carbon

Home

Search

Collections

Journals

About

Contact us

My IOPscience

Layer-by-layer assembly of TiO2 nanowire/carbon nanotube films and characterization of their photocatalytic activity

This article has been downloaded from IOPscience. Please scroll down to see the full text article. 2011 Nanotechnology 22 195701 (http://iopscience.iop.org/0957-4484/22/19/195701) View the table of contents for this issue, or go to the journal homepage for more

Download details: IP Address: 160.114.21.16 The article was downloaded on 29/03/2011 at 12:02

Please note that terms and conditions apply.

IOP PUBLISHING

NANOTECHNOLOGY

Nanotechnology 22 (2011) 195701 (9pp)

doi:10.1088/0957-4484/22/19/195701

Layer-by-layer assembly of TiO2 nanowire/carbon nanotube films and characterization of their photocatalytic activity ´ Mária Darányi1 , Tamás Csesznok1 , Akos Kukovecz1, 1 1 Zoltán Kónya , Imre Kiricsi , Pulickel M Ajayan2 and Robert Vajtai2,3 1

Department of Applied and Environmental Chemistry, University of Szeged, H-6720 Szeged, Rerrich B. tér 1, Hungary 2 Department of Mechanical Engineering and Materials Science, Rice University, 6100 Main Street, Houston, TX 77005, USA E-mail: [email protected]

Received 22 November 2010, in final form 31 January 2011 Published 23 March 2011 Online at stacks.iop.org/Nano/22/195701 Abstract We report on the layer-by-layer (LbL) formation of TiO2 –MWNT–TiO2 coatings on quartz with either trititanate derived TiO2 nanowires or Degussa P25 as the photocatalytically active material. The optimized deposition sequence is discussed in detail and the morphology of the prepared coatings is analyzed by SEM and XRD. The heterogeneous photocatalytic performance of the coatings was tested in the methyl orange oxidation reaction. The apparent first order rate constant fell in the 0.01–0.20 h−1 range over a 2.5 × 2.5 cm2 film depending on the type and the thickness of the titanate coating. Building a multiwall carbon nanotube layer into the middle of the layer improved the photocatalytic activity for each material for all of the studied thicknesses. P25 based films performed 2–5 times better than TiO2 nanowire films; however, the pores in the P25 based films were largely blocked because the isotropic P25 nanoparticles form closely packed layers by themselves and even more so with the comparably sized multiwall carbon nanotubes. Therefore, films derived from titanate nanowires appear to be more suitable for use as multifunctional, photocatalytically active filtration media. S Online supplementary data available from stacks.iop.org/Nano/22/195701/mmedia

Specialized filtration processes are gradually replacing traditional thermodynamic separation methods in certain applications in today’s energy conscious chemical industry. A remarkably rapidly evolving field is that of photocatalytic membranes capable of oxidizing filtered contaminants and thus maintaining a constant permeability over time [1–3] and providing an important physical representation of heterogeneous photocatalysis [4]. In a series of studies the possibilities of H2 production by photocatalytic water dissociation on strontium titanate [5] and iron oxide [6] surfaces were investigated by the Somorjai group. Photocatalytic water splitting remains a hot

topic and a potential application field for TiO2 nanotubes [7, 8] and nanostructures [9] along with other nanomaterials such as vanadium oxide [10] or iron oxide [11]. A successful membrane candidate should possess (i) an open pore system, (ii) tunable pore size distribution and (iii) controllable photocatalytic activity. This last prerequisite effectively rules out several potential network forming materials like silica and alumina nanowires, polymer and ceramic fibers and one-dimensional metallic nanorods [12]. Trititanate (Nax H2−x Ti3 O7 ·n H2 O) nanotubes [13] and nanowires [14] synthesized in a simple, readily scalable hydrothermal process with yields close to 100% have received considerable attention recently [15–17]. Although they are not photocatalytically

