CA2389726C - Analytical devices based on diffusion boundary layer calibration and quantitative sorption - Google Patents

Abstract

This invention relates to a method and apparatus of determining the concentration of analytes in sample using an extraction device whereby the concentration of analytes of interest can be determined from the diffusion coefficient for said analytes. Various designs of the extraction devices are considered including fibres, membranes, coated vessels and stirring mechanisms. To facilitate well-defined boundary layer, agitation means are incorporated in the design of the devices.

Description

BACKGROUND OF THE INVENTION
FIELD OF INVENTION
This invention relates to a method and apparatus of determining the concentration of analytes in sample using an extraction device whereby the concentration of analytes of interest can be determined from the diffusion coefficient for said analytes.
More particularly, this invention relates to a method and apparatus for determining the concentration of organic and inorganic compounds in liquid and gaseous samples.
DESCRIPTION OF THE PRIOR AR'h It is known to use solid phase microextraction (SPME) and polydimethylsiloxane 1o (PDMS) coated fibers to extract volatile organic compounds (VOC's) in environmental samples. PDMS is the most widely used coating for extracting nonpolar volatile analytes as well as many polar analytes. However, the sensitivity of mixed phase SPME
coatings, such as PDMS/DVB and Carboxen/PI~MS was reported to be much higher compared to PDMS coating for extracting VOC's (see Mani et al, Applications of Solid Phase 15 Microextraction, RSC, (:ornwall, U.K., 1999, Chapter 5). Mixed phase coatings have some complementary properties compared to PDMS and are more suitable for sampling highly volatile species (see Pawliszyn, Solid-Phase Microextraction: 'Theory and Practice, Wiley-VCH, Inc., New York, 1997, Chapter 4). Mixed phase SPME fibers have been used for sampling and quantifying target VOC's present in indoor air at the part per 2o billion level and even at the part per trillium level.
Indoor air quality and its potential impact on human health is an increased concern to the public and government environmental agencies. Many VOC's, such as formaldehyde, aromatic compounds, and halogenated hydrocarbons have been found to be highly toxic to humans. Large-scale air quality testing by conventional air sampling 2s methods can often be time-consuming and expensive. Solid phase microextraction coupled with gas chromotography has been previously successfully applied to analyze various air samples. Chai et al., Analyst, 1993, 118,1541 reported the determination of the presence of volatile chlorinated hydrocarbons in air by SPME in 1993.
Martos et al.
developed a new method using a linear temperature-programmed retention index method 3o to calibrate SPME devices for fluctuations in sampling temperature, and for air analysis of total VOC's with (PDMS) fibers (see Martos et al., Analytical Chemistry, 1997, 69, 206 and 402). Grote et al used SPMI? for fast quantitative analysis of acetone, isoprene and ethanol in human breath with SPME fibers in Analytical Chemistry, 1997, 69, 587.
It is known that the syringe-like SPME device is portable and can be easily used for field analysis. When coupled with a field-portable gas chromatograph, both SPME
sampling and instrumental analysis can be conducted at the test site without the need for sample preservation (see Koziel et al, Analytical Chemistry, Acta, 1999, 400(1-3), 153.
In mixed phase coatings, the majority of interaction on porous polymer particles is determined by the adsorption process. W ith mixed phase coatings, the molecules can be attracted to a solid surface via van der Waals, dipole-dipole, and other weak to intermolecular forces (see Gorecki et al, Applications of Solid Phase Microextraction, RSC, Cornwall, UK, 1999, Chapter 7 j. Hydrophobic interaction and electrostatic interaction also occur when extracting analytes from water and ionizable analytes from aqueous phase, respectively. Compared to the dilTusion coefficient in liquid coatings of PDMS or PA, the diffusion coefficients of VOC's in divinylbenzene and Carboxen are so small that, within the frame of SPME analysis, essentially all the molecules remain on the surface of the coating. Therefore, the fundamental difference between adsorption and absorption is that in adsorption molecules bind directly to the surface of a solid phase while, in absorption, they dissolve into the bulk of the liquid phase.
'The Langmuir adsorption isotherm is one of most important adsorption theories. The 2o Langmuir model assumes there is only a limited number of surface sites that can be occupied by analyte molecules, all sites are equivalent, and there is no interaction between adsorbate molecules on adjacent sites. The Langmuir adsorption isotherm was used to describe the adsorption equilibrium on PDMS/DVB and Carbowax/DVB
coatings. A linear function is found to exist only if the affinity of an analyte toward the coating is low or its concentration in the sample is very low. In a real sample matrix, for example, air, there are usually more than two components. Since different components have different affinities towards the active sites, the presence of mufti-components must affect the adsorption of one other. Unlike the non-competitive absorption process in liquid coatings, adsorption process onto porous polymer coatings in a mufti-component 3o system is a competitive process and therefore displacement effect is expected. Sampling conditions, the sample matrix composition and concentration can largely affect the amount of analytes extracted by mixed phase fibers. From a practical point of view, this makes quantitative analysis using porous polymer SPME coatings more difficult.
The majority of adsorption models are based on the equilibrium theory. In SPME, however, the equilibrium time ranges from a few minutes to a couple of hours depending on the nature of the analytes and the sampling conditions (see Ai et al., Applications of Solid Phase Microextraction, RSC, Cornwall, UK, 1999, Chapter 2). For porous solid coatings, the equilibrium time for the same analyte is usually much longer than that in liquid coatings. It may be impractical to wait for partition eduilibrium of all of analytes in the matrix if the equilibrium times for some analytes are too long.
to In the direct SPME system, such as sampling in air ur in water, the analyte movement proceeds in two steps. The first step consists of the mass transfer of analytes from the bulk sample matrix to the surface to the SPME polymer coating followed by diffusion of the analytes within the coating. Fick's first law of diffusion (equation 1;) can describe the rate of mass diffusion in the sample matrix in the coating as follows:
~' - -DS(aCs ) (eq.l) c'~ x Where F is the flux of analyze in the direction x from the sample matrix bulk to the SPME fiber surface.
(i) DS is the diffusion coefficient of the analyte in the sample matrix, (ii) CS is the analyte concentration in the sample bulk.
2o In a static gas system, mass movement results only from molecular diffusion due to intermolecular collisions. In practice, both molecular diffusion and bulk fluid movement must be considered. 'rhe extent of fluid movement (agitation), reflects the access of analytes to the surface and is frequently described as a theoretical parameter called the boundary layer (Cooper et al, Air Pollution Control: A Design Approach, Waveland Press Inc., Prospect Heights, 1994, Chapter 1 ~).
According to the boundary layer theory, a laminar sublayer or the sample matrix film is formed when a fluid passes a fixed object. The only way that the analyte can pass from the air bulk phase to the surface of the coating is via molecular diffusion across the boundary layer. In the liquid/solid interface, the thickness of the boundary layer is determined by the agitation conditions and the viscosity of the fluid (see Pawliszyn, Solid Phase Microextraction: Theory and Practice. Wiley/VCH, Inc., New York, 1997).
In a gas system, air wind velocity is a very important factor in mass transfer process.
Because the value of wind velocity represents the degree of bulk air movement, wind velocity will influence the overall mass transfer rate in the bulk of fluid.
Based on mass transfer theories, the mass transfer rate of an analyte is proportional to the mass diffusivity, and inversely proportional to the Thickness of gas film at the interface.
Many factors such as temperature, pressure, molecular structure and molecular weight can directly affect the molecular diffusion coefficients of VOC's (see Lugg, G.A., Analytical Chemistry, 1968, 40 (7), 1072). Since accurate experimental measurement of the diffusion coefficient is difficult, relatively few values for organic compounds in gas systems are available from the literature. A number of methods have been proposed for estimation of diffusion coefficients of VOC"s in air systems. 'the method by Fuller, ~ 5 Schettler and Giddings (FSG method) was reported to be most accurate for non-polar organic gases at low to moderate temperature (see Lyman et al, Handbook of Chemical Property Estimation Method, ACS, McGraw-Hill, Inc., New York, 1982, Chapter 17).
Minimal error is associated with the aliphatics and aromatics. FSG model describes that the molecular diffusion coefficient of an analyte is directly proportional to temperature, and inversely proportioned to air pressure. The relative humidity of air is another factor that can affect VOC extraction on SPME fibers because water molecules participate in the adsorption process.
SUMMARY OF THE 1NVEN'rION
A method of determining a concentration of analytes of interest in a sample using a solid phase microextraction device having a surface containing an extraction coating comprises bringing the sample into contact with the coating while highly agitating the sample under controlled conditions to maintain a substantially constant boundary layer between the sample and the coating. 'The method further comprises limiting a time of contact between the sample and the coating and siring the coating so that all analytes that 3o pass through the boundary layer are adsorbed by the coating. The method further comprises terminating the contact, determining the amount of each anaiyte of interest in the coating and calculating the concentration of each analyte of interest in the sample by using the diffusion coefficient for that analyte.
A method of determining the concentration of analytes of interest in a sample uses an extraction device having a membrane. The method comprises bringing the sample into contact with the membrane for sufficient time to allow tnicroextraction to occur. The method further comprises choosing a membrane with a large surface area and limiting the time of contact so that all analytes that contact the membrane are adsorbed by the membrane. The method further comprises separating the membrane from the sample, 1o determining the amount of each analyte of interest in the membrane and calculating the concentration of each analyte of interest in the sample using the diffusion coefficient for that analyte.
A device for determining the concentration of analytes of interest in a sample, the device comprises a solid phase microextraction device having a surface containing an 15 extraction coating characterized by a large surface area to adsorb all analytes that contact said coating and agitation means to highly agitate the sample under controlled conditions during microextraction.
In a further embodiment, a device for determining the concentration of analytes of interest in a sample uses a membrane having a large surface area to adsorb all analytes 2o that contact a surface of the membrane in a time allowed for extraction.
The membrane is sized and shaped to fit into an injection port of an analytical instrument.
BRIEF DESCRIPTION OF 'THE DRAWINGS
Figure 1 is a flow diagram of an air sampling system;
Figure 2 is a further embodiment of an air sampling system for samples having 25 varying humilities;
Figure 3 is a graph showing the effect of wind velocity on adsorption;
Figure 4 is a graph showing the extraction time profiles of toluene under different wind velocities;
Figure 5 is a graph of the slopes of toluene extraction time profiles and wind velocity;

