Exposure to Fibres, Crystalline Silica, Silicon Carbide and Sulphur Dioxide in the Norwegian Silicon Carbide Industry

ABSTRACT

Objectives: The aim of this study was to assess personal exposure to fibres, crystalline silica, silicon carbide (SiC) and sulphur dioxide in the Norwegian SiC industry.

Methods: Approximately 720 fibre samples, 720 respirable dust samples and 1400 total dust samples were collected from randomly chosen workers from the furnace, processing and maintenance departments in all three Norwegian SiC plants. The respirable dust samples were analysed for quartz, cristobalite and non-fibrous SiC content. Approximately 240 sulphur dioxide samples were collected from workers in the furnace department.

Results: The sorting operators from all plants, control room and cleaning operators in Plant A and charger, charger/mix and payloader operators in Plant C had a geometric mean (GM) of fibre exposure above the Norwegian occupational exposure limit (OEL) (0.1 fibre cm−3). The cleaner operators in Plant A had the highest GM exposure to respirable quartz (20 μg m−3). The charger/mix operators in Plant C had the highest GM exposure to respirable cristobalite (38 μg m−3) and the refinery crusher operators in Plant A had the highest GM exposure to non-fibrous SiC (0.65 mg m−3). Exposure to the crystalline silica and non-fibrous SiC was generally low and between 0.4 and 2.1% of the measurements exceeded the OELs. The cleaner operators in Plant A had the highest GM exposure to respirable dust (1.3 mg m−3) and total dust (21 mg m−3). GM exposures for respirable dust above the Norwegian SiC industry-specific OEL of 0.5 mg m−3 were also found for refinery crusher operators in all plants and mix, charger, charger/mix and sorting operators in Plant C. Only 4% of the total dust measurements exceeded the OEL for nuisance dust of (10 mg m−3). Exposure to sulphur dioxide was generally low. However, peaks in the range of 10–100 p.p.m. were observed for control room and crane operators in Plants A and B and for charger and charger/mix operators in Plant C.

Conclusion: Workers in the SiC industry are exposed to a mixture of several agents including SiC fibres, quartz, cristobalite, non-fibrous SiC and sulphur dioxide. Exposure levels were generally below the current Norwegian OELs; however, high exposure to fibres and respirable dust still occurs in the furnace department.

MATERIALS AND METHODS

SiC appears in two different crystalline forms. The hexagonal α-SiC is the main product, while the cubic β-SiC is formed at lower temperatures and is used in the metallurgic industry or recycled into the furnace mix. SiC is produced as either green or black crystals from a mixture of high-grade quartz sand and petrol coke. In the production of black SiC product, reclaimed furnace mix and aluminium oxide may be added to the furnace mix, and sawdust is sometimes added to improve porosity of the furnace mix. The furnace mix is transported to the furnace building and loaded into an electrical resistance furnace with a graphite core in the centre. The furnace mix is heated electrically by the graphite core that functions as a resistance element. Quartz (SiO2) will react with carbon and form α-SiC and carbon monoxide (CO) at temperatures >1700°C according to the chemical reaction:

After the completion of a furnace cycle (8–10 days), unreacted material is removed and returned to the mix area, while the crude SiC is transported to the sorting area. In the sorting area β-SiC is removed from α-SiC. α-SiC is then crushed and transported to the processing department where it is crushed further and treated chemically with pine oil, sulphuric acid and sodium hydroxide to remove unreacted crystalline silica, silicon and carbon. Metallic impurities are removed by magnetic separation. SiC is then sieved and classified into size fractions (grits) with a mass median particle size ranging from 0.1 to 880 μm. The high temperatures in the furnace will transform quartz into cristobalite, another crystalline form of silica. Sulphur impurities in the coke will lead to emission of sulphur dioxide.

A SiC plant can be divided into three different departments: furnace department where the crude SiC is produced, processing department where the SiC grits are manufactured and maintenance department responsible for maintenance work in all parts of the plant, see Fig. 1 and Table 1.

Plant characteristics

Exposure assessment was performed in all three Norwegian SiC plants. One plant is located in central Norway, while the two other plants are located in the southern part of Norway. The three plants employ a total of ∼350 production and maintenance workers. The furnaces are located inside a furnace hall building. Differences between plants are described under the job groups in Table 1.

Sampling strategy

The agents that were measured were fibres, respirable quartz, respirable cristobalite, respirable non-fibrous SiC, respirable dust, total dust and sulphur dioxide. Other agents known to be present in the work atmosphere in the SiC industry are carbon monoxide, PAHs and amorphous silica. Carbon monoxide was not measured, as this gas was not expected to induce chronic respiratory effects. Amorphous silica was not analysed due to analytical limitations. PAHs were not quantified due to relative low levels reported in other studies (Dufresne et al., 1987; Petry et al., 1994).

