Sorption kinetics of chlorinated hydrophobic organic chemicals

Research Articles

Sorption Kinetics of Chlorinated Hydrophobic Organic Chemicals, Part II

Sorption Kinetics of Chlorinated Hydrophobic Organic Chemicals Part Ih Desorption Experiments* S. Marca Schrap 1, Gerard L. G. Sleijpen 2, Willem Seinen 1, Antoon Opperhuizen 3 t Research Institute of Toxicology, Environmental Chemistry Group, University of Utrecht, P.O. Box 80.058, 3508 TB Utrecht, The Netherlands. 2 Department of Mathematics, University of Utrecht, P.O. Box 80.010, 3508 TA Utrecht, The Netherlands. 3 Ministry of Transport, Public Works and Water Management, Tidal Waters Division, P.O. Box 20.907, 2500 EX The Hague, The Netherlands

* Part [ "The Use of First-Order Kinetic Multi-Compartment Models" is published in Vol. 1, issue 1, pp. 21 - 2 8 (1994) of ESPR. Corresponding author: Dr. S. Marca Schrap

Abstract This is the second of a two-part series describing the sorption kinetics of hydrophobic organic chemicals. Part I "The Use of First-Order Kinetic Multi-Compartment Models" is published in issue 1 of this journal, pp. 21 - 28. Sorption kinetics of chlorinated benzenes from a natural lake sediment have been investigated in gas-purge desorption experiments. Biphasic desorption curves, with an initial "fast" part and a subsequent "slow" part, were found for all tested chlorobenzenes. From these results first-order sorption uptake and desorption rate constants were calculated with a two-sediment compartment model, which is presented in the first paper. in three sets of experiments the sorption uptake period and sediment/water ratio were varied. Rate constants are not influenced by these experimental conditions, which supports the partitioning concept for the sorption of hydrophobic organic chemicals in sediments.

1

Introduction

The fate of micropollutants in the aquatic environment can be affected significantly by sorption in sediments and soils. Sorption in suspended particles or dissolved colloids may serve as a mechanism for enhanced or retarded mobility of the chemicals [e.g. 1,2]. Rates of chemical or biological degradation of the chemicals in the sorbed phase have been shown to differ from those in the aquatic phase [3]. In addition, the bioavailability of pollutants in relation to bioaccumulation and toxicity in aquatic organisms can be influenced by sorption [e.g. 4 - 8]. An understanding of the sorption processes is therefore of importance to describe the fate of pollutants in aquatic systems. Sorption of hydrophobic organic chemicals in soils, sediments or dissolved organic materials is often described as a partitioning process [9 - 13]. According to this concept, sorption can be seen as a liquid-liquid distribution process, in which the chemical dissolves in the (organic carbon of the) sediment. At equilibrium there is a constant ratio between the concentration of the chemical in the sediment and in the water: ~:p = c 7 / c~~

(1)

where Kp is the sorption partition coefficient (L/kg) and C~q and C~q are the concentrations of the chemicals in the sediment (/Jg/kg) and the water (/ag/L) respectively at equilibESPR-Environ. Sci. & Pollut. Res. 1 (2) 8 1 - 92 (1994) 9 ecomed publishers, D-86899 Landsberg, Germany

rium. The exchange of the chemicals between the sediment and the water compartment is described with first-order kinetics. Although in this partition model the sediment is assumed to be one homogeneous compartment, a twocompartment approach for the sediment phase seems to be a better description. This conclusion is based on the observation of a two-stage approach to equilibrium: a fast, short initial period followed by a much slower extended period [ 1 4 - 19]. Also terms such as reversible and resistant [20] and labile and non-labile [21] sorbed fractions are related to the two-compartment approach. Different methods of measuring sorption kinetics of organic chemicals have been reported. A common method is the batch method [e.g. 18, 2 2 - 2 4 ] in which known amounts of sediment, water and chemicals are continuously mixed and allowed to equilibrate for hours (or days or months). By diluting or by replacing the aqueous phase (or a part of it) by clean water, one can measure desorption from the release of the chemicals into the aqueous phase. This release can be optimized by adding an adsorbent as a sink for the desorbed chemicals [ 1 4 - 1 6 , 19]. Another method for measuring sorption kinetics is the column method. In this method dissolved chemicals are passed through columns packed with sediment or soil [e.g. 25 - 29]. Sorption and desorption can be determined from the differences between the aqueous concentration of the chemical in the inlet and the outlet of the column. These two methods may be subject to experimental artefacts in that the measured aqueous concentration of the chemical may consist of a freely dissolved fraction as well as a fraction sorbed in the remaining sediment parts [30 - 32]. These sediment parts, which are present in the aqueous phase, can be dissolved or they can be of a particulate nature. However, they cannot be completely separated from the aqueous phase by phase separation techniques (e.g. filtration, centrifugation). Consequently, measured concentrations in the aqueous phase are higher than the actual dissolved concentrations. This experimental artefact can result in nonsingularity or hysteresis of sorption-uptake and desorption data [e.g. 31, 33, 34]. However, also (too) short experimental sorption-uptake and desorption periods result in a non-equilibrium state, and lead to an apparent sorption non-singularity. This observation is then often misinterpreted as irreversibility or hysteresis, 81

