Annals of Occupational Hygiene Advance Access originally published online on April 28, 2008
Annals of Occupational Hygiene 2008 52(4):239-247; doi:10.1093/annhyg/men014
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
GABIE and Perkin Elmer Passive Sampler Performance under Fluctuating Concentration Conditions
Institut National de Recherche et de Sécurite (INRS), 50 avenue de Bourgogne, BP27, 54500 Vandoeuvre les Nancy, France
* Author to whom correspondence should be addressed. Tel: +33 (0)383 502 025; fax: +33 (0)383 502 060; e-mail: eddy.langlois{at}inrs.fr
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONEQUIPMENT AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Passive sampling is an approved and accurate method for the assessment of organic compound exposure over long sampling time. This method could be very convenient for the short-time exposure assessment, but passive samplers have to be validated for this use. In this article, the behaviour of two commercial passive samplers (GABIE and Perkin Elmer) under fluctuant concentration conditions is studied. Artificial atmospheres were produced in the laboratory and passive samplers were exposed to different concentration profiles. Both theoretical and experimental results detailed in the paper underline the capability of these two samplers to assess pollutant exposure either when the concentration is unsteady or when the sampling time is short. Then, a suitable sampling strategy is proposed for the assessment of short-term exposure, based on the association of a direct reading photoionization device and passive sampler.
Keywords: exposure assessment • fluctuating concentration • passive sampling • peak exposure • short-term exposure
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONEQUIPMENT AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Diffusive sampling is known to be an accurate alternative to the active sampling method of assessing organic compound exposure in workplaces. Based on the principle of molecular diffusion, passive samplers are small in size, lightweight and easy to both use and wear (Gorecki and Namiesnik, 2002). The passive sampling method can be very convenient for performing long-term monitoring, large-scale exposure assessment or defining homogeneous exposure groups because it is not very expensive and does not require technically trained personnel, unlike active sampling. Many studies have validated the use of passive sampling either in a laboratory-generated atmosphere or under the workplace conditions for long-term exposure assessment (Hickey and Bishop, 1981; Melcher, 1985; Ullrich, 1992).
Many industrial operations cause large variations in pollutant concentration and workers can therefore be exposed to high concentrations over short periods. From a toxicological standpoint, peak exposure may be associated with acute effects because it produces a high dose in target tissues or organs. Implementation of a long-term sampling strategy (8 h) tends to produce an average concentration over the whole period and therefore underestimates the potential acute toxic effects (Preller et al., 2004). Short-term exposure assessment is consequently often more accurate when operations creating peak exposures are identified. However, hygienists find implementation of this kind of sampling strategy quite difficult using the usual active methods. This context highlights the advantages of passive sampling, as previously described, and passive sampler usage would considerably simplify the sampling procedure.
Controlled test atmospheres are created for assessing sampler performance and accuracy when developing sampling and analytical methods or undertaking experimental uptake rate determination. In all these methods, INRS (2008), OSHA (2008), NIOSH (1994), ISO 16107 (2007), samplers are exposed in continuous, steady-state pollutant concentration atmospheres for a reference period, frequently 8 h (Kennedy et al., 1995). While this laboratory step is necessary, it is insufficient because workplace exposure is rarely as uniform as in this kind of controlled test atmosphere. In most real cases, the pollutant profile corresponds to intermittent emission of pollution peaks induced by the industrial process concerned.
There is very little available information on the behaviour of sampling systems in peak exposure situations and their capacity for determining pollutant concentration. Under these circumstances, sampling systems are often assumed to be as accurate as in stable, continuous pollutant concentration atmospheres and their performance has seldom been studied in relation to measuring pollutant concentrations featuring sharp peaks. While active sampling works at a constant, established flow rate, passive samplers are known to be subject to a transient period prior to diffusion flow steady state. This transient period corresponds to boundary limit establishment in the diffusion path (Hori and Tanaka, 1993) and depends mainly on the diffusion path length and the nature of the attenuation layer covering the sampler shape (Hearl and Manning, 1980; Bartley et al., 1983). Uptake rates for passive samplers are determined under validation steady-state conditions described above. Transient period is related to the sampler geometric parameters, so uptake rates could influence results depending on the total sampling time. In this case, it is important to evaluate transient period influence on sampler performance to characterize the minimum peak ‘size’ (height, time, etc.), which could be accurately measured. Simpson et al. (2003) have suggested using a thermal desorption sampler to assess peak exposure and have found that active and passive methods agree well, based on European standard requirements (EN 482, 2006). However, thermal desorption is not a universal method and has a number of drawbacks, so using a solvent desorption passive sampler should constitute an interesting alternative.
