CERAMICSINTERNATIONAL

Available online at www.sciencedirect.com

http://dx.doi.org/0272-8842 & 20

nCorrespondinnnCorrespondE-mail addre

nobremal@fct.un

Ceramics International 40 (2014) 9525–9534www.elsevier.com/locate/ceramint

Photodegradation of phenol red on a Ni-doped niobate/carbon composite

Silvania Lanfredia,n, Marcos A.L. Nobrea,n, Paula G.P. Moraesb, Juan Matosc,nn

aFaculdade de Ciências e Tecnologia – FCT, Universidade Estadual Paulista – UNESP, P.O. Box 467, Presidente Prudente, SP 19060-900, BrazilbInstituto de Biociências, Letras e Ciências Exatas, Universidade Estadual Paulista – UNESP, P.O. Box 6154, São José do Rio Preto, SP, Brazil

cDepartment of Photocatalysis and Alternative Energies, Venezuelan Institute for Scientific Research (IVIC), Apartado 20632, Caracas 1020-A, Venezuela

Received 10 January 2014; received in revised form 6 February 2014; accepted 7 February 2014Available online 17 February 2014

Abstract

The photocatalytic activity of a new nanostructured Ni-doped niobate KSr2(Ni0.75Nb4.25)O15�δ was studied using the phenol red dye as a testmolecule and the influence of amorphous carbon deposits upon the photoactivity of the niobate-based materials was also verified.KSr2(Ni0.75Nb4.25)O15�δ powder was prepared by a high energy ball milling method and the C-KSr2(Ni0.75Nb4.25)O15�δ composite by thepartial pyrolysis of the niobate dispersed in a polyester matrix. Materials were characterized by FTIR spectroscopy and X-ray diffraction (XRD).The diffraction line profile and the refinement of the structural parameters of KSr2(Ni0.75Nb4.25)O15�δ were derived by the Rietveld method. Bothsamples showed similar phenol red photodegradation under steady-state kinetic conditions. However, amorphous carbon seems to beneficiallyaffect the reaction mechanism which followed first order kinetics. In terms of KSr2(Ni0.75Nb4.25)O15�δ concentration a clear enhancement in thephotoactivity of the niobate in the presence of amorphous carbon by a factor 4.7 was found suggesting a synergy effect between both solids.We conclude that C-KSr2(Ni0.75Nb4.25)O15�δ composite can be employed for the photocatalytic degradation of diluted pollutants.& 2014 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

Keywords: Amorphous carbon; Phenol red dye; Photocatalysis; TTB-type structure

1. Introduction

Ferroelectric oxides with the tetragonal tungsten bronze(TTB)-type structure have been the forefront of both researchand industrial applications. Ferroelectric TTB-type structurecan be considered a derivative of the classical perovskitestructure and can be described by the chemical formula(A1)2(A2)4C4Nb10O30 where A1, A2, and C denote differentsites in the crystal structure. All A1 cavities have cuboctahe-dral coordination, the A2 cavities have a pentacapped penta-gonal prismatic coordination, and C cavities exhibit atricappedtrigonal prismatic coordination. A wide variety ofcation substitutions have been possible due to the presence ofseveral interstices in the TTB-type structure [1–3]. The cationsize and the replacement fraction of the cations at different

10.1016/j.ceramint.2014.02.02614 Elsevier Ltd and Techna Group S.r.l. All rights reserved.

g authors.ing author.sses: silvania@fct.unesp.br (S. Lanfredi),esp.br (M.A.L. Nobre), jmatos@ivic.gob.ve (J. Matos).

sites in the TTB-type structure might modulate physicalproperties of these materials such as electro-optic, nonlinear,elasto-optic, pyroelectric and electrical properties [4]. In thissense, the engineering of non-stoichiometry compounds bydoping via a non-isovalent substitution cation is an easieralternative. In recent years, several catalytic and photocatalyticapplications have been broadly studied in Nb-based materialsdue to their redox, strong acidic and photosensitivity properties[5,6]. For example, oxide materials based on the niobate haveshowed activity and high selectivity in different catalyticprocesses such as oxidative dehydrogenation of alkanes [7],oxidative coupling of methane [8], dry methane reforming(DMR) [9], ethylene homologation reaction (EHR) [9] andphotocatalysis [6,10]. Although TiO2 oxide is widely used asphotocatalytic material, currently niobate-based materialssuch as KNb3O8, K6Nb10.8O30, K4Ce2Nb10O30, NiNb2O6,K4Nb6O17, and NiO–KTiNbO5 have been widely studiedbecause of their excellent photocatalytic properties [11,12].Niobate-based materials have presented high photocatalyticactivity also in the field of water decomposition [13]. Up to

S. Lanfredi et al. / Ceramics International 40 (2014) 9525–95349526

now, the photocatalytic activity of KSr2Nb5O15 has beenreported in the photodegradation of organic pollutants in anaqueous phase [14]. However, the potential application inheterogenous photocatalysis of the non-stoichiometric niobatewith tetragonal tungsten bronze (TTB)-type structure withformula close to KSr2(Ni0.75Nb4.25)O15�δ dispersed in anamorphous carbon matrix is reported here for the first time.

2. Experimental

2.1. Synthesis

Single phase and crystalline powder of KSr2(Ni0.75Nb4.25)O15�δ non-stoichiometric were prepared using the mechanicalmixture of oxides via the high-energy ball milling method(HEBM) [15–17]. The preparation was done using a high-energy milling type Attritor (Netzsch). The starting high purityreagents were Nb2O5 � 4H2O, K2CO3, SrCO3 and Ni2O3. Themixture of the starting materials was carried out in isopropylicalcohol using stabilized zirconia balls of 1.2 mm in diameter.A powder to ball weight ratio of 1:16 was used. The mixturewas stirred with a Molinex-type agitator shaft with eccentricradial disks that accelerated the grinding media, which gives anextra radial impulse during each rotation of the shaft with amotor of 1/3 hp. The milling was performed with a rate of1200 rpm for 5 h. After milling of reagents, the material wasdried in a glove box with forced air flow at 373 K. Singlephase powders were obtained after calcination at 1373 K for10 h. The precursor was calcined in a tube furnace under anintegral oxygen atmosphere. An oxygen flow of 300 mL/minwas maintained during a complete thermal cycle. Powderwas de-agglomerated in an agate mortar with a 350-meshminimum. At this process the particles showed to be fragile.

