Determination of polycyclic aromatic amines in natural and synthetic

Jul 23, 1981 - (9) Ranger, C. B. Anal. Chem. 1981, 53, 20 A-32 A. .... imate column dimensions were 50 cm X 0.66 cm2 cross sectional area. A special a...
1 downloads 0 Views 808KB Size
Anal. Chem. 1982, 54, 91-96

Snyder, L. R. Anal. Chim. Acta 1980, 114, 3-18. Betteridge, D. Anal. Ctiem. 1978, 50, 832 A-846 A. Ruzicka, J.; Hansen, E:. H. Anal. Cbim. Acta 1980, 114, 19-44. Ranger, C. B. Anal. Chem. 1981, 53, 20 A-32 A. Karlberg, 6.; Thelandeir, S. Anal. Cbim. Acta 1978, 98, 1-7. (1 . 1), Filho. H. B.: Medeiros, Jl. X.: Reis, 6. F.: Zaaatto, E. A. G. Anal. Chin?. (6) (7) (8) (9) (10)

Acta 1978, 701, 9-16, (12) Kina, K.; Shiraishl, K.; lshlbashi, N. Talanfa 1978, 25, 295-298. (13) Karlberg, B.; Johansson, P. A,; Thelander, S. Anal. Cbim. Acta 1979, 104. -21-28. . I

(14) Karlberg, B.; Thelander, S. Anal. Cbim. Acta 1979, 174, 129-136. (15) Kawase, J.; Nakae, 14.; Yamanaka, M. Anal. Cbem. 1979, 57, 1640-1 643. (16) Kawase, J. Anal. Chert?. 1980, 52, 2124-21127.

91

(17) Novotny, M.; Lee, M. L.; Bartle, K. D. J . Chromatogr. Sci. 1974, 12, 606-6 12. (18) Natusch, D. F. S.; Tomkins, B. A. Anal. Chem. 1978, 50, 1429-1434. (19) Hertz, H. S.; Brown, J. M.; Chesler, S. N.: Guenther, F. R.; Hllpert, L. R.; May, W. E.; Parris, R. M.; Wise, S. A. Anal. Cbem. 1980, 51, 1650- 1657. (20) Tijssen, R. Anal. Cbim. Acta 1980, 174, 71-89.

RECEIVED for review July 23,1981. Accepted October 12,1981. This work was supported in part by grants from the Office of Naval Research and the TAMU Center for Energy and Mineral Resources (18757).

Determination of Pollycyclic A,romatic Amines in Natural and Synthetic Crudes B. A. Tomkins" and C.-h. Ho Analytical Chemistry Division, Oak Ridge National Laboratory, P.O. Box X, Oak Ridge, Tennessee 37830

The amine content of 21 fossil fuel sample Is determined by using an extensive acildic extractlon and adsorption plus gel chromatographic purlfication procedure. After the amlne Isolate is lrifluoroacetylated, the derlvatlred products are identlfled by GC/MS. The procedure lglves a precision of approximately f30% RSD at a level of 10 pg/g for amines ranging from 1-aminoriaphthalene to 1-aminopyrene, and a detectlon llmlt of 2 pg/g. The specles 2-aminonaphthalene may be a useful Indlcdor of the polycycllc aromatlc amine content of fossil fuel niaterlals.

Polycyclic aromatic hydrocarbons (PAH) have long been one focus of concern to health- and environmental-effects researchers working with developing synthetic fuels industries. More recently, however, even more highly mutagenic polycyclic aromatic amines (PAA) have been detected (1-5) in some synfuels materials and have been found (3,6-8) to be the determinant mutagens in some coal-derived materials. These species include recognized or suspected carcinogenssuch as 1-aminonaphthalene (1-AN), 2-amlinonaphthalene, 4aminobiphenyl, 1-aminoanthracene, and 2-aminoanthracene (9-11). Thus, the identification and quantitation of PAA in synthetic fuels is important to the development of safe and environmentally sound synthetic fuels industries. We have recently relported (5) a procedure for isolating an amine-enriched fraction from synthetic crudes. This procedure, while effective, was not suitable for rapid analytical-scale determinations. Aromatic amines have been determined in cigarette smoke using column-chromatographic isolation procedures followed by derivatization with pentafluoropropionic anhydride 1(12). Derivatization of amines with fluorinated anhydrides, which is effective for all stericalliy unhindered amines (13-15), is a useful step in any amine analysis. For example, ~t~ifluoroacetylation introduces a foreign element which is appropriate for selective spectroscopic identification procedures (16). In this paper, we report the isolation and quantitation of PAA in fossil fuel products. The amines are isolated via extensive extraction a n d chromatographic purification. These species are identified and quantitated by GC-MS and gas chromatography, respectively, after the isolate is derivatized with trifluoroacetic anhydride (TFAA) I