The aim of our study is firstly to investigate the behaviour of two different passive samplers during exposure to pollutant concentration peaks under controlled test atmospheres. The first sampler is the thermally desorbed Perkin Elmer sampler and the second is the INRS-designed and -developed GABIE sampler containing activated charcoal to be desorbed with solvent after exposure. The latter sampler has been validated for many substances, in particular, volatile organic compounds (Delcourt and Sandino, 2000). Performance of both these samplers has been compared to reference active sampling.
Passive sampling is subsequently validated for monitoring not only highly fluctuating atmospheres but also short-period peak emissions for comparison of concentration values with short term exposure limit (STEL) values. We then propose a suitable strategy, based on simultaneous use of photoionization detectors (PIDs) for following the global concentration trend and a passive sampler for method selectivity in the case of multi-pollution.
In the final section, we discuss the advantages, drawbacks and limitations of both these passive sampling systems to provide advice to hygienists, who need to perform short-term exposure assessment, and to help in selecting the method most suited to the specific character of the exposure situation.
|
EQUIPMENT AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONEQUIPMENT AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Atmosphere generation
Controlled atmosphere generations were carried out in a test chamber, in which temperature, pressure and pollutant concentration were accurately known (Fig. 1). The main airflow temperature and the moisture were first adjusted and a small calibrated polluted airflow was then injected into this main flow. The total flow circulated within the exposure chamber, in which tubes, badges and PIDs were placed. Passive samplers require a minimum air velocity of 0.2 m s–1 to be efficient, so the atmospheric air velocity was fixed by a recycling loop incorporating a fan such that an air velocity of 0.5 m s–1 was maintained.
|
A programmable syringe pump was used to introduce liquid pollutants in the form of peak emissions to simulate real workplace conditions.
Pollutant concentrations were monitored using an online gas chromatography (GC) giving the sequential concentration every 4 min and a photoacoustic infrared detector (PAIR) signalling a continuous concentration value. Both systems had been previously calibrated with respect to a gas cylinder containing pollutant at a certified concentration.
In addition to providing a continuous display of the concentration profile, the PAIR has the advantages of good linearity within a wide concentration range, so that it is particularly suitable for peak concentration monitoring. Overall uncertainty of the PAIR concentration evaluation is <5%.
Toluene was used for experiments in this first part of the study. Table 1 gives the profile characteristics of the nine experiments. Reference atmosphere (A) was sampled with a constant toluene concentration at 50 p.p.m. for a period of 4 h (Experiment A). All subsequent atmospheres (B–G) contained the same quantity of toluene injected in square-wave profiles for the same overall exposure time (4 h). Final Experiment I was conducted for a short time (15 min) to monitor a single peak lasting 6 min. The purpose of this was to simulate a short sampling exercise intended for STEL value comparison. Figure 2 shows the modelled concentration profile and the PAIR-measured actual concentration profile for the controlled atmospheres used in this first part of the study.
|
|
In the second part of the study, controlled test atmospheres were performed featuring peaks of three different pollutants to investigate the influence of pollutant type and the effect of co-pollution. Perchloroethylene and isoflurane therefore complemented toluene simply to simulate co-pollution. In the first atmosphere, the three pollutants were injected under steady-state conditions (4 h, 200 p.p.m. of each pollutant), the three pollutants were then injected as four 15-min peaks and, finally, as a single 15-min peak.
Sampling and analysis
Active sampling was performed using individual pump (Gillian LFS 113 or Gillair 3 fitted with a low-flow adaptor). Flow rates were measured at the start and end of the sampling period using a soap bubble flow meter and SKC®(SKC)-activated charcoal tubes (100 + 50 mg). The sampling flow rate was 
100 cm3 min–1, but this was accurately measured at the start and end of the sampling period.
Passive sampling was performed with GABIE charcoal samplers (ARELCO 2008, France) and Perkin Elmer Tenax TA tubes. Nominal workplace uptake rates were initially used and these were corrected for ambient temperature and pressure.
Six charcoal tubes, six GABIE samplers and five Perkin Elmer tubes were simultaneously exposed (except for Experiments A and B, in which only three Perkin Elmer badges were exposed) in each artificial atmosphere.
