The Phylmar Group, Inc.
This study describes a comparison of worker exposure to total and inhalable dust, inorganic arsenic, and borates using two types of particulate sampling assemblies as part of a comprehensive industrial hygiene evaluation in a borate mining and processing facility. Employees were segmented into similar exposure groups (SEG) based on work location within the facility, job classification, and type of chemical agent. Approximately 10% of the employees from each SEG wore two personal sampling devices simultaneously for the purpose of collecting total and inhalable particulate fractions using a closed face, 37mm mixed cellulose ester matched-weight filters (MMW) and Institute of Occupational Medicine (IOM) sampling assembly, respectively. Sample results indicated that the IOM concentrations were consistently higher than the corresponding MMW concentrations for all three agents. An analysis was performed to investigate a relationship between MMW and IOM. The data revealed correlation coefficient values of 0.72, 0.82 and 0.84 for total dust (n=197), inorganic arsenic (n=137), and borates (n=194), respectively. These positive correlation coefficients indicate that the IOM and MMW measurements are consistent with each other, and can be used for predicting exposure levels. The total dust and borate large mean ratios should be considered in developing inhalable fraction-based regulatory standards.
With the introduction of the Institute of Occupational Medicine (IOM) sampling assembly for inhalable dusts, questions have been raised as to whether the 37mm cassette assembly or the IOM sampling assembly is better at approximating the amount of particles that enter the respiratory tract, and whether exposures can be estimated when comparing the results between the sampling assemblies.
Currently, most Occupational Safety and Health Administration (OSHA) and National Institute of Occupational Health (NIOSH) sampling methods for particulates rely on the use of a 37mm filter cassette assembly. In particular, arsenic and borate sampling requires the use of a 37mm 0.8 micron mixed cellulose ester (MCE) filter in-line with personal sampling pump, whereas total dust sampling requires a 37mm 5.0 micron polyvinyl chloride (PVC) filter in-line. However, when sampling arsenic, borate, and total dust simultaneously, it is recommended that a 37mm MCE matched weight sampling (MMW) assembly is used. The MMW allows for the gravimetric analysis of total dust, as well as the speciation of arsenic via atomic absorption spectrophotometry, and boron via inductively coupled plasma.
Inhalable particles form the fraction of particulates with diameters of 100 microns or less, which can be inhaled through the upper respiratory system. This range of particulates can actually be deposited anywhere in the respiratory tract. The IOM sampling assembly was designed to simulate particle collection behavior that occurs during breathing.
The IOM sampler is a personal sampling device that has a cylindrical body 37mm in diameter and 27mm long. The sampler has a 15mm circular orifice with a thin lip protruding outward, minimizing sample variations due to loss of particulates from the outer surfaces of the sampler. The orifice and lip are an integral part of the filter cassette assembly. When analyzing a total dust sample, the filter and cassette are weighed together before and after sample collection. All of the particles that are deposited on the filter and filter cassette are analyzed by the laboratory. The MMW sampling assembly is a 37mm plastic cassette containing two MCE filters supported on a cellulose pad (see Figure 1). During gravimetric analysis, the weight of the unexposed second filter is tared from the filter that is exposed to the environment, thereby minimizing the effect of the hygroscopic nature of MCE filters. At 5mm, the orifice of the MMW is substantially smaller than that of the IOM. When collecting a total dust sample with a MMW cassette, only the filter is weighed before and after collection. Therefore, any particulate that has been deposited on the surface of the cassette wall will not be analyzed, resulting in an underestimation of exposure.
Samples were colleted at U.S. Boron facility, a major boron mining, processing, and distribution center. Employees in the Boron, California facility were grouped in accordance to similar exposure groups (SEGs). A SEG is a group of employees with similar exposure to an environmental agent. Owing to the exposure similarity within each group, a subgroup of randomly selected individuals, representative of the exposure distribution, were used to evaluate trends within each SEG.
SEG characterization was based on three parameters: 1) work location (department) within the facility, 2) job classification, and 3) chemical agent monitored. All monitored employees within a given SEG were selected randomly by U.S. Borax personnel.
Each selected employee was monitored for personal airborne exposures to total dust and borate. In addition, if an employee had a potential exposure to arsenic, then he/she was monitored for inorganic arsenic along with total dust and borate. Personal air samples were collected by attaching a MSA Escort personal sampling pump to each monitored employee’s belt. The sampling pump was calibrated with an in-line 37mm closed-face 0.8-micron MMW sampling cassette assembly to a flow rate of 2 liters per minute (lpm) using a Gilibrator, a primary calibration standard. Borates, total dust, and arsenic were collected simultaneously on a MMW sampling cassette. During sample collection, the MMW was positioned in the breathing zone of the employee.
Approximately 10% of the employees in each SEG were required to wear the MMW and IOM filter assemblies simultaneously in order to compare the variation in sampling methods. The personal IOM air samples were collected by attaching a MSA Escort personal sampling pump to each monitored employee’s belt. The sampling pump was calibrated with an in-line 25mm 0.8 micron IOM sampling assembly to a flow rate of 2 lpm with a Gilibrator. Borates, total dust, and arsenic were collected simultaneously on the IOM that was placed in the breathing zone of the employee.
