lable at ScienceDirect

Journal of Environmental Management 92 (2011) 23e30

Contents lists avai

Journal of Environmental Management

journal homepage: www.elsevier .com/locate/ jenvman

A procedure to design a Permeable Adsorptive Barrier (PAB) for contaminatedgroundwater remediation

A. Erto a,*, A. Lancia a, I. Bortone b, A. Di Nardo b, M. Di Natale b, D. Musmarra b

aDipartimento di Ingegneria Chimica, Università di Napoli Federico II, P.le Tecchio, 80 e 80125 Napoli, ItalybCIRIAM, Centro Interdipartimentale di Ricerca in Ingegneria Ambientale, Dipartimento di Ingegneria Civile, Seconda Università di Napoli, via Roma, 29 e 81031 Aversa (CE), Italy

a r t i c l e i n f o

Article history:Received 16 February 2010Received in revised form27 July 2010Accepted 28 July 2010Available online 16 September 2010

Keywords:Permeable Adsorptive BarrierActivated carbonAdsorptionGroundwater contaminationTetrachloroethylene

* Corresponding author. Tel.: þ39 081 7682246; faxE-mail address: aleserto@unina.it (A. Erto).

0301-4797/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.jenvman.2010.07.044

a b s t r a c t

A procedure to optimize the design of a Permeable Adsorptive Barrier (PAB) for the remediation ofa contaminated aquifer is presented in this paper. A computer code, including different routines thatdescribe the groundwater contaminant transport and the pollutant capture by adsorption in unsteadyconditions over the barrier solid surface, has been developed. The complete characterization of thechemicalephysical interactions between adsorbing solids and the contaminated water, required by thecomputer code, has been obtained by experimental measurements. A case study in which the proceduredeveloped has been applied to a tetrachloroethylene (PCE)-contaminated aquifer near a solid wastelandfill, in the district of Napoli (Italy), is also presented and the main dimensions of the barrier (lengthand width) have been evaluated. Model results show that PAB is effective for the remediation of a PCE-contaminated aquifer, since the concentration of PCE flowing out of the barrier is everywhere alwayslower than the concentration limit provided for in the Italian regulations on groundwater quality.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

The accidental discharge of solid waste landfill leachate oftenleads to groundwater contamination, both by inorganic and organiccompounds (Jun et al., 2009; Kjeldsen et al., 2002). A majorcontribution to groundwater pollution at solid waste landfillscomes from chlorinated organic compounds, mainly originatingfrom plastic material degradation due to rainfall water leaching.Tetrachloroethylene (PCE) can be considered as one of the mostdangerous pollutants, due to its high toxicity and persistence grade(EPA, 1988).

Directive 1999/13/EC (Solvent Emissions Directive) requires EUMember States to implement controls on the emissions of volatileorganic compounds, including PCE. Furthermore, under directives2000/60/EC (Water Framework Directive) and 2006/118/EC(Groundwater Directive), this compound is expressly listed amongthemost dangerous pollutants. Consequently, the Italian regulatorylimit for groundwater quality has recently been set at 1.1 mg l�1.

In aquatic environments, PCE can be present both in NAPL form(non-aqueous phase liquid) and dissolved in water. The environ-mental fate of this species strictly depends on its main physical andchemical properties; PCE is more dense and less viscous than

: þ39 081 5936936.

All rights reserved.

water, which promotes its migration from superficial water andcontaminated soils towards groundwater. In these conditions, theabsence of light prevents photolytic degradation and the highpressure reduces volatility, thus increasing residence times inaquifers (ATSDR, 1997). Moreover, the hydraulic characteristics ofthe surrounding media greatly influence its mobility and persis-tence in groundwater.

Several remediation technologies have been developed forgroundwater depuration that include ex-situ treatments, coupledwith pump and treat technique, or in-situ treatments, performedwith Permeable Reactive Barriers (PRB). There is a growing interestfor PRB installations as an effective alternative to classic remedia-tion methods, thanks to its low operating and maintenance costsfor groundwater remediation, where there are often very highvolumes and low pollutant concentrations.

