en insertion in titanium dioxidefor optimal photocatalytic degradation of atrazine
Emy Marlina Samsudin,a Sharifah Bee Abd Hamid,*a Joon Ching Juan,a
Wan Jefrey Basirun,a Ahmad Esmaielzadeh Kandjanib and Suresh K. Bhargavab
Introducing defects into the intrinsic TiO2 structural framework with nitrogen enhanced the photocatalytic
response towards the degradation of atrazine, as compared to undoped TiO2. Both catalysts, which were
prepared in an analogous manner, demonstrated high crystallinity and anatase phase dominant with well
defined {101} facets, which serves as a pioneer platform for good photocatalytic activity. The
introduction of nitrogen increased the stability of the crystal structure which leads to the formation of
pure active anatase phase. Although the optical response was shifted towards the visible region, initiated
by the formation of new absorption defects and interstate energy levels, the chemical state of nitrogen
in the doped TiO2 controls the overall catalyst photoreactivity. In this study, it was found that the surface
area and degree of band gap reduction played a lesser role for photocatalysis enhancement, although
they partly contributed, than the concentration of surface charge traps and the type of structural
framework formed during nitrogen incorporation. The enhancement in the photocatalytic degradation of
atrazine clearly was influenced by the loading and nature of the nitrogen dopant, which in turn,
governed the types of chemical and optical properties of the final catalyst product.
Introduction
Since the green revolution introduced by Nobel laureate NormalBorlaug in 1968, the usage of synthetic chemicals inmaximizingcrop productions increased rapidly and made a signicantimpact towards decreasing the prevalence of starvation world-wide. However, despite the positive changes in sustainingglobal needs, the green revolution has resulted in unintendedserious public health and environmental issues.1 Atrazine, an s-triazine group containing compound, is a herbicide that is mostcommonly used in Malaysia to control the growth of broad leafand grassy weeds2 (Fig. 1). Due to the stable s-triazine ring ofatrazine, it has poor biodegradability and great potential toleach through the soil and enter the ground water system.3
According to McMurray et al.,4 atrazine has been repeatedlyfound in drinking water supplies and has been detected abovethe recommended level of 0.1 ppb throughout Europeancountries and the United States.
The technologies for water pollutant abatement, whichinclude absorption, ultrasonic destruction, microbial, chemicaloxidation and ltration treatments are effective to a certaindegree; however, they lack the ability to degrade very lowpollutant concentrations. Atrazine is usually found in part per
nter, University of Malaya, 50603 Kuala
.edu.my; Fax: +60 379676956; Tel: +60
l Chemistry, RMIT University, Melbourne
hemistry 2015
billion levels in water sources and can be efficiently degradedinto its harmless intermediates via photocatalysis. Unlikeadsorption, the hazardous pollutants are destroyed rather thanbeing transferred to another phase.5 Photocatalysis also offers amore economical approach as there is no requirement for wastedisposal or the usage of a large amount of energy, and it is lesssensitive towards environmental changes, i.e. the pH ofwastewater.
The photocatalytic degradation of atrazine using 2.0 g L�1 ofTiO2 irradiated under a solar lamp was intensively studied byPellizzetti et al.3. The photocatalytic efficiency was proven as ittook only 10 minutes to degrade 2 ppb of atrazine to less than0.1 ppb. However, a longer time was required for the completemineralization of the intermediate products. Following thisdiscovery, many have attempted to degrade atrazine using TiO2
nanoparticulates. C. Minero et al.6 evaluated the photocatalyticdegradation of 10 ppm atrazine using a 0.2 g L�1 TiO2
suspension, and complete degradation was achieved in less
a Source of N from triethylamine precursor. b Amount of N successfully doped into TiO2 (EDX analysis). c Amount of N successfully doped into TiO2(CHNS analysis).
RSC Advances Paper
Publ
ishe
d on
15
May
201
5. D
ownl
oade
d by
Uni
vers
ity o
f M
alay
a on
18/
05/2
015
03:1
3:29
. View Article Online
than 120 min. I. K. Konstantinou et al.7 observed partialdegradation of 10 ppm atrazine using 0.1 g L�1 TiO2 suspen-sions with a half-life (t1/2) of 18.6 min. O. Zahraa et al.8 degraded10 ppm of atrazine using 1 g L�1 TiO2 suspension, which took200 min for the complete removal of atrazine. C. A. Ruslimieet al.9 evaluated the degradation of 5 ppm atrazine using 0.2 gL�1 TiO2, which took more than 240 min for 70.6% atrazinedegradation. The rate of photocatalytic degradation of atrazinediffers in each experiment, which could be due to the differenttypes of reactor congurations and TiO2 synthesis methods.Hence, an accurate evaluation of the photodegradationprocesses could not be done. In all of these works, completemineralization of atrazine was not observed, and a nal stableend product of cyanuric acid was obtained.
One of the main factors that control the rate of photo-catalytic degradation is the physicochemical properties of thecatalyst used.10,11 The application of intrinsic TiO2 is limited dueto the wide band gap, which requires a substantial amount ofphotons for electron/hole excitation.12 Furthermore, theabsorption efficiency of intrinsic TiO2 covers only 4–5% of theUV region, leaving the major constituents of solar light unhar-nessed.13 Inducing defects in the TiO2 lattice using non-metalions eliminates the drawbacks of doping using transitionmetals, such as multiple charge carrier recombination centerswithin the band gap due to deep localized d states.14,15 Nitrogendoped TiO2 was introduced by Asahi et al.,14 where TiO2�xNx
lms were prepared by sputtering TiO2 in a N2 (40%)/Ar gasmixture and subsequently annealed at 550 �C in N2 gas for 4hours. The results showed the ability of the TiO2�xNx lms toabsorb light below 500 nm, which was not observed for nakedTiO2 lms. Furthermore, the hydrophilicity of the TiO2�xNx
lms was enhanced, an important criteria for a good
Fig. 2 FESEM images of (a) N0–TiO2, (b) N1–TiO2, (c) N3–TiO2 and (d)
44042 | RSC Adv., 2015, 5, 44041–44052
photocatalyst system.16 Many works in N–TiO2 have since beenreported13,17–21 concluding the positive effect of nitrogen incor-poration into TiO2 in terms of particle homogeneity, widerabsorption of solar light, narrowing of the band gap, alternativepathways for electron/hole recombination, retardation of crystalphase changes and enhanced photocatalytic degradationactivity.
