How does contamination move in groundwater? | iFLUX · flux data raise up to 13 g/m2/day for perchloroethylene in the plume zone. Calculated PFM flux averaged concentra-tion data

1 23

Environmental Science and PollutionResearch ISSN 0944-1344Volume 20Number 7 Environ Sci Pollut Res (2013)20:4621-4634DOI 10.1007/s11356-012-1417-8

Field demonstration and evaluation of thePassive Flux Meter on a CAH groundwaterplume

G. Verreydt, M. D. Annable,S. Kaskassian, I. Van Keer, J. Bronders,L. Diels & P. Vanderauwera

1 23

Your article is protected by copyright and

all rights are held exclusively by Springer-

Verlag Berlin Heidelberg. This e-offprint is

for personal use only and shall not be self-

archived in electronic repositories. If you wish

to self-archive your article, please use the

accepted manuscript version for posting on

your own website. You may further deposit

the accepted manuscript version in any

repository, provided it is only made publicly

available 12 months after official publication

or later and provided acknowledgement is

given to the original source of publication

and a link is inserted to the published article

on Springer's website. The link must be

accompanied by the following text: "The final

publication is available at link.springer.com”.

RESEARCH ARTICLE

Field demonstration and evaluation of the Passive Flux Meteron a CAH groundwater plume

G. Verreydt & M. D. Annable & S. Kaskassian &

I. Van Keer & J. Bronders & L. Diels & P. Vanderauwera

Received: 5 July 2012 /Accepted: 10 December 2012 /Published online: 3 January 2013# Springer-Verlag Berlin Heidelberg 2012

Abstract This study comprises the first application of thePassive Flux Meter (PFM) for the measurement of chlori-nated aliphatic hydrocarbon (CAH) mass fluxes and Darcywater fluxes in groundwater at a European field site. ThePFM was originally developed and applied to measurementsnear source zones. The focus of the PFM is extended fromnear source to plume zones. For this purpose, 48 PFMs of1.4 m length were constructed and installed in eight differ-ent monitoring wells in the source and plume zone of aCAH-contaminated field site located in France. The PFMswere retrieved, sampled, and analyzed after 3 to 11 weeks ofexposure time, depending on the expected contaminant flux.PFM evaluation criteria include analytical, technical, andpractical aspects as well as conditions and applicability.PFM flux data were compared with so-called traditional soiland groundwater concentration data obtained using activesampling methods. The PFMs deliver reasonable results forsource as well as plume zones. The limiting factor in thePFM applicability is the exposure time together with thegroundwater flux. Measured groundwater velocities at the

field site range from 2 to 41 cm/day. Measured contaminantflux data raise up to 13 g/m2/day for perchloroethylene inthe plume zone. Calculated PFM flux averaged concentra-tion data and traditional concentration data were of similarmagnitude for most wells. However, both datasets need tobe compared with reservation because of the different sam-pling nature and time. Two important issues are the PFMtracer loss during installation/extraction and the deviation ofthe groundwater flow field when passing the monitoringwell and PFM. The demonstration of the PFM at a CAH-contaminated field site in Europe confirmed the efficiencyof the flux measurement technique for source as well asplume zones. The PFM can be applied without concerns inmonitoring wells with European standards. The acquiredflux data are of great value for the purpose of site charac-terization and mass discharge modeling, and can be used incombination with traditional soil and groundwater samplingmethods.

Keywords Passive FluxMeter . Contaminant mass flux .

VOCs . Groundwater . Site assessment

Introduction

The increasing interest in the measurement of groundwatercontaminant mass fluxes (mass of contaminants passing perunit time per unit area) and contaminant mass discharges (sumof all mass fluxes across an entire plume control plane) showsthe need for in situ flux determination techniques (Goltz et al.2007; ITRC 2010; Newell et al. 2011; Triplett Kingston et al.2012). Especially the characterization and management oflarge-scale volatile organic carbon (CAH) groundwater con-taminations requires a thorough understanding of the contam-inant mass flux profiles. Contaminant flux profiles from sourcezones provide necessary information for the characterization ofthe source, while flux profiles from down-gradient contami-nant plumes allow (1) the verification of environmental risks

Responsible editor: Hailong Wang

G. Verreydt (*) : I. Van Keer : J. Bronders : L. DielsFlemish Institute for Technological Research (VITO),Boeretang 200,2400 Mol, Belgiume-mail: [email protected]

