Ann. occup. Hyg., Vol. 48, No. 2, pp. 159-170, 2004
© 2004 British Occupational Hygiene Society
Published by Oxford University Press
Exposure to Sulfuric Acid in Zinc Production
1 University of Bergen, Section for Occupational Medicine, Department of Public Health and Primary Health Care, Kalfarveien 31, N-5018 Bergen; 2 Outokumpu Norzink AS, N-5750 Odda, Norway
Received 20 May 2003; in final form 28 August 2003
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This study characterized workers’ exposure to sulfuric acid in two cell houses of a zinc production plant. We also aimed at estimating previous exposure to sulfuric acid by simulating the process conditions from before 1975 to produce exposure data for an epidemiological study on cancer in this industry. Further, we compared different sampling methods for aerosols in the cell houses. Personal sampling with a 37 mm Millipore cassette showed that the geometric means of the exposure levels for the workers in the two cell houses were 0.07 mg/m3 (range 0.01–0.48 mg/m3) and 0.04 mg/m3 (range 0.01–0.15 mg/m3). Norway’s newly revised limit value of 0.1 mg/m3 was exceeded in 39.0 and 12.9% of the samples in the two cell houses. After the foam layer was removed from the electrolyte surface to simulate the production process from before 1975, the concentration of sulfuric acid increased from 0.11 to 6.04 mg/m3 in stationary measurement by the Millipore sampler. Stationary sampling showed that the Millipore sampler and the inhalable fraction of the Respicon® impactor underestimated the sulfuric acid concentration by factors of 1.5 and 2.1 compared with the Institute of Occupational Medicine (IOM) sampler. Sampling with the Respicon impactor showed that the respirable, tracheobronchial and extrathoracic fractions constituted 3.0, 18.7 and 71.7% of the inhalable sulfuric acid aerosol, respectively. Today’s exposure levels are lower than those reported to be associated with an increased prevalence of laryngeal cancer in other industries, but the levels prior to 1975 seem to have been much higher. By mass, most of the inhalable aerosol was in the size fractions considered to be highly relevant for the effects of sulfuric acid on the respiratory system. The risk of cancer among the cell house workers should be investigated in an epidemiological study.
Keywords: acid mist; acid aerosol; exposure assessment; Respicon; sulfuric acid; zinc
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Sulfuric acid irritates the mucous epithelium of the respiratory system (Steenland, 1997), decreases the mucociliary clearance rate of the bronchi at 0.1 mg/m3 (Leikauf et al., 1984) and reduces the activity of pulmonary macrophages at concentrations >0.75 mg/m3 (Zelikoff et al., 1994). Further, epidemiological studies have concluded that occupational exposure to sulfuric acid is associated with an increased risk of laryngeal cancer (Sathiakumar et al., 1997). Thus, the tracheobronchial and extrathoracic fractions of sulfuric acid aerosols are presumably of greatest importance for the biological effects on the respiratory system.
The workers in the electrolytic cell houses of the Norwegian zinc production plant investigated are exposed to sulfuric acid originating from the electrolyte, which mainly consists of a solution of zinc sulfate and sulfuric acid. Gas bubbles bursting at the electrolyte surface produce an acid mist. Since 1975, a surfactant that forms a foam layer on the electrolyte surface has been added to reduce aerosol production. Because of its low vapor pressure and hygroscopic properties, sulfuric acid is mainly present as a mist or aerosol in the working atmosphere. In the zinc industry, the only published study on personal exposure to sulfuric acid is from Finland, where the exposure level was in the range 0.5–1.0 mg/m3 (Roto, 1980). The average exposure to sulfuric acid mist in industrial processes such as metal pickling, electroplating and other acid treatments of metal frequently exceeds 0.5 mg/m3, whereas the levels are usually lower in the manufacture of lead–acid batteries and the production of phosphate fertilizer (IARC, 1992).
Norway’s recommended limit value for sulfuric acid was recently reduced from 1.0 to 0.1 mg/m3 (Norwegian Directorate of Labour Inspection, 2001). Consequently, the industry needs a strategy for more accurate exposure assessment. In Norway, the closed-face Millipore cassette is normally used for sampling ‘total’ aerosol to compare with recommended limit values. However, personal sampling in field studies has shown that the Institute of Occupational Medicine (IOM) cassette, which is designed to collect the inhalable, health-related fraction of the aerosol, collects 1.5 to 3.0 times as much aerosol as the Millipore sampler (Lidén et al., 2000). However, no comparison between these samplers has been published for sulfuric acid aerosol. Laboratory tests have shown that the IOM sampler follows the sampling convention for inhalable dust better than the Millipore sampler (Kenny et al., 1997) and is therefore often used as a reference sampler for the inhalable aerosol fraction. For size-selective sampling, the Respicon® virtual impactor has shown reasonable accuracy in collecting respirable, thoracic and inhalable particles (Li et al., 2000; Koch et al., 2002; Tatum et al., 2002). This also enables the tracheobronchial and the extrathoracic fractions of the aerosol to be calculated, which seems to be especially relevant for the carcinogenic effects of sulfuric acid.
