Synthesis and characterization of gadolinium-doped nanotubular titania for enhanced photocatalysis

Journal of Alloys and Compounds 617 (2014) 756–762

Contents lists available at ScienceDirect

Journal of Alloys and Compounds journal homepage: www.elsevier.com/locate/jalcom

Synthesis and characterization of gadolinium-doped nanotubular titania for enhanced photocatalysis Liang Shi a, Lixin Cao a,b,⇑, Rongjie Gao b, Yanling Zhao b, Huibin Zhang a, Chenghui Xia a a b

College of Chemistry and Chemical Engineering, Ocean University of China, Qingdao 266100, PR China Institute of Materials Science and Engineering, Ocean University of China, Qingdao 266100, PR China

a r t i c l e

i n f o

Article history: Received 20 April 2014 Received in revised form 11 August 2014 Accepted 12 August 2014 Available online 21 August 2014 Keywords: Gadolinium Titania nanotubes Hydrothermal method Photocatalytic activity

a b s t r a c t Gadolinium-doped titanium dioxide nanotubes were fabricated with a facile two-step route. Precursors Gd-doped titania nanoparticles were synthesized by a traditional sol–gel method. Hydrothermal process and acid treatment were employed afterwards, and Gd-doped titania nanotubes were finally obtained after calcination. The nominal doping concentration was expressed by Gd/Ti atomic ratio, ranged from 0% to 5.0%. Both the precursors and nanotubes were characterized by X-ray photoelectron spectra, inductively coupled plasma-atomic emission spectrometry, transmission electron microscopy, scanning electron microscopy, X-ray diffraction, energy dispersive X-ray spectrometer, UV–vis diffusion reflection spectra and N2 absorption–desorption experiment. The photocatalytic activities were investigated using methyl orange as the model pollutant. The results indicated that Gd-doped titania nanotubes with nominal Gd/Ti of 0.5% possessed the optimal photocatalytic activity in our study. Ó 2014 Elsevier B.V. All rights reserved.

1. Introduction Over the past few decades, titanium dioxide nanostructures have attracted much attention due to a wide range of applications in photocatalysis [1,2], hydrogen generation [3], sensors [4] and dye-sensitized solar cells [5]. Compared with titania nanoparticles, nanotubular titania exhibits better performance in photocatalytic activities because of a larger specific surface area and carriers confinement [6]. Many approaches have been applied to prepare nanotubular titania, including anodic oxidation [7,8], template-based method [9] and hydrothermal method [10,11]. Numerous papers focused on the hydrothermal method for its simplicity and lower cost, and the dissolution/recrystallization process [12] was widely accepted by most of researchers. In order to improve the photocatalytic performance, many modification processes were often conducted after nanotubular titania was hydrothermally produced. Unfortunately, this method is lack of homogeneity as the hydrothermal reaction was difficult to control. Sol–gel method has always been preferred by researchers [13,14] synthesizing homogeneous nanoparticles, and the modification process can be simultaneously achieved. ⇑ Corresponding author at: College of Chemistry and Chemical Engineering, Ocean University of China, Qingdao 266100, PR China. Tel.: +86 532 6678 1901. E-mail address: [email protected] (L. Cao). http://dx.doi.org/10.1016/j.jallcom.2014.08.121 0925-8388/Ó 2014 Elsevier B.V. All rights reserved.

Lanthanides ions have the special electronic structure of 4fx5dy [15], which would tend to form complexes with various Lewis bases (i.e., amines, aldehydes, alcohols, thiols, etc.) through interaction between functional groups and f-orbital of the lanthanides [16]. Thus, it could promote adsorption of the organic pollutants on the semiconductor surface. According to other reports, incorporation of lanthanides ions in titania matrix could increase the surface area of photocatalyst and retard the transformation from anatase to rutile phase, which were beneficial to the photocatalytic activity of titania [17]. In recent years, many researchers have focused on rare-earth doped TiO2 nanoparticles and their photocatalytic activities, investigating the different effects on photocatalysis for a diversity of rare-earth elements [18]. Among the lanthanides, gadolinium possessed a half-full 4f stable electronic structure [19], thus the synthesis of gadoliniumdoped titania attracted much attention. Mohamed and coworkers found that Gd-doped titania nanoparticles possessed the minimum grain size and band gap, the largest specific surface area and the best photocatalytic activity compared with other rare earth doped titania synthesized by sol–gel route [14,20]. Guo claimed that Gddoped titania nanocomposites ranked first in the decreasing order of amount of hydroxyl oxygen and adsorbed oxygen which were significant in photocatalytic process, leaving Pr/Eu/Nd/Y doped titania behind [18]. Although Gd-doped TiO2 nanoparticles and other nanostructures such as nanobelts and nanotubes have already been studied

L. Shi et al. / Journal of Alloys and Compounds 617 (2014) 756–762

[21,22], the work using homogeneous Gd-doped titania nanoparticles as raw material for further hydrothermal treatment to fabricate nanotubes has not been done yet. In this paper, Gd-doped titania nanoparticles were first synthesized by sol–gel method as raw material, and then Gd-doped titania nanotubes were fabricated by hydrothermal process followed by calcination. In addition, the photocatalytic activities of these nanostructures including Gd-doped titania nanoparticles and nanotubes were also systematically investigated.

