Use of Two Different Acidic Aerosol Samplers To Measure Acidic Aerosols in Hsinchu,Taiwan

ABSTRACT

Acidic aerosol concentrations measured by an annular denuder system (ADS) and a honeycomb denuder system (HDS) in Hsinchu, Taiwan, were compared. Aerosols were also sampled by a MOUDI (micro-orifice uniform deposit impactor) and analyzed by an ion chromatograph to determine the size distributions of different species. Using the measured aerosol size distribution, theoretical analysis showed that positive HNO₃ artifact due to volatiliza-tion of NH₄NO₃ is generally negligible for both samplers.4 3Comparing two different denuder samplers, the average concentration of HNO₃ measured by the ADS was found3to be lower than that measured by the HDS, while the difference between the two samplers for the average concentration of other species was found to be within ±15%. A possible cause of the difference in HNO₃ con-3centrations is due to a greater loss of HNO₃ in the cyclone3 used by the ADS than in the impactor used by the HDS. The study also showed incomplete absorption of the evaporated HCl and HNO₃ from the particles on the Teflon3filter by the first nylon filter in the filter pack of the ADS. Collection efficiency and capacity of HCl and HNO₃ by3the nylon filters need further investigation.     

Simultaneous sampling of NH3, HNO3, HC1, SO2 and H2O2 in ambient air by a wet annular denuder system

A new sampling device is described for the simultaneous collection of NH₃, HNO₃, HCl, SO₂ and H₂O₂ in ambient air. The apparatus is based on air sampling by two parallel annular denuder tubes. The gases are collected by absorption in solutions present in the annulus of the denuder tubes. After a sampling time of 30 min at flow rate of 32 ℓ min⁻¹ the solutions are extracted from the denuders and analyzed off-line. The detection limits of NH₃, HNO₃, HCL and SO₂ are in the order of 0.1–0.5 μm⁻³. For H₂O₂ the detection limit is 0.01 μm⁻³. The reproducibility is 5–10% at the level of ambient air concentrations. Comparison of this novel technique with existing methods gives satisfactory results. The compact set-up offers the possibility of field experiments without the need of extensive equipment.     

Comparison of the Annular Denuder System and the Transition Flow Reactor for Measurements of Selected Dry Deposition Species

An intensive field study was conducted in Research Triangle Park, North Carolina in the fall of 1986. Ambient concentrations of the following constituents were obtained: nitric acid, nitrous acid, nitrogen dioxide, sulfur dioxide, ammonia, hydrogen ion, and particulate nitrate, sulfate, and ammonium. Results collected using the annular denuder system (ADS) and the transition flow reactor (TFR) are presented and compared.

Both types of samplers had operational detection limits on daily (22-hour) samples that were generally below 1 μg m_-3 suggesting that both samplers can provide sensitive measurements for most of the constituents of interest. Both the ADS and TFR show reasonable (>25 percent) within-sampler precision for most of the measured species concentrations, except TFR fine particulate nitrate measurements where results were frequently negative (The TFR fine particulate nitrate measurement is calculated using subtraction of positive numbers).

Comparison of ADS and TFR daily results showed good agreement for total particulate sulfate, the sum of total (coarse plus fine) particulate and gaseous nitrate, and ammonia. As a result of different inlet particle collection efficiencies, the ADS fine particulate sulfate exceeded the TFR (5 percent). In the absence of a filter to collect volatilized particulate ammonium in the ADS, the sum of total particulate and gaseous ammonium in the TFR exceeded that in the ADS. Of potentially more importance, ADS measurements of SO₂ and H⁺ exceeded those of the TFR, while TFR measurements of HNO₃ exceeded those of the ADS. Results of this study suggest that the TFR may provide biased measurements of SO₂, H⁺, HNO₃, and Fine NO₃ ^- that cannot be corrected without modifications to the fundamental design of the sampling system.     