EXPERIMENTAL SECTION Solvents and Chemicals. Pentane, methylene chloride, and benzene were either purchased from Burdick and Jackson Laboratories (Muskegan, MI) and used without further purification or else redistilled from ACS reagent grade stock purchased from Fisher Scientific, Inc. (Fairlawn, NJ). The following reagents were also purchased from Fisher Scientific, Inc., and were used as received: dimethyl sulfoxide (Me2SO),2-propanol, methanol, acetone, diethyl ether, potassium hydroxide pellets, and sodium bicarbonate. Trifluoroaceticanhydride (TFAA)was obtained in 99+% purity (Gold Label grade) from the Aldrich Chemical Coo (Milwaukee,WI) and was used as received. ACS reagent grade hydrochloric acid was purchased from Hi-pure Chemicals, Inc. (Nazareth, PA) and was used as received. Sephadex LH-20, particle size 25-100 pm, was procured from Pharmacia, Inc. (Piscataway,NJ), and swollen in 2-propanol for at least 24 h before use. Basic alumina, AG 10, 100-200 mesh, pH 10-10.5, wm acquired from Bio Rad Laboratories (Richmond, CA). All of the amines used were purchased in reagent or better grade from the Aldrich. Chemical Co. (Milwaukee,WI), RFR Inc. (Hope, RI), Fisher Scientific Inc., and the Sigma Chemical Co. (St.Louis, MO) and were used as received. Tracers. Radioactive benzidine [14C(U)],specific activity 25.7 mCi/mmol, was purchased from New England Nuclear (Boston, MA) and used as received. The radioactive tracer l-aminonaphthalene-l-14Chydrochloride, 4.75 mCi/mmol, was customprepared by California Bionuclear Corp. (Sun Valley, CA). The hydrochloride salt was dissolved in methylene chloride and extracted with 1M sodium hydroxide solution. After the organic layer was dried, the solvent was removed. The residue, which was redissolved in was the free base l-amin~naphthalene-I-~~C, toluene to yield a solution with 1.5 X lo6 dpm/mL. Special Equipment. The chromatographic columns, which were fabricated in-house, consisted of a 25-mL buret equipped with a Teflon stopcock and a 300-mL solvent reservoir. The reservoir was topped with a i40/25 socket joint. The approximate column dimensions were 50 cm X 0.66 cm2cross sectional area. A special adaptor, employing a J 40/25 ball joint, was used whenever nitrogen (approximately 15-20 psi) was applied to the column. The columns were filled with one of the following: (a) 20 g of basic alumina slurried in 50% (v/v) methylene chloride in benzene, or (b) a slurry of Sephadex LH-20 in 2-propanol to a bed volume of 25 mL. Volumetric flasks with capacities of 0.1, 0.3, or 0.5 mL were obtained from SGA Scientific, Inc. (Bloomfield,NJ), stock number JM-6165. The continuous extractors used were similar to those described by Coles (17) and were fabricated in-house.

