Quantitation of polycyclic aromatic hydrocarbons in ... - ACS Publications

---

2129

Anal. Chem. 1984, 56, 2129-2134

Quantitation of Polycyclic Aromatic Hydrocarbons in Diesel Exhaust Particulate Matter by High-Performance Liquid Chromatography Fractionation and High-Resolution Gas Chromatography H.Y.Tong a n d F. W.Karasek* Department of Chemistry, University of Waterloo, Waterloo, Ontario, Canada N2L 3Gl

Seventy-slx components in the extracts of three different dlesel exhaust particulate samples were Identifled by GC/MS and quantlfled uslng WCOT column GC wlth F I D after they had been Isolated from other compounds by HPLC fractlonatlon. A slngie characterlstlc F I D response factor obtained from the average of response factors of 22 PAH compound standards wlth an RSD of 5.7% was used to quantify PAH components In the samples. The quantltathm results obtained show hlgh reproduclbUtty, prectdon, and good reliablltty. Data obtalned for PAH components in these three samples ranged from 30 to 8000 ng/mg of extract.

Numerous polycyclic aromatic hydrocarbons (PAH) have been linked to mutagenic and carcinogenic hazards (1, 2). Those known and potential health hazards enter the human environment in an increasing amount from industrial waste, combustion products, and automotive exhaust (3). A number of studies of PAH compounds in environmental samples have identified many PAH (4-9),but studies of PAH compounds associated with environmental toxicology and experimental carcinogenesis require accurate quantitation of PAH compounds. So far, limited success has been reported in this work. The environmental samples are very complex so that extensive quantitation of PAH usually involves the analysis of an extremely complex mixture of organic compounds. One of the prime concerns in these difficult analyses is to isolate PAH from the complex matrix to minimize the interferences present. Different procedures have been applied for this purpose. Among the methods used for separation, high efficiency, good recovery, and reproducibility are more easily obtained with high-performance liquid chromatography (HPLC). By use of the HPLC separation, PAH compounds have been isolated from the sample matrix of airborne particulate matter, diesel exhaust particulate, and fly ash emitted from municipal incinerators (10-12). Another major factor limiting the development of extensive quantitation of PAH is the heavy demand on the available reference standard compounds. Usually compounds in a sample can be precisely quantified only when the standard compounds are available. There are numerous PAH compounds and their isomers present in environmental samples, but the number of the available reference standard compounds is inadequate. For the quantitative analysis of PAH, GC with flame ionization detection (FID) is one of the most widely used techniques. In past years, some investigations have been carried out in the study of FID responses of PAH compounds. Disagreements among the different literature studies reported have led to the conclusion that an individual standard compound must be used for accurate quantitation (13). Lao determined the FID response factor of a number of PAH standard compounds on packed column GC with vaporization 0003-2700/84/0356-2129$01.50/0

injection (14). His results indicated that the FID response factor, in terms of response per unit weight of chemical between two-ring PAH and six-ring PAH compounds, differs by 100%. However, in the past, the accuracy and precision of the FID response factor determination have been greatly limited by GC performance. The use of WCOT GC columns, combined with on-column injection, has greatly improved the quality of GC analysis. It has recently been shown that accurate and precise determination of FID response factors for a number of PAH standard was possible (15). The results obtained indicated that PAH and their alkyl-substituted derivatives exhibit very similar FID responses and a single characteristic response factor with a small standard deviation can be used for the quantitation of PAH compounds on GC using WCOT columns and a flame ionization detector. Therefore, the GC analysis of the PAH fraction from an HPLC separation may be one of the most feasible procedures for the accurate quantitation of PAH compounds in a complex mixture. In this study, this quantitation procedure was applied to the extracts of three different diesel exhaust particulate samples. A number of organic components in the extracts were separated into different compound classes. Components in PAH fractions were identified or tentatively identified with GC/MS and their characteristic retention indexes and then quantified on high-resolution GC/FID with on-column injection. A single characteristic FID response factor with a small relative standard deviation of 5.7% was obtained from multiple injections of 22 PAH standards and used for the quantitation of other PAH components. The reliability of results obtained from GC analysis was checked by GC/MS with selected ion monitoring (GC/MS/SIM) technique. This HPLC-GC/FID quantitation procedure shows good reproducibility.

