Health and Safety Executive

Testing of high flow rate respirable samplers to assess the technical feasibility of measuring 0.05 mg.m-3 respirable crystalline silica

Prepared by the Health and Safety Laboratory for the Health and Safety Executive 2010

RR825 Research Report Health and Safety Executive

Testing of high flow rate respirable samplers to assess the technical feasibility of measuring 0.05 mg.m-3 respirable crystalline silica

Peter Stacey and Andrew Thorpe Health and Safety Laboratory Harpur Hill Buxton Derbyshire SK17 9JN

Testing of high flow rate samplers to assess the technical feasibility of measuring 0.05 mg.m-3 respirable crystalline silica.

This report describes testing of five personal respirable dust samplers operating with flow rates of 4 l/min or greater, available in 2008. Three were commercially available, one a prototype and one adapted at HSL to operate at a higher flow rate. Testing compared these samplers with a reference sampler, operating at 2.2 l/min, to ascertain if an increase in the mass of dust sampled could improve the reliability of measurements of respirable crystalline silica (RCS). None of the samplers satisfied all of the success criteria for the project, which included, the ability to maintain the specified flow rate over 4-hours, ease of use in the workplace, and an improvement in the measurement precision without additional complications caused by the increased mass of sampled dust. Infrared analysis is not recommended for samples with dust mixtures, because it was difficult to obtain a reliable result when the loading exceeds 1 mg. The samplers with the best performance were the PGP10 and the modified GK2.69 samplers. The other samplers tested either under-sampled or there was lost sample during transfer onto the analysis filter. When field tests were conducted, air sampling pumps operating with the modified GK2.69 samplers failed to maintain a consistent flow rate, and the PGP 10 samplers were heavy and caused discomfort for the workers The report recommends the use of the PGP 10 and GK2.69 samplers after further work to resolve the minor issues and changes in the sampling and measurement strategies to accommodate new procedures for use of higher flow rate samplers.

This report and the work it describes were funded by the Health and Safety Executive (HSE). Its contents, including any opinions and/or conclusions expressed, are those of the authors alone and do not necessarily reflect HSE policy.

HSE Books © Crown copyright 2010

First published 2010

All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means (electronic, mechanical, photocopying, recording or otherwise) without the prior written permission of the copyright owner.

Applications for reproduction should be made in writing to: Licensing Division, Her Majesty’s Stationery Office, St Clements House, 2-16 Colegate, Norwich NR3 1BQ or by e-mail to [email protected]

ii

CONTENTS

1 INTRODUCTION...... 1 1.1 Background ...... 1 1.2 Principles of the analysis techniques...... 1

2 STAGE 1: SELECTION OF EQUIPMENT INCLUDED IN THE STUDY..... 4 2.1 Selection of Samplers...... 4 2.2 Selection of filters ...... 4 2.3 Selection of sampling trains...... 4

3 CALIBRATIONS ...... 10 3.1 Calibrations for x-ray diffraction ...... 10 3.2 Calibrations for infrared analysis ...... 10 3.3 Discussion ...... 10

4 STAGE 2B: ABSORPTION AND DEPTH EXPERIMENTS...... 12 4.1 Evaluation of absorption and depth effects in x-ray diffraction analysis. 12 4.2 Evaluation of the effect of absorption on FTIR analysis...... 20

5 STAGE 3: ASSESSMENT OF THE BIAS OF SAMPLERS...... 23 5.1 Sampling tests with Arizona road dust...... 23 5.2 Gravimetric Analysis...... 24 5.3 RCS Analysis by X-ray Diffraction ...... 25 5.4 RCS Analysis by Direct on Filter Infrared Analysis ...... 26 5.5 Recovery for the PGP 10 cyclone using cellulose nitrate filters...... 27

6 STAGE 4: ASSESSMENT OF BIAS OF ANALYTICAL TECHNIQUES .. 29 6.1 Workplace tasks examined...... 29 6.2 Particle SIZE distribution of the generated aerosol...... 30 6.3 Gravimetric analysis ...... 30 6.4 X-ray Diffraction analysis ...... 33

7 FIELD TRIALS...... 36 7.1 Approach ...... 36 7.2 Analytical Results ...... 36 7.3 Practical experience ...... 39 7.4 Comments recorded from the Workers...... 39

8 OVERALL PERFORMANCE OF SAMPLERS ...... 44 8.1 Modified GK 2.69 cyclone...... 44 8.2 IOM sampler with foam separator...... 44 8.3 IPP Impactor...... 44 8.4 PGP 10 Cyclone ...... 45 8.5 CIP 10 SAMPLER ...... 45

-3 9 PRECISION OF XRD MEASUREMENTS AT 0.05 MG.M ...... 46

iii

10 PROJECT SUCCESS CRITERIA...... 49

11 REFERENCES ...... 52 12 APPENDIX 1: PRESSURE DROP WITH FLOW RATE ACROSS AN MIXED CELLULOSE ESTER FILTER...... 54

13 APPENDIX 2: INSTRUMENTAL PARAMETERS...... 56 13.1 X-ray Diffraction...... 56 13.2 FTIR...... 56

14 APPENDIX 3: CALIBRATIONS FOR X-RAY DIFFRACTION ...... 57 14.1 SIMPEDS Calibrations...... 57 14.2 GK 2.69 cyclone Calibrations ...... 57 14.3 IOM sampler wih foam separator...... 58

15 APPENDIX 4: CALIBRATION FOR THE INDIRECT ANALYSIS PROCEDURE ...... 59 16 APPENDIX 5: CALIBRATIONS FOR INFRARED – DIRECT ON-FILTER ANALYSIS ...... 60 16.1 SIMPEDS ...... 60 16.2 GK 2.69 cyclone ...... 60 16.3 IOM sampler with foam separator...... 61

17 APPENDIX 6: CALIBRATIONSii FOR INFRARED – INDIRECT ANALYSIS ...... 62

18 APPENDIX 7: XRD SCANS OF ABSORPTION TEST MATERIALS ... 64

19 APPENDIX 8: INFRA RED ABSORNACES ...... 67 20 APPENDIX 9: PARTICLE SIZE DISTRIBUTIONS FROM SIMULATED WORK TASKS...... 69 21 APPENDIX 10: GRAVIMETRIC ANALYSIS WITH SIMULATED WORK ACTIVITIES...... 71

22 APPENDIX 11. DUST LOSSES FROM PVC FILTERS ...... 72

23 APPENDIX 12: STAGE 4 XRD COMPARISON WITH SIMPEDS ...... 75

24 APPENDIX 13 STAGE 4: COMPARISON WITH PGP 10 CYCLONES 77

25 SITE 1: FOUNDRY VISIT...... 79

26 SITE 2: POT/BRICK MANUFACTURE ...... 94

27 SITE 3: CERAMICS MANUFACTURE...... 105

28 SITE 4 CONSTRUCTION...... 115

iv

29 SITE 5: QUARRY VISIT...... 123

v

EXECUTIVE SUMMARY

Objectives

This project investigated the technical feasibility of using high volume personal respirable dust samplers (> 4 l/min) to improve the reliability of measurements of respirable crystalline silica (RCS) using x-ray diffraction (XRD) and Fourier Transform Infra Red (FTIR) analysis in order to support further reductions in the United Kingdom’s (UK) Workplace Exposure Limit (WEL).

The objective was to evaluate the samplers against the following criteria:

• Can the nominal flow rate of the sampler be maintained over typical sampling periods (within ± 5%)?

• Does the sampler have a performance comparable with the SIMPEDS respirable dust sampler recommended in the Health and Safety Executive (HSE) method MDHS 14 for dust sampling (HSE 2000)?

• Is the sampler be comfortable to wear without interfering with the activity of the worker and is it applicable for use in UK workplaces?

• Does the increased mass of dust collected with the sampler affect RCS measurement by XRD and FTIR?

• Does the increased mass of dust collected with the sampler increase interferences?

• Would the precision of measurement at the proposed WEL of 0.05 mg.m-3 be equal to or better than the precision of ± 12% (2σ) at the current WEL of 0.1 mg.m-3?

Five high flow rate (> 4 l/min) personal samplers were evaluated in this project: a GK 2.69 cyclone, modified to use 25 mm rather than 37 mm diameter filters, operating at 4 l/min; the PGP 10 cyclone, operating at 10 l/min; a prototype IPP impactor, operating at 8 l/min; an IOM sampler with a foam separator selected to operate at 4 l/min; and the CIP 10 sampler, operating at 10 l/min.

Main Findings

Sampling

a) Pump performance

• For a flow rate of 4 l/min through a filter with a 0.8 μm pore size, the sampling pump needs to cope with a backpressure greater than 25 inches of water under load.

• The sampling pump used with the modified GK 2.69 cyclone failed to maintain the nominal flow rate within ± 5% in the field tests when conducting 4 hour sampling.

• The pumps used with the PGP 10 and IPP samplers in the field trails maintained their nominal flow rates during the sampling periods.

b) Respirable dust collection

• The GK 2.69 cyclone, operating at 4 l/min, and the PGP 10 cyclone, operating at 10 l/min, gave comparable results to the SIMPEDS for respirable dust.

vi

• The prototype IPP impactor tended to slightly under sample respirable dust compared with the SIMPEDS (~15%), possibly associated with the potential for loss of dust from handling the filter.

• The IOM sampler with a foam separator for respirable dust selected to operate at 4 l/min tended to under sample compared with the SIMPEDS (~22%) at high loadings (> 1.26 mg) and results were more variable than for other samplers. It is suspected that the collection of dust eventually changes alters the performance of the foam separator.

• In these tests, the CIP 10 sampler tended to under sampled respirable dust compared with the SIMPEDS (28 – 34%). c) Practical use

• The PGP 10 cyclones and their pumps were heavy and interfered with the worker’s activities.

• The IPP impactor is small and compact but required the same heavy pumps that were used for the PGP 10 cyclones.

• The GK 2.69 cyclones and their pumps were lighter compared with the PGP 10 cyclone and pump, and did not interfere with the workers activities.

Measurement a) X-ray diffraction

• The GK 2.69 cyclone and the PGP 10 cyclone gave RCS results that closely matched those obtained with the SIMPEDS.

• The IOM sampler with foam separator gave RCS results that closely matched the SIMPEDS, when the sample loading was < 1.26 mg. However, the XRD results reported less respirable quartz.

• The IPP impactor obtained RCS results that were comparable to those obtained with the SIMPEDS when measuring Arizona Road Dust (ARD), which contains about 70% crystalline silica. However, when sampling in simulated workplace conditions it tended to report less crystalline quartz than the SIMPEDS, PGP 10 cyclone and GK 2.69 cyclone. This could have been due to losses during recovery of dust for analysis.

• Using a silver filter can reduce the calibration uncertainty of by about half.

• XRD response to crystalline silica can be reduced for samples with heavy dust loadings (> 2 mg) because of absorption by sample matrix. This study found that accurate corrections for the intensity are possible using the reflections from a silver filter as an internal standard with an indirect analysis approach, in which the dust on the sample filter is recovered and deposited onto a silver filter for analysis using filtration apparatus with 15 mm diameter funnel.

• The study also found that the response from the measurement of 2% respirable quartz, in a material with a mass absorption coefficient of 156 cm2/g, across a 15 mm diameter deposit on a filter was a linear up to 2 mg of dust. The absorption and depth effects will

vii

not be as significant when measuring RCS in an air sample filter with the deposit spread over an active diameter of 20-22 mm. This finding indicates that the current guidance in HSE method MDHS 101 (HSE 2005) is unnecessarily cautious for XRD measurements and should be increased to 3 mg.

• The response of 2% quartz in a material with mass absorption coefficient similar to quartz (34.8 g/cm2) was linear up to about 3 mg when analysing a 15 mm diameter deposit with the instrumental parameters used at HSL.

• Little benefit is gained by using high volume samplers when the percentage of quartz in a matrix with an absorption coefficient of 71 cm2/g is less than 1% since no improvement in intensity is made despite the greater mass of dust. This is the almost same percentage of quartz in an air sample containing 0.05 mg.m-3 of RCS in 4 mg.m-3 of respirable dust.

b) Infrared analysis

• Infrared analysis should not be used as an analysis technique for samples with high loadings (> 2 mg) because the absorbance, when using the direct on-filter analysis procedure, is dependent on the matrix.

• FTIR gave comparable results to XRD when analysing ARD.

Can a lower WEL for RCS be measured reliably with the available apparatus?

None of the samplers fully met all of the success criteria, but not all the difficulties are insurmountable.

The two samplers that had most consistent performance in these tests, were the PGP 10 cyclone and the modified GK 2.69 cyclone.

The PGP 10 cyclone provides reasonably accurate results when the process of recovery from the sample filter is done carefully. However, the sampler and pump are heavy and its sample train can interfere with the workers activity.

The modified GK 2.69 cyclone gives a very similar measurement performance in laboratory tests, but the pumps used with it failed to maintain a consistent flow (± 5%) in the field tests.

The IPP impactor tended to under sample respirable dust compared with the SIMPEDS. This sampler could meet the respirable dust sampling convention prescribed in EN 481 and ISO 7708, if used carefully to avoid sample losses when handling the sampling filters, but greater confidence in its performance would be needed before it could be put in widespread use. The IPP is compact but requires use of the same heavy pump as the PGP 10 cyclone.

Options

Four possible options for are outlined below taking into account the following considerations:

• The performance of the samplers in the tests.

• The potential additional costs and consequences of HSE changing its measurement philosophy.

viii

• Potential future developments.

Option 1: Retaining the existing samplers and using a silver filter for sampling could achieve a further reduction in the WEL

Sampling using silver filters could halve measurement uncertainty. In turn, this could enable a reduction in the WEL towards 0.05 mg.m-3 if short sampling times (< 4 hours) were no longer permitted for making measurements to assess compliance with the WEL. This option would allow HSE to retain its current measurement philosophy of using the direct on-filter analysis technique specified in MDHS 101 (HSE 2005). The use of silvers filters and a direct on-filter analysis procedure is already used in Italy where an insurance limit of 0.05 mg.m-3 exists.

However, many of the pumps used at present might not operate successfully with the backpressures experienced when samplers are used with silver filters and some industries might have to purchase new pumps. Furthermore, the exact improvement in measurement precision with silver filters has not yet been fully quantified and the improvement could be sampler dependent. The use of FTIR analysis is not possible with silver filters and many small laboratories cannot afford to replace their FTIR spectrometers with XRD instruments.

Option 2: Use the modified GK 2.69 cyclone

The flow rate used with the modified GK 2.69 cyclone is sufficient to collect enough dust to achieve a significant improvement in measurement precision and the use of silver filters would provide a further enhancement by allowing accurate corrections for heavy sample loadings. The use of the modified GK 2.69 cyclone will allow HSE to retain its current measurement philosophy of using direct on-filter analysis. The use of a silver filter with a larger pore size (1.2 µm) could reduce backpressure and cooperation with industry may improve pump performance. A newly available pump, known as the SG 10–2, should be able to work with the backpressures experienced by the modified GK 2.69 cyclone, but at the time of the publication of this report this pump would cost about £1k per unit.

The success of using this option is dependent on:

a) A manufacturer adopting the design and making the sampler commercial available

b) The availability of a pump to operate with the backpressures experienced.

Both factors will require cooperation with industry.

Option 3: Use the unmodified GK 2.69 cyclone

This sampler is commercially available and the use of a 37 mm filter reduces the backpressure.

However, an indirect analysis procedure is needed which could introduce inaccuracies due to sample losses and HSE would have to change its measurement philosophy, which would incur additional costs.

Option 4: Use the PGP 10 cyclone

The flow rate is more than sufficient to collect enough dust to improve the measurement precision and the sampler and pump are commercially available. It is however, very expensive and the purchase of a single unit with pump (> £2k) would represent a significant cost to a small company. The sampler is also considered heavy, interferes with the worker’s activity and received many adverse comments from workers. Many of the concerns raised could be

ix

overcome with the use of a harness and backpack to support the sampling equipment on the worker.

Recommendations

Recommendations for policy

There is a history of occupational hygienists taking half shift (~4 hour) samples in the UK when assessing a worker’s exposure to RCS. HSE should work with the British Society (BOHS) to encourage occupational hygienists to take full shift samples, whenever possible, even if the task involving work with silica containing materials is for less than 4 hours. The differences in strategies for task specific sampling exercises and the taking measurements to demonstrate compliance with the law should be emphasised in HSE’s guidance. Measurements taken for less than a full shift should not be considered as sufficient to check compliance with a WEL. Sampling for the majority of the working day will reduce the uncertainty of the reported result and a longer sampling period is more representative of an individual’s exposure as it makes better consideration of failures of controls from other processes or contamination.

Additional time is needed to assess if the manufacturers of pumps can improve the performance of pumps operating under load, to enhance the options for sampling in future and reduce the incidence of pump failure experienced at present. HSL also needs to work with the manufacturers of the GK 2.69 cyclone, when and if a suitable pump is available, to assess if the modified version can be made commercially produced. A short ANOVA evaluation of the improved precision when using silver filters with the existing equipment is underway and the results will soon be available. The use of these samplers with silver filters, combined with longer sampling times and better pump performance, could be sufficient to reduce the WEL for RCS. This would be the most cost effective approach for HSE and industry.

If the pumps cannot be improved to cope with the backpressure of the modified GK 2.69 cyclone and a manufacture cannot be found then HSE should encourage the use of the unmodified GK 2.69 cyclone and adopt a different measurement strategy.

However, change to a measurement strategy reliant on the use of samplers with filters greater than 25 mm in diameter would inevitably increase costs for industry/HSE because:

a. An indirect analysis approach is required and the analytical process is more time consuming.

b. The analytical procedure could require the purchase of additional specialised equipment to ash or to weight 37 mm diameter filters in most laboratories.

c. Additional funding would be needed for HSE to communicate its expectations to the occupational hygiene community and industry.

The PGP 10 cyclone is commercially available but the introduction of this sampler into widespread use is unlikely, since the equipment procurement costs are significant and its use is more limited than the GK 2.69 cyclone. The PGP 10 cyclone will probably find a role with specialist activities such as task-specific sampling with short sampling times (< 2 hours), static and verification of enforcement sampling by HSE, where the levels of dust are low. Further work is needed to assess if the difficulties of using this device could be overcome through the use of a harness a new pump.

x

Recommendations for sampling and analysis

Consideration should be given to creating a stronger link between the sampling environment, working patterns, selection of the sampler and the type of analysis procedure required to measure the samples, so that the sampler and measurement approach can be linked to specific sampling applications. HSE should consider the development of a guidance document for sampling and analysis, similar to HSG 248 "Asbestos: The analysts' guide for sampling, analysis and clearance procedures'.

The CIP 10 sampler should not be used in the UK since it can potentially under sample respirable dust compared with the SIMPEDS when carrying out some common workplace tasks.

The Workplace Analysis Scheme for Proficiency (WASP) proficiency-testing programme should trial the use of silver filters to confirm any improvement in the reliability of measurements between laboratories.

If direct on-filter XRD analysis is used, sampling should be carried out using 25 mm diameter, 0.8 µm or 1.2 µm pore size, silver filters, since these offer advantages over the use of the GLA 5000 PVC filters.

FTIR analysis should not be used for the analysis of RCS except in specific specialist applications (e.g. for foundry samples and for coal samples, where the carbon is anthracite) and then only when the loading of the sample is < 1 mg.

The HSE method for the analysis of RCS should be rewritten to exclude FTIR analysis. Originally, the XRD and FTIR analysis methods were separate before they were combined as MDHS 101 (HSE 2005). The FTIR and the XRD analysis procedure should be separate MDHS methods as they were previously, so that the different analytical procedures for different sampling equipment with XRD analysis and the limitations of the FTIR analysis procedure are more clearly defined. FTIR analysis can still have a specialist role.

xi xii

1 INTRODUCTION

‘The Commission considered …………… a lower level (WEL for respirable crystalline silica) was not practicable at present, if it could not be measured, but advised that HSE should be aiming to achieve a further reduction’. (Health and Safety Commission, Minutes 2006)

1.1 BACKGROUND

This document details the results from a project to investigate the technical feasibility of using high volume samplers to improve the measurement precision to enable the Health and Safety Executive (HSE) to reduce the Workplace Exposure Limits (WEL) for respirable crystalline silica (RCS) from its present level of 0.1 mg·m-3 to 0.05 mg·m-3.

HSE document EH 74/4 (HSE 2004) describes the risk of developing after 15 years of exposure to RCS as about 0.5% (1 person in 200) at an air concentration of 0.04 mg.m-3 and 2.5% at 0.1 mg.m-3. Other studies indicate a similar risk to worker health at lower concentrations (NIOSH 1998). This has increased the pressure on legislative authorities to reduce exposure limits to a level where the studies suggest that the risk of damage to the health of an individual is minimal. The European Scientific Committee for Occupational Exposure Limits (SCOEL) has recommended that, to eliminate silicosis, European occupational exposure standards should be set below 0.05 mg.m-3. However, evidence presented to the Health and Safety Commission (HSC) and HSE’s Advisory Committee on Toxic Substances (ACTS) demonstrated the poor reliability of measurements of concentrations of RCS in air at 0.05 mg.m3 made when collecting samples at the flow rates of the samplers then available (HSE 2006). HSC therefore accepted HSE’s recommendation that the WEL for RSC should be lowered from its earlier value of 0.3 mg.m-3 to 0.1 mg.m-3, rather than 0.05 mg.m-3. If a WEL is set at a level where the analytical method approaches its limit of quantification, exposure measurements in this region would be unreliable because of high analytical variability. To enforce the law the regulatory authorities have to be confident that measurements between laboratories are reliably consistent. This is especially important when considering implementing lower exposure limits than were previously enforced. Most samplers in the United Kingdom (UK) operate at flow rates between 1.7-2.2 l/min. This project investigated the use of higher flow rate samplers to collect more dust for measurement. Samplers to be tested were selected on the basis that they should be able to double the amount of dust collected, so potentially halving measurement precision. Therefore, samplers with flow rates less than 4 l/min were excluded from this project. The sensitivity of the instrumental techniques used to determine the mass of RCS in an air sample is dependent on the physical properties of the dust collected, such as the depth of the deposition, the absorption of radiation by the sample matrix, the distribution of the dust across the filter and the particle size. Some of these factors can reduce, rather than increase, the sensitivity of the analysis. This study investigated the performance of measurements for quartz (the most common polymorph of RCS) when using the high volume personal samplers available in 2008.