The SKC tubes and GABIE sampler were desorbed with CS2 and analysed by GC according to the reference method. Calibrations were performed prior to experiment analyses (INRS, OSHA, NIOSH).
The Perkin Elmer tubes were analysed by thermal desorption on an ATD 400 Perkin Elmer analytical device. Calibration was performed before each experiment analysis by filling tubes with the same calibrated gas cylinder used for online GC and PAIR calibration in the case of toluene and from the controlled atmosphere in the case of perchloroethylene and isoflurane.
A ToxiRAE PID was used to monitor concentration in the controlled atmosphere device and was isobutylene calibrated prior to monitoring.
Collected quantity calculation
For a steady-state concentration profile, theoretical recovery of substance i by both active and passive sampling systems is given by equations (1) and (2), respectively.
 | (1) |
 | (2) |
According to Fick's second law, concentration in the diffusion zone actually varies with respect to both time and distance in the diffusion length, so that equation (2) becomes equation (3):
 | (3) |
Concentration variations with distance in the diffusion path can then be neglected and the ambient concentration has to be integrated merely with respect to time in both equations (1) and (3). The compound quantity can therefore be calculated using equations (4) and (5).
 | (4) |
 | (5) |
As long as the concentration profiles are available, the integral term featuring concentration with respect to time can be determined. A trapezoidal approximation method was used to solve this integral.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONEQUIPMENT AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
Single pollutant atmosphere
Toluene quantities can be calculated based on the concentration profile, uptake rates and sampling time for each type of sampler. Analysed quantities were compared with these calculated values. Mean values for the six measurements are given in Table 2. Results are expressed in terms of bias between the pollutant predicted value and the quantities determined by active and passive sampler analysis.
|
Bias between the GABIE sampler, and to a lesser extent the Perkin Elmer tubes, and the active sampler (charcoal tube) is quite large, even for the reference atmosphere (
5 and 4%, respectively, in Experiment A). We decided to standardize all sampler results based on Experiment A, featuring continuous constant exposure, to avoid confusion between bias due to uptake rate uncertainty and the bias due to the peak exposure profile effect. The mean values of the different sampler results were then adjusted to ensure an unbiased fit with Experiment A, as shown in Fig. 3. Although this study does not consider uptake rate accuracy, it is worth noting that the necessary standardization correction was <5% for all samplers; this implies good agreement not only between active and passive sampling but also with the uptake rates quoted by the manufacturer for each passive sampler.
|
Once corrected, all analysed toluene quantities for all exposure profiles and sampling methods agreed closely with the predicted values. Figure 3 shows the bias variation between predicted and sampler medium-measured pollutant values for each artificial atmosphere (corrected as above). The artificial atmosphere profiles have been plotted at the top of figure. The measurement mean value for each atmosphere profile is plotted in standard deviation (SD) bar form.
Passive sampler results are more scattered than for active sampling charcoal tubes. Intraclass SDs are approximately five times smaller for active sampling charcoal tubes than for passive samplers. This scatter is mainly due to the many environmental parameters, which can significantly influence uptake rate values. Active sampling tube and GABIE sampler results were derived by applying the same analytical method but the pollutant quantity trapped on the badge is two or three times less than for the charcoal tube because of the different sampling rates. Furthermore, solvent quantity for desorption is five times greater for the GABIE sampler than for the charcoal tube because of the very large quantity of charcoal. The analysed pollutant concentration in solvent for the GABIE sampler is an order of magnitude less than that observed with active sampling tubes. This could explain the higher SD depending on the analytical procedure implemented.
The intermediate group scatter of the Perkin Elmer tube results was very high (from –10.4 to 7.5%). While the combined results appear to fall within the overall uncertainty, the results of Experiments D–I reveal a trend. After a slight drop in the first experiments, the Perkin Elmer sampler results seem to rise rapidly, when pollution peaks are shorter and concentration values increase. However, calibration process (spiking tubes with calibrated bottles) uncertainty could also give the high SDs in this intermediate group.
The bias between charcoal tubes and Perkin Elmer sampler fluctuates widely and exceeds 10% of the last experiment, while the bias between charcoal tubes and GABIE sampler revealed an insignificant trend with respect to exposure profile and remained within 5% in all experiments.
