Annals of Occupational Hygiene Advance Access originally published online on April 9, 2008
Annals of Occupational Hygiene 2008 52(4):249-257; doi:10.1093/annhyg/men009
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Diffusive Sampling of C7–C16 Hydrocarbons in Workplace Air: Uptake Rates, Wall Effects and Use in Oil Mist Measurements
Health and Safety Laboratory, Harpur Hill, Buxton, SK17 9JN, UK
* Author to whom correspondence should be addressed. Tel: +44 (0)1298 218000; fax: +44 (0)1298 218570; e-mail: andrew.simpson{at}hsl.gov.uk
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONSFUNDINGREFERENCES |
|---|
The measurement of semi-volatile hydrocarbons in workplace air is complicated by their readiness to condense to form aerosols or adsorb on to surfaces. The diffusive sampling and analysis by thermal desorption of alkanes up to hexadecane was investigated with the aim of quantifying vapour from petroleum distillate fractions and possibly differentiating particles from vapour in oil mist measurements of light mineral oil-based metalworking fluids. Diffusive uptake rates were measured on Perkin Elmer thermal desorption tube samplers packed with Tenax TA, and the potential for deposition within the tubes was examined. Hydrocarbon vapour was found to adsorb on the oxide layer that can develop on the sampler's internal walls. General measurements of mixed hydrocarbon vapours (i.e. petroleum distillate fractions) should not be unduly affected if concentrations are greater than
5 mg m–3 and the tubes are in good condition. For the purposes of differentiating light mineral oil mist and vapour from a total hydrocarbon measurement, it is unlikely that measuring the vapour separately could be used to calculate mist concentrations <3 mg m–3 with sufficient accuracy.
Keywords: diffusive • hydrocarbon • oil mist • sampling • semi-volatile
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONSFUNDINGREFERENCES |
|---|
Analysis of solvent vapour by diffusive monitoring and thermal desorption is now a well-established technique, having many advantages over pumped sampling. However, the accuracy of the method depends upon using the correct diffusive uptake rate to calculate the airborne concentration. Theoretical values can be calculated but experimentally determined values are more appropriate. Many compounds have experimentally derived values quoted in the literature (HSE, 1995). However, values for aliphatic hydrocarbons have been published only for alkanes C5–C10. Alkanes are the major component of many petroleum distillate fractions. White spirit contains components from C8 to C12, kerosene contains mostly C10–C16 alkanes and motor gasoline up to C14 alkanes. It would be advantageous to have experimentally determined values for these hydrocarbons.
As molecular mass and boiling point increase, the volatility decreases until a point where the compound may be described as semi-volatile (a boiling point of 240–260°C). Alkanes containing 14 carbon atoms or more could thus be described as semi-volatile organic compounds (SVOCs). For such compounds, airborne vapour may condense and aerosol particles may evaporate (Soderholm, 1988). Measuring airborne concentrations of such substances presents a problem relevant to a wide range of chemicals of interest to hygienists including amines and phthalates (Perez and Soderholm, 1991). It is recommended for such substances that both the inhalable particulate and vapour phases are sampled together (BSI, 2001), and this has been successfully done for polycyclic aromatic hydrocarbons (Scobbie et al., 1998) and hydroquinone (Scobbie and Groves, 1999).
One potential application of diffusive sampling of hydrocarbon vapour could be in the measurement of oil mists containing significant quantities of SVOCs (e.g. in low viscosity oils). Traditional filter sampling of the mist loses oil by evaporation, but with the addition of a backup sorbent tube to collect original and newly evaporated vapour, total airborne hydrocarbon can be determined. However, in the case of oil mist, the particulate phase is considered to be of greater concern than the vapour, even though the vapour is often present in much higher concentrations. One way to resolve this problem is to determine the vapour concentration separately using a thermal desorption diffusion tube sampler in parallel with the total airborne hydrocarbon measurement and determine the mist concentration by difference. There may be a problem in sampling SVOC vapour by diffusion if there is premature condensation of the hydrocarbon on the internal walls of the steel tube sampler. During previous work on methods for quantifying mineral oil mist and vapour mixtures (Simpson, 2003), hydrocarbons in the range C12–C18 were found inside control tube samplers containing no sorbent and exposed to oil mist and vapour. The distribution of the components suggested that the majority of material collected on the internal walls was condensed vapour rather than impacted aerosol droplets.
