Mercury removal from aqueous solutions of HgCl2 by heterogeneous photocatalysis with TiO2

Mercury removal from aqueous solutions of HgCl2 by heterogeneous photocatalysis with TiO2

Applied Catalysis B: Environmental 104 (2011) 220–228 Contents lists available at ScienceDirect Applied Catalysis B: Environmental journal homepage:...

832KB Sizes 2 Downloads 28 Views

Applied Catalysis B: Environmental 104 (2011) 220–228

Contents lists available at ScienceDirect

Applied Catalysis B: Environmental journal homepage: www.elsevier.com/locate/apcatb

Mercury removal from aqueous solutions of HgCl2 by heterogeneous photocatalysis with TiO2 ˜ a,∗ , J. Aguado a , A. Arencibia b , R. Pascual a M.J. López-Munoz a b

Department of Chemical and Environmental Technology, ESCET, Universidad Rey Juan Carlos, C/Tulipán s/n, 28933 Móstoles, Madrid, Spain Department of Chemical and Energy Technology, ESCET, Universidad Rey Juan Carlos, C/Tulipán s/n, 28933 Móstoles, Madrid, Spain

a r t i c l e

i n f o

Article history: Received 12 November 2010 Received in revised form 18 March 2011 Accepted 24 March 2011 Available online 2 April 2011 Keywords: Hg(II) removal Adsorption Heterogeneous Photocatalysis TiO2 Organic donors

a b s t r a c t The photocatalytic removal of Hg(II) from aqueous solutions of HgCl2 using TiO2 as catalyst was studied. The influence of pH and the addition of methanol, formic acid and oxalic acid as sacrificial additives on the extent of Hg(II) adsorption and photocatalytic reduction was investigated. The results showed that the overall process strongly depended on pH, being enhanced as the pH was increased. At pH 10, an efficient removal of Hg(II) was achieved even in the absence of organic additives, attaining final mercury concentrations in solution at trace levels (␮g L−1 ). In acidic conditions, the addition of sacrificial organic molecules significantly increased the rate and extent of aqueous Hg(II) removal. The nature and distribution of mercury products deposited on the catalyst was dependent on the reaction conditions. In the absence of additives, Hg2 Cl2 and Hg0 were respectively identified in acidic and neutral/alkaline media as main reduced species on the titania surface. The addition of organic additives enhanced the photocatalytic reduction to Hg0 . Comparison between adsorption and reaction results evidenced that it cannot be established a direct correlation between Hg(II) dark adsorption on the TiO2 surface and the efficiency of Hg(II) photoreduction achieved. © 2011 Elsevier B.V. All rights reserved.

1. Introduction Pollution of water with mercury and its compounds deserves special attention due to the high toxicity of these species and their tendency to bio-accumulate, even at very low concentrations. Industries such as chlorine-alkali, chemical, petrochemical, pharmaceutical, metallurgical, paint, oil-refining fertilizer and electrical extensively use mercury, which can be found at significant degree in the resulting manufacturing wastewater streams [1,2]. As disposal of mercury in aqueous systems constitutes a major environmental concern, worldwide regulations are very restrictive about the concentration levels for mercury discharge into water and quality standards values in bodies of surface water [3–7]. The currently available technologies for the treatment of mercury-polluted aqueous solutions include precipitation as sulphide, membrane filtration, ion exchange, electrodeposition, adsorption and coagulation [2]. Most of those conventional methods, however, show some drawbacks to achieve the low mercury levels required nowadays. Hence, many studies are devoted to improve the established methods or to develop novel treatment

∗ Corresponding author. Tel.: +34 91 6647464. ˜ E-mail address: [email protected] (M.J. López-Munoz). 0926-3373/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.apcatb.2011.03.029

procedures for the removal of mercury from contaminated surface and ground waters [8–10]. In that respect, heterogeneous photocatalysis with titanium dioxide can be a promising alternative because it has demonstrated its effectiveness in the removal and recovery of a variety of metal ions from aqueous effluents [11–13]. The process is based on the irradiation of a semiconductor, usually titanium dioxide, to promote electrons from its valence band to the conduction band with the simultaneous generation of holes in the former band. Holes and electrons can either recombine in the bulk or at the semiconductor surface or react with adsorbed species to bring about redox reactions. In aqueous TiO2 suspensions, the potentially oxidisable species are water molecules, hydroxyl ions, and organic compounds whereas photogenerated electrons can reduce either adsorbed oxygen molecules to yield superoxide radicals or metallic ions as far as the potential of the conduction band edge of the semiconductor is more negative than the reduction potential of the Mn+ /M(n−m)+ couple [14]. If reduction of metal ions proceeds, metal is deposited on the titania surface from where it can be extracted through chemical or physical procedures [15–17]. The photocatalytic reduction of Hg(II) in aqueous solution has been previously reported [18–23], although the number of studies on this subject is relatively scarce. As an example, Serpone et al. [18] investigated the removal of mercury(II) chloride and methylmercury(II) chloride from aqueous media in air-equilibrated

M.J. López-Mu˜ noz et al. / Applied Catalysis B: Environmental 104 (2011) 220–228