3 Author to whom any correspondence should be addressed.

0957-4484/11/195701+09$33.00

1

© 2011 IOP Publishing Ltd Printed in the UK & the USA

Nanotechnology 22 (2011) 195701

M Darányi et al

Figure 1. Illustration of the sequence used for the preparation of a photocatalytic TiONW–MWNT–TiONW coating. Sequential steps in the top row from left to right: surfactant-assisted deposition of TiONWs onto the quartz substrate, removal of surfactant by UV illumination and annealing in air at 300 ◦ C, deposition of the carbon nanotube layer, surfactant-assisted deposition of the second TiONW layer. The sequence is concluded by another UV + thermal cleaning step (middle right). The sandwich catalyst is finally placed in the reaction vessel containing the methyl orange solution (large central drawing). The reactor is illuminated by evenly distributed UV light from above. The reaction scheme on the left shows the widely accepted mechanism of TiO2 based photooxidation in aqueous media. Gray stained cylinders, white cylinders and hexagonal network cylinders represent surfactant-covered TiONWs, cleaned TiONWs and MWCNTs, respectively. (This figure is in colour only in the electronic version)

active in their as-synthesized form, they are readily transformed into active catalysts either by doping [18] or by thermal conversion into anatase TiO2 [19, 20]. Composites of one-dimensional titanium oxide nanostructures and carbon nanotubes have been investigated recently as, e.g., fibers [21], fuel cell catalyst supports [22] and H2 storage materials [23]. Several authors reported that carbon nanotubes (CNTs) improve the photocatalytic efficiency of titanium oxides. This effect is believed to originate either from the favorable adsorption properties of CNTs or their ability to trap electrons and thus reduce the electron–hole recombination rate [24]. Charge transfer processes involving TiO2 were also investigated by Somorjai [25] from a catalytic point of view. Bouazza et al have suggested that CNTs may also act as heat sinks in CNT–TiO2 hybrid catalysts tested in propene photooxidation [26]. For a recent review of the topic we refer to Woan et al [27]. The controlled deposition of nanostructures is crucial for the successful preparation of networks with tunable pore characteristics [28]. In our earlier publication we discussed the optimization of trititanate nanowire layers on a glass surface [29] and also investigated the pore size engineering of multiwall carbon nanotube buckypapers [30]. In this work we are creating titania–MWNT–titania sandwich coatings on quartz plates using layer-by-layer (LbL) deposition, combining and synergistically enhancing these effects. The LbL method is a versatile bottom-up method offering superior structural control for the preparation of multilayer coatings [31] and was successfully applied previously for the preparation of,

e.g., Ag/titanate nanotube films [32] and photocatalytic TiO2 coatings [33]. Figure 1 is a quick overview of the work reported in this contribution. Our primary goal is to investigate the factors governing the photocatalytic activity of titania–MWNT–titania sandwich multilayers to be used as advanced, multifunctional coating materials [34]. The sandwich structure is favored instead of a simple mixture of TiO2 and MWCNTs because it offers more control over the morphology of the resulting film. Moreover, even though the LbL coatings studied here were all prepared on quartz supports, our recent experiments have shown that it is possible to reproduce the sandwich structure in the form of a self-supporting standalone filter membrane. Self-supporting membranes made of simple TiONW–MWCNT mixtures exhibited inferior mechanical properties under similar conditions.

1. Results and discussion Figure 2 depicts the characteristic SEM images and XRD profiles of the starting trititanate and titanate nanomaterials used in this study; representative SEM images of the reference photocatalyst Degussa P25 TiO2 , consisting of anatase and rutile in approximately 3:1 ratio, as well as those of MWNT samples are included in supplementary figure 2 (available at stacks.iop.org/Nano/22/195701/mmedia). The P25 particles were spherical with an average particle diameter of 30.7 ± 8.7 nm. The specific surface area of P25 was 44.2 m2 g−1 . The 2