In this work, experimentally derived uptake rates were determined for alkanes likely to be present in the vapour from petroleum-based products, and the nature and impact of deposition on the internal walls of the tube sampler was investigated. The mechanisms of deposition investigated included condensation and adsorption on residual sorbent in recycled empty tubes and active sites caused by scratches, or directly on the sampler walls.
|
EXPERIMENTAL |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONSFUNDINGREFERENCES |
|---|
Diffusive uptake rate
The diffusive uptake rates for n-alkanes in the range C7–C16 on to Perkin Elmer automated thermal desorption (ATD) tubes (89 x 6.4 mm od, 5.0 mm id) packed with 230 mg Tenax TA were determined by sampling dynamic test atmospheres of single compounds at 20°C and 50% relative humidity. The Tenax TA sorbent is held behind a metal gauze 14 mm from the end of the tube (the diffusion gap). During diffusive sampling, the tubes were fitted with standard diffusion caps (without membrane) which incorporate a metal grid. The grid defines the end of the diffusion gap and reduces wind turbulence within the air gap. The alkane vapour was introduced using the syringe injection technique (HSE, 1990), with the addition of electrical heating tape wound around the apparatus to aid vaporization. The vapour produced (5 l min–1) was diluted with a mixture of wet and dry filtered air to its final concentration (20 l min–1) which was fed into a 14-l temperature-controlled glass chamber. The concentration was set at approximately 170 mg m–3 for C7–C10 alkanes and at
2.5% of the saturated vapour concentration (SVC) for individual C10–C16 alkanes, to prevent condensation within the test atmosphere. Six diffusive tubes were exposed to the test vapour simultaneously with six pumped Tenax ATD tubes (5 ml min–1) for a period of 4 h. The pumped tubes were to provide a reference value for calculating the uptake rate. Both sets of tubes were analysed by ATD–gas chromatography with flame ionization detection (ATD–GC–flame ionization detection). The ATD primary desorption was at 300°C for 5 min with a low flow trap ("Air Monitoring Trap", Perkin Elmer part no. L427-5108) set at –30°C. The secondary desorption was at 350°C for 5 min. The split ratio varied over the range 0.2–6%, depending on the amount of alkane present.
Deposition on internal sampler walls
Deposition on to the internal walls of ATD tubes (either in empty tubes used as controls or in the diffusion gap of packed tubes) was investigated by exposing empty tubes and Tenax-packed tubes to test atmospheres of hydrocarbon vapour for 4 h at 20°C and 50% relative humidity.
Changes in the level of deposition of alkane vapour by molecular mass (C12–C16) were investigated by exposing six empty new tubes (fitted with standard internal gauze) to the same 4-h test atmospheres used to determine uptake rates (i.e. at 2.5% SVC). The reversibility of the process was investigated with a back diffusion test, where an additional six empty tubes were simultaneously exposed to the vapour for the 4 hours of the test, followed by clean air for 2 h before capping. The empty tubes were analysed using the same analytical method used for the Tenax-packed tubes, but with lower ATD split ratios for greater sensitivity. The tests were repeated with all alkane concentrations at 5 mg m–3, rather than as a function of the SVC.
Deposition mechanism
The mechanism by which the oil was deposited on the sampler walls was considered to be either condensation or adsorption, possibly on residual sorbent in the recycled empty tubes, on active sites caused by scratches, or just generally on the sampler walls.
Deposition of vapour on to seven new empty tubes was compared to seven recycled empty tubes originally packed with Tenax by exposing to test atmospheres of dodecane and tetradecane for 4-h periods. Before testing, the old tubes were prepared by brushing clean with a pipe cleaner, being ultrasonicated for 5 min while standing upright in methanol, fitting with gauze meshes and then being purged with helium at 300°C on the ATD tube cleaning programme.