In a PRB treatment, the barrier is commonly built with reactivematerials whose hydraulic conductivity is higher than that of thesurrounding soils, so that the contaminated groundwater is forcedto pass through the barrier itself, moving under natural hydraulicgradient. The mechanism of action of a PRB depends on the choiceof the reactive material used to build the barrier. The most widelyused reactive material for PCE-contaminated groundwater is zero-valent iron (’D’Andrea et al., 2005; EPA, 1998; Farrell et al., 2000;Higgins and Olson, 2009; Jun et al., 2009; Moon et al., 2005;Plagentz et al., 2006; Vogan et al., 1999) in which organic pollut-ants are degraded through a series of reduction reactions. These

A. Erto et al. / Journal of Environmental Management 92 (2011) 23e3024

reduction reactions are greatly unselective andmay frequently leadto solid precipitation deriving from inorganic reduction. Thisundesired event can seriously affect the barrier efficiency as itdrastically reduces its porosity and conductivity; it can also giverise to preferential flow paths that lead to a reduction of the contacttime of the contaminated water with the reactive material(Kamolpornwijit et al., 2003; Li et al., 2006; Mackenzie et al., 1999).Another limitation of zero-valent iron PRBs is that dechlorinationreactions are very slow at high water flow-rates, hence barriersshould be very thick in order to extend the contaminant residencetime in the system (Moon et al., 2005).

PRBs made of alternative materials, such as zeolites, fly ash andactivated carbons, where contaminants are removed by adsorptioninstead of reduction reactions, may be valid solutions for long-livedbarriers (Ake et al., 2003; Czurda and Haus, 2002; Di Natale et al.,2008; Komnitsas et al., 2006; Lorbeer et al., 2002).

In particular, the removal of chlorinated organic compounds,such as PCE, from polluted water and wastewater can be efficientlyachieved by adsorption combining a good efficiency with a verysimple process configuration (Bembnowska et al., 2003; Pelechet al., 2003; Suzuki, 1990). As adsorption phenomena take place,the pollutant is immobilized into the barrier, avoiding any precip-itation phenomena. In this sense, the Permeable Adsorbing Barrier(PAB) can be considered as a particular case of PRB made ofadsorptive material.

There are several sorbents that can be used for organiccompound removal, including fly ash, natural materials or wastematerials, but industrial activated carbons appear to be an optimalsolution in as much as they have much greater adsorption capac-ities and can therefore be used in lower amounts to build thebarrier and the in-situ intervention can be minimized.

The design and optimization of an in-situ effective depurationtechnology must also take into account the hydrological andgeotechnical properties of the entire polluted aquifer. Hence,a complete characterization of the site chemical, hydro-geologicaland geotechnical properties is required.

The main goal of the present work is the development ofa procedure to optimize the design of a PAB for the remediation ofa PCE-contaminated aquifer. The procedure includes the charac-terization of the chemicalephysical interactions between adsorb-ing solids and the contaminated water and the development ofa computer code able to describe the groundwater contaminanttransport and capture by the barrier. The procedure has beenapplied to a PCE-contaminated aquifer near a solid waste landfill inthe district of Napoli (Italy) and the main dimensions of the barrier(length and width) are evaluated.

2. Permeable Adsorptive Barrier design

The design of a Permeable Adsorptive Barrier for a pollutedaquifer, for which a hydraulic, geotechnical and contaminantcharacterization has been previously performed, mainly consists inthe definition of the barrier location, orientation and dimensions.The problem cannot be approached by direct calculation and aniterative procedure has to be applied. First the location and size ofthe barrier have to be chosen and then it is necessary to checkwhether the choice allows a thorough pollutant capture during thewhole lifetime of the barrier. Moreover, the minimum dimensionsof the barrier have to be identified for it to be cost-effective. Sincethe calculation procedure may be very time-consuming, an opti-mization criterion is necessary.

Eventually, as the pollutant concentration at barrier inlet mayvaryduring the barrierworkingperiod, the occurrence of desorptionphenomena within the barrier must also be taken into consider-ation. Adsorption continues until the pollutant concentration at the

barrier inlet remains lower than the equilibrium values corre-sponding to the amount of pollutant adsorbed on carbon, but if thepollutant inlet concentration decreases, the pollutant adsorbedmaybe desorbed from the barrier solid, giving rise to a contaminatedplume at the exit of the barrier itself. When desorption occurs,a thicker barrier ensures a slower release of the pollutant adsorbed,avoiding critical outbound concentrations. Therefore, the barriermust be designed both to retain intense concentrationpeaks and forlong term performances, also considering the occurrence of anydesorption phenomena.

