Benzene Removal In Horizontal Subsurface Flow Constructed Wetlands Treatment

Abstract

Although much information is available about conventional water quality parameters in subsurface flow (SSF) constructed wetlands, few data are available regarding specific contaminants. In this paper we focus on the behavior of Benzene in three types of constructed wetlands (CWs) (two of them were planted with two different species of macrophytes -Phragmitesaustralis and Typhalatifolia-and the third was used as a control -unplanted), located in Southern Italy (Apulia region, Lecce). The objectives of this study are to compare hydraulic behavior of the CWs with the trend of the model by varying the hydraulic conditions, to evaluate the effect of the clogging and then to assess the efficiency of the different species of macrophytes in removing Benz ene. At the beginning of the experience and after 24 months, the results show a good correlation in the hydraulic behavior between model and physical data by modifying input parameters as a consequence of the clogging. The Benzene removal from the planted fields is higher than the unplanted one.

Full text

International Journal of Applied Engineering Research
ISSN 0973-4562 Volume 10, Number 6 (2015) pp. 14603-14614
© Research India Publications
http://www.ripublication.com
Benzene Removal In Horizontal Subsurface Flow
Constructed Wetlands Treatment
Ezio Ranieria, Angela Gorgoglionea*, Andrea Petrellaa, Valentina Petruzzellia,
PetrosGikasb
aDICATECh, Polytechnic University of Bari, viaOrabona 4, Bari70125, Italy;
E-mails: ezio.ranieri@poliba.it; andrea.petrella@poliba.it;
valentina.petruzzelli@poliba.it
bDepartment of Environmental Engineering, Technical University of Crete, Chania
campus, Chania74100, Greece;
E-mail: petros.gikas@enveng.tuc.gr
*Corresponding author: Tel. +39 3356781687; Fax: +39 0805481395; email:
angela.gorgoglione@poliba.it
Abstract
Although much information is available about conventional water quality
parameters in subsurface flow (SSF) constructed wetlands, few data are
available regarding specific contaminants. In this paper we focus on the
behavior of Benzene in three types of constructed wetlands (CWs) (two of
them were planted with two different species of macrophytes –
Phragmitesaustralis and Typhalatifolia- and the third was used as a control -
unplanted), located in Southern Italy (Apulia region, Lecce). The objectives of
this study are to compare hydraulic behavior of the CWs with the trend of the
model by varying the hydraulic conditions, to evaluate the effect of the
clogging and then to assess the efficiency of the different species of
macrophytes in removing Benzene. At the beginning of the experience and
after 24 months, the results show a good correlation in the hydraulic behavior
between model and physical data by modifying input parameters as a
consequence of the clogging. The Benzene removal from the planted fields is
higher than the unplanted one.
Keywords:Benzene, Constructed wetlands, Phragmitesaustralis, Typhalatifolia
Introduction
Benzene, toluene, xylene, phenol, halogenated aromatic compounds, chloroform and
trichloroethylene are the major products of the petroleum and fine chemical industries
and the most frequently used organic solvents (Yeom and Yoo, 1999). Exploration,

