F2C

DBCa

b

c

d

e

f

a

ARRAA

KCG2FMC

1

villttptm

(UT

0h

Applied Surface Science 288 (2014) 373– 380

Contents lists available at ScienceDirect

Applied Surface Science

j ourna l ho me page: www.elsev ier .com/ locate /apsusc

unctionalisation of mesoporous silica gel with-[(phosphonomethyl)-amino]acetic acid functional groups.haracterisation and application

ario Caldarolaa,b, Dimitar P. Mitevb, Lucile Marlinc,d, Ekaterina P. Nesterenkod,rett Paullb, Barbara Onidaa,e, Maria Concetta Bruzzoniti f, Rosa Maria De Carlo f,orrado Sarzanini f, Pavel N. Nesterenkob,∗

Department of Applied Science and Technology, Politecnico di Torino, Corso Duca degli Abruzzi 24, 10129 Torino, ItalyAustralian Centre for Research on Separation Sciences (ACROSS), University of Tasmania, Hobart, Tasmania 7001, AustraliaEcole Nationale Superieure des Ingenieurs en Arts Chimiques et Technologiquesm, Toulouse, FranceIrish Separation Science Cluster, Dublin City University, Dublin, IrelandCR-INSTM for Materials with Controlled Porosity, ItalyAnalytical Chemistry Department, University of Torino, Via P. Giuria 5, 10125 Torino, Italy

r t i c l e i n f o

rticle history:eceived 26 April 2013eceived in revised form 1 October 2013ccepted 5 October 2013vailable online 14 October 2013

eywords:

a b s t r a c t

A new complexing adsorbent was prepared by chemical modification of mesoporous silicaKieselgel 60 (dp = 37–63 �m, average pore size 6 nm, specific surface area 425 m2 g−1) with3-glycidoxypropyltrimethoxysilane and 2-[(phosphonomethyl)amino]acetic acid (PMA), commonlyknown as glyphosate. The prepared adsorbent was fully characterised using elemental analysis, thermalgravimetric analysis (TGA), acid–base potentiometric titration, Fourier Transform Infrared spectroscopy(FTIR), X-ray photoelectron spectroscopy (XPS), nitrogen adsorption isotherms at 77 K (BET), scanning

hemically modified silicalyphosate-[(Phosphonomethyl)amino]acetic acidunctionalisationetal removal

helation

electron microscopy with energy dispersive X-ray spectroscopy (SEM-EDS). The concentration of bondedPMA groups calculated from the nitrogen content was 0.38 mmol per gram. The adsorption of transitionmetal ions on PMA functionalised silica (HEPMAS) was studied from aqueous solutions having differentpH and the following selectivity was established, Zn(II) < Co(II) < Cd(II) < Mn(II) < Ni(II) < Cu(II). The cal-culated values of distribution coefficients D for the adsorption of ecotoxic metal ions on HEPMAS are5.0 × 104, 4.9 × 105 and 2.6 × 104 for Zn(II), Pb(II) and Cd(II), respectively.

. Introduction

Chemically modified silicas (CMS) represent a significant class ofersatile adsorbents, used intensively over the past decades for var-ous applications, including stationary phases in different modes ofiquid chromatography, adsorbents for solid phase extraction, cata-ysts, uses within sensors, and many other such areas [1]. Obviously,he properties of CMS are defined by the type and porous struc-ure of the silica matrix and the bonded layer, and by the chemical

roperties of the immobilised functional groups. The immobilisa-ion of various complexing or chelating ligands at the surface of

esoporous silica was explored in detail in the pioneering work of

∗ Corresponding author at: Australian Centre for Research on Separation ScienceACROSS), School of Chemistry, Faculty of Science Engineering and Technology,niversity of Tasmania, Private Bag 75, Hobart, Tasmania 7001, Australia.el.: +61 3 6226 2156; fax: +61 3 6226 2858.

E-mail address: Pavel.Nesterenko@utas.edu.au (P.N. Nesterenko).

169-4332/$ – see front matter © 2013 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.apsusc.2013.10.035

© 2013 Elsevier B.V. All rights reserved.

Leyden et al., published in 1975 [2,3], following which many suchmaterials have been prepared as highly selective adsorbents for thepre-concentration and separation of various metal ions from com-plex mixtures [4–7]. There are several reviews which have beenpublished on the application of chemically modified silicas in inor-ganic analysis [8–11].

One of the most effective means of exploiting the superior selec-tivity and separation efficiency of complexing CMS, is via theirapplication as stationary phases in high-performance chelation ionchromatography (HPCIC) [12,13], which has now become a use-ful ion chromatographic method for the determination of tracemetals in high salinity samples. Since the establishment of thisseparation method in 1992 [14], the chromatographic proper-ties of silica bonded with ethylenediamine, 8-hydroxyquinoline,N,N′-iminodiacetic acid, amidoxime, N-hydroxamic acid, and ˇ-

diketone, together with many other such chelating reagents,have all been studied and documented [12,13]. As a rule, com-plexing adsorbents containing immobilised acidic ligands (e.g.iminodiacetic acid or IDA) have provided significantly higher

3 urface

ctbToffaga[Aogiwogs

amaasDuies9assrpcfatwatweAhc0tomiu[f

fcetcauaf

74 D. Caldarola et al. / Applied S

hromatographic efficiencies when applied to the separation ofransition and heavy metals, as compared with complexing adsor-ents having immobilised basic ligands (e.g. 8-hydroxyquinoline).he reason for this effect is related to the higher diffusion speedsf metal cations from the “external” solution, towards the sur-ace of the adsorbent with bonded acidic or negatively chargedunctional groups, and with notably faster kinetics of complexssociation/dissociation. As a result, silica based adsorbents withrafted IDA [15,16], amino acids (Asp, Glu) [17,18], nitrilotri-cetic acid [19], (poly)itaconic [20] and (poly)butadiene-maleic18] acids, demonstrated superior separation efficiency in HPCIC.ll of the above listed reagents are classed as amino carboxylicr dicarboxylic ligands, and each contain at least one carboxylicroup within the molecule. For this reason, selectivity exhib-ted by this class of complexing adsorbents often appears similar,

hich provides a significant reason to explore the preparationf complexing adsorbents with other types of acidic functionalroups, which can provide alternative and perhaps more specificelectivity.