suspensions of TiO2 (2 g L−1 ) irradiated by AM1 simulated Sunlight. They found that HgCl2 was efficiently removed from water under natural pH conditions whereas photoreduction and deposition of mercury from methylmercury(II) solutions needed addition of methanol (20% v/v). Aguado et al. [19] investigated the photocatalytic treatment of Hg(NO3 )2 solutions and observed that Hg(II) reduction was first order with respect to the Hg(II) concentration. They concluded that temperature had no great influence, oxygen competed with the mercury(II) for the reduction process and the reaction was faster at alkaline pH. The effect of pH in the 2–4.5 range and the addition of formic acid and chlorides on the photocatalytic efficiency was studied by Wang et al. [20]. They concluded that the reduction rate of Hg(II) was controlled by the amount of Hg(II) adsorbed onto TiO2 . Formic acid had a beneficial influence on the process, explained by means of its role as hole scavenger, whereas the presence of increasing chloride concentrations decreased the Hg(II) photo-reduction rate. All previous works indicated the formation of Hg(0) as reduction product deposited on TiO2 but no investigation on the nature of other possible mercury species was done. This aspect was further studied by Botta et al. [21], who reported the formation of metallic Hg, HgO and Hg2 Cl2 during the photocatalytic treatment of HgCl2 with TiO2 Degussa P25. Even though the previous studies clearly demonstrated the feasibility of heterogeneous photocatalysis for uptake of aqueous Hg(II), the analytical methods used in most of them were not sensitive enough to analyze mercury at trace levels (␮g L−1 ). Taking into consideration that the concentrations currently legislated for mercury in water range from 100 ␮g L−1 Hg (maximum acceptable discharge concentration) to 3–50 ng L−1 Hg (water quality criterion in surface waters) [3–7], it is important to determine the suitability of the photocatalytic treatment for achieving those Hg concentrations levels. On this basis, the purpose of the present study was to investigate the scope of heterogeneous photocatalysis as a plausible technique for the treatment of waters polluted with mercury compounds. Special attention has been paid to the determination at trace levels of final mercury concentration achieved upon the different experimental conditions evaluated. The research has been focused on the photocatalytic treatment of aqueous solutions of HgCl2 with TiO2 . The effect of pH and the presence of sacrificial organic additives, namely methanol, oxalic and formic acid, on the extent of Hg(II) adsorption and photocatalytic reduction has been investigated. The nature and distribution of mercury deposits on the catalyst has also been analysed by X-ray Diffraction (XRD) and Environmental Scanning Electron Microscopy (ESEM).

2. Experimental 2.1. Reagents All reagents were used as received without further purification. Mercuric chloride (HgCl2 , puris. Riedel-de Haën) was employed as the source of Hg(II). Formic acid (HCOOH, ∼98%) was obtained from Fluka, methanol (CH3 OH, reagent grade), and oxalic acid (C2 H2 O4 , puris.) were purchased to Scharlau. Standard aqueous solution of mercury (1000 mg L−1 ) for analysis was obtained from Merck. Sodium hydroxide (NaOH, reagent grade, Scharlau) and nitric acid (HNO3 , reagent grade, Scharlau) were used for adjusting the initial pH. Solutions were prepared in Milli-Q water (resistivity 18.2 M cm). The commercial titanium dioxide Degussa P25 was used as photocatalyst. It consists of anatase and rutile crystalline phases in a ratio of 4:1. It is a non-porous solid, with a BET surface area of 50 m2 g−1 and a mean particle size of ca. 30 nm although larger polycrystalline aggregates are formed in aqueous suspension.

221

2.2. Dark adsorption experiments Experiments to determine the extent of mercury adsorption on the titania surface under different conditions were performed by keeping a 1 L of aqueous solution 100 mg L−1 (0.5 mM) HgCl2 in contact with TiO2 (concentration interval from 0.2 to 4 g L−1 ) at 20 ◦ C for 1 h in the dark. Afterwards, the solution was filtrated through a 0.22 ␮m nylon membrane to remove the suspended catalyst before being analysed. The adsorption of mercury was taken as the difference between the initial Hg(II) concentration and the amount of mercury remaining in the solution after the equilibration period. 2.3. Photocatalytic reactions procedure The photoreaction runs were carried out in a cylindrical Pyrex batch reactor of 1 L as effective solution volume, provided on the top with three ports for inflow and outflow of gases and withdrawal of aliquots for analysis. The initial concentration of Hg(II) in the aqueous solution was 100 mg L−1 (0.5 mM) and the catalyst was suspended under magnetic stirring with a load of 2 g L−1 . Prior to irradiations, the suspension was kept in the dark for 1 h under the same conditions of pH, organic additives, gas bubbling and stirring used in the photocatalytic reaction to allow the adsorption of mercury on the catalyst. UV-irradiation was performed with a 150 W medium pressure mercury lamp (Heraeus TQ-150) axially immersed within the reactor and surrounded with a cooling tube where an 0.01 M aqueous solution of copper sulphate was circulated to prevent overheating of the suspension and cut off radiation with wavelengths below 320 nm. In order to achieve a stabilised radiation emission, the lamp was switched on 15 min before being fitted into the reactor. Along the photocatalytic runs aliquots were withdrawn at time intervals, following filtration through 0.22 ␮m nylon membranes to remove the suspended catalyst before being analysed. 2.4. Analytical techniques Analysis of mercury concentration in the range of concentration of 0–10 mg L−1 was carried out by a Varian Vista AX inductive coupled plasma-atomic emission spectrometer (ICP-AES). Two emission mercury lines (194 and 253 nm) were used according to the standard EPA method for mercury analysis [24]. Mercury concentrations lower than 0.1 mg L−1 were determined by cold vapour atomic fluorescence spectroscopy (CVAFS). Analyses were carried out in a PSA Analytical Millennium Merlin equipment following the EN 13506 standard procedure [25]. X-ray diffraction (XRD) patterns of the solids recovered at the end of the reactions were acquired on a Philips X’PERT MPD appara˚ Scans were made in the 2 tus using Cu K␣ radiation ( = 1.54059 A). range 5–90◦ with a step size of 0.04 and a step time of 1 s, enough to obtain a good signal-to-noise ratio in all the studied reflections. The morphology of the solids was examined by environmental scanning electron microscopy (ESEM) using a XL30 ESEM.FEI Philips equipment. 2.5. Speciation modeling in aqueous solution In order to study the influence of the mercury chemical state on the adsorption and photocatalytic process, speciation diagrams of aqueous mercury were calculated as a function of pH, mercury concentration and presence of organic species using the chemical equilibrium modeling software MINEQL+ [26]. The values for stability constants of mercury complexes used in these calculations are shown in Table 1 [26–28].

