Desorption of Polycyclic Aromatic Hydrocarbons from Soot Surface

Feb 19, 2010 - ... et Environnement (ICARE), CNRS, 45071 Orléans Cedex 2, France and Ecole des Mines de Douai, ... Telephone: +33 238255474. ... that...
0 downloads 0 Views 310KB Size
Download PDF
J. Phys. Chem. A 2010, 114, 3533–3539

3533

Desorption of Polycyclic Aromatic Hydrocarbons from Soot Surface: Five- and Six-Ring (C22, C24) PAHs Yuri Bedjanian,*,† Mai Lan Nguyen,† and Ange´lique Guilloteau†,‡ Institut de Combustion, Ae´rothermique, Re´actiVite´ et EnVironnement (ICARE), CNRS, 45071 Orle´ans Cedex 2, France and Ecole des Mines de Douai, De´partement Chimie et EnVironnement, BP 838, 59508 Douai, France ReceiVed: December 23, 2009; ReVised Manuscript ReceiVed: February 5, 2010

The kinetics of thermal desorption of five soot-bound nonvolatile (C22, C24) polycyclic aromatic hydrocarbons (benzo(ghi)perylene (BghiP), indeno(1,2,3-cd)pyrene (IdP), anthanthrene (Antha), dibenz(ah)anthracene (DBahA), dibenzo(ae)pyrene (DBaeP)) identified in laboratory-generated kerosene soot were studied over the temperature range 335-370 K in a low pressure flow reactor. The PAH desorption rate constants were measured using off-line HPLC monitoring of their concentrations in soot samples as a function of desorption time. The following Arrhenius expressions were determined for the desorption rate constants of the sootbound PAHs studied: kdes ) 8.4 × 1015 exp[-(129.7 ( 4.4)/RT], 1.0 × 1016 exp[-(130.6 ( 6.1)/RT], 1.1 × 1016 exp[-(131.6 ( 3.2)/RT], 1.0 × 1016 exp[-(128.0 ( 3.9)/RT], and 1.4 × 1016 exp[-(134.2 ( 10.7)/RT] (kdes are in s-1 and activation energies in kJ mol-1) for BghiP, IdP, Antha, DBahA, and DBaeP, respectively. Analysis of the present and previous experimental data showed that PAH-soot binding energies linearly correlate with the number of PAH carbon atoms. The present data and those from previous studies are discussed in the frame of the existing theoretical gas-to-particle partitioning model. 1. Introduction

2. Experimental Section

Polycyclic aromatic hydrocarbons (PAHs), which principally originate from incomplete combustion and pyrolytic processes at high temperatures, are recognized as important pollutants with carcinogenic and mutagenic properties.1 Partitioning of PAHs between particulate and gas phases is a very important factor determining the atmospheric fate of these compounds: transport, reactivity, deposition processes, as well as their health and climate impact and their influence on chemical composition of the atmosphere.1-3 Soot particles being a coproduct of the incomplete combustion of fossil fuels and biomass represent an important source of particulate PAHs. Considering that PAHs have a high affinity to carbonaceous surfaces, adsorption of PAHs onto the soot fraction of atmospheric aerosols may be an important mechanism affecting PAHs partitioning between particulate and gas phases in the atmosphere.4 There were only a few laboratory studies dealing with the distribution of PAHs between gas phase and carbonaceous aerosol.5-7 The present paper is the third one in a series on a systematic study of the kinetics and thermodynamics of a large set of PAHs desorption from kerosene soot surface carried out in our laboratory. In previous papers,8,9 the experimental approach used for the study of PAHs desorption from the laboratory-generated soot samples has been detailed and the results from a kinetic study of the desorption of a set of three- to five-ring PAHs (C14, C20) from soot have been reported. The present paper reports the results of experimental measurements of kinetic and thermodynamic parameters for desorption of heavier, five- and six-ring (C22, C24), soot-bound PAHs.

