Zr

RDa

b

c

d

a

ARRA

KZMWA

1

sftceStcfabaattbs

rT

1d

Chemical Engineering Journal 178 (2011) 210– 216

Contents lists available at SciVerse ScienceDirect

Chemical Engineering Journal

jo u r n al hom epage: www.elsev ier .com/ locate /ce j

eolites in a permeable reactive barrier (PRB): One-year of field experience in aefinery groundwater. Part 2: Zeolite characterization

odolfo Vignolaa, Roberto Bagatina,∗, Alessandra De Folly D’Aurisb, Elisabetta Previde Massarac,anila Ghisletti c, Roberto Millini c, Raffaello Sistod

Department of Environmental Technologies, Eni S.p.A., Istituto eni Donegani, Via G. Fauser 4, I-28100 Novara, ItalyDepartment of Chemistry, La Sapienza Università di Roma, Piazzale A. Moro 5, I-00185 Roma, ItalyEni S.p.A., Refining & Marketing Division, San Donato Milanese Research Center, Physical Chemistry Dept. Via F. Maritano 26, I-20097 S. Donato Milanese, Milano, ItalyStudies & Researches Department, Renewable Energy & Environment Technologies, eni S.p.A. Piazzale E. Mattei 1, I-00144 Roma, Italy

r t i c l e i n f o

rticle history:eceived 4 August 2011eceived in revised form 14 October 2011ccepted 18 October 2011

a b s t r a c t

Zeolite filters composed by ZSM-5 and Mordenite extrudated with 20 wt% of clay and alumina, respec-tively, were used in a long term tests of decontamination of groundwater containing a variety ofhydrocarbons and significant concentrations of dissolved inorganic salts. Extensive physical–chemicalcharacterization of the samples taken after 6 and 12 months of tests confirmed that zeolites did not

eywords:SM-5ordeniteater treatment

dsorption properties

undergo any modification, the original microporous volume being restored by thermal treatment. Depo-sition of gypsum and iron sulphide occurred without altering the sorption capacity of the adsorbents.The only macroscopic effect of the long permanence in groundwater was identified in the modification ofthe textural properties of the extrudates, with the increase of the mesopore volume, probably associatedwith the reconstruction of the binders due to the loss of Al in the case of ZSM-5 and to its deposition in

Mordenite samples.

. Introduction

The zeolite advantages concerning with organic contaminantelectivity, rapid kinetics and no salt and humic substances inter-erence (also at concentration of tens grams per liter) and mainly,he effective capacity in the removal of almost all the organicompounds present in groundwater of petrochemical and refin-ry sites, were demonstrated in previous laboratory works [1–3].pecific zeolites were identified for efficient removal of methyl-ertbutylether (MTBE), 1,2-dichloroethane (1,2-DCA) and vinylhloride (VC). In particular, ZSM-5 zeolites turned out suitableor aliphatic, halo-aliphatic and mono-aromatic molecules, suchs benzene, toluene, ethylbenzene and xylenes (BTEX) and halo-enzene derivatives, while Mordenite was more appropriate forromatic molecules with two or more aromatic rings, halo- andlkyl-substituted ones, and ethers such as MTBE. Furthermore,hese microporous adsorbents have been proposed for the decon-

amination of groundwater with the use of permeable reactivearriers (PRB), as described in the first part of this study [4]. Inuch case, the system of one or two zeolites placed in series forms

∗ Corresponding author at: Eni S.p.A., Istituto Eni Donegani, Department of Envi-onmental Technologies, Via G. Fauser, 4, I-28100 Novara, Italy.el.: +39 0321447310; fax: +39 0321447506.

E-mail address: roberto.bagatin@eni.com (R. Bagatin).