2.1. PAB design procedure

The first step of the procedure consists in the definition of thesize of the calculation domain that must be large enough to includethe whole pollutant plume and longer than the plume itself alongthe direction of the natural groundwater flow. Then the calculationgrid (i.e., the dimensions of one 3D calculation cell) must be chosen,also considering that larger domains or smaller calculation cellsproduce more precise results but calculation is more time-consuming. Then the size of the calculation grid must be definedspecifically for the barrier, by refining the calculation cells in order toreach a more detailed concentration profile inside the barrier itself.

The second step is the preliminary choice of barrier location anddimensions, i.e., distance from the pollutant plume (E), orientationwith respect to the North (b), length (L), height (H) and width (W)andbarrier adsorbingmaterial. Barriers are generally built at a depththat somewhat over-encompasses the vertical dimensions of thecontaminant plume, as a safety factor. The preliminary choice ofbarrier width can be made by considering the following inequality:

Wub

> ðkcaÞ�1 (1)

where kc represents the mass transfer coefficient for adsorptionreactions, a represents the external specific surface area of theadsorbent particles and ub is the average groundwater flow velocitythrough the barrier, which must be previously identified bya complete hydro-geological characterization.

Eq. (1) imposes that the contaminated flow travel time throughthe barrier is long enough for adsorption process to take place.Values in Eq. (1) for a and kc can be determined by a preliminaryadsorbent characterization, often provided by the vendor.

Once the geometrical properties of the barrier together with theadsorbingmaterial have been chosen, the barrier performancemustbe checked by evaluating the evolution in time of the pollutantplume over the calculation domain considering that the pollutantconcentration downstream the barrier (CW) has to be everywherelower than the fixed limit value (Clim) for the whole barrier lifetime.

Eventually, barrier dimension and location will be optimized inorder to find the solution that complies with the fixed concentra-tion limit value and is the least expensive (e.g., has the minimumvolume of adsorbing material, VPAB).

The procedure, schematically shown in the flow chart of Fig. 1,requires the development of a computer routine describing thetransport of pollutant and its interactions with the adsorbingmaterials (PAB DESIGN TOOL�); furthermore a visualization code isrequired to present the results (AMBSIT�).

The main characteristics of the routine (PAB DESIGN TOOL�)included in the procedure are described in the following.

2.2. PAB design tool

The main goal of the PAB DESIGN TOOL� routine is the evalu-ation of the contaminant plume migration in the aquifer and its

Fig. 1. PAB design flow chart.

A. Erto et al. / Journal of Environmental Management 92 (2011) 23e30 25

evolution in time. To this purpose the dissolved contaminant masstransport equation has to be solved. The aquifer is a porousmediumand the pollutant transport is due to advectionedispersionprocesses; moreover inside the barrier adsorption/desorptionphenomena have to be taken into account.

The design procedure was conceived under the hypothesis ofconstant pollutant concentration profiles throughout the height ofthe aquifer (i.e., z-dimension). Hence, for a two-dimension system(x; y, coinciding with the top plane of the aquifer), the dissolvedcontaminant mass transport equation may be written as follows(Bear, 1979):

vCvt

þ rbnb

vu

vtþ u!VC

nb� VðDhVCÞ ¼ 0 (2)

In Eq. (2), C represents concentration in fluid, u! the unit fluxvector, u the contaminant concentration on solid, rb the dryadsorbing material bulk density, nb the soil porosity.

The hydrodynamic dispersion coefficientDh can be expressed as:

Dh ¼ Dþ D�d (3)

In Eq. (3), D is the tensor of mechanical dispersion and D*d is the

coefficient of molecular diffusion (a scalar). The components of themechanical dispersion tensor may be expressed as follows:

D ¼

26666664

aLu2xjuj�!þ aT

u2y

juj�! ðaL � aT Þuxuy

juj�!

ðaL � aTÞuxuy

juj�! aLu2y

juj�!þ aTu2xjuj�!

37777775

(4)

In Eq. (4), aL and aT represent longitudinal and transverse dis-persivity coefficients, respectively. The unit flux vector ð u!Þcan bedetermined by the application of the Darcy equation, written as:

u! ¼ �Ks$Vh (5)

in which Ks is the hydraulic conductivity of the soil and h is thehydraulic load which can be calculated starting from the Laplaceequation:

�v2hvx2

� v2hvy2

� v2hvz2

¼ 0 (6)

The Eq. (6) can be integrated with appropriate boundaryconditions. The second term on the left hand side of (2) reads as:

rbnb

vu

vt¼ kca

hC � C*ðuÞ

i(7)

In Eq. (7), C*(u) derives from the adsorption isotherm ofcontaminant on adsorbent material.

