TOP Purses Seance ano TeenoLocy or Abeasced Maes ‘Sci: Tetmol, Ad Mate. 13 (012) 015008 (pp) 0.188 46996371508 Functionalized diatom silica microparticles for removal of mercury ions Yang Yu', Jonas Addai-Mensah' and Dusan Losie!? “Tan Wark Research Institute, Univesity of South Australia, Mawson Lakes, SA 5095, Australia 2 School of Chemical Engineering, The University of Adelaide, Adelaide, SA S005, Australia E-mail: dasa Josie @ delide eda. au Received 7 August 2011 Accepted for publication 15 December 2011 Published 9 February 2012 Online at stacks.iop.ory/STAM/13/015008 Abstract Diatom silica microparticles were chemically modified with self-assembled monolayers of 3-mereaptopropyl-trimethoxysilane (MPTMS), 3-aminopropyl-trimethoxysilane (APTES) and »n-(2-aminoethy)-3-aminopropyl-trimethoxysilane (AEAPTMS). and their application for the adsorption of meseury ions (Hg(ID) is demonstrated. Fourier transform infrared spectroscopy and x-ray photoelectron spectroscopy analyses revealed that the functional groups (-SH or -NH,) were successfully grafted onto the diatom silica surface. The kinetics and efliciency of He(ll) adsorption were markedly improved by the chemical functionalization of diatom microparticles, The relationship among the type of functional groups, pH and adsorption efficiency of mercury ions was established. The Hg(Il) adsorption reached equilibrium within {60;min with maximum adsonption capacities of 185.2, 131.7 and 169.5 mg g~' for particles functionalized with MPTMS, APTES and AEAPTMS, respectively. The adsorption behavior followed a pseudo-second-order reaction model and Langmuirian isotherm, These results show that mereapto- of amino-functionalized diatom micropasticles are promising natural, cost-effective and environmentally benign adsorbents suitable forthe removal of mercury ions from aqueous solutions Keywords: diatoms, diatom silica, organosilane mocifcations, 3-mereaptopropyl-trimethoxysilane (MPTMS), mercury adsorption 1. Introduction ‘The fabrication of microscale materials with three dimensional (3D) morphology and well-defined nanoscale features presents one of the biggest challenges in material science and nanotechnology. Nature, on the other hand, hhas developed a wide variety of ‘highly sophisticated Diomineralization processes to create solid structures with complex 3D morphology and hierarchical organization with nanoscale precision [1]. Diatoms, single-cell photosynthetic algae having an extraordinary 3D porous structure with miero- to nanoscale dimensions of their silica shes (rustules) are recognized as an outstanding example of nature-designed nanomaterials [2,3]. Their distinet pore structures with unique mechanical, transport and photonic properties, moderate Hos. 09961201500861183300 sueface area and good biocompatibility made this biomaterial highly atractive for many applications, including photonics, biosensing, filtration, adsorption, microfluidic, catalysis, drug delivery and nanofabrication [3-5]. Diatom silica material is available from two sources, which include diatom cell cultures (small quantities) and diatomaceous earth mineral (DE, fossilized diatoms) in large quantities. Hence, it can be considered as an inexpensive biomaterial with nanoscale dimensions offering great perspectives for the development of new materials with advanced properties, Considerable research efforts have been devoted to modifying diatom silica structure into technologically more suitable functional materials such as inorganic oxides (MgO, TiO>, zeolites), metals (Au) and polymers (polyaniline) [3-7]. In £2012 Nuon Ista for Materials Science Pit in the UK‘Sci Tet Ad, Mat. 13,2012) 015008 Yara! ‘our previous work, we showed that characteristic optical, catalytic and magnetic properties can be generated by the replication or structural and surface modification of diatom architecture with titania, gold or magnetic nanoparticles and used for diverse applications including molecular separation, catalysis and drug delivery [5, 8-13] Mercury is recognized as one of the most harmful pollutants in the environment, because of its high toxicity, Volatility and bioaecumulation. ‘There are many sources ‘of mercury pollution including natural processes such as volcanoes and