Anal. Chem. 1983, 55, 2132-2137
2132
(14) Guerin, M. R.; Ho, C.-.; Rao, T. K.; Clark, B. R.; Epler, J. L. Int. J . Environ. Anal. Chem. 1980, 8 , 217-225. (15) Mefford, I.; Keller, R. W.; Adams, R. N.; Sternson, L. A,: Yllo, M. S. Anal. Chem. 1977, 49,683. Tomkins, B. A.; Ostrum, V. H.; Caton, J. E. Anal. Chim. Acta 1982, 134, 301-311. Toste, A. P.; Sklarew, D. W.; Pelroy, R. A. J . Chromatog. 1982, 249, 267-282. Haugen, D. A.; Peak, M. J.; Suhrbler, K. M.; Stamoudis, V. C. Anal. Chem. 1982, 54,32-37. Masuda, Y.; Hoffmann, D. Anal. Chem. 1989, 4 7 , 650-652. Masuda, Y.; Hoffmann, D. J . Chromatogr. Sci. 1989, 7, 694-697. Tomkins, B. A.; Feldman, C. Anal. Chem. Acta 1980, 779,283-290. Tomkins, B. A.; Ho, C.-h. Anal. Chem. 1982, 54,91-96. Later, D. W.;Lee, M. L.; Wilson, B. W. Anal. Chem. 1982, 54, 117-123. Campbell, J. A.; Grlmsrud, E. P.; Hageman, L. R. Anal. Chem. 1983, 55, 1335-1340. Buchanan, M. V. Anal. Chem. 1982, 54,570-574. Later, D. W.; Lee, M. L.; Bartle, K. D.; Kong, R. C.; Vassilaros, D. L. Anal. Chem. 1981, 53, 1612-1620.
(27) Tedjamulia, M.; Tominaga, Y.; Lee, M. L.; Castle, R. N. Proceedings of the Seventh International Symposium on Polynuclear Aromatic Hydrocarbons: Columbus. OH. Oct 26-28. 1982. (28) Slhlbom, L. Acta Chem. Scand. 1954, 8 , 529-533. (29) de Kok, A.; Roorda, I . M.; Frel, R. W.; Brlnkman. U. A. Chromatographia 1981, 14,579-566. (30) Vogel, A. In "Practical Organic Chemistry, 3rd Ed.; Wiley: New York, 1956; pp 498-1074. (31) Vassllaros, D. L.; Kong, R. C.; Later, D. W.; Lee, M. L. J . Chromatogr. 1982, 252, 1-20.
RECEIVED for review June 10, 1983. Accepted July 21, 1983. This work was supported by the Department of Energy, Division of Biomedical and Environmental Research, Contract No. DE-AC02-79EV101237,with Brigham Young University and Department of Energy, Contract No. DE-AC06-76RLO1830, with Pacific Northwest Laboratory.
Determination of Alkyllead Salts in Water and Whole Eggs by Capillary Column Gas Chromatography with Electron Capture Detection Donald S. Forsyth and William D. Marshall*
Department of Agricultural Chemistry and Physics, Macdonald Campus of McGill University, Ste. Anne-de-Bellevue, Quebec H9X 1CO, Canada
Alkyllead salts (R3Pb+ and R,Pb2+, R = Me or Et) are recovered from water or whole eggs by complexometrlc extraction wlth dlthlzone. The dlthironates are phenylated and speclated by caplllary column GC wlth electron capture detectlon. The method Is sensltlve to low parts per billion levels of lead salts In 2.5 g egg homogenate. At these levels methyllead salts (but not ethyllead salts) interact strongly wlth the sample matrix. Treatment of the matrix wlth llpases and proteases releases them.