Multi-pollutant atmosphere
PAIR monitoring proved ineffective for quantifying pollutant mixing (toluene, perchloroethylene and isoflurane). The predicted quantity of each pollutant cannot be calculated in this way. For this second part of the study, results have therefore been expressed as differences between active and passive sampling methods (bias versus tube in Table 3).
|
With regard to the GABIE sampler, the differences from the active method observed in Experiment A, during the first part of the study, are confirmed for toluene and are greater for isoflurane and perchloroethylene (see Table 3). As in the first part of the study, uptake rates are subsequently corrected with respect to the reference exposure (Experiment K, 4 h, 50 p.p.m. of each pollutant) for the other two experiments (L and M).
Corrected GABIE sampler results remain unaffected by the peak profile for any substance. On the other hand, bias between active sampling and the Perkin Elmer sampler is subject to highly variable discrepancies, and underestimations ranging from 16 to 25% for Experiment M (bolded results in table 3). In this last experiment, the exposure to the pollutants was followed by a long exposure time to clean air (225 min). These results highlight the importance of back diffusion for Tenax Perkin Elmer sampler compared to charcoal GABIE sampler, as previously observed (Oury et al., 2006).
Although PID results agree well with predicted values for single pollutant exposure assessment, the overall signal in multi-pollutant experiment does follow the peak profile, but cannot obviously provide quantitative information.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONEQUIPMENT AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
In this study, peak duration was limited by technical constraints, the artificial atmosphere production systems implemented being incapable of generating peaks shorter than a few minutes. However, the results obtained can be easily extrapolated to shorter peaks based on the following considerations.
In a theoretical study about the behaviour of ideal sorbent passive sampler, Bartley et al. (1983) calculated the relaxation time 
relax by integrating Fick's second law. This value of relaxation time, given by equation (6) is the equilibrium time in the diffusion area, assuming that there is no substance loss from the sorbent:
 | (6) |
However, in real samplers, sorbent are subject to back diffusion; as a consequence, the equilibrium time can be much longer due to the loss from the sorbent. In order to take the back diffusion into consideration, Tompkins and Goldsmith (1977) defined the average residence time in the diffusion area at the steady-state conditions as the substance mass hold-up in the sorbent divided by the diffusive transport rate, given by equation (7):
 | (7) |
Hearl and Manning (1980) have shown that, during the transient period, the concentration profile throughout the diffusion path is non-linear, unlike under steady-state conditions, but the equilibrium is reached when the value of the dimensionless time , given by equation (8), exceeds 1.5. Hori and Tanaka (1993) have theoretically and experimentally confirmed the fact that the concentration profile is non-linear in the diffusion area during the transient period and that the equilibrium is reached within 1.4 times the average residence time at the steady state (). But even in this case, surplus collected matter during increasing concentration at the beginning of a peak is balanced by the deficiency of collected quantities when the peak is decreasing such that the mass balance is nil. The correction factor proposed by Hearl and Manning (1980) is therefore not needed.
 | (8) |
The required time for the equilibrium, calculated by solving equation (9), is equivalent to three times the value of residence time defined by Tompkins and Goldsmith (1977); values of relaxation time and equilibrium times are given in Table 4 for the two samplers applied for toluene; this time does not exceed 13 s for GABIE and 77 s for Perkin Elmer. The differences between ideal sorbent sampler relaxation time and residence time evaluations show the complexity of adsorption and back diffusion phenomena. However, it is reasonable to think that actual passive sampler response times are likely to be within these two extreme values.
 | (9) |
|
As foreseen by this ‘equilibrium time’ calculation, the transient period in our experiment had no significant influence on results as long as the peak duration exceeded a couple of minutes. According to a longer diffusion area and the weaker sorbent, for shorter peaks (<1 min), the Perkin Elmer's equilibrium time will affect results especially for multi-pollutant exposure.
GABIE and Perkin Elmer passive samplers are therefore suitable for assessing concentration in a fluctuating concentration atmosphere over both long and short periods. Overall measurement accuracy remains in all cases in agreement with the European standard requirements for sampling and analytical methods for the workplace exposure assessment (EN 482, 2006). Equilibrium time should not be influenced by the type of pollutant because it depends on the diffusion constant and geometric parameters. This hypothesis was partially proven in the multi-pollutant experiment, but requires confirmation through field experiments.
For short-term measurement, usual sampling methods comprise associating a direct-reading sampler (photoionization), which provides an instantaneous global air concentration trend, which is not specific in the multi-pollutant atmosphere case, and active sampling methods, in which collecting tubes are placed close to the PID. Active sampling and its subsequent analysis allow airborne pollution qualification (Poirot et al., 2004; Preller et al., 2004). Use of passive samplers instead of active tubes, proposed by Simpson et al. (2003), is very convenient in the case of multiple assessments or sampling in severe conditions (electromagnetic field, sterile location, etc.). This sampling technique needs to be selected based on the assumption that the material collected on the sampler agrees with the detection limit level used in the analytical process. In other words, hygienists need to establish a compromise between sampling task simplification and threshold value for the collected quantity of pollutant.