flooding of river basins, industrial activity such as gaseous emissions of fossil fuel combustion, and industrial processes including chloralkali production, paint, pulp, fertilizer, pharmaceutical, rubber and paper industries, oil refineries and municipal solid waste teatment [14-16 Commonly accepted methods for the removal of mercury from wastewater include sulfide and chelate precipitation, iton and aluminum coagulation, ion exchange, reverse ‘osmosis, membrane filtration, electrodeposition, activated carbon adsomption, bioadsorption and photoreduetion [17-19] ‘Among them, the adsorption using solid porous materials as adsorbents is considered as the most reliable, simple, elfcient and cost-effective method. Numerous synthetic and natural adsorbents have been explored, including activated carbon, polymer resin, chitosan, starch, zeolite clays, perlite, fly ash and synthetic: mesoporous sitiea (MCM-41- and ‘SBA-15) [20-22]. However, since commercially activated cearbons and synthetic mesoporous materials ate expensive or their production involves the use of toxie materials, the search for alternative and cheaper natural sorbents is a particularly attractive option. For efficient mercury removal from water, porous adsorbents. should have good chemical affinity to both inorganic and organic mercury, high adsorption speed and adsorption ellicieney for low and high concentrations, high surface area, high mechanical strength, good thermal stability and low cost, Diatom silica in the form of porous microparticles from natural mineral, which retains diatoms’ skeletal (lrustules) structures with distinctive shapes and intsicate mesoporosity, has been recognized as a superior and inexpensive adsorbent of heavy metals from water [5, 6] I is am inexpensive (~$200 per ton), chemically inert, biocompatible and environmentally friendly materia, panicularly suitable for environmental applications. A umber of studies were performed to investigate the potential of DE as an absorbent for removing radionuclides, heavy metal ions (e.g. Pb(ID, Zn(lf, Cu(Ml), Cad) and Hg¢tD) and textile dyes from wastewater [22-24], While these studies have demonstrated a promising potential of DE for ‘wastewater treatment, most of them used raw DE materia, ‘which is not optimized for specific applications. DE has a mxterate surface atea (20-100 m? 2~!) and excellent porosity Its surface chemistry is based on hydroxyl groups including ngle silanol groups (-SiOH), geminal silanol groups (-Si(OH)2) and -Si-O-Si bridges, which have a limited adsorption capacity for specific metal ions such as Hg(I). ‘The importance of having specific metal-binding functional ‘groups (eg. -SH) at the DE surface was recognized by Fowler era, who functionalized an (NH, )sHF-pretreated DE. surface with 3-mereapfopropyt-trimethoxysilane (MPTMS) and vinyluriethoxysilane (VTES) using a sol-gel procedure [25]. While the adsorption capacity of Hg(II) was improved in comparison with that of untreated DE, the results were not impressive, showing that only 3% of -SH sites on the functionalized DE surface were accessible to Hg(ll) species [25]. This low efficiency can be explained by the oxidation of -SH groups or extensive unwanted ‘organosilane polymerization on the DE surface. Hence, further studies are necessary to improve the functionalization ‘of DE and provide a high density of active Functional groups. In this work, we explored the surface functionalization ‘of diatom microparticles with self-assembled monolayers of selected organosilanes containing mereapto (SH), amino (NHL), and ethylenediamino groups (-NH-(CH2)>-NH2), aiming 10 demonstrate the improved adsorption properties Of functionalized DE for the removal of Hg(ll) ions from ‘aqueous solutions. Our proposed approach, showing the typical structure of diatom microparticles and scheme of allached organosilane molecules, is presented in figure 1 We have studied the performance of organosilane-modified diatom panicles forthe adsorption capacity efficiency toward Hg(ll) ions, dependence of functional groups, contact time, initial metal ion concentration and pH. Mercurie chloride (HgCla) was used as the source of mercury ions (Hg(ID), a8 chlori waters and among the inorganie mercury complexes ¢ ions are almost