Development and Application of a Needle Trap Device for Time

Anal. Chem. 2008, 80, 7275–7282

Development and Application of a Needle Trap Device for Time-Weighted Average Diffusive Sampling Ying Gong,† In-Yong Eom,† Da-Wei Lou,†,‡ Dietmar Hein,§ and Janusz Pawliszyn*,† Department of Chemistry, University of Waterloo, Waterloo, Ontario, N2L 3G1 Canada, and PAS Technology, Magdala, Germany A simple, cost-effective analysis combining solventless extraction, thermal desorption, and determination of volatile organic compounds (VOCs) was developed and validated. A needle trap device (NTD) packed with the sorbent Carboxen1000 was used as a time-weighted average (TWA) diffusive sampler to collect target compounds by molecular diffusion and adsorption to the packed sorbent. This process can be described with derivations of Fick’s first law of diffusion, which expresses the relation between the TWA concentrations to which the passive sampler is exposed and the mass of analytes adsorbed to the packed sorbent in the sampler. The effects of experimental factors such as temperature, pressure, humidity, and face velocity were taken into account in applying diffusive sampling under nonideal conditions. This study demonstrates that NTD is effective for air analysis of benzene, toluene, ethylbenzene, and o-xylene (BTEX), due to the good adsorption/desorption quality of Carboxen 1000 and to the special geometric shape of the needle with a small cross section avoiding the need for calibration. Storage tests showed good storage stability for BTEX. Verification of the theoretical model showed good agreement between theoretical and experimental sampling rates. Method validation done against NIOSH method 1501, SPME, and NTD active sampling revealed good agreement between those methods. Automated NTD sample introduction to a gas chromatograph facilitates the use of this technology for industrial hygiene applications.

sampling methods to monitor VOCs at trace levels in the workplace or in ambient air. Conventional sampling methods involve active and passive sampling by using various samplers, such as the diffusive dosimeter, activated carbon cloth, badges and tubes, as well as commercial active and diffusive samplers manufactured by PerkinElmer, Supelco, or other companies.4,5 Each of these methods is based on the sorption of the target compounds by a solid or liquid sorbent, followed by either solvent extraction or thermal desorption prior to analysis by gas chromatography. The needle trap device (NTD), which can be used as a timeweighted average (TWA) diffusive sampler, is an extraction trap containing a sorbent within a needle.6 There are two approaches for NTD to obtain a TWA concentration.7 One is to determine the concentrations of a large number of samples obtained at different time intervals and then to average the concentrations over total sampling time. The alternative is to collect one sample for a certain sampling period with one sampler. The latter approach is preferable because the number of samples and analyses is considerably reduced. Some earlier investigations have mentioned that a NTD can be used for TWA diffusive sampling but provided little experimental support for this conclusion.8 In this study, a series of systematical theoretical and experimental investigations were conducted to show that NTD with Carboxen 1000 is an effective diffusive sampler for BTEX analysis and that it is not significantly affected by environmental factors such as temperature, pressure, face velocity, etc.

Volatile organic compounds (VOCs) are found as contaminants in both indoor air and the environment. The presence of these compounds has resulted in various health problems and has created the need to determine and monitor VOCs in air.1-3 Therefore, it is necessary to develop accurate and convenient

EXPERIMENTAL SECTION Chemicals and Materials. Benzene, toluene, ethylbenzene, o-xylene, Carboxen 1000 and gold wire (0.1 mm diameter) were purchased from Sigma-Aldrich (Ontario, Canada). All 22-gauge needles were purchased from Dyna Medical Corporation (Ontario, Canada). All side holes (0.4 mm i.d.) and personal sampler holders were made by the machine shop at the University of Waterloo. Luer-lock type 22-gauge needles were provided by the PAS technology (Magdala, German). The timer and digital humidity

* Corresponding author. Phone: +1-519-888-4641. Fax: +1-519-746-0435. E-mail: [email protected]. † Department of Chemistry, University of Waterloo. ‡ Current address: Jilin Institute of Chemical Technology, Jilin, Jilin, 132022, China. § PAS Technology. (1) Elke, K.; Jermann, E.; Begerow, J.; Dunemann, L. J. Chromatogr., A 1998, 826, 191–200. (2) Shojania, S.; Oleschuk, R. D.; McComb, M. E.; Gesser, H. D.; Chow, A. Talanta 1999, 50, 193–205. (3) McComb, M. E.; Oleschuk, R. D.; Giller, E.; Gesser, H. D. Talanta 1997, 44, 2137–2143. 10.1021/ac800884f CCC: $40.75  2008 American Chemical Society Published on Web 09/03/2008

(4) Gesser, H. D.; Giller, E. Environ. Int. 1995, 21 (6), 839–844. (5) Brown, R. H.; Wright, M. D. Analyst 1994, 119, 75–77. (6) Koziel, J. A.; Odziemkowski, M.; Pawliszyn, J. Anal. Chem. 2001, 73, 47– 54. (7) Martos, P. A.; Pawliszyn, J. Anal. Chem. 1999, 71, 1513–1520. (8) Wang, A.; Fang, F.; Pawliszyn, J. J. Chromatogr., A 2005, 1072, 127–135. (9) Koziel, J.; Jia, M.; Pawliszyn, J. Anal. Chem. 2000, 72, 5178–5186.