Employees were monitored for at least 7 hours of an 8-hour workshift. Whenever a workshift exceeded 8 hours, monitoring time was extended so that a sample representative of 88% of the work shift was collected. The employees had "zero" or negligible exposure during non-sampled time.
U.S. Borax employees performed routine duties and tasks that were assigned at the plant. The monitored employees were observed throughout the sampling period in order to collect detailed information regarding exposure conditions, such as daily tasks conducted, production/equipment run, production equipment used, any visible elevated dust conditions, personal protective equipment worn, engineering controls in place, and the predominant species of borate dust to which the worker was exposed. These data were collected to document and characterize the sampling conditions for each sample.
All of the MMW and IOM samples were sent to NATLSCO, a laboratory accredited by the American Industrial Hygiene Association. Arsenic, boron, and total dust were analyzed in accordance with NIOSH method 7900 (atomic absorption with arsine flame generation), OSHA 125G (inductively coupled plasma), and modified NIOSH method 0500 (gravimetric analysis), respectively. NIOSH method 0500 was modified for total dust sampling in order to allow for the collection of arsenic, boron and total particulates on the same filter. Unlike a PVC filter that is specified in NIOSH 0500, a MMW filter can easily be digested for additional agent speciation of a total particulate sample.
RESULTS AND DISCUSSION
Laboratory results for arsenic, boron, and total dust were provided in micrograms (mg) of agent detected on the filter. A borate correction factor was applied to the boron analytical results in order to take into account the hydration state of the borate species monitored. This information was used to calculate an 8-hour time-weighted average (TWA) for each employee. All results presented in this section are based on 8-hour TWA values.
Pairs of MMW and IOM measurements collected in each similar exposure group (SEG) were compared to investigate a relationship between MMW and IOM so that observed MMW values could be converted to equivalent IOM values. Total dust and borate personal exposure data were collected for 197 U.S. Borax employees. Of these employees, 137 were also monitored for inorganic arsenic.
The ratio of IOM/MMW, or the slope (S), has been used to measure the relationship between IOM and MMW, assuming that individual exposure measurements are random and normally distributed. The personal exposure results from industrial hygiene sampling methods using MMW and IOM are always positive. Particulate deposition results often follow a lognormal distribution. Therefore, prior to using the S measurement, MMW and IOM data for the total dust, inorganic arsenic, and borate were examined in order to assess whether a normal or lognormal distribution should be applied to the data sets. It was found that the statistical distributions for MMW and IOM for the three agents were closer to a lognormal than to a normal distribution. Thus, for the analyses it was assumed that MMW and IOM are lognormally distributed. Under this assumption, natural logarithms of the measurements MMW and IOM can be assumed to follow normal distributions. In this study ln(MMW) and ln(IOM) are represented by MMWL and IOML, respectively.
To investigate a predictive relationship between MMW and IOM, linear regression analyses were considered on the pairs (MMWL, IOML). However prior to performing the regression analyses, pairs (MMWL, IOML) were evaluated statistically for identification of outliers. Outlier pairs were identified by using the characteristic of a standard normal distribution (Z, with mean = 0, standard deviation =1) where more than 99% probability accounts for Z values within this interval (-3,3) around the means. To apply this characteristic, MMWL and IOML were standardized so that they are comparable to a standard normal distribution. These standardized values were labeled as MMWz and IOMz. Then the pair (MMWL, IOML) was identified as an outlier at a 99% level of confidence. Identified outliers from the data of the three agents were excluded from the regression analysis. The results are shown in Table 1.
The least square method was used to predict average IOML given a MMWL value representative of a SEG. The predicted value IOMp is derived from the linear regression equation IOMp = a + b MMWL, where ‘a’ is the regression intercept and ‘b’ is the regression coefficient representing the slope of the regression line. Multiple correlation coefficients (R) between zero and one from these analyses were used to assess the strength of the linear regression relationship between IOML and MMWL. A value of R closer to one indicates that prediction of IOML given MMWL can be made with a high degree of confidence. Number of employees, the intercepts, regression coefficients, and multiple correlations are shown in Table 2. These regression equations and plots are shown in Figures 1 through 3.
A reverse transformation of the regression equation to the original scale of data shows that the given MMW value of the SEG and the average predicted IOM for the SEG can be expressed by exp [a+b ln (MMW)]. These equations are shown in Table 2.
Table 1: Identification of Outliers
|Number of Employees in SEG||197||143||197|
|Number of Outliers Identified||5||3||3|
Table 2: Regression Analysis and Prediction Equation
Prediction Equation: IOMp = a MMWb where IOMp = Predicted IOM given MMW = X
|Number of Employees||192||143||194|
|Regression Intercept ‘a’||1.1496||-0.1552||1.8254|
|Regression Coefficient ‘b’||1.0071||0.8344||0.8314|
|Prediction Equation for IOM given MMW||exp [1.1496 + 1.0071 ln(MMW)]||exp [-0.1552 +0.8344 ln (MMW)]||exp [1.8254 + .8314 ln (MMW)]|
· The exposure monitoring results are linear in the logarithmic scale.
· A regression equation can be used to predict an average IOML for a given MMWL .
· Factors that may contribute to the variability of monitoring results (i.e. placement of sampling mechanism, potential splash situations, etc.) should be considered when developing inhalable fraction based regulatory standards.
This article reprinted from Biological Trace Element Research, Vol. 66, 1998, with permission from Humana Press, Inc.