It is undeniable that the successive preparation of N–TiO2
photocatalysts has generated successful visible-energy lightconversion and improved its photocatalytic response indegrading organic pollutants. However, there is still a gap inunderstanding the factors which inuence the photoactivity,especially by varying the concentration of nitrogen that entersthe TiO2 framework and creating either coordinated or substi-tutional doping. The degree of impurity imposed by introducingdifferent loadings of nitrogen was studied in this work, and wasconnected carefully to the photocatalytic degradation of atra-zine. In addition, very little of the aforementioned work evalu-ated the photodegradation of atrazine using N–TiO2, which isdiscussed in this paper, followed by the degradation pathwayand the nal formation of the stable end product of cyanuricacid.
ExperimentalMaterials
Titanium(IV) isopropoxide (TTIP, 97%), hydrochloric acid (37%),ethanol (95%), anhydrous ethanol (99%), triethylamine (99%)and atrazine (98%) were purchased from Sigma Aldrich. Allchemicals were used as received without further pre-treatment.Milli-Q deionized water was used throughout the wholeexperiment.
Five different nitrogen loading contents of N–TiO2 photo-catalyst were synthesized using amodied sol–gel method22 andtitanium(IV) isopropoxide (TTIP) as the metal precursor, pre-dissolved in anhydrous ethanol. A solution of deionized waterand hydrochloric acid in the molar ratio of 15 : 0.5 was addeddropwise to obtain a clear sol, followed by the addition of tri-ethylamine in molar ratios from 0.5 to 2.5. The sol was le toage prior to solvent evaporation at 80 �C overnight. The driedsol–gels were further decomposed at 500 �C for 6 hours under aconstant air ow rate of 20 mL min�1 to obtain white N–TiO2
powder. A similar method without the addition of the nitrogenprecursor was employed for preparing un-doped TiO2.
Characterization of N–TiO2
The surface morphology and elemental composition of thesuccessfully nitrogen doped TiO2 were observed with a FESEMQuanta FEI 200F and energy dispersive X-ray spectroscopy(INCA Soware). A CHNS elemental analyzer (Perkin Elmer,2400 series II) was also used to provide secondary informationon the content of nitrogen successfully doped into TiO2, using
Fig. 4 HR-TEM images of TiO2 powders for (a) N0–TiO2 and (b) N1–TiO
sulfamethazine (Sigma Aldrich) as the standard. The specicBET surface area and porosity were measured using a multi-point nitrogen adsorption–desorption analyzer at a relativepressure range of 0.01–0.90 P/Po. The crystallinity and crystalphase of the samples were characterized using powder X-raydiffraction (XRD, Bruker AXS D8) using Cu Ka radiation. Highresolution transmission electron microscopy (HR-TEM, JEM2100-F) was performed at an accelerating voltage of 200 kV andwas used to determine the microstructure of the catalyst. Thesamples were examined using 40 kV and 40 mA with a step sizeof 0.002 from 10� to 80�. The Raman spectra and photo-luminescence (PL) measurements were performed using aLabRam confocal Raman microscope. Samples were excitedusing the 325 nm line of a continuous He–Cd laser at roomtemperature. Analysis of the functional groups and types ofchemical bonds was performed using Fourier TransformedInfra-red, FTIR Bruker Vertex 80/80v, from 4000 to 400 cm�1.The absorption edge of the prepared photocatalyst wasmeasured using an ultraviolet–visible spectroscope (Agilent,Cary 100) with a diffuse reectance accessory. X-ray photoelec-tron spectroscopy (XPS) surface analyses were carried out using
2 and (c) the lattice spacing of the anatase phase for N1–TiO2.
a at gold surface (Si/10 nm Ti/200 nm Au) as the substrate andreference. A Thermo Scientic K-alpha instrument with an un-monochromatized Mg Ka radiation (photon energy 1253.6 eV)source and a vacuum better than 10�9 Torr, as well as a spectralresolution of 0.1 eV, was used for XPS studies. The XPS corelevels were aligned to the C 1s binding energy (BE) of 285 eV.The weight fractions of the crystal phase were calculated usingthe Spurr equation, fA ¼ 1/(1 + 1.26 � IR/IA), where fA is theweight fraction of anatase, fR ¼ 1 � fA is the weight fraction ofrutile, IA is the intensity of the most intense anatase phase peak{101}, IR is the intensity of the most intense rutile phase peak{110} and 1.26 is the scattering coefficient.23 For undoped TiO2,IR and IA are 5.57 and 108.44, respectively.
Photocatalytic reactor and degradation activity
A custom-built stirred tank photo-reactor (STR) consisted of six250 mL quartz tubes of equal dimensions were systematicallyarranged to receive equal light intensity. The source of light wasa xenon lamp (150 W) with a UV light block lter (l < 420 nm),positioned in the center of the reactor. The mass transferbetween the photocatalyst and atrazine was enhanced viaturbulent mixing at 450 rpm. Undoped TiO2 and N–TiO2 powderwere used to degrade an initial atrazine concentration of 0.5 mgL�1. The initial molar ratios of nitrogen in the doped TiO2 were0.5, 1.0, 1.5, 2.0 and 2.5, with an initial catalyst loading of
Fig. 5 Nitrogen adsorption–desorption linear isotherm (type IV) for (a) N
44044 | RSC Adv., 2015, 5, 44041–44052
0.5 g L�1. Prior to the experiment, the photocatalyst was le inthe aqueous suspension in darkness for 70 minutes to achieveadsorption equilibrium between the catalyst and the atrazinemolecules. The experiment was performed in batch operationmode with a constant temperature maintained at 25 � 2 �C.Throughout the experiments, 5.0 cm3 liquid samples were takenat various time intervals. The resultant aqueous solution wasltered prior to UV-vis analysis (Perkin Elmer, Lambda 35). Theaverage wavelength of atrazine used for UV-vis analysis was222.3 nm, with a calibration coefficient, r2, of 0.9998.The degradation efficiency, h, was described using the equation[h ¼ (co � c)/co � 100%¼ (Ao � A)/Ao � 100%]; where co and c/Aoand A represent atrazine concentration/absorption intensitiesinitially and aer photocatalytic activity for a certain time.
