Environ. Sci. Technol. 2001, 35, 1481-1486
Field Sampling and Determination of Formaldehyde in Indoor Air with Solid-Phase Microextraction and On-Fiber Derivatization JACEK A. KOZIEL,† JAPHETH NOAH, AND JANUSZ PAWLISZYN* Department of Chemistry, University of Waterloo, Waterloo, Ontario, N2L 3G1Canada
A new sampling and analysis method for formaldehyde in indoor air was tested in several indoor air surveys. The method was based on the use of solid-phase microextraction (SPME) poly(dimethylsiloxane)/divinylbenzene,65-µm fiber and gas chromatography. Indoor air surveys included grab and time-weighted average (TWA) sampling and were completed at six locations using (a) the SPME method employing on-fiber formaldehyde derivatization with o-(2,3,4,5,6-pentafluorobenzyl)hydroxylamine hydrochloride and (b) the conventional National Institute for Occupational Safety and Health (NIOSH) 2451 method. Sampling time for SPME fiber ranged from 10 min for grab sampling to 8 h for TWA sampling. Sampling locations included a residential house, a rental apartment, an office building, and industrial workplaces. The air concentrations measured by SPME ranged from 10 to 380 ppbv and correlated well with those estimated by the NIOSH method. Results also indicated that in some cases the formaldehyde concentrations measured in residential air could be much higher than those allowed in occupational settings. The SPME method proved to be accurate, fast, sensitive, and cost-efficient in field sampling applications. This research should be of interest to research, industrial, and regulatory agencies as well as to the general public concerned with indoor air quality.
Introduction In the past two decades, an increased awareness of the importance of indoor air quality (IAQ) and its potential impact on human health has stimulated an interest in indoor air research (1-3). A number of studies reported a direct relationship between poor IAQ and health problems (4, 5). Volatile organic compounds (VOCs) are recognized as one of the major contributors to poor IAQ. Typical VOCs include aldehydes and aromatic and halogenated hydrocarbons. Typical aldehydes found in indoor air include formaldehyde, acetaldehyde, and glutaraldehyde. Formaldehyde (chemical formula, HCHO) is an ubiquitous indoor air pollutant due to its use in many materials and processes. Examples include building materials, e.g., plywood, particleboard, medium-density fiberboard, adhesives, glues, paints, insulation, and carpeting (6). Formaldehyde can be found in cosmetic products, tobacco smoke, ozone genera* Corresponding author phone: (519)888-4641; fax: (519)746-0435; e-mail:
[email protected]. † Current address: Texas Agricultural Experiment Station, Texas A&M University, Amarillo, TX 79106. 10.1021/es001653i CCC: $20.00 Published on Web 03/03/2001
2001 American Chemical Society
tors, permanent-pressed clothing and draperies, laboratories, hospitals, mortuaries, and others. Formaldehyde is classified as a probable human carcinogen by the U.S. EPA, as a known animal carcinogen by the NIOSH and OSHA, and as a suspected human carcinogen by the American Conference of Governmental Industrial Hygienist (ACGIH). Human responses to concentrations as low as 100 parts per billion by volume (ppbv) include burning sensations and watery eyes, itching skin, nausea, and difficulty in breathing. Acceptable limit concentrations range from 16 to 100 ppbv in occupational settings (7). However, because of the widespread use of formaldehyde, it is very likely that in some cases formaldehyde concentrations in residential indoor air, i.e., where no regulations exist, may be much higher than those in a regulated occupational setting. Studies that attempted to establish a link between exposure to indoor air formaldehyde and observed health effects have been only marginally successful. The prohibitive cost of developing a proper sampling protocol with respect to the number of samples and the appropriate detection levels has limited these studies. Only a few methods can detect formaldehyde in the 5-50 ppbv range that is typical of nonoccupational indoor air. Conventional methods for airborne formaldehyde use colorimetry, polarography, and liquid and gas chromatography (7-11). Because of formaldehyde reactivity, many of these methods use derivatization. Most popular methods use active sampling with solid adsorbent coated with the derivatizing agent dinitrophenyl hydrazine (DNPH) or (hydroxymethyl)piperidine (HMP) (7). Passive sampler techniques can also be used in occupational settings (7, 8, 11). However, in many cases, conventional sampling methods may require longer sampling time (2-3 h) to be suitable for indoor air sampling, particularly in cases where formaldehyde concentrations are in the low ppbv range. The aforementioned limitations in indoor air sampling and determination of formaldehyde can be addressed with SPME technology. SPME uses a syringe-like device that combines sampling, preconcentration, and the direct