The experimental approach used in the study of PAHs desorption from the laboratory-generated soot samples, including procedure of soot production, extraction of PAHs from soot samples, and their concentration measurements as well as the employed kinetic method, was detailed in our previous study.8 Therefore, only a brief description of the experimental tools and of the procedure of the measurements will be given here. A flat-flame burner was used for the preparation and deposition of soot samples from premixed flames of the mixture of hydrocarbons (decane: propylbenzene: propylcyclohexane ) 74: 15:11) referred to as kerosene in the paper.8,10 Soot particles from stabilized premixed flames with a richness near 2.0 (the fuel/oxygen ratio multiplied by stoichiometric coefficient of oxygen) were sampled at 4 cm above the burner surface and deposited on the outer surface of a cylindrical Pyrex tube (0.9 cm O.D.) thermostabilized at 45 °C, which was rotated and moved through the flame. This procedure of soot sample preparation was shown to provide soot samples with reproducible (within 15%) concentrations of particulate PAHs.8 Desorption experiments were carried in a flow reactor using a coaxial configuration,8,9 Pyrex tube with deposited soot sample being introduced into the main reactor along its axis. Kinetic measurements consisted in HPLC off-line concentration measurements of PAHs present in solvent extractions of soot samples as a function of a duration of the desorption experiment (soot sample residence time in the reactor). For extraction of PAHs from soot samples we employed the ultrasonic assisted extraction method with acetonitrile as a solvent.8 Analysis of the extraction efficiency has shown that the employed method provided the extraction of at least 90% of extractable concentrations of PAHs.8 Soot solvent extracts were separated from soot particles using Teflon filters with pore diameter of 0.2 µm and analyzed for PAHs content by means

* To whom correspondence should be addressed. Telephone: +33 238255474. Fax: +33 238696004. E-mail: [email protected]. † Institut de Combustion. ‡ Ecole des Mines de Douai.

10.1021/jp912110b  2010 American Chemical Society Published on Web 02/19/2010

3534

J. Phys. Chem. A, Vol. 114, No. 10, 2010

Bedjanian et al.

TABLE 1: List of Kerosene Soot-Bound PAHS Considered in the Present Study PAH

abbreviation

detectora

concentrationb (µg × mg-1)

concentrationc (10-6 mol m-2)

benzo(ghi)perylene indeno(1,2,3-cd)pyrene anthanthrene dibenz(ah)anthracene dibenzo(ae)pyrene

BghiP IdP Antha DBahA DBaeP

F UV F F F

1.83 ( 0.09 0.61 ( 0.03 2.39 ( 0.14 0.49 ( 0.02 1.29 ( 0.01

0.055 0.018 0.072 0.015 0.036

a F, fluorescence detector; UV, multiwavelength UV/Visible detector. b Error represents one standard deviation from mean value from five replicates. c Calculated using specific surface area of 120 m2 g-1.10

(1 h). A similar phenomenon was observed in our previous studies for much longer pumping times (up to 17 h).8,9 In specific experiments focused on this noncomplete PAHs desorption,8 it was shown that PAHs are desorbing homogeneously from the soot surface and the concentrations of the undesorbed PAH molecules are similar throughout the soot sample volume, that is, there is no radial gradient of the concentrations of the undesorbed PAHs (more molecules in underlying soot layers) in the soot sample. The PAHs trapping in the pores of soot samples due to slow diffusion retarded by PAHs readsorption was proposed as the most probable reason for this noncomplete desorption of PAHs. Finally, in the present study as in previous ones, the undesorbed molecules were not considered when determining the desorption rate constants. To determine the firstorder desorption rate coefficients (kdes), experimental plots were fitted with an exponential function approaching the plateau

Figure 1. Kinetics of soot-bound dibenz(ah)anthracene desorption at different temperatures.