0003-2700/82/0354-0091$01.25/0 0 1981 American Chemical Society

02

ANALYTICAL CHEMISTRY, VOL. 54, NO. 1, JANUARY 1982

,

C? J C I A 2 S - 4 5 E N ES

DEP 'VA'IZAT

ON

>ATH T F A A

Figure 1. Outline of the isolation procedures for amines and azaarenes. Instrumentation. All gas chromatograms were recorded by using either a Perkin-Elmer 3920 or Sigma I gas chromatograph (Perkin-Elmer Corp., Norwalk, CT). The reporting feature of the Sigma I was used not only to identify but also to quantitate appropriate peaks. The gas chromatographic column used in both instruments was a 3.7 m X 3.2 mm 0.d. (12 ft x l/g in. 0.d.) glass column packed with 3.0 % (w/w) Dexsil 400 on Supelcoport, 100/120 mesh. The injector and detector temperatures were 300 and 320 "C, respectively, while the oven was programmed from 90 "C (hold for 8 min) t o 320 "C (hold for 16 min) at a rate of 2 "C/min. The carrier gas, helium, flowed at a rate of 35 mL/min. The hydrogen and air pressure for the FID were identical, 20 psig. Each determination used 5 FL of sample. Samples. All gasifier tar samples, which were collected at the low Btu coal gasifier a t the University of Minnesota a t Duluth, were supplied by B. R. Clark of this Laboratory. Three of the materials, a coal oil, a crude shale oil, and a petroleum crude, were Comparative Research Materials (CRMs) which are available (18) for biological assay and analytical methods development research through the US. Environmental Protection Agency/Department of Energy Fossil Fuels Research Materials Facility (18,19). The remaining coal oil was obtained from the National Bureau of Standards (Washington, DC) through the joint DOE/NBSsponsored Analytical Characterization Group. Recovery Measurement. A 3-10% volume aliquot of the amine isolate was added to 10 mL of a scintillation cocktail prepared by dissolving 4 g of Omnifluor (New England Nuclear, Boston, MA) in 1 L of toluene. Liquid scintillation counting of the carbon-14 tracers was performed for 10 min a t room temperature using the carbon-14 counting channel of a Tri-Carb Liquid Scintillation Counter, Model C-2425 (Packard Instrument Go., Downers Grove, IL). All sample counts were corrected for scintillation quenching by using the calibrated Automatic External Standard option of the instrument. Procedure. (a) Isolation. All samples had to be reasonably free of solid particles before beginning the acid extraction step. Gasifier tar samples, which are semisolids, were dissolved in methylene chloride as completely as possible and spiked with 1 mL of radioactive tracer solution. A 10-fold excess by volume of pentane was then added, and the entire solution fitered through a medium-porosity fritted funnel. The filtrate was then taken as the sample solution. Coal oils, shale oil, and petroleum crude were not subjected to this pretreatment step. Ten grams of sample (except tar and petroleum crude) were dissolved in 100 mL of benzene, spiked with 1 mL of tracer solution, and extracted four times with 75 mL of 6 M HC1 (20 g of petroleum crude was spiked with 1mL of tracer and extracted continuously for 24 h (17)with 50 mL of 6 M HC1 and pentane). The pooled acid extracts were then neutralized to pH 14 with solid

potassium hydroxide pellets. Ice was added frequently to keep the temperature below 50 OC. The basic emulsion was then extracted four times with 150 mL of ether. A 200-mL portion of benzene was added te the pooled ethereal layers, to form an azeotrope with the water, and the soivents were removed by rotary evaporation to yield the ether-soluble base (ESB) fraction. The ESB fraction was transferred quantitatively, with less than 5 mL of 50% (v/v) methylene chloride in benzene, to the basic alumina column. After the body of the column was wrapped with aluminum foil to exclude light, the ESB was eluted under gravity feed with 50% (v/v) methylene chloride in benzene until quinoline eluted from the column. This first cut, the "crude azaarene fraction", was collected, taken to dryness by rotary evaporation, and set aside. The column was then cleaned with 250 mL of methanol. This second fraction, the "crude amine fraction", was collected and also taken to dryness using rotary evaporation. The residue was redissolved in not more than 5 mL of 2-propanol. The 2-propanol solution of the crude amine isolate was transferred to the Sephadex LH-20 column and eluted under 20 psi nitrogen pressure feed with 2-propanol until aniline emerged from the column (typically one column volume of 2-propanol was used). Then the column was washed with 250 mL acetone. The acetone fraction was collected, taken to dryness as noted previously, and weighed to yield the "purified amine fraction", which was redissolved in exactly 1 mL of methylene chloride. The crude azaarene fraction was subjected to a M e 8 0 partition procedure, as described previously (20). This residue is the "purified azaarene fraction". (b) Derivatization and Analysis. Aliquots of both the purified amine and azaarene fractions were taken for recovery measurement. A known volume of the purified amine fraction was then derivatized with trifluoroacetic anhydride (TFAA), as described in ref 16, and concentrated to exactly 0.10 mL. Portions of the derivatized and underivatized isolates and a derivatized standard were chromatographed under the conditions listed under "Instrumentation". The overall isolation and quantitation procedure is presented in Figure 1. The constituents of the derivatized amine isolates were identified with GC/MS. Positional isomers were identified by using authentic derivatized standards where such materials were available. The response factors for all derivatized amines beyond Nmethylaniline (except 2-aminoanthracene which readily decomposes) listed in Table I were averaged to give R, the average response factor. ( R typically exhibited a relative standard deviation of & E % . ) The concentration of the individual amines could then be calculated by using the following equation: Fg/g = (A/RmY)(Vf/vinj)(vT/ v d ) (1) where A = peak area, R = average response factor, Fg/area, VT