Sorption Kinetics of Chlorinated Hydrophobic Organic Chemicals, Part II whereas the actual case is the lack of equilibrium [33, 34]. So one needs to know how much time is required to reach equilibrium, i.e. one needs knowledge of sorption kinetics to be able to distinguish between irreversibility and nonequilibrium. A third method, the gas-purge method [21, 25, 35], seems to be a suitable method for measuring sorption kinetics in sediment/water systems. The advantage of this technique is that it does not lead to experimental artefacts generated by incomplete phase separation. In a gas-purge experiment gas is bubbled through a solution or sediment suspension containing chemicals. Hereby it is assumed that only the dissolved chemicals will be removed from the suspension by volatilization. Thus stripping the chemical from the water phase induces desorption from the sediment. The desorption can then be followed by measuring the amounts of volatilized chemicals (trapping the chemicals in the gas stream) or by following the disappearance of the chemicals from the suspension (measuring the decrease in chemical concentration in the suspension). In the present study the desorption kinetics of some chlorinated benzenes from a natural lake sediment are investigated, using the gas-purge technique. The desorption of these compounds from a natural sediment is examined in three sets of experiments. The sorption-uptake period prior to the desorption experiment as well as the sediment/water ratio of the sediment suspension are varied to examine the influences of these factors on the desorption kinetics.

2 2.1

Materials and M e t h o d s

Sediment

A natural lake sediment from the Netherlands (Oostvaardersplassen) was wet sieved (1 mm) and stored under water at 5 ~ The sediment was oven dried at 200 ~ before use. The organic carbon content was determined and found to be 3 % (w/w) based on dry weight ( _+ 40 % (w/w)). Background contamination was less than 10 ng/g for 1,2,3-tri-, 1,3,5-tri-, and 1,2,3,4-tetrachlorobenzene and less than 1 ng/g for penta- and hexachlorobenzene. 2.3

Chemical Analyses

Water samples of 5 ml were extracted with 1 ml 2,2,4-trimethylpentane and analyzed with GC-ECD. Spiking clean water samples (1 ml) with a solution of the chemicals in hexane resulted in recoveries of more than 95 % for all test compounds used. Sediment suspension samples of 5 ml, with an additional 20 ml of water and 50 ml of hexane, were extracted by heating under reflux (90 minutes). For the extraction of the

82

sediment a mixture of hexane and water was chosen, because aqueous mixtures seem to be the most effective for extraction of halogenated compounds from sediment, as was shown by WAHLE et al. [36]. The hexane layer was removed after centrifugation ( • 900 g, 20 minutes) and concentrated to about 5 ml under a gentle nitrogen stream. For clean-up this extract was eluted first through a silica-H2SO4 column and then through a silica-NaOH column. The sample was concentrated again before analysis. Spiking clean sediment suspension samples (5 ml) with 1 ml of a solution of the test compounds in hexane resulted in recoveries of more than 95 % for all the test compounds. Water samples were analyzed on a TRACOR-550 gas chromatograph equipped with a 63Ni electron capture detector (Cp-Sil-8 CB column, 10 m, film thickness 0.12 pm, inner diameter 0.32 mm, outer diameter 0.45 mm, helium as carrier gas). The detector was connected to a Shimadzu C-RIA Chromatopac integrator. Sediment suspension samples were analyzed on a Carlo Erba HRGL 5300 gas chromatograph equipped with a 63Ni electron capture detector (DBS-column, 30 m x 0.318 mm, film thickness 0.25 am, helium as carrier gas). Samples were analyzed with a Carlo Erba AS-550 autosampler. The detector was connected to a PC for data processing (program: Baseline 810, Waters, Millipore Corporation, Milford, U.S.A.). An external standard method was used for quantification. The detection limit varied from approximately 1 pg for the trichlorobenzenes to approximately 0.01 pg for hexachlorobenzene. Uncertainties in sample volume and GC quantification yielded an estimate of • 10 % error in the calculated concentrations.

Chemicals

1,2,3-tri-, 1,3,5-tri- (Aldrich, Milwaukee, WI), 1,2,3,4tetra-, penta- (Fluka AG, Switzerland) and hexachlorobenzene (BDH, England) were used as test chemicals. All chemicals were more than 95 % pure, as confirmed by GC-ECD. n-Hexane and 2,2,4-trimethylpentane (Baker Inc., Phillipsburg, NJ) were used as organic solvents. 2.2

Research Articles

2.4

Experimental

Figure I gives a schematic representation of the experimental system that was used in the gas-purge experiments (modified from KAVaCKHOFF[21]). The system consisted of one glass tube of + 55 cm containing water (I.D. + 2 cm) to saturate the purge gas (A), and two smaller glass tubes of + 35 cm (I.D. +_ 2 cm) so that the experiments could be run in duplicate (B). The tube bottoms consisted of glass flit, which served to distribute the purge gas uniformly across the tubes and to keep the sediment in suspension. Nitrogen gas was used as purge gas at flow rates of 300 to 500 ml/min. Small glass tubes (I.D. + 0.5 cm) filled with 2 cm TenaxTA adsorbent (60 - 80 mesh, chrompack) were used to clean the gas before purging (C) and to remove the volatilized chemicals from the purge gas (D). The Tenax-TA columns were renewed three times in each of the experiments. After extraction with 2 ml 2,2,4-trimethylpentane, Tenax-TA was analyzed for the total amount of the chemicals adsorbed. At the beginning of the purge experiments 70 ml of an aqueous solution or sediment suspension, containing a mixture of the test chemicals, was added to each of the two glass tubes. At several time intervals 5 ml samples were taken from the aqueous solution or sediment suspension and analyzed for the concentration of the chemicals. The efficiency of stripping the chemicals from the aqueous solution was tested by purging a chemical solution without

ESPR-Environ. Sci. & PoUut. Res. 1 (2) 1994

Research Articles

Sorption Kinetics of Chlorinated Hydrophobic Organic Chemicals, Part II

,