In samples of workplace air in cell houses of the zinc industry, sulfuric acid cannot be separated from the sulfates of the metal elements by routine methods. In such work environments, a correction factor has been proposed to estimate the sulfuric acid in the air samples by assuming that the sampled aerosol has the same ratio of sulfuric acid to total sulfate as the emission solution (Krämer et al., 2002).
The main objective of this study was to characterize the exposure to sulfuric acid in the cell houses of a zinc production plant, to produce exposure data for an epidemiological study on cancer in this industry. Within this context, we also aimed at estimating previous exposure to sulfuric acid by simulating the process conditions from before 1975 by removing the foam layer from the cells. Different sampling methods were also compared. Particle size-selective sampling was done to determine the proportion of sulfuric acid aerosol in the extrathoracic and tracheobronchial fractions, which are considered to be most relevant for the described effects on the respiratory system.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Zinc plant
Outokumpu Norzink AS is located in western Norway and employs about 350 workers. This zinc production plant was established in 1929 and underwent gradual modernization until 1986. It now has two roasting plants, one leaching department, two electrolytic cell houses and one furnace. The nominal yearly capacity is 150 000 tons of zinc.
Cell houses
Cell house series IV (S-IV), from 1970, is 42 m long, 35 m wide and 5.6–7.4 m high, and consists of 144 electrolytic cells divided into 12 cascades, each having 12 cells. The individual cell has 40 Al cathode plates and 41 Pb anode plates. S-IV has natural ventilation through openings in the roof.
Cell house series V (S-V), from 1986, is 102 m long (north end, 33 m), 26 m wide (north end, 39 m) and 6 m high and has mechanical updraft ventilation of 2 000 000 m3/h. This cell house has two rows of cells with 44 cells in each row, 100 Al cathodes and 101 Pb anodes.
Electrolyte composition
The electrolyte in both cell houses is composed of sulfuric acid (175–200 g/l) and sulfates of zinc (48–52 g/l), manganese (5–15 g/l), magnesium (7–10 g/l) and sodium (2–3 g/l). Other metallic elements, such as calcium, cobalt, nickel, cadmium, antimony and thallium, are present in lower concentrations. Total sulfate is 300–330 g/l and chloride is 0.15–0.30 g/l. The temperature of the electrolyte is 37°C. In order to form a foam layer to reduce aerosol production, a soap surfactant (20 mg/l) is added to the electrolyte in S-IV, whereas in S-V a licorice extract (1 mg/l) is added.
Work shifts and practices
Cell house S-IV
The two main job groups in S-IV are the Zn strippers (n = 20) and the Mn workers (n = 6). The Zn strippers are organized into five teams of four strippers, and there are three daily shifts. They start by loading cathode plates from the electrolytic cell onto a transport carriage, which is sent to two automatic stripping machines located in a hall next to the cell house. Each machine is operated by one of the strippers. Then the stripped cathode plates are returned to the cell house. When the Zn plates are stripped off, they fall through an opening in the floor onto a conveyor belt for transport to the storage site before entering the smelter. This storage site, which is located under the cell house, is operated by one of the strippers. The Zn strippers rotate between these job tasks through the shift.
The Mn workers in S-IV are organized in one team working only day shifts. They have four different tasks and do not normally rotate between these jobs: cell cleaning (two workers), anode and cathode replacement (two workers), truck driving (one worker) and anode cleaning (one worker). When the anodes and cathodes are replaced, the lead anodes are pulled out from the cells. Then the truck driver transports the anodes to the satellite, next to the cell house. In the satellite, the anode cleaning comprises operation of the anode maintenance machine, which cleans the anode by using high pressure spraying to remove the deposited MnO2. Cell cleaning in the cell house implies removing the Mn sludge from the electrolytic cells by a vacuum system. Following this operation, cell cleaning also involves adjusting the distance between the anode and the cathode before the electrolytic process starts.
Cell house S-V
In S-V the Zn strippers (n = 20), Mn workers (n = 10) and inspectors (n = 5) are organized into five teams each comprising four Zn strippers, two Mn workers and one inspector. There are three daily shifts.
The Zn strippers operate the two cathode cranes, which pull the cathode plates from the cell, and the two automatic Zn stripping machines. The strippers also change damaged cathode plates. One of the strippers drives the truck on the feeding floor under the cell house. He receives Zn plates from the elevator to feed the smelter or for storage. The four Zn strippers in each shift rotate between these job tasks about every hour.
The tasks for the two Mn workers in each shift are either cell cleaning or anode cleaning. During anode cleaning, which is normally carried out during the afternoon and night shifts, the Mn workers operate the anode crane. MnO2 is removed from the anodes by high pressure spraying, and damaged anodes are repaired or changed. Cell cleaning is done mainly during day shifts by pumping the Mn sludge from the electrolytic cells by a vacuum system.
The inspector supervises the processes in both cell houses and most of the time stays in the control room of S-V and in other areas outside the cell houses.
The workers in the cell houses do not normally use personal respiratory protection.