222

M.J. López-Mu˜ noz et al. / Applied Catalysis B: Environmental 104 (2011) 220–228

Table 1 Stability constants of mercury species used in the MINEQL+ speciation calculations. Reaction −

Log K

Hg + Cl  HgCl Hg2+ + 2Cl−  HgCl2 Hg2+ + 3Cl−  HgCl3 − Hg2+ + 4Cl−  HgCl4 2− Hg2+ + Cl− + OH−  HgClOH Hg2+ + OH−  HgOH+ Hg2+ + 2OH−  HgOH2 Hg2+ + 3OH−  Hg(OH)3 − Hg2+ + NO3 −  HgNO3 + Hg2+ + 2NO3 −  Hg(NO3 )2 Hg2+ + HCOO−  HgHCOO+ Hg(OH)2 −  H2 O + HgO↓ 2+

+

7.3 14.0 14.2 15.5 10.44 10.9 22.3 21.46 0.77 1.00 3.5 3.6

3. Results and discussion 3.1. Adsorption of Hg(II) on TiO2 3.1.1. Influence of pH Adsorption experiments were performed in the dark using a solution of 100 mg L−1 of HgCl2 to assess the capacity of TiO2 P25 for Hg(II) adsorption at different pH values. Fig. 1 shows the plots of Hg(II) adsorption percentage at initial pH 2, 4.5 (natural pH), 7, and 10 as a function of titania concentration. As it can be observed, the adsorption of Hg(II) on titania strongly depended on the solution pH, being increased with increasing pH in the range of TiO2 concentration evaluated. Mercury showed a little affinity for the oxide surface at acidic pH values, whereas adsorption percentages above 60% of Hg(II) initially present in solution were attained at pH 7 and 10. No significant increase of mercury adsorption was found at pH 7 in the interval from 2 to 4 g L−1 of TiO2 . By contrast, at pH 10 the uptake of Hg(II) on titania was increased over the evaluated range of TiO2 dosage, attaining a maximum Hg(II) removal of 98% with a titania concentration of 4 g L−1 . To the best of our knowledge, there are scarce references in the literature that report data on mercury adsorption on TiO2 P25 from HgCl2 aqueous solutions. Moreover, it is noticeable that some dissimilar results are given. As an example, Botta et al. [21] studied the degree of adsorption of Hg(II) nitrate, chloride and perchlorate onto TiO2 surface at initial pH values of 3, 7 and 11. For a titania concentration of 1 g L−1 and Hg(II) concentration of 0.5 mM, adsorption percentages of around 27% at pH 3, 34% at pH 7 and 30% at pH 11

Fig. 1. Adsorption of aqueous Hg(II) as a function of TiO2 concentration at different pH values. C0 Hg(II) = 100 mg L−1 .

Fig. 2. Aqueous speciation diagram of HgCl2 (C0 Hg(II) = 100 mg L−1 ).

were found in the case of HgCl2 ,therefore showing no strong differences between the different pH values. On the other hand, Serpone et al. [18] reported an adsorption percentage of Hg(II) ca. 30% at an initial pH 4.65 and ca. 85 to 90% at pH 7.0 for a HgCl2 concentration of 0.5 mM in the presence of 2 g L−1 of TiO2 P25. Wang et al. [20] found Hg(II) adsorption percentages of 1.8% at pH 2.5, 4.5% at pH 3 and 12.3% at pH 4, using a TiO2 P25 concentration of 2 g L−1 and Hg(II) concentration of 0.56 mM and Chen et al. [16] reported a 10.4% of Hg(II) adsorption at natural pH for an initial concentration of 0.6 mM of HgCl2 and a TiO2 dosage of 2 g L−1 . Both the charge of titanium dioxide surface and the speciation of mercury compounds are significantly affected by the pH of the solution, therefore influencing the process of Hg(II) adsorption. For TiO2 Degussa P25, zero point charge (z.p.c.) values of 6.8 [29] and 7.0 [30] have been reported. Consequently, at pH 2 and 4.5 titania particles are positively charged according to the following equilibrium: Ti–OH + H+ ↔ TiOH2 +

(1)

At pH 7, Ti–OH entities are expected to be the main species on the catalyst surface whereas TiO− groups must prevail at pH 10 due to the deprotonation of surface hydroxyl groups above the isoelectric point: Ti–OH + OH− ↔ TiO− + H2 O

(2)

On the other hand, the nature of Hg(II) species present in the solution drastically varies with pH, as it can be observed in the aqueous speciation diagram displayed in Fig. 2. At pH 2, HgCl2 are the major species present in solution and therefore the main Hg(II) adsorbate. As the pH is increased up to 4.5 a small contribution of HgClOH has to be also considered. The aqueous speciation changes significantly at pH 7, where HgClOH and Hg(OH)2 are the coexisting solution species, becoming mercuric hydroxide predominant at pH 10. The calculations for speciation predict the potential precipitation of montroydite (HgO) at pH 10. However, no solid formation was observed upon adjusting to 10 the pH of the 100 mg L−1 HgCl2 solution, so these species are not included in the diagram neither in the subsequent discussion of results. The comparison of Hg(II) speciation with the adsorption results obtained at pH 2 and 4.5 (Fig. 1) reveals the lack of affinity of HgCl2 for the TiO2 surface. In agreement, it has been previously reported that Hg(II) adsorption on goethite (␣-FeOOH), ␥-alumina (␥-Al2 O3 ) and bayerite (␤-Al(OH)3 ) was inhibited in the presence of increasing chloride concentration [31,32]. This inhibition was attributed to the formation of HgCl2 complexes which are characterized by their high stability in solution what makes them to be no prone to adsorption. Neither the small fraction of HgCl+ species coexisting