G. Verreydt : L. DielsUniversity of Antwerp, Antwerp, Belgium

G. Verreydt : P. VanderauweraArtesis University College of Antwerp, Antwerp, Belgium

M. D. AnnableUniversity of Florida, 2500 Southwest 2nd Avenue,Gainesville, FL 32607, USA

S. KaskassianBurgeap, 13 Rue du Parc,67205 Oberhausbergen, France

Environ Sci Pollut Res (2013) 20:4621–4634DOI 10.1007/s11356-012-1417-8

Author's personal copy

(Marsman et al. 2006; Caterina et al. 2009; Brouyère et al.2010; Verreydt et al. 2012), (2) the evaluation of the remedi-ation technology performance (Annable 2008; Brooks et al.2008; Brusseau et al. 2011), and (3) the assessment of naturalattenuation rates (Basu et al. 2006; SNRL 2006). To date, fluxprofiles are usually estimated from individual measurements/calculations of the Darcy water fluxes combined with thecontaminant concentrations. High uncertainty arises from suchindirect and often snapshot estimation of contaminant fluxes.The Passive Flux Meter (PFM) is the only in situ time-integrating passive sampling device for groundwater thatallows the simultaneous determination of contaminant massfluxes and Darcy groundwater fluxes (Hatfield et al. 2004,Verreydt et al. 2010). Segmentation within the PFM togetherwith the vertical superposition of several PFMs yields differ-entiated vertical flux profiles. Although considerable research,mainly in the USA, has been carried out with respect to theapplication of the PFMs (ITRC 2010), a lack of experience andconfidence still exists in Europe. Furthermore, all previousapplications and investigations have predominantly focusedon groundwater near source zones (Annable et al. 2005; Basuet al. 2006, 2009; Brooks et al. 2008; Klammler et al. 2012).The assessment of less contaminated plume zones with PFMsis still limited. Moreover, a PFM application under tidal influ-ence is completely new. This study aims to (1) demonstrate andevaluate the application of the PFM in Europe at a CAH-contaminated field site under tidal influence, (2) extend thefocus of the PFMs from near source zones to the plumeincluding investigating the possibilities for long-term monitor-ing with a PFM in the expected lower concentration contam-inated plume zones, and (3) provide valuable data for sourceand plume characterization in addition to traditional ground-water monitoring information. Theoretical and analytical aswell as technical and practical aspects will be discussed, to-gether with the challenges related to PFM implementation atthe study site.

Site description

The PFM is demonstrated and evaluated at a former indus-trial site near a tidal River in France. The site is contami-nated with petroleum hydrocarbons, chlorinated solvents,and sulfates. The PFM field campaign focused on the pri-mary pollutants perchloroethylene (PCE), trichloroethylene(TCE), and the degradation products cis+ trans-1,2-dichloroethylene (DCE) and vinyl chloride (VC). Multiplesource zones resulted from spills or leaks at a chemicalstorage facility between 1970 and 1988. The largest source(Fig. 1) resulted from an overturned rail car containing80 m3 of perchloroethylene and trichloroethylene. Thehydrogeology of the site is characterized by an alluvialaquifer consisting of red sand mixed with gravel and flint

with a mean hydraulic conductivity of 0.00231 m/s. Thealluvial aquifer ranges in thickness from 12 m to 2 m asapproaching the river (Fig. 2). The source zone (pure prod-uct+surrounding high contaminant concentration zone) cov-ers roughly 20 by 30 m with a depth of 8 m below groundsurface (bgs) into the silt and alluvial sand formations andfrom 16 to 23 m bgs into the sandy clay pseudo-substratumformation of the aquifer which ends 50 m bgs. Figure 3shows a vertical profile of the soil contamination in thesource zone. The dissolved CAH plume is approximately350 m long, 250 m wide, and is mainly captured in thesaturated alluvial aquifer. In 2009, the total volume of thesource zone is estimated at about 2,000 m3 and likely con-tains about 28 tons CAH (Burgeap, France). The total vol-ume of the contaminated groundwater is estimated at about350,000 m3. The source zone and the dissolved plume wererecently monitored by traditional sampling methods. Thesite is equipped with 69 monitoring wells, 60 of it wereinstalled during the PFM study period (November 2008–June 2010). In 2012, site investigations were finished andplans for remediation of the site are being developed.

Tidal effects are observed in the surficial aquifer due tothe proximity of the river. Water table fluctuations as well asshifting groundwater flow directions occur. The intensity ofthe effects increases near the river. The source zone issubject to daily water table fluctuations in the range of25 cm while the plume zone notes daily fluctuations up to1.5 m (Fig. 4). Groundwater modeling based on intensivehydraulic head monitoring yields an average Darcy ground-water flux through the hydraulic aquifer of 0.035 to 0.17 m/day, directed versus east-south-east.

Materials and methods

Passive Flux Meter (PFM) characteristics

The PFM is a passive sampling device that provides simul-taneous in situ point measurements of a time-averaged con-taminant mass flux, Jc, and water flux, q0 (Hatfield et al.2004; Annable et al. 2005; Basu et al. 2006). The device isplaced in a monitoring well or borehole where it interceptsthe groundwater flow and captures the contaminants. ThePFM consists of a permeable sock which is packed with apermeable sorbent matrix (Fig. 5). Each PFM has a length of140 cm and a diameter equal to the internal diameter of theselected monitoring well. Each PFM sock is divided into 30-cm-long segments to allow vertical flux differentiation. Toprevent vertical water flow within the PFM these segmentsare separated from each other using rubber washers. Acenter tube serves as a backbone for the PFM and facilitatesthe installation and retrieval allowing water bypass. Thesorbent matrix is impregnated with known amounts of one

4622 Environ Sci Pollut Res (2013) 20:4621–4634


or more water-soluble resident tracers. These tracers areleached from the PFM at rates proportional to groundwaterflux. The tracer retardation factor, Rd, is a measure of thedelay of a specific tracer as it passes through the sorbent. Itexpresses the ratio between the time that the tracer needs forpassing the sorbent phase versus the time it needs for pass-ing the same distance in the groundwater. Rd can be mea-sured using a column elution test (Hatfield et al. 2004;Appelo and Postma 2007). The selected sorbent matrix is agranular silver-impregnated activated carbon (Si-AC) char-acterized by a bulk density of 500 g/l, a mesh size of 5×30(2.36–0.60 mm), and a moisture content of 5 %. The selec-tion was based on batch and column sorption tests con-ducted on a range of possibly suitable sorbents for CAH inwater. Five tracer alcohols were selected for impregnatingthe activated carbon in order to cover different magnitudesof groundwater fluxes. The retardation function, i.e., the rateat which each tracer was eluted from the activated carbon,

was determined on a small laboratory column packed withthe tracer-equilibrated activated carbon used in the fieldtests. Retardation factors, Rd (–), of respectively 6 for meth-anol, 33 for ethanol, 145 for 2-propanol, and 266 for tert-butanol were derived from the executed column elutiontests. The internal tracer 2,4-dimethyl-3-pentanol (2,4-DMP) has a very high retardation and serves as a perfor-mance release compound (PRC) (Huckins et al. 2005) dur-ing the exposure. Table 1 summarizes the characteristics ofthe constructed PFMs.