As to initial conditions, contaminant liquid concentration in allthe area is previously determined throughout the entire flowdomain and the initial contaminant concentration on sorbentmaterial is assumed to be zero:

C ¼ C0 t ¼ 0 cx cyu ¼ 0 t ¼ 0 cx cy (8)

The boundary conditions are stated as follows:

C ¼ 0 x ¼ 0 cy ct

vCvt

þ u!VCnb

� VðDhVCÞ ¼ 0 x ¼ X cy ct

C ¼ 0 y ¼ 0 cx ct

C ¼ 0 y ¼ Y cx ct

(9)

assuming a reference frame coinciding with the boundary of thedomain, X is the distance between the barrier and the origin, and Yis the size of the domain in y direction.

The PAB DESIGN TOOL performs the numerical integration ofEqs (2)e(7) with the initial and boundary conditions (8)e(9) bymeans of a first order finite difference implicit scheme.

Results are reported in terms of contaminant concentration allover the domain and are graphically represented by means ofAMBSIT� (by CIRIAM, Seconda Università di Napoli), a GIS (Geo-graphic Information System) application specifically developed toimprove contour plots of pollutant concentration. It allows to drawconcentration contour plots with different interpolation algorithms.

2.3. Laboratory sorbent characterization

In Eq. (7), the term C* ¼ C*(u) describes the thermodynamicequilibrium between the concentration of the pollutant in theliquid phase and the amount adsorbed over the solid surface(adsorption isotherm). This term is of great importance to properlydescribe the behavior of an adsorbing barrier because it defines thedriving force which promotes pollutant capture or pollutantdesorption. Furthermore, the C* ¼ C*(u) relationship is typical ofthe couple pollutant/adsorbing solid and includes the dependenceon temperature and chemistry of solution (i.e., pH, salinity, ionicstrength, presence of other pollutants); therefore a reliableC* ¼ C*(u) relationship has to be obtained by ad hoc experimentalmeasures. The procedure to obtain experimentally the adsorptionisotherm of contaminant on adsorbing material is described ina simplified manner in the following.

2.3.1. Materials and methodThe adsorbing solid chosen for the barrier set-up is a commer-

cially available non impregnated granular activated carbon (GAC),the Aquacarb 207EA� (Sutcliffe Carbon), and the pollutantconsidered is tetrachloroethylene (PCE).

A complete solid characterization by means of mercury poros-imetry (Carlo Erba Porosimeter, 2000), BET (Carlo Erba SORPTO-MATIC 1900) and SEM (Environmental Scanning ElectronMicroscope

A. Erto et al. / Journal of Environmental Management 92 (2011) 23e3026

Philips XL30) has been carried out. This material has a BET surfacearea of 950 m2 g�1 and a micropore volume of 0.249 cm3 g�1 mainlycentered in the pore size region of 8e18Å. The dry bulk density (rb) isabout 500 kg m�3, the porosity (nb) is 0.4 m3/m3 and its hydraulicconductivity is about 0.001 m s�1.

A commercial mineral water, whose composition can be repre-sentative of groundwater, has been used for sample preparation. Ithas a pH ¼ 8.0 and an ionic strength of 0.0046 M; a complete list ofchemical properties is reported in Erto et al. (2010).

Humic acid (Fluka) has been used as a model component tosimulate the co-presence of natural organic matter (NOM) intogroundwater. Humic acid, as a dry powder, is a complex mixture oforganic materials including proteins, polysaccharides, phenols andcarboxylic acids with M.W. ranging between 2000e500,000 amu.The ultimate analysis has shown a total organic carbon (TOC)content of 46.6%.

PCE adsorption isotherms have been conducted in a PIDcontrolled thermostatic oven, using glass vessels as batch reactors.The sample solutions have been prepared by adding PCE (SigmaAldrich, 99%) and activated carbon to 200 ml amber stained,headspace-free glass vessels of mineral water. Then, they have beensealed with a Teflon cap and covered with an aluminium sheet toprevent photodegradation phenomena. The initial PCE concentra-tions and the carbon mass used in each run have been selected sothat the equilibrium concentrations fall in the range 0e2 mg l�1,typical of groundwater contamination (ATSDR, 1997). The initialsolution pH has been adjusted by adding nitric acid (0.1 M) orsodium hydroxide (0.1 M) to the stock solutions and it has not beenfurther altered during the experimental runs. Similarly, the salinityof stock solutions has been modified by adding sodium chloride.

