1930
Anal. Chem. 1987, 59, 1930-1937
lengths are used for the eight Anacin-3 capsules. The smallest amount of KCN placed in these capsules, 9 mg, caused the capsule in which it was placed to appear 5.96 SDs from the training set in 4-D (4-wavelength) space.
CONCLUSIONS The Quantile BEAST method, when used in conjunction with NIRA data a t only four wavelengths, is able to quickly detect a wide variety of contaminants in capsules, obviating the need to open them. The ability of this technique to detect the absence of components that should be present as well as the presence of components that should be absent enables it even to signal the presence of contaminants that have no near-infrared absorption. Good results are achieved on a simple filter-based instrument, without the need for complex wavelength-selectionprocedures. The detection limit for KCN in capsules that has been obtained in this work is as much as 2 orders of magnitude below the lowest reported lethal dose (13).Substances other than KCN in capsules could also be detected at low concentrations. Selecting analytical wavelengths near the absorption features of components of interest should improve detection limits beyond those observed in these experiments. A representative training set composed of unadulterated samples is required to train the BEAST algorithm to recognize a good sample. In this research, collecting the training-set spectra took less than 1h and training the BEAST algorithm required less than 5 s. All of the training-set spectra were collected in a single day, but the tampering experiments took place over a period of 2 weeks. Nevertheless, repeated runs of validation samples showed that the calibration remained
stable throughout the duration of the experiment. More samples, it seems collected over time, would only enhance the reliability of the method. Registry No. Fez03,1309-37-1;Al, 7429-90-5; NaF, 7681-49-4; Asz03, 1327-53-3;NaCN, 143-33-9; KCN, 151-50-8.
LITERATURE CITED Tim, S. Time 1982, 120(0ct. l l ) , 18. Church, G. J. Tlme 1982, 720(0ct. 18), 16-18 Church, G. J. Time 1982, IPO(N0v. 8). 27. Wolnlk, K. A.; Fricke, F. L.; Bonnln, E.; Gaston, C. M.;Satzger, R. D. Anal. Chem. 1984, 56, 466A-474A. (5) Waldhole. M. Wall Street Journal 1986, (Feb. 14), 3. ( 6 ) Davidson, S. T i m 1986. 127(Feb. 24), 22. (7) Wall Street Journal 1988, (Jun. 19), 2 and 18. (8) Shenon, P. New York Times 1968, (Mar. 21). A1 and D19 (9) U . S . News and WorM Report 1986, IOO(Mar. 31), 8. (10) Greenwald, J. 6uslness Insurance 1986, Feb. 24, 2. (1 1) Andresky, J. Forbes 1988, 137(Apr. 28), 76-77. (12) Borrnan, S. A. Anal. Chem. 1982, 5 4 , 1474A. (13) Sax, N. I. Dangerous Propertks of IndustdalMaterlals;Van NostrandReinhold: New York, 1984. (14) Reese, K. M. Chem. €47. News 1982, 6O(Dec. 13), 82. (15) Hadzija, B. W.; Mattock, A. M. Forenslc Sci. Int. 1983, 23, 143-147. (16) Wetzel, D. L. Anal. Chem. 1983, 55, 1165A-1176A. (17) Lodder, R. A.; HleftJe. G. M. Anal. Chem., submitted for publication. (18) Efron, B. BiOmetr&a 1981. 68(3),569-599. (19) Honigs, D. E.; Hlrschfeld, T. B.;Hieftje, G. M. Appl. Spectrosc. 1985, 39, 1062-1065. (20) Honigs. D. E.; HleftJe, G. M.; Hirschfeld, T. B. Appl. Spectrosc. 1984, 38, 844-847. (21) Mark, H. L.; Tunneil, D. Anal. Chem. 1985, 57, 1449-1456. (22) Mark, H. L. Anal. Chem. 1986, 58, 379-384. (1) (2) (3) (4)
RECEIVED for review December 3, 1986. Accepted April 6, 1987. This work has been supported in part by the National Science Foundation through Grant CHE 83-20053, by the Office of Naval Research, and by the Upjohn Company.
Effect of Primary Ion Beam Parameters on the Secondary Ion Emission of Biomolecules from Liquid Matrices Richard B. Cole,’S2Christian Guenat,’ J. Ronald H a ~ sand , ~ Richard W. Linton*2 Laboratory of Molecular Biophysics, N.I.E.H.S., P.O. Box 12233,Research Triangle Park, North Carolina 27709, Department of Chemistry, University of North Carolina, Chapel Hill, North Carolina 27514, and Triangle Laboratories, Inc., P.O. Box 13485,Research Triangle Park, North Carolina 27709 A liquid metal ion source (Ga+) Interfaced to a double-focuskrg mass spectrometer (VG U S - 2 F ) was used to Investlgate the secondary ion mass spectrometry (SIMS) of blomolecules SIMS studies were performed on three anatyte compounds smponded In llquld matrlces. Varlations In the secondary ion profile of the peptide methlonlne enkephalin were probed vla monWhg of both the protonatedmolecular Ion (m/z 574) and fragment Ion ( m / z 120) olgnak Temporgl fluctuations In signal intensity as a function of prlmary Ion current denstty (controkd by the prlmary beam focus) were measured. Varlations In the relathre dgnal lntendtles of the two ions, as a function of local primary ion dose (controlled by the rastering speed), were characterlzed. The ablllty to obtain Images of both the protonated molecular ion and fragment Ions was also demonstrated for methbnlne enkephalln In a Bqdd m a t h protruding thrwgh a mlcroscoplc grld.
Liquid
ion (LM1) emitters are
achieving
prominent stature as primary ion sources for secondary ion mass spectrometry (SIMS). Their utility has been demonstrated in the application of spatially resolved inorganic SIMS to metal and semiconductor analysis (1)and to the analysis of alkali and alkaline-earth metals in biological specimens (2). The feasibility of molecular level organic compound analysis using LMI sources has also been demonstrated for a limited number of applications using a low current density primary beam directed at compounds dissolved in a glycerol matrix (3)and for the two-dimensional molecular imaging of a relatively stable organic liquid in an ion microprobe system (4). A unique feature of the LMI source is the small “virtual” source size, of the order of 10 nm in diameter (5))which enables electrostatic focusing of the primary ion beam to microscopic dimensions. Fqrthermore, primary beam current densities can be controlled (as will be described) over several orders of magnitude, from the low density conditions used in “static” SIMS experiments (
