Deniz Altunoz Erdogana,
Touradj Soloukib and
Emrah Ozensoy*a
aDepartment of Chemistry, Bilkent University, 06800, Ankara, Turkey. E-mail: ozensoy@fen.bilkent.edu.tr; Fax: +90-312-266-4068; Tel: +90-312-290-2121
bBaylor University, Department of Chemistry & Biochemistry, Waco, TX 76798, USA. E-mail: touradj_solouki@baylor.edu; Fax: +1-254-710-4272; Tel: +1-254-710-2678
First published on 20th May 2015
A simple sol–gel synthesis method is presented for the production of micron-sized buckyball-like TiO2 architectures using naturally occurring Lycopodium clavatum (LC) spores as biotemplates. We demonstrate that by simply altering the calcination temperature and titanium(IV) isopropoxide:ethanol volume ratio, the crystal structure and surface composition of the buckyball-like TiO2 overlayer can be readily fine-tuned. After the removal of the biological scaffold, the unique surface morphology and pore structure of the LC biotemplate can be successfully transferred to the inorganic TiO2 overlayer. We also utilize photocatalytic degradation of Rhodamine B dye samples to demonstrate the photocatalytic functionality of these micron-sized buckyball-like TiO2 architectures. Moreover, we show that the photocatalytic activity of TiO2 overlayers can be modified in a controlled manner by varying the relative surface coverages of anatase and rutile domains. These results open a potential gateway for the synthesis of a variety of bio-inspired materials with unique surface properties and shapes comprised of reducible metal oxides, metal sulfides, mixed-metal oxides, and/or perovskites.
Sol–gel chemistry can be utilized to develop simple, versatile, and inexpensive synthetic methods to grow inorganic (e.g., metal oxide) thin film coatings on biological scaffolds.1,3,6 By conserving the morphology/geometry of the underlying biological architectures, ordered overlayers with a wide range of thicknesses, ranging from a few nanometers to micrometers, can be manufactured.1,3,6 For instance, to replicate the fine-structural details of a biotemplate, first, an inorganic precursor can be brought into contact with the self-assembled entities on the surface of the template. After the deposition/loading process, an organic–inorganic hybrid material can be obtained. Finally, this process can be followed by the removal of the biotemplate and transfer of the morphology/shape/geometry of the nascent biological scaffold to the inorganic overlayer structure. Calcination process is often employed to remove the organic template.12 The use of such a thermal process for elimination of the biotemplate also offers an opportunity to fine-tune the structural properties of the inorganic film. It should be emphasized that during the thermal treatment process, undesirable deformation of the organic–inorganic hybrid material may also occur.13–15 Thus, for an optimal biotemplating architecture, material properties of the organic (biological) and inorganic components should be structurally compatible. Furthermore, an ideal biotemplate should be inexpensive, mechanically and chemically adaptable, non-toxic, and abundant in nature. Based on the aforementioned requirements and considerations, botanical material platforms are excellent candidates for biotemplating. In particular, pollens and spores of various plants reveal moderately robust outer layers;16 these biomaterials often display unique surface morphologies and pore structures, in the nanometer to micrometer range and can be readily utilized for biotemplating.