Figure 6 shows extraction time profiles for difference concentrations of benzene at high wind velocities;
Figure 7 is a graph of the amount of benzene adsorbed and concentration after one minute of sampling time;
Figure 8 is a graph of the amount of adsorption with temperature;
Figure 9 is a graph showing the relationship between the amount adsorbed at an exposure time of 5 seconds with increasing temperature;
Figure 10 is a graph showing the amount of adsorption with temperature at a 10 second sampling time with a different fiber than that used for Figure 9;
to Figure 11 is a graph showing the amount of benzene adsorbed with time at different humilities;
Figure 12 is a graph showing the amount of toluene adsorbed with time at different humilities;
Figure 13 is a graph showing the amount of p-Xylene adsorbed with time at different l5 humilities;
Figure 14 shows an extraction device having an electric blower to provide constant air agitation;
Figure 1 SA is a schematic side view o P an extraction device having a fiber with an extraction coating thereon;
2o Figure 1 SB is a schematic side view of a flaw tube having an extraction coating on an inner surface thereof;
Figure 1 SC is a schematic side view of a vessel having an extraction coating on an interior surface of the vessel;
Figure 15D is a schematic side view of a vessel containing particulates with an 2s extraction coating thereon;
Figure 1 SE is a schematic side view of a vessel having a stirrer with an extraction coating on the stirrer;