Evaluation of the filter pack for long-duration sampling of ambient air

A 14-week filter pack (FP) sampler evaluation field study was conducted at a site near Bondville, IL to investigate the impact of weekly sampling duration. Simultaneous samples were collected using collocated filter packs (FP) from two independent air quality monitoring networks (CASTNet and Acid-MODES) and using duplicate annular denuder systems (ADS). Precision estimates for most of the measured species are similar for weekly ADS and composited FPs. There is generally good agreement between the weekly CASTNet FP results aggregated from weekly daytime and weekly nighttime samples and those aggregated from daily 24 h Acid-MODES samples; although SO₂ is the exception, and CASTNet concentrations are higher than Acid-MODES. Comparison of weekly ADS results with composited weekly FP results from CASTNet shows good agreement for SO^2-₄. With the exception of the two weeks where the FP exceeded the ADS, both HNO₃ and the sum of particulate and gaseous NO^-₃ show good agreement. The FP often provides good estimates of HNO₃, but when used to sample atmospheres that have experienced substantial photochemical reactivity, FP HNO₃ determinations using nylon filters may be biased high. It is suggested that HNO₂ or some other oxidized nitrogen compound can accumulate on a regional scale and may interfere with the FP determination of HNO₃. FP particulate NO^-₃ results are in fair agreement with the ADS. Since FP SO₂ results are biased low by 12–20%, SO₂ concentration in the CASTNet data archive should be adjusted upward. Nylon presents problems as a sampling medium in terms of SO₂ recovery and specificity for HNO₃. Additional comparative sampler evaluation studies are recommended at several sites over each season to permit comprehensive assessment of the concentrations of atmospheric trace constituents archived by CASTNet.     

Performance of the Annular Denuder System with Different Arrangements for HNO3 and HNO2 Measurements in Taiwan

ABSTRACT

Experiments on different annular denuder system (ADS)arrangements for sampling nitrous acid (HNO₂) and ni-tric acid (HNO₃) gases were conducted in this study toevaluate their sampling artifacts. The evaluation basis isthe one that employed one sodium chloride denuder forsampling HNO₃ gas and two sodium carbonate (Na₂CO₃)denuders for sampling HNO₂ gas, which is a commonlyemployed ADS arrangement in many field applicationsin the United States. A field study was conducted inHsinchu, Taiwan, and the results indicated that this ADSarrangement may yield over 80% relative errors for HNO₃gas. It also showed that the relative errors for HNO₂ gascan be less than 10% as sampled with only one Na₂CO₃denuder. This is attributed to the fact that the ambientHNO₃ concentration measured in this study was relativelylow while the HNO₂ concentration was high, as comparedto typical concentrations of these two gases measured inthe United States.

The sampling error of HNO₃ gas may be due to highconcentrations of N-containing interfering speciespresent in Taiwan’s atmosphere. Because the relative sam-pling errors of HNO₃ and HNO₂ gases depend mainly ontheir concentrations in the atmosphere as well as con-centrations caused by interfering species, the risk for higherror while measuring low HNO₂ concentrations by onlyone Na₂CO₃ denuder is also possible. As a result, it is sug-gested that pretests are necessary to evaluate possiblesources and degrees of sampling errors before fieldsampling of HNO₂ and HNO₃ gases. The sampling errorsof these two gases can, therefore, be minimized with abetter arrangement of the ADS.     

The performance of a gas and aerosol monitoring system (GAMS) for the determination of acidic water soluble organic and inorganic gases and ammonia as well as related particles from the atmosphere

A new application of the quasi-simultaneous gas/particle phase sampling and analysis principle first proposed by Simon and Dasgupta (Anal. Chem. 34 (1995) 71) is described. For the first time, a gradient chromatograph is used in connection with such a sampling system to allow the simultaneous determination of major organic (formic, acetic, propionic, oxalic, malonic and succinic) and inorganic (SO₂, HNO₂, HNO₃, HCl and H₂F₂) acidic gases and related particles. Another addition to the previous systems is the analysis of cations other than ammonium from the particulate phase. The time resolution of the instrument still remains high, 1 h, during which gaseous water-soluble acidic compounds, ammonia, as well as related anionic particles and inorganic major cations are analysed. Sampling is based on diffusion in a wetted parallel plate denuder for gases and on growth in supersaturated water vapour for particles. The determination limits range from 2 ppt (acetate) to 0.4 ppb (ammonia) in the gas phase and 0.01 μg m⁻³ (citric acid) to 0.79 μg m⁻³ (calcium) for particulate matter. Collection efficiencies for gas and aerosol sampling were quantified and the supersaturation in the aerosol sampling apparatus investigated. The system has been used for field measurements at a background station; selected results of these measurements are presented.     

A Fabric Denuder for Sampling Semi-Volatile Species

ABSTRACT

A new style of diffusion denuder has been evaluated specifically for sampling HNO₃. A coated fabric is used as the denuder substrate, which can be loaded directly into a standard filter holder. This approach allows direct denuder sampling with no additional capital costs over filter sampling and simplifies the coating and extraction process.