EXPERIMENTAL SECTION Solvents and Chemicals. All solvents used were "distilled in glass", UV grade (Caledon Laboratories, Ltd., Toronto, Ontario, Canada). PAH standards were purchased from either Aldrich Chemical Co. (Montreal, Quebec, Canada) or Chem. Service, Inc., (West Chester, PA) and their purities are 95-99%. After being cleaned by ultrasonic agitation in detergent, rinsed with deionized water, and dried at 250 "C for 3 h, all glassware was rinsed several times with benzene and methylene chloride immediately before its use. Sample Collection and Extraction. Three dichloromethane (CHZCI,) extracts of diesel particulate matter collected from in-use diesel automobiles (Volkswagen) were received solvent free from the New York State Department of Environmental Conservation. These samples came from different testing vehicles, driving cycles, and fuel and lubricant combinations. The extracts were labeled as VW-1, VW-2, and VW-3, and stored in the freezer at -15 O C . The values of soluble organic fraction (w/w) of particulate matter for these three diesel particulate samples were VW-1 11.1%, VW-2 14.7%, and VW-3 12.1%, respectively. 0 1984 American Chemical Society

2130

ANALYTICAL CHEMISTRY, VOL. 56, NO. 12, OCTOBER 1984

Table I. Compounds Identified and Tentatively Identified in HPLC Fraction 2 of Diesel Particulate Extract

no. 1 2

3 4 5 6 7 8 9 10 11 12

13 14 15 16 17 18

19 20 21 22

23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70

71 72 73 74

compound acenaphthylene trimethylnaphthalene trimethylnaphthalene trimethylnaphthalene trimethylnaphthalene fluorene dimethylbiphenyl C4-naphthalene (&-naphthalene C4-naphthalene trimethylbiphenyl dibenzothiophene phenanthrene anthracene methyldibenzothiophene methyldibenzothiophene methyldibenzothiophene 3-methylphenanthrene 2-methylphenanthrene 2-methylanthracene 4H-~yclopenta[deflphenanthrene ethyldibenzothiophene 2-phenylnaphthalene 9- or 2-ethylphenanthrene dimethylphenanthrene dimethylphenanthrene dimethyl(phenanthrene/anthracene) dimethyl(phenanthrene/anthracene) fluoranthene benzo[defldibenzothiophene benzacenaphthylene pyrene ethylmethyl(phenanthene/anthracene) ethylmethyl(phenanthrene/anthracene) methyl(fluoranthene/pyrene) methyl(fluoranthene/pyrene) methyl(pyrene/fluoranthene) methyl(pyrene/fluoranthene)

benzo[a]fluorene benzo[bJfluorene 1-methylpyrene methyl-substituted PAH benzo[b]naphtho[ 2,l-d]thiophene cyclopentapyrene benzo[ghi]fluoranthene benzonaphthiophene benz [a]anthracene chrysene or triphenylene phenyl(phenanthrene/anthracene) 1,2-binaphthyl 9-phenylphenanthrene methylbenz[a]anthracene 3-methylchrysene 1-phenylphenanthrene 2,2-binaphthyl phenyl(anthracene/phenanthrene) phenyl(anthracene/ phenanthrene) unknown PAH unknown PAH“ benzofilfluoranthene benzo[blfluoranthene benzo[k]fluoranthene benzo[e]pyrene benzo[a]pyrene benz[a,h]anthracene unknown PAH indeno(1,2,3-cd)pyrene unknown PA” unknown PAHf unknown PAHf unknown PAHf benzo[ghi]perylene unknown PAHg dibenzopyrene or dibenzo[def,p]chrysene

mol wt retention time, min 152 170 170 170 170 166 182 184 184 184 196 184 178 178 198 198 198 192 192 192 190 212

204 206 206 206 206 206 202 208 202 202 220 220 216 216 216 216 216 216 216 242 234 226 226 234 228 228 254 254 254 242 242 254 254 254 254 250 278 252 252 252 252 252 278 264 276 276 276 276 276 276 288 302

17.40 20.30 20.60 21.30 21.90 22.46 23.63 25.90 26.80 27.08 28.30 28.65 29.75 29.99 32.24 32.62 32.97 33.75 33.92 34.41 34.72 35.73 36.40 37.49 37.81 37.90 38.60 38.85 39.25 39.53 39.91 40.87 41.97 42.16 42.93 43.51 43.63 43.77 44.39 45.12 45.32 47.30 48.50 48.80 48.87 49.70 50.54 50.86 51.60 51.80 52.20 52.80 53.97 54.71 54.97 55.47 55.80 56.40

concn,h ng/mg of ex:tract vw-2 vw-3

vw-1 30 50

30 50

50

50

50

30 50

50 168 83 50 149 152 50 246 4883 356 323 205 244 1287 1481 1522 1033 159 1336 464 524 548 140 86 7321 262 1643 8002 317 400 552 233 467 717 990 538 443 50 53 1671 418 30 1076 1529 93 50 94 50 192 163 283 116 93 95

57.50

30

58.71 58.80 59.35 60.43 60.72 62.50 64.04 66.20 66.80 67.37 67.47 67.91 69.30 72.50 75.38