2. Experimental 2.1. Materials Tetrabutyl titanate (Ti(C4H9O)4) and gadolinium nitrate hexahydrate (Gd(NO3)36H2O) were purchased from Aladdin (Shanghai) Reagent Co., Ltd., and ethanol, nitric acid (HNO3), sodium hydroxide (NaOH), hydrochloric acid (HCl) and methyl

757

orange (MO) were purchased from Sinopharm Chemical Reagent Co., Ltd. All regents were all of analytical grade and used as received without further purification. The high purity water used in the experiment was double distilled and then purified with the Milli-Q system. 2.2. Preparation method 2.2.1. Synthesis of Gd-doped titania nanoparticles The synthesis process was conducted following the sol–gel method. Typically, 10.35 mL of tetrabutyl titanate was diluted in ethanol (22.00 mL), and the mixture was kept under continuously magnetically stirring for 30 min, forming a light-yellow solution A. Then 0.80 mL of HNO3 and 2.20 mL of distilled water were diluted in ethanol (22.00 mL), resulting transparent solution B. Desired amount of Gd(NO3)3 was also added to solution B which determined the atomic ratio of Gd/Ti = x% (x = 0, 0.1, 0.2, 0.5, 1.0, 2.0, 5.0). Finally, the solution B was gradually added to solution A under continuously stirring, and Gd-doped titanium sol was obtained after 6 h. The sols were aged for 2 days and dried for 10 h in an oven at 353 K, resulting yellow gels. The dried gel was ground and calcinated at 973 K for 6 h to get the Gd-doped titania nanoparticles, which can be labeled as xGd-TNPs according to nominal Gd/Ti.

Fig. 1. TEM images of (a) 0.5Gd-TNPs, (b) 0Gd-TNTs, (c) 0.5Gd-TNTs, (d) 2.0Gd-TNTs and HRTEM images of nanoparticles in 0Gd-TNTs, (e) rutile, (f) anatase.

758

L. Shi et al. / Journal of Alloys and Compounds 617 (2014) 756–762

Fig. 2. XRD patterns of Gd-doped titania nanoparticles with different nominal content (a) x = 0, (b) x = 0.1, (c) x = 0.2, (d) x = 0.5, (e) x = 1.0, (f) x = 2.0 and (g) x = 5.0.

2.2.2. Fabrication of Gd-doped titania nanotubes Gd-doped titania nanotubes with uniform diameter and length were synthesized via a hydrothermal chemical process [23,24]. Typically, 1.00 g of Gd-TNPs was mixed with 50.00 mL of 10 M NaOH aqueous solution under continuous magnetic stirring for 2 h. Then the mixture was transferred in an autoclave and kept for 24 h in an oven at 423 K. After it was cooled down to ambient temperature, the white sediment was centrifuged and washed with distilled water several times until pH of supernatant approached 7.0. Then the sediment was immersed in 250 mL of hydrochloric acid solution under continuously stirring for 2 h, and centrifuged and washed again. Finally, the sediment was dried for 6 h in an oven at 353 K and ground for further use. In order to investigate the photocatalytic activities of the samples, calcination procedure was needed by treating nanotubes at 673 K for 90 min, thus Gd-doped titania nanotubes were obtained, and labeled as xGd-TNTs 2.3. Characterization X-ray diffraction (XRD) experiments were performed with a BRUKER D8 ADVANCE X-ray diffractometer fitted with Cu Ka radiation over the 2h ranges from 10° to 70°, and the scanning speed was 5°/min. Morphology of the samples was observed using H-7000 transmission electron microscope (TEM). Scanning electron microscope (SEM) inspection and element analysis were conducted using a JEOL6700F equipped with an Oxford INCA energy dispersive X-ray (EDX) analyzer. The actual gadolinium level in the titania matrix was measured by inductive coupled plasma atomic emission spectrometry (ICP-AES, ICAP 6300, Thermo Fisher). X-ray photoelectron spectroscopy (XPS) analysis was performed by an Amicus XPS spectrometer (Kratos Analytical Ltd., England). Specific surface area was measured on an ST-08A apparatus (Beijing Analysis Instruments Co.). The optical absorption spectra and diffuse reflection spectra (DRS) were obtained using an ultraviolet– visible spectrophotometer (Shimadzu UV-2550, Japan) with an integrating sphere. 2.4. Photocatalysis testing Photocatalytic experiments were carried out in OCRS-I photochemical reactor under ambient temperature. A 300 W high pressure mercury lamp was used as light source, which was put in a cylindrical quartz vessel with a recycling water quartz jacket and circulating water. In a general procedure, 100 mg of sample was added