ubiquitous in natural and industrial 2, Experimental section 2. Materials DE samples (white rocks, ~2em in size) were provided by Mount Sylvia Pry Lid (Queensland, Austaia). They were mechanically cnished, pulverized, ‘sieved (20 em sive) washed with Mili-Q water and dried [12], yielding fine (<20,um) diatom pariles. Mercury <(Il) chloride was supplied by Merck and toluene by Chem Supply (Australia). Organosilanes, including" MPTMS, S.aminopropybtrimetboxysilane (APTES) anda aminoethy)-aminopropyLtimethoxyslane_ (AEAPTMS) were purchase from Sigma Aldrich (Ausra). Hiah-purty Milli-Q water (specific conductivity <0.5 Sem", surface tension 72.8mN mand pH = 5.6 at 20°C) was used forthe reparation of ll the aqueous solutions 2.2. Surface modification of diatom microparticles Diatom particles. modified with organosilanes (MPTMS, APTES or AEAPTMS) were prepared by a standard silanization procedure used forthe deposition of organosilane monolayers on mesoporous substrates [26]. Brielly, 3g. of diatom particles was suspended in 150ml of toluene in a three-necked $00 ml round-bottomed flask, which was fitted witha thermometer a reflux condenser and a Nz gas tube and maintained under a dry No atmosphere. Milli-Q water (1.6m) was then added into the mixture and sted for 2h at ambient ‘temperature 1o allow the water to disperse through the diatom‘Sci Tet Ad, Mat. 13,2012) 015008 Yara! Diatom ~e Figure 1. Surface modifiation of diatom microparticle for removal of mercury ions: SEM images of diatom structure and schemes of ‘organosilane molecules covalently attached on datom silica surface particles and make a fully hydroxylated silica surface, Then, 8.5 ml of organosilane (MPTMS, APTES or AEAPTMS) was added into the mixture, which was taken to reflux for 6 at 660°C. Afier that, the mixture was cooled and washed with toluene, 2-propaniol and Milli-Q water 3 dimes, The samples ‘modified with MPTMS, APTES or AEAPTMS were dried in 4 vacuum desiccator for 3 days 2.3. Character Scanning electron microscopy (SEM, XL-30, Philips) was used to observe the morphology of diatoms before and alter ‘modifications. The size disibtion of diatom microparticles in water solution (1% w/v) was determined using the Malvern Masiesizer X, Malvern (UK), The zeta potentials were measured using the Nano-ZS Zetsizer in the elect phoretic light-scattering mode (Malvern Instruments Ltd Worcestershire, UK) using. uted. panicle suspensions. We prepared 0.01 wt suspensions by adding 0252 of dy pristine or rganosilane-modiied (MPTMS, APTES or AEAPTMS) diatom particles 10 50m! of a 10-?M KNOs solution followed by magnetically stirring the suspensions for 20min. The suspensions were then allowed to stand for S min andthe colloidal particles (< yam in size) inthe supematan were siphoned off forthe zeta potential measurement in the pH range of 2-11 Fourier vansform infared spectuoscopy (FTIR) analysis of diatom particles was used t determine functional groups fon their surface after modification. FTIR spectra were recorded witha Nicolet Magna-IR 750 spectrometer. KBr pellet was prepared using a 1/100 proportion of diatom panieles/KBr. For each specirum, 128 scans were recorded at em”! resolution, in the range of 4000-400em-" in ‘uansmitance mode, X-ray photoelectron specuoscopy (XPS) analysis was performed using an AXIS Ultra DLD spectrometer (Kratos Analytical Li) equipped witha monochromatic Al Ka source (1486.6eV). Elements present were identiied from survey spectra and high-resolution spectra, which were collected at 160 and 40eV, respectively. Data were recorded in the VAMAS format and processed using the CasaXPS software package X-ray powder diffraction (XRD, Panalytical PW 3040) was used (0 characterize the crystallographic structure of diatom samples. X-ray fluorescence spectroscopy (XRF. ‘Siemens SRS3000) and inductively coupled plasma mass spectrometry (ICP-MS, Perkin Elmer Optima 9000) analyses zzave the elemental composition in terms of the oxide ‘components ofthe samples and solution metal concentration, ‘Thermogravimetric analysis (TGA) was carried out using ‘a Hi-Res Modulated TAG 2950 analyzer at a heating rate of 10°C min-" from 30 to 800°C in a nitrogen gas flow. ‘The specific surface area (Brunauer-Emmett—Teller, BET analysis) of