14604 Ezio Ranieri
production, refining, storage, transportation, distribution and utilization of petroleum
hydrocarbons, accidental spills, improper practices and leaching landfills resulted in
the frequent occurrence of these anthropogenic organic compounds in air, water and
soil (Tang et al., 2009). Losing these substances to the receiving environments may
lead to an adverse impact and might endanger public health and welfare. Therefore,
considerable research has been conducted to remove these compounds from
contaminated environments (Nickelsen and Cooper, 1992; Lu et al., 2002).
In recent years, an increasing number of full-scale constructed wetlands (CWs) have
been put into operation (Garcìa, 2004; Puigagut at al., 2007). This rise in the
implementation of such treatment systems is because CWs have several advantages
over conventional wastewater treatment systems, particularly in small villages (≤
2000 population equivalent): staff do not require specific training, the operation and
maintenance costs are lower, and they integrate well into the surrounding landscape
(Vymazal, 2005; Rousseau et al., 2008; Pedescoll et al., 2009;Raboni et al., 2014).
Constructed wetlands are designed to use natural wetland processes that are
associated with wetland hydrology, soils, microbes and plants to treat wastewater
(Lee et al., 2005; Tang at al., 2009). CWs are artificial wastewater treatment systems
consisting of shallow (usually less than 1 m deep) ponds or channels which have been
planted with aquatic plants and which rely upon natural microbial, biological,
physical, and chemical processes to treat wastewater. They typically have impervious
clay or synthetic liners, and engineered structures to control the flow direction, liquid
detention time, and water level. Depending on the type of system, they may or may
not contain an inert porous media such as rock, gravel, or sand (Ranieri et al., 2014).
CWs can also remove Benzene from wastewater. This topic has been studied
increasingly in recent years (Eke and Scholz, 2008; Ranieri et al., 2013a).Benzene is a
hydrocarbon that occurs as a volatile liquid that is capable of rapidly evaporating,
colorless and highly flammable. In the atmosphere the most significant source of
benzene is represented by vehicular traffic, mainly from the exhaust gases of gasoline
powered vehicles. Therefore the highest concentrations are found in the proximity of
areas of intense traffic and large parking lots.
Otherwise hydraulic considerations have a significant role in prediction of the actual
removal percentages for every contaminant. This study assesses and elaborates the
hydraulic performance in the pilot-scale horizontal subsurface flow constructed
wetlands (HSFCWs) and observes trends over time. Design parameters such as aspect
ratio, size of the porous media, and hydraulic loading rate can improve the hydraulic
behavior of constructed wetland systems by imparting a hydraulic flow behavior that
approaches that of an ideal flow system (Ranieri, 2012; Ranieri and Young, 2012).
The experiments were conducted using tracer tests (KBr), which provided the
residence time distribution (RTD). Particularly, after 24 months of operating,
clogging conditions in experimental HSFCWs result in a lower hydraulic conductivity
values (Ranieri et al., 2013b).
The objectives of this study are: 1) to evaluate the hydraulic behavior of constructed
wetlands not planted and planted with different species (Phragmitesaustralis and
Typhalatifolia), by varying hydraulic conductivity; 2) to assess the correlation of the
experimental RTD curves with the curve of the model, as a function on the variation

Benzene Removal In Horizontal Subsurface Flow Constructed Wetlands 14605
of hydraulic characteristics and clogging; 3) to evaluate Benzene removal as a
function of the hydraulic residence time (HRT).
Material and Methods
Experimental Constructed Wetlands
The experimental area includes three constructed wetland fields, two containing
different species of plants and the third serving as a control reactor. Water is supplied
to the fields from four high density polyethylene (HDPE) tanks; samples are obtained
from eighteen sampling ports and effluent is stored in two lagoon ponds. Fig. 1
depicts the plan view of the site (Fig. 1A) and the longitudinalsection of one planted
constructed wetland bed (Fig. 1B). Each wetland has a planted area equal to 15 m2 (3
x 5 m), a water depth ranging from 0.6 m to 0.65 m and a resulting total volume of
approximately 9.4 m3. The constructed wetlands have a bottom slope of 1% to
facilitate the outflow of water by gravity. The stability of the side banks is ensured by
providing a 45° inclination. Five perforated tubes with 200 mm internal diameter are
positioned within each field to permit collection of water samples and control of water
levels. The bottom of each reactor is waterproofed with a bentonite liner that is
permeable to plant roots but largely impermeable to water. The liner consists of three
layers: an upper geotextile 220 g/m2, a lower geo-textile 110 g/m2, and sodic
powdered bentonite 4670 g/m2, containing approximately 90% montmorillonite. The
total weight of the geo-composite is 5000 g/m2 and its total dry thickness is 6 mm.
The hydraulic conductivity of the installed liner is k < 10-11 m/s.
Raw water is supplied at the reactor inlet and passes slowly through the filtration
medium under the surface of the bed in a generally horizontal path until it reaches the
outlet zone where it is collected and discharged to the lagoon. The filtration medium
consists of three layers: 0.1 m of soil, 0.2 m of stones, and 0.30 - 0.35 m gravel as
shown in Fig. 1B. The mineral composition of the substrate is 59% calcium carbonate,
32% silica and 9% iron oxide for the rocks and gravel. The soil is a mixture of red
clay and organic matter. The unplanted bed served as a control reactor to isolate the
impact of macrophytes and any microorganisms associated with their root mass on
Benzene removal. At no time during the study was the water depth above the top of
the media in any of the reactors, i.e., all flow was subsurface.
Residence time distribution (RTD) curves were assessed by introducing 30 L of a 10
g/L solution of lithium bromide along the first cross section of each wetland unit as a
conservative tracer.
Water samples were collected every 30 min from each of the sampling points. The
experiments were carried out from March to June during the vegetation period of the
macrophytes. Daily average temperatures during the four months of experimentation
range from 14 °C to 24 °C.