1.2 PRINCIPLES OF THE ANALYSIS TECHNIQUES

The two most frequently used instrumental techniques for the analysis of respirable crystalline silica are x-ray diffraction (XRD) and Fourier Transform Infra Red (FTIR) spectrometry.

1

1.2.1 X-ray diffraction

Powder XRD measures the x-ray radiation reflected from a material. When a crystalline material is placed in focal point of an x-ray beam and the orientation of a plane of atoms within the structure of crystals in the material fulfils the Bragg equation nλ=2dSinθ, where λ is the wavelength of the radiation, d is the lattice spacing and θ is the diffraction angle (the angle between the plane of atoms and the detected signal), constructive interference of the radiation occurs and a peak is observed above the background signal. Amorphous or semi-crystalline materials influence the background in a variety of ways but do not produce scans with defined peaks. The penetration of the x-ray beam into the sample is dependent on the intensity of the radiation, the density of the material and how the material absorbs the radiation. For example, copper radiation generally penetrates 100 µm into the surface of a powder sample. Devices called slits and masks are used to control the area of the sample illuminated and the amount and type of radiation that is accepted by the detector. Figure 1 shows the arrangement of the optics in XRD spectrometer.

Figure 1: Optics in an XRD spectrometer (reproduced with the permission of Panalytical)

1.2.2 Infra red analysis

Infra red analysis measures the amount of infrared energy that is absorbed by the sample. When chemical bonds within a molecule are subjected to infrared radiation many will vibrate or stretch and absorb the infrared energy. Each bond vibration or stretch has distinct absorption within the infrared spectrum and a chemical is identified from the spectrum pattern. Respirable quartz has two distinct absorbencies at 800 cm-1 and 780 cm-1. FTIR instruments use an interferometer to simultaneously scan the infrared range. This enables the FTIR to have a fast analysis time compared with XRD analysis and the instrument is cheaper to manufacture and

2

maintain. However, a sample needs to be thin enough to allow the transmission of the infrared radiation through the sample, so heavily loaded air sample filters cannot be analysed reliably. Also the infrared beam measures a small proportion of the total area of the filter (7-8 mm) and repeatability of the measurement is dependent on the homogeneity of the dust distributed on the air sample filter or in a disc of potassium bromide.

1.2.3 Analytical approaches

Generally, the analytical procedure employed to analyse RCS in airborne dust is dependent on the type of sampler used to collect the aerosol. In countries where samplers using filters with a diameter greater than 25 mm, or where foam is used as the dust collection medium, methods were developed to recover the dust from the filter or foam and to concentrate the dust onto a smaller filter, so that all the collected dust deposit is analysed by the instrument. These procedures are referred to as indirect analysis procedures since they involve a process to recover the dust from the air sample filter. In many countries where samplers with a 25 mm diameter filter are used, the air filter is analysed for RCS without any pre-treatment. This is referred to as direct on-filter analysis and is the approach adopted by HSE in its published method MDHS 101 (HSE 2005). If smaller diameter filters are used in samplers that operate at higher flow rates this can lead to higher backpressures that can strain the pump and cause failures. Therefore, designers of higher flow rate devices generally opt to reduce the backpressure by using a filter with a larger surface area.

3

2 STAGE 1: SELECTION OF EQUIPMENT INCLUDED IN THE STUDY

2.1 SELECTION OF SAMPLERS

The samplers selected for study are shown in Table 1.

When this project started, there were few commercially available samplers designed for personal sampling of respirable dust that operated at flow rates of 4 l/min or greater. Two of the five samplers included in the study were adapted at HSL. The GK 2.69 cyclone was modified to operate with a 25 mm diameter filter for use with a direct on-filter analysis approach; and the IOM sampler was used with a foam separator, selected at HSL to operate at 4 l/min rather than 2 l/min. The other samplers tested were a prototype IPP impactor, operating at 8 l/min, which was loaned to HSL by the manufacturer, and the PGP 10 cyclone and the CIP 10 sampler, both operating at 10 l/min,. The reference sampler used in the study was the Safety in Mines Personal Dust Sampler (SIMPEDS), manufactured by Casella Ltd, a version of the Higgins-Dewell cyclone, operating at 2 l/min. The SIMPEDS was used as the reference sampler because it is the sampler recommended for respirable dust in MDHS 14/3 (HSE 2000).

Manufacturers, other institutes and HSL have tested the samplers evaluated in this study to assess their conformance with the ISO/CEN convention for respirable dust (ISO 1995, CEN 1993) and all have a reasonable match to the respirable convention. The purpose of this study, therefore, was not to assess this, but to examine differences in measurements of RCS in the dust collected by the samplers. In most experiments, two samplers of each type were used to sample each test material. However, only one prototype IPP impactor was available.

2.2 SELECTION OF FILTERS

Both 25 mm diameter 0.8 µm pore size, silver filters and 25 mm diameter, 5 µm pore size, Gelman GLA 5000 PVC filters were included in the study for those samplers able to be used with the direct on-filter analysis approach, such as the SIMPEDS, the adapted GK 2.69 cyclone and the IOM sampler with the modified foam separator. The filters or foams recommended by the manufacturers were used for the other samplers. These included 37 mm diameter, 5 µm pore size, Gelman GLA 5000 PVC filters for the IPP impactor and 37 mm diameter, 8 µm pore size, cellulose nitrate filters for the PGP 10 cyclone.

2.3 SELECTION OF SAMPLING TRAINS

2.3.1 Assessment of backpressure on filters at 4 litres/minute

In order to assess the ability of sampling pumps to operate with the backpressure caused by a 25 mm diameter filter operating at 4 l/min, pressure differentials were measured for 0.8 µm, 1.2 µm and 3 µm pore size mixed cellulose ester (MCE) filters mounted in an in-line filter holder with a calibrated micro-manometer connected to each side (Figure 2).

4

Table 1: Samplers tested in the project Apparatus Make Nominal Flow Comment Picture Rate (l/min) Adapted GK 2.69 HSL/BGI 4 GK 2.69 cyclone cyclone adapted at HSL for use with 25 mm diameter filters

IOM sampler HSL/SKC 4 IOM sampler with foam with foam separator selected by HSL to separate respirable dust at 4 l/min

IPP impactor SKC 8 Impactor onto (prototype) PVC filter

PGP 10 cyclone GSA 10 Specified for use Messgerätebau with celluose GmbH nitrate filters

CIP 10 sampler Arelco 10 Spinning foam in head Integral pump unit with battery

5

Figure 2 Apparatus for the testing of backpressure across filters

Charts showing the relationship between pressure drop and flow rate through each pore size of MCE filter are shown in Appendix 1.

Table 2 lists the pressure drop, in inches of water, derived from the charts in Appendix 1, for the different pore sizes of MCE filter, when subject to a flow rate of 4 l/min.

Table 2: Pressure drop at 4 l/min Filter pore size Pressure drop (µm) (inches of water) 0.8 23 1.2 16.4 3.0 9.6

The results indicate that if filters with a pore size of 0.8 µm are used then pumps that cannot cope with a backpressure of 23 inches of water are not likely to achieve or maintain a flow rate of 4 l/min. This eliminated from the project the majority of pumps currently used at HSL. Three sets of pumps from three different manufacturers were run with the 3 µm pore size MCE filter at 4 l/min and of these only one set of pumps ran for more than six hours. This highlights the need for high quality pumps.

2.3.2 Pressure drop across silver filters

Silver filters offer advantages of lower background, lower calibration line uncertainty and better measurement repeatability when compared with the membrane filters based on cellulose or other polymers that are also used for XRD analysis. Their disadvantage is that they are relatively expensive compared with the other membrane filters, cannot be used with infrared analysis and slowly absorb sulphur dioxide and chlorine from the environment, so slightly increasing their weight from day to day. Silver filters of a number of different pore sizes are available but the particle collection efficiency for respirable dust of silver filters with larger pore sizes might not be adequate (Lee K and Ramamurthi M 1993) Also it is thought that for silver filters of larger pore size some dust could be masked from the x-ray beam if the dust is drawn into the pores of the material. It was therefore decided to test the backpressure associated with 6

0.8 µm pore size silver filters with a view to using these in this work. Figure 3 shows the results of the tests concerned.

16.00

14.00 y = 0.0031x R2 = 0.9782 12.00

10.00

8.00

6.00

4.00 Pressure drop (Inches water)

2.00

0.00 0 1000 2000 3000 4000 5000 Flow rate (cm3 min-1)

Figure 3 Pressure drop across a 0.8 µm pore size silver filter

Surprisingly, the back pressure on 0.8 µm pore size silver filters operating at a flow rate of 4 l/min was only 12.4 inches of water, which is not much greater than 3 µm pore size silver filter and better than 0.8 µm pore size MCE filters. They were therefore fit for purpose.

2.3.3 Performance of sampling train under load

Pumps were purchased that operated at the required backpressure and a series of tests were performed to assess now these pumps would also operate under load at the nominal flow rates of the samplers. Buck Libra pumps were used for the devices with a flow rate of 4 l/min and Leyland legacy pumps from SKC were used with the IPP impactor and PGP 10 cyclone. The samplers , fitted with pre-weighed filters and operated at their nominal flow rates, were exposed to an aerosol of fine Arizona Road Dust (ARD) in a small dust tunnel (Figure 4) at an approximate concentration in air of 40 mg.m-3, estimated using a real time dust monitor.

7

Location of the samlpers

. Figure 4 Samplers tested in dust tunnel

The flow rate was checked periodically with a Gilibrator bubble flow meter and the mass collected on each filter weighed to assess the loading at which the sampling train would not be able to meet the requirement, prescribed in EN 1232 (BSI 1997), that the flow rate should be maintained within 5% of the quoted value.

The results are shown in Figure 5

17.50

) 15.00 -1 12.50 10.00 7.50 5.00 2.50

% drop in flow rate (l min 0.00 -2.500.00 5.00 10.00 15.00 20.00 25.00 Dust load (mg) GK2.69 modified - 0.8µm silver filter IOM with foam - 0.8µm silver filter

PPI08 impactor - 5µm PVC filter PGP 10 - 8 µm cellulose nitrate filter

SIMPEDS - 0.8 µm silver filter . Figure 5 Drop in flow rate with dust loading

All the samplers, except the IOM sampler with foam separator, maintained their required flow rates up to a loading of about 5 mg or greater. The reduced performance of the IOM sampler was believed to be due to the collection of dust on both the foam separator and the filter increasing the backpressure to a greater extent than just the filter. One modified GK 2.69 cyclone failed the criteria with just 1.5 mg of dust on the filter but it was decided to 8

continue its use because in another run more than 5 mg of dust was collected on the filter before the flow rate drop exceeded 5% and the recorded back-pressure at maximum loading was well within the pump’s specification of 55 inches of water. Figure 6 shows the backpressures obtained with each sampler during the loading of the filters. The performance of the new pumps purchased for the modified GK 2.69 cyclone and IOM sampler with foam separator, applying the criteria developed in 2.3.1 that they should cope with a backpressure of 23 inches of water, was better than the existing equipment at HSL.% Table 3 list the samplers, the filter medium and the maximum loading before failure.

Table 3: Maximum loading of sampling trains Sampler Medium Pore size Maximum loading (µm) (mg) GK 2.69 cyclone Silver filter 0.8 ~5 IOM with foam separator Silver filter 0.8 ~ 2 PGP 10 PVC filter 5 ~ 8.3 PGP 10 Cellulose nitrate 8 > 10 PPI impactor PVC filter 5 > 15

35.00

30.00

25.00

20.00

15.00

10.00

5.00 Pressure drop (inches water) 0.00 0.00 5.00 10.00 15.00 20.00

GK modified - 0.8µmDust silver load filter (mg) IOM with 75ppi foam - 0.8µm silver filter PPI08 impactor - 5µm filter PGP10 - 5 µm pore size PVC filter SIMPED - 0.8 µm pore size silver filter

Figure 6 Change in backpressure with loading

9

3 CALIBRATIONS

3.1 CALIBRATIONS FOR X-RAY DIFFRACTION

3.1.1 Quartz x-ray diffraction pattern

Quartz has four reflections in its x-ray diffraction pattern that are commonly used by laboratories for quantification. When using copper Kα radiation, these are: 1) the 100 reflection (20.9 degrees 2θ) with a relative intensity of about 25%; 2) the 101 reflection (26.6 degrees 2θ) with a relative intensity of 100%; 3) 112 reflection (50.1 degrees 2θ) with a relative intensity of about 9-14%; and 4) the 211 reflection (60.1 degrees 2θ) with a relative intensity of about 7-9%. The XRD instrument was calibrated for the 100, 101 and 112 reflections. A programme, established in x-pert industry software (Appendix 2), was applied to all the calibration samples so that all the samples were analysed using the same instrumental parameters.

3.1.2 Direct on-filter analysis

Calibration samples were prepared and analysed for the SIMPEDS, the modified GK 2.69 cyclone and the IOM sampler using the procedure described in MDHS 101 (HSE 2005). Separate calibrations were generated for the 25 mm diameter, 0.8 µm pore size, silver filters and the 25 mm diameter, 5 µm pore size, GLA 5000 PVC filters. XRD calibrations for both filter types are shown in Appendix 3.

3.1.3 Indirect filter analysis

Separate calibrations were established for the IPP impactor and the PGP 10 cyclone following the procedure outlined in NIOSH 7500 (NIOSH 2003). A suspension of HSE quartz standard A9950 in isopropanol was prepared and aliquots of the suspension were filtered onto 25 mm diameter, 0.45 µm pore size, silver filters or 25 mm diameter, 5 µm pore size, GLA 5000 PVC filters to produce calibration samples with loadings in the range 1 mg – 6 mg. The calibrations established by the analysis of these samples are shown, corrected for depth effects using the reflection from a silver filter as an internal standard, and uncorrected, in Appendix 4.

3.2 CALIBRATIONS FOR INFRARED ANALYSIS

The calibration samples prepared on GLA 5000 PVC filters for indirect filter analysis by XRD (3.1.3) were also analysed on a Perkin Elmer FTIR spectrometer using the principal quartz infrared absorptions of 800 cm-1 and 780 cm-1 to establish separate calibrations for each sampler type. The instrumental parameters used in the analysis are given in Appendix 3 and the calibrations for the direct on-filter analysis methods and for the indirect analysis methods are shown in Appendix 5 and Appendix 6.

3.3 DISCUSSION

It has been suggested that the absorption of copper x-ray radiation by silver might reduce the intensity of the signal detected by the XRD instrument when using silver filters. However, it can be seen that the trend lines in Appendix 3 are very similar. The silver filter calibrations have slightly better regression coefficients than the GLA 5000 PVC filter calibrations, except for the samples from the IOM sampler. The IOM sampler has a metal sample cassette and difficulties extracting silver filters without damage could have led to sample loss . It is also possible that the IOM cassette might not seal as effectively when using a silver filter as when compared with other filter types and this could cause migration of dust to the edges of the filters and subsequent loss when the top part of sampling cassette is removed (in a circular motion) to gain access to 10

the filter beneath. All results from the FTIR analysis gave straight trend lines except for the calibrations for the indirect analysis procedure which were curved and more variable as the depth of dust on the calibration samples reaches the point at which transmittance of the infrared radiation is so small that the relationship of absorbance with mass on the filter is no longer linear (the Beer-Lambert law).

11

4 STAGE 2B: ABSORPTION AND DEPTH EXPERIMENTS

4.1 EVALUATION OF ABSORPTION AND DEPTH EFFECTS IN X-RAY DIFFRACTION ANALYSIS

4.1.1 Factors that influence the measurement of crystalline silica using X-ray diffraction

Two major factors that influence the measurement of crystalline silica in a significant amount of another dust on a filter by XRD are the depth of the sample and the absorption of the matrix. The current HSE method for the analysis of respirable crystalline silica (HSE 2005) assumes the depth of the sample is so thin that the x-ray radiation penetrates the whole sample and reflected radiation is not absorbed. This assumption is satisfactory if samplers operating at 2 l/min are used and if the sample does not include a high absorbing element, such as the absorption of copper radiation by iron. However, when an increased mass of dust is collected this assumption is broken. A limited study was developed to examine these effects on the measurement of a small amount of crystalline silica in a matrix because crystalline silica is a trace component of many mineral products.

4.1.2 Materials

Three different matrix materials were selected for study that:

a. were minerals encountered in industry

b. offered a range of absorption coefficients (an absorption coefficient is a measure of the ability of the mineral to absorb x-ray radiation, with high numbers ~200 cm2/g indicating a very strong absorbance)

c. did not have peaks that interfere with the 101 reflection of quartz.

Table 4 provides details of the minerals selected as matrix materials and of the quartz standard used in the study.

Table 4: Minerals involved in the absorption study Mineral Formula Mass Absorption Coefficient Origin 2 Olivine (MgFe)2SiO4 154.8 cm /g Single Crystal Richard Taylor Minerals 2 Calcite CaCO3 71.5 cm /g Alfa Aesar 99.99% 2 Talc Mg3Si4O10(OH)2 31.4 cm /g BDH Chemicals Ltd fine Talc purified by acids 2 Quartz SiO2 34.8 cm /g HSL Standard A9950

XRD scans of each of the selected matrix materials (see Appendix 7)did not show the presence of crystalline silica.

The olivine and talc samples were ground for more than 8 hours in an attempt to obtain powder within the respirable size range. This was not completely effective as a proportion of the ground material had particle sizes larger than 20 µm, although the larger particles sizes could have been due to unresolved agglomerations. The calcite was not ground because, when examined through a microscope and compared against a calibrated micrometer, it appeared to be within the

12

respirable size range. However, results from particle size measurements performed using a laser light scattering instrument (an Horiba 390L manufactured by Particle Analysis Ltd) showed a distribution of particles that was slight larger than the respirable range. The median particle sizes for each mineral are shown in Table 5.

Table 5: Median particle size Mineral Median Particle Size (µm) Calcite 23.2 Olivine 8 Talc 13.9 Quartz 2

4.1.3 Procedure

Two suspensions of each matrix material in isopropanol were prepared by following the procedure outlined in NIOSH 7500 (NIOSH 2003). A known mass of mineral dust was mixed with an amount of HSE quartz standard A9950 and the mixture was then dispersed in a known volume of isopropanol (determined from the weight and the density of isopropanol). The proportion of HSE standard A9950 used in each of the two mineral dust suspensions was equivalent to quartz at the current WEL (0.1 mg.m-3) and at half the WEL (0.05 mg.m-3) in matrix material at the WEL for respirable dust (4 mg.m-3). This resulted in suspensions containing approximately 1% and 2% crystalline silica. Fortuitously, these percentages of quartz are levels that are sometimes encountered in commercially available products of these minerals .

Aliquots of each suspension were then taken to produce loadings within the range from 1 mg – 6 mg on each filter. Each aliquot of suspension was filtered onto a 25 mm diameter, 0.45 µm pore size, silver filter for analysis by XRD, or onto a 25 mm diameter, 5 µm pore size, GLA 5000 PVC filter for analysis by FTIR. However, the deposit did not extend across the whole filter because the filter funnel diameter was only approximately 15 mm. The expected mass of crystalline silica on each filter was calculated from the total mass of dust deposited and the percentage of quartz mixed with the matrix material. An example of the pipetting accuracy is given in Figure 7. The slope of the regression line for the preparation of a series of test samples of 2% RCS in talc is very close to the ideal value of 1.00, indicating the suspension was mixed uniformly and that the pipetting of aliquots produced repeatable test samples.

13

6

Recorded mass = 1.0129x 5 R2 = 0.9982

4

Results 3 Linear (1:1 Relationship) 2 Recorded Mass (mg) Recorded

1

0 0123456 Predicted Mass (mg)

Figure 7 Accuracy of loading test samples

4.1.4 XRD results for 1% quartz in talc

Figure 8 compares the mass of quartz recorded by the XRD instrument for the measurement of approximately 1% crystalline quartz in talc, with the expected value of quartz in the test sample derived from the weighed mass of dust collected on the filter and the percentage of quartz standard in the mineral suspension. There are two data sets shown on the chart in Figure 8. One data set shows the results uncorrected and the other the results corrected for depth and absorption effects. The measurement for quartz was corrected following the procedure outlined in NIOSH 7500 (NIOSH 2003) using the silver line at 38.12 degrees 2θ as an internal standard. The recorded value is the average of the results from the 100, 101 and 112 reflections. Occasionally, the 100 reflection was excluded because of inconsistency with the results from the other two reflections and a suspected interference.

14

60

Corrected = 0.9981x 50 2 R = 0.9824

40

30

20 Recorded Value (µg) 1 % Quartz in Talc Corrected

10 1% Quartz in Talc Uncorrected Linear (1:1 Relationship)

0 0 102030405060 Expected Value (µg)

Figure 8 Measurement of 1% quartz in talc

Figure 9 compares the results for a mixture of dust containing 2.45% crystalline quartz in talc. Only the 101 and 100 reflections were used to obtain the recorded value because the reflection at 211 was subjected to interference.

