Health
Effects of Occupational Exposure
to
Respirable Crystalline Silica
2.1 Chemical and Physical Properties
In the crystalline state, one silicon atom and four oxygen atoms are arranged in an ordered, repetitive array of three-dimensional tetrahedrons. The silicon atom is the center of the tetrahedron. Each of the four corners consists of a shared oxygen atom.
Exposure to changes in temperature and pressure, either natural or synthetic, may cause the crystalline structure to change [Iler 1979; Klein and Hurlbut 1993; Navrotsky 1994; Hemley et al. 1994; IARC 1997]. An example of a naturally occurring pressure change is the transformation of alpha quartz to coesite in a rock subjected to the impact of a large meteorite [Iler 1979; Klein and Hurlbut 1993; IARC 1997]. Alpha quartz and beta quartz are the respective designations given to the low- and high-temperature crystal structures. Quartz changes from the alpha to the beta form at 573 °C (1,063 °F) [Ampian and Virta 1992; NIOSH 1983a; Virta 1993; Guthrie and Heaney 1995].
The solubility of quartz in water at room temperature ranges from 6 to 11 micrograms per cubic centimeter (µg/cm3) (6 to 11 parts per million [ppm]) as SiO2 [Coyle 1982; Iler 1979]. Quartz is slightly soluble in body fluids, where it forms silicic acid and is excreted by the urinary system [IARC 1987]. The amount of silica dissolved depends on various factors, including particle size, shape, and structure; solution temperature; viscosity; pH; the proportion of dust to liquid; and the presence of trace minerals [King and McGeorge 1938; King 1937; Iler 1979; Wiecek 1988; IARC 1997; Guthrie 1997]. However, the dissolution of quartz does not contribute substantially to its clearance or to changes in its biological activity [IARC 1997; Heppleston 1984; Vigliani and Pernis 1958].
2.2 Number of Workers Potentially Exposed
NIOSH [1991] estimates that at least 1.7 million U.S. workers are potentially exposed to respirable crystalline silica. This estimate is based on information from the National Occupational Exposure Survey (NOES) [NIOSH 1983b] and the County Business Patterns 1986 [Bureau of the Census 1986]. Table 1 lists the nonmining industries (excluding agriculture) and mining industries with the largest numbers of workers potentially exposed to respirable crystalline silica. In addition, an undetermined portion of the 3.7 million U.S. agricultural workers [Bureau of the Census 1997] may be exposed to dust containing a significant percentage of respirable crystalline silica [Linch et al. 1998].
2.3 Dust-Generating Activities, Uses, and Potential Exposures
Crystalline silica (quartz) is a component of nearly every mineral deposit [Greskevitch et al. 1992]. Thus most crystalline silica exposures are to mixed dust with variable silica content that must be measured by dust collection and analysis [Wagner 1995; Donaldson and Borm 1998].
Workers in a large variety of industries and occupations may be exposed to crystalline silica because of its widespread natural occurrence and the wide uses of the materials and products containing it. OSHA compliance officers found respirable quartz in 6,779 personal samples (8-hr TWA) taken in 255 industries that were targeted for inspection (excluding mining and agriculture). In 48% of the industries, average overall exposure exceeded the PEL for respirable quartz [Freeman and Grossman 1995]. Linch et al. [1998] applied an algorithm to OSHA compliance data from the period 1979-1995 and County Business Patterns 1993 data [Bureau of the Census 1993] to estimate the percentage of workers by industry (excluding mining and agriculture) exposed to defined concentrations of respirable crystalline silica (e.g., >0.05 mg/m3) in 1993. Area samples and samples involving complaints to OSHA were excluded from the analysis. Although data limitations could have resulted in underestimating or overestimating the number of workers exposed, the authors found 5 three-digit standardized industrial classification (SIC) codes in which an estimated number of workers were exposed to concentrations at least 10 times the NIOSH REL:
Masonry and plastering Heavy construction Painting and paper hanging Iron and steel foundries Metal services Additional three-digit SICs had a number of workers with crystalline silica exposures that were two or five times higher than the NIOSH REL [Linch et al. 1998].
