). Crystalline characterization was carried out by X-ray diffraction (XRD, X’Pert Pro, PANalytical/Phillips) using Cu Kα incident radiation (λ=1.54 nm) with a tube voltage of 45 kV and a current of 40 mA. The scan range was 2θ: 23˚ – 85˚with a scanning rate of 0.2˚/min.
|Length(μm)||Outer radius(nm)||Top WallThickness(nm)|
|Bottom Wall Thickness(nm)||Roughness Factor||Porosity(%)|
Table 3. Structural parameters of amorphous titania nanotubes. The values above were calculated using SEM images such as those shown in Figure 3.
3 Results and Discussion
3.1 Basic Morphology
After anodization, the morphology of the nanotubes was studied through SEM imaging. Figure 3 depicts images of the tubes from sample pan IO59 02 (annealed at 280 ˚C for 90 minutes), taken from the side, top, and bottom.
It is evident that the tubes grow in highly ordered arrays, parallel to each other and perpendicular to a base plane. From these images, the structural parameters of the nanotubes were estimated. A summary of these estimates is provided in Table 3. The roughness factor and porosity were calculated under the approximate assumption that the wall thickness increases linearly from top to bottom, and that the tubes exhibit 2-D hexagonal close packing. (The linearity of thickess ) In this case, if the outer radius is r0, the top inner radius is r1, the bottom inner radius is r2, and the length is l, the following equations can be derived:
A larger substrate surface area enables more dye to be coated on the substrate. Since the dye allows the electrons to absorb a wider range of the electromagnetic spectrum and initiates the electron transport chain, a larger amount of it is indicative of more efficient DSSCs . Therefore, the remarkably high roughness factor exhibited by these nanotubes provides another significant advantage over other nanoscale structures.
3.2 The Peak Annealing Temperature and Crystallization
Annealing is a gradual process rather than an instantaneous reaction, and so it is inevitable that both destabilization and crystallization occur simultaneously as the sample is heated.
Since a higher degree of crystallization implies higher-performing DSSCs, finding the ideal balance point between these two counteracting reactions is crucial to optimal solar energy harvesting. In light of this, the relationship between the peak annealing temperature and crystallization was investigated. First, in order to prevent double counting and overlapping, the Kα2 radiation spectrum was removed from the original data using an XRD analysis software. Then, the 2θ peaks and their heights were studied and compared to the reference data for pure anatase, as shown in Table 4.
|Reference 2θ (˚)||Relative Intensity||(h, k , l)|
|25.308||100||(1, 0, 1)|
|36.95||8.8||(1, 0, 3)|
|37.79||27.5||(0, 0, 4)|
|38.571||10.6||(1, 1, 2)|
|48.046||45||(2, 0, 0)|
|53.884||31||(1, 0, 5)|
|55.071||31.2||(2, 1, 1)|
|62.115||6.1||(2, 1, 3)|
|62.69||26.1||(2, 0, 4)|
|68.754||12.1||(1, 1, 6)|
|70.301||13.4||(2, 2, 0)|
|74.048||1.4||(1, 0, 7)|
|75.051||21.4||(2, 1, 5)|
|76.049||5.9||(3, 0, 1)|
|80.731||1.2||(0, 8, 8)|
|82.17||1.5||(3, 0, 3)|
|82.684||11.5||(2, 2, 4)|
|83.171||4.8||(3, 1, 2)|
Table 4. The XRD peak data for pure anatase. θ represents the scattering angle, while the numbers h, k, and l, also known as the Miller indices, define a crystalline plane. Each index represents the number of sections into which the crystalline plane divides each axis of the unit cell.
Figure 4 depicts the X-ray diffraction spectra of sample pans IO59 02 and 03, annealed at 280˚C and 250˚C, respectively. It is clear that the two spectra share their characteristic 2θ peak values, suggesting that they also share the same crystalline structure. To further confirm that the 250˚ spectrum is that of anatase, the deviation of the 2θ value for each peak was calculated, as shown in Table 5.
|Reference 2θ (˚)||Observed 2θ (˚)||Deviation(Observed – Expected, ˚)|
Table 5. A comparison between the reference peaks and the peaks observed at 250 ˚C. The small deviations further support the observation made in Figure 4 that the two samples are essentially composed of the same crystalline structure.
While previous reports on the crystallization temperature of anatase were at or above 280˚C [19, 23, 26, 27, 28], these comparisons clearly show that anatase may be formed under annealing temperatures as low as 250˚C.
Based on this initial result, the lower temperature bound for anatase formation was determined by comparing the XRD spectra for sample pans Dalmau 01 (annealed at 200˚C), 02 (220˚C), and 03 (240˚C), which are shown in Figure 5. Even though there are signs of a broad, nascent peak around 25˚, which corresponds to the major peak (h, k, l) = (1, 0, 1) of anatase, these spectra deviate significantly from the clear peaks exhibited at temperatures above 250˚C.