The effect of available surface area was investigated by exposing new empty tubes with and without internal gauzes (seven of each) simultaneously to an atmosphere of tetradecane for 4 h. In the same experiment, the effect of internal scratches was investigated by including a further six new empty tubes, three of which contained scratches made by running an ATD retaining spring up and down inside the tube three times and three of which were unscratched. In other respects, these six tubes were very close to the original condition when received.
To investigate the presence of residual sorbent, old and new empty tubes were investigated further by exposing seven of each tube type to the headspace of a beaker of hexane for 5 min, followed by 2 h of clean air before analysis. Hexane was expected to quickly saturate and remain tightly bound to any sorbent present. Following GC analysis, the old tubes were heated at 400°C for 30 min while purging the tube with helium to degrade any residual sorbent particles. The headspace experiment was then repeated.
In an additional experiment using hexane, the effect of adsorption on to a chromium oxide layer on the inside of the old tubes was investigated. The tubes were first sanded with metallographic grinding paper (P1200 grit) and then polished with 6 µm diamond paste, followed by further polishing with 1 µm diamond paste. The tubes were then rinsed with water, ultrasonicated for 5 min while standing upright in methanol and cleaned on the ATD at 300°C before being re-exposed to the hexane headspace.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONSFUNDINGREFERENCES |
|---|
Diffusive uptake rate
The diffusive uptake rates determined in this work (equation 1) are recorded in Table 1 and presented with values from MDHS 80 (HSE, 1995) and from previous work (Simpson, 2000) in Fig. 1. The latter obtained uptake rates >8 h using a mixture of all the alkanes at concentrations between 3.64 mg m–3 (C10) and 0.064 mg m–3 (C16).
 | (1) |
|
|
- Where m = mass of analyte collected on the diffusive sampler (ng),
- C = concentration of analyte (mg m–3) and
- t = exposure time (min).
- C = concentration of analyte (mg m–3) and
It can be seen that up until dodecane, the diffusive uptake rate when expressed in units of ml min–1 falls gradually with increasing mass. After dodecane, the steady decrease cannot be clearly seen; discernment of any trend is hindered by the greater variability of the measurements at these concentrations. The diffusive uptake rates for decane to tetradecane agree reasonably well with those determined during previous work on oil mist; however, those for pentadecane and hexadecane are rather lower in this study, although both values lie within each others range of standard uncertainty. Consensus values are presented in Table 2. The uptake rates for heptane to decane are also somewhat lower than those quoted in MDHS 80. Use of a lower diffusive uptake rate will result in higher calculated airborne concentrations.
|
Petroleum-based solvents contain a great many hydrocarbons. As well as the straight chain alkanes used in these tests, there are many branched and cyclic alkanes. Although uptake rates may differ between alkane isomers, they will be of a similar magnitude. It would be too impractical to consider each individual compound separately. A pragmatic approach would be to use the same uptake rate for all alkanes of the same mass or eluting with similar retention times in GC on a non-polar column. Alkane isomers appear close together in the chromatogram, with branched alkanes generally eluting just before the straight chain isomer. For substances with relatively narrow boiling point ranges, it may be appropriate in some cases to use the uptake rate for the largest mid-range alkane and apply it to the whole hydrocarbon envelope. But it may be inappropriate if there was a significant aromatic fraction present.