M Darányi et al

Figure 2. Characteristic SEM images of the starting materials: (a) as-synthesized trititanate nanowires, (b) titanate nanowires annealed at 600 ◦ C. The right panel depicts the XRD profiles of the materials used as follows: (a) multiwall carbon nanotubes, (b) Degussa P25, (c) as-synthesized trititanate nanowires, (d) nanowires annealed at 600 ◦ C.

multiwall carbon nanotubes used here measured 25–30 nm in diameter and were typically over 2 μm in length. The specific surface areas of the MWNTs, trititanate nanotubes and titanate nanowires were 186 m2 g−1 , 80 m2 g−1 , and 15 m2 g−1 , respectively. Degussa P25 is the de facto reference material for photocatalytic experiments today [35]. Its XRD profile features the characteristic anatase reflections at 25.3◦ (101), 37.8◦ (004), 48.0◦ (200), 53.8◦ (105), 54.9◦ (211) and 62.5◦ (204) (JCPDS card no. 21-1272). Moreover, weak rutile peaks at 27.3◦ (100), 36.1◦ (101) 41.2◦ (111) and 56.6◦ (211) are also identifiable (JCPDS card no. 21-1276). MWNT XRD profiles exhibit the graphitic reflections at 26.2◦ (002) and 44◦ (002). Peaks related to Fe and Co particles are absent from the XRD profiles, indicating that the purified carbon nanotubes were free of transition metal catalyst impurities. The absence of leftover catalyst particles was also confirmed by energy dispersive xray spectroscopy performed in the SEM instrument. It is well known that the alkaline hydrothermal conversion of anatase yields trititanate nanotubes first by oriented crystal growth [16]. Carrying the reaction on in a rotating autoclave results in the oriented self-organization of nanotubes into trititanate nanowires, as depicted in figure 2(a). The average diameter of these nanowires is 63 ± 16 nm, their length is in the 500–2500 nm range and they possess a trititanate structure as confirmed by the characteristic reflections observable in the XRD profile (figure 2(c)) at 9.9◦ (001), 11.4◦ (200), ¯ and 48.6◦ (020) (JCPDS 26.1◦ (110), 29.6◦ (211), 34.3◦ (312) card no. 41-00192). The morphology is preserved even after calcination at 600 ◦ C (figure 2(b)); however, the heat treatment transforms the trititanate structure mostly into anatase, as evidenced by figure 1(d). More precisely, the XRD profile indicates that the dominant phase is anatase TiO2 and that the annealed material does not contain rutile TiO2 . In addition, the annealed material contains at least one more phase characterized by weak reflections at 14.2◦ , 44.0◦ and 57.8◦ . This phase can be identified as TiO2 -B (JCPDS card no. 46-1238) [36]. Fitting of the figure 2(c) XRD profile indicates that the amount of TiO2 -B is 10.8 ± 6.1% in the sample. TiO2 -B nanotubes have recently been shown to exhibit some photocatalytic activity of their own [37, 38]. The TiO2 nanowires (TiONW) with the figure 2(c) XRD signature were used in all coating experiments. Spin coating is a simple and fast method for creating thin films. We tested its applicability for the preparation of TiONW, TiONW–MWNT and TiONW–MWNT–TiONW

