Determination of non-certified levoglucosan, sugar polyols and ergosterol in NIST Standard Reference Material 1649a

Atmospheric Environment 84 (2014) 332e338

Contents lists available at ScienceDirect

Atmospheric Environment journal homepage: www.elsevier.com/locate/atmosenv

Determination of non-certified levoglucosan, sugar polyols and ergosterol in NIST Standard Reference Material 1649a Donatella Pomata a, *, Patrizia Di Filippo a, Carmela Riccardi a, Francesca Buiarelli b, Valentina Gallo b a b

Inail, Research, Certification and Control Division, DIPIA, Via di Fontana Candida 1, 00040 Monteporzio Catone, Rome, Italy Department of Chemistry, University of Rome “Sapienza”, P.le Aldo Moro, 5-00185 Rome, Italy

h i g h l i g h t s
Levoglucosan and xylitol were used as tracers for biomass burning emissions.
Arabitol, mannitol and ergosterol were used as biomarkers for airborne fungi.
A matrix effect was observed for xylitol, mannitol and ergosterol.
The above mentioned analytes were determinated in SRM 1649a and PM10 samples by GC/MS.
This study represent an important starting point to future inter-comparison studies.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 1 July 2013 Received in revised form 25 November 2013 Accepted 27 November 2013

Organic component of airborne particulate matter originates from both natural and anthropogenic sources whose contributions can be identified through the analysis of chemical markers. The validation of analytical methods for analysis of compounds used as chemical markers is of great importance especially if they must be determined in rather complex matrices. Currently, standard reference materials (SRM) with certified values for all those analytes are not available. In this paper, we report a method for the simultaneous determination of levoglucosan and xylitol as tracers for biomass burning emissions, and arabitol, mannitol and ergosterol as biomarkers for airborne fungi in SRM 1649a, by GC/MS. Their quantitative analysis in SRM 1649a was carried out using both internal standard calibration curves and standard addition method. A matrix effect was observed for all analytes, minor for levoglucosan and major for polyols and ergosterol. The results related to levoglucosan around 160 mg g1 agreed with those reported by other authors, while no comparison was possible for xylitol (120 mg g1), arabitol (15 mg g1), mannitol (18 mg g1), and ergosterol (0.5 mg g1). The analytical method used for SRM 1649a was also applied to PM10 samples collected in Rome during four seasonal sampling campaigns. The ratios between annual analyte concentrations in PM10 samples and in SRM 1649a were of the same order of magnitude although particulate matter samples analyzed were collected in two different sites and periods. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Levoglucosan Xylitol Arabitol Mannitol Ergosterol Standard reference material

1. Introduction Natural sources and biomass combustion contribute to the complex carbonaceous component of atmospheric aerosol. Such contributions can also be identified and quantified through the analysis of chemical markers in collected aerosols. Levoglucosan (1,6-anhydro-b-D-glucopyranose) is a dehydromonosaccharide derivate exclusively formed from the thermal

* Corresponding author. E-mail address: [email protected] (D. Pomata). 1352-2310/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.atmosenv.2013.11.069

breakdown of cellulose during combustion and so it is regarded as an unambiguous tracer of biomass burning emissions (Kuo et al., 2008; Louchouarn et al., 2009; Schkolnik and Rudich, 2006; Simoneit et al., 1999). The determination of levoglucosan atmospheric concentrations became even more important with the development of wood as renewable energy for domestic heating. In fact, many researchers demonstrated the increase, during recent years, of atmospheric particulate matter load, due to domestic biomass combustion in developed countries (Piot et al., 2011). As other saccharides, xylitol is suggested as a possible tracer for surface soil dust (Simoneit et al., 2004a) although it may also be emitted significantly by thermal stripping during burning in

