The rock dusts used in these experiments were an untreated and treated limestone (treatment with proprietary hydrophobic stearate coating is representative of other similar treatments), and an untreated marble and a treated product consisting of a blend of that marble (at 87.5%) with another, surface-modified marble product. These are referred to as untreated limestone (UL), treated limestone (TL), untreated marble (UM), and treated marble (TM), respectively. Dust source and treatment are proprietary.

The selected rock dusts were aerosolized by a fluidized bed aerosol generator (Model 3400, TSI Inc., Shoreview, MN, USA). The generated aerosols were introduced into an aerosol chamber with a special aerosol nozzle that allows the airflow to enter the chamber in a radial direction. The testing setup is similar to one that has been described previously.^{(13, 14)} Six high-flow rate respirable size-selective samplers, FSP10 cyclones (GSA Messgerätebau GmbH, Ratingen, Germany), which operated at 11.2 L min⁻¹, loaded with polyvinyl chloride filters (PVC, 5-μm pore size, 37 mm; GLA5000, SKC Inc., Eighty Four, PA, USA) were placed in the aerosol chamber, and rock dusts were collected in three repetitions. Prior to sample collection, the PVC filters were equilibrated for a minimum of 72 hours in a weighing room at constant relative humidity (50% ± 2) and temperature (26 °C ± 2). Filters were pre-weighed with a micro balance (XP6U, Mettler-Toledo, Columbus, OH, USA; readability, 0.1 μg).⁽¹⁵⁾ The filters were passed through an electrostatic bar (Mettler-Toledo) before weighing to dissipate static charge. A single measurement for each filter was made after allowing exactly 180 seconds for balance stabilization. The flow rates through the samplers were calibrated with a mass flow meter (Model 4040, TSI Inc.). The flow rates were calibrated before and after each sampling session to confirm that they did not change significantly (all remained within ±5%). The filters were equilibrated in the weighing room for a minimum of 72 hours before post-weighing. The respirable dust mass concentration was determined on the basis of obtained mass, pre and post flow rate, and sampling time.

Particle size-segregated samples were obtained with a Micro-Orifice Uniform Deposit Impactor (MOUDI™ Model 110, MSP Corp., Minneapolis, MN, USA). The MOUDI consists of an inlet stage (with a 50% cut-off aerodynamic diameter [d₅₀] of 18 μm) and additional impaction stages (with d₅₀ of 10.0, 5.6, 3.2, 1.8, 1.00, 0.56, 0.32, 0.18, 0.10, and 0.056 μm). Gravimetric analyses of the PVC filters loaded at each stage of the MOUDI, was as described for cyclone filters above. The size distributions represent the dust mass concentration per unit width of size channel at different aerodynamic diameters. The MOUDI data were obtained with the equation (1)dMdlog(dp)=miQMTMΔlog(dp) where

m_i = the mass of the dust collected on the filter of stage i of the MOUDI;

Q_M = the sampling flow rate of the MOUDI (nominal flow rate is 30 L min⁻¹);

T_M = the total sampling time of the MOUDI for one test; and

Δlog(d_p)= the width of the size channel for the MOUDI.

It is assumed that the inlet stage has an upper size limit of 100 μm and the final filter stage has a lower size limit of 0.01 μm. The particle size distribution of each rock dust was averaged using three replicate measurements.

Quartz was measured according to the NIOSH Manual of Analytical Methods (NMAM) 7603 (Quartz in Coal Mine Dust, by IR [redeposition]).⁽¹⁶⁾ Briefly, the filter with dust catch was treated with hydrochloric acid such that any carbonates in the dust would be dissolved. The acid solution was rinsed from the filter with 2-propanol leaving the remainder of the catch in place. The filters with remaining dust catch were air dried and transferred to loosely covered crucibles (Model 5013, Avogadro's Lab Supply Inc., Miller Place, NY, USA) and ashed in a muffle furnace (Model F6010, Barnstead/ThermoLyne, Dubuque, IA, USA) for 2 hours at 600°C. The residue from the ashed filters was then gently re-suspended with 10 mL of 2-propanol and sonication for 5 minutes. The re-suspended ashed residues were re-deposited on a PVC filter. Although NMAM 7603 requires the residue to be redeposited onto a 10 mm diameter area of a vinyl/acrylic copolymer (DM-450) filter, these filters are currently unobtainable, and in the present study the re-suspended ashed sample was redeposited on a 5-μm pore size PVC filter. Information regarding the performance of the replacement PVC filters for respirable quartz measurement by infrared spectroscopy has been described previously.⁽¹⁷⁾ The quartz mass was quantified by FTIR (Nicolet 6700, Thermo Fisher Scientific, Waltham, MA, USA with the Omnic software package Version 8.1.11), reading peak height at 800 cm⁻¹ (with baseline between 820 and 670 cm⁻¹). In the previous study noted above we demonstrated that the recovery in spiked samples of respirable α-quartz suspension in isopropyl alcohol ranged from 109-118% (coefficient of variation 1.8-4.5%). ⁽¹⁷⁾

Three different net masses (5, 10, and 15 mg) of each rock dust were weighed on a micro balance (AG245, readability 0.01 mg, Mettler-Toledo) without size-selective sampling (as-received; unknown particle size distribution), and the quartz mass of each sample was then determined as described above. Note that bulk sample analysis using calibration standards of respirable-size quartz can lead to differences in results between infra-red and x-ray methods, due to their different particle size-dependent responses.⁽¹⁸⁾