transfer of the analytes into a standard gas chromatograph (GC) (12). The analytes in air are adsorbed or absorbed to the thin polymeric film coated on a fused silica fiber. After an appropriate amount of sampling time, the fiber is pulled inside the needle and transferred into the injector of a gas chromatograph, where the SPME fiber coating is exposed to the hot carrier gas and completely desorbed. All desorbed analytes are introduced into the column and a GC detector (12, 13). Several research studies have focused on the application of SPME to air sampling and analysis (13-23). To date, SPME sampling methods have been developed for total volatile organic compounds (TVOCs) and formaldehyde in air (17, 18). The latter method is based on the use of on-fiber derivatization and can be used for formaldehyde in air and for headspace and water sampling. A recent study indicated that SPME might also be used for TWA sampling (19). These fundamental studies included some limited field applications and validation with conventional methods. The primary objective of this research was to demonstrate the viability of SPME for the sampling and determination of formaldehyde in indoor air. Specific objectives included the adaptation of an existing on-fiber derivatization method for the quantitative estimation using grab and TWA sampling. These methods were then applied to several indoor air surveys. Air concentrations estimated with SPME were also compared with those obtained using the NIOSH 2451 method. VOL. 35, NO. 7, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1481
dependent on the formaldehyde concentration in air (18):
ν ) CHCHOK*
FIGURE 1. Schematic of the reaction between formaldehyde and the derivatizing agent (PFBHA) occurring on the SPME fiber. Finally, this research identified the advantages and limitations of SPME technology for indoor air surveys for formaldehyde.
Theory On-Fiber Derivatization with SPME. Concentrations of airborne formaldehyde can be estimated using a relatively new method based on the use of 65 µm of poly(dimethylsiloxane)/divinylbenzene (PDMS/DVB) adsorptive coating and on-fiber derivatization (18). First, the derivatizing agent, i.e., PFBHA, is loaded on the SPME fiber for several minutes using headspace extraction immediately before ambient formaldehyde sampling. Next, during actual sampling, the airborne formaldehyde reacts with the derivatizing agent forming a very stable formaldehyde-PFBHA oxime (Figure 1). The amount of oxime formed on the fiber is proportional to the formaldehyde concentration in air, and it can be analyzed using a conventional GC equipped with a flame ionization detector (FID). The overall rate of oxime formation on solid SPME fiber coating as a function of the formaldehyde concentration can be described by a four-step process (18): k1,k-1
PFBHA + S 9 7 8 PFBHA*S
k2,k-2
HCHO + S 798 HCHO*S
K*
(adsorption/desorption) (1) (adsorption/desorption) (2)
HCHO + PFBHA*S 798 oxime*S k3
oxime*S 798 oxime + S
(reaction)
(3)
9
where ν is the rate of oxime formation (ng/s) (sometimes referred to as a “sampling rate” by those who use passive dosimeters), CHCHO is the formaldehyde concentration in air (ppbv), and K* is the apparent first-order rate constant (ng/ (ppbv × s)). Therefore, quantitative analysis of an unknown formaldehyde concentration is possible when K* is established experimentally and the amount of the derivatizing agent consumed is negligible, i.e., for short sampling times (18). Time-Weighted Average Sampling with SPME. Longterm sampling for gaseous formaldehyde can be accomplished with SPME fibers (19). In this case, the fiber coating loaded with PFBHA is retracted inside the needle. The distance between the fiber opening and the fiber tip serves as a diffusion path for formaldehyde, and the sampling rate is reduced as compared to the exposed fiber arrangement. Similarly to the grab sampling with exposed fiber, an excess of PFBHA must be available to react with formaldehyde and maintain the first-order reaction rate. Retracted fiber coating works as a TWA sampler. It has been shown that the amount of analyte on the sorbent (nf) (ng) is proportional to the integrated gas-phase concentration over the sampling time (t) (s), the gas-phase molecular diffusion coefficient (Dg) (cm2/ s), and the ratio of the needle opening area (A) (cm2) to the diffusion path length (Z) (19):
nf ) Dg
A Z
∫C (t) dt g
(6)
Equation 6 can be used for the estimation of TWA concentration (Cg(t)) (ng/mL) since nf can be established from the detector response and t is measured. The diffusion path length can be reduced or increased providing the flexibility of extending or reducing the sampling time in the field, based on the estimate of an analyte concentration from a fast, “screening” sample with an exposed fiber. The term DgA/Z is often referred to as the first-order rate K′ for TWA sampling and can be estimated from theory or established experimentally as K′ ) nf/(Cg(t) × t) in (ng/(ppbv × s)) (19).