D. Pomata et al. / Atmospheric Environment 84 (2014) 332e338

wildfires (Simoneit et al., 2004b). Moreover, some researchers consider that xylitol may be a secondary species produced during the upward transport of aerosols (Gao et al., 2003; Pio et al., 2008). Ergosterol, mannitol and arabitol are regarded as biomarkers for the determination of fungal component of aerosol (viruses, bacteria, fungi, algae, fern spores, pollens) well known to cause several respiratory allergic diseases (Deguillaume et al., 2008; Srikanth et al., 2008; Womiloju et al., 2003). The steroid ergosterol was first considered as a marker for fungal abundance in environmental samples (Lee et al., 1980; Miller and Young, 1997; Volker et al., 2000), and it is still considered a reliable biomarker for airborne fungi (Di Filippo et al., 2013; Lau et al., 2006; Lee et al., 2007; MilleLindblom et al., 2004; Zhao et al., 2008). In fact ergosterol in fungi and fungal spores is bound to the cell membrane in both free and esterified form (Mille-Lindblom et al., 2004). The two sugar alcohols, arabitol and mannitol, constitute an important fraction of the dry weight of fungi, and in particular mannitol can contribute between 20 and 50% of the mycelium dry weight (Vélëz et al., 2007). Although other natural sources of these two polyols (lower plants, bacteria, insects and algae) (Graham et al., 2003; Vélëz et al., 2007; Zhang et al., 2010) are also known to exist, in some cases, arabitol and mannitol can be associated to fungal spores in airborne particulate matter, as biomarkers (Carvalho et al., 2003; Di Filippo et al., 2013). Analysis of levoglucosan, polyols and ergosterol is frequently performed with Gas ChromatographyeMass Spectrometry (GCe MS) technique after derivatization (Burshtein et al., 2011; Piot et al., 2011; Simpson et al., 2004). Despite many methods of extraction, purification and analysis have already been developed to determine these analytes in particulate matter samples, there is still a lack of thorough validation procedures. The validation of analytical methods for analysis of compounds used as chemical markers is of great importance especially if they must be determined in rather complex matrices, as in this case. Currently, standard reference materials (SRM) with certified values for all those analytes are not available, and National Institute of Standard and Technologies (NIST) only provides the SRM 1649a with reference value for levoglucosan concentration, expressed as mass fraction. The aim of this study was to simultaneously determine, in addition to levoglucosan, other not-certified compounds, such as three sugar polyols (xylitol, mannitol and arabitol) and a plant sterol (ergosterol) in SRM 1649a, following our previously published protocol of analysis by GC/MS (Buiarelli et al., 2013; Di Filippo et al., 2013). Due to the lack of published studies concerning the other compounds, only levoglucosan results could be compared with NIST reference mass concentration and with those already obtained by other authors (Kuo et al., 2008; Larsen et al., 2006; Louchouarn et al., 2009; Orasche et al., 2011). Quantitative analysis of levoglucosan, sugar polyols and ergosterol by GCeMS in SRM 1649a samples was conducted both by internal standard calibration curves method and by the standard addition method to provide an internal confirmation of our results. Results obtained in this study represent an important starting point to future inter-comparison studies. 2. Materials and method 2.1. Chemicals and reagents All the chemical and chromatographic reagents used were HPLC or analytical grade. HPLC RS-Plus Methanol, anhydrous Ethanol and Toluene were purchased by ROMIL (Romil, Cambridge, UK).

333

N,O-bis(trimethylsilyl)trifluoroacetamide (BSTFA), trimethylchlorosilane (TMCS) and pyridine were obtained from Sigma Aldrich (SigmaeAldrich, St. Louis, MO). Levoglucosan, xylitol, mannitol, arabitol, ergosterol, levoglucosan-13C (IS), xylitol-13C (IS), mannitol-13C (IS), and dehydrocholesterol (IS) were purchased from Sigma Aldrich. Stock standard solution of unlabeled analytes (multistandard solution) and internal standard solution were prepared by dissolving each compound in methanol (1 mg mL1) and storing them at 20  C in the dark. 2.2. Materials Standard Reference Material 1649a prepared by the U.S. National Institute of Standards and Technology (Gaithersburg, MD, USA) was collected in urban settings and is thus also referred to as urban dust SRM. The first urban dust was collected in 1976e1977 in the Washington, D.C. area over a period in excess of twelve months. The material was then sieved through a fine-mesh sieve (<125 mm) and, in 1982, issued as SRM 1649, as a standard for polycyclic aromatic hydrocarbon analyses in aerosols. This material was recertified in 1999 for additional organic and inorganic contaminants as well as radiocarbon and total organic carbon content and issued as SRM 1649a. Strata NH2 cartridges (200 mg mL1) were obtained from Phenomenex (Torrance, CA). DB-5MS capillary column (30 m  0.25 mm i.d., 0.25 mm film thickness) was obtained from Agilent Technologies Inc. (Palo Alto, CA). A high purity, inert diatomaceous earth sorbent (hydromatrix) was obtained from Dionex (ThermoFisher Scientific, Sunnyvale, CA). 2.3. Sample preparation For the analysis of SRM 1649a, the standard material was weighed using an analytical electronic balance (Sartorius MC-5, Dm  0.001 mg) after conditioning in a climatic chamber (Activa Climatic Cabinet, Aquaria MI, according to UNI EN 12341/2001; UNI CHIM 285/2003 and D.M. 60/2002) for 24 h, at T ¼ 20  1  C and at 50  5 relative humidity and deposited on clean Teflon filters. The extraction of analytes (Buiarelli et al., 2013) was carried out with ethanol by Accelerated Solvent Extraction (ASE 200eDionex, ThermoFisher Scientific, Sunnyvale, CA), operating at high pressure (1500 psi) and temperature (100  C) using two static cycles, after spiking samples with levoglucosan-13C, xylitol-13C, mannitol-13C and dehydrocholesterol as internal standards (IS). Hydromatrix was used to fill the void space in the pressurized fluid extraction vessels. As part of the optimization method, already published in detail (Buiarelli et al., 2013), extraction tests were run on blank Teflon filters spiked with a multi-standard solution containing all target analytes, to evaluate the extraction efficiencies. Methanol 100% and ethanol 100% gave similar results, but ethanol was chosen as extraction solvent, since the subsequent purification involved the use of the amino-cartridge, that does not retain the analytes of interest if solubilized in methanol. The appropriate number of extraction cycles was also tested and, while the second extraction cycle still contained an appreciable amount of analytes, the third one was free of them. A test checking the third cycle of extraction with ethanol was also performed on filter added with an aliquot of SRM 1649a and spiked with a solution of IS. The following GC/MS analysis of this third extract proved the absence of the target analytes. In all the tests performed, the IS solution was added just before derivatization step. The additions were carried out in order to obtain final concentrations of 13C-labeled and unlabeled analytes within the range of linearity found for each compound and the