Analytical Chemistry, Vol. 80, No. 19, October 1, 2008

7275

meter were purchased from VWR International (Ontario, Canada). Hydrogen, nitrogen, helium, compressed air, and carbon dioxide were obtained from Praxair Canada Incorporated (Ontario, Canada). The 5-min epoxy glue was purchased from Henkel Canada (Ontario, Canada). The narrow-neck liner was purchased from SGE Analytical Science (Texas). Permeation tubes were purchased from Kin-Tech Laboratories (Texas). The in-line impinger trap, the 75 µm Carboxen/PDMS fiber and holder, ORBO-32 tubes, gas purifiers, syringes, Thermgreen septa, and vials were purchased from Supelco (Ontario, Canada). Personal air pumps and Mini-buck calibrator were purchased from A. P. Buck (Florida). Standard Gas Generator. A standard-gas generator (model 491M-B, Kin-Tech Laboratories, Texas) was used to generate the benzene, toluene, ethylbenzene, and o-xylene (BTEX) standard gas concentrations. The traceable permeation tubes were placed inside a glass cylinder held in a temperature-controlled oven and swept with a constant flow of compressed air. Different concentrations of BTEX were achieved by adjusting both permeation oven temperature and dilution air flow rate. A long glass cylinder with three different diameters was installed downstream from the main chamber to obtain BTEX at different velocities. An in-line impinger trap was connected to the standard gas system to generate standard gases with different levels of humidity.9 The digital humidity meter was used to measure relative humidity and temperature. Preparation of Needle Trap Device. The NTD consisted of a 22-gauge needle and a sorbent positioned at a distance Z from the tip of the needle (Figure 1, part A). Carboxen 1000 (mesh size 60/80 and surface area 1200 m2/g) was used as a sorbent in this study. The procedure to prepare a NTD is as follows: a metal wire was used to indicate the position inside the needle where the sorbent would be packed; about 1 cm of gold wire was tightly coiled into a 22-gauge needle and fixed in the required position; then an aspirator was used to aspirate the sorbent about 0.6 mg (packing density around 0.45 g/mL) into the needle until it came to the required position. A small quantity of 5-min epoxy glue was used to immobilize the sorbent; to prevent the glue from blocking the needle, the aspirator was connected to the needle to draw air through the NTD before the glue solidified. The NTD was conditioned in a GC injector at 300 °C for 2 h to remove impurities. Finally, the two ends and side hole were sealed with Teflon caps and tubing. Six luer-lock type needle trap devices (nominal z-paths of 3 mm) were prepared as explained above, but there was no side hole. Instrumentation. A Varian CP-3800 gas chromatograph (GC) equipped with a Varian 2000 mass spectrometer (Ontario, Canada) was used to separate and analyze the target compounds extracted by the needle trap device. A narrow-neck glass liner (i.d. 0.8 mm) was connected to a VF-5 ms GC capillary column (30 m × 0.25 mm, 0.25 µm film thickness). Figure 1 (part B) shows a schematic diagram of the system. The injector was maintained at 300 °C. The helium carrier gas entered the side hole of the NTD, flowed through the sorbent, and carried the desorbed compounds into the GC column. The column temperature program was maintained at 50 °C for 1.5 min and then ramped at 20 °C/min to 120 °C and held at that temperature for 1 min. The carrier gas flow rate was set at 2.5 7276


Figure 1. (A) Schematic of a needle trap device diffusive sampler. (B) Schematic of a gas chromatograph, needle trap device, narrowneck glass liner, capillary column, and mass spectrometer.