Experimental Section (desorption)
(4)
where S is the surface of the sorbent available for binding; k1 and k-1 are rates of PFBHA adsorption and desorption from the fiber coating; k2 and k-2 are rates of formaldehyde adsorption and desorption from the fiber coating; K* is the rate of reaction between airborne formaldehyde in air and sorbed PFBHA; and k3 is the rate of oxime desorption, which only occurs significantly inside a heated GC injector. Experimental data showed that k1 . k-1, i.e., the desorption rate of PFBHA is negligible (18). The rate of formaldehyde adsorption to the unoccupied surface sites (k2) is also negligible because the coating is nearly saturated and all sites are occupied by PFBHA. As a result, formaldehyde in air is free to react with the hydroxylamine moiety (Figure 1). The reaction between formaldehyde and PFBHA is favored by using short and long sampling times for high and low concentrations, respectively. As a result, only a small amount of available PFBHA is consumed, while a detectable mass of oxime is produced. The rate of formaldehyde diffusion toward the sorbent is negligible as compared to the reaction rate K* that is controlling the sampling process. The consumption of PFBHA has a negligible effect on the overall rate of reaction. Thus, the on-fiber derivatization process is controlled by first-order kinetics, where the overall rate of reaction is 1482
(5)
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 7, 2001
Standard Gas Sampling System. A Kin-Tek standard gas generator (Kin-Tek, La Marque, TX) (model 491MB) and a certified permeation tube were used to generate standard gas concentrations of formaldehyde (20). The standard gas concentrations were estimated based on the NIST-certified emission rate for the formaldehyde permeation tube at the incubation temperature (80 °C) and flow rates of the ultrahigh purity (UHP) grade nitrogen, measured using a mini-Buck primary flow calibrator (A. P. Buck, Orlando, FL). The formaldehyde concentration was adjusted for the actual temperature inside the sampling chamber and the atmospheric pressure. The resulting standard formaldehyde concentrations ranged from 5 to 3000 ppbv. All calibration was completed at temperatures ranging from 21 to 25 °C, i.e., close to the typical temperatures in indoor air environments. A new sampling chamber for the standard gas generator was constructed and installed downstream from the standard gas generators (20). This sampling chamber provided a steady-state mass flow of formaldehyde at constant temperatures typical of indoor air and enabled both the SPME and the NIOSH method calibration. Temperature changes within approximately (5 °C of 25 °C and air velocities typical of indoor air do not significantly affect the amount of oxime produced (18, 21). This assumption is valid for typical indoor air sampling conditions.
GC Methods. The SPME samples were preserved and analyzed in the laboratory less than 1 h following formaldehyde sample collection. A Varian 3400 GC equipped with a SPB-5 (30 m × 0.25 mm, 1.0 µm film thickness) capillary column (Supelco, Mississauga, ON) and a FID were used for all formaldehyde analyses. The GC oven temperature was programmed from 45 °C (held for 2 min) to 200 °C at a ramp of 30 °C/min followed by a ramp of 50 °C/min to 290 °C and held for 4 min. The septum-programmable injector (SPI) was set to 210 °C, and the detector temperature was set to 300 °C. The carrier gas was UHP-grade hydrogen at 26 psi. Liquid injections and the analyses of NIOSH-based samples were completed using the same GC and column. The GC oven temperature was programmed from 50 °C (held for 1 min) to 240 °C at a ramp of 10 °C/min. The SPI injector was temperature programmed from 45 to 250 °C at 300 °C/min. The injector was cooled with liquid CO2 between injections. SPME Fiber Selection. The 65-µm PDMS/DVB fiber coating (Supelco, Mississauga, ON) was used in this method (18). This fiber coating was selected for its ability to retain the derivatizing