334

D. Pomata et al. / Atmospheric Environment 84 (2014) 332e338

results were between 90% and 95% for all compounds. The high recovery efficiency indicated that high pressure and temperature did not affect the extraction method. A filter was spiked only with the internal standard before the derivatization step and the small contributes detected in the un-spiked blank were successively subtracted. The extract was then reduced to w1 mL under nitrogen using an Glas-Col SE 500 automated evaporation system (Glas-Col, Terre Haute, IN), passed through an amino-cartridge using a vacuum manifold (Alltech12-Port model SPE Vacuum manifold) (Grace, Deerfield, IL) and divided into two parts: the apolar fraction was immediately released from the cartridge in ethanol, while polar compounds, retained by the cartridge, were eluted with a mixture of methanol-water (80:20). The two fractions, ethanolic (containing ergosterol) and methanolic (containing sugars) were evaporated to dryness under a slight nitrogen flow and stored in a dryer, prior to derivatization by BSTFA containing 1% of TMCS and 20% of pyridine at 70  C for 1 h. After cooling to room temperature, the two fractions were dried again in order to evaporate excess of pyridine affecting GC analysis. In order to guarantee the analysis of trace amounts of compounds, in a complex matrix like particulate matter, the clean-up step was crucial and the SPE cartridge allowed to both isolate the compounds of interest in two different fractions and to remove those substances that might interfere with the analysis. The purification step resulted in chromatographic peaks with increased signal to noise ratios in tests carried out by extracting spiked SRM 1649a portions.

Positive identification of analytes was carried out using retention time and by comparing the relative abundance of quantification/confirmation ions from each sample to those produced by authentic standards. 2.5. Quantitative analysis Concerning levoglucosan in SRM 1649a samples, in some studies reported in literature (Harrison et al., 2012; Kuo et al., 2008; Larsen et al., 2006; Louchouarn et al., 2009), quantitative analysis was carried out with linear response curves obtained by injecting standard solutions of levoglucosan and its internal standard in a GC/MS system. Standard addition method applied to reference filters (PM samples) by in-situ derivatization thermal desorption gas chromatography time-of-flight mass spectrometry (IDTD-GCTOFMS) was only used by Orasche et al. (2011), that measured SRM 1649a levoglucosan concentration by interpolation on standard addition plot built with reference filters. Due to the complexity of the particulate matter as analytical matrix, in the present study quantitative analysis of urban dust SRM 1649a samples was carried out using both internal standard calibration curves and the standard addition method. The ratio of analyte area to that of internal standard versus the ratio of analyte concentrations to that of internal standard was used to make either calibration plots or standard addition curves. In the first case, extracted and purified SRM 1649a samples were quantified through standard calibration curves; in the second case the SRM amount of target analytes was obtained as x-intercept of the standard addition plots.

2.4. GC/MS equipment and conditions A HP 6890 gas chromatograph fitted with HP 7683 autosampler and connected to HP5973 was utilized (Agilent Technologies, Palo Alto, CA). A single quadrupole mass-selective detector was used for GCeEIeMS analysis. GC separation was achieved in 12 min on a DB5MS capillary column. The temperature program was: 100  C initial temperature, ramped at 25  C min1 to 180  C, then ramped at 40  C min1 to 300  C and held for 8 min. Samples (1 ml) were injected in splitless mode. The injector temperature was set at 280  C. The helium carrier gas was maintained at a constant flow of 1.0 mL min1. The quadrupole and ion source temperatures were set at 150  C and 230  C, respectively. The mass spectrometer was operated in EI mode at 70 eV, using the Agilent MSD ChemStation D.01.00 software. Acquisitions were performed in selected ion monitoring (SIM) mode. The mass fragments monitored, the expected relative ion abundances, and retention times of both target analytes and their internal standards are summarized in Table 1. For each compound, three mass fragments were monitored with one fragment used for quantification (target ion) and the other two fragments (qualifier ion) used for additional confirmation of identity.

2.5.1. Standard calibration curves Standard calibration curves were constructed by adding increasing volumes (from 10 to 100 ml) of the multi-standard solution and a constant volume (50 ml) of the internal standard solution, for six calibration levels, on clean Teflon filters, in a concentration range from 28 to 300 mg L1 for levoglucosan, from 28 to 1480 mg L1 for arabitol, from 14 to 1500 mg L1 for xylitol, from 14 to 860 mg L1 for mannitol, and from 30 to 830 mg L1 for ergosterol. The small volumes, added before extraction in the ASE vessels, were immediately adsorbed by diatomaceous earth. The thus prepared samples were processed following the whole procedure to take account of any losses during sample treatment and injected three times in the GCeMS system. 2.5.2. Standard addition method The standard addition method was used to overcome any potential matrix effects. For this aim, two samples of 1.47 and 10.43 mg of urban dust SRM 1649a, deposited on two clean Teflon filters, were extracted. Both extracts were divided into seven aliquots, each corresponding to the extract of respectively 0.21 and 1.49 mg of urban dust.