Deposition on internal sampler walls
The tests made comparing the levels of hydrocarbon detected on diffusive Tenax ATD tubes with empty ATD tubes were made using relatively new Tenax tubes and brand new empty tubes. Due to limited numbers, the individual new empty tubes were used more frequently than the Tenax-packed tubes. The results are presented in Table 3 and graphically in Figs. 2 and 3. Figure 2 shows the data from the test atmospheres set at approximately 2.5% SVC. The amount collected on the empty tubes was fairly constant over the whole range and did not vary with the airborne concentration. The back diffusion test tube results appear to drop in comparison to the empty tubes for dodecane, tridecane and tetradecane. Significance testing gives single-tailed t-test results of P = 0.059, 0.047 and 0.027, respectively (df = 8). Back diffusion of the more volatile hydrocarbons would not be surprising.
|
|
|
Figure 3 shows the results from tests performed at similar concentration (
5 mg m–3). The dodecane and tridecane results are comparable to the 2.5% SVC data; the results for tetradecane, pentadecane and hexadecane are marginally higher and appear relatively constant. The back diffusion test results were all similar to the corresponding empty tube results, suggesting that the molecules were relatively tightly bound. As the volatility decreased, the amount of hydrocarbon on the empty tubes may have been expected to increase if condensation was involved. Despite the variation in hydrocarbon airborne concentrations, the decrease in the SVC of the hydrocarbon as molecular weight increased was not reflected in an increase in empty tube results and so condensation appears to be an unlikely mechanism for deposition.
Deposition Mechanism
The test results investigating wall effects are summarized in Table 4.
|
Comparison of old and new ATD tubes.
The tests comparing new empty ATD tubes with old empty recycled ATD tubes showed distinct differences in both mean loading and in variation of the results, with the older tubes collecting more hydrocarbon with much higher variation. This happened for both tetradecane (F-test P = 0.001, single-tailed t-test P = 0.002, df = 9) and the more volatile dodecane (F-test P < 0.001, single-tailed t-test P = 0.009, df = 6).
The new empty tubes were all purchased at the same time and had a similar history of employment. The histories of the older tubes are unknown, but they would have been of differing ages (likely to be >10 years old) and used in different applications. All the older tubes once contained sorbent. This difference in histories is likely to have affected both the differences in mean hydrocarbon loading and the variation.
The back diffusion test using dodecane showed a decrease in sample load over the 2 h exposed to fresh air for the new tubes (P = 0.028, df = 12), but not for the old tubes (P = 0.359, df = 11) in single-tailed t-tests.
The test comparing new tubes with and without the retaining mesh gauzes showed decreased hydrocarbon levels when the gauze was not present (single-tailed t-test P = 0.003, df = 12). This suggests that the degree of adsorption in the tube is primarily dependent on the surface area available rather than variable factors such as scratches, etc.
By ratioing the two analytical results and using the measured surface area of the tube walls (1398 mm2), a value was calculated for the effective surface area of the gauze (616 mm2). This was then used to calculate a surface area ratio of 0.11 between the Tenax tube diffusion gap (220 mm2) and the empty tube.
Impact of tube wall defects (scratches).
When the scratched tubes (mean 0.056 µg tetradecane adsorbed) were compared to the unscratched tubes (mean 0.066 µg) in a two-tailed t-test, there was found to be no statistical difference (P = 0.629, df = 4); however, when compared to a reference set of ‘new’ tubes (0.111 µg), they were significantly lower (single-tailed t-test P = 0.003, df = 8). Hydrocarbon adsorption on unscratched tubes was also lower than for the reference set (single-tailed t-test P = 0.001, df = 8). Both sets of new tubes were recently purchased and had not been recycled; the tubes used in the scratch test were unused at that point, whereas the other new tubes had been used in other tests. It appears that the scratches have made no difference to the level of adsorption, but even the limited use of the reference set of new tubes during this work has increased the potential for adsorption. They were regularly being desorbed at 300°C in the ATD with a helium purge, and cleaned at 300°C in either the ATD or in an oven with a nitrogen purge. This may account for the increased retention of dodecane in the back diffusion test at 5 mg m–3.
Examination for traces of sorbent.
In follow-up work exposing empty tubes to a static headspace of hexane vapour, it was initially intended to rule out the presence of minute quantities of sorbent remaining in the old tubes. It was anticipated that hexane would saturate any sorbent but be relatively unretained by scratches etc. It was thought that heating the tube to 400°C in the ATD tube cleaning programme would degrade the sorbent (although not necessarily removing it), without unduly changing the internal surface of the tube. Heat treatment was expected to decrease the adsorptive properties of any sorbent and possibly produce aromatic degradation products.