of a Jasco high performance liquid chromatograph with multiwavelength Jasco MD-2010 UV/Visible and Jasco FP-2020 fluorescence detectors. Particulate PAHs concentrations were quantified using calibrated solutions of PAHs mixtures. The PAHs studied in the present work are shown in Table 1. 3. Results Fresh soot samples were used in all kinetic experiments. It was first verified8,9 that the PAH surface concentrations did not significantly change (due to desorption or photochemical degradation) during soot sample handling (10 - 20 min) between its preparation and introduction into the flow reactor. The experiments showed no significant changes in PAH concentrations during up to 1 h exposure of the soot samples to air under ambient laboratory conditions (P ) 1 atm, T ) 298 ( 3 K).8,9 Desorption experiments were carried out in a flow reactor at a total pressure of 0.5-1 torr of helium. To minimize the possible impact of the PAHs readsorption to soot surface,8 the measurements of PAH desorption rates were carried out using high linear flow velocities in the reactor (3850-4300 cm s-1) and soot samples with relatively low masses: 0.2-0.5 mg of soot, homogeneously distributed on a few cm length of the support tube. Example of experimental plots of particulate benzo(ah)anthracene desorption measured at different temperatures in the reactor is shown in Figure 1. The desorption kinetics measured at T ) 365 K clearly points to an incomplete PAH desorption: the kinetics is reaching a nonzero plateau, indicating that a part of sorbed PAHs is remaining on the soot surface and is not released into the gas phase on the time scale of the experiments

[PAH] ) [PAH]plateau + ([PAH]0 - [PAH]plateau) × exp(-kdest) (1) where [PAH]0, [PAH], and [PAH]plateau are particulate PAH concentrations at t ) 0, t, and at the plateau region, respectively. The experimental uncertainty on the procedure of the determination of kdes was estimated to be within 15-25%. However, it could be much higher (up to 50%) for the lowest desorption rates measured (corresponding to a few percent decrease of the particulate PAH concentration). Temperature dependence of the desorption rate constant was described by the Arrhenius equation:

kdes ) A exp(-EA /RT)

(2)

where A is the pre-exponential frequency factor, EA is the activation energy for desorption, and R is the molar gas constant. Temperature dependences of kdes observed for different compounds studied are shown in Figures 2 and 3, where continuous lines represent the best Arrhenius expression fit to the experimental data. Resulting Arrhenius parameters determined for all the compounds studied are summarized in Table 2. Kinetic parameters for desorption of other soot-bound PAHs measured in our previous studies8,9 are also given for comparison. The temperature range of the measurements was defined by the condition of the measurability of the PAH desorption rate under experimental conditions used (monitoring time of the desorption kinetics from a few minutes to one hour). One can note high uncertainty of the determination of the pre-exponential factors, which is due to the relatively narrow temperature range used in the measurements. Discussion Pre-Exponential Factors. In many desorption studies the accent is put on the activation energies and the pre-exponential

Desorption of PAHs from Soot Surface

J. Phys. Chem. A, Vol. 114, No. 10, 2010 3535 TABLE 2: Arrhenius Parameters of the Rate Constant for Desorption of PAHs from Soot Surfacea

Figure 2. Temperature dependence of the desorption rate constants of soot-bound dibenz(ah)anthracene and indeno(1,2,3-cd)pyrene.

a Data from the present study (BghiP to DBaeP) and from ref 9, except for Fluoranthene (Fla) and Pyrene (Pyr).8 b Shown in parentheses is the uncertainty factor on the measured values of A. c The quoted error represents 1σ statistical uncertainty.

Figure 3. Temperature dependence of the desorption rate constants of soot-bound benzo(ghi)perylene, anthanthrene, and dibenzo(ae)pyrene.

factors are usually assumed to have a “typical” value of 1013 s-1, although the values of the A-factor can be higher by orders of magnitude.11,12 Fichthorn and Miron11 using molecular dynamics simulations and transition-state-theory calculations of the rate constant for thermal desorption of an n-alkane series from Au (111), have shown that the pre-exponential factors increased with increasing chain length of the alkanes (from 1.5 × 1012 s-1 for methane to 2.8 × 1016 s-1 for dodecane desorption) and for desorption of large molecules can be significantly higher than typical estimates for small molecules. The authors pointed out that the phenomena is general and should be characteristic for the adsorption and desorption of large molecules from solid surfaces. The experimental data from the present study show that this conclusion holds at least for the process of PAHs desorption from soot surface. A trend of increase of the A-factor for larger PAHs molecules is observed: pre-exponential factor (although measured with a high degree of uncertainty) increases from ∼5 × 1014 s-1 for three-ring PAHs to the value of ∼1016 s-1 for the six-ring PAHs studied.