15

120

Corrected = 0.9392x 100 2 R = 0.9772

80

60 2.5 % Quartz in Talc Uncorrected 2.5 % Quartz in Talc 40 Corrected Linear (1:1 Relationship) Recorded Mass of Quartz (µg) 20 Linear (2.5 % Quartz in Talc Corrected)

0 0 20 40 60 80 100 120 Expected Mass of Quartz (µg)

Figure 9 Measurement of 2.5% quartz in talc

Although, many of the corrected results do lie on the ideal 1:1 relationship line it can be seen that a number of results slightly under record the mass of RCS in the sample. It is thought that some of the variability is due to the inconsistency of the deposition of dust on the filter surface after filtration. Some of the deposits are slightly off centre. However, a correlation of 0.98 for the analytical range 20–120 µg should be considered reasonably good because this range is at the lower end of the validated calibration range for this procedure of 20–2000 µg (NIOSH 2003).

4.1.5 XRD results for calcite

4.1.5.1 XRD results for 1% quartz in calcite

The XRD results for 1% quartz in calcite, which has a mass absorption coefficient of 71.5 g/cm2, twice that of talc, have been separated into different charts for the 100 and 101 with the 211 reflections. This has been done to demonstrate the unexpected difficulties encountered when measuring quartz in this matrix. Figure 10 shows results for the 100 reflection of quartz (secondary peak at 20.9 degrees with a relative intensity of 25% compared with the 101 reflection) and Figure 11 shows the combined results for the 101 and 211 reflection (primary quartz peak at 26.67 and tertiary peak at 50.1 degrees) for a mixture of 0.93% crystalline silica in calcite. The results for the 101 and 211 were more consistent, although produced slightly low values compared with the expected results. The results for the 100 reflection produced very high results.

16

1% Quartz in Calcite 120 100 reflection corrected

2 1% quartz in calcite R = 0.9695 100 reflection 100 uncorrected Linear (1:1 Relationship) 80 Linear (1% Quartz in Calcite 100 reflection corrected) 60 Linear (1% quartz in calcite 100 reflection uncorrected) 40

Mass of Quartz Measured (µg) 20

0 0 50 100 150 Expected Mass (µg)

Figure 10 XRD response for the 100 reflection (1% quartz in calcite)

1 % quartz in calcite 101 120 and 112 Corrected

1% quartz in calcite 101 and 112 Uncorrected 100 Linear (1:1 Relationship)

80 Linear (1 % quartz in calcite 101 and 112 Corrected) Linear (1% quartz in 60 calcite 101 and 112 y = 0.8874x R2 = 0.9748 40 Mass of Quartz Measured (µg) 20

0 0 20 40 60 80 100 120 Expected Mass (µg)

Figure 11 XRD response for the 101 and 112 reflections (1% quartz in calcite)

17

4.1.5.2 XRD results for 2% quartz in calcite

The results for the analysis of 2% quartz in calcite are shown in Figure 12

2.5 % quartz in 120 calcite corrected 2.5 % quartz in calcite uncorrected 1:1 Relationship 100 Linear (1:1 Relationship) 80

60

40

Recorded Mass of Quartz (mg) 20

0 0 20406080100120 Expecetd Mass of Quartz (mg)

Figure 12 Measurement of 2% quartz in calcite

In Figure 12, the two most consistent results used to record the measured value were those from the 100 and 211 reflections. The most sensitive 101 reflection generally recorded a lower value compared with the other two results.

4.1.6 XRD results for 2% quartz in olivine

Figure 13 shows the results obtained when analysing 2% quartz in olivine, which has an absorption coefficient of 154 g/cm-1, four times that of quartz. The results were obtained from the 101 reflection since this reflection was the only peak relatively free from interference.

18

120 2 % quartz in olivine corrected

100 2 % quartz in olivine - uncorrected Linear (1:1 80 relationship)

60

40 Recorded Mass 101 Reflection 20

0 0 20406080100120 Expected Mass

Figure 13 measurement of 2% quartz in olivine

4.1.7 Discussion

It proved very difficult to grind all the crystals in the powders below the respirable size, and this could have affected absorption because of differences in the packing density of the particles on the filters. Larger particles could have ‘shadowed’ the smaller particles of respirable quartz powder or build deposits with a greater depth, since they pack less efficiently. This could have led to an excessively pessimistic impression of the effect of absorption and sample depth than is in fact the case.

The mixture of quartz in Talc represents the simplest case in this study because the absorption coefficient of Talc is similar to that of quartz (see Table 4). Effectively, the results shown in Figure 8 and Figure 9 show the effect of depth on the measurement of quartz rather than depth and absorption. The results for the line, corrected for depth and absorption, have a correlation coefficient of 0.982, which is very reasonable considering proximity the measurement to the limit of quantification of the analytical procedure. The results tend to deviate from the ‘ideal’ 1:1 relationship between the expected and the recorded loading when approximately 3.5 mg of dust are loaded on the filter. This is slightly above the calibration range specified for XRD analysis of crystalline silica in air in NIOSH 7500 (NIOSH 2003). The good correlation for Talc indicates that accurate results for the measurement of RCS are obtained when analysing a material with a mass absorption coefficient similar to quartz and when the matrix causes few interferences.

The corrected relationships for trace amounts of quartz in calcite are more variable. This is partly attributable to the effect of calcite peaks on the background of the adjacent silver reflections used for absorption correction and also to the proximity of the large calcite peak to the 100 reflection. An additional factor for the mixture containing 1% quartz in calcite was the low intensity of the quartz peaks measured in every sample, which added to the variability of

19

these measurements. Few results recorded greater than 20 µg for the measurements of the test samples containing 1% quartz in calcite, even for the most intense 101 reflection. The uncorrected trend lines for 1% quartz in calcite (see Figure 10 and Figure 11) do not show a significant relationship with the expected mass, indicating that, without correction, the limit of detection cannot be described in terms absolute mass but rather as a percentage of quartz in a matrix. This is because, although each test sample contained more quartz, the intensity of the quartz peaks did not change significantly.

The uncorrected lines for the 101 reflections in the charts for calcite and olivine with 2% quartz are similar (see Figure 12 and Figure 13). The uncorrected line tends to move away from the ideal relationship when there is about 40 µg of quartz in 2 mg of dust, which matches the upper limit of the calibration range given in the NIOSH 7500. This contrasts with the results in Figure 9, in which uncorrected results from samples with 60 µg of quartz in 3 mg of talc gave the expected result.

The test samples used in this work were prepared using an aliquot of suspension deposited onto a filter using a filter funnel with a diameter of about 15 mm. In practice, therefore, the dust loading at which absorption correction is required would be different from that determined because the diameter of the deposit is approximately 20 - 22 mm for air samples. A 20 - 22 mm diameter deposit has almost twice the area of a 15 mm diameter deposit, so absorption effects would be less for the same mass of particles on the filter. Theoretically, therefore, if the dust were uniformly spread over a filter with an active area of 20 - 22 mm diameter, then approximately 4 mg would need to be collected before absorption would affect the measurement result. However, the dust is not usually uniformly spread over the filter due to the sampling characteristics of the apparatus.

The results obtained in this work confirm the upper limits of the calibration range for a heavily loaded filter using the indirect analysis approach specified in NIOSH 7500 (NIOSH 2003).

4.2 EVALUATION OF THE EFFECT OF ABSORPTION ON FTIR ANALYSIS

4.2.1 Materials

The same materials were used in the evaluation of the effect of absorption on FTIR analysis as in the experiments carried out to evaluate absorption and depth effects in XRD analysis (see 4.1.2).

4.2.2 Procedure

Test samples containing dust mixtures of 2% quartz in each of the minerals and 1% quartz in calcite were prepared on GLA 5000 PVC filters for FTIR analysis. Each filter was scanned before loading using the same instrumental parameters as the for calibration samples. The test samples were prepared with loadings of about 1–6 mg of dust. Figure 14 plots the peak height of the 780 cm-1 absorbance obtained for each mineral.

20

0.12 2.5% quartz in talc 0.1 2.5 % quartz in olivine 1 % quartz in calcite 2.5% quartz in calcite 0.08 R2 Talc = 0.9806 0.06 R2 Calcite = 0.9016 0.04 Expected Mass of Quartz (µg) 0.02

R2 Olivine = 0.9287 0 0 20 40 60 80 100 120 140 Absorbance (cm-1)

Figure 14 infrared responses at 780 cm-1

140

120

100

80

60 2.5 % quartz in talc 2.5 % quartz in olivine 40 1 % quartz in calcite Expected Mass of Quartz (µg) 20 2.5 % quartz in calcite

0 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Absorbance (cm-1)

Figure 15 FTIR responses at 800 cm-1

4.2.3 Discussion

Appendix 8, compares a scan of a calibration sample containing 60 µg of A9950 quartz on a filter with scans obtained from test samples of a similar mass of quartz on a filter in a matrix of 21

talc, calcite or olivine. Figure 14 and Figure 15 show that the sensitivity of the peak height absorbance is dependent on the matrix of the material and that, although straight trend lines were obtained, for a successful analysis the FTIR calibration has to be matrix specific because of the different response of quartz in each mineral. The straight-line relationships shown in Figure 15 were obtained for the 800 cm-1 absorbance for quartz, despite an uncertainty about where to place the background points for the calculation of the peak height (Appendix 8). The ratio of the 780 cm-1 and 800 cm-1 absorbances for talc were not the same as for the calibration sample and this suggests possible talc interference at 800 cm-1. It is possible that the slope of the background had an effect on results and that use of software to deconstruct the absorbance of silica from the matrix could have produced a more reliable measurement. Although all the scans were corrected for the filter absorbance, this probably had little effect because of the magnitude of the absorbance of the matrix. These data indicate that FTIR analysis would not be a suitable analytical technique for very heavily loaded filters unless the calibrations were matched to the sample matrix.

22

5 STAGE 3: ASSESSMENT OF THE BIAS OF SAMPLERS

5.1 SAMPLING TESTS WITH ARIZONA ROAD DUST Aerosols to challenge the performance of the samplers were generated in the calm air dust chamber originally developed at HSL by Kenny et al (1999) for work on inhalable samplers (Figure 16). This is now the standard apparatus used to test aerosol samplers at HSL and is similar to the apparatus used by manufacturers of dust monitoring equipment. The calm air test chamber comprised two steel sections (1m x 1m x 1m) stacked on top of the other. The sections were earthed to reduce any effects caused by charge on the aerosol. Typically, the volume flow rate was set to 220 l.min-1, which equated to an air velocity through the chamber of 0.4 cm.s-1. The samplers were therefore tested within calm air conditions (assumed to be air movements of < 10 cm.s-1). Although the temperature and relative humidity inside the chamber were not regulated, they were fairly constant between 21-23°C and 30-35%, respectively, throughout the tests. The dust was introduced into the chamber using the rotating brush generator model RBG 1000 manufactured by PALAS GmbH.

Figure 16 Calm air dust chamber

The main disadvantage of this method is that the resultant aerosol can have an extremely high charge produced by tribo-electrification effects; however, an ionizing fan on the chamber helped to reduce the charge on the aerosol. The main advantage is that it can produce a very constant and reproducible feed of dust. The aerosol is generated and mixed with the air in the top of the chamber. Honeycomb sections of foil between each section help create a lamina flow of air and aerosol towards the samplers in the bottom chamber. The samplers are rotated at the bottom to reduce the effects of any variability of the concentration of dust in the aerosol. The homogeneity

23

of the aerosol in the chamber, when tested gravimetrically, using six GK 2.69 cyclones at three different concentration levels, was calculated as ± 6%. Arizona road dust (ARD) supplied by Powder Technology Incorporated was selected for sampling tests because the powder conforms to the requirements of ISO 12103 Part 1 (1997), is commonly used as a test material for respirable samplers and contains about 70% crystalline silica.

The samplers were challenged to six different concentrations of ARD by maintaining all the parameters except the length of the sampling time of each sampling exercise.

5.2 GRAVIMETRIC ANALYSIS

The results for the gravimetric determination of the mass of aerosol collected by each sampler are shown in Figure 17, which compares the concentration of dust collected by each sampler with the concentration of dust collected by the reference sampler ((SIMPEDS).

Comparison of respirable samplers

16.00 PGP10 = 1.0073x GK2.69 IOM foam 14.00 R2 = 0.997

PGP10 PPI8 GK2.69 = 0.9219x 12.00 R2 = 0.9972 CIP(10) PPI = 0.853x 10.00 R2 = 0.9785

CIP10 = 0.6689x 8.00 R2 = 0.9984

6.00

4.00

Sampler concentration (mg m-3) IOM Foam = 0.9622x 2.00 R2 = 0.9622

0.00 0.00 5.00 10.00 15.00 Reference SIMPEDS Concentration (mg m-3)

Figure 17 Gravimetric comparison of respirable samples

The results shown on Figure 17 indicate that, when challenged to the ARD used in this exercise, the PGP 10 cyclone had the best relationship with the SIMPEDS and the CIP 10 sampler under estimated the gravimetric dust by about 35%. It is known that the CIP 10 sampler can under sample very small respirable dust sizes ( P Görner 2001). The trend lines for the modified GK 2.69 cyclone and the IPP impactor were possibly influenced by the underestimation at the highest air concentration. The reason for the underestimation with the GK 2.69 cyclone is attributed to the losses from the highest loaded filter material.

24

5.3 RCS ANALYSIS BY X-RAY DIFFRACTION

Figure 18 shows the results for RCS obtained using XRD and the calibrations developed in Appendix 3 and Appendix 4.

14

12 GK2.69 = 1.0657x GK IOM PPI CIP1 CIP2 R2 = 0.9957 IOM = 1.0042x

-3 R2 = 0.9599 10

PPI = 0.9329x R2 = 0.9656 8 CIP1 = 0.564x R2 = 0.9487 6

CIP2 = 0.5439x 4 R2 = 0.9414 Sampler Air Concentration mg.m

2

0 024681012 Reference SIMPEDS Sampler Air Concentration mg.m-3

Figure 18 Measurement of quartz by x-ray diffraction

The filters from the IPP impactor were ashed in a furnace brought to a temperature of 400°C for 4 hours. Those samples from the PGP 10 cyclone and CIP 10 sampler were ashed following the French standard method NF X 43-295 (AFNOR 1995) One SIMPEDS sample sustained slight damage on its surface as it was removed from the filter cassette in the sampler and was excluded from the evaluation.

The modified GK 2.69 cyclone and the IOM sampler with foam separator obtained the closest relationships with the reference sampler (SIMPEDS). The IOM sampler with foam separator obtained a slightly poorer correlation coefficient, probably due to one very low result The IPP impactor obtained a better slope than with the gravimetric tests, suggesting a better recovery when analysed by XRD. Its results closely matched the IOM sampler and SIMPEDS. The CIP 10 sampler under recorded the mass of quartz in the respirable sample by almost 45% on average compared with the SIMPEDS, when exposed to the fine ARD. This suggests that the CIP 10 sampler under sampled respirable quartz or that 10-15% of dust was lost during the analytical procedure. The results for the PGP 10 cyclone are not shown because the recovery of dust using the procedure at HSL was very poor and they were too variable.

25

5.4 RCS ANALYSIS BY DIRECT ON FILTER INFRARED ANALYSIS

The three GLA 5000 PVC filters were analysed using the direct on-filter FTIR analysis approach described in MDHS 101 (HSE 2005). Figure 19 shows the relationship obtained for the modified GK 2.69 cyclone and the IOM sampler with foam separator against the values obtained for the reference sampler (SIMPEDS). Only the absorbance at 780 cm-1 was used because the peak from the 800 cm-1 absorbance was very small and did not compare well with the profile of the calibration test samples.

4 GK2.69 IOM Heads 3.5 Linear (1:1 Relationship) 3

2.5

2

1.5

1

0.5 Sampler Air Concentration (mg.m-3) 0 01234 Reference Sampler (SIMPEDS) Air Concentration (mg.m-3)

Figure 19 Comparison of FTIR analysis results

The RCS concentration obtained using the 780 cm-1 absorbance is compared with the RCS concentration obtained from the average XRD result obtained from the 100, 101 and 112 reflections in Figure 20. The slope of the relationship line is not significantly different from 1 (95% confidence level is from 0.77 to 1.16) indicating no significant bias is observed between the XRD and the FTIR results.

26

Comparison All Data= 1.0093x 4 2

) R = 0.9512

-3 GK2.69 Results 3.5 IOM Heads Results SIMPEDS Results 3 Linear (1:1 Relationship)

2.5

2

1.5

1

0.5 Air Concentration Obtained by FTIR (mg.m 0 0 0.5 1 1.5 2 2.5 3 3.5 4 Air Concentration Obtained by XRD (mg.m-3)

Figure 20 Comparison of analysis techniques

5.5 RECOVERY FOR THE PGP 10 CYCLONE USING CELLULOSE NITRATE FILTERS

The results for the PGP 10 cyclones are not shown in Figure 18 because they were extremely variable due to the very poor recovery of some samples. The recommended filter for the PGP 10 cyclone causes recovery problems because it is fabricated from cellulose nitrate, which can ‘pop’ causing a loss of sample when placed in a furnace or in a low temperature asher. As no published procedure for the recovery of dust from cellulose nitrate filters for RCS analysis exists at the time of writing this report, the procedure described in the French method NF 243 (AFNOR 1995) for collapsing plastic foams from CIP 10 samplers was used. This procedure involves ignition of the air sample filter in isopropanol before placing the residue in a furnace or low temperature asher to complete the combustion process. The purpose of the ignition process is to melt and burn the plastic foam at a relatively low temperature. However, if burning of the filter is not completely effective any remaining filter material will ignite rapidly causing a loss of sample. Markus Mattenklott of the Institut für Arbeitsschutz der Deutschen Gesetzlichen Unfallversicherung (IFA), who has worked with the PGP 10 cyclone, was approached for assistance with this problem. The method employed at IFA uses 1,3 butanediol to burn the filter as the furnace is heated to the temperature required, removing the ash. This process is smoky and the furnace needs suitable ventilation. Initially, low recoveries were still obtained because some of the material was ‘sticking’ to the crucible. However, improved recoveries were obtained when the crucible with sample was placed in a beaker with isopropanol and put in an ultrasonic bath in isopropanol for about 5 minutes to disperse agglomerates rather than the usual practice of washing the contents of the crucible into the beaker before using ultrasound and filtering. After this treatment it was filtered in the usual manner, which is to wash the contents of the crucible into the beaker several times and then disperse the agglomerates. Figure 21 shows the gravimetrically determined recoveries following this adapted method. The slope

27

coefficient suggests a slight under recovery of about 2%, which suggests that the improved procedure is capable of giving satisfactory results for the PGP 10 cyclone in future tests.

10

8

Recovered = 0.9811x R2 = 0.9983

6

4 Linear (1:1

Mass Recovered (mg) Relationship)

2

0 0246810 Mass Loaded (mg)

Figure 21 Recovery for cellulose nitrate filters

28

6 STAGE 4: ASSESSMENT OF BIAS OF ANALYTICAL TECHNIQUES

6.1 WORKPLACE TASKS EXAMINED

The calm air chamber was again used to contain and sample dust generated from two simulated work tasks in experiments to assess the bias of the analytical techniques used in RCS measurement. The purpose of the exercise was to generate dust with two different particle size distributions in order to assess whether this influenced the performance of the analytical techniques. Two workplace tasks were simulated on two materials in the top of the calm air chamber and the samplers involved in this study were used to collect the aerosolised dust in the bottom chamber. The first tool used was a power chisel, but this caused health concerns (noise and vibration issues) and it was replaced with a hammer hand drill. The other tool used in the experiments was an angle grinder. Two tests were performed with the angle grinder, two tests with the hammer drill on the kerbstone, two tests were performed with the angle grinder and one test with the hammer drill on the sandstone. Figure 22 and Figure 23 show the types of tools and materials employed in these experiments.

Figure 22 Angle grinder and kerbstone in the top of the calm air chamber

Figure 23 Hammer Drill and kerbstone in the top of the calm air chamber

29

Except for the IPP impactor, two samplers of each type were placed in the sampling chamber. For those samplers with 25 mm diameter filters that could be used for the direct on-filter analysis procedure, one sampler contained a 5 µm pore size Gelman GLA 5000 PVC filter and the other contained a 0.8 µm pore size silver filter.

6.2 PARTICLE SIZE DISTRIBUTION OF THE GENERATED AEROSOL

The particle size distribution of the aerosol generated from each task and test material was measured by collecting a sample of aerosol onto an open-faced filter with a conductive cowl. The dust collected on the filter was then analysed for its particle size distribution using an AeroSizer and AeroDisperser (Amherst Process Instruments Inc, [API], MA, USA), which is an aerodynamic particle sizer incorporating a time-of flight aerosol beam spectrometer.

The results are shown in Table 6 and figures showing the distributions of the particles are shown in Appendix 9.

Table 6: Summary of mean particle sizes for dusts produced by different work activities and different materials Activity/Material Run No Aerodynamic particle size (µm) (Aerosizer) Number Volume Mean STDEV Mean STDEV Drilling Concrete 175 4.06 2.54 18.47 1.54 Cutting Concrete 187 3.94 2.22 14.69 1.61 Drilling Sandstone 179 4.17 2.37 20.77 1.68 Cutting Sandstone 183 4.51 2.78 29.29 1.79 Arizona Road Dust 193 4.51 2.56 26.74 1.58 Activity/Material Run No Geometric particle size (µm) (Aerosizer) Number Volume Mean STDEV Mean STDEV Drilling Concrete 175 2.64 2.57 12.06 1.53 Cutting Concrete 187 2.54 2.23 9.56 1.61 Drilling Sandstone 179 2.75 2.48 13.79 1.67 Cutting Sandstone 183 2.93 2.78 19.03 1.88 Arizona Road Dust 193 2.73 2.59 16.43 1.57

Unexpectedly, the simulated work tasks produced very similar bimodal distributions in terms of number of particles for each size fraction. The mean values for particle size in Table 6 are all probably located near the mid point of the bimodal distribution (in the dip). The study was not able to ascertain the particular size fraction of the respirable quartz particles in the concrete dusts, so it has to be assumed that they are evenly distributed over the whole range and that the two tools have similar mechanical actions that cause a bimodal distribution of particles.