Titanium dioxide (TiO2) has been widely used in the field of photocatalysis due to its high activity, chemical stability, environmentally friendly nature, and low-cost.8,17,18 Here, we utilize Lycopodium clavatum (LC) spores19,20 as efficient biotemplates, decorated with TiO2 as an inorganic overlayer, and demonstrate the synthesis of a hierarchically-ordered novel material platform (i.e., micron-sized buckyball-like TiO2 architectures). LC is a commercially available, affordable, abundant, non-toxic, and versatile biomaterial (e.g., it is commonly used in latent fingerprint development agents for forensic science applications).21 In this study, we show that inorganic thin films such as TiO2 can be coated on a LC biotemplate to mimic the pore structure and geometry of the underlying substrate. Furthermore, structural properties (e.g., type and relative abundance of various polymorphs) of the TiO2 overlayer can be fine-tuned via a simple calcination process, during the removal of the LC biotemplate. In addition, we demonstrate that by varying the synthesis parameters employed in the sol–gel process as well as the calcination protocol, functional properties of the bio-inspired final product can be controlled. By utilizing the TiO2–LC hierarchical architectures in the photocatalytic degradation of Rhodamine B samples under UV illumination, we establish functional versatilities of these bio-inspired products. Current study opens a potential gateway for the synthesis of a large variety of future material platforms comprised of reducible metal oxide (e.g., TiO2, CeO2, ZrO2, ZnO, Fe2O3, Fe3O4 etc.), metal sulfide (e.g., CdS, PbS etc.), mixed-metal oxide (e.g., TiO2–Al2O3, TiO2–ZrO2, CeO2–ZrO2, TiO2–ZnO etc.), and/or perovskite (e.g., LaCoO3, LaMnO3 etc.) systems with unprecedented surface/electronic/photonic/structural properties. These new materials could potentially play important roles in catalysis, energy, biology, medicine, and nanotechnology applications. The current study is also relevant to metal oxide growth mechanisms in biological templates and natural bio-mineralization processes.22
The crystallographic structures of the samples were analyzed by using a X-ray diffractometer (Rigaku, Japan) equipped with a Miniflex goniometer where a monochromatic X-ray source (CuKα, λ = 0.15405 nm, 30 kV, 15 mA) was utilized. For the XRD measurements, samples were scanned within a 2θ range of 10–60° with a scan rate of 0.02° s−1. Diffraction patterns were assigned using Joint Committee on Powder Diffraction Standards (JCPDS) cards supplied by the International Centre for Diffraction Database (ICDD).
Raman spectroscopic measurements were performed on a LabRAM HR800 spectrometer (Horiba Jobin Yvon, Japan) equipped with a Nd:YAG laser (λ = 532.1 nm) operated with a power of 20 mW and an integrated confocal Olympus BX41 microscope. Prior to conducting Raman measurements, the powder samples were mechanically dispersed onto a single-crystal Si substrate. The Raman spectrometer was regularly calibrated by adjusting the zero-order position of the grating and using the reference Si Raman shift at 520.7 cm−1. Raman spectra were recorded in the range of 100–1500 cm−1 with a spectral resolution of 4 cm−1.
Fig. 1 SEM image and the corresponding EDX spectrum of an uncoated commercial Lycopodium clavatum (LC) biotemplate sample. Dashed region depicts the region where the EDX spectrum was acquired. |
The biopolymer network on the surface of the LC biotemplate is capable of forming complexes with metal-alkoxide functionalities.27,28 Thus, a simple sol–gel synthetic approach can be employed to deposit TIP on the outer shell of LC spores. By controlling the hydrolysis–condensation kinetics of TIP and the subsequent formation of the TiO2 overlayer, it is feasible to coat the LC exine capsule without any major changes in the size/geometry, pore structure, and morphology of the biotemplate. Fig. 2 contains SEM and EDX data from analyses of a TIP-coated LC spore (i.e., LcTi(2:1)-25) after aging at room temperature (i.e., before calcination). As will be discussed in the next sections, this particular TIP loading revealed the highest photocatalytic activity. General characteristics and morphology of the samples with other TIP/EtOH ratios (data not shown) were rather similar. SEM images given in Fig. 2a and b show that a relatively uniform TiOx/Ti(OiPr)4 overlayer was deposited on the LC spore without the existence of neither extremely large (>100 nm) agglomerates of Ti-containing domains nor large patches of uncoated/bare LC biotemplate. This is also visible in the EDX line scan of the Ti signal across the TIP-coated LC spore (Fig. 2c) as well as the Ti and O elemental EDX mapping results given in Fig. 2e and f, respectively.