Personal sampling
Personal sampling of aerosol (n = 135) was carried out for inspectors, Zn strippers and Mn cleaners in the two cell houses from 14 to 18 and 26 to 29 November 2001. For each of the job groups, at least three sets of full shifts were sampled for the complete day, evening and night shifts (Table 1). For each worker, the number of hours spent inside the respective cell houses was recorded by interview after the shift. According to the inspectors, production was normal during the sampling periods. The outdoor temperature was between 0 and 9°C, and it was mostly cloudy, light wind or breeze, with some showers or rain.
|
Aerosol samples were collected on Fluoropore membrane filters (Type 1.0 µm FA, Millipore catalog no. FALP03700) placed in 37 mm closed-face Millipore samplers, using a flow rate of 2.0 l/min. The sampling filters were placed in the breathing zone of the workers, and the sampling time was 5.5–7.5 h. The sampling time was considered representative for the whole work shift. The flow was controlled before and at the end of the sampling period. Six pumps stopped during sampling and these samples were eliminated from the study. Samples of the electrolyte were taken on the days of personal monitoring.
Stationary sampling
Parallel sampling with a closed-face Millipore sampler, IOM sampler and Respicon virtual impactor was carried out simultaneously in six positions along the walkways of the operators, at a height of 1.5 m in cell house S-V (13 March 2002, sampling time 8 h) and in cell house S-IV (14 and 15 March 2002, sampling time 4.5 h each day). In S-IV two sets of samplers were placed along the central walkway, two sets were located at the rear end of the cell house and two sets were placed at the front end of the cell house, i.e. between the cells and the transport system. In cell house S-V three sets of samplers were placed along the central walkway, two sets were situated between the central walkway and one of the sidewalls and one set was placed on the stripping machine.
In S-V and on the first day of measurement in S-IV, the production conditions were normal, whereas on the second day in S-IV the foam layer on the electrolyte surface was removed. The management did not allow more than 5 h of production with the foam completely removed. Thus, the sampling time in S-IV was correspondingly reduced compared with S-V. In addition to these three main sampling days, results were also obtained from two sets of parallel sampling in S-IV performed on a day when the foam layer was reduced (19 March 2002, sampling time 7 h). On this day the two sets of samplers were placed on the central walkway and at the rear end of the cell house, respectively.
The weather conditions on the days of stationary sampling were 2–5°C, clear skies and calm. Samples of the electrolyte were taken from the respective cell houses on the days of stationary sampling.
The Millipore samplers were prepared as described previously for personal sampling. The IOM samplers had 25 mm Fluoropore membrane filters (Type 1.0 µm FA, Millipore catalog no. FALP02500) and sampling was done at a flow rate of 2.0 l/min. The Respicon samplers had 37 mm Fluoropore membrane filters (Type 1.0 µm FA, Millipore catalog no. FALP03700) and the pump flow rate was 3.1 l/min. Flow in all samplers was controlled before and at the end of the sampling period. The accelerating nozzles and the annular slit inlet of the Respicon sampler were cleaned with pressurized air between each series of measurement.
The Respicon three-stage sampler is designed to simultaneously collect the three aerosol fractions defined and agreed upon by the International Organization for Standardization (ISO), the European Committee for Standardization (CEN) and the American Conference of Governmental Industrial Hygienists (ACGIH) (Koch et al., 1999). The aerosol mass on the three filters in the Respicon allows the determination of aerosol concentration for the respirable (CI, cut-point 4 µm), thoracic (CT, cut-point 10 µm) and inhalable (CR, cut-point 100 µm) fractions, as given by Li et al. (2000):
Mass1–3 represent aerosol mass on the top filter (1), middle filter (2) and bottom filter (3) and t represents sampling time. Q1 is the sampling flow rate for the top filter (2.66 l/min), Q2 is the sampling flow rate for the middle filter (0.33 l/min) and Q3 is the sampling flow rate for the bottom filter (0.11 l/min).
The tracheobronchial fraction (CTB) was calculated by subtracting the respirable fraction from the thoracic fraction. The extrathoracic fraction (CET) was calculated by subtracting the thoracic fraction from the inhalable fraction. The manufacturer’s instructions suggest a factor of 1.5 for calculating the corrected extrathoracic fraction (CET corr) to account for sampling biases affecting very coarse particles (TSI, 1999):
This correction also implies a corrected value for the inhalable fraction (CI corr):
The results systematically showed outlier values for the bottom filter (stage 3) for two particular Respicons in all three series of parallel samplings when they were used, resulting in sulfuric acid concentrations close to zero on this filter. The percentages of the extrathoracic fraction to the inhalable aerosol were 0.0, 1.6, 2.7, 6.0, 7.2 and 10.4% for these two samplers, whereas for the other four Respicon samplers it was between 50.0 and 89.5% (n = 14). The total flow for the two Respicons was not reduced, but the flow could not be checked on the individual stages after sampling started. These results indicate that stage 3 of these two Respicons was clogged after a short period of sampling. Because these two Respicons deviated systematically in the three different sets of measurements, the results from these two Respicons were eliminated from the study.