agent and its affinity for the PFBHAformaldehyde oxime. Caution: Some recent (1999) batches of PDMS/DVB 65-µm fibers may contain interferences. The new fibers produce a large impurity peak, which coelutes and covers the derivatization product (oxime), precluding the use of the 65-µm PDMS/DVB fiber for formaldehyde sampling. These interferences were not present in the older fiber batches, i.e., when the original method was actually developed. Thus, particular care must be exercised with new batches of 65-µm PDMS/DVB fibers. This problem has been identified and documented, and the SPME fiber manufacturer was notified. In this study, only older PDMS/DVB fibers that did not contain impurities were used for the formaldehyde sampling. Loading of Derivatizing Agent. A 2-min loading of PFBHA by headspace extraction from an aqueous solution (17 mg/ L) was used. Approximately 2 mL of this solution was prepared in a 4-mL vial, capped with a Teflon-coated septum, and stirred at 1800 rpm using a VWR 400HPS stirrer. The PFBHA loading time of 2 min and the PFBHA concentration 17 g/L were used to ensure that a relatively large amount of derivatizing agent was loaded onto the fiber and that the depletion of PFBHA during the reaction with formaldehyde did not affect the reaction rate. The loading time and concentration were consistent with the original method presented in ref 18. Selection of Sampling Time. Linear responses were previously reported for the reaction product of the airborne formaldehyde and PFBHA for sampling times ranging from 10 s to 12 min and formaldehyde concentrations ranging from 15 to 640 ppbv (18). In this study, a 10-min sampling time was selected to maximize the amount of the reaction product (oxime) and increase the method sensitivity. A typical calibration curve for a 65-µm PDMS/DVB fiber with PFBHA and 10-min sampling time for standard formaldehyde concentrations ranging from 10 to 300 ppbv is presented in Figure 2. NIOSH Sampling. Simultaneously with SPME-based grab sampling using a PFBHA-loaded 65-µm PDMS/DVB fiber, the conventional NIOSH-2541 sampling method was used for the determination of standard formaldehyde concentrations and validation of the SPME method. For the NIOSH2541 method, ORBO-25 adsorbent tubes and an A. P. Buck industrial hygiene pump were deployed for 4 h at 50 mL/ min. The adsorbent tubes used (hydroxymethyl)piperidine (HMP) coated on a cross-linked polymeric XAD-2 sorbent for derivatization of airborne formaldehyde. Standard (theoretical) concentrations were estimated using the NISTcertified formaldehyde emissions rate and the standard gas flow rate, adjusted for the actual temperature and pressure.
FIGURE 2. Calibration curve for formaldehyde using a 10-min sampling time with the 65 µm of PDMS/DVB fiber.
FIGURE 3. Schematic of field sample preservation. SPME needle is capped with a narrow bore Teflon cap and stored in a cooler. The NIOSH-based samples with adsorbent were collected in the outlet of the sampling chamber. The sampling flow rate was verified with the mini-Buck primary gas calibrator (A. P. Buck, Orlando, FL) and was set to 100 mL/min. The sampling time ranged from 4 to 8 h. Field Sampling Plan. Six locations were selected for indoor air surveys. These locations included a house and a rental apartment, an elementary school with portable classroom units, an office building, and two semi-industrial workplaces (an engineering shop and a chemical store) in the Waterloo area. Samples were collected in at least three rooms within each site. All surveys were conducted in the summer and were generally completed in 1 day. Indoor air surveys at each site included both grab and TWA sampling with SPME fibers. For grab sampling, at least three 10-min samples were collected at each room. For TWA sampling at least one duplicate 8-h sample per site was collected. The NIOSH sampling consisted of approximately 5-10% of the total samples collected and in most cases was completed sideby-side with the TWA SPME sampling.