2.2. Pyrolysis method

C-containing KSr2(Ni0.75Nb4.25)O15�δ composite was pre-pared by a partial pyrolysis method similar to the processdeveloped by the Pechini method consisting in the formation ofa polymeric resin produced by polyesterification between a metalchelate complex from hydroxycarboxylic acid and a polyhydrox-yalcohol such as ethyleneglycol. Citric acid (C6H8O7 �H2O),ethyleneglycol (HOCH2CH2OH), and KSr2(Ni0.75Nb4.25)O15�δ

were the starting reactants.The citric acid was dissolved inethyleneglycol (mass ratio of 40:60) with continuous magneticstirring at 70 1C to promote polymerization. After the polyester-ification reaction, a polymeric gel was obtained. The polymermaintained in the beaker was submitted to a primary calcinationin a box-type furnace. The composite was obtained by a one-stepcalcination heating cycle under static air atmosphere fromroom temperature up to 300 1C at constant heating rate ofabout 1 1C min�1. The composite was deagglomerated in theagate mortar with a 350-mesh minimum and denotedC-KSr2(Ni0.75Nb4.25)O15�δ.

2.3. Photocatalytic tests

Photocatalytic activity of KSr2(Ni0.75Nb4.25)O15�δ andC-KSr2(Ni0.75Nb4.25)O15�δ composite was studied by follow-ing the kinetics of phenol red disappearance as the testmolecule. A low-flow dynamic and open to air photocatalyticreactor was used (Fig. S1, Supplementary material). 1 Lvolume with 3.53� 10�5 mol L�1 initial phenol red concen-tration in the presence of 100 mg or 200 mg weight for theniobate-based materials was used. A low power (15 W) UVlamp (285 nm) was used as an irradiation source. In order towell-establish the photocatalytic behavior of the presentniobate-based materials, both the kinetic of phenol reddegradation by direct photolysis in the absence of solids(Figs. S2 and S3, Supplementary material) and the kinetic ofphenol red disappearance in the presence of the amorphouscarbon matrix (Figs. S4 and S5, Supplementary material) werefollowed. The concentration of phenol red in aqueous solutionwas determined by the changes in absorbance at 480 nm with aColorimeter. The disappearance of phenol red was reported interms of the conversion (X) obtained by the followingexpression:

X ¼ ½ðCo�CtÞ=Co�U100� ½ðAo�AtÞ=Ao�U100 ð1Þwhere Co is the initial concentration of phenol red dye, Ct isthe concentration at the reaction time t, and Ao and At are theinitial absorbance and the absorbance at reaction time t.

2.4. Characterization

Texture characterization of samples was performed byadsorption–desorption N2 isotherms at 77 K. The full iso-therms in the range of 4� 10�3

–84 kPa were measured inMicromeritics ASAP-2020. Equivalent surface area, microporevolume, and pore diameters were obtained by Brunauer–Emmet–Teller (BET), Harkins–Jura (HJ) and Horvath–Kawa-zoe (HK) methods, respectively. HJ and HK methods wereemployed because they are better suited algorithms for TTB-type structures.Structural characterization of the KSr2(Ni0.75Nb4.25)O15�δ

powders was carried out by X-ray diffraction (XRD).A Shimadzu (model XRD-6000) diffractometer with Cu-Kα1radiation (λ¼1.54056 Å) and a graphite monochromator wasused. Measurements were carried out over an angular range of51r 2θr 801 with slower scanning rate equal to 0.5o/minand with a scanning step of 0.02o and 2.40 s as fixed countingtime. Divergence, scattered and receiving radiation slits were11, 11 and 0.2 mm respectively.The structures were refinedaccording to the Rietveld method [17–19] using the Fullprofprogram [20]. Refinement parameters and variables adoptedwere the background coefficients, profile coefficients, scalefactor, lattice parameters, atomic coordination, occupancyfactors and isothermal parameters for five species of atoms(Kþ , Sr2þ , Ni2þ , Nb5þ and O2-). The background level wasfitted with a five-order polynomial function, and the peakshape, with a pseudo-Voigt function. The angular dependenceof the peak full-width at half-maximum (H) was defined by the

S. Lanfredi et al. / Ceramics International 40 (2014) 9525–9534 9527

function determined by Caglioti et al. [21]. Powder data andexperimental conditions of Ni-doped KSr2Nb5O15 are listed inTable 1. From atomic positions derived in the refinement step,the Ni-doped KSr2Nb5O15 unit cell was built using theDiamond software package. The average crystallite size (D)and the lattice strain of KSr2(Ni0.75Nb4.25)O15�δ of thenanostructured powder were estimated from X-ray diffractionline broadening. The crystallite size was estimated by Scher-rer's equation using the Jade 8 Plus software:

D¼ k:λ

β: cos θð2Þ

where β is the broadening of the diffraction line measured athalf of the maximum intensity, λ is the wavelength (Cu-Kα), θis the Bragg angle for a given diffraction, and k is a constant,which is in general equal to 0.9 for powders. The instrumentalbroadening effect was eliminated by subtracting the full widthat half-maximum (βo) of a standard sample (SiO2) from β ofthe respective Bragg peaks.The lattice strain was estimatedusing the Williamson–Hall approach:

βcos ðθÞ

λ¼ 1

Dþ4ε

sin ðθÞλ

ð3Þ

where D represents the crystallite size and ε represents thelattice strain. The characteristic Williamson–Hall plot corre-sponds to the graph of βcos(θ)/λ as a function of 4sin(θ)/λ.Here, 4sin(θ)/λ is on the x-axis. Each point is assigned to aspecific diffraction line. After point collection, a linear

Table 1Powder data and experimental conditions.