The results from the three parallel samplers were compared by linear regression using the relations:
CIOM, C37, CI and CI corr represent the sulfuric acid concentrations measured by the IOM cassette, the 37 mm Millipore cassette and the uncorrected and the corrected inhalable fraction of the Respicon impactor, respectively. S37, SI and SI corr are the slopes of the regression lines between the IOM sampler and the 37 mm Millipore sampler and between the IOM sampler and the uncorrected and the corrected inhalable fraction of the Respicon, respectively.
In such parallel sampling, the highest values will contribute more to the regression than the lowest values. The data were therefore estimated using weighted least squares to ‘stabilize’ the variance across the range of the data set as recommended by Tsai and Vincent (2001). The intercepts of the regression lines on the y-axis were found to be non-significant for both weighted and unweighted sets of data. By assuming a linear regression line through the origin, these S values were used as conversion factors to convert measured values from the Millipore or the Respicon samplers to corresponding values for the IOM sampler.
Analysis of samples
Total sulfate on the filters was analyzed based on the US Occupational Safety and Health Administration ID-113 method (OHSA, 1991). The sampling filters were desorbed with 5 ml of ultrafiltrated, ion-exchanged water, shaken and treated with ultrasound for 10 min. The sulfate concentration on the filters and the electrolyte was determined with a Dionex DX-120 Ion Chromatograph with Dionex IonPac AS14 column and conductive detector. A solution of 3.5 mM Na2CO3 and 1.0 mM NaHCO3 was used as the eluent. The total sulfate from the air sample filters and from the electrolyte was determined by a calibration curve with standards of 0.2, 1.2, 10, 20 and 40 mg/l. The detection limit for the analytical method for sulfate was 0.1 mg/l, corresponding to 0.5 µg sulfate on the filter. The uncertainty of the sulfate analysis at the measured levels was <10%. Three blind filters of 37 mm and three blind filters of 25 mm were also analyzed.
The concentration of sulfuric acid in the electrolyte was determined by titration with sodium carbonate, and methyl orange (2%) was used as the indicator.
In agreement with Krämer et al. (2002), it was assumed that the sampled aerosol had the same ratio of sulfuric acid to total sulfate as the emission solution (electrolyte). For each sampling day a correction factor (K) was determined as the ratio of the concentration (C) of sulfuric acid to total sulfate in the electrolyte: K = Csulfuric acid in electrolyte/Ctotal sulfate in electrolyte.
The arithmetic mean values for K were 0.609 (SD 0.021) for the 8 days of personal sampling in S-V and 0.610 (SD 0.023) for the 9 days of personal sampling in S-V.
The correction factor for the respective sampling days in the cell houses was used to calculate the concentration of sulfuric acid in the air samples taken: Csulfuric acid in air sample = K x Ctotal sulfate in air sample.
Statistics
The SPSS program (SPSS Inc., Chicago, IL) was used for statistical analysis. Groups were compared using t-tests and analysis of variance on log-transformed data, since cumulative probability plots showed that the data were best described by a log-normal distribution. The results are given as geometric means (GM) and geometric standard deviations (GSD). Linear regression analysis was carried out to determine the relationships between variables. Significance was set at P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Personal sampling with Millipore cassettes
In S-IV the GM for the exposure to sulfuric acid was 0.07 mg/m3 (Table 2). There was no significant difference in the level of exposure between the Mn workers and the Zn strippers (P = 0.47). Among the Mn workers, the cell cleaners and the anode and cathode replacers were the most highly exposed job groups, but exposure did not exceed that of the Zn strippers (P = 0.59). A total of 39% of the samples taken in cell house S-IV exceeded Norway’s limit value of 0.1 mg/m3. The highest value detected was 0.48 mg/m3 for one of the Zn strippers.
|
The exposure to sulfuric acid in S-V was lower than in S-IV (P = 0.002) (Table 2). This was also true when the inspectors, the job group with the lowest exposure in S-V, were omitted from the test (P = 0.029). In this cell house, the Mn workers had significantly higher exposure than the Zn strippers (P = 0.006). The Mn workers had significantly higher exposure when performing cell cleaning than anode cleaning (P = 0.013). For the Zn strippers, exposure was lower during the day shift (GM 0.03, n = 15) than the evening and night shifts (GM 0.07, n = 12; GM 0.06, n = 12) (ANOVA, P < 0.001). For the other main job groups exposure did not differ between the shifts in any of the cell houses. Only 13% of the samples in S-V exceeded Norway’s limit value of 0.1 mg/m3, and the highest value detected was 0.15 mg/m3.
The sulfuric acid exposure and time spent inside the respective cell houses were significantly correlated (Table 3). The truck drivers (n = 3) in S-IV were omitted because they could not estimate how long they had been inside the cell house. The regression equations given in Table 3 show that the predicted time to reach Norway’s limit value (0.1 mg/m3) would be 
2 h and 33 min in cell house S-IV and 5 h and 24 min in cell house S-V. The regression model also predicts that, at 0 h spent in the cell houses, the exposure to sulfuric acid would be 0.05 and 0.02 mg/m3 in S-IV and S-V, respectively. Further, for 8 h inside the cell houses, the predicted exposure in S-IV and S-V would be 0.53 and 0.21 mg/m3.