6.3 GRAVIMETRIC ANALYSIS

6.3.1 Comparison of results to the SIMPEDS

The figures in Appendix 10 compare the gravimetric results obtained using each of the tested samplers with those obtained with the SIMPEDS reference sampler.

The mass of dust loaded onto each filter was greater than in the tests carried out in Stage 3 of the project. Gravimetrically, the PGP 10 cyclone, the IPP impactor and the GK 2.69 cyclone 30

were the samplers giving results closest to the average SIMPEDS value , whilst the IOM sampler with a foam separator and CIP 10 sampler gave results with a consistent negative bias.

6.3.2 Comparison of results obtained with silver filters and GLA 5000 PVC filters

Silver filters and GLA 5000 filters used in the same type of sampler were exposed to the same concentration of dust in a series of sampling exercises. Figure 24 compares the gravimetric results obtained on silver filters with those obtained on GLA 5000 PVC filters.

3.5

3

2.5

2 SIMPEDS

1.5 GK2.69 IOM Mass on Silver Filters (mg) 1 Linear (1:1 Relationship)

0.5 0.511.522.533.5 Mass on PVC Filters (mg)

Figure 24 Comparison of silver with PVC filters Caption The results obtained with the SIMPEDS exhibited an almost 1:1 relationship comparing use of the two filter types. Excluding one pair of results for which one of the filters was recorded as damaged, the slope for the SIMPEDS was 1.03 with a correlation coefficient of 0.956, suggesting a slight positive bias towards the silver filters. On the whole, the silver filters appeared to collect slightly more dust. . The results obtained for the GK 2.69 cyclone and the IOM sampler with foam separator were more variable

6.3.3 Comparison of results from weighing of filters and weighing of filters in plastic filter cassettes

Several of the samplers tested have an internal cassette that is intended to be sent to the laboratory without unloading the filter. Wall deposits form part of the sample and the filter and filter cassette are intended to be weighed together rather than the filter being removed from the cassette and weighed separately. However, due to the difficulty of obtaining reproducible results when weighing filter cassettes, filters are often weighed separately to improve accuracy. Figure 25 compares the results obtained for the GK 2.69 cyclone and SIMPEDS when weighing the filter and filter cassette together and when weighing the filter removed from the cassette. The results are highly scattered around the 1:1 relationship and this is a reflection of the difficulty of weighing filter and filter cassette together, something that is further illustrated by the fact that some negative values were obtained. It should also be noted that it would be 31

anticipated that the loss of sample reported in 6.3.4 would lead to higher results from weighing of the filter and filter cassette together. However, such a pattern cannot be discerned in Figure 25, suggesting that a larger proportion of the variability of the weighing is from the plastic cassettes rather than the visually observed losses.

4

3

2

1 (mg)

0 SIMPEDS PVC Filters 00.511.522.533.5SIMPEDS Silver Filters GK2.69 PVC Filters -1 GK2.69 Silver Filters Linear (1:1 Relationship) Mass from filter and cassette weighing

-2 Mass from filter weighing (mg)

Figure 265 Weighing of filters in plastic cassettes

6.3.4 Loss of sample in filter cassettes before XRD analysis There was visual evidence of sample loss onto the internal surfaces of the cassettes containing the GLA 5000 PVC filters. Photographs illustrating these losses and comparing them to the situation for cassettes containing silver filters are shown in Appendix 11. Some of the cassettes containing GLA 5000 PVC filters showed signs of dust having fallen into the space behind the filter and some showed dust on the o-ring that secures the filter in position in the cassette. There was no indication of similar losses occurring for the cassettes containing silver filters, but, occasionally, the edge of a silver filter was distorted, if it was put in the cassette slightly askew. This could theoretically cause a problem if the surface of the filter was uneven during analysis but it was possible to press the metal back into shape without affecting the filter deposit. There are two potential reasons for the loss of sample from the surface of the GLA 5000 PVC filters: 1. Loss of sample due to the filters flexing as they are removed from the cassettes for weighing. 2. Filters becoming charged during sampling, leading to increased deposition of aerosol on to the internal surfaces of the cassette. The problem of filter charge with GLA 5000 PVC filters is known and it has been recommended to soak the filters in a 1% solution of surfactant and dry before use (Blackford et al 1985). The action of the machine tools on the test materials could have charged the particles generated in this work and there was no mechanism employed to eliminate the static charge in the dust as it was generated. However, the tasks simulated in this exercise are typical workplace 32

processes and the experiments are valid because charge is not eliminated from the aerosol when sampling in the field.

6.4 X-RAY DIFFRACTION ANALYSIS

6.4.1 Comparison of RCS results from direct on-filter analysis of silver filters and GLA 5000 PVC filters by XRD

Figure 26 shows the relationship, for the SIMPEDS, GK 2.69 cyclone and IOM sampler, between the average RCS mass recorded on the silver filters and the average RCS mass recorded on the GLA 5000 PVC filters.

2 GK2.69 SIMPEDS 1.8 IOM Sampler 1.6 Linear (1:1 Relationship) 1.4 1.2 1 0.8 0.6 0.4 0.2 Mass Recorded on Silver Filters (mg) 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Mass Recorded on PVC Filters (mg)

Figure 27 Comparison of XRD results on silver filters and GLA 5000 PVC filters

Figure 26 shows that the majority of results on silver filters were higher than on GLA 5000 PVC filters. It was therefore decided not to include the results from the GLA 5000 PVC filters when comparing the performance of different samplers for the XRD determination of respirable quartz because the potential losses of dust from the GLA 5000 PVC filters could compromise any conclusions. Unfortunately, the variability of results from the GLA 5000 PVC filters also meant that it was not possible to assess differences in the performance of the FTIR and XRD techniques because silver filters cannot be used for FTIR analysis.

6.4.2 Comparison of results to the SIMPEDS

The figures in Appendix 12 compare XRD results for RCS obtained using each of the tested samplers with those obtained using the SIMPEDS reference sampler. All results are corrected for recovery and absorption, using the formula stipulated in the NIOSH 7500 (NIOSH 2003).

The XRD results compared to the SIMPEDS are not as good as the gravimetric comparison (6.3.1), which is probably due to dust losses occurring between the two analyses being

33

performed. The best regressions are for the PGP 10 cyclone (1.02) and the modified GK 2.69 cyclone (0.94).

Another approach to comparing the samplers was to examine the correlations between pairs of results. Pearson correlation values and their probabilities of no relationship existing, shown in Table 7, were obtained for each pair of samplers, excluding one pair of results for which the silver filter from the SIMPEDS was damaged. The smaller the probability the better the relationship between each set of paired results. A value of p = 0.01 represents a probability level of about 99% that a relationship exists.

Table 7: Pearson correlation coefficients of paired x-ray diffraction results

Sampler SIMPEDS PGP 10 GK 2.69 IOM IPP CIP 10

0.89 0.91 0.45 0.78 0.76 SIMPEDS (p=0.035) (p=0.022) (p=0.15) (p=0.024) (p=0.037)

0.98 0.20 0.81 0.87 PGP 10 (p=<0.01) (p=0.34) (p=0.01) (p=0.014)

0.32 0.79 0.86 GK 2.69 (p=0.24) (p=0.017) (p=0.015)

0.48 0.40 IOM (p=0.14) (p=0.21)

0.96 IPP (p=0.01)

Poor correlations (< 0.5) have been shaded and indicate the IOM sampler had a problem in these experiments. The best Pearson correlation coefficients are obtained between the PGP 10 cyclone and the GK 2.69 cyclone, which suggests these two samplers have a very similar performance. The CIP 10 sampler has reasonable correlations with PGP 10 cyclone, the GK 2.69 cyclone and the IPP impactor, which is a different conclusion when compared with the results from the gravimetric analysis.

Due to the uncertainty in the results from the SIMPEDS, the results from the XRD analysis of quartz were compared with those obtained by the PGP 10 cyclone using the indirect analysis procedure. Figures for these comparisons are shown in Appendix 13. These figures suggest that better agreement is obtained between the PGP 10 cyclone and the GK 2.69 cyclone, the CIP 10 sampler and the IPP impactor. The information in the figures also suggests that, for the relationship between the PGP 10 and the SIMPEDS, a single point in the results for the SIMPEDS could have affected the evaluation of results in Appendix 12. The figures in Appendix 13 also show an increase in the efficiency of the foam separators of the IOM samplers at high loadings, leading to a reduced mass of quartz measured by XRD.

6.4.3 Correction for absorption when using the direct on-filter procedure

Figure 27 shows relationship between the results obtained on the samples corrected for depth and absorption effects and the results uncorrected.

34

2.0

1.5

1.0 Linear (1:1 Relationship)

0.5

Corrected Result GK2.69 sampler (mg) 0.0 0.0 0.5 1.0 1.5 2.0 Uncorrected Result GK 2.69 sampler (mg)

Figure 28 Comparison of results with and without absorption correction

It can be seen that an absorption correction for the samples included in this study is not necessary and the result is unlikely to be significantly different from the corrected value, even though one of the samples had a gravimetric loading of about 3 mg.

35

7 FIELD TRIALS

7.1 APPROACH

The modified GK 2.69 cyclone, the IPP impactor and the PGP 10 cyclone were selected for field trails because of their similar performance. The purpose of the field trials was to assess their practical use and to validate the measurement and sampling methodology developed during the tests.

One visit was planned to a site from each of five key areas where exposure to RCS is common. The selected industry areas included in the study were, brick making, foundry works, ceramics manufacture, stone masonry (quarry work) and construction.

The number of available samplers and pumps limited the number of samples taken at each site. Three GK 2.69 cyclones with Libra Buck pumps, one IPP impactor, with a Leyland Legacy pump, and two PGP 10 cyclones, with Leyland legacy pumps, were used for each visit. The samplers were calibrated using TSI 4100 Series (4146) primary calibrator.

7.2 ANALYTICAL RESULTS

7.2.1 Overview

The results obtained in all site visits are summarised in Table 8. Generally, the mass of dust collected on the filters from the sites visited was low (from 0.1-3 mg). Even the samplers operating at flow rates of 10 l/min often collected less than 3 mg of dust. Of the fourteen personal samples taken, six peak exposures (30%) of RCS were above 0.05 mg.m-3 and three of these were (20%) were above 0.1 mg.m-3. Of the personal exposures reported, six had time- weighted averages (TWA) for the working day greater than 0.05 mg.m-3, of which five TWAs (30%) were greater than 0.1 mg.m-3. Although the field scientist took gravimetric dust measurements for respirable dust, these are not discussed because large negative values were obtained for the cellulose nitrate filters, probably because the temperature and humidity were not well controlled at the field laboratory. Cellulose nitrate filters are susceptible to changes in humidity and it should be noted that PVC filters can be used as an alternative in the PGP 10 cyclone.

7.2.2 Site 1: Foundry

The occupational hygienist's report is attached in Appendix Site 1 Foundry Visit. An XRD scan showed several major interferences with the primary quartz reflection of 101 (26.6 degrees 2θ) , so results obtained using the secondary 100 and tertiary 211 reflections were reported. The samples taken on this visit represent half-shift sampling in a foundry in which zircon Zr(SiO4) was the major component of the respirable dust in the environment. Two results obtained with the modified GK 2.69 cyclone recorded results greater than 0.05 mg.m-3.

In this field exercise, the occupational hygienist was unable to get the pumps assigned to the modified GK 2.69 cyclones to operate satisfactorily with a flow rate of 4 l/min. It was only possible to achieve flow rates of about 1.4 l/min or 1.7 l/min, which is well below the required rate. Furthermore, the deposit of dust on the filters for the modified GK 2.69 cyclone was not distributed as evenly as on the filters generated in the laboratory and this suggests that mistakes could have been made in preparing the equipment for sampling. This is attributed to the hygienist’s unfamiliarity with the apparatus. The results obtained with the modified GK 2.69 cyclone should be considered an over estimate because the sample deposit was concentrated in the middle of the filter and this distribution produces a higher signal than if the sample was 36

spread evenly over the whole sample. Zircon and another component of the dust, iron silicate oxide (Al2FeO2)SiO3, caused a large interference on the trace amounts of quartz present and therefore the less sensitive secondary and tertary quartz peaks were used for quantification. One of the advantages of an indirect analysis is the ability to include some sample treatment as part of the analytical procedure. Zircon could have been removed using phosphoric acid (Talivite 1951). However, this procedure was not used as it can also remove or etch the crystalline silica and is not effective on other common minerals (Miles 1999). The gravimetric results from the PGP 10 cyclone and the IPP impactor suggest that the percentage of quartz in the recovered respirable dust was about 2-3%. These results are consistent with each other and suggest that they are more reliable than those obtained with the modified GK 2.69 cyclone on this occasion. The PGP 10 cyclone gave air concentrations of about 0.026 mg.m-3, for a 4-hour sample containing 2-3% quartz.

7.2.3 Site 2: Brick manufacture

Static and personal samples were taken from the site of a major concrete and clay brick manufacturer. Clay chimney pots were manufactured at the site. A scan of one of the samples showed the presence of quartz and a mica (possibly muscovite); and possibly a trace of tri- calcium silicate in the airborne dust. The occupational hygienist’s report is attached in Appendix Site 2; Pot and Brick manufacture. The samples were taken for about 4 hours, representing a half-shift sampling exercise. Only one sample taken with the PGP 10 cyclone recorded a peak exposure greater than 0.05 mg.m-3 from a mass of 126 µg on the filter. The SIMPEDS operating at 2.2 l/min would have collected about 28 µg for the same air concentration, which is about 78% less than the PGP 10 cyclone.

Two of the three modified GK 2.69 cyclones failed to maintain a flow rate within ± 5% of the nominal value throughout the sampling period, as required by European standard EN 1232 (1997). The flow rate of one of the modified GK 2.69 cyclones dropped by as much as a half. Under these circumstances, the cyclones would no longer have been sampling to the respirable convention described in EN 481 (CEN 1993) and their results are not valid. The modified GK 2.69 cyclone, which did maintain a consistent flow rate, gave a similar percentage of quartz in the respirable dust as the PGP 10 cyclone.

According to the gravimetric analysis, the IPP impactor collected a large mass of dust (1.5 mg). However, no quartz was measured in this sample. The reason for this cannot readily be explained, although it is possible that the sample collected with the IPP impactor could have had a different composition to samples collected with the other samplers if it was not located in close proximity to them.

7.2.4 Site 3: Ceramics manufacture

A well-known manufacture of chinaware was visited. Analysis of one of the samples showed the presence of quartz (SiO2), aluminium silicate oxide Al2(SiO4)O, possibly a trace Feldspar. The occupational hygienist’s report is attached as Appendix Site 3: Ceramics manufacturer. The GK 2.69 cyclone samples were collected over a 4-hour sampling period representing a half shift sampling exercise. However, the PGP 10 cyclone and IPP impactor samples were only collected over a 2-hour sampling period, suggesting a task specific sampling exercise. Two of the six peak results reported were above 0.05 mg.m-3. One of these was a static with a GK 2.69 cyclone at a small hollowware glazing (SHG) unit with a result of 0.069 mg.m-3 from a mass of 55 µg on the filter. The other was a personal sample collected with a PGP 10 cyclone from the slip house team leader, gave a peak result of 0.15 mg.m-3 from a mass of 206 µg on the filter. The other personal samples included one collected with an IPP impactor on the operator of the small hollowware glazing unit flow line, which the did not record any detectable quartz, and another

37

collected on the operator of a ram press using a PGP 10 sampler, which recorded a result of 0.027 mg.m-3 from a mass of 44 µg measured on the filter. The other samples were statics taken with the modified GK 2.69 cyclone and did not record any significant RCS.

The flow rate drop for the GK 2.69 cyclone was a re-occurring problem. The flow rate of all three GK 2.69 cyclones dropped by more than ± 5% during this sampling exercise, which is contrary to the requirements in EN 1232. To some extent the samplers might self compensate by sampling more larger particles. However, a change of flow rate alters the sampling characteristics for this type of cyclone and can lead to inaccuracies in the measured values for RCS, since the response of XRD is dependent, to some extent, on the particle size of the dust.

Despite sampling almost 0.5 mg of respirable dust, quartz was not detected in the sample for the IPP impactor. This result is unusual, since it is a personal sample and a static GK 2.69 cyclone in close proximity to the IPP impactor, monitoring the same process, recorded a result of 0.069 mg.m-3.

The PGP 10 cyclones recorded 0.7 mg and 1 mg of dust recovered after ashing and gave results of 40 µg and 205 µg of RCS on the filter.

7.2.5 Site 4: Construction Site

Static and personal samples were taken on a construction site where the workers were scabbling concrete with a machine tool. This visit represents a task based sampling exercise because the maximum sampling time was 40 minutes. The work task took place in the open air, the weather was inclement and the concrete involved in the work was wet. Despite the wet weather all the samplers, operating in close proximity to each other, recorded a peak exposure to RCS above 0.05 mg.m-3, except the IPP impactor. A scan of one sample showed the presence of quartz, kaolinite, tricalcium silicate and calcite. A modified GK 2.69 cyclone and a PGP 10 cyclone was attached to a single worker primarily employed on the main scrabbling task. The GK 2.69 cyclone recorded a result of 0.15 mg.m-3 from a mass of 36 µg measured on the filter and the PGP 10 recorded a result of 0.12 mg.m-3 from a mass of 50 µg measured on the filter. It is well known that different the results obtained will depend on the relative position of the sampler to the dominant side of the worker and the spatial separation of the samplers. It is expected that the GK 2.69 cyclone would give the higher result because this sampler was positioned on the right hand side of the worker and the worker was right handed, whereas the PGP 10 cyclone was positioned more centrally. These two results are therefore comparable, and even recorded the same percentage of crystalline silica in the respirable dust, despite potential variability of the measurement result from the GK 2.69 cyclone due to the small mass of RCS measured as a result of the short sampling time. It is worth noting that a sampler using the current flow rate of 2.2 l/min would collect about 10 µg of RCS during the same sampling time, which is on the limit of detection for these measurements and would not allow an accurate quantification. All GK 2.69 cyclones maintained their flow rates within EN 1232 requirements.

7.2.6 Quarry/Stonemasons

A quarry was also visited where the cutting of sandstone under wet suppression took place. The sampling at this site represents a half shift sampling exercise because the sampling period was slightly less than four hours. An analysis of one sample showed the presence of quartz and clay; possibly kaolinite. The two GK 2.69 cyclones assigned to the workers in the workshops gave very high peak exposure results of 0.33 mg.m-3 from a measurement of 77 µg of RCS on the filter and 0.45 mg.m-3 from a measurement of 101 µg of RCS on the filter. All the other samplers used in a static position inside the cabins of vehicles or in the manufacturing area containing saws with wet suppression recorded results less than 0.02 mg.m-3 for RCS. The two GK 2.69 cyclones that obtained the high personal exposure results were not able to maintain 38

their flow within EN 2123 (1997) requirements. The hygienist was not able to get one of these samplers to start with the correct flow rate whilst the other reduced its flow rate by 9% over the sampling period.

7.3 PRACTICAL EXPERIENCE

The occupational hygienist made the following comments and proposed solutions.

7.3.1 PGP 10 cyclone and IPP impactor sampling kit (SKC Legacy Pumps)

The main advantage of the sampling train for the PGP 10 cyclones was that these pumps maintain their set flow over the selected period of sampling time. The battery bar shows a full battery after 4 hours of sampling and flow rate does not drop

The potential problems identified were:

1. The pump was too heavy and bulky for personal sampling and caused problems for the worker when moving and when fixing to the belt. 2. There was no attachment available on the PGP 10 cyclone unit to help attach it to a person. The cyclone does not easily stay in position unless sufficiently taped, due to its size and weight. 3. There was no attachment available on the IPP impactor to help attach it to a person. 4. The hygienist was unable to discover how to lock the keys on the pump to avoid it being tampering with.

The solutions offered by the hygienist were:

1. Harness / small rucksack is needed to make the apparatus easier to hold / wear during personal sampling. 2. An attachment / design a clip(s) for the cyclone, to enable the cyclone to be worn in breathing zone and stay in position.

7.3.2 GK 2.69GK 2.69 cyclone sampling kit

The main advantages of the sampling train for the modified GK 2.69 cyclone proposed by the occupational hygienist were that these samplers are a lot lighter and easier to wear than the PGP 10 pumps. They were also less easily tampered with or stopped.

The following potential problems were identified:

1. The pumps need to be charged for approximately 3 days prior to sampling in order to obtain a 4 l/min sampling rate. 2. The pumps loose charge (and therefore sampling rate) quickly. Pumps have no battery left after 3.5-4 hours of sampling. 3. Clips to attach to a person are present on sampling head; however, the sampling heads are heavy and do not stay in position unless taped.

7.4 COMMENTS RECORDED FROM THE WORKERS

The following is a list of comments from the workers using the apparatus. No comments were received for the GK 2.69 cyclones.