1367 1098 289 946 558 96 176 93 94 619 187 989 1050 93 254

100

30 50

89 146 50

129 2186 224 231 101

188 929 1099 517 517 151 650 422 494 443 585 315 3399 254 820 3652 304 286 252 244 288 567 541 216 197 30 30 869 217

463 657 50 30 30 30 50 89 89 50 30 30 30 492 421 163 510 270 50

89 30 30

166 89 397 443 50 169

30 30 30 50 166 91 50 97 109 50 172 2821 155 294 154 280 970 1142 782 650 179 889 388 523 1046 482 357 3748 333 791 3532 289 432 341 224 420 570 610 175 144 91 52 914 227 126 495 873 30 30 30 30 50 90 91 30 30 30 30 680 593 91 487 208 50 136 50 50 266 136 553 661 50

136

ANALYTICAL CHEMISTRY, VOL. 56, NO. 12, OCTOBER 1984

75 76

dibenzopyrene or dibenzo[def,p]chrysene coronene

302 300

76.16 77.47

a a, d

171 521

89

301

2131

90 307

OIdentified by sample mass spectra. *Identified by retention index published in ref 18 and 24. CIdentifiedby standard injected into GC/MS and GC. dCan be found in ref 8, 11, 22, and 23. eTentatively identified as benzo[b]chrysene. /The possible compounds are ideno [ 1,7-ab]pyrene, ideno[1,7,6,5-cdeflchrysene, ideno [ 5,6,7,l-defg]chrysene, benzo [e]cyclopenpk]pyrene, cyclopenta[cd]perylene, anthanthrene. g Tentatively identified as 1,12-methylenebenzo[ghi]perylenein ref 23. Note: concentrations of less than 50 ng/mg extract were obtained by approximate calculation.

HPLC Separation. Each dry sample received was redissolved in a mixture of CH2C12and acetone (31)to give a concentration of ca. 20 mg of extract/mL. The instrument used for the HPLC separation was a Spectra-Physics SP-8000 HPLC equipped with SP-8400 UV/vis detector and an SP-4100 integrator. The monitoring wavelength was 254 mm. A lO-rrn, semipreparative Spherisorbsilica column (25 cm X 9.4 mm i.d., Terochem, Toronto, Ontario, Canada) was employed with a 140-pL sample loop. A modified gradient elution program was used (16). It consisted of 100% n-C&14 for 20 min, programmed to 100% CHzClz over 30 min and held at 100% CH2C12for 20 min, programmed to 100% CH,CN over 10 rnin and held at 100% CH,CN for 1 min, programmed back to 100% CH,C12 in 5 min, and finally to 100% n-C6H1, in another 5 min. During the gradient program, six separate fractions were collected at elution times of 0 min to the start of first peak and at 20,40, 50, 70, and 91 min. They were designated as fraction 1 to fraction 6. The flow rate was 5 mL/min. To avoid column overload and obtain enough sample material for the GC and GC/MS analysis, each sample was fractionated four times and the correspondingfractions were combined. Each combined fraction was evaporated to dryness by rotary evaporation under aspirator vacuum and by blowing a gentle stream of high-puritynitrogen over the sample. The residuals were finally dissolved in a mixture of CHzClzand benzene (70%) to give an accurate volume of 200 MLfor fractions 1 , 2 , and 6, and 100 pL for the other fractions. The fractions were stored in the freezer at -15 "C. A blank of HPLC separation procedure was obtained in the same manner as above but with injection of benzene instead of sample. Six fractions of the benzene blank run were used to test for impurities in HPLC separation procedure. The sample VW-3 was subjected to the HPLC separation procedure described above in duplicate, as a test of the reproducibility of the method used. Fraction 2 collected from this HPLC separation procedure was subjected to PAH analysis. WCOT Gas Chromatographic/Mass Spectrometric Analysis. The GC/MS analyses were performed on two GC/ MS/DS systems. Most of GC/MS analyses were done on a Hewlett-Packard HP5992 GC/MS/calculator. Two bonded phase fused silica capillary columns, a 50 m X 0.32 mm i.d. SE-54 cross-linked column (Hewlett-Packard Co., Avondale, PA) and a 30 m X 0.32 mm i.d. DB-5 (J&W Scientific Inc. Rancho Cardova, CA), were used. The GC conditions were as follows: on-column injection a t less than 50 "C; column temperature held at 80 "C for 1 min and then programmed to 275 "C or 300 "C a t a rate of 4 "C/min depending on the column used. The helium carrier gas rate was 3 mL/min at rcom temperature. An HP59916A glass capillary effluent splitter interfacing the column to the MS allowed a flow of approximately 0.5 mL/min to enter the MS analyzer. For the compound identification the mass spectra search, probability based matching (PBM) and self-training interpretive retrieval system (STIRS) were used by a stand-alone terminal phone-linked to Cornel1 University. Also, a Hewlett-Packard HP5987A GC/MS system with a HPlOOO data system and HP7914 Winchester disk drive was used for the confirmation of compound identification. This GC/MS system has a PBM and STIRS libarary search system based on over 70000 reference compounds. Also, an open split GC/MS interface, cool on-column injector, and electron impact ionization with 70 eV were used for HP5987A GC/MS in this study. A 30 m X 0.32 mm i.d. DB-5 fused silica column and the same GC conditions described above were used. WCOT Gas Chromatographic Analysis. GC analyses were done on a Hewlett-Packard HP5880A gas chromatograph equipped with flame ionization detection (FID) and cool oncolumn injector for capillary column. A 30 m X 0.32 mm i.d. DB-5