Figure 15F is a schematic side view of a vessel having a stirring bar with an extraction coating on said bar;
Figure 16 is a schematic perspective view of a boundary layer surrounding a silica rod having an extraction coating thereon a.nd a graph of concentration profile;
Figure 17 is a perspective view of an expanded membrane attached to a handle;
Figure 18 is a perspective view of the membr~me of F figure 17 rolled around the handle;
Figure 19 is a schematic side view of a holder having three fibers; and Figure 20 is a schematic side view of a device for determining the concentration of 1 o analytes in a liquid.
DESCRIPTION OF A PREFERRED EMBODIMENT
Compared to the diffusion coefficient and liquid coatings of PDMS or PA, the diffusion coefficients of VOC's in divinylbenzene and C",arboxen are so small that, with the frame of SPME analysis, essentially all of the molecules remain on the surface of the 15 coating. Sampling conditions, the sample matrix composition and concentration can largely effect the amount of analytes extracted by mixed phase fibers. From a practical point of view, this makes quantitative analysis using porous polymer SPME
coatings more difficult.
It has been found that with a very short exposure time (for example, one minute), 20 when using SPME with PDMS/DVB coating fibers for fast sampling and analysis of VOC's in indoor air, there is a linear relationship between adsorption and concentration.
Within a one minute sampling period, airborne benzene, toluene, ethylbenzene and p-xylene (BTEX) extracted on a PDMS/DVB fiber increased linearly with the sampling time. The short exposure time before equilibrium produces an advantage due to the fact 25 that the adsorption rate is controlled by diffusion coefficients of analytes rather than their distribution constants. Because the differences between the diffusion coefficients of VOC's are much smaller compared to the differences of the distribution constants are much smaller than the differences between the distribution constants, all target VOC's with similar molecular weights produce similar extraction rates when using a short sampling time.
The mass transfer parameters through a boundary layer should include both molecular diffusion and bulk fluid movement. When using a porous polymer SPME coating for air sampling, it can be reasonably assumed that all available analyte molecules are mobilized within a very short exposure period. In other words, when the concentration of analyte on the coating surface is far from the saturation point, all of the target molecules are immediately adsorbed as soon as they contact the surface of the porous solid extraction coating. If the matrix composition and sampling conditions are kept constant, it has been 1o found that the rate of mass diffusion of analyte will be proportional to its mass diffusion coefficient in the sample bulk within this short time period. Also, it has been found that there exists a quantitative relationship between the amount of analyrtes adsorbed and concentration depending on the diffusion coefficient of analytes when using a very short exposure time that occurs well before equilibrium.
1 s In the gas-solid interface, the thickness of the gas film is largely affected by the air movement or by the wind velocity and the nature of air. In a gas system, air wind velocity is a very important factor in the mass transfer process. Since the value of wind velocity represents the degree of bulk air movement, wind velocity will influence overall mass transfer rate in the bulk of the fluid. Based on mass transfer theories, the mass 2o transfer rate of an analyte is proportional to the mass diffusivity and inversely proportional to the thickness of the gas film at the interface. ~Cherefore, when considering air sampling with porous SPME fibers, air wind velocity is a very important factor related to the adsorption process, especially for pre-equilibrium extraction.
With the present invention, some of the critical factors, including air/wind velocity, 25 sampling temperature and air relative humidity have been investigated in relation to the adsorption process of VOC's onto porous polymer SPME coatings under non-equilibrium conditions.
Extraction Model Development. 'the solid SPME fiber coating can be modeled as a long cylinder with length L, and outside and inside diameters of b and a, respectively 30 (Figure I). When the coating is exposed to moving air, an interface (or boundary layer) with thickness b develops between the bulk of air and the idealized surface of the fiber.
The analytes are transported from the bulk air to the surface of the coating via molecular diffusion across the boundary layer. In most cases, the molecular diffusion of analytes across the interface is the rate-limiting step in the whole adsorption process.
5 The analyte concentration in the bulk air (Cg) can be considered constant when a short sampling time is used, and there is a constant supply of an analyte via convection.
These assumptions are true for most cases of SPME air sampling, where the volume of air is much greater then than the volume of the interface, and the extraction process does not affect the bulk air concentration. In addition, the SPME solid coating can be treated 1o as a perfect sink. The adsorption binding is instantaneous and the analyte concentration on the coating surface (CO) is far from saturation and can be assumed to be negligible for short sampling times and relatively low analyte concentrations in a typical air. These concentrations range from parts-per-trillion (by volume) to parts-per-million (by volume) for most VOCs of interest and typical industrial hygiene, indoor and ambient air concentrations. The analyte concentration profile can be assumed to be linear from Cg to C0. In addition, the initial analyte concentration on the coating surface (CO) can be assumed to be equal to zero when extraction begins. Diffusion inside the pores of a solid coating controls mass transfer from b to a.
The mass of extracted analyte with sampling time can be derived using the 2o analogy of heat transfer in a cylinder with inside and outside diameters of b and b, respectively, with a constant axial supply of heat. The steady-state solution to heat transfer can be translated into a mass transfer solution by replacing temperatures with concentrations, heat with flux of mass and heat transfer coefficient with gas-phase molecular diffusion coefficient. As a result, the mass of extracted analyte can be estimated from the following equation:
2rcD 1.
) In b~+8 ,~(:"x (t)dt b ~ 1 () where: n is the mass of extracted analyte over sampling time (t) in ng; Dg is the gas-phase molecular diffusion coefficient (em2/s); b is the outside radius of the fiber coating (cm); L is the length of the coated rod (cm); b is the thickness of the boundary layer surrounding the fiber coating (cm); and Cg is analyte concentration in the bulk air (ngJmL). It can be assumed that the analyte concentration is constant for very short sampling times and therefore Equation 1 can be further reduced to:
C r In b+bl x b /, (2) where t is the sampling time (s). The fiber length and th.e outside diameter of the fiber coating are constant for each type of the fiber. The nominal length for the 65 p,m 1o PDMS/DVB and the 75 p,m Carboxen'rM/PDMS coatings is L = 1 cm, and the outside diameter 2b = 0.0240 cm (~10%) and 2b = 0.026() cm (~l 0%), respectively.
It can be seen from Equation 2 that the amount of extracted mass is proportional to the sampling time, Dg for each analyte, bulk air concentration, and inversely proportional to 8. This in turn allows for quantitative air analysis. Equation 2 can be t 5 modified to estimate the analyte concentration in the air in ng/mL for rapid sampling with solid SPME coatings:
n 1nC b + 8 a _ b 2~Dg Lt (3) The amount of extracted analyte (n) can be estimated from the detector response.
2o For a special case, where the thickness of the boundary layer is much smaller than the outside radius of the fiber (8<<b), the general solution can be reduced to a flat plate problem. For such condition, ln(1+8/b) ~ b/b, 2~cbl, _ ~~, and Equation 2 simplifies to:

D A
n(t) = S Cg t (4) where A is the surface area of the sorbent. Equation 4 is analogous to the mass uptake model for the TWA sampling with retracted SPME fiber, where the distance between the needle opening and the fiber (Z) is replaced by 8.11,12 Under equal conditions, the amount of extracted mass will be greater for an analyte with a greater gas-phase molecular diffusion coefficient (Dg). This is consistent with the fact that the analyte with a greater Dg will cross the interface and reach the surface of the fiber coating faster. Values of Dg for each analyte can be found in the t o literature or estimated from physicochemical properties. A number of methods have been proposed for estimation of diffusion coefficients of VOCs in air systems. The method by Fuller, Schettler and Giddings (FSG) was reported to be the most accurate for non-polar organic gases at low to moderate temperatures:
0.001x'1.75 x 1._+._._~__ Mair Mo_oc Dg - r . _-.17 P~~, fair ~3 +~~ vvnc'3 (5) where Dg is expressed in cm2/s; T is the absolute temperature (K); Mair, Mvoc are molecular weights for air and VOC of interest (g/mol); p is the absolute pressure (atm);
Vair, Vvoc are the molar volumes of air and the VOC of interest (em3/rnol).
According to the FSG model, Dg is directly proportional to temperature and inversely proportional to air pressure. Because the atmospheric pressure changes are relatively low, the air temperature is a more important factor than pressure when considering air sampling.
Regardless, both atmospheric pressure and air temperature are routinely monitored during conventional air sampling.
The thickness of the boundary layer (8) is a function of sampling conditions.
The most important factors affecting 8 are SPME coating radius, air velocity, air temperature and Dg for each analyte. The effective thickness of the boundary layer is determined by both rate of convection and diffusion. As the analyte approaches the sorbent surface, the overall flux is increasingly more dependent of diffusion than convection. The analyte flux in the bulk sample is assumed to be controlled by convection, whereas the analyte flux inside the boundary layer region is assumed to be controlled by diffusion. The effective thickness of the boundary layer can be described as the location where this transition occurs, i.e., where the flux towards cS (controlled by convection) is equal to the flux towards the surface of the SPME coating (controlled by diffusion). In the Nernst model, the matrix within the boundary layer is stationary. Experimental research indicated that convection was also present inside the boundary layer. However, its t o effects decreased with the distance to the solid surface. The effective thickness of the boundary layer can be estimated using Equation 6, adapted tiom the heat transfer theory for an SPME fiber in a cross flow:
8 = 9.52 ~' (6) Reo'~2Sco~38 where Re is the Reynolds number = 2ublY, a is the linear air velocity (cm/s);
v is the t5 kinematic viscosity for air (em2/s); Sc is the Schmidt number = vlDg. The effective thickness of the boundary layer in Equation 6 is a surrogate (or average) estimate and does not take into account changes of the thickness that may occur when the flow separates and/or a wake is formed. Equation 6 indicates that the thickness of the boundary layer will decrease with an increase of the linear air velocity.
Similarly, when 2o air temperature (Tg) increases, the kinematic viscosity also increases.
Since the kinematic viscosity term is present in the numerator of Re and in the denominator of,Sc, the overall effect on 8 is small.
The gas-phase molecular diffusion coefficient (DR) for each analyte is also an important parameter controlling & As illustrated in Equation 6, the effective thickness of 25 the boundary layer will be reduced for analytes with lower D~. This can be explained considering that, analytes with low molecular weight will reach the coating surface faster then the less volatile analytes under equal experimental conditions and therefore the point at which the diffusion is a primary mode of analyte transport to the coating is located further away from the surface. The reduction of the boundary layer and the increase of 1 ~1 the mass transfer rate for m analyte can be achieved in at least two ways, i.e., by increasing the air velocity and by increasing the air temperature. However, the temperature increase will reduce the solid sorbent efficiency. As a result, the sorbent coating may not behave as a zero sink for all analytes.
Chemicals and Supplies. The volatile organic compounds under study, i.e., benzene, toluene and p-xylenes were purchased from Sigma-Aldrich (Mississauga, ON). All VOC
standards had purities >_ 98.0% and used for calibrating GC/FID response factors.
National Institute of Standards and Technology (NIST) traceable certified permeation tubes of benzene, toluene and p-xylene were purchased from Kin-Tech (La Marque, TX), 1o and used for the generation of a standard gas mixture. Ultrahigh purity hydrogen, nitrogen, air were purchased from Praxair (Waterloo, ON). SPME fibers with 65 p,m PDMS/DVB, 75pm Carboxen/PDMS and SPME holders were purchased from Supelco (Oakville, ON).
Standard Gas. A standard gas-generating device with a flow-through sampling chamber, was constructed to provide a wide range of target VOC concentrations at constant temperature. Ultrahigh purity air (zero gas) was supplied from a Whatman air generator (Haverhill, MA) and maintained at 50 psi head pressure. Permeation tubes of benzene, toluene and p-xylene were held inside a glass permeation tube adapter (Kin-Tech, La Marque, TX) and swept with a constant flow of dilution air. The adapter was 2o placed inside a cylindrical aluminum coven, which was heated by two heating elements ( 100 W), and its temperature was controlled by K-type thermocouple (Omega ~~M, Stamford, CT) and an electronic heat control device (Science Shops, the University of Waterloo, ON). The air flow rate was controlled by two Sidetrack ~~M mass flow controllers (Sierra Instruments, Monterey, CA) placed on both the primary and the dilution loops in the system. Wide ranges of concentration for target VOC's were obtained by adjusting both the air flow rate and the permeation tube incubating temperature.
Design for Air Wind Velocity Study. As shown in Figure l, an air sampling system 2 consists of a main cylindrical glass chamber 4 and four additional cylindrical glass 3o chambers 6, 8, 10, 12. All of the glass chambers 4, 6, 8, 10, 12 have different diameters with the chamber 4 being the largest. The chamber 6 has the smallest diameter with the chambers 8, 10, 12 each increasing in diameter in chronological order. The main chamber 4 contains an SPME device 14, which is described in more detail subsequently.
A 1.0 L glass sampling bulb 16 (Supelco, Oakville, ON;1, was constructed and installed 5 downstream from the standard gas generator. The gas flow rate varied from 1,000 standard cubic centimeters per minute (scan) to 4,000 scan for generating a wide range of air wind velocities. This new sampling system can provide both a dynamic airflow with different wind velocities and a static gas mixture. Experiments for estimation of the range of air velocities in typical indoor environments were conducted in a mechanically 10 ventilated building using an OMEGA ~ M HHFS 1 Temperature and Air Velocity Meter (OMEGA, Stamford, CT). The indoor air velocities were found to vary with the distance between the measured location and the air vent, which is shown in the Table 1.
The average indoor air velocity varied from 0 to 10 cm/s, and the average wind velocity at ventilating zones (near vent) varied from 15 to 40 cm/s. The data listed in Table 1 were 15 consistent with the values reported by Wasiolek et al. ( 1999) Indoor Air-International Journal Indoor Air Quality and Climate, 9 (2) 125, who found that the average indoor wind velocities (at 19 locations in a workroom) varying from 1.4 to 9.7 cm/s, and the average wind velocity at the breathing-zone height varying from 9.9 to 35.5 cm/s by using an accurate three-dimensional sonic anemometer. T'he glass chambers 4, 6, 8, 10, 12 allowed for sampling under dynamic flow conditions, and the chamber 16 allowed for static sampling when stopcocks 18 and 20 were closed. A stopcock 22 opens and closes an exhaust line 24. The average wind velc>cities were calculated by dividing the airflow rate by the cross-section area of gas sampling chambers. When the air flow rate was set at 1,000 scan, the air velocities ranged from 0.2 to 20.8 cm/s. When the air flow rate was increased to 4,000 scan, the air velocities ranged from 0.8 to 83.2 cm/s.
Since the Reynolds numbers in all chambers were less than 1,200, the air flow in the sampling chambers was in a laminar flow ecrndition. A 65 qm PDMS/DVB fiber was used to sample the VOC gas mixture in each sampling port under different average air velocities.
A short exposure time of 20 seconds was used to examine the effect of wind velocity on 3o the VOC adsorption process onto PDMS/DVB fiber. The extraction time profiles of airborne BTEX were also constructed under various wind velocities.

