Fluorescence Spectroscopy of Complex Aromatic Mixtures - Analytical

Structural Characterization of Large Polycyclic Aromatic Hydrocarbons. Part 2: ... Analytical techniques for high-mass materials. Rafael Kandiyoti , A...
0 downloads 17 Views 83KB Size
Anal. Chem. 2004, 76, 2138-2143

Fluorescence Spectroscopy of Complex Aromatic Mixtures Barbara Apicella,* Anna Ciajolo, and Antonio Tregrossi

Istituto di Ricerche sulla Combustione - CNR, P. le Tecchio, 80-80125 Napoli, Italy

The contribution of two- to seven-ring polycyclic aromatic hydrocarbons (PAH) and of larger aromatic structures contained in complex PAH-laden mixtures collected in flames was evaluated by fluorescence spectroscopy. A composition procedure of the fluorescence spectra of individual PAHs, analyzed by gas chromatography/mass spectrometry (GC/MS) was applied for the evaluation of their contribution to the fluorescence spectra of PAHladen mixtures. In this way, it was possible to put in evidence the contribution to the total fluorescence spectrum of high molecular weight aromatic species present in the PAH-laden mixtures and not detectable by GC/MS. Qualitative and quantitative interpretation of synchronous and conventional fluorescence spectra of PAH-laden mixtures formed in combustion processes was proposed. The composition procedure was showed to be reliable in the UV-visible region for samples dissolved in cyclohexane solutions, but failed in the UV region when the solvent contained heavy atoms, as in the case of dichloromethane. However, the heavy-atom solvent effect was not sufficient to explain the depression of the UV fluorescence signal. Energy transfer interaction between fluorene and other fluorescing PAHs was suggested to be also responsible for this effect on the basis of fluorescence studies performed on single PAHs and their mixtures in cyclohexane, methanol, and dichloromethane. Fluorescence spectroscopy is a useful diagnostic tool to evaluate polycyclic aromatic hydrocarbons (PAH) in complex PAH-laden mixtures as separated by high-pressure liquid chromatography or even, without separation, by means of direct selective methods, such as selective modulation and synchronous fluorescence.1 In conventional fluorometry, the fluorescence intensity is recorded as a function of the emission wavelength while the excitation wavelength is held at a constant value. Excitation fluorometry involves the reverse situation that is the variation of the excitation wavelength at a constant emission wavelength.2 The synchronous excitation fluorometry gives a third possibility whereby neither excitation or emission wavelength is kept constant, with the constraint that the difference between the * Corresponding author. Phone: +39/0817682254. Fax: +39/0815936936. E-mail: [email protected]. (1) Lee, M. L.; Novotny, M. V.; Bartle, K. D. Analytical Chemistry of Polycyclic Aromatic Compounds; Academic Press: New York, 1981. (2) Lloyd, J. B. F. Nat. Phys. Sci. 1971, 231, 64-65.

2138 Analytical Chemistry, Vol. 76, No. 7, April 1, 2004

excitation and emission wavelengths or frequencies remains constant throughout the spectrum. Synchronous fluorescence introduced by Lloyd2 has been used to characterize complex mixtures3,4 providing fingerprints of complex samples, such as crude oils of different origins5,6 and extracts and pyrolysis tars of coals.7,8 A methodology for the application of synchronous fluorescence has been further developed and applied to a synthetic mixture of PAHs.9 However, some limitations of the technique have to be taken into account for when using synchronous fluorescence for the characterization of complex PAH mixtures. Energy transfer processes have been considered responsible for the depressive interference on the fluorene peak found in the synchronous fluorescence of mixtures of PAHs dissolved in cyclohexane by Latz et al.,10 but the authors underlined that, even though these effects occur also in conventional fluorometry, in this case, they stand out and, therefore, can be accounted for by variation in experimental conditions, separation procedures, etc. They attributed the decrease of fluorene synchronous signal to the presence of pyrene; however, their results have been criticized by Lloyd, who has ascribed their findings to the use of very concentrated solutions leading to “a massive and unacceptable inner filter effects”.11 Nevertheless, the observations of Latz and his group have been confirmed by Katoh et al.8 in the practical application of synchronous fluorescence to the analysis of a coalderived liquid dissolved in methanol. The effect of heavy-atom solvents on synchronous fluorescence spectra of complex mixtures has to be considered important in qualitative and quantitative analysis, as noted by Lloyd,3 since it causes a quenching of some signals and the appearance of new emissions. The assessment of the occurrence of these effects is generally important for a correct interpretation of fluorescence PAH spectra and, particularly, when fluorescence spectroscopy is used for the characterization of very complex aromatic mixtures not easily analyzable with conventional chromatographic techniques. In this work, these effects have been studied for a correct interpretation (3) (4) (5) (6) (7) (8) (9) (10) (11)