Walk-through surveys of the plants were performed by one of the authors and information on jobs and tasks was collected. Workers were then divided into job groups performing similar tasks in similar work conditions. The jobs were categorized as described in detail in Table 1.

A random sample of workers from each group was invited to participate in the study and all except one agreed to participate. Exposure to dust and gas was determined by means of personal sampling. The aim was to measure two or more agents for each person for at least 2 days. Workers were interviewed after sampling for their perception of the work conditions and respirator use. There were no criteria given for stating normal/worse/better working conditions other than the workers’ own perception of the work on that specific day.

The sampling duration for sulphur dioxide, total dust full-shift samples and respirable dust was close to a full work shift (6–8 h). Sampling duration for total dust short-term and fibre samples was limited to 0.5–3.5 h to avoid particle overload of the filter used for fibre analyses.

The sampling was performed between November 2002 and December 2003 and included two sampling periods of two work weeks (total 10 days). One sampling period was during autumn/winter (November–February) and the second period during spring/summer (May–August).

The work was organized as daytime only, two-shift schedule per day (morning and afternoon) or three-shift schedule per day (morning, afternoon and night). Sampling was evenly divided on morning and afternoon shifts. In addition, four night shifts per agent per job group were included in each plant.

Sampling methods and analysis

Total dust full-shift samples were collected on 37-mm cellulose acetate filters (Millipore Corporation, Bedford, NY, USA), with a 5.0-μm pore size, whereas the total dust short-term samples were collected on 37-mm Teflon (polytetrafluoroethylene) filters (Millipore Corporation). Both filters were fitted in 37-mm closed-faced aerosol filter cassettes (Millipore Corporation) applying a sampling flow rate of 2 l min−1. Respirable dust was collected on 37-mm cellulose acetate filters with a pore size of 5.0 μm (Millipore Corporation) using a cyclone (Casella T13026/2, London, UK) at a sampling flow rate of 2.2 l min−1. The particle mass was measured with a microbalance Sartorius Micro MC 210 P (Sartorius AG, Goettingen, Germany), with a detection limit of 0.06 mg. The detection limit with 8-h sampling time was therefore 0.063 mg m−3 for total dust and 0.057 mg m−3 for respirable dust. The amount of quartz, cristobalite and non-fibrous SiC in the respirable dust was determined by the use of X-ray powder diffractometry, applying Philips PW1729 X-ray generator, Phillips PW 1710 diffractometer control and Phillips APD software. The crystalline silica was determined by the use of standard methods (Bye, 1983; NIOSH, 1998) with some modifications due to the presence of graphite in samples from the furnace hall. Graphite interferes with quartz in the analyses and was removed by high-temperature ashing (700°C). The detection limit for quartz was 5 μg, which amounts to 5.2 μg m−3 with 8-h sampling time, and for cristobalite 10 μg, which amounts to 10.4 μg m−3 with 8-h sampling time. Non-fibrous SiC was determined by a corresponding X-ray method developed in our laboratory (E Bye et al, in press). Pure SiC products from the three plants were used for calibration purposes. The detection limit for SiC was 12 μg which amounts to 0.013 mg m−3 with 8-h sampling time. Due to the detection limits of the XRD analytical methods, dust samples were combined if there was not enough dust to ensure sufficient material for analysis (>0.7 mg). Samples were combined within plant and job groups, preferably from the same persons. The detection limits apply to the combined samples and the detection limit for the individual samples would be lower depending on the amount of dust in the sample. The total number of samples was 680 and they were combined into 272 analyses.

Fibres were collected on 25-mm cellulose acetate filters (Millipore Corporation) with a pore size of 1.2 μm using an open-face aerosol filter cassette of conducting polypropylene (Gelman Sciences, Ann Arbor, MI, USA) at a sampling flow rate of 1 l min−1. The fibres were counted with a light microscope according to World Health Organization (WHO) counting criteria (WHO, 1997) with a detection limit of four fibres which amounts to 0.016 fibres cm−3 with a sampling time of 2 h.

The samples were collected in parallel with a cyclone and a total dust cassette, or a fibre cassette and a short-term total dust cassette placed side by side on the worker. The two parallel cassettes were connected to the same high-flow pump through a hose with a Y-passage (SKC Inc., Eighty Four, PA, USA).

Sulphur dioxide was measured with direct-reading electrochemical sensors with a data-logging facility built into the instrument (PAC III Dräger Aktiengesellschaft, Lübeck, Germany). An averaging period of one reading every 10 s was selected. The detection limit was 0.2 p.p.m. for each 10-s period.
Quality control

One field blank was taken to the plants per day for every 10 particulate samples, with at least one blank per day. The average mass change of 1-day blanks were subtracted from the mass change for samples collected that day. The quality control procedures for the gravimetric measurements also included measuring two weights, at the beginning of each weighing session. The Norwegian Metrology Service calibrated the balances annually. The response factors of the electrochemical sensors were calibrated before each sampling period with calibration gas obtained from Hydrogass Norge AS, Oslo, Norway. Crystalline silica analyses were controlled by participation in an inter-laboratory proficiency-testing programme (Grunder, 2003).