Selection of monitoring wells for PFM installation

The monitoring wells for the PFM installation were careful-ly selected to (1) allow an easy PFM installation and retriev-ing, (2) allow a reliable convergence/divergence factordetermination, and (3) provide mapping of a vertical fluxdifferentiation in the source zone and a general downstream

N

Fig. 1 Map of the studied site with indication of the contamination

Environ Sci Pollut Res (2013) 20:4621–4634 4623


flux distribution in the plume zone. The selection criteria formonitoring wells eligible for PFM installation are listed inTable 2. Well and aquifer specifications should be known inorder to estimate the flow convergence/divergence factor(“Flow convergence or divergence around the monitoringwell and PFM”).

Seven monitoring wells along several transects down-gradient of the suspected source zone were selected, apartfrom one single deep well in the source zone. All selectedwells were recently installed and developed. The monitoringwell details are presented in Table 3.

PFM application

The PFMs were constructed in the laboratory and trans-ported to the field. During construction, a sample of sorbentmatrix was collected for the analysis of the initial tracerconcentrations (Fig. 6d). In order to monitor potential vol-atile losses of tracers during storage and transportation toand from the site, one PFM blank was transported to andfrom site and stored during deployment (Fig. 6b). AnotherPFM blank was also installed in a monitoring well andimmediately retrieved in order to map the installation/re-trieval effect on the flux measurements. PFMs were super-imposed in wells to cover the large well screen intervals.The completed PFM was then lowered into the well down to

the desired screen interval (Fig. 6a). Every PFM was at-tached to an individual stainless steel cable (Fig. 6c). After aspecified period of exposure to groundwater flow, the PFMs

Fig. 2 Geological cross-sectionof the study site

0,1 10 10000

5

10

15

20

25

30

0

5

10

15

20

25

300,1 10 1000

Dep

th b

elow

gro

und

surf

ace

(m b

gs)

CPCE Soil (mg/kg dry soil)

TCEPCE

Fig. 3 Vertical profile of soil contamination in Mw200 (source zone)



were retrieved. The PFMs in the source zone were exposedfor 3 weeks while the PFMs in the plume were exposed togroundwater flow for 9 weeks. Subsequently, the PFMsorbent was carefully extracted to quantify the mass of allintercepted contaminants and the residual masses of allresident tracers. The contaminant masses are used to calcu-late cumulative time-averaged contaminant mass fluxes,while residual resident tracer masses are used to calculatecumulative time-averaged groundwater fluxes (Hatfield etal. 2004).

Sorbent Analysis

Tracers and contaminants were extracted from the sor-bent in two extraction steps: a first step involved anextraction with isobutyl alcohol (IBA), subsequently anextraction with an acetone/hexane (AH) mixture fol-lowed. From the IBA extract, all alcohols, including2,4-DMP, were analyzed using an Interscience Trace

Gas Chromatograph equipped with an automated liquidinjection and flame ionization detector. The CAHs wereanalyzed from the AH extract using an InterscienceTrace Gas Chromatograph equipped with an automatedliquid injection and a mass spectrograph (MS). Extrac-tion efficiencies measured in laboratory batch tests were52±4 % for PCE, 94±4 % for TCE, 82±8 % for DCE,44±6 % for VC, and 100±6.5 % for the selectedalcohol tracers. More analytical details are provided inAnnable et al. (2005).

Theory

Flux calculations

A contaminant mass flux can be defined as the total amountof contaminant, expressed as mass, passing per unit time perunit area through a control plane that is orthogonal to the

Fig. 4 Groundwater levelmonitoring data from a well inthe CAH source and plumezone near the river bank

10 cm

Fig. 5 Concept of a PFM



mean groundwater flow direction (Bear 1988; Bear andVerruijt 1998; Basmadjian 2004; Newman et al. 2005).

Jc ¼ C:q0 ¼ m

A:tð1Þ

where Jc is the contaminant mass flux (g/m2/day), C is themean concentration of the contaminant in the groundwater(g/m3), q0 is the Darcy groundwater flux (m3/m2/day), m isthe mass of contaminant (g), A is the area of a control planeorthogonal to the groundwater flow direction (m2), and t isthe time (day).

A PFM measures local contaminant mass and Darcywater fluxes, bounded by the length and diameter of thePFM section. A PFM, installed along a control plane andexposed to a contaminated groundwater flux, will adsorbcontaminants from the groundwater while tracer elutes fromthe sorbent. If an irreversible complete contaminant sorptionon the activated carbon is assumed, i.e., for highcontaminant-sorbent retardation factors and/or short

exposure periods, the contaminant mass flux (Jc) can becalculated as (Hatfield et al. 2004):

Jc ¼ mc

2:a:r:b:tð2Þ

where mc is the mass of contaminant sorbed by the PFM (g),α is the convergence or divergence factor of flow around thePFM (–), r is the radius of the PFM cylinder (m), and b is thelength of the sorbent matrix or the vertical thickness of theaquifer interval sampled (m).

The Darcy water flux (q0) through the aquifer can bedetermined by:

q0 ¼ 1:67:r:Rd:ð1�4RÞa:t

ð3Þ

where Rd is the retardation coefficient of the resident traceron the sorbent (–) and ΩR is the relative mass of the residenttracer remaining in the PFM sorbent at the particular welldepth (–). More details are provided in Hatfield et al. (2004).