385-8947/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.cej.2011.10.052

© 2011 Elsevier B.V. All rights reserved.

the active medium of the barrier, situated in situ perpendicular tothe groundwater flow. In fact, the decontamination of the pollutedstream passing through it occurs by immobilizing the contaminat-ing species. Generally, in literature and practical PRB application,the selection of the reactive medium is based on the targetedcontaminants and on the hydro-geological assessment of the site.Zero-valent iron (ZVI) is the most common medium used in PRBs,followed by granulated activated carbon (GAC). Examples of otherreactive media include micro-organisms, natural zeolites, peat,phosphate, limestone, and amorphous ferric oxide. The treatmentprocesses that occur within the trench consist into degradation,adsorption, and/or precipitation of the contaminant. ZVI, used inPRB since the nineties, appears to be ineffective on irreduciblecompounds such as 1,2-DCA and chlorobenzenes as well as hydro-carbons [5]. Further efforts have been placed on finding innovativereactive systems either following the reductive approach by intro-ducing other metals (e.g. Pd, Ni) or using more effective adsorbents(e.g. ion-exchange polymers of AmberliteTM family). On the otherhand, the use of the GAC was encouraged by economics and by thepossibility of removing a wide range of contaminants, such as chlo-rinated solvents and hydrocarbons. However, GAC presents severaldisadvantages due to the physical–chemical characteristics of its

surface: pore plugging, interactions with humic substances andinorganic ions, adverse effect of pH on the adsorption of organ-ics. Additionally, GAC has been shown to be slightly effective intreatment of water containing very soluble compounds, such as

R. Vignola et al. / Chemical Engineering Journal 178 (2011) 210– 216 211

Table 1Main characteristics of zeolites.

Characteristics ZSM-5 Mordenite

SiO2/Al2O3 (mol/mol) 2100 230Na2O (wt%) <0.01 <0.05

oV

vciwpp

zwot(dfac

2

2

t

we

fi

(ib

t

3

alws(

tSf

pdsfla

5 10 15 20 25 30 35 40 45 502-The ta [°]

(a)

(b)

(c)

ions, samples of zeolites recovered from the top and the bottom ofthe filters were subjected to an extensive physical–chemical char-acterization.

5 10 15 20 25 30 35 40 45 50

(a)

(b)

(c)

Binder (wt%) Clay (20) Alumina (20)Pellets size (diameter × length (mm)) 1.5 × 3 1.5 × 3

xygenated organics, or low molecular weight compounds, such asC [6–8].

This paper discusses the use of hydrophobic zeolites as an inno-ative adsorbent material suitable for the in situ abatement ofontaminants in groundwater through PRB and mainly evidencests use in an industrial site situated on the coast. The field treatment,

ith the duration of one-year, was divided into two operationalhases different for water flow, the first one to verify the materialerformance and the second one to verify the system limits.

The main aim of present work is to show that channels ofeolite adsorbents are not affected by fouling occurred duringorking time. In this respect, physical–chemical characterization

f aged materials was performed: X-ray powder diffraction (XRD)o determine any structural changes, scanning electron microscopySEM) coupled with energy dispersive X-ray spectroscopy (EDS) toetect the presence of compounds adhering to the external sur-ace, thermo-gravimetric (TG) and gas-chromatographic (GC–MS)nalyses to verify the immobilization of contaminants within thehannels.

. Experimental

.1. Samples

Zeolites were supplied by Tosoh Corp. as extrudated cylinders;he characteristics are reported in Table 1.

The field tests were performed as previously described [4]. Twoorking phases, six months long each, were conducted with differ-

nt set of filters:Phase A: 120 kg ZSM-5 filter (Z) followed by 150 kg Mordenite

lter (M) in a Z + M sequence.Phase B: new filters of 120 kg ZSM-5 (Z′) and 150 kg Mordenite

M′) were added so each previous filter was followed by a new onen a sequence Z + Z′ + M + M′. Groundwater went through filters fromottom to top (up flow).

At the end of the test, samples were taken from the top and fromhe bottom of each filter and opportunely treated for the analyses.

. Methods

Structural analysis was carried out on powdered sample by XRD; Philips X’PERTTM vertical diffractometer equipped with a scintil-ator detector and a secondary mono-chromator was used. Data

ere collected stepwise in the 5 ≤ 2� ≤ 50◦ angular region, with atep size of 0.02◦ 2� and 7 s/step accumulation time, using the Cu K�� = 1.54178 A) radiation.

