The different polymorphs of crystalline silica exhibit distinct absorption patterns with primary and secondary absorption lines [68]. Alpha-quartz exhibits a characteristic doublet at 798-790 and 779-780 cm-1 and secondary peaks at 694, 512, 460, 397 and 370 cm-1 . The 694 cm-1 peak may be used for quantification in cases where mineral interferences overlap with the primary doublet. Alpha-quartz may be quantified in the presence of amorphous silica by using the 694 cm-1 peak or, alternatively, a phosphoric acid digestion may be used prior to analysis. Cristobalite exhibits characteristic peaks at 798, 623, 490, 385, 297 and 274 cm-1. The 623 cm-1 peak may be used for quantification in cases where mineral interferences overlap with the customarily used 798 cm-1 peak. Tridymite exhibits characteristic peaks at 793, 617 and 476 cm-1. It may not be possible to quantify tridymite in the presence of alpha-quartz or cristobalite [69]. Matrix effects from interferences with other silicates (such as kaolinite) present analysis problems. There is a potential for bias when correcting for matrix absorption effects, with the bias increasing at low levels of quartz. In coal mining the assumption is made that alpha-quartz is the only polymorph present due to the geological processes involved in coal formation. The only mineral interference found in coal is kaolinite. MSHA P-7 gives a spectral correction procedure for eliminating IR interference from kaolinite [65]. Although most analytical chemists are familiar with the IR technique as applied to organic analyses, mineralogical samples require additional knowledge of geology and mineralogy to correctly interpret crystal structure, matrix interferences and mineral transformation for the laboratory client. In addition, data bases of IR spectra for minerals and inorganic compounds can be valuable resources [e.g., 70].
Colorimetric Methods. The colorimetric methods for crystalline silica [e.g., 71] are significantly less precise than either the X-ray or the infrared methods. The colorimetric analytical technique exhibits a nonlinear dependence on the mass of crystalline silica present [72]. There is a limited linear range and significant blank values (20 µg silica or higher) are common [39,72,73]. High intralaboratory variability (up to twice as high as XRD or IR) of the colorimetric methods has been noted in studies of Proficiency Analytical Testing (PAT) program results [40]. The colorimetric method is less precise than IR or XRD methods and has no advantage over these two techniques of analysis; therefore the colorimetric methods should no longer be used for routine measurement of exposure to crystalline silica.
Quality Assurance. Because of the complex nature of crystalline silica analysis, it is essential to have a quality assurance program which incorporates strict adherence to standardized procedures. The most important requirement should be following the analytical methods exactly as written. Any modifications made in the methods in daily laboratory practices should be accompanied by validation data demonstrating equivalency of the modified method. Other factors which are important in measurement accuracy and precision and should be tracked are sample preparation, calibration and proficiency testing.
In methods which redeposit the sample onto an analytical filter, redeposition techniques can be difficult, especially for low sample loadings, and require good intralaboratory precision. For methods involving redeposition of the sample, the entire analysis is dependent upon the uniform deposition of silica material onto the analytical filter for laboratory analysis. The analyst's ability to perform this step quantitatively and repeatably should be determined initially and periodically thereafter.