Lloyd, J. B. F. Analyst 1974, 99, 729-738. John, P.; Soutar, I. Anal. Chem. 1976, 48, 520-524. Lloyd, J. B. F. J. Forensic Sci. Soc. 1971, 11, 153-157. Clark, E. R.; Darwent, J. R.; Demirci, B.; Flunder, K.; Gaines, A. F.; Jones, A. C. Energy Fuels 1987, 1, 392-397. Li, C.; Wu, F.; Cai, H.; Kandiyoti, R. Energy Fuels 1994, 8, 1039-1048. Katoh, T.; Yokohama, S.; Sanada, Y. Fuel 1980, 59, 845-850. Vo-Dinh, T. Anal. Chem. 1978, 50, 396-401. Latz, H. W.; Ullman, A. H.; Winefordner, J. D. Anal. Chem. 1978, 50, 21482149. Lloyd, J. B. F. Anal. Chem. 1980, 52, 189-191. 10.1021/ac034860k CCC: $27.50

© 2004 American Chemical Society Published on Web 03/06/2004

of synchronous and conventional spectra of complex PAH-laden mixtures, such as flame-formed samples derived from fuel-rich combustion processes. The contribution of the fluorescence of two- to seven-ring PAHs to the fluorescence spectra of flame-sampled PAH mixtures has been evaluated by means of a composition procedure which requires the measurement of the single PAH content in the sample, obtained by gas chromatography/mass spectrometry (GC/MS), and the measurement of the fluorescence of the single PAH. The procedure was set up and tested for mixtures of standard PAHs in dichloromethane solution and for an aromatic oil whose chemical composition has been completely determined by GC/MS. The reduction of the fluorescence signal in the UV wavelength region typical of fluorene has been observed. The effects of the interaction between fluorene and the other PAHs and of a heavy-atom solvent effect on the fluorescence behavior have been studied by systematic measurements of synchronous and fixed excitation wavelength (λexc ) 266 nm) spectra of standard PAH mixtures, neat fluorene, an aromatic oil (anthracene oil), and flame-formed PAH-laden mixtures in dichloromethane and cyclohexane. EXPERIMENTAL SECTION Fluorescence Measurements. Emission and synchronous fluorescence spectra of samples were measured by a Perkin-Elmer PE LS-50 B luminescence spectrometer. Fluorescence measurements were carried out by diluting the sample solutions with suitable solvent amounts such that the total concentration of PAH did not exceed 0.01 mg/L. Fluorescence spectra of diluted dichloromethane solutions (0.1-0.01 mg/L) of the most abundant and fluorescent PAHs identified in the flame samples have been also measured in order to evaluate the contribution of each of them to the total fluorescence of the flame samples and, by summing these contributions, to obtain a composite spectrum that defines the contribution of all GC-PAHs to the total fluorescence. The procedure for determining the contribution of PAHs to the fluorescence spectra of the samples has been applied in a previous work12 to reconstruct absorption spectra of PAH standard mixtures and PAH-rich flame-formed samples. The fluorescence contribution of identified PAHs has been computed with the following expression in matrix/vector notation

ICf ) yPAHIPAH XW f

where ICf is the nλ × 1 vector of the computed fluorescence intensity (nλ ) number of wavelengths in the 260-500 nm range), yPAH is the weight fraction of total identified PAH in the PAH is the nλ × npah matrix of the PAH fluorescence, XW sample, IPAH f is the npah × 1 vector of weight fraction of each PAH in total identified PAH, and npah is the number of PAHs. PAH Analysis. Polycyclic aromatic hydrocarbons from naphthalene (128 Da) up to coronene (300 Da), collectively called GCPAHs, were quantitatively analyzed in the PAH-laden mixture collected in flame and in an anthracene oil sample by GC/MS on (12) Tregrossi, A.; Ciajolo, A.; Barbella, R. Proceedings of III National Congress of Informatic Chemistry, Naples, Italy,1997, pp 295-300.

