Evaluation of DustTrak DRX to estimate respirable quartz during hydraulic fracturing

Two particulate matter (PM) size range concentrations, ≤ 4µm (PM4) and 1-4µm (PM1-4), measured from a DustTrak DRX 8533 (TSI Inc, Shoreview, MN), were compared to respirable quartz concentrations from a GK2.69 cyclone (Mesa Labs, Park Place Butler, NJ) and a nylon cyclone (TSI Inc, Shoreview, MN) during hydraulic fracturing in the shale plays of Haynesville-Bossier (Louisiana), Eagle Ford (Texas), and Permian Basin (New Mexico). The goal of the study was to evaluate the use of DustTrak respirable dust measurements as a surrogate for respirable quartz concentrations during hydraulic fracturing. DustTrak PM4 (17 samples), DustTrak PM1-4 (17 samples), GK2.69 cyclone (17 samples) and nylon cyclone (17 samples) measurements were log normally distributed. A one-way ANOVA suggested significant effect on respirable quartz geometric means at the p\u3c0.05 level for DustTrak PM 4, DustTrak PM1-4, GK2.69 cyclone respirable quartz and nylon cyclone respirable quartz geometric means. Post hoc comparisons using the Tukey HSD test indicated that the mean measurement for the DustTrak PM 1-4 condition (GM = 0.017 mg/m3, GSD = 1.244) was significantly different (p \u3c 0.001) from GK2.69 cyclone respirable quartz measurement (GM = 0.076 mg/m3, GSD = 0.949), significantly different (p \u3c 0.001) from nylon cyclone respirable quartz measurements (GM = 0.046 mg/m3, GSD = 0.966) and significantly different (p = 0.001) from DustTrak PM4 measurements (GM = 0.198 mg/m3, GSD = 0.635). However, the DustTrak PM4 measurements did not significantly differ from the GK2.69 cyclone respirable quartz measurement (p = 0.990) or nylon cyclone respirable quartz measurements (p = 0.697). GK 2.69 respirable quartz measurement were not significantly different from respirable quartz measurements from a nylon cyclone (p = 0.852). In the simple linear regression analysis, the highest coefficient of correlation for respirable quartz was found when comparing the nylon cyclone and DustTrak PM1-4 (R = 0.847), followed by the nylon cyclone and DustTrak PM4 (R = 0.798). The GK2.69 respirable quartz correlations were R = 0.735 when compared to the DustTrak PM1-4 measurements and R = 0.636 when comparing to the DustTrak PM4 measurements. Error of predicted values confirmed that the DustTrak PM1-4 comparison with respirable quartz nylon had the smallest standard error value (SE = 0.29), followed by the DustTrak PM4 comparison with respirable quartz nylon cyclone (SE = 0.36). Results in this study suggest that measurements using DustTrak may be useful to estimate respirable quartz concentration in real time due to high correlations and low predicted standard error values