International Conference on WAter POllution in natural POrous media at different scales. Assessment of fate, impact and indicators. WAPO2
REACTIVE TRANSPORT EXPERIMENTS OF LINEAR ALKYLBENZENE SULFONATE IN LABORATORY SOIL COLUMNS N. Boluda1, V. Cases1, V.M. León1, V. Gomis1 and D. Prats2, (1) Department of Chemical Engineering. University of Alicante. Apdo. 99, E-03080 Alicante. Spain. e-mail:
[email protected],
[email protected],
[email protected],
[email protected]. (2) Instituto Universitario del Agua y de las Ciencias Ambientales (IUACA). University of Alicante. Apdo. 99, E-03080 Alicante. Spain. e-mail:
[email protected].
ABSTRACT Laboratory column experiments aids in the understanding of hydrogeochemistry dynamics of different linear alkylbenzene sulfonate homologues when they interact with different sand/soil proportions. Previously it was necessary to characterize the soil columns with a tracer. In this paper, the method to obtain breakthrough curves for the conductivity and chloride concentration of a tracer is verified, and also permits the calculation of transport parameters with an easy-to-use interface, ACUAINTRUSION. The experimental results obtained for LAS concentration in two columns using the same type of core have demonstrated the reproducibility of the method. In these experiments, LAS sorption is more important for longer alkylic chain homologues, especially for C13LAS that was only detected at very low concentrations in the outlet of the soil column. However, C10LAS is less retained by the soil column due to its lower hydrophobicity. In the two experiments with different sand/agricultural soil proportions, the LAS concentrations values were quite similar, and therefore a higher soil column content seems to have no significant effect on LAS homologue detected levels, even when a high LAS concentration pulse occurs (100 ppm). Key words: Column experiment, LAS, reactive transport, breakthrough curves, surfactant.
1. INTRODUCTION Linear Alkylbenzene Sulfonate (LAS) is the highest volume anionic surfactant used in domestic and commercial detergent formulations. LAS enters the environment primarily through treated and untreated sewage effluents and sewage sludge. The reuse of wastewater in agricultural practices is an absolute necessity in regions with serious water shortages, such as the Community of Valencia in Spain (Prats, 2000). The practice of applying sewage sludge to agricultural land is yet another way for LAS to enter the environment (Branner et al., 1999). Moreover, the environmental fate of surfactants in groundwater is also of interest because surfactants are being investigated as potential agents for enhancing the remediation of aquifers, contaminated with dense non aqueous phase liquids (Abriola et al., 2002). Therefore, given the possibility of surfactants showing up in groundwater, whether as micro pollutants or remediation additives, it is important to understand surfactant behaviour in subsurface environments (Krueger et al., 1998). For the purpose of identifying the physicochemical processes that take place in the environment, it is necessary to carry out laboratory tests in soil cores under controlled conditions (Gomis et al., 2000). The heart of the experimental apparatus consists of a stainless steel column connected to a HPLC pump (Gomis et al., 1997). In this paper, we present the results obtained in continuous flow tests with 100 ppm LAS pulse injection to characterize the hydrogeochemistry dynamics of LAS homologues (commercial product) when they interact with varying sand/soil proportions in columns with different transport parameters such as porosity, Péclet number, etc.
2. EXPERIMENTAL PROCEDURE 2.1. Materials and methods Several column experiments were carried out to obtain LAS homologue concentration profiles into which a pulse of 100 ppm total LAS containing 12.1% C10, 34.1% C11, 30.6% C12 and 23.2 % C13
1
BARCELONA (SPAIN) APRIL 11th – 13th, 2007
International Conference on WAter POllution in natural POrous media at different scales. Assessment of fate, impact and indicators. WAPO2
(donated by PETRESA) was introduced. The columns were filled with agricultural soil (CaCO3: 38.3%, organic carbon: 0.78%, sand: 23.6%, slime: 38.0% and clay: 38.4%) and purified commercial sea sand (Merck). The experiments were performed using sterilized irrigation water (table 1), to prevent any biodegradation processes from taking place, in columns containing different sand/agricultural soil proportions. Table 1. Irrigation water concentration Concentration
Na+
K+
Ca+
Mg+
HCO3-
SO42-
Cl-
ppm
100
2.8
85
46
230
70
275
The experimental set-up (figure 1) consisted of a thermostated stainless steel column (22.4 cm length, 2.5 cm internal diameter), operating at room conditions ( 25 ºC, 1 atm), connected to a HPLC pump (Shimadzu LC 9A). A conductivity detector (Shimadzu CDD-6A) was used to record changes in tracer conductivity (CaCl2 0.01 - 0.04 M corresponding to 3800 – 6600 µS/cm). The effluent was collected continuously, in 10mL test tubes, at the outlet of the column (fraction collector, Frac-920 Amersham).
