High and low flow rate respirable size selective samplers including the CIP10-R (10 l min^{−1}), FSP10 (11.2 l min^{−1}), GK2.69 (4.4 l min^{−1}), 10-mm nylon (1.7 l min^{−1}), and Higgins-Dewell type (2.2 l min^{−1}) were compared via side-by-side sampling in workplaces for respirable crystalline silica measurement. Sampling was conducted at eight different occupational sites in the USA and five different stonemasonry sites in Ireland. A total of 536 (268 pairs) personal samples and 55 area samples were collected. Gravimetric analysis was used to determine respirable dust mass and X-ray diffraction analysis was used to determine quartz mass. Ratios of respirable dust mass concentration, quartz mass concentration, respirable dust mass, and quartz mass from high and low flow rate samplers were compared. In general, samplers did not show significant differences greater than 30% in respirable dust mass concentration and quartz mass concentration when outliers (ratio <0.3 or >3.0) were removed from the analysis. The frequency of samples above the limit of detection and limit of quantification of quartz was significantly higher for the CIP10-R and FSP10 samplers compared to low flow rate samplers, while the GK2.69 cyclone did not show significant difference from low flow rate samplers. High flow rate samplers collected significantly more respirable dust and quartz than low flow rate samplers as expected indicating that utilizing high flow rate samplers might improve precision in quartz measurement. Although the samplers did not show significant differences in respirable dust and quartz concentrations, other practical attributes might make them more or less suitable for personal sampling.

The US Occupational Safety and Health Administration (OSHA) proposes lowering the permissible exposure limit to 0.05 mg m^{−3} with an action level of 0.025 mg m^{−3} for respirable crystalline silica (RCS) as part of a new comprehensive standard (^{−1}. The 10-mm nylon cyclone can be used to collect samples containing at least 10 μg of respirable quartz at 0.025 mg m^{−3} for 8-h and even 4-h samples, which will allow quantification. However, it cannot be used to quantify concentrations <0.025 mg m^{−3} over periods <8-h sampling periods. In addition, if the proposed OSHA action level is ratified, users would prefer a method able to assure the user that a sample is below this concentration. Performance of respirable size selective samplers operating at high flow rates (flow rate > 4 l min^{−1}) for RCS measurement has been compared to that of low flow rate samplers (1.7–2.2 l min^{−1}) under laboratory and field conditions. These studies have shown that samples collected with high flow rate samplers could provide precise analytical results, i.e. significantly above the limit of detection (LOD) and/or LOQ by increasing the mass collection on filters (

Eight different sites in the USA were included in this study. Work processes underway on these sites included construction (masonry, demolition, and concrete drilling), silica sand production, and metal mining (area sampling only).

The two types of low flow rate samplers employed were (i) 10-mm nylon cyclone (Sensidyne, Clearwater, FL, USA), designed for particle collection with 5 μm pore size 37-mm polyvinyl chloride (PVC) filter (GLA5000, SKC Inc., Eighty Four, PA, USA) and a sampling flow rate of 1.7 l min^{−1}, and (ii) Higgins-Dewell type cyclone [model: BGI4L (nickel plated aluminum body and aluminum grit pot), BGI USA Inc., Waltham, MA, USA] used with 5 μm pore size 37-mm PVC filter (GLA5000, SKC Inc.) and a sampling flow rate of 2.2 l min^{−1}. Three high flow rate samplers employed were (i) CIP10-R sampler (Arelco ARC, Paris, France) with particle collection on a polyurethane foam in rotating cup and a sampling flow rate of 10 l min^{−1}, (ii) FSP10 cyclone (GSA Messgerätebau GmbH, Ratingen, Germany) with particle collection on 5 μm pore size 37-mm PVC filter (GLA5000, SKC Inc.) at a sampling flow rate of 11.2 l min^{−1}, and (iii) GK2.69 cyclone (BGI Inc., Waltham, MA, USA), particle collection by 5 μm pore size 37-mm PVC filter (GLA5000, SKC Inc.), sampling flow rate of 4.4 l min^{−1}. In order to minimize particle deposition on sampling cassette walls, conductive polypropylene cassettes (SKC Inc.) were used for 10-mm nylon, BGI4L, and GK2.69 cyclones (

