Aarhus Universitets segl

No. 128: Passive sampling of polychlorinated biphenyls (PCB) in indoor air: Towards a cost-effective screening tool

Vorkamp, K. & Mayer, P. 2014. Passive sampling of polychlorinated biphenyls (PCB) in indoor air: Towards a cost-effective screening tool. Aarhus University, DCE – Danish Centre for Environment and Energy, 118 pp. Scientific Report from DCE – Danish Centre for Environment and Energy No. 128. http://dce2.au.dk/pub/SR128.pdf

Summary

Polychlorinated biphenyls (PCBs) are non-flammable and chemically stable substances with a large variety of former industrial and commercial applications.  Their production and use peaked in the 1960s and 1970s, concurrently with a time period of high building activity in Denmark. Although ultimately banned in Denmark in 1986, PCBs are still being released from construction materials today, thus potentially causing exposure of people staying in buildings with PCB containing materials. The objective of this study was to develop a robust and inexpensive screening tool based on passive sampling, which allows an initial assessment of indoor air concentrations in relation to the cut-off values of 300 and 3000 ng PCB/m3 air set by the National Board of Human Health.

The work proceeded in three phases combining a literature review, laboratory studies and field work. The literature review of the suitability of passive sampling formats for the purpose of this project resulted in recommendations of semipermeable membrane devices (SPMDs), silicone-coated vials and stir bar sorptive extraction (SBSE) for further work. As the use of performance reference compounds (PRCs) in an indoor setting was declined SPMDs were eventually deselected and replaced by silicone-coated petri dishes.

The second project phase focused on the effect of non-standardized uptake conditions in terms of variable air velocities on the partition kinetics of PCBs between air and the sampler, on detection limits, precision and questions of practical handling. Elimination experiments showed a substantial dependency of elimination rates on air velocity: Variations from 0.1 to 0.3 m/s or 0.3 to 1 m/s led to changes in elimination rates by up to a factor of 3, with implications for accuracy. Detection limits and precision were considered satisfactory, i.e. detection limits were generally < 1 ng/m3 or even < 0.1 ng/m3 based on uptake rates from the literature and an injection volume of 0.1 % of the final extract. Relative standard deviations between duplicates were generally < 10 %, with few exceptions.

Results of the first two project phases were presented to experts in the field of PCB containing construction materials, e.g. consulting engineers, laboratories, researchers and other interested parties, who were supportive of a third project phase with focus on measurements in buildings with potential PCB sources. There was consensus to test the silicone-coated petri dishes as a kinetic sampler as well as silicone-coated paper sheets with a view to equilibrium sampling.

These two formats, i.e. petri dishes with a thin layer of silicone and silicone-coated paper, were tested in the third project phase, in terms of two time series and ten 24-hour-measurements in buildings alongside conventional active measurements taken by the companies Rambøll and Grontmij A/S. The sampling experiments were run with several replicates which confirmed the high precision observed in project phase 2. The time series showed linear uptake of lower chlorinated PCB congeners on the silicone in petri dishes. The micrometer thin silicone layer on paper reached equilibrium between 1-10 days, depending on the PCB congener. This means that the 24-hour-measurements still covered the kinetic phase and sampling periods of 1-2 weeks will be required for equilibrium sampling. The samplers were calibrated by comparison with results from five active measurements, resulting in sampling rates for CB-28, CB-52, CB-101, CB-118 and CB-153. Using these sampling rates, concentrations were calculated for the remaining five locations and compared with the results of the active measurements (validation). The kinetic sampling showed a tendency of overestimating the concentrations obtained by active sampling, by up to roughly a factor of 3. For two of the petri dishes, the concentrations of the active measurements were exceeded by a factor of 10. It was possible to calculate low concentrations of CB-118 and CB-153, which were below detection limits in the active measurements.

The results of the third project phase were presented to a group of experts again, with the main conclusions that both methods were precise and the silicone-coated paper in particular showed high sensitivity, but more work would be needed to evaluate the paper under equilibrium conditions. Both passive samplers are generally robust and easy to handle. Accuracy remains the main challenge, but might be considered acceptable for the purpose of an initial screening.