Industrial hygienist generally place air sampling probes on the mid to upper torso of workers when attempting to determine their inhaled concentrations. A review of the literature does not supply convincing evidence that concentrations measured on the chest are equal to inhaled concentrations. To determine how well surrogate locations matched concentrations at the mouth (Cmouth), low concentrations of ethanol in nitrogen were released between the hands of standing and seated subjects while they sat or stood in a wind tunnel. Experimental results for a manikin were compared to experimental results for human subjects and to concentrations predicted from simulations made using computational fluid dynamics. In all cases the subjects stood or sat with their backs to the airflow because previous studies had established that exposures facing downstream produced exposures that were more than 100 times exposures when facing upstream or sideways to the flow with the source in the subjects' hands. Sampling probes were placed at the subject's mouth, nose, forehead, neck, both collars, center chest and both lapels. Airflow was drawn to 3L sampling bags using sampling pumps at 0.1L/min. Concentrations were measured using a gas chromatograph with a photo-ionization detector, which was calibrated daily over the range of sample concentrations. Test conditions included 5 levels of cross-draft velocities (11, 27, 48, 82 and 104 ft/min), two levels of body heat (unheated/heated), two levels of hair length ("2" and bald), and two levels of posture (sitting/standing). The manikin was anthropometrically correct and could be heated to natural temperatures. In addition, body-temperature air could be "inhaled" and "exhaled" through its nose in a realistic manner. The manikin either wore a wig or was "bald." The human subjects sat or stood while doing make-work tests that involved moving their hands over the tracer gas source. Three dimensional transient computational fluid dynamics (CFD) was used to simulate exposure conditions for the manikin with and without heating. For the manikin tests, wind tunnel velocity, heating, and posture each had a statistically significant effect for all sampling locations. Hair length had a significant effect for some locations but not others. For the unheated manikin, concentrations for all sampling locations declined monotonically with wind tunnel velocity. However, for the heated conditions, concentrations varied with an inverted- V relationship with wind tunnel velocity. For heated conditions, concentrations at the mouth were always higher for standing than sitting. Concentrations measured at the chest and shoulder levels were higher than mouth concentrations for the standing posture and were lower than mouth concentrations for sitting. Concentrations measured at the forehead location were always lower than concentrations measured at the mouth for both sitting and standing. For the 13 human subjects the test conditions included cross-draft velocities of 11, 27, 48, and 103 ft/min, breathing/not breathing, and sitting/standing. Effects of not breathing were simulated by having the subject breathe through a long tube. Results for the human subjects were consistent with the results for the manikin in some ways but not others. As with manikins, concentrations varied with wind tunnel velocity (p< 0.01) in an inverted V-shape and posture was significantly related to concentrations at all locations (p<0.01). However, unlike manikins, whose concentrations were roughly twice as high for standing as sitting, the ratio was inverted for humans. Breathing through a tube appeared to reduce the concentrations at most sampling sites but was not statistically significant. This appears to contradict the finding that breathing was important for manikins. However, the lack of statistical significance for humans may be due to the higher variability in results across subjects for humans than was found for manikins. If diverse manikins had been used, they may or may not have exhibited the same variability. Concentrations for different human subjects ranged over a 4 to 1 range for the same conditions. It is likely that differing body sizes, body shapes, and lengths of hair contributed to the differences among subjects. Concentrations at the mouth were often substantially different from concentrations elsewhere. Concentrations at the humans' foreheads averaged about 85% of the concentration at the nose. The center chest averaging 136% of the nose. However, the 10% of the ratios of center chest to mouth were below 58% and the 10% were above 259%. The lapels were only slightly better. The corresponding values for collars were 75% and 126%. For the neck it was 80% and 140% and for the Nose it was 80% and 106%. Ratios of concentrations measured below the collars were much more variable than the ratio of concentrations at the nose to the mouth. The geometric standard deviation for Cchest/Cmouth was 2.12, a disturbingly high value. The dramatic effects of velocities and postures suggest that exposure studies should consider multiple postures and cross-draft velocities. For human subjects, the forehead, collars, and adjacent to the nose were good surrogates (i.e., concentrations mostly within +20% of the mouth). The nose and forehead could be corrected by fixed amounts to estimate concentrations at the mouth. The collars were more accurate without correction but had higher standard deviations (i.e., lower precisions). The CFD results closely matched the manikin results for the same conditions. In addition, the following was noted: 1) The heat flux from the body significantly affects the flow field and the subsequent contaminant concentration field at low Reynolds numbers; 2) the free stream turbulence plays an important role in the variation of exposure measurements at low Reynolds numbers; 3) results calculated with the Large Eddy Simulation (LES) illustrate the turbulence structure in the wake of the manikin and indicate that the flow unsteadiness plays an important role in the variation of exposure measurements; 4) calculations with various body shapes suggests that oversimplified body shapes may lead to inaccurate predictions in worker exposure assessment; 5) the concentrations measured at the lapel were very different than the concentrations measured near the mouth. The results of this study raise strong concerns about the accuracy of samples taken anywhere below the collars when the subject is close to the source and flow is from the rear. The results from the manikin also cast strong doubts on the use of a manikin as a surrogate for humans unless it is heated and breathes realistically. The dramatic effects of posture on concentrations suggest that exposure studies should include both sitting and standing as well as variations of each. It is troubling that the effect of posture was reversed for humans and the manikin. The effects of body size and shape and hair length should be investigated in future studies.
Steven E. Guffey, Ph.D., CIH, West Virginia University, College of Egineering and Mineral Resources, Dept. of Industrial Management and Systems Engineering, PO Box 6070, 353C Mineral Resources Engineering, Morgantown, WV 26506-6070