Potential denuder materials and coatings were evaluated in the laboratory to test the removal efficiency. NaCl coatings were used to assess more than 20 materials for HNO₃ collection efficiency. Particle retention, which would cause a denuder to have a positive bias for gas concentration measurements, was evaluated by ambient air sampling using particulate sulfate as the reference aerosol. Particle retention varied from 0 to 15%, depending on the denuder material tested. The best performing material showed an average particle retention of less than 3%.

Denuder efficiency of four fabric materials was tested under ambient conditions to determine removal efficiency. The fabric denuder method was compared with a long path-length Fourier transform infrared (FTIR) spectrometer, a tunable diode laser absorption spectrometer (TDLAS), and a denuder difference sampler to independently measure HNO₃. HNO₃ collection efficiency was typically 90% for the denuders, whether coated with NaCl or not. For 10-L/min sampling rates with the fabric denuder, the square of the correlation coefficient with the FTIR spectrometer was 0.73, compared to 0.24 with the TDLAS.     

HNO3 losses within the cyclone inlet of a diffusion-denuder system under simulated marine environments

We evaluated the loss of HNO₃ within a Teflon-coated aluminum cyclone of an annular diffusion denuder atmospheric sampling system (ADS) under simulated marine conditions. To simulate marine environment, the cyclones were pre-coated with NaCl aerosol droplets. Loss of vapor-phase HNO₃ within the NaCl-coated cyclone was generally greater than 30% at relative humidities (RH) of 60 and 80% and as large as 67% when the cumulative HNO₃ dosages were lower than 3 μg. In contrast, there was little loss of HNO₃ (<8%) in cyclones with no NaCl coating at RHs ranging from 0 to 80%, at HNO₃ air concentrations of 4.3±1.6 μg m⁻³, and at cumulative HNO₃ dosages of greater than 5 μg. However, at lower HNO₃ cumulative dosages (<3 μg), losses in the non-coated cyclones were strongly influenced by RH, ranging from 9% in dry air to 58% at 80% RH. The enhanced loss of HNO₃ in the NaCl-coated cyclone was most likely caused by the reaction between HNO₃ and NaCl on the cyclone wall.     

Size mass distribution of water-soluble ionic species and gas conversion to sulfate and nitrate in particulate matter in southern Taiwan

A Micro-Orifice Uniform Deposition Impactor (MOUDI) and a Nano-MOUDI were employed to determine the size-segregated mass distributions of ambient particulate matter (PM) and water-soluble ionic species for particulate constituents. In addition, gas precursors, including HCl, HONO, HNO₃, SO₂, and NH₃ gases, were analyzed by an annular denuder system. PM size mass distribution, mass concentration, and ionic species concentration were measured during the day and at night during episode and non-episode periods in winter and summer. Average total suspended particle (TSP) concentrations during episode days in winter were as high as 153?±?33 μg/m³, and PM mass concentrations in summer were as low as one-third of that in winter. Generally, PM concentration at night was higher than that in the daytime in southern Taiwan during the sampling periods. In winter during the episode periods, the size-segregated mass distribution of PM mass concentration was mostly in the 0.32–3.2-μm range, and the PM concentration increased significantly in the range of 0.32–3.2 μm at night. Ammonium, nitrate, and sulfate were the dominant water-soluble ionic species in PM, contributing 34–48 % of TSP mass. High concentrations of ammonia (12.9–49 μg/m³) and SO₂ (2.6–27 μg/m³) were observed in the gas precursors. The conversion ratio was high in the PM size range of 0.18–3.2 μm both during the day and at night in winter, and the conversion ratio of episode days was 20 % higher than that of non-episode days. The conversion factor was high for both nitrogen and sulfur species at nighttime, especially on episode days.     