Consequently flux averaged concentrations (CJ) can be cal-culated based on the measured contaminant and water fluxes:

CJ ¼ Jcq0

ð4Þ

Flow convergence or divergence around the monitoring welland PFM

The streamlines around the PFM will deviate because of thehigher permeability of the device towards the permeabilityof the aquifer. This also applies for the streamlines aroundthe monitoring well and the gravel filter around the wellfilter (Basu et al. 2006; Börke 2007). The water flux (q)through the PFM is directly proportional to the water flux(q0) in the aquifer. This can be expressed in:

q ¼ a:q0 ð5Þwhere α is the convergence or divergence of the groundwa-ter flow in the vicinity of a monitoring well or passive

Table 1 PFM characteristics

Characteristic Parameter Value/description

Dimensions Length 140 cm

Diameter 80 mm/103 mm

Weight 4.5 kg/6.5 kg

Sections Amount 4–5

Length 25–30 cm

Material Inner tube HDPE

Fixing ring and cable Stainless steel

Sock 100 % polyester

Outer mesh Nylon

Sorbent Material Silver-impregnatedactivated carbon

Mesh size 5 × 30

Hydraulic conductivity 0.00386 m/s

Tracers Methanol, ethanol, 2-propanol,tert-butanol, and2,4-dimethyl-3-pentanol

Table 2 Monitoring well selection criteria

Type Parameter Specifications Comments

Well specific Internal diameter 25–104 mm Optimal diameter is 41 mm

Depth 1.5–50 m Depends on well filter location, maximum installationdepth depends on PFM weight

Filter slit >0.3 mm Filter permeability preferably >6× aquifer permeability

Gravel filter Gravel permeability≥aquifer permeabilityfilter thickness<1/8 well diameter

Flow bypass through the gravel filter has to be taken into account

Well head Undamaged, straight PFM slides directly from transport tube into well

Installation date Recent Clogging and biofouling will influence the flux measurement

Site specific Location Source/plume one well/severalwells along control plane

Exact location/selection of wells is based onexisting traditional groundwater sampling data



sampler. α can be calculated from the potential theory(Drost et al. 1968; Klammler et al. 2007), based on thehydraulic conductivities of the aquifer, the gravel filtersurrounding the monitoring well, the well filter screen, andthe PFM. These hydraulic conductivities can be determinedin the laboratory. The potential theory supposes a constant αfactor and horizontal streamlines through the monitoringwell and the PFM. Vertical short circuit currents may havea significant influence on the convergence/divergence of thestreamlines.

Results

Convergence/divergence factor α

Theα-factor estimates for the present study are summarized inTable 4. All calculations were made through the potentialtheory as presented by Klammler et al. (2007). The hydraulicconductivities (k values) of the gravel filters (included inTable 4) and the PFM (kS=0.386 cm/s) were measured with

a permeameter (Klute 1986; Eijkelkamp 2008). The hydraulicconductivities of the well screens were estimated based onborehole dilution tests in a laboratory setup (Lamontagne et al.2002; Massey et al. 2007). According to in situ boreholeimpeller-type flowmeter measurements, the hydraulic conduc-tivities of the surrounding aquifer materials were estimated tobe 0.045–0.27 cm/s for well Mw200 and 0.0045–0.85 cm/sfor the remaining wells (Kaskassian et al. 2012).

PFM blanks

Two PFM blanks were added to the test series. A firsttransport blank, which consists of four sections, shows nosignificant overall losses of tracers nor contaminant loadingduring storage and transportation. One end section shows aminor discrepancy, resulting in higher coefficients of varia-tion (CV) between PFM sections. The second PFM blank, adeployment blank, indicates significant tracer loss and con-taminant loading on all five sections (Table 5). The deploy-ment PFM blank was installed in Mw200 (well in sourcezone), down to 25 m bgs, and immediately retrieved. In

Table 3 Monitoring well details

Well ID Installation date Location Diameter (mm) Screen interval (m bgs) Slit size (mm) Filter grain size (mm)

200 08/06/2009 Source 103 4–22.5 0.3 0.75

121 03/06/2009 Plume 80 12.5–25.5 1 2/3

132 09/07/2009 Plume 103 12.5–19 0.3 0.75

134 16/07/2009 Plume 80 11.8–20.5 0.5 0.6/1.2

137 22/07/2009 Plume 80 10–17.5 0.5 0.6/1.2

135 31/08/2009 Plume 80 14–20 0.5 0.6/1.2

122 03/06/2009 Plume 80 15–25.5 1 2/3

138 01/09/2009 Plume 80 12–17.5 0.5 0.6/1.2

Fig. 6 PFM application in thefield. a Installation of a PFM inwell 200 by pushing it downwith Geoprobe® rods; b threePFMs in their transport tubes; ctop view of well 200, equippedwith 14 PFMs, maximum threecables per cap hole; and dsubsampling of the Si-AC froma PFM in a 20-mL vial withisobutyl alcohol



total, approximately 279 L of groundwater was pushedthrough the center tube of the PFM blank during deploy-ment/extraction. The tracer loss of the deployment blankcorresponds to a Darcy velocity of 1.4 cm/day (Eq. 3, t =1 day). The water deployment/extraction volume of an80 mm PFM (Table 1) amounts to 8 L per meter deploymentdepth below groundwater level. A 103-mm PFM requires11 L of water displacement per meter deployment depthbelow groundwater level. Consequently, Darcy water fluxdeployment errors per meter groundwater deployment depthcan be calculated for both PFM types (Eq. 3, t = 63 days andrespectively 21 days). For an 80-mm PFM, it is estimated tobe 0.048 cm/day/m groundwater deployment depth; for a103-mm PFM, it is 0.067 cm/day/m groundwater deploy-ment depth. The contaminant loading of the deploymentblank respectively corresponds to calculated contaminantmass fluxes of respectively 37 mg PCE/m2/day, 35 mgTCE/m2/day, 200 mg DCE/m2/day, and 1.9 mg VC/m2/day.