diatom samples was Micromerities ASAP 2000 system, sasured using 24, Adsonption experiments Adsorption isotherm experiments were carried out by varying the inital Ha(l}) concentration from 1 to 400mgI~! to investigate the adsorption behavior of diatom particles modified with MPTMS, APTES or AEAPTMS. Standard solutions of Hg(I) were prepared by diluting of 4()meI! Heil) (HCl; in solution) with Milli-Q water. Then, 0.01 2 fof adsorbent. was dispersed in 40m! of Hg(ll) solution ‘The mixture was magnetically sired for 60min to ensure equilibrium. The residual particle suspensions were filtered through a 0-45 jm membrane and the filtered solutions were analyzed for Hg(ll) concentration using ICP-MS. (Perkin Elmer Elan 9000), The effects of pH, contact time and initial Ha) concentration on the Hai) adsorption kinetics and isotherms were studied and all the experiments were replicated 3 times. Ha(ll) adsorption capacity q (mg.g~) and clciency were calculated using the following equations [9] u-Ov z o‘Sci Tet Ad, Mat. 13,2012) 015008 ‘igure 2, (a) SEM image of diatom particles from DE material ‘sed inthis works (b,c) SEM images of single diatom particles botox and after surace modification with MPTMS-DE.(e) Particle size distribution of pristine diatoms and diatoms modified with MPIMS, APTES or AEAPTMS Adsorption efficiency = x 100%, Q) where Co and C are the initial and equilibrium concentrations ‘of He(tl) (mel-"), respectively, V is the volume of solution (and W isthe dry mass of adsorbent (2) 3. Results and discussion 3.1, Structural and chemical characterization of pristine and ‘modified diatom microparticles ie 2 presents SEM images of the typical morpholo structure of the diatom microparticles (frustules) used in this work, showing their perforated cylindrical shell with aan opening at one end [12, 13]. Regularly spaced rows ‘of pores of ~300-S00nm diameters are located along the frustule shell wall (figure 2(b)). SEM images of diatoms modified with MPTMS, APTES or AEAPTMS confirmed that their moephologies and pore structures are the same as in the pristine diatoms (figure 20). The 10th, 50th and 90th percentile particle sizes of the pristine diatom Yara! fay spa “ps Absorbance » ah t V/ -—4)44Hs—§~— ee +) 4 TERE ‘Wavenumber cn" Figure 3. FTIR spectra of pristine diatoms and diatoms madiied with MPTMS, APTES or AEAPTMS, sample were 1.2 5.8 and 12.7 jm, respectively (igure 2d) indicating that major diatom particles are whole frustule structures with minor fractures and aggregations. Again, no siamiticant change of particle size distribution was observed in the modified diatoms (MPTMS-DE, APTES-DE and AEAPTMS-DE, confining the formation of an uluathin layer of organesilanes on the diatom’ surface without aggregations. The BET surface area of pristine diatom 12més, whieh miroparicles was determined as 20.3 is a moderate surface area for their proposed application in the adsorption of metal ions. The average pore diameter and pore volume of the pristine diatom particles were 12:41 and (L0S2mlg~!, respectively. The BET specific surface area (~19.3n0 g"), pore diameter (~11.51nm) and volume (O049mig~!) of orwanosilane-mositied diatom particles were similar to those ofthe pristine diatoms. XRD analysis showed that the diatom microparticles mostly consist of amorphous silica with a ace amount of erystalline quartz, which is consistent with our previous studies (12, 13], XRF analysis revealed that SiO» (~90.2%) was the main oxide component in pristine diatoms. with impurities such as AlsOs (~2.34), Fe:Os (~0.8%), Mn, Mg. and Ca as trace elements, A similar composition, but sith a noticeable increase in C, $ or N concentration was deduced from the XRF analysis of organosilane-modifed samples, Figure 3 shows the infrared absorption spectra of the pristine and organosilane-modified diatoms (MPTMS-DE,‘Sci Tet Ad, Mat. 13,2012) 015008 inline Ena Figure XPS spectra (a)§ 2p for MPTMS-DE, (b)N Is for APTES-DE an (c) N Is for AEAPTMS-DE samples ‘Table 1. Atomic composition ofthe studied diatom particles termined from XPS analysis, Sample fone composition OCs) PrisineDE 174 S61 234 00 MPrMS. 