Results and Discussions Sample Preservation. Stability of the PFBHA-formaldehyde oxime over time on the 65-µm PDMS/DVB fibers was studied before actual sampling in the field. Previous study has shown that the formaldehyde-PFBHA oxime is extremely stable and that the on-fiber derivatization is terminated when no formaldehyde reaches the fiber (21). The main concern is actually oxime gain due to continued reaction rather than oxime loss. Thus, to ensure that no formaldehyde can react with PFBHA after the sample collection, the SPME needle opening should be sealed off to prevent the ambient formaldehyde diffusion (Figure 3). In this study, it was determined that the best technique consisted of retracting the fiber approximately 10 mm inside the SPME needle, followed by sealing the needle with a narrow-bore Teflon cap (Figure 3) and placing it on a bed of dry ice. Several VOL. 35, NO. 7, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1483
FIGURE 4. Schematic of a modified SPME device for TWA sampling. 65-µm PDMS/DVB fiber coatings loaded with PFBHA were exposed for 10 min to the standard 50 ppbv gas mixture, capped, and analyzed after 15 h and compared to samples that were analyzed immediately after sample collection. Sample gains were below 10% of the initial concentration at room temperature. This gain should be lower when capped SPME needles are stored at low temperatures because of the gas-phase diffusion coefficient reduction. All SPME-based samples in this research were analyzed in less than 1 h after sample collection. Validation of Standard Gases with NIOSH Method. The NIOSH-2541 method using analysis on a GC/FID was selected for validation of SPME sampling. The theoretical and measured (using NIOSH-2541) formaldehyde concentrations correlated well to approximately 100 ppbv, i.e., below the recommended method detection limit (MDL). Longer sampling times were used so that concentrations lower than 100 ppb could also be detected. However, the concentration variance was significant, rendering the NIOSH 2541 method less reliable for concentrations below 100 ppbv. The SPMEbased method was sensitive to formaldehyde concentrations below 5 ppbv. It should be noted that other conventional methods based on adsorbent cartridges and high-performance liquid chromatography (HPLC) can reasonably achieve detection limits close to 0.5 ppbv. Method Detection Limits. The MDL associated with 10 min SPME sampling with on-fiber derivatization was estimated as the product of the standard deviation of 10 replicate samples at a 5 ppb level and the two-tailed t-statistics for n ) 9 degrees of freedom at the 95% confidence interval. The estimated MDL was approximately 1 ppbv. Estimation of First-Order Rate for TWA Sampling. The estimation of formaldehyde first-order rate (K′) was completed using modified TWA SPME sampling devices (Figure 4) (13). Three standard formaldehyde concentrations were used, i.e., 10, 50, and 500 ppbv. Two diffusion path lengths (Z) were initially tested, i.e, 3 and 10 mm. The TWA sampling time ranged from 30 to 960 min. The K′ values were estimated based on standard (theoretical) formaldehyde concentrations, sampling time, and the amount of oxime formed. The resulting K′ value was not consistent when Z ) 3 mm likely because the diffusion path from the needle opening is relatively short and varying. These variations were reduced when Z ) 10 mm (RSD ) 5.3%). Thus, Z was set to 10 mm for all TWA measurements in this study, and the average K′ value for Z ) 10 mm was used for concentration estimates. The amount of depleted PFBHA was always less than 10% of the initially loaded mass. Figure 5 presents a typical oxime loading over time. In addition, indoor air temperatures, pressures, and relative humidities were measured and recorded for every location in the indoor air surveys (Table 1). Formaldehyde Concentrations. The summary of SPME grab samples and the 8-h TWA concentrations using simultaneous SPME and the NIOSH-2541 method is presented in Table 2. The SPME-based concentrations were generally close to the NIOSH-based concentrations. Concentrations for the 1484
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 7, 2001
FIGURE 5. Mass loading of oxime during TWA sampling for formaldehyde.
TABLE 1. Environmental Parameters Measured during Indoor Air Surveysa site chemical stores
residential house office building
engineering shop elementary school
rental apartment
a
room
T (°C)
RH (%)
p (mmHg)
main office waste receiving solvent storage outdoor air master bedroom playroom garage main office copying room small office outdoor air paint shop grinding shop carpenter shop classroom new portable unit old portable unit outdoor air living room master bedroom bathroom
24.0 19.5 19.5 27.5 23.0 22.5 24.0 21.5 22.0 23.5 32.0 23.5 22.0 21.5 26.5 22.0 23.5 26.5 28.5 26.5 28.0
57 73 72 42 76 70 72 62 62 55 35 34 35 37 60 44 40 67 62 67 63
727.5 728.2 728.2 728.2 726.2 726.2 725.4 727.5 727.5 728.2 728.2 732.0 732.8 732.8 727.5 728.2 727.5 727.5 715.5 716.3 717.8
T ) temperature; RH ) relative humidity; p ) atmospheric pressure.
