New Air Sampling System for Long Term Sampling

摘要

A continuing challenge in occupational hygiene is that of accurately estimating a worker's exposure to the multitude of airborne chemicals found in the workplace and surrounding community. Hence, the development of new methods that will permit more effective sampling of contaminants in the workplace is essential to ensure that accurate exposure assessments are completed. Over the last five years a capillary flow controller was designed and tested in laboratory and field studies. The very low flow rate provided by the capillary flow controller permits a sample to be collected over an extended period of time when used with a small portable evacuated canister. The initial research focused on the development of the capillary-canister and the evaluation of its ability to collect representative air samples. However, capillary flow controllers exhibit a characteristic drop in flow rate with time. The time dependent flow rate coupled with concentration changes in the sampling environment could result in a sampling bias. The focus of this study was performance evaluation based on field studies and prediction of the sampling bias. As part of this study, the capillary canister system was redesigned to make it less expensive, more compact, and more acceptable to the person wearing the sampler. The performance of the capillary canisters was evaluated by conducting field experiments at two different locations. Twenty-four paired samples were collected using canisters and diffusive badges. The paired t-test performed on the personal samples showed no statistical difference between the two methods. The model developed to predict the sampling bias associated with capillary canisters was experimentally validated. A series of predicted worst-case scenarios involving peak concentrations were experimentally simulated at three sampling time durations: 8 hrs, 24 hrs and 40 hrs. Three canisters, a gas chromatography (GC), and photo ionization detector were used to monitor the concentration profile during the experiments. The bias between the canisters and the GC was calculated from the experimental values. When analytical errors were considered, the experimental values approached the predicted model values.