Textual analyses were performed through adsorp-ion/desorption cycles of N2 at 77 K with a ThermoQuestorptomatic 1990TM porosimeter. Samples were analyzed inorm of pellets previously treated at 383 K for 2 h under vacuum.

Morphological and compositional characterization of the sam-les was carried out by means of SEM) equipped with an EDSetector. A Jeol JSM 7600FTM with “in lens” thermal field emis-

ion gun emitter (FEG) was used. SEM had an accelerating voltagerom 0.1 kV to 30 kV, a SEI resolution of 1.0 nm at 15 kV, upper andower (semi-in lens) detectors for secondary electrons, retractablend low angle backscattered electron detectors (RBE and LABE).

Fig. 1. XRD patterns of ZSM-5 samples: (a) fresh, (b) filter Z bottom and (c) filter Ztop.

The EDS spectrometer, from Oxford Instruments, has a resolutionof 130 eV and a crystal area detector of 20 mm.

The samples, previously treated at 343 K for 4 h under vacuumto eliminate the volatile components, were analyzed in differentforms:

– pellets: the samples were milled and the powder pelletized inan infra-red spectroscopy press to obtain pellets suitable for thedetermination of the average chemical composition. For eachpellet, 5 different areas were analyzed by EDS at relatively lowmagnifications (500×) and the results averaged;

– as made: the extrudates were analyzed without any manipulationin order to inspect the surface morphology and composition;

– embedded: the extrudates were embedded, either in vertical andhorizontal positions, in epoxy resin then polished with abrasivepapers of decreasing size; in this way, it was possible to obtaintransversal and longitudinal sections which allowed the inspec-tion of the surface and the internal morphology as well as thedistribution of the main elements through the acquisition of X-ray maps and profiles.

4. Results and discussion

In order to verify the stability of the zeolite adsorbents left forlong time (from 6 to 12 months) in groundwater contaminated byorganic species and with a significant amount of dissolved inorganic

2-The ta [°]

Fig. 2. XRD patterns of Mordenite samples: (a) fresh, (b) filter Z′ bottom and (c) filterZ′ top.

212 R. Vignola et al. / Chemical Engineering Journal 178 (2011) 210– 216

Table 2Bulk chemical composition of ZSM-5 samples from SEM/EDS analysis (data in wt%normalized to 100%).

Fresha Z Z′

Top Bottom Top Bottom

SiO2 94.44 92.13 92.72 92.23 93.79Al2O3 4.03 3.69 3.89 3.82 3.39SARb 39.79 42.36 40.40 40.93 47.02NaCl (from Cl content) 0 0.31 0.34 0.21 0.23Na2O (from residual Na) 0.47 0.29 0.29 0.38 0.32K2O 0.09 0.09 0.10 0.12 0.10MgO 0.56 0.56 0.54 0.53 0.49FeS2 (from Fe content) 0 1.75 1.53 1.74 1.37CaSO4·2H2O (from residual S) 0 1.18 0.51 0.97 0.14CaO (from residual Ca) 0.29 0 0.08 0 0.17Fe2O3

c 0.12 0 0 0 0˙ 100 100 100 100 100

a

oozbmamlr

asp

4

ft

oosZm

TBw

Cl and S not detected.b SiO2/Al2O3 molar ratio.c Considered only when S is not present.

The possibility to perform a field test has represented a uniquepportunity to collect data otherwise difficult to obtain at a lab-ratory level thus allowing the determination of the capability ofeolites in the removal of the organic contaminants, keeping themelow the threshold limits. Furthermore, through the structural,orphological and compositional characterization of the adsorbent

t the end of the field test, it has been possible to determine theodifications, if any, induced by the long permanence of the zeo-

ites in groundwater and to verify if the adsorption capacity can beecovered by ex situ treatments.

All these data have been useful for estimating the performancesnd the life of the adsorbent system (i.e. the parameters neces-ary for evaluating the economics and the applicability of the entirerocess).

.1. Structural characterization

XRD analysis was performed on samples taken from the top androm the bottom of each filter, milled and dried at room tempera-ure.