mL/min. The compounds were then separated by the GC column and detected by mass spectrometry. For automation analysis, the NTD-TWA-autosampler was integrated on a GC/FID (Acme 6100 GC, Young Lin Instrument, Republic of Korea) equipped with a nonpolar capillary column (HR1, 30 m × 0.32 mm, i.d. 0.25 µm, Shinwa Chemical Industries Ltd., Japan). A programmable GC injector (OPTIC 3, ATAS GL, Veldhoven, The Netherlands) kept a carrier gas flow rate at 3.5 mL/min and controlled the injector temperature as programmed. Figure 2 shows the schematic of a newly developed GC-autosampler (Concept, PAS Technology, Germany) and its flow configuration for desorption of analyte trapped in a luer-lock NTD. Luerlock NTDs were assembled manually with luer-lock adapters on a NTD sampler tray after sampling. The NTDs were transferred one by one to the GC injector by the magnetic arm, and then the closure (pneumatic arm) closed the NTD to provide a carrier gas through a lure-lock needle head (Figure 2b). After all preset parameters (carrier gas flow rates, temperature of the injector, etc) were ready, the control software, CONCEPT 1.1, started the programmable GC injector and the GC data acquisition simultaneously. The initial temperature of the GC injector was kept at 20 °C to minimize sample loss before the injector ramping (16 °C/min) up to 300 °C. Because of the high temperature of the injector, a cryogenic refocusing trap (ATAS GL, Veldhoven, The

which exposes the needle trap sampler both continuously and intermittently to the standard gas chamber.13 For continuous exposure, the NTD was exposed to the standard gas chamber for 60 min; for intermittent exposure, it was exposed to the standard gas chamber for 20 min and then to clean air for 20 min. This process was repeated three times. If the sorbent were to exhibit adequate sorption efficiency, the mass of each analyte on the sorbent would be the same for both continuous and intermittent exposures. Three NTDs with different diffusion lengths 0.37, 0.48, and 0.65 cm, respectively, were used to perform the zero sink tests. The average mass percentages of intermittent to continuous exposure were 99%, 101%, 103%, and 98% individually. The results indicate that Carboxen 1000 behaves like a zero sink for BTEX, presumably because of its strong affinity and capability. The second prerequisite is that the NTD should have a short response time to changes in sample concentration at the face of the sampler. Fick’s first law assumes steady state conditions. In practice, the ambient analyte level is likely to change considerably over time. As has been previously demonstrated, the diffusive needle trap sampler produces a truly integrated response and does not miss short-lived transients before the analytes are trapped by the sorbent.10,14,15 Response time, that is, how long it takes for a molecule to diffuse into the sorbent from the needle tip, is calculated by the following equation:16

t) Figure 2. Schematic of a GC-autosampler in the desorption mode for NTDs (a) and a more detailed schematic description of the juction of adapter and luer-lock head of the NTD. The PTFE disk (∼100 µm thick) is placed in an adapter for sealing purposes.

Netherlands) was placed right after the GC injector to trap the desorbed analytes back to the sorbent. The instrument was checked daily for calibration using a midpoint standard gas mixture of BTEX. Any deviation in area counts greater than 15% required reinjection of that standard; if the deviations were still great than 15%, the instrument was calibrated with a five-point calibration curve. RESULTS AND DISCUSSION Three Prerequisites for Applying Fick’s First Law. The process of collecting target analytes using a needle trap diffusive sampler can be described in derivation of Fick’s first law.7,10-12 In order to achieve successful diffusive TWA sampling, three prerequisites must be satisfied. First, the sorbent inside the NTD should be zero sink for the target analytes. The sampling rate of a diffusive needle trap sampler depends on the sorption efficiency of the sorbent. For a sorbent with high sorption efficiency, the concentration of analyte adsorbed at the sorbent surface, Csorbent, will be very small compared with the concentration at the needle tip, Cface. In this situation, it is assumed that the sorbent is a zero sink for the adsorbed analyte. Zero sink ensures that once the analyte is adsorbed, it does not change the mass loading rate of additional analyte. Zero sink can be validated by a simple test, (10) Hearl, F. J.; Manning, M. P. Am. Ind. Hyg. Assoc. J. 1980, 41 (11), 778– 783. (11) Young, M. S.; Monat, J. P. Am. Ind. Hyg. Assoc. J. 1982, 43 (12), 890–896. (12) Rose, V. E.; Perkins, J. L. Am. Ind. Hyg. Assoc. J. 1982, 43 (8), 605–621.

Z2 2D

(1)