coatings by varying the deposited amount (20, 60 and 100 droplets per layer) and the rotational speed (110, 140 and 170 rpm). The sample was allowed to dry for 10 s between each droplet. Comparative study of the spin and the dip coating processes on the example of the TiONW–MWNT bilayers (supplementary figure 3 available at stacks.iop.org/Nano/22/ 195701/mmedia) showed evidently that the centrifugal force acting on the bottom TiONW layer results in a highly anisotropic morphology, where all titanate nanowires are aligned radially. When the second layer is deposited on the top, carbon nanotubes are filtered by the TiONWs instead of forming their natural random pore system [39]. The resulting ‘hairy’ structure contains radially elongated openings of up to 0.5 μm × 3 μm area and, therefore, it is not suitable for acting as a filter in the direction normal to the deposition plane. On the other hand, the 1D nanoobjects in the bilayer created by dip coating (supplementary figure 3(b) available at stacks.iop.org/Nano/22/195701/mmedia) are deposited in a random fashion; thus useful nanopores in the 50–200 nm range are formed. For multifunctional films only dip coated multilayers were proven to be useful and tested further in this study. An even higher level of control over the morphology of TiO2 nanotube arrays can be achieved by the electrochemical anodization of titanium, as described recently by Wang and Lin [40, 41]. Electrolyte temperature and anodization potential can be used to tailor product structure [42]. Ruthenium dye sensitization of the resulting nanotubes enabled the rational surface engineering of TiO2 to yield solar cells with enhanced performance [43]. On the basis of previous investigations aimed at the controlled formation of titanate nanowire [29] and MWNT layers [44], a complex protocol was devised for the reproducible preparation of multilayer ‘sandwich’ structures. In general, this includes first pre-treating the quartz surface as detailed in section 3 and then depositing the desired titanium oxide and/or carbon nanotube layers in multiple dip coating steps separated by adhesion control (PDDA dip) and washing (water rinse) phases (for a summary of the sequences for deposition steps see table 1). The thickness of the resulting coating can be finely controlled by varying the number of dipping cycles. A280 denotes the UV absorbance of the film at 280 nm. Wu et al [45] have established a method for estimating the thickness of double sided TiO2 coatings on substrates from the UV light A280 absorbance of the film measured at 280 nm 3


M Darányi et al

Table 1. Sequences used for the preparation of the layered photocatalysts discussed in this study. Films were built by traversing in each column from top to bottom and repeating each dipping cycle as many times as defined in the corresponding column. In the finishing step, all samples were cured as described in section 3. In and out denote the vertical dipping and lifting speed of the plate, respectively. P25 and TiONW suspension, in: 20 mm min−1 , out: 10 mm min−1 , dwell time: 480 s, drying: 180 s. Curing under a 80 W Hg lamp for 2 h followed by annealing the sample at 300 ◦ C in air for 1 h. PDDA solution, in: 20 mm min−1 , out: 10 mm min−1 , dwell time: 480 s, drying: 180 s. Washing in distilled water, in: 20 mm min−1 , out: 10 mm min−1 , dwell time: 3 s, drying: 3 min. MWNT suspension, in: 20 mm min−1 , out: 10 mm min−1 , dwell time: 600 s, drying: 300 s. Sample type P25 Sample# Process sequence TiO2 PDDA Washing MWCNT PDDA Washing TiO2 A 280 ‘h ’ calculated thickness (nm)

P25–MWNT–P25

1

2

3

4

5

6

2

3

5

5 1 2

5 1 2

5 1 2

7

3 1 2 2 1 2 1 2 5 1 0.23 0.49 0.62 0.73 2.55 3.72 0.49 16.1 35 43.8 51.5 180 263 35

