Slow desorption of volatile organic compounds from soil: evidence of desorption step limitations

https://doi.org/10.1016/S0304-3894(98)00266-0Get rights and content

Abstract

Transient adsorption and desorption of 1,2 dichloroethane and toluene on dry Yolo silt loam soil were studied by continuously measuring the composition of the effluent from a soil-packed chromatography column with a mass spectrometer. After obtaining complete breakthrough at approximately 30% relative saturation of one chemical in nitrogen, pure nitrogen feed was initiated and maintained for several hours. Of the material adsorbed at breakthrough, 9.7% of the 1,2 dichloroethane and 14.2% of the toluene were highly resistant to desorption and remained sorbed on the soil even after 5 h of nitrogen flow. When a second chemical with a higher adsorption affinity was introduced into the soil column (water following toluene or toluene following 1,2 dichloroethane), the majority of the first chemical was quickly desorbed and began leaving the soil column before breakthrough of the second chemical. Conversely, when a second chemical with a smaller adsorption affinity was introduced into the soil column, only a small amount of the first chemical was displaced and began leaving the soil column after breakthrough of the second chemical. The results of this study indicate that the desorption step itself may be the rate-limiting step for sorbate which remains after prolonged exposure to sorbate-free gas.

Introduction

In the vadose zone, vapor-phase sorption is an important process which controls the movement of volatile organic compounds (VOCs). Recently, a number of studies have found that after a VOC adsorbs on soil, a significant portion does not readily desorb 1, 2, 3, 4. This phenomena, which we refer to as slow desorption, has been described in the literature by a variety of other terms such as recalcitrant, rate-limiting, slowly reversible, nonequilibrium, hysteresis, and irreversible adsorption.

At this time, the majority of the studies in the literature dealing with slow desorption have used soils that are either water saturated or contain significant amounts of water. An excellent review of these studies is presented by Pignatello and Xing [1]. Although studies dealing with dry soils are less abundant, slow desorption has been observed with dry soils. In this paper, dry soil is used in order to avoid the complications introduced by the presence of water so the mechanism of slow desorption in dry soils may more clearly be understood.

In the literature, slow desorption on dry soils has been attributed to either a slow intraparticle diffusion process or the actual desorption step. Although possible in dry soils, slow intraparticle diffusion is more likely to be responsible for slow desorption in water saturated soils because liquid-phase diffusivities are approximately four orders of magnitude smaller than vapor-phase diffusivities [5]. This difference was clearly seen by Grathwohl and Reinhard [2]who found that even after a long venting time, the removal rate of VOCs from columns of wet soil was independent of the carrier stream's flow rate while for oven-dry soil it was proportional to the carrier stream's flow rate. Intraaggregate limitations were much greater in the wet soil columns than in the dry soil columns. It has also been shown that the micropores of many clays have sheets of molecular dimensions [6]which could severely restrict diffusion leading to diffusion limitations. Satterfield et al. [7]showed that a molecule to pore diameter ratio of 0.2 can cause a 59% reduction in the diffusion rate. Differences in sorbate molecular sizes may also be important in the lamellae region. Keyes and Silcox [8]suggest that dodecane does not penetrate the lamellae of montmorillonite but toluene does. For this reason, toluene was thought to desorb more slowly, perhaps due to a chemical adsorption process or restrictive diffusion in the lamellae region.

Tognotti et al. [3]found a considerable amount of slow desorption of toluene vapor from oven-dry Spherocarb (Foxboro Analabs). Montmorillonite was also found to exhibit slow desorption of toluene, although to a lesser extent while Carbopack did not.

Using carbon tetrachloride and the same sorbents, Spherocarb was again found to have a significant amount of slow desorption but montmorillonite and Carbopack did not. Tognotti et al. [3]found that the slowly desorbing sorbate constituted less than a monolayer of coverage in these sorbents and hypothesized that this sorbate was more strongly bound and therefore more slowly desorbed due to capillary condensation and intercalation of sorbate molecules in micropores.

Another possible cause of slow desorption is kinetic limitations from certain adsorption sites. Pinnavaia and Mortland [9]have shown with spectroscopic evidence that toluene adsorption on montmorillonite clays is caused by chemical adsorption due to coordination through the π-electrons of the aromatic rings [9]. It has also been found that clay surfaces may behave as Lewis acids and therefore can strongly interact with organic species containing either double bonds or aromatic rings [10]. Intercalated organic species have also been shown to often exhibit multiple types of adsorption sites with different bonding [11]. The higher adsorption energies associated with chemical adsorption could result in much slower desorption compared to physical adsorption.