However, it is also important to note the discrepancy between the heights of the major (1, 0, 1) peak in samples IO59 02 and IO59 03. While pan 02, annealed at 280˚C, exhibits a major peak height of 1494, that of pan 03 is only 841, indicating incomplete crystallization. Thus, it is clear that even though 250˚C is above the crystallization threshold, crystallization at this temperature does not proceed as fast as it does at higher temperatures. This result led to the investigation into the next parameter: the duration of isothermal scanning.
3.3 The Duration of Isothermal Scanning and Crystallization
Once anatase crystallization was confirmed to occur at temperatures as low as 250˚C, the duration of isothermal scanning at this temperature was varied, with a focus on the relationship between this duration and the degree of crystallization. Figure 6 shows the XRD spectra for sample pans Dalmau 8, 9, 10, 11, and 12, a collection of nanotubes annealed at 250˚C for 5, 30, 90, 120, and 180 minutes, respectively.
Upon examination, it becomes clear that annealing for 30 minutes or less did not have a significant impact on the crystalline nature of the nanotubes. In addition, the relative peak heights of pans 10, 11, and 12 reveal that prolonged isothermal scanning indeed leads to a higher degree of crystallization. Therefore, provided that the nanotubes undergo a sufficiently long isothermal scanning period, it may be possible to produce anatase with enough crystalline purity to be directly applied to DSSCs.
3.4 Intermediary Changes
While the crystallization of titania nanotubes has been extensively studied in numerous papers, not much attention has been directed towards changes in their morphologies. Figure 7 shows an image of the nanotubes after being heated to 270˚C at 0.5˚C/min and annealed for 90 minutes.
The unique geometric pattern observed in the bases of the tubes provide insight into the collapsing process. It has already been reported that when more than a certain amount of thermal or electrical destabilization is applied to the nanotubes, the surface morphology of the tubes breaks down . Figure 7 shows that such external collapsing starts with the condensing of surface crystals into several discrete blocks.
In addition, in varying the duration of isothermal scanning, small nuclei of anatase crystals were observed on the walls of the titania nanotubes after 30 minutes at 250˚C (Fig. 8).
This confirms that anatase crystallization has multiple origins. Based on Figure 8, a possible mechanism for crystallization is proposed: spontaneous nucleation happens at various locations on the tube above a certain threshold temperature, and the crystalline structure spreads simultaneously from these multiple origins.
3.5 Future Studies
Future studies may aim to prove the conjectures provided in this study by producing nearly pure anatase through prolonged annealing at 250˚C. In addition, the physical nature of the titania
nanotubes as well as the exact functions of the two intermediary structures identified in section 3.4 may be further studied through other methods, such as transmission electron microscopy. Finally, nanotubes annealed under varying conditions may be employed to DSSCs for efficiency measurements and comparisons.
Titanate nanotubes were synthesized via electrochemical anodization, and their structural parameters were evaluated using SEM imaging. Investigation into the various parameters related to the formation of anatase revealed that crystallization occurs at a previously unreported temperature of 250˚C, and prolonged periods of isothermal scanning resulted in more complete crystallization. Evidence of nucleation was observed with an annealing period of 30 minutes, giving rise to a proposed crystallization mechanism centered on this process.
Their economic advantage and viable efficiency make dye-sensitized solar cells a crucial part of the novel technologies currently under development with the goal of securing clean, reliable energy sources. The optimization of DSSCs may go a long way in improving the efficiency of solar energy harvesting and contribute to the search for renewable energy.
- M. Gratzel and B. O’Reagan: A low-cost, high-efficiency solar cell based on dye-sensitized colloidal TiO2 films, Nature 353 (1991) 737-740.
- C. Ratanatawanate, A. Chyao, and K. Balkus: S-Nitrosocysteine-Decorated PbS QDs/TiO2 Nanotubes for Enhanced Production of Singlet Oxygen, JACS 133 (2011) 3492-7.
- J. Bisquert, D. Cahen, G. Hodes, S. Ruhle, and A. Zaban: Physical Chemical Principles of Photovoltaic Conversion with Nanoparticulate, Mesoporous Dye-Sensitized Solar Cells, J. Phys. Chem. B 108 (2004) 8106-18.
- H. Rensmo, S. Lunell, and H. Siegbahn: Absorption and electrochemical properties of ruthenium(II) dyes, studied by semiempirical quantum chemical calculations J. of Photochemistry and Photobiology A: Chemistry 114 (1998) 117-24.
- Sol Ideas Technology Department. (2009). How Dye-Sensitized Solar Cells Work [Image], Retrieved July 19, 2011, from: http://www.solideas.com/solrcell/howworks.html
- American Chemical Society: Ultrathin, Dye-sensitized Solar Cells Called Most Efficient To Date, ScienceDaily, 20 September 2006.
- O. Schultz, R. Preu, S. Glunz, and A. Mette: Silicon Solar Cells with Screen-Printed Front Side Metallization Exceeding 19% Efficiency, 22nd European Photovoltaic Solar Energy Conference and Exhibition, 9 July 2007.
- K. Tennakone, G. Kumara, I. Kottegoda, and V. Perera: An efficient dye-sensitized photoelectrochemical solar cell made from oxides of tin and zinc, Chemical Community 15 (1999).