M.J. López-Mu˜ noz et al. / Applied Catalysis B: Environmental 104 (2011) 220–228

with HgCl2 at pH 2 are adsorbed on TiO2 , what can be explained in terms of electrostatic repulsions with the positively charged oxide surface. On the other hand, the small Hg(II) uptake observed at pH 4.5 must be related to the presence of HgClOH species. Weerasooriya et al. [33] reported some affinity of HgClOH for gibbsite which was attributed to the charge asymmetry of HgClOH. Nevertheless, the affinity of HgClOH species for the positively charged titania surface is scarce, according to the low adsorption percentages obtained in the TiO2 concentration range studied. The significant rise in Hg(II) adsorption with increasing pH in the range 7–10 can be attributed to the increase in concentration of Hg(OH)2 species, which become the main Hg(II) adsorbate. This trend is in accordance with previous studies focused on the adsorption of mercury on a variety of metal oxides and mineral surfaces. As an example, in the adsorption of Hg(II) on quartz and gibbsite it was observed that the pH at which maximum Hg(II) adsorption occurred was comparable to the pKa for the hydrolysis of Hg2+ to form Hg(OH)2 [34]. Similar conclusions were obtained in a study focused on Hg(II) adsorption on goethite, being the major Hg(II) retention correlated to the formation of Hg(OH)2 species [35]. Accordingly, the plateau of ca. 60% of Hg(II) removal achieved at pH 7 in the range 2–4 g TiO2 L−1 (Fig. 1) can be explained taking into account that only a fraction of aqueous mercury is present in the solution as Hg(OH)2 species (Fig. 2). The experimental results indicate that a loading of 2 g TiO2 L−1 is enough to uptake the whole mercury species that can be readily adsorbed on the titania surface and therefore, upon increasing the titania loading from 2 to 4 g TiO2 L−1 , no enhancement of mercury adsorption can be achieved. The surface complexation model has been proposed to describe the adsorption of Hg(II) in solution onto natural oxides and oxide-like surfaces: gibbsite [34]; ␣-SiO2 (quartz) and amorphous hydrous ferric oxide [36]; amorphous silica and goethite [37] and kaolinite [38]. In this approach it is assumed that mercury adsorption takes place preferentially onto hydroxyl surface sites and gives rise to the formation of a surface complex between the metal cation and the hydroxyl groups. Accordingly, Sarkar et al. [38] indicated that for those pH values where Hg(OH)2 is the predominant solution species, the Hg(II) adsorption process on kaolinite could be illustrated by the reaction: ≡ XOH + Hg(OH)2 → ≡ XOHg(OH)2 − + H+

(3)

where (≡XOH) represented the silanol and aluminol groups on the surface. Following similar considerations, the adsorption of Hg(II) on TiO2 at pH values in the range 7–10 might be described by: ≡ TiOH + Hg(OH)2 → ≡ TiOHg(OH)2 − + H+

(4)

This is consistent with the decrease of the initial pH of the solution from 10 to 8.5 and from 7 to 6 detected along the mercury adsorption process. Nevertheless, the high capacity of TiO2 for Hg(II) adsorption at pH 10 indicates that not only hydroxyl groups but also TiO− entities must account for the adsorption process, most likely through: ≡ TiO− + Hg(OH)2 → ≡ TiOHg(OH)2 −

(5)

3.1.2. Influence of organic additives Experiments were conducted to investigate the influence of the presence of organic species on the adsorption of Hg(II) on TiO2 . Formic acid, oxalic acid and methanol were chosen on the basis of their role as sacrificial electron donors in photocatalytic processes. Based in the results shown in Fig. 1, TiO2 concentration was selected to be 2 g L−1 to avoid the ca. 98% of Hg(II) adsorption observed at pH 10 for the higher titania loadings. Hg(II) uptake percentages achieved at the different initial pH values after equilibration are displayed in Table 2.

223

Table 2 Influence of organic additives on adsorption of aqueous Hg(II) on TiO2 . Organic addedb

Hg(II) adsorption percentagea pH 2 pH 4.5 pH 7

pH 10

No organic Methanol Formic acid Oxalic acid

0.00 3.84 1.96 5.25

82.81 85.57 88.25 79.74

a b

4.22 0.01 1.45 21.46

58.60 56.40 64.06 75.34

C0 Hg(II) = 100 mg L−1 ; C (TiO2 ) = 2 g L−1 . C0 = 2 mM.

The most significant variations in Hg(II) adsorption on TiO2 occurred upon addition of oxalic acid at initial pH 4.5 and 7, conditions where an enhancement of ca. 17% in mercury adsorption was brought about (from 4.22% to 21.46% and from 58.60 to 75.34% respectively). At pH 2 the presence of oxalic acid was also beneficial, although in a quite lesser extent (5.25% of adsorption enhancement) whereas a slight diminution in Hg(II) adsorption was promoted at pH 10. In the presence of formic acid and methanol, the changes in the extent of Hg(II) adsorption were, in general, less significant. In this case, the major influences observed were, on one side an enhancement in Hg(II) adsorption of ca. 5% at pH 7 and 10 promoted by formic acid and on the other, a decrease of ca. 4% in mercury adsorption induced by the presence of methanol. In order to determine the potential presence in the solution of cationic mercury species or the formation of complexes between mercury and the organic additives, the aqueous speciation diagrams of mercury as a function of the pH and the presence of each organic additive were calculated. The results indicated that no variations in the speciation of mercury shown in Fig. 2 should be induced by the presence of 2 mM methanol or 2 mM formic acid. Therefore, the slight decrease in adsorption observed at pH 4.5 in the presence of these organics, cannot be attributed to their ability to complex mercury in solution, but to some competition for adsorption centres on the titania surface. In contrast, aqueous mercury oxalate (HgC2 O4 ) is expected to be formed in the pH range 2–7 by adding 2 mM oxalic acid to the HgCl2 solution. Even though these complexes are stable in aqueous solution, the enhancing of Hg(II) uptake observed in the presence of oxalic acid points out to the formation of ternary surface complexes [TiO2 surface-oxalate–mercury], postulation which is consistent with the strong tendency of oxalic acid to be chemisorbed on titania [39,40]. As for the influence of formate on Hg(II) adsorption at pH 7 and 10, it has been previously reported that the presence of HCOO− ions enhanced the adsorption of cadmium ions, being this increase attributed to the role of formate ions as “interfacial anchor” for Cd(II) to adsorb onto the TiO2 surface [41]. Similarly, some chelating effect can be proposed for mercury species, although, according to the small increase in the Hg(II) adsorption percentage, this interaction might not be very significant. 3.2. Photocatalytic reduction of Hg(II) 3.2.1. Influence of pH Accordingly to the results above described, the adsorption of mercury on the catalyst was allowed to be established in the dark previously to the photocatalytic reactions, in order to clearly distinguish between photocatalytic Hg(II) uptake and removal of dissolved Hg(II) by adsorption. Also based in the results shown in Fig. 1, a TiO2 concentration of 2 g L−1 was selected. The photocatalytic reactions were performed in anoxic conditions as the presence of oxygen was observed to inhibit or drastically reduce the activity under all experimental conditions evaluated in the present work. Fig. 3 depicts the kinetic profiles of normalized Hg(II) concentration (C/C0 ) vs. time obtained upon irradiation of HgCl2 solution