A similar trend of the increase of the pre-exponential factor with adsorbate size was observed by Zacharia et al.12 in their study of PAH desorption from graphite. Incorrect assumption of the relatively low value for the preexponential factor can result in an underestimation of the desorption energies of the surface bound species11 that are often determined using thermal programmed desorption technique (TPD) usually combined with mass spectrometric detection of the desorbing species. Talley et al.,13 measuring the PAHs desorption from mineral and organic surfaces, noted that TPD provides a method to compare the relative strength of PAH binding to different solid surfaces, however it does not allow for quantitative estimates of the desorption energy due to the problem associated with determining an unique set of parameters (pre-exponential factor and desorption energy) in a twoparameter model. They concluded that an independent measurement of the desorption rate at a fixed temperature is required in combination with TPD in order to assess the binding energies for sorbed PAHs by this method. In this respect, the quantitative data obtained in the present study (although with relatively high uncertainty) for the pre-exponential factors of desorption of different PAHs from soot surface seem to be of practical interest, even if the A-factor values can be dependent on the material of the solid surface.

3536

J. Phys. Chem. A, Vol. 114, No. 10, 2010

Bedjanian et al. TABLE 3: Thermodynamic Parameters of Selected PAHs Studied: Present Work and Refs 8 and 9 PAH

∆Hdes (kJ/mol)a

∆Hsub (kJ/mol)b

bPc

log KPd

pLo (Pa)e

Phe Ant Fla Pyr BaA Chr BeP BbF BkF BaP BghiP DBahA IdP Anth

85.6 88.1 93.9 95.2 113.9 114.9 119.9 118.7 120.8 121.8 129.7 128 130.6 131.6

88.9 99.7 98.3 97.9 115.5 118.8 117.9 119.2 130.0 122.5 129.9 134.1 nf f 135

-18.32 -18.76 -18.35 -18.52 -19.32 -19.50 -19.30 -19.36 -19.33 -19.38 -19.74 -19.82 -19.82 -19.84

-3.53 -3.53 -2.10 -2.05 0.43 0.43 1.50 1.23 1.63 1.75 2.78 2.40 2.86 3.01

8.20 × 10-2 7.64 × 10-2 6.72 × 10-3 4.16 × 10-3 1.86 × 10-4 1.70 × 10-4 7.90 × 10-6 1.28 × 10-5 1.26 × 10-5 7.00 × 10-6 7.78 × 10-7 1.05 × 10-6 1.18 × 10-6 5.79 × 10-7

This study and refs 8 and 9: ∆Hdes ) EA. b Nass et al.,28 except for benzo(k)fluoranthene29 and anthanthrene.30 c Calculated for T ) 290 K using eq 8. d Calculated for T ) 298 K using eqs 6-8. e Yamasaki et al.,31 calibrated by Lei et al.32 (from phenanthrene to benzo(a)pyrene) and Offenberg et al.33 (for last four PAHs). f nf, data not found in literature. a

Figure 4. Dependence of the activation energy for desorption of PAHs from carbonaceous surfaces on the number of PAH carbon atoms: 4, Steiner and Burtscher;14 ], Zacharia et al.;12 0, Aubin and Abbatt;7 O, this work and refs 8 and 9. The uncertainties shown are those given in references.