ANALYTICAL CHEMISTRY, VOL. 54, NO. 1, JANUARY 1982

93

Table I. Identities anci Concentrations of Derivatized Standard PAAa re1 retention timeb before derivatization 0.06

com p ou nd aniline-TFAA o-toluidine-TFAA rn-toluidine-TFAA N-ethylaniline-TFAA N-methylaniline-TFAA quinoline" 2,5-dimetliylaniline-TFAA 2,4,6-trime thylaniline-TFAA 2-aminobiphenyl-TFAA 1-aminonaphthalene-TFAA 2-aminonaphthalene-TFAA acridined 4-aminobiphenyl-TFAA 2-aminofluorene-TFAA 1-aminoanihracene-TFAA 2-aminoanthracene-TFAAC' 1-aminopyrene-TFAA 6-aminochrysene-TFAA

0.1 1 0.1 1

0.15 0.1 1 0.30 0.20 0.30 0.70

0.67 0.70

1.00 0.93 1.19 1.41 1.45 1.79 2.06

after derivatization 0.15 0.19 0.22 0.24 0.26 0.30 0.32 0.44 0.72 0.74 0.82 1.00

1.04 1.28 1.39 1.48 1.73 1.94

concn, mg/mL 0.62 0.23 0.49 0.56 0.36 0.65 0.58 0.57 0.33 0.33 0.47 0.43 0.37 0.36 0.41 0.38 0.43 0.31

re1 response factor 1.13 1.60 2.40 1.73 0.93 1.87 1.13 1.00 1.00

0.93 1.00

1.00 0.93 0.87 1.13

2.33 1.07 1.07

Relative a A chromatogram of' these species is given in Figure 2C. The peaks are identified in order, from left to right. to acridine. Relative )to 2-aminonaphthalene-TFAA. Azaarene compound. Response factor not used in calculation of R. e Response factor not used in calculation of R. See text. = final volume of amine isolate in pL, Vd =: volume, in pL of VT which was derivatized, Vf = final volume of derivatized amine isolate, pL, V , = volunie injected into GC, pL, w = weight of sample, g, an2 y = overall yield determined by scintillation counting. In the special case of L .aminonaphthalene, where the mass of the tracer was significant compared to the mass of the native species, eq 1 had to be corrected as indicated below:

1-AN, pg/g = (A/EEwy)(Vf/V,,)(VT/Vd)

-

2 3 . 5 ~ / ~(2)

where 23.5 pg is the mms of radioactive tracer added to each sample.

RESULTS AND DISCUI3SION Extraction studies which employed a variety of one- to four-ring PAA demonstrated that the recovery of individual PAA is dependent upon the concentration of HCl used in the extraction. While both 1 M and 6 M HCl recovered one- to three-ring aromatic amines equally well (minimum 75% recovery), 1 M HC1 was capable of extracting only 22% of the 1-aminopyrene present. The overall recovery of l-aminopyrene in the analytical procedure described is 87%)as shown in Table 11, thus demonstrating a 4-fold improvement in the recovery of these four-r ing, highly mutagenic species. Under normal conditions, batch extraction of the sample is quite sufficient for removing IPAA from the sample. In the case of the petroleum crude (C'IIM-3) sample, however, an emulsion formed which did not break even upon standing for 48 h. For this reason, continuous extraction was substituted for batch extraction with this sample. H o et al. (21)indicated that basic alumina had the ability to resolve azaarene and m i n e constituent in a shorter period of time than neutral alumina. Because both acidic alumina and Florisil, which a b o is acidic (22)) irreversibly adsorb amines, they are unsuitable for this application. The quinoline cutpoint is used becauw it is the last azaarene to elute from a basic alumina column (21). However, the use of basic alumina leads to two difficulties. First, it irreversibly adsorbs many monocyclic aromutic amines, thus leading to reduced values for these relatively nonmutagenic species. Secondly, the use of this adsorbent, precludes TFAA derivatization before column separation because the adsorbent, acting as a strong base, will hydrolyze the imide linkage (23). On the other hand, basic alumina does not irreversibly adsorlb PAA. Postcolumm