PCE carbon capture (u) has been determined by PCE materialbalance as:

u ¼ C0 � C*

m$V (10)

where C* is the PCE equilibrium concentration, C0 is the PCE initialconcentration, m is the solid mass and V is the solution volume.

Further details on experimental procedure and analyticalmethods are reported in Erto et al. (2010).

2.3.2. Adsorption isothermsThe effect of groundwater temperature, pH and salinity and the

co-presence of NOM have been investigated by adsorptionisotherms determination. These parameters may vary to a large

C, g/μ l

0 500 1000 150 000

ω, m

g/g

0

50

100

150

20010°C20°C

A B

0 2

Fig. 2. Adsorption isotherms of PCE onto Aquacarb 207EA� GAC as a function of temperatuLangmuir model results (lines).

extent, for example when leachate infiltrations occur at solid wastelandfills (Zaporozec, 2002; Kjeldsen et al., 2002), and the removalefficiency of adsorption processes is affected accordingly(Benjamin, 2002).

In Fig. 2A, the PCE adsorption isotherms on Aquacarb 207EA�GAC at 10 and 20 �C, and a constant value of pH¼ 7 are reported. Asexpected, adsorption shows the characteristic exothermic trendand PCE adsorption capacity values are typical of activated carbons(Bembnowska et al., 2003; Pelech et al., 2003).

The effect of pH on PCE adsorption at 20 �C is reported in Fig. 2B,which shows that the PCE adsorption capacity does not depend onpH. Experimental runs focused on the effects of ionic strength(NaCl) and NOM presence (Humic acid) show that neither salinitynor humic acid affects the adsorption capacity, therefore they arenot reported.

A data regression analysis has been performed on experimentaldata at two temperature levels; the Langmuir model is the best forPCE adsorption description, as can be observed in Fig. 2A.

All the regression analyses were carried out on the unmodifiedisotherm equations, rather than their linearized formulation, sincethis approach is generally considered as the most appropriate(Montgomery, 2001; Benjamin, 2002). The least residual sum-of-squares has been used as a criterion for the determination of thebest fitting model parameters.

The analysis includes the determination of the mean value andstandard error of parameters, the coefficient of determination (R2),the Student-test (T) for each parameter, the normality test (P) andthe Fisher-test (F) for the regression. The real meaning of theseparameters is reported in classical statistical works (Montgomery,2001).

The values of regression parameters of Langmuir model, withtheir error of determination, are reported in Table 1.

2.4. Case study

The procedure described in Fig. 1 has been applied to the designof a PAB for the remediation of a contaminated aquifer near a solidwaste landfill in Giugliano in Campania, a town in the metropolitanarea North of Napoli (Italy). In this area, many solid waste landfillsexist and over the past 20 years, about eight million tons of urbanand special waste have been deposited in these landfills. Apreliminary site characterization, including geotechnical, hydraulicand pollution properties, is available. The groundwater aquifer iscontaminated by a large number of pollutants, both inorganic and

C, g/μ l

0 500 1000 1500 2000

ω, m

g/g

0

50

100

150

200pH=3pH=7pH=10

re (A) and equilibrium pH (B). Comparison between experimental data (symbols) and

Table 1Langmuir model parameters for PCE adsorption regression analysis.

T (�C) Langmuirparametersu ¼ umax

KC1þKC

Value R2 P-test F-test

Mean Std. error T-test

10 umax (mg g�1) 913.9 121.9 14.60 0.9976 <0.0001 1489.8DG (kJ mol�1) �23.27 0.372 32.73K (l mol�1) 19830 3140 11.23

20 umax (mg g�1) 567.4 35.44 16.01 0.9927 <0.0001 2319.8DG (kJ mol�1) �24.95 0.254 97.95K (l mol�1) 28190 2949 9.56

Table 2Case study: aquifer characteristic, PAB properties and numerical model parameters.