Groundwater and contaminant fluxes

Figure 7a–c visualizes the PFM groundwater flux (q0) andcontaminant flux (JPCE, JTCE, JDCE, and JVC) data for threemonitoring wells with different filter screen lengths and atdifferent distances from the river (Mw200, Mw121, andMw135; see Fig. 1). The measured Darcy groundwaterfluxes range from 2 to 41 cm/day. The contaminant massfluxes vary between 0 and respectively 13,000 mg/m2/dayfor PCE, 12,000 mg/m2/day for TCE, 6,200 mg/m2/day forDCE, and 58 mg/m2/day for VC. According to Fig. 7b–c,high contaminant fluxes are related to high groundwaterfluxes. In Mw200, however, the TCE flux peak at a depthof 12 m bgs is not related to an increased Darcy water flux(Fig. 7a) and thus may indicate the presence of a TCE pureproduct spot. A significantly increased contaminant fluxvalue that does not correspond to an increased Darcy waterflux may indicate an upstream source. Since DCE and VC,both degradation products of PCE and TCE, show signifi-cant fluxes in source as well as in plume zone, the process ofnatural attenuation of the chlorinated solvent pollution hasbeen proven to be started. Although ethylene productionwas observed up to 40 μg/L within the source zone, acomplete degradation up to ethane has not been proven yet.

PFM data versus traditional monitoring data

Concentration data

The traditional groundwater samples were taken with asubmersible pump intake located in the well filter screen.Only PFM measurements from the same depth as the sub-mersible pump were used for comparison. The PFM fluxdata were converted into flux averaged concentrations (CJ,see eq. 4) in order to provide comparative data. Traditionalpumping and PFM flux averaged concentrations are bothshown in Fig. 8a–d for all eight selected wells.

The PFM and traditional concentration data are of similarmagnitude for most wells. Mw137 draws the largest aber-rance with respect to DCE. Mw200 also notes a magnitudedifference for PCE and TCE. The PFM flux averaged con-centrations are generally higher than the traditional concen-trations. For monitoring well Mw200 in the source zone,flux averaged concentration data can also be plotted versusgroundwater concentration data gathered by Multi PressureLevel Sampling (MPLS) (Nilsson et al. 1995; Schirmer et al.1995). Figure 9 shows the vertical profiles of the PCE, TCE,DCE, and VC concentrations in the groundwater surround-ing Mw200. In the soil phase (Fig. 3), two major PCE/TCE

Table 4 α-Factor estimates

Well diameter inner/outer (mm) Borehole diameter(mm)

Slit size (mm) Grain size gravelfilter (mm)

k value wellscreen (m/s)

k value gravelfilter (cm/s)

α factor (–)

103/114 178 0.3 0.75 0.09 0.422 0.42–0.84

103/114 140 0.3 0.75 0.09 0.422 0.32–0.73

80/90 140 1 1–3 0.30 1.53 0.24–0.52

80/90 140 0.5 0.6–1.2 0.15 0.462 0.26–0.51

Table 5 Results of PFM blanks

Compound Deployment blank Transport blank

Rel. conca (–) CVb Rel. conca (–) CVb

Methanol 0.38 0.13 1.06 0.17

Ethanol 0.85 0.03 1.03 0.09

2-Propanol 1.01 0.02 1.01 0.11

Tert-butanol 1.04 0.03 0.94 0.09

Conc. (μg/gAC) CVb Conc. (μg/gAC) CVb

PCE 31 0.25 0.00 0.00

TCE 30 0.22 0.00 0.00

DCE 167 0.58 0.00 0.00

VC 2 0.27 0.00 0.00

a Relative concentration, i.e., tracer mass on Si-AC after exposureversus initial tracer mass on Si-ACb Coefficient of variation between the PFM sections



Fig. 7 PFM flux data for monitoring well. a Mw200 (source), b Mw121 (plume), and c Mw135 (plume)



sources can be distinguished. The first zone is spread fromground level to about 10 m bgs; the second zone reaches

from 15 to at least 22 m bgs. Both source zones are reflectedin the groundwater concentration data of PFMs (Fig. 9). The

Fig. 9 Groundwater (GW) concentration data for monitoring well Mw200

a b

c d

Fig. 8 Comparison between local PFM flux averaged and traditional concentrations for a perchloroethylene, b trichloroethylene, c dichloro-ethylene, and d vinylchloride



PFM data also reveal a higher concentration zone from 13 toabout 16 m bgs. The MPLS groundwater datasets for PCEand TCE confirm the likely presence of the first majorsource zone in the soil phase. Increased concentrations ofthe daughter products DCE and VC are also recorded in thedeeper MPLS data. The PFM concentration dataset addi-tionally reveals a thin, high-concentration zone for TCE andVC at 11 m bgl, which is not noticed by MPLS or soil coresampling.

It is very difficult to compare concentration data of dif-ferent groundwater monitoring methods. Only a generalcomparison can be made. The sampling methods produceintrinsically different datasets. Table 6 highlights the mostimportant differences between PFM, traditional groundwa-ter, and MPLS sampling.

Darcy fluxes and hydraulic conductivities

Darcy’s law (Darcy 1856) defines the linear relationshipbetween the Darcy groundwater flux (q0) and the hydraulicconductivity (k) of the aquifer material. Figure 10 presentsboth q0 and k for four monitoring wells at different distancesfrom the tidal river. The Darcy groundwater flux is deter-mined using the PFMs while the hydraulic conductivity dataare obtained from in situ borehole impeller-type flowmetermeasurements (Kaskassian et al. 2012). The location of thealluvial aquifer is indicated at the figures.

Overall, the q0 and k data series show good similarities.The location of the very permeable alluvial aquifer (redalluvial sand at Fig. 2) is reflected in the groundwater fluxgraphs and the hydraulic conductivity graphs by means ofincreased q0 and k values. The tidal effect of the river on the

surrounding aquifer (Fig. 4) is not clearly noticed in thePFM water flux data.

