Maria J. Sampaioa,
Luisa M. Pastrana-Martíneza,
Adrián M. T. Silva*a,
Josephus G. Buijnstersb,
Changseok Hanc,
Cláudia G. Silvaa,
Sónia A. C. Carabineiroa,
Dionysios D. Dionysiou*c and
Joaquim L. Fariaa
aLCM – Laboratory of Catalysis and Materials – Associate Laboratory LSRE-LCM, Faculdade de Engenharia, Universidade do Porto, Rua Dr Roberto Frias s/n, 4200-465 Porto, Portugal. E-mail: adrian@fe.up.pt
bDepartment of Precision and Microsystems Engineering, Research Group of Micro and Nano Engineering, Delft University of Technology, Mekelweg 2, 2628 CD, Delft, The Netherlands
cEnvironmental Engineering and Science Program, University of Cincinnati, Cincinnati, OH 45221, USA. E-mail: dionysios.d.dionysiou@uc.edu
First published on 23rd June 2015
Titanium dioxide (TiO2) has been under intensive investigation for photocatalytic degradation of cyanobacterial toxins. In order to develop more efficient photocatalysts, TiO2 and oxidized nanodiamonds (NDox) were combined as a composite catalyst (NDox–TiO2), which was tested in the photocatalytic oxidation of microcystin-LA (MC-LA), a cyanotoxin frequently found in freshwater. NDox–TiO2 and neat TiO2 photocatalysts were prepared by a liquid phase deposition method. A wide variety of analytical techniques, including physical adsorption of nitrogen, X-ray diffraction (XRD), UV-Vis and IR diffuse reflectance spectroscopies (DRUV-Vis and DRIFT), X-ray photoelectron spectroscopy (XPS), scanning electron microscopy (SEM) and transmission electron microscopy (TEM), were used to characterize the materials. The performance of the photocatalysts was studied under both simulated solar and visible light. The kinetic results show remarkable efficiency for the NDox–TiO2 composite under simulated solar light irradiation with a synergistic factor of more than 15 relative to neat TiO2, while negligible photocatalytic activity was observed for the degradation of MC-LA when NDox–TiO2 was used under visible light illumination due to the wide band gap of the composite material. The photocatalytic efficiency of NDox–TiO2 was ascribed to the good dispersion of both phases in the composite material, facilitating the possible electronic interaction at the heterojunction interface between NDox and TiO2.
Recently, nanodiamonds (NDs) have been used as alternative carbon materials towards the development of efficient photocatalysts upon combination with TiO2.12 Diamonds (carbon with sp3 hybridization) are potentially applicable candidates for composite synthesis due to their unique chemical, structural, mechanical, biological and optical properties.13,14 In fact, these materials are being increasingly used in a wide variety of applications (medicine, biotechnology, catalysis, among others). The cost of NDs is mostly dictated by the technique employed in their production. Nowadays, they are obtained on a large scale, by relatively inexpensive detonation processes of carbon-containing explosives.15 NDs produced by detonation of carbon explosive materials are commonly defined as diamonds with small sizes (typically 4–5 nm) that have high specific surface areas (around 300 m2 g−1) allowing to create large amounts of reactive chemical surface groups.16
To the best of our knowledge, there is only one study12 where some of us describe the application of NDs–TiO2 composites for photocatalytic water treatment. In such report a significant improvement in the photocatalytic activity was observed for degradation of an organic pharmaceutical water pollutant (diphenhydramine) when NDs were combined with TiO2, using an optimal content of oxidized NDs (i.e., 15 wt% of NDox).
Microcystins (MCs), the most widespread cyanotoxins present in diverse aqueous environments, cause water quality problems for fisheries, aquaculture and sanitary hazard for human and animal.17–19 The most common and studied MC found in freshwaters is the variant MC-LR, which has been widely employed as model cyanotoxin in photocatalysis.20–22 Other commonly detected variants in freshwaters include MC-YR, MC-RR and MC-LA,23 which are less studied. In this regard, the present work is focused on the degradation of MC-LA (the microcystin most similar to MC-LR) in water under simulated solar light and employing a NDox–TiO2 composite photocatalyst with 15 wt% NDox content.