SPME grab sampling represent an average value for all samples collected at the same location during the same survey, i.e., n ) 2 or 3. Measured (by SPME) concentrations ranged from approximately 10 to almost 380 ppbv. The maximum concentration was measured at a residential house, where a new set of particleboard furniture was installed approximately 1 month before the survey. Higherthan-average formaldehyde concentrations (103 ppbv) were also measured in another room within the same house. Formaldehyde concentrations at the rental apartment were much lower and averaged approximately 21 ppbv. Both residential locations had formaldehyde concentrations higher than the NIOSH threshold value of 16 ppbv for occupational settings. This finding suggests that a significant fraction of population may be exposed to residential formaldehyde concentrations greater than those allowed in occupational settings. In general, the average concentrations at workplaces surveyed were between 10 and 50 ppbv. The 8-h TWA concentrations were always significantly higher than those associated with the grab (10-min) sampling. The true reason for this is unknown. It is not likely that the formaldehyde concentrations were significantly changing during sampling events. Instead, possible causes could be associated with uncertainties associated with the estimation of K′ and K*, particularly its dependence on air temperature and air velocity. If significant, these effects could be a major limitation to this method. However, effects of these variables
TABLE 2. Summary of Formaldehyde Concentrations (in ppbv) during Indoor Air Surveysa SPME site chemical stores
residential houseb office building
engineering shop elementary school
rental apartment f
a
Age of buildings as noted:
b
room
NIOSH-2541 (8 h)
grab 10 min
main office waste receiving solvent storage outdoor air master bedroom playroom garage main office copying room small office outdoor air paint shop grinding shop carpenter shop classroomc new portable unitd old portable unite outdoor air living room master bedroom bathroom outdoor air
33
15 12 19 11 376 103 60 28 14 36 17 14 24 23 11 13 13 12 20 19 22 10
30 yr. c 43 yr.
d
214
81 59
16 19 66
TWA 8 h 37
200
112 57
15 (5 h) 22 (7 h) 39
1 yr. e 9 yr. f 28 yr.
were not studied in this research. It should be noted that much progress has been made recently in the area of theory and applications of fast SPME sampling in moving air and diffusion-based calibration for VOCs and adsorptive SPME fibers (22, 23). In this method, air velocity and air temperature were accounted for in the model, and the predicted concentrations were consistent with the experimental data (22, 23). Thus, it is reasonable to expect that the limitations associated with K′ and K* could be addressed with future research. The outdoor concentrations detected with the SPME device ranged from approximately 10 to 17 ppbv. These concentrations are consistent with the values listed in the literature (24-27). It should be emphasized that the Waterloo area is likely affected by air quality problems typical to Southern Ontario, caused by concentrations of industrial sources, elevated levels of ozone in summer associated with transportation and geography, and by the long-range transport of pollutants (27). Other factors, such as the use of ethanol in gasoline, may also affect the levels of formaldehyde in ambient air (26). In general, air sampling with SPME devices proved to be an alternative to the NIOSH-based field sampling of formaldehyde, particularly where low-cost screening for formaldehyde is needed. The total sampling cost for consumables is less than 50 cents per sample, considering that all derivatizing agent and oxime is completely desorbed and that the same SPME fiber can be used about 200 times. The time required to complete the analytical process has been reduced by at least an order of magnitude as compared to conventional NIOSH techniques. SPME-based sampling proved to be reliable for both grab and TWA sampling. The sampling times ranged from 10 min to 8 h, rendering the SPME device useful for a wide range of exposure assessments. The same derivatizing agent also works for other aldehydes in air other than those studied here, e.g., acetaldehyde and glutaraldehyde. There is great potential for using the SPME method for quantitative analysis for these analytes in industrial hygiene, occupational, and residential sampling. Preliminary research showed that the formaldehyde-PFBHA oxime could be also detected on a photoionization detector (PID). Thus, it may
be possible to combine on-fiber derivatization with on-site analysis using a portable GC/PID. One of the limitations of this coupling is associated with generally higher detection limits achievable with portable GCs. The limitations of this methodology are associated with the need to load the derivatizing agent each time onto the SPME fiber before sampling, possible fiber coating impurities that are common in some recent batches of 65-µm PDMS/DVB, and method uncertainties. These uncertainties are associated with the lack of a comprehensive assessment of several important variables, such as air velocity and air temperature, on formaldehyde sampling rate. In addition, preliminary studies have shown that on-fiber derivatization of formaldehyde can be also used for drinking water sampling. The 65-µm PDMS/DVB coating loaded with PFBHA (via headspace extraction) and headspace extraction proved more sensitive than other methodologies tested, such as direct extraction or spiking of PFBHA into water and direct extraction. The MDLs for 5 min headspace extraction from half-full 40 mL vials stirred at 1,200 rpm with Teflon-coated magnetic stir-bars was approximately 3 ppbv. This method could be used for routine formaldehyde analysis of drinking water, which is receiving increased regulatory attention due to the generation of aldehydes as byproducts of ozonation.