Intermediate identication – LC/MS/MS
The atrazine degradation intermediates were determined usingtriple quadrupole liquid chromatography mass spectroscopy,LCMS/MS (Agilent 6400 Series), using electrospray ionization (ESI)and a Zorbax StableBond-C18 reversed phase column (4.6 mm ID� 250 mm, 5 mm). A full scan mode (1 spectrum per s) with a owrate of 0.3 mL min�1 and runtime of 15 minutes was used tocollect the MS and MS/MS data. The mobile phase consisted ofltered deionized water and acetonitrile in a 9 : 1 volume ratio,and the analysis was done under isocratic elution at pH 3.0.
EDX, as CHNS measures the total bulk nitrogen while EDXmeasures the nitrogen present on the surface of N–TiO2. In bothof the prepared catalysts, no change of color was observed, andthey remained as white powders. This implies that coordinatednitrogen doping or chemisorbed g-N2 is probably present inTiO2, as color changes are generally related to the red shi ofthe absorption edge due to substitutional doping.24 The FESEMimages of the undoped and doped TiO2 possessed agglomeratesof spherical particles with an average particle diameter of 27 nmto 37 nm (Fig. 2). The EDX spectra can be observed in Fig. 3.Although it was observed that the N-precursor did not inuencethe size and shape of the particles much, there was a slightincrease of particle size for the N–TiO2 prepared at a lowernitrogen loading. This could be partially due to the larger size ofthe nitrogen atomic radius as compared to the oxygen atomicradius that occurred during dopant incorporation. This obser-vation contradicts with the work done by M. Sathish et al.,18
where a signicant decrease in the particle size was obtainedaer nitrogen doping. At higher nitrogen loadings, less particleagglomeration appeared; thus the homogeneity and dis-persibility of the particles were enhanced. This is clearlyrevealed in the HR-TEM images (Fig. 4) where N1–TiO2 appearsmore dispersed as compared to N0–TiO2. Both samples showedirregular crystal size ranging from 11 nm to 22 nm, with largercrystallite size observed in the doped TiO2. The BET specicsurface areas and porosities presented in Table 2 showed thatnitrogen doping had less inuence on the change of porosity, ascompared to the specic surface area. The nitrogen absorption–desorption isotherms in Fig. 5 showed a showed a stronginteraction of isotherm type IV and a narrow distribution ofuniform mesoporous pores in the range of 2 to 50 nm.25
Crystallinity and crystal phase. XRD analysis was used todetermine the crystallinity, crystal phase and weight fraction ofeach phase. The undoped TiO2 consisted of two crystal phases,namely anatase and rutile, with a mass ratio of 94 : 6 (Table 2).Interestingly, all doped TiO2 samples showed predominance ofthe anatase phase and enhanced crystallinity under allsynthesis conditions. A major peak corresponding to the {101}anatase phase was observed at an angle of 25�, followed byminor peaks at 37�, 47�, 54�, 55�, 62�, 69�, 70� and 75�, based onJCPDS 731764 (Fig. 6). During the sol–gel process, the growthrate of the crystal exceeded the nucleation rate due to theaddition of aqueous amine solution, which affects the overallpH of the solution. The presence of nitrogen in TiO2 inhibits thecondensation of the spiral chains of anatase TiO6 octahedra tolinear chains of rutile TiO6 octahedra. This can be seen from thedisappearance of the rutile major plane {110} at 27� aer theaddition of nitrogen. A similar observation was reported20 wherethe addition of nitrogen improved the thermal stability of thecatalyst by increasing the temperature required for rutile phaseformation. It was further observed that the presence of nitrogenaffects the full width at half maximum (FWHM), which is anindication of crystal size changes and possible lattice distortion(Fig. 7). At low nitrogen loading, there was a signicant nar-rowing of the FWHM, and the crystal size increased from 23 nmto 34 nm. The drastic change in the crystal size, however, didnot affect the lattice spacing as shown in Table 2.26 All crystallite
Fig. 8 Raman spectra of (a) N0–TiO2, (b) N1–TiO2, (c) N2–TiO2, (d)N3–TiO2, (e) N4–TiO2 and (f) N5–TiO2.
Fig. 9 Localized Raman spectra of (a) N0–TiO2, (b) N1–TiO2, (c) N2–TiO2, (d) N3–TiO2, (e) N4–TiO2 and (f) N5–TiO2.
RSC Advances Paper
Publ
ishe
d on
15
May
201
5. D
ownl
oade
d by
Uni
vers
ity o
f M
alay
a on
18/
05/2
015
03:1
3:29
. View Article Online
sizes were calculated using Debye-Scherer’s formula27 using thehighest intensity peak of the predominant phase.