X-ray diffraction (XRD) analysis was performed in a PANalytical X'Pert MPD equipped with an X'Celerator detector and secondary monochromator (Cu Kα λ = 0.154 nm, 50 kV, 40 mA; data recorded at a 0.017° step size, 100 s per step). Rietveld refinement with a PowderCell software was applied for identification of the crystallographic phases and to calculate the crystallite size. The morphology of the materials was analyzed by scanning electron microscopy (SEM) using a Philips XL 30 ESEM-FEG apparatus. Transmission electron microscopy (TEM) was performed in a Philips CM20 equipment.
Diffuse reflectance infrared Fourier transform (DRIFT) spectroscopic analysis of the materials was performed on a Nicolet 510P FTIR Spectrometer, converting the interferograms to equivalent absorption units in the Kubelka–Munk scale. Diffuse reflectance UV-Vis spectra (DRUV-Vis) of the materials were measured on a JASCO V-560 UV-Vis spectrophotometer equipped with an integrating sphere. The spectra were recorded in diffuse reflectance mode and transformed to equivalent absorption Kubelka–Munk units. The band gap of the photocatalyst was determined using the Kubelka–Munk units as a function of the energy.
X-ray photoelectron spectroscopy (XPS) was performed in a Kratos AXIS Ultra HSA apparatus using a monochromatic Al Kα X-ray source (1486.7 eV), operating at 15 kV (90 W), in FAT mode (Fixed Analyser Transmission), with a pass energy of 40 eV for regions of interest and 80 eV for survey. Multi-region spectra were recorded at C 1s, O 1s and Ti 2p photoelectron peaks.
The concentration of MC-LA was evaluated by High Performance Liquid Chromatography (HPLC) using an Agilent 1100 Series apparatus equipped with a photodiode-array detector (PDA) set at 238 nm. The stationary phase consisted in a C18 Discovery HS (Supelco) column (150 mm × 2.1 mm, 5 μm particle size) working at 313 K with a flow rate of 0.2 mL min−1 and an injection volume of 50 μL. The method of analysis is described elsewhere.21
The X-ray powder diffractograms of TiO2, NDox and NDox–TiO2 composite are shown in Fig. 1. The presence of crystalline anatase TiO2 is confirmed by XRD measurements, where the peaks at 25.1° and 47.6° correspond to the lattice plane of (101) and (004), respectively. The anatase crystallite size, determined by Rietveld refinement from the XRD data is around 8 nm. The XRD analysis of NDox shows two signals at 44.0° and 75.3°, resulting from the diamond reflection (111) and (220) planes, respectively.26 The crystallite dimension found for NDox was 5 nm. The diffractogram of NDox–TiO2 composite was similar to that obtained in the case of TiO2 (Fig. 1) with a small contribution of (111) reflection of NDox. No significant change on the crystallite size of NDox was observed for NDox–TiO2 composite. Nevertheless, the crystallite dimension of TiO2 slightly increases to 11 nm, which could be related with the lower SBET of the composite in comparison with the respective nominal value. This increase in the crystallite TiO2 dimensions has been previously reported for oxidized carbon nanotube–TiO2 composites, being attributed to the competition of the TiO2 precursor species to the oxidized sites at the surface of the carbon material during the synthesis process, affecting the size of the TiO2 crystallites.27
DRIFT analysis was performed for investigating possible interactions between TiO2 and NDox, and the results are depicted in Fig. 2. The DRIFT spectrum recorded for TiO2 shows a broad band located between 2500 and 3800 cm−1 ascribed to the stretching vibrations of hydrogen-bonded surface water molecules and hydroxyl groups. These observations are confirmed by the presence of a weak band centered at 1640 cm−1 associated to the bending vibration of water molecules as well as the presence of Ti–OH bonds.28,29 A typical band of TiO2 materials around 970 cm−1 due to the Ti–O vibration was also observed,28 while the sharp peak at 1420 cm−1 can be assigned to the lattice vibrations of TiO2.30,31