In the test after the heat treatment, all the old tubes collected more, rather than less, hexane. The mean tube loading increased from 0.059 to 0.099 µg with individual increases ranging between 30 and 141%. The wide variation in the condition of the old tubes prevented a paired t-test for means; however, regression analysis (omitting the highest result) gave a line of best fit whose 95% confidence limit for the gradient (1.22–1.33) excluded a value of 1 (Fig. 4). The chromatograms produced in analyses after heat treatment were virtually identical to those produced prior to heat treatment, i.e. there was no sign of any breakdown products. Although the results for the new tubes exposed in parallel also appeared to show an increase in hexane levels (mean increase 44%), they were at very low concentrations in both analyses and should be considered semi-quantitative.
|
There were no airborne concentration measurements made of the headspace, but hexane would have been present in overwhelming excess. No evidence for the presence of any sorbent was found but the data suggest that heating the tubes at high temperatures (higher than would normally be experienced) increased the subsequent level of adsorption.
Investigation of tube oxidation.
Considering the earlier results, it was thought that the increase in empty tube loadings over time could not be attributed to either minute quantities of sorbent or scratches to the internal surfaces, as these would, if present, have remained constant. There were no opportunities for either sorbent contamination or scratches to have occurred inside the new unused tubes. Therefore, some other change to the surface was suspected. When purchased, the ATD tubes are clean and shiny; however, over time during thermal cycling, the external surfaces start to oxidize to a bronze-like colour. During thermal desorption and cleaning, the air inside the tube is purged with inert helium or nitrogen before heating takes place and so there should be little or no oxidation. If, however, traces of oxygen were still present during heating or entered after heating while cooling, there may be slow oxidation over time. The rate of change would have been dependent on how regularly the tubes were analysed or cleaned and the temperatures used. The temperature used in this work (300°C) is close to the practical upper limit for Tenax TA and would be unsuitable for some other polymeric sorbents such as Chromosorb 106, but was chosen because of the high boiling points of the hydrocarbons being considered. Any oxide layer should be removable with abrasion, returning the tube to a state close to its original condition.
The results from the tube-cleaning experiment (sanding and polishing) showed a drop in the mean amount of hexane adsorbed by the tubes from 0.010 to 0.003 µg, shown to be significant in a single-tailed paired t-test (P = 0.003, df = 6). The individual decreases ranged between 56 and 96% (mean 80%). The pre-treatment mean value appears low in comparison to that in the heat treatment test (0.059 µg); however, the ranges are comparable when the two higher values in the heat treatment test are excluded (0.003–0.028 µg here compared to 0.003–0.030 µg). Hexane was undetected in the untreated new tubes run in parallel.
The sanding and polishing processes applied did not completely restore the tubes to a condition comparable with the new tubes. The process of cleaning the tubes was somewhat intuitive as it was impossible to view how effective the work was. However, the decrease in adsorption was significant, showing that the oxide layer was responsible (Fig. 5). Graphically, the change in the amount of hydrocarbon appears to be more uneven than in the heat treatment test, probably due to the lack of reproducibility in the cleaning process.
|
Impact of surface adsorption on measurement.
SVOC vapour may adsorb on the diffusion cap or in the tube's diffusion gap before it reaches the sorbent. If the vapour remains adsorbed then it will have an effect on the measurement. Vapour adsorbing on the sampler walls will shorten the mean path length and the uptake rate will be higher. If the metal surfaces become saturated, the diffusion gradient will stabilize. Vapour adsorbed to the sampler walls on capping would contribute to the analytical measurement, but vapour adsorbed on the diffusion cap would not. The measurements in this study were carried out using tubes with standard diffusive sampling caps with no membrane, so any bias from adsorption on the caps would have been included in the value calculated. The caps do not undergo any heating during use, so their effect is likely to be low.