Chemical composition. Raman spectroscopy was employedto study the crystalline structure of the undoped and dopedTiO2. All the prepared catalysts showed a signicant anatasephase with major bands at 144 cm�1 (Eg), 395 cm�1 (B1g), 515cm�1 (A1g) and 637 cm�1 (E1g) (Fig. 8). A weak, but apparentRaman band of anatase was shown at 195 cm�1 (Eg). In all thedoped TiO2 samples, no signicant rutile band at 446 cm�1 (Eg)was observed, which supports the single crystal phase observedin the X-ray diffraction (XRD) spectra. Although 6 wt% of therutile phase was present in the undoped TiO2, no signicantrutile band can be seen in the Raman spectra. This shows thatRaman spectroscopy is suitable to identify the presence of TiO2
but not for distinguishing the different phases of anatase andrutile, especially at a low mass fraction. G. Yang et al.28 preparedN–TiO2 at different N : Ti atomic ratios using a solvothermal
44046 | RSC Adv., 2015, 5, 44041–44052
method and obtained predominant anatase bands in theRaman spectra. It was observed that the amount of nitrogenloaded into TiO2 inuences the size of the particles in a linearproportion. Larger particle size caused a blue shi towardslower band values in the Raman spectra. Interestingly, in thiswork, the Raman shi patterns were similar to doped TiO2
prepared with a nitrogen loading below 11 wt% (Fig. 9).However, as the nitrogen loading increased, a Raman red shiwas observed, indicating smaller particle size. This observationis in good agreement with those shown in the XRD data. FTIRanalysis was used to analyze the presence of nitrogen in TiO2
prior to doping at different loadings. In Fig. 10a and b, similarspectra were observed for both the undoped and doped TiO2
around 2363–2360 cm�1 and 700–400 cm�1, which correspondto absorbed CO2,29 Ti–O stretching and the Ti–O–Ti bridgingstretchingmode.30 The enlarged FTIR spectrum from 1700 cm�1
to 1200 cm�1 revealed signicant differences between thecatalysts (Fig. 10c). The doped TiO2 possessed multiple peakswithin this region, which correspond to surface absorbednitrogen at 1463–1384 cm�1 (ref. 28) and lattice nitrogen in theTiO2 at 1632 cm�1, 1546 cm�1, 1338 cm�1 and 1255 cm�1.31
Chemical state. XPS analysis was used to determine thesurface chemistry of the prepared photocatalysts. In Fig. 11, nonitrogen signal was detected in the undoped TiO2. The bindingenergy, Eg, of the N 1s spectra was observed at 400.4 eV, and noother Eg was present between 395 eV and 399 eV or above 400.4eV. From previous work, interstitial nitrogen showed Eg valuesof around �400 to 402 eV,13,32 399.6 eV (ref. 33) and 399.8 eV.34
In addition, molecularly chemisorbed g–N2 shows an Eg ofaround �400 to 402 eV.35,36 The nature of the N 1s present at anEg of �400 to 402 eV is still contentious, as to whether thenitrogen is doped interstitially, present as chemisorbed N2 orboth. The differences of Eg and XPS intensities in each reportedwork were inuenced by the selection of dopant20 and catalystpreparation routes.33 In this work, triethylamine was used as thenitrogen precursor source. J. Ananpattarachai et al.20 used asimilar dopant in preparing N–TiO2 and obtained an Eg of 402.5eV, which was assigned to nitric oxide or nitrogen monoxide(NO) and nitrite (NO2
�), corresponding to interstitially dopednitrogen in TiO2 due to the formation of localised states with p
characteristics. L. Hu et al.35 also incorporated triethylamine asthe N-precursor in synthesizing N–TiO2. Interestingly, it wasreported that the point of dopant introduction during thehydrolysis process determines the nal state of the N 1s formedin TiO2. Molecularly chemisorbed N2 is generally present if thecrystallization process of TiO2 occurs before the introduction ofnitrogen and shows an N 1s Eg of around 400.5 and 401.4 eV,while substituted nitrogen results when the introduction ofnitrogen occurs rst. Broad N 1s XPS emission similar to thatshown in Fig. 11 was also observed in other reported work,37–39
however a strong correlation could not be made as differentpreparation procedures were employed. B. Viswanathan et al.40
reported that low nitrogen loading in TiO2 generally results in N1s spectra with an Eg of �396–397 eV, while at high nitrogenloading it is about �400 eV, however no quantication wasmentioned. In this work, the reported Eg of the N 1s peak at�400.4 eV is attributed to the interstitial nitrogen in the TiO2
Fig. 10 FTIR spectra of (a) N0–TiO2, (b) N1–TiO2 and (c) enlarged FTIR spectrum of (b) from 1200 cm�1 to 1700 cm�1.
Paper RSC Advances
Publ
ishe
d on
15
May
201
5. D
ownl
oade
d by
Uni
vers
ity o
f M
alay
a on
18/
05/2
015
03:1
3:29
. View Article Online
matrix with possible Ti–O–N or Ti–N–O linkages or chem-isorbed molecular g-N2. For the latter, TiO2 was doped at a highnitrogen loading, however, no signicant change in the crystalphase and lattice spacing was observed (Table 2). This indicatesthe presence of chemisorbed molecular g-N2 on the surface ofTiO2. As the amount of nitrogen doped into TiO2 increased, thecrystallinity of TiO2 decreased. Observations made from theFTIR spectra (Fig. 10) support the presence of probable Ti–O–Nor Ti–N–O linkages, as a clear demarcation of lattice nitrogen isobserved, especially in the magnied FTIR spectrum from1200–1700 cm�1. The N–H bond and NOX are not present in N1–TiO2 due to the absence of an Eg around 398.7 eV and above 403
Fig. 11 N 1s, Ti 2p and O 1s XPS spectra of N0–TiO2 and N1–TiO2.
eV respectively.33 This is further supported by the FTIR spec-trum in Fig. 10c as no frequency corresponding to ammonia(N–H) was present. The probability of N–Ti–O lattice substitu-tion can be ruled out as no Eg was observed below 397.5 eV, acharacteristic of substituted nitrogen.20 Futhermore, a slightnegative shi of the Eg in Ti 2p3/2 between the undoped TiO2
and N–TiO2 was observed in previous work when nitrogensubstituted for the oxygen atom in a TiO2 lattice.41 The changeof Eg is due to the difference of electronegativity of the nitrogenand oxygen atoms in a N–Ti–O linkage, causing partial electrontransfer from N to Ti as oxygen is more electronegative thannitrogen, thus increasing the electron density of Ti. However,
Fig. 12 (a) Absorption spectra and (b) localized absorption spectra for all the prepared photocatalysts at different nitrogen loadings.