Table 1 Mass fragments monitored, relative ion abundances, and retention times. Analytes

MW

MW-(TMS)n

Target Ions m/z (ab. %)

Qualifier Ions m/z (ab. %)

13

168 162 157 152 152 188 182 384

384 378 517 512 512 620 614 496

206 204 310 217 217 323 319 351

338 333 323 307 307 220 217 325

396

468

363 [M-105] (100)

C-Levoglucosan (ISTD) Levoglucosan 13 C-Xylitol (ISTD) Xylitol Arabitol 13 C-Mannitol (ISTD) Mannitol Deidrocholesterol (ISTD) Ergosterol

[M-178] [M-174] [M-187] [M-295] [M-295] [M-297] [M-295] [M-145]

(100) (100) (100) (100) (100) (100) (100) (100)

[M-46] (16) [M-45] (19) [M-194] (77) [M-205] (21) [M-205] (39) [M-400] (53) [M-397] (50) [M-171] (99)

337 [M-131] (68)

tR 220 217 220 319 319

[M-164] [M-161] [M-297] [M-193] [M-193]

(99) (95) (94) (16) (32)

205 [M-409] (69) 456 [M-40] (14)

5.12 5.12 5.13 5.13 5.17 5.83 5.83 11.10

468 [M] (11)

11.54

D. Pomata et al. / Atmospheric Environment 84 (2014) 332e338

335

Table 2 Concentration ranges, equations and correlation coefficients of standard calibration curves for all analytes. Compounds

Concentration ranges (mg L1)

13

220 28e300 200 28e1480 14e1500 190 14e860 400 30e830

C-Levoglucosan Levoglucosan 13 C-Xylitol Xylitol Arabitol 13 C-Mannitol Mannitol Deidrocholesterol Ergosterol

The extract corresponding to 0.21 mg was chosen to simulate a sample coming from a multistage impactor whose filters collect an amount of particulate matter ranging from about 0.05 to 0.7 mg if the sampling period is two weeks (Di Filippo et al., 2010). The extract corresponding to 1.49 mg was chosen based on the quantity of particulate matter averagely collected with a sampler sampling at a flow rate of 2.3 m3 h1, for 24 h. From 10 to 100 ml of the methanolic standard solutions were added, after extraction, to each aliquot, in order to obtain final concentrations of 13C-labeled and unlabeled analytes consistent with those of standard calibration curve points in a concentration range from 10 to 300 mg L1 for levoglucosan, arabitol and mannitol, from 70 to 460 mg L1 for xylitol and from 30 to 250 mg L1 for ergosterol. The resulting solutions were processed following the whole procedure and then injected three times in the GCeMS system, obtaining for each analyte two standard addition plots (PLOT1 and PLOT2 respectively for 0.21 and 1.49 mg of urban dust). The concentrations (Cx) of analytes in the SRM 1649a were obtained as x-intercept of the regression lines. Small contributes due to blank filters and protocol of extraction, clean-up, and derivatization were subtracted by the chromatographic signals obtained. Concentration values obtained for each analyte were also corrected for the recovery of the ASE extraction. 3. Results and discussion 3.1. Levoglucosan, sugar polyols and ergosterol concentrations in urban dust SRM 1649a Good linearities with correlation coefficients (R2) not lower than 0.993, for both internal standard calibration curves (Table 2) and standard addition plots (Table 3), were obtained for all analytes of interest.

Equation of calibration curves

R2

y ¼ (0.625  0.029) x þ (0.099  0.022)

0.993

y ¼ (4.456  0.019) x þ (0.139  0.062) y ¼ (3.177  0.048) x þ (0.198  0.016)

0.999 0.997

y ¼ (0.794  0.029) x þ (0.057  0.017)

0.996

y ¼ (0.792  0.013) x  (0.014  0.004)

0.995

As for quantitative analysis of urban dust SRM 1649a carried out with standard calibration curves, concentrations (mg g1) of levoglucosan, polyols and ergosterol (expressed as mean  standard deviation of three repeated tests) obtained from the extraction of three samples of about 0.6e2 mg are shown in Table 4. Table 4 also shows the concentrations (mg g1) of the target analytes extrapolated with standard addition plots (expressed as extrapolated values  their standard deviations (sxE )). As seen from Table 4, for quantitative analysis of levoglucosan, both standard addition plots were used and the results obtained with two extrapolations were 173.3  27.7 mg g1 (PLOT1) and 160.3  16.2 mg g1 (PLOT2) with a mean of 166.78 and percentage differences of 7.51%. Arabitol, mannitol and ergosterol concentrations in SRM 1649a were only extrapolated with PLOT2 since the concentration of these compounds in 0.21 mg of SRM 1649a was lower or equal to the analytical LOD, as expected given their low atmospheric concentrations. On the contrary only PLOT1 was used to extrapolate xylitol concentration in SRM 1649a since the concentration obtained with PLOT2 showed a very high percentage standard deviation (see Table 4) probably due to the higher number of interferents present in 1.49 mg of SRM 1649a and the high amount of xylitol, as expected for its averagely higher atmospheric concentration (Di Filippo et al., 2013). 3.2. Evaluation of the matrix effect Matrix effect can both reduce and enhance the detector response when compared to response of the standards in neat solvent. The matrix effect depends on the instrument (mainly GCinjection system), the type and amount of matrix (grams of matrix per milliliter of extract), the sample pre-treatment procedure

Table 3 Concentration ranges, equations and correlation coefficients of standard addition plots for all analytes. Compounds

Concentration ranges (mg L1)

13

220 7e300 7e300 200 70e460 70e460 8e300 8e300 190 6e260 6e260 400 30e250 30e250

C-Levoglucosan Levoglucosan (PLOT1) Levoglucosan (PLOT2) 13 C-Xylitol Xylitol (PLOT1) Xylitol (PLOT2) Arabitol (PLOT1) Arabitol (PLOT2) 13 C-Mannitol Mannitol (PLOT1) Mannitol (PLOT2) Deidrocholesterol Ergosterol (PLOT1) Ergosterol (PLOT2)