224

M.J. López-Mu˜ noz et al. / Applied Catalysis B: Environmental 104 (2011) 220–228

Fig. 3. Photocatalytic uptake of aqueous Hg(II) as a function of time. Conditions: C0 Hg(II) = 100 mg L−1 ; C (TiO2 ) = 2 g L−1 . Fig. 4. XRD diffraction patterns of: (a) fresh catalyst; (b) and (c) catalyst recovered after photocatalytic reactions performed at initial pH 2 and 10 respectively.

in the presence of TiO2 at different initial pH values. First order kinetics respect to mercury concentration gave the best fitting of the experimental values with correlation factors higher than 0.97. The kinetic constants were calculated by fitting the experimental concentration changes in logarithm format vs. reaction time using a least squares linear regression algorithm. The values obtained at the different pH conditions were the following: kpH 2 = 0.0028 min−1 ; kpH 4.5 = 0.0168 min−1 ; kpH 7 = 0.0814 min−1 and kpH 10 = 0.1248 min−1 . In agreement with previous studies [17,20,21] it can be clearly seen the strong influence of pH on the photocatalytic Hg(II) uptake, being the process enhanced by increasing the initial pH of the solution. The aqueous mercury concentration attained in the photocatalytic treatment at the different experimental conditions evaluated in the present work is displayed in Table 3 (values for organic additives addition are discussed below). As seen, the highest efficiency was obtained at pH 10, conditions where the concentration of Hg(II) measured in the solution after 180 min of irradiation was 112 ␮g L−1 . By contrast, at pH 2 a mercury concentration of 60 mg L−1 remained in the solution after the same reaction time. At the end of the photocatalytic reactions all the catalysts exhibited a dark gray coloration, indicative of Hg0 formation, being the colour as much intense as higher the initial pH conditions. Furthermore, the XRD analysis of the filtered solids (Fig. 4) revealed the co-deposition on the titania surface of calomel (Hg2 Cl2 ) in those reactions carried out under acidic and neutral conditions (Fig. 4b). Peaks attributable to HgO were identified in the XRD patterns recorded following reaction at pH 10 (Fig. 4c). The latter mercury phase is mostly formed upon oxidation of metallic Hg once produced on the titania surface, process which is reported to be favoured at high pH [21]. The distribution of mercury on the titania particles was examined by ESEM microscopy. Fig. 5 shows the micrographs of samples recovered after the reactions performed at pH 2 (a) and 10 (b). As it can be observed, mercury is quite homogeneously dispersed on the

catalyst being the size of mercury moieties deposited on titania in the range of 1–3 ␮m. The higher density of mercury deposits found in the sample obtained at pH 10 is in agreement with the higher Hg(II) uptake achieved in these conditions in comparison to pH 2. Most likely the mercury deposits appearing in the micrographs correspond to calomel (Fig. 5a) and mercuric oxide (Fig. 5b). They showed a high stability during the analysis, in contrast to metallic mercury which evaporates under vacuum conditions inside the microscope. Three main factors may account for the differences in kinetics of Hg(II) reduction observed by varying the pH of the solution: (i) the potential of TiO2 conduction band electrons; (ii) the distinct redox potentials of the different mercury species present at each pH solution; and (iii) the differences in dark adsorption above discussed. The potential of the conduction band of TiO2 is pH dependent and shifts to more cathodic potentials by 59 mV per pH unit [42]: E CB = −0.05 − 0.059 pH

(at 25 ◦ C)

(6)

According to Eq. (6) the driving force of conduction band electrons increases with pH as follows: −0.168 V at pH 2; −0.315 V at pH 4.5; −0.463 V at pH 7; and −0.64 at pH 10. The standard potentials for reduction of HgCl2 , Hg2 Cl2 and Hg(OH)2 species (no data have been found in literature for HgClOH species) are [43]: Hg(OH)2 + 2H+ + 2 e− → Hg0 (l) + 2 H2 O (aq) −



0

2 HgCl2 + 2 e → Hg2 Cl2 + 2 Cl (aq) −



0

HgCl2 + 2 e → Hg (l) + 2 Cl (aq) −

0

Hg2 Cl2 + 2 e → 2 Hg (l) + 2 Cl



E 0 = 1.034 V

(7)

E = 0.63 V

(8)

0

(9)

E = 0.41 V 0

E = 0.268 V

(10)

Therefore, the greater efficiency for Hg(II) uptake obtained at initial pH values of 7 and 10 as compared to acidic conditions can be considered derived from (i) the higher driving force of conduction band electrons; (ii) the thermodynamically feasibility for

Table 3 Hg(II) concentration remaining in the aqueous solution after photocatalytic treatment at different initial pH. Organic addeda

pH 2 [Hg(II)] (mg L−1 )

pH 4.5 [Hg(II)] (mg L−1 )

pH 7 [Hg(II)] (mg L−1 )

pH 10 [Hg(II)] (mg L−1 )

No additive Methanol Formic acid Oxalic acid

59.89 0.627 0.554 0.183

4.405 0.460 0.392 0.051

0.447 0.358 0.020 0.061

0.112 0.042 0.036 0.071

a

C0 = 2 mM.

M.J. López-Mu˜ noz et al. / Applied Catalysis B: Environmental 104 (2011) 220–228

225

Fig. 5. ESEM micrographs of catalysts recovered after photocatalytic reactions performed at: (a) initial pH 2; and (b) initial pH 10.