Fig. 1 compares the XRD pattern of the fresh ZSM-5 with thosef the samples taken from the bottom and from the top of filter Z

perated for the entire test (12 months). The main difference con-ists in the transition from the orthorhombic symmetry of the freshSM-5 sample to the monoclinic one characterizing all the usedaterials. Being a high-silica form (SAR = 2100, Table 1), essentially

able 3ulk chemical composition of Mordenite samples from SEM/EDS analysis (data int% normalized to 100%).

Fresha M M′

Top Bottom Top Bottom

SiO2 79.67 74.94 73.69 75.70 76.24Al2O3 20.28 21.77 22.06 20.91 21.26SARb 6.67 5.84 5.67 6.14 6.09NaCl (from Cl content) 0.03 0.38 0.37 0.41 0.40Na2O (from residual Na) 0.02 0.52 0.54 0.44 0.46K2O 0 0.25 0.29 0.23 0.19MgO 0 0.24 0.22 0.21 0.15FeS2 (from Fe content) 0 0.50 0.75 0.76 0.39CaSO4·2H2O (from residual S) 0 1.40 2.08 1.34 0.91CaO (from residual Ca) 0 0 0 0 0Fe2O3

c 0 0 0 0 0˙ 100 100 100 100 100

a K, Mg, Ca, Fe and S not detected.b SiO2/Al2O3 molar ratio.c Considered only when S is not present.

Fig. 3. SEM micrographs collected in BE mode of the surface of ZSM-5 extrudatestaken from the top of filter Z: (a) extended region showing the presence of gyp-

sum and iron sulphide deposits (bright areas) showed in details in (b) and (c),respectively.

a Silicalite-1 phase, the monoclinic symmetry would be expectedalso for the fresh sample since the transition from the meta-stablehigh-temperature orthorhombic form occurs at 350–380 K [9]. Inour experience, it is not uncommon to observe that the orthorhom-bic symmetry of MFI-type zeolites with a high SAR is maintainedeven at room temperature as a consequence, for instance, of arapid cooling or quenching from high temperature treatments; theorthorhombic to monoclinic transition is achieved when the sam-ples are exchanged with an ammonium acetate solution and finallycalcined to be transformed in acid form.

Anyway, it has to be underlined that the orthorhombic to mon-oclinic transition has a minor importance when considering theapplication of ZSM-5 in PRB since the two crystalline forms are

R. Vignola et al. / Chemical Engineering Journal 178 (2011) 210– 216 213

terior

oi

pCo

aatC

Fa

Fig. 4. SEM micrographs collected in BE mode of the in

nly slightly different and their adsorption properties do not varyn a significant manner.

The other difference between the fresh and the used sam-les concerns the presence, in the latter, of trace amounts ofaSO4·2H2O (gypsum), detected in all the samples analyzed. Nother crystalline phases were detected in all the used samples.

As far as Mordenite is concerned, XRD analysis did not evidenceny significant difference among the samples (Fig. 2). The patternsre dominated by the reflections of MOR-type zeolite and only in

he sample taken from the bottom of filter M, trace amounts ofaSO4·2H2O (gypsum) were detected.

ig. 5. SEM/EDS line profile showing the distribution of Fe ( ) and S ( )long the path indicated by the white line in the micrograph of a ZSM-5 extrudate.

of the ZSM-5 extrudates taken from the top of filter Z.

4.2. Morphology and composition

In order to verify the modifications induced by the longpermanence of the adsorbents in groundwater, an extensive mor-phological and compositional characterization was performed onall the samples taken from the bottom and from the top of the fourfilters.

Attention was firstly put on the overall elemental compositionof the sample, with the focus on the inorganic elements retained bythe adsorbents. Being both the phases at high SAR, we did not expectany significant contribution of ion-exchange processes; instead,as evidenced by XRD analysis with the presence of gypsum, wechecked the possible presence of other components distributed inform of phases with characteristics not suitable for being detectedby XRD.

To collect reliable analytical data SEM/EDS analyses were per-formed on pellets. The analytical data indicated that the elements

present in the samples were: Si, Al, Na, Mg, K, Cl, S, Ca and Fe.For better describing the overall situation, instead of reporting the

Table 4Textural properties of the samples.