Desorption Energies. As one would expect, a trend in the ordering of PAHs with respect to the desorption thermodynamics is observed (Table 2): desorption energies (which are equivalent to the binding energies for a nonactivated desorption) are very close for PAHs with the same number of carbon atoms and increase with increasing number of PAH carbon atoms. These results are in line with available literature data on PAHs desorption from carbonaceous materials7,12,14 and are consistent with the additivity of the van der Waals forces that dominate the PAHs-soot interaction. Figure 4 presents the activation energies for PAHs desorption measured in the present study and previous studies from our laboratory8,9 and those from literature7,12,14 as a function of the number of carbon atoms in corresponding PAHs. One can note a very good linear correlation between the activation energies for PAH desorption from carbonaceous surface and the number of carbon atoms in PAH (similar correlation holds when PAH molecule mass is used instead of the number of carbon atoms). The data from Zacharia et al.12 correspond to the interaction of benzene and three PAHs molecules with the basal plane of graphite studied using thermal desorption (TD) - mass spectrometry method. The desorption activation energies of 48.2, 82.0, 135.1, and 202.6 kJ mol-1 were found for benzene, naphthalene, coronene, and ovalene, respectively. These values represent the average of the data obtained with two approaches to analyze the TD traces: using Redhead’s peak maximum method15 and an isothermal analysis introduced by Falconer and Madix.16 It can be noted that the activation energies obtained in the study of Zacharia et al.12 for desorption of benzene and naphthalene might be somewhat overestimated, since relatively high values of the pre-exponential factors (compared with the trend observed in the present study, Table 2) were estimated for these relatively small molecules and used in the calculations. Aubin and Abbatt7 measured the adsorption isotherms of benzene, ethylbenzene, naphthalene, acenaphtylene, and acenaphthene on n-hexane combustion soot. The adsorption enthalpies of 37.5 and 55.9 kJ mol-1 were determined for benzene and ethylbenzene, respectively, from the van’t Hoff equation using the equilibrium constants obtained from the Langmuir type isotherms observed for these species. For three PAHs (naphtalene, acenaphtylene, and acenaphthene)

the BET isotherm was used, and the experimentally determined parameter was the difference between desorption and vaporization enthalpies, ∆Hdes - ∆Hvap, which is related to the BET constant, CBET, as follows17

CBET ≈ exp((∆Hdes-∆Hvap)/RT)

(3)

Consequently, the values of ∆Hdes for naphtalene, acenaphtylene, and acenaphthene (88.2, 89.9, and 97.5 kJ mol-1, respectively) were determined using experimental data for ∆Hdes - ∆Hvap and the values of ∆Hvap known from literature. These three values fairly deviate from the linear correlation shown in Figure 4. However, it should be noted that the analysis of the values of ∆Hvap used by Aubin and Abbatt7 in their calculations shows that, probably, the literature values of ∆Hsub were used instead. The use of the true values of ∆Hvap would lead to a decrease of the reported values of ∆Hdes by 15-20 kJ mol-1, resulting in their much better agreement with the correlation line in Figure 4. The experimental data for the desorption activation energies can be compared with the sublimation enthalpies of corresponding species (Table 3), which are relevant to binding energy for desorption from PAH multilayer coverage or when the activation energies for desorption from monolayer and multilayer coverage are similar.18 Figure 5 shows the activation energies for PAHs desorption measured in the present study as well as in refs, 8, 9, 12, and 14, plotted against the sublimation enthalpies of the corresponding compounds. The straight line in Figure 5 represents a zero-intercept linear fit to the data from our laboratory (this study and refs 8 and 9) and corresponds to the equation EA ) 0.97 × ∆Hsub (with the correlation coefficients R2 ) 0.99). The data from refs 12 and 14 (not considered in the linear regression) measured on different carbonaceous materials and with different methods fall on the straight line, supporting the similarity of EA and ∆Hsub values for the processes of PAHs desorption from carbon-like materials. Similarity of the activation energies for PAHs desorption from soot surface with the corresponding sublimation enthalpies seem to be consistent with the structural resemblance between planar PAH molecules and graphitic sheets of soot. Ulbricht et al.,18

Desorption of PAHs from Soot Surface

J. Phys. Chem. A, Vol. 114, No. 10, 2010 3537

KP )

(F/TSP) A

(4)

where A and F (in ng m-3) are equilibrium concentrations in the gas and particulate phases, respectively, and TSP (in µg m-3) is the total suspended particulate matter. Gas-solid adsorptive partitioning theory predicts that:21

KP )

Figure 5. Activation energy for desorption of PAHs from carbonaceous surfaces as a function of the sublimation enthalpy of corresponding compounds: +, Steiner and Burtscher;14 ], Zacharia et al.;12 O, this work and refs 8 and 9. The uncertainties shown are those given by the authors.