8

9

5 1 2 2 1 2 2 1.96 139

5 2 5 1 2 2 1 2 5 3.72 0.26 1.11 263 18.5 78.3

A280 α

10

11

TiONW-MWNT-TiONW

12

13

14

5 1 2

5 1 2

5 1 2

15

4 1 2 2 1 2 1 3 5 1 1.28 1.80 3.72 1.06 90.3 127 263 75.1

16

17

18

19

5 1 2 2 1 2 1 1.31 92.7

5 1 2 2 1 2 2 2.04 144

5 1 2 2 1 2 3 2.35 166

5 1 2 2 1 2 5 3.73 264

because the pore formation itself is a random phenomenon occurring as a result of local impurity remnants on the substrate surface. On the other hand, the formation of 50–200 nm diameter nanopores is an intrinsic property of random networks of one-dimensional nanoobjects. Therefore, the pores in the bottom TiONW layer (figure 3(d)) are distributed evenly in the layer and their diameter stays within the range defined by the diameter, the length and the deposited amount of nanowires. The diameter of MWNTs matches that of the P25 particles very well. This is the reason why the P25–MWNT bilayer (figure 3(b)) exhibits a very closely packed structure, whereas its TiONW based counterpart (figure 3(e)) is able to maintain most of the original porosity of the underlying TiONW layer. In this latter case, the thinner and more flexible carbon nanotubes adapt to the structure of the dominant TiONWs. Deposition of the top P25 layer (figure 3(c)) replicates the phenomenon observed for the bottom one, except that now the P25 particles are deposited on a less uniform surface and, therefore, random irregularities are emphasized. Similarly, the top TiONW layer (figure 3(f)) mirrors the appearance of the bottom one: nanoporous openings are present and are uniformly distributed over the coated surface. Summarizing, TiO2 nanowires are more suitable for the preparation of multilayer composite coatings than P25 if the goal is to prepare permeable films with tunable pore system characteristics. Before the photocatalytic experiments two control measurements were carried out. First, a clean uncoated quartz plate was immersed into the reaction mixture and illuminated. Second, a plate coated with P25 was kept in the MO solution in total darkness. Neither of these experiments indicated any MO decomposition activity, thus confirming that MO decomposition is the direct consequence of the photocatalytic activity of the titanium oxide coatings. All coatings exhibited photocatalytic activity in the test reaction to some extent. In figure 4(a) the decomposition of methyl orange as a function of time measured on sample #17 TiONW–MWNT–TiONW trilayer catalyst prepared by dip coating is monitored using

using equation (1): 2h = 2.303

TiONW

(1)

where α = e12 is the absorption coefficient of the film and h (cm) is the thickness of the coating on one side of the substrate. We used this equation for estimating the thickness of all films discussed in this report. This calculation is valid for the single P25 films, as those films are very similar to the original coatings studied in [45]. On the other hand, it certainly implies assumptions when applied to films containing TiO2 nanotubes and multiwall carbon nanotubes. The two most important ones are (i) assuming that the absorption coefficient is the same for all three materials and (ii) assuming that all coatings behave in the same way from the point of view of porosity. Although these assumptions look strong, we have found experimentally that the A280 values correlate well with the thickness of the multilayer coatings measured by SEM; therefore, we suggest extending the validity of equation (1) to the titanate nanowire–multiwall carbon nanotube coatings as well. Usage of this thickness gives a good approximation for the real thickness, which would be very difficult to determine by other methods, e.g. it would need many SEM measurements on the same sample for averaging, while the optical measurement is nondestructive and integrative for the sample size. Representative SEM images corresponding to the individual formation steps of two exemplar sandwich layers—for P25–MWNT–P25 sample #8 ( A280 = 1.958) and for TiONW– MWNT–TiONW sample #17 ( A280 = 2.040)—are presented in figure 3. The morphology of sandwich TiO2 –MWNT–TiO2 coatings prepared by layer-by-layer deposition is governed by the morphology of the titanium oxide nanoparticles (figure 4). The spherical P25 particles form a semi-continuous layer (supplementary figure 2(a) available at stacks.iop.org/Nano/ 22/195701/mmedia). Nanopore size distribution and pore location distribution in this layer are both random, presumably 4


M Darányi et al

Figure 3. SEM images of the formation steps of P25 ((a)–(c)) and TiONW ((d)–(f)) based titanium-oxide–carbon nanotube coatings. The top, middle and bottom rows correspond to the single layer (P25: (a), TiONW: (d)), the bilayer (P25–MWNT: (b), TiONW–MWNT: (e)) and the trilayer (P25–MWNT–P25: (c), TiONW–MWNT–TiONW: (f)) stages, respectively. The coating sequences for these films were identical with those of sample #8 for the P25 series ((a)–(c)) and sample #17 for the TiONW series ((d)–(f)).