39

1. Too bulky – needs removing when manually unloading 2. Constant vibrating gets annoying 3. Sampler too heavy for gaffer tape 4. Needs clip on impactor cassette 5. The sampler is too heavy 6. The sampler puts a strain on your back and neck 7. The waist belt kept popping open 8. Pump is very heavy on one side 9. Piece near my neck keeps catching my face 10. Metal piece is sticking in my collar bone 11. My ear plugs (on a string) get caught near the strap/metal

40

Table 8: Summary of results from the site visits Site Dust recovered Quartz Recovered Sampling Sampling RCS in air Sampler Sampling from the filter (mg) (µg) dust (%) time (mins) volume (l) (mg.m-3) Foundry GK 2.69 Static 31.1 14 210 299.25 0.083 GK 2.69 Static 32.4 23 218 388.04 0.070 GK 2.69 Personal < 8 <0.8 225 477 < 0.021 IPP Personal 2.279 45.4 2 215 1716.78 0.026 PGP 10 Personal 1.939 57.4 3 227 2208.71 0.026 PGP 10 Static 0.833 24.6 3 221 2176.85 0.011 Brickmaking GK 2.69 Personal 19.93 2 192 566.4 0.035 GK 2.69 Static 12.29 9 185 738.15 0.017 GK 2.69 Static < 8.00 < 4 175 678.12 < 0.012 IPP Personal 1.474 < 8.00 < 0.5 105 841.05 < 0.010 PGP 10 Personal 1.039 125.72 12 165 1621.65 0.078 PGP 10 Personal 0.683 87.12 13 198 1920.6 0.045 Ceramics GK 2.69 Personal 55.44 8 210 802.2 0.069 manufacture GK 2.69 Static 14.91 2 225 883.1 0.017 GK 2.69 Static 14.84 4 243 945.7 0.016 IPP Personal 0.473 9.98 2 120 955.2 0.010 PGP 10 Personal 1.269 205.76 16 140 1407 0.146 PGP 10 Personal 0.515 39.75 8 145 1499.3 0.027

41

Table 8: Summary of results from the site visits Site Dust recovered Quartz Recovered Sampling Sampling RCS in air Sampler Sampling from the filter (mg) (µg) dust (%) time (mins) volume (l) (mg.m-3) Construction GK 2.69 Static < 8.00 < 3 30 117 0.068 GK 2.69 Personal 24.53 17 40 160 0.153 GK 2.69 Static 36.42 7 35 140 0.260 IPP Static 1.086 11.30 1 30 240 0.047 PGP 10 Personal 0.293 49.94 17 40 400 0.125 PGP 10 Static 0.37 48.35 13 40 402 0.120 Quarry GK 2.69 Static < 8.00 < 0.4 225 225 < 0.036 GK 2.69 Personal 77.13 27 235 235 0.328 GK 2.69 Personal 100.78 40 225 225 0.448 IPP Static 0.07 < 8.00 <11 210 1680 0.005 PGP 10 Static 0.243 17.87 7 215 2169.35 0.008 PGP 10 Personal/Static 0.185 48.35 26 235 2315.92 0.02

42

43

8 OVERALL PERFORMANCE OF SAMPLERS

8.1 MODIFIED GK 2.69 CYCLONE

In laboratory tests the performance of this cyclone compared well with the reference SIMPEDS and the PGP 10 cyclones. However, in the field tests the pump selected to operate with the modified GK 2.69 cyclone failed to maintain a flow rate of 4 l/min within ± 5%. According to the manufacturers literature, the pump is able to operate with a backpressure of about 55 inches of water for about 7 hours and the loadings obtained suggested that the pumps were failing at about 25–30 inches of water. The selection of a suitable pump can be resolved by selecting another that can operate with a greater backpressure, if one exists. Some additional work is needed to retest the sampling train under load and to discuss the issues with the manufacturer of the pump.

8.2 IOM SAMPLER WITH FOAM SEPARATOR

Whilst the IOM sampler with foam separator operating at 4 l/min operated well in the first laboratory test with ARD (slope = 1.0), this sampler under performed in the tests with simulated work activities (slope = 0.78). It is thought that as larger amounts of dust collect in the foam separator change the particle selection characteristics. i.e. the foam becomes more efficient at collecting particles because the dust collected in the foam reduces the size of the foam’s pores. Also, some dust was found to have dropped from the foam on top of one of the filters adding to the recorded mass of dust collected. These factors added to the variability of results for this sampler.

8.3 IPP IMPACTOR

This prototype sampler shows promise because it is compact and samples to a relatively high flow rate (8 litres/minute). The IPP impactor compared well with the reference SIMPEDS in the laboratory tests with the ARD (slope = 0.93) and gave a performance similar to the PGP 10 and modified GK 2.69 cyclones in the simulated work activity tests although had a tendency to under sample compared with the SIMPEDS (slope = 0.81 and 0.79). In the limited number of field tests the PPI samplers did not record the presence of any quartz when other samplers with lower flow rates detected significant amounts, even thought 1 or 2 mg of dust was collected on the filter after recovery, but this could be due to sampling issues rather than a failure of the sampler itself. The main factor in the field test was that the pump used with this compact sampler was relatively heavy. Figure 28 shows a dust deposit from a worksite sample where the dust seems to have built up, into a mound. The cleared spots in the centre of the mounds are a consistent feature of filters from this sampler and may indicate a secondary effect causing some particles to disperse again within the section for the air sample filter.

44

Figure 29 Deposit on a filter from an IPP impactor

8.4 PGP 10 CYCLONE

In the gravimetric tests the PGP 10 performed well against the SIMPEDS reference sampler with both the ARD (slope = 1.01) and the simulated work tasks (slope = 1.07). Once initial problems with the recovery of dust from the air sample filter for XRD analysis were overcome, the performance of this sampler also compared well with the modified GK 2.69 cyclone. The occupational hygienist and workers on site in the field tests identified a number of operational difficulties, such as the heavy pump and cyclone. These problems might be overcome with the use of a harness to hold the equipment and use of PVC filters instead of the cellulose nitrate filters recommended by the manufacturer. The analytical process with cellulose nitrate filters requires the burning of 1,3 butadiol, which is a very smoky process and potentially risky if the sample is placed in a preheated furnace rather brought up to temperature. However, PVC filters could be used, which burn cleanly, without the need for additional treatment with solvents. The main disadvantage of this apparatus is its cost. Purchase of one pump and sampler will cost an occupational hygienist or a company several thousands of pounds.

8.5 CIP 10 SAMPLER

This sampling instrument was shown to under sample quartz in the ARD in the first tests, by about 35 – 45%. When sampling the simulated work tasks this sampler also performed poorly against the SIMPEDS reference sampler in the gravimetric tests (slope = 0.72) but compared well with the PGP 10 (r2=0.87), modified GK 2.69 (r2 =0.86) and IPP impactor (r2=0.96) when the XRD quartz analysis results were evaluated. Work by Gömer et al (2001) indicates that this sampler under samples the smaller particle sizes of the respirable dust. One of the limiting factors of the work in this report is that we were unable to generate a dust with a majority of larger sized respirable particles (> 4µm). However, the generation of the size range of respirable particles is though to depend on the type of tool and the energy applied to the work task. Samplers must be able to perform in all types of environments because the sampler is likely to encounter a range of different work tasks in the same work place. As these experiments have demonstrated, the CIP 10 sampler does not have a comparable performance with the SIMPEDS when sampling some very common work tasks.

45

9 PRECISION OF XRD MEASUREMENTS AT 0.05 mg.m-3

A Higgins-Dewell cyclone operating a 2.2 l/min for a 4-hour sampling period, from an air concentration of 0.05 mg.m-3 collects 26 µg of RCS on a filter. Devices operating at 4 l/min collect 48 μg and those operating at 10 l/min collect 120 μg, of RCS on the air sample filter, assuming a full 4 hours was sampled. Increasing the flow rate collects more dust, but does not necessarily improve the measurement precision by the same proportion because the sensitivity of the calibration is dependent on how uniformly the dust is distributed across the filter. The different samplers have different deposition characteristics and the X-ray radiation is not uniformly detected over the whole area of a filter. This affects the sensitivity of the measurement. Figure 29 shows the signal detected from an alumina reference plate SRM 1976a that was masked with holes of different diameters punched from a silver filter. The y-axis is the intensity of the counts measured for a set time per unit area and the x-axis is the area of the exposed plate. When the full area of the plate is exposed (~23 mm) the counts detected per unit area have dropped by about half. The figure shows that most of the x-ray intensity from the sample is obtained from the centre of the filter.

60 50 0.04 solar slit 40 30 0.04 degree slits = -0.0568x + 47.486 20 R2 = 0.9419 10 Counts/Unit Area 0 0 100 200 300 400 500 Illuminated Area (mm2)

Figure 30 Response detected across the alumina reference plate

Figure 30 shows the sensitivities of the responses from the most intense XRD reflection for quartz at 26.67 degrees 2θ for the PGP 10 cyclone, the modified GK 2.69 cyclone and the SIMPEDS. If the intensity of the x-ray radiation were uniformly detected over the whole of the filter area the trend line for each sampler would coincide because they would have the same sensitivity. However, there are three trend lines of different sensitivities indicating differences in the deposition of the material on the filters.

46

250 GK 2.69 = 0.2828x PGP 10 = 0.4413x

200

SIMPEDS = 0.3139x 150

100 Intensity (cps) PGP 10 modified GK2.69 SIMPEDS 50

0 0 200 400 600 800 1000 Mass (µg)

Figure 31 Sensitivity of the 101-quartz reflection

The test samples from the PGP 10 obtain the most counts per second per unit mass (0.44) because it is a deposition method where the dust is recovered from the air sample and deposited onto another filter for analysis. The diameter of the filter funnel used to deposit the sample on the analysis filter was 15 mm (an effective area of 176 mm2). The sample from the modified GK 2.69 cyclone is spread more uniformly over more than twice the area (~ 415 mm2) of the dust deposited from the 15 mm diameter filter funnel. Subsequently, the response from the modified GK 2.69 cyclone, in terms of counts per second per unit area, is about a third less. This gives the PGP 10 cyclone an advantage in terms of measurement sensitivity, and therefore precision, so long as, the recovery of dust onto the measurement filter is efficient, the deposit of dust is not sufficient enough to require correction for absorption and the diameter of the deposit is about 15 mm. The calibration sensitivity of the modified GK 2.69 cyclone is slightly poorer than that obtained with the SIMPEDS cyclone. For example, a measurement with intensity of 15.7 counts/s will produce a result of 50 μg applying the calibration for the SIMPEDS cyclone and 55.5 μg applying the calibration for the modified GK 2.69 cyclone, suggesting that more sample is needed to achieve the same counts with the SIMPEDS cyclone. The use of silver filters, rather than GLA 5000 PVC filters, can improve the precision of measurement for the direct on-filter analysis technique because of their greater weight stability, if the sample is collected and weighed on the same day. The uncertainty of the measurement of test samples of quartz on GLA 5000 PVC filters and silver filters was determined as part of project JS20004271 Silica Method Development. Figure 31 and Figure 32 show the calibration uncertainty determined following the procedure detailed in ISO 24095 (2009) and demonstrates an improvement in precision when using silver filters when using the metal Higgins-Dewell cyclone. The calibration uncertainty in with silver filters is about half that when using GLA 5000 PVC filters between 25–50 µg.

47

70 60 100 Reflection (20.9 degrees) 50 101 Reflection (26.6 degrees) 40 112 Reflection (50.1 degrees) 30 20 10 Percentage Uncertainty 0 0 50 100 150 200 250 300 350 Mass (µg)

Figure 32 Uncertainty of calibration line on PVC GLA 5000 filters

70 60 50 100 reflection (20.9 degrees) 101 Reflection (26.6 degrees) 40 211 Reflection (50.1 degrees) 30 20

Percentage Uncertainty 10 0 0 50 100 150 200 250 300 350 Mass (µg)

Figure 33 Uncertainty of calibration line on silver filters

48

10 PROJECT SUCCESS CRITERIA

The two candidate samplers that met most of the project success requirements were the PGP 10 and the modified GK 2.69. The remaining problems associated with the introduction of these samplers to help lower the WEL for RCS are not insurmountable

The following success criteria were agreed at the start of the project.

• Can the nominal flow rate of the sampler be maintained over typical sampling periods (within ± 5%)?.

The PGP 10 fully meets this condition but the modified GK 2.69 does not with the pump currently selected. It should be possible to identify a pump from another manufacturer that will operate under a high backpressure and to make an appropriate recommendation or to work with the manufacturers to improve pump performance.

• Does the sampler have a performance comparable with the SIMPEDS respirable dust sampler recommended in the Health and Safety Executive’s method MDHS 14 for dust sampling (HSE 2000)

Both samplers fully met this criterion.

• Is the sampler comfortable to wear without interfering with the activity of the worker and is it applicable for use in UK workplaces?

Although, no worker refused to wear the PGP 10 they made a significant number of comments. No comments were made against the modified GK 2.69 cyclone

• Does the increased mass of dust collected with the sampler affect RCS measurement by XRD and FTIR?

High loadings do affect the XRD response but good corrections for absorption were obtained in samples with a trace amount of quartz, so long as interference did not occur. For FTIR, the relationship between absorbance and expected mass of quartz was linear over the range studied, but the sensitivity of this relationship is dependent on the matrix.

• Does the increased mass of dust collected with the sampler increase interferences?

Data shows that the reliability of a measurement of 1% quartz in a matrix similar to calcite will be subject to error (Figures 10 and 11), however, this problem is also encountered in the existing method and use of silver filters or an indirect analysis procedure may allow the opportunity to treat the sample to remove the calcite with acids. Experience with the minerals used in this project suggests that the problems with interference will be no worse than the current method.

49

• Would the precision of measurement at the proposed WEL of 0.05 mg.m-3 will equal of be better than the current analytical precision of ± 12% (2σ) at the current WEL of 0.1 mg.m3.

When analysing quartz dust the PGP 10 would collect about 125 µg of RCS for 4 -hour sampling periods from an air concentration of 0.05 mg.m-3. This mass of dust is greater than the 100 µg that would be collected by current samplers operating at 2 l/min from an air concentration of 0.1 mg.m-3 and will achieve the same or better measurement precision than ± 12% (2σ), so long as recovery of dust from the air filter does not result in loss of sample. Figure 21 shows that good recoveries from the cellulose nitrate filters used with the PGP 10 cyclone are obtained with care. The modified GK 2.69 should collect almost double the mass collected by a Higgins-Dewell SIMPEDS cyclone, however, the sensitivity of the measurement is slightly poorer. The precision of the measurement on samples from the modified GK 2.69 can be improved further by the use of silver filters as the air sample filter. However, this has not yet been completely verified.

50

Table 9 summarises the advantages and disadvantages of the PGP 10 and modified GK samplers when compared against the success criteria and other factors. Table 9: Summary o f PGP 10 and modified GK 2.69 cyclone performance

Sampler PGP 10 Modified GK 2.69

Flow rate Passed Passed in the laboratory but failed in specification the field tests. May have to use the pump for the PGP 10.

Sampling Both samplers have a performance comparable with the SIMPEDS reference Characteristics sampler in these tests.

Analysis Longer preparation procedure. The Short analysis procedure procedure laboratory will need additional equipment and there is a risk of losing Sample can be recovered from silver sample if not done correctly. filter if necessary for treatment.

XRD Passed May pass if silver filters are used Precision

Correction for Passed Corrections are possible if the dust is matrix effects sampled on a silver filter. For Accurate corrections can be obtained kaolinite, and common carbonaceous in matrices with trace percentages of interferences such as coal dust, the RCS. analyst may be able to ash the dust on the filter without disturbing its surface.

Impediment to Causes discomfort to the worker, Passed worker although this may be resolved with a activity design of harness.

Cost Very expensive cyclone and pump > Approximately £1000 for new £2000 cyclone and pump

Analysis costs Approximately 3 times the cost of a Same as current costs direct on filter analysis.

51

11 REFERENCES

AFNOR (1995b) Norme NF X 43-295 De´termination par rayons X de la concentration de de´poˆt alve´olaire de silice cristalline—Echantillonnage par dispositif a` coupelle rotative. 1er tirage 95-06 AFNOR 1995.

Blackford (1985) The Reduction Of Dust Losses Within The Cassette Of The SIMPEDS Personal Dust Sampler D Blackford, G Harris and G. Revell Ann. occup. Hyg., Vol. 29, No. 2, pp. 169-180, 1985

(BS 1997) BS EN 1232 Workplace atmospheres – Pumps for personal sampling of chemical agents – Requirements and test methods, European Committee for (CEN), Brussels 1997, ISBN 0580 283283

CEN. (1993) Comite Europe´en de Normalisation (CEN) Workplace atmospheres: size fraction definitions for measurement of airborne particles in the workplace. CEN standard EN 481. Brussels 1993.

Görner P (2001), Study of fifteen respirable aerosol samplers used in occupational hygiene, Görner P, Wrobel R, Miĉka V, Skoda V, Denis J and Fabriès J-F Ann occup Hyg, Vol 45, No 1 pp 43-54, 2001

HSE (2000) Health and Safety Executive (HSE): MDHS 14/3 Methods for the determination of hazardous substances, General methods for sampling and gravimetric analysis of respirable and inhalable dust. HSE Books, Sudbury, Sufflok, 2000, ISBN 07176 –1749-1

HSE (2002) Health and Safety Executive (HSE): Respirable crystalline silica: Phase 1, Variability in fibrogenic potency and exposure-response relationships for silicosis; Assessment Document, Environmental Guidance Note EN 75/4. Sudbury, UK: HSE Books, 2002; ISBN 07176 23742.

HSE (2005) Health and Safety Executive (HSE): MDHS 101 Methods for the determination of hazardous substances, respirable crystalline silica in air, direct on filter analyses by infrared spectroscopy and x-ray diffraction, HSE Books, Sudbury, Sufflok, 2005, ISBN 07176 –2897 – 3

HSC (2006) Health and Safety Commission (HSC) Advisory Committee on Toxic Substances (ACTS) Measuring low masses of respirable crystalline silica (RCS) at proposed workplace limits. A Paper by: Peter Stacey, HSL 20/2006 http://www.hse.gov.uk/aboutus/meetings/iacs/acts/301106/acts202006.pdf (January 2010)

ISO (1995) International Organisation for Standardisation (ISO) 7708 Particle size definitions for health related sampling, ISO Geneva, Switzerland, 1995

ISO (1997) International Organisation for Standardisation (ISO) 12103-1 Road vehicles -- Test dust for filter evaluation -- Part 1: Arizona test dust, ISO Geneva, Switzerland, 1997

ISO (2009) International Organisation for Standardisation (ISO) 24095 Workplace air- Guidance for the analysis of respirable crystalline silica, ISO Geneva, Switzerland, 2009

Kenny LC, Gusmann RA. (1996) Characterization and modeling of a family of cyclone aerosol preseparators. J Aerosol Sci; 28: 677–88. 52

Kenny LC, Aitken RJ, Baldwin PEJ et al. (1999) The sampling efficiency of personal inhalable aerosol samplers in low air movement environments. J Aerosol Sci; 30 627–38

Lee K and Ramamurthi M (1993) Aerosol measurement, principles, Techniques and Application, Filter Collection, eds P Baron and K Willeke, pp 196 – 197, Chapman and Hall, London 1993 ISBN 0-442-00486-9

Miles (1999), Issues and controversy, The measurement of crystalline silica, review papers on analytical methods, AIHAJ, Vol 60, pp 396 – 402 (1999)

NIOSH (2004) National Institute for Occupational Safety and Health (NIOSH): Health Effects of Occupational Exposure to Resiprable Crystalline Silica: NIOSH Hazard Review. Cincinnati, OH, USA: NIOSH, 2002; Publ. No. 2002–129

NIOSH (2003) Method 7500, Silica, Crystalline by XRD, Manual of Analytical Methods 4th Edition, National Institute of Occupational Safety and Health, Cincinnati, OH, United States, 2003 Issue 3

Talavite (1951). Determination of quartz in the presence of silicates using phosphoric acid. Anal Chem (23) 623-626. 1951

53

12 APPENDIX 1: PRESSURE DROP WITH FLOW RATE ACROSS AN MIXED CELLULOSE ESTER FILTER

Filter pressure drop (0.8 µm MCE filter)

16.00

14.00 y = 0.0057x R2 = 0.9937 12.00

10.00

8.00

6.00

4.00 Pressure drop (inches water) drop (inches Pressure 2.00

0.00 0 500 1000 1500 2000 2500 3000 Flow rate cm3 min-1)

Filter pressure drop (1.2 µm MCE filter)

14.00

y = 0.0041x 12.00 R2 = 0.9984

10.00

8.00

6.00

4.00 Pressure drop (Inches water) drop (Inches Pressure 2.00

0.00 0 500 1000 1500 2000 2500 3000 3500 Flow rate (cm3 min-1)

54

Filter pressure drop (3µm MCE filter)

9.00

8.00 y = 0.0024x R2 = 0.9974 7.00

6.00

5.00

4.00

3.00

Pressure drop (Inches water) drop (Inches Pressure 2.00

1.00

0.00 0 500 1000 1500 2000 2500 3000 3500 4000 Flow rate (cm3 min-1)

55

13 APPENDIX 2: INSTRUMENTAL PARAMETERS

13.1 X-RAY DIFFRACTION

The samples were analysed by XRD on a Panalytical Xpert Pro MPD using the following specifications optimised for intensity rather than resolution.

• Bragg-Brento semi-focusing geometry

• A broad focus tube with copper target operating at 55Kv and 49 mA to give a power output of 2.7 watts

• Automatic divergence and antiscatter slits set at 18mm

• Spinner set at 1 revolution per second

• Automatic sample changer

• Array detector set on continuous scan mode with a window area of 2.12 degrees.

Table 3: Scan parameters

Angle (2θ) Scan range Step size Counts per step (seconds)

20.9 19.9 – 21.9 0.05 600

26.6 25.65 – 27.65 0.05 420

50.1 49.1 – 51.1 0.05 600

13.2 FTIR

The following parameters were established on a Perkin Elmer FTIR 2000 System.

