Fabrication of Al-Doped TiO 2 Visible-Light Photocatalyst for Low-ConcentrationMercury Removal

1 Graduate Institute of Engineering Science and Technology, National Kaohsiung First University of Science and Technology, No. 2 Jhuoyue Road, Nanzih, Kaohsiung 811, Taiwan 2 Department of Environmental, Safety, and Health Engineering, Tungnan University, Section 3, 152, Peishen Road, Shenkeng, New Taipei 222, Taiwan 3 Institute of Environmental Engineering and Management, National Taipei University of Technology, Section 3, No. 1, Chung-Hsiao E. Road, Taipei 106, Taiwan


Introduction
Mercury (Hg) releases from nature and anthropogenic sources have been the major focus of environmental studies owing to the toxicity and bioaccumulative behaviors [1].Hg species in gaseous phase from emission sources in general exist in three main forms: elemental (Hg 0 ), oxidized (Hg 2+ ), and particle-bound (Hg p ). Hg 2+ and Hg p can be easily removed by air pollution control devices such as wet flue gas desulfurization and electrostatic precipitators.Nevertheless, Hg 0 is highly volatile, insoluble in water, and thus difficult to remove from a gas stream.
Using titanium dioxide (TiO 2 ) photocatalysts as adsorbents and catalysts has been advised as a novel technique to effectively remove Hg 0 [2][3][4][5][6][7][8][9][10][11].Wu et al. [2] used in situ produced TiO 2 particles to remove Hg under UV irradiation.Upon irradiation with UV light, active sites become available on the TiO 2 particle surface and effectively adsorbed Hg to form a complex with TiO 2 .The manufacturing procedures strongly influence the purity and surface characteristics of resulting TiO 2 nanoparticles, which afterward affect the photocatalytic properties.Sol-gel method has been widely used in the bench-scale TiO 2 nanoparticles fabrication due to its simplicity to perform [12][13][14][15].Nevertheless, the processing temperature of sol-gel method syntheses is relatively low.A multistep fabrication procedure, that is, sample synthesis and subsequent calcination, was thus needed to transform the obtained TiO 2 into anatase or rutile.Thermal plasma has been shown to possess advantages to develop nanoparticles with clean surface and narrow particle size distribution.Using thermal plasma as a heating source may directly vaporize Ti metal having a high melting point at 1941 K and induce the high-purity TiO 2 formation in a single step.
A decisive obstruction in the successful application of TiO 2 is the band gap energy of 3.2 eV causing the TiO photocatalysts have therefore obtained great attention in recent years.Several studies have indicated that an improved TiO 2 photocatalyst excited by VL sources can be prepared by substitutional doping with metal atoms, such as Fe [16,17], Er [18], and Al [19][20][21][22][23].Because the ionic radii for Al and Ti are similar (0.053 nm for Al 3+ and 0.061 nm for Ti 4+ ), Al can easily fill into a regular cation position and form a substitutional solid solution.Numerous studies have shown that Al-doped TiO 2 nanoparticles can be manufactured via vapor-phase.Lee et al. [19] prepared Al-doped TiO 2 with thermal plasma using TiCl 4 and AlCl 3 as precursors.The absorption band of synthesized Al-doped TiO 2 nanoparticles shifted from the UV region to the VL region.Choi et al. [23] prepared Al-doped TiO 2 nanoparticles using a citrate-nitrate autocombustion system with Ti solution and Al(NO 3 ) 3 solution as precursors.The authors also demonstrated that Aldoped TiO 2 gas sensor was more selective and sensitive to CO and O 2 under 600 • C.
In our previous study, high-purity TiO 2 nanoparticles using Ti metal as a precursor were successfully manufactured with a transferred plasma torch [24].Nevertheless, using transferred DC plasma torch as the heating source has defects on low nanoparticles yield.Another plasma system was established in our earlier study using a nontransferred DC plasma torch as the heating source [25].Our preliminary test has shown that Al-doped TiO 2 nanoparticles can be successfully formed in a single step via this non-transferred DC thermal plasma system.However, the anatase/rutile ratio of the formed Al-doped TiO 2 was below 58.2 wt%, which may be due to the higher plasma power (i.e., 8 kW).In this study, Al-doped TiO 2 photocatalyst with a broad absorption spectrum was developed under a lower plasma power (i.e., 6 kW).The photocatalyst fabricated from this innovative single-step procedure was tested for Hg 0 capture under both UV and VL irradiation.We expected that a reducing plasma power can increase the anatase ratio of the formed Aldoped TiO 2 crystal.The greater content of anatase phase in TiO 2 may enhance the transformation of Hg 0 into Hg 2+ , which could subsequently enhance the removal effectiveness of TiO 2 .The Hg 0 removal effectiveness of Al-doped TiO 2 in the presence of O 2 , H 2 O, and light irradiation was further discussed.Notably, few studies have examined the VL photocatalytic effects of Al-doped TiO 2 on removal of Hg 0 at an extreme low concentration, namely, μg Nm −3 level.