Discussion

The present study evaluates the PFM with respect to appli-cability (source versus plume zone) and to technical as wellas practical aspects. Forty-eight PFMs contributed to thepresented Darcy groundwater and contaminant mass fluxprofiles for the eight selected monitoring wells.

The installation of a PFM requires the alignment of thetransport tube on top of the monitoring well. The non-horizontal top ends of the monitoring well in combinationwith the presence of aboveground metal well protectiontubes complicated the PFM installation. Finally, one out of49 PFMs were damaged during deployment and could notbe used. All PFMs were attached to individual cables.Because a 100-mm-diameter PFM of 1.40 m length is surelyheavy (Table 1), retrieving the 14 PFMs from Mw200 waschallenging. Smaller monitoring wells and thus less heavyPFMs are preferred. The sampling and analysis of the PFMsorbents went perfectly according to the method mentionedin “Sorbent analysis”.

A PFM requires a monitoring well in “good status”. Thismeans a monitoring well with a proper surrounding gravelfilter and open filter slits (<0.3 mm) since clogging andbiofouling influence the flux measurement. Therefore, arecently installed monitoring well is preferred. The optimalinternal diameter of a well for PFM application rangesbetween 41 and 80 mm. The well top needs a horizontal,undamaged cut-off since the PFM slides directly from a

Table 6 Comparison between PFM, traditional groundwater, and MPLS sampling

Factor PFM Traditional pumping MPLS

Sampling force Advection by natural groundwaterflow+adsorption

Submersible pump Mini pressure pump

Monitoring well opening Completely filled with PFM(s) Open well hole with pump Completely filled with MPLS liner

Sampling time 3–9 weeks Snapshot Snapshot

Sampling area All groundwater passing PFM byadvection during sampling

Withdrawal zone of the pump,surrounding the well filter,mainly from the most permeableaquifer that crosses the well filter

Groundwater from gravel filter,surrounding the monitoringwell

Sampled volume ofgroundwater

20–28.000 L 20 mL (+ 3× well volume for purgingmonitoring well)

20 mL (+ purging of MPLStube)

Sampling speed = Groundwater flow rate ca. 200 mL/min ca. 5 mL/min

Volatilization duringsampling/analysis

None Possible Minor

Output data Contaminant flux Snapshot concentration Groundwater fluxGroundwater flux

Flux averaged concentration

Representativeness ofdata for aquifer

Very good, with detailed verticaldifferentiation

Average, no vertical differentiation Average, with verticaldifferentiation



transport tube into the well. One of the most importantissues concerning the application of the PFM is the devia-tion of the groundwater flow paths when passing a monitor-ing well equipped with a PFM, expressed as the α-factor(convergence/divergence) (“Flow convergence or diver-gence around the monitoring well and PFM”). The charac-teristics of the well and the gravel filter (diameter, length,and hydraulic conductivity) should be known in order tocalculate the flow convergence or divergence through wellfilter and PFM.

Both types of 80-mm-diameter wells at the study sitepossess very low α values (0.24 and 0.26) due to a largegroundwater bypass through their large surrounding gravelfilters. The 80-mm wells were installed in 140-mm-diameterboreholes. Low α factors increase the errors on the calcu-lated Darcy groundwater fluxes (cf. Eq. 3). Local differ-ences in filter pack thicknesses as well as clogging of wellscreens will also result in different α values. Thus, theoret-ical α calculations may differ from the field situation. Morepractical research on this issue is needed. However, the flux-averaged concentration is unaffected by the α factor.

The analytical results show that a PFM can be installednear source as well as in plume zones. In this respect, theexposure time is the determining factor together with thegroundwater flow rate. Saturation of the sorbent needs to beavoided and at least one tracer alcohol must still be presentat the sorbent phase at greater than 32 % of its initialconcentration to enable a correct water flux calculation(Hatfield et al. 2004). The analysis of the PFM itself com-prises two extraction steps as mentioned in “Sorbent analy-sis”. Recovery percentages of 94±6 % for TCE and 100±

6.5 % for alcohol tracers prove the proper analytical proto-col. However, the analysis is a time- and solvent-consumingprocess.

An important analytical issue concerns the significanttracer loss and contaminant loading of the deploymentblank. The tracer loss corresponds to 12 % of the averagemeasured Darcy water flux in MW200 on this study sitewhile the contaminant loading draws up to about 7–15 % ofthe measured contaminant flux in the same monitoring well.Deployment error percentages can be taken into account.However, the influence of the deployment/extraction alsodepends on the characteristics of the monitoring well filterand the surrounding gravel filter. Contaminant loading isparticularly a source zone problem when shorter PFM ex-posure times are used and higher contaminant amounts arepresent in the groundwater. Therefore, “cleaning” the wellfirst by lowering a PFM dummy without washers allowingvertical water flow may solve this problem in source zones.Tracers with higher retardation factors (e.g., isopropanol andtert-butanol) and larger PFM exposure times may solve thetracer loss problem if combined with higher PFM exposuretimes (Bondehagen 2010).

The complexity of the site, especially due to the tidaleffects of the nearby river, unfortunately complicates theevaluation of the PFM results. Tidal water level fluctuationswere observed in the monitoring wells during the PFMexposures (Fig. 4) and may cause slight vertical shifts ofthe flux profiles. However, they were not clearly noticed inthe PFM flux data. Insufficient simultaneous piezometrichydraulic heads were available for the theoretical calcula-tion of all local Darcy water fluxes based on the hydraulic

Fig. 10 Darcy velocity and hydraulic conductivity data for wells Mw121, 138, 135, and 122



aquifer conductivities. Nevertheless, PFM flux averagedconcentration data and traditional concentration data are ofsimilar magnitude for most wells. It should be noted thatboth datasets need to be compared with reservation becauseof the different sampling nature and time (summarized inTable 6). Active groundwater pumping from the middle ofthe filter screen usually samples the nearest and most per-meable aquifer layers. MPLS groundwater concentrationdata should be more comparable with the PFM flux aver-aged concentration data because of their vertical differenti-ation and low pressure purging. Only a rudimentarycorrelation can be made between the PFM and MPLS datadue to the snapshot character of the MPLS. Finally, the PFMflux data can be combined with additional monitoring datato refine the groundwater pollution map for this site.