Table 11. Overall Recovery of Amines

compound 2-aminonaphthalene 4-aminobiph enyl 1-aminoanthracene 2-aminoanthracene 1-aminopyrene

mass added, a mg 1.15 1.49 1.03 1.19

0.80

mass recovered, mg 1.15 1.49 0.75 0.89 0.71

a Added to 5 g of CRM-2 (crude shale oil). of duplicate determinations.

av recovery, %

-- 100 100 73 75 89 Average

derivatization with TFAA does offer advantages, as will be explained later. The crude amine fraction typically contains a small quantity of apparently polymeric material which can be removed by the Sephadex gel column. The initial acid partition removes most, but not all, of the aliphatic hydrocarbons from the ether-soluble base fraction. These aliphatics coelute with the azaarenes in the basic alumina purification step and interfere with both the proper concentration and gas chromatography of this fraction. The use of MezSO liquid/liquid partition removes these interferences and yields a purified isolate. The overall procedure yields two fractions containing basic constituents, namely, the purified azaarene fraction and the purified amine fraction, which may be analyzed independently. Azaarenes such as quinoline, however, may photodecompose as they pass through the basic alumina column. Wrapping the column with aluminum foil minimizes decomposition. The direct GC analysis of the purified amine fraction is not straightforward for two reasons. First, many free amines are poorly resolved under the gas chromatographic conditions employed herein. In Figure 2A, which is the gas chromatogram of the purified amine fraction of coal oil A, 1-and 2aminonaphthalene (peaks a' and b', respectively) are poorly resolved both from each other and from a series of methylaminonaphthalenes which immediately follow. Secondly, it is extremely difficult to distinguish free amines from their methylated azaarene isomers using GC/MS because the parent ion will be the same in both cases (24). The standard procedure of derivatizing the amine fraction with trifluoroacetic anhydride (TFAA) simultaneously overcomes both difficulties. First, the resolution of amines is

94

ANALYTICAL CHEMISTRY, VOL. 54, NO. 1, JANUARY 1982

B

35

3

15

I C

20

TIME chr1

Figure 2. Gas chromatograms of the underivatized (A) and derivatized (B) amine fraction of coal oil A compared to a set of standard dervatized amines (C). Peak identities: a’, 1-aminonaphthalene; b’, 2aminonaphthalene; a, 1-aminonaphthalene-TFAA; b, 2-aminonaphthalene-TFAA; c, derivatized methyl aminonaphthalenes; d, derivatlzed aminobiphenyl;e, derivatized C,-methylaminobiphenyl; f, de-

rivatized aminoanthracenes; g, derivatized methyiaminoanthracenes. greatly improved after TFAA derivatization. Table I demonstrates this clearly for at least two groups of compounds. o-Toluidine, rn-toluidine, and N-methylaniline coelute without derivatization but are clearly resolved upon trifluoroacetylation. Similarly, 2-aminobiphenyl, 1-aminonaphthalene, and 2-aminonaphthalene are all base line resolved after derivatization but are poorly resolved as free amines. In Figure 2B, which is the gas chromatogram of the derivatized, purified amine fraction, the TFAA derivatives of 1- and 2-aminonaphthalene (peaks a and b, respectively), are better resolved than the corresponding underivatized peaks in Figure 2A. Furthermore, the collection of derivatized isomeric methylaminonaphthalenes (peak c) is also better resolved. In our experience, TFAA derivatization of the amine fraction leads to slightly longer retention times, compared to underivatized amines, for derivatized amines with one to three aromatic rings, while retention times for derivatized amines with four or more aromatic rings are slightly shorter than those of their underivatized counterparts. Secondly, an amine of mass M which is derivatized with TFAA shows a characteristic M + 96 peak in its mass spectrum. Hence, TFAA derivatization provides a “tag” by which amines and azaarenes may be spectroscopically distinguished. A packed gas chromatographic column was used in this study because quantitation using such a column is a wellestablished, routine procedure. A capillary gas chromatographic column would have provided significantly improved resolution; however, the quantitative data obtained would have been less reliable. The mass spectra of the derivatized amines were sufficiently clean (after background subtraction) to permit easy location of both the M 96 and confirming M - 1 mass peaks and subsequent generic identification of the derivatized amine.