14606 Ezio Ranieri
Figure 1: Constructed wetlands pilot plant in Lecce, Italy, shown in (A) plan view
and (B) longitudinal section.
Tracer injection
Sampling for HRT measurements
The tracer used in the experimental plant was the potassium bromide (KBr), because
it is highly soluble, non-degradable, relatively inexpensive, and can be measured
quantitatively in very low concentrations. Tracer solution was added in 10 min mixed
with wastewater flow in order to reduce sinking effects related to density differences.
Composite samples of the effluent from each constructed wetland were collected in
500 mL amber glass bottles using an auto sampler. Effluent grab samples were taken
approximately every 12 h from the morning of day 3 until the evening of day 9. From
the morning of day 10 to the morning of day 12, samples were taken every 24 h. Tests
finished on the fourth day after a total sampling period of approximately 330 h. For a

Benzene Removal In Horizontal Subsurface Flow Constructed Wetlands 14607
time of approximately 300 h, the tracer concentration was not detected and, therefore,
a period of time of 300 h was enough to obtain a complete response of the tracer
injection. RTD curves were assessed by introducing 6 kg/m3solution of KBr in 10 min
along the first cross-section of each wetland unit as a conservative tracer.
Benzene sampling and analysis
Benzene solution was conveyed to the CWs from the supply tanks containing tap
water at a constant initial concentration of 0.5 mg/L, respectively, for each compound,
for all tests. Composite samples of the effluent from each constructed wetland were
collected in 500 mL amber glass bottles every 6 h using an auto sampler for a time
period of 220 d. Samples were collected at inlet and outlet two times per week and
were kept refrigerated at 4°C until analyses. Samples were analyzed according to
Standard Methods (APHA, 2005) using an HP 5890 series II Gas Chromatograph
equipped flame ionization detector and a split/splitless injector. Standard deviation
(SD) was calculated for each measurement series and was less than 5% for each
compound considered. For all measurements, standard Quality Control (QC) was
performed. QC samples consisted of triplicate samples and spiked samples.
Plug flow with dispersion reactor model
The RTD curves have been calculated using the plug flow with dispersion reactor
(PFDR) model by adjusting the HRT (θ) and the reactor Peclet number to minimize
the sum of the squared errors between the experimental bromide concentration data
and the analytical solution to the PFDR model given by Levenspiel and Smith (1975):
(1)
where iCi t is the area under the RTD curve, Pe is Peclet number, and θ is the HRT.
The equation has been modified from its original dimensionless form by multiplying
the summation of CiΔt, which approximates the area under the RTD curve.
Results and Discussion
Effect of Clogging
In the pilot HSFCWs, experimental curves have been collected at time t = 0 and at
time t = 24 months to the aim of evaluating the different clogging conditions. Table 1
reports the hydraulic parameters adopted for the experimental HSFCWs and utilized
in the model at the beginning of the experience and after two years for both planted
and unplanted fields.

14608 Ezio Ranieri
Table 1:Values of parameters that define hydraulic behavior of the tested
experimental plants and after 24 months (planted and unplanted).
Parameter
Symbol
Parameter Name
Value (at the
beginning)
Value (after
24 months for
planted fields)
Value (after 24
months for
unplanted field)
K
Hydraulic
Conductivity
30 [m/d]
25 [m/d]
30 [m/d]
Hout
Hydraulic head at
the outlet
0.6 [m]
0.6 [m]
0.6 [m]
αL
Longitudinal
Dispersion
0.2 [m]
0.35 [m]
0.35 [m]
αT
Tranversal
Dispersion
0.02 [m]
0.02 [m]
0.02 [m]
DKBr
Diffusion
2.02-5 [cm2/s]
2.02-5 [cm2/s]
2.02-5 [cm2/s]
p
Porosity
0.16
0.16
0.15
Fig. 2(A) shows the comparison between the experimental and the simulated RTD
curves, measured at the beginning of the experience in the experimental plant. A good
correspondence, R2 = 0.98, according to well-recognized model evaluation techniques
(Moriasi et al., 2007) between the simulation curve and the tracer test has been found.
A less-pronounced plug flow behavior in the Lecce plant is probably due to lower
porosity of the substrate. Phragmites HSFCWs behavior is quite similar to the Typha
ones, whereas the unplanted field showing a clearer plug flow behavior. The variation
from the ideal plug flow behavior and the coefficient correlation (R2) was equal to
0.92 for the unplanted field, 0.87 for the Phragmites field and 0.86 for the Typha
field.
(a)
0
5
10
15
20
25
020 40 60 80 100 120 140 160
Br [mg/L]
Time [h]
Bromide concentration at time t2= 24 months
(Phragmites and Typha)
Model t1
Model t2
Phragmites
Typha