A possible alternative to aminocarboxylic and dicarboxyliccid containing chelating ion-exchangers for the separation ofetal ions, could be silica with immobilised phosphonic and

minophosphonic acid derivatives. Sumskaya et al. reported thepplication of N-propylimino(dimethylphosphonic) acid bondedilica for the micro-column chromatography of copper (II) [21].ue to the large sized particles, 40–100 �m, and a short col-mn length, the authors studied the elution of only one metal

n a frontal mode of column liquid chromatography. Garcia-Vallst al. [22] however used a chromatographic column of dimen-ions 150 mm × 4.6 mm I.D., packed with silica particles of diameter.5–11.0 �m, with covalently attached ethylphosphonic acid, andpplied this to the separation of seven heavy lanthanides in aeparation time of 15 min. The column demonstrated very goodelectivity but quite poor efficiency, in this instance once againelated to the relatively large particle size and additionally a broadarticle size distribution. Shaw et al. used a dual phosphonate- andarboxylate-functionalised ion-exchange column, Dionex CS-12A,or the separation of alkaline-earth, transition and heavy metalsnd lanthanides [23,24], and Nesterenko et al. [25–27] have inves-igated the retention of various metal ions on columns packedith 5 �m silica particles bonded with aminomethylphosphonic

cid silica (APAS). In the latter example, under optimised condi-ions both excellent selectivity and high separation efficienciesere achieved, as demonstrated by the baseline separation of

ight metal ions on a 250 mm × 4.6 mm I.D. column in 20 min.PAS was synthesised via the reaction of aminopropylsilica withypophosphorous acid and formaldehyde, resulting in a surfaceoncentration of 0.10 mmol g−1 of phosphonic acid groups and.35 mmol g−1 of unreacted aminopropyl groups. The authors pos-ulated the presence of the latter groups were responsible for thebserved broad peak shape of copper(II), due to possible involve-ent in the coordination of Cu2+ ions. More recently, an impressive

socratic separation of 11 metal ions in 20 min was obtainedsing a 150 mm × 4.0 mm I.D. column packed with 3 �m APAS12], however once again a relatively broad peak was recordedor Cu2+.

Given the above observations with APAS, clear potential existsor the development of new improved schemes for surface modifi-ation of silica particles with such ligands, to minimise secondaryquilibria between separated metals and chelating groups, and forhe characterisation and immobilisation of new phosphonic acidontaining ligands, for application to selective metal ion complex-

tion. In relation to the latter challenge, the following paper reportspon the use of 2-[(phosphonomethyl)amino]acetic acid (PMA) as

potential chelating ligand, and its immobilisation on a silica sur-ace. The aim therefore of the present work was to explore the

Science 288 (2014) 373– 380

preparation and characterisation of such a new silica based adsor-bent, with covalently bonded PMA (glyphosate).

2. Experimental

2.1. Chemicals

All reagents were of analytical reagent grade obtained fromSigma–Aldrich (Sydney, NSW, Australia) and were used as sup-plied unless stated otherwise. Solutions were prepared in Milli-Qwater (Millipore, Bedford, MA, USA.). The 37% solution of HCl waspurchased from Scharlau (Gillman, SA, Australia). Stock solutionof cadmium, lead, beryllium and zinc were prepared from Titrisolstandard solutions (Merck, Darmstadt, Germany) or Spectrosolstandard solutions (BDH Laboratory Supplies, Poole, UK).

Aqueous solutions of metals were prepared using copper, nickel,cadmium, cobalt, zinc, and manganese nitrates or chlorides, pur-chased from Sigma–Aldrich (Dublin, Ireland). Glacial acetic acid,nitric acid sodium acetate trihydrate, boric acid and sodiumhydroxide (all from Sigma–Aldrich) were used for preparation ofacetate and borate buffers. 4-(2-pyridylazo) resorcinol (PAR) waspurchased from Acros Organics (Dublin, Ireland) and a 0.025% solu-tion in 0.05 M borate buffer solution (pH 9.18) was used for thespectrophotometric determination of metals. Deionised water wassupplied from a Milli-Q system (Millipore, Bedford, USA). 20 mMdipicolinic acid (Sigma–Aldrich, Dublin, Ireland) was used for theremoval of adsorbed metal ions if required.

2.2. Instrumentation and methods

Scanning electron microscopy (SEM) images were obtainedwith a JEOL JXA-8600 (JEOL Italy, Milan, Italy). FTIR spectra ofthe PMA bonded silica or, more correctly, 2-hydroxyethyl-PMAbonded silica (HEPMAS), were obtained with a Bruker Equinox 55(Milan, Italy) spectrometer equipped with a mercury cadmium tel-luride (MCT) cryodetector at a resolution 2 cm−1. Silica powderswere pressed into thin self-supporting wafers, then placed intoa quartz cell and degassed for 2 h under dynamic vacuum (resid-ual pressure/10−3 mbar) at 150 ◦C before spectra recording. X-rayphotoelectron spectroscopy (XPS) analysis was performed with aVersaProbe 5000 (Physical Electronics Instruments, Chanhassen,MN, USA). The source of radiation was the monochromatic Al K�radiation.

Low-temperature nitrogen adsorption/desorption measure-ments were performed using a Quantachrome Autosorb 1 (BoyntonBeach, FL, USA) instrument, following a 1 h degassing stepat 150 ◦C. Pore size distribution was calculated from nitrogenadsorption/desorption measurements using the DFT/Monte-Carlomethod. The Brunauer–Emmett–Teller (BET) method was used tomeasure specific surface area and mesopore volumes of bare silicaand the HEPMAS.