+

It should be noted that it is not feasible to identify the derivatized amines on the basis of retention time data alone because standards for all of the amines of interest are not commercially available. Hence, GCf MS data are essential for providing the qualitative identifications. Furthermore, the mass spectrum for each GC peak usually indicated a single fragmentation pattern, thereby demonstrating that a given peak was representing either a single compound or, a t worst, isomers of the same compound. Quantitation could therefore be achieved by using a flame ionization detector. We have observed that either derivatized or underivatized amine fractions begin decomposing after approximately 2 weeks, even after storage in the dark at 4 “C. For this reason, it is recommended that all amine analyses be performed immediately after isolation and that standards be checked and replaced frequently. The decomposition of the derivatized amines is due, in part, to the slow hydrolysis of the amide linkage. The species 2-aminoanthracene is particularly sensitive to degradation and therefore is somewhat unreliable as a standard compound. Of the two radioactive tracers used in this study, 1aminonaphthalene [ 1-I4C]is preferred to benzidine [14C-(U)] for two reasons. First, benzidine (4,4’-diaminobiphenyl) has two amino groups in it, while the major amines found thus far in fossil fuels have only one. Secondly, derivatized benzidine will slowly precipitate in either acetonitrile or methylene chloride solutions of the isolate, thereby reducing the apparent recovery, whereas derivatized 1-aminonaphthalene will not. The sole advantage of the radioactive benzidine is that it is commercially available and does not have to be custom synthesized. The overall recovery of the radioactive tracer is correlated, to a great extent, with the viscosity of the sample. Thus, the recovery of aminonaphthalene label in the gasifier tar, a semisolid (viscosity exceeds 300 cSt), is 50-60%, that in the crude shale oil, a somewhat viscous liquid (48.8 cSt), is 75-85%, and finally, that in the coal oil, which is a freeflowing liquid (83. cSt), is 85-95%. The efficiency and accuracy of the overall isolation and analytical procedure are illustrated by the recoveries of five, two- through four-ring PAAs which were spiked into CRM-2, a crude shale oil. The level of the spike was considerably above that of the native PAA in this sample. The recoveries, measured in duplicate (Table 11), range from 73 to 100%. They indicate that PAA in the range from two rings through a t least five rings may be identified and measured by this procedure. For most samples, the gas chromatographic base line remains fairly linear, and the integration of peaks of interest is straightforward. This is not the case for gasifier tar samples, where even freshly derivatized isolates can exhibit a considerable amount of unresolved material. For this reason, a “valley-to-valley” criterion was used to establish the base line for each peak. The aromatic amine profile of coal oil A, both underivatized and derivatized, is shown in Figure 2A,B, respectively. The identities and concentrations of the standard compounds are given in Table I. Table I11 lists the identities, which were established by GC/MS, and concentrations of the amine species found in three gasifier tars and coal oil CRM-1. The CRM-1 data represent the average of quadruplicate results. Typically, the standard deviation of the amine concentrations was 3 pg/g, and the mean 10 pg f g, thus yielding a relative standard deviation of nominally 30%. This precision value is acceptable, considering the trace levels of each component, the complexity of the matrix, and the degree of sample handling required for the analytical method. In some of the samples tested, anthracenecarbonitrile, dimethylindole, and tetrahydroquinoline were tentatively identified in the deriv-

ANALYTICAL CHEMISTRY, VOL. 54, NO. 1, JANUARY 1982

Table 111. Composition of the Amine Fractions of Three Gasifier Tars and CRM-1 Coal Oil

compound

tar 17a

concentrations, p g / g tar tar lBa 83a CRM-lbgC

aniline toluidine C,-aniline

5 1 3 5 i 58 2 0 i 10 37 i 30

C,-aniline

1 3 + 11 7

9

C,-aminonaphthalene 3 14 1-aminonaphthalened 2-amin~naphthalene~ 41 5 C, -aminonaphthalene 27 15 25 15 C,-aminonaphthalene

18 55 15 7

14 27 12

21

16 16 31 17 6

18 31 18 10

5

52i. 3 69i 2 19i6 32i 9 6i:3 30i 7 9i.2

12 C, -aminobiphenyl C,-aminobiphenyl C, -aminobiphenyl

12 11

28i 7 3 i 2 7 i 3 4+2 18+ 5 6

C,-amino biphenyl aminofluorene acridined C,-aminonaphthalene aminofluorene C,-amino biphenyl C,-aminofluorene C,-aminofluorene

25 25

C,-aminofluorene 1Faminoanthracene 2-aminoanthracene C, -aminoanthracene

10

14+9 22+ 8 6i 3 24 18

4

25 3 i 2

8 8 11

3 3

7

42

repository sample no.