Sample Specific surfacearea (m2 g−1)

Specific porevolume (cm3 g−1)

Total Microporousa

ZSM-5 fresh 296 0.18 0.12 (0.16)Z top 383 K 185 0.14 0.08 (0.10)Z top 823 K 310 0.22 0.12 (0.16)Mordenite fresh 397 0.38 0.16 (0.19)M bottom 383 K 217 0.30 0.07 (0.09)M bottom 823 K 367 0.47 0.16 (0.19)

a Specific pore volume normalized to the zeolite content in parentheses.

214 R. Vignola et al. / Chemical Engineering Journal 178 (2011) 210– 216

rior of

ea

(

(

F)e

(

Fig. 6. SEM micrographs collected in BE mode of the inte

lemental composition derived from EDS analysis, the followingssumptions were considered:

a) Fe was considered present as sulphide (FeS2), an assumptionbased on the SEM/EDS observations of as made and embeddedsamples (see below);

b) on the basis of the structural analysis, the residual S (i.e.not involved in the formation of FeS2), was attributed to

CaSO4·2H2O (gypsum);

(c) the residual Ca was considered as CaO, an indication of the pos-sible presence of other phases containing the earth-alkali metalion (e.g. CaCO3);

ig. 7. SEM/EDS line profile showing the distribution of Fe ( ) and S ( along the path indicated by the white line in the micrograph of a Mordenitextrudate.

the Mordenite extrudates taken from the top of filter M.

d) when S was not detected, Fe was considered as Fe2O3;(e) Cl was considered to be present as NaCl;(f) the excess of Na was indicated as Na2O;(g) for all the other elements, the most common oxidic species were

considered: SiO2, Al2O3, K2O and MgO.

The results of the SEM/EDS analysis for the ZSM-5 and Mordenitesamples are reported in Tables 2 and 3, respectively.

As far as the filters Z and Z′, constituted by ZSM-5, are con-cerned, the composition of the fresh sample reflected the use ofunidentified clay as binder (Table 1). In fact, the relatively highconcentration of Al as well as the presence of alkali (Na andK), earth-alkali (Mg and Ca) metal ions and even Fe should beattributed to the binder phase. Upon comparison of the analyticaldata collected on all the samples, it emerges that:

– there is small but significant decrease of the Al2O3 content inthe used samples (SAR = 40.4–47.0) respect to the fresh one(SAR = 39.8);

– excluding sodium chloride, whose precipitation occurs duringthe evaporation of the residual groundwater present in the extru-dates, the inorganic phases deposited are iron sulphide andgypsum;

– it is substantially excluded the precipitation of significantamounts of carbonatic phases (e.g. CaCO3).

Iron sulphide was detected in all samples, with a slightly higherconcentration on the top of the filters, corresponding to the inletof the water flow. Gypsum is also present in all the samples but

its deposition prevailed on the top of the filters (content ≥ 1.0 wt%)with respect to the bottom (content ≤ 0.5 wt%, Table 2).

SEM observations performed on the extrudates as such revealedthat iron sulphide and gypsum are mainly deposited on the external

R. Vignola et al. / Chemical Engineering Journal 178 (2011) 210– 216 215

Fig. 8. Adsorption ( � ) and desorption ( � ) isotherms measured on (a) fresh ZSM-5, (b) ZSM-5 taken from the top of filter Z, (c) fresh Mordenite and (d) Mordenitetaken from the bottom of filter M.

sAacm

tvbpflomtwwcpte

ivZoaa

urface, in the region corresponding to the bright areas of Fig. 3a. more detailed inspection revealed that gypsum is present eithers large isolated crystals (Fig. 3b) or in star-like aggregates of smallrystallites, while iron sulphide is present on form of aggregates oficronic and sub-micronic crystallites (Fig. 3c).Interesting was also to verify if the deposition occurred only on

he surface or even in the interior of the extrudates. The obser-ations performed on the embedded samples evidenced that theinder and the ZSM-5 crystals form a porous network with theresence of meso-macropores which assure the diffusion of theuid within the extrudate (Fig. 4a). It was also evident the presencef large silica particles, with dimensions in the range 10–100 �m,ost likely part of the binder (Fig. 4b). Deposits are present on

he external surface as well as in the interior of the extrudates,here isolated or small aggregates of iron sulphide crystallitesere observed (Fig. 4c). Notably, this phase was found to be mainly

oncentrated in the large cavities sometimes present in the internalart of the extrudates (Fig. 4d). Not only, line EDS profiles indicatedhat there was an enrichment of Fe and S in the external 5 �m layer,ven in the absence of discrete iron sulphide particles (Fig. 5).