Acknowledgments The authors thank the Center for Indoor Air Research for funding this study and Dr. Perry Martos for helpful discussions.
Literature Cited (1) Molhave, L. Environ. Int. 1982, 8, 117-127. (2) Molhave, L.; Bach, B.; Pederson, O. F. Environ. Int. 1986, 12, 167-175. (3) Wallingford, K. M. Energy Technol. 1988, 15, 436-444. (4) Kerns, W. D.; Donofrio, D. J.; Pavkov, K. L. Cancer Res. 1983, 43, 4382-4392. (5) Hays, S. M.; Gobbell, R. V.; Ganick, N. R. Indoor Air Quality, Solutions and Strategies; McGraw-Hill: New York, 1995. (6) Kelly, T. J.; Smith, D. L.; Satola, J. Environ. Sci. Technol. 1999, 33, 81-88. VOL. 35, NO. 7, 2001 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
1485
(7) National Institute for Occupational Safety and Health. NIOSH Manual of Analytical Methods, 4th ed.; NTIS: Springfield, VA, 1998. (8) Kring, E. V.; Ansul, G. R.; Basilio, A. N.; McGibney, P. D.; Stephens, J. S.; O’Dell, H. Am. Ind. Hyg. Assoc. J. 1984, 45 (5), 318-324. (9) Woo, C. S.; Barry, S. E.; Zaromb, S. Environ. Sci. Technol. 1998, 32, 169-176. (10) Klendiest, T. E.; Corse, E. W.; Blanchard, F. T.; Lonneman, W. A. Environ. Sci. Technol. 1998, 32, 123-130. (11) Gillett, R. W.; Kreibich, H.; Ayers, G. P. Environ. Sci. Technol. 2000, 34, 2051-2056. (12) Pawliszyn, J. Solid-Phase Microextraction: Theory and Practice; Wiley-VCH: New York, 1997. (13) Koziel, J.; Pawliszyn, J. J. Air Waste Manage. Assoc. 2001, 51, 173-185. (14) Chai, M.; Pawliszyn, J. Environ. Sci. Technol. 1995, 29, 693-701. (15) Grote, C.; Pawliszyn, J. Anal. Chem. 1997, 69, 587-596. (16) Eisert, R.; Pawliszyn, P.; Barinshteyn, G.; Chambers, D. Anal. Commun. 1998, 35, 187-189. (17) Martos, P.; Pawliszyn, J. Anal. Chem. 1997, 69, 206-215. (18) Martos, P.; Pawliszyn, J. Anal. Chem. 1998, 70, 2311-2320. (19) Martos, P.; Pawliszyn, J. Anal. Chem. 1999, 71, 1513-1520.
1486
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 35, NO. 7, 2001
(20) Koziel, J.; Jia, M.; Khaled, A.; Noah, J.; Pawliszyn, J. Anal. Chim. Acta 1999, 400, 153-162. (21) Martos, P. Ph.D. Dissertation, University of Waterloo, Canada, 1998. (22) Koziel, J.; Jia, M.; Pawliszyn, J. Anal. Chem. 2000, 72, 51785186. (23) Augusto, F.; Koziel, J.; Pawliszyn, J. Anal. Chem. 2001, 73, 481486. (24) Seinfeld, J. Atmospheric Chemistry and Physics of Air Pollution; John Wiley & Sons: New York, 1986. (25) Grosjean, E.; Grosjean, D.; Frasier, M. P.; Cass, G. R. Environ. Sci. Technol. 1996, 30, 2687-2703. (26) Gaffney, J. S.; Marley N. A.; Martin, R. S.; Dixon, R. W.; Reyes, L. G.; Popp, C. J. Environ. Sci. Technol. 1997, 31, 3053-3061. (27) McLaren, R.; Singleton, D. L.; Lai, J. Y. K.; Khouw, B.; Singer, E.; Wu, Z.; Niki, H. Atmos. Environ. 1996, 30, 2219-2232.
Received for review September 11, 2000. Revised manuscript received January 29, 2001. Accepted January 30, 2001. ES001653I
