Exposure of Miners to Diesel Exhaust Particulates in Underground Nonmetal Mines

ABSTRACT 

A study was initiated to examine worker exposures in seven underground nonmetal mines and to examine the precision of the National Institute for Occupational Safety and Health (NIOSH) 5040 sampling and analytical method for diesel exhaust that has recently been adopted for compliance monitoring by the Mine Safety and Health Administration (MSHA). Approximately 1000 air samples using cyclones were taken on workers and in areas throughout the mines. Results indicated that worker exposures were consistently above the MSHA final limit of 160 micrograms/m3 (time-weighted average; TWA) for total carbon as determined by the NIOSH 5040 method and greater than the proposed American Conference of Governmental Industrial Hygienists TLV limit of 20 micrograms/m3 (TWA) for elemental carbon.

A number of difficulties were documented when sampling for diesel exhaust using organic carbon: high and variable blank values from filters, a high variability (+/- 20%) from duplicate punches from the same sampling filter, a consistent positive interference (+26%) when open-faced monitors were sampled side-by-side with cyclones, poor correlation (r 2 = 0.38) to elemental carbon levels, and an interference from limestone that could not be adequately corrected by acid-washing of filters. The sampling and analytical precision (relative standard deviation) was approximately 11% for elemental carbon, 17% for organic carbon, and 11% for total carbon. An hypothesis is presented and supported with data that gaseous organic carbon constituents of diesel exhaust adsorb onto not only the submicron elemental carbon particles found in diesel exhaust, but also mining ore dusts. Such mining dusts are mostly nonrespirable and should not be considered equivalent to submicron diesel particulates in their potential for adverse pulmonary effects. It is recommended that size-selective sampling be employed, rather than open-faced monitoring, when using the NIOSH 5040 method.


METHODS

A sampling protocol initially was developed for use at five non- metal mines, and subsequently two additional salt mines chose to participate using a slightly modified protocol. The protocol for the initial five mines included sampling underground mine maintenance and production workers for 5 consecutive days. Personal and area air samples were taken using cyclone samplers to minimize possible interference with the NIOSH 5040 method by ore dust being mined. A special impactor described in the NIOSH 5040 method was not used, because it was not commercially avail- able at the time of this study.

Twenty to 30 underground workers were monitored each day for a full shift (approximately 8 hours). Prior to entering the mine, workers were equipped with air sampling pumps calibrated at 1.7 L/min connected to a MSA (Mine Safety Appliance Co., Pitts- burgh, Pa.) nylon cyclone equipped with a 37-mm quartz filter. The cyclone was placed in the breathing zone of each worker. Air sampling pumps were calibrated before and after monitoring, and samples that had a postcalibration flow rate that varied by more than 5% were voided.

Approximately 20 production and maintenance workers involved in surface operations at each mine were monitored for two to three shifts during the study. The equipment they wore and the duration of sampling was identical to those of underground workers. An effort was made to sample surface workers who had occasional exposure to diesel exhaust, but most workers had no routine exposure to diesel exhaust.

Air sampling baskets were placed in various areas of the mines. Five or six air sampling pumps, each equipped identically to the personal air sampling trains, were placed in metal baskets and located approximately 4 to 6 ft off the ground. The purpose of these baskets was to measure the precision of the NIOSH 5040 method in each specific mining environment.

Some baskets included an additional air sampling pump equipped with a closed-faced poly- vinyl chloride filter cassette to measure total dust levels. Field blanks were also put into air sampling baskets and consisted of an air sampling pump equipped with a cyclone and quartz filter, but these pumps were never turned on. Air samples from area baskets were typically taken for 6 to 8 hours during a shift. All pumps were calibrated prior to sampling, and samples were voided if post- calibration flow rates varied by more than 5%.

A total of 48 underground mine workers were asked to wear two air sampling pumps, each equipped with a cyclone and quartz filter. The purpose was to examine differences due to variations inherent in the 5040 method and microenvironmental differences, either of which could cause imprecision in estimating exposure values for these workers.

To evaluate the NIOSH 5040 method protocol for DPM monitoring, nine air-sampling pumps with open-faced quartz filter cassettes were placed in area baskets along with those equipped with a cyclone and quartz filter. The objective was to evaluate differences in elemental and OC levels due to particle-size selective sampling.

Industrial hygienists who conducted the monitoring remained in the mine for the duration of each shift along with the workers. Efforts were made to monitor air sampling equipment worn by every employee, but some employees worked at remote areas of the mine and were only seen at the beginning and end of each shift. Equipment in area air sampling baskets was monitored periodically during the shift.

Job classifications for underground workers were noted, but for the purposes of this study each worker was placed into one of two categories: maintenance positions (these include shops distant from production or elsewhere) and production workers (primarily working in areas where there was active mining).