Proposed sampling methodology for peak exposure assessment based on the simultaneous use of the GABIE passive sampler and PID has now to be validated through field experiments, in which peaks and co-pollution conditions are more complex and random.
It is also important to pay particular attention to using measurements made under space- and time-based concentration fluctuations. These types of exposure profile are characterized by a log-normal distribution featuring a high SD. Recent studies have highlighted the marked rise in uncertainty of a single measurement in a fluctuating concentration atmosphere (Grzebyk and Sandino, 2005). In this work, the authors have shown that day-to-day concentration variations lead to significant discrepancies for two different-day samples and they emphasize the importance of increasing the number of samples to give a more confident result. On a smaller scale, concentration fluctuations within a day are therefore considered to have the same influence on two independent samples. Large numbers of active samples also lead to repetition at the flow rate checking stage, which not only is time consuming but also allows the tubes to be exposed to the previous or next peak. Use of a passive sampler is undoubtedly more convenient for large series of samples.
It is important to notice that the analytical stage differs significantly in both principle and aim despite the fact that both samplers are validated for short-term exposure assessment. The following factors must therefore be taken into account when opting for a sampling strategy. While sampling process performance is undoubtedly more convenient with passive samplers, more attention has to be given to selecting the analytical method. Thermal desorption is a very sensitive analytical technique and is therefore more suitable for trace analysis, for sharp peaks or for substances with low limit values. However, recent studies have revealed the importance of back diffusion of hydrocarbons trapped on thermal desorption, Tenax TA (Oury et al., 2006) or graphitized carbon (Pennequin-Cardinal et al., 2005) samplers, when they are no longer exposed to a pollutant. This problem of back diffusion due to weaker interactions between sorbent and pollutant could be a serious drawback in short-peak exposure followed by lengthy non-exposure and may lead to underestimation of pollutant concentration.
Another drawback of thermal desorption analysis is the awkward calibration process, often based on the pollutant spiking principle. As far as solvent desorption is concerned, a calibration process based on preparing a liquid calibration solution is more accurate. Depending on its purpose, the user then has to balance method sensitivity against result accuracy.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONEQUIPMENT AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
GABIE and Perkin Elmer passive samplers give good results when exposed to peak profiles and the sampling rate experimentally determined for constant concentration profiles is sufficiently accurate to be confidently used. The diffusion path length for the Perkin Elmer sampler amplifies the influence of the transient period, so this badge should be used in the case of ‘gentle’ peak profiles.
Relative SD is significantly higher for passive samplers than for active samplers. Slight negative bias is generally obtained for sharp peak exposure profiles under active sampling, compared with positive bias under passive sampling. Bias is more difficult to estimate because of the higher SD under passive sampling. In both cases, bias and SD agree closely with EN 482 standard requirements governing conventional occupational exposure assessment measurement (30 and 50%, depending on the concentration level). Agreement with active sampling results is good. Peak size has no significant influence.
Perkin Elmer sampler results exhibit important concentration underestimations when exposed for long times to clean air just after a short-time pollutant exposure which can be explained by back diffusion. Even if this phenomenon is not very important in the case of short-term exposure assessment, it could be emphasized in some cases of long-term exposure assessment. Hygienists, who need to assess highly fluctuating concentration atmospheres or short-term exposures, are therefore advised to combine the use of passive samplers and direct-reading photoionization devices. This sampling strategy must be suited to both concentration levels and peak occurrence frequencies. A thermally desorbed Perkin Elmer sampler may be preferred for low concentration levels because of its very high analytical sensitivity. In all other cases, a solvent desorbed sampler may be used without specific equipment.
The transient period depends essentially on the geometrical shape of the sampler; thus, the results of this toluene-based study could be extrapolated to other substances.
This capacity confirms the use of passive samplers for either long-term exposure with peak emissions or short-term exposure assessment for STEL comparison, when the peak occurrence time is well known, for example, when opening a solvent tank or reactor.