|
Stationary sampling
Under normal conditions in the cell houses the concentration of sulfuric acid was slightly higher (but not significantly different) in S-IV than in S-V when measured by IOM, Millipore or Respicon samplers (Table 4). After the foam layer was completely removed in S-IV, the concentration of sulfuric acid in the air samples increased considerably (Table 4). The Millipore sampler measured a 55-fold increase in concentration, from 0.11 to 6.04 mg/m3 (Table 4).
|
When all sets of data were included (Table 4), the IOM sampler showed higher concentrations of sulfuric acid than the Millipore sampler (paired t-test, P < 0.001, n = 20) or the uncorrected inhalable fraction of the Respicon sampler (P < 0.001, n = 14). The Millipore sampler showed higher values than the uncorrected values of the Respicon sampler (P < 0.001, n = 14). The corrected results of the Respicon sampler differed significantly from the results of the IOM sampler (P = 0.011, n = 14). The variability in the parallel sets taken by the respective samplers was relatively low, as expressed by the GSD (Table 4).
Linear regression with the line forced through the origin on the complete data set showed significant correlations (P < 0.001) between the results of the IOM and the Millipore samplers as well as between the IOM and the Respicon samplers (Fig. 1 and Table 5). Conversion factors of 1.48 and 2.12 were found between the IOM and the Millipore samplers and between the IOM and the uncorrected values of the Respicon sampler, respectively (Table 5). By using the weighted least squares function of SPSS, the most appropriate weightings were found to be 1/(Millipore)2 or 1/(Respicon)2 for correlating the IOM sampler with the Millipore and Respicon samplers, respectively. The regression analysis also showed an association (P < 0.001) between the results of the IOM and the Millipore samplers and between the results of the IOM and the Respicon samplers for the weighted data (Table 5).
|
|
Particle size-selective sampling with Respicon shows relatively small differences in the ratio of the respirable, the tracheobronchial or the extrathoracic fractions to the inhalable fraction of sulfuric acid between the four sets of parallel sampling (Table 6). The different sets were not compared statistically because each sampling set had few samples. When all data sets of the Respicon sampler were pooled together, the geometric means of the respirable, tracheobronchial and extrathoracic mass fractions were 3.0, 18.7 and 71.7% of the inhalable sulfuric acid, respectively (Table 6 and Fig. 2).
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
For several of the job groups in zinc production the personal exposure to sulfuric acid for cell house workers was close to Norway’s recently revised recommended limit value of 0.1 mg/m3 when sampled with Millipore cassettes. The exposure was highest for the workers in the oldest cell house (S-IV). Among the Mn workers in S-IV, the cell cleaners and the anode and cathode replacers (four workers in each shift) had the highest exposures, with GM of 0.09 and 0.11 mg/m3. For these job tasks, a considerable fraction (one-third and two-thirds) of the samples exceeded the limit value. The Zn strippers in S-IV were not less exposed than the Mn workers even though they spent most of their time in the stripping room next to the cell house. The gate between these rooms was always open, presumably leading to considerable aerosol transport between the cell house and the stripping room. The relatively high value for the predicted exposure at 0 h spent inside S-IV also indicates that a considerable fraction of their total exposure is obtained in rooms next to the actual cell house. In the newest cell house (S-V), from 1986, the exposure was generally lower. However, the cell cleaners had a GM of 0.08 mg/m3, and
43% of the samples were over the limit value for this job group. Removing the foam from the cells in S-IV to simulate the working environment before 1975 was expected to increase the concentration of sulfuric acid in the workplace air considerably. In cooperation with the workers, the management decided that the employees should not stay inside this cell house when the foam was removed. Thus, only stationary area sampling was possible under these conditions.
Measurements with a 37 mm Millipore cassette showed that after the foam was removed, the concentration of sulfuric acid increased from 0.11 to 6.04 mg/m3, which is 60 times higher than Norway’s current limit value for 8 h personal exposure. The IOM sampler and the Respicon impactor mounted in parallel with the Millipore sampler found similar increases. The foam layer appears to reduce aerosol emission by 98%, which is the same reduction as previously shown in laboratory studies after a surfactant was added to an analogous electrolyte system (Van Dusen and Smith, 1989). In the present study, the relatively low variability within each of the samplers both before and after foam removal indicates that the aerosol was quite evenly distributed in the cell house. The area sampling performance of the Respicon matches the inhalable convention fairly well (Li et al., 2000). However, the 37 mm Millipore and the IOM samplers were designed as personal samplers and depend more on such factors as wind speed, wind direction and particle size and are considered to be less appropriate for stationary sampling (Li et al., 2000). During normal conditions in the respective cell houses, the time-dependent regression equations for personal exposure predict that, for 8 h work inside S-V and S-IV, the exposure to sulfuric acid would be 2.4 and 4.8 times higher than the levels monitored by stationary sampling using the Millipore sampler. Although care should be taken in comparing stationary and personal sampling, these results indicate that personal exposure before 1975 was likely to be at least as high, as indicated by the stationary samples when the foam was removed. Other modifications in the production process since 1975 might have contributed to the reduction in aerosol emission, but the process engineers considered the addition of foam to be the main factor. They also stated that the workers did not spend less time within the cell houses at that time.