Hg(OH)2 reduction to Hg(0), reaction which is the most readily favoured according to the standard potentials given for reactions (7–10); and/or (iii) the effective adsorption of Hg(OH)2 species, in contrast to HgCl2 , indicative of their great affinity for the titania surface. As for the results obtained under acidic conditions, the aqueous speciation diagram depicted in Fig. 2 shows that HgCl2 are the dominant aqueous mercury species at pH 2 and 4.5. The redox potential value of these species indicates that it is thermodynamically feasible their photocatalytic reduction with TiO2 to either Hg0 (reaction (9)) or Hg(I) as Hg2 Cl2 (reaction (8)), both of them detected as mercury species deposited on the titania surface after reactions carried out at initial pH 2 and 4.5. A further reduction of Hg2 Cl2 to Hg0 (reaction (10)) should also be expected, however, the high stability of calomel and the low driving force of the process must prevent it to occur. Consequently, both one-electron and complete reduction of Hg(II) take place simultaneously under acidic conditions. It is worthy to note the distinct rate obtained at initial pH 2 and 4.5 (Fig. 3) despite the similar scarce Hg(II) adsorption achieved in both conditions (Fig. 1). The results point out that the photocatalytic uptake of Hg(II) from the solution has not a direct correspondence with the amount of Hg(II) pre-adsorbed. Hence, it can be concluded that under acidic conditions the efficiency of the photocatalytic uptake of mercury is mainly controlled by the driving force of the electrons photogenerated in the conduction band instead of the affinity and extent of adsorption of mercury species on the titania surface. 3.2.2. Influence of organic additives The influence of sacrificial electron donors on the overall efficiency for Hg(II) uptake was investigated by performing the photocatalytic reactions in the presence of 2 mM methanol, formic acid and oxalic acid respectively. Fig. 6 displays the time course of photocatalytic reduction of Hg(II) in the presence of the organic additives at the evaluated pH values. The aqueous mercury concentrations finally attained in the different experimental conditions are summarized in Table 3. The three organics had a beneficial effect on the photocatalytic Hg(II) reduction, enhancing the rate and the extent of mercury uptake. The nature of the organic determined their specific positive influence on the process, being so much more distinctive as pH decreased. At pH 2 it follows the order of oxalic acid > formic acid > methanol for both kinetics (Fig. 6) and final mercury concentration attained (Table 3). In these pH conditions it is worthy to notice the significant effect of the three additives that resulted in a final Hg(II) concentration remaining in solution below 0.6 mg L−1 as compared to 60 mg L−1 when no organic was added. At pH 4.5 and pH 7 the effect of formic and oxalic acid on the

kinetics is quite similar, better than for methanol. The differences between both organic acids arise in the final mercury level achieved at each pH (51 ␮g L−1 with oxalic acid and 392 ␮g L−1 with formic acid at natural pH; 61 ␮g L−1 with oxalic acid and 20 ␮g L−1 with formic acid at pH 7). At pH 10, no substantial differences were found between the three organics, all of them improving the photocatalytic performance for Hg(II) uptake, achieving final concentrations below 70 ␮g L−1 . Considering the aqueous mercury limits currently legislated [3–7], the results obtained point out the crucial role of organic additives and pH conditions to attain mercury concentrations below 100 ␮g L−1 . The comparison of the results obtained for the dark adsorption (Table 2) and photocatalytic uptake of Hg(II) in the presence of organic additives (Fig. 6 and Table 3) evidences that it cannot be established a direct correlation between Hg(II) dark adsorption on the TiO2 surface and the efficiency of photoreduction bring about by the sacrificial donor. The better results in terms of final mercury concentration attained at pH 4.5 with oxalic acid as compared with formic acid could be related to the enhancement in Hg(II) adsorption induced by oxalic acid. However, there is not a general trend to support a correlation between the extent of adsorption and photocatalytic Hg(II) reduction. As an example, the slight effect on mercury adsorption exerted at pH 2 by methanol, oxalic and formic acid does not correlate with the effective and distinct photoreduction achieved in the presence of each additive. Neither the greater enhancement in Hg(II) adsorption induced by the presence of oxalic acid as compared to formic acid at pH 7 can be directly associated with the reaction results. The catalysts recovered by filtration once the reaction had been finished showed a dark gray coloration indicative of metallic mercury deposition. Moreover, the comparison of the samples recovered after the reactions performed at acidic conditions with and without organic additives revealed a higher colour intensity, hence a higher amount of deposited Hg(0), in the former ones. The XRD diffraction patterns (not shown) revealed calomel (Hg2 Cl2 ) as the only crystalline mercury phase deposited on the titania surface, exclusively in the samples recuperated from the reactions performed at pH 2 in the presence of methanol or formic acid. By contrast, no Hg2 Cl2 deposition on the solid was detected in the reaction performed at pH 2 upon oxalic acid addition neither in any of the catalysts recovered after reactions carried out at pH 4.5, 7 or 10. The ESEM analysis of the solids revealed that the quantity of calomel deposited on titania in reactions at pH 2 with formic acid or methanol addition was much poorer than that obtained in the absence of organic additives (Fig. 7a). Greater mercury aggregates, probably related to oxalate species were observed in the sample obtained from the reaction carried out at pH 2 in the presence of oxalic acid (Fig. 7b).

226

M.J. López-Mu˜ noz et al. / Applied Catalysis B: Environmental 104 (2011) 220–228

Fig. 6. Photocatalytic reduction of aqueous Hg(II) as a function of time in the presence of ( ) methanol; () formic acid; () oxalic acid and () no organic, at initial pH values: (a) 2; (b) 4.5; (c) 7; and (d) 10.