It is worth noting that the thickness of the TiOx/Ti(OiPr)4 overlayer can be conveniently modified by varying the amount of the Ti-precursor and/or immersion time of the biotemplate in the TIP/EtOH solution. For example, when the immersion time was decreased below 30 min, uncoated regions on the surface of the biotemplate were detected via EDX measurements (data not shown). On the other hand, for longer immersion times (e.g., >60 min) local aggregations/clusters of TiOx/Ti(OiPr)4 overlayer were observed in SEM images (data not shown). Thus, an optimal immersion duration of 30 min was utilized in the synthesis protocol. To demonstrate the influence of the amount of Ti on the photocatalytic performance and chemical composition of the overlayer, precursor solutions with different TIP loadings (i.e., LcTi(3:2), LcTi(2:1), and LcTi(3:1)) were used in the material synthesis. It was observed that by increasing TIP loading, LC surfaces became coarser at the nanometer scale and the thickness of the partitions or walls separating the polygon-shaped hierarchical cavities increased from ∼350 nm to ∼750 nm (e.g., compare SEM images in Fig. 1 and 2b); while lower TIP loadings led to 2D islands/patches (that is existence of uncoated biotemplate domains). Thus, TIP loadings were varied between LcTi(3:1)–LcTi(3:2).
Calcination process was employed to transform the amorphous TiOx/Ti(OiPr)4 overlayer, obtained after room temperature TIP/EtOH deposition and successive aging, into various ordered polymorphs of TiO2 and remove the underlying LC biotemplate. To prevent major structural deformation of the biotemplate, before the formation of the micron-sized buckyball-like TiO2 architectures, calcination parameters (i.e., annealing ramp rate, calcination temperature, and external gaseous environment) were carefully optimized.
SEM images in Fig. 3 show that when LC spores are coated with a TiOx/Ti(OiPr)4 overlayer (i.e., for LcTi(3:2) or LcTi(2:1)) and calcined at elevated temperatures (e.g., 800–900 °C), micron-scale structural details of the pollen substrate are still preserved. Dimensions of the micron-sized buckyball-like TiO2 architectures after calcination are also comparable to dimensions of the original LC spores. However, upon calcination, TiO2 overlayer inside the cavities (or pockets) displayed sporadic cracks and holes (possibly due to the mechanical stress inflicted on the TiO2 film during the high temperature treatment and removal of the LC substrate (Fig. 3a–c, f and g)). For example, the SEM image shown in Fig. 3f, corresponding to the LcTi(3:2)-800 sample, confirms that upon the high-temperature calcination, inner polysaccharide core of the LC structure as well as the exine capsule comprising of sporopollenin are eliminated to a large extent (though not entirely), revealing a hollow TiO2 buckyball-shell. This is also spectroscopically confirmed by EDX analysis of the LcTi(2:1)-800 sample as shown in Fig. 3d and e. Fig. 3d and e show that C and O EDX signals originating from the core of the biotemplate drastically diminish upon calcination. Furthermore, Ti signal due to the TiO2 buckyball-shell becomes significantly prominent. Fig. 3e also shows an EDX line scan of the elemental Ti signal illustrating that the Ti signal coincides with the corrugations on the LC spore which is consistent with the presence of a rather uniform TiO2 coating on the LcTi(2:1)-800 sample surface. It is worth noting that the typical specific surface area of the LcTi(2:1)-800 sample obtained via Brunauer–Emmett–Teller (BET) method was ∼7.5 m2 g−1.