Figure 4. (a) UV–vis spectra of methyl orange, showing the progression of the photocatalytic degradation reaction over a TiONW–MWNT–TiONW catalyst (Sample #17, A 280 = 2.040). The curves from top to bottom correspond to measurements taken at start, 1 h, 3 h, 5 h, 7 h and 9 h, respectively. (b) Apparent methyl orange photocatalytic degradation rate constant (kapp ) as a function of the measured absorbance of the film. The open and solid squares denote pure TiONW and TiONW–MWNT–TiONW composite films, respectively. The open and solid triangles denote pure P25 and P25–MWNT–P25 composite films, respectively. The dashed and dotted lines connecting the points indicate linear fits (the parameters of the fits are included in the supplementary materials (available at stacks.iop.org/ Nano/22/195701/mmedia)) for P25 and TiONW based films, respectively. The equivalent total thickness of the films (double the single-sided calculated thickness ‘h ’ from table 1) calculated using equation (1) is given on the top axis.

5


M Darányi et al

have found colloidal anatase prepared from sodium trititanate nanotube suspension to be more photoactive in methyl orange bleaching than P25 at exposure times up to 1.5 h [50], and Mozia et al reported on titanate nanotubes performing better by a factor of 2.5 than P25 in photocatalytic methane production [51]. Concerning the overall positive effect of MWNTs on the activity, our results confirm those of Yu et al [24] and Bouazza et al [26]. At this stage it is not yet possible to unambiguously decide whether the enhancement in our system is due to the favorable adsorption properties or the electron–hole recombination blocking behavior of the carbon nanotubes. However, the recent XPS and band-gap study by Akhavan et al on the photocatalytic activity of carbon nanotube doped sol–gel TiO2 revealed that a high temperature annealing step is necessary for the formation of Ti–C and Ti–O–C bonds that can modify the electron–hole recombination rate to an appreciable extent [52]. Since the coatings studied in the present work were also subjected to high temperature annealing, it seems likely that charge transfer and electron–hole recombination effects do contribute significantly to the improved photocatalytic activity of the MWCNT–TiO2 – MWCNT sandwich structures as well. The most peculiar finding of our study is that in TiONW based films the MWNT enhancement is more pronounced in thick coatings, whereas in P25 based ones the activity improvement is inversely proportional to the total thickness. A possible interpretation for this phenomenon can be offered on a morphological basis, which is not uncommon in the TiO2 photocatalysis literature [53]. TiONW based coatings possess a highly permeable nanopore system where the MO diffusion rate is not a decisive factor (figure 3(f)). Therefore, the synergy between the TiONW active centers and the MWNTs is unbroken even in the thicker films that contain more photocatalytically active material and, consequently, a larger apparent MO degradation rate is measured. On the other hand, the narrow pores of P25 based coatings (see figure 4(b) in particular) can amplify the importance of diffusion within the multilayer film. Thin P25–MWNT–P25 multilayers are more permeable as they contain less material (shorter diffusion channels); thus the whole cross section of the coating works and the overall behavior is governed by the beneficial effect of the carbon nanotubes. However, diffusion in thick P25– MWNT–P25 coatings is largely hindered. Therefore, the photocatalytic reaction takes place primarily in the topmost P25 layer which is shielded from the positive influence of the underlying MWNTs. This is the reason why thick P25 layers perform similarly in the presence and absence of carbon nanotubes.