The addition of oxidizable additives is generally reported to enhance the efficiency of photocatalytic reduction processes and also, in some cases, to make feasible the photoreduction of certain metal ions that cannot be reduced directly by the conduction band electrons [44]. The sacrificial organic donors can trap photogenerated holes either directly (reaction (11)) or indirectly (reaction (12)) therefore improving the photochemical quantum yield by minimizing charge carriers recombination: R + h+ VB → R



(11) •

R + • OH → R + OH−

(12)

The direct photogenerated hole transfer is considered to occur preferentially in the case of chemisorbed organic compounds such as carboxylates, whereas the primary oxidation mediated by trapped holes is believed to take place with weakly or nonadsorbed molecules such as alcohols [45]. Besides minimizing (e− –h+ ) pairs • recombination, both processes may generate radical derivates (R ) with strong oxidizing power that can either inject an electron into the conduction band of the semiconductor (current doubling effect) [14] or react with other species adsorbed on the titania surface. Specifically, the photocatalytic degradation of oxalate is considered to occur through the cleavage of the C–C bond to form • CO2 plus the strong reducing species CO2 − [14], radicals also proposed to be formed upon photocatalytic oxidation of formate ions

Fig. 7. ESEM micrographs of catalysts recovered after photocatalytic reactions performed at (a) initial pH 2 in the presence of methanol; and (b) initial pH 2 in the presence of oxalic acid.

M.J. López-Mu˜ noz et al. / Applied Catalysis B: Environmental 104 (2011) 220–228 •

[46]. The potential for the CO2 /CO2 − couple has been reported as −1.84 V vs. NHE [47]. As for the reaction of methanol molecules with hydroxyl radicals, it proceeds through abstraction of a hydrogen atom from the C–H bond therefore generating ␣-hydroxymethyl • radicals, CH2 OH [48], species that are also strong reductants since • the redox potential of the CH2 OH+ / CH2 OH couple has been estimated as −0.74 V (vs. NHE). • Consistent to their respective redox potential values, both CO2 − • and CH2 OH radicals have sufficient reducing power to promote Hg(II) reduction reactions (7)–(10). It is, therefore, justified the increase in the rate and efficiency of Hg(II) photocatalytic uptake from the solution derived from the addition of methanol, formic and oxalic acids as sacrificial agents. The experimental results evidence that not only the kinetics for mercury removal are enhanced but the chemical form of mercury photodeposited on the titania surface is modified by the presence of the organic compounds. In general, all of them induced in acidic pH conditions a yield to metallic mercury formation better than achieved without organic additives. That means that photocatalytic reduction of HgCl2 and Hg2 Cl2 to Hg(0) (reactions (9) and (10)) is enhanced in the presence of the sacrificial additives, what supports the role of the organic compounds not only as hole traps but as sources of strong reducing radicals. Moreover, the lesser effect on the reduction rate exerted by methanol compared to the carboxylic acids agrees with the smaller reduc• ing power of ␣-hydroxymethyl radicals in comparison to CO2 − species. Besides the double function as sacrificial electron donor and as a source of reducing radicals, in the case of oxalic acid, it is plausible to propose that the better results for Hg(II) uptake achieved in acidic conditions are likely related to its specific chemical affinity for mercury. The presence of HC2 O4 − and or C2 O4 2− species would allow the stabilization of intermediate monovalent mercury Hg(I) as oxalate deposits on the titania surface (Fig. 7) instead of calomel (Hg2 Cl2 ).

4. Conclusions The photocatalytic removal of aqueous Hg(II) from 100 mg L−1 HgCl2 solutions using TiO2 as catalyst was studied. The results obtained evidence that adsorption of mercury species on the titania surface is strongly dependent on the solution pH, due to both the significant variations in speciation of mercury and the charge of titania surface with pH. Whereas HgCl2 species are hardly adsorbed on titania surface, Hg(OH)2 species show a high affinity for TiO2 most likely by interaction with surface – TiOH and – TiO− entities. Also the rate and extent of Hg(II) photocatalytic removal from aqueous solution strongly depended on pH and the presence or absence of organic donors. At pH 10, an efficient removal of Hg(II) was achieved under all experimental conditions evaluated, attaining final mercury concentrations at trace levels (␮g L−1 ). By contrast, at acidic conditions an efficient removal of mercury could be attained only upon addition of the three evaluated organic additives, namely formic acid, oxalic acid and methanol. The nature of the organic compound determined the rate and extent of mercury uptake achieved at each pH. It has been demonstrated that it is plausible to attain by heterogeneous photocatalysis the mercury disposal limits currently legislated (100 ␮g L−1 ) by selecting the suitable organic additive and pH conditions. The comparison of results obtained for adsorption and photocatalytic uptake of Hg(II) from the aqueous solution evidenced that, in general, it cannot be established a direct correlation between Hg(II) dark adsorption on the TiO2 surface and the efficiency of photoreduction achieved. The nature and distribution of mercury products deposited on the catalyst, analysed by XRD and ESEM was dependent on the

227

reaction conditions. In the absence of additives, Hg2 Cl2 and Hg0 were respectively identified in acidic and alkaline media as main reduced species on the titania surface. The addition of organic additives enhanced the photocatalytic reduction to Hg0 what supports their role not only as hole traps but as source of strong reducing radicals.

Acknowledgements Authors thank the Regional Government of Madrid for the financial support through the projects URJC-CM-2008-CET-3756 and REMTAVARES (S-2009/AMB/1588) and to the “Ministerio de Ciencia e Innovación” for the financial support through the projects CTM2009-08649 and the CONSOLIDER INGENIO TRAGUA Network CSD2006-44.