fused silica capillary column (J&W Scientific Inc., Rancho Cardova, CA) was used. The GC conditions were similar to those used in GC/MS analysis with the exception of a temperature programming rate of 3 OC/min, The detector temperature was 350 "C and the helium carrier gas flow rate was 3 mL/min measured at room temperature. A user-developedsoftware is stored in the HP5880A terminal to calculate PAH retention indexes based on the reference system of Lee using naphthalene, phenanthrene, chrysene, and picene as standards (12, 17,18). A modified injection technique called "four segment injection" was used for on-column injection of small sample volumes on the WCOT column at room temperature in both GC and GC/MS analyses (19). Qualitative Analysis. Compound identification in GC/MS analysis was achieved by matching the mass spectra of sample components to those of references. The data of reference mass spectra were obtained from the standard compounds injected, publications (8,11,22,23),and the atlas of mass spectra (20,21), the computer library search of the two GC/MS systems. The PAH retention index data obtained from GC analysis was also used to facilitate identification of compounds. Retention index data using the PAH reference system for more than 200 standards of PAH and their derivatives were published (18,24). In this study GC retention index data used for qualitative analysis were based on triplicate injections. Quantitative Analysis. Since similar column and chromatographic conditions were used in both GC and GC/MS analyses, the GC/FID trace and totalion current (TIC) trace of each sample in both analyses were qualitatively similar and the corresponding peaks in each trace are easily located. The basic data for component quantification in this study are the integrated peak areas on the chromatogram obtained from the HP5880A GC with cool on-column injection. Triplicate injections were made for each sample and the averaged peak area was used for quantitation. Twenty-two of PAH standard compounds were multiinjected to determine the unique characteristic FID response factor for PAH compounds. Some selected components in diesel particulate extract were also quantified by GC/MS/SIM analysis. The integrated peak areas of their characteristic ions were used to compare with that of an individual external standard compound injected under the same condition.

RESULTS AND DISCUSSION The HPLC separation procedure described in the Experimental Section was designed to separate PAH compounds in diesel exhaust particulate extract from other compound classes. The type of major compounds found in each of six fractions were aliphatic hydrocarbon (fraction l),PAH and sulfur-containing PAH (fraction 2), oxygenated PAH, nitrogen-containing PAH, and nitro-PAH (fraction 3), small amounts of oxygenated PAH and phthalate (fraction 4 and 5), and oillike compounds (fraction 6). PAH and their alkyl-substituted derivatives were separated into fraction 2. They were predominant in this fraction. In addition, eight sulfur-containing PAH (S-PAH) were identified in this fraction. Compound identification was primarily based on the GS/MS data as described in the Experimental Section. Numerous isomers of PAH and their alkyl-substituted derivatives are present in the samples. Many of these isomers have very similar mass spectra, making it impossible t o differentiate them by their mass spectra alone. In these cases the PAH retention index (RI) suggested by Lee was used to

2132

ANALYTICAL CHEMISTRY, VOL. 56, NO. 12, OCTOBER 1984

I50 r

20

200

,

250

3 0 0 I@othrrnrl TEMP('C)

I

30

40

50

60

80

10

TIME(MIN)

Figure 1. Gas chromatogram of PAH fraction (fraction 2) of diesel exhaust particulate extract. HC designates aliphatic hydrocarbon compound.