Prior to sample collection, filters and foams for the samplers were equilibrated for a minimum of 72 h in the weighing room at constant relative humidity (50% ± 2) and temperature (26°C ±2). Pre-weighing of filters and rotating cups with foams was performed with a microbalance (XP6U, Mettler-Toledo, Columbus, OH, USA; readability 0.1 μg). Filters and rotating cups with foams were passed through an electrostatic bar (Mettler-Toledo) before weighing to dissipate static charge. A single measurement for each filter and rotating cup was made after allowing exactly 180 seconds for balance stabilization. Average coefficients of variation of blank PVC filters and rotating cups were 0.046 and 0.002%, respectively. The pre-weighed filters were assembled in conductive (static-dissipative) polypropylene cassettes and leak-checked using a field cassette leak tester (SKC Inc.). Aircheck PCXR-4 pumps (SKC Inc.) were used with 10-mm nylon and BGI4L cyclones. SG 10–2 pumps (GSA Messgerätebau GmbH) and Legacy pumps (SKC Inc.) were connected to the FSP10 and GK2.69 cyclones, respectively. The flow rates through the samplers were calibrated using a BIOS DryCal Meter (BIOS International Corporation, Butler, NJ, USA). The flow rates were calibrated before and after each sampling session to confirm that they did not change significantly (all remained within ±5%). The flow rate of the CIP10-R sampler was initially calibrated with a CIP10 Calibration bench (Arelco, ARC) against a tachometer, and the rotational speed of the cup was checked in the field before and after sampling.

Side-by-side personal sampling with six sampler combinations (CIP10-R/10-mm nylon, FSP10/10-mm nylon, GK2.69/10-mm nylon, CIP10-R/BGI4L, FSP10/BGI4L, GK2.69/BGI4L) was conducted with volunteer workers. The participants were asked to wear commercial back-braces (Safe-T-Lift, Style No. 70-110543, FLA Orthopedics Inc. Charlotte, NC, USA) or safety vests (Model SV7O5X, Radians, Memphis, TN, USA) and the high and low flow rate samplers were located in the breathing zone of the worker, one on each side, with sides randomized for different pairs. The pumps were attached to the back-braces around the waist of the participants or in the pockets of vests. Sampling duration was between 10 and 390 min and most samples were collected between 180 and 240 min.

Area sampling was also conducted at four of the sites including metal mining, concrete drilling and construction, and bricklayer training due to the limited number of workers available for personal sampling. A stationary Lippman-type sampling apparatus (

The filters and rotating cups with foams were equilibrated in the weighing room for a minimum of 72 h before post-weighing. The respirable dust mass concentration was determined using obtained mass, pre-and post-flow rate, and sampling time.

Sampling procedures used in Ireland have been described in a previous publication (^{−1}) was used instead of BGI4L cyclone and (ii) personal side-by-side samples were only collected for FSP10/10-mm nylon and FSP10/SIMPEDS pairs due to limited numbers of workers [other pairs were collected as area samples by placing the samplers as near to the worker as physically possible (0.5–15 m from the worker) at a height of 1.5 m using a tripod]. Both SIMPEDS and BGI4L cyclones are based on Higgins-Dewell design and thus similar performance can be assumed (

Major activities and the number of samples collected for each site are shown in

XRD analysis was carried out by an American Industrial Hygiene Association (AIHA) accredited laboratory according to the NIOSH Manual of Analytical Method (NMAM) 7500 [SILICA, CRYSTALLINE, by XRD (filter redeposition)] (

Results of area and personal sampling were combined for each pair of the samplers and data were analyzed using SAS/STAT software, Version 9.3 of the SAS System for Windows (SAS Institute, Cary, NC, USA). Data were transformed using the natural log prior to analysis. Geometric means (GMs) and confidence intervals are back transformed into their natural units for presentation. Sampler types were compared to one another using mixed model analyses of variance carried out with Proc Mixed. Sampling site and sampling pair were considered random variables. Slopes were determined using Proc Reg and making measures from the high flow rate samplers the response variable.

Differences in frequency of below and above LOD and LOQ of quartz mass collected with the high and low flow rate samplers were determined using McNemar’s test (

Eleven sets (total of 55 individual samples) of area samples and 268 pairs of personal samples (536 samples) were collected. The ratios of respirable dust concentration, quartz mass concentration, respirable dust mass, and quartz mass between high and low flow rate samplers showed a log normal distribution (Shapiro–Wilk test). These data were described using the GM with 95% levels of confidence. Negative respirable mass, due to low respirable dust mass concentrations, was found in 17 and 6% of samples from low and high flow rate samplers, respectively. Additionally, four samples were lost due to pump failure.