Recently a renewed interest in the speciation of lead in environmental samples has resulted from several diverse lines of investigation. Organolead compounds have been detected in cod, lobster, mackerel, and flounder meal (10 to 90% of the total lead burden) ( I ) , freshwater fish (2),air (3-6),street dust and human brains (8). A steady input of organoleads into the environment results from the continued use of tetraalkylleads as antiknock additives. In addition evidence for the chemical (9-12) and biological (13-16) alkylation of organolead salts or of lead(I1) salts have been obtained. Although organoleads may make only a small contribution to the total lead intake of an organism, it has been demonstrated that trialkyllead salts arising in tissues from the degradation of tetraalkylleads are important in lead toxicity. The conversion of R4Pb to R,Pb+ occurs rapidly in liver homogenates from rats and rabbits (17,18). Acute toxicities of tetraalkylleads and of trialkyllead salts are similar (17,19) and are at least an order of magnitude greater than dialkyllead salts or inorganic lead salts. Dialkyllead salts cause symptoms of toxicity similar to those produced by inorganic lead salts and they exhibit an affinity for thiol compounds ( 1 7 ) . Triorganolead salts inhibit oxidative phosphorylation and have been reported to bind to proteins (20, 21).
(a,
0003-2700/83/0355-2 132$01.50/0
Several methods for the determination of tetraalkyllead have been proposed which depend on a combination of gas chromatography (GC) with selective detectors (22). To our knowledge there have been few reports on the direct determination by GC of trialkyllead salts (23) and none for dialkyllead salts. Our approach to the determination of these salts has been to further alkylate them with a Grignard reagent to convert them to their tetraalkyl analogues prior to GC analysis. This paper describes a method for the isolation of trialkyl- and dialkyllead salts at the low parts per billion level from water and from whole eggs (minus shells) and their determination, after derivatization, by capillary column GC with electron capture detection. The pretreatment of chicken whole egg homogenate with a mixture of proteases and lipases in an effort to dislodge bound residues of these salts is also described. EXPERIMENTAL SECTION Apparatus. A Varian 3700 gas chromatograph equipped with a 63Nielectron capture detector, a 30 m X 0.327 mm DB-1 fused silica capillary column, and a splitless injector was used in this investigation. Operating conditions were as follows: carrier gas, helium, 3 mL/min; detector makeup gas, nitrogen, 30 mL/min; detector temperature, 300 "C; injector temperature, 200 "C; temperature program, isothermal at 50 OC for 2 min followed by linear ramping to 200 O C at 4 OC/min. Quantitation was performed by external standards with a Shimadzu C-RIB data processor. Reagents. Solvents were distilled in glass grade from Caledon Laboratories Ltd. Inorganic reagents including diphenylthiocarbazone (dithizone) were ACS reagent grade or better. The ammoniacal buffer comprised 11.3 g of diammonium citrate, 2.0 g of potassium cyanide, and 12.0 g of sodium sulfite in 250 mL of HzO. The pH was adjusted with concentrated ammonium hydroxide. Alkyl- and phenylmagnesium chloride were obtained from Alfa Products and the phenylmagnesium bromide was obtained from Aldrich Chemical Co. 0 1963 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 55, NO. 13, NOVEMBER 1983
Trimethyl- (Me3PbC1)and triethyllead chloride (EbPbC1)were prepared by halogenation of the tetraalkyl analogues with anhydrous HC1 using standard methods (24). Dimethyllead dichloride (Me2PbCl12)was prepared from tetramethyllead by oxidation in chloroform at -20 “C with chlorine. Diethyllead dichloride was purchased from Alfa Products. Trialkyllead salts were purified by irepeated crystallization from ethyl acetate/ hexane. Dialkyllead salts were recrystallized from methanol/ diethyl ether. Purity was monitored by thin-layer chromatography (TLC). An optimum separation was achieved on Eastman Kodak 6061 silica gel plates (0.2 inm) using acetone:hexane:propionic acid as eluent (3:7:0.2, v/v/v). Relative molbilities ( R f )were as