Conclusions

For the first time in Europe, the application of the PFM wasdemonstrated for site assessment at a CAH-contaminatedfield site. The focus of the PFM was extended from nearsource to plume zones. PFM evaluation criteria includeapplicability and analytical, technical as well as practicalaspects. Darcy groundwater flux and contaminant mass fluxprofiles were compared with more traditional groundwaterdata. PFM installation and transport blanks were included inthe study. Application of the PFM in Europe has been verysuccessful. Overall, the PFM yields representative flux datafor a VOC source as well as a plume zone. Groundwaterplume zones can be sampled with the current PFM config-uration as long as the PFM exposure time is extended. Forlong-term PFM applications (>2.5 months), a mix of inter-nal tracers with a higher retardation factor (>150) and lowbiodegradability is recommended. The most importantadvantages of the PFM technology can be summarized asfollows: (1) the simultaneous evaluation of both water andcontaminant fluxes under natural gradient conditions; (2)less sensitivity to daily fluctuations in groundwater flow orcontaminant concentrations because of the cumulative mea-surement over longs contact periods (3 to 9 weeks); (3) thevery large and detailed dataset including vertical differenti-ation in measured horizontal fluxes; (4) easy to apply in astandard monitoring well without the need of pumping orelectrical power; and (5) minimal changes in the hydraulicflow field, mass loss through volatilization, or cross-contamination.

There are many challenges in PFM implementation at afield site. Tracer loss during the installation and deviation ofthe groundwater flow field through the monitoring well andPFM are the two most important points of concern. How-ever, the acquired flux data are of great value. The PFMresults included in the present paper forms the basis for a

complete site characterization study including mass dis-charge modeling.

Acknowledgments This study is part of the PhD research of GoedeleVerreydt and funded through the Flemish Institute for TechnologicalResearch (VITO NV). Dr. Hansjörg Weiss (Innovative Messtechnik Dr.Weiss) is gratefully acknowledged for his free provision of the MPLStechnology.

References

Annable, M.D. 2008. Mass flux as a remedial performance metric atNAPL contaminated sites. In: Methods and techniques for clean-ing up contaminated sites. NATO Science for Peace and SecuritySeries C—Environmental Security, 177–186

Annable MD, Hatfield K, Cho J, Klammler H, Parker BL, Cherry JA,Rao PSC (2005) Field-scale evaluation of the passive flux meterfor simultaneous measurement of groundwater and contaminantfluxes. Environ Sci Technol 39(18):7194–7201

Appelo CAJ, Postma D (2007) Geochemistry, groundwater and pollu-tion, Secondth edn. Balkema, Amsterdam, Reprint withcorrections

Basmadjian D (2004) Mass transfer: principles and applications. CRC,Boca Raton

Basu NB, Rao PSC, Poyer IC, Annable MD, Hatfield K (2006) Flux-based assessment at a manufacturing site contaminated with tri-chloroethylene. J Contam Hydrol 86:105–127

Basu NB, Rao S, Poyer I, Nandy S, Mallavarapu M, Naidu R, Davis G,Patterson B, Annable MD, Hatfield K (2009) Integration of tra-ditional and innovative characterization techniques for flux-basedassessment of dense non-aqueous phase liquid (DNAPL) sites. JContam Hydrol 105:161–72

Bear J (1988) Dynamics of fluids in porous media. Dover edition.Originally published: New York: American Elsevier, 1972. Re-print with corrections. New York: Dover

Bear J, Verruijt A (1998) Modeling groundwater flow and pollution.First published 1987. Reprinted with corrections. Reidel,Dordrecht

Bondehagen D (2010) Measuring groundwater and contaminant flux:passive flux meter field applications and issues with alcoholdegradability. Air Soil Water Res 3:23–35

Börke P (2007) Untersuchungen zur Quantifizierung der Grundwasser-immisson von Polyzyklischen Aromatischen Kohlenwasserstof-fen mithilfe von Passiven Probennahmesystemen. Ph.D. Diss.,Technische Universität Dresden, Dresden, Germany (In German)

Brooks MC, Wood AL, Annable MD, Hatfield K, Cho J, Holbert C,Rao RSC, Enfield CG, Lynch K, Smith RE (2008) Changes incontaminant mass discharge from DNAPL source mass depletion:evaluation at two field sites. J Contam Hydrol 102(1–2):140–153

Brouyère S, Jamin P, Dolle F, Chisala B, Orban PH, Popescu IC,Hérivaux C, Dassargues A (2010) A regional flux-based riskassessment approach of contaminated sites on groundwater bod-ies. GQ10: Groundwater Quality Management in a RapidlyChanging World (Proc. 7th International Groundwater QualityConference held in Zurich, Switzerland, 13–18 June 2010)

Brusseau ML, Carroll KC, Allen T, Baker J, DiGuiseppi W, Hatton J,Morrison C, Russo A, Berkompas J (2011) Impact of in situchemical oxidation on contaminant mass discharge: linkingsource-zone and plume-scale characterizations of remediationperformance. Environ Sci Technol 45(12):5352–5358

Caterina D, Brouyère S, Jamin P, Batlle-Aguilar J, Dassargues A,Dejonghe W, Diels L, Crèvecoeur S, Thomé JP, Dujardin J,