an HP5890 gas chromatograph coupled with an HP5989A mass spectrometer. Chemicals. Dichloromethane, cyclohexane, and methanol (Aldrich, fluorescence grade) and PAH compounds (Aldrich) were used as received. Samples. The aromatic oil (anthracene oil) is a dark tarry liquid obtained by coal tar distillation, mainly consisting of threering alkyl-substituted aromatics. PAH-laden mixtures were isokinetically sampled by means of a sampling line consisting of a stainless steel water-cooled probe (i.d. ) 2 mm), a condenser, and a filter in an atmospheric-pressure fuel-rich premixed C2H4/O2 flame (C/O ) 0.8). More experimental details on the sampling procedure are reported in previous works.13,14 RESULTS AND DISCUSSION Evaluation of the PAH Fluorescence Contribution to PAHLaden Samples and to Standard PAH Mixtures in Dichloromethane. The PAH-laden mixtures were obtained by extraction of the flame-sampled total particulate with dichloromethane (DCM), which has been shown to be the best solvent for their extraction and solubilization.1 By consequence, the fluorescence spectra of the PAH-laden mixtures have been measured in DCM solutions, even though most of the literature spectroscopic data15-17 are relative to PAH solutions in spectroscopically transparent solvents, such as cyclohexane and ethanol. The excitation wavelength (λexc ) 266 nm) chosen to measure the fluorescence emission is in the range where PAH mixtures show a high absorbance and corresponds also to a typical excitation wavelength of a Nd:YAG laser recently used for laserinduced fluorescence (LIF) measurements applied directly in the flame for PAH determination.14 In synchronous fluorometry, that is, the simultaneous scanning of excitation and emission wavelengths with a fixed wavelength difference of ∆λ, a signal is observed only when ∆λ matches the interval between one absorption band and one emission band. A relatively small ∆λ would favor the emissions from aromatic structures with strong 0-0 transitions, whereas a relatively big ∆λ appears to favor the emission from aromatic structures without strong 0-0 transitions.9 On the basis of the results from trial experiments, a ∆λ of 10 nm, used in this work, appeared to be a good compromise under these constraints. The GC/MS analysis of the flame samples allowed the identification and quantification of unsubstituted PAHs of two rings (naphthalene) to seven rings (coronene), which account for ∼3040% of the total mass. The distribution of PAHs contained in the flame-sampled PAH mixture is reported in the Supporting Information. The large abundance of two-to-three-ring PAHs typically found in fuel-rich combustion conditions18 is evident. (13) Ciajolo, A.; Barbella, R.; Tregrossi, A.; Bonfanti, L. Proc. Int. Symp. Combust. 1998, 27, 1481-1487. (14) Ciajolo, A.; Ragucci, R.; Apicella, B.; Barbella, R.; de Joannon, M.; Tregrossi, A. Chemosphere 2000, 42/3-4, 383-389. (15) Berlman, I. B. Handbook of Fluorescence Spectra of Aromatic Molecules, 2nd ed.; Academic Press: New York, 1971. (16) Karcher, W.; Fordham, R. J.; Dubois, J. J.; Glaude, P. G. J. M.; Lighart, J. A. M. Spectral Atlas of Polycyclic Aromatic Compounds; D. Reidel Publishing Company: Dordrecht, 1983; Vol. 1. (17) Karcher, W. Spectral Atlas of Polycyclic Aromatic Compounds; Kluwer Academic Publishers: New York, 1988; Vol. 2.

Analytical Chemistry, Vol. 76, No. 7, April 1, 2004

2139

Figure 1. Measured and composite fluorescence emission spectra at fixed (λexc ) 266 nm) (a) and synchronous (∆λ ) 20 nm) (b) excitation wavelengths of a flame-sampled PAH-laden mixture in dichloromethane solution.

The measured UV-excited (λexc ) 266 nm) fluorescence spectra of a PAH-laden mixture collected in a premixed sooting ethylene/ oxygen flame is reported in Figure 1a in comparison with the composite PAH spectrum obtained with the procedure described in the Experimental Section. It can be noted that the composite PAH spectrum overcomes the measured one in the 300-340-nm wavelength region where two- to three-ring compounds, and in particular fluorene, are known to be the main contributors to the UV fluorescence.15-17 By contrast, the fluorescence profile of the PAH sample in the high-wavelength region (from 340 to 500 nm), that is, in the region where large aromatic ring systems preferentially fluoresce, is only partially accounted for by the composite spectrum. The differences between the measured and the composite fluorescence spectra are enhanced by applying a synchronous technique, since the synchronous fluorescence technique is known to narrow the spectral bands and improve the selectivity by spectral simplification. This is clearly exhibited in Figure 1b, where (18) Haynes, B. S. In Fossil Fuel Combustion; Bartok, W., Sarofim, A. F., Eds.; John Wiley & Sons:New York; p 261, 1991.