follows: Me,PbCl (0.71), Et3PbC1 (0.88), Me2PbC12 (0.321, Et2PbC12(0.55), PbC12 (0.00). The plates were activated at 120 “C for 30 min prior to use. Plates from other sources resulted in less efficient jeparations. Products were visualized by fluorescence quench and confirmed by overspraying with 0.1% dithizone in chloroform (Pb(II), pink; R2PbC12,salmon; RSPbC1, yellow). Phenylated alkyllead standards were preprved by two different methods. I3bCl2was treated with phenylmagnesium chloride, alkylmagnesium chloride, and alkyl bromide. For trialkylphenylleads (R,PbPh) the inolar ratio of PhMgCl/RMgCl/PbClz was 1:3:1; for the dialkylcliphenylleads (R:$PbPhz)and for alkyltriphenylleads (RPbPh3)the ratio was adjmted to 2:2:1. PbClz (10 g) was added in small portions to ice cold mixtures of vigorously stirred alkylmagnesium chloride and alkyl bromide in tetrahydrofuran (under N2). The reaction was allowed to proceed at 5 “C for 10 min; the reactants were then allowed to warm to room temperature over a period of 1h. The crude reaction mixture was diluted with hexane and washed three times with distilled water. The organic fraction was dried, reduced in volume, and fractionated by vacuum distillation by use of a 6-cm heated Vigreux column. Alternately Et3PbC1 was treated with phenyllithium at 3 “ C under anhydrous conditions. Product fractionation was as described above. No impurities in alkylphenyl leads prepared by redistillation of the above products were detected by spectroricopic or elemental analysis. The identity of each standard wa13verified by capillary column GC/MS. Transalkylation. Equimolar quantities of Me2PbPhzand Et2PbPh2were refluxed for up to 6 h in isooctane-conditions known to cause extensive transdkylation of related tetraalkylleads (25). Thermally induced transalkylation of trialkyllead salts was achieved by refluxing equimolar quantities of Me,PbCl and h. TLC of the crude reaction mixture EgPbCl in acetone for followed by visualization with dithizone indicated that small quantities of mixed trialkyllead salts had formed but gave no indication of diallryllead salt or inorganic lead salt. Methods. Hydrolysis. ‘Wholeegg homogenate was incubated at 37 “C for 24 h in 60 mL of 5% ethanol/O.l M phosphate buffer (pH 7.5) containing 30 mlg of Lipase Type I11 and 30 mg of Protease Type XIW. Preparations were obtained from Sigma Chemical Co. Extraction. Method 1 (used for water samples). Ammoniacal buffer (pH 8.5,10 mL) was added to the sample (60 mL) which was then extracted three times with 0.005% (w/v) dithizone (10 mL) in 50% benzene/hexane. The organic extracts were combined, reduced in volume to 0.5 mL, and tlerivatized directly. Method 2 (used for egg homogenate). A.bsolute ethanol (15 mL for experiments B and D or 22 mL for experiments E and F), and ammoniacal buffer (pH 9.5, 10 mL) were added to the sample. The mixture was extracted three times with 0.05% (w/v) dithizone (10 mL) in 50% benzene/hexane. The organic extracts were combined, centrifuged and back extracted three times with 10 mL of HNO, (0.15 M). The aqueous waeihes were combined, neutralized with NaOH, and further basified with 5 mL ammoniacal buffer (pH 9.5). The alkyllead salts were recovered by extracting the aqueous phase three times with 0.01% dithizone (10 mL) in 50% benzene/hexane. These washes were combined and centrifuged and the organic layer was reduced in volume to 0.5 mL. Derivatization. Anhydrous THF (2 mL) and 0.5 mL of phenylmagnesium bromide (3 M) were addecl to the concentrated extracts under nitrogen. The solution was stirred for 30 min at room temperature then transferred quantitatively to a centrifuge tube (15 mL). The volume was adjusted to 10 mL with water and
2133