Precision and Accuracy in the Determination of Sulfur Oxides,Fluoride, and Spherical Aluminosilicate Fly Ash Particles in Project MOHAVE

The precision and accuracy of the determination of particu-late sulfate and fluoride, and gas phase SO₂ and HF are estimated from the results obtained from collocated replicate samples and from collocated comparison samples for high-and low-volume filter pack and annular diffusion denuder samplers. The results of replicate analysis of collocated samples and replicate analyses of a given sample for the determination of spherical aluminosilicate fly ash particles have also been compared. Each of these species is being used in the chemical mass balance source apportionment of sulfur oxides in the Grand Canyon region as part of Project MOHAVE, and the precision and accuracy analyses given in this paper provide input to that analysis. The precision of the various measurements reported here is ±1.8 nmol/m³ and ±2.5 nmol/m³ for the determination of SO₂ and sulfate, respectively, with an annular denuder. The precision is ±0.5 nmol/m³ and ±2.0 nmol/m³ for the determination of the same species with a high-volume or low-volume filter pack. The precision for the determination of the sum of HF(g) and fine particulate fluoride is ±0.3 nmol/m³. The precision for the determination of aluminosilicate fly ash particles is ±100 particles/m³. At high concentrations of the various species, reproducibility of the various measurements is ±10% to ±14% of the measured concentration. The concentrations of sulfate determined using filter pack samplers are frequently higher than those determined using diffusion denuder sampling systems. The magnitude of the difference (e.g., 2-10 nmol sulfate/m³) is small, but important relative to the precision of the data and the concentrations of particulate sul-fate present (typically 5-20 nmol sulfate/m³). The concentrations of SO₂(g) determined using a high-volume cascade impactor filter pack sampler are correspondingly lower than those obtained with diffusion denuder samplers. The concentrations of SO_x (SO₂(g) plus particulate sulfate) determined using the two samplers during Project MOHAVE at the Spirit Mountain, NV, and Hopi Point, AZ, sampling sites were in agreement. However, for samples collected at Painted Desert, AZ, and Meadview, AZ, the concentrations of SO_x and SO₂(g) determined with a high-volume cascade impactor filter pack sampler were frequently lower than those determined using a diffusion denuder sampling system. These two sites had very low ambient relative humidity, an average of 25%. Possible causes of observed differences in the SO₂(g) and sulfate results obtained from different types of samplers are given.     

A theoretical evaluation on the HNO3 artifact of the annular denuder system due to evaporation and diffusional deposition of NH4NO3-containing aerosols

A mathematical model was developed to evaluate HNO₃ artifact of the annular denuder system due to evaporation and diffusional deposition of nitrate-containing aerosols. The model performance was validated by comparing its numerical solutions with laboratory and numerical data available in the literature for evaporation and diffusional deposition of monodisperse and polydisperse NH₄NO₃ aerosols. Measurement artifacts were evaluated by varying typical sampling ranges of ambient temperature, HNO₃ gas concentration, aerosol number concentration, aerosol mass median diameter, and nitrate mass fraction of <2.5 μm aerosols to see their respective effects. Potential application of the present model on estimating HNO₃ artifacts was demonstrated using literature data sampled in USA, Taiwan, Netherlands, Korea and Japan. Significant measurement artifact could be found in Taiwan and Netherlands due either to low HNO₃ gas concentration and high nitrate concentration in <2.5 μm aerosols or to high ambient temperature.     

Laboratory and field investigations of a new and simple design for the parallel plate denuder

Diffusion denuders are commonly used for collection of water soluble atmospheric gases. Denuder plates made of glass with silica coating were however difficult to handle in field measurements, and also had wettability problems. In this paper, we describe a new design of parallel plate denuder consisting of perfluoralkoxy (PFA) plates and polyester sheets. The parallel plate denuder was coupled on-line to ion chromatography for analysis of atmospheric HNO₃ and SO₂. With time resolution of 20 min, the detection limits for both gases were 6 ppt(v). Memory effect was found to be significant for HNO₃ especially at high humidity with the long sample inlet tube. Particle losses due to diffusion and electrostatic effects in the denuder were investigated and compared with theory. Particle loss in the sample inlet tube was found to play an important role by reducing the number of charged particles entering the denuder. The instrument was applied in two European Union field measurement campaigns, and the intercomparison results are presented in this paper.     