Experimental Details
2.1.Synthesis of Al-Doped TiO 2 Nanoparticles.The DC plasma torch apparatus used for preparing TiO 2 nanoparticles is illustrated in Figure 1.The system comprised a non-transferred plasma torch connected to a DC power supply (Model PHS-15C, Taiwan Plasma Corp., Taiwan), a stainless steel reaction chamber (i.d.= 30 cm; length = 100 cm), a stainless steel powder feeder, a powder filter, a buffer tank, and a vacuum pump (GVD-050A, ULVAC) for shifting the particles floating in the exhaust gas.The plasma torch consisted of a water-cooled copper alloy electrode.The system was performed at 30 A and 200 V. Titanium powder (99.8% purity), Al 2 O 3 powder (99.9% purity), and ultrahigh-purity (UHP) O 2 were used as precursors.A mixture of UHP Ar and O 2 was used as the plasma gas.The flow rate was 60 L min −1 at Ar : O 2 = 3 : 1 by volume.UHP Ar with a flow rate of 2 L min −1 was also used as the carrier gas of the Ti and Al 2 O 3 powder feedstock.The Al 2 O 3 /Ti mass ratio was controlled at 0, 0.1, 0.3, and 0.5.The powder feeding rate was fixed at 0.2 g min −1 .The gas stream containing the formed TiO 2 nanoparticles was passed through the stainless steel powder filter and a buffer tank induced by the vacuum pump.

Al-Doped TiO 2
Nanoparticle Characterization.The particle size and morphology of TiO 2 nanoparticles were examined with a transmission electron microscope (TEM, Philips CM-200).Powder X-ray diffraction (XRD, Rigaku Rinet 200) with Cu κα radiation (λ = 1.5405Å) was used for crystal structure identification.The JCPDS database was used for powder crystalline phase identification.The mass fractions of anatase to rutile in formed TiO 2 nanoparticles were calculated by [26], where f A is the mass fraction of anatase, I A is the intensity of (101) reflection of anatase, and I R is the intensity of (110) reflection of rutile.The diffuse reflectance UV-visible spectra (UV-Vis) of TiO 2 nanoparticles were measured from 300 to 800 nm using a spectrophotometer (Hitachi U-3010).X-ray photoelectron spectroscopy (XPS, ULVAC-PHI 1600) was used for Ti, O, Al, and Cu bonding patterns identification.
The obtained XPS spectra were deconvoluted with the XPSPEAK software.