then extracted three times with hexane (3 mL). Centrifugation followed each extraction to hasten phase separation. For experiments A-D, the combined hexane extracts were washed once with water ( 5 mL) and twice with 30% H20/CH3CN(2 mL) and then diluted to 15 mL with hexane. For experiments E and F, the hexane extracts were washed once with water (5 mL), diluted to 15 mL with hexane, and then washed three times with CH&N (1 mL). The hexane was readjusted to 15 mL and analyzed directly. R E S U L T S A N D DISCUSSION Rationale. The method reported herein was developed to speciate lead in Herring Gull (Larus argentatus) eggs. This species is considered to be an excellent biological indicator of pollution within the Great Lakes in that it occupies a position high on the aquatic food chain and has a quite diverse diet. The Canadian Wildlife Service had made available to us several samples of whole egg homogenate which were known to contain relatively high levels of total lead. The method was developed by using chicken whole egg homogenate that could be solubilized with concentrated sodium chloride solutions. Increased salt concentration improved the efficiency of extraction of alkyllead salts by organic solvents (benzene, ethyl acetate, or methyl isobutyl ketone). Alkali and alkaline earth halides from several sources, however, contained appreciable quantities of inorganic lead. The possibility of transalkylation of alkyllead salts (25) precluded heat treatment in any proposed method. As we wished to eventually modify these procedures to include determinations of inorganic lead, solvent extraction using complexing agents seemed t o be the most promising procedure. The added benefit of such a strategy would be that no column cleanup would be necessary. We wished to incorporate an optional hydrolysis procedure which might allow us to differentiate between “readily extractible” and “dislodgeable” alkyllead salts. Chromatography. The commercial splitless injector system was unsuitable for these analyses because of persistent sample decomposition despite exhaustive silylation of the inserts and of the injector body with dimethyldichlorosilane in boiling xylene. On-column silylating agents (Silyl-8 or hexamethyldisilazane) were not appreciably better. The problem was overcome by using an all-glass insert and the injector modifications that are detailed in Figure 1. The salient features of the modified system are an increased internal volume and a decreased metal surface area. The major modification corisists of an adaptor, a 1 / 4 in. Swagelok nut in SSI male stainless steel which has been silver soldered to a connector. The interior of the adaptor has been drilled out t o accept a 6.25 mm 0.d. borosilicate glass insert (2 mm i d . X 82 mm). A gastight seal between this insert and the adaptor is made by means of a 1/4 in. diameter disk of lead foil (2 mm thickness). The borosilicate insert is held in place by tightening the adaptor on to the bulkhead fitting in the roof of the chromatographic oven. In so doing both the vespel/graphite ferrule (C) and the lead foil (D) are slightly deformed to assure gastight seals. Three capilliary columns were examined for separation, bleed, and inertness characteristics. A 30-m borosilicate column coated dynamically with SE-30 (Canadian Capillary Co.) developed significantly more bleed and produced more peak tailing of organoleads during programming than did a 25-m methylsilicone fused silica column (Hewlett-Packard Co.). A 30-m fused silica DB-1 column (J and W Scientific Co.), however, developed a superior separation of the alkyllead standards from coextractives. It was employed in all subsequent studies. Helium was chosen as carrier gas over nitrogen because it increased resolution of the mixture and nitrogen was chosen as makeup gas relative to 5% argon methane because it resulted in more stable detector operation. Nitrogen doped with 10 ppm oxygen caused increased detector response
2134
ANALYTICAL CHEMISTRY, VOL. 55, NO. 13, NOVEMBER 1983
Table I. Absolute Retention Times and Retention Indexes (Based on Retention Times) Relative to a Homologous Series of n-Bromoalkanes
a
’
analyte
absolute retention time, min
retention indexa
n-C,H,,Br Me ,PbPh n-C,,H 21 Br EtMe,PbPh Et,MePbPh n-Cl,H 25Br
11.47 16.15 18.13 19.25 22.11 24.29
800 941 1000 1036 (1033)b 1129 (1125)b 1200
Formula used, I = lOON
+
Et,PbPh n-C 14H29Br Me,PbPh , EtMePbP h Et ,PbPh, n-C18H3,Br 1 0 0 r ~ [ ( R T- ~R~T )R ( N ) / ( R T R ( N + -~ RTR(N))]. )
Me,PbPh Et,PbPh Me,PbPh, Et,PbPh, a
linear regression eqa
absolute retention time
retention index
24.76 29.84 30.87 33.11 35.14 39.47
1217 14 00 1443 1536 (1532)b 1620 1800
Predicted retention index (26).