Crystallographic dataFormula KSr2(Ni0.75Nb4.75)O15-δ

Crystal system TetragonalSpace group P4bm(no. 100)a (Å) 12.4635(3)c (Å) 3.9361(9)V (Å) 611.43(3)Z 2

Data collectionTemperature (oC) 26Wavelength [CuKα] (Å) 1.5418Monochromator GraphiteMeasuring range (deg) 5r2θr80Step (o2θ) 0.02Integration time (s) 10

Rietveld dataProgram FULLPROFFunction for background level Polynomial – 5 orderFunction for peak shape Pseudo-Voigt(H2¼U tan2θþV tanθþW)U 0.096 (4)V �0.047 (2)W 0.027 (4)RBragg (%) 6.06RF (%) 5.03cRp (%) 13.4cRwp (%) 19.0cRexp (%) 8.68χ2 4.39

regression should give a linear fit. The crystalline size wasextracted from the y-intercept of the fit. The lattice strain η wasextracted from the slope of the fit curve. The size and micro-morphology of the KSr2(Ni0.75Nb4.25)O15�δ precursor powdercalcined at 1373 K for 10 h were also investigated usingscanning electron microscopy (SEM).Chemical bonds were analyzed by infrared spectroscopy

(FTIR). The sample was diluted in KBr in 1:100 ratio.Measurements were carried out with an instrument resolutionof 1 cm�1 in the range of 900–400 cm�1 for 100 scans using aFourier transform spectrometer, Model Digilab Excalibur (FTS3100 HE series). This range was selected because it fits in themid-infrared region where the characteristic bands assigned toNb–O bonds appear.

3. Results and discussion

3.1. Structural properties

Fig. 1 shows the X-ray diffraction pattern of theKSr2(Ni0.75Nb4.25)O15�δ powder with the observed and calcu-lated XRD patterns, as well as the difference between them.Crystalline and single phase based on the KSr2Nb5O15 host-structure exhibited only a set of diffraction lines ascribed tothe TTB-type structure. Structural parameter refinements bythe Rietveld method were carried out by taking into accountthe non-centro symmetric space group P4bm(#100) which iscompatible with the existence rule [(0 k l) k¼2n], as describedelsewhere [17,22]. The lattice parameters derived for KSr2(Ni0.75Nb4.25)O15�δ by the Rietveld method were a= 12.4635(3) Å and c= 3.9361 (9) Å, while both the volume and thetetragonal ratio c/a were 611.4291(3) Å3 and 0.316, respec-tively. These parameters values have been reported recently byour group [17]. The best theoretical adjustment for theKSr2(Ni0.75Nb4.25)O15�δ phase was obtained by assuming thateach pentagonal site [4c (x, xþ1/2, z] was statisticallyoccupied by equal amounts of Kþ and Sr2þ ions, thetetragonal site [2a (0,0,z)] occupied by Sr2þ ions and eachoctahedral site was partially occupied by Ni2þ ions. Thetrigonal site was considered vacant. The statistic mixing of

Fig. 1. Rietveld plot for KSr2(Ni0.75Nb4.25)O15�δ powder.

Table 2Atomic coordinates, isotropic atomic displacement parameters B(eq)(Å)

2, and relative occupancies P.

Atoms WyckoffPosition

x/a y/b z/c B(eq) (Å)2 P

Sr(1) 2a 0 0 �0.016(6) 3.6320 (3) 0.16K(1) 4c 0.1707(3) 0.6707(3) �0.022(7) 3.2372 (4) 0.32Sr(2) 4c 0.1707(3) 0.6707(3) �0.022(7) 3.2372 (4) 0.32Ni(1) 2b 0 1/2 1/2 0.7895 (2) 0.28Nb(2) 8d 0.07569(12) 0.2126(2) 0.490(5) 1.6028 (14) 1.00O(1) 8d 0.1345(9) 0.0630(9) 0.574(6) 6.5534 (3) 1.03O(2) 8d 0.3384(13) 0.0012(8) 0.535(10) 6.5534 (3) 1.03O(3) 8d 0.0716(14) 0.2004(14) 0.080(7) 6.5534 (3) 1.03O(4) 4c 0.2791(12) 0.7791(12) 0.493(18) 6.5534 (3) 0.50O(5) 2b 0 0.50000 0.175(15) 6.5534 (3) 0.28

Fig. 2. SEM images of the KSr2(Ni0.75Nb4.25)O15�δ material.

S. Lanfredi et al. / Ceramics International 40 (2014) 9525–95349528

Nb5þ and Ni2þ cations at 2b and 8d sites did not produce agood adjustment. The atomic parameters derived by therefinement of the KSr2(Ni0.75Nb4.25)O15�δ are listed inTable 2. It must be pointed out that the crystallite size ofKSr2(Ni0.75Nb4.25)O15�δ obtained from Scherrer´s equationwas equal to 35 nm.

Fig. 2 shows two SEM images of KSr2(Ni0.75Nb4.25)O15�δ

sample. From the analysis of these images it can be inferred thatthe niobate-based material is constituted by spherical particles withseveral sizes, indicating nanoparticle coalescence and well definedaggregate development. SEM images show that the size of primaryparticles is in the order of 45 nm because coalescence actuationsecondary particles with sizes ranging from 80 to 100 nmwere formed. Thus, the XRD and SEM data clearly indicate thatultrafine crystalline KSr2(Ni0.75Nb4.25)O15�δ nanostructuredpowders were prepared. The slight difference between the crystal-lite size calculated using X-ray diffraction data (35 nm) and the

particle size obtained by SEM (45 nm) was due to the fact that theparticles were composed of several crystalline domains, whichwere observed by X-ray diffraction, while the whole particle wasobserved in SEM. Fig. 3 shows the graphic representation of theunit cell obtained for the Ni-doped KSr2Nb5O15 powder along theac plane from data listed in Table 2. Niobium and nickel atomsare coordinated by oxygen atoms in a ratio 1:6; four oxygen atomsare explicitly located in the same plane as the niobium and nickelatoms, and other two are located above and below the plane,respectively. Specifically, nickel cations (Ni2þ ) occupy Nb(1)positions, which is due to the ionic ratio and octahedral occupationpreferential [17]. This occupation results in some distortion degreeof [NbO6] polyhedra, which appears to be necessary for thecomplete structural accommodation of the nickel. In this sense,both rotation and tilting phenomena of the octahedra can behypothesized for such an accommodation to occur. After solvingthe structural characterization, the influence of the amorphouscarbon matrix on the KSr2(Ni0.75Nb4.25)O15�δ powder wasdiscussed.Fig. 4 shows the room temperature XRD pattern of theKSr2(Ni0.75Nb4.25)O15�δ embedded in the carbon matrix, wherepoorly defined crystalline diffraction peaks were superimposed tohaloes coming from the amorphous carbonmatrix. The observedpeaks can be indexed as the Bragg reflections of KSr2Nb5O15 withtetragonal symmetry and space group P4bm (#100). The largecontribution of the carbon matrix and the low intensity ratiobetween the diffraction peaks and the background makes theRietveld analysis of the XRD pattern difficult; hence, a preciseestimation of the average crystallite size of KSr2(Ni0.75Nb4.25)O15�δ is not possible although the large broadening of thediffraction peaks suggests average particle sizes in the nanometerlength scale. An XRD pattern in the C-KSr2(Ni0.75Nb4.25)O15�δ