comparing their data for activation energies of monolayer and multilayer desorption of PAHs, have noted that as the size of the PAH increased, its interaction with graphite surface became less and less distinguishable from the corresponding bulk behavior. This feature was attributed to the increasing similarity of the substrate and adsorbate’s chemical and physical properties for larger polyaromatic compounds. Paserba and Gellman,19 measuring the desorption kinetics of a set of n-alkanes (CnH2n+2, 5 < n < 60) from the surface of single crystalline graphite have found that ∆Hdes had a nonlinear dependence on chain length and took the empirical form ∆Hdes ) a + bn0.5. Fichthorn and Miron11 have shown that the binding of a large molecule to a solid surface can involve many local minima. They explained the less-than-linear increase in the desorption energy of alkanes by increase of the number of these minima, as the chain length increases, which makes them increasingly accessible at desorption temperatures. The linear relationship between the experimentally determined activation energy EA and the number of carbon atoms of the PAH displayed in Figure 4 seems to reveal a much more narrowly distributed ground state structure of a PAH adsorbed on kerosene soot than in the linear alkane chain where many more molecular configurations prevail in the adsorbed state. The quantitative experimental data on the desorption of PAHs from solids is rather scarce. In this respect, the linear correlations between desorption activation energy and number of PAH carbon atoms or corresponding enthalpy of sublimation observed in the present study seem to be of practical interest, since they can serve as a first approximation for the prediction of the desorption energy of other (not studied) PAHs from carbonaceous surfaces. Atmospheric Implications. The experimental data obtained in the present study can be directly applied to the calculations of the PAHs partitioning in the atmosphere using a theoretical partitioning framework developed in previous studies. The distribution of PAHs between the two phases depends on the physicochemical properties of PAH (saturation vapor pressure), those of the particulate phase (surface area, composition), on the environmental conditions (temperature), and is usually described by a partition coefficient KP (m3 µg-1):20-22

NSaTSPTe(∆Hdes-∆Hvap)/RT 16P°L

(5)

where NS is the concentration of surface adsorption sites (mol cm-2); aTSP is the specific surface area of the particulate matter (cm2 µg-1); ∆Hdes and ∆Hvap are the enthalpies (kJ mol-1) of desorption and vaporization, respectively; R is the molar gas constant; T is the temperature (K); and P°L is the subcooled liquid vapor pressure (torr). The experimental data on ∆Hdes from the present and previous studies obtained for PAHs desorption from carbonaceous particles (which seem to dominate gas-particle partitioning of PAHs,4 at least at polluted urban areas) can be used for direct calculations of KP via eq 1 for given atmospheric aerosol loading (aTSP) and using tabulated physical parameters (∆Hvap, P°) L of the species of interest. Laboratory data for PAHs desorption can be confronted to some parameters derived from the field data on PAHs partitioning as well as to theoretical predictions. For instance, for a variety of PAHs sampled in Osaka, Yamasaki et al.20 have observed a linear dependence of log KP on 1/T:

log KP ) mP /T + bP

(6)

Pankow21 has shown that for adsorptive partitioning:

mP ) ∆Hdes /2.303R-T/4.606 bP ) log

aTSPt0 275(M/T)1/2

+ 1/4.606

(7)

(8)

where t0(s) is a molecular vibration time, and M is the molecular weight of the compound of interest. The y-intercept bP is supposed to be similar for similar compounds. Pankow22 applied the linear regression (6) to the field data of Yamasaki et al.20 and determined the values of mP and bP for each of the compounds or groups of compounds studied by Yamasaki et al.20 Resulting values of bP were spread between -15.44 and -21.83 and the mean bP value of -18.48 was chosen as a common (for the PAHs) y-intercept. Using eq 8 we have calculated the bP values for the compounds considered in the present study (with aTSP ) 0.025 cm2 µg-1,21 T ) 290 K and using experimentally determined 1/A instead of t0). Results of these calculations are presented in Table 3. The calculated values of bP are decreasing from -18.32 for phenanthrene to -19.84 for anthanthrene, with the mean value being -19.29. This trend is due to a decrease of t0 ) 1/A for higher PAHs molecules (see discussion on the pre-exponential factor). Let us note that in earlier studies the value of t0 ) 10-13 to 10-12 s was supposed to be similar for all compounds,21,22 although Pankow21 noted the possible considerable uncertainty concerning an accurate, representative value for t0 for atmospheric organic compounds. As one can note, the results obtained in our laboratory study