Table 11. FID Response Factors of PAH Standards'

compound naphthalene 2-methylnaphthalene 2,6-dimethylnaphthalene 2,3-dimethylnaphthalene acenaphthene fluorene 9,lO-dihydroanthracene phenanthrene anthracene 2-methylanthracene 9-methylanthracene 1,2,6,7-tetrahydropyrene

fluoranthene pyrene 1,2-benzofluorene 1,l-binaphthyl chrysene triphenylene 7,12-dimethylbenz[a]anthracene benzo[e ]pyrene benzo[a ]pyrene dibenzanthracene

mol wt

RF, area counts/ ng

128 142 156 156 154 166 180 178 178 192 192 206 202 202 216 254 228 228 256 252 252 278

14.02 13.75 13.62 13.63 14.16 13.66 13.60 13.77 13.85 13.41 14.28 11.86 13.51 13.72 12.60 13.92 12.88 14.09 12.86 12.80 14.81 11.54

Table 111. Comparison in Quantification of Some Selected Compounds in Sample VW-1 Using Different Techniques RSD,b % 4.6 4.2 4.2 0.7 4.3 4.2 4.2 4.3 1.7 3.8 4.4 7.7 4.0 4.0 2.0 3.6 3.8 1.2 1.3

4.2 8.6 4.2

OAverage RF of 22 PAH standards, 13.47; relative standard deviation, 5.7%. bRelative standard deviation based on a total of four injections at two concentration levels (two injections at the 100 ng level and two iniections at the 20 ng level). facilitate compound identification (18). Many of the isomers do have different retention indexes. For example, fluoranthene and pyrene have identical mass spectra but have retention indexes of 344.98and 352.66, respectively, which were determined in fraction 2 by WCOT GC. The published retention indexes for fluoranthene and pyrene standards are 344.01 and 351.22, respectively (18). On the basis of both the mass spectra and RI values, the earlier eluate was identified as fluoranthene and the latter as pyrene. On combination of the data of mass spectra and retention index, many isomers in fraction 2, such as phenanthrene and anthracene, 3methylanthracene, benzofluoranthene isomers, benzo[e]pyrene, and benzo[a]pyrene were distinguishable. The identification of some compounds in the samples was confirmed by the injection of standard compounds. The data obtained from related literature were also used as auxiliary information for compound identification. Table I lists the compounds identified or tentatively identified in the PAH fractions of three diesel particulate extracts. The methods used for compound identification are also listed in this table.

compound fluorene phenanthrene methylanthracene methylphenanthreneb fluoranthrene pyrene dibenzothiophene

analysis results, ng/wL of test solution ( % RSD)" GC/FID GC/MS/SIM 9 (7) 259 (3) 147 (4) 81 (10) 387 (1) 423 (2) 13 (4)

13 (23) 283 (20) 163 (27) 104 (17) 402 (16) 419 (12) 12 (39)

a Relative standard deviation based on three determinations. Summarv of two isomers.

Table IV. Reproducibility of the Quantitative Analysis of Selected Compounds in Sample VW-3 Using HPLC-GC/FID

compound

overall result av, retention ng/mg of time, min extract RSD," 7'0 28.65 29.99 33.92 34.72

173 143 1140 685

4.6 18.6 5.2 7.7

dimethyl(phenanthrene/anthr-

35.73 38.60

189 463

11.9 6.8

acene) fluoranthene pyrene methyl(fluoranthene/ pyrene) benzo [a]fluorene benz[a]anthracene benzo[e ]pyrene benzo[a]pyrene unknown PAH with mol wt 276 unknown PAH with mol wt 276 benzo[ghi] perylene

39.25 40.87 42.93 44.39 50.54 60.43 60.72 67.37 67.91 69.30

3939 3429 341 553 477 487 204 266 557 668

5.5 9.7 17.5 12.9 7.1 4.4 3.8 15.7 4.4 2.8

dibenzothiophene anthracene 2-methylphenanthrene 4H-cyclopenta[deflphenanthrene ethyldibenzothiophene

"Based on a total of five GC determinations for HPLC run 1 and run 2 (three injections for HPLC run 1 and two injections for HPLC run 2). The identification procedures used in this study provide an identification with varying degrees of certainty. The only way to provide a more certain identification is to directly obtain the mass spectrum and retention behavior of pure standards for each compound. This is possible for some compounds but is not possible for most other compounds treated here.