Design for Temperature Study. The gas flow rate was maintained at 1,000 sccm, and the permeation tubes were incubated at 60 °C. The main sampling chamber 4 in Figure 1 was used to provide a steady-state mass flow of VOC's at different temperatures. The air temperature in the vicinity of the SPME fibers was maintained within ~ 0.3°C at the range of room temperature. A 65 pm PDMS/DVB fiber and a 75 pm Carboxen/PDMS fiber were used to sample the VOC gas mixture in the chamber.
The temperature of the air stream in the chamber varied from 22 to 40°C:. The SPME
fiber exposure times were 5 seconds and 1 U seconds, respectively.
Design for Air Humidity Study. As shown in Figure 2, to create a dynamic airflow 1o under different humidities, an in-line impinger trap 26 (Supelco, Oakville, ON), and a humidity meter 28 (Radio Shack, Waterloo, ON) were installed in the air sampling system. The components of Figure 2 that are identical to the components of Figure 1 are described using the same reference numerals as those used in F figure 1 without further description. Relative humidities of 47% and 75% were obtained by maintaining the water level in the impinger trap at 1.0 cm and 8.0 crn height, respectively. A
65 pm PDMS/DVB fiber was used to sample VOC's in the gas mixture under different humidities.
Gas Chromatography. A Varian 3400 GC (Varian Associates, Sunnyvale, CA), equipped with a FID and a carbon dioxide-cooled septum programmable injector, was 2o used to analyze air samples extracted by SPME fibers and liquid samples of standard compounds. An SPB-5 capillary column (30 m x 0.25 mm i.d., 1.0 ~m film thickness) was installed in the GC, and UHP helium was used as the carrier gas with a flow rate of 2.0 mL/min at 26 psi head pressure. 'The oven temperature program was 50 °C for 1 min, 15 °C/min to 240 °C and held for 2 min. For SPME fiber desorption, the injector temperature was isothermally set at 300 °C for Carboxen/PDMS fibers, and at 250 °C for PDMS/DVB fibers. For liquid injection, the injector was programmed from 45 °C to 225 °C at a ramp of 300 °C /min. The quantification of target VOC's in standard gases was based on the response factors obtained from the F1D signals by liquid injection of VOC
standards in the test range.