Blank filters were submitted to the laboratory at the rate of approximately 10% of the total number of air samples taken. Two types of blank samples were submitted: the field blank previously described, and filter cassettes supplied by the laboratory and submitted along with the field samples. In each case the analytical laboratory personnel were blinded as to whether a filter cassette was a personal, area, or blank sample.

The NIOSH 5040 method is a destructive analytical method, but only requires the use of a small (1.5 cm2) punch from each filter. Submitted to the laboratory were 165 selected samples, with instructions to return 1 or 2 additional punches. Of these punches, 97 were spiked with an amount of sucrose that approximately equaled the amount of OC determined on the filter punch. Due to a lack of a suitable standard, EC was not spiked onto these filter punches. When the duplicate and spiked punches were received from the laboratory, they were given new identification numbers and returned to the laboratory in a blinded manner similar to the field samples. This effort allowed for an evaluation of the precision of the analytical method, uneven deposition of aerosol on the filter, and the potential of ores from the mining environment to interfere with the OC present from DPM.

As previously described, all air samples for diesel particulate were analyzed using the NIOSH 5040 method by an analytical laboratory accredited by AIHA. Total aerosol samples collected on polyvinyl chloride filters were analyzed according to NIOSH Method 0500.


RESULTS

Personal Exposure Monitoring


A total of nearly 800 personal air samples was taken at the seven mines, and the results are summarized in Table II. Data from each facility were examined using a Shapiro-Wilk normality test, and it was determined that the data were neither lognormally nor normally distributed. Therefore, the data are summarized using both arithmetic and geometric means and standard deviations.

The arithmetic mean of EC levels varied from across mines from 41-405 g/m3 for underground miners and 1-39 g/m3 for above-ground miners. The arithmetic mean of OC levels varied from across mines from 45-153 g/m3 for underground miners and 43-79 g/m3 for above-surface jobs. The arithmetic mean of total carbon (TC) levels varied across mines from 86-558 g/m3 for underground miners and 50-118 g/m3 for above-surface jobs. The ratios of mean EC to mean TC levels ranged from 36-75% in underground jobs and from 2-33% in surface jobs. For individual underground samples the EC:TC ratio ranged from 0.7-87%, and for individual above ground samples this ratio ranged from 0.2-80%.

Area Basket Sampling

The results of area sampling at the seven underground mines are shown in Table III. Six to 20 baskets of air samples were taken at each facility, each containing five or more DPM samplers. Efforts were made to take area air samples where diesel exhaust was expected to be present in moderate to high quantities. Results indicated that a few baskets were placed in areas with minimal levels of EC, but most baskets were in areas where DPM levels exceeded the MSHA final standard of 160 g/m3 for TC and the proposed TLV of 20 g/m3 for EC.

Relative standard deviations (RSD: standard deviation/mean, sometimes referred to as coefficient of variation) obtained from area baskets are shown in Table III. For all baskets with four or more valid samplers, the RSD averaged about 11% for TC with a standard error of the estimate (SEE) of 8%, 11% for EC (SEE of 10%), and 17% for OC (SEE of 13%). Total dust levels varied from nondetectable at one mine to as high as 15 mg/m3 at the limestone mine. The ratio of EC to TC varied from 0.02-0.90, probably due in some cases to the fact that samples were taken where little EC was present.


The results of nine sets of air samples from two mines obtained using open-faced filter cassettes placed side-by-side with cyclones are shown in Table IV. The EC results from the two sampling methods were nearly identical, differing by an average of only 2%. By contrast, open-faced samples for OC were at least 7% higher than cyclones for every sample, and they averaged 26% higher over all of the samples.

All air samples were blank corrected for the mean field blank values obtained from each mine. The results for EC and OC blank samples are presented in Table V. In the case of EC, 67-100% of field samples for each mine were nondetectable. OC blanks yielded highly variable values that, with one exception, were all above the limit of detection. The mean OC blank values for the individual mines ranged from 7-21 g/filter with standard deviations rang- ing from 3.4-9.4.

For five of the mines, additional filter punches were analyzed in a single-blinded manner, as previously described. Some of these punches were spiked with sucrose in quantities similar to the levels of OC originally found on the sample. The percentage recovered was then calculated. These data are displayed in Table VI. For Mines 3-5 there were significant amounts of variability between sample punches from the same filter, and this variability was similar in magnitude to the variability observed from different samplers taken in the same area baskets as displayed in Table III. Recoveries of the sucrose spikes were generally close to 100% for samples from three of the five mines.

Forty-eight side-by-side DPM samplers were worn by miners at four of the seven mines. The results of these side-by-side com- parisons are displayed in Table VII. Mean differences between samples taken on the same individual at the same time were 14% for EC and 20% for OC. The ratio of EC to TC varied by an average of 8% in these samples.

