Anal. Chem. 2007, 79, 4222-4227
Systematic Investigation of the Sorption Properties of Polyurethane Foams for Organic Vapors Ireen Kamprad and Kai-Uwe Goss*
Department of Biogeochemistry and Pollutant Dynamics, ETH Zurich, Universita¨tsstrasse 16, 8092 Zurich, Switzerland
Polyurethane foams (PUFs) have long been used for the sampling of airborne organic pollutants. However, a comprehensive investigation of their sorption properties for various organic compounds is still lacking. Here we present a systematic sorption study for a diverse set of up to 100 compounds at temperatures between 15 and 95 °C, at various humidities, and for various types of PUFs. The results can be summarized as follows: The performance of PU-ethers is not susceptible to the humidity of the sampled air. Variability between different types of PU-ethers is small, and the compound variability in the sorption constants as well as the enthalpies of sorption can accurately be described with linear free energy relationships (LFERs) presented in the text. PU-esters revealed an inferior sorbent behavior as compared to PUethers: sorption was kinetically hindered and was affected by humidity. Sorption varied between different types of esters and prediction with LFERs was less accurate than for the ethers. Open-cell polyurethane foams (PUFs) are among the most frequently used sorbents for actively or passively sampling airborne organic compounds. Passive samplers require a constant diffusive flux from air to the sorbent over the sampling period. This implies a constant concentration gradient, a condition that can only be achieved if the amount of PUF is chosen such that at the end of the sampling period the PUF is still far from sorption equilibrium with the air. Such a decision requires knowledge of the sorption constant of the target compound on the used PUF. In active air sampling, the amount of PUF that is employed must be sufficiently large that no breakthrough of the target compound can occur at a given sampling volume. By definition the sorption coefficient of a compound equals the sampling volume per weight of sorbent at which the breakthrough is 50%. Typically, the sampling efficiency will be g90% if the sampling volume is smaller than half the 50% breakthrough volume.1 Hence, for active and passive sampling it is essential to know the sorption coefficient, Ki, of the target compound i on the PUF at ambient temperature in order to chose the right amount of sorbent: * Author to whom correspondence should be addressed. Current address: UFZ Helmholtz Centre for Environmental Research, Permoserstr. 15, 04318 Leipzig, Germany. Phone: ++49 341 235 2408. Fax: ++49 341 235 2625. E-mail:
[email protected]. (1) Pankow, J. F. Atmos. Environ. 1989, 23, 1107-1111.
4222 Analytical Chemistry, Vol. 79, No. 11, June 1, 2007
Ki )
equilibrium concn of i on the PUF (mol/g) (1) equilibrium concn of i in the air (mol/L)
Currently, the sorption constants relevant for active or passive air sampling must be determined experimentally for every new analyte. Only if these data are already known for other, similar compounds can Ki be estimated by using a log-linear correlation of the form:
log Ki ) a log pi,L + c
(2)
where pi,L is the saturated liquid vapor pressure of compound i, and a and c are coefficients to be determined by a regression analysis from existing experimental data.1,2 The use of such equations is very limited because separate regressions have to be set up for every compound class.3 Furthermore, information on the variability of the sorption capacity of various types of PUF and on temperature and humidity dependence of sorption to PUF is not yet available in the literature. Here we set out to systematically study all factors that may affect sorption to PUFs so that eventually we could come up with a reliable, predictive tool for all required sorption coefficients. To this end we have carried out sorption experiments for different temperatures, humidities, types of PUF, and flow velocities of air and for a wide variety of polar and nonpolar organic compounds. The experimental sorption coefficients of a set of various analytes at a given set of conditions (temperature, humidity, type of PUF) were evaluated by a polyparameter linear free energy relationship (pp-LFER) of the following form:
log Ki ) lPUFLi 16 + sPUFSi + aPUFAi + bPUFBi + vPUFVi + cPUF (3) The capital letters are compound descriptors for the various types of interactions a compound can undergo. Li16 is the logarithm of the hexadecane/air partition constant at 25 °C in units of (mair3/ mhexadecane3), and Vi is the McGowan characteristic volume in units of (cm3/mol)/100. These descriptors are needed to describe nonspecific interactions (van der Waals interactions and cavity formation) between the analyte and the sorbent. The remaining (2) You, F.; Bidleman, T. F. Environ. Sci. Technol. 1984, 18, 330-333. (3) Goss, K.-U.; Schwarzenbach, R. P. Environ. Sci. Technol. 2001, 35, 1-9. 10.1021/ac070265x CCC: $37.00
© 2007 American Chemical Society Published on Web 04/26/2007
Table 1. Summary of the Used Polyurethane Foams and Their Properties name
supplier
bulk density [kg/m3]
compression force deflection [kPa]
porosity
pore size23
PU-ether LM 20 33 THE 25 19 TEX 40 30
Foampartnera Foampartnera Foampartnera
20 25 40
33 19 30
0.98 0.98 0.96
35-50 35-50 35-50
PU-ester PU Ester R Ester CH R 200 U P 100 Z Willsorp
Foampartnera Foampartnera Illbruckb Illbruckb Illbruckb
0.97 0.97 0.97
80 80 70 ( 10 100 85
30 30 29
remark
reticulated reticulated
a Foampartner, Fritz Nauer AG, Oberwolfhauserstr. 3, CH-8633 Wolfhausen, Switzerland. b Illbruck Foamtec, 3800 Washington Ave. N., Minneapolis, MN 55412.