Figure 1. Experimental set-up
2.1. Column transport parameters Before starting the experimental study of LAS, the hydrogeochemical parameters in the different soil columns was characterized using CaCl2 as a tracer. Samples were collected at the outlet of the column and analyzed to determine chloride concentrations. At the same time, conductivity values were recorded on a computer. The obtained breakthrough curves were utilized to calculate transport parameters with ACUAINTRUSION(Boluda et al., 2006). This graphical user interface calculates the best fit of the experimental data (chloride concentration (mmol/L) or conductivity versus experimental time (h)) with the analytical solution of the convection-dispersion equation (Lapidus and Amundson, 1952) being C r (L, t ) = C i +
(C 0 − C i ) ⎡ 2
⎛ L − vt ⎞ ⎛ ⎞⎤ ⎟ + exp⎛⎜ vL ⎞⎟erfc⎜ L + vt ⎟⎥ ⎢erfc⎜ ⎜D ⎟ ⎜ 4D t ⎟ ⎜ 4D t ⎟ ⎥ ⎢⎣ ⎝ L⎠ L ⎠ L ⎠⎦ ⎝ ⎝
2
BARCELONA (SPAIN) APRIL 11th – 13th, 2007
International Conference on WAter POllution in natural POrous media at different scales. Assessment of fate, impact and indicators. WAPO2
where Cr(L,t), is the resident solute concentration at the output stream of the column, Ci, initial concentration of the solute, C0, concentration of solute in the continuous supply of tracer, L, column length, t, time, v, interstitial water velocity in the direction of propagation, and DL, longitudinal dispersion coefficient To obtain the best fit, the quadratic mean deviation between experimental and calculated compositions have been minimized. After that calculation, the program provides the calculated transport parameters: Péclet number (vL/D), residence time TR, Darcy velocity u, porosity ε, interstitial velocity v, dispersion D and dispersivity α.
2.3. LAS column experiment Once the transport parameters had been obtained, the porous media was conditioned with filtered and sterilized irrigation water until the sediment was saturated. The LAS concentration experiments were performed utilizing a 4 hour pulse feed at the inlet of the column. The feed bottle was kept in a vessel at 25 ºC and constantly mixed to avoid LAS molecules from migrating to the water/air interface and bottle wall. Several LAS pulse column experiments were carried out with different sand/soil proportions because a high clay content may produce low soil permeability. The LAS experiments were also performed simultaneously to check system reproducibility. The effluent was collected continuously, as samples of 10 mL each , at the outlet of the column with a fraction collector. Subsequently, the samples were analysed by a HPLC (Agilent 1100) with a UV detector (254 nm) to determine the LAS concentration. As the stationary phase, a Lichrocart (150x4.6 mm) RP-8 (5 mm) was used and as the mobile phase a solution of MeOH/H2O (85/15) with 0.5 M of NaClO4·H2O, with a flow rate of 0.8 mL/.min (isocratic regime), was employed. To determine the sodium, calcium, magnesium and potassium concentrations, the samples were analyzed by ICP-MS.
3. RESULTS 3.1. Tracer breakthrough curves Figure 2 shows breakthrough curves, obtained from values of conductivity and chloride concentration, presented as dimensionless (C/Co vs pore volumes = [time/mean elution time of conservative chloride]), for two experiments with same core (50% sand + 50% soil). When C/Co is 0.5, the pore volume is equal to 1, that is, the time necessary for the tracer to traverse the column when no interactions occur between tracer and sediment (TR, residence time). It can be seen on figure 2 that the conductivity and concentration chloride data coincide. Therefore, both methods are efficient when employed to calculate column parameters in step experiments. Similarly, a comparison of the shapes of the chloride concentration curves obtained from the different experiments (I and II), demonstrates the reproducibility of the method.