Fig. 13 Photoluminescence spectra for all the prepared photo-catalysts at different nitrogen loadings.
RSC Advances Paper
Publ
ishe
d on
15
May
201
5. D
ownl
oade
d by
Uni
vers
ity o
f M
alay
a on
18/
05/2
015
03:1
3:29
. View Article Online
this was not observed in this study, hence the probability ofsubstituted nitrogen in the TiO2 lattice is again eliminated.Based on the Ti 2p spectra, the Eg of the photoelectron peakillustrates the existence of Ti4+ species in the TiO2 nano-structures.42 The two deconvoluted O 1s peaks remainedunchanged for both undoped and doped TiO2 at 530.0 eV and531.7 eV. The former represents metallic oxide (Ti–O) and thelatter represents surface hydroxyl, OOH and is in agreement withthe IR frequencies observed for N0–TiO2 and N1–TiO2 at 3437.8cm�1 and 3442 cm�1 respectively. No absorbed H2O at �532.7eV was detected in the XPS peaks.43 The presence of nitrogen inthe TiO2 increased the number of hydroxyl sites and enhancedthe hydrophilicity of the catalyst. The hydroxyl peaks in both theFTIR (Fig. 10) and XPS (Fig. 11) analyses were more signicantin N1–TiO2 as compared to N0–TiO2. In agreement with T.Morikawa et al.,16 the addition of nitrogen increased thewettability of the catalyst surface, thus enhancing the photo-catalytic activity.
44048 | RSC Adv., 2015, 5, 44041–44052
Optical properties. The UV-vis spectra of the undoped anddoped TiO2 catalysts prepared at different nitrogen loadings arepresented in Fig. 12. The corresponding band gap was obtainedusing the Kubelka–Munk function34 calculated using the diffusereectance spectrum, with an indirect band gap coefficient of0.5 (ref. 44) (Table 2). A signicant increase of light absorptionat lower photon energy levels was observed for all the dopedTiO2 samples, extending the absorption coverage towards 400–550 nm of the visible light region. This is due to the additionalimpurity level created within the band gap, thus causing a shiof the Fermi level closer to the conduction band and conse-quent narrowing of the band gap. This claim is further sup-ported by the work of Xiang et al.,25 where from rst principledensity functional theory (DFT) calculations, an interstitial N-precursor can induce local states above the valence band andis responsible for the visible light response. According to B. Tianet al.,29 the incorporation of nitrogen in the TiO2 matrix forms anarrow N 2p band, which overlaps the with O 2p orbital,promoting greater electron mobility from the valence band tothe conduction band during photoexcitation. However, nocontinuum interstate overlapping was observed in this work, asthe XPS data clearly shows no presence of substituted nitrogenin the TiO2 lattice forming Ti–N–Ti linkages, but merely inter-stitial doping and chemisorbed g-N2. Apart from that, theformation of oxygen vacancies to compensate for the overallcharge balance played an important role towards the narrowingof the catalyst band gap as well. The formation of oxygenvacancies in doped TiO2 enhances the wider light absorptionsabove 500 nm (ref. 45) and is in good agreement with the blueshi of the Ti 2p binding energy obtained in Fig. 11. Further-more, the work done by H. Wu et al.46 suggests that theformation of oxygen vacancies in the presence of nitrogen iseasier as compared to naked TiO2. Despite the remarkableimprovement of photonic efficiency, the fate of the electronsand holes created during catalyst photoexcitation determinesthe overall photoactivity mechanism. The incorporation ofnitrogen into TiO2 shows broad visible emission around530 nm, which consequently impacts the electron/hole
recombination rate (Fig. 13). At high nitrogen loading, thephotoluminescence intensity increases due to the increasingnumber of oxygen vacancies and defect sites. Excessive oxygenvacancies act as electron/hole recombination centres, thusreducing the amount of active radicals generated and giving riseto high photoluminescence intensity.29 It is observed thatelectron/hole recombination is inhibited at low nitrogenloading due to fewer charge trapping sites, hence facilitatingbetter photocatalytic activity. In addition, the PL peaks at 450nm and 500 nm indicate the presence of defects in the dopedTiO2, which could act as new adsorption centers47 in the visiblerange, while interstate overlap of N 2p with O 2p that could beresponsible for the visible light activity was not observed in thisstudy. Furthermore, the PL spectra showed broad peaks, indi-cating the presence of oxygen vacancies, demonstrating that notall of the doped nitrogen participates in forming Ti–O–N or Ti–N–O linkages, but also exists as amorphous nitrogen. Thisobservation is supported by the absence of changes in thecrystal structure as shown in Fig. 6.