sample is in agreement with the features in SEM images fromFig. 5 where excessive deposits of amorphous carbon can beobserved in the case of C-KSr2(Ni0.75Nb4.25)O15�δ sample sug-gesting that further studies concerning to reach an optimal amountof amorphous carbon in the composite must be done.Figs. S2 and S3 in the Supplementary material show the N2

adsorption–desorption isotherm of KSr2(Ni0.75Nb4.25)O15�δ andC-KSr2(Ni0.75Nb4.25)O15�δ. Both samples showed very similartrends and a summary of textural parameters is given in Table S1.As expected, the N2 adsorption–desorption isotherms obtained on

Fig. 3. Graphical representation of the unit cell obtained for the KSr2(Ni0.75Nb4.25)O15�δ.

Fig. 4. XRD pattern of the C-KSr2(Ni0.75Nb4.25)O15�δ.

Fig. 5. SEM images of the C-KSr2(Ni0.75Nb4.25)O15�δ composite.

S. Lanfredi et al. / Ceramics International 40 (2014) 9525–9534 9529

KSr2(Ni0.75Nb4.25)O15�δ and C-KSr2(Ni0.75Nb4.25)O15�δ areII-type with a remarkably vertical hysteresis. The isothermscorrespond to non-porous or relatively large pore size materialsin agreement with the low surface area about 25 and 36 m2 g-1 forKSr2(Ni0.75Nb4.25)O15�δ and C-KSr2(Ni0.75Nb4.25)O15�δ, respec-tively, with a negligible micropore volume about 0.001 and0.003 cm3 g�1 for KSr2(Ni0.75Nb4.25)O15�δ and C-KSr2(Ni0.75Nb4.25)O15�δ, respectively and with the large mean porediameter of about 58 and 62 nm, for KSr2(Ni0.75Nb4.25)O15�δ

and C-KSr2(Ni0.75Nb4.25)O15�δ respectively. This apparent hys-teresis phenomenon can be attributable to aggregation ofnanoparticles as previously reported for non-porous niobate-basedmaterials [9]. It can be concluded from these results that carbondeposits influence on the niobate-based material is negligible inagreement with the amorphous morphology observed in the SEMimage from Fig. 5. It should be also noticed that it was impossibleto obtain a well-defined N2 adsorption–desorption isotherm forthe amorphous carbon alone because the non-porous frameworksuggested that a further thermal treatment is required to activatethis carbonaceous material.

3.2. Chemical bond analysis

Fig. 6 shows the FTIR spectra of KSr2(Ni0.75Nb4.25)O15�δ

and C-KSr2(Ni0.75Nb4.25)O15�δ powders. These spectra showa systematic absence of peaks assigned to particular vibrations.It is impossible to achieve further information about the bandintensity. The spectrum of KSr2(Ni0.75Nb4.25)O15�δ shows fivebands between 400 and 900 cm�1, being three of them broad,asymmetric and of middle intensity centered on 656, 724 and840 cm�1, respectively. Other two bands are sharp with lowintensity being centered at 594 and 424 cm�1. The absorptionband centered at 840 cm�1 has been assigned to the (Nb–O)stretching of the niobium bond with the apical oxygen in NbO6

octahedra [23]. For the C-KSr2(Ni0.75Nb4.25)O15�δ composite,

Fig. 6. FTIR transmittance spectrum of the niobate-based powders.

Table 3FTIR band positions and characteristics of niobate and amorphous carbonmatrix-niobate composite.

Assigments Wavenumber (cm�1) Intensity Ref.

νas (Nb–O)a 560–419 low [23,24]νs (O–Nb–O)a 725–640 medium [23,24]νs(Nb–O)b 838–880 low [23]νs (C–O), δ(C–H) 1281–1462 medium [25]νs (C¼O) 1756 high [25]νs (C¼O) 2548 low [25]νs (C–H) 2949 high [25][νs (O–H)]Adsorvido 3000–3800 high [25]

aSlightly distorted NbO6 Octahedra.bDistorted NbO6 Octahedra.

S. Lanfredi et al. / Ceramics International 40 (2014) 9525–95349530

the bands identified between 400 and 900 cm�1 were 458,567, 663, 731 and 840 cm�1. All bands have been assigned toNb–O bonds [23,24]. The sharp absorption positioned around731 and 458 cm�1 and the band centered at 663 cm�1 hasbeen assigned to Nb–O stretching [23,24]. Parameter bandpositions and band characteristics are listed in Table 3. Theanalysis of the carbon matrix effects on KSr2(Ni0.75Nb4.25)O15�δ shows that above 900 cm�1 the peaks are assigned tothe carbon matrix. The intense absorption peak at 1743 cm�1

corresponds to the stretching vibration of C¼O [25] and theabsorption at 1603 cm�1 is assigned to the stretching vibrationof C¼C. The broad peaks observed at around 1263 cm�1 andat 1447 cm�1 could be assigned to C–O stretch and also toC–C bonds. The band about 2953 cm�1 can be associated withCH2 or CH3 stretching while the band at 2585 cm�1 can beassigned to the C¼O stretching vibrations. The IR absorptionin the range of 3000–3800 cm�1 with an intense absorptionpeak at 3636 cm�1 is associated with hydroxyl stretchingvibration of adsorbed water or surface carboxylic groups [26].In short, amorphous carbon seems to influence slightly theabsorption peaks due to the Nb–O stretching. In a similar way,our group has already reported this influence in the case ofthe interaction between nanoporous activated carbons and TiO2

[27,28]. In this sense, the photocatalytic activity of UV-irradiated

KSr2(Ni0.75Nb4.25)O15�δ and C-KSr2(Ni0.75Nb4.25)O15�δ is dis-cussed as follows.