Equation of standard addition plots

R2

y ¼ (0.679  0.008) x þ (0.129  0.005) y ¼ (0.524  0.001) x þ (0.545  0.007)

0.997 0.995

y y y y

¼ ¼ ¼ ¼

(3.313 (2.639 (2.157 (3.681

   

0.030) 0.108) 0.036) 0.012)

x x x x

þ þ þ þ

(0.216 (0.231 (0.115 (0.217

   

0.022) 0.140) 0.027) 0.009)

0.999 0.995 0.996 0.999

y ¼ (1.184  0.004) x þ (0.035  0.003) y ¼ (1.811  0.008) x þ (0.171  0.005)

0.999 0.999

y ¼ (0.668  0.048) x þ (0.016  0.030) y ¼ (0.563  0.010) x þ (0.040  0.003)

0.993 0.999

336

D. Pomata et al. / Atmospheric Environment 84 (2014) 332e338

Table 4 Mean values (mg g1)  standard deviation and extrapolated values (mg g1)  standard deviations (sE) of target analyte concentrations in SRM 1649a measured with both internal standard calibration method and standard addition method. Analytes

Levoglucosan Xylitol Arabitol Mannitol Ergosterol

Calibration curves 148.54 107.86 16.58 15.63 0.61

    

18.39 23.46 3.91 0.86 0.14

Standard addition PLOT1 Concentration [mg g 173.23  27.68 130.96  12.44 n.d. n.d. n.d.

1

]

Standard addition PLOT2 160.3 86.07 12.40 19.81 0.41

    

16.22 18.85 0.96 0.98 0.06

n.d.: not detected. sxE : standard deviation of extrapolated value, calculated according to the following P formula: s2xE ¼ S2y=x =b2 ½ð1=nÞ þ y2 =b2 ðxi  xÞ2  b ¼ slope of standard addition plot. x and y ¼ centroids of the points xi, yi n ¼ number of standard addition points. sy/x ¼ standard deviation of y/x.

(derivatization), the analytes and their concentration (Pizzutti et al., 2012). Matrix-induced signal suppression in GC typically occurs when coextracted matrix components, accumulated into the gas chromatographic system generate new active sites which cause the analyte to be adsorbed and/or degraded (Pizzutti et al., 2012). In the case of matrix-induced signal enhancement, coextracted matrix components are competitively adsorbed to the active sites of the gas chromatography system (inlet and column), instead of target analytes (Pizzutti et al., 2012; Poole, 2007). In this study, the matrix effect was calculated as the slope ratios between standard calibration curves and standard addition plots. For compounds showing slope ratios between 0.9 and 1.1, matrix effect was considered to be negligible. For arabitol and levoglucosan in PLOT1, a negligible matrix effect, with slopes ratio of 0.9, was obtained. The slope ratio of 1.2, obtained for levoglucosan in PLOT2, indicates the presence of a small matrix effect, probably due to ion suppressions caused by the higher number of interferents present in 1.49 mg of SRM 1649a. A negative matrix effect (slope ratio > 1) was found for xylitol and ergosterol (respectively ratios ¼ 1.3 and 1.4), while a positive matrix effect (slope ratio < 1) was found for mannitol (ratio ¼ 0.4). Negative matrix effect observed for xylitol and ergosterol may originate from the competition between the analytes and the presence of coeluting, undetected matrix components, as also demonstrated by higher slopes of PLOT1 on respect PLOT2. For ergosterol the measured matrix effect may also be due to the use of a homologue rather than the stable isotope-labeled analyte as internal standard. Positive matrix effect observed for mannitol may be due to the presence of interferences that compete with the ionization and ion evaporation processes, effectively increasing (ion enhancement) the formation efficiency of the desired mannitol ions, as also demonstrated by higher slope of PLOT2 on respect PLOT1.

sampler (FAI Instruments, Rome), an automatic system for simultaneous sampling of airborne particles on two independent channels. The instrument was equipped with two size selective inlets and used to collect PM10 on two different filters, at a flow rate of 2.3 m3 h1. The sampling time for each PM10 sample was 24 h, with a final volume of sampled air equal to about 55 Nm3. The amount of particulate matter collected on each filter was averagely 1.5e2.0 mg and the filters were spiked with the internal standard solution and handled as described above for SRM 1649a in the paragraph on sample preparation. It is noteworthy that levoglucosan and xylitol, as marker of combustion processes, preferentially distribute in fine fraction particles (aerodynamic diameter  2.5 mm), while arabitol, mannitol and ergosterol preferentially distribute in particles with aerodynamic diameter 50 mm. Therefore their concentrations, expressed as mg g1, in the two matrices were not comparable since SRM 1649a was composed by size segregated particles 125 mm. This is confirmed by 20e30 times higher concentrations found in PM10 on respect SRM 1649a. For that reason, we studied and compared the ratios among concentrations of target analytes in Rome PM10 samples (yearly averaged) and in SRM 1649a. The results found are shown in Fig. 2. It was very interesting to observe that the ratios between analytes in PM10 samples and in SRM 1649a were of the same order of magnitude. The small differences were probably caused by the different emission sources due to the different geographic areas. Therefore these ratios could be used as diagnostic ratios to identify different source contributions to atmospheric particulate matter, such as biomass combustion, airborne fungi, etc. In order to confirm this possibility, more extensive studies must be conducted on determination of these markers and their ratios in different seasons, in sites of different typology (urban, rural, remote, industrial), in sites where wildfires have occurred. 3.4. Consistency of the results obtained for levoglucosan Levoglucosan mass concentration in SRM 1649a samples was already measured and published by other authors by using different methods of extraction and analysis (Harrison et al., 2012; Kuo et al., 2008; Larsen et al., 2006; Louchouarn et al., 2009; Orasche et al., 2011). In Kuo, Larsen and Louchouarn studies the samples were extracted via pressurized fluid extraction (PFE) with an accelerated solvent extractor (Dionex ASE 200), the first one using a mixture of methylene chloride (DCM) e methanol (90:10), the other two ethyl acetate and both mixtures. Soxhlet extraction was also used by Larsen with a mixture of DCM: acetone (80:20). In Harrison and Orasche the sample extraction was carried out in an