Photocatalytic activity
Decomposition of atrazine under visible light. Prior to thephotocatalytic degradation, the catalyst was le in the dark toreach absorption–desorption equilibrium (Fig. 14). In thisstudy, 70 min was sufficient to reach the absorption–desorptionsaturation point. The surface area and porosity of the catalystplayed an important role in controlling the amount of atrazineabsorbed, as well as the surface charge of the catalyst andsubstrate at the given pH. Relatively, higher absorption ofatrazine was observed for the undoped TiO2 and highestloading of doped TiO2 due to the smaller particle size andgreater porosity (Table 2). However, these properties are notsufficient to justify the strength of a photocatalyst, and manyother factors are involved, such as morphology, crystal structureand optical and electronic properties.48 In addition, the catalystsurface charge is an important factor and is predominantlycontrolled by pH. The initial measured pH of the atrazinesolution was 7.5, and no further adjustment of the pH was donethroughout the experiment. Under these conditions, the chargeof both catalysts is presumed to be negative as the isoelectricpoint of TiO2 is from 6.25 to 6.90.24 The attraction between thenegatively charged catalyst and the atrazine substrate wasfavored, as atrazine is protonated by the presence of H+, thusfacilitating the absorption of atrazine. In return, this provided agood combination for the photocatalytic activity as the gener-ated electron/hole pairs on the catalyst surface could readilyreduce and oxidize the absorbed atrazine into its degradationintermediates. As observed, the presence of nitrogen in TiO2
signicantly improved the photocatalytic degradation activity ofatrazine, when compared to the undoped TiO2 (Fig. 15) undervisible light irradiation. As shown in Fig. 15, the degradationrate constants are dependent on the amount of nitrogen loadedonto the TiO2, however, it does not represent a linear mathe-matical function. The function ts a rst order of reaction witha coefficient, R2 ¼ 0.98, whereby the rate of degradation isdependent on the loading of nitrogen (Fig. 16). The highest
Fig. 17 Proposed degradation pathway of atrazine using TiO2 doped with nitrogen.
RSC Advances Paper
Publ
ishe
d on
15
May
201
5. D
ownl
oade
d by
Uni
vers
ity o
f M
alay
a on
18/
05/2
015
03:1
3:29
. View Article Online
removal of atrazine was achieved in less than 150 min at thelowest nitrogen loading in doped TiO2. Higher nitrogen loadingdoes not necessarily favor the photocatalytic reaction49 due tothe rapid electron/hole recombination as represented in Fig. 13,although the absorption of atrazine in the dark was greater.This phenomenon explains the necessity of slow electron/holerecombination as well as enhanced photon absorption for agood photocatalysis process. Subsequently, the undoped TiO2
showed the weakest performance as the absorption of photonsin the visible region is very poor, hence less light can beabsorbed to generate active radicals for photodegradation.
Degradation mechanisms. The intermediates of the photo-degradation of atrazine were studied aer 150 min of reactionusing 0.5 g L�1 doped TiO2. Based on LCMS/MS peaks, a total of8 degradation intermediates were observed, which enables theplotting of a potential degradation pathway (Fig. 17). Theobtained peak atm/z 215.7 showed the presence of atrazine evenaer 150 min of photodegradation activity, and indicates theincomplete removal of atrazine. The degradation pathwayinvolves the dehalogenation of chlorine at position two with ahydroxyl group, oxidation of the alkyl side chain, furtherdealkylation and deamination. The degradation processesinclude the formation of hydroxyatrazine, deethylatrazine,
Table 3 Atrazine degradation intermediates by LCMS/MS
desethylhydroxyatrazine, deisopropylatrazine, deethyldeisopro-pylatrazine, ammelide and ammeline(3,4,6), which are listed inTable 3. The degradation pathway leads to a nal amino groupdisplacement with hydroxyl and the formation of a stableintermediate of cyanuric acid at m/z 129.1. I. K. Konstantinouet al.5 referred to the photodegradation mechanism of s-triazineas a photo-Kolbe decarboxylation process, for which similarformation patterns of alcohol, aldehyde and acid derivativeswere observed .
Conclusion
Undoped and nitrogen doped TiO2 were synthesized via a sol–gel technique using titanium(IV) isopropoxide and triethylamineas the Ti and N precursors sources, respectively. A singleanatase phase with less particle agglomeration was observed forall the doped TiO2 samples. There was no dramatic change inthe particle size of the doped TiO2 as the atomic radius of anitrogen atom is 6% larger than the atomic radius of oxygen.Based on the FTIR and XPS data, the hydrophilicity of the TiO2
surface increased in the presence of nitrogen, which is bene-cial to any photocatalytic degradation of water pollutants.Nitrogen is doped interstitially forming Ti–O–N or Ti–N–O
linkages, and induces local states 0.23 to 0.26 eV above the valenceband, which was responsible for the visible light responsebetween 400 to 550 nm and the effective narrowing of the bandgap. In addition, the absence of signicant change in the crystalstructure, coupled with the N 1s XPS data and high nitrogenloading, conrms the presence of chemisorbed g-N2 as well. In allthe prepared photocatalysts, the doped TiO2 with the lowestnitrogen loading yielded the highest rate of atrazine degradation,and superseded the performance of the undoped TiO2. At lownitrogen loading, the number of electron–hole trapping sites arereduced, which results in a higher concentration of active radicalsfor atrazine degradation and mineralization. The degradation ofatrazine followed a rst order rate of reaction, where the degra-dation activity is greatly inuenced by the atrazine concentration.The degradation pathway includes dehalogenation of chlorine atposition two with a hydroxyl group, oxidation of the alkyl sidechain, further dealkylation and deamination to form the stablenon-toxic by-product, cyanuric acid.
Acknowledgements
This research was funded by High Impact Research Grant (HIR-F000032) and IPPP Postgraduate Research Grant (PG062-2013B).
References
1 D. Pimentel, Green revolution agriculture and chemicalhazards, Sci. Total Environ., 1996, 7(188), 86–98.
2 A. Ahmad, L. Tan and S. Abd Shukor, Dimethoate andatrazine retention from aqueous solution by nanoltrationmembranes., J. Hazard. Mater., 2008, 151(1), 71–77.
3 E. Pelizzeti, V. Mauriono, C. Minero, V. Carlin, E. Pramauro,O. Zerbinati and M. L. Tosato, Photocatalytic Degradation ofAtrazine and Other s-Triazine Herbicides, Environ. Sci.Technol., 1990, (20), 1559–1565.
4 T. McMurray, P. Dunlop and J. Byrne, The photocatalyticdegradation of atrazine on nanoparticulate TiO2 lms, J.Photochem. Photobiol., A, 2006, 182(1), 43–51.