description

CRM-1

coal oil A

CRM-2

crude shale oil A

CRM-3

petroleum crude A gasifier tar 83

NBS coal oil

2-AN concn, av concn,a RSD, pg/g

71 67 68 71 2.2 1.5 4.3 3.6

pg/g

69

i

2.1

%

3

2.9 + 1.3 44

ND b 12.9 2 7 + 15.4 19.7 48.4 25.0 117c 117c

58

naphthalene, if only gas chromatography is used, and slightly lower if GC/MS is employed. One fossil fuel material may be analyzed in quadruplicate in 10 man-days. The analytical procedure described herein is clearly useful for profiling amines in such diverse matrices as petroleum-, shale oil, and coal-derived materials. Furthermore, this procedure demonstrates a useful multicomponent property in that all amines with two aromatic rings and greater (through a t least four rings) may be successfully identified and quan-

titated.

10

aminopyrene % yield of tracer

Table IV. Concentration of 2-Aminonaphthalene (2-AN) in Fossil Fuel Samples

a Represents mean i. standard deviation. Not detect. ed: duplicate analyses. One analysis. Relative standard deviation.

10

amino biphenyl

95

36

65

4 i 2 17i 2 20i 2 < 2 i 0.5 2i 1 3+2 3 %1 5 ? 1.3 9+5 8 0 + 15

Benzidine [ ‘C-( U) ] tracer used for recovery correction: one determination. 1-Aminonaphthalene-1-14C tracer used for recovery correction. Represents mean + standard deviation of four determinations. Values without standard deviation indicate compound appeared in only one or two of four nliquots. GC/:MS data confirmed by GC retention time, others tentatively identified by GC/MS only. a

atized extract. However, these species were chromatographically resolved from the derivatized amines. The data in Table Ill, which gives the concentration of amines in three tars arid a coal oil, suggest that 1- and 2aminonaphthalene and the collection of methylaminonaphthalenes might be useful indicators of the entire class of PAA in some coal-der ived materials, because these species are present in higher concentrations than m y others and they parallel the total measurable PAA concentration. Table IW lists the concentration OF 2-aminonaphthalene determined b y this procedure (up to four times) in two coal oils, a shale oil, a petroleum crude, and gasifier tar no. $3. In two of these samples, crude shale oil A and petroleum crude A, the %AN was the only measurable PAA. The detection limit of this procedure is approximately 2 kg/g based on 2-amino-

ACKNOWLEDGMENT The authors thank C. A. Pritchard, M. V. Buchanan, and G. Olerich for performing the GC/MS analyses mentioned in this work. LITERATURE CITED (1) Guerin, M. R.; Ho, C.-h.; Rao, T. K.; Clark, B. R.; Epler, J. L. Envlron. Res. 1980. 23. 42-53. (2) Wilson, B. ‘W.;’Pelroy, R. A.; Cresto, J. T. Mutat. R ~ S .1980, 79, 193-202. (3) Guerin, M. R.; Rubln, I. B.; Rao, T. K.; Clark, B. R.; Epler, J. L. Fuel 1981, 60, 282-288. (4) Buchanan, M. V.; Ho, C.-h.; Guerin, M. R.; Clark, B. R. I n “Chemlcel

Analysis and Blological Fate: Polynuclear Aromatlc Hydrocarbons”; Cooke, M., Dennis, A. J., Eds.; Batelle Press: Columbus, OH, 1981; pp 133-144. (5) Ho, C.-h.; Clark, B. R.; Guerin, M. R.; Ma, C. Y.; Rao, T. K. Pfepr. Pap.-Am. Chem. SOC., Div. Fuel Chem. 1979, 24 (l), 281-291. (6) Pelroy, R. A.; Gandolfi, A. Mutat. Res. 1980, 72, 329-334. (7) Wilson, B. W.; Pelroy, R. A.; Lee, M. L.; Later, D. W. Proceedlngs of the 29th Conference on Mass Spectrometry and Allled Toplcs, 1981, Abst. No. MPM085. (8) Wilson, B. W.; Craun, J. C.; Pelroy, R. A.; Peterson, M. R.; Felix, D. W. Proceedings, 2nd USDOE Environmental Control Symposium 1980, 7,