where t is the response time (s), Z is the length of the diffusion path (cm), and D is the analyte diffusion coefficient (cm2/s). In this study, which uses a NTD whose diffusion length is less than 1.0 cm and where the smallest analyte diffusion coefficient among BTEX is 0.073 cm2/s, the response time should be less than 7.0 s.17 Obviously, the sampling time of a diffusive needle trap sampler is much longer than the response time, which ensures that the sample collected represents a real TWA concentration. The NTD must meet the third prerequisite, that the bulk analyte concentration, Cbulk, must equal the analyte concentration at the tip of the needle, Cface. The sampling rate of the diffusive needle trap sampler relies on diffusion length, Z, the crosssectional area, A, and the diffusion coefficients of the target analytes. Theoretically, the diffusion length is the distance between the sorbent surface and the external face of the needle tip. In practice, however, the effective diffusion length is not identical to the theoretical length and may be greater or less depending on circumstances.18,19 When face velocity is low, there may be insufficient air to carry analyte molecules from the site of removal by diffusion to the (13) Chen, Y.; Pawliszyn, J. Anal. Chem. 2003, 75, 2004–2010. (14) Brown, R. H.; Charlton, J.; Saunders, K. J. Am. Ind. Hyg. Assoc. J. 1981, 42, 865–869. (15) Hori, H.; Tanaka, I. Am. Ind. Hyg. Assoc. J. 1993, 54, 95–101. (16) Lautenberger, W. J.; Kring, E. V.; Morello, J. A. Am. Ind. Hyg. Assoc. J. 1980, 41 (10), 737–747. (17) Rowe, R. K.; Mukunoki, T.; Sangam, H. P. J. Geotech. Geoenviron. Eng. 2005, 131 (10), 1211–1221. (18) Underhill, D. W.; Feighley, C. E. Anal. Chem. 1991, 63, 1011–1013. (19) Fan, Z.; Jung, K. H.; Lioy, P. J. Environ. Sci. Technol. 2006, 40 (19), 6051– 6057.


7277

Table 1. Sampling Rates of BTEX with Carboxen 1000 at Different Velocities, Average Sampling Rates, and RSDs sampling rate, mL/min v ) 1.44 (cm/min) v ) 4.29 (cm/min) v ) 32.48 (cm/min) average RSD (%)

benzene

toluene

ethylbenzene

o-xylene

0.0068 0.0066 0.0067 0.0067 1.5

0.0065 0.0065 0.0065 0.0065 0.3

0.0062 0.0061 0.0061 0.0061 1.5

0.0055 0.0056 0.0058 0.0056 2.0

sampler surface. In this case, the effective diffusion length may be higher than the theoretical length due to the occurrence of a boundary layer between the static air within the needle and the moving air outside, which contributes to the effective diffusion length. The boundary layer decreases with increased face velocity.20 Thus, diffusive sampling with a minimum face velocity is required; once minimum face velocity is satisfied, the sampling rate will not be sensitive to velocity over a wide range.13 When face velocity is high, the static air layer within the needle may be disturbed. It results in a decrease in effective diffusion length. Thus, a sampler with a small cross section and long nominal diffusion length will be relatively unaffected by air velocity, provided that the sampler’s ratio of diffusion length to diameter is greater than 2.5-3.21 Because the needle used in this study had a diameter of 0.04 cm and the nominal diffusion length was greater than 0.37 cm, the ratio (0.37/0.04 ) 9.2) was larger than 2.5-3.0. Therefore, high face velocity was not a factor. In order to investigate the effect of face velocity on sampling rate, the NTD (Z ) 0.93 cm) was inserted into the sampling chamber, perpendicular to the gas flow direction, and sampled for 1 h at 25 °C prior to thermal desorption and GC/MS analysis. BTEX sampling rates were determined at 1.44, 4.29, and 32.48 cm/min, respectively. Table 1 summarizes the experimental results for sampling rates of BTEX at different face velocities. These results indicate that there was no significant impact of face velocity on BTEX sampling rates within the given range. In reality, face velocities in both indoor and outdoor are larger than 1.44 cm/min.22,23 Therefore, the results of this study demonstrate that the NTD with Carboxen 1000 can be used for TWA diffusive sampling of BTEX without having to take into account the face velocity. Effect of Environmental Factors on Sampling Performance. Before use of the NTD diffusive sampler in field sampling, it was necessary to determine the effects of environmental conditions like temperature, pressure, and relative humidity on its performance. For a diffusive sampler, the average concentration of analyte during the exposure period is calculated by integration of Fick’s first law10

C)

nZ DAt

7278


diffusion path, n is the number of moles of target analyte (mol), A is the diffusion path cross-sectional area (cm2), and t is the duration (s) of the sampling period. The mass loading rate of diffusive sampler, U, can be derived from eq 2 U)

n DAC ) t Z

(2)

(3)

The diffusion coefficient of a given compound depends substantially on temperature and pressure, a relationship shown by eq 424,25 3

KT 2 D) P

(4)

where K is a constant related to molecular weight and volume, T is the ambient temperature (K), and P is absolute pressure (Pa). The concentration of the given compound can be derived from the ideal gas law, C ) P/(RT). Therefore, the conclusion can be drawn U∝T1/2, namely, the mass loading rate is proportional to the square root of the temperature and independent of pressure. This mathematical relationship between mass loading rate and temperature was proved by determining the mass loading rate at 25, 30, and 35 °C. The experimental results, presented in Figure 3, show that BTEX mass loading rates increased only slightly with increasing temperature. This trend is consistent with the theoretical prediction. In practice, the effect of temperature can be eliminated by calibrating the diffusion coefficient in accordance with the sampling temperature. The diffusion coefficients of many VOCs in air at 25 °C are known. If sampling is performed at a different temperature, T (K), the coefficient for this temperature can be adjusted using the following equation26 T 298

( )

DT ) Dg

where C is the analyte concentration (mol/cm3), D is the analyte diffusion coefficient (cm2/s), Z is the length (cm) along the (20) Brown, R. H. J. Environ. Monit. 2000, 2, 1–9. (21) Coleman, S. R. Am. Ind. Hyg. Assoc. J. 1983, 44, 929–936.