3538

J. Phys. Chem. A, Vol. 114, No. 10, 2010

Bedjanian et al.

are very similar to those coming from analysis of the field measurements data. The mean value of bP of three- to five-ring PAHs (Table 3) is -19.01, which is very close to -18.48 derived for these compounds in ref 19. Concerning this comparison, it is important to note that calculated values of bP depend (although not strongly, since ∆bP ) ∆log(aTSP)) on the choice of aTSP. Using eq 7 and the values of mP determined from the linear regression to the field data, Pankow22 calculated the following 95% confidence intervals for the enthalpies of desorption of the various compounds from particulate matter (in kJ mol-1): 60.6-98.6, phenanthrene + anthracene; 70-101, fluoranthene; 64.3-97.7, pyrene; 96.0-128.7, chrysene + benzo(a)anthracene; 86.6-133.2, benzo(b)fluoranthene + benzo(k)fluoranthene; 66.6-122.3, benzo(a)pyrene + benzo(e)pyrene. One can note that laboratory ∆Hdes data reported in Table 3 for corresponding compounds are consistent with those from field measurements, even if particulate matter in the field study certainly differed from the laboratory-generated soot. Assuming the value of ∆Hdes - ∆Hvap to be constant within a homologous series of molecules, eq 5 leads to a linear relationship between log KP and log P°: L

log KP ) m log P°L + b

(9)

which was widely used to interpret the PAHs partitioning observed under variety of environmental conditions (for examples, see refs 23-26). The assumption of a constant value of ∆Hdes - ∆Hvap for PAH molecules, initially based on the analysis of field data of Yamasaki et al.20 by Pankow,21 was verified in our previous paper.8 Examination of the ∆Hdes ∆Hvap data for 10 three- to five-ring PAHs (ranging from 8.5 to 15.2 kJ mol-1 with mean value of (10.9 ( 2.5) kJ mol-1) supported (at least for PAHs molecules) the widely used assumption that the value of ∆Hdes - ∆Hvap can be considered as constant within a homologous series of molecules. Unfortunately, the values of ∆Hdes - ∆Hvap for the compounds studied in the present work could not been examined, since we have not found in the literature the corresponding values of the enthalpy of vaporization, ∆Hvap. In our PAHs desorption studies, we have not determined the values of the partition constants KP. However these values can be calculated using eqs 6-8 with the measured ∆Hdes data. The values of KP at T ) 298 K calculated in this way are plotted in Figure 6 against the log of subcooled liquid vapor pressure of the corresponding compounds. Linear regression between these two quantities is observed in agreement with theoretical prediction (eq 9). The value of -1.26 obtained for m from the slope of the regression line differs from that predicted by theory, m ) -1, however it agrees with the values of m derived from field measurements data (up to -1.49).26 Goss27 pointed out that, due to thermodynamic reasons, the slope of plots of log KP vs log P°L may differ significantly from unity depending on the sorption medium. Considering (i) the high uncertainties on the subcooled liquid vapor pressure data and that (ii) partitioning constants were not measured but were calculated in the present study, the agreements between theoretical predictions and linear regression in Figure 6 can be considered as very satisfactory. The reported kinetic and thermodynamic data obtained for desorption of a number of PAHs from laboratory-generated soot surface seem to represent a significant and useful support for the theoretical and field studies of the PAHs distribution