Effects of Air Velocity. Figure 3 shows the effect of wind velocity on the adsorption of benzene, toluene, p-xylene and ethylbenzene on a 75 micrometers Carboxen/PDMS coating for Ss sampling of airborne BETEX. Each data point represents a normalized mass, i.e., the ratio of adsorbed mass and the analyte concentration in air, and is shown with ~ one standard deviation for three samples. Figure 3 clearly indicates that two distinct regimes of mass transfer are present: regime (1) where the extracted amount depends on the air velocity and regime (2) where the air velocity has a less significant effect on the amount of extracted mass ("semi- plateau" region).
The two zone phenomena can be explained by considering an interface between l0 air and the porous solid sorbent. 'The first region in Figure 3 describes diffusion of analytes through the static, well-developed boundary layer surrounding the SPME
coating. In this region, the increase in air velocity causes a reduction in the boundary layer thickness and more of each analyte can be extracted per unit of time.
This finding is consistent with the theory summarized by Equation 2. In the second region, above 15 some critical velocity, the thickness of the boundary layer is further reduced, but it is small enough that the mass transfer is controlled by the diffusion inside the pores of the SPME coating. Therefore the increase in air velocity has only a small effect on the amount of extracted analyte.
The critical velocity for which the effects of the boundary layer thickness are 2o negligible is approximately 10 cm/s for the analytes in this study.
Although this range is lower than the average air velocities in ambient air, the critical velocity is close to the range of measured air velocities in typical indoor air. Reported average indoor air velocities at the breathing-zone height varied from 9.9 to 3S.S cm/s, with the average of 19 locations in a workroom varying from 1.4 to 9.7 cm/s. Particular care must be taken to 25 ensure the reproducibility of extraction conditions with porous SPME fibers in field sampling. This is because a small change in air velocity in the vicinity of solid SPME
fiber can have a significant effect on the amount of adsorbed analyte, particularly in the first mass transfer region (Figure 3).
Considering the fact that the amount of extracted mass for solid SPME fibers can 3o be enhanced when sampling is conducted at greater air velocities, i.e., in the "semi-plateau" region (Figure 3), an external fan or an attachment to an air sampling pump can be used to provide greater rate of mass transfer. Such a device could be used by air sampling professionals wishing to equalize the extraction conditions and provide reproducible effective thickness of the boundary layer for each sample. The use of a higher air velocity for sampling with solid SPME coatings leads to enhanced sensitivity.
Preliminary results indicate that the use of solid PDMS/DVB 65 ~m fiber coating, 30 s sampling and average air velocity of 1 m/s allows for detection of BTEX at 10 ppt (by volume) range.
The greatest amount of mass was adsorbed for benzene, followed by toluene, p-xylene and ethylbenzene. This finding is consistent with theory presented in Equation 2, i.e., the mass of adsorbed analyte using rapid sampling is proportional to the Dg for each analyte, when all other sampling conditions are equal. The 75 p,m CarboxenTM/PDMS
coating was acting as a zero sink for short sampling times. The ratio of normalized masses in Figure 3 for benzene and toluene was close to the ratio of their Dg s estimated by the FSG method. Normalized masses for ethylbenzene andp-xylene were smaller than expected. This discrepancy is likely associated with experimental errors.
Figure 4 shows the extraction time profiles of toluene using a 65pm PDMSIDVB
fiber under different wind velocities ranging from 0.8 to 83.2 cm/s. These curves illustrate the variation of toluene uptake within the whole range of indoor air wind speed.
The toluene mass loading on PDMS/DVB fiber linearly increases with the sampling time within a short period of time ( 1 min). Furthermore, within this short sampling time, the toluene uptake rises with the increase of wind velocity. However, the equilibrium mass loading of toluene generally decreases with the increase of wind speed. 'This is caused by the fact that other, more strongly bound compounds extract faster as well resulting in a faster occurring displacement effect.
A further examination of the wind speed effect indicates that there are different influences on toluene adsorption on PI)MS/DVB fiber when sampling at different wind velocities. Generally, the slope of the toluene extraction time profiles increases with the increase of wind velocity. if we plot the slopes of toluene extraction time profiles against 3o the average wind velocities applied, we can obtain Figure 5. This figure demonstrates l9 that the slope increases approximately linearly as the wind speed increases from 0.8 to 8.7 cm/s. This means that the toluene mass loading on PDMS/DVB fiber was significantly affected by the variation within the average indoor wind velocity range (0-em/s), and an approximately linear increase of mass loading can be expected within this wind range. Only a slight increase of toluene extraction was found as the wind speed increased from 8.7 to 83.2 cm/s. This range of wind velocity is usually found at indoor air ventilating zones. This indicates that the mass loading of toluene is only slightly affected by the variation of wind speed within indoor air ventilating zones (>
10 em/s).
Air Sampling under Wind Velocities Equal to or above a Critical Air Velocity.
to Figures 4 and 5 suggest that air sampling with a PDMS/DVB fiber should be conducted above some critical air velocity, above which the mass transfer and the analyte uptake are not affected by air velocity variation. Figure 6 shows extraction time profiles for benzene using a 65 p.m PDMS/DVB fiber to sample a standard VOC gas mixture at wind velocities equal to or above 10.2 cm/s. Three VOC concentrations were obtained by setting different incubation temperatures (35, 40 and 60"C",) for VOC
permeation tubes and a constant air flow at 2,000 scan. Another VOC concentration was generated by maintaining the incubation temperature at 60°C and increasing the air flow rate to 4,000 sccm. These curves illustrate that the benzene uptake increased with the sampling time before reaching its equilibrium level. The higher the concentration, the less time was 2o required for the PDMS/DVB fiber to reach the equilibrium. However, only within a very short sampling time ( 1 min), was benzene mass loading approximately linear with sampling time. Figure 7 shows that benzene uptake or response increased linearly with the concentration when 1 min sampling time was used under an air velocity equal to or above 10.2 cm/s. For other target VOC's, i.e., toluene andp-xylene, similar results were also observed (data not shown).
Temperature Effect on VOC Adsorption on Porous SPME Fibers. Figure 8 shows that amounts of toluene andp-xylene adsorbed on the PDMSIDVB fiber for a Ss exposure increase linearly as the temperature increases from 22 to 26°C, while benzene uptake remains almost constant in this temperature range. As the temperature increases 3o continuously, the amounts of toluene and p-xylene adsorbed increase slightly, while the amount of benzene adsorbed decreases. This indicates the displacement of benzene molecules byp-xylene or toluene molecules, which have higher affinities to the PDMS/DVB coating than benzene. T'he results indicate that increasing temperature within a certain range will enhance the adsorption of VOC's on PDMS/DVB