In the case of filters M and M′, the elemental analyses were facil-tated by the fact that the binder used (alumina) did not contain theariety of elements characterizing the clay employed as binder of

SM-5. Except for Si, from Mordenite, and Al, mainly from alumina,nly trace amounts of Na and Cl were in fact detected in the freshdsorbent (Table 3). The analyses of the samples taken from the topnd the bottom of filters M and M′ evidenced that:

– there is small but significant increase of the Al2O3 content inthe used samples (SAR = 5.67–6.09) respect to the fresh one(SAR = 6.67);

– excluding sodium chloride and possibly potassium, whoseprecipitation occurs during the evaporation of the residualgroundwater present in the extrudates, the inorganic phasesdeposited are mainly gypsum and in lesser extent iron sulphide;trace amounts of Mg were also detected;

– as in the case of ZSM-5, it is substantially excluded the precipi-tation of carbonatic phases (e.g. CaCO3).

Under the qualitative point of view, the behaviour of Morden-ite filters can be considered similar to that of ZSM-5. Some smallbut significant differences exist when considering the quantitativedata. In particular, the higher amount of gypsum detected and thelower amount of iron sulphide are deposited on Mordenite respectto ZSM-5. But still more interesting is the slight but significantincrease of the Al content (better appreciated by considering theSAR’s), a phenomenon which can be associated to the deposition ofAl species deriving from the partial dissolution of the binder of the Zand Z′ filters. It is worth noting that the decrease of the SAR is morepronounced in the case of filter M than for filter M′, as a consequence

of the different length of permanence in groundwater.

Gypsum and iron sulphide deposits were observed on the sur-face of the extrudates, which appeared to be substantially similarto that depicted in Fig. 3.

2 ineerin

psDcats

ofEis

4

tgMatnat

sioromsstimstdoat

pia

[

[

[

[

[

[

[

[

16 R. Vignola et al. / Chemical Eng

Even the Mordenite extrudates were characterized by theresence of large and diffused macroporosity (Fig. 6), neces-ary condition for assuring the penetration of the groundwater.ifferently from ZSM-5 extrudates, where the zeolite crystals werelearly identified (Fig. 4), in the Mordenite ones zeolite crystalsnd alumina binder formed a 3D network, which embedded rela-ively large alumina-rich particles (Fig. 6a–c) and sometimes smallpherical particles of unidentified phases (Fig. 6d).

Isolated iron sulphide crystals (sometimes agglomerated) werebserved in the interior of the agglomerated, though with lowerrequency with respect to ZSM-5 extrudates. Even in this case, lineDS profiles indicated that there was an enrichment of Fe and Sn the external 5 �m layer, even in the absence of discrete ironulphide particles (Fig. 7).

.3. Textural analysis

Textural analyses were performed in order to verify the varia-ions possibly induced by the long permanence of the adsorbents inroundwater. Two samples (ZSM-5 taken from the top of filter Z andordenite from the Bottom of filter M) were selected because char-

cterized by the highest amount of deposits of inorganic phases andogether with the fresh adsorbents were subjected to the determi-ation of the specific surface area and specific pore volume eitherfter mild thermal treatment (383 K) and after calcination at 823 Ko remove all the species adsorbed.