Fig. 3. XRD patterns of Gd-doped titania nanotubes with different nominal content (a) x = 0, (b) x = 0.1, (c) x = 0.2, (d) x = 0.5, (e) x = 1.0, (f) x = 2.0 and (g) x = 5.0.

to 60 mL of methyl orange solution. The mixture was stirred in darkness for 30 min to reach adsorption–desorption equilibrium. Then the light was on and 5 mL aliquot of suspension were taken out every 30 min, 7 in all. After suspensions were centrifuged, supernatant was collected and analyzed to quantify the residual concentration of methyl orange.

3. Results and discussion 3.1. Morphology analysis It can be seen from the representative TEM images in Fig. 1 that Gd-TNPs possess an agglomerate morphology of nanoparticles, while Gd-TNTs show a tubular structure. As shown in Fig. 1(a), 0.5Gd-TNPs exhibits uniform size of about 25–30 nm in spherical shape, and an extremely granular aggregate structure was observed due to calcination treatment. Much attention should be paid to the morphology of 0Gd-TNTs (Fig. 1(b)), where both nanoparticles about 10 nm in diameter and nanotubes several hundred nanometers in length could be observed. The reason for this notable phenomenon is that rutile phase was more stable in the hydrothermal process [25], and thus the nanoparticles were incompletely reacted rutile crystals, as we found in Fig. 1(e). Moreover, it is worthy to note that anatase nanoparticles were also found in 0Gd-TNTs, as shown in Fig. 1(f), which were transformed by imperfect titanate nanosheets caused by lower reactivity. As discussed in the composition section below, nanotubes alone were obtained when Gd/Ti ratio exceeded over 0.5%, and the corresponding nanoparticles were entirely in anatase phase, shown in Fig. 1(c) and (d). However, an ambiguous nanotubular morphology was observed when doping amount was excessively large because

Table 1 Structural and component parameters of both Gd-TNPs and Gd-TNTs. Nominal xGd/Ti (%)

0 0.1 0.2 0.5 1.0 2.0 5.0

Gd-TNPs

Gd-TNTs

Anatase (%)

Crystalline size (nm)

Anatase (%)

Surface area (m2/g)

0 8.9 82.7 94.3 93.5 100 100

55.8 49.1 32.8 28.3 22.4 12.5 11.2

71.2 86.1 90.1 100 100 100 100

170 181 257 281 143 135 118

759

L. Shi et al. / Journal of Alloys and Compounds 617 (2014) 756–762

Fig. 4. Morphology of the 5.0Gd-TNTs on large scale (a) 30k (b) 80k magnification. Element mapping of agglomerate zone composed by massive nanotubes was shown in the bottom below (4k magnification).

of low crystallinity of precursor nanoparticles, as shown in Fig. 1(d). 3.2. Structure and composition analysis Fig. 2 shows XRD patterns of bare and Gd-doped titania nanoparticles calcinated at 973 K for 6 h. No other phase except titania was detected by X-ray diffraction due to low doping amount [26] in our experiment. As shown in Fig. 2(a), pure rutile phase was obtained when bare titania gels underwent a calcination treatment. The intense peaks located at 2h = 27.4°, 36.1°, 39.2°, 41.2°, 44.0°, 54.3°, 56.6°, 62.7°, 64.0°, 69.0° and 69.7° can be indexed to the (1 1 0), (1 0 1), (2 0 0), (1 1 1), (2 1 0), (2 1 1), (2 2 0), (0 0 2), (3 1 0), (3 0 1) and (1 1 2) diffractions lines of rutile (JCPDS card No. 219-1276). Nevertheless, anatase began to form and its transformation to rutile could be suppressed by rare earth elements doping [27], so the proportion of anatase rose with an increasing Gd doping level, as shown in Fig. 2(b)–(g). The diffractions at 2h = 25.3°, 36.9°, 37.8°, 38.6°, 40.0°, 53.9°, 55.1°, 62.7° and 68.8° can be attributed to the (1 0 1), (1 0 3), (0 0 4), (1 1 2), (2 0 0), (1 0 5), (2 1 1), (2 0 4) and (1 1 6) atomic planes of Anatase (JCPDS card No. 21-1272). From XRD patterns of Fig. 2(d)–(g) where x was greater than 0.5, rarely detected rutile and lower crystallinity further demonstrated the prohibition of grain growth and anatase to rutile transformation. The changes in grain size and phase composition of titania nanoparticles can be directly observed in Table 1. Kk Scherrer formula B ¼ b cos using a K factor of 0.93, was h employed to estimate the crystalline size of the as-prepared samples [28], where B is crystalline size (nm), k = 1.54056 Å for Cu Ka radiation, b denotes FWHM (full width at half max) of respective line in radians, h is angular position of the Bragg reflection. The percentages of anatase and rutile were determined by using the integral intensities of (1 0 1) and (1 1 0) peaks for anatase and rutile, respectively, following the equation [29]:

W R ð%Þ ¼

AR  100 0:8844AA þ AR

Fig. 3 shows XRD patterns of Gd-doped titania nanotubes synthesized by hydrothermally treating the corresponding titania nanoparticles, and calcinating at 673 K afterwards. Gd components were still not detected in the samples either. The results indicated that all the as-prepared nanotubes were composed by anatase, though some rutile was also detected, as shown in Fig. 3(a)–(c). The detected rutile components were probably ascribed to unreacted rutile nanoparticles, which agreed with previous reports [30] and observations in Fig. 1. Meanwhile, many researchers found that titania nanotubes could be obtained by pure rutile [31] except for lower kinetics [32], and this can be confirmed by the synthesis of 0Gd-TNTs. Although different Gd-doped titania nanoparticles possessed different crystallinity, a similar composition of nanotubes were obtained when Gd/Ti value over 0.5%, as shown in Fig. 3(d)–(g). The specific surface area of the undoped and Gd-doped titania nanotubes measured by Brunauer–Emmett–Teller (BET) method were listed in the Table 1. From data in Table 1 it can be seen that the specific surface area increased as the Gd doping level went up in the range of 0–0.5%, and this result can be ascribed to the proportion deduction of the rutile phase. On the contrary, the specific surface area dropped again when Gd dopant exceeded 0.5% in the raw materials, caused by poor morphology of nanotubes, which agreed with our explanation above.

3.3. Elemental analysis In order to exclude the influence of Gd-TNPs on account of nanoparticles with rutile phase that were not fully converted to nanotubes and reduced Gd content during transformation, we used

Table 2 Actual Gd/Ti ratios in xGd-TNTs calculated from ICP results. Nominal x (%) Actual Gd/Ti (%)

0.1 0.09

0.2 0.13

0.5 0.28

1.0 0.71

2.0 1.74

5.0 4.72

L. Shi et al. / Journal of Alloys and Compounds 617 (2014) 756–762

600

500

400

300

200

100

0

700

600

Ti 2p3/2

464

462

460

458

456

454

466

464

400

300

200

100

0

Ti 2p3/2

462

O 1s

(f) 5.0Gd-TNTs

458

456

454

O 1s Ti-O

Intensity (a.u.)

O-H

460

Binding Energy (eV)

Ti-O

Intensity (a.u.)

130

Ti 2p1/2

468

Binding Energy (eV)

(e)0.5Gd-TNTs

500

(d)5.0Gd-TNTs

Ti 2p1/2

466

140

Gd 4d

Binding Energy (eV)

Intensity (a.u.)

Intensity (a.u.) 468

150

(b)5.0Gd-TNTs

Binding Energy (eV)

(c)0.5Gd-TNTs

Ti 2p 160

Ti 3s Ti 3p

Ti 3s Ti 3p

(a)5.0Gd-TNTs 700

130

Gd 4d

140 Gd 4d

C 1s

150

C 1s

160

Ti 2s

Intensity (a.u.)

Ti 2s

Intensity (a.u.)

Ti 2p

O 1s

O 1s

760

O-H

535 534 533 532 531 530 529 528 527 526 525

535 534 533 532 531 530 529 528 527 526 525

Binding Energy (eV)

Binding Energy (eV)

Fig. 5. XPS spectra of 0.5Gd-TNTs and 5.0Gd-TNTs (a and b) survey scan (inset: detail of XPS scan of Gd 4d) (c and d) Ti 2p (e and f) O 1s.