A statistical comparison (McNemar’s test) for frequency of below and above LOD (5 μg; NMAM 7500) and LOQ (15 μg; NMAM 7500; a rough estimation from LOD × 3) of quartz mass collected with the high and low flow rate samplers was made (

GM with 95% lower and upper confidence intervals of (i) respirable dust mass concentration ratio, (ii) quartz mass concentration ratio, (iii) respirable dust mass ratio, and (iv) quartz mass ratio for each pair of samplers are shown in

Since the job tasks in Ireland were similar to each other, they have been grouped as one site. Differences between the sites were not tested for significance because it is a random effect variable rather than fixed effect variable and sample sizes for most sites are too small to have power for statistical analysis.

Sampler pairs of the CIP10-R/10-mm nylon, FSP10/10-mm nylon, and GK2.69/HD type showed significant differences in respirable dust mass concentration and sampler pairs of the CIP10-R/10-mm nylon and GK2.69/HD type showed significant differences in quartz mass concentration (

The FSP10 and CIP10-R samplers collected significantly more respirable dust and quartz mass than low flow rate samplers. The GK2.69 cyclone collected significantly more respirable and quartz mass than low flow rate samplers when the outliers were removed from the analysis. The CIP10-R, FSP10, and GK2.69 samplers are expected to collect 5.9, 6.6, and 2.6 times more mass compared to 10-mm nylon cyclone, respectively, and they are expected to collect 4.7, 5.1, and 2.0 times more than HD type cyclone, respectively. GM of respirable dust mass ratios and quartz mass ratios for each pair of the samplers were closer to the expected respirable dust mass ratio and quartz mass ratios without outliers (

Linear regression analysis results of respirable mass concentrations and quartz concentrations for each pair of samplers with and without outliers are shown in

Performance of high flow rate samplers was previously investigated in laboratory experiments (

While the most important consideration in air sampling is the accuracy and precision in measuring the intended size fraction, i.e. inhalable, thoracic, or respirable, there are other important issues including cost of sampling, worker acceptance (comfort and placement of sampling device) and industrial hygienist concerns (ease of calibration and sample analysis; for example, the jar necessary for calibration of the 10-mm nylon cyclone is cumbersome in the field). The FSP10 cyclone has been calibrated to provide a respirable sample at 11.2 l min^{−1}, which would provide adequate sensitivity for most purposes (^{−1} (^{−1}, and therefore provide a sample loading of 12 or 13 μg RCS at a concentration one-half of 0.025 mg m^{−3} over a 4-h sample. Since it uses a 37-mm filter, the pressure-drop is within the range of some lower-cost personal sampling pumps already in common use for taking an 8-h sample. The GK2.69 cyclone showed no significant difference in number of samples above the LOD and borderline significance above the LOQ compared to low flow rate samplers (^{−3} has a consequence of imposing a limit for the respirable fraction of particles not otherwise specified (or regulated) of 2.5 mg m^{−3}. However, a full-shift sample of respirable dust with a GK2.69 cyclone at 2.5 mg m^{−3} will approximately have a mass of 5 mg, so it may be prudent to restrict high flow rate sampling in very dusty environments.

Variation in RCS analysis determined from the AIHA Proficiency Analytical Testing (PAT) results between 2005 and 2013 [average relative standard deviation (RSD) 20%] was reduced compared to the period between 1990 and 1998 (average RSD 29%), but it is still higher than for other occupational samples (^{−1} to collect samples at the proposed action limit. The new analysis suggested this trend was a result of sample preparation procedure, rather than analytical capability, but it remains the case that low flow rates at the action level concentration produce sample loadings below the current range of PAT samples. High flow cyclones provide higher loadings, which would fall within the PAT sample range.

Performance of high flow rate samplers for respirable size selective sampling including CIP10-R, FSP10, and GK2.69 was compared to that of low flow rate samplers in occupational environments. The high flow rate samplers did not generally show significant differences in respirable dust and quartz mass concentration. The high flow rate samplers allow for greater respirable quartz mass collection over shorter sampling periods affording improved levels of precision. However, higher flow rate samplers may have other attributes including cost and size and weight of both sampler and pump that may influence the decision as to whether to use them for routine personal sampling. A cyclone operated at 1.7 l min^{−1} for 4 h at the proposed action level would collect 10 μg of silica, which is the limit of quantitation for many laboratories. The present study confirms the conclusions of previous studies that respirable size selective samplers operating with high flow rate can be used to reliably quantify silica concentrations below 0.025 mg m^{−3} over sampling periods <8 h or 0.025 mg m^{−3} for sampling periods <4 h.

