Ceramic Dosimeter Design

The ceramic dosimeter consists of a ceramic tube with a diameter of 1cm and a wall thickness of 1.5 mm (manufactured by UFS Schumacher, Crailsheim, Germany). The pore size of the inner wall coating of the tube is 5 nm. The length of the tube can be varied, but in most applications so far a length of 5 cm was used. The tube is filled with water-saturated sorbent and closed at either end with caps made of, e.g., polytetrafluoroethylene (PTFE) (Fig. 12.1).

12.2.1 Ceramic membrane

The ceramic membrane serves both as a diffusion barrier and as a container to hold the sorbent material. Key characteristics of this membrane with regard to quantitative time-integrated passive sampling are its porosity and inertness, its inner pore size and its thickness. Porosity and inertness

The porosity (e) of the ceramic membrane was measured by Piepenbrink [4] by means of a capillary pyknometer and determined to be 0.305 (or 30.5%). The membrane can easily be saturated with water. Water saturation does not lead to swelling that can occur with some organic polymers. With the high porosity and water saturation, a steady-state concentration profile of chemicals within the membrane

Fig. 12.1. Dosimeter design and cross section through the ceramic tube filled with sorbent material.

can be accomplished quickly with no significant lag times relative to the usual sampling periods, which are typically several weeks to months. This is also related to the inertness of the membrane, which is typical for ceramics in general. Piepenbrink [4] modeled the breakthrough of phenanthrene through the ceramic membrane and found a quasi-steady-state to develop within an hour of exposure. Pore size

The inner wall of the ceramic membrane has a pore size of 5 nm. With this characteristic, the ceramic membrane prevents microorganisms from entering the interior of the ceramic tube (cut-off for microorganisms is a pore size of about 200 nm). The small pore size also minimizes flow of water and limits solute transfer to diffusion only. Thickness

The wall of the ceramic tube is 1.5 mm thick. With this thickness, the ceramic tube forms a rate-limiting diffusion barrier because it is larger than any diffusion boundary barrier so far described under environmentally relevant conditions. Thus, even under low-flow conditions, where a significant boundary layer may form outside of the sampling device, the sampling behavior can be assumed to be determined by diffusion through the ceramic membrane alone. Gale [7], for example, reported an aqueous boundary layer thickness of 100-400 mm in a quiescent aqueous system. This is about 1/15 to 1/4 of the thickness of the ceramic membrane. The great advantage of the ceramic membrane therefore is that uptake of chemicals is predictable based on its characteristics. Along these lines, uptake of chemicals is independent of hydrodynamic flow because effects of flow velocity on the diffusion boundary layer around the sampling device have no impact on the diffusion barrier formed by the ceramic membrane. Taken together, the thickness of the ceramic membrane forms the basis for quantitative analysis of TWA contaminant concentrations without the necessity to calibrate for chemical uptake in the laboratory or otherwise account for varying exposure conditions.

12.2.2 Sorbent material

Upon diffusion of solutes through the ceramic membrane, they need to be trapped onto a receiving phase. This step removes the chemical from the aqueous phase inside the ceramic tube. The role of this is to maintain a steep concentration gradient between the exterior and the interior of the sampler in order to ensure continuous diffusion along the gradient into the sampling device. Thus, the sorbents need to have a high affinity as well as a high capacity for the uptake of the chemicals to be sampled. In the ceramic dosimeter, this is accomplished by means of solid sorbent beads. In as much as the ceramic dosimeter is used under water-saturated conditions, the beads need furthermore to be easily wetted by water and non-swelling. Additionally, in order to allow for chemical analysis after sampling, they are required to yield high recovery rates of the target chemicals by means of solvent extraction or, potentially, thermo-desorption.

Three different bead materials have thus far been identified to fulfill these criteria. These are Amberlite IRA-743, which has proven suitable for sampling of PAHs [2-4], Biosilon, which has been applied in the ceramic membrane-based Toximeter for sampling PAHs (see Chapter 18), as well as Dowex Optipore L-493, the sorbent material of choice for BTEX, naphthalenes [5] and CHCs [5,6]. In contrast, activated carbon (F100) as well as XAD resin (XAD8), which sorb many organic chemicals very well, showed either very low recovery rates or cannot easily be handled in water [8].

For chemical analysis of the sorbent materials after sampling, a simple solvent extraction is desired. The materials mentioned above can be extracted two to three times using acetone. The pooled acetone extracts are subsequently spiked with deuterated internal standards for a direct GC-MS analyses [2,4,5]. Based on this method, recovery rates for Dowex Optipore L-493 upon 2 weeks of contact with the chemicals in mixture were 92%, 95% and 96% for BTEX, naphthalenes and CHSs, respectively [5]. Recovery rates for Amberlite IRA 743 upon 4 months of contact with a mixture of PAHs were 101% for naphthalene, 98% for acenaphthene, 100% for fluorene, 97% for phenanthrene, 94% for fluo-ranthene, 83% for benzo[a]anthracene and 63% for benzo[a]pyrene (n — 3 for each PAH; [3]). Although recovery rates have not explicitly been tested in the laboratory upon exposures beyond 4 months of contact time, the results obtained in field exposures, using Dowex Optipore L-493 for 3 months [5] or Amberlite IRA 743 for up to 12 months [2], are in support of the high extraction efficiency and recovery rates obtained in the laboratory.

12.2.3 Determination of time-weighted average chemical concentrations

The diffusive flux of chemicals across the ceramic membrane is defined by Fick's first law of diffusion [9]. This law states

0 0

Post a comment