Hg Removal Experiments.
Al-doped TiO 2 synthesized at the Al 2 O 3 /Ti mass ratio of 0.5 was evaluated for the removal effectiveness of low-concentration gaseous Hg 0 .Gaseous Hg 0 was generated with a certificated Hg 0 permeation tube (VICI Metronics) which heated at 70 ± 0.1 • C to ensure a constant Hg 0 diffusion rate.Gaseous Hg 0 with a known concentration was mixed with N 2 , O 2 , and water vapor which was generated by streaming N 2 passed through a water bubbler.All gas mixing and Hg 0 injection occurred within a temperaturecontrolled chamber and heated tubes/lines to prevent water condensation.The generated Hg 0 -containing gas with a concentration of 10-15 μg Hg 0 Nm −3 and a flow rate of 1.5 L min −1 flowed through the photochemical reactor with 30 mg Al-doped TiO 2 were irradiated with UV or VL light.The photochemical reactor was a cylindrical quartz tube with an i.d.= 25 mm and a length = 150 mm.The nanoparticles were uniformly coated onto a glass slide, which was placed horizontally in the tube.The UV and VL light sets was located 1 cm above the tube.The photochemical tube reactor was operated at 25 • C and atmospheric pressure.The effluent gas from the photochemical reactor flowed through a moisture trap (i.e., a nefion tube) to remove H 2 O from the gas stream and thus to minimize the interference in Hg detection.The tail gas then flowed through a gold amalgamation column held by a heating coil (Brooks Rand model AC-01) where the Hg 0 in the gas was adsorbed.The Hg 0 that was concentrated on the gold was then thermally desorbed and sent as a concentrated Hg stream to a cold-vapor atomic fluorescence spectrophotometer (Model III, Brooks Rand Lab) for analysis.Six minutes were needed to complete a test run; ten runs were performed for each test condition.Finally, the exhaust was passed through a carbon trap before it was effluent into the fume hood.XRD powder patterns for the nanoparticles fabricated at various Al 2 O 3 /Ti mass ratios are presented in Figure 3.All the peak intensity of powder diffraction was normalized by anatase (101).The experimental results suggested that most of the diffraction peaks could be designated as the presence of anatase and rutile phases.However, the diffraction peaks standing for Al 2 O 3 appeared as Al 2 O 3 /Ti = 0.5, based on 2θ = 35.15• , 43.35 • , 52.5 • , and 57.5 • that are indexed to the Al 2 O 3 diffraction pattern.These observations are consistent with our previous work [25] and suggest that at Al 2 O 3 /Ti = 0.5 loading, the thermal plasma was less effective to vaporize the Al 2 O 3 powders and to induce the interactions between Ti, O, and Al atoms.Moreover, the relative content of anatase, depicted with the f A value (Table 1), noticeably reduced with increasing Al 2 O 3 /Ti due to consequent transformation into rutile at an elevating temperature [27].Notably, the TiO 2 was fabricated at a fixed plasma power of 6 kW.The plasma temperature change at various Al 2 O 3 /Ti ratios should be small and may not markedly influence the transformation of anatase into rutile.The enhancement in transformation of anatase into rutile may also attribute to the increase in Al doping [19,21,22,28].