The SEM and EDX data for LcTi(3:2)-400 sample (Fig. 4) demonstrate that low calcination temperatures such as 400 °C are insufficient to remove the polysaccharide core of the LC system. Fig. 4a shows two different regions: one located inside the inner core of the hollow capsule (marked with an empty red circle, corresponding to the underlying intact biotemplate substrate below the TiO2 overlayer) and a second region corresponding to the outer surface of the hollow capsule (marked with an empty blue circle, on the periphery). EDX spectrum corresponding to the inner red zone (i.e., inside the hollow capsule) is dominated by C and O signals without a significant contribution from the Ti signal; conversely, the EDX spectrum for the outer blue zone (i.e., outermost surface) is dominated by Ti signals. These results are consistent with the thermogravimetric analysis (TGA) measurements in the literature,29 which reported that while TiO2 revealed a negligible gravimetric loss within 25–800 °C, uncoated Lycopodium spores underwent almost 60% weight loss within 250–450 °C due to thermal decomposition/degradation/oxidation processes.
Fig. 4 SEM image (a) of the LcTi(3:2)-400 sample; corresponding EDX spectra representing (b) the interior seed part of the LC bio-template and (c) external TiO2 buckyball-like shell. |
Calcination process used for the removal of the LC biotemplate after the formation of the inorganic overlayer can be utilized as a tool to fine-tune the chemical composition and the crystallographic structure of the outermost layer. Such compositional properties were also characterized in detail via XRD as a function of the calcination temperature (as well as TIP/EtOH ratio (Fig. 5)). XRD patterns revealed that anatase (ICDD card no.: 00-021-1272) signals (designated as “A” in Fig. 5) were the only prominent diffraction signals at T ≤ 500 °C and became sharper with increasing temperatures suggesting ordering and increasing average particle size. At T ≥ 600 °C, rutile (designated as “R” in Fig. 5) diffraction signals (ICDD card no.: 00-021-1276) started to appear and dominate the XRD patterns at elevated temperatures. When calcination temperatures below 400 °C were utilized, samples were found to contain mostly amorphous/disordered TiO2/TiOx phases.
The average crystallite sizes of the anatase and rutile phases were calculated based on the main XRD peaks corresponding to anatase (101) and rutile (110) signals using Scherrer equation as a function of precursor loading and calcination temperature (Fig. 6). As can be noted from the stacked column chart in Fig. 6, the crystallinity of TiO2 domains typically increase with increasing calcination temperature. For all samples analyzed, anatase phase had a characteristically smaller average crystallite size than the rutile phase. Fig. 6 clearly demonstrates that the extent of crystallization depends both on the calcination temperature and precursor loading. It is also apparent that anatase to rutile phase transformation temperatures increase with increasing TIP loading in the initial precursor mixture. Relative mass fractions of anatase versus rutile phases were also calculated via Spurr and Myers approach (Table 1).30 LcTi(2:1)-700 and LcTi(2:1)-800 samples (marked with bold numerals in Table 1, columns two and three), which exhibited two of the highest photocatalytic activity values, revealed a phase composition where anatase and rutile phases had similar mass percentiles (i.e., ∼50% anatase and ∼50% rutile).