2140

Analytical Chemistry, Vol. 76, No. 7, April 1, 2004

the measured synchronous spectrum of the PAH-laden flame sample is compared with the composite spectrum. Indeed, it can be observed that the overestimation of the composite spectrum in the low wavelength region and the underestimation in the long wavelength region are much more evident by using synchronous technique. The underestimation of fluorescence in the visible range can be attributed to the presence of species larger than those analyzed by gas-chromatography/mass spectrometry analysis whose contribution is not included in the composite PAH spectrum that includes only PAHs up to seven rings. The chemical characterization of the heaviest fraction of flame-sampled PAH-laden mixtures is still under study both by using spectroscopic analysis13,14 and by size exclusion chromatography (SEC).19,20 To investigate the reliability of the spectral composition procedure used for establishing the contribution of GC/MS-PAH to the total fluorescence, it was applied to standard PAH mixtures. The mixtures consisted of the most abundant PAHs detected in the flame samples. Their composite and measured spectra at fixed (λexc ) 266 nm) and synchronous (∆λ ) 10 nm) excitation wavelengths are reported in the Supporting Information. It appeared evident that the long-wavelength region of the PAH fluorescence profile was well-reproduced in the composite spectrum, confirming that the underestimation of the visible emission shown before (Figure 1a-b) is due to the presence in the PAHladen flame sample of other fluorescing species above the mass detection limit of gas chromatography/mass spectrometry (300 Da). The composition procedure was also applied to an aromatic oil, anthracene oil, which is not a laboratory mixture of PAHs completely analyzable by GC/MS. The composite PAH spectrum of anthracene oil (see the Supporting Information) reproduced quite well the measured one from 340 nm up to the visible, confirming the reliability of the composition procedure in this region of the spectrum. As previously observed for the PAH-laden samples (Figure 1), the composition procedure applied both to the standard mixture and to the anthracene oil showed that the measured fluorescence at λem ) 310-340 nm was much lower than that calculated on the basis of the fluorescence measured on single PAH solutions. The reduction of the UV fluorescence peak with respect to the composite spectrum occurred to the same extent for the standard and flame samples (∼30% less than the computed signal). The effect of the heavy atom in the solvent or the possible interactions between the PAHs present in the mixtures could be invoked to explain the depression of the UV signal in both combustion samples and PAH mixtures. Interaction with Other PAHs. The PAHs responsible for the UV emission are two-to-three-ring PAHs, and among those present in the PAH-laden samples and in the standard mixtures, the PAH responsible for the UV emission is fluorene, since its quantum efficiency is very high (0.76, as measured in this work), in agreement with the literature15 with respect to the other two-tothree-ring PAHs. Thus, the UV emission reduction in the PAH (19) Apicella, B.; Ciajolo, A.; Suelves, I. I.; Morgan, T. J.; Herod, A. A.; Kandiyoti, R. Combust. Sci. Technol. 2002, 174 (11-12), 403-417. (20) Apicella, B.; Barbella, R.; Ciajolo, A.; Tregrossi, A. Chemosphere 2003, 51/ 10, 1063-1069.

Table 1. Change in the Fluorescence Peak Height at 310 nm (λexc ) 266 nm)

fluorene standard mixture flame-sampled PAH mixture anthracene oil

from DCM to methanol (%)

from DCM to cyclohexane (%)

from methanol to DCM (%)

from cyclohexane to DCM (%)