• Resolution = 8.00 cm-1 J stop resolution = 3.7 cm-1

• Apodication = strong

• Gain = 1

• OPD velocity = 1 cm/s

• Interferogram = Bi directional double sided

• Phase correction = self 256

• Number of scans = 32

• Range = 400 – 1200 cm-1 Interval = 1.0 cm-1

56

14 APPENDIX 3: CALIBRATIONS FOR X-RAY DIFFRACTION

14.1 SIMPEDS CALIBRATIONS

PVC Qtz 50 = 0.0265x + 0.0001 R2 = 0.9969 1 Ag Qtz50 = 0.0266x + 0.0194 R2 = 0.9993 0.9 0.8 Ag Qtz 21 = 0.0174x + 0.0066 0.7 R2 = 0.9991 0.6 PVC Qtz 21 = 0.0181x + 0.0296 R2 = 0.9983 0.5 Ag Qtz21 Ag Qtz 26 = 0.0032x + 0.0067 Ag Qtz26 0.4 2 R = 0.9996 Ag Qtz50 Mass (mg) 0.3 PVC Qtz 26 = 0.0034x + 0.0059 PVC Qtz21 0.2 R2 = 0.9995 PVC Qtz26 0.1 PVC Qtz 50 0 0 50 100 150 200 250 300 Intensity (cps)

14.2 GK 2.69 CYCLONE CALIBRATIONS

Ag Qtz 50 = 0.031x + 0.0406 R2 = 0.9993

3.5 PVC Qtz 50 = 0.0303x + 0.016 2 3 R = 0.9966 2.5 Ag Qtz 21= 0.0233x + 0.0024 GLA Qtz 26 2 R = 0.998 Ag Qtz 26 2 PVC Qtz 21 = 0.022x + 0.0642 GLA Qtz 21 R2 = 0.9911 1.5 Ag Qtz 26 = 0.0042x - 0.0228 Ag Qtz 21 2 Mass (mg) 1 R = 0.9993 GLA Qtz 50 PVC Qtz 26 = 0.004x + 0.034 Ag Qtz 50 2 0.5 R = 0.9941 0 0 200 400 600 800 1000 Intensity (cps)

57

14.3 IOM SAMPLER WIH FOAM SEPARATOR

Qtz 50 PVC = 0.0293x + 0.0016 R2 = 0.9986 3 Qtz 50 Ag = 0.0311x + 0.0502 R2 = 0.9913 2.5 Qtz 21PVC = 0.0228x - 0.0419 R2 = 0.9983 Qtz 26 PVC 2 Qtz 21 Ag = 0.0232x + 0.0199 Ag Qtz 26 2 R = 0.9875 Qtz 21 PVC 1.5 Qtz21 Ag Qtz50 PVC Mass (mg) 1 Qtz 26 PVC = 0.004x - 0.0126 Qtz50 Ag R2 = 0.996 0.5 Qtz 26 Ag = 0.004x + 0.0718 R2 = 0.9826 0 0 100 200 300 400 500 600 Intensity (cps)

58

15 APPENDIX 4: CALIBRATION FOR THE INDIRECT ANALYSIS PROCEDURE

Uncorrected for depth effects 9 PVC Qtz 50 = 0.013x1.1324 8 R2 = 0.9982 Ag Qtz 50 = 0.0176x1.0752 7 2 R = 0.9993 PVC Qtz 26 6 Ag Qtz 26 PVC Qtz 21 = 0.0068x1.2333 PVC Qtz 21 5 2 R = 0.991 Ag Qtz 50 1.2161 4 Ag Qtz 21 = 0.0074x Ag Qtz 21 R2 = 0.9899 1.1811 PVC Qtz 50 Mass (mg) 3 PVC Qtz 26 = 0.0011x R2 = 0.9967 2 Ag Qtz 26 = 0.0008x1.2263 1 R2 = 0.994 0 0 500 1000 1500 2000 Intensity (cps)

Deposition Calibration on Gelman and Silver Filters Corrected for Absorption PVC Qtz 50 = 0.0216x + 0.0007 9 R2 = 0.9984 8 Ag Qtz 50 = 0.0223x + 0.0827 R2 = 0.9964 7 Qtz 26 PVC PVC Qtz 21 = 0.0172x - 0.1115 Qtz 26 Ag 6 R2 = 0.9869 Qtz 21 PVC 5 Qtz 21 Ag Ag Qtz 21 = 0.0169x + 0.1001 Qtz 50 PVC 4 R2 = 0.9794 Qtz 50 Ag Mass (mg) 3 Ag Qtz 26 = 0.0031x - 0.0252 R2 = 0.9875 2 1 PVC Qtz 26 = 0.003x - 0.0818 R2 = 0.9967 0 0 500 1000 1500 2000 2500 3000 Intensity (cps)

59

16 APPENDIX 5: CALIBRATIONS FOR INFRARED – DIRECT ON-FILTER ANALYSIS

16.1 SIMPEDS

1 0.9 780 cm -1= 1.7216x 0.8 R2 = 0.996 0.7 800 cm-1 = 1.4726x R2 = 0.9949 0.6 0.5 800 cm-1 0.4

Mass (mg) 780 cm-1 0.3 0.2 0.1 0 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Absorbance (cm-1)

16.2 GK 2.69 CYCLONE

2.5 780 cm-1 = 2.9173x R2 = 0.9994 2 800 cm-1 = 2.1947x + 0.0483 R2 = 0.999 1.5 800 cm-1 Abs 780 cm-1 Abs 1 Mass (mg)

0.5

0 0 0.2 0.4 0.6 0.8 1 1.2 Absorbance (cm-1)

60

16.3 IOM SAMPLER WITH FOAM SEPARATOR

2.5

2 780 cm-1 = 3.1674x - 0.0352 R2 = 0.9942 1.5 800 cm-1 = 2.5047x - 0.0452 Abs 800 R2 = 0.9923 Abs 780

Mass (mg) 1

0.5

0 0 0.2 0.4 0.6 0.8 Absorbance (cm-1)

61

17 APPENDIX 6: CALIBRATIONS FOR INFRARED – INDIRECT ANALYSIS

Deposition Methods

-1 0.8003 780 cm-1 = 1.984x1.0057 800 cm = 1.8435x 9 2 R2 = 0.9908 R = 0.9661 8 7 6 5

4 800 cm-1 3

Mass (mg) 780 cm-1 2 1 0 01234567 Absorbance (cm-1)

62

63

18 APPENDIX 7: XRD SCANS OF ABSORPTION TEST MATERIALS

The amorphous peak at about 45 and 50 degrees is cellulose from a material used to support the sample during analysis

Pattern List

Visible Ref.Code Score Compound Name Displ.[°2Th] Scale Fac. Chem. Formula * 00-019-0770 53 Talc-2\ITM\RG -0.053 0.595 Mg3 Si4 O10 ( O H )2

64

Pattern List

Visible Ref.Code Score Compound Name Displ.[°2Th] Scale Fac. Chem. Formula * 01-086-2334 97 Calcite 0.013 0.922 Ca (C O3)

65

Pattern List

Visible Ref.Code Score Compound Name Displ.[°2Th] Scale Fac. Chem. Formula * 01-079-1193 90 Olivine 0.084 0.951 Mg1.8 Fe.2 ( Si O4 )

66

19 APPENDIX 8: INFRA RED ABSORNACES

Figure 34 IR Difference scan, containing 60 µg of A9950 quartz

Figure 35 IR Difference scan, containing 74 µg of A9950 quartz in 3 mg of Talc

67

Figure 36 IR Difference scan, containing 61 µg of A9950 quartz in 2.8 mg of calcite

Figure 37 IR Difference scan, containing 65 µg of A9950 quartz in 2.8 mg of olivine

68

20 APPENDIX 9: PARTICLE SIZE DISTRIBUTIONS FROM SIMULATED WORK TASKS

Arizona Road Dust

Volume distribution (Aerodynamic) Number distribution (Aerodynamic)

0.0007 2.5

0.0006 2 0.0005

0.0004 1.5

0.0003 1

0.0002 % of total number of total % 0.5 0.0001

0 0 0.1 1 10 100 1000 0.1 1 10 100 1000 Aerodynamic diameter (µm) Aerodynamic diameter (µm)

Concrete dust generated with an angle grinder

Volume distribution (Aerodynamic) Number distribution (Aerodynamic)

6 0.7

5 0.6

0.5 4 0.4 3 0.3 2 0.2 % of total volume of total % % of total number of total % 1 0.1

0 0 0.1 1 10 100 1000 0.1 1 10 100 1000 Aerodynamic diameter (µm) Aerodynamic diameter (µm)

Sandstone dust generated with an angle grinder

Volume distribution (Aerodynamic) Number distribution (Aerodynamic)

2 6 1.8

5 1.6 1.4 4 1.2

3 1 0.8 2 0.6 % of total volume % of total number of total % 1 0.4 0.2 0 0.1 1 10 100 1000 0 0.1 1 10 100 1000 Aerodynamic diameter (µm) Aerodynamic diameter (µm)

69

Concrete dust generated with a hammer drill

Volume distribution (Aerodynamic) Number distribution (Aerodynamic)

8 2.5

7 2 6

5 1.5

4 1 3 % of total number of total % 2 0.5

1 0 0 0.1 1 10 100 1000 0.1 1 10 100 1000 Aerodynamic diameter (µm) Aerodynamic diameter (µm)

Sandstone dust generated with a hammer drill

Volume distribution (Aerodynamic) Number distribution (Aerodynamic)

6 2.5

5 2

4 1.5 3

1 2 % of total volume % of total number of total % 1 0.5

0 0 0.1 1 10 100 1000 0.1 1 10 100 1000 Aerodynamic diameter (µm) Aerodynamic diameter (µm)

70

21 APPENDIX 10: GRAVIMETRIC ANALYSIS WITH SIMULATED WORK ACTIVITIES

GK2.69 versus SIMPEDS PGP10 versus SIMPEDS

40 35

35 y = 1.0735x

30 ) ) 2 -3

-3 y = 0.8953x R = 0.988 30 R2 = 0.9882 25

25

20 20

15 15

10 10 PGP10 Conc (mg m Conc PGP10

GK2.69 Conc (mg m GK2.69 5 5

0 0 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40 -3 Reference SIMPEDS Conc (mg m-3) Reference SIMPEDS Conc (mg m )

IOM foam versus SIMPEDS CIP10 versus SIMPEDS

35 30

) 30 -3 25 )

y = 0.776x -3 y = 0.7224x 25 2 2 R = 0.7212 20 R = 0.9777

20 15

15

10 10

CIP10 Conc (mg m 5 5 IOM foam Conc (mg m

0 0 0 5 10 15 20 25 30 35 40 0 5 10 15 20 25 30 35 40 Reference SIMPEDS Conc (mg m-3) Reference SIMPEDS Conc (mg m-3)

PPI8 versus SIMPEDS

35

30 y = 0.8439x ) 2 -3 R = 0.8998 25

20

15

10 PPI8 Conc (mg m 5

0 0 5 10 15 20 25 30 35 40 Reference SIMPEDS Conc (mg m-3)

71

22 APPENDIX 11. DUST LOSSES FROM PVC FILTERS

DUST

Figure 38 Wall depositions inside sample cassettes containing a PVC filter .

Back of cassette

Filter Support

Figure 39 Dust on the back of the cassette containing a PVC filter

72

Loose Dust

Figure 40 The PVC filter and cassette showing loose dust on the filter and dust losses on the cassette

Figure 41 Cassette lid with a silver filter containing a similar mass of dust showing no loose dust or dust losses in the cassette

73

Crimpled filter edge

Figure 42 Cassette lid contained silver filter with a crimpled edge caused by the cassette

74

23 APPENDIX 12: STAGE 4 XRD COMPARISON WITH SIMPEDS

18 Linear (1:1 16 Relationship) 14 12 y = 1.0263x 10 8 6 4 5 7 9 11 13 15 Air Concentration PGP 10 mg.m-3 PGP Concentration Air Reference SIMPEDS Air Concentration mg.m-3

18 Linear (1:1 16 Relationship) 14 12 y = 0.9557x 10 8 6 4 5 7 9 11 13 15 Air Concentration GK2.69 mg.m-3 GK2.69 Concentration Air Reference SIMPEDS Air Concentration mg.m-3

18 Linear (1:1 16 Relationship) 14 12 y = 0.9399x 10 mg.m-3 8 6 4 Air Concentration IOM Heads IOM Concentration Air 5 7 9 111315

Reference SIMPEDS Air Concentration mg.m-3

75

18 16 Linear (1:1 Relationship) 14 12 10 8 y = 0.8367x 6 4

Air Concentration CIP 10 mg.m-3 CIP Concentration Air 5 7 9 111315 Reference SIMPEDS Air Concentration mg.m-3

18 Linear (1:1 16 relationship) 14 12 10 mg.m-3 8 y = 0.7816x 6 4 Air Concentration IPP Impactor Impactor IPP Concentration Air 579111315 Reference SiMPEDS air Concentration mg.m-3

76

24 APPENDIX 13 STAGE 4: COMPARISON WITH PGP 10 CYCLONES

16 14 12 y = 0.9178x 10 8 6 Linear (1:1 Relationship)

sampler mg.m-3 sampler 4 Linear 2 (GK2.69) Air Concentration GK2.69 GK2.69 Concentration Air 0 0 5 10 15 20 Air Concentration PGP 10 Sampler

16 IOM Heads

14 Linear (1:1 12 Relationship)

10

mg.m-3 8

6

4 Air Concentration IOM Heads IOM Concentration Air 0 5 10 15 20 Air Concentration PGP 10 Sampler mg.m-3

16 CIP 10 14 Linear (1:1 12 Relationship) 10 Linear (CIP y = 0.8072x 8 10) 6 4 2 0 Air Concentration CIP 10 mg.m-3 CIP Concentration Air 0 5 10 15 20 Air Concentration PGP 10 mg.m-3

77

16 Linear (1:1 Relationship) 14 Linear (IPP 12 Impactor) y = 0.7649x 10 8 6 4

Sampler mg.m-3 2 0

Air Concentration IPP Impactor Impactor IPP Concentration Air 0 5 10 15 20 Air Concentration PGP 10 Sampler mg.m-3

16 14 Linear (1:1 Relationship) 12 10 8 6

Sampler mg.m-3 4 2 Air Concentration Simpeds Simpeds Concentration Air 0 0 5 10 15 20 Air Concentration PGP 10 Sampler mg.m-3

78

25 SITE 1: FOUNDRY VISIT

Harpur Hill, Buxton Derbyshire, SK17 9JN T: +44 (0)1298 218000 F: +44 (0)1298 218590 W: www.hsl.gov.uk

HSL Occupational Hygiene Section Site Visit Report

High Volume Silica Monitoring Project

Report Number OH/2009/LET/65 Job Number JN0003804 / A1050 Report author Helen Ashpool Company name & address Site 1 Foundry Persons seen Bill Stevens, Health and Safety Manager Thomas Riley, Health and Safety Officer Ian Thompson, Health and Safety Officer Date of visit 8th September 2009 HSL/HSE staff present Helen Ashpool, HSL Tariq Khan, HSE Luke Messenger, HSE Date of report January 2010 Report distribution Peter Stacey, HSL Tariq Khan, HSE FOD Luke Messenger, HSE FOD Harvey Wild, HSE FOD Phil Smith, HSE Molten Metals & Minerals Section Manufacturing Sector Tim Roff, Head of HSL OHU NEDB OHU Archive Report approval V Sandys

79

Introduction

1. This report describes a site visit on 8th September 2009. The purpose of the visit was to trial new monitoring equipment in the field for the sampling of respirable crystalline silica (RCS). HSE/HSL personnel on site were Tariq Khan, HM Inspector of Health and Safety, Luke Messenger, HM Inspector of Health and Safety, and Helen Ashpool, HSL. Representing the company were Bill Stevens, Health and Safety Manager, Thomas Riley, Health and Safety Officer and Ian Thompson, Health and Safety Officer.

2. The foundry undertakes “Lost Wax Investment Casting” to make various shaped and sized vanes and blades for the automotive and turbine industries. In this type of process, the wax is lost during firing in an autoclave at the end of the process leaving behind the shell.

3. Personal and static location samples were taken in the shell shop on the production line and at the HAZMAC finishing booth. The shell shop manufactured the shells for the moulds for the products, which would eventually be made up in metal. The production line used various siliceous products including at the rain sander (“flour shower”) and the HAZMAC booth, which undertook the finishing off (fettling) of items after firing. Operatives in both areas could potentially be exposed to RCS.

4. The health hazard associated with the inhalation of RCS is silicosis. This is a serious chronic disease where scar tissue is formed in the lungs resulting in a reduction in lung capacity. RCS is also a suspected human carcinogen.

5. RCS is currently assigned a workplace exposure limit (WEL) of 0.1 mg/m3; however, part of the purpose of the overall project was to find out whether the WEL could be lowered to 0.05 mg/m3.

Process description

6. The shell shop manufactured the shells for the moulds for vane and blade products using a “Lost Wax Investment Casting Process”. The wax would eventually be lost during firing in autoclave at the end of the process leaving behind the shell, which was coated onto the wax.

7. Wax moulds were manufactured in an adjacent area and then transferred into the shell shop for coating.

8. There were two production lines in the shell shop; however, only one line was running during the monitoring period due to work requirements.

9. The production line was operated by four personnel per shift and was kept running 24 hours per day.

10. One operative was involved with the mixing of substances as well as operating the production line. The substances mixed all contained a siliceous product according to the material safety data sheets (MSDS) supplied. Substances included a Collodial silica liquid product (REMASOL SP-25) containing 25% amorphous silica; a Zircosil filler, containing 95-100% zirconium silicate; and a Mulcoa filler (Mulcoa 60) containing 77% bauxitic kaolin clay, <23% amorphous silica and 0.3% cristobalite. The operative reported to undertake mixing approximately three times per shift.

11. Mixing was undertaken on a mezzanine level adjacent to and above the rain sander. The mixer was enclosed except for an opening in the top at which a manoeuvrable local 80

exhaust ventilation (LEV) arm was aligned (see photo 1). The bag tipping position was not served with LEV. The operative wore an air fed visor whilst undertaking this task (see photo 2). The area (surfaces and floor) was covered with significant deposits.

12. The wax items (hung on a “hanger”) were coated by being transported from the main carousel line by a robot arm to the mixing tanks where they were dipped (see photo 4) prior to being coated in the rain sander (see photo 5). The rain shower also utilised Zircosil 200M as described in paragraph 10 above. The rain sander area (surfaces and floor) was observed to be coated with significant deposits; however, the area could only be entered by personnel for maintenance operations. Operations were undertaken remotely from a booth overlooking the area (see photo 3).

13. After coating (some items were required to need more than one coating and may be in the shell shop for 2.5 days), the items were placed back on the main carousel by the robot arm, which were then transported to the HAZMAC booth for finishing operations (see photo 6). Coated items could potentially be on the conveyor drying for 4 hours. The items were dried whilst moving along the carousel by tall dryers, which rotated in order to dry each side of the item (see photo 7).

14. One operative worked at the HAZMAC booth using various tools to undertake finishing / fettling operations (see photo 8). Tools used included a pneumatic handheld grinding wheel (see photo 9), which was used to chip off excess coating material. It was reported that between 3 and 15 minutes could be spent on each hanger in the booth. This was dependent on the product.

15. A compressed airline was used to blow off small areas of coating deposits during inspection of items on the carousel. It was reported that the pressure of a vacuum would “suck off” or disturb the layers and / or the wax.

81

Exposure controls

16. Exposure controls were not examined in detail as this was outside the scope of the project; however, some details are noted below.

17. The production line was operated from a partially enclosed booth opposite the rain shower. The control panel area was surrounded by clear plastic screens, which helped protect the operative from airborne particulates (see photo 3).

18. The HAZMAC booth was served with local exhaust ventilation (LEV); however, it was observed to blow back particulate into the booth which had previously been extracted, when it was switched off. The dust deposits may have been deposited in the ductwork before the DCE (filter / fan) unit due to low transport velocities.

19. Operatives wore 3M 800 series air fed visors with 3M Jupiter battery packs during the mixing process (bag tipping) and in the HAZMAC booth. Disposable respirators were reported to have been previously used; however, after trialling the visors, all operatives opted to use the visors. Respiratory protective equipment (RPE) was not worn during any other operation.

Measurement strategy.

20. Air sampling was undertaken in accordance with methodology described in MDHS101 Crystalline Silica in Respirable Airborne Dusts. Samples were collected at varying flow rates dependent on pump type as described by Peter Stacey, Inorganic Measurement Section, HSL. Details can be seen in table 1 overleaf.

21. Respirable particulate air samples were collected in accordance with MDHS14/3 “General methods for the sampling and gravimetric analysis of respirable and inhalable dust” using the same sampling media as that used for measuring RCS.

22. Two types of sampling pumps were used during the monitoring period, with three different sampling heads. These are detailed in table 1 overleaf.

23. RCS samples were analysed by HSL’s Analytical Sciences Unit, based at HSL’s Buxton laboratory. The respirable particulate concentrations were determined gravimetrically at the HSL Field Unit, Luton. Results can be found in the BOHS tables overleaf.