The DRIFT spectrum of NDox shows characteristic bands of surface oxidized carbon materials, with a broad band around 3400 cm−1 that can be usually attributed to vibration of C–OH groups of carboxylic acids and phenols, and to adsorbed water as confirmed by the presence of a OH bending mode at 1640 cm−1. A small shoulder at ca. 3000 cm−1 attributed to C–H stretching is also observed. The band at 1800 cm−1 can be assigned to the vibration of CO bonds in carboxylic acids, carboxylic anhydrides, quinones and lactones, while the band peaking at 1470 cm−1 corresponds to CC aromatic bending.25 The bands in the range 1000 to 1300 cm−1 are characteristic of C–O stretching vibrations from anhydrides and lactones. As expected, previously reported temperature programmed desorption (TPD) analysis of NDox revealed that carboxylic acid groups are not present at the surface of NDox since the oxidation treatment was carried out at a temperature (703 K) higher than the decomposition temperature of these functional groups (around 503–649 K).12,22,32 Therefore, carboxylic anhydrides, lactones, phenols and carbonyl/quinone are the main oxygen groups present on the surface of NDox.
Regarding the DRIFT spectrum of the NDox–TiO2 composite, the broadening of the intense band centered at 970 cm−1 (characteristic of Ti–O) can be assigned to Ti–O–C bonds,33 thus suggesting the creation of a heterojunction at the interface between NDox and TiO2. The bands at ca. 1640 and 1470 cm−1 became more intense in the spectrum of the composite, resulting from the additive contribution of the vibration bands observed for both TiO2 and NDox phases. Yet, the NDox band at 1800 cm−1 corresponding to carbonyl group vibration practically disappeared, while the broad band at 3400 cm−1 became less intense, indicating the existence of an interphase interaction between NDox and TiO2.
XPS deconvolution spectra of the of Ti 2p, C 1s and O 1s binding regions of TiO2 and the NDox–TiO2 composite were performed in order to identify possible chemical interactions between the elements in the near surface range. The XPS spectrum of Ti 2p region for NDox–TiO2 (Fig. 3a) shows two peaks at binding energies of 458.8 and 464.5 eV, corresponding respectively to Ti 2p3/2 and Ti 2p1/2 spin-orbital splitting photoelectrons (5.7 eV) in the Ti4+ chemical state,34,35 as also observed for neat TiO2. In contrast, the spectra of O 1s core level for TiO2 (Fig. 3b) and NDox–TiO2 (Fig. 3c) were slightly different (insight of Fig. 3c is shown for better comparison). While neat TiO2 exhibits two peaks centered at 529.9 eV and 531.1 eV in the O 1s spectrum (Fig. 3b), which are due to the lattice of Ti–O bond and –OH groups, respectively, the deconvolution fitting for the NDox–TiO2 composite shows three constituents centered at 529.9, 531.1 and 532.7 eV (Fig. 3c), that could be attributed to the presence of Ti–O, –OH and carbon phase (C–O) bonds, respectively.4,36 In the case of the C 1s core level spectra (not shown), carbon bonds were identified for neat TiO2, probably resulting from some organic species that were not removed by the thermal treatment (N2, 473 K). Even so, it was possible to conclude that C–O (287.0 eV)37,38 is the main bond in the composite (as expected due to the known carbon sp3 hybridization in diamonds), followed by C–C/C–H (285.0 eV) and CO (288.9 eV),37,38 while other type of chemical bonds cannot be confirmed by this characterization technique. These results are in agreement with TPD analysis,12 were phenol or ether (C–OH, C–O–C; 286.3 eV (ref. 36)) and quinone (CO) groups were detected, while carboxylic acids (–COOH), carboxylic anhydrides (–C(O)2O) and lactones (–COO) were not identified in the composite.