3.3. Levoglucosan, sugar polyols and ergosterol concentrations in PM10 samples Since Standard Reference Material 1649a is representative of an urban particulate and it is a time-integrated sample, having been collected over a period in excess of 12 months, we applied the method described to PM10 samples collected in a suburban area, 15 km South-East from the city center of Rome [Latitude 41500 2200 N; Longitude 12 380 5000 E], during four seasonal sampling campaigns: JulyeAugust 2009, December 2009, April and October 2010, for a total of thirty-six filters. Particulate matter with diameter less than 10 mm (PM10) was collected with a dual channel

Fig. 1. Comparison among levoglucosan mass concentrations in SRM 1649a obtained in this and in other studies (Larsen et al., 2006 (a) refers to a sample stored at room temperature; Larsen et al., 2006 (b) to sample freezer stored).

D. Pomata et al. / Atmospheric Environment 84 (2014) 332e338

337

Fig. 2. Comparisons among concentration ratios of target analytes in Rome PM10 samples and in SRM 1649a. (L: levoglucosan; X: xylitol; A: arabitol; M: mannitol; E: ergosterol).

ultrasonic bath with respectively DCM and DCM/methanol (1:1, v/ v). In all the studies reported the derivatization process was carried out using BSTFA containing 1% TMCS. Only Harrison used a second silylation mixture containing 99% of N-methyl-N-trimethylsilyltrifluoroacetamide (MSTFA) and 1% TMCS. The analysis of samples extracted was carried out in GC/MS o GC/MSeMS, except in Orasche study in which it was carried out with IDTD-GC-TOFMS. In the present study, levoglucosan concentrations obtained (173.3  27.7 mg g1 and 160.3  16.2 mg g1) agreed with those reported by Larsen et al. (2006) (162  8 mg g1, freezer-stored samples-(b)), Kuo et al. (2008) (163.9  11.8 mg g1), Orasche et al. (2011) (165.0  0.8 mg g1), and Louchouarn et al. (2009) (160.5  4.7 mg g1) who also analyzed SRM 1649b samples (160.5  5.0 mg g1) (Fig. 1). SRM 1649b is a recertification of SRM 1649a recently performed after sieving through a 63 mm sieve. Higher standard deviations observed in present study can be considered acceptable since the method of standard additions is an extrapolation method, and it is generally less precise than interpolation technique that was used by all the other authors (Miller and Miller, 1993). One discrepancy exists, however, in the study by Larsen et al. (2006), who, in addition to 162  8 mg g1 value obtained for freezer-stored samples, reported an anomalously low value for a sample of SRM 1649a stored at room temperature for 25 years-(a) (81.1  9.4 mg g1). Harrison et al. (2012) also estimated an anomalous lower levoglucosan concentration in samples of SRM 1649b. In order to explain the low value, Larsen hypothesized that levoglucosan might be unstable in solid samples over long periods. However, levoglucosan has been shown to resist both photochemical oxidation and acid catalyzed hydrolysis in atmospheric aerosols (Fraser and Lakshmanan, 2000; Louchouarn et al., 2009), as well as diagenesis during long term preservation in sedimentary deposits spanning over 7000 years (Elias et al., 2001). We agree with Louchouarn et al. in believing that the twofold difference between the cluster of SRM replicates and the lower anomalous value may instead be due to the compromised nature of that specific SRM or to an analytical artifact, such as low efficient extraction. Larsen used ethyl-acetate and DCM/acetone mixture and Louchouarn underlined that high extraction efficiencies of polar compounds, such as anhydrosugars, may only be achieved in the presence of at least a small proportion of a polar solvent (e.g. methanol) to ensure their full extraction from the inner pores of aerosol particles, and retention in the solvent mixture. Conversely, the higher value obtained by Larsen from a low temperature frozen sample (162  8 mg g1) may be due to the low temperature freezing (80  C) providing possible particle pore disruption and thus increasing extraction efficiency even in the absence of a polar solvent.