Benzene Removal In Horizontal Subsurface Flow Constructed Wetlands 14609
(b)
(c)
Figure 2: Bromide concentration trends vs HRT at time t=0 (A) and t=24 months (B)
and (C).
After 24 months, tracer tests were performed and the results are illustrated in Fig.
2(B). The correlation between the model and the experimental data was less evident.
In particular, while the unplanted field still maintains a good plug flow behavior, the
Phragmites and the Typha plants show a decrease of the peak with lower concavity of
the curve and a higher distance from the model interpolation curve. This is probably
due to the lower hydraulic conductivity measured in the field hydraulic conductivity
decreasing from 30 to 25m/d for the Phragmites field and 25.2 m/d for the Typha
field, as reported in previous experiences. The RTDs for the unplanted wetland have
been assessed after 24 months. Results are shown in Fig. 2(C). It is observed that the
model curve interpolate very well the tracer experimental data in the unplanted field
where the hydraulic conductivity still remain equal to 30 m/d and only the porosity
decrease from 0.16 to 0.15 as measured in the field. The RTD curves have been
calculated using the PFDR model. All of the PFDR model fits to the tracer data had
0
5
10
15
20
25
020 40 60 80 100 120 140 160
Br [mg/L]
Time [h]
Bromide concentration at time t1=0
Phragmites
Typha
unplanted
model
0
5
10
15
20
25
020 40 60 80 100 120 140 160
Br [mg/L]
Time [h]
Bromide concentration at time t2= 24 months
(unplanted field)
unplanted
model
unplanted

14610 Ezio Ranieri
R2 values greater than 0.975. The differences between the RTD curves for the planted
reactors are probably related to the different root structures of the two species. The
roots of Phragmitesaustralis penetrate to a depth of approx. 51 cm, while
Typhalatifolia roots are not likely to extend beyond about 29 cm, according to
previous experience with these species (Bedessem et al., 2007). The differential root
penetration depth is likely to be responsible for the slightly different flow regimes in
the two wetland beds. Clogging was more significant in the Phragmites bed, and this
favors the development of preferential flow paths and causes the slightly shorter HRT
compared to the Typha one bed and the slightly lower Peclet number of 26.7 in
contrast to Pe for the Typha bed of 29.7. The unplanted bed had a Peclet number of
24.9.
Benzene Removal
The residual concentrations at the sampling points for Phragmites field ranged
between 0.88 mg/L at the out and 0.77 mg/L at the lagoon pond. The final residual
concentration in the Typha field was 0.83 mg/L, while in the pond it was 0.75 mg/L.
In the unplanted field the residual concentrations were 1.09 mg/L at the out and 0.74
mg/L at the pond (Fig. 3). Based on the above, the average removal efficiencies are
equal to 39.78% for the Phragmites field, 35.14% for the Typha field and 29.67% for
the unplanted one (Fig.4).
Figure 3: Benzene concentration at the sampling points in the Phragmites field, in the
Typha field and in the unplanted one.
0
0.5
1
1.5
2
2.5
Benzene (mg/l)
Phragmites
Typha
Soil