Aquifer characteristicPolluted area total extent, A 1.1 km2

Aquifer bed height, H 12.5 mPiezometric gradient, J 0.01 mm�1

Porosity, ns 0.25Dry soil bulk density, rs 1500 kg m�3

Hydraulic conductivity, Ks 5 � 10�5 m s�1

Longitudinal dispersivity, aL 1 mTransverse dispersivity, aT 0.01 mMolecular diffusion coefficient, D�

d 10�8 m2 s�1

Numerical model parametersHorizontal space step, Dx 6 mTransversal space step, Dy 6 mVertical space step, Dz 12.5 mTime step, Dt 60 days

PAB propertiesReactive media Activated carbonDry bulk density, rb 520 kg m�3

Porosity, nb 0.4Hydraulic conductivity, Kb 0.01 ms�1

Horizontal space step, Dxb Dx/100Transversal space step, Dyb DyBarrier width, W 3 mBarrier length, L 900 mDistance from contaminant plume, E 6 mBarrier distance from domain boundary, X 1266 mTransversal domain extension, Y 996

A. Erto et al. / Journal of Environmental Management 92 (2011) 23e30 27

organic. In particular a severe PCE pollution has been found. Thesoil composition can be approximated with a unique mineral type(tuff), whose hydraulic conductivity is 5 � 10�5 m s�1. Thegroundwater flux lines are EW oriented under a piezometricgradient of 0.01. An impermeable layer is present at the bottom ofthe aquifer, hence a horizontal movement of water in the aquiferwith the pollutants can be predicted.

Fig. 3 shows the preliminary characterization in the currentconditions in the form of PCE iso-concentrations, together with thepiezometric lines of the aquifer. This figure shows that PCE pollu-tion is widespread over the area with a PCE maximum concentra-tion value that is about 20 times higher than Italian regulatory limitfor groundwater quality, fixed at 1.1 mg l�1.

In Table 2 the properties of the aquifer, the main characteristicsof the adsorbing material chosen for the barrier and the numericalparameters used in the calculation are reported.

For the sake of simplicity, the total volume of pollutedgroundwater has been assumed to be constant during the moni-toring time of the aquifer and the dissolution of further amounts ofpollutant or the dilution of its concentration due to rainfall havebeen neglected. Therefore, the evolution of PCE concentrationprofiles is predicted by considering just advectiveeconvectivephenomena as well as adsorption phenomena inside the barrier.

Fig. 3. PCE iso-concentration and iso-piezom

2.5. Numerical results

The design procedure described above has been applied to theaquifer described in Fig. 3 by using the parameter reported in Table2. As previously said, the optimal barrier position and dimensionshave to be identified by a trialeerror procedure and a great numberof numerical simulations are necessary to optimize the barrierparameters. In particular, the length and the height of the barrier

etric lines for the case study considered.

A. Erto et al. / Journal of Environmental Management 92 (2011) 23e3028

mainly depend on the extension of the contaminant plume; as to itswidth, considering that the higher the width, the lower the barrieroutflow pollutant concentration (CW), it is defined so as to guar-antee CW is lower than the regulatory limit (Clim) for PCE (cf. Fig. 1).The best results are obtained for a barrier 3 mwide and 900m long,having a direction almost coincident with North heading direction,as shown in Fig. 4.

The results of the numerical simulation are reported in Fig. 4which shows the evolution over run time of pollutant spots inthe form of snapshots taken every 10 years. Fig. 4 shows thatpollutant spots move towards the barrier where they are captured;indeed in any run time the PCE concentration flowing out of thebarrier (CW) is always lower than the concentration limit (Clim)fixed at 1.1 mg l�1 Fig. 4 also shows that run time has been prolonged

Fig. 4. PCE iso-concentration into the aqui

to 60 years in order to assess barrier performances. This very longtime is due to the hydraulic properties of the aquifer considered.

Eventually, results in Fig. 4 show that thewhole pollutant plumepasses through the barrier and no overflow is observed at thebarrier ends.

It is worth noticing that as adsorption takes place, the barrierstarts to saturate; therefore, the barrier width has to be largeenough to ensure the capture of the maximum PCE concentration.Moreover, when the PCE concentration approaching the barrierdecreases with time the desorption of previously captured PCEmayoccur. In order to better clarify this consideration, Fig. 5 shows, forthe observation point S indicated in Fig. 4, the PCE concentration atthe inlet (CIN) and at the outlet (CW) of the barrier, as a function ofrun time.

fer as a function of the working time.

Fig. 5. Breakthrough curve for PCE concentration in correspondence of the point S indicated in Fig. 3.