ANALYTICAL CHEMISTRY, VOL. 56, NO. 12, OCTOBER 1984

Some components showed typical mass spectra of PAH compounds but could not be identified. These compounds were labeled as unknown PAH. The compound behavior on HPLC provided additional information for the identification of compound class. T h e component peaks shown on the chromatogram of fraction 2 in Figure 1 are reasonably separated after reducing the complexity of the mixture by HPLC fractionation. It is possible to quantify components based on their GC peaks area. In recent work, it was indicated that PAH and their alkylsubstituted derivatives have very similar behavior in FID response and it is possible to use a unique FID response factor, in terms of area counts/ng, to quantify all PAH which are suitable for GC analysis (15). Table I1 lists the response factors determined for 22 PAH standards. An average FID response factor obtained from multiple injections of 22 PAH standards was used to quantify all the PAH and their alkyl-substituted derivatives in the PAH fractions of diesel samples. The relative standard deviation (RSD) of the average FID response factor is 5.7 % and is considerably smaller than many experimental errors commonly involved in trace analysis of complex organic mixtures. The FID response factor of dibenzothiophene was determined to be 11.84 area counts/ng with a relative standard deviation of 1.2% based on four injections. I n this study it was assumed that the FID response factors of dibenzothiphene and its derivatives are quite similar. Therefore, the response factor of dibenzothiophene was used for the quantitation of dibenzothiophene and its derivatives. The PAH fraction of each diesel sample was injected in triplicate with cool on-column injection. T h e averaged peak area of each component was used for quantitative analysis. Good reproducibility was obtained in triplicate injections. The quantitative results of PAH, their alkyl-substituted derivatives, and some S-PAH compounds in three diesel particulate extracts are listed in Table I. T o check the reliability of the results of GC analyses, a few components in fraction 2 of VW-1 were selected to be quantified by GC/MS/SIM using individual external standards. The compounds selected in the sample cover a wide concentration range. Better selectivity of analysis is obtained by GC/MS/SIM, because there is less background interference than in GC. Table I11 compares the quantitative results and precision obtained by each method. Considering the complexity of the samples studied, the two methods give consistent results, but GC analysis showed better precision. T h e reproducibility of the HPLC separation and GC quantitation was shown by two HPLC separations of sample VW-3, which were described in the Experimental Section. These two HPLC separations are labeled HPLC run 1 and HPLC run 2. Triplicate and duplicate GC quantitative analyses were performed on HPLC run 1 and HPLC run 2, respectively. The results for some selected compounds covering a large concentration and boiling point range are listed in Table IV. The RSD of 2.8 to 18.6% in this table demonstrate the reproducibility in this quantitation procedure. On-column injection combined with the "four segment injection" technique is important to obtain these results. A study in the recovery of HPLC fractionation step showed a recovery range of 81% to 100% for some selected standards including PAH, oxy-PAH, and nitro-PAH compounds (25). No significant impurities were found in fraction 2 of the HPLC blank run. The data in Table I show that the distribution of PAH components in these three diesel particulate samples is qualitatively similar, but significantly different in quantity. Additionally, among the PAH components shown in Table I, PAH compounds with three and four rings and their alkyl-substituted derivatives are the most abundant.

2133

CONCLUSION Using the combined techniques of HPLC separationGC/MS identification-GC/FID quantitation, the present results demonstrate that accurate, precise, and extensive quantitation of PAH in a complex mixture can be achieved using a characteristic FID response factor. Applying a single, characteristic FID response factor for the quantitation of all PAH and their alkyl-substituted derivatives simplifies these analyses, reduces the heavy demand on the number of pure standards, and still produces quantitative results with good accuracy and precision.

ACKNOWLEDGMENT This work was supported by the National Research Council of Canada. The valuable assistance of R. E. Gibbs and J. D. Hyde of the New York State Department of Environmental Conservation, who supplied the diesel particulate extracts and the related data, is gratefully acknowledged. Registry No. Acenaphthylene, 208-96-8;trimethylnaphthalene, 28652-77-9; fluorene, 86-73-7; dimethylbiphenyl, 28013-11-8; trimethylbiphenyl, 30581-97-6; dibenzothiophene, 132-65-0; phenanthrene, 85-01-8; anthracene, 120-12-7;methyldibenzothiophene, 30995-64-3; 3-methylphenanthrene, 832-71-3; 2methylphenanthrene, 2531-84-2; 2-methylanthracene, 613-12-7; 4H-~yclopenta[def]phenanthrene, 203-64-5; ethyldibenzothiophene, 79313-22-7; 2-phenylnaphthalene, 612-94-2; ethylphenanthrene, 30997-38-7; dimethylphenanthrene, 29062-98-4; fluoranthene, 206-44-0;benzo[defldibenzothiophene,30796-92-0; benzacenaphthylene, 76774-50-0; pyrene, 129-00-0; benzo[a]fluorene, 30777-18-5; benzo[b]fluorene, 30777-19-6; l-methyl239-35-0; pyrene, 2381-21-7; benzo[b]naphtho[2,1-d]thiophene, cyclopentapyrene,83381-96-8;benzo[ghi]fluoranthene,203-12-3; benzonaphthiophene, 61523-34-0; benz[a]anthracene, 56-55-3; 1,2-binaphthyl, 4325-74-0; 9-phenylphenanthrene, 844-20-2; methylbenz[a]anthracene, 43178-22-9; 3-methylchrysene, 3351-31-3; 1-phenylphenanthrene, 4325-76-2; 2,2-binaphthyl, 612-78-2; benzolj]fluoranthene, 205-82-3; benzo[b] fluoranthene, 205-99-2; benzo[k]fluoranthene, 207-08-9;benzo[e]pyrene, 192-97-2;benzo[a]pyrene, 50-32-8; dibenz[a,h]anthracene, 53-70-3; indeno(1,2,3-cd)pyrene,193-39-5;benzo[ghi]perylene,191-24-2;coronene, 191-07-1.