Table 2 lists the main industries around the world in which silica exposure has been reported. Virtually any process that involves movement of earth or disturbance of silica-containing products such as masonry and concrete may expose a worker to silica (see Table 3 for uses of industrial silica sand and gravel). Table 4 presents, from selected States, the most frequently recorded occupations of U.S. residents aged 15 or above whose death certificates list silicosis as an underlying or contributory cause of death [NIOSH 1996a]. In addition, Table 5 lists published case reports of silicosis in workers from other industries and occupations.
2.4 Sampling and Analytical Methods
Historically, several methods have been used to measure worker exposure to airborne crystalline silica (quartz, cristobalite, or tridymite). These methods differ primarily in the analytical technique employed, although they all rely on a collection procedure that uses a cyclone for size-selective sampling. Airborne samples are collected using a cyclone to remove nonrespirable particles and an appropriate filter medium (e.g., polyvinyl chloride) to retain the respirable dust fraction. Preparation of the sample for crystalline silica determination differs depending on the type of analytical technique used. One of three analytical techniques is typically used for the quantitative determination of crystalline silica: X-ray diffraction (XRD) spectrometry, infrared absorption (IR) spectrometry, or colorimetric spectrophotometry. XRD and IR are the most common techniques used for crystalline silica analyses. The quantitative limit of detection for these methods ranges from 5 to 10 µg per sample; but the accuracy is poor, particularly at the low filter loadings (<30 µg per sample) that are typically collected when workplace concentrations of airborne crystalline silica are near the NIOSH REL of 50 µg/m3 (or 0.05 mg/m3).
Current sampling methods for crystalline silica involve the use of a cyclone attached to a filter cassette to collect the respirable fraction of the airborne particulate. To minimize measurement bias and variability, these samplers should conform to the criteria of the International Organization for Standardization (ISO), the European Standardization Committee (CEN), and the American Conference of Governmental Industrial Hygienists (ACGIH) for collecting particles of the appropriate size [ISO 1991; CEN 1992; ACGIH 2001]. Also, the cyclone should exhibit sufficient conductivity to minimize the electrostatic effects on particle collection. Cyclones typically used for crystalline silica measurements include the Dorr-Oliver 10-mm nylon cyclone and the Higgins-Dewell conductive cyclone. These cyclones have been evaluated for their compliance with the ISO/CEN/ACGIH respirable aerosol sampling convention. Flow rates of 1.7 L/min for the Dorr-Oliver cyclone and 2.2 L/min for the Higgins-Dewell cyclone provide minimum bias for a wide range of particle size distributions that are likely to occur in the workplace [Bartley et al. 1994]. The Dorr-Oliver 10-mm cyclone is required by MSHA, and the Higgins- Dewell cyclone is used in the United Kingdom. Recently, the GK2.69 cyclone [Kenny and Gussman 1997] has become available with a sampling rate equal to 4.2 L/min. The GK2.69 cyclone is expected to be at least as adequate as the nylon cyclone for conforming to the ISO/ CEN/ACGIH respirable aerosol sampling convention; and it may be preferable for silica sampling since it is conductive, has well-defined dimensional characteristics, and can be used at higher flow rates for better mass sensitivity. Because each type of cyclone exhibits specific particle collection characteristics, the use of a single cyclone type for each application would be advisable until evidence becomes available indicating that bias among cyclone types will not increase laboratory-to-laboratory variability.
Cyclones and filter cassettes should be leak tested to avoid gross failure in the field. The cyclones may be tested using a simple pressure- (or vacuum-) holding test. The filter cassette should also be checked for leakage while attached to the cyclone. Two approaches to testing the cassettes have been used. A micromanometer has been used to measure the pressure drop across a single type of cassette and compare it with the average pressure drop across well-sealed cassettes [Van den Heever 1994]. An alternative approach uses a particle counter to measure the penetration of submicrometer ambient aerosol through the cassette, with the percentage of penetration serving as an indicator of leakage [Baron 2001]. Measurement of cassette leakage by several laboratories indicates that significant leakage can occur in certain situations. Cassettes should be assembled using a press, and they should be routinely checked for leakage.