SEM combined with element mapping technique to investigate the distribution of Gd element on titania nanotubes. Sample 5.0GdTNTs were selected as representative due to low amount of rutile in the precursor as mentioned above, which could eliminate the influence of Gd-TNPs as much as possible before we conducted the element mapping experiment. As shown in Fig. 4(a) and (b), 5.0Gd-TNTs exhibited interlacing morphology with agglomerates on large scale, indicating a high yield of nanotubular products [33]. Element mapping was employed to mark the Gd element on a bulky region composed by massive nanotubes demonstrated by SEM results, and results revealed a uniform distribution of Gd, as well as Ti and O elements. Interestingly, the gadolinium contents in Gd-TNTs calculated by EDX results were smaller than those of Gd-TNPs, suggesting the gadolinium loss during the nanotubes formation process. As EDX is a semi-quantitative technique in characterization of element content, we employed ICP-AES to further detect the authentic content of gadolinium element. We dissolved our xGdTNTs in hot sulfuric acid first, and then measured the concentration of Gd ions in the liquid, the actual Gd/Ti ratio was calculated according to the standards and shown in Table 2.

According to the results in Table 2, the actual Gd/Ti ratio of all the Gd-TNTs are lower than the nominal value, indicating an incomplete transformation from raw materials to the final samples, which can be ascribed to mass loss in the reactions and transfer procedure. 3.4. XPS analysis Fig. 5(a) presents survey scan spectra of typical sample 0.5GdTNTs, with detail XPS spectrum of gadolinium in the inset, and no obvious Gd 4d peak was found. However, a distinct Gd 4d peak centered at 142.7 eV was observed in excessively modified sample 5.0Gd-TNTs, as shown in Fig. 5(b). Although it cannot be attributed to Gd2O3, whose binding energy is about 1 eV higher than this value, the formation of Gd–O–Ti is plausible according to previous research [34]. It can be seen from Fig. 5(c) and (d) that the peaks at about 458.5 eV and 464.3 eV were the typical 2p3/2 and 2p1/2 peaks of Ti (IV) in anatase, respectively [35]. Furthermore, it seems that there are no differences between the samples, indicating Gd doping had no effects on Ti states. As shown in Fig. 5(e), the single O 1s peak at 529.8 eV can be resolved into two peaks at about

761

L. Shi et al. / Journal of Alloys and Compounds 617 (2014) 756–762

Fig. 6 shows UV–vis absorption spectra of all the calcinated GdTNTs with various Gd/Ti ratio, and no obvious red-shift was observed, in contradiction with Xu et al. [15]. On the contrary, there seemed great blue-shift for Gd-doped titania nanotubes compared with bare TiO2 nanotubes synthesized with the same route. 0Gd-TNTs display a little wider light-absorption range than other samples, as shown in Fig. 6(a), due to increased amount of rutile component, because rutile has a slightly smaller bandgap than anatase by 0.2 eV [37]. In order to quantify the edge shift, we obtained the bandgap (Eg) of the Gd-TNTs by calculation according to Tauc’s equation [38], and the values were 3.45, 3.48, 3.50, 3.51, 3.58, 3.60, and 3.60 eV for 0–5.0GdTNTs, respectively. Therefore, the obvious edge blue-shift was attributed to the increasing bandgap of the Gd-TNTs, which was caused by decreasing crystalline. This hypothesis was in accordance with our XRD results, which Gd dopants affected the crystallization process both in the calcination and hydrothermal treatment. In the previous report by Parida and Sahu [39], rare earth lanthanum doped titania enhanced light absorption in the UV range, a similar phenomenon was found in our case, as Fig. 6(b)–(d) shows. As could be seen from Fig. 6(e), when excessive gadolinium amount was introduced in the titania matrix, Gd-TNTs exhibited a poorer absorption ability due to reduced light penetration of titania component. This change could be more evident when it comes to Fig. 6(f)–(g), and the absorption values were even smaller than those of bare titania nanotubes.