LITERATURE CITED Freudenthal, R. I., Jones, P. W., Eds. "Polynuclear Aromatic Hydrocarbons: Chemistry, Metabolism and Carcinogenesis"; Raven: New York, 1976; Vol. 1.. Gelboin, H. V., Ts'o, P. 0. P., Eds. "Polycycllc Hydrocarbons and Cancer"; Academic Press: New York, 1978. Siebert, P. C.; Craig, C. A.; Coffey, E. 8. "Preliminary Assessment of the Sources, Control and Population Exposure to Airborne Polycyclic Organic Matter (POM) as Indicated by Benzo(a)pyrene (BaP)"; Final Report, EPA Contract No. 68-02-2836, Environmental Protection Agency, 1978. Lee, M. L.; Novotry, M.; Bartle, K. D. Anal. Chem. 1978, 48, 1566. Harrison, R . M.; Perry, R.; Welllngs, R. A. Water Res. 1975, 9, 331. Giger, W.; Blumer, M. Anal. Chem. 1974, 46, 1663. Howard, J. W.; Fazlo, T. J . Agrlc. Food Chem. 1989, 17,527. Yu, M.-L.; Hltes, R . A. Anal. Chem. 1981, 53,951. Eiceman, G. A.; Clement, R. E.; Karasek, F. W. Anal. Chem. 1979, 51,2344. Choudhury, D. R.; Bush, B. Anal. Chem. 1981, 53, 1351. Schuetzle, D.; Lee, F. S.-CI; Prater, T. J. I n t . J . Envlron. Anal. Chem. 1981, 9 , 9 3 . Tong, H. Y.; Shore, D. L.; Karasek, F. W.; Heliand. P.; Jelium, P. J. Chromatogr. 1984, 285, 423. Lee, M. L.; Novotny, M. V.; Bartle, K. D. "Analytical Chemistry of Polycyclic Aromatic Compounds"; Academic Press: Toronto, 1981. Lao, R. C.; Thomas, R. S.;Oja, H.; Dubois, L. Anal. Chem. 1973, 45, 908. Tong, H. Y.; Karasek, F. W. Anal. Chem. 1984, 56, 2124. Karasek, F. W.; Sweetman, J. A.; Clement, R . E.; Vance, F. P.; Tong, H. Y. "Chemical Characterizatlon of Organic Compounds Associated with Diesel Engine Particulate Emissions"; Contract Report to Ontario Ministry of Environment, Aprll 1982. Van den Dool, H.; Kratz, P. D. J. Chromatogr. 1963, I f , 463. Lee, M. L.; Vassllarus, D. L. Anal. Chem. 1979, 51,768. Tong, H. Y.; Sweetman, J. A,; Karasek, F. W. J. Chromatogr. 1983, 264. 231. EPAiNIH Mass Spectral Data Base, U S . Government Printing Office, Washington. DC. 1978. "Eight Peak Index of Mass Spectra", 2nd ed.; Mass Spectrometry Data Centre: Aldermaston, UK, 1974.

2134

Anal. Chem. 1984, 56,2134-2137

(22) Howard, A. G.; Mills, G. A. Int. J . Environ. Anal. Chem. 1983, 74, 43. (23) Stenberg, U.; Alsberg, T.; Blomberg, T.; Wannman, T. "Polynuclear Aromatic Hydrocarbons"; Jones, P. W., Leber, P., Eds.; Ann Arbor Science: Ann Arbor, MI, 1979; p 313. (24) Vassllarus, D. L.; Kong, R. C.; Later, D. W.; Lee, M. L. J . C h r o ~ t o g r . 1982, 252. 1.

(25) Tong, H. Y.; Sweetman, J. A,; Karasek, F. W.; Jellum, E.; Thorsrud, A. K. J. Chromatogr., in press.

for review February 16, 1984. Accepted May 21, 1984.