Determination of PM2.5 Sulfate and Nitrate with a PC-BOSS Designed for Routine Sampling for Semi-Volatile Particulate Matter

ABSTRACT

Ambient particles contain substantial quantities of material that can be lost from the particles during sample collection on a filter. These include ammonium nitrate and semi-volatile organic compounds. As a result, the concentrations of these species are often significantly in error for results obtained with a filter pack sampler. The accurate measurement of these semi-volatile fine particulate species is essential for a complete understanding of the possible causes of health effects associated with exposure to fine particles. Past organic compound diffusion denuder samplers developed by the authors (e.g., the Brigham Young University Organic Sampling System [BOSS]) are not amenable to routine field use because of the need to independently determine the gas-phase semi-volatile organic material efficiency of the denuder for each sample. This problem has been eliminated using a combined virtual impactor, particle-concentrator inlet to provide a concentrated stream of 0.1-2.5-μm particles. This is followed by a BOSS diffusion denuder and filter packs to collect particles, including any semi-volatile species lost from the particles during sampling. The samp ler (Particle Concentrator-Brigham Young University Organic Sampling System [PC-BOSS]) contains a post-denuder multifilter pack unit to allow for the routine collection of several sequential samples. The PC-BOSS can be used for the determination of both fine particulate nitrate and semi-volatile organic material without significant “positive” or “negative” sampling artifacts. Validation of the sampler for the determination of PM_2.5 sulfate and nitrate based on comparison of results obtained at Riverside, CA with collocated PC-BOSS, annular denuder, and Chem Spec samplers indicates the PC-BOSS gives accurate results for these species with a precision of ±5-8%. An average of 33% of the PM_2.5 nitrate was lost from the particles during sampling for both denuder and single filter samplers.     

Seasonal variations in atmospheric SOx and NOy species in the adirondacks

This paper reports the results of over 2 years of measurements of several of the species comprising atmospheric SO_x (=SO₂+SO₄²⁻) and NO_y (=NO+NO₂ + PAN + HNO₃+NO₃⁻+ organicnitrates + HONO + 2N₂O₅ …) at Whiteface Mountain, New York. Continuous real-time measurements of SO₂ and total gaseous NO_y provided data for about 50% and 65% of the period, respectively, and 122 filter pack samples were obtained for HNO₃, SO₂ and aerosol SO₄²⁻, NO₃⁻, H⁺ and NH₄⁺. Concentrations of SO₂ and NO_y were greatest in winter, whereas concentrations of the reaction products SO₄²⁻ and HNO₃were greatest in summer. The seasonal variation in SO₄²⁻ was considerably more pronounced than that of HNO₃and the high concentrations of SO₄²⁻ aerosol present in summer were also relatively more acidic than SO₄²⁻ aerosol in other seasons. As a result, SO₄²⁻ aerosol was the predominant acidic species present in summer, HNO₃was predominant in other seasons. Aerosol NO₃⁻ concentrations were low in all seasons and appeared unrelated to simultaneous NO_y and HNO₃concentrations. These data are consistent with seasonal variations in photochemical oxidation rates and with existing data on seasonal variations in precipitation composition. The results of this study suggest that emission reductions targeted at the summer season might be a cost-effective way to reduce deposition of S species, but would not be similarly cost-effective in reducing deposition of N species. kwAcid deposition, seasonal variation, sulfate, nitrate, nitric acid, sulfur dioxide, oxides of nitrogen, hydrogen peroxide, ozone, air pollution, Adirondack Mountains     

Causes of the elevated nitrate aerosol levels during episodic days in Taichung urban area,Taiwan

The purpose of this study is to explore the possible reasons accounting for elevated nitrate aerosol levels during high particulate days (HPD) in Taichung urban area of central Taiwan. To achieve this goal, simultaneous measurements of particulate and gaseous pollutants were carried out from September 2004 to April 2005 using an annular denuder system (ADS). The formation rate of NO₂ to nitrate aerosol, calculated using the relevant chemical reactions, was employed to interpret enhanced nitrate aerosol concentrations during HPD. The observations showed that nitrate concentration during HPD was 14 times higher than that during low particulate days (LPD). The average formation rate during HPD was 4.0% h^?1, which was 3.1 times higher than that during LPD. The quantitative analysis showed that the formation rate was mainly influenced by temperature and relative humidity. Lower temperature and higher relative humidity led much nitrate aerosol formation in HPD. Moreover, the residence time analysis of air masses staying over the studied area showed that the slow-motion air retained high nitrate concentrations due to more nitrate aerosol converted from the precursors in NOx-rich areas.     