Figure 3. Effect of temperature on diffusive mass loading rates of BTEX using the NTD with Carboxen 1000 (Z ) 0.45 cm).

3 2

(5)

(22) Teitel, M.; Tanny, J. Flow, Turbul. Combust. 2005, 74, 21–47. (23) Baldwin, P. E. J.; Maynard, A. D. Ann. Occup. Hyg. 1998, 42 (5), 303–313. (24) Paganelli, C. V.; Ar, A.; Rahn, H.; Wangensteen, O. D. Respir. Physiol. 1975, 25, 247–258. (25) Piiper, J.; Worth, H. Respir. Physiol. 1980, 41, 233–240. (26) Batterman, S.; Metts, T.; Kalliokoski, P. J. Environ. Monit. 2002, 4, 870– 878.

Table 2. Comparison between the Experimental and Calibrated Mass Loading Rates of BTEX at Different Temperatures 25 °C

30 °C

mass loading rate (ng/min)



experimental

calibrated

experimental

calibrated

experimental

calibrated

0.0693 0.0723 0.0285 0.0309

0.0704 0.0714 0.0290 0.0305

0.0707 0.0727 0.0296 0.0311

0.0722 0.0732 0.0298 0.0313

0.0717 0.0737 0.0310 0.0322

0.0740 0.0750 0.0305 0.0321

benzene toluene ethylbenzne o-xylene

Table 3. Comparison of BTEX Sampling Rates with Carboxen 1000 between Experimental (at RH 17%, 42%, and 53%) and Theoretical Sampling Rates sampling rate (mL/min) RH, 17% RH, 42% RH, 53% theoretical

35 °C

benzene

toluene

ethylbenzene

o-xylene

0.016 0.016 0.015 0.016

0.016 0.016 0.015 0.015

0.015 0.015 0.013 0.013

0.013 0.012 0.012 0.013

where DT is the diffusion coefficient (cm2/s) at temperature T, Dg is the diffusion coefficient (cm2/s) at temperature 298 K, and T is the temperature (K). Table 2 lists comparisons between experimental and calibrated mass loading rates of BTEX, calculated using calibrated diffusion coefficients at specific sampling temperature of 25, 30, and 35 °C. No significant difference was observed from the experimental results at different temperatures; therefore, the temperature effects on mass loading rates of BTEX can be eliminated by calibrating the diffusion coefficients of BTEX to the sampling temperature. The effect of relative humidity was also investigated. A series of experiments was performed to determine sampling rates at different relative humidities including 17%, 42%, and 53%. The results, shown in Table 3, indicate that at a relative humidity less than 45%, the sampling rates of BTEX were very close; while higher than 45%, the sampling rates of BTEX were slightly affected by relative humidity. Compared with theoretical sampling rates of BTEX, the biggest difference was found between experimental and theoretical sampling rate of ethylbenzene; and this difference resulted in 14% of relative error of the result, which is within the acceptable range of ±20% suggested by U.S. Environmental Protection Agency Compendium Method TO-17.27 Therefore, there is no significant impact of relative humidity on the sampling rates of BTEX within this range. The possible reason is that Carboxen 1000, which behaves as a hydrophobic sorbent, was far from saturation during TWA diffusive sampling; so the competition between the target analytes and water molecules for occupying the micropores was not strong enough to affect the adsorption of target analytes.13 Effect of Sampling Duration on Sampling Performance. The mass loading rate of a diffusive sampler may be constant for a period of exposure time. As analyte loading increases, the mass loading rate may decrease with increasing sampling time. At this point, a small quantity of analyte may accumulate on the surface (27) U.S. Environmental Protection Agency. Determination of Volatile Organic Compounds in Ambient Air Using Active Sampling onto Sorbent Tubes; Compendium Method TO-17, 2nd ed.; EPA/625/R-96/010b; Office of Research and Development, Cincinnati, OH, 1999.