0Gd-TNTs 0.1Gd-TNTs 0.2Gd-TNTs 0.5Gd-TNTs 1.0Gd-TNTs 2.0Gd-TNTs 5.0Gd-TNTs P25

1.0

(a) 0.8

0.6 Percentage of degradation (%)

3.5. Optical properties

of titania, electrons in the valence band (VB) can be activated and promoted to conduction band (CB), leaving holes in the valence band. Subsequently, the electrons and holes migrated to the surface of titania photocatalysts, and were trapped by surface species. The highly oxidizing holes can directly oxidize organic compounds or react with water molecule to form OH radicals. Meanwhile, the reducing electrons can be captured by molecular oxygen absorbed on the surface, generating reactive oxygen species (O2  /HO2 and H2O2), which can finally form OH radicals [42]. The oxidizing potential of the OH radicals is 3.06 V [43], leading organic compounds to be oxidized without selectivity. In order to evaluate the photocatalytic activity of the Gd-TNTs in our experiment, methyl orange was employed as a target pollutant, and photodegradation reaction was conducted under ultra violet illumination. The results revealed an increased photodegradation activity together with Gd/Ti ratio from 0% to 0.5%, reaching about 85% degradation of methyl orange solution, which was comparable with P25, as shown in Fig. 7(a). But this value decreased seriously when 1.0Gd-TNTs sample was investigated, and 2.0GdTNTs was even worse as well as 5.0Gd-TNTs, displaying a similar capability as that of 0Gd-TNTs. The repeatability of photocatalysts were also tested using 0.5Gd-TNTs for example, and the results were shown in inset of Fig. 7(a). The rate constants for the photocatalytic reaction were calculated from the slope of linear fitting curves in Fig. 7(b). Since photocatalytic degradation follows pseudo-first-order kinetics

C/C0

529.6 eV and 531.6 eV [18], where the former can be ascribed to the lattice oxygen and the latter to chemisorbed hydroxyl groups. Interestingly, the peak of hydroxyl groups chemisorbed on titania surface became more obvious when Gd dopants were increased, as shown in Fig. 5(f). Several papers mentioned Ti (IV) diffuses into gadolinium oxides and occupies sites in tetrahedron or octahedron positions, causing charges imbalance and enhanced adsorption of hydroxyl groups [36].

0.4

0.2

3.6. Photocatalytic activities By far, the most active application of titania-based photocatalysts is the photodegradation of various organic pollutants, including dyes [40]. Although the precise photocatalytic mechanism is still under discussion, the OH radical oxidizing process [41] is widely accepted by most researchers. When the titania photocatalysts are irradiated by photons with energy high than the bandgap

0.0 0

f g

40 20 0

0

1

2

3

4

5

Trial Number

30

60

90

120

150

180

-1

1.5

ln (C0/Ct)

Absorption (a.u.)

0.2

60

0 (k=0.00356min ) -1 0.1(k=0.00626min ) -1 0.2(k=0.00847min ) -1 0.5(k=0.00986min ) -1 1.0(k=0.00412min ) -1 2.0(k=0.00322min ) -1 5.0(k=0.00249min )

0.3

b c d a e

Repeatability

80

Irradiation time (min)

2.0

(a) 0Gd-TNTs (b) 0.1Gd-TNTs (c) 0.2Gd-TNTs (d) 0.5Gd-TNTs (e) 1.0Gd-TNTs (f ) 2.0Gd-TNTs (g) 5.0Gd-TNTs

100

1.0

(b)

0.5

0.1

0.0 0 0.0

30

60

90

120

150

180

Irradiation time (min) 200

300

400

500

600

700

Wavelength (nm) Fig. 6. UV–vis absorption spectra of Gd-TNTs calcinated at 673 K with different Gd/ Ti ratios.

Fig. 7. (a) Plots of 20 mg/L methyl orange removal by Gd-TNTs with various Gd/Ti ratios (inset: repeatability of 0.5Gd-TNTs, tested by 6 batches of samples. Trial number 0 represented the initial experiment, and the other were repeated ones). (b) Linear fitting of 20 mg/L methyl orange removal by Gd-TNTs with various Gd/Ti ratios.

L. Shi et al. / Journal of Alloys and Compounds 617 (2014) 756–762

Mass of degraded MO (mg)

3

ln (C0/Ct)

2

18

100

16

80

14

60

12 40 10 20 8 10

20

30

40

50

Percentage of degraded MO (%)

762

Acknowledgement

MO=10mg/L(*) MO=20mg/L MO=30mg/L MO=40mg/L MO=50mg/L

This work was supported by the National Natural Science Foundation of China (NSFC 51372234). References

Initial MO concentration (mg/L)

1

0

0

30

60

90

120

150

180

Irradiation time (min) Fig. 8. Influence of the initial MO concentration on photodegradation rate of the MO (the inset shows the mass of degraded MO and degradation rate as function of initial MO concentration) experimental conditions: 100 mg photocatalyst and 60 mL MO solution * 10 mg/L MO was completely degraded after 150 min, so the last data point was omitted.