coating, 5 especially for analytes with higher affinity to the coating. Similarly to Figure 8, Figure 9 shows that the mass of toluene and p-xylene adsorbed onto the PDMS/DVB fiber for a lOs exposure increase with the increase of temperature from 22 to 25°C.
Benzene adsorbed remains almost constant from 22 to 25°C, but decreases as the temperature increases further ftom 25 to 40°C. In fact, this situation should be expected because the active surface sites become saturated as the adsorption on the coating proceeds, and some benzene molecules are possibly displaced by toluene or p-xylene molecules.
In Figure 10, amounts of all three analytes adsorbed on the CarboxeWPDMS fiber increase linearly as the temperature increases from 22 to 25°C when using l Os sampling time. As the temperature increases continuously, the mass of analytes adsorbed only ~5 increases slightly. Generally, a similar adsorption behavior between PDMS/DVB and Carboxen/PDMS was observed. However, the Carboxet~/PDMS fiber has a higher adsorption capacity than PDMS/DVB fiber for extraction of benzene, the smallest molecule among the analytes. Unlike the DVB particle consisting of mainly mesopores and a smaller fraction of macropores and micropores, the Carboxen polymer particle has 2o an even distribution of micro, meso, and macro pares. Therefore, Carboxen particles are better for sampling smaller molecules (CZ-C~2) compared to PDMS/DVB fiber.
Unfortunately, the weakness of CarboxenLPDMS is the difficulty for analyte desorption.
Peak tailing is often observed even with the GC injector temperature of 300 °C.
As one of the most important experimental parameters in SPME sampling, the extraction temperature has been discussed in several previous papers related to SPME air sampling. When a pure-phase liquid SPME fiber is used, an increase in extraction temperature usually causes an increase in extraction rate, but simultaneously a decrease in the distribution constant. Since the extraction by the SPME coating is an exothermic process, a decrease in mass loading at equilibrium is usually expected as the extraction 3o temperature increases. In contrast to liquid fiber, however, an opposite trend of temperature effect was found in this study when mixed-phase porous SPME fibers were used. Within a very short sampling time (far .from equilibrium), VOC analyzes on a porous SPME fibre can linearly increase as the extraction temperature increases in a narrow range. Since VOC', adsorption on a porous SPME fiber is controlled by the diffusion process or diffusion coefficient rather than the extraction equilibrium or distribution constant, an increase in diffusion coefficients with an elevated temperature should increase VOC uptake on the solid SPME coating.
Effect of Humidity on VOC Adsorption onto PDMS/DVB Fiber. Figures 11-13 are the extraction time profiles of benzene, toluene and ~-xylene using a 65 p.m PDMS/DVB fiber to extract a standard VOC gas mixture at different humidities.
These to figures indicate that a humidity level of 7S% resulted in a significant decrease in the VOC uptake on the PDMS/DVB coating at equilibrium, especially for smaller molecules, e.g., benzene. Due to the high affinity to the PDMS/DVB porous polymer coating, water molecules compete with other VOC molecules and occupy a portion of active surface sites on the coating surface. Therefore, fewer active suri:ace sites are available to VOC
15 molecules, especially to smaller molecules with lower affinity to the coating. However, within a very short sampling time, i.e., 1 minute, no significant difference was observed among the conditions with different humidities. This indicates that the active surface sites are not saturated within a very short extraction time, and still available to VOC
molecules. Thus, a short sampling time (far before equilibrium) minimizes the effect of 2o humidity on adsorption of VOC's on the PDMS/DVB coating. Table 4 shows the effect of relative humidity for decreasing VOC uptake onto the YDMS/DVB coating. The humidity effect can be neglected if using PDMS/DVB fiber for a very short time air sampling in a low humidity (< 50%). However, the result suggests that a mass loading decrease of VOC on PDMS/DVB can be expected if the relative humidity is above 50%.
25 Air wind velocity and temperature are important parameters related to the diffusion process on porous polymer SPME fibers, particularly at non-equilibrium conditions.
Wind speed or bulk air movement significantly affects the VOC mass transfer process from the bulk air to the fiber in a certain range. This indicates that the thickness of the gas-phase boundary layer between the fiber and air is diminished as the wind speed 3o increases, and the mass transfer rate was accelerated between 0 and S cm/s.
This wind speed range is typical to average air velocities in indoor air. Therefore, air sampling with porous polymer coated SPME fibers should be conducted above some critical air velocity, above which the mass transfer and the analyte uptake are not affected, and the extraction can be reproduced.
When using a liquid SPME fiber, a decrease in mass loading at equilibrium is usually expected as the extraction temperature increases because the extraction is an exothermic process. An opposite trend of temperature effect was found in this study when mixed-phase porous SPME fibers were used with a short sampling time. Within a very short sampling time far from equilibrium, analytes extracted on a porous polymer coated SPME fiber increased linearly as the extraction temperature increased in a narrow range from 22 to 25°C. In this case, the adsorption process is controlled by diffusion coefficients instead of distribution constants of analytes. The effects of wind and temperature on adsorption by porous polymer coated SPME fibers under non-equilibrium conditions have not been addressed by previous researchers. The analytical data indicate ~ 5 that there is a direct relationship between the rate of mass transfer and the analyte diffusion coefficients. Therefore appropriate diffusion coefficients obtained either from the literature, calculated or experimentally determined can be used to calibrate the relationship between amount of analyte extracted versus concentration for given extraction time. During the experiment constant agitation condition are necessary, or if different agitation conditions are used than the appropriate adjustment coefficients needs to be calculated. Figure 14 shows the example of a simple device 34 based on an electric blower 36 having a fan (not shown) that sucks air into an inlet 38 on a cylindrical head 40 and exhausts the air through an outlet 42. The flower 36 is able to provide constant air agitation. An SPME device 14 has a holder 44 mounted in an SPME insert 46. An O
ring 48 is located between the holder 44 and insert 46. A sleeve 49 extends through the head 40 to position a fiber 50 within the head 40. The blower 36 has a handle 52. The reported data can be extended to liquid sample analysis. In this case, diffusion of analytes in a liquid matrix can be used to calibrate the response. Analogous devices to one showed on Figure 14 for air analysis can be designed to provide constant agitation of 3o the sample matrix. For solid samples, indirect headspace or liquid extraction can be used.