XRD methods used for crystalline silica determination include NIOSH Method 7500 [NIOSH 1998], OSHA Method ID-142 [OSHA 1996], MSHA Method P-2 [MSHA 1999], and the Health and Safety Executive (HSE) Method for the Determination of Hazardous Substances (MDHS) 51/2 [HSE 1988]. Details of these methods are presented in Table 6. XRD is capable of distinguishing the three prevalent polymorphs of crystalline silica (quartz, cristobalite, and tridymite) and can simultaneously analyze for each polymorph while correcting for interferences that may be present on the sample [Madsen et al. 1995]. Although most samples collected in industrial workplaces are relatively free of mineral interferences, an XRD scan of some samples should be performed to ensure the absence of interferences through confirmation of the correct peak ratios for the three largest peaks.
IR methods used for crystalline silica determination include NIOSH Methods 7602 and 7603 [NIOSH 1994a,c], MSHA Method P-7 [MSHA 1994], and MDHS 37 and 38 [HSE 1987, 1984]. Details of these methods are presented in Table 7. Although IR is less specific than XRD (IR methods cannot readily distinguish crystalline silica polymorphs), the technique is less expensive and can be optimized for measuring quartz in well-defined sample matrices [Madsen et al. 1995; Smith 1997; Hurst et al. 1997]. Samples that contain other silicates (such as kaolinite) and amorphous silica can present interferences in the analyses. Also, a potential for bias exists when correcting for matrix absorption effects, with an increasing risk of bias at lower quartz concentrations.
2.4.2.3 Colorimetric Spectrophotometry
The NIOSH colorimetric method for crystalline silica (NIOSH Method 7601) [NIOSH 1994b] is significantly less precise than IR or XRD methods. The colorimetric analytical method exhibits a nonlinear dependence on the mass of crystalline silica present [Eller et al. 1999a]. The linear range of the method is limited, and the blank values for samples can be high (20 µg silica or higher) [Talvitie 1951, 1964; Talvitie and Hyslop 1958]. High intra-laboratory variability of the method (up to twice that of IR or XRD) has been noted in studies conducted in the Proficiency Analytical Testing Program (PAT) [Shulman et al. 1992]. The colorimetric method cannot distinguish between silica and silicates, since it is based on the measurement of silicon.
2.4.2.4 Factors Affecting the Sensitivity and Accuracy of Analytical Techniques
Samples prepared for XRD analyses are measured directly (MDHS 51/2) or are redeposited onto 25-mm silver membrane filters (NIOSH Method 7500 and OSHA Method ID-142). IR samples can be measured directly (MDHS 37), redeposited on an acrylic copolymer membrane filter (NIOSH Method 7603 and MSHA Method P-7 ), or incorporated into a potassium bromide (KBr) pellet (NIOSH Method 7602 and MDHS 38). Techniques used for redepositing the sample (both IR and XRD) are difficult to perform at low sample loadings and require the laboratory analyst to demonstrate good intra-laboratory reproducibility. However, these techniques can be optimized by preparing multiple working standards from multiple suspensions of calibration standards and by ensuring that the sample is redeposited evenly as a thin layer on the filter. No statistically significant difference has been observed between ashing the filter (muffle furnace and low-temperature asher) and dissolving the filter by tetrahydrofuran before redepositing the sample [Eller et al. 1999a].
The instrument response of all three analytical techniques is influenced by the size of the particles in the sample. With XRD, the diffraction intensity (as measured by peak height) can vary considerably with particle size, with smaller particles showing lower intensities [Bhaskar et al. 1994]. The sensitivity of IR analyses decreases with increasing particle size. The colorimetric method requires the use of a precisely timed heating step with phosphoric acid to digest amorphous silica and silicates during sample preparation, causing a possible loss of some small crystalline silica particles [Eller et al. 1999a]. Since particle size affects the sensitivity of all three analytical techniques, the particle size distribution of the calibration standard should closely match the size of the particles retained on the collected sample.