k3-*
Table 11. Linear Regression Analysis of Calibration Curves for Alkylphenylleads
analyte
analyte
coefficient of correla- data tion points
+ 1843 + 494.4 + 3163.0 + 778.7
y = 415.9~ y = 238.9~ Y = 662.5~ y = 547.1~
0.9998 0.9997 0.9996 0.9988
15 15 15 15
y = peak area (pV s) and x = pg of Pb.
to coextractives but did not enhance analyte response appreciably. Five operating parameters were each varied systematically to optimize the detector response (a combination of sensitivity and reproducibility) to each of the four analytes. The results are summarized in Figure 2 in which detector response for approximately equimolar alkylphenyllead standard is plotted as a function of (a) injector temperature, (b) carrier gas flow rate, (c) column depth inside injector, and (d) detector makeup gas flow rate. Each data point is the average of three replicate determinations and the standard deviation is recorded as an “error bar”. Each parameter was varied separately while the other four parameters were maintained constant: makeup gas, 30 mL min-l; injector, 200 “C; 2.0 cm column depth; 2.0 min initial hold. On the basis of sensitivity and reproducibility a carrier gas flow rate of 3.0 mL m i d and a makeup gas flow rate of 30 mL m i d nitrogen were judged optimal. At column flow rates above 3.0 mL m i d the resolution of analytes from coextractives rapidly deteriorated. More importantly there was a pronounced deterioration in the analyte peak shape. Although increased sensitivity was observed with lower makeup gas flow rates, the poorer reproducibility and the increased susceptibility to contamination of the detector (particularly metal deposits) did not warrant this modest improvement. When determinations of egg homogenate were attempted with 20.0 mL m i d
I-H Flgure 1. Splitless injector composed of (B) in 0.d. borosilicate insert, (C) I/.,in. Vespel/graphite ferrule, (D) lead foil pierced by a in. male pinhole, (E) ’1, in. Swagelok nut silver soldered to a connector (SSI),(F) 0.4 mm graphite ferrule, (G) in. Swagelok nut, and (H) fused sillca capilliary column. The insert Is secured in the Injector housing by tightening (E) to bulkhead fitting (A) inside the oven.
makeup gas, frequent decontamination of the detector (high-temperature baking) was necessary. An isothermal hold time between 1 and 3 min did not alter response or reproducibility. No septum purging was found to be necessary during chromatographic runs. Under these conditions the observed retention times for the four alkyphenyllead standards and their corresponding retention indexes relative to a series of n-bromoalkanes (26) are recorded in Table I. Under our programming conditions the log of the adjusted retention time vs. carbon number was distinctly curvilinear (7 = 0.965) and a better correlation was obtained by using the actual retention
Table 111. Mean Recoveries of Alkyllead(1V) Salts from Water, Buffer, and Whole Egg Homogenate by Using Different Extraction Procedures procedure % mean recovery of alkylphenyllead method spiked exp t matrix at ppba hydrolysis 1 2 Me,PbCl Et,PbCl Me,PbCl, Et,PbCl, 7 2 i 4 go* 5 103i 7 93i 3 A distilled waterb 14-19 J 60+3 9 9 + 13 85i3 92* 5 B distilled waterb 0.9-1.5 C phosphate bufferb 14-19 89i 3 10026 70i2 86 i 5 400-525 J J 9oi3 8524 62+2 78 i 4 D whole egg homogenate J 35i6 83i. 26 2 3 i 15 1 1 5 + 22 E whole egg 30-50 homogenate J 8 0 i 12 79i 4 7 6 i 15 1 1 7 i 15 F whole egg 30-50 J homogenate
$
a
Reported as ppb of Pb.
Volume 60 mL.
Weight 2.5 g.
no. of replicates 6 3 3 3 5 6
ANALYTICAL CHEMISTRY, VOL. 55, NO. 13, NOVEMBER 1983
2135
I
d 1.0
1.1
2.0
2.1
3.0
COLUMN fLOU M T C I M L / M I Y I
c 1oo-j 4 27s:
I
I
I I
____ -
-....4
*..a.