In a spectroscopic study of the rate of accumulation of 1,2 dichloroethane (DCA) on dry montmorillonite and dry kaolinite, Aochi and Farmer [4]showed that the DCA becomes associated with the minerals in both a liquid and a labile vapor state. It was further shown that there is a third sorbed species which accumulates at a much different rate than the other two sorbed species. They hypothesized that this additional sorbed species is sorbed in areas of the mineral to which access was limited. For this reason, accumulation of this species was somewhat delayed and increased with time even after desorption was initiated [4].

Dogu et al. [12]found a significant amount of slow desorption when studying benzene adsorption on dry soil. They determined, based on adsorption energy calculations, that the reversible fraction of the adsorption was physical while the irreversible binding was caused by either chemisorption or entrapment in soil micropores.

Keyes and Silcox [8]studied the desorption of toluene from montmorillonite particles. Toluene exhibited a considerable amount of slow desorption which they concluded was caused by either a slow desorption process related to breaking chemical bonds in the lamellae region or by a slow diffusion process. They concluded that local desorption kinetics in pores larger than those in the lamellae region was not rate-limiting because the desorption was found to be independent of particle size.

At this time, the rate-limiting mechanism which causes slow desorption in dry soil, is poorly understood. However, the importance of slow desorption can be seen from a study in which 1,2 dibromoethane (EDB) was detected in agricultural topsoil 19 years after application even though it is degradable and volatile [13]. Although this study involved soil with significant amounts of water, the implications of organic chemicals remaining decades after exposure are clear.

In this paper, the fundamental cause of slow desorption from dry soil is investigated. Adsorption of two common VOCs, 1,2 dichloroethane (DCA) and toluene, on dry Yolo silt loam soil was studied using both batch and chromatographic experiments. The effect of exposing a soil with a slowly desorbing VOC to a second VOC or water is also examined.

Section snippets

Soil and sorbates

The Yolo silt loam soil used in all experiments was obtained from a site near Davis, CA which has never been treated with pesticides or other organic chemicals. This soil is a mixed fine-silty, non-acidic, thermic, typic xerorthent. The characteristics of this soil as determined by the Division of Agricultural and Natural Resources (DANR) Analytical Laboratory at the University of California at Davis and from Amali et al. [14]are given in Table 1. The soil was first ground to pass through a 2.0

Microbalance equilibrium studies

After exposing a virgin soil to a given partial pressure of a VOC and then allowing the VOC to desorb, it was found some of the adsorbed VOC did not desorb. When exposed a second time to the same partial pressure, the soil was able to adsorb only as much as was desorbed after the first exposure. Fig. 3 shows the adsorption isotherm when virgin soil was exposed to toluene and the apparent isotherm for the second exposure. The term `apparent isotherm' is used because this isotherm uses the

Conclusions

The data presented in this work demonstrate that desorption of toluene, water, and DCA on Yolo silt loam soil occurs on two distinct time scales, one very fast and one much slower. Upon first exposure, up to 16.3% of the adsorbed sorbate is involved in a slow desorption process. Subsequent exposures do not cause a significant increase in the amount of slowly desorbing sorbate. The sorbate retained due to slow desorption is highly resistant to desorption into dry nitrogen or a vacuum.

Acknowledgements

This publication was made possible by grant number 2 P42 ES04699 from the National Institute of Environmental Health Sciences, NIH. We would also like to thank Eric Soroker for his assistance with the microbalance kinetic experiments.

References (21)

  • S. Amali et al.

    J. Haz. Mat.

    (1994)
  • J.J. Pignatello et al.

    Environ. Sci. Technol.

    (1996)
  • P. Grathwohl et al.

    Environ. Sci. Technol.

    (1993)
  • L. Tognotti et al.

    Environ. Sci. Technol.

    (1991)
  • Y.O. Aochi et al.

    Environ. Sci. Technol.

    (1995)
  • W.J. Lyman, W.F. Reehl, D.H. Rosenblatt, Handbook of Chemical Property Estimations Methods, ACS, Washington DC,...
  • R.M. Barrer, Zeolites and Clay Minerals as Sorbents and Moleular Sieves, Academic Press, London,...
  • C.N. Satterfield et al.

    AIChE J.

    (1973)
  • B.R. Keyes et al.

    Environ. Sci. Technol.

    (1994)
  • T.J. Pinnavaia et al.

    J. Phys. Chem.

    (1971)
There are more references available in the full text version of this article.
View full text