100 S267 31 0. APIES-DE 184 48286 031 AEAPTMS-DE 188 399 241 033 APTES-DE and AEAPTMS-DE). The vibrations observed at 468.3 and L101.1/em™" correspond to asymmetric stretching modes of Si-O-Si bonds whilst the AI-O-Si stretching ‘bration is indicated at $00.8 em" for both the pristine and ‘onganosilane-modifed samples 27, 28). Aditional peaks at 1476.8, 1585.1, 2942.9, 2871.7 and 2566,8em™"! are observed inthe MPTMS:DE, APTES-DE and AEAPTMS-DE samples. ‘The vibrations at 2942.9 and 2871.7em~! are associated with the asymmetric and symmetric stretching moses of the CH. moiety, respectively, which is directly related to the earbon chain of organosilane molecules [29, 30]. In the spectra of MPTMS-DE, the ~SH stretching vibration occurs at about 2566.8em~! and is associated with the functional group of the adsorbed MPTMS [29]. The scissor vibration ofthe -NH2 terminal group (1585.lem~") and the deformation mode of the -NH; group (1476.8cm") are observed in the spectra of APTES-DE and ABAPTMS-DE samples. Both these 'moxies are associated withthe -NH groups from APTES or AEAPTMS organosilanes (31). These FTIR results clearly indicate the presence of -SH or -NHz functional groups ‘on the diatom surface originaing from MPTMS, APTES or |AEAPTMS. ‘The XPS analyses ofthe elemental surface compositions of pristine and MPTMS-DE, APTES-DE and AEAPTMS-DE panicles confirmed changes in the atomic concentrations of Si, C, Nand S asa result of the moditication process (igure + and table 1). The 23.4% C concentration in the pristine diatoms is due to the surface contamination. The inerease in carbon concentration from 23.4 to 26.7, 28.6 and 24.1% in functionalized diatoms may correspond (0 the introduction ‘of carbon chains from the MPTMS, APTES and AEAPTMS dorganosilanes, respectively, 10 the diatom surface (32, 33] ‘The surface functionalization by organosilanes was confirmed by high-resolution XPS measurements. The increase in the atomic concentration of sulfur from O to 3.1% in the MPTMS-modified sample indicates a suecessful surface funetionalization with -SH groups [32]. The S 2p XPS signal (igure 4(a)) appears only in the MPTMS-teeated sample, at around 163.18eV with two peaks $ 2p! at 163.07eV and 'S2p!? at 164.25eV), which ean be attributed to the reduced sulfur (-SH) in the MPTMS mereapto group [34]. The N 1s peak (figures 4b) and (c) is only present in the APTES-DE, and AEAPTMS-DE samples and the atomic concentration ‘of nitrogen inereases to 3.1 and 3.3%, respectively. The NN Is peak was decomposed into two components: one at about 3992eV was attributed to the -NH group and the ‘other component at about 401.7eV indicated the presence of protonated amines (NHS) (31, 32]. These XPS cesuts confirm thatthe diatom siiea surface is covered with -SH or =NH2 groups from covalently attached MPTMS, APTES or AEAPTMS molecules TGA analyses provide further supporting evidence of the grafting of MPTMS, APTES and AEAPTMS onto the at pH > 5, These species will affet the adsorption‘Sci Tet Ad, Mat. 13,2012) 015008 1g) removal ficiency 96 Hig) removal efciency "6 ‘AEAPTMS-DE |'"" ls i removal ficiency 96 [Figure 8. Etfect of pl onthe extraction of Hel) ions by (a) MPTMS.DE, (6) APTES-DE and (¢) AEAPTMS-DE particles Inia concentration of Hg is Lg" (60min of eoatac time), cllicieney of Hg(II) in the solution upon interaction with ‘protonated and deprotonated surface sites, which are affected by pl. Figure 8 shows the pH dependence of the adsorption capacity and efficiency obtained for organosilane-moditied diatoms inthe pH range from | to 9. These graphs show that the adsorption capacity for Hg(ID) was remarkably affected by pH, and the maximum adsorption by diatoms modified with MPTMS, APTES or AEAPTMS is observed in the pH range of 5-9. At pH <5, decreasing trends in adsomption were observed as the HgCl; species became dominant in the Solution [4]. These results reveal that the nature of Heg(ll) species in solution at various pHs plays an important role in the interactions of the functional groups. (SH, =NHz and -NH-(CH2)-NH2) with the Hg(D) species, which determine the adsorption mechanisms. Hg(Il) is known to have a strong affinity for halogens and ligands donating sulfur or nitrogen. At pH <5, the decreasing trend in the ausorption of Ha() is due to the compet the formation of mereury-sulfur or mereury-nitiogen bonds during the adsorption onto diatom particles modified with MPTMS-, APTES- or AEAPTMS and the stabilization of the chloro-compleses (gC) in solution [45, 46]. At pH > 5 Where Hg(OH)> prevails, Ha) species can be effetvely removed fom solution by their specific interactions withthe functional groups (SH, NH: and -NH-(CH.):-NH2) of MPTMS, APTES or ABAPTMS-modifcd diatoms [4 At pH'=6, MPTMS-DE shows the highest adsorption efcieney (02.1%) of Hp(l) compared with APTES-DE and ‘AEAPTMS-DE. This result indicates that sulu-donating Tigands (-SH) have a stronger capability of capturing Hyd) than niteogen-donaing ligands) NH; and -NH-(CH:)2-NH).