82

Table 1 Pump Type Sampling Head Flow Rate Number of Type Samples Buck / Libra GKM 2.69 4 L/min 3

PGP 10 10 L/min 2

SKC / Legacy Impactor 8 L/min 1

83

Results BOHS British Occupational Hygiene Society ENVIRONMENTAL MONITORING DATA Author H Ashpool Tel No. 01582 444230 File reference Ref. to related records Date of sampling Total no of people Agent Respirable Respirable JN0003804 /A1050 08/09/09 on site-Approx 300 Crystalline Silica Particulate (RCS) (RP) Occupier of premises CAS No. - - Address of premises/Location/Identity Units mg/m3 mg/m3 Sampling/Analysis MDHS 101 MDHS 14/3 Details XRD Gravimetric Postcode MSDS Ref - - Department/Area – Coating / Mixing production lines Males exposed 4 4 Building/Room- Coating / Mixing production lines Females exposed 0 0 Reference Sample type / Sample description Male NI no. Sample Duration Result TWA Result TWA Number Sample head (eg name/task/process/equipment Female Personal period (Mins) type 06616/09 SL Static to shelves by LEV grille in N/a - 13:05 – 16:35 210 0.111* - <1.0 - GK 2.69 HAZMAC booth 06617/09 SL Static on gangway above rain sander N/a - 12:55 – 16:33 218 0.083* - <1.0 - GK 2.69 06618/09 PL Production line operative M - 12:45 – 16:30 225 <0.02 <0.03 3.31 4.76 GK 2.69 06622/09 SL Static to rear wall in HAZMAC booth. N/a - 13:00 – 16:35 215 0.026 - 1.78 - Impactor 06626/09 PL HAZMAC booth & production line M - 12:40 – 16:27 227 0.026 0.037 1.52 2.19 PGP 10 operative 06627/09 SL Static by mixing vessels on mezzanine N/a - 12:50 – 16:30 221 0.011 - 0.97 - PGP 10 level above mixing tanks Limits of detection (LOD): LOD defined as 8ug/ filter. LOD in air concentration (mg/m3) calculated as follows: GK 2.69: RCS = 0.02 mg/m3 & RP =1.0 mg/m3 for a 388L air sample. Impactor: RCS= 0.005 mg/m3 & RP = 0.08 mg/m3 for a 1717L air sample. PGP 10: RCS= 0.005 mg/m3 & RP = 0.8 mg/m3 for a 2178L air sample. LTEL (8 Hour) 0.1 mg/m3 4 mg/m3 Exposure limits 3 3 STEL (15 minute) Guidance- 3x LTEL 0.3 mg/m 12 mg/m

84

Industry and SIC Code 29.56 Mould for foundry (manufacture) Comments on origin of sampled material e.g. product name Reason for monitoring Trial of new sampling equipment for RCS monitoring. *Caution should be applied to the interpretation of these results marked with *. The Project Work. deposit on the filter did not match the deposit of the calibration filters and the results are Biological monitoring None positively bias because the dust is located in a small circle in the centre of the filter. These Exposure details Control measures Related records results marked with a * are not meaningful. Conditions Frequency Metabolic RPE LEV rate ** TWA calculated based on 11 hour 30 minutes at work location in a 12-hour shift. One official break of 30 minutes was taken with additional tea breaks taken as and when it was Normal Intermittent N/A N/A Y*** required or time was available, but time spent was variable dependent on work requirements. Normal Continuous N/A N/A N Comments on exposure modifiers, e.g. skin contact, other relevant jobs, confounding factors, biological monitoring Normal Continuous- Moderate Yes- during N/A production line mixing only Intermittent- Mixing Normal Intermittent N/A N/A Y*** *** LEV was present on machine; however, the assessment of the LEV systems was Normal Intermittent- Moderate to Yes- during Y*** outside the scope of this work. It is unknown whether these systems were performing to a HAZMAC High operation of satisfactory standard. booth operation. the Continuous - HAZMAC Production Line booth Normal Continuous N/A N/A N/A

85

Photographs

Photo 1: Mixer in Shell Shop with LEV aligned at top opening. Mixer fed from arm on left hand side from bag tipping position (see photo 2 overleaf).

87

Photo 2: Bag tipping / Mixing operations in Shell Shop. Air fed visor worn during mixing operations.

Photo 3: Production Line Control Booth

88

Photo 4: Wax mould blades transported along line by robot to mixing barrels.

Photo 5: Rain Sander (Flour Shower) coating blades.

89

Photo 6: Carousel of coated items being transported to HAZMAC booth for finishing operations.

90

Photo 7: Dryers on carousel line.

91

Photo 8: Hangers of coating items within HAZMAC booth ready for finishing.

92

Photo 9: Finishing activities in HAZMAC booth.

93

26 SITE 2: POT/BRICK MANUFACTURE

Harpur Hill, Buxton Derbyshire, SK17 9JN T: +44 (0)1298 218000 F: +44 (0)1298 218590 W: www.hsl.gov.uk

HSL Occupational Hygiene Section Site Visit Report

High Volume Silica Monitoring Project

Report Number OH/2009/LET/68 Job Number JN0003804 / A1050 Report author Helen Ashpool Company name & address Site 2: Pot / Brick manufacture Persons seen Keith Jones, Area Health and Safety Advisor Alan Bowman, Site Health and Safety Co-ordinator Date of visit 4th November 2009 HSL/HSE staff present Helen Ashpool, HSL David LeFever, HSE Date of report January 2010 Report distribution Peter Stacey, HSL David LeFever, HSE Phil Smith, HSE Molten Metals & Minerals Section Manufacturing Sector Tim Roff, Head of HSL OHU NEDB OHU Archive Report approval V Sandys

94

Introduction

1. This report describes a site visit on 4th November 2009. The purpose of the visit was to trial new monitoring equipment in the field for the sampling of respirable crystalline silica (RCS). HSE/HSL personnel on site were David LeFever, HM Inspector of Health and Safety, and Helen Ashpool, HSL. Representing the company were Keith Jones, Area Health and Safety Advisor, and Alan Bowman, Site Health and Safety Co- ordinator.

2. The site in Swadlingcote manufacture a range of chimney and roofing products and a range of bricks all of which contain a proportion of free silica, which when dry can produce respirable crystalline silica (RCS).

3. Personal samples for RCS and respirable particulate were undertaken throughout the manufacturing areas. Where possible samples, at the client’s request, were taken on personnel who had previously undertaken routine personal exposure monitoring so they would be able to provide a comparison of the new sampling equipment with equipment used in the past, and so the company could compare results with the previous monitoring exercises. Some problems with the new equipment and possible solutions as noted during sampling operations can be seen in appendix 1.

4. The health hazard associated with the inhalation of respirable crystalline silica (RCS) is silicosis. This is a serious chronic disease where scar tissue is formed in the lungs resulting in a reduction in lung capacity. RCS is also a suspected human carcinogen.

5. RCS is currently assigned a workplace exposure limit (WEL) of 0.1 mg/m3; however, part of the purpose of the overall project was to find out whether the WEL could be lowered to 0.05 mg/m3.

Process description

6. Manufacture a range of chimney and roofing components and a range of bricks. Personal exposure monitoring was undertaken in a variety of areas across the site from the start of the manufacturing process in the clay preparation area, mixing, pressing / moulding machines and at the TK1 machines where operatives removed final products from the conveyors (see photos 1-8).

7. All process involved wet products except at the clay preparation area where the digger driver fed dry raw materials into hoppers, which in turn fed the various machines as required.

8. The mixer was automatic and a single operative worked on a platform at a control panel overlooking the process. The mixer used wet materials; however, it was noted that the area was covered with significant dust deposits (see photo 5).

9. One operative worked at the Spengler machine, which manufactured airbricks for cavity walls. The process was enclosed in the machine. The operative operated the machine via a control panel, which was located adjacent to the pressing position in the machine.

10. One operative worked at the KVM machine, which pressed and fired large brick products. The machine process was enclosed behind a screen except at the exit conveyor where the bricks were stacked (see photos 3 – 4). The operative also worked on a forklift truck, removing the completed products on pallets from the end of the line. 95

11. One operative worked at the “TK1 setters” and one at the “TK1 drawing pots” areas. The operative at the “TK1 setters” position was involved with the removal of products from machine conveyors and stacking on the kiln belt ready for firing. The operative at the “TK1 drawing pots” position was involved with the removal of products from the kiln belt after firing. Both areas had some dust deposits on surfaces. The areas were not served with local exhaust ventilation (LEV).

Exposure controls

12. Exposure controls were not examined in detail as this was outside the scope of the project; however, some details are noted below.

13. The digger driver was protected by ensuring doors and windows to the cab were kept closed when inside. The cab was also provided with air conditioning.

14. All other machines including the mixers used a wet mix. There was therefore limited airborne respirable (fine) particulate as all products were damp.

15. LEV was provided to the Spengler machine, the KVM machine and the mixer; however, these were not assessed as this was outside the scope of the project.

16. Respiratory protective equipment (RPE) was not worn during any operation.

Measurement strategy.

17. Air sampling was undertaken in accordance with methodology described in MDHS101 Crystalline Silica in Respirable Airborne Dusts. Samples were collected at varying flow rates dependent on pump type as described by Peter Stacey, Inorganic Measurement Section, HSL. Details can be seen in table 1 below.

18. Respirable particulate air samples were collected in accordance with MDHS14/3 “General methods for the sampling and gravimetric analysis of respirable and inhalable dust”.

19. Samples were taken for as long a period as possible, but due to work requirements, some sampling periods were shorter than others.

20. Two types of sampling pumps were used during the monitoring period, with three different sampling heads. These are detailed in table 1 below.

21. RCS samples were analysed by HSL’s Analytical Sciences Unit, based at HSL’s Buxton laboratory. The respirable particulate concentrations were determined gravimetrically at the HSL Field Unit, Luton. Results can be found in the BOHS tables overleaf.

Table 1 Pump Type Sampling Head Flow Rate Number of Type Samples Buck / Libra GKM 2.69 4 L/min 3

96

PGP 10 10 L/min 2

SKC / Legacy Impactor 8 L/min 1

97

Results BOHS British Occupational Hygiene Society ENVIRONMENTAL MONITORING DATA Author H Ashpool Tel No. 01582 444230 File reference Ref. to related records Date of sampling Total no of people Agent Respirable Respirable JN0003804 /A1050 04/11/09 on site- approx 150 Crystalline Silica Particulate (RCS) (RP) Occupier of premises Site 2 Pot / Brick manufacture CAS No. - - Address of premises/Location/Identity Units mg/m3 mg/m3 Sampling/Analysis MDHS 101 MDHS 14/3 Details XRD Gravimetric Post Code DE12 7EL MSDS Ref - - Department/Area- Manufacturing Areas Males exposed 150 150 Building/Room- Manufacturing Areas Females exposed 0 0 Reference Sample type / Sample description Male NI no. Sample Duration Result TWA Result TWA Number sample head (eg name/task/process/equipment Female Personal period (Mins) type 08349/09 PL , KVM Concrete Machine M - 10:10 – 13:22 192 0.035 0.034 0.64 0.62 GK 2.69 Operative 08350/09 PL TK1 Drawing Pots M - 10:30 – 13:35 185 0.017 0.016 <0.18 <0.17 GK 2.69 operative 08351/09 PL Mixer Number 7 operative M - 10:35 – 13:30 175 <0.012 <0.012 0.19 0.18 GK 2.69 08352/09 PL . TK1 Setters operative M - 10:20 – 12:05 105 <0.010 <0.010 9.63 9.32 Impactor 08353/09 PL Spengler Machine M - 09:55 – 11:55 165 0.078 0.076 <0.06 <0.06 PGP 10 operative 12:30 – 13:15 08354/09 PL Digger Driver, Clay M - 10:00 – 13:18 198 0.045 0.044 <0.06 <0.06 PGP 10 Preparation Area Limits of detection (LOD): RCS: LOD defined as 8ug/ filter. LOD in air concentration (mg/m3) calculated as follows: GK 2.69: RCS = 0.01 mg/m3 & RP = 0.18 mg/m3 for a 661L air sample. Impactor: RCS= 0.01 mg/m3 & RP = 0.24 mg/m3 for an 841L air sample. PGP 10: RCS= 0.005 mg/m3 & RP = 0.06 mg/m3 for an 1771L air sample LTEL (8 Hour) 0.1 mg/m3 4 mg/m3 Exposure limits STEL (15 minute) Guidance- 3x LTEL 0.3 mg/m3 12 mg/m3

98

Industry and SIC Code 26.26 Bricks made of silica (manufacture) Comments on origin of sampled material e.g. product name 26.40 Chimney liners & pots made of clay (manufacture) Reason for monitoring Trial of new sampling equipment for RCS monitoring. Sampling undertaken during the manufacture of bricks and other roofing products. Project work. Biological monitoring None Operatives were only working a 3-day (30 hour) week due to current work requirements at the time of monitoring. There was a variation in shift length per day; Exposure details Control measures Related records however, the majority of personnel worked from 06:00 hours to 14:30 hours. The Conditions Frequency Metabolic RPE LEV TWA has therefore been calculated based on 7 hour 45 minutes spent at the work rate position in an 8 hour 30 minute shift.

Normal Continuous Moderate No Yes **

Normal Continuous Moderate No No Comments on exposure modifiers, e.g. skin contact, other relevant jobs, confounding factors, biological monitoring Normal Continuous Moderate No Yes ** Normal Continuous Moderate No No ** LEV was present on machine; however, the assessment of the LEV systems was outside the scope of this work. It is unknown whether these systems were Normal Continuous Moderate No Yes ** performing to a satisfactory standard. Normal Continuous Moderate No No

99

100

Photographs

Photo 1: Area of sample RCS 005. Spengler machine, manufacturing airbricks for cavity walls.

Photo 2: Area of sample RCS 006. Ingredient hoppers in Clay Preparation Area.

101

Photo 3: Area of sample RCS 001. KVM concrete machine exit conveyor (leading to photo 4).

Photo 4: Area of sample RCS 001. KVM concrete machine exit / stacking position.

102

Photo 5: Area of sample RCS 003. Mixer No.7 machine.

Photo 6: Area of sample RCS 004. TK1 Setters.

103

Photo 7: Area of location of sample RCS 002. TK1 Drawing Pots. At control panel / operative’s position.

Photo 8: Area of location of sample RCS 002. TK1 Drawing Pots. From operative’s position by control panel looking at incoming conveyor.

104

27 SITE 3: CERAMICS MANUFACTURE

Harpur Hill, Buxton Derbyshire, SK17 9JN T: +44 (0)1298 218000 F: +44 (0)1298 218590 W: www.hsl.gov.uk

HSL Occupational Hygiene Section Site Visit Report

High Volume Silica Monitoring Project

Report Number OH/2009/LET/66 Job Number JN0003804 / A1050 Report author Helen Ashpool Company name & address Site 3 Ceramics manufacture Persons seen Gary Pearson, Health and Safety Manager Mick Clarke, Health and Safety Officer Date of visit 15th October 2009 HSL/HSE staff present Helen Ashpool, HSL Date of report January 2010 Report distribution Peter Stacey, HSL Noelle Walker, HSE FOD Phil Smith, HSE Molten Metals & Minerals Section Manufacturing Sector Tim Roff, Head of HSL OHU NEDB OHU Archive Report approval V Sandys

105

Introduction

1. This report describes a site visit on 15th October 2009. The purpose of the visit was to trial new monitoring equipment in the field for the sampling of respirable crystalline silica (RCS). HSL personnel on site was Helen Ashpool, and representing the company were Gary Pearson, Health and Safety Manager and Mick Clarke, Health and Safety Officer.

2. Pottery manufacture of ceramic kitchenware and tableware items such as pots, plates, cups and bowls.

3. Clay, slip and glazes used in the ceramics industry all contain silica, which when dry can produce respirable crystalline silica (RCS). RCS can be found in various materials in the ceramics industry in varying proportions. See http://www.hse.gov.uk/pubns/guidance/cr0.pdf for details.

4. Samples for RCS and respirable particulate were positioned where possible, at the client’s request, on personnel who had previously undertaken routine personal exposure monitoring so they would be able to provide a comparison of the new sampling equipment with equipment used previously. Some problems with the new equipment and possible solutions as noted during sampling operations can be seen in appendix 1.

5. Static samples were where possible, positioned in areas where sampling had previously been undertaken so the company would be able to compare results with the previous monitoring exercises.

6. The health hazard associated with the inhalation of respirable crystalline silica (RCS) is silicosis. This is a serious chronic disease where scar tissue is formed in the lungs resulting in a reduction in lung capacity. RCS is also a suspected human carcinogen.

7. RCS is currently assigned a workplace exposure limit (WEL) of 0.1 mg/m3; however, part of the purpose of the overall project was to find out whether the WEL could be lowered to 0.05 mg/m3.

Process description

8. Monitoring was undertaken at various stages in the manufacturing process from the clay preparation area to the pressing and hand fettling of products.

9. The process began in the clay preparation area / slip house, where clay was extruded for use in the manufacturing areas (see photos 4, 5 and 6). Personnel in this area operated various automatic machines.

10. The extruded clay was cut and then pressed in various manufacturing areas and then placed in adjacent dryer carousels (see photos 1 and 2). Monitoring was undertaken in the Ram Press Area where wet extruded clay was cut using a guillotine before transferring into a press. In this area, the clay was pressed into a flat dish with a rim shape and then transferred into a drier. The press was wiped clean by the operative after product removal and the cycle started again.

11. Exposure monitoring was also undertaken in the Small Hollowware Glazing (SHG) area. Products were removed from dryer carousels and hand fettled within the confines

106

of a small table mounted enclosure served with local exhaust ventilation (LEV) (see photo 3).

Exposure controls

12. Exposure controls were not examined in detail as this was outside the scope of the project; however, some details are noted below.

13. All processes monitored dealt with wet materials except for the hand-fettling position at the SHG area. Here the products were dried before fettling was undertaken; however, the enclosure used for undertaking this activity was served with LEV.

14. As the clay was wet in all other monitored areas there would be less potential for generating airborne particulates.

15. LEV served the dryers in the Ram Press Area.

16. No RPE was worn by any operative in any of the monitored areas.

Measurement strategy.

17. Air sampling was undertaken in accordance with methodology described in MDHS101 Crystalline Silica in Respirable Airborne Dusts. Samples were collected at varying flow rates dependent on pump type as described by Peter Stacey, Inorganic Measurement Section, HSL. Details can be seen in table 1 overleaf.

18. Respirable particulate air samples were collected in accordance with MDHS14/3 “General methods for the sampling and gravimetric analysis of respirable and inhalable dust” using the same sampling media as that used for measuring RCS.

19. Two types of sampling pumps were used during the monitoring period, with three different sampling heads. These are detailed in table 1 overleaf.

20. RCS samples were analysed by HSL’s Analytical Sciences Unit, based at HSL’s Buxton laboratory. The respirable particulate concentrations were determined gravimetrically at the HSL Field Unit, Luton. Results can be found in the BOHS tables overleaf.

107

Table 1 Pump Type Sampling Head Flow Rate Number of Type Samples Buck / Libra GKM 2.69 4 L/min 3

PGP 10 10 L/min 2

SKC / Legacy Impactor 8 L/min 1

108

Results BOHS British Occupational Hygiene Society ENVIRONMENTAL MONITORING DATA Author H Ashpool Tel No. 01582 444230 File reference Ref. to related records Date of sampling Total no of people Agent Respirable Respirable JN0003804 /A1050 15/ 10/09 on site- 267 Crystalline Silica Particulate (RCS) (RP) Occupier of premises CAS No. - - Address of premises/Location/Identity Units mg/m3 mg/m3 Sampling/Analysis MDHS 101 MDHS 14/3 Details XRD Gravimetric Post Code MSDS Ref - -

Department/Area – Clay Preparation and Manufacturing Areas Males exposed 175 175 Building/Room- Clay Preparation and Manufacturing Areas Females exposed 92 92 Reference Sample type / Sample description Male NI no. Sample Duration Result TWA Result TWA Number sample head (eg name/task/process/equipment Female Personal period (Mins) type 07806/09 SL Static at Flowline D Trim, Small N/a - 10:05 – 13:35 210 0.069 - <1.1 - GK 2.69 Holloware Glazing Area (SHG) 07807/09 SL Static by walkway in Ram Press Area N/a - 09:55 – 13:40 225 0.017 - <1.1 - GK 2.69 07808/09 SL Static on workbench in Slip House, Clay N/a - 09:25 – 13:30 245 0.016 - <1.1 - GK 2.69 Preparation Area 07809/09 PL Flowline D Trim Operative F - 10:10 – 12:10 120 0.010 0.010 0.97 0.99 Impactor 07810/09 PL Slip House Team Leader, Clay Preparation M - 09:30 – 11:50 140 0.146 0.149 <0.05 <0.05 PGP 10 Area 07811/09 PL Ram Press Operative F - 09:50 – 12:15 145 0.026 0.026 <0.05 <0.05 PGP 10 Limits of detection (LOD): RCS: LOD defined as 8ug/ filter. LOD in air concentration (mg/m3) calculated as follows: GK 2.69: RCS = 0.009 mg/m3 & RP =1.1 mg/m3 for an 880L air sample. Impactor: RCS= 0.008 mg/m3 & RP = 0.10 mg/m3 for a 955L air sample. PGP 10: RCS= 0.005 mg/m3 & RP = 0.05 mg/m3 for an 1490L air sample LTEL (8 Hour) 0.1 mg/m3 4 mg/m3 Exposure limits STEL (15 minute) Guidance- 3x LTEL 0.3 mg/m3 12 mg/m3

109

Industry and SIC Code 26.21 Kitchenware made of ceramics (manufacture) Comments on origin of sampled material e.g. product name Reason for monitoring Trial of new sampling equipment for RCS monitoring. Sampling undertaken during the manufacture of ceramic kitchenware and tableware Project Work. products such as pots, plates, cups and bowls. Biological monitoring None *TWA calculated based on 8 hours 10 minutes at the work location in a 9-hour shift. Exposure details Control measures Related records Conditions Frequency Metabolic RPE LEV rate

Normal Continuous N/a N/a Yes **

Normal Continuous N/a N/a Yes ** Comments on exposure modifiers, e.g. skin contact, other relevant jobs, confounding factors, biological monitoring Normal Continuous N/a N/a No Normal Continuous Moderate No Yes ** ** LEV was present on machine; however, the assessment of the LEV systems was outside the scope of this work. It is unknown whether these systems were performing to a Normal Continuous Moderate No No satisfactory standard. Normal Continuous Moderate No Yes **

110

111

Photographs

Photo 1: Location of sample RCS 006. Ram Press operative.