of the sorbent and result in a nonlinear increase of adsorbed mass. Finally, the adsorbed mass will stop increasing and breakthrough will occur. The effect of sampling duration on mass loading rates of BTEX was investigated by exposing a NTD (with Carboxen 1000 and a diffusion length of 0.52 cm) to the BTEX gas mixture in the sampling chamber for diffusive sampling times ranging from 1 to 95 h. BTEX concentrations in the sampling chamber were 4.94, 5.52, 2.21, and 2.31 ng/mL, respectively. The mass of BTEX adsorbed onto the NTD increased linearly (R2 ranged from 0.990 to 0.992). The mass linear ranges of BTEX of the sampling duration were 3.9-35.4, 3.4-32.8, 1.1-12.2, and 1.6-13.7 ng, respectively. These results indicate that Carboxen 1000 acts as a zero sink for BTEX over a sampling duration of 1-95 h. Storage Stability and Reusability. Storage stability is very important for field TWA sampling. If storage is unstable, analytes adsorbed inside the sampler may be lost, resulting in experimental error. The storage stability of the NTD with Carboxen 1000 for BTEX compounds was tested by active sampling. The NTD was used to sample a 4.0-mL sample of BTEX from the standard gas system, which was then injected into the GC/MS and analyzed. The mass of BTEX were 17.98, 19.68, 9.34, and 10.08 ng, respectively. Next, the same 4.0-mL volume of BTEX was sampled and sealed into the NTD with Teflon tips and tubing for 24 h, after which it was injected into the GC/MS and analyzed. The determined BTEX mass were 17.61, 19.40, 9.19, and 10.06 ng, respectively. The results show that the needle trap sampler with Carboxen 1000 has good storage stability for BTEX. Reusability is one of the advantages of a needle trap diffusive sampler. The idea of reusability is based on the fact that there should be no target analyte adsorbed by the sorbent after thermal desorption. The needle trap diffusive sampler should be sealed effectively after injection in case some compounds in air diffuse into the sampler, resulting in any inaccurate result. In addition, the sampler should be conditioned at the temperature 20 °C over the desorption temperature. Three NTDs were used to sample the 4.0-mL sample of BTEX from the standard gas system 15 times, respectively, to investigate their reusability. Table 4 summarizes the average concentrations of BTEX determined by these three NTDs, their relative standard deviations, and comparison between the concentrations analyzed and the standard. It indicates that the NDT can be used repeatedly many times and maintain adsorption and desorption effectively. In addition, the limit of detection (LOD) and limit of quantitation (LOQ) of BTEX were determined based on the standard calibration curve. LOD ) 3 × RSD/b and LOQ ) 10 × RSD/b, where RSD is residual standard deviation of the regression line Analytical Chemistry, Vol. 80, No. 19, October 1, 2008

7279

Table 4. Average Concentrations of BTEX Determined Using Different Needle Trap Diffusive Samplers (n)15), RSDs, and Standard Concentrations of BTEX benzene

toluene

ethylbenzene

o-xylene

C (ng/mL)

RSD (%)

C (ng/mL)

RSD (%)

C (ng/mL)

RSD (%)

C (ng/mL)

RSD (%)

19.9 20.1 20.0 19.8

4 3 5

19.2 20.0 19.6 19.7

4 2 5

9.0 8.7 8.7 8.9

3 5 5

9.1 8.9 9.1 9.2

5 4 6

sampler 1 sampler 2 sampler 3 standard

Table 5. Theoretical and Experimental Sampling Rates of BTEX of Six NTDs with Different Diffusion Lengthsa sampling rate (mL/min) (B) diffusion needle length (cm) 1 2 3 4 5 6 a

0.44 0.63 0.78 0.73 0.58 0.49

sampling rate (mL/min) (T)

sampling rate (mL/min) (E)

sampling rate (mL/min) (X)

RT

RE

RT

RE

RT

RE

RT

RE

0.016 0.012 0.009 0.010 0.013 0.015

0.016 0.011 0.010 0.010 0.011 0.014

0.015 0.011 0.009 0.009 0.011 0.014

0.015 0.010 0.010 0.010 0.010 0.013

0.013 0.009 0.008 0.008 0.010 0.012

0.015 0.011 0.009 0.010 0.010 0.011

0.013 0.009 0.007 0.008 0.010 0.012

0.013 0.009 0.008 0.006 0.009 0.010

RT, theoretical sampling rate; RE, experimental sampling rate.