the UV–vis spectra of the MO solution. It is clear from the lack of peak shifts and/or appearance of new spectral features that the photocatalytic degradation was the only reaction taking place in the cell. The C(t) MO concentrations were calculated from the 466 nm light absorbance of the methyl orange solution using the Lambert–Beer law. Figure 4(b) represents a typical case; similar series were measured for all other coatings studied here. The photocatalytic decomposition of methyl orange can be considered a pseudo first rate reaction [46]. Methyl orange concentration C(t) as a function of time t[h] is given by equation (2): C(t) = C0 e−kapp t (2) where C0 is the initial methyl orange concentration [M] and kapp [h −1 ] is the apparent rate constant of the pseudo first rate kinetics. Rate constants are summarized in figure 4(b). It is interesting to note in the figure that there exists a linear correlation between kapp and the absorbance of the coating. The corresponding fitted parameters of all four catalyst series are given in the supplementary table (available at stacks.iop.org/ Nano/22/195701/mmedia). It should also be noted that the r 2 fit parameters also show the possibility of a nonlinear dependence. Nevertheless, in the studied coating thickness range the linear fit serves as a good guide for eyes and describes the behavior of the system adequately in the resolution limits of the experimental results. A closer inspection of figure 4(b) reveals the following: (i) the photocatalytic degradation of methyl orange is faster on P25 based films than on the corresponding TiONW based ones, (ii) thicker coatings are more active in the reaction than thin ones, (iii) the presence of MWNTs invariably improves the photocatalytic activity (the enhancement factor is between 10 and 130%) and (iv) the effect of MWNTs is more pronounced in thicker TiONW based films and in thinner P25 based films. These findings are somewhat counterintuitive at first since it is highly unlikely that the molecular level kinetics of the MO photooxidation could depend on the thickness of the catalyst film. However, kapp is the measure of the overall performance of the system and, as such, it is determined by several other factors besides the reaction itself. The coating thickness effect can be explained by realizing that thicker films consist of more photocatalytically active nanoparticles accessible for the MO solution through the pore system. This effect is in an interesting competition with the hindrance caused by the reduction of photon flux reaching the lower layers due to the total coating thickness increase. However, more studies are required to decide whether the overall result of the two competing effects can explain the observed linear kapp versus A280 relationships or not. The superior action of P25 agrees with common literature wisdom well. Indeed, P25 was found to outperform titanate-derived TiO2 materials in, e.g., photocatalytic H2 production [47] or methylene blue photobleaching [48]. However, the preferential adsorption of the anionic methyl orange organic dye molecule on the P25 surface as compared to the TiONW surface discussed recently by Bavykin et al may also play a role in the observed differences [49]. It is also interesting to note that in a very recent work Church et al

2. Conclusions This study discussed the creation of TiO2 –MWNT–TiO2 sandwich films using spin coating and layer-by-layer deposition. Sequences for the controlled and reproducible deposition of such films on quartz plates were reported and the morphology of the formed layers was discussed. TiO2 nanowires synthesized from trititanate nanowires are more suitable as building blocks of multilayer coatings than Degussa 6


M Darányi et al

on a Quantachrome NOVA 2200 instrument at −196 ◦ C and evaluated using the multi-point BET method in the 0.05 < prel < 0.3 relative pressure range. Samples were degassed in vacuum at 150 ◦ C for 2 h before the N2 adsorption measurement. UV–vis absorption spectra of the coatings and methyl orange solutions were measured on a Hitachi U-2001 instrument. The emissivity map was measured using an Ocean Optics USB4000 spectrometer and an R400-7-SR fiber optic sampling accessory.

P25 nanoparticles because the pore system of the deposited nanowire network is more uniform and tunable. The coatings were tested as heterogeneous photocatalysts in the degradation of methyl orange. The apparent first order reaction rate constant of the photooxidation as a function of coating type and thickness was measured. The activities of P25 and TiONW based systems were comparable but P25 based films outperformed the nanowire based ones by a factor of 2–5. The presence of multiwall carbon nanotubes was beneficial for the photocatalytic performance of all systems. The MWNT-related enhancement was more pronounced in thicker TiONW and in thinner P25 systems. Photocatalytic activities were explained on the basis of literature results and new morphological insights. On the basis of these results is seems possible to create TiONW–MWNT–TiONW coatings, filters or even selfsupporting films featuring a permeable nanopore network and photocatalytic activity commensurable with Degussa P25 for advanced filter applications.