The acid–base potentiometric titration of adsorbents wasperformed with Titrando 809 autotitrator (Metrohm, Herisau,Switzerland) equipped with a combined glass electrode, Metrohm6.0262.100, having an built-in reference Ag/AgCl electrode. Themass of the adsorbent used for titration was 0.3 g in all experimentsand 26.12 mmol L−1 HCl was used for the titration in the presenceof 0.5–1.0 M NaCl.

To evaluate the adsorption capacity of the adsorbents, tests wereperformed by suspending 100 mg of silica substrate into 50 mL ofsodium acetate buffer solution (pH 4.75) containing 0.18 mg L−1 of

Be2+, 1.3 mg L−1 of Zn2+, 4.14 mg L−1 of Pb2+ and 2.14 mg L−1 of Cd2+.The adsorbent-metal mixtures were kept on a mixer platform for24 h to allow the powder to adsorb the maximum quantity of metalpossible and to reach a condition close to equilibrium.

rface Science 288 (2014) 373– 380 375

aVepcc

Wmdn

Uu

2

bGh(iggrrit

tPT(osls3tTaitfatoTisnb

sdlG

cc(tmP

6,0 6,5 7,0 7,5 8,0 8,5

-0,04

0,00

0,04

Conduciv

ity

time (mi n)

residual mixture

weed kil ler

pur ified glyph osat e

Fig. 1. Comparison of the chromatograms obtained for the 50 mg L−1 solutions of

D. Caldarola et al. / Applied Su

The solutions were then centrifuged and the supernatant sep-rated. To calculate the residual metal concentrations, a 710-ESarian (Lexington, MA, USA) inductively coupled plasma–opticalmission spectrometer (ICP-OES) was used. The spectrometer isrovided with a radio frequency (RF) generator (40 MHz), a glassyclonic spray chamber, a concentric glass nebuliser, and a chargeoupled device (CCD) detector.

The HPLC instrument used for the PMA purity check was aaters model 2695 Separation Module, coupled with a Watersodel 430 conductivity detector (Waters, Milford, MA, USA) for

irect (non-suppressed) detection of PMA and isopropylammo-ium.

For the absorbance measurements, a Shimadzu UV/mini 1240V/Vis spectrophotometer (Shimadzu, Duisburg, Germany) wassed.

.3. Isolation of PMA from commercial herbicide formulation

One litre of liquid glyphosate based herbicide (GBH), producedy Multicrop Ltd. (Bayswater, Vic, Australia) under thetrade name,lyphosate 360 Weed Killer, was obtained. Note: As a commercialerbicide with concentration limit in drinking waters of 0.7 mg L−1

US EPA), some environmental issues should be considered dur-ng its use to limit any possible environmental impacts [28]. Purelyphosate or PMA is relatively expensive (around 200 US$ perram), so the isolation procedure of PMA from the GBH prepa-ation was considered to reduce the cost of the synthesis of theequired adsorbent (particularly if larger batches are to be producedn future work). Common GBH costs around 10–25 US$ for 1 litre,hat corresponds to less than 0.05 US$ per gram of purified PMA.

According to the GBH product label, this particular prepara-ion contained 360 g L−1 solution of the isopropylammonium salt ofMA, which after calculation corresponds to 266 g L−1 of free PMA.he difference between the solubility of isopropylammonium salt500 g L−1) in water, and the solubility of acidic/protonated formf PMA (0.05 g L−1) in ethanol was used for the isolation of pureubstrate [29]. This purification of PMA is required, as isopropy-amine can interfere with the immobilisation reaction at the silicaurface, through competitive reaction with the oxirane ring from-glycidoxypropyltrimethoxysilane (see further discussion on syn-hesis in Section 3.4). The purification procedure included 4-steps.he first step consisted of the precipitation obtained following theddition of two volumes of ethanol, which caused a dramatic dropn the solubility of the isopropylammonium salt of PMA. Followinghis, concentrated HCl was added to the resultant suspension until aull clear supernatant was observed, together with the formation of

thick white precipitate at the bottom of beaker. Acidification leadso further decrease in PMA solubility and promotes the formationf its protonated form from the original isopropylammonium salt.he third step involved washing the white precipitate with ethanoln a volume close to the volume of the precipitate, with vigoroustirring of the mixture to remove the rest of the isopropylammo-ium chloride. Finally, the ethanol was slowly evaporated from theeaker with the mixture thermostated at 50 ◦C in a water bath.

The washed precipitate was re-crystallised from water after dis-olution at 50 ◦C for 1 h. The obtained crystal powder was finallyried at room temperature for 1 day. This method allowed the iso-

ation of approximately 25 g of pure PMA from 100 mL of originalBH, which contained 26.6 g of PMA according to the manufacturer.

To investigate the actual isolation of PMA from the purifi-ation steps, the residual reactive solution was analysed by ionhromatography (IC) using a Dionex AS16 anion-exchange column

Thermo Scientific, formally Dionex, Sunnyvale, CA, USA) accordingo the method described by Guo et al. [30]. The obtained chro-

atogram was compared to those recorded for GBH and for theMA extract from GBH (Fig. 1). The peak of PMA is no longer present

residual reactive mixture, weed killer and purified glyphosate. Sample volume:10 �L. Eluent: 20 mM citric acid. Flow rate: 1 mL/min. Detection: conductimetric,without suppression.

within the chromatogram of the solution left after its isolation,demonstrating the completeness of the proposed isolation proce-dure.

2.4. Synthesis of HEPMAS

The mesoporous silica gel, Kieselgel 60 (Merck, Germany), withparticles size 37–63 �m, was used as a matrix for the synthesis ofthe adsorbent. At the first step, 20 g of silica was activated by theoverhead stirring in of 800 mL of boiling deionised water for 24 h,in order to increase the concentration of reactive silanol groups atthe surface. Then silica was filtered, washed and dried on a filter atroom temperature, ready for further modification.