Particle morphology analysis was conducted with a combination of scanning electron microscopy (SEM, S-4800-2, Hitachi High Technologies America Inc.) under computer control and energy-dispersive x-ray spectroscopy (EDX) analysis (EDX, PGT Inc., Princeton, NJ, USA). Three cowls (BestChek^® cassettes, SKC Inc.) loaded with polycarbonate filters for SEM/EDX were placed in the aerosol chamber, and dusts were collected with simultaneous respirable size-selective sampling. Transmission electron micrographs of rock dust particles in respirable size fraction were obtained on a JEOL TEM 1220 (JEOL USA, Peabody, MA, USA) at a working voltage of 80 kV. TEM images were photographed by placing a drop of diluted sample on a formvar-coated copper grid to dry. Several TEM images were analyzed to identify at least 5-10 individual particles per image to estimate approximate dimensions.

A total of 72 individual samples of respirable dust were collected (4 different rock dusts × 6 size-selective samplings × 3 repetitions). One sample was lost during handling. The respirable dust mass concentrations and quartz concentrations (values shown in parentheses here) determined with the FSP10 cyclone ranged from 1.72 to 4.70 mg m⁻³ (8.80 to 49.0 μg m⁻³) for UL; 11.3 to 17.9 mg m⁻³ (78.6 to 118 μg m⁻³) for TL; 4.17 to 7.97 mg m⁻³ (2.86 to 3.64 μg m⁻³) for UM; and13.2 to 23.9 mg m⁻³ (0.29 to 31.8 μg m⁻³) for TM. Quartz was detected in only 2 of 18 respirable dust samples of UM because of its low bulk quartz content.

Normalized mass-weighted size distributions of the aerosolized rock dusts, determined by MOUDI cascade impactor, are shown in FIGURE 1. In general, the MOUDI data showed reasonably good repeatability. The particle size distribution was similar for three of the types of rock dust whereas TM particles had a smaller mass median aerodynamic diameter (MMAD, 1.25 μm). This difference in MMAD between UM and TM is not likely the result of the addition of the surface-modified marble even if it were very much finer than the untreated marble because a bimodal distribution is not observed. The change is more likely the result of a further grinding step or further size classification involved in the preparation of TM.

The MMAD, GSD, airborne total dust concentration, quartz concentration, and quartz content (%) from each test using the MOUDI are shown in TABLE I. All particle size distributions showed a lognormal distribution, with the MMAD between 1.17 μm (TM #3) and 5.26 μm (UL #3) and the GSD less than 2.78 (TM #1; as shown in TABLE I). All the size distributions show that a large fraction of the airborne rock dust was respirable (smaller than 10 μm). Note that the normalized particle size distribution determined with laboratory samples might differ from actual mining industry situations because the low velocity at the exit of the fluidized bed generator prevents the output of particles with aerodynamic diameters greater than 50 μm.⁽¹⁹⁾

FIGURE 2 shows the average and standard deviation of quartz content (%) in bulk samples, the airborne total dust from the MOUDI, respirable dust from the FSP10, and the calculated respirable fraction from MOUDI of UL, TL, UM, and TM rock dusts. Quartz percentage contents for each category were calculated as follows:

Quartz content in bulk samples: QuartzmassDustmassfrombulksamples×100

Quartz content in airborne total dust from MOUDI: ΣiQuartzmassiΣiAirbornetotaldustmassi.

Quartz content in respirable dust: QuartzmassRespirabledustmass×100

Calculated respirable quartz content according to the published sampling efficiency of the FSP10⁽¹³⁾ from size distribution found by the MOUDI: ΣiRespirablefractionofFSP10cyclonei×QuartzmassiΣi=stageAirbornetotaldustmassi (i is stage of MOUDI; respirable fraction of FSP10 cyclone as previously reported).

The percentage of quartz in airborne total dust from the MOUDI is significantly (p < 0.05) enriched in UL and TL compared to the percentage of quartz in the nonaerosolized bulk samples, but the difference is smaller for only the respirable fractions. The percentage of quartz in UL and TL airborne samples collected by the FSP10 respirable samplers is significantly (p < 0.05) greater than that in the respective bulk materials, with an increase around 40-50%. The percentage of quartz in the respirable fraction from the FSP10 is greater than the percentage of quartz in the respirable fraction calculated from the MOUDI. This may be an artefact of dividing the quartz into smaller amounts of sample on the MOUDI stages. The percentage of quartz in both bulk and respirable UM and TM is extremely low and there is no significant enrichment in the respirable fraction over the bulk.

SEM images of each rock dust and energy-dispersive x-ray analysis results are presented in FIGURE 3, images (a) through (d). Images and x-ray spectra are typical of the four kinds of rock dust studied. By visual inspection, it is apparent that the particle sizes range from < 1 μm to 15 μm in most of these images. The dust spectra are predominantly from limestone or marble (calcium carbonate). There are some spectra from clay (aluminosilicate) particles (as shown in FIGURE 3 (d) for illustration) and gypsum (calcium sulfate) particles. No free quartz particles were observed because of the low quantity of quartz and the limited number of particles studied. Considerably more particles (thousands) would need to be examined to determine accurately the characteristics of a small percentage of particles.⁽²⁰⁾ In FIGURE 4, TEM analysis show that each rock dust in respirable size fraction contains irregularly-shaped particles, with size range from 0.2 to 3 μm. Size distributions for quantitative comparison would have required the measurement of many more particles as noted with SEM.