3.3. Photocatalytic activity of KSr2(Ni0.75Nb4.25)O15�δ

and C-KSr2(Ni0.75Nb4.25)O15�δ

Preliminary kinetic studies of phenol red adsorption in darkon KSr2(Ni0.75Nb4.25)O15�δ and C-KSr2(Ni0.75Nb4.25)O15�δ

were performed. The phenol red quantity adsorbed after60 min on KSr2(Ni0.75Nb4.25)O15�δ and C-KSr2(Ni0.75Nb4.25)O15�δ was about 5% and 10%, respectively. These valuesadsorbed in dark can be considered as negligible and accord-ingly the kinetics of phenol red disappearance underUV-irradiation were performed without preliminary kineticsof adsorption. However, it is important to note that phenol redconversion detected by direct photolysis and in the presence ofneat amorphous carbon was preliminary performed (Supple-mentary material). Fig. S4 shows that about 65% of phenol redwas converted after 4 h UV-irradiation. Fig. S5 shows thelinear regression of the kinetic data from Fig. S4. A goodlinearity suggesting a proper fit with a first-order mechanismfor the phenol red degradation by direct photolysis can be seen.The first-order apparent rate-constant obtained from directphotolysis (Fig. S5, Supplementary material) was about0.0040 min�1 (Table 4). Fig. S6 (Supplementary material)shows that phenol red disappearance increases up to 80% after4 h irradiation in the presence of amorphous carbon. Having inmind that amorphous carbon is non-photoactive [29–33], thepresent result suggests that 100 mg amorphous carbonadsorbed about 15% phenol red. For comparative purposesthe apparent rate constant of phenol red disappearance on neatamorphous carbon was estimated from the linear regression(Fig. S7, Supplementary material) and this value was about0.0078 min�1 (Table 4).On the other hand, Fig. 7 shows the phenol red conversion

as a function of time by using two different KSr2(Ni0.75Nb4.25)O15�δ concentrations. It can be seen from Fig. 7 that phenolred conversion decreases with increasing of catalyst concen-tration from 100 to 200 mg/L. After 3 h irradiation phenol redconverted on KSr2(Ni0.75Nb4.25)O15�δ was up to about 85%using 100 mg/L while using 200 mg/L conversion was onlyabout 60%. As discussed above, direct photolysis was about65% and the phenol red adsorbed on KSr2(Ni0.75Nb4.25)O15�δ

was only 5% suggesting that up to 15% phenol red wasphotodegradated on UV-irradiated KSr2(Ni0.75Nb4.25)O15�δ.It indicates that phenol red disappearance can be ascribed toan intrinsically photocatalytic activity of the Ni-doped niobatematerial. The kinetics of the photocatalytic process underUV-irradiated KSr2(Ni0.75Nb4.25)O15�δ can be described bythe following general reactions:

KSr2ðNi0:75Nb4:25ÞO15� δ-hvKSr2ðNi0:75Nb4:25ÞO15� δðe�CBþhþ

VBÞð4Þ

hþVBþOH�

ads-OHd ð5Þ

Table 4Summary of kinetic parameters for the phenol red dye photodegradation.

Catalyst concentration (mg L-1) Apparent rate constant, kapp (min-1) Correlation coefficient, R t0.5 (min)

KSr2(Ni0.75Nb4.25)O15�δ

Photolysis 0.004 0.9952 173.3100 0.0088 0.9842 78.4200 0.0061 0.9793 113.6

C-KSr2(Ni0.75Nb4.25)O15�δ

Carbon 0.0078 0.9752 88.9100 0.0084 0.9716 82.5

Fig. 7. Conversion of phenol red dye on UV-irradiated KSr2(Ni0.75Nb4.25)O15�δ as a function of catalysts concentration. (a) 100 mg/L; (b) 200 mg/L.

Fig. 8. Linear regression Ln(Co/Ct)¼ f(t) from the kinetic data of Fig. 7.Catalyst concentration: (a) 100 mg/L; (b) 200 mg/L.

S. Lanfredi et al. / Ceramics International 40 (2014) 9525–9534 9531

e�CBþO2� ads-O� d

2� ads ð6Þ

KSr2ðNi0:75Nb4:25ÞO15� δðe�BCþhþBVÞ-KSr2ðNi0:75Nb4:25ÞO15� δþΔ

ð7ÞA kinetic explanation both for the photoactivity ofKSr2(Ni0.75Nb4.25)O15�δ and the influence of the amorphouscarbon matrix can be based on the Langmuir–Hinshelwoodmechanism [31–33] with the reaction rate being proportional tothe surface coverage θ varying as follows:

r ¼ kθ ¼ k½Kads:Ceq=ð1þKads:CeqþΣKiCiÞ�: ð8Þwhere Kads and Ki correspond to the adsorption constants ofphenol red and the intermediate i, Ceq and Ci are the phenol redand intermediate concentrations in solution after achieving theequilibrium adsorption in dark. However, we have alreadypointed out that the phenol red adsorption in dark both onKSr2(Ni0.75Nb4.25)O15�δ and on C-KSr2(Ni0.75Nb4.25)O15�δ

can be considered negligible to the intrinsic photoactivity ofthe niobate-based material and therefore both terms Kads.Ceq

and ΣKiCi are much lower than unit and Expression (8) issimplified to

r� k½Kads:Ceq� ð9ÞThis approximation agrees with an apparent first ordermechanism of reaction because k.Kads corresponds to the firstorder apparent rate-constant [29,31–33] denoted as kapp. This

value is obtained from the linear regression Ln(Co/Ct)=f(t) ofthe kinetic data by