A. Erto et al. / Journal of Environmental Management 92 (2011) 23e30 29

Results in Fig. 5 show that the inbound PCE concentrationchanges during the run time having maxima and minima, theoutbound PCE concentration (CW) follows this behavior with valuesthat are always far lower than the concentration limit (Clim).Moreover, Fig. 5 shows that the barrier attenuates the oscillation ofpollutant concentration and shifts it forward with respect to runtime; moreover, at the end of the run time the barrier has a residualadsorption capacity that may represent a protectionwith respect toother pollutant waves.

3. Conclusions

In this work, a procedure for the design of an activated carbonPermeable Adsorptive Barrier (PAB) for the remediation of a PCE-polluted aquifer has been developed. The procedure has beenapplied to a contaminated aquifer near a solid waste landfill in thedistrict of Napoli (Italy), which can be considered as a case study,obtaining the best geometrical parameters for the barrier. Numer-ical simulations have been able to describe the migration of thecontaminated plume towards the barrier under the naturalhydraulic gradient and the adsorptive action inside the barrier andhave been used to assess barrier performances.

Numerical results showed that PAB is a valid remediationtechnology for the in-situ treatment of a PCE-contaminated aquifer,as the concentration of PCE flowing out of the barrier resultedeverywhere always lower than the stated regulatory concentrationlimit on groundwater quality.

Finally, adsorbing material blinding effects, over a very longtime, and competitive adsorption by other pollutants deserve moreinvestigation.

Notation

a adsorbing material external surface area, m2 m�3

C liquid concentration, mg l�1

C* equilibrium liquid concentration, mg l�1

C0 initial liquid concentration in batch experiments, mg l�1

CIN barrier inflow pollutant concentration, mg l�1

Clim pollutant regulatory limit value, mg l�1

CW barrier outflow pollutant concentration, mg l�1

D tensor of mechanical dispersionD�d molecular diffusion coefficient, m2 s�1

Dh hydrodynamic dispersion coefficient, m2 s�1

E distance between barrier and pollutant plume, mF-test F-Fisher statistical testH barrier height, mh hydraulic load, mJ Piezometric gradient, m m�1

K Langmuir constant, l mol�1

Kb barrier hydraulic conductivity, m s�1

Kc mass transfer coefficient, s�1

Ks hydraulic conductivity, m s�1

L barrier length, mm activated carbon mass in batch experiments, gnb barrier porosityns soil porosityP-test normality testR2 coefficient of determinationT absolute temperature, KT-test T-Student statistical testub groundwater flow velocityV liquid volume in batch experiments, lVPAB barrier adsorbing material volume, m3

V limPAB minimum barrier adsorbing material volume, m3

W barrier width, maL longitudinal dispersivity, maT transversal dispersivity, mb barrier orientation, �

DG Gibbs free energy, kJ mol�1

Dx horizontal space step, mDy transversal space step, mDz vertical space step, mrb activated carbon density, kg m�3

rs dry soil bulk density, kg m�3

u activated carbon adsorption capacity, mg g�1

uMAX maximum carbon adsorption capacity, mg g�1

References

Ake, C.L., Wiles, M.C., Huebner, H.J., McDonald, T.J., Cosgriff, D., Richardson, M.B.,Donnelly, K.C., Phillips, T.D., 2003. Porous organoclay composite for the sorptionof polycyclic aromatic hydrocarbons and pentachlorophenol from groundwater.Chemosphere 51, 835e844.

ATSDR (Agency for Toxic Substances and Disease Registry), 1997. ToxicologicalProfile for Tetrachloroethylene (PCE). U.S. Public Health Service, U.S. Depart-ment of Health and Human Services, Atlanta, GA.

Bear, J., 1979. Hydraulics of Groundwater. McGraw-Hill, New York.

A. Erto et al. / Journal of Environmental Management 92 (2011) 23e3030

Bembnowska, A., Pelech, R., Milchert, E., 2003. Adsorption from aqueous solutionsof chlorinated organic compounds onto activated carbon. J. Colloid Interface Sci.265, 276e282.

Benjamin, M., 2002. Water Chemistry. McGraw Hill, New York.Czurda, K.A., Haus, R., 2002. Reactive barriers with fly ash zeolites for in situ

groundwater remediation. Appl. Clay Sci. 1e2, 3e20.D’Andrea, P., Lai, K.C., Kjeldsen, P., Lo, I.M., 2005. Effect of groundwater inorganics

on the reductive dechlorination of TCE by zero-valent iron. Water Air Soil Poll.162, 401e420.