Sample | Weight percent of anatase (%) | Weight percent of rutile (%) |
---|---|---|
LcTi(3:2)-400 | 100.0 | 0.0 |
LcTi(3:2)-500 | 100.0 | 0.0 |
LcTi(3:2)-600 | 95.4 | 4.6 |
LcTi(3:2)-700 | 85.9 | 14.1 |
LcTi(3:2)-800 | 77.0 | 23.0 |
LcTi(3:2)-900 | 35.0 | 65.0 |
LcTi(2:1)-400 | 100.0 | 0.0 |
LcTi(2:1)-500 | 100.0 | 0.0 |
LcTi(2:1)-600 | 85.6 | 14.4 |
LcTi(2:1)-700 | 54.5 | 45.5 |
LcTi(2:1)-800 | 41.7 | 58.3 |
LcTi(2:1)-900 | 8.4 | 91.6 |
LcTi(3:1)-400 | 100.0 | 0.0 |
LcTi(3:1)-500 | 100.0 | 0.0 |
LcTi(3:1)-600 | 96.6 | 3.4 |
LcTi(3:1)-700 | 96.1 | 3.9 |
LcTi(3:1)-800 | 91.7 | 8.3 |
LcTi(3:1)-900 | 65.8 | 34.2 |
As a complementary characterization technique, Raman spectroscopy was also employed for the structural analysis of the micron-sized buckyball-like TiO2 architectures as a function of calcination temperature and TIP loading. In general, Raman spectra presented in Fig. 7 were in very good agreement with the XRD data (Fig. 5). It is well-known that anatase phase has six (1A1g, 2B1g and 3Eg) and the rutile phase has five (B1g, multi-proton process, Eg, A1g and B2g) characteristic Raman active modes.31 Due to spectral overlap, poorly ordered phases with relatively smaller crystallite sizes, and contributions from phonon bands originating from the polymer network of the residual LC biotemplate (i.e., features marked with ♣ symbol in Fig. 7), only four prominent anatase peaks at 142 cm−1 (Eg), 393 cm−1 (B1g), 514 cm−1 (A1g), and 638 cm−1 (Eg) and two rutile peaks at 446 cm−1 (Eg) and 609 cm−1 (A1g) were discernible (Fig. 7). Consistent with the XRD results shown in Fig. 5, Raman data also suggested that lower calcination temperatures favored anatase phase, while the rutile content increased with increasing calcination temperatures.
Fig. 8a and b illustrates the relative decolorization performances of the photocatalysts. Before the UVA illumination, RhB dye was kept in contact with the photocatalyst under dark conditions for 1 h. It was observed that after the adsorption–desorption equilibrium was reached (under dark conditions), only a small amount of RhB dye (∼0.9–2.0% of the initial RhB concentration) was adsorbed on the photocatalyst surface. Thus, photodegradation results were corrected using this “dark” control experiment. As a second control experiment, we also checked the self-degradation of RhB under UV illumination in the absence of any photocatalysts. Hence, RhB self-degradation was also taken into account for reporting the final photodegradation results.
From the data presented in Fig. 8a it can be concluded that among all of the investigated micron-sized buckyball-like TiO2 architectures, LcTi(2:1)-800 sample has the highest k′ value. Fig. 8c illustrates that the photocatalytic performance of the LcTi(2:1) sample increases with increasing calcination temperature and reaches its highest value at 800 °C. After this optimum temperature, photocatalytic activity falls in a drastic manner. Presented data from XRD and Raman spectroscopy in Fig. 5–7, suggest that there is an optimum anatase/rutile weight ratio (viz., 1:1) that leads to an optimum photocatalytic performance. This optimum phase composition is reached at a calcination temperature of 800 °C; at higher temperatures than 800 °C, TiO2 domains become enriched in rutile and lose their activities.
Fig. 8d demonstrates the effect of TIP precursor loading for the photocatalysts calcined at the optimum calcination temperature of 800 °C. It is shown that for low TIP/EtOH ratios (i.e., LcTi(3:2)-800), there is simply not enough active sites. For the intermediate TIP/EtOH value, the photocatalytic activity is maximized and for higher TIP/EtOH ratios, photocatalytic activity starts to decline. Drop in the photocatalytic activity at higher TIP loadings can presumably be attributed to sintering of the TiO2 domains and deviations in the relative anatase:rutile compositional ratio from the optimal value.
Fig. 9 represents the % photodegradation efficiency values (i.e. ((C0 − C)/C0) × 100) obtained after 330 min of irradiation for each run. The red data point in Fig. 9 corresponds to the performance of the fresh catalyst whose behaviour was presented earlier (Fig. 8b) while the blue data points represent the successive runs where the fresh catalyst was re-used multiple times. As can be seen from Fig. 9, catalytic performance of the catalyst is conserved to a great extent after multiple runs without a significant indication of catalytic deactivation.
This journal is © The Royal Society of Chemistry 2015 |