Fig. 3 XPS peak deconvolution of the binding energy regions: (a) Ti 2p for TiO2 and NDox–TiO2, (b) O 1s for TiO2, (c) O 1s for NDox–TiO2 (inset: comparison of O 1s spectra). |
Several authors39,40 have reported that detonated NDs when dispersed in liquid medium induce strong particle aggregation, attributing the phenomenon to the harsh conditions in the detonation chamber that leads to the creation of dandling bonds on the ND surface. The free electron surfaces can cooperate via intermolecular surface forces such as van der Waals and hydrogen bonding creating covalent bonds between the primary particles.39 This finding combined with the air oxidation treatment of NDs, which has been proved to increase the functional groups at the surface of NDs,12 can promote the interphase interaction between NDox and TiO2.
Fig. 4 shows the DRUV-Vis spectra of the photocatalysts. A characteristic absorption sharp edge rising at 400 nm was observed for TiO2. For this material, the valence band (VB) is composed of O 2p states and the conduction band (CB) is composed of Ti 3d states. The 330 nm absorption band is attributed to the charge transfer from O 2p to Ti 3d.41 In case of the NDox–TiO2 spectrum it was noticed that the TiO2 band is blue-shifted by the incorporation of NDox and a decrease in the shoulder tail in the UV range being also observed. These observations can be attributed to the creation of site defects at the TiO2 crystalline structure, which can be linked with the carbon phase.
Fig. 4 DRUV-Vis spectra of TiO2 and NDox–TiO2, and plot of Kubelka–Munk units as a function of the light energy (inset). |
The transformed Kubelka–Munk function was plotted as a function of the energy of light (Fig. 4 inset) for the determination of the bandgap of the semiconductor materials. The band gaps for TiO2 and NDox–TiO2 were estimated as 3.26 and 3.41 eV, respectively. Quantum size and electronic interphase effects may constitute the major contributions for the increase of the bandgap energy of the composite material. Also, the high band gap of ND (5.5 eV (ref. 42 and 43)) when mixed with TiO2 may contribute to the increase of the band gap energy of the resulting composite material.
The representative SEM micrographs of TiO2, NDox, and NDox–TiO2 are shown in Fig. 5a, c and e, respectively (higher magnifications are shown as insets). The structural morphology of neat TiO2 (Fig. 5a) significantly differs from the NDox–TiO2 composite (Fig. 5e). The neat TiO2 presents spherical-like particles aggregated to form larger (micron-size) particles and consisting of small anatase crystallites as determined by XRD analysis. It is known that when NDs are functionalized with oxygen-containing functional groups, the microscale morphology consists of small nanoparticles forming porous aggregates (as observed in Fig. 5c) due to hydrogen bonding and van der Waals forces between the particles.42 NDox–TiO2 shows a convoluted morphology, where particles of very small dimensions are grouped forming composite clusters (Fig. 5e).
Fig. 5 SEM (a, c and e) and TEM (b, d and f) images of neat TiO2, NDox, and NDox–TiO2, respectively. |
TEM micrographs of TiO2, NDox and NDox–TiO2 are shown in Fig. 5b, d and f, respectively. The micrograph of bare TiO2 (Fig. 5b) suggests that the already observed TiO2 spherical particles are quite dense.12 The TEM micrographs of NDox (Fig. 5d) reveal that the aggregates are formed by small (∼5 nm) nanoparticles of NDs confirming the value obtained by XRD analysis. The TEM micrograph of the NDox–TiO2 composite reveals a homogeneous distribution of TiO2 and NDox particles forming small clusters of okenite-like structures (Fig. 5f).