LnðCo=CtÞ ¼ kapp:t ð10Þ

Fig. 8 shows the linear regression Ln(Co/Ct)¼ f(t) of the kineticdata from Fig. 7. A summary of the kinetic parameters obtained inthe phenol red photodegradation on UV-irradiated KSr2(Ni0.75Nb4.25)O15�δ and C-KSr2(Ni0.75Nb4.25)O15�δ composite is givenin Table 4. The near to unit value of the correlation coefficient (R)showed in Table 4 indicates that phenol red photodegradationswith 100 and 200 mg L-1 photocatalysts concentration follow a firstorder kinetics. It is important to have in mind that the first orderapparent rate-constants obtained on KSr2(Ni0.75Nb4.25)O15�δ andC-KSr2(Ni0.75Nb4.25)O15�δ in Table 4 contain the direct photolysisand adsorption contributions, respectively. However, both con-tributions are constant and therefore kapp values in Table 4 suggestthat the phenol red disappearance is about 2.2 and 1.5 times higherthan that obtained with direct photolysisby using 100 mg L�1and200 mg L�1 KSr2(Ni0.75Nb4.25)O15�δ, respectively. These resultsdemonstrate that the present novel material KSr2(Ni0.75Nb4.25)O15�δ is intrinsically photoactive under UV-irradiation. Photo-activity is about 1.5 times higher by using 100 mg L-1 than thatobtained by using 200 mg L-1 of UV-irradiated KSr2(Ni0.75Nb4.25)O15�δ. The decrease in the photocatalytic activity at the higherconcentration could be caused by shielding or scattering of UVphotons consequence by the aggregation of KSr2(Ni0.75Nb4.25)

Fig. 10. Linear regression Ln(Co/Ct)¼f(t) of the kinetic data from Fig. 9.(a) C-KSr2(Ni0.75Nb4.25)O15�δ composite and (b) KSr2(Ni0.75Nb4.25)O15�δ.

S. Lanfredi et al. / Ceramics International 40 (2014) 9525–95349532

O15�δ nanoparticles previously reported by our group on TiO2

nanoparticles during the photodegradation of phenol [29,31,32] and4-chlorophenol [27,30,33]. In these works a clear correlationbetween the aggregation of TiO2 nanoparticles and the surfacepH of nanoporous carbons on C-TiO2 binary materials wasreported. The more acidic the carbons surface pH the higher theaggregation of TiO2 nanoparticles and a lower photocatalyticactivity [30,33]. In the present case, it is clear to expect animportant aggregation of KSr2(Ni0.75Nb4.25)O15�δ nanoparticles athigh concentration of photocatalysts because of the acidic nature ofniobate-based materials [9]. The influence of the amorphous carbonmaterial is discussed as follows. C-KSr2(Ni0.75Nb4.25)O15�δ com-posite has about 80% weight amorphous carbon and about 20%weight KSr2(Ni0.75Nb4.25)O15�δ. In other words, 100 mg L�1 ofC-KSr2(Ni0.75Nb4.25)O15�δ contains about 80 mg amorphouscarbon and about 20 mg KSr2(Ni0.75Nb4.25)O15�δ. Fig. 9 showsa comparison of the photocatalytic activity of 100 mg L-1 ofC-KSr2(Ni0.75Nb4.25)O15�δ and KSr2(Ni0.75Nb4.25)O15�δ. It canbe seen from Fig. 9 that between 30 and 150 min reaction theC-KSr2(Ni0.75Nb4.25)O15�δ composite showed a higher phenol redconversion than that on pristine KSr2(Ni0.75Nb4.25)O15�δ. Thesephenomena can be ascribed to a higher mass diffusion of phenolred induced from solution to the amorphous carbon playing thecatalytic support role. However, after 180 min reaction the phenolred disappearance detected on C-KSr2(Ni0.75Nb4.25)O15�δ wassimilar to that detected on KSr2(Ni0.75Nb4.25)O15�δ. Several worksof phenol [32,34,35], 4-chlorophenol [27,28,30,33], and methyleneblue photodegradations [36–38] on UV-irradiated TiO2 havereported that this apparent inhibition in photoactivity can beassociated with an accumulative effect of the main intermediatesproducts formed during irradiation. As discussed above, due to thelow adsorption in dark of phenol red, it can be suggested thatthe intermediate products could compete with the phenol red bythe photoactive sites on KSr2(Ni0.75Nb4.25)O15�δ. Fig. 10 showsthe linear regression Ln(Co/Ct)¼ f(t) from kinetic data ofFig. 9 suggesting that the disappearance of phenol red on C-KSr2(Ni0.75Nb4.25)O15�δ composite also follows a first orderkinetic which is confirmed by the nearly to unit correlationcoefficient (Table 4). The first-order apparent rate-constant obtained

Fig. 9. Conversion of phenol red dye. (a) C-KSr2(Ni0.75Nb4.25)O15�δ compo-site and (b) KSr2(Ni0.75Nb4.25)O15�δ.

on 100 mg L-1 C-KSr2(Ni0.75Nb4.25)O15�δ composite is slightlylower than that on KSr2(Ni0.75Nb4.25)O15�δ (0.0084 min-1 against0.0088 min-1) but it is still 2.1 higher than that obtained by directphotolysis. Having in mind that only 20% weight (about 20 mg/L)in the composite corresponds to KSr2(Ni0.75Nb4.25)O15�δ, thephotocatalytic activity can be re-estimated in terms of theconcentration with values about 0.0884 L min-1 g-1 and 0.4195 Lmin-1 g-1 for KSr2(Ni0.75Nb4.25)O15�δ and C-KSr2(Ni0.75Nb4.25)O15�δ composite, respectively. The photoactivity of theC-KSr2(Ni0.75Nb4.25)O15�δ composite is clearly higher than thatobtained on KSr2(Ni0.75Nb4.25)O15�δ by a factor about 4.7suggesting that amorphous carbon plays a synergy effect withKSr2(Ni0.75Nb4.25)O15�δ in the C-KSr2(Ni0.75Nb4.25)O15�δ com-posite. This synergy effect could be ascribed to an enhancement inthe mass diffusion of phenol red from the carbon deposits to theniobate-based materials as reported in previous works [27–33] andin agreement with the higher surface area of the composite (TableS1, Supplementary material). Moreover, having in mind that FTIRspectra of the C-KSr2(Ni0.75Nb4.25)O15�δ composite (Fig. 6)showed that the surface of carbon deposits contains oxygenatedfunctional groups, the present results suggest that amorphouscarbon matrix influence the reaction mechanism of phenol reddye photodegradation on UV-irradiated KSr2(Ni0.75Nb4.25)O15�δ

probably by a photoassisting role as reported in previous works ofphenol [29,32], 4-chlorophenol [27,28,30,31,33], 2,4-dichlorophe-noxicetic acid [31] and more recently in the photooxidation of 2-propanol in gas-solid-regime [39].It should be pointed out that due to the relative sensitivity of

phenol red to be degradated under UV irradiation, to verify apossible scale-up of niobate-based materials, a deep study ofmethylene blue photodegradation under solar irradiation isbeing performed under several types of niobate-based ceramicmaterials. Methylene blue is a more refractory molecule harderto suffer photooxidation in comparison to phenol red becausebesides two dimethyl amino groups, one sulfur and onenitrogen atoms are included in the central aromatic ring inthe methylene blue and thus, the genolysis of S- andN-bonding to C atoms in the aromatic structure is requirebefore oxidation.

