EFFECT OF FLOW RATE ON THE RETENTION OF THREE HEAVY METALS (ZN, CD AND PB) IN A FLUVIO-GLACIAL DEPOSIT COLUMN
Laureline FEVRIERa,*, Thierry WINIARSKIa, Cécile DELOLMEa, Jean-Paul GAUDETb
a Laboratoire des sciences de l'Environnement (LSE), ENTPE, 2 rue Maurice Audin, 69518 Vaulx-en-Velin Cedex, France
b Laboratoire d'étude des Transferts en Hydrologie et Environnement (LTHE), INPG-UJF-CNRS (UMR 55 64), BP 53, 38041 Grenoble, France
*
corresponding author, e-mail : fevrier@entpe.fr
ABSTRACT
In the Lyon
area (France), the use of stormwater infiltration basins led to an infiltration
of polluted waters in a high permeability fluvio-glacial deposit (a commonly
found in the area). Thus it was of major interest to study the fate of
stormwater pollutants in such a soil. The objective of our study was to
evaluate the effect of flow rate on the retention of lead, zinc and cadmium in
this soil by monitoring soil column leaching experiments.
Hydrodynamic parameters of the column were
first determined by flow tracer experiments (Br- used as tracer) and
by fitting the bromide breakthrough curves with a two region transport model
(MIM). The fate of the three metals was analysed through the shape of their
elution curves, which are used for modelling, and by the determination of the
amount of metals retained in the soil. Low flow rate favoured metals retention,
even if equilibrium was never attained in the column.
INTRODUCTION
Over these
last few decades, the city of Lyon (France) has widely promoted the use of
stormwater infiltration basins. Thus the pollutants contained in the stormwater
(hydrocarbons, heavy metals, ...) penetrate the soil under the basin, without
any qualitative control. The soil encountered in this area is a fluvio-glacial
deposit with high permeability due to its coarse texture and a high amount of
carbonates (Tab. 1). Transfer of pollutants in this soil should be influenced
by these two properties. Nevertheless, the association of hydraulic and
chemical processes in such a soil have been rarely studied.
Thus our
study was designed on a laboratory scale to study the effects of hydraulic
mechanisms on the retention of lead, cadmium and zinc (species commonly found
in stormwater) in a fluvio-glacial deposit column.
Tab. 1: Soil characteristic
|
Particle
size distribution |
|
Physico-chemical
parameters
|
|
|
|
(% dry soil) |
pH |
8,65 |
|
gravel (2mm<X<10mm) |
45,06 |
carbonate content (%) |
25 |
|
sand (0,08mm<X<2mm) |
53,25 |
organic matter (%) |
2 |
|
silt and clay (<0,08mm) |
1,69 |
CEC (mol.g-1) |
2,75.10-5 |
METHODS
Fig. 1
schematically shows the experimental set-up. The columns were made from
transparent PVC and had a length of 30 cm (length of soil: 20 cm) and a
diameter of 10 cm. They were filled with 3 kg of soil humidified to 9%. Then
the soil was saturated slowly with NaNO3 (0,01 M). Solutions were
supplied at a constant flow using a peristaltic pump. Steady-state flow
conditions were established before all experiments by leaching one or two pore
volumes of NaNO3 (0,01 M). Samples were automatically taken with a
fraction collector.
Fig. 1: Experimental set-up
Tab.2: Soil column leaching characteristics
|
soil sieved (mm) |
10 |
|
soil density rd (g.cm-3) |
1,75 |
|
soil height (cm) |
20 |
|
water content q (cm3.cm-3) |
0,27 |
|
flow rate (cm.min-1) |
0,13 0,013 |
|
Br- 's pulse (Vo) |
2,9 |
|
metals' pulse (Vo) |
8 |
|
[Zn] (mol.l-1) |
10-3 |
|
[Cd] (mol.l-1) |
10-3 |
|
[Pb] (mol.l-1) |
10-3 |
An anionic
tracer (Br- - 0,01 M) was used for flow tracer experiments. Since
previous research had already studied the leaching of single metal in such a
soil [Plassard, 2000, Crosnier, 1999], we introduced the three metals together
([Pb] = [Cd] = [Zn] = 10-3mol.l-1). For each leaching
experiment the ionic strength was adjusted at 0,01 M. All solutes were
introduced as a pulse in the column. Table 2 gives further details about
leaching experiment conditions.
Metals
concentrations were measured by F-AAS whereas bromide concentrations were
measured by ionic chromatography.
Mass
balances of solutes (Br- and each metals) were estimated by the zero
order moment of their BTCs. The mean travel time and the retardation factor (R)
of Br- were estimated by the first order moment of the BTC [Leij and
Dane,1992], whereas they were estimated by the median time of the BTC for the
metals (because of the low amount of metals eluted) [Das and Kluitenberg,
1996]. Amounts of metals in soil at the end of the leaching experiments were
estimated to complete mass balance analyse. The extraction was made with aqua
regia in a microwave closed system.