Benzene Removal In Horizontal Subsurface Flow Constructed Wetlands 14611
Figure 4:Percentage removal of Benzene measured in each sampling point from the
input in the Phragmites field, in the Typhafield and in the unplanted one.
The removal efficiencies determined in the present work are lower or similar
compared with a similar experimental work (Eke and Scholze, 2008;Machate et al.,
1999). The latter may be attributed to the fact that the inflow concentration at the
present concentration was lower (0.5 mg/L here, instead of 2 mg/L) (Gikas et al.,
2013). The observed removals in the Phragmites field were, on average, 11.67%
higher than the Typha field and 25.42% higher than the unplanted field. However,
because of the low affinity of the Benzene compounds with plant tissues, the direct
effect of vegetation should be less significant compared with the net effect of sorption
(Mitsch and Gosselink, 2000). However the effect of macrophytes is more evident
approx after 30 h HRT – piezometer E and D – where the difference between
percentage values of planted and unplanted is higher than 20% (Fig. 4). Higher
removal is probably due to the microbial communities associated with the plant
rhizosphere which create an environment conducive to degradation for many volatile
organic compounds (Schnoor et al., 1995). Benzene overall removal after pond
settlement is around 60 %.
Conclusions
CWs offer a potential for the removal of more than 50% of Benzene from wastewater
at HRT higher than 100 h, however, the latter correlation should be evaluated as a
function of inlet concentrations. Percentages removal at the outlet of plants varies
from 45.4% (unplanted) to 55.6% (Phragmites) and 58.6 (Typha).
Hydraulic residence time have been also evaluated in CWs experimental plants.
Model shows a good agreement with experimental data for plants. After 24 months
the hydraulic conductivity is varied from 30 m/d to 25 m/d for both Phragmites and
Typha plants in HSFCWs due to clogging and the model curve is capable of
0
10
20
30
40
50
60
70
Piez. A
Piez. B
Piez. C
Piez. D
Piez. E
Out
Pond
% removal
Phragmites
Typha
Soil

14612 Ezio Ranieri
interpolating this hydraulic behaviour variation. The lack of the vegetation in the
unplanted constructed wetland results in a constant value of the hydraulic
conductivity after 24 months of operating. In the unplanted field only a slightly
decrease of the porosity has been evidence. This causes a shift of the experimental
curve towards lower HRTs. The field measurement of hydraulic conductivity appears
to be one crucial parameter useful to predict the actual hydraulic behaviour.
Further large-scale experimental tests should be carried out to validate the results
presented in this paper.
Acknowledgements
This research has been partially financed by the PRIN Program, Italian Minister of
University.
Conflict of Interests
The authors declare that there is no conflict of interests regarding the publication of
this paper.
References
[1]. APHA-AWWA-WPCF, 2005, Standard Methods for the Examination of
Water and Wastewater twenty; American Public Health Association,
Washington DC.
[2]. Bedessem, M.E., Ferro, A.M., and Hiegel, T., 2007,“Pilot-scale
constructed wetlands for petroleum-contaminated groundwater”, Water
Environ. Res., 79, pp. 581–586.
[3]. Eke, P.E., andScholz, M., 2008,“Benzene removal with vertical-flow
constructed treatment wetlands”, J. Chem. Technol. Biotechnol., 83, pp.
55–63.
[4]. García, J.,2004, Constructed wetlands for pollution control: perspective on
a technology in expansion. In: New Criteria for Design and Operation of
Constructed Wetlands: A Low-Cost Alternative for Wastewater Treatment,
3rdedn; García, J., Morató, J., Bayona, J.M.; Ediciones CPET, Barcelona,
Spain, pp. 7–16.
[5]. Gikas, P., Ranieri, E., andTchobanoglous, G., 2013,“Removal of iron,
chromium and lead from waste-water by horizontal subsurface flow
constructed wetlands”, J. Chem. Technol. Biotechnol.,88, pp. 1906–1912.
[6]. Lee, B.H., Scholz, M., and Horn, A., 2005,“Constructed wetlands for the
treatment of concentrated stormwater runoff (part A)”, Environmental
Engineering Science, 23, pp. 191–202.
[7]. Levenspiel, O., and Smith, W.K., 1957,“Notes on the diffusion type model
for the longitudinal mixing of fluids in flow”, Chem. Eng. Sci., 6, pp. 227–
235.