It is noteworthy that levoglucosan concentration adopted by NIST as reference value in SRM 1649a refers to Larsen value (81.1  9.4 mg g1) and does not take account of any results obtained afterward. Results obtained from 2006 to today suggest that levoglucosan can be reproducibly extracted from particulate matter samples under different laboratory conditions and encourage the potential use of SRM 1649a and SRM 1649b as working standards for levoglucosan analysis with an updated reference value of 163.6  4.6 mg g1, obtained as average between the five comparable studies (see Fig.1), including Larsen et al. (2006) (for the freezer-stored sample). 3.5. Conclusions Levoglucosan and xylitol as tracer of biomass combustion processes and arabitol, mannitol and ergosterol as biomarkers of fungal component of aerosol, were simultaneously determined in SRM 1649a, since so far no standard reference material certifying their concentrations exists. Only levoglucosan concentration is indicated as reference value (81 mg g1), even if many authors suggested a higher concentration in the SRM 1649a. We found concentrations around 160 mg g1 for levoglucosan, 120 mg g1 for xylitol, 15 mg g1 for arabitol, 18 mg g1 for mannitol, and 0.5 mg g1 for ergosterol. Levoglucosan concentration found in the present study is in agreement with the authors that have shown higher concentrations. Moreover, being this standard reference material an urban dust, namely a complex mixture of many constituents, the matrix effect was calculated as the slope ratios between target analyte standard calibration curves and standard addition plots. The latter were built starting from two different amounts of SRM 1649a, one being seven times higher than the other, and demonstrating a matrix effect dependent from lower or higher levels of interferences. We showed standard addition curves for the quantitation of unknown samples were suggested for all the analytes except for arabitol. Accordingly we applied the present method to PM10 samples collected in Rome during four seasonal sampling campaigns and compared the ratios between annual analyte concentrations in PM10 samples with those measured in SRM 1649a. Although particulate matter samples analyzed were collected in two different sites and in so different periods the results were rather similar. Since currently standard reference materials with certified values for all those analytes are not available, levoglucosan, sugar polyols and ergosterol concentrations reported in this study represent an important starting point to future inter-comparison studies.

338

D. Pomata et al. / Atmospheric Environment 84 (2014) 332e338

Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.atmosenv.2013.11.069. References Buiarelli, F., Canepari, S., Di Filippo, P., Perrino, C., Pomata, D., Riccardi, C., Speziale, R., 2013. Extraction and analysis of fungi spore biomarkers in atmospheric bioaerosol by HPLCeMSeMS and GCeMS. Talanta 105, 142e151. Burshtein, N., Lang-Yona, N., Rudich, Y., 2011. Ergosterol, arabitol and mannitol as tracers for biogenic aerosols in the eastern Mediterranean. Atmos. Chem. Phys. 11, 829e839. Carvalho, A., Pio, C., Santos, C., 2003. Water-soluble hydroxylated organic compounds in German and Finnish aerosols. Atmos. Environ. 37, 1775e 1783. Deguillaume, L., Leriche, M., Amato, P., Ariya, P.A., Delort, A.-M., Poschl, U., Chaumerliac, N., Bauer, H., Flossmann, A.I., Morris, C.E., 2008. Microbiology and atmospheric processes: chemical interactions of primary biological aerosols. Biogeosciences 5, 1073e1084. Di Filippo, P., Riccardi, C., Pomata, D., Buiarelli, F., 2010. Concentrations of PAHs, and nitro- and methyl-derivatives associated with a size-segregated urban aerosol. Atmos. Environ. 44, 2742e2749. Di Filippo, P., Pomata, D., Riccardi, C., Buiarelli, F., Perrino, C., 2013. Fungal contribution to size-segregated aerosol measured through biomarkers. Atmos. Environ. 64, 132e140. Elias, V.O., Simoneit, B.R.T., Cordeiro, R.C., Turcq, B., 2001. Evaluating levoglucosan as an indicator of biomass burning in Carajas, Amazonia: a comparison to the charcoal record. Geochim. Cosmochim. Acta 65 (2), 267e272. Fraser, M.P., Lakshmanan, K., 2000. Using levoglucosan as a molecular marker for the long-range transport of biomass combustion aerosols. Environ. Sci. Technol. 34 (21), 4560e4564. Gao, S., Hegg, D.A., Hobbs, P.V., Kirchstetter, T.W., Magi, B.I., Sadilek, M., 2003. Water-soluble organic components in aerosols associated with savanna fires in southern Africa: identification, evolution, and distribution. J. Geophys. Res. 108, 8491. Graham, B., Guyon, P., Taylor, P.E., Artaxo, P., Maenhaut, W., Glovsky, M.M., Flagan, R.C., Andreae, O., 2003. Organic compounds present in the natural Amazonian aerosol: characterization by gas chromatography mass spectrometry. J. Geophys. Res. 108 (D24), 4766. Harrison, R.M., Beddows, D.C.S., Hu, L., Yin, J., 2012. Comparison of methods for evaluation of wood smoke and estimation of UK ambient concentrations. Atmos. Chem. Phys. 12, 8271e8283. Kuo, L.J., Herbert, B.E., Louchouarn, P., 2008. Can levoglucosan be used to characterize and quantify char/charcoal black carbon in environmental media? Org. Geochem. 39, 1466e1478. Larsen, R.K., Schantz, M.M., Wise, S.A., 2006. Determination of levoglucosan in particulate matter reference materials. Aerosol Sci. Technol. 40, 781e787. Lau, A.P.S., Lee, A.K.Y., Chan, C.K., Ming, F., 2006. Ergosterol as a biomarker for the quantification of the fungal biomass in atmospheric aerosols. Atmos. Environ. 40, 249e259. Lee, C., Howarth, R.W., Howes, B.L., 1980. Sterols in decomposing Spartina alterniflora and the use of ergosterol in estimating the contribution of fungi to detrital nitrogen. Limnol. Oceanogr. 25, 290e303. Lee, A.K.Y., Lau, A.P.S., Cheng, J.Y.W., Fang, M., Chan, C.K., 2007. Source identification analysis for the airborne bacteria and fungi using a biomarker approach. Atmos. Environ. 41, 2831e2843. Louchouarn, P., Kuo, L.J., Wade, T.L., Schantz, M., 2009. Determination of levoglucosan and its isomers in size fractions of aerosol standard reference materials. Atmos. Environ. 43, 5630e5636.