[44], the slope of linear regression represents the apparent reaction rate constant k. As can be seen from Fig. 7(b), the constant k exhibits an increasing trend together with Gd content between 0% and 0.5%, and reaches the maximum for 0.5Gd-TNTs sample. The maximal photocatalytic activity was attributed to the optimal Gd amount in this experiment conditions, integrating good crystallinity, preponderance of anatase phase, large specific surface area and enhanced light absorption. Although none of Gd–Ti bonds were detected in our experiment, induced defects by Gd ions in the Gd-TNTs can act as electron scavengers [45], promoting electron transfer between Gd 4f and Ti 2p [15], and retarding excitons recombination in the photocatalytic reaction [46]. Moreover, larger Gd amount, for Gd/Ti ratios over 1.0% gave a little worse crystallinity and poorer light absorption in the ultraviolet range, resulting in lower photocatalytic activity. Fig. 8 illustrates the influence of the initial MO concentration on it for the best photocatalyst, 0.5Gd-TNTs. As mentioned above, the photocatalytic process in the concentration range we discussed followed pseudo-first-order kinetics, and the linear fitting result improved as initial MO concentration increased. It is obvious (from the inset of Fig. 8) that the mass of degraded MO can be described by a crest-like curve while the percentage of degradation exhibited monotonic decrease as the increase of initial MO concentration. A larger initial MO concentration could provide photocatalyst a better driving force, while excessive initial MO concentration lowered the migration of the degraded product and transmittance of the mixture. Therefore, the most appropriate initial MO concentration for the best photocatalyst was found to be 30 mg/L. 4. Conclusions Gd-doped titania nanotubes were synthesized by the hydrothermal treatment of precursory Gd-doped titania nanoparticles, which were primarily fabricated by traditional sol–gel method. Gd-TNTs possessed a tubular morphology with length of several hundred nanometers, and their composition and optical properties were found to be altered along with the adjustment of doping level. Appropriate Gd doping increased considerably photocatalytic activity, and the optimal degradation capability belonged to 0.5GdTNTs (85%), which properly integrates better crystallinity and larger specific surface area. A drop in photocatalytic activity occurred for further increased the amount of Gd dopant, because of the worse crystallinity and weaker light absorption.