The degree of bias experienced by sample tubes depends on the amount of hydrocarbon present on the metal surfaces and the amount collected on the sorbent, and should have a greater effect at lower airborne concentrations and shorter exposure times. However, the bias experienced by the empty tubes would be much larger due to the larger surface area exposed. The empty tube result could be scaled to the level in the sample tubes using the 0.11 factor calculated here.
An approximation of the bias can be obtained by comparison of the scaled empty tube result against the corresponding diffusive tube result. The true bias is likely to be lower due to the concentration gradient within the free space in the diffusion gap. Hence, the uncertainty of measuring the hydrocarbon can be calculated as the bias plus twice the relative standard deviation of the diffusive tube result, thus indicating the likely possibility of diffusive sampling being successfully used at the tube loadings indicated (Table 5). Uncertainty as used here is similar to the ‘overall uncertainty’ of the former EN 482 where the bias is known approximately but not corrected for (BSI, 1994). Bias is interpreted differently in the 2006 revision of EN 482. A bias component that is unknown is treated as another variance. Where known and correctable, bias is excluded from uncertainty (BSI, 2006).
|
Adsorption had a higher proportional effect for lower airborne concentrations. The values calculated for dodecane to tetradecane, when scaled, represent <2% of the sample collected and so would have minimal effect on diffusive tube measurements. The bias was slightly larger for pentadecane and hexadecane. Relative standard deviation was more significant in calculating uncertainty. Measurements of aliphatic hydrocarbon vapour for occupational hygiene purposes are generally concerned with concentrations of >200 mg m–3 because they are of relatively low toxicity. The SVC of pentadecane, however, is only 23 mg m–3 at 20°C and that of hexadecane 10 mg m–3. The effect of adsorption on pentadecane and hexadecane, combined with the lower precision of measurements at low concentrations, resulted in unacceptable overall uncertainty (>50%) for measurements by diffusive sampling of the individual compounds at low concentration. However, they are only likely to be present in the workplace as minor components in hydrocarbon mixtures. In such a case, their contribution to the uncertainty of a measurement of the whole vapour mixture would be reduced, and the whole would be at a more acceptable level.
The impact may be larger if the purpose of the measurement was ultimately to use the vapour concentration to determine a particulate concentration from a total hydrocarbon measurement of a light mineral oil mist. In such cases, the bulk of the airborne oil is likely to be in the form of vapour (Simpson, 2003). For illustration, if a total hydrocarbon measurement was 10 mg m–3 and the accompanying separate vapour measurement was 9 mg m–3, then an acceptable error on the vapour concentration of e.g. 0.5 mg m–3 would be less acceptable on the calculated oil mist concentration of 1 mg m–3. The standard deviation of the diffusive tube data in Table 3 expressed as an airborne concentration ranged from 0.02 to 0.94 mg m–2 and was typically 0.4 mg m–3. Taking an estimate of the standard deviation of the total hydrocarbon concentration at 10 mg m–3 from previous work (Simpson, 2003) to be 0.5 mg m–3, the combined standard uncertainty will be approximately 0.64 mg m–3. This would correspond to an acceptable combined expanded uncertainty (i.e. 50%) at 2.56 mg m–3, appropriate for a limit value of 26 mg m–3 (BSI, 2006).
It should be noted that the values obtained here were from relatively new Tenax TA-packed tubes and new empty tubes. Older tubes that have undergone numerous cycles at higher temperatures would be subject to greater bias. Without knowledge of the age and history of the individual tube, it would be impossible to know if the adsorption effects in the diffusion gap had increased to an unacceptable degree. It is also impracticable to remove the oxide layer from a packed ATD tube without also repacking it. In principle, the use of silico-steel-lined tubes would remove most adsorption sites. Glass tubes packed in-house would be unsuitable for diffusive sampling due to variability in the diffusion gap. Consequently, the choice of sampling media for similar measurements would be important, especially if the analytes were of low volatility and at low concentration. In order to reduce oxide build up on the tube surface, it is recommended that the ATD tubes are desorbed at the minimum temperature necessary for the hydrocarbon in question, and silico-steel tubes should be considered for use with semi-volatiles.