The reaction mixture for the modification process consisted of21 g of PMA powder (isolated as described in Section 2.3), 13 mL of3-glycidoxypropyltrimethoxysilane and 34.5 mL of 50% (w/w) solu-tion of NaOH (both from Sigma–Aldrich, Sydney, NSW, Australia),added to keep the pH at 11. The mixture was stirred at 70 ◦C for1 h to allow the reaction of secondary amino groups within thePMA molecule with the epoxy groups from the silane. Followingthis, the pH of the reaction mixture was adjusted to a value ofaround 5 by titration with glacial acetic acid. The resulting mixturewas finally added to a beaker with 20 g of the previously activatedsilica gel and 160 mL of water and stirred at 80 ◦C for 24 h. The pre-pared adsorbent was filtered and washed by water and acetone anddried. A reaction scheme is shown in Fig. 2. The resultant bondedchemistry can now be correctly referred to as N-(2-hydroxyethyl)-N-(2-[phosphonomethyl)amino]acetic acid (HEPMA).

2.5. Investigation of the adsorbtion dependence on pH

The metal salt working solutions were prepared by adding 1 mLof 0.4 mM metal stock solution to 9 mL of the 0.2 M acetate buffersolution (pH 6.87, 5.95, 5.00, 3.94, 3.14) or HNO3 solutions with pH2.0 and 1.4. Calibration solutions were prepared by adding 1 mL of0.3, 0.2, 0.1, 0.05, 0.02 or 0.01 mM metal solution and 1 mL of PARto 8 mL of water.

The HEPMAS (0.103 g) was placed in a glass 5 mm I.D. columnwith a frit at the bottom (column bed height ∼6 mm), and prior

to experiments washed with 20 mL of 20 mM dipicolinic acid fol-lowed by 150 mL of 0.1 M HNO3 and deionised water. After this thecolumn was equilibrated with the 10–20 mL of the buffer, adjustedto the required pH. Then 10 mL of the metal-containing sample of

376 D. Caldarola et al. / Applied Surface Science 288 (2014) 373– 380

PMA

+Si OH + O

HC CH2

O

Si

CH3O OCH3

CH3O

O

N

P

O-

O-

OO-

Si O SiO

HO

O

NH

P

O-

O-

OO-

of HEP

trToarmpaP

3

3

(wpimpg[olF

Iaetsiw

3

aTsfip

lowing contents for nitrogen, carbon and hydrogen: 0.53% N;5.25% C; 1.12% H. These values correspond to 0.38 mmol g−1

bonded HEPMA groups (calculation based on nitrogen content)

O

HN

O

O O

O

NO

O O

O

OH

IDA HEIDA

O

HN P

O OO

OH

ON P

O OO

OH

O

OH

HEPM APMA

Fig. 2. Scheme

he same pH was passed through the prepared column at a flowate 1 mL/min. The effluent was collected in a plastic test-tube.hen column was washed twice with 10 mL of 1 M HNO3. Eachf these fractions were also collected in two separate containers;fter that the column was washed twice with water and finallye-equilibrated at the next desired pH. Prior to the photometriceasurements the pH of all the collected samples was adjusted to

H ≈ 9 and the absorbance of all samples was measured at 540 nmgainst the blank which was prepared by mixing 1 mL solution ofAR and 9 mL of water.

. Results and discussion

.1. Choice of the ligand for immobilisation

Glyphosate or 2-[(phosphonomethyl)amino]acetic acid (PMA)see the structure of PMA in Fig. 3) is a broad-spectrum herbicide,hich is widely used in gardening for killing weeds. Glyphosate isroduced in large scale, inexpensive and is available commercially

n the form of concentrated 360 g L−1 solutions of isopropylam-onium salt in water. PMA has three functionalities, namely,

hosphonic group with pKa1 = 0.78 and pKa3 = 5.69, carboxylicroup with pKa2 = 2.34, and secondary amino group with pKa4 = 10.231]. The presence of the latter group in the molecule allows the usef oxirane chemistry for the immobilisation of PMA on hydroxy-ated silica surfaces, according to the reaction scheme shown inig. 2.

PMA has a structure similar to the popular complexing reagentDA (Fig. 3), except containing a phosponic acid group in place of

carboxylic acid. Similarly to IDA, PMA is a relative strong acid,xhibiting a high complexing affinity towards most common transi-ion metal ions (Table 1). However, from the stability constants datahown in Table 1, it can be seen that PMA exhibits a higher selectiv-ty towards aluminium, iron (III), cadmium and zinc, as compared

ith IDA.

.2. Immobilisation of PMA on silica surface

Mesoporous silica gel Kieselgel 60, having a developed surfacerea of 411 m2 g−1 and a pore diameter of 8.2 nm (see data inable 2), was used for the immobilisation of PMA. Such a porous

tructure should provide grafting at concentration levels sufficientor the further characterisation of the prepared adsorbent. An SEMmage of the HEPMAS is shown in Fig. 4. Shape and size of thearticles look the same as for the bare silica (BS), meaning that

HEPMAS

MAS synthesis.

the PMA bonding process does not change the morphology ofadsorbent.

In this study comparative physico-chemical measurementswere performed for three materials namely bare silica (BS), acti-vated silica (or hydroxylated) via boiling in water for 24 h (AS)and HEPMAS. The elemental analysis of HEPMAS gave the fol-

Fig. 3. Structures of iminodiacetic acid (IDA), 2-[(phosphonomethyl)amino]aceticacid (PMA), and corresponding structures N-hydroxyethyliminodiacetic (HEIDA)and N-(2-hydroxyethyl)-N-(2-[phosphonomethyl)amino]acetic (HEPMA) inimmobilised-to-silica form.

D. Caldarola et al. / Applied Surface Science 288 (2014) 373– 380 377

Table 1Stability constants of complexes of metal ions with IDA and PMA at 25 ◦C [31,37].