A two
region (mobile-immobile water) transport model with first-order exchange of
solutes (MIM) was used to fit Br- BTCs, in order to determine
hydrodynamic parameters of the column. The MIM equations are [Van Genuchten and
Wieranga, 1976]:
and 
with 
where Cm
and Cim (mol.l-1) are the concentrations respectively in
the mobile and immobile regions, Dm the dispersion coefficient (cm2.min-1),
qm and qim the volumetric
water content in the mobile and immobile regions (cm3.cm-3)
and a the solute exchange rate (min-1)
between these two regions.
The
analytical solution for a pulse input of solute with a third-type boundary
condition [Toride et al, 1993] was used. Thus, the parameters fitted were : Dm,
a and qm
A transport
model for reactive solute, based on the MIM model including adsorption and a
degradation rate constant, was used to fit the metal BTCs.
RESULTS AND
DISCUSSION
Br-
BTCs showed identical behaviour at both flow rates when they are plotted in
dimensionless parameter (Fig. 2). They were very asymmetric, with an early
breakthrough and a long tailing.
Br-
mass balances were always close to unity. But the retardation factors were less
than 1, suggesting an anionic exclusion processes despite the little amount of
clay in our soil [Schoen et al, 1999]. Thus, a new water content, corresponding
to the volume of water "used" by the bromide, qBr = R.q was defined. And the Br- BTCs were fitted with the MIM model, according the
use of qBr instead of q as input in the model.
The MIM
model provided good results (Fig.3), since hydrodynamic parameters were
coherent between different experimental conditions. The fitted dispersivity (l = Dm.qm /q = 4cm) was quite high due to the structure
of the fluvio-glacial deposit. And the values of the Peclet number (Pe = q.L/(Dm.qm)) indicated that both convection and
dispersion were responsible for the transport in our column [Sardin et al,
1991].
The fitted
fraction of immobile water was very high (qim/q = 48%), in agreement with the early breakthrough of bromide and the
diffusive desorption front. This value was found independent of the flow rate ,
whereas the fitted solute exchange rate was correlated to the pore water
velocity.
Fig. 2: Br- BTC at two flow rate
Fig. 3: Br- BTC fitted with the MIM
model
Mass
balances calculated on the metals BTCs were far from unity (Fig. 4), indicating
that degradation or irreversible retention (precipitation) occurred during the
transport. This was particularly true for lead, which was not eluted at all.
Fig. 4: Zn and Cd BTCs at two flow rates
The amount
of metals in the column after leaching experiments (Fig. 5) showed different
trends. Lead was mainly retained at the top of the column whereas Zn and Cd
were more equally retained throughout the column.
Fig. 5:
Distribution of Pb, CD and Zn in the soil after leaching experiments for both
flow rate
The effect
of flow rate was significant on the amount of Cd and Zn eluted (28 % eluted at
high flow rate and 9 % at low flow rate) and on the retardation factors of
these two metals (R = 6 and R = 8
respectively for high and low flow rate). Given the metals profile in the
column (Fig. 5), this effect was significant only for the first twelve
centimeters of the soil.
Moreover,
the nonsimilarity of the BTCs for low and high flow rate (specially the
adsorption front) indicated that kinetic effects occurred during Zn and Cd
elution and that equilibrium was never obtained in the column. The asymmetric
shape of the BTCs was caused by nonlinear adsorption of metals [Hinz and Selim,
1994]. This was confirmed by our failure to model these curves with a MIM model
including a linear adsorption and a degradation rate constant [data not shown].
Thus more
research is needed to obtain a model which would the hydraulic behaviour of our
soil take into account and the non linearity of cadmium and zinc adsorption.
Thus,
chemical properties of a soil are far from being sufficient to estimate its
pollutants retention capacity. Our results show that hydraulic parameters,
especially flow velocities, play a major role in the fate of lead, cadmium and
zinc in a fluvio-glacial deposit. But more research is needed in this
particular domain of great environmental interest.
REFERENCES
Crosnier J
(1999), Devenir de la pollution métallique drainée par les eaux pluviales.
Influence du compartiment microbien et des alternances de
dessication/réhumectation sur le transfert du zinc dans la zone non saturée du
sol. Thèse Doct. Univ.Claude Bernard-Lyon I.
Das BS,
Kluitenberg GJ (1996), Soil Sci. Soc. Am. J. 60: 1724-1731
Hinz C,
Selim HM (1994), Soil Sci. Soc. Am. J. 58: 1316-1327
Leij FJ,
Dane JH (1992), Soil Sci. Soc. Am. J. 56: 667-674
Plassard F,
Winiarski T, Petit-Ramel M (2000), J. Cont. Hydrol. 42: 99-111
Sardin M,
Schweich D, Leij FJ, Van Genuchten MTh (1991), Water Resour. Res. 9: 2287-2307
Schoen R,
Gaudet JP, Elrick DE (1999), J. Hydrol. 215: 82-93
Toride N,
Leij FJ, Van Genuchten MTh (1993), Water Resour. Res. 7: 2167-2182
Van
Genuchten MTh, Wieranga PJ (1976), Soil Sci. Soc. Am. J. 40: 473-480