-‘The adsorption eflcicey of AEAPTMS-DE (82.5%) is higher than that of APTES-DE (65.2%), which may be explained by the formation of the more stable diamino chloro-Haill) complex between the Hail) species and ethylenediamino groups originating from the AEAPTMS-DE surface [47] Since all three adsorbents show the maximum adsorption efficiency for Het) at pH=6, this pH was selected for the other adsorption experiments and subsequent analyses, 3.3, Adsomption kinetics ‘The adsorption kinetics of diatom microparticles modified with MPTMS, APTES or AEAPTMS as function of different intial concentrations of Hg(ll) showed that adsorption occurred extremely rapidly a the beginning and reached equilibrium within 60min (figures 9a) and (©). “The equilibrium adsorption capacity (qe) ofall the three adsorbents increased significantly with increasing initial concentration of Ha(l}) from 1 10 30mg. This is due © the increase in inital concentration of Hai, providing a greater driving force for mass transfer and subsequent surface adsorption oto the diatom particles [48] Both pseudo-irst-ordr and pscudo-second-order emp tical power law models maybe used to evaluate the adsorption of Hal) onto the surface of diatoms modified with MPTMS, APTES or AEAPTMS. Empirical first- and second-order kinetic models are expressed with equation (1) for reaction exponents = I and 2 espectively [49] et ” Here g, (mgg"!) and qe (mgg~!) represent adsorption eapacity at contact ime # and at equilibrium, respectively: ky and Ky are the ise and second-order kinetic rate constants, respectively. Inegrating equation (4) with the boundary conditions # =0, 4, =0 and =F, 4 =4, tothe linear form ives equations (5) and (6) forn = I and 2, respectively logta—40) “ tote o‘Sci Tet Ad, Mat. 13,2012) 015008 Yara! ae) eof MPTMS-DE | oe a = 5mgi "> 40. Dome : Simo zs comet * 29 o & © oi a ° Timeimin ne) ooo % Timelmin AEAPTMS-OE =tmngtt 5mgi Homgt, soma 3b Tevelmin Figure 9, Etec of iil concentration onthe adsorption capacitis of diatom particles modiid with (a) MPTMS, (6) APTES and (¢) AEAPTMS. ‘The rate constants i and ks may be determined from the slope of Tinear plots of log(q,—q,) versus # and the intercept of the linear plot of 1/4, versus, respective! ‘AS can be seen from tables 2a) and (b), the pseudo-second-order model fits the experimental data better than the pseudo-firs-order model. The calculated ge values (Cable 2(b)) from the pseudo-second-order model were also in good agreement with the experimental values 4. (exp) ‘Thus, an empirical second-order model is suitable for the description of the adsorption kinetics of Hg(I}) species onto modified diatom panicles. The ks values (table 2(b)) indicate that the initial adsomption rate for organosilane-moditied diatoms is much higher at a low Hg(II) concentration of TmgI' and the adsomption can reach equilibrium faster than in the solution with higher Hg(I1) concentrations, ‘The rate constant ky is larger for the meteapto-erminated (MPTMS-DE) than amino-terminated diatoms, This result may be explained by the stronger metal-binding capability of the ~SH functional groups from the MPTMS-DE surface that allows the adsorption reaction to proceed faster. As for amino. and ethylenediamino-terminated adsorbents, the rte constant ky of AEAPTMS- DE is higher than that of APTES-DE, which may be attributed to the stronger alfinity between the Hg(I}) species and ethylenediamino groups from the AEAPTMS-DE surface 3.4, Adsonption isotherms Langmuir and Freundlich isotherms are commonly used for molecular adsorption at interfaces and for the prediction of equilibrium parameters [49]. The Langmuir