Use of badge-type diffusive samplers would avoid the problem of adsorption within the sampler; however, precautions must be made to avoid aerosol particle impaction or splash contamination. Bias due to the ingress of mist particles into diffusive ATD tubes is likely to be low relative to the vapour, but it would be recommended to direct diffusive tubes downwards during sampling and to consider using empty control tubes to identify any problems (Simpson, 2003). High wind speeds should not affect tubes fitted with diffusive sampling caps.
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONSFUNDINGREFERENCES |
|---|
Mixed hydrocarbon solvent vapours up to hexadecane can be measured by diffusive sampling and analysis by thermal desorption providing concentrations are reasonably high (e.g. >5 mg m–3) and the tubes are in reasonable condition. Pentadecane and hexadecane could not be determined as individual compounds.
When used to calculate oil mist concentrations from total airborne hydrocarbon concentrations, the oil vapour measurement uncertainty would adversely impact on determining oil mist concentrations less than 3 mg m–3.
Semi-volatile components can become adsorbed on the internal metal surfaces of ATD tubes during exposure measurements. Back diffusion tests suggest that the molecules can be relatively tightly bound. The level of adsorption depends upon the history of the tube and is relative to the build up of an oxide layer caused by regular heating to high temperatures. In order to reduce oxide build-up, it is recommended that the ATD tubes are desorbed at the minimum temperatures necessary for the hydrocarbon in question.
|
FUNDING |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONSFUNDINGREFERENCES |
|---|
UK Health and Safety Executive.
Received October 8, 2007; in final form February 20, 2008
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEXPERIMENTALRESULTS AND DISCUSSIONCONCLUSIONSFUNDINGREFERENCES |
|---|
British Standards Institution (BSI). BS EN 482:1994 Workplace atmospheres—general requirements for the performance of procedures for the measurement of chemical agents (1994) London: British Standards Institution.
British Standards Institution (BSI). ENV 13936:2001 Workplace atmospheres. Measurement of chemical agents present as mixtures of airborne particles and vapour. Requirements and test methods (2001) London: British Standards Institution.
British Standards Institution (BSI). BS EN 482:2006 Workplace atmospheres—general requirements for the performance of procedures for the measurement of chemical agents (2006) London: British Standards Institution.
HSE. ‘Generation of test atmospheres of organic vapours by the syringe injection technique. Portable apparatus for laboratory and field use.’ MDHS 3 Methods for the determination of hazardous substances. HSE Books ISBN 0 7176 0228 1. Available at www.hse.gov.uk/pubns/mdhs/index.htm (1990).
HSE. (1995) ‘Volatile organic compounds in air, laboratory method using diffusive solid sorbent tubes, thermal desorption and gas chromatography’. MDHS 80 Methods for the determination of hazardous substances. HSE Books ISBN 0 7176 0913 8. Available at www.hse.gov.uk/pubns/mdhs/index.htm.
Perez C, Soderholm SC. Some chemicals requiring special consideration when deciding whether to sample the particle, vapour, or both phases of an atmosphere. Appl occup Environ Hyg (1991) 6:859–64.
Scobbie E, Dabill DW, Groves JA. The development of an improved method for the determination of coal tar pitch volatiles (CTPV) in air. Ann Occup Hyg (1998) 42:45–59.
Scobbie E, Groves JA. Determination of hydroquinone in air by high performance liquid chromatography. Ann Occup Hyg (1999) 43:131–41.
Simpson AT. Assessment of exposure to light mineral oil based metal working fluids (2000) Health and Safety Laboratory Report HSL/2000/22. HSE. Available at www.hse.gov.uk/research/hsl/workenvn.htm.
Simpson AT. Comparison of methods for the measurement of mist and vapour from light mineral oil based metalworking fluids. Appl Occup Environ Hyg (2003) 18:865–76.[Medline]
Soderholm SC. Aerosol instabilities. Appl Ind Hyg (1988) 3:35–40.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||