1 ....... ..--....._.
column flow rate, (C) column Flgure 2. Detector response for equimolar alkylphenyllead standards as a function of (A) Injector temperature, (6) detector makeup gas flow rate. depth inside injector, and (0)
time ( 7 = 0.997). These data were used t o predict (27)retention times of EtMezPbPh, EtzMePbPh, and EtMePbPhz. The predicted retention times and the observed values (from thermally induced transalkylation studies) are also recorded in Table I. Calibration curves were generated for each of the alkylphenyllead standards using chromatographic conditions described above. A linear increase in detector response was
observed with increasing analyte concentrations (range 4-500 pg) of each of the four alkylphenyllead standards. The relevant parameters are summarized in Table 11. Extraction. Two separate extraction procedures were developed; the single complexometric extraction was suitable for water samples whereas the double extraction procedure provided additional cleanup for the more complex whole egg
2136
ANALYTICAL CHEMISTRY, VOL. 55,
NO. 13, NOVEMBER
30-
I
272U21h
F
18-
v
3
0
0
5
10
15
20
25
30
35
UO
US
50
ELAPSED TIME (HOURS1
LEGEND: DATE
-FEE
16
.__----. FEE 23
Figure 3. The release of amino nitrogen (mg) from egg homogenate (2.5 g) by enzymatic hydrolysis.
homogenate. Residual droplets of aqueous phase in organic centrifugates were conveniently removed by placing the capped centrifuge tube in a freezer (-20 "C)for 30 min. After this time the organic phase was decanted leaving a pellet of ice behind. Enzymatic Hydrolysis. The course of the enzymatic hydrolysis of whole egg homogenate is presented in Figure 3 in which micrograms of N were monitored by ninhydrin (28) as a function of time. Three replicate trials were run on each of two separate days. Within experimental error no differences were observed between 24 h and 36 h. This technique was found to be 72 f 9% effective after 24 h and 78 f 9% effective after 48 h relative to classical acid hydrolysis. Relative to the
1983
protease alone a combination of protease and lipase resulted in a more homogeneous hydrolysate which had less tendency to emulsify when extracted with organic solvents. Derivatization. The derivatization sequence was performed directly on the dithizonates because of small but persistent losses of the free alkyllead salts during solvent removal. It was imperative that the extracts not be taken to dryness (freeze-drying of the lead salts or of their corresponding dithizonates resulted in appreciable loss). Interestingly phenylation yields were also consistently better on the dithizonates than on the free salts. It is postulated that the dithizone acts as a keeper in this system. Recoveries. Table I11 summarizes the results of six recovery trials. By use of the simplified extraction procedure (method 1) or the double extraction technique (method 2) recoveries from distilled water or from phosphate buffer (spiked at 1-20 ppb Pb) were consistently high. No consistent differences in recoveries from the two methods were detected. Only method 2 was attempted for samples of whole egg homogenate. The hydrolysis when incorporated into the determination sequence had no substantial effect on the mean recovery of ethyllead salts; however, precision was considerably improved (experiment E vs. experiment F). In contrast the mean recovery of methyllead salts was dramatically improved. These results suggest that the methyllead salts are rapidly bound up by the egg matrix. The determinations were commenced 15 min after each sample was spiked. The enzymatic pretreatment apparently releases the bound analytes. The possibility that transalkylation of one lead salt by another might occur during the isolation procedures or upon chromatography of the derivatives was investigated in the following studies. Gas chromatography/mass spectrometry identified the following mixed alkylphenyl leads, EtMezPbPh, EtMePbPh, and EtMePbPhz as well as biphenyl and terphenyls in the crude reequilibration reaction mixtures. We have not been able to detect any of these "mixed alkylleads during recovery trials of alkylphenyllead standards or during recovery trials using trialkyllead chlorides or dialkyllead chlorides. Figure 4 records chromatograms of (A) a synthetic mix of alkylphenyllead standards, (B) one of the replicate recovery trials from experiment "F" (equivalent to 30-50 ppb P b in whole egg homogenate), and (C)a control recovery trial from whole egg homogenate. Because recoveries from water, phosphate buffer, and hydrolyzed egg homogenate were consistently high and because there were no significant peaks in the chromatograms which could have been attributed to transalkylation products, i t is concluded that transalkylation
A
1 1
2,
Figure 4. Chromatograms of: (A) synthetic mixture of 15.0 pg of trimethylphenyllead (l), 15.0 pg of triethyiphenyllead (2), 15.9 pg of dimethyldiphenyllead (3), and 17.1 pg of diethyldiphenyllead (4); (B) recovery trial from whole egg homogenate (spiked with 30-50 ppb alkyllead salts); (C) recovery trial from control whole egg homogenate.