3
BARCELONA (SPAIN) APRIL 11th – 13th, 2007
International Conference on WAter POllution in natural POrous media at different scales. Assessment of fate, impact and indicators. WAPO2 Breakthrought curves Column 50% Sand + 50% Soil
1.2 1
C/Co
0.8 0.6 Chloride (Exp I) Conductivity (Exp I)
0.4
Chloride (Exp II) Conductivity (Exp II)
0.2 0 0
0.5
1
1.5
2
Pore Volume
Figure 2. Breakthrough curves in different experiments with the same core (50% sand + 50% soil).
3.2. LAS experimental concentrations Figure 3 shows the LAS homologue experimental data when a 100 ppm pulse total LAS was introduced, simultaneously, into two different columns filled with 50% sand + 50% soil. The reproducibility of the method is demonstrated, as the trends clearly show on the figure. Pulse 100 ppm LAS (Reproducibility) Column: 50% Sand + 50% Soil
10 C10LAS (exp 1)
Concentration (ppm)
9
C10LAS (exp 2)
8
C11LAS (exp 1)
7
C11LAS (exp 2)
6
C12LAS (exp 1)
5
C12LAS (exp 2)
4 3 2 1 0 0
20
40
60
80
100
120
140
Time (hours)
Figure 3. LAS homologue concentration of two experiments (realized simultaneously)in two different columns
It can be seen that the plotted peaks for each LAS homologue show the same order that the homologues follow when they leave the column, (C10 < C11 < C12 ). The shorter alkylic chain homologues travel faster through the core, while other others like C12, which has a longer linear chain, stay in the column for longer periods of time. These results support previous experiments that have shown that hydrophobic interaction is the main process by which LAS sorption on soil or sediments takes place (Hand & Williams, 1987), as does the higher sorption capacity for longer alkylic chain or for isomers with longer alkylic fragments. And what is more, homologues collected in independent experiments reach peak values at similar pore volumes, obtaining the same results in different columns filled with similar cores.
4
BARCELONA (SPAIN) APRIL 11th – 13th, 2007
International Conference on WAter POllution in natural POrous media at different scales. Assessment of fate, impact and indicators. WAPO2
In the case of C13LAS, peaks were not detected during experiments until after long periods of time (pore volume up to 50) had transpired.This means the homologue could stay inside the column for an extensive period of time due to its higher sorption capacity as compared to the rest of the LAS homologues. Experiment 1 (50% sand+50% soil), showed in figure 3, is plotted with other experimental data obtained from a 75% sand-25% soil column in figure 4. The shape is similar in both cases but not very different concentrations are observed. PULSE 100 PPM LAS: 75% SAND - 25% SOIL and 50% SAND - 50%SOIL 10 C10LAS(COL. 75/25)
9
C11LAS(COL. 75/25)
ppm LAS ::
8
C12LAS(COL. 75/25)
7
C10LAS(COL. 50/50)
6
C11LAS(COL. 50/50)
5
C12LAS(COL. 50/50)
4 3 2 1 0 0
10
20
30
40
50
Pore Volume
Figure 4. Two experiments with different columns filled with 75% sand - 25% soil and with 50% sand - 50% soil.