model is based ‘on the monolayer adsorption onto homogeneous active sites ‘on adsorbents and can be expressed as [50], cn © Goan” Gos o Here, qe is the adsorption capacity (mg s~"), Cay is Hg) concentration (gi!) at equilibrium: and’ gay. ate the Langmuir constant (Img~!) and maximum adsorption capacity (mg ¢~") determined by the intercept and slope of the linear plot of Cay/q versus Cy, respectively. ‘The Freundlich isotherm is based on the adsorption onto a heterogeneous surface with uniform energy with no restriction to the formation of a monolayer. It may be expressed as [51] where Ky is the Freundlich adsorption constant (Img!) eee eas eee cooficients R® are significantly higher for the Langmuir model (R® =0,99-1.00) than for the Freundlich model (8 =0.89-0,96), suggesting that the Langmuir adsorption isotherm is more suitable for describing the adsorption of HiIl) onto MPTMS-DE, APTES-DE and AEAPTMS-DE (igure 10). The process may be assumed to occur at substantially homogeneous functional groups and binding sites on the surface of the adsorbents up to a monolayer ‘coverage [47]. Compared with amino-terminated adsorbents {APTES-DE, AEAPTMS-DE), the -SH functionalized adsor bent MPTMS-DE shows the highest maximum adsorption capacity of 185.2mg_g~! for Hg(ID). The adsorption eapacity for Hg(II) of MPTMS-DE is considerably higher than that (of other thiol-DE-based materials (@.0-13.4mg 2~!) prepared using the sol-gel process in ar 25]. The maximum adsorbed amount of Heil) (0.926mMs-") indicates that up to 95.6% of the thiol centers could bind the He(Il) species, which is much higher than @ previously reported value for MPTMS-DE (3%) [25]. Ths significant improvement isthe result of a higher density of active -SH groups on the DE surface and can be explained by the preparation using 4 nonaqueous solvent (toluene) under nitrogen atmosphere. ‘The grafting of organosilane layers of MPTMS in aqueous‘Sci Tet Ad, Mat. 13,2012) 015008 Yara! “Table 2 (a) Pseudo-fst-ordr ad (b)preudo-second-onler kinetic parameters of He(}) adsorption onto mexified diatom microparticles. Sample Cole) _(@) Peendo-rstorder model (mgmt) ke getea ie!) tge"!) MPTMS-DE 1 019 6 Oe 3 Ol | 276 OS O17? 539s 30 oNSt tats APTES-DE 1 Oe usd ol Ste a ani 10 046550988 3 Os 723 OTe ABATMS-DE 1 Ol! | om oa Sus ass ons 1 032 3.8666 30 Old 6880S Sample Collet ge exp) _(b) Pseudo sevond-ores (ga) be qe) (emg mio") (ngs) MPTNS-DE 1 361 1793 35909) 5 960313 ws 1.00 0 am (Olah 2778 1.00, w 6T6t 0.068 6136 098 APTES DE, 1 232 O86 2490.99 5 eo O36 0.99 1561 0.087 Loo 304001 086 09 AEAPTMS-DE 1 328 1.108 099 5 1562228 00 0 1939.6. 099 30-5200 a8 099) “Table 3. Langmuir and Freundlich isotherm parameters Sample anginal model Freundlich model » ae ak ” (onlme"") mee) (ime) MPINSDE 0077 1852100 2261 246089 APTES:DE 00291316 0996651. 1817095 AEAPTMSDE 0037 1695-099 10.732 1.926 0.96 solution in air could lead to the formation of oxidized disulfide moieties and multilayer formation that could significantly decrease the density of active ~SH. sroups ‘The adsorption capacity of functionalized DE. materials reported in this study’ is relatively high compared with those of other low-cost adsorbents. (20.6-89.1 mz g~) [24, 52-4}, modified mesoporous-silca-based adsorbents 20.8-249,0mee~!) [35-59] and modified activated carbon (67.8-121,6mg g~1) (21, 43, 60]. For instance, the kinetics oF H(t) adsorption onto MPTMS-DE particles is considerably faster (-92.1% recovery in ISmin) and the equilbyium adsorption capacity (3.6-185.2mgg~') is at least twice {greater than those reported for MPTMS-modiied. zeolite particles in similar mercury adsorption studies (24. It shoal be noted that although DE particles have mach 4 lower BET specific surface area than the conventional oP Figure 10, Adsorption isotherms of HU) onto diatom particles ‘mexitied with MPTMS, APTES or AEAPTMS‘Sci Tet Ad, Mat. 13,2012) 015008 Yara! “Table 4. Comparison of adsorption capacities for Hat) of ifferent nano- to mesoporous adsorbents. Reterence ‘Adsorbent ‘Shey__ Maximum adsorption capacity (xe) mee gm) [5] MPIMS-functionaized DE 90 90-138 Los) [52] Carbon arose! 