Because of simultaneous exposure to sulfates and sulfuric acid, an indirect method was used to estimate the concentration of sulfuric acid in air samples. One alternative method of determining the concentration of sulfuric acid in the air samples could have been to analyze the filter for total sulfate and for the major metallic elements followed by subtraction of the sulfate associated with these elements from the total sulfate detected on the filter. If the elements had no source other than the electrolyte, this ‘remainder sulfate’ would represent sulfuric acid. However, in this work environment other processes, such as Zn stripping, are a probable source of elements, and this method was rejected. Hethmon and Ludlow (1997) tried to differentiate between the sulfate bound in sulfuric acid from that found in sulfate salts on the sampling filters by extracting the sulfuric acid in isopropanol. Although promising results were found, they concluded that the method needs to be validated further.
Our method for analyzing and calculating the concentration of sulfuric acid in the samples was based on the assumption made by Krämer et al. (2002) that the sampled aerosol has the same ratio of sulfuric acid to total sulfate as the electrolyte. The results from sampling with an Anderson impactor at a height of 1.5 m over zinc electrolysis cells showed that >50% of the aerosol particles by mass exceeded 7 µm and 90% were larger than 3 µm (Rondia and Closset, 1985). These authors reported that the composition of the aerosol was not fully comparable with the composition of the electrolyte. They found an overall ratio of sulfuric acid to total sulfate of 0.36 in the aerosol and 0.42 in the electrolyte. For particles larger than 7 µm this ratio was close to the corresponding ratio in the electrolyte, but the ratio was considerably lower for particles <3 µm. The interpretation of these results is difficult since there is no information on the variability of the results, the number of samples taken or the addition of any foam to the electrolyte surface. However, results from sampling with the Respicon impactor show that by mass the aerosol in the present study mainly consisted of larger particles. If Rondia and Closset’s conclusions for particle composition are valid in our study, the ratio of total sulfate to sulfuric acid in the acid mist is more comparable to the electrolyte than if the aerosol particles were in the respirable size fraction.
In our study, the personal sampling of ‘total’ aerosol was carried out with closed-face 37 mm Millipore samplers with a 4 mm circular orifice. Alternatively, the IOM sampler could have been used to collect the inhalable fraction of the aerosol to come closer to the health-related fraction of the aerosol. However, for personal sampling the IOM sampler was rejected because of the risk of electrolyte splashing directly onto the filter through its 15 mm diameter entry facing horizontally outwards.
The results from the stationary area sampling indicate that the 37 mm Millipore sampler underestimates the sulfuric acid concentration by a factor of about 1.5 compared with the IOM cassette. This factor is comparable to the range of previous observations during personal sampling of aerosol in nickel alloy production (Tsai and Vincent, 2001), lead smelting (Spear et al., 1997) and wood processing (Harper and Muhler, 2002), but higher factors are often found (Lidén et al., 2000). However, since the sampling performances of the IOM and Millipore cassettes are highly dependent on wind orientation and speed, the relationship between these samplers is expected to be different when they are used as stationary compared with personal samplers (Tsai and Vincent, 2001).
The Millipore sampler has a very low sampling efficiency for particle sizes larger than 45 µm (Kenny et al., 1997). Thus, the coarseness of the aerosol strongly determines the difference between the sampling methods (Spear et al., 1997). Many investigators also gravimetrically quantify the aerosol deposited on the internal walls of the IOM sampler, but this is not always done (Demange et al., 2002). Normally, internal wall deposits are not included for the Millipore sampler (Tsai et al., 1996). Kenny et al. (1997) found that the percentage of the sample mass on the filter of IOM samplers decreased from 100% at small particle sizes to 75% at 100 µm. However, they also reported that, at moderate wind speeds (0.5 m/s), the IOM filter deposit alone gave good agreement with the inhalable convention (Kenny et al., 1997). Our study did not take this fraction of the aerosol deposited on the wall of the IOM into account, which might partly explain why the ratio between the IOM and Millipore samplers was in the lower range compared with studies in which both the internal wall and the filter deposits of the IOM sampler were quantified. The Respicon impactor undersampled the inhalable fraction of sulfuric acid by a factor of about 2.1 compared with the IOM sampler in parallel stationary samples. Stabilizing the variance across the range of the data sets to reduce the contribution of the highest concentrations did not produce any significant changes in the regression coefficient between the IOM and Millipore or Respicon samplers in the present study. A correction factor of 1.5 in calculation of the inhalable fraction of the Respicon sampler, as suggested by the Respicon manufacturer and later confirmed for stationary sampling in the forest product industry (Tatum et al., 2002), was not sufficient to comply with the results from the IOM sampler when analyzed by linear regression. Koch et al. (2002) found that a higher correction factor (1.8) was needed to calculate the inhalable fraction in stationary Respicon sampling in a nickel refinery. In contrast, Li et al. (2000) reported that the Respicon sampler provided a reasonable match of the inhalable convention without any correction factor during stationary sampling of monodisperse particles. Apparently, more research is needed to evaluate the sampling properties of the Respicon impactor, including liquid aerosols.