+25 +38 not measured not measured

+19 +35 +33 +34

-22 not measured not measured not measured

-24 not measured not measured not measured

mixtures has been considered as the reduction of fluorene fluorescence. The reduction of fluorene emission when mixed with other PAHs has been observed in earlier work10 and attributed to pyrene interference. This hypothesis was contrasted by Lloyd,11 who attributed this finding to “inner filter effects” due to the quite high PAH concentrations used for the emission measurements by the authors.10 However, all the fluorescence measurements described in this current report were performed in very dilute conditions (=0.01 mg/L) such that inner filter effects could be excluded21 and an energy transfer process occurring between fluorene and other PAH could be invoked for justifying fluorescence quenching of fluorene. The energy transfer processes should occur by a radiative process involving a photon emission by the donor (fluorene) and absorption by the acceptor molecule, whereas radiationless transfer has to be excluded, since it involves dipoledipole or collisional interactions, which generally occur in much more concentrated solutions.22 The overlap between the emission spectrum of fluorene and the excitation spectrum of pyrene suggests that the radiation emitted by fluorene (the donor) can be actually absorbed by pyrene, which could act as acceptor. However, in the present study, it was found (spectra not reported) that the UV fluorescence reduction occurs also for PAH mixture 3 (in which pyrene is absent; see Supporting Information), and it is very similar to the reduction observed for the other two PAH mixtures, in which the fluorene/pirene concentration ratio is 1 and 0.5, respectively. Fluorene was mixed with individual PAHs (benzo[e]pyrene, indeno[1,2,3-cd]pyrene, benzo[ghi]perylene) that have an absorption spectrum slightly overlapped with the fluorene emission spectrum, and the suppressive effect was still observed. By consequence, it was concluded that in a PAH mixture, the fluorene fluorescence reduction depends in a complex way on many parameters (the concentration ratio between donor and acceptor; the degree of overlapping between emission and absorption spectra of donor and acceptor, respectively; the presence of different acceptors in solution; etc.). Solvent Effect. To determine if the complex effect observed in DCM solutions between fluorene and the other PAHs occurs also in other solvents, measurements were performed by diluting a concentrated DCM mixture of 18 PAHs (mixture 1) widely used in spectroscopic studies with methanol and cyclohexane. The dilution of concentrated DCM solution was necessary because of the lower solubility of sample components in methanol and cyclohexane, which hindered the direct dissolution in these solvents. (21) Parker, C. A. Photoluminescence of Solutions; Elsevier Publishing Company: Amsterdam, 1968. (22) Birks, J. B. Photophysics of Aromatic Molecules; Wiley-Interscience: London, 1970.

Figure 2. Comparison of fluorescence emission spectra (λexc ) 266 nm) measured in cyclohexane and in dichloromethane solution of a flame-sampled PAH-laden mixture.

The oxygen quenching of the fluorescence signal could be neglected in this study because the oxygen solubility in cyclohexane, methanol, and dichloromethane are quite similar3 (slightly higher in cyclohexane), and the oxygen quenching is scarcely effective on fluorene. In fact, oxygen is known to be a quencher of fluorophores having decay times in excess of 20 ns,23 whereas the fluorene decay time is 10 ns.15 The change (increase is indicated with a positive sign, decrease with a negative sign) of the fluorescence signal at 310 nm observed by diluting the PAH mixtures, the anthracene oil, and the flamesampled PAH mixture is reported in Table 1. Overall, it can be noted that the dilution with cyclohexane and methanol enhances the UV peak. To investigate the effect of DCM on fluorene fluorescence, dilute fluorene solutions (0.01 mg/L) in cyclohexane and methanol prepared from concentrated DCM solutions (10 mg/ L) of fluorene were compared to dilute fluorene solutions in DCM (0.01 mg/L) prepared from concentrated cyclohexane and methanol solutions of fluorene (10 mg/L). In this case, the reduction of the fluorescence signal by dichloromethane dilution was ∼20%. The enhancement of the UV fluorescence signal by dilution with cyclohexane can be clearly seen in the spectra of a flamesampled PAH-laden mixture (Figure 2) and also in the spectra of the anthracene oil sample (reported in the Supporting Information) both in pure DCM and diluted in cyclohexane. Moreover, the composition procedure applied to the standard PAH mixture 1 dissolved in cyclohexane reproduces completely (23) Sharma, A.; Schulman, S. G. Introduction to Fluorescence Spectroscopy; WileyInterscience: London, 1999.

Analytical Chemistry, Vol. 76, No. 7, April 1, 2004

2141

Figure 3. Measured and composite fluorescence emission spectra at λexc ) 266 nm of PAH mixture 1 (a) and of a flame-sampled PAHladen mixture (b) in cyclohexane.

the measured spectrum both in the UV and in the visible region, as shown in Figure 3a. The composition procedure applied to the flame-sampled PAH-laden mixture in cyclohexane, reported in Figure 3b, confirms that no signal reduction occurs in the UV region; i.e., the contribution to the UV fluorescence is due to GCPAH. On the other hand, the underestimation of the visible fluorescence is still present in cyclohexane, and it supports our claim that there is no solvent effect, and species not detectable by GC/MS are responsible for the visible fluorescence of the flame-sampled PAH mixture. From these observations, it appears that the reduction of UV fluorescence in DCM solutions is mainly due to the occurrence of a heavy-atom solvent effect. The heavy-atom solvent effect, which was first studied by Kasha,24 is known as the external heavy-atom effect.3,22 It can dramatically affect the intensity of fluorescence, but does not have an appreciable effect on the frequency of transition.23 More in detail, high nuclear charges of heavy atoms, such as iodine or (24) Kasha, M. J. Chem. Phys. 1952, 20, 71-74.