[1] R.M. Mohamed, E.S. Aazam, J. Alloys Comp. 595 (2014) 8–13. [2] Y.H. Tang, G. Zhang, C.B. Liu, S.L. Luo, X.L. Xu, L. Chen, B.G. Wang, J. Hazard. Mater. 252–253 (2013) 115–122. [3] P. Zheng, H.J. Wu, J.L. Guo, J.H. Dong, S.P. Jia, Z.P. Zhu, J. Alloys Comp. 615 (2014) 79–83. [4] S.F. Jin, Y.Z. Li, H. Xie, X. Chen, T.T. Tian, X.J. Zhao, J. Mater. Chem. 22 (2012) 1469–1476. [5] X.L. Liu, J. Lin, Y.H. Tsang, X.F. Chen, P. Hing, H.T. Huang, J. Alloys Comp. 607 (2014) 50–53. [6] C.K. Lee, K.S. Lin, C.F. Wu, M.D. Lyu, C.C. Lo, J. Hazard. Mater. 150 (2008) 494– 503. [7] W.T. Chang, Y.C. Hsueh, S.H. Huang, K.I. Liu, C.C. Kei, T.P. Perng, J. Mater. Chem. A 1 (2013) 1987–1991. [8] M. Erol, T. Dikici, M. Toparli, E. Celik, J. Alloys Comp. 604 (2014) 66–72. [9] T.B. Foong, Y.D. Shen, X. Hu, A. Sellinger, Adv. Funct. Mater. 20 (2010) 1390– 1396. [10] A. Nakahira, T. Kubo, C. Numako, Inorg. Chem. 49 (2010) 5845–5852. [11] P.Y. Dong, Y.H. Wang, B. Liu, L.N. Guo, Y.J. Huang, S. Yin, Appl. Surf. Sci. 258 (2012) 7052–7058. [12] M.J. Li, Z.Y. Chi, Y.C. Wu, J. Am. Ceram. Soc. 95 (2012) 3297–3304. [13] C. Leostean, M. Stefan, O. Pana, A.I. Cadis, R.C. Suciu, T.D. Silipas, E. Gautron, J. Alloys Comp. 575 (2013) 29–39. [14] R.M. Mohamed, I.A. Mkhalid, J. Alloys Comp. 501 (2010) 143–147. [15] J.J. Xu, Y.H. Ao, D.G. Fu, C.W. Yuan, Colloid Surf. A 334 (2009) 107–111. [16] I. Cacciotti, A. Bianco, G. Pezzotti, G. Gusmano, Chem. Eng. J. 166 (2011) 751– 764. [17] X.G. Sun, C.M. Li, L.Y. Ruan, Z. Peng, J.M. Zhang, J.J. Zhao, Y.T. Li, J. Alloys Comp. 585 (2014) 800–804. [18] J.H. Li, X. Yang, X.D. Yu, L.L. Xu, W.L. Kang, W.H. Yan, H.F. Gao, Z.H. Liu, Y.H. Guo, Appl. Surf. Sci. 255 (2009) 3731–3738. [19] M. Mahalakshmi, B. Arabindoo, M. Palanichamy, V. Murugesan, J. Nanosci. Nanotechnol. 7 (2007) 3277–3285. [20] Z.M. El-Bahy, A.A. Ismail, R.M. Mohamed, J. Hazard. Mater. 166 (2009) 138– 143. [21] H.J. Liu, G.G. Liu, G.H. Xie, M.L. Zhang, Z.H. Hou, Z.W. He, Appl. Surf. Sci. 257 (2011) 3728–3732. [22] H.J. Liu, G.G. Liu, Q.X. Zhou, G.H. Xie, Z.H. Hou, M.L. Zhang, Z.W. He, Microporous Mesoporous Mater. 142 (2011) 439–443. [23] T. Kasuga, M. Hiramatsu, A. Hoson, T. Sekino, K. Niihara, Langmuir 14 (1998) 3160–3163. [24] M.H.T. Kasuga, A. Hoson, T. Sekino, K. Niihara, Adv. Mater. 11 (1999) 1307– 1311. [25] Y. Lan, X.P. Gao, H.Y. Zhu, Z.F. Zheng, T.Y. Yan, F. Wu, S.P. Ringer, D.Y. Song, Adv. Funct. Mater. 15 (2005) 1310–1318. [26] A.W. Xu, Y. Gao, H.Q. Liu, J. Catal. 207 (2002) 151–157. [27] S. Bingham, W.A. Daoud, J. Mater. Chem. 21 (2011) 2041–2050. [28] N. Bouazza, M. Ouzzine, M.A. Lillo-Ródenas, D. Eder, A. Linares-Solano, Appl. Catal. B: Environ. 92 (2009) 377–383. [29] M. Lillorodenas, N. Bouazza, A. Berenguermurcia, J. Linaressalinas, P. Soto, A. Linaressolano, Appl. Catal. B: Environ. 71 (2007) 298–309. [30] G. Dawson, W. Chen, T.K. Zhang, Z. Chen, X.R. Cheng, Solid State Sci. 12 (2010) 2170–2176. [31] C.L. Li, J. Yuan, B.Y. Han, L. Jiang, W.F. Shangguan, Int. J. Hydrogen Energy 35 (2010) 7073–7079. [32] D.S. Seo, J.K. Lee, H. Kim, J. Cryst. Growth 229 (2001) 428–432. [33] L. Miao, S. Tanemura, T. Jiang, M. Tanemura, K. Yoshida, N. Tanaka, G. Xu, Superlattices Microstruct. 46 (2009) 357–364. [34] V. Mendez, S. Daniele, J. Nanosci. Nanotechnol. 11 (2011) 9237–9243. [35] A. Danon, K. Bhattacharyya, B.K. Vijayan, J. Lu, D.J. Sauter, K.A. Gray, P.C. Stair, E. Weitz, ACS Catal. 2 (2012) 45–49. [36] Q.Y. Wang, H.Q. Jiang, S.Y. Zang, J.S. Li, Q.F. Wang, J. Alloys Comp. 586 (2014) 411–419. [37] D.M. Tobaldi, A.S. Škapin, R.C. Pullar, M.P. Seabra, J.A. Labrincha, Ceram. Int. 39 (2013) 2619–2629. [38] J. Tauc, R. Grigorovici, A. Vancu, Phys. State Solid 15 (1966) 627–637. [39] K.M. Parida, N. Sahu, J. Mol. Catal. A: Chem. 287 (2008) 151–158. [40] D.Y. Liang, C. Cui, H.H. Hu, Y.P. Wang, S. Xu, B.L. Ying, P.G. Li, B.Q. Lu, H.L. Shen, J. Alloys Comp. 582 (2014) 236–240. [41] T. Hirakawa, K. Yawata, Y. Nosaka, Appl. Catal. A: Gen 325 (2007) 105–111. [42] X.Z. Li, C.C. Chen, J.C. Zhao, Langmuir 17 (2001) 4118–4122. [43] H. Tang, D. Zhang, G.G. Tang, X.R. Ji, C.S. Li, X.H. Yan, Q. Wu, J. Alloys Comp. 591 (2014) 52–57. [44] W.H. Ching, M. Leung, D.Y.C. Leung, Sol. Energy 77 (2004) 129–135. [45] B. Yang, Z.P. Zhu, Y. Zhou, C.B. Xia, J. Sci. Conf. Proc. 1 (2009) 243–246. [46] C.X. Lv, Y. Zhou, H. Li, M.M. Dang, C.C. Guo, Y.C. Ou, B. Xiao, Appl. Surf. Sci. 257 (2011) 5104–5108.