5 I. K. Konstantinou and T. A. Albanis, Photocatalytictransformation of pesticides in aqueous titanium dioxidesuspensions using articial and solar light: intermediatesand degradation pathways, Appl. Catal., B, 2003, 42, 319–335.
6 C. Minero, E. Pelizzetti, S. Malato and J. Blanc, Large solarplant photocatalytic water decontamination: degradation ofatrazine, Sol. Energy, 1996, (5), 411–419.
7 I. K. Konstantinou, T. M. Sakellarides, V. A. Sakkas andT. A. Albanis, Photocatalytic Degradation of Selected s-Triazine Herbicides and Organophosphorus Insecticidesover Aqueous TiO2 Suspensions, Environ. Sci. Technol.,2001, 35(2), 398–405.
8 O. Zahraa, L. Sauvanaud, G. Hamard and M. Bouchy,Kinetics of atrazine degradation by photocatalytic processin aqueous solution, Int. J. Photoenergy, 2003, 05(14), 87–93.
9 C. A. Ruslimie, H. Razali and W. M. Khairul, Catalytic Studyon TiO2 Photocatalyst Synthesised via MicroemulsionMethod on Atrazine, Sains Malays., 2011, 40(8), 897–902.
10 Q. Chen, H. Shi, W. Shi, Y. Xu and D. Wu, Enhanced visiblephotocatalytic activity of titania–silica photocatalysts: effectof carbon and silver doping, Catal. Sci. Technol., 2012, 2(6),1213.
11 H. S. Kibombo, R. Peng, S. Rasalingam and R. T. Koodali,Versatility of heterogeneous photocatalysis: syntheticmethodologies epitomizing the role of silica support inTiO2 based mixed oxides, Catal. Sci. Technol., 2012, 2(9),1737.
12 Y. Alivov, V. Singh, Y. Ding, L. J. Cerkovnik and P. Nagpal,Doping of wide-bandgap titanium-dioxide nanotubes:optical, electronic and magnetic properties, Nanoscale,2014, 6(18), 10839–10849.
13 S. J. Darzi, A. R. Mahjoub and S. Sar, Visible-Light-ActiveNitrogen Doped TiO2 Nanoparticles Prepared by Sol–GelAcid Catalyzed Reaction, Iran. J. Mater. Sci. Eng., 2012, 9(3),17–23.
14 R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki and Y. Taga,Visible-light photocatalysis in nitrogen-doped titaniumoxides, Science, 2001, 293(5528), 269–271.
15 A. S. Weber, A. M. Grady and R. T. Koodali, Lanthanidemodied semiconductor photocatalysts, Catal. Sci.Technol., 2012, 2(4), 683.
16 T. Morikawa, R. Asahi, T. Ohwaki, K. Aoki, K. Suzuki, andY. Taga, Visible-light Photocatalyst-Nitrogen-dopedTitanium Dioxide, Research Report for R&D Review of ToyotaCRDL, 2005, vol. 40, no. 3, pp. 45–50.
17 M. Mrowetz, W. Balcerski, A. J. Colussi, M. R. Hoffmann,C. Fisica and V. Uni, Oxidative Power of Nitrogen-DopedTiO2 Photocatalysts under Visible Illumination, J. Org.Chem., 2004, 108(45), 17269–17273.
18 M. Sathish, B. Viswanathan, R. P. Viswanath andC. S. Gopinath, Synthesis, Characterization, ElectronicStructure, and Photocatalytic Activity of Nitrogen-DopedTiO2 Nanocatalyst, Chem. Mater., 2005, 17(10), 6349–6353.
19 Y. U. Huang, Z. Xuxu, Y. I. N. Zhongyi and T. A. O. Feng,Preparation of Nitrogen-doped TiO2 Nanoparticle Catalystand Its Catalytic Activity under Visible Light, Chi, 2007,15(6), 802–807.
20 J. Ananpattarachai, P. Kajitvichyanukul and S. Seraphin,Visible light absorption ability and photocatalyticoxidation activity of various interstitial N-doped TiO2
prepared from different nitrogen dopants, J. Hazard.Mater., Aug. 2009, 168(1), 253–261.
21 K. Zhao, Z. Wu, R. Tang and Y. Jiang, Preparation of HighlyVisible-Light Photocatalytic Active N-Doped TiO2
Microcuboids, J. Korean Chem. Soc., 2013, 57(4), 489–492.22 X. Chen, L. Liu, P. Y. Yu and S. S. Mao, Increasing solar
absorption for photocatalysis with black hydrogenatedtitanium dioxide nanocrystals, Science, 2011, 331, 746–750.
23 S. C. Padmanabhan, S. C. Pillai, J. Colreavy, S. Balakrishnan,D. E. Mccormack, T. S. Perova, Y. Gun, S. J. Hinder andJ. M. Kelly, A Simple Sol–Gel Processing for theDevelopment of High-Temperature Stable PhotoactiveAnatase Titania, Chem. Mater., 2007, 19(11), 4474–4481.
24 J. Sun, L. Qiao, S. Sun and G. Wang, Photocatalyticdegradation of Orange G on nitrogen doped TiO2 catalysts
under visible light and sunlight irradiation., J. Hazard.Mater., 2008, 155(1–2), 312–319.
25 Q. Xiang, J. Yu, W. Wang and M. Jaroniec, Nitrogen self-doped nanosized TiO2 sheets with exposed {001} facets forenhanced visible-light photocatalytic activity., Chem.Commun., 2011, 47(24), 6906–6908.
26 J.-Y. Bae, T.-K. Yun, K.-S. Ahn and J.-H. Kim, Visible-photoresponsive Nitrogen-Doped Mesoporous TiO2 Filmsfor Photoelectrochemical Cells, Bull. Korean Chem. Soc.,2010, 31(4), 925–928.
27 V. Chaudhary, A. K. Srivastava and J. Kumar, On the Sol–gelSynthesis and Characterization of Titanium OxideNanoparticles, Mater. Res. Soc., 2011, 1352, 10–24.