S. Lanfredi et al. / Ceramics International 40 (2014) 9525–9534 9533

4. Conclusions

We do believe that the present results and discussion of thekinetic of disappearance of phenol red on UV-irradiationdemonstrate and report by the first-time the photocatalyticactivity of the new KSr2(Ni0.75Nb4.25)O15�δ material. Amor-phous carbon beneficially influence the photoactivity of theniobate-based material and C-KSr2(Ni0.75Nb4.25)O15�δ com-posite showed a higher photocatalytic activity thanKSr2(Ni0.75Nb4.25)O15�δ. This effect was ascribed as asynergy effect with a clear enhancement in the photoactivityof the niobate by a factor 4.7 suggesting a clear potential of theC-KSr2(Ni0.75Nb4.25)O15�δ composite to be employed indiluted conditions of pollutants. Also, due to the majorgranulometry of the composite is possible to reclaim thephotocatalyst by simple mechanic filtration which is anotheradvantage of this C-KSr2(Ni0.75Nb4.25)O15�δ composite.

Acknowledgments

FAPESP: Contracts 2001/13421–7, 2002/05997-9 and2007/03510-9. CNPq Contracts 308924//2009-6 and 472773/2009-7 for the financial support.

Appendix A. Supplementary material

Supplementary data associated with this article can be foundin the online version at http://dx.doi.org/10.1016/j.ceramint.2014.02.026.

References

[1] P.R. Slater, J.T.S. Irvine, Synthesis and electrical characterisation of thetetragonal tungsten bronze type phases, (Ba/Sr/Ca/La)0.6MxNb1�xO3�δ

(M¼Mg, Ni, Mn, Cr, Fe, In, Sn): evaluation as potential anode materialsfor solid oxide fuel cells, Solid State Ion. 124 (1–2) (1999) 61–72.

[2] B. Tribotté, J.M. Haussonne, K2Sr4Nb10O30-based dielectric ceramicshaving the tetragonal tungsten bronze structure and temperature-stablehigh permittivity, J. Eur. Ceram. Soc. 19 (6–7) (1999) 1105–1109.

[3] S.C. Abrahams, P.B. Jamieson, J.L. Bernstein, Ferroelectric tungstenbronze‐type crystal structures: III. Potassium lithium niobateK(6�x�y)Li(4þx)Nb(10þ y)O30, J. Chem. Phys. 54 (6) (1971) 2355–2364.

[4] N. Wakiya, J.K. Wang, A. Saiki, K. Shinozaki, N. Mizutani, Synthesisand dielectric properties of Ba1�xR2x/3Nb2O6 (R: rare earth) withtetragonal tungsten bronze structure, J. Eur. Ceram. Soc. 19 (6–7) (1999)1071–1075.

[5] K. Tanabe, Application of niobium oxides as catalysts, Catal. Today 8 (1)(1990) 1–11.

[6] R.M. Navarro Yerga, M.C. Alvarez Galván, F. del Valle, J.A. Villoria dela Mano, J.L.G. Fierro, Water splitting on semiconductor catalysts undervisible-light irradiation, ChemSusChem 2 (6) (2009) 471–485.

[7] O. Desponds, R.L. Keiski, G.A. Somorjai, The oxidative dehydrogena-tion of ethane over molybdenum–vanadium–niobium oxide catalysts: therole of catalyst composition, Catal. Lett. 19 (1) (1993) 17–32.

[8] H.M. Swaan, Y. Li, K. Seshan, J.G. Van Ommen, J.R.H. Ross, Theoxidative coupling of methane and the oxidative dehydrogenation ofethane over a niobium promoted lithium doped magnesium oxidecatalyst, Catal. Today 16 (3-4) (1993) 537–546.

[9] J. Matos, P.S. Poon, S. Lanfredi, M.A.L. Nobre, Nobre. functionalnanostructured catalysts based on the niobates to the dry methane

reforming and ethylene homologation reactions, Fuel 107 (1) (2013)503–510.

[10] K. Domen, A. Kudo, A. Tanaka, T. Onishi, Overall photodecompositionof water on a layered niobiate catalyst, Catal. Today 8 (1) (1990) 77–84.

[11] G. Zhang, F. He, X. Zou, J. Gong, H. Tu, H. Zhang, Q. Zhang, Y. Liu,Hydrothermal synthesis and photocatalytic property of KNb3O8 withnanometer leaf-like network, J. Alloys Compd. 427 (1–2) (2007) 82–86.

[12] M. Tian, W. Shangguan, J. Yuan, L. Jiang, M. Chen, J. Shi, Z. Ouyang,S. Wang, K4Ce2M10O30 (M¼Ta, Nb) as visible light-driven photocata-lysts for hydrogen evolution from water decomposition, Appl.Catal. A309 (1) (2006) 76–84.

[13] H. Takahashi, M. Kakihana, Y. Yamashita, K. Yoshida, S. Ikeda,M. Hara, K. Domen, Synthesis of NiO-loaded KTiNbO5 photocatalystsby a novel polymerizable complex method, J. Alloys Compd. 285 (1-2)(1999) 77–81.

[14] Z. Gaoke, L. Yiqiu, W. Junting, T. Haibin, Y. Xinyi, J. Mater. Sci. Ed. 24(2009) 742–746.