and b is the slope of the standard calibration curve.28 The LODs of BTEX were 0.78, 0.70, 0.60, and 0.67 ng, respectively, and the LOQs of BTEX were 2.60, 2.35, 2.02, and 2.23 ng, respectively. Verification and Validation. Ideally, according to Fick’s first law, the theoretical sampling rate of a diffusive sampler, which depends on the geometry of the sampler and on an individual analyte with a particular diffusion coefficient in air, should be equal to the experimental sampling rate. In practice, a variety of factors may affect sampling rate, including face velocity, sorbent strength, response time, atmospheric temperature and pressure, relative humidity, and sampling duration. If the theoretical sampling rate is equal to the experimental sampling rate under experimental conditions, then the concentration of the target analyte can be calculated directly, without having to determine the experimental sampling rate, by using the theoretical sampling rate, the adsorbed mass, and sampling time. In order to use eq 2 for determining the concentration of target analyte, the ratio of theoretical sampling rate to experimental sampling rate should be close to 1. The theoretical sampling rate, RT (cm3/min), of a needle trap sampler, where diffusion length, Z (cm), and cross-sectional area of the surface, A (cm2), are known, was calculated using RT ) (DA)/Z, where D is the analyte diffusion coefficient (cm2/s). The diffusive sampler was then used to do TWA diffusive sampling of a standard BTEX gas mixture for a determined length of time. The experimental sampling rate was calculated using RE ) n/(Ct); the concentration C (ng/mL) of each analyte was calculated from the permeation rate of the analyte and the flow rate of the carrier air; n is the loading mass (ng) of the analyte, and t is the sampling time (s). Verification of the theoretical model was conducted using six different NTDs with different diffusion lengths. The concentrations of BTEX in the standard gas system were 4.33, 4.80, 2.18, and 2.42 ng/mL, respectively; and the sampling time was 1 h. The

results (Table 5) show good agreement between the theoretical and experimental sampling rates; the experimental and theoretical sampling rates differ slightly, perhaps due to factors like nonideal sorbent, error in diffusion length measurement, humidity, etc. TWA diffusive sampling with a NTD is a new method. It is therefore essential to compare it with other, established methods. Validation tests were conducted in a standard gas system against charcoal tube sampling using NIOSH method 1501, SPME, and NTD active sampling.29 In the standard gas system, the temperature was 25 °C, relative humidity was 21%, air velocity was 3.05 cm/min, and BTEX concentrations were 1.57, 1.74, 0.79, and 0.88 ng/mL, respectively. The results are presented in Figure 4. BTEX concentrations determined by NTD TWA diffusive sampling were 1.52, 1.69, 0.81, and 0.89 ng/mL. The % RSDs (n ) 4) of BTEX for NTD TWA diffusive sampling were 2, 3, 6, and 4%; and the relative error, compared to the standard concentrations of BTEX in the standard gas system, were 5, 4, 0.01, and 8%. No significant

(28) Sonali, P.; Lata, K.; Asha, T.; Deshpande, A. D. Indian J. Pharm. Sci. 2007, 69 (4), 525–528.

(29) National Institute for Occupational Safety and Health. Manual of Analytical Methods, 4th ed.; U.S. Department of Health and Human Service, 1994.

7280


Figure 4. Comparison between needle trap TWA diffusive sampling and other methods, as well as the standard gas system.

Figure 5. Schematic of needle trap device diffusive samplers: (A) badgelike sampler and (B) penlike sampler.

differences were found between NTD TWA diffusive sampling and the other methods. TWA Diffusive Sampler and Field Sampling. In this study, two types of simple and portable TWA diffusive samplers were developed to monitor toxic gas and vapor concentrations in both occupational and community environments. The badgelike sampler includes two components, a sampler holder and a NTD. On the back of the sampler holder there is clip which can be used to fix the sampler to a front pocket or under a shirt collar. The sampler holder consists of one metal plate and four Teflon chips. There is a hole in the center of each chip, which is used to seal the side hole and the tip of the needle. Figure 5 (part A) is the schematic diagram of the back of the holder, sealing position, and sampling position of the needle trap diffusive sampler. The type of diffusive sampler is a penlike sampler, which can also be used in automation sampling. Figure 5 (part B) shows two positions of the sampler. When the button at the end of the pen is pressed, the needle is in the forward position and the tip of needle is sealed by the Teflon tip fixed to the pen-tip; this position is the sampler’s sealing status. Before sampling, the needle is retracted by pressing the button. The tip of the needle is exposed to the space in the pen, where air can move in and out through three windows on each side of the pen; this position is the sampler’s sampling position. This type of sampler is lighter