Results and Discussion
Figure 4 demonstrates the UV-visible spectra of the Aldoped TiO 2 nanoparticles synthesized at various Al 2 O 3 /Ti mass ratios over the wavelength range of 300-800 nm.The commercial Degussa P-25 photocatalyst was also tested for comparison.The experimental results showed that the absorption edge was at ∼390 nm for P-25 photocatalyst.The TiO 2 nanoparticles synthesized in thermal plasma at the Al 2 O 3 /Ti mass ratio of 0 to 0.3 possessed an absorption edge at ∼400 nm.The absorption spectra of Al-doped TiO 2 slightly shifted from UV to VL region with reference to an increase in Al 2 O 3 /Ti mass ration can be assigned to the band gap narrowing relation to the interstitial Al species in the TiO 2 crystal [19,29,30].Especially, TiO 2 synthesized at Al 2 O 3 /Ti = 0 also had band gap absorption at ∼400 nm, which was comparable to that of Al-doped TiO 2 formed at Al 2 O 3 /Ti = 0.1.The extent of red shift and band broadening may be attributed to the presence of oxygen vacancies in TiO 2 crystal formed in the high-temperature plasma flame.Numerous studies have shown the appearance of the visiblelight activity was attributed to the newly formed oxygen vacancy state in the TiO 2 band structure [24,31].Further International Journal of Photoenergy    peaks are indications to the presence of TiO 2 (Ti 4+ 2p1/2 and Ti 4+ 2p3/2 ) and Ti 2 O 3 (Ti 3+ 2p1/2 and Ti 3+ 2p3/2 ), respectively [32].For the Al-doped TiO 2 , the Al 2p peaks at a binding energy of 75.5 eV can be attributed to the presence of Al 3+ .The results suggested that the Al 3+ content enhanced with an increase in Al 2 O 3 addition.The calculated Ti 3+ /(Ti 3+ + Ti 4+ ) ratios for Al-doped TiO 2 at Al 2 O 3 /Ti = 0, 0.1, 0.3, and 0.5 were 3.1, 17.1, 17.5, and 33.2%, respectively, based on the deconvoluted peak area.These data indicated that Ti 3+ concentration greatly enhanced with increasing Al 2 O 3 addition owing to the transformation of TiO 2 into Ti 2 O 3 .We suspected that Ti 4+ /Al 3+ ionic substitution may take place during the Al doping.If this assumption held, when Ti 4+ (ionic radius = 0.061 nm) is substituted by Al 3+ (ionic radius = 0.053 nm) from TiO 2 crystal, the lattice mismatch occurs as a result of that the ionic radius of Ti 4+ is larger than that of Al 3+ .To atone for the smaller ionic radius of Al 3+ , a Ti species having an ionic radius larger than Ti 4+ is needed.Consequently, Ti 4+ is reduced to Ti 3+ (ionic radius = 0.067 nm).The observed Ti 3+ peaks in the present study are consistent with Steveson et al. [33] and our previous work [25].In addition, the formal charge generated by the substitution of Ti 4+ with Al 3+ can also be compensated via the formation of O − from O 2− , which resulted in the oxygen vacancies found in Aldoped TiO 2 [21,33,34].Notably, about 1.95-4.62at.%Cu was found in the formed Al-doped TiO 2 based on the XPS analysis (Table 2).The Cu impurity was from the vaporization of plasma torch made of Cu alloy under the high-temperature plasma environment.Associated with the results from the UV-visible  analysis, the doped Cu and the oxygen vacancy may synergistically contribute to the observed red shift in the UVvisible absorption spectrum for the non-Al-doped TiO 2 (i.e., Al 2 O 3 /Ti = 0).Nevertheless, Zhang et al. suggested that Cu content < 38 at.% in TiO 2 had negligible effects on the Ti 2p binding energy of XPS examinations [35].Therefore, isomorphous substitution due to Al doping into TiO 2 crystal should be the major contribution of the increasing Ti 3+ .The red shift in the absorption spectra of TiO 2 nanoparticles thus primarily was attributed to the Al doping and generated oxygen vacancy (Figure 4).The Hg 0 adsorption breakthrough results for Al-doped TiO 2 nanoparticles synthesized at Al 2 O 3 /Ti = 0.5 are shown in Figure 7.The experimental parameters included O 2 concentration, humidity, and the type of light sources.These parameters were tested alternately by gradually increasing the O 2 concentration from 0% to 12% combined with introducing H 2 O (20% relative humidity) and UV/VL irradiation to the photocatalytic reactor.The experimental results showed that Hg 0 capture was very small with rapid breakthrough at the 0% O 2 , dry (H 2 O < 0.1 vol%), and dark condition (runs 0-10 in Figure 7), manifesting that the synthesized Al-doped TiO 2 (Al 2 O 3 /Ti = 0.5) was less effective in removal of Hg 0 under the test condition.This result was anticipated because Hg 0 appeared to the main Hg species at the 0% O 2 , dry, and dark condition and was not easy to form a strong binding with the surface of TiO 2 .However, an increase in the Hg 0 capture to 25% was apparently observed when UV irradiation was applied (breakthrough down to approximately 0.75; runs [11][12][13][14][15][16][17][18][19][20].The significant enhancement in Hg 0 removal for Al-doped TiO 2 nanoparticles under UV irradiation is a strong indication that this sample had a good photocatalytic potential to transform Hg 0 into Hg 2+ that enhanced the adsorption onto Al-doped TiO 2 .It is also noteworthy that without the presence of O 2 , VL was less effective in photocatalytic oxidation than UV (runs 21-30).In addition, a significant decrease in Hg 0 capture was observed at the humid condition (runs 31-60).This result suggest that H 2 O competitively adsorbs onto the TiO 2 active sites, causing the reemission of adsorbed Hg species, which can be Hg 0 or Hg 2+ needed to be further examined.The experimental results presented here are in agreement with those found in earlier studies [6,7,24,36].Li and Wu reported that the physically adsorbed Hg 0 can be desorbed from the surface of a SiO 2 -TiO 2 composite by water vapor at high concentration, which suggested that Hg 0 is only weakly adsorbed on the sorbent surface [6].Dissimilar to H 2 O, O 2 notably improved the adsorption of Hg 0 ; increasing O 2 concentration strongly enhanced the Hg 0 capture of Aldoped TiO 2 to up to 40% (equivalent to breakthrough of 0.6; Figure 7).It is noteworthy that when O 2 was > 6%, Hg 0 capture was similar for Al-doped TiO 2 under either UV or VL irradiation (runs 131-150 and 221-240).These data not only revealed the importance of O 2 in enhancing Hg 0 capture, but also verified the visible-light activity of the synthesized Aldoped TiO 2 nanoparticles on Hg 0 oxidation/adsorption.