Photo 2: Location of sample RCS 002. Static location sample by Ram Press Dryer 03.

112

Photo 3: Location of sample RCS 004 and RCS 001. Flowline D Trim operative and static location sample at Flowline D Trim. Small Holloware Glazing Area.

Photo 4: Location of sample RCS 003. Static location sample on workbench in Slip House. Clay Preparation Area.

113

Photo 5: Slip House. Clay Preparation Area. Sample RCS 005 located on, team leader, working throughout this area.

Photo 6: Slip House. Clay Preparation Area. Clay is extruded by machine and stacked before transferral into main production area. Sample RCS 005 was located on, team leader, working throughout this area.

114

28 SITE 4 CONSTRUCTION

Harpur Hill, Buxton Derbyshire, SK17 9JN T: +44 (0)1298 218000 F: +44 (0)1298 218590 W: www.hsl.gov.uk

HSL Occupational Hygiene Section Site Visit Report

High Volume Silica Monitoring Project Site 4 Construction

Report Number OH/2009/LET/67 Job Number JN0003804 / A1050 Report author Helen Ashpool Company name & address Site 4 Construction Persons seen Mark Hotson Date of visit 21st October 2009 HSL/HSE staff present Helen Ashpool, HSL John Berezansky, HSE Mark Lucas, HSE Date of report January 2010 Report distribution Peter Stacey, HSL John Berezansky, HSE Construction Sector Mark Lucas, HSE Tim Roff, Head of HSL OHU NEDB OHU Archive Report approval V Sandys

115

Introduction

1. This report describes a site on 21st October 2009. The purpose of the visit was to trial new monitoring equipment in the field for the sampling of respirable crystalline silica (RCS). HSE/HSL personnel on site were John Berezansky, HM Inspector of Health and Safety, Mark Lucas, HSE FOD Litigation Officer and Helen Ashpool, HSL. Representing the company were Mark Hotson, Project EHS Manager,

2. Large-scale construction work is being undertaken were the principle contractor operating on the site undertaking building work worked with another contractor, who were involved in concrete cutting activities as and when it was required.

3. Monitoring was undertaken on contractors during the cutting of concrete surfaces and scabbling activities.

4. The weather was wet during the monitoring period. Sampling was undertaken during light rain. The ground was also significantly damp from previous rainfall.

5. Concrete contains crystalline silica, which is released during concrete cutting activities such as scabbling (see http://www.hse.gov.uk/pubns/cis36.pdf for more information).

6. The health hazard associated with the inhalation of respirable crystalline silica (RCS) is silicosis. This is a serious chronic disease where scar tissue is formed in the lungs resulting in a reduction in lung capacity. RCS is also a suspected human carcinogen.

7. RCS is currently assigned a workplace exposure limit (WEL) of 0.1 mg/m3; however, part of the purpose of the overall project was to find out whether the WEL could be lowered to 0.05 mg/m3.

Process description

8. Sampling was undertaken on a construction site at RAF Wyton where various contractors were undertaking work that included scabbling and concrete cutting operations (see photos 1 and 2).

Concrete Cutting

9. Concrete cutting was undertaken by an operative using a Husquarna concrete cutter. Cutting was undertaken on a concrete surface, wet from earlier rainfall. The machine cut lines in the concrete surface along marked lines measured out by contractors

10. The Husquarna concrete cutter was operated by pushing along the surface, hand guided along the indicated lines. The machine was fitted with a water barrel which sprayed water onto the surface prior to cutting.

11. A limited amount of concrete cutting could be monitored due to work requirements.

Scabbling

12. An operative undertook scabbling, cutting the edges of a concrete ridge / kerb (see photo 2) as required using a hand-held tool. A limited amount of scabbling could be

116

monitored due to hand-arm vibration safety issues. The operative had already undertaken some scabbling before the monitoring began.

Exposure controls

13. Exposure controls were not examined in detail as this was outside the scope of the project; however, some details are noted below.

14. No were utilised during scabbling; however, the concrete cutting machine utilised damping down techniques. Water from a barrel mounted on the machine was automatically sprayed at the cutting position on the ground.

15. No respiratory protective equipment (RPE) was worn by the concrete cutter operative; however, a disposable FFP2 respirator was worn by the scabbling operative. This type of RPE was unvalved and the operative reported to find it hard to work in this type of respirator due to the high metabolic rate required for this type of work. It was reported that normally the RPE provided was of specification FFP3 and was valved; however, the stores had run out of this type, and the unvalved type was the only available.

Measurement strategy.

16. Air sampling was undertaken in accordance with methodology described in MDHS101 Crystalline Silica in Respirable Airborne Dusts. Samples were collected at varying flow rates dependent on pump type as described by Peter Stacey, Inorganic Measurement Section, HSL. Details can be seen in table 1 overleaf.

17. Respirable particulate air samples were collected in accordance with MDHS14/3 “General methods for the sampling and gravimetric analysis of respirable and inhalable dust” using the same sampling media as that used for measuring RCS.

18. A limited amount of sampling could be undertaken in both areas monitored. The concrete cutter had only a limited amount of work to undertake, and the scabbling operative had already undertaken a significant amount of scabbling in the morning prior to the visit and was unable to undertake much more work due to hand-arm vibration safety issues.

19. Two types of sampling pumps were used during the monitoring period, with three different sampling heads. These are detailed in table 1 below.

20. RCS samples were analysed by HSL’s Analytical Sciences Unit, based at HSL’s Buxton laboratory. The respirable particulate concentrations were determined gravimetrically at the HSL Field Unit, Luton. Results can be found in the BOHS tables overleaf.

Table 1 Pump Type Sampling Head Flow Rate Number of Type Samples Buck / Libra GKM 2.69 4 L/min 3

PGP 10 10 L/min 2

SKC / Legacy Impactor 8 L/min 1

117

118

Results BOHS British Occupational Hygiene Society ENVIRONMENTAL MONITORING DATA Author H Ashpool Tel No. 01582 444230 File reference Ref. to related records Date of sampling Total no of people Agent Respirable Respirable JN0003804 /A1050 21/10/09 on site- variable Crystalline Silica Particulate ~200 (RCS) (RP) Occupier of premises Site 4: Construction CAS No. - - Address of premises/Location/Identity Units mg/m3 mg/m3 Sampling/Analysis MDHS 101 MDHS 14/3 Details XRD Gravimetric Post Code MSDS Ref - - Department/Area- Construction Area Males exposed (directly) 2 2 Building/Room- Construction Area Females exposed 0 0 Reference Sample type / Sample description Male NI no. Sample Duration Result TWA Result TWA Number sample head (eg name/task/process/equipment Female Personal period (Mins) type 07791/09 SS Static to ground adjacent to concrete N/a - 09:40 – 10:10 30 <0.068 - 2.1 - GK 2.69 cutting line 07792/09 PS Personal undertaking scabbling activities M - 11:00 – 11:40 40 0.153 0.077 * 1.2 <1.1 GK 2.69 07793/09 SS Static to right hand side of Husquarna N/a - 09:35 – 10:10 35 0.260 - 3.6 - GK 2.69 concrete cutter beneath handle. 07794/09 SS Static to ground adjacent to concrete N/a - 09:40 – 10:10 30 0.047 - 11.68 - Impactor cutting line 07795/09 PS Personal undertaking scabbling activities M - 11:00 – 11:40 40 0.125 0.063 * 6.0 3.0 * PGP 10 07796/09 SS Static to right hand side of Husquarna N/a - 09:30 – 10:10 40 0.120 - 7.1 - PGP 10 concrete cutter beneath handle. Limits of detection (LOD): RCS: LOD defined as 8ug/ filter. LOD in air concentration (mg/m3) calculated as follows: GK 2.69: RCS= 0.057 mg/m3 & RP = 1.1 mg/m3 for a 140L air sample. Impactor: RCS= 0.033 mg/m3 & RP = 0.36 mg/m3 for a 239.85L air sample. PGP 10: RCS= 0.02mg/m3 & RP = 0.21 mg/m3 for a 400L air sample. 3 3 LTEL (8 Hour) 0.1 mg/m 4 mg/m Exposure limits STEL (15 minute) Guidance- 3x LTEL 0.3 mg/m3 12 mg/m3

119

Industry and SIC Code 45.25 Diamond drilling of concrete and asphalt Comments on origin of sampled material e.g. product name 45.21/3 Constructional Engineering Reason for monitoring Trial of new sampling equipment for RCS monitoring. Sampling undertaken during scabbling activities and cutting of concrete by contractors Project Work. working on the site. *TWA calculated based on 4 hours spent at the working position in a 10-hour shift. Biological monitoring None The concrete cutting operative worked only on the site as and when it was required. This Exposure details Control measures Related records was reported to be from one hour in a day or for a full day if required. A limited amount of sampling could be undertaken in both areas monitored. The concrete Conditions Frequency Metabolic RPE LEV rate cutter had only a limited amount of work to undertake, and the scabbling operative had already undertaken a significant amount of scabbling in the morning prior to the visit and

was unable to undertake much more work due to hand-arm vibration safety issues. Normal Intermittent N/a N/a No

Normal Intermittent High No No Comments on exposure modifiers, e.g. skin contact, other relevant jobs, confounding factors, biological monitoring Normal Intermittent N/a N/a No

Normal Intermittent N/a N/a No Sampling was undertaken during light rain. The ground was also damp from previous Normal Intermittent High No No heavy rain. The surface water would have damped down any dust released. It was thought that on a dry day there would have been considerably more airborne particulate present. Normal Intermittent N/a N/a No

The scabbling operative reported never to undertake scabbling operations for more than 4 hours in one shift due to hand-arm vibration safety issues. If more scabbling was required, then a second operative would undertake the operation. The results indicate exposure for scabbling only and not other operations undertaken throughout the shift, which may generate more respirable particulate and / or RCS.

120

Photographs

Photo 1: Area of concrete cutting. Samples RCS001, 003, 004 & 006.

121

Photo 2: Scabbling operations

122

29 SITE 5: QUARRY VISIT

Harpur Hill, Buxton Derbyshire, SK17 9JN T: +44 (0)1298 218000 F: +44 (0)1298 218590 W: www.hsl.gov.uk

HSL Occupational Hygiene Section Site Visit Report

High Volume Silica Monitoring Project Site 5: Quarry

Report Number OH/2009/LET/69 Job Number JN0003804 / A1050 Report author Helen Ashpool Company name & address Quarry/Stonemasons Persons seen Shaun Berry, Works Manager Nicola Bullas, Head of Administration Date of visit 19th November 2009 HSL/HSE staff present Helen Ashpool, HSL Jackie Ferguson, HSE Date of report January 2010 Report distribution Peter Stacey, HSL Chris Keen, HSL Jackie Ferguson, HSE FOD Phil Smith, HSE Molten Metals & Minerals Section Manufacturing Sector Tim Roff, Head of HSL OHU NEDB OHU Archive Report approval V Sandys

Introduction

123

1. This report describes a site visit on 19th November 2009. The purpose of the visit was to trial new monitoring equipment in the field for the sampling of respirable crystalline silica (RCS). HSE/HSL personnel present on site were Jackie Ferguson (HM Inspector of Health and Safety) and Helen Ashpool (HSL). Representing the company were Shaun Berry, Works Manager and Nicola Bullas, Head of Administration.

2. A stonemasonry workshop was located at the site as well as a quarry. The workshop utilised various machines to cut the quarried stone as well as new robots. The workshop did not utilise traditional dry hand cutting, chipping and grinding techniques.

3. The company manufactures various products of all sizes for architectural masonry, cladding, internal flooring, paving, landscape / feature stonework, and walling. See http://www.myersgroup.co.uk/jwqc/default.asp for details.

4. Two types of sandstone are quarried on the site and are known as “Woodhead Natural York Stone” and “Crosland Hill Hard York Stone”. See http://www.myersgroup.co.uk/jwqc/default.asp for details of the stone types.

5. Sampling for respirable crystalline silica and respirable particulate was undertaken both in cabs of excavators in the quarry and on personnel operating machines in the workshop. Static location samples were also undertaken in the new robot area. Samples were positioned at the client’s request, in areas and on personnel where annual routine monitoring had previously been undertaken. This would provide a comparison of the new sampling equipment with equipment used in the past, and the company could also compare results with the previous exercises. Some problems with the new equipment and possible solutions as noted during sampling operations can be seen in appendix 1.

6. The health hazard associated with the inhalation of respirable crystalline silica (RCS) is silicosis. This is a serious chronic disease where scar tissue is formed in the lungs resulting in a reduction in lung capacity. RCS is also a suspected human carcinogen.

7. RCS is currently assigned a workplace exposure limit (WEL) of 0.1 mg/m3; however, part of the purpose of the overall project was to find out whether the WEL could be lowered to 0.05 mg/m3.

8. Sandstone typically contains greater than 70% free silica http://www.hse.gov.uk/PUBNS/guidance/st0.pdf

124

Process description

9. The stonemasonry workshop and quarry were both visited at request of the client.

10. Various vehicles were used in the quarry to excavate the stone (see photos 1 and 2).

11. Excavated stone was transferred from the quarry to the stonemasonry workshop where various automatic wet cutting machines cut the quarried stone to size and shape. A new robot production line was also in operation. The stonemasonry workshop did not utilise traditional dry hand cutting, chipping and grinding techniques.

12. Personnel operated the various automatic cutting machines from adjacent control panels or from the adjacent control room for the robot production line (see photos 3- 6).

Exposure controls

13. Exposure controls were not examined in detail as this was outside the scope of the project; however, some details are noted below.

Quarry

14. All operatives working in this area were located within cabs of various excavator vehicles. All vehicle cabs were air-conditioned and all doors and windows remained closed when the operative was inside. Sample pumps were positioned within the cabs and not on the operative as it was felt that these would hinder movement. Sample pumps were positioned where they would not be accidentally knocked, with the sampling head as near to the breathing zone as possible (i.e. on the cab frame / window, at the height of the operative). Vehicles can be seen in photos 1 & 2.

Workshop

15. All machinery within the workshop utilised wet cutting techniques. Dust that was created was damped down and therefore minimising operative exposure. The operative’s who were monitored, operated machines from control panels adjacent to the machines (see photos 5 & 6) in the main workshop, or from a control room adjoining in the robot production line area (see photos 3 & 4).

Respiratory Protective Equipment (RPE)

16. No RPE was worn by any operative in the quarry or in the stone masonry workshop.

125

Measurement strategy.

17. Air sampling was undertaken in accordance with methodology described in MDHS101 Crystalline Silica in Respirable Airborne Dusts. Samples were collected at varying flow rates dependent on pump type as described by Peter Stacey, Inorganic Measurement Section, HSL. Details can be seen in table 1 below.

18. Respirable particulate air samples were collected in accordance with MDHS14/3 “General methods for the sampling and gravimetric analysis of respirable and inhalable dust” using the same sampling media as that used for RCS.

19. Two types of sampling pumps were used during the monitoring period, with three different sampling heads. These are detailed in table 1 below.

20. RCS samples were analysed by HSL’s Analytical Sciences Unit, based at HSL’s Buxton laboratory. The respirable particulate concentrations were determined gravimetrically at the HSL Field Unit, Luton. Results can be found in the BOHS tables overleaf.

Table 1 Pump Type Sampling Head Flow Rate Number of Type Samples Buck / Libra GKM 2.69 4 L/min 3

PGP 10 10 L/min 2

SKC / Legacy Impactor 8 L/min 1

126

Results BOHS British Occupational Hygiene Society ENVIRONMENTAL MONITORING DATA Author H Ashpool Tel No. 01582 444230 File reference Ref. to related records Date of sampling Total no of people Agent Respirable Respirable JN0003804 /A1050 19/11/09 on site Crystalline Particulate Approx 80 Silica (RCS) (RP) Occupier of premises Site 5 CAS No. - - Address of premises/Location/Identity Units mg/m3 mg/m3 Sampling/Analysis MDHS 101 MDHS 14/3 Details XRD Gravimetric Post Code MSDS Ref - - Department/Area- Quarry and Stonemasonry workshop Males exposed 80 80 Building/Room- Quarry and Stonemasonry workshop Females exposed 0 0 Reference Sample type / Sample description Male NI no. Sample Duration Result TWA Result TWA Number sample head (eg name/task/process/equipment Female Personal period (Mins) type 08872/09 SL Static in cab of loading shovel operated by, N/a - 11:10 – 14:55 225 <0.036 - <1.5 - GK 2.69 Quarry 08873/09 PL , VV1 & 2 operative. Workshop M - 11:25 – 15:20 235 0.328 0.390 <1.5 <1.8 GK 2.69 08874/09 PL , Loffler MLS4 operative. Workshop M - 11:30 – 15:15 225 0.448 0.504 <1.5 <1.8 GK 2.69 08875/09 SL Static on windowsill in centre of robot N/a - 11:40 – 15:10 210 <0.005 - 7.49 - Impactor control room. Workshop 08876/09 SL Static opposite robot production line. N/a - 11:35 – 15:10 215 0.008 - <0.08 - PGP 10 Workshop 08877/09 SL Static in cab of Liebherr 954 excavator N/a - 10:50 – 14:45 235 0.020 - <0.08 - PGP 10 operated by Limits of detection (LOD): RCS: LOD defined as 8ug/ filter. LOD in air concentration (mg/m3) calculated as follows: GK 2.69: RCS = 0.036 mg/m3 & RP =1.5 mg/m3 for an 838L air sample. Impactor: RCS = 0.005 mg/m3 & RP = 0.05 mg/m3 for a 1680L air sample. PGP 10: RCS = 0.004 mg/m3 & RP = 0.08 mg/m3 for a 2243L air sample. LTEL (8 Hour) 0.1 mg/m3 4 mg/m3 Exposure limits STEL (15 minute) Guidance- 3x LTEL 0.3 mg/m3 12 mg/m3

127

Industry and SIC Code 14.11 Quarrying of ornamental and building stone & Comments on origin of sampled material e.g. product name 45.25 Stonemasonry (building) Reason for monitoring Trialling of new sampling equipment for RCS monitoring. Exposure monitoring undertaken during quarrying and stonemasonry activities using Project Work. two types of sandstone, “Woodhead Natural York Stone” and “Crosland Hill Hard York Stone”. Biological monitoring None

Exposure details Control measures Related records * TWA calculated on 9 hours 30 minutes at work location in a 10 hour shift. Conditions Frequency Metabolic RPE LEV rate

Normal Continuous N/a N/a No

Normal Continuous Moderate No No Comments on exposure modifiers, e.g. skin contact, other relevant jobs, confounding factors, biological monitoring Normal Continuous Moderate No No Normal Continuous N/a N/a No The weather was extremely windy in the quarry areas and very wet, although it was not raining during the sampling. Airborne particulates would have been damped Normal Continuous N/a N/a No down and therefore releasing lower amounts of contaminants that would have been Normal Continuous N/a N/a No when dry.

128

129

Photographs

Photo 1: Location of sample RCS 006 inside cab of Liebherr 954 excavator operated by. Quarry.

Photo 2: Location of sample RCS 001 inside cab of loading shovel 180D operated by Quarry.

130

Photo 3: Location of sample RCS 004, on windowsill in centre of Robot Control Room. Workshop.

131

Photo 4: Location of sample RCS 005, opposite Robot Production Line. Workshop.

Photo 5: Location of sample RCS 003., Loffler MLS4 (4 headed diamond blade circular saw) operative. Workshop.

132

Photo 6: Location of sample RCS 002., Van Vorden (VV) 1 & 2 (diamond blade profile saw) operative. Workshop.

133 Published by the Health and Safety Executive 10/10 Health and Safety Executive

Testing of high flow rate respirable samplers to assess the technical feasibility of measuring 0.05 mg.m-3 respirable crystalline silica

Testing of high flow rate samplers to assess the technical feasibility of measuring 0.05 mg.m-3 respirable crystalline silica.

This report describes testing of five personal respirable dust samplers operating with flow rates of 4 l/min or greater, available in 2008. Three were commercially available, one a prototype and one adapted at HSL to operate at a higher flow rate. Testing compared these samplers with a reference sampler, operating at 2.2 l/min, to ascertain if an increase in the mass of dust sampled could improve the reliability of measurements of respirable crystalline silica (RCS). None of the samplers satisfied all of the success criteria for the project, which included, the ability to maintain the specified flow rate over 4-hours, ease of use in the workplace, and an improvement in the measurement precision without additional complications caused by the increased mass of sampled dust. Infrared analysis is not recommended for samples with dust mixtures, because it was difficult to obtain a reliable result when the loading exceeds 1 mg. The samplers with the best performance were the PGP10 and the modified GK2.69 samplers. The other samplers tested either under-sampled or there was lost sample during transfer onto the analysis filter. When field tests were conducted, air sampling pumps operating with the modified GK2.69 samplers failed to maintain a consistent flow rate, and the PGP 10 samplers were heavy and caused discomfort for the workers The report recommends the use of the PGP 10 and GK2.69 samplers after further work to resolve the minor issues and changes in the sampling and measurement strategies to accommodate new procedures for use of higher flow rate samplers.

This report and the work it describes were funded by the Health and Safety Executive (HSE). Its contents, including any opinions and/or conclusions expressed, are those of the authors alone and do not necessarily reflect HSE policy.

RR825 www.hse.gov.uk