7010 0s ea MPTMS-zeoit = = (561 [-fory! thiourea urea-funetionalized SBA-15 3780 032 [57] Ammonium-functionalized magnetic mesoporous silica 427.3 os (58) MPIMS-MCM-AL ass 2490 oss (59) 2-mereapothiazoline-fentionalized SBA- 5070-2200 os 19) 2-mereapothiazoline functionalized MCM-41 S040 1400. 017 ei [Lacylthiosemicarbazide-modifed activated carbon . ons = {60} Hybrid igand-modified activated carbon 6500 1216 019 13) APTES sctivated earn 6500 975 ous ‘This work _ MPTMS-DE 1931882 96 synthetic mesoporous adsorbents, the maximum adsorption References capacity for Hg(ll) at an equal surface area (9.64 is significantly higher for modified DE than for modified mesoporous silica or activated carbon (0.05-0.55 mgm~), 'A comparison of our results and previously reported data is presented in table 4. It suggests that the functionalized DE panicles have superior adsorption efficiency for Hg(II) per unit surface area, which may be ascribed {0 superior ‘organosilane grafting and increased density of accessible active sites on the DE surface. Studies are underway using natural water samples and industrial wastewaters containing several ion species to reveal more information on Heil) adsorption behavior, kinetics, adsorption isotherms and selectivity of functionalized diatom particles. 4. Conclusions Surface modification of diatom panicles with MPTMS, APTES and AEAPTMS was successfully carried out and their application as an effective adsorbent of Hg(Il) was demonstrated. Diatom particles modified with mercapto ‘groups (MPTMS-DE) show higher adsorption efficiency ‘than diatoms modified with amino groups (APTES-DE and AEAPTMS-DE), confirming a stronger capability of sulfur-donating functional groups (-SH) for capturing H(II) ions. The adsorption eapacity for Hg( Il) was affected by pH in the solution and the maximum adsorption was observed inthe pH range of $-9. The Hg(I}) adsorption reached equilibrium ‘within 60 min, resulting in maximum adsorption capacities of 185.2, 131.7 and 169.5mgg~? for MPTMS-DE, APTES-DE and AEAPTMS-DE, respectively. The organosilane-based diatom material is a promising adsorbent for the removal of He ions from aqueous solutions. The approach presented in this work can be used for designing diatom adsorbents of various metal ions, Acknowledgment ‘The authors acknowledge the financial support by the Australian Research Council (ARC LP 0989229) and Mt Sylvia Pty Led [1 Sarikaya M, TamerlerC, Jen A KY, Schulten K and Baneys F "2003 Nat, Mater 2577 [2] Hildebrand M 2008 Chem. Res. 108 4855 [3] Sumper Mand Brunper E2006 Adi: Funct, Mater 16 17 [4] Gordon R, LosieD. Tiffany M A. Nagy §S and Sterrenburg F A'S 2009 Tren Biotechnol. 27 116 [S] LosieD, Mitchell 1G and Voeleker N H 2009 Ad: Mater 212917 [6] Lopez PJ, Descles J, Allen A E and Bowler C 2005 Cure ‘Opin. Biotechnol. 16 180 [7] Bao Zt af 2007 Nature 446 [8] LosieD, Mitchell G and Voeleker NH 2006 Nev J. Chem 30908 [91 LosieD, Mitchell G, La Rand Voelker N H 2007 Ad Funct Mater 17 2439 [10] LosieD, Trini G, Evans PJ, Atanacio A, Mitchell Gand Voelcher N'H 2006 J. Mater Chem, 16-4029 [11] LosieD. Ya ¥, Aw MS, Simovie 8, Thisry B and "Addai-Mensah J 2010 Chem. Commun. 46 6323 [12] Aw MS, Simovie S, Addsi-Mensth J and Lose D 2011 ‘Nanomedicine 6 1159, 113] Yu Y, Addai-Mensab J and Losie D 2010 Langmuir 2614068 [14] Fitzgerald W F and Clarkson T'W 1991 Environ. Health Perspect. 96 159 [15] Boening D W 2000 Chemosphere 0 1335 [16] Bodaly R A, St Louis V L, Paterson MJ. Fudge RJ, Hall BD, Rosenbere D Mand Rodi W 1997 Mer fons Biol. Sys 3250) [17] BesshousseH, Rhlalow , Verchie J Fand Lebrun L 2008 Membr Si, 32597 [18} Veglio F and Beolchini F 1997 Hydrometalrgy 44301 [19 Mictaky P and Ciel A F 2009.4. Hacard. Mater 167 10 [20 BudinosaT, Petrov N, Parra J and Baloutaow V 2008 J. Biron, Manag. 88 165 [21] Gao R et af 2008.1, Hear, Mater 172.324 [22] Merield J, Davids WG, MacRae J D ad Amisbaman A "2004 Wier Res. 383132 [23] Ma X. 11 ¥, YeZ. Yang L, Zhou Land Wang L 2011 1 flaca. Mater 188 148 [24] Zhang X Y. Wang Q C. Zhang § Q. Sun X J and Zhang ZS 2009 J. Hasan Mater 168 1575 [25] Fowler CE, BuchborC, Lebeau B, Patri J, Deladte Cand ‘Walearis A 2007 Appl. Sur. Sei, 283 5485 126] Feng X. Frsxell GE, Wang LQ, Kim AY, LivJ and Kemer KIM 1097 Science 276923 [27] Sheng G er al 2009 Colloids Sur A.339 159 128} Huang J. Liu ¥, Jin Q. Wang X and Yang 2007 J. Haar Mater 143 S31