Development and Validation of an Air Monitoring Method for 1,3=Dichloropropene, trans - 1,2,3-Trichloropropene, cis- 1,2,3-Trichloropropene, 1,1,2,3-Tetrachloropropene, 2,3,3-Tr ic hlor0-2-p ropen- 1-0I, and 1,1,2,2,3-Pentachloropropane Mark A. Leiber and Howard C. Berk* Research Department, Monsanto Agricultural Products Company, 800 North Lindbergh Boulevard, St. Louis, Missouri 63167

A procedure for the sensitive and specific determination of 1,3dlchloropropene, c/s-1,2,3-trlchloropropene, traffs-1,2,3trlchloropropene, 1,1,2,3-tetrachloropropene, 2,3,3-trlchloro2-propen-1-01, and 1,1,2,2,3-pentachloropropane In air has been developed and evaluated. Tenax-GC sampling tubes were used for sample collection followed by solvent desorption and sample analysts by ceplllary GC/ECD. The method was laboratory valldated for the range 1 ppb to 6 ppm and overall average recoveries were 98 % at the parts-permllllon level and 106% at the parts-per-bllllon level for all analytes except 1,3-dlchloropropene. For 1,3-dlchloropropene the range was 0.4-4.0 ppm with a mean recovery of 100 %. The pooled CVs ranged from 0.03 to 0.09. Fleld tests were conducted at a productlon plant.

of a variety of organic compounds (1-11). Tenax has been shown to be efficient in trapping chlorinated hydrocarbons (2, 3,6-8). Solvent desorption of specific compounds from Tenax-GC has also been shown to be a useful technique (4, 5). For analysis, electron capture detection has been well documented as a selective and sensitive gas chromatographic detection system in connection with industrial hygiene methodology (12-14). The combination of the resolution of capillary columns and the sensitivity of electron capture detection has been found to be a powerful technique for the quantitation of organohalides (15). In this work we utilized Tenax-GC as the solid adsorbent to retain the series of chloropropenes and chloropropanes outlined above. The analysis technique involved desorption of the Texax-GC with isooctane a t 90 O C and then analysis by capillary GC with electron capture detection.

EXPERIMENTAL SECTION The development of appropriate air monitoring methods for all Monsanto herbicides and the key intermediates used in their production is an ongoing program in Monsanto Agricultural Products (MAP) Research. The methods are utilized to define the levels of these materials in work place air and to assure there are no locations where there may be unacceptable worker exposure to chemicals of concern. In addition to providing assurance that good industrial hygiene practices are followed at all manufacturing facilities, these air monitoring procedures may also be used to assess plant fugitive emissions when the need arises. Triallate, S-(2,3,3-trichloroally1)diisopropylthiocarbamate, is the active ingredient in FAR-GO herbicide which is used to control wild oats in wheat, barley, peas, and lentils. In the manufacture of triallate several chloropropenes and chloropropanes are used as process intermediates. In this study a method was developed to define the levels of 1,3-dichloropropene, cis-1,2,3-trichloropropene,trans-1,2,3-trichloropropene, 1,1,2,3-tetrachloropropene,2,3,3-trichloro-2propen-1-01, and 1,1,2,2,3-pentachloropropanein work place air. Tenax-GC, a polymer of 2,6-diphenyl-p-phenylene oxide, has been reported to be an effective absorbent for the analysis 0003-2700/84/0356-2134$01.50/0

Reagents. 1,3-Dichloropropene,cis-1,2,3-trichloropropene, trans-1,2,3-trichloropropene, 1,1,2,3-tetrachloropropene, 2,3,3trichloro-2-propen-l-ol, and 1,1,2,2,3-pentachloropropanewere obtained from Monsanto Agricultural Products Co. (St. Louis, MO). 1,3,5-Tribromobenzenewas obtained from Eastman Organic Chemicals,Rochester, NY. 2,2,4-Trimethylpentane (isooctane) was MCB-pesticidegrade. Tenax-GC (60-80 mesh) from Applied Science Laboratories, Inc., was used as the solid sorbent and no preparation or cleaning was needed. Apparatus. Air sampling pumps (Model P4000) were from E. I. du Pont de Nemours and Co., Inc. The gas chromatograph was a Varian Model 3700 GC, equipped with a Varian 8000 series autosampler, a 63Nielectron capture detector, an SE-30 30 m X 0.25 mm fused silica open tubular capillary column obtained from J and W Scientific, and a Hewlett-Packard 3390A recording integrator. High purity nitrogen was used as the carrier gas for the analyses. To generate standard parts per million level test atmospheres, a standards generator from Analytical Instruments Development, Inc. (Model 350), equipped with three chamber temperature settings was used. The chamber was operated at 50 "C for 1,3dichloropropene, trans-1,2,3-trichloropropene, and cis-1,2,3-trichloropropene and at 70 "C for 1,1,2,2,3-pentachloropropane, 1,1,2,3-tetrachloropropene, and 2,3,3-trichloro-2-propen-l-o1. Diffusion tube neck inside diameters used were as follows: 1,30 1984 American Chemical Society