Anal. Chem. 1983, 55, 2137-2142
was not a major source of loss in the methods. The application of the method to Herring Gull whole egg homogenate was tested by analyzing three samples. The procedures followed were exactly those for experiment F of the recovery trials. No practical difficulties were encountered with the methods indicating that chicken egg homogenate was a suitable substitute matrix for these samples. Nu indication of the presence of lead salts in these samples was obtained and it is concluded that, a t the time of analysis, these analytes were present at less than 1-4 ppb in Herring Gull egg samples. Recovery trials directly from Herring Gull egg homogenate were not attempted for lack of suitable samples; thus binding to this matrix maLy be different from thoise interactions observed in the case of chicken egg homogenate. ACKNC)WLEDGMENT Our sincere appreciation to M. Ihnat, CBRI, Agriculture Canada, for microanalyses and to R. Norstrom, Canadian Wildlife Service, for GC/MS analyses. Registry No. Me3PbC1, 1520-78-1; Et3PbC1, 1067-14-7; MezPbClz,1520-77-0;Me,Pb, 75-74-1;EtzPbC1,, 13231-90-8;Pb, 7439-92-1; lipase, 9001-611-1; protease, 9001-92-7; Me3PbPh, 19040-53-0;Et3PbF’h,878-50-2;MezPbPh2,421169-20-0;EhPbPhz, 4692-79-9. LITERATURE CITED Slrota, 6. R.; Uthe, J. F. Anal. Chem. 1977, 49, 823-825. Chau, Y. K.; Wong, P. T. S.; Kramer, 0.; Bengert, (3. A.; Cruz, R. B.; Klnrade, J. 0.; Lye, J.; Van Loon, J. C. Bull. Envlron. Contam. Toxl. col. 1980, 2 4 , 265-269. Harrison, R. M.; Perry, R. Atmos. Environ. 1977, 1 1 , 847-852. De doughe, W R. A.; /\dams, F. C. Atmos. Envlron. 1980, 14, 1177-1 180. Reamer, D. C.; Zoller, W. H.; O’Haver, T. C. Anal. Chem. 1978, 5 0 , 1449-1455. Rohbock, E.; Georgil, H. W.; Muller, J. Atmcs. Envlron. 1980, 14, 89-98.
2137
(7) Harrison, R. M. J . Environ. Sci. Health, Part A 1978, A 1 I , 419-423. (8) Nlelsen, T.; Jensen, K. A.; Grandjean, P. Nature (London) 1978, 274, 602-603. (9) Ahmad, J.; Chan, Y. K.; Wong, P. T. S.; Carty, A. J.; Taylor, L. Nature (London) 1981, 287, 70-71. (IO) Reisinger, K.; Stoeppler, M.; Nurnberg, W. H. Nature (London) 1981, 291, 228-230. (11) Cralg, P. J. Envlron. Techno/. Lett. 1980, 7 , 17-20. (12) Jarvie, A. W. P.; Markall, R. N.; Potter, H. R. Nature (London) 1975, 255, 217-218. (13) Chau, Y. K.; Wong, P. T. S.“Lead in the Marine Environment”; Branlca, M., Konrad, Z.,Eds.; Proceedings of the International Expert Dlscussion on Lead Occurrence, Fate and Pollution in the Marine Envlronment, Yugoslavia, 1977; Pergamon Press: New York, 1980. (14) Wong. P. T. S.; Chau, Y. K.; Luxon, P. L. Nature (London) 1975, 253, 263-264. (15) Schmidt, U.; Huber, F. Nature (London) 1978, 259, 157-158. (16) Thayer, J. S. “Occurrence of Biological Methylation of Elements in The Environment”; Amerlcan Chemical Society: Washlngton, DC, 1978; ACS Advances in Chemlstry Series No. 182, pp 188-205. (17) Cremer, J. E. Br. J . Ind. M e d . 1959, 16, 191-199. (18) Bolanowska, W.; Wisniewska-Knypie, J. M. Blocbem. Pharmacol. 1971, 2 0 , 2108-2110. (19) Grandjean, P.; Nielsen, T. Resklue Rev. 1979, 7 2 , 97-148. (20) Bolanowska, W. Br. J . Ind. M e d . 1988, 2 5 , 203-208. (21) Byington, K. H.; Yates, D. A.; Mulllns, W. A. Toxlcol. Appl. Pharmacol. 1980, 5 2 , 379-385. (22) Crompton, T. R. “Gas Chromatography of Organometallic Compounds”; Plenum Press: New York, 1982; pp 361-451. (23) Estes, S. A.; Uden, P. C.; Barnes, R. M. Anal. Cbem. 1981, 5 3 , 1336-1340. (24) Shapiro, H.; Frey, F. W. “The Organlc Compounds of Lead”; Interscience: New Ynrk, 1968. (25) Caiingaert, G.; Beatty, H. A.; Soroos, H. J. Am. Chem. SOC. 1940, 62, 1099-1104. (26) Pacholec, F.; Poole, C. F. Anal. Cbem. 1982, 5 4 , 1019-1021. (27) Kovlts, E. sz. Adv. Chromatogr. 1985, 1 , 229-247. (28) Yemm, E. W.; Cocking, E. C. Analyst (London) 1955, 8 0 , 209-213.