Particle size-selective sampling showed that much of the inhalable sulfuric acid aerosol was in the extrathoracic fraction (71.7%). These results indicate that, by mass, the coarse particles dominate in both cell houses. This was also the case close to sulfuric acid tanks in the titanium dioxide industry, where the extrathoracic fraction constituted 61% of the inhalable sulfuric acid aerosol (Hery et al., 1992). The particles in the extrathoracic fraction are likely to be deposited in the upper airways, which comprise all structures, including the larynx, above the trachea. The tracheobronchial fraction, which constituted 18.7% of the inhalable sulfuric acid, is expected to deposit in the trachea and the bronchial tree as far as the terminal bronchioles and includes the airway with ciliated lining. The respirable fraction of the sulfuric acid aerosol, which may enter the non-ciliated bronchioles and the alveoli, was small in the present study (3%). After foam was removed in S-IV, the respirable fraction of the aerosol decreased from 14.3 to 0.6%, indicating very low emission of the smallest particles from the electrolyte in those process conditions. Relatively small differences were found in the ratio of the respirable, the tracheobronchial or the extrathoracic fractions to the inhalable fraction between cell house S-IV and S-V under normal conditions and between the different process conditions in S-IV, indicating a similar size distribution of sulfuric acid aerosol in the cell houses. However, because few samples were taken in each set of measurements with the Respicon sampler, we cannot exclude potential significant differences in the size distribution of aerosols between the cell houses and between the different process conditions.
Previous authors (Koch et al., 1999; Li et al., 2000) have strongly recommended cleaning the accelerating nozzles after each measurement because of the risk of clogging of the nozzles. Even though this was done in the present study, we had to reject the results from two of the Respicons. Clogging was presumably the reason for the outlier values systematically found for two of the six Respicons. However, we could not determine why these particular Respicons were affected. These problems also indicate that care should be taken in interpreting the results from particle size-selective sampling with the Respicon impactor.
The International Agency for Research on Cancer (IARC, 1992) stated that sufficient evidence existed to determine that occupational exposure to sulfuric acid mist is carcinogenic to humans. Laryngeal cancer associated with exposure to sulfuric acid was reported for workers in steel pickling, soap production and ethanol production (Sathiakumar et al., 1997). Little evidence supports a causal relationship between acid exposure and lung cancer (Sathiakumar et al., 1997). In steel pickling, the average personal exposure to sulfuric acid associated with laryngeal cancer was 0.19 mg/m3, and the average duration of exposure was 9.2 yr (Steenland, 1997). In soap production, the exposure to sulfuric acid was reported to be 0.64–1.12 mg/m3, and the latency time for laryngeal cancer was >10 yr (Forastiere et al., 1987). Lower exposures, like those found in the phosphate fertilizer and lead acid battery industries, have weaker associations with laryngeal cancer (Sathiakumar et al., 1997). Possible mechanisms for the carcinogenic effect of sulfuric acid include irritation of epithelial cells in conjunction with cigarette smoking or a direct genotoxic effect from modification of cellular pH (Steenland, 1997).
Leikauf et al. (1984) reported that concentrations of sulfuric acid similar to those we found during personal sampling in the cell houses might affect the defense mechanisms of the ciliated airways by reducing mucociliary clearance. Personal respiratory protection might be needed to avoid such effects in the cell houses. Today’s exposure levels are somewhat lower than those other authors previously reported to be associated with laryngeal cancer. However, the stationary measurements strongly indicate that the exposure levels prior to 1975 were much higher. Size-selective sampling shows that 18.7% of the inhalable aerosol was in the tracheobronchial fraction, suggesting deposition mainly in the ciliated airways, whereas the major proportion (71.7%) of the inhalable aerosol was in the extrathoracic fraction, suggesting deposition in the upper airways, above the trachea. An epidemiological study should be designed to compare the risk of laryngeal cancer for workers employed before and after the production process changed in 1975.
Acknowledgements—We gratefully acknowledge financial support from the Confederation of Norwegian Business and Industry. We thank the management and the workers at Outokumpu Norzink AS for their cooperation in the study.
|
FOOTNOTES |
|---|
* Author to whom correspondence should be addressed. Tel: +47-55-58-60-73; fax: +47-55-58-61-05; e-mail: magne.bratveit{at}isf.uib.no
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Demange M, Gorner P, Elcabache JM, Wrobel R. (2002) Field comparison of 37-mm closed-face cassettes and IOM samplers. Appl Occup Environ Hyg; 17: 200–8.[CrossRef][Medline]
Forastiere F, Valesini S, Salimei E, Magliola E, Peruzzi CA. (1987) Respiratory cancer among soap production workers. Scand J Work Environ Health; 13: 258–60.[Web of Science][Medline]
Harper M, Muhler BS. (2002) An evaluation of total and inhalable samplers for the collection of wood dust in three wood products industries. J Environ Monit; 4: 648–56.[CrossRef][Web of Science][Medline]
Hethmon TA, Ludlow SK. (1997) Air sampling methodologies for sulphuric acid mist in copper electrowinning: field evaluation and development of a new method. Appl Occup Environ Hyg; 12: 1041–6.