three descriptors stand for various specific intermolecular interactions: Ai characterizes the H-donor (or electron-acceptor) property of the compound, Bi stands for the H-acceptor (or electron-donor) property, and Si is called the dipolarity/polarizability descriptor. The small letters represent the complementary system descriptors which here stand for the sorption properties of the used PUF. pp-LFERs such as the one in eq 3 have been shown to successfully describe the variance of sorption coefficients for very diverse compounds in all kinds of partition media (e.g., solvents,4 polymers,5 surfaces,6 humic material7). The compound descriptors required in eq 3 are tabulated for a large number of compounds in the literature.8-10 The system descriptors are typically determined from regressing eq 3 to a diverse set of experimentally determined sorption coefficients for compounds with known compound descriptors. Equation 3 is a modification of the wellknown Abraham equation where an E-descriptor (for dispersive interactions) is used instead of the V-descriptor in eq 3. Equation 3 is as powerful as the original Abraham equation in fitting various partition data but has the advantage that it can be used within a thermodynamic cycle.16 EXPERIMENTAL SECTION Method. The sorption constants were determined by dynamic sorption experiments where the PUFs served as stationary phases in a gas chromatographic system. When a substance i is injected onto such a PUF-column under isothermal conditions, the retention observed is a measure of its sorption intensity. This method directly allows to study the dependence of sorption on flow rate (4) Acree, W. E.; Abraham, M. H. Fluid Phase Equilib. 2002, 201, 245-258. (5) Braun, J.-M.; Guillet, J. E. Adv. Polym. Sci. 1976, 21, 107-145. (6) Goss, K.-U. Crit. Rev. Environ. Sci. Technol. 2004, 34, 339-389. (7) Niederer, C.; Goss, K.-U.; Schwarzenbach, R. P. Environ. Sci Technol. 2006, 5374-5379. (8) Abraham, M. H.; Andonian-Haftvan, J.; Whiting, G. S.; Leo, A.; Taft, R. S. J. Chem. Soc., Perkin Trans. 2 1994, 2, 1777-1791. (9) Abraham, M. H.; Chadha, H. S.; Whiting, G. S.; Mitchell, R. C. J. Pharm. Sci. 1994, 83, 1085-1100. (10) Abraham, M. H. J. Chromatogr. 1993, 644, 95-139. (11) Conder, J. R.; Young, C. L. Physicochemical Measurement by Gas Chromatography; Wiley: New York, 1979. (12) Gray, D. G. Prog. Polym. Sci. 1977, 5, 1-60. (13) Papirer, E.; Balard, H.; Rahmani, Y. Chromatographia 1987, 23, 639-647. (14) Du, Q.; Hattam, P.; Munk, P. J. Chem. Eng. Data 1990, 35, 367-371. (15) Hierlemann, A.; Ricco, A. J.; Bodenho¨fer, K.; Go¨pel, W. Anal. Chem. 1999, 71, 3022-3035. (16) Goss, K.-U. Fluid Phase Equilib. 2005, 233, 19-22.
which is an important information for active air sampling. For passive air sampling, equilibrium sorption constants are the relevant entity. These can also be derived from the observed retention provided that the carrier gas velocity is small enough to allow for the compounds to completely penetrate the cell walls of the PUF. In this case, equilibrium sorption coefficients can be derived from the observed retention volume Vi according to eq 4:11
Ki ) (Vi - Vtracer)/MPUF
(4)
where, Vtracer is the elution volume of a non-retained tracer and MPUF is the amount of PUF in the column. The volumes Vi and Vtracer are determined from the retention time of the respective peaks which are marked by their first statistical moment and the volumetric flow rate of the mobile phase corrected for the pressure drop.11 Measurements without the PUF revealed that sorption to the stainless steel system (capillaries, frits, and column walls) was negligible. Such dynamic sorption experiments have been routinely used to determine adsorption of organic compounds from the gas phase on surfaces or absorption into thin organic films or polymers in various scientific disciplines.12-15 The experiments were conducted with stainless steel highpressure liquid chromatography columns (1.0 cm length, 0.3 cm inner diameter) filled with cylinders of the respective PUF. The column was connected to the injector and flame ionization detector of a chromatograph (Carlo Erba 5300, Milan, Italy) with short stainless steel capillary tubings (0.5 mm i.d.). Nitrogen with a relative humidity