Batelaan O, Canters F, Hérivaux C (2009) Flux-based risk assess-ment of the impact of contaminants on water resources andecosystems “FRAC-WECO”. Final Report. Brussels: BelgianScience Policy 2009—59 p. (Research Programme Science for aSustainable Development)

Darcy H (1856) Détermination des lois d’écoulement de l’eau à traversle sable. pp. 590–594. In: Les Fontaines Publiques de la Ville deDijon. Dalmont, Paris

Drost W, Klotz D, Koch A, Moser J, Neumaier F, Rauert W (1968)Point dilution method measuring groundwater flow by means ofradioisotopes. Water Resour Res 4:125–146

Goltz MN, Kim S, Yoon H, Park J (2007) Review of groundwatercontaminant mass flux measurement. Environ Eng Res 12:176–193

Eijkelkamp (2008) Operating Instructions 09.02 Laboratory permea-meter. M1.09.02.E

Hatfield K, Annable M, Cho J, Rao PSC, Klammler H (2004) A directpassive method for measuring water and contaminant fluxes inporous media. J Contam Hydrol 75:155–181

Huckins JN, K Booij, WL Cranor, DA Alvarez, RW Gale, BartkowME, Robertson GL, Clark RC, Stewart RE (2005) Fundamentalsof the use of performance reference compounds (PRCs) in passivesamplers. In: Proceedings of SETAC North America 26th AnnualMeeting, November 13–17, 2005, 1–24. Pensacola, USA: SETAC

ITRC (Interstate Technology & Regulatory Council) (2010) Use andmeasurement of mass flux and mass discharge. MASSFLUX-1.Washington, D.C.: Interstate Technology & Regulatory Council,Integrated DNAPL Site Strategy Team. http://www.itrcweb.org

Kaskassian S, Gleize T, Chastanet J, Côme JM (2012). ProjetATTENA 2—Module 3 Démonstration, Tâche 3.1.2 : Mise enœuvre du protocole MACAOH par BURGEAP sur le Site 1bis(solvants chlorés). Rapport Final confidentiel ADEME (ADEMEconfidential final report for project ATTENA 2) (in French)

Klammler H, Hatfield K, Annable MD, Agyei E, Parker BL, CherryJA, Rao PSC (2007) General analytical treatment of the flow fieldrelevant to the interpretation of passive fluxmeter measurements.Water Resour Res 43. doi:10.1029/2005WR004718,W04407

Klammler H, Hatfield K, Luz J, Annable M, Newman M, Cho J,Peacock A, Stucker V, Ranville J, Cabaniss S, Rao PSC (2012)Contaminant discharge and uncertainty estimates from passiveflux meter measurements. Water Resour Res 48, W02512,doi:10.1029/2011WR010535

Klute A (1986) Methods of soil analysis, part 1: physical and miner-alogical methods. SSSA Book Series: 5

Lamontagne S, Dighton J, Ullman W (2002) Estimates of groundwatervelocity in riparian zones using point dilution tests. CISRO Tech-nical Report 14/02, pp. 16

Marsman A, Valstar J, ter Meer J (2006) Risk analysis of groundwatercontamination at the megasite port of Rotterdam. (Proceedings ofModelCARE’2005, held at The Hague, The Netherlands, June2005) IAHS Publication 304, pp. 309

Massey AJ, Friesz PJ, Carlson CS (2007) Borehole-dilution tests tomeasure natural flushing of screens in wells monitored withpassive diffusion sampling. USGS. Geological Society of AmericaAbstracts with Programs 39(1):101

Newell CJ, Farhat SK, Adamson DT, Looney BB (2011) Contaminantplume classification system based on mass discharge. GroundWater 46(6):914–919

Newman M, Hatfield K, Hayworth J, Rao PSC, Stauffer T (2005) Ahybrid method for inverse characterization of subsurface contam-inant flux. J Contam Hydrol 81:34–62

NilssonB, Luckner L, SchirmerM (1995) Field trials of active andmultiportsock samplers in gravel-packed wells. J Hydrol 171:259–289

Schirmer M, Jones I, Teutch G, Lerner DN (1995) Developmentand testing of multi sock samplers for groundwater. J Hydrol171:239–257

SNRL (Savannah River National Laboratory) (2006) Passive flux-meters: application as a characterization/monitoring tool for mon-itored natural attenuation of chlorinated solvents. WashingtonSavannah River Company, Washington D.C., WSRC-STI-2006-00315 http://www.osti.gov

Triplett Kingston JL, Dahlen PR, Johnson PC (2012) Assessment ofgroundwater quality improvements and mass discharge reduc-tions at five in situ electrical resistance heating remediationsites. Ground Water Monit Remediat. doi:10.1111/j.1745-6592.2011.01389.x

Verreydt G, Bronders J, Van Keer I, Diels L, Vanderauwera P (2010)Passive samplers for monitoring VOCs in groundwater and theprospects related to mass flux measurements. Ground WaterMonit Remediat 30(2):114–126

Verreydt G, Van Keer I, Bronders J, Diels L, Vanderauwera P (2012)Flux-based risk management strategy of groundwater pollutions:the CMF approach. Environ Geochem Health. Special issue onSEGH2011. doi: 10.1007/s10653-012-9491-x



http://www.itrcweb.org

http://dx.doi.org/10.1029/2005WR004718,W04407

http://dx.doi.org/10.1029/2011

http://www.osti.gov

http://dx.doi.org/10.1111/j.1745-6592.2011.01389.x

http://dx.doi.org/10.1111/j.1745-6592.2011.01389.x

http://dx.doi.org/10.1007/s10653-012-9491-x

Documents

How does contamination move in groundwater? | iFLUX · flux data raise up to 13 g/m2/day for perchloroethylene in the plume zone. Calculated PFM flux averaged concentra-tion data