Metal ion IDApKa1 = 1.77pKa2 = 2.98

PMApKa1 = 0.78pKa2 = 2.09pKa3 = 5.69

log ˇ1 log ˇ2 log ˇ1 log ˇ2

Na+ 0.36 – 1.4 –Mg2+ 2.98 4.85 3.28 5.47Ca2+ 2.60 5.90 3.29 5.87Mn2+ 4.72 7.82 5.50 7.80Cd2+ 5.71 10.10 7.29 10.91Fe2+ 5.80 10.10 6.87 11.18Co2+ 6.97 12.22 7.23 11.12Zn2+ 7.15 12.40 8.74 11.69Ni2+ 8.30 14.60 7.90 12.30Cu2+ 10.56 16.30 11.86 15.94Fe3+ 10.72 19.15 16.09 23.0La3+ 5.88 9.97 6.7 10.1VO2+ 9.0 – 10.69 15.9Al3+ 8.18 15.33 13.7 16.2

Table 2Parameters of porous structure evaluated for BS, AS and HEPMAS by using BET andDFT models.

BS AS HEPMAS

BET Area, m2 g−1 411 409 280C factor 116 118 65

a3ea

a(miio

100 20 0 300 40 0 500 600 70 0 800

80

90

100

Absorb

ance

-1

Pore size, nm 8.2 10.9 9.8Pore volume, cm3 g−1 0.86 0.98 0.67

nd 0.17 mmol g−1 of diol groups, from covalently bonded-glycidoxypropyltrimethoxysilane residues having unreactedpoxy-groups. The latter value was calculated from carbon contentfter subtraction of carbon included in bonded HEPMA groups.

Thermogravimetric analysis shows two mass losses for HEPMASt temperature ranges from 80 to 120 ◦C, and again at 250 to 400 ◦CFig. 5). The first loss of 5–7 mass % is observed both for bare and

odified silica, and can be attributed to the evaporation of phys-cally adsorbed water. The second mass loss of approximately 8%s due to thermal decomposition of bonded HEPMA groups. Thebtained value is slightly higher than the total mass of carbon,

Fig. 4. SEM image of HEPMAS.

Wavenumbe rs (cm )

Fig. 5. Thermo gravimetric analysis plot of AS (dotted line) and HEPMAS (solid line).

nitrogen and hydrogen (∼7 mass %), as detected by elemental anal-ysis of HEPMAS. Obviously, the higher value of mass loss in TGAincludes the weight of chemisorbed water removed from silicasurface at higher temperatures.

Other techniques including acid–base potentiometric titration,FTIR, isotherms of nitrogen adsorption at 77 K, and XPS were alsoused for physico-chemical characterisation of HEPMAS and its com-parison with properties of bare silica.

3.3. Characterisation of prepared adsorbent

3.3.1. Acid-base potentiometric titrationThe acid-base properties of ionogenic molecules can be changed

after immobilisation on silica. Herein, conversion of secondaryamino groups within PMA, to tertiary amino groups within itsbonded-to-silica state (see HEPMA structure in Fig. 3) are one suchexample. Additionally, the chemical heterogeniety of the surfaceof porous silica, and possible hydrogen bonding and electrostaticinteractions between bonded ligands themselves, and with polarsilanols groups, can also effect both acidity and basicity of func-tional groups [32]. For this reason titration in the presence of strongelectrolytes is recommended to suppress secondary equilibria inthe bonded layer of adsorbent and to increase the accuracy of theresultant titration data.

Acid-base titration of BS and HEPMAS in the presence of0.5–1.0 M NaCl was performed, for the determination of both con-centration and the pKa3 value of the bonded PMA residues. Theresults are shown in Fig. 6. The calculated concentration of bondedPMA was found to be 0.23 mmol g−1 and the pKa3 value for HEP-MAS at approximately 5.1. The autotitrator measures pKa valuesas the pH at the half of the volume of titrant used to achieve theEnd Point (EP), while differention of the titration curves provides amore clearly defined EP.

3.3.2. FTIR spectroscopyFig. 7 shows FTIR spectra of both AS and HEPMAS. It was not pos-

sible to observe the presence of the phosphonic group itself due tothe spectral saturation arising from the silanols under 1300 cm−1,but the presence of several clear peaks provided evidence of thepresence of PMA and 3-glycidoxypropyltrimethoxysilane. Firstly,aliphatic CH2 groups belonging to both compounds have 2 observ-

able bands: stretching is clearly visible at 2873 cm−1 (symmetric)and 2935 cm−1 (asymmetric), with symmetric bending giving asmall shoulder at 1435 cm−1. Secondly, the carbonyl group fromthe PMA residue can be detected as three bands, at 1732 cm−1

378 D. Caldarola et al. / Applied Surface Science 288 (2014) 373– 380

pH

2.5

3.0

3.5

4.0

4.5

5.0

5.5

6.0

V, ml

0 2 4 6 8

ER

C (

eq

uiv

ale

nce

po

int re

co

gn

itio

n c

rite

rio

n)

0

10

20

30

40

50

60

70

1

2

3

4

5

6

Fp6

(1se

3

tItcf

w

396398400402404

50

100

150

200

250

300

Binding Energy (eV)

co

un

ts

ig. 6. Titration curves of bare silica in presence of 1 M NaCl (3) and HEPMAS inresence of 0.5 M NaCl (2) and 1 M NaCl (1). The differential form of curves is marked, 5 and 4, respectively.

C O of free COOH), 1434 cm−1 (COO− symmetric stretching) and605 cm−1 (COO− asymmetric stretching). In general, the FTIRpectrum observed corresponded well to the data reported in lit-rature for PMA (reported as glyphosate) [33].

.3.3. X-ray photoelectron spectroscopyX-ray photoelectron spectroscopy (XPS) was performed to prove

he presence of key elements, such as nitrogen, within HEPMAS.n fact, the nitrogen can be detected in the adsorbent solely dueo the immobilised PMA, and the high resolution of XPS analysisan provide additional information on the chemical form of the

unctional groups.