Data analysis

Using cumulative probability plots, the exposure data were found to be best described by lognormal distributions and were log10 transformed before the statistical analyses. Standard measures of central tendency and distributions [arithmetic mean (AM), geometric mean (GM) and geometric standard deviation] were calculated. The GM was also calculated using mixed effect models, as was the 95th percentile. The mixed effects models were constructed with the exposure as the dependent variable. Exposure determinants were treated as fixed effects, whereas worker was treated as random effects. For sulphur dioxide measurements the highest value recorded for a 10-s averaging period within a work shift was registered as the maximum peak value.

Values below the limit of detection were treated as follows: readable values above the background noise level were directly applied in calculations and modelling, while non-readable values were substituted with the lowest readable value divided by the square root of two (Eduard, 2002).

The significance of differences in exposure levels among the job groups and plants was evaluated using post hoc tests with Bonferroni adjustment. In order to investigate whether the short-term samples were representative of full-shift exposure, we calculated the ratio of the adjusted GMs of short-term and full-shift total dust samples for each job group in all three plants.

The software package SPSS version 15.0 for Windows (SPSS Inc., Chicago, IL, USA) and SAS version 9.1 (SAS Institute Inc., Cary, NC, USA) was used for statistical analysis.

RESULTS

All measurements were carried out between November 2002 and December 2003. Most of the workers (77%) were monitored on more than one occasion. Results are shown in Tables 2–9 and Figs 2 and 3, and the GM referred to in the following text is the GM adjusted with mixed effect models.

Fibres

Most of the fibre sampling (90%) was initiated during the first half of the shift due to practical considerations. A total of 40% of the samples were below the detection limit. However, the fibre count was zero in only 9% of the samples. The fibre exposure levels are shown in Fig. 2 and Table 2. Highest GM of fibre exposures was found in the furnace and sorting areas in all plants, and the cleaning operators in Plant A had the highest exposure to fibres (2.7 fibres cm−3). The control room, cleaning and sorting operators in Plant A, sorting operators in Plant B and charger, charger/mix, payloader and sorting operators in Plant C had all GM exposures of 0.1 fibres cm−3 or more. The sorting operators had significantly higher exposure in Plant C compared to the other plants (P < 0.05). Control room operators had significantly higher exposure to fibres in Plant A compared to the two other plants (P < 0.05). The Norwegian occupational exposure limit (OEL) for SiC fibres of 0.1 fibre cm−3 was exceeded by 53% of the samples from the furnace department and 17% of the samples from the maintenance department (The Norwegian Labour Inspection Authority, 2007). Only 0.2% of the samples from the processing department exceeded the OEL.

Crystalline silica

The cleaning operators in Plant A had the highest GM exposure to respirable quartz (20 μg m−3). The GM exposure of the mix operators in Plants A and C and charger/mix and charger operators in Plant C varied between 13 and 8.0 μg m−3, while all other job groups had a GM exposure of <5 μg m−3 (Table 3). The sorting operators in Plant C had a significantly higher exposure to quartz than the sorting operators in the two other plants (P < 0.05). The quartz exposures were generally low and <1% of the samples exceeded the OEL of 100 μg m−3 (The Norwegian Labour Inspection Authority, 2007). The samples exceeding the OEL were all from the maintenance department in Plant B.

The job group exposed to the highest levels of respirable cristobalite was the charger/mix operators in Plant C (GM = 35 μg m−3) (Table 4). GM exposures >10 μg m−3 were found among the cleaning operators in Plant A, sorting operators in Plants B and C and the mix, charger and payloader operators in Plant C. The mix operators, crane and sorting operators had significantly lower exposure in Plant A compared to the two other plants (P < 0.01). The OEL of 50 μg m−3 was exceeded in 2.1% of the samples (The Norwegian Labour Inspection Authority, 2007).

The crystalline silica exposures for workers in the processing department and in the maintenance department were generally low. More than 90% of the cristobalite samples from each of these departments were below the detection limit and the corresponding results for quartz exposure levels in these departments were 65 and 58%, respectively.

Respirable non-fibrous SiC

The highest GM exposure to respirable non-fibrous SiC was found among the crusher operators (GM = 0.39–0.65 mg m−3) and the cleaning operators in Plant A (GM = 0.49 mg m−3) (Table 5). The mix, sorting and fines operators had a significantly lower exposure in Plant A compared to the two other plants (P < 0.05). The control room operators had a significantly lower exposure in Plant C compared to the two other plants (P < 0.001). Norway does not have a specific OEL for respirable non-fibrous SiC. The American Conference of Industrial Hygienists has recommended a threshold limit value of 3 mg m−3 (ACGIH, 2007) and only 0.4% of the measurements exceeded this limit.