Figure 6. Plot of the calculated (see text) values of log KP vs log P°. L

in the atmosphere. However, a better understanding of the PAHs partitioning requires similar precise laboratory data for other solid supports of atmospheric relevance. Acknowledgment. This study has been carried out within the program PRIMEQUAL 2 funded by French Ministry of Ecology and Sustainable Development. A.G. is very grateful to Ecole des Mines de Douai and European Structural Funds for cofinancing her Ph.D. grant. References and Notes (1) Finlayson-Pitts, B. J.; Pitts, J. N. J. Chemistry of the Upper and Lower Atmosphere: Theory, Experiments and Applications; Academic Press: San Diego, 2000. (2) Calvert, J. G.; Atkinson, R.; Becker, K. H.; Kamens, R. M.; Seinfeld, J. H.; Wallington, T. J.; Yarwood, G. The Mechanisms of Atmospheric Oxidation of Aromatic Hydrocarbons.; Oxford University Press: New York, 2002. (3) Bidleman, T. F. EnViron. Sci. Technol. 1988, 22, 361. (4) Dachs, J.; Eisenreich, S. J. EnViron. Sci. Technol. 2000, 34, 3690. (5) Niessner, R.; Wilbring, P. Anal. Chem. 1989, 61, 708. (6) Hueglin, C.; Paul, J.; Scherrer, L.; Siegmann, K. J. Phys. Chem. B 1997, 101, 9335. (7) Aubin, D. G.; Abbatt, J. P. EnViron. Sci. Technol. 2006, 40, 179. (8) Guilloteau, A.; Nguyen, M. L.; Bedjanian, Y.; Le Bras, G. J. Phys. Chem. A 2008, 112, 10552. (9) Guilloteau, A.; Bedjanian, Y.; Nguyen, M. L.; Tomas, A. J. Phys. Chem. A 2010, 114, 942. (10) Lelie`vre, S.; Bedjanian, Y.; Pouvesle, N.; Delfau, J. L.; Vovelle, C.; Le Bras, G. Phys. Chem. Chem. Phys. 2004, 6, 1181. (11) Fichthorn, K. A.; Miron, R. A. Phys. ReV. Lett. 2002, 89, 196103. (12) Zacharia, R.; Ulbricht, H.; Hertel, T. Phys. ReV. B 2004, 69, 155406. (13) Talley, J. W.; Ghosh, U.; Furey, J. S.; Tucker, S. G.; Luthy, R. G. EnViron. Eng. Sci. 2004, 21, 647. (14) Steiner, D.; Burtscher, H. K. EnViron. Sci. Technol. 1994, 28, 1254. (15) Readhead, A. Vacuum 1962, 12, 203. (16) Falconer, J. L.; Madix, R. J. J. Catal. 1977, 4, 8–262. (17) Brunauer, S.; Emmett, P. H.; Teller, E. J. Am. Chem. Soc. 1938, 60, 309. (18) Ulbricht, H.; Zacharia, R.; Cindir, N.; Hertel, T. Carbon 2006, 44, 2931. (19) Paserba, K.; Gellman, A. Phys. J. Chem. Phys. 2001, 115, 6737. (20) Yamasaki, H.; Kuwata, K.; Miyamoto, H. EnViron. Sci. Technol. 1982, 16, 189. (21) Pankow, J. F. Atmos. EnViron. (1967) 1987, 21, 2275. (22) Pankow, J. F. Atmos. EnViron. Part A. General Topics 1991, 25, 2229. (23) Gustafson, K. E.; Dickhut, R. M. EnViron. Sci. Technol. 1997, 31, 140.

Desorption of PAHs from Soot Surface (24) Simcik, M. F.; Franz, T. P.; Zhang, H.; Eisenreich, S. J. EnViron. Sci. Technol. 1998, 32, 251. (25) Ngabe, B.; Poissant, L. EnViron. Sci. Technol. 2003, 37, 2094. (26) Sitaras, I. E.; Bakeas, E. B.; Siskos, P. A. Sci. Total EnViron. 2004, 327, 249. (27) Goss, K.-U. EnViron. Sci. Technol. 1997, 31, 3736. (28) Nass, K.; Lenoir, D.; Kettrup, A. Angew. Chem., Int. Ed. Engl. 1995, 34, 1735. (29) Shiu, W.-Y.; Ma, K.-C. J. Phys. Chem. Ref. Data 2000, 29, 41.

J. Phys. Chem. A, Vol. 114, No. 10, 2010 3539 (30) Inokuchi, H.; Shiba, S.; Handa, T.; Akamatu, H. Bull. Chem. Soc. Jpn. 1952, 25, 299. (31) Yamasaki, H.; Kuwata, K.; Kuge, Y. Nippon Kagaku Kaishi 1984, 1324. (32) Lei, Y. D.; Chankalal, R.; Chan, A.; Wania, F. J. Chem. Eng. Data 2002, 47, 801. (33) Offenberg, J. H.; Baker, J. E. EnViron. Sci. Technol. 1999, 33, 3324.

JP912110B