and more user-friendly than the first; however, it is more complicated to load the NTD into the holder with this sampler. Although laboratory validation may determine some parameters that might influence the diffusive sampling performance of the NTD, it is often not possible or feasible to simulate practical field sampling conditions. Therefore, it is necessary to compare the field performance of the NTD diffusive sampler with that of NIOSH method 1501. Field validation was conducted by monitoring BTEX concentrations in a newly painted apartment using the NTD diffusive samplers (shown in Figure 5) and charcoal tubes from NIOSH method 1501. Four TWA diffusive samplers and seven charcoal tubes were used to sample the air. The sampling time was 8 h, and the sampling volume was 8 L per charcoal tube. BTEX concentrations determined by NTD diffusive sampling were 0.21, 0.13, 0.07, and 0.02 ng/mL; those determined by charcoal tube sampling were 0.22, 0.14, 0.08, and 0.02 ng/mL. The difference between these two results is not significant. This validation test demonstrates that the NTD with Carboxen 1000 can perform successful diffusive sampling of BTEX not only under controlled conditions but also under field conditions. Field diffusive sampling using the NTDs was also performed in the garage, living room, bedroom, and shed for 24 or 40 h. The results, shown in Table 6, indicate that benzene was only detected in the shed and that concentrations of toluene, ethylbenzene, and o-xylene were much higher than that in the air monitored in other places. Concentrations of toluene, ethylbenzene, and o-xylene in the living room and bedroom were lower than those in the garage. NTD-TWA-Autosampler Determinations. One of the main advantages for the needle type sampler is the ease of providing automation and online coupling to GC.30 This combination of the NTD-TWA diffusive sampler and automation analysis eliminated the error caused by manual injection and reduced the analysts’ labor and time. Repeatability tests were performed by the automation analysis. Six different Carboxen 1000 packed NTDs (Z ) 0.3 cm) were used to sample the BTEX mixture from the standard generation system. TWA sampling time was 3 h and the trapped BTEX were analyzed by the NTD-TWA-GC-autosampler. As shown in Table 7, the most % RSDs for an intra-NTD were quite good for the BTE components with less than 5% RSD, while the % RSD of o-xylene ranged 2-11% RSD due to the coelution of an unknown contaminant. The highest % RSD for intra-NTD analysis was 12%. With a very low amount of absolute extracted amount of the BTEX (ranging from 10-30 ng), this level of % RSD was acceptable especially for an on-site field application. CONCLUSION A needle trap device with Carboxen 1000 in conjunction with GC/MS by thermal desorption was developed for sampling and analysis of VOCs. It offers several advantages over other sampling devices. It is inexpensive, reusable, solventless, does not require empirical calibration, and easy to operate and automate; it also integrates sample preparation and separation into a single step, reduces analysis time, and obviates losses of target compounds that can occur during transportation from the sampler to the GC. (30) Demeestere, K.; Dewulf, J.; Witte, B. D.; Langenhove, H. V. J. Chromatogr., A 2007, 1153, 130–144.


7281

Table 6. Determined Concentrations of BTEX in the Garage, Living Room, Bedroom and Shed

a

sampling place

C (ng/mL) (B)

garage living room bedroom shed

nda nda nda 4.75 ± 0.01

C (ng/mL) (T)

C (ng/mL) (E)

C (ng/mL) (X)

0.40 ± 0.01 0.33 ± 0.02 0.29 ± 0.004 55.2 ± 0.18

0.55 ± 0.02 0.36 ± 0.04 0.03 ± 0.005 8.49 ± 0.50

0.43 ± 0.05 0.18 ± 0.03 0.06 ± 0.008 8.90 ± 0.57

nd, not determined.

Table 7. Repeatability Test from the Same NTD (intra-NTD) and between Six Different NTDs (inter-NTDs) by the NTD-TWA-Autosampler Analysis % relative standard deviation CAR NTD1 (n ) 4) CAR NTD2 (n ) 4) CAR NTD3 (n ) 4) CAR NTD4 (n ) 4) CAR NTD5 (n ) 4) CAR NTD6 (n ) 4) all (CAR NTD 1-6) (n ) 24)

benzene

toluene

ethylbenzene

o-xylene

2.8 1.8 6.2 3.6 2.2 1.7 11.9

4.2 5.1 2.3 2.8 1.9 1.8 10.4

4.8 6.4 4.0 1.8 3.1 3.8 7.6

11.1 5.0 5.1 9.8 2.0 7.0 10.0

The NTD is shown to be a successful diffusive sampler for monitoring TWA concentrations of BTEX under experimental conditions. The GC-autosampler for NTD determinations eliminated the manual injection error and reduced the time and labor of the analyst. Further efforts will be made to monitor air quality under different environmental conditions and sampling models

7282


and apply this technique to a wider range of analytes. The NTD diffusive sampler combined with automated GC determination facilitates convenient monitoring of both personal exposure to VOCs in occupational environments and ambient air quality. ACKNOWLEDGMENT The authors thank the Natural Sciences and Engineering Research Council of Canada (NSERC) and Shinwa Chemical Industries Ltd. for financial support. NOTE ADDED AFTER ASAP PUBLICATION This article was released ASAP on September 3, 2008 and the correct version was posted September 6, 2008 with additional corrections.

Received for review April 30, 2008. Accepted August 8, 2008. AC800884F

Development and Application of a Needle Trap Device for Time

Recommend Documents