References [1] United Nations Environment Programme, Chemicals, Global Mercury assessment, Report no. 54790-01, Geneva, Switzerland, 2002. [2] U.S. EPA, Mercury Study Report to Congress, EPA-452/R-97-003, 1997. [3] Directive 2008/105/EC of the European Parliament and of the Council of 16 December 2008 on environmental quality standards in the field of water policy. [4] U.S. EPA, National Primary Drinking Water Standards, EPA-816/F-09-0004, 2009. [5] Council Directive 98/83/EC on the quality of water intended for human consumption. [6] U.S. EPA, Great Lakes National Program Office, Great Lakes binational toxics strategy, Management assessment for mercury. Chicago, IL and Environment Canada, 2006. [7] Decreto 130/2003, DOGC 29 de Mayo 2003, Ley 10/1993, de 26 de Octubre. BOE n. 312 de 30/12/1993 (Modified by Decreto 57/2005). [8] J. Aguado, J.M. Arsuaga, A. Arencibia, Micropor. Mesopor. Mater. 109 (2008) 513–524. [9] R.S. Vieira, E. Guibal, E.A. Silva, M.M. Beppu, Adsorption 13 (2007) 603–611. [10] I. Wagner-Döbler, H. von Canstein, Y. Li, K.N. Timmis, W.D. Deckwer, Environ. Sci. Technol. 34 (2000) 4628–4634. ˜ [11] M.J. López-Munoz, J. Aguado, R. van Grieken, J. Marugán, Appl. Catal. B: Environ. 86 (2009) 53–62. ˜ [12] R. van Grieken, J. Aguado, M.J. López-Munoz, J. Marugán, Gold Bull. 38 (2005) 180–187. [13] E. Borgarello, N. Serpone, G. Emo, R. Harris, E. Pelizzetti, C. Minero, Inorg. Chem. 25 (1986) 4499–4503. [14] F. Forouzan, T.C. Richards, A.J. Bard, J. Phys. Chem. 100 (1996) 18123–18127. [15] M.I. Litter, Appl. Catal. B: Environ. 23 (1999) 89–114. [16] D. Chen, A.K. Ray, Chem. Eng. Sci. 56 (2001) 1561–1570. [17] M.R. Prairie, L.R. Evans, B.M. Stange, S.L. Martínez, Environ. Sci. Technol. 27 (1993) 1776–1782. [18] N. Serpone, Y.K.A. You, T.P. Tran, R. Harris, E. Pelizzetti, H. Hidaka, Solar Energy 39 (1987) 491–498. [19] M.A. Aguado, S. Cervera-March, J. Giménez, Chem. Eng. Sci. 50 (1995) 1561–1569. [20] X. Wang, S.O. Pehkonen, A.K. Ray, Electrochim. Acta 49 (2004) 1435–1444. [21] S.G. Botta, D.J. Rodríguez, A.G. Leyva, M.I. Litter, Catal. Today 76 (2002) 247–258. [22] H.E. Byrne, D.W. Mazyck, J. Hazard. Mater. 170 (2009) 915–919. [23] J. Bussi, M.N. Cabrera, J. Chiazzaro, C. Canel, S. Veiga, C. Florencio, E.A. Dalchielec, M. Belluzzi, J. Chem. Technol. Biotechnol. 85 (2010) 478–484. [24] EPA, Method 200.7. Trace elements in water, solids and biosolids by inductively coupled plasma-atomic emission spectroscopy, EPA-821-R-021-01-010, 2001. [25] UNE-EN-13506, Water quality – Determination of mercury – Method using atomic fluorescence spectrometry. [26] W.D. Schecher, D.C. McAwoy, MINEQL+: A Chemical Equilibrium Modelling System. Version 4.5, Environmental Research Software, Hallowell, ME, 2001. [27] L. Pezza, Chem. Rev. 75 (1975) 585–602. [28] M. Molina, C.B. Melios, M. de Moraes, J.O. Tognoli, Talanta (1996) 1697–1704. [29] J. Zhao, H. Hidaka, A. Takamura, E. Pelizzetti, N. Serpone, Langmuir 9 (1993) 1646–1650. [30] L.A. García Rodenas, A.D. Weisz, G.E. Magaz, M.A. Blesa, J. Colloid Interface Sci. 230 (2000) 181–185. [31] N.J. Barrow, V.C. Cox, J. Soil Sci. 43 (1992) 295–304. [32] C.S. Kim, J.J. Rytuba, G.E. Brown Jr., J. Colloid Interface Sci. 270 (2004) 9–20. [33] R. Weerasooriya, H.J. Tobschall, W. Seneviratne, A. Bandara, J. Hazard. Mater. 147 (2007) 971–978. [34] D. Sarkar, M.E. Essington, K.C. Misra, Soil Sci. Soc. Am. J. 63 (1999) 1626–1636. [35] E.A. Forbes, A.M. Posner, J.P. Quirk, J. Colloid Interface Sci. 168 (1974) 87–93. [36] C. Tiffreau, J. Lützenkirchen, Ph. Behra, J. Colloid Interface Sci. 172 (1995) 82–93. [37] P. Bonnissel-Gissinger, M. Alnot, J.P. Lickes, J.J. Ehrhardt, P. Behra, J. Colloid Interface Sci. 215 (1999) 313–322. [38] D. Sarkar, M.E. Essington, K.C. Misra, Soil Sci. Soc. Am. J. 64 (2000) 1675–1968.

228

M.J. López-Mu˜ noz et al. / Applied Catalysis B: Environmental 104 (2011) 220–228

[39] A.D. Weisz, L. García Rodena, P.J. Morando, A.E. Regazzoni, M.A. Blesa, Catal. Today 76 (2002) 103–112. [40] S. Hug, B. Sulzberger, Langmuir 10 (1994) 3587. [41] C.R. Chenthamarakshan, K. Rajeshwar, Electrochem. Comm. 2 (2000) 527–530. [42] M.D. Ward, J.R. White, A.J. Bard, J. Am. Chem. Soc. 105 (1983) 27–31. [43] D.C. Harris, Quantitative Chemical Analysis, sixth ed., W.H. Freeman, Houndmills, England, 2002.

[44] Y. Ming, C.R. Chenthamarakshan, K. Rajeshwar, J. Photochem. Photobiol. A: Chem. 147 (2002) 199–204. [45] M.E. Calvo, R.J. Candal, S. Bilmes, Environ. Sci. Technol. 35 (2001) 4132–4138. [46] L.L. Perissinotti, M.A. Brusa, M.A. Grela, Langmuir 17 (2001) 8422–8427. [47] W.H. Koppenol, J.D. Rush, J. Phys. Chem. 91 (1987) 4429–4430. [48] L. Sun, J.R. Bolton, J. Phys. Chem. 100 (1996) 4127–4134.