A high-resolution XPS spectrum showing the N 1s line is seenithin Fig. 8. The deconvolution of the high-resolution peak gave

4000 350 0 300 0 250 0 200 0 150 0

0.5

1.0

1.5

2.0

2.5

Absorb

ance

Wavenumbers (cm-1)

1600 1500 1400

0.5

1.0

Fig. 7. FTIR spectra of AS (dotted line) and HEPMAS (solid line).

Fig. 8. N 1s line in XPS spectrum of HEPMAS. High resolution signal (solid line) anddeconvoluted components (dotted lines).

two main components at 399.7 and 401.9 eV. These energies cor-respond to the presence of a deprotonated amine group and aprotonated amine group, respectively, [34]. The presence of thecharged amine provides evidence for the ionic form of the ligand atthe interface, which would increase the complexation capabilitiesand retention characteristics of the modified sorbent.

The XPS method was also used to evaluate the concentrationof bonded HEPMA, which resulted in a value of 0.468 mmol g−1.This value is higher than that calculated from the elemental anal-ysis data, as the XPS response is representative of the compositionof only the very surface layer, of only a few hundred nanometersthickness or depth. This means that the functionalisation of silicaoccurred with greater efficiency at outer surface of HEPMA-silica,but not as completely inside of the silica’s pores.

3.4. Porous structure and surface area

The nitrogen adsorption isotherms obtained at low tempera-tures (Fig. 9) have the shape typical for mesoporous materials,

0,15,00,0

0

200

400

600

Volu

me

P/P0

Fig. 9. N2 adsorption isotherms of as BS (empty circles), AS (filled squares) andHEPMAS (crosses).

D. Caldarola et al. / Applied Surface Science 288 (2014) 373– 380 379

Table 3Concentration limits and potential effects of the studied metals (US EPA).

Metal MLC(mg L−1)

Potential health effects from long-term exposureabove the MCL

Pb 0 Infants and children: Delays in physical or mentaldevelopment; children could show slight deficitsin attention span and learning abilitiesAdults: Kidney problems; high blood pressure

Be 0.004 Intestinal lesions

era

ntog2spfh

wswsTh

3

dctol

t4oafou

ftpoi

TD

t

Cd 0.005 Kidney damageZn 5 Secondary contaminant, cause metallic taste and

turbidity

xhibiting the hysteresis loop around 0.7–0.8 P/P0 values. The cor-esponding calculated parameters characterising porous structurere presented in Table 2.

A decrease of both surface area and of mean pore diameter wasoted for HEPMAS, as compared with AS, indicating the modifica-ion of the surface within pores. In fact, the steric hindrance effectf the grafted chain makes the surface less available for the nitro-en adsorption. Higher values of BET constant C were noted for the

unmodified silicas, confirming the same information about theurface. HEPMAS has indeed a lower value, indicative of a lowerolarisation of the surface due to the presence of the alkyl chainrom silane molecules, which convert the hydrophilic silanols toydrophobic groups.

The same consideration can be applied to the pore volume value,hich is lower for the HEPMAS. Interestingly, the pore volume mea-

ured for AS was higher than for BS, meaning boiling of silica inater for 24 h may not only increase the concentration of reactive

ilanols at the surface, but also can modify the porous framework.his hypothesis is confirmed by the changes in pore size, which wasigher for AS.

.5. Metal removal tests

The complexing properties of the prepared adsorbent wereemonstrated through the adsorption of zinc, lead, beryllium andadmium, which are of interest due to their well known toxicityo a range of living organisms. Table 3 listed the potential effectsf these metals at long exposure and their maximum contaminantevels (MLCs) according to US EPA regulations.

The adsorption tests were performed using 20 �mol L−1 solu-ions for each metal or 0.18 mg L−1 for Be(II), 1.3 mg L−1 for Zn(II),.14 mg L−1 for Pb(II) and 2.14 mg L−1 for Cd(II). Prior to use, 100 mgf HEPMAS was washed with 200 mL of 20 mM HNO3 to removeny pre-adsorbed metals, followed by water and then dried at 40 ◦Cor one day. Table 4 shows the results obtained for the adsorptionf the above metals on the HEPMAS, as compared with that fornmodified BS.

The Table shows no adsorption of selected metals on BS, exceptor the beryllium. According to the results of a previous study [35],

he adsorption of beryllium from aqueous solutions on silica takeslace at pH > 6.0. However, there is a possibility that some amountf beryllium acetate was precipitated from the acetate buffer usedn these experiments. HEPMAS shows excellent removal results

able 4ata on adsorption of metals from aqueous solutions using HEPMAS.

Metals BS AS HEPMAS HEPMAS, R

Be(II) 24.7* 23.3 100 99.0Zn(II) 1.1 0.0 99.0 94.8Pb(II) 0.0 0.0 99.9 95.4Cd(II) 0.0 0.0 98.1 91.3

* – was calculated as 100 × (C0 − C1)/C0, where C0 and C1 are equal to the concen-ration of metal ion in solution before addition of HEPMAS and after, respectively.

Fig. 10. pH dependences for metal adsorption on HEPMAS.

for all of the 4 metals, with an order of affinity that appears tobe, Be(II) > Pb(II) > Zn(II) > Cd(II). The results on adsorption of Zn(II)and Cd(II) are in agreement with the stability constants shown inTable 1. The second round of adsorption-desorption (see repeatdata for HEPMAS, R in Table 4) shows slightly lower outcomes for allthe metals, although with recovery still close to 100%. The affinityorder was the same, confirming the presence of PMA in the sorbentand adsorption by a chelating mechanism.

The prepared adsorbent has a strong affinity towards to all stud-ied metals and can be used for their removal from contaminatedwaters. The calculated values of distribution coefficients D for theadsorption of toxic metal ions on HEPMAS (see data in Table 4)are 5.0 × 104, 4.9 × 105 and 2.6 × 104 for Zn(II), Pb(II) and Cd(II),respectively.

