Research Projects

Project at a Glance

Project Status: complete

Title: Near-source measurement of crystalline silica concentrations in California: Pilot study.

Principal Investigator / Author(s): Holmen, Britt A.

Contractor: UC Davis

Contract Number: 98-348

Research Program Area: Emissions Monitoring & Control, Atmospheric Processes

Topic Areas: Monitoring, Stationary Sources


Quantitative determination of crystalline silica (CS) concentrations in air samples downwind of industrial sources is required to determine the general population's exposure to this potentially toxic air contaminant. A Pilot Study was carried out to develop methods to characterize near-source CS concentrations in air samples collected at multiple distances downwind of a stationary source. The sampling and analysis involved the following: (1) collection of PM2.5, PM10, and size-resolved PM samples downwind of a representative CS stationary source in California, (2) collection of bulk source material and determination of the composition of PM10 derived from the source material, (3) analysis of the CS in the near-source air samples and the bulk source material. Analytical techniques included X-ray diffraction (XRD) techniques that are specific for CS, proton-induced X-ray emission (PIXE) of PM10 and PM2.5 filter samples to identify a trace element "fingerprint" of the source material, and scanning electron microscopy (SEM). Light detection and ranging (lidar) was also used to monitor the dust plume characteristics downwind of the source.

Note: quartz was the only crystalline silica mineral detected in the Pilot Study samples. Therefore, throughout this report, all references to “quartz” should be considered equivalent to “crystalline silica” (CS).


a. A sand and gravel plant was selected as a representative CS stationary source. The site was selected because the site layout enabled positioning of multiple PM samplers downwind of the Main Plant operation to a downwind distance of approximately 500m and had a suitable location for an upwind sampler.

b. Sampling was conducted over six days in June 2000 at one upwind and four locations downwind of the plant. Meteorological conditions were similar for all tests with wind directions from the NW and average test wind speeds at 2m height that ranged from 2.6 to 4.2 m s -1 . Sample test durations ranged from 2.7 to 12.5 hours and for all sampling periods, the following data was collected: lidar scans (except 6/20 and 6/21), meteorological data, PM2.5, PM10, size-resolved PM using DRUM collecting 5 size cuts, soil samples, and rangefinder positions of all sites.

c. The PM2.5 and PM10 samples were analyzed for gravimetric mass and elemental composition (PIXE, XRF). Selected PM2.5 and PM10 and DRUM samples were analyzed for crystalline silica by X-ray diffraction (XRD) and the grain size and morphology of DRUM size fractions was determined by Scanning Electron Microscopy (SEM). Grains having a pure silica composition were not detected by SEM and therefore use of the SEM for quantifying silicate mineral abundances was not a reliable technique for the sand and gravel source PM samples. The SEM did indicate a wide size range in Stage 1 ( > 8.54 µm) DRUM samples that reflects the fact that near-source fugitive dust from these facilities is generally coarse.

d. Crystalline silica (quartz) was detected in downwind DRUM samples (all sizes) at concentrations up to approximately 120 µg m -3 above the levels detected at the upwind site. For the Teflon filters, quartz concentrations in PM10 averaged 6 µg m -3 at the upwind site and 33.7 µg m -3 at the four downwind sites. PM10 and PM2.5 mass concentrations at all sites ranged from 26 to 1026 and 0 to 62 µg m -3 , respectively. Mean PM10 and PM2.5 mass concentrations were higher at the downwind sites [191 (± 181) and 16.9 (± 20.6) µg m -3 , respectively] than the upwind site [35.2 (± 7.6) and 1.6 (± 4.5) µg m -3 , respectively]. On average, approximately 15% of PM10 was attributed to quartz at both the upwind and downwind sites. Quartz was very difficult to detect in the PM2.5 Teflon filter samples, chiefly because most PM2.5 filters had small mass loadings. In the two samples measured, concentrations were < 0.64 µg m -3 at the upwind U1 site and estimated to be between 3.8 Ś 5.3 µg m -3 at the downwind D1 site.

e. The methods used here demonstrate that as few as 2.7 hours of sampling using a flow rate of ~ 1 Lpm allowed XRD determination of the mass of quartz in size-resolved DRUM samples. Quartz was, however, below the X-ray diffraction detection limit in size cuts smaller than ~ 2 µm aerodynamic diameter in all samples analyzed. Also, the DRUM sampler Stage 1 ( > 8.54 µm) XRD results must be interpreted with caution because the large grain sizes observed by SEM may affect the applicability of the quartz standard curve used for XRD quantitation. A more detailed investigation would be required to examine this issue.

f. Both silicon and PM10 mass concentrations showed linear correlations with XRD quartz concentrations measured in PM10 Teflon filter samples for the 15 samples analyzed by XRD. This suggests the possibility of developing empirical relationships between easily-determined Si concentrations and sample mineralogy for PM compositions similar to that at the Pilot Study site.

g. The Pilot Study upwind quartz concentrations were significantly lower than the concentrations measured downwind of the plant. Therefore, most of the quartz detected downwind could be attributed to the source operation. X-ray diffraction analysis of resuspended bulk source material (=D1 site soil) showed identical mineralogy to the downwind D1 site PM10 sample, therefore elemental analysis of the Teflon PM10 samples collected at D1 were assumed to be representative of the source elemental profile. Elemental signatures of the upwind and downwind PM10 and PM2.5 samples were examined to identify a unique marker that could be used to distinguish background PM from source PM. These comparisons identified S/Fe , K/H and S/H ratios as significantly different between upwind and downwind PM samples with the differences being greater for PM2.5 than PM10.The elemental data also revealed that the farthest downwind site, D4, had an elemental signature more similar to the upwind site, U1, than the other downwind sites.

h. Horizontal lidar scans indicated that plumes from the Pilot Study sand and gravel operation extended over 400 m, on average, from the edge of the main plant operation at an optical intensity that generally decreased by a factor up to ~ 2. The lidar vertical scan data showed that fugitive dust plumes generated by the plant routinely exceed heights of 100 meters and therefore the ground-level point samplers are only sampling a small portion of the entire dust plume. Therefore, the crystalline silica concentrations reported here for locations downwind of the plant should be considered representative of only the ground-level concentrations at those locations and should not be used to estimate the total emissions from the plant.

i. While both the impactor and Teflon filter sampling methods had specific limitations (see Section 4), downwind quartz concentrations consistently exceeded those measured in upwind samples. Future near-source studies should be conducted at other types of CS sources and should compare the Teflon filter method used here to NIOSH Method 7500.


For questions regarding research reports, contact: Heather Choi at (916) 322-3893

Stay involved, sign up with ARB's Research Email Listserver