The results, reported in Table 4, evidenced that there were aignificant reduction of the specific surface area and pore volumen the samples treated at 383 K as a consequence of the presencef organic material adsorbed in the pores. As a matter of fact, theeduction of the total pore volume was exclusively due to the lossf micropore (i.e. zeolitic) volume. After calcination at 823 K, theicropore volume was fully restored at the original values, while a

ignificant increase of the total pore volume was observed for bothamples (Table 4). This was associated to the modification of theexture of the materials evidenced by the different hysteresis loops,ndicative of the formation of additional mesopores during the per-

anence in groundwater (Fig. 8). Most likely, this phenomenonhould be associated to the reconstruction of the binders due tohe loss of Al in the case of ZSM-5 and to its deposition in Mor-enite samples. On the contrary, no modification was expectedn zeolite structures, since the microporous volume did not showny variation, being completely recovered after calcination at highemperature.

These data confirmed also that the deposition of the inorganichases (mainly gypsum and iron sulphide) did not have any signif-

cant influence on the textural properties of the adsorbents, whosedsorption properties can be easily restored by thermal treatment.

[

g Journal 178 (2011) 210– 216

5. Conclusion

The availability of samples of zeolite adsorbents used in fieldtests of purification of groundwater contaminated by organic pol-lutants gave an unique opportunity of verifying the modificationinduced by the long permanence in groundwater rich in inorganicsalts.

The extensive physico-chemical characterization of samplestaken from the filters demonstrated that zeolites were not affectedby the tests, maintaining unaltered their structure and pore char-acteristics. Though a significant deposition of inorganic phases(mainly gypsum and iron sulphide) was observed on the surfaceand in minor extent in the interior of the extrudates, they did notinfluence the textural properties of the adsorbents.

The most significant phenomenon observed was related to themodification of the texture of the extrudates, with the increase ofthe specific volume of the mesopores in both samples. This featurewas likely a consequence of the extraction of Al from the binder ofZSM-5 (first filter) and its deposition on Mordenite (second filter).It was verified that even this phenomenon had no influence on themicroporous characteristics of the adsorbents, which can be easilyrestored by calcination at 823 K.

References

1] R. Vignola, U. Cova, F. Fabiani, G. Grillo, M. Molinari, R. Sbardellati, R. Sisto,Remediation of hydrocarbon contaminants in groundwater using specific zeo-lites in full-scale pump and treat and demonstrative permeable barrier tests, in:Proceedings of the 4th International FEZA Conference, Paris, France, 2008.

2] R. Vignola, G. Grillo, R. Sisto, G. Capotorti, P. Cesti, M. Molinari, Use of syntheticzeolites as sorbent material in permeable reactive barriers for in situ treatmentof groundwater at industrially contaminated sites, in: Proceedings of the Inter-national Symposium on Permeable Reactive Barriers, Belfast, Northern Ireland,2005.

3] M. Occelli, H.E. Robson, Zeolite Synthesis, ACS Symposium Series, AmericanChemical Society, Washington, DC, 1989.

4] R. Vignola, R. Bagatin, A. De Folly D’Auris, C. Flego, M. Nalli, D. Ghisletti, R. Millini,R. Sisto, Zeolites in a permeable reactive barrier (PRB): one year of field experi-ence in a refinery groundwater. Part 1: the performances, Chem. Eng. J. 178 (c)(2011) 204–209.

5] R. Klein, H. Schad, Results from a full scale funnel and gate system at the BEKA sitein Tübingen (Germany) using zero-valent iron, in: Proceedings of the 7th Inter-national FZK/TNO Conference on Contaminated Soil, Leipzig, Germany, 2000.

6] R. Gonzalez-Olmos, M. Iglesias, Thermodynamics and kinetics of fuel oxy-genate adsorption into granular activated carbon, J. Chem. Eng. Data 53 (2008)2556–2561.

7] J.E. Kilduff, T. Karanfil, Trichloroethylene adsorption by activated carbonpreloaded with humic substances: effects of solution chemistry, Water Res. 36(7) (2002) 1685–1698.

8] R. Pełech, E. Milchert, R. Wróbel, Adsorption dynamics of chlorinated hydrocar-

bons from multi-component aqueous solution onto activated carbon, J. Hazard.Mater. 137 (3) (2006) 1479–1487.

9] B.F. Mentzen, J.-M. Letoffe, P. Claudy, Enthalpy change and temperature of thereversible monoclinic–orthorombic phase transition in MFI type zeolitic mate-rials, Thermochim. Acta 288 (1–2) (1996) 1–7.