Miller, J.C., Miller, J.N., 1993. Errors in instrumental analysis; regression and correlation (Ch.5; Sec-5.8). In: Statistics for Analytical Chemistry, third ed. Ellis Horwood Limited, Chichester, England, pp. 117e120. Miller, J.D., Young, J.C., 1997. The use of ergosterol to measure exposure to fungal propagules in indoor air. Am. Ind. Hyg. Assoc. J. 58, 39e43. Mille-Lindblom, C., Wachenfeld, E., Tranvik, L.J., 2004. Ergosterol as a measure of living fungal biomass: persistence in environmental samples after fungal death. J. Microbiol. Methods 59, 253e262. Orasche, J., Schnelle-Kreis, J., Abbaszade, G., Zimmermann, R., 2011. Technical Note: in-situ derivatization thermal desorption GC-TOFMS for direct analysis of particle-bound non-polar and polar organic species. Atmos. Chem. Phys. 11, 8977e8993. Pio, C.A., Legrand, M., Alves, C.A., Oliveira, T., Afonso, J., Caseiro, A., Puxbaum, H., Sanchez-Ochoa, A., Gelencser, A., 2008. Chemical composition of atmospheric aerosols during the 2003 summer intense forest fire period. Atmos. Environ. 42, 7530e7543. Piot, C., Jaffrezo, J.L., Cozic, J., Pissot, N., El Haddad, I., Marchand, N., Besombes, J.L., 2011. Quantification of levoglucosan and its isomers by High Performance Liquid ChromatographyeElectrospray Ionization tandem Mass Spectrometry and its applications to atmospheric and soil samples. Atmos. Meas. Tech. Discuss. 4, 4539e4560. Pizzutti, I.R., de Kok, A., Cardoso, C.D., Reichert, B., de Kroon, M., Wind, W., Righi, L.W., da Silva, R.C., 2012. A multi-residue method for pesticides analysis in green coffee beans using gas chromatographyenegative chemical ionization mass spectrometry in selective ion monitoring mode. J. Chromatogr. A. 1251, 16e26. Poole, C.F., 2007. Matrix-induced response enhancement in pesticide residue analysis by gas chromatography. J. Chromatogr. A. 1158, 241e250. Schkolnik, G., Rudich, Y., 2006. Detection and quantification of levoglucosan in atmospheric aerosols: a review. Anal. Bioanal. Chem. 385, 26e33. Simoneit, B.R.T., Schauer, J.J., Nolte, C.G., Oros, D.R., Elias, V.O., Fraser, M.P., Rogge, W.F., Cass, G.R., 1999. Levoglucosan, a tracer for cellulose in biomass burning and atmospheric particles. Atmos. Environ. 33, 173e182. Simoneit, B.R.T., Kobayashi, M., Mochida, M., Kawamura, K., Lee, M., Lim, H.J., Turpin, B.J., Komazaki, Y., 2004a. Composition and major sources of organic compounds of aerosol particulate matter sampled during the ACE-Asia campaign. J. Geophys. Res. 109, D19S10. Simoneit, B.R.T., Elias, V.O., Kobayashi, M., Kawamura, K., Rushdi, A., Medeiros, P.M., Rogge, W.F., Didik, B.M., 2004b. Sugars-Dominant water-soluble organic compounds in soils and characterization as tracers in atmospheric particulate matter. Environ. Sci. Technol. 38, 5939e5949. Simpson, C.D., Dills, R.L., Katz, B.S., Kalman, D.A., 2004. Determination of levoglucosan in atmospheric fine particulate matter. J. Air Waste Manage. Assoc. 54, 689e694. Srikanth, P., Sudharsanam, S., Steinberg, R., 2008. Bio-aerosols in indoor environment: composition, health effects and analysis. Indian J. Med. Microbiol. 26 (4), 302e312. Vélëz, H., Glassbrook, N.J., Daub, M.E., 2007. Mannitol metabolism in the phytopathogenic fungus Alternaria alternata. Fungal Genet. Biol. 44, 258e268. Volker, E.J., Di Lella, D., Terneus, K., Baldwin, C., Volker, I., 2000. The determination of ergosterol in environmental samples. J. Chem. Educ. 77 (12). Womiloju, T.O., Miller, J.D., Mayer, P.M., Brook, J.R., 2003. Methods to determine the biological composition of particulate matter collected from outdoor air. Atmos. Environ. 37, 4335e4344. Zhang, T., Engling, G., Chan, C.Y., Zhang, Y.N., Zhang, Z.S., Lin, M., Sang, X.F., Li, Y.D., Li, Y.S., 2010. Contribution of fungal spores to particulate matter in a tropical rainforest. Environ. Res. Lett. 5, 024010 (9pp). Zhao, Z., Sebastian, A., Larsson, L., Wang, Z., Zhang, Z., Norback, D., 2008. Asthmatic symptoms among pupils in relation to microbial dust exposure in schools in Taiyuan. China. Pediatr. Allergy Immunol. 19, 455e465.