28 G. Yang, Z. Jiang, H. Shi, T. Xiao and Z. Yan, Preparation ofhighly visible-light active N-doped TiO2 photocatalyst, Chin.J. Chem. Eng., 2010, 2015(25), 5301.
29 B. Tian, Y. Qian, B. Hu, J. Sun and Z. Du, Synthesis andcharacterization of nitrogen-doped titanium dioxidenanomaterials derived from nanotube sodium titanateprecursor, J. Mater. Res., 2012, 27(18), 2408–2416.
30 A. A. Umar, M. Yusri, A. Rahman, S. K. Saad andM. M. Salleh, Effect of NH3 Concentration on thePerformance of Nitrogen doped TiO2 PhotoelectrochemicalCell, Int. J. Electrochem. Sci., 2012, 7, 7855–7865.
31 S. Livraghi, M. C. Paganini, E. Giamello, A. Selloni, C. DiValentin and G. Pacchioni, Origin of Photoactivity ofNitrogen-Doped Titanium Dioxide under Visible Light, J.Am. Chem. Soc., 2006, 128(21), 15666–15671.
32 A. V. Emeline, V. N. Kuznetsov, V. K. Rybchuk andN. Serpone, Visible-Light-Active Titania Photocatalysts: TheCase of N-Doped TiO2s—Properties and SomeFundamental Issues, Int. J. Photoenergy, 2008, (258394), 1–19.
33 J. Wang, D. N. Tafen, J. P. Lewis, Z. Hong, A. Manivannan,M. Zhi, M. Li and N. Wu, Origin of Photocatalytic Activityof Nitrogen-Doped TiO2 Nanobelts, J. Am. Chem. Soc., 2009,131, 12290–12297.
34 N. R. Khalid, E. Ahmed, Z. Hong, Y. Zhang and M. Ahmad,Nitrogen doped TiO2 nanoparticles decorated on graphenesheets for photocatalysis applications, Curr. Appl. Phys.,2012, 12(6), 1485–1492.
35 L. Hu, J. Wang, J. Zhang, Q. Zhang and Z. Liu, An N-dopedanatase/rutile TiO2 hybrid from low-temperature directnitridization: enhanced photoactivity under UV-/visible-light, RSC Adv., 2014, 4(1), 420–427.
36 T. M. Orikawa, R. A. Sahi and T. O. Hwaki, Band-GapNarrowing of Titanium Dioxide by Nitrogen Doping, Jpn.Soc. Appl. Phys., 2001, 40(6A), 561–563.
44052 | RSC Adv., 2015, 5, 44041–44052
37 H. Irie, Y. Watanabe and K. Hashimoto, Nitrogen-Concentration Dependence on Photocatalytic Activity ofTiO2�xNx Powders, J. Phys. Chem. B, 2003, 107, 5483–5486.
38 Y. Nosaka, M. Matsushita, J. Nishino and A. Y. Nosaka,Nitrogen-doped titanium dioxide photocatalysts for visibleresponse prepared by using organic compounds, Sci.Technol. Adv. Mater., 2005, 6, 143–148.
39 Y. Lin, C. Weng, H. Hsu, Y. Lin and C. Shiesh, TheSynergistic Effect of Nitrogen Dopant and CalcinationTemperature on the Visible-Light-Induced Photoactivity ofN-Doped TiO2, Int. J. Photoenergy, 2013, (268723), 1–13.
40 B. Viswanathan and K. R. Krishanmurthy, NitrogenIncorporation in TiO2: Does It Make a Visible Light Photo-Active Material?, Int. J. Photoenergy, 2012, (269654), 1–10.
41 Y. Cong, J. Zhang, F. Chen and M. Anpo, Synthesis andCharacterization of Nitrogen-Doped TiO2
Nanophotocatalyst with High Visible Light Activity, J. Phys.Chem. C, 2007, (111), 6976–6982.
42 J. L. Sullivan and I. Bertoti, Effect of ion and neutralsputtering on single crystal TiO2, Vacuum, 1991, 42(18), 1203.
43 F. Dong, Y. Sun and M. Fu, Enhanced Visible LightPhotocatalytic Activity of Cluster Modied N-Doped forDegradation of Toluene in Air, Int. J. Photoenergy, 2012,(569719), 1–10.
44 U. Pal, E. S. Mora and A. E. Morales, Use of diffusereectance spectroscopy for optical characterization of un-supported nanostructures, Rev. Mex. Fis., 2007, 53(5), 18–22.
45 W. Zhang, L. Zou, R. Lewis and D. Dionysio, A Review ofVisible-Light Sensitive TiO2 Synthesis via Sol–Gel N-Dopingfor the Degradation of Dissolved Organic Compounds inWastewater Treatment, J. Mater. Sci. Chem. Eng., 2014, 2,28–40.
46 H. Wu, Y. Lin and S. Lin, Mechanisms of Visible LightPhotocatalysis in N-Doped Anatase TiO2 with OxygenVacancies from GGA + U Calculations, Int. J. Photoenergy,2013, (289328), 1–7.
47 D. Li, H. Haneda, S. Hishita, N. Ohashi andN. K. Labhsetwar, Fluorine-doped TiO2 powders preparedby spray pyrolysis and their improved photocatalyticactivity for decomposition of gas-phase acetaldehyde, J.Fluorine Chem., 2005, 126, 69–77.
48 T. Mishra, M. Mahato, N. Aman, J. N. Patel and R. K. Sahu, Amesoporous W, N co-doped titania nanomaterial withenhanced photocatalytic aqueous nitrate removal activityunder visible light, Catal. Sci. Technol., 2011, 1(4), 609.
49 S. Bingham and W. A. Daoud, Recent advances in makingnano-sized TiO2 visible-light active through rare-earthmetal doping, J. Mater. Chem., 2011, (21), 2041–2050.