- S. Ferrere, A. Zaban, and B. Gregg: Dye Sensitization of Nanocrystalline Tin Oxide by Perylene Derivatives J. Phys. Chem. B. 101 (1997) 4490-3.
- K. Sayama, H. Suguhara, and H. Arakawa: Photoelectrochemical Properties of a Porous Nb2O5 Electrode Sensitized by a Ruthenium Dye, Chemical Matter 10 (1998) 3825.
- M. Paulose, K. Shankar, O. Varghese, G. Mor, and C. Grimes: Backside illuminated dye-sensitized solar cells based on titania nanotube array electrodes, Journal of Physics D: Applied Physics 39 (2006) 2498-503.
- Y. Zhang, Z. Xie, and J. Wang: Highly efficient dye-sensitized solar cells of thick meso- porous titania films derived from supramolecular templating IOP Publishing: Nanotech- nology 20 (2009) 505602. 16
- T. Peng, A. Hasegawa, J. Qiu, and K. Hirao: Fabrication of Titania Tubules with High Surface Area and Well-Developed Mesostructural Walls by Surfactant-Mediated Tem- plating Method, Chemistry of Materials 15 (2003) 2011-6.
- M. Adachi, Y. Murata, I. Okada, and S. Yoshikawa: Formation of Titania Nanotubes and Applications for Dye-Sensitized Solar Cells, Journal of the Electrochemical Society 8 (2003) G488-93.
- B. Poudel, W. Wang, C. Dames, J. Huang, S. Kunwar, D. Wang, D. Banerjee, G. Chen, and Z. Ren: Formation of crystallized titania nanotubes and their transformation into nanowires Institute of Physics Publishing: Nanotechnology, 16 (2005) 1935-40.
- M. Wei, Y. Konishi, H. Zhou, H. Sugihara, and H. Arakawa: Utilization of Titanate Nan- otubes as an Electrode Material in Dye-Sensitized Solar Cells J. of the Electrochemical Society 153 (2006) A1232-6.
- Y. Zhu, H. Li, Y. Koltypin, Y. Hacohen, and A. Gedanken: Sonochemical synthesis of titania whiskers and nanotubes Chem. Commun. 24 (2001) 2616-27.
- S. Chu, K. Wada, S. Inoue, and S. Todoroki: Synthesis and Characterization of Titania Nanostructures on Glass by Al Anodization and Sol-Gel Process Chem. Mater., 14 (2002) 266-72.
- M. Neupane, I. Park, T. Bae, H. Yi, F. Watari, M. Lee: Synthesis and Morphology of TiO2 Nanotubes by Anodic Oxidation Using Surfactant Based Fluorinated Electrolyte, J. Electrochem. Soc. 158 (2011) C242-C245.
- I. Flores, J. Nei de Freitas, C. Longo, M. De Paoli, H. Winnischofer, and A. Nogueira: Dye-sensitized solar cells based on TiO2 nanotubes and a solid-state electrolyte J. Pho- tochemistry and Photobiology A: Chemistry 189 (2007) 153-160.
- V. Mahajan, M. Misra, K. Raja, and S. Mohapatra: Self-organized TiO2 nanotubular arrays for photoelectrochemical hydrogen generation: effect of crystallization and defect structures J. Phys. D: Appl. Phys. 41 (2008) 125307.
- S. Albu, A. Ghicov, S. Aldabergenova, P. Drechsel, D. LeClere, G. Thompson, J. Macak, and P. Schmuki: Formation of Double-Walled TiO2 Nanotubes and Robust Anatase Membranes, Advanced Materials 20 (2008) 4135-9.
- O. Varghese, D. Gong, M. Paulose, C. Grimes, and E. Dickey: Crystalization and high- temperature structural stability of titanium oxide nanotube arrays, J. Mater. Res. 18 (2003) 156-165.
- S. Rohani and A. El-Ruby: Synthesis of Titania Nanotube Arrays by Anodization, De- partment of Chemical and Biochemical Engineering, University of Western Ontario. 17
- J. Zhao, X. Wang, R. Chen, and L. Li: Fabrication of titanium oxide nanotube arrays by anodic oxidation, Solid State Communications 134 (2005) 705-10.
- K. Yanagisawa, Y. Yamamoto, Q. Feng, and N. Yamasaki: Formation mechanism of fine anatase crsytals from amorphous titania under hydrothermal conditions, Journal of Materials Research 13 (1998) 825-9.
- G. Mor, K. Shankar, M. Paulose, O. Varghese, and C. Grimes: Use of Highly-Ordered TiO2 Nanotube Arrays in Dye-Sensitized Solar Cells, Nano Letters 6 (2006) 215-8.
- T. Stergiopoulos, A. Ghicov, V. Likodimos, D. Tsoukleris, J. Kunze, P. Schmuki, and P. Falaras: Dye-sensitized solar cells based on thick highly ordered TiO2 nanotubes pro- duced by controlled anodic oxidation in non-aqueous electrolytic media IOP Publishing: Nanotechnology 19 (2008) 235602.