3.6. pH-dependence for metal adsorption

The adsorption of Mn(II), Co(II), Cd(II), Ni(II), Zn(II) and Cu(II) onHEPMAS was studied under dynamic conditions by passing dilutedsolutions of metal through glass column of internal diameter 5 mm.The height of the bed formed by 0.103 g of HEPMAS was 6 mm andflow rate was about 1.0 mL/min in all experiments.

The adsorption of metals increases with pH increase in all cases(Fig. 10). The complete adsorption (85–99%) occurred at pH of solu-tions above 3. According to obtained data, HEPMAS interacts withstudied metal ions within a narrow pH range indicating possibil-ity of their separation under isocratic elution with more efficientchromatographic column. The pH values of 50% adsorption ofmetals show the selectivity of metal binding with chelating groupsfrom HEPMAS. In general, the observed order Zn(II), Cd(II) > Ni(II),Co(II) > Mn(II) > Cu(II) correlates with the order of increasing sta-bility constants of metal complexes with PMA (Table 1). HEPMASshows a stronger binding capability for Mn(II), which may beattributed to difference in the structures of PMA and HEPMA. Forthe latter ligand the additional coordination of metal cation by

OH- group is possible. A similar change in the selectivity has beenobserved for 2-hydroethyl-iminodiacetic acid (HEIDA) comparedwith iminodiacetic acid (IDA) [36]. Unfortunately, there are noreported stability constants for HEPMA. The desorption of all metals

3 urface

fi

aswt1cm

4

muafts

lem

A

el0R

R

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

[

80 D. Caldarola et al. / Applied S

rom the column can be achieved by flushing with 20 mM dipicol-nic acid or with 1 M HNO3.

HEPMAS adsorbs quantitatively Mn(II), Co(II), Cd(II), Zn(II), Ni(II)nd Cu(II) from aqueous solutions at pH > 3.0, which indicates atronger binding capability of this chelating substrate as comparedith IDA-silica [9]. At the same time the pH values corresponding

o 50% adsorption of these metals are in narrow pH range from.0 to 2.5. This is an important property of HEPMAS to be used ashelating stationary phase for the chromatographic separation ofetals.

. Conclusions

A procedure for the extraction of PMA (glyphosate) from a com-ercial weed killer has been developed. The purified PMA was

sed for the surface modification of silica particles. The prepareddsorbent was characterised by several methods showing success-ul immobilisation of PMA. The prepared HEPMAS was tested forhe extraction of four metals of environmental concern and demon-trated a high affinity and adsorption capacity.

Further work will be focused on in situ modification of mono-ithic mesoporous silica column with PMA glyphosate and on thevaluation of its chromatographic performance in high perfor-ance chelation ion chromatography.

cknowledgements

Central Science Laboratory of the University of Tasmania for thelemental analysis of the prepared adsorbent. The authors wouldike to acknowledge Science Foundation Ireland (Grant Number8/SRC/B1412) and Discovery grant DP110102046 of Australianesearch Council.

eferences

[1] G.V. Lisichkin, A.Y. Fadeev, A.A. Serdan, P.N. Nesterenko, D.B. Furman, Chemistryof Surface Grafted Compounds, Fizmatlit, Moscow, 2003, pp. 592.

[2] D.E. Leyden, G.H. Luttrell, Preconcentration of trace metals using chelatinggroups immobilized via silylation, Anal. Chem. 47 (1975) 1612–1617.

[3] D.E. Leyden, G.H. Luttrell, T.A. Patterson, Silica gel with immobilized chelatinggroups as an analytical sampling tool, Anal. Lett. 8 (1975) 51–56.

[4] A. Badiei, H. Goldooz, G.M. Ziarani, A novel method for preparation of 8-hydroxyquinoline functionalized mesoporous silica: aluminum complexes andphotoluminescence studies, Appl. Surf. Sci. 257 (2011) 4912–4918.

[5] A. Arencibia, J. Aguado, J.M. Arsuaga, Regeneration of thiol-functionalizedmesostructured silica adsorbents of mercury, Appl. Surf. Sci. 256 (2010)5453–5457.

[6] B.J. Gao, F.Q. An, K.K. Liu, Studies on chelating adsorption properties of novelcomposite material polyethyleneimine/silica gel for heavy-metal ions, Appl.Surf. Sci. 253 (2006) 1946–1952.

[7] R.K. Dey, T. Patnaik, V.K. Singh, S.K. Swain, C. Airoldi, Attachment of linearpoly(amido amine) to silica surface and evaluation of metal-binding behavior,Appl. Surf. Sci. 255 (2009) 8176–8182.

[8] G.V. Lisichkin, G.V. Kudryavtsev, P.N. Nesterenko, Chemically modified silicasand their use in inorganic analysis, J. Anal. Chem. USSR 38 (1983) 1288–1307.

[9] M. Zougagh, J.M.C. Pavon, A.G. de Torres, Chelating sorbents based on silica geland their application in atomic spectrometry, Anal. Bioanal. Chem. 381 (2005)1103–1113.

10] J.F. Biernat, P. Konieczka, B.J. Tarbet, J.S. Bradshaw, R.M. Izatt, Complexing andchelating agents immobilized on silica gel and related materials and their appli-cation for sorption of inorganic species, Sep. Purif. Methods 23 (1994) 77–348.

11] P.K. Jal, S. Patel, B. Mishra, Chemical modification of silica surface by immobi-lization of functional groups for extractive concentration of metal ions, Talanta62 (2004) 1005–1028.

12] P.N. Nesterenko, P. Jones, Recent developments in the high-performance chela-tion ion chromatography of trace metals, J. Sep. Sci. 30 (2007) 1773–1793.

[

Science 288 (2014) 373– 380

13] P.N. Nesterenko, P. Jones, B. Paull, Chelating Stationary Phases, Royal Society ofChemistry, Cambridge, 2011, pp. 35–115.