41 5-4 18. (9) Hoffmann, D.; Wynder, E. L. I n “Chemical Carcinogenesis”; Searle, C. E., Ed.; American Chemical Society: Washington, DC, 1976, pp. 324-365. (10) Clayson, D. G.; Garner, R. C. I n “Chemical Carclnogenesis”, Searle, C. E., Ed.; Amerlcan Chemical Society: Washington, DC, 1976; pp 366-462. ( 1 1 ) Shaikh, B.; Hallmark, M. R.; Hallmark, R. K.; Manning W. G.; Pinnock, A.; Kawalik, J. C. J . Chromafogr. 1980, 795, 392-397. (12) Patriankos, C.; Hoffmann, D. J. Anal. Toxlcol. 1979, 3. 150-154. (13) Masuda, Y.; Hoffmann, D. J . Chromatogr. Sci. 1989, 7, 694-697. (14) Blau, K.; Klng, G. S. “Handbook of Derivatives for Chromatography”; Heyden: London, 1978; pp 104-151. (15) Knapp, D. R. “Handbook of Analytical Derlvatization Reactions”; Wlley: New York, 1979; pp 65-103. (16) Tomkins, B. A.; Feldman, C. Anal. Chim. Acta 1980, 119, 283-290. (17) Coles, G. V. J. SCl. FOOdAgric. 1956, 7 , 11-17. (18) Grlest, W. H.; Coffin, D. L.; Guerin, M. R. Fossll Fuels Research Matrix

-

Proaram. ORNL/TM-7346 Oak Ridoe National Laboratorv: Oak Ridae. - . TN,-June 1980. (19) Coffin, D. L.; Guerln, M. R.; Griest, W. H. “Proceedings of the Symposium on the Potential Health and Environmental Effects of Fossil Fuel

96

Anal. Chem. 1982, 54, 96-101

Technologies," Gatlinburg, TN, Sept 25-28, 1978, CONF-780903, Oak Ridge National Laboratory: Oak Ridge, TN, 1979; p 153. (201 Natusch. D. F. S.: Tomklns. B. A. Anal. Chem. 1978. 50. 1429-1434. i21) Ho, C.-h.; Guerln, M. R.; Clark, B. R.; Rao, T. K.; Epier, J L J Anal. Toxicol. 1081. 5. 143-147. (22) Heftmann, E., Ed. "Chromatography", 3rd ed.; Van Nostrand-Reinhoid: New York, 1975; p 56. (23) Vogel, A. I. "Practical Organic Chemistry", 3rd ed.; Wiley: New York, 1956; pp 798-799, 1074-1076.

(24) Buchanan, M. V. Oak Ridge National Laboratory, personal communlcation, 1980.

RECEIVED for review April 16, 1981. Accepted October 13, 1981, This work was funded by the U.S, Department of Office of Health and Environmental ReSHU'ch under Contract No. W-7405-eng-26 with the Union Carbide Corp.

Collision-Induced Dissociation with Fourier Transform Mass Spectrometry R. B. Cody, R. C. Burnier, and B. S. Frelser" Department of ChemMy, Purdue University, West Lafayette, Indiana 47907

Colllslon-induced dissociation is demonstrated on a number of primary and secondary Ions uslng a Nicolet prototype Fourler transform mass spectrometer. Like the triple quadrupole technlque, collision-Induced dissociation uslng FT-MS Is a relatively low energy and efflclent process. The ablllty to study a wlde range of Ion-molecule reaction products is exemplified by results on proton-bound dimers and transltlon metal containing ionic specles. Variation of colllslon energy by varying the RF Irradiation level can provlde information about product distributions as a functlon of energy as well as yleld ion structural Informatlon. Like the triple quadrupole technlque, no slits are employed and vlrtually all of the fragment ions formed by the CID process may be detected. Unlike all previous mass spectrometric techniques for studying CID, a tandem Instrument Is not required, and dlfferent experiments are performed by making software modifications rather than hardware modifications.

Collision-induced dissociation (CID) remains one of the most useful and widely employed mass spectrometric techniques for ion structure determination and forms the basis for the powerful MS/MS technique for complex mixture analysis. In general, the CID technique consists of accelerating a given ion into a collision gas thereby imparting energy to the ion and inducing fragmentation. The ionic fragments are then mass analyzed, yielding essentially a "mass spectrum" of the precursor ion. While high kinetic energies (3-30 keV) are required to observe CID using mass-analyzed ion kinetic energy spectrometry (MIKES) in reverse-geometry mass spectrometers (1-3),Yost and Enke using a triple quadrupole mass spectrometer recently demonstrated the unexpected result that a low-energy (