3.3. Film preparation P25 or titanate nanowire suspensions were prepared by sonicating 0.2 g of the material in 100 cm3 distilled water for 3 min using a 1000 W Hielscher ultrasonic horn. The resulting suspension was centrifuged at 2700 rpm using a 22 cm diameter rotor for 9 min. The supernatant was adjusted to pH 8 using 25 wt% NH4 OH and used freshly in the coating experiments. The multiwall carbon nanotube suspension was prepared by sonicating 0.1 g carboxyl functionalized MWNT and 3 g sodium dodecylbenzenesulfonate (SDBS) in 100 cm3 distilled water for 30 min using a low power (20 W) ultrasonic bath. Spin coating was carried out on a Novocontrol SCV20 unit, dip coating was carried out using a KSV LMX2 instrument. Both coating types were applied to 2.5 cm × 2.5 cm quartz plates. The plates were subjected to the following pre-treatment before the coating experiments: (i) 15 min sonication in ethanol, (ii) 15 min sonication in distilled water, (iii) 30 min sonication in 0.2 M NaOH solution, (iv) rinsing with water, (v) six dipping cycles into 2 wt% hexadecylpyridinium chloride monohydrate (HDPCl) solution (in: 20 mm min−1 , out: 10 mm min−1 , dwell time: 480 s, drying: 180 s) and (vi) two washing cycles in distilled water (in: 20 mm min−1 , out: 10 mm min−1 , dwell time: 3 s, drying: 3 min). A major disadvantage of the LbL method is that the surfactants used in the coating process stick to the deposited species and may interfere with the adhesion of subsequent layers as well as with the photocatalytic performance of the final multilayer structure. In the case of the (TiO2 , MWNT) system the main contaminant is poly(diallyldimethylammonium) chloride (PDDA), a cationic surfactant used for improving the interlayer adhesion characteristics of titanium oxide nanoparticles. We have found it possible to remove PDDA by curing the samples under an 80 W Hg lamp for 2 h followed by annealing the sample at 300 ◦ C in air for 1 h. The decomposition of PDDA was confirmed by IR spectroscopy and SEM (not shown). We established a protocol for removing PDDA in the described manner after each titanium oxide layer deposition. The steps of coating formation are illustrated in figure 1 and defined in table 1.

3. Methods 3.1. Synthesis Multiwall carbon nanotubes (MWNTs) were synthesized by catalytic chemical vapor deposition (CCVD) from a C2 H4 /N2 mixture over a (2.5 wt% Fe, 2.5 wt% Co)/CaCO3 catalyst at 650 ◦ C and purified using oxidative and acidic washing steps as detailed earlier [54]. The nanotubes were annealed at 1000 ◦ C in 100 cm3 min−1 N2 flow for 12 h after purification to equilibrate their surface chemistry [55]. This method is estimated to yield a MWNT sample with over 95% purity [56]. In order to improve their solubility, the nanotubes were functionalized with –COOH groups before use by sonicating 1 g MWNT in 100 cm3 HNO3 solution at 70 ◦ C for 30 min, followed by cooling to room temperature, filtration and washing to pH 7 with distilled water. Trititanate nanowires were synthesized using the low temperature rotating autoclave alkaline hydrothermal route [14]. 20 g anatase TiO2 was mixed into 1000 cm3 10 M NaOH solution at room temperature, then the suspension was transferred into a Teflon lined steel autoclave and kept at 125 ◦ C for 48 h while maintaining a rotational speed of 28 rpm. The product was filtered, washed with distilled water until the filtrate pH reached 7 and dried at 80 ◦ C in air. Finally, the trititanate nanowires were annealed at 600 ◦ C in air for 6 h (denoted as TiONW from now on). The TiO2 reference material Aeroxide® P25 by Evonik Degussa GmbH was purchased through Sigma-Aldrich at >99.5% purity and used without further purification. 3.2. Characterization All materials were characterized by scanning electron microscopy (SEM) using a Hitachi S-4700 Type II FE-SEM. Samples for SEM were coated with a thin (