14] O.J. Challenger, S.J. Hill, P. Jones, N.W. Barnett, Application of chelating exchangeion chromatography to the determination of trace metals in high ionic strengthmedia, Anal. Proc. 29 (1992) 91–93.

15] P. Nesterenko, P. Jones, The comparative investigation of several stationaryphases containing iminodiacetic functional groups for the high-performancechelating exchange chromatography, J. Liq. Chromatogr. Relat. Technol. 19(1996) 1033–1045.

16] E. Sugrue, P. Nesterenko, B. Paull, Ion exchange properties of monolithic andparticle type iminodiacetic acid modified silica, J. Sep. Sci. 27 (2004) 921–930.

17] A.I. Elefterov, M.G. Kolpachnikova, P.N. Nesterenko, O.A. Shpigun, Ion-exchangeproperties of glutamic acid-bonded silica, J. Chromatogr. A 769 (1997) 179–188.

18] P.A. Kebets, K.A. Kuz’mina, P.N. Nesterenko, The heats of adsorption of metalcations on silica gel with covalently fixed polyaspartic acid, Russ. J. Phys. Chem.76 (2002) 1481–1484.

19] Y. Inoue, H. Kumagai, Y. Shimomura, T. Yokoyama, T.M. Suzuki, Ion-chromatographic separation of rare-earth elements using a nitrilotriacetate-type chelating resin as the stationary phase, Anal. Chem. 68 (1996) 1517–1520.

20] W. Bashir, E. Tyrrell, O. Feeney, B. Paull, Retention of alkali, alkaline earth andtransition metals on an itaconic acid cation-exchange column – eluent pH, ionicstrength and temperature effects upon selectivity, J. Chromatogr. A 964 (2002)113–122.

21] N.R. Sumskaya, Y.V. Kholin, V.N. Zaitsev, Microcolumn frontal high-performance liquid chromatography of copper(II) chloride on amin-odiphosphonic acid-modified silica, Zhurnal Fizicheskoi Khimii 71 (1997)905–910.

22] R. Garcia-Valls, A. Hrdlicka, J. Perutka, J. Havel, N.V. Deorkar, L.L. Tavlarides,M. Munoz, M. Valiente, Separation of rare earth elements by high performanceliquid chromatography using a covalent modified silica gel column, Anal. Chim.Acta 439 (2001) 247–253.

23] M.J. Shaw, P.N. Nesterenko, G.W. Dicinoski, P.R. Haddad, Selectivity behaviourof a bonded phosphonate-carboxylate polymeric ion exchanger for transitionand heavy metal cations with varying eluent ionic strength and pH, J. Chro-matogr. A 997 (2003) 3–11.

24] M.J. Shaw, P.N. Nesterenko, G.W. Dicinoski, P.R. Haddad, Retention character-istics of lanthanide ions on a mixed phosphonic acid – carboxylic acid cationexchanger, Aust. J. Chem. 56 (2003) 201–206.

25] P.N. Nesterenko, O.S. Zhukova, O.A. Shpigun, P. Jones, Synthesis and ion-exchange properties of silica chemically modified with aminophosphonic acid,J. Chromatogr. A 813 (1998) 47–53.

26] P.N. Nesterenko, M.J. Shaw, S.J. Hill, P. Jones, Aminophosphonate functionalisedsilica: a versatile chromatographic stationary phase for high-performancechelation ion-chromatography, Microchem. J. 62 (1999) 58–69.

27] M.J. Shaw, S.J. Hill, P. Jones, P.N. Nesterenko, Determination of beryllium in astream sediment by high-performance chelation ion chromatography, J. Chro-matogr. A 876 (2000) 127–133.

28] A. Paganelli, V. Gnazzo, H. Acosta, S.L. Lopez, A.E. Carrasco, Glyphosate-basedherbicides produce teratogenic effects on vertebrates by impairing retinoic acidsignaling, Chem. Res. Toxicol. 23 (2010) 1586–1595.

29] W.J. Fu, Q.L. Guo, J.D. Wang, Solubility of glyphosate in ethanol + water, 1-propanol + water, and 2-propanol + water from (293 to 333) K, J. Chem. Eng.Data 55 (2010) 3915–3917.

30] Z.X. Guo, Q.T. Cai, Z.G. Yang, Determination of glyphosate and phosphate inwater by ion chromatography – inductively coupled plasma mass spectrometrydetection, J. Chromatogr. A 1100 (2005) 160–167.

31] A.E. Martell, R.M. Smith, NIST Critically Selected Stability Constants of MetalComplexes, NIST standard reference database 46.Version 8.0, 2004.

32] Y. Kholin, V. Zaitsev, Quantitative physicochemical analysis of equilib-ria on chemically modified silica surfaces, Pure Appl. Chem. 80 (2008)1561–1592.

33] A. Piccolo, G. Celano, Modification of infrared-spectra of the herbicideglyphosate induced by pH variation, J. Environ. Sci. Health B 28 (1993)447–457.

34] J. Sheals, S. Sjoberg, P. Persson, Adsorption of glyphosate on goethite: molec-ular characterization of surface complexes, Environ. Sci. Technol. 36 (2002)3090–3095.

35] I. Nukatsuka, K. Sakai, R. Kudo, K. Ohzeki, Studies on the adsorption of traceamounts of beryllium(II) on the surface of silica fibres from an aqueous phasefor the development of a novel enrichment technique for electrothermal atomicabsorption spectrometry, Analyst 120 (1995) 2819–2822.

36] E. Nesterenko, P. Nesterenko, B. Paull, M. Melendez, J. Corredor, Fast direct

determination of strontium in seawater using high-performance chelation ionchromatography, Microchem. J. 111 (2013) 8–15.

37] K. Popov, H. Ronkkomaki, L.H.J. Lajunen, Critical evaluation of stabilityconstants of phosphonic acids (IUPAC technical report), Pure Appl. Chem. 73(2001) 1641–1677.