Hery M, Hecht G, Gerber JM, Bemer D, Gorner P. (1992) Exposure to sulphuric acid and sulphur dioxide in the manufacturing of titanium dioxide. Ann Occup Hyg; 36: 653–61.
IARC. (1992) Occupational exposure to mists and vapours from strong inorganic acids; and other industrial chemicals. IARC Monogr Eval Carcinog Risk Chem Hum; 54: 1–130.
Kenny LC, Aitken R, Chalmers C et al. (1997) A collaborative European study of personal inhalable aerosol sampler performance. Ann Occup Hyg; 41: 135–53.
Koch W, Dunkhorst W, Lodding H. (1999) Design and performance of a new personal aerosol monitor. Aerosol Sci Technol; 31: 231–46.[CrossRef]
Koch W, Dunkhorst W, Lodding H et al. (2002) Evaluation of the Respicon as a personal inhalable sampler in industrial environments. J Environ Monit; 4: 657–62.[CrossRef][Web of Science][Medline]
Krämer W, Bender HF, Leuppert G, Fischer P, Gusbeth K, Breuer D. (2002) Messung von schwefelsäure in verschiedenen Arbeitsbereichen. Gefahrstoffe Reinhaltung Luft; 62: 45–51.
Leikauf GD, Spektor DM, Albert RE, Lippmann M. (1984) Dose-dependent effects of submicrometer sulfuric acid aerosol on particle clearance from ciliated human lung airways. Am Ind Hyg Assoc J; 45: 285–92.[Web of Science][Medline]
Li S-N, Lundgren DA, Rovell-Rixx D. (2000) Evaluation of six inhalable aerosol samplers. Am Ind Hyg Assoc J; 61: 506–16.
Lidén G, Melin B, Lidblom A, Lindberg K, Norén JO. (2000) Personal sampling in parallel with open-face filter cassettes and IOM samplers for inhalable dust—implications for occupational exposure limits. Appl Occup Environ Hyg; 15: 263–76.[CrossRef][Medline]
Norwegian Directorate of Labour Inspection. (2001) Administrative levels for pollution in the working atmosphere. Guidelines [in Norwegian]. Oslo: Tiden Norsk Forlag.
OSHA. (1991) Sulfuric acid in workplace atmospheres, ID-113. In: Manual of analytical methods, 2nd edn, Part Two, Vol. 1. Washington, DC: OSHA.
Rondia D, Closset J. (1985) Aerosol versus solution composition in occupational exposures. Sci Total Environ; 46: 107–12.[CrossRef][Medline]
Roto P. (1980) Asthma, symptoms of chronic bronchitis and ventilatory capacity among cobalt and zinc production workers. Scand J Work Environ Health; 6: 1–49.[Web of Science][Medline]
Sathiakumar N, Delzell E, Amoateng-Adjepong Y, Larson R, Cole P. (1997) Epidemiologic evidence on the relationship between mists containing sulfuric acid and respiratory tract cancer. Crit Rev Toxicol; 27: 233–51.[Web of Science][Medline]
Spear TM, Werner MA, Bootland J et al. (1997) Comparison of methods for personal sampling of inhalable and total lead and cadmium-containing aerosols in a primary lead smelter. Am Ind Hyg Assoc J; 58: 893–9.[Web of Science][Medline]
Steenland K. (1997) Laryngeal cancer incidence among workers exposed to acid mists. Cancer Causes Control; 8: 34–8.[CrossRef][Web of Science][Medline]
Tatum V, Ray AE, Rovell-Rixx D. (2002) Performance of the RespiCon personal aerosol sampler in forest products industry workplaces. Am Ind Hyg Assoc J; 63: 311–6.
Tsai PJ, Vincent JH. (2001) A study of workers’ exposures to the inhalable and "total" aerosol fractions in the primary nickel production industry using mannequins to simulate personal sampling. Ann Occup Hyg; 5: 385–94.[CrossRef]
Tsai PJ, Vincent JH, Wahl GA, Maldonado G. (1996) Workers exposures to inhalable and total aerosol during nickel alloy production. Ann Occup Hyg; 40: 651–9.
TSI. (1999) Model 8522 Respicon particle sampler. Operation and service manual. St Paul, MN: TSI Inc.
Van Dusen J, Smith JW. (1989) Comparison of the effectiveness of bubble coalescence and foamed surfactant in controlling the acid mist formed by electrowinning cells. Am Ind Hyg Assoc J; 50: 252–6.
Zelikoff JT, Sisco MP, Yang Z, Cohen MD, Schlesinger RB. (1994) Immunotoxicity of sulfuric acid aerosol: effects on pulmonary macrophage effector and functional activities critical for maintaining host resistance against infectious diseases. Toxicology; 92: 269–86.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  D. H. BROUWER, I. H. M. LINKS, S. A. F. DE VREEDE, and Y. CHRISTOPHER Size Selective Dustiness and Exposure; Simulated Workplace Comparisons Ann. Hyg., July 1, 2006; 50(5): 445 - 452. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. MWAISELAGE, M. BRATVEIT, B. MOEN, and M. YOST Variability in Dust Exposure in a Cement Factory in Tanzania Ann. Hyg., August 1, 2005; 49(6): 511 - 519. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||