A comparison of filter,denuder, and real-time chemiluminescence techniques for nitric acid determination in ambient air

Measurements of gas-phase nitric acid were made by four separate techniques during a 7-day summertime period at a near-coastal site on Long Island, NY. Results from methods intercomparison data for HNO₃, and their relationship to particulate NO₃⁻ and other odd N and oxidant species show the following: (a) high-volume filter pack HNO₃ concentrations are well correlated with diffusion denuder difference (DD) results, except for small absolute losses with the former; (b) daytime real-time two-channel chemiluminescence HNO₃ levels correlated well with DD results, but were higher during night-time periods; (c) results by a new Al₂(SO₄)₃ denuder/thermal evolution technique were not in agreement with other techniques. Based on HNO₃ and paniculate NO₃ results reported herein, it appears that negative errors in HNO₃filter-sampling techniques resulting from HNO₃ loss by sorption generally exceed positive errors NH₄NO₃ volatilization at this site.     

Investigation of the positive artifact formation during sampling semi-volatile aerosol using wet denuders

Wet denuders are used in several steam-based semi-continuous aerosol monitors to avoid gaseous absorption artifacts and pre-humidify the air stream, while simultaneously allowing measurements of water-soluble gaseous species. Unlike dry denuders, wet denuders saturate the sample air stream with water vapor, which can lead to re-partitioning of water-soluble volatile species to the aerosol phase, thereby causing a positive artifact in aerosol measurements. This paper investigates the magnitude of the positive artifact formation occurring in wet denuders using modeling techniques. Gaseous nitric acid was used as an example of volatile water-soluble gas in both flat and annular wet denuders. We have also verified the occurrence of the positive artifact in a flat wet denuder through a laboratory experiment. The model results indicate that the magnitude of the artifact is rather limited under typical conditions being less than 2.5% of ambient nitric acid concentration for the flat denuder and less than 0.6% for the annular denuder. The magnitude of the artifact increases with condensational sink of the aerosol (i.e. with the mean aerosol size and number concentration) and aerosol water solubility. While the artifact is relatively small in the absolute sense, it could be substantial for aerosol nitrate measurements, especially in ammonia limited conditions, when the concentration of the nitric acid is high and the concentration of nitrate is low. Therefore, we recommend that the artifact is assessed regularly by replacing the wet denuder with a dry denuder.     

Seasonal variations of acidic air pollutants in Seoul,South Korea

The annular denuder system (ADS) was used to characterize seasonal variations of acidic air pollutants in Seoul, South Korea. Fifty- four 24 h samples were collected over four seasons from October 1996 to September 1997. The annual mean concentrations of HNO₃, HNO₂, SO₂ and NH₃ in the gas phase were 1.09, 4.51, 17.3 and 4.34 μg m^-3, respectively. The annual mean concentrations of PM_2.5(d_p≤2.5 μm in aerodynamic diameter, 50% cutoff), SO^2-₄, NO^-₃ and NH⁺₄ in the particulate phase were 56.9, 8.70, 5.97 and 4.19 μg m^-3, respectively. All chemical species monitored from this study showed statistical seasonal variations. Nitric acid (HNO₃) and ammonia (NH₃) exhibited substantially higher concentrations during the summer, while nitrous acid (HNO₂) and sulfur dioxide(SO₂) were higher during the winter. Concentrations of PM_2.5, SO^2-₄, NO^-₃ and NH⁺₄ in the particulate phase were higher during the winter months. SO^2-₄, NO^-₃ and NH⁺₄ accounted for 26–38% of PM_2.5. High correlations were found among PM_2.5, SO^2-₄, NO^-₃ and NH⁺₄. The mean H⁺ concentration measured only in the fall was 5.19 nmole m^-3.     

Simultaneous Sampling of Gas- and Aerosol-Phase TDI with a Triple Filter System

Abstract

A new triple filter system sampler/model is proposed for the precise and accurate simultaneous sampling and determination of gas- and aerosol-phase 2,4-toluene diisocyanate (TDI). The system consists of two front Teflon filters for sampling aerosol-phase TDI and a final coated glass fiber filter to collect gas-phase TDI. The aerosol-phase TDI is collected on the first Teflon filter, while the second Teflon filter is used to estimate gaseous TDI adsorbed by the first. According to the gas adsorption test of two Teflon filters in series, the TDI gas adsorption fraction of the two filters is almost the same. Results of the evaporation test using pure TDI aerosols collected on the Teflon filter show that significant evaporation of the compound does not occur during sampling. These two findings allow the use of a model to estimate accurate gasand aerosol-phase TDI concentrations. The comparison test with an annular denuder shows that the triple filter system can minimize the TDI sampling bias between the dual filter and the annular denuder systems.