Fig. 6 (a) Normalized concentration of MC-LA (C/C0) using the different materials under simulated solar light. (b) MC-LA removal under dark conditions and both visible and solar light irradiation. |
The apparent first order constants (kapp), determined by non-linear curve fitting to the experimental data, indicate that no appreciable MC-LA degradation when the pristine NDox sample was used (kapp = 9.8 × 10−4 min−1), while the introduction of NDox to the TiO2 matrix produced a significant increase in the photocatalytic activity. In fact, NDox–TiO2 (kapp = 4.4 × 10−1 min−1) exhibits a kapp more than fifteen times larger than that for neat TiO2 (kapp = 2.8 × 10−2 min−1).
Superior surface area and narrow band gap are normally required attributes for enhanced efficiency of photocatalysts. However, these factors are not likely to be the main features contributing to the high efficiency observed for the NDox–TiO2 composite when compared to the sole NDox and TiO2 phases. Despite the high band gap of diamond, Nebel44 inferred that this material could be a good candidate for photocatalytic applications due to its inert chemistry and stability in liquids. The functional groups on the surface of air oxidized NDs, such as oxygen and hydroxyl species, can create a surface dipole layer that could affect their interfacial electron affinity. In other works, researchers even demonstrated that the presence of hydrogen bonds on H-terminated diamond surfaces could shift the conduction band above the vacuum energy level and the electrons can be directly transferred to the surrounding medium via surface states.45,46
In view of above findings, the good dispersion of NDox and TiO2 phases during the synthesis of the composite material can promote the creation of strong electronic interphase interactions, as inferred from its characterization. Therefore, beneficial synergies and cooperative effects between the semiconductor and the carbon phase are expected, where electrons could be transferred from the NDox conduction band to the TiO2 conduction band as well as the holes transferred from the TiO2 valence band to the NDox valence band under light excitation. Thus, the separation of the photogenerated charges is facilitated, decreasing the occurrence of electron–hole recombination and leading to an increase of the NDox–TiO2 photocatalytic activity.
Experiments under dark conditions were also performed to establish the adsorption equilibrium between MC-LA and the photocatalysts. As shown in Fig. 6b, no significant adsorption was observed after 30 min, with a decrease of only 8%, 1% and 3% of the initial MC-LA concentration being observed using neat TiO2, NDox, and NDox–TiO2, respectively. In addition, it is important to refer that 30 min was proved to be sufficient to reach the adsorption equilibrium, as previously confirmed by performing dark adsorption runs for 5 h.
The results also show that the removal of MC-LA using TiO2 or NDox–TiO2 is negligible (Fig. 6b) under visible light illumination (30 min), which is attributed to the large band gap energies of these materials44 and thus low light absorption in the visible spectral range. No conversion was observed, even after 5 h of reaction (not shown), the apparent MC-LA removal being essentially related to adsorption. The wide band gap of NDs (∼5 eV)42–44 may also explain the inactiveness of NDox when irradiated at λ > 400 nm. TiO2-based composites using other carbon materials, such as graphene oxide (GO–TiO2) and carbon nanotubes (CNT–TiO2), have shown improved efficiency under simulated solar light and visible light radiation.47,48 A synergistic effect was found between the carbon and the TiO2 phase, but low band gap energies were determined for these materials (2.95 eV and 3.04 eV, respectively). In particular, the lower band gap energy of GO–TiO2 resulted in the best performance under visible light illumination. In contrast, negligible activity was observed when NDox–TiO2 was tested under visible light due to the higher band gap of this composite (3.41 eV). Yet, improved efficiency was observed under solar irradiation (which includes UV-photoexcitation). Therefore, the low efficiency of NDox–TiO2 under only visible illumination seems to be related with the optical properties of both TiO2 and NDox materials. However, the NDox–TiO2 composite shows to be highly efficient under real field conditions (i.e., solar light irradiation), of major relevance for solar applications. Thus, it will be now of interest to explore the combination of NDs with other TiO2 materials presenting different crystalline phases and particle sizes.
This journal is © The Royal Society of Chemistry 2015 |