For air analysis, high humidity was found to decrease VOC uptake on the PDMS/DVB coating. However, the humidity effect can be minimized by using a very short exposure time, in which case the active surface sites are not saturated.
The humidity effect can be neglected when using PDMS/UVB fiber for a very short time air-sampling in a low humidity, but a mass loading decrease of VOC's on PDMS/DVB can be expected if air samples are taken from a high humidity environment when the calibration might be necessary. It is expected that both temperature and humidity will be monitored during the field measurement and appropriate correctic>r~ coefficients will be calculated, if necessary to adjust the response.
1o The proposed diffusion based extraction and calibration approach is expected to be the fastest possible sampling/sample preparation approach in the field and in the laboratory. Various different arrangements of the extraction phase can be used to practically implement this technology (see Figures 15A to 1 SE).
In Figure I SA, a vessel 60 containing a sample 62 has an extraction phase coating 63 15 on a tubular member 64. In Figure 1 SB a tube 66 has an extraction phase coating 63 on an inner surface thereof. Sample 62 contacts the extraction phase coating as it flows through the tube 66.
In Figure 15C, the vessel 60 has an extraction phase coating 63 lining an inner surface thereof. The sample 62 contacts the coating 63 when it is contained in the vessel.
20 In Figure 15D, the vessel 60 contains the sample 62. Particles 72 are located within the sample and each particle is surrounded by extraction phase coating 63. In Figure 15E, the vessel 60 has a sample 62 with a stirrer 74 extending into the sample. The stirrer has paddles 76 with extraction phase coating 63 on the paddles. In Figure 15f, there is shown a vessel 60 containing a sample 62 with a stirring bar 78 located within the 25 sample. The stirring bar 78 contains extraction phase coating 63.
In Figure 16, there is shown a schematic perspective view of a silica rod 60 surrounded by a solid extraction phase coating 82 that is porous and contains pores 86. A
cylindrically shaped boundary layer 84 surrounds the cylindrically shaped coating 82 and analytes, designated by Dg pass through the boundary layer and are adsorbed by the 3o coating 86. A graph at the bottom of Figure 16 shows that the boundary layer has a thickness delta, the rod 80 has a radius a and a radius b equals the distance from a center of the rod 80 to the exterior surface of the coating 86. 'The concentration at Co is the concentration at the interface between the boundary layer and the coating and the concentration at Cg is the concentration of the gas.
In Figure 17, a membrane 90 is supported on a handle 92 and the membrane is in an unfolded position. In Figure 18, the membrane 90 is rolled around the handle 92. Figure l7 shows the collection mode where the handle and membrane are brought into contact with a sample (not shown). After rapid extraction has occurred, the membrane is moved out of contact with the sample and the membrane 90 is rolled around the handle 92 to 1o make it more compact. 'The membrane and handle can then be inserted into a cylindrical sheath (not shown), which is airtight. The membrane can then be transferred to an analytical instrument, which can be located in the field where the sample has been taken or to an instrument located away from the test site. The sheath prevents the membrane from becoming contaminated during transport.
In Figure 19, there is shown a further embodiment of the invention where a brush 100 has bristles 102 extending therefrom and each bristle has an extraction phase coating 104 thereon. In Figure 20, a device 107 has a SPME syringe 108 supporting a fibre 109 with a coating 110. A motor l 12 powers a stirrer 114 in a bracket 116.

Claims (5)

1. A device for determining the concentration of analytes of interest in a sample, said device comprising a membrane having a large surface area to adsorb all analytes that contact a surface of said membrane in a time allowed for extraction, said membrane having a handle, said membrane being folded by rolling the membrane around the handle, there being a sheath into which the rolled up membrane can be inserted said membrane being sized and shaped to fit into an injection port of an analytical instrument.

2. A device as claimed in claim 1 wherein the membrane has a rectangular shape with a large surface area relative to the analytes of interest that can be potentially adsorbed during the time of exposure.

3. A device as claimed in claim 1 wherein the sheath can be sealed.

4. A device as claimed in claim 3 wherein the device is portable.

5. A device fin determining the concentration of analytes of interest in a sample, said device comprising a membrane having a large surface area to adsorb all analytes that contact a surface of said membrane in a time allowed for extraction, said membrane being a plurality of fibers connected to a holder, said fibers being extendable and retractable into and out of said holder respectively, said fibers being sized and shaped to fit into an injection port of an analytical instrument.