The results obtained with toluene in controlled test atmospheres need to be validated by field experiments and this will represent the next phase of this study.
|
ACKNOWLEDGEMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONEQUIPMENT AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
The author wishes to thank in particular Blandine Castel, Catherine Lefevre, Yves Morele and Bruno Galland (INRS) for their analytical support throughout this study.
Received August 13, 2007; in final form March 3, 2008
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONEQUIPMENT AND METHODSRESULTSDISCUSSIONCONCLUSIONACKNOWLEDGEMENTSREFERENCES |
|---|
ARELCO. (2008) Gabie sampler provider. Available at http://www.arelco.fr. Accessed April 2008.
Bartley DL, Doemeny LJ, Taylor DG. Diffusive monitoring of fluctuating concentrations. Am Ind Hyg Assoc J (1983) 44:241–7.[Web of Science]
Delcourt J, Sandino J-P. Performance assessment of a passive sampler in industrial atmospheres. Int Arch Occup Environ Health (2000) 74:49–54.[CrossRef][Web of Science]
EN 482. (2006) European Standard: workplace atmospheres – general requirements for the performance of procedures for the measurement of chemical.
Gorecki T, Namiesnik J. Passive sampling. Trends Anal Chem (2002) 21:276–91.[CrossRef]
Grzebyk M, Sandino J-P. Statistical aspects and influence of measurement uncertainty on evaluation of occupational exposure to chemical agents. Hyg Sécur Trav (2005) 200:9–22.
Hearl FJ, Manning MP. Transient response of diffusion dosimeters. Am Ind Hyg Assoc J (1980) 41:778–83.[Web of Science][Medline]
Hickey JLS, Bishop CC. Field comparison of charcoal tubes and passive vapor monitors with mixed organic vapors. Am Ind Hyg Assoc J (1981) 42:264–7.[Web of Science]
Hori H, Tanaka I. Response characteristics of the diffusive sampler at fluctuating vapor concentrations. Am Ind Hyg Assoc J (1993) 54:95–101.[Web of Science]
INRS. INRS MetroPol database (2008) Available at http://www.inrs.fr/metropol/sommet.htm. Accessed 4 April 2008.
ISO 16107. Protocol for evaluating the performance of diffusive samplers (2007) Geneva, Switzerland: International Organization for Standardization.
Kennedy ER, Fischbach TJ, Song R, et al. Guidelines for air sampling and analytical method development and evaluation (1995) Cincinnati, OH: DHHS (NIOSH).
Melcher R. Laboratory field validation of solid sorbent samplers. In: Sampling and calibration for atmospheric measurements—Taylor J, ed. (1985) Boulder, CO: ASTM.
NIOSH. NIOSH manual of analytical methods (NMAM). Available at http://www.cdc.gov/niosh/nmam/. Accessed April 2008.
OSHA. OSHA guideline. Evaluation guidelines for air sampling methods utilizing chromatographic analysis. Available at http://www.osha-slc.gov/dts/sltc/methods/chromguide/index.html. Accessed 14 January 2008.
Oury B, Protois J, Lhuillier F, et al. Behaviour of the GABIE, 3M, PERKIN ELMER Tenax TA and Radiello 145 diffusive sampler exposed over a long period to a low concentration of VOC. J Occup Environ Hyg (2006) 3:547–57.[CrossRef][Web of Science][Medline]
Pennequin-Cardinal A, Plaisance H, Locoge N, et al. Dependence on sampling rates of Radiello(R) diffusion sampler for BTEX measurements with the concentration level and exposure time. Talanta Opt Waveguide Anal (2005) 65:1233–40.
Poirot P, Subra I, Gerardin F, et al. Determination of short-term exposure with a direct reading photoionization detector. Ann Occup Hyg (2004) 48:75–84.
Preller L, Burstyn I, De Pater N, et al. Characteristics of peaks of inhalation exposure to organic solvents. Ann Occup Hyg (2004) 48:643–52.
Simpson A, Hardwick K, Walsh P, et al. Evaluation of diffusive samplers and photoionisation detectors for measuring very short peak exposure in the workplace. J Environ Monit (2003) 5:732–8.[CrossRef][Web of Science][Medline]
Tompkins FC, Goldsmith RL. A new personal dosimeter for the monitoring of industrial pollutants. Am Ind Hyg Assoc J (1977) 38:371–7.[Web of Science][Medline]
Ullrich D. Diffusive sampling of organic compounds. In: Clean air at work—Brown R, Curtis M, Saunders K, et al, eds. (1992) Lewis Pub: Royal Society of Chemistry.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||