Quantified crystalline components

The respirable dust in the furnace department contained on average 18% SiC, 1.1% quartz and 2.1% cristobalite. In the respirable dust from the processing department, we found 57% SiC, 0.2% quartz and 0.1% cristobalite, while the respirable dust in the maintenance department contained 21% SiC, 0.5% quartz and 0.2% cristobalite.

Respirable dust

The respirable dust exposure levels are shown in Fig. 3 and Table 6. The highest GM exposures to respirable dust (>0.5 mg m−3) were found among cleaning operators in Plant A, mix, charger, charger/mix and sorting operators in Plant C and crusher operators in all plants. When comparing plants, sorting operators in Plant A had significantly lower exposure levels than the sorting operators in the other plants (P < 0.05) and the mechanics in Plant B had a significantly higher exposure than the other two plants (P < 0.05). The Norwegian OEL for mixed respirable dust is 0.5 mg m−3 in the furnace department and furnace-related areas of the SiC industry (The Norwegian Labour Inspection Authority, 2007). A total of 26% of the samples from the furnace department and 15% of the samples from maintenance workers performing maintenance in the furnace department exceeded this limit. Since there is no specific OEL for respirable dust in the processing department, the OEL for nuisance respirable dust of 5 mg m−3 was applied here (The Norwegian Labour Inspection Authority, 2007). Only 0.4% of the samples in the processing department and none of the samples from maintenance workers performing maintenance work in the processing department exceeded this limit.

Total dust

Cleaning and crusher operators in Plant A and charger, charger/mix and sorting operators in Plant C had GM exposures to total dust of 4 mg m−3 or higher (Table 7). The cleaning operators had four times higher GM exposure than any of the other job groups, but this observation was based on only two measurements. The sorting operators had a significantly lower exposure in Plant A compared to Plants B and C (P < 0.001) and the crane operators in Plant B had a significantly higher exposure compared to crane operators in the two other plants (P < 0.001). As there is no specific OEL for total dust in the SiC industry, the Norwegian OEL for nuisance total dust of 10 mg m−3 was applied here (The Norwegian Labour Inspection Authority, 2007). This OEL does not take into account that the dust might contain harmful components. Four per cent of the samples were above the OEL and two-thirds of these were from the fines or refinery crusher operators.

Table 8 summarizes the total dust short-term exposure levels. The job group GM ratios of short-term to full-shift total dust samples varied from 0.19 to 1.4 in Plant A, 2.3 to 0.55 in Plant B and 0.4 to 1.2 in Plant C. However, the GM ratios for the separate plants were close to 1; 0.9 for Plant A and C and 1.0 for Plant B.

Sulphur dioxide

Sulphur dioxide exposure was assessed for job groups in the furnace hall. The mean exposure to sulphur dioxide over a full shift was generally low. The charger and charger/mix operators had the highest measured GM (0.37 p.p.m.), which is one-fifth of the OEL of 2 p.p.m. (Table 9) (The Norwegian Labour Inspection Authority, 2007). The highest GM for maximum peak value was found among the control room operators in Plant B (13 p.p.m.) (Table 9). The crane and control room operators in Plant B had a significantly higher maximum peak value compared to the same job groups in the other plants (P < 0.05). The sorting operators in Plant A had a significantly lower exposure than the sorting operators in the two other plants (P < 0.001).

Use of respirators

Respirators were available for all workers. The use of respirators was mandatory for workers in the furnace hall and for some operations in the refinery department (e.g. packing and cleaning). Different types of respirators were used [i.e. disposable half-masks with P2 or P3 particulate filters, half-masks with particulate filter and gas filter for acid gases (SO2), powered air-purifying respirators, compressed air-fed respirators and self-contained breathing apparatus (when concentrations of CO were high)]. Most of the workers (74%) reported using respirators some or all of the time during the sampling. The use of respirators varied between plants, and 79% of the workers in Plant C used respirators all or some of the time compared to ∼50% of the workers in Plants A and B. The use also varied within plants with 78% of the workers in the furnace department using respirators some or all of the time compared to 46% in refinery and maintenance departments. When measurements exceeded the OEL, between 79 and 100% of the workers had used respirators some or all of the time depending on component. The GM exposure for total dust was 79% higher among workers using respirators all the time, and 65% higher among workers using respirators some of the time compared to workers not using respirators. Similar trends were seen for all other agents.

Work conditions

Ninety-three per cent of the workers reported their perception of the work conditions of the shift. Of these, 84% reported that the conditions were normal, 6% reported it to be worse than normal and 10% reported better than normal.