RECEIVED for review November 29,1982. Resubmitted July 12, 1983. Accepted July 12, 1983. Financial support in the form of an operating grant (W.D.M.) and a postgraduate scholarship (D.S.F.) from the National Science and Engineering Research Council is gratefully acknowledged.
Table-Driven Procedure for Infrared Spectrum Interpretation Mark 0. Trulson’ and Morton E. Munk*
Department of Chemistry, Arizona State University, Tempe, Arizona 85287
An infrared spectrum interpreter program lis described which is based on a table-driven procedure. Eighteen commonly encountered carbonyl-containing groups were selected for program development and testlng. The knowledge base includes only information about those regions of the spectrum considered to be diagnostic for each of the 18 classes. It was derived from an experimental data file of about 2000 infrared spectra compiled specifically for this study. The program requires peak position, Intensity, and shape. Its performance was evaluated on a test set of 146 spectra.
An important rnethod of the elucidation of the structure of organic compounds depends on an analysis of their spectral and chemical properties. CASE (I),a linked set of computer programs, is being developed to assist the chemist in the execution of the three major components of the process: (a) interpretation, the reduction of the chemical and spectral properties of the unknown compound to tlheir structural imPresent address: Department of Chemistry, University of California, Berkeley, Berkeley, CA 94704. 0003-2700/83/0355-2 137$01.50/0
plications; (b) molecule assembly, the generation of molecular structures compatible with the inferred structural information; and (c) spectrum simulation and comparison, the ranking of the generated structures on the basis of the fit between predicted and observed spectral properties. An advanced molecule assembler, program ASSEMBLE (2), has been developed to accept the structural inferences drawn by the interpreter, program INTERPRET, on the basis of multispectral data. ASSEMBLE generates all molecular structures consistent with these inferences and any additional information provided by the chemist. Program SIMULATE (3), like INTERPRET, is at an early stage of development. In this paper the role of infrared spectrometry in the development of INTERPRET is discussed. Three approaches to automated spectrum interpretation have received wide attention: pattern recognition (4-7), library search (6, 8-13), and artificial intelligence (14-18). More recently, hierarchical clustering techniques based on “average spectra” as representations of specific structure-spectra correlations have been studied (19, 20). Our initial effort in automated infrared spectrum interpretation resulted in program INFRARED, an interactive heuristic program designed for application to multifunctional 0 1983 Amerlcan Chemical Society