Benzene Removal In Horizontal Subsurface Flow Constructed Wetlands 14613
[8]. Lu, C., Lin, M.R., and Chu, C., 2002,“Effects of pH, moisture, and flow
pattern on trickle-bed air biofilter performance for BTEX removal”,
Advance in Environmental Research, 6, pp. 99–106.
[9]. Machate, T., Heuermann, E., Schramm, K.W., and Kettrup, A.,
1999,“Purification of fuel and nitrate contaminated ground water using a
free water surface constructed wetland plant”, J. Environ. Quality, 28, pp.
1665–1673.
[10]. Mitsch, W.J., andGosselink, J.G., 2000, Wetlands; Wiley, New York.
[11]. Moriasi, D.N., Arnold, J.G., Liew, M.W.V., Bingner, R.L., Harmel, R.D.,
andVeith, T.L., 2007,“Model evaluation guidelines for systematic
quantification of accuracy in watershed simulations”, Trans. Am. Soc.
Agr. Biol. Eng., 50, pp. 885–900.
[12]. Nickelsen, M.G., and Cooper, W.J., 1992,“Removal of benzene and
selected alkyl-substituted benzenes from aqueous solution utilizing
continuous high-energy electron irradiation”, Environmental Science and
Technology, 26, pp. 144–152.
[13]. Pedescoll, A., Uggetti, E., Llorens, E., Granés, F., García, D., andGarcía,
J., 2009,“Practical method based on saturated hydraulic conductivity used
to assess clogging in subsurface flow constructed wetlands”, Ecol. Eng.,
35, pp. 1216–1224.
[14]. Puigagut, J., Villaseñor, J., Salas, J.J., Bécares, E., andGarcía, J.,
2007,“Subsurface-flow constructed wetlands in Spain for the sanitation of
small communities: a comparative study”, Ecol. Eng., 30 (4), pp. 312–319.
[15]. Raboni, M., Gavasci, R., andUrbini, G., 2014,“UASB followed by Sub-
Surface Horizontal Flow Phytodepuration for the Treatment of the Sewage
Generated by a Small Rural Community”, Sustainability 6(10), pp. 6998-
7012; doi:10.3390/su6106998.
[16]. Ranieri, E., 2012 “Chromium and nickel control in full and small scale
subsuperficial flow constructed wetlands”, Soil Sediment Contam., 21, pp.
802–814.
[17]. Ranieri, E., Gikas, P., andTchobanoglous, G., 2013a,“BTEX removal in
pilot-scale horizontal subsurface flow constructed wetlands”,Desalin.
Water Treat., 51, pp. 3032–3039.
[18]. Ranieri, E., Gorgoglione, A., Solimeno, A, 2013bAcomparisonbetween
model and experimental hydraulic performances in a pilot-scale
horizontalsubsurface flow constructedwetland . Ecological Engineering,
60, pp 45-49.
[19]. Ranieri, E., Gorgoglione, A., Montanaro, C., Iacovelli, A., and Gikas, P.,
2014,“Removal capacity of BTEX and metals of constructed wetlands
under the influence of hydraulic conductivity”, Desalin. Water Treat., 1–8,
http://dx.doi.org/10.1080/19443994.2014.951963.
[20]. Ranieri, E., and Young, T.M., 2012,“Clogging influence on metals
migration and removal in sub-surface flow constructed wetlands”, J.
Contam. Hydrol., 129–130, pp. 38–45.

14614 Ezio Ranieri
[21]. Rousseau, D.P.L., Lesage, E., Story, A., Vanrolleghem, P.A., and De
Pauw, N., 2008,“Constructed wetlands for water reclamation”,
Desalination, 218, pp. 181–189.
[22]. Schnoor, J., Litch, L.A., McCutcheon, S.C., Wolfe, N.L., andCarreira,
L.H.,1995,“Phytoremediation of organic and nutrients contaminants”,
Environ. Sci. Technol., 29(7), pp. 318A–323A.
[23]. Tang, X., Eke, P.E., Scholz, M., and Huang, S., 2009,“Processes impacting
on benzene removal in vertical-flow constructed wetlands”, Bioresource
Technology, 100(1), pp. 227–234.
[24]. Vymazal, J., 2005,“Horizontal sub-surface flow and hybrid constructed
wetlands systems for wastewater treatment”, Ecol. Eng., 25, pp. 478–490.
[25]. Yeom, S.H., andYoo, Y.J., 1999,“Removal of benzene in a hybrid
bioreactor”, Process Biochemistry, 34, pp. 281–288.