4. DISCUSSION The breakthrough curves obtained allow the calculation of transport parameters with ACUAINTRUSION. Table 2 provides results obtained from experimental breakthrough curves along with those of conductivity and chloride concentration. It goes without saying that there is a perfect coincidence of numerical data, that permits, on one hand, the validation of results and, on the other, the use of both methods to characterize the core. Table 2. Transport parameters calculated by following conductivity and chloride concentration breakthrough curves. Column
Péclet Number
Residence Time
Darcy`s velocity
(h)
(cm/h)
Porosity
Intersticial velocity
Dispersion coefficient
Dispersivity
(cm/h)
(cm2/h)
(m×103)
CHLORIDE
177
2.16
5.85
0.56
10.4
1.31
1.26
CONDUCTIVITY
182
2.11
5.85
0.55
10.6
1.31
1.23
Table 3 shows column parameters obtained from two different cores. It can be seen that porosity increases with proportion of soil in the column, and hence the tracer needs more time to travel through the column. In this case, the Péclet number increases, and the peaks representing the LAS homologues on figure 3 are less dispersed, indicating higher concentration values. The higher soil column content seems not to have as much significance when a heavy concentration of LAS pulse is introduced into
5
BARCELONA (SPAIN) APRIL 11th – 13th, 2007
International Conference on WAter POllution in natural POrous media at different scales. Assessment of fate, impact and indicators. WAPO2
the column. In these experiments no interactions with cations were observed, in contrast to those seen in a step experiment (Boluda, 2005). Table 3. Transport parameters calculated by using chloride concentration breakthrough curves in two different kind cores. Column
Péclet Number
Residence Time
Darcy`s velocity
(h)
(cm/h)
Porosity
Intersticial velocity
Dispersion coefficient
Dispersivity
(cm/h)
(cm2/h)
(cm)
75% SAND + 25% SOIL
113
1.91
5.74
0.49
11.7
2.32
0.198
50% SAND + 50% SOIL
163
2.16
5.78
0.56
10.4
1.43
0.138
5. CONCLUSIONS The breakthrough curves of conductivity and chloride concentration obtained, coincide perfectly and both methods employed allow the calculation of transport parameters with an easy-to-use interface, ACUAINTRUSION. Similar LAS concentrations obtained in two columns with the same core permits the validation of the column results. LAS sorption is more important for longer alkylic chain homologues, especially for C13LAS, that was detected at low concentrations in the outlet of the soil column. However, C10LAS is less retained by the soil column due to its lower hydrophobicity, and consequently higher quantities could gain access to deeper soil layers. In two experiments with different sand/agricultural soil proportions, LAS concentration values are quite similar, and therefore a higher soil column content seems to have no significant effect on LAS homologue detected levels, even when a high LAS concentration pulse occurs (100 ppm).
6. BIBLIOGRAPHY Abriola, L.M., Drummond, C.D., Lemke, L.D. and Rthfelder, K.M. (2002): Surfactant enhanced aquifer remediation: application of mathematical models in the design and evaluation of a pilot-scale test. In: Groundwater Quality. Natural and Enhanced Restoration of Groundwater Pollution, IAHS Publ., 275: 303-309. Boluda, N.; León, V.; Prats, D.; Chorro, M.C. (2005): Characterization of transport and reaction processes of Linear Alkylbenzene Sulfonate in agricultural soil columns. In: 10th Mediterranean Congress of Chemical Engineering. Sociedad Española de Química Industrial e Ingeniería Química Barcelona (Spain), 389. Boluda, N., Gomis, V. and Pedraza, R. (2006): ACUAINTRUSION - A graphical user interface for a hydrogeochemical seawater intrusion model. In: 1st SWIM-SWICA. Cagliary (Italy). Branner, U., Mygind, M. and Jorgensen, C. (1999): Degradation of Alkylbenzene Sulfonate in Soil Columns. Environ. Toxicol. Chem., 18, 8: 1772-1778. Gomis, V., Boluda, N. and Ruiz, F. (1997): Column displacement experiment to validate hydrogeochemical models of seawater intrusions. J. Cont. Hydrol., 29: 81-91. Gomis-Yagües, V., Boluda-Botella, N. and Ruiz-Beviá, F. (2000): Gypsum precipitation as an explanation of the decrease of sulphate concentration during seawater intrusion. J. Cont. Hydrol., 228: 48-55. Hand, V.C., Williams, G.K. (1987): Structure-activity relationships for sorption of linear alkylbenzenesulfonates. Environ. Sci. Technol., 21, 370-373. Krueger C.J., Barber, L.B., Metge, D.W. and Field, J.A. (1998): Fate and transport of Linear Alkylbenzenesulfonate in a Sewage-Contaminated Aquifer: A Comparison of Natural-Gradient Pulsed Tracer Test. Environ. Sci. Technol., 32: 1134-1142. Lapidus, L. and Amundson, N.R. (1952): Mathematics of adsoption in beds. VI. The effect of longitudinal diffusion in ion-exchange and chromatographic columns. J. Phys.Chem. 56: 984-988. Prats, D. 2000. Abastecimiento de agua en cuencas deficitarias. Actuaciones posibles y problemas asociados. In: Congreso Nacional Gestión del Agua en Cuencas Deficitarias, Centro del Investigación del Bajo Segura “Alquibla”, Orihuela (Spain), 297-307.
6
BARCELONA (SPAIN) APRIL 11th – 13th, 2007