2142

Analytical Chemistry, Vol. 76, No. 7, April 1, 2004

chlorine, in solvent molecules of the primary solvent cage of solute molecule causes spin and orbital electronic angular momentums of the solute to interact strongly with each other. The interaction of spin and orbital electronic motions, which is very weak in the absence of heavy atoms, results in the loss of total molecular spin angular momentum as a well-defined molecular property (quantum number). As a result, the selection rule forbidding changes of spin angular momentum in electronic transitions is partially removed, with the consequence that the probabilities of singletf triplet absorptive transitions, singlet-triplet intersystem crossing from the lowest excited singlet state, phosphorescence, and triplet-singlet intersystem crossing from the lowest triplet to the ground state are significantly increased. Consequently, molecules fluorescing in ordinary solvents (e.g., water, alcohols, hydrocarbons) can fluoresce less strongly or may even have their fluorescence completely quenched in heavy-atom solvents. Fluorescence quenching can be static (e.g., complex formation) or dynamic (e.g., collision quenching).25 The halogen-containing substances generally act as collision quenchers. In the case of collision quenching, the quencher must diffuse to the fluorophore during the lifetime of the excited state. Upon contact, the fluorophore returns to the ground state, without emission of a photon. All quenching processes that depend on the varying distances between excited and other molecules will obey the Stern-Volmer equation in their dependence on the number of “effective collisions”, regardless of the specific mechanism by which the excitation energy is consumed. Assuming that an excited molecule, once in contact with a quencher, cannot react with a second quencher approaching from another direction, the quenching will still be increased by an increase in the quencher concentration, although it will no longer obey the Stern-Volmer equation, particularly in the limiting case when the excited molecules are permanently and completely surrounded by “quenching molecules”, namely, the molecules of solvent.26 However, the heavy-atom effect cannot be the only reason for the UV fluorescence quenching, since the spectrum of fluorene used for the composition procedure was obtained in DCM, and the UV quenching was still observed. The UV quenching in sample mixtures dissolved in DCM is ∼33 to 38%, that is, higher than the quenching observed in a DCM solution of pure fluorene (about 20%), suggesting that the external heavy-atom effect acts synergically or favors the interaction between fluorene and the other fluorescing species present in solutions. On this complex effect, further study will be, however, necessary. It is worth noting that the results presented in this work have been found by using both fixed excitation wavelength and synchronous fluorescence techniques. This confirms, in agreement with Lloyd,4,11 that the occurrence of the fluorene quenching is a limit of both synchronous and conventional fluorescence techniques. The solvent effect and the interaction between PAHs have to be taken into account for a correct interpretation of fluorescence spectra of complex PAH mixtures, such as those formed in (25) Lakowicz, J. R. Principles of Fluorescence Spectroscopy; Plenum Press: New York, 1983. (26) Pringsheim, P. Fluorescence and Phosphorescence; Interscience Publishers: New York, 1949.

combustion processes and those originating from heavy fossil fuels and coal products. CONCLUSIONS In this work, a composition procedure for the interpretation of synchronous and conventional fluorescence spectra of complex PAH-laden mixtures was proposed. The occurrence of an “external heavy atom effect” when the solvent contains heavy atoms, such as dichloromethane, has been found that can act sinergically with an energy transfer between fluorene, mainly responsible for UV fluorescence, and other fluorescing PAHs. Therefore, caution should be used in comparing fluorescence spectra of samples when different solvents are used.

The interpretation of fluorescence spectra of PAH samples in the UV region needs to take into account solvent effects and the possible interaction between PAHs. SUPPORTING INFORMATION AVAILABLE Experimental details and resulting spectra are available as Supporting Information. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review July 28, 2003. Accepted February 2, 2004. AC034860K

Analytical Chemistry, Vol. 76, No. 7, April 1, 2004

2143