The authors would like to thank the workers who volunteered to be sampled for the study and Don Anderson (Bricklayers Local 1 PA-DE), Emanuele Cauda (NIOSH/OMSHR), Lorenzo Cena (NIOSH/HELD), Alan Echt (NIOSH/DART), Jim Kinateder (Fred Kinateder Masonry Inc.), Eun Gyung Lee (NIOSH/HELD), and Jhy-Charm Soo (NIOSH/HELD) for technical support of field sampling.

The findings and conclusions in this report are those of the author(s) and do not necessarily represent the official position of the Centers for Disease Control and Prevention/the Agency for Toxic Substances and Disease Registry.

Area sampling apparatus for collection of respirable dust.

Geometric means (95% lower and upper confidence intervals) of high/low flow rate (a) mass concentration and (b) net mass ratios including outliers. *Significantly difference between two samplers in accordance with mixed model analyses of variance (

Geometric means (95% lower and upper confidence intervals) of high/low flow rate (a) mass concentration and (b) net mass ratios without outliers. *Significantly difference between two samplers in accordance with mixed model analyses of variance (

Box plot of quartz content ratio between high and low flow rate samplers. The horizontal lines in the box plot from bottom to top indicate 10th, 25th, 50th (median), 75th, and 90th percentiles. The circles indicate the 5th (lower circle) and 95th (upper circle) percentiles. HD type is Higgins-Dewell type. *Significantly difference between two samplers in accordance with Mann–Whitney rank sum test (

Sampling site, major process, number of sample pairs collected, number of outliers for each sites samples, respirable dust mass [geometric mean (95% lower and upper confidence interval)], and quartz mass concentration [geometric mean (95% lower and upper confidence interval)] for each site

Sampling site | Major process | Number of sample collected (pairs) | Number of outliers in ratios concentration (pairs)
| Respirable dust mass concentration (mg m^{−3}) | Quartz mass concentration (μg m^{−3}) | |
---|---|---|---|---|---|---|

Quartz | Respirable dust | |||||

1 | Cutting, tuck-pointing | 22 | 5 | 5 | 0.350 (0.224–0.548) | 63.5 (43.1–93.7) |

2 | Jack hammering | 28 | 5 | 6 | 0.096 (0.064–0.144) | 14.8 (8.53–25.5) |

3 | Cutting, tuck-pointing | 8 | 2 | 2 | 1.07 (0.322–3.57) | 216 (57.7–805) |

4 | Drilling | 12 | 0 | 0 | 2.24 (0.906–5.52) | 304 (128–721) |

5 | Crushing, mining, milling, loading | 54 | 2 | 7 | 0.097 (0.074–0.129) | 18.9 (15.4–23.1) |

6 | Bagging, milling, loading | 30 | 1 | 5 | 0.085 (0.068–0.106) | 45.6 (34.5–60.4) |

7 | Cutting | 114 | 45 | 52 | 7.45 (5.84–9.51) | 4062 (2722–5881) |

Frequency of quartz mass below and above limit of detection and limit of quantification collected with high and low flow rate samplers

Pair of samplers | Sample number below LOD | Sample number above LOD | Sample number below LOQ | Sample number above LOQ | ||
---|---|---|---|---|---|---|

10-mm nylon | 19 | 36 | <0.0001 | 30 | 25 | <0.0001 |

CIP10-R | 3 | 52 | 9 | 46 | ||

10-mm nylon | 21 | 35 | 0.0008 | 32 | 24 | <0.0001 |

FSP10 | 8 | 48 | 13 | 43 | ||

10-mm nylon | 18 | 37 | 0.103 | 26 | 29 | 0.059 |

GK2.69 | 14 | 41 | 21 | 34 | ||

HD type | 11 | 43 | 0.033 | 21 | 33 | 0.0005 |

CIP10-R | 5 | 49 | 9 | 45 | ||

HD type | 16 | 40 | 0.0039 | 25 | 31 | 0.0002 |

FSP10 | 6 | 50 | 11 | 45 | ||

HD type | 16 | 41 | 0.317 | 28 | 29 | 0.052 |

GK2.69 | 14 | 43 | 21 | 36 |

Higgins-Dewell type, including BGI4L and SIMPEDS.

Limit of detection.

Limit of quantification.