Di Natale, F., Di Natale, M., Greco, R., Lancia, A., Laudante, C., Musmarra, D., 2008.Groundwater protection from cadmium contamination by Permeable Adsorb-ing Barriers. J. Hazard. Mater. 160 (2e3), 428e434.

Updated Health Assessment Document for Tetrachloroethylene, 1988. EPA (U.S.Environmental Protection Agency). EPA/600/8e82/005B.

Permeable Reactive Barrier Technologies for Contaminant Remediation, 1998. EPA(U.S. Environmental Protection Agency). EPA/600/R-98/125.

Erto, A., Andreozzi, R., Di Natale, F., Lancia, A., Musmarra, D., 2010. Experimental andstatistical analysis of trichloroethylene adsorption onto activated carbon. Chem.Eng. J. 156, 353e359.

Farrell, J., Melitas, N., Kason, M., Li, T., 2000. Electrochemical and column investi-gation of iron-mediated reductive dechlorination of trichloroethylene andpercholoroethylene. Environ. Sci. Technol. 34, 2549e2556.

Higgins,M.R.,Olson, T.M., 2009. Life-cycle case studycomparisonofpermeable reactivebarrier versus pump-and-treat remediation. Environ. Sci. Technol. 43, 9432e9438.

Jun, D., Yongsheng, Z., Weihong, Z., Mei, H., 2009. Laboratory study on sequencedpermeable reactive barrier remediation for landfill leachate-contaminatedgroundwater. J. Hazard. Mater. 161 (1), 224e230.

Kamolpornwijit, W., Liang, L., West, O.R., Moline, G.R., Sullivan, B.A., 2003. Prefer-ential flow path development and its influence on long-term PRB performance:column study. J. Contam. Hydrol. 66, 161e178.

Kjeldsen, P., Barlaz, M.A., Rocker, A.P., Baun, A., Ledin, A., Christensen, T.H., 2002.Present and long-term composition of MSW landfill leachate: a review. Crit.Rev. Env. Sci. Tech. 32 (4), 297e336.

Komnitsas, K., Bartzas, G., Paspaliaris, I., 2006. Modeling of reaction front progressin fly ash permeable reactive barriers. Environ. For. 7, 219e231.

Li, L., Benson, C.H., Lawson, E.M., 2006. Modeling porosity reductions caused bymineral fouling in continuous wall permeable reactive barriers. J. Contam.Hydrol. 83 (1e2), 89e121.

Lorbeer, H., Starke, S., Gozan, M., Tiehm, A., Werner, P., 2002. Bioremediation ofchlorobenzene-contaminated groundwater on granular activated carbonbarriers. Water Air Soil Poll. Focus. 2, 183e193.

Mackenzie, P.D., Horney, D.P., Sivavec, T.M., 1999. Mineral precipitation and porositylosses in granular iron columns. J. Hazard. Mater. 68, 1e17.

Montgomery, D.C., 2001. Design and Analysis of Experiments, fifth ed.. John Wiley &Sons Inc., New York.

Moon, J.W., Moon, H.S., Kim, H., Roh, Y., 2005. Remediation of TCE-contaminatedgroundwater using zero valent iron and direct current: experimental resultsand electron competition model. Environ. Geol. 48 (6), 805e817.

Pelech, R., Bembnowska, A., Milchert, E., 2003. Adsorption of hydrocarbon chloro-derivatives onto DTO commercial carbon from multi-component aqueoussolutions. Adsorp. Sci. Technol. 21 (8), 707e720.

Plagentz, V., Ebert, M., Dahmke, A., 2006. Remediation of groundwater containingchlorinated and brominated hydrocarbons, benzene and chromate by sequen-tial treatment using ZVI and GAC. Environ. Geol. 49 (5), 684e695.

Suzuki, M., 1990. Adsorption Engineering. Elsevier, Amsterdam.Vogan, J.L., Focht, R.M., Clark, D.K., Graham, S.L., 1999. Performance evaluation of

a permeable reactive barrier for remediation of dissolved chlorinated solventsin groundwater. J. Hazard. Mater. 68, 97e108.

IHP-VI, Series on groundwater N� 2. In: Zaporozec, A. (Ed.), GroundwaterContamination Inventory: A Methodological Guide. UNESCO, Paris, France.