For all analytical techniques, strict adherence to standardized procedures is necessary to produce accurate results. Specifically, appropriate calibration of the technique has been shown to be critical in the accurate measurement of crystalline silica [Eller et al. 1999b]. Also, it is essential that only standard reference materials from the National Institute of Standards and Technology (NIST) (for which particle size and phase purity has been established) be used to prepare calibration curves for quartz (1878a) and cristobalite (1879a) [Eller et al. 1999a]. No standard reference material for tridymite is available, since this silica polymorph rarely exists in the workplace. However, a well-characterized sample of tridymite of the appropriate particle size is available from the U.S. Geological Survey* and can be used as a reference standard.
Direct-on-filter techniques are used by the United Kingdom, the European Union, and Australia [Madsen et al. 1995]. These techniques require less time and labor than others and are amenable to both XRD and IR analyses [Lorberau et al. 1990]. However, direct-on-filter techniques are affected by the manner in which the particles are deposited on the filter sample (particle deposition may be nonuniform). Thus care must be taken when choosing the area of the filter to measure so that results can be compared with other methods. Sample overloading is possible for a sample collected over a full work shift.
*Tridymite reference material may be obtained from Dr. Stephen A. Wilson, U.S. Geological Survey, Box 25046, MS 973, Denver, CO 80225 (telephone: 303-236-2454; FAX: 303-236-3200; e-mail: swilson @usgs.gov; Web site: http://minerals.cr.usgs.gov/geochem).
2.4.3 Feasibility of Measuring Crystalline Silica at Various Concentrations
The efficacy of sampling and analytical methods for measuring concentrations of hazardous materials may be established using the NIOSH accuracy criterion [NIOSH 1995b], which requires better than 25% accuracy at concentrations of expected method application. Accuracy, as a percentage of true concentration values, is defined in terms of an interval expected to contain 95% of (future) measurements. To account for uncertainty in method evaluations, the upper 95% confidence limit on the accuracy is measured and used in the criterion. Generally, the accuracy of a method is measured over a range of concentrations bracketing the OSHA PEL. Use of a range of measurements means that accuracy is assured both at concentrations below the PEL (for possible use in action level determinations) and, more significantly, at the PEL (where method results must be legally defensible).
NIOSH has evaluated both the XRD silica method (NIOSH Method P&CAM 259, the forerunner to NIOSH Method 7500) [NIOSH 1979b] and an IR silica method (MSHA Method P-7, equivalent to NIOSH Method 7603) in a collaborative test among several laboratories [NIOSH, BOM 1983]. One result of the test was that the accuracy of the methods was estimated by evaluating the intralaboratory variability at various filter loadings. The concentrations to which these filter loadings correspond depend on the flow rate of the pre-sampler used. Experimental conditions and results relevant to the derivation of these estimates are summarized in Tables 8 and 9. The results of the collaborative tests indicate that both the XRD and IR methods tested meet the NIOSH accuracy criterion [NIOSH 1995b] over the range of filter loadings measured. Currently, OSHA uses the 10-mm nylon cyclone at a sampling rate of 1.7 L/min for sampling crystalline silica. The concentrations relevant to the collaborative test conditions are listed in Tables 10 and 11 and assume an 8-hr sampling period. As indicated in Tables 10 and 11, the traditional nylon cyclone meets the accuracy criterion over a range of concentrations bracketing 100 µg/m3.
Since the GK2.69 cyclone is expected to conform to the ISO/CEN/ACGIH respirable aerosol sampling convention, the NIOSH intralaboratory collaborative tests can be used to establish confidence limits on its accuracy. The results of the collaborative tests indicate that the GK2.69 cyclone meets the accuracy criterion over a range of concentrations bracketing 50 µg/m3, as illustrated in Tables 10 and 11.
| <previous | Table of Contents | Next> |