Conclusion
Al-doped TiO 2 nanoparticles were successfully synthesized in a single step using Ti powders, Al 2 O 3 powders, and O 2 by a nontransferred plasma torch system.TiO 2 nanoparticles formed at Al 2 O 3 /Ti mass ratios 0 to 0.5 were approximately between 10 and 105 nm.The crystal phases of the formed TiO 2 nanoparticles were mainly in anatase and rutile forms.However, increasing the Al 2 O 3 addition caused the ratio of anatase to rutile decreased.The presence of oxygen vacancy and the substitution of Ti 4+ with Al 3+ were suspected to cause the slight red shift in the absorption edge to lower energy due to band gap narrowing.Al doping and oxygen vacancy in the TiO 2 crystal may also result in the phase transformation from TiO 2 to Ti 2 O 3 .Hg breakthrough results showed that the Hg 0 removal with formed Al-doped TiO 2 in a dry condition was greater than that in a humid condition when light irradiation was applied.Hg capture was also found to be markedly enhanced by increasing O 2 concentration.Nevertheless, H 2 O showed deteriorating effects on the adsorption of Hg through competition for active sites on the Al-doped TiO 2 surface.Results presented here suggest that Hg 0 removal using synthesized Al-doped TiO 2 nanoparticles may be greatly affected by the extent of catalytic transformation of Hg 0 into Hg 2+ and amphoteric (hydrophilic-hydrophobic) surface properties of Al-doped TiO 2 .
We envision that the photocatalyst can be successfully used to capture Hg from coal-derived flue gas.Nevertheless, it is imperative to note that this study employed gas streams consisting of Hg 0 in a moisture-oxygen-nitrogen mix.Coalderived flue gas is a complex mixture also containing fly ash particles, moisture, CO, and many acid gases.For example, a typical untreated flue gas derived from the combustion of a US low sulfur eastern bituminous coal can contain 5-7% H 2 O, 3-4% O 2 , 15-16% CO 2 , 1 ppbv total Hg, 20 ppm CO, 10 ppm hydrocarbons, 100 ppm HCl, 800 ppm SO 2 , 10 ppm SO 3 , 500 ppm NO x , and balance N 2 [9][10][11].The influences of the flue gas components on Hg removal using TiO 2 photocatalysts are thus highly needed to be further investigated.

Figure 1 :
Figure 1: Schematic diagram of non-transferred DC plasma torch system for synthesis Al-doped TiO 2 photocatalysts.

Figure 2
Figure2illustrates the TEM images of the nanoparticles produced at various Al 2 O 3 /Ti mass ratios.TEM results indicated that the synthesized Al-doped TiO 2 was homogeneous, without significant phase separation or coating on the surface.It was also noticed that the synthesized nanoparticles was in hexagonal or spherical shapes; adding Al 2 O 3 powder had insignificant effects on the shape of the formed nanoparticles.The darker areas in the TEM micrographs showed the agglomeration of Al-doped TiO 2 nanoparticles.The powder size of the feedstock Ti and Al 2 O 3 powder were about 5-15 μm.However, nanoparticles formed at Al 2 O 3 /Ti mass ratios of 0 to 0.5 were approximately between 10 and

Figure 7 :
Figure 7: Hg breakthrough ratio of Al-doped TiO 2 in the continuous adsorption experiment under various test conditions.

Table 1 :
Crystal phases of Al-doped TiO 2 synthesized at various Al 2 O 3 /Ti mass ratio.These observation results indicate that the injected Ti and Al 2 O 3 powders successfully vaporized at the high flame temperature and subsequently synthesized Al-doped TiO 2 nanoparticles via the recombination of vaporized Ti, O and Al atoms in the thermal plasma environment.

Table 2 :
Atomic percentage of synthesized TiO 2 nanoparticles based on XPS examinations.