[15] S. Lanfredi, D.H.M. Gênova, I.A.O. Brito, A.R.F. Lima, M.A.L. Nobre,Structural characterization and Curie temperature determination of asodium strontium niobate ferroelectric nanostructured powder, J. SolidState Chem. 184 (5) (2011) 990–1000.

[16] S. Lanfredi, D.H.M. Gênova, I.A.O Brito, A.R.F. Limaand,M.A.L. Nobre, Erratum to Structural characterization and Curie tempera-ture determination of a sodium strontium niobate ferroelectric nanos-tructured powder, [J. Solid State Chem. 184 (2011) 990–1000]; J. SolidState Chem. 190 (2012) 309.

[17] S. Lanfredi, G. Palacio, C.V. Colin, M.A.L. Nobre, Thermistor behaviourand electric conduction analysis of Ni-doped niobate ferroelectric: therole of multiple β parameters, J. Phys. D 43 (2012) 435302.

[18] M.A.L. Nobre, E.R. Leite, E. Longo, J.A. Varela, Synthesis and sinteringof ultra fine NaNbO3 powder by use of polymeric precursors, Mater. Lett.28 (1-3) (1996) 215–220.

[19] H.M. Rietveld, Line profiles of neutron powder-diffraction peaks forstructure refinement, Acta Crystallogr. 22 (1) (1967) 151–152.

[20] J.R. Carvajal. An introduction to the Program FullProff 2000, CEA/Saclay, France, 2008.

[21] G. Caglioti, A. Paoletti, F.P. Ricci, Choice of collimators for a crystalspectrometer for neutron diffraction, Nucl. Instrum. 3 (4) (1958) 223–228.

[22] S. Lanfredi, C.X. Cardoso, M.A.L. Nobre, Crystallographic properties ofKSr2Nb5O15, Mater. Sci. Eng. B 112 (2–3) (2004) 139–143.

[23] J.M. Jehng, I.E. Wachs, Structural chemistry and Raman spectra ofniobium oxides, Chem. Mater. 3 (1) (1991) 100–107.

[24] F.I. Farrel, V.A. Maroni, T.G. Spiro, Vibrational analysis for Nb6O198-

and Ta6O198- and the the Raman intensity criterion for metal–metalinteraction, Inorg. Chem. 8 (8) (1969) 2638–2642.

[25] M.T. Martinez, M.A. Callejas, A.M. Benito, M. Cochet, T. Seeger,A. Anson, J. Schreiber, C. Gordon, C. Marhic, O. Chauvet, J.L.G. Fierro,W.K. Maser, Sensitivity of single wall carbon nanotubes to oxidativeprocessing: structural modification, intercalation and functionalisation,Carbon 41 (12) (2003) 2247–2256.

[26] A.M. Puziy, O.I. Poddubnaya, A. Martinez-Alonso, F. Suarez-Garcia,J.M.D. Tascon, Synthetic carbons activated with phosphoric acid:I. Surface chemistry and ion binding properties, Carbon 40 (9) (2002)1493–1505.

[27] J. Matos, A. Garcia, T. Cordero, J.M. Chovelon, C. Ferronato, Eco-friendly TiO2-AC photocatalyst for the selective photooxidation of4-chlorophenol, Catal. Lett. 130 (3–4) (2009) 568–574.

[28] J. Matos, A. García, P.S. Poon, Environmental green chemistry applica-tions of nanoporous carbons, J. Mater. Sci. 45 (2010) 4934–4944.

[29] J. Matos, J. Laine, J.M. Herrmann, Synergy effect in the photocatalyticdegradation of phenol on a suspended mixture of titania and activatedcarbon, Appl. Catal. B 18 (1998) 281–291.

[30] T. Cordero, C. Duchamp, J.M. Chovelon, C. Ferronato, J. Matos,Influence of L-type activated carbons on photocatalytic activity ofTiO2 in 4-chlorophenol photodegradation, J. Photochem. Photobiol. A191 (2–3) (2007) 122–131.

S. Lanfredi et al. / Ceramics International 40 (2014) 9525–95349534

[31] J. Matos, J. Laine, J.M. Herrmann, Effect of the type of activated carbonson the photocatalytic degradation of aqueous organic pollutants byUV-irradiated titania, J. Catal. 200 (2001) 10–20.

[32] J. Matos, J. Laine, J.M. Herrmann, D. Uzcategui, J.L. Brito, Influence ofactivated carbon upon titania on aqueous photocatalytic consecutive runsof phenol photomineralization, Appl. Catal. B 70 (1–4) (2007) 461–469.

[33] T. Cordero, J.M. Chovelon, C. Duchamp, C. Ferronato, J. Matos, Surfacenano-aggregation and photocatalytic activity of TiO2 on H-type activatedcarbons, Appl. Catal. B 73 (2007) 227–235.

[34] L.F. Velasco, J.B. Parra, C.O. Ania, Role of activated carbon features onthe photocatalytic degradation of phenol, Appl. Surf. Sci. 256 (17) (2010)5254–5258.

[35] L.F. Velasco, V. Maurino, E. Laurenti, C.O. Ania, Light-induced generationof radicals on semiconductor-free carbon photocatalysts, Appl. Catal. A 453(1) (2013) 310–315.

[36] J. Matos, A. García, L. Zhao, M.M. Titirici, Solvothermal carbon-dopedTiO2 photocatalyst for the enhanced methylene blue degradation undervisible light, Appl. Catal. A 390 (1–2) (2010) 175–182.

[37] J. Matos, M. Rosales, A. García, C. Nieto-Delgado, J.R. Rangel-Mendez,Hybrid photoactive materials from municipal sewage sludge for thephotocatalytic degradation of methylene blue, Green Chem. 13 (12)(2011) 3431–3439.

[38] T.J. Bandosz, J. Matos, M. Seredych, M.S.Z. Islam, R. Alfano, Photo-activity of S-doped nanoporous activated carbons: A new perspective forharvesting solar energy on carbon-based semiconductors, Appl. Catal. A445–446 (1) (2012) 159–165.

[39] J. Matos, E. García-López, L. Palmisano, A. García, G. Marci, Influenceof activated carbon in TiO2 and ZnO mediated photo-assisted degradationof 2-propanol in gas–solid regime, Appl. Catal. B 99 (2010) 170–180.