Comparison of linear regression analysis of respirable mass concentration and quartz concentration between samples from all US sites (Irish samples were removed) and samples from all sites when outliers were removed

Pair of samplers | Respirable dust mass concentration
| Quartz concentration
| ||
---|---|---|---|---|

Samples from all US site (Irish sample were removed) | Sample for all site (outliers removed) | Sample from all US site (Irish sample were removed) | Sample from all site–8 (outliers removed) | |

CIP10-R/10-mm nylon | 1.04 | 1.08 | 1.02 | 0.96 |

FSP10/10-mm nylon | 0.94 | 1.02 | 0.90 | 1.05 |

GK2.69/10-mm nylon | 0.97 | 1.03 | 0.91 | 1.08 |

CIP10-R/Higgins-Dewell type | 1.02 | 0.94 | 1.01 | 0.89 |

FSP10/Higgins-Dewell type | 1.01 | 0.92 | 1.01 | 0.96 |

GK2.69/Higgins-Dewell type | 0.94 | 0.99 | 0.94 | 1.01 |

Significantly different from 1:1 relationship (

Linear regression analysis of respirable mass concentration and quartz concentration for each pair of samplers with and without outliers

Samplers pair | Slope (^{2}) w/respirable dust mass concentration | Slope (^{2}) w/respirable dust mass concentration (outlier removed) | Slope (^{2}) w/quartz mass concentration | Slope (^{2}) w/quartz mass concentration (outlier removed) |
---|---|---|---|---|

CIP10-R/10-mm nylon | 0.89 (0.71) | 1.08 (0.94) | 0.89 (0.81) | 0.96 (0.93) |

FSP10/10-mm nylon | 1.1 (0.84) | 1.02 (0.95) | 1.2 (0.92) | 1.05 (0.94) |

GK2.69/10-mm nylon | 0.89 (0.72) | 1.03 (0.96) | 1.1 (0.96) | 1.08 (0.97) |

CIP10-R/Higgins-Dewell type | 0.85 (0.84) | 0.94 (0.96) | 0.77 (0.89) | 0.89 (0.98) |

FSP10/Higgins-Dewell type | 0.87 (0.89) | 0.92 (0.95) | 0.87 (0.87) | 0.96 (0.97) |

GK2.69/Higgins-Dewell type | 0.68 (0.81) | 0.99 (0.95) | 0.66 (0.85) | 1.01 (0.96) |

Significantly different from 1:1 relationship (

Comparison of average respirable dust mass concentration ratio and quartz mass concentration ratios between high and low flow rate samplers. The average ratios from the laboratory studies (

Pair of samplers | Present study
| ||||
---|---|---|---|---|---|

Respirable dust mass concentration ratio | Quartz mass concentration ratio | Quartz mass concentration ratio | Respirable dust mass concentration ratio | Quartz mass concentration ratio | |

CIP10-R/10-mm nylon | 1.10 (0.193) | 1.17 (0.555) | 1.15 (0.936–1.42) | 1.26 (1.04–1.53) | |

FSP10/10-mm nylon | 1.22 (0.168) | 1.46 (0.762) | 1.11 (0.945–1.31) | 1.05 (0.841–1.31) | |

GK2.69/10-mm nylon | 1.09 (0.183) | 1.32 (0.602) | 1.11 (0.960–1.28) | 1.15 (0.995–1.34) | |

CIP10-R/Higgins-Dewell type | 1.03 (0.207) | 1.05 (0.504) | 0.88 | 0.891 (0.766–1.04) | 1.27 (1.09–1.48) |

FSP10/Higgins-Dewell type | 1.14 (0.172) | 1.27 (0.535) | 1.07 | 1.07 (0.891–1.28) | 1.15 (0.977–1.35) |

GK2.69/Higgins-Dewell type | 0.999 (0.171) | 1.13 (0.412) | 1.00 | 1.02 (0.867–1.19) | 1.13 (0.997–1.29) |

Linear regression analysis comparison in respirable dust mass concentration between high and low flow rate samplers

Samplers pair | Present study | ||
---|---|---|---|

CIP10-R/10-mm nylon | 1.02 | 1.01 | 1.12 |

FSP10/10-mm nylon | 1.19 | 1.21 | 1.05 |

GK2.69/10-mm nylon | 1.06 | 1.03 | 1.04 |

CIP10-R/Higgins-Dewell type | 0.96 | 0.84 | 0.97 |

FSP10/Higgins-Dewell type | 1.14 | 0.99 | 0.98 |

GK2.69/Higgins-Dewell type | 1.02 | 0.84 | 0.95 |