2251
Anal. Chem. 1984, 56,2251-2253
Table I. Least-Squares Statistics for Fits of First-Derivative Signal vs. Concentration" of Nd(II1) smoothing function
slope i std
intercept
devb
h std devb
std err est
corr coeff
0.81 i 2.5 13.6 f 14 -2.2 h 2.1
3.3 18.6 2.8
0.9998 0.992 0.9999
6.04 h 0.07 none unmodified S-Gc 4.75 f 0.42
modified S-Gc
6.15 i 0.06
OFour samples: 0.6, 1.2, 2.4, and 6.3 mg mL-'. bRelativeunits. Five-point smooth.
with amplitude of 1.0. Whereas the modified function gives a smooth decrease on either side of the attenuated peak value, the unmodified function generates a response with significant character associated with it. It is this feature which caused problems with our attempts to use the smoothing functions for derivative spectroscopy because derivatives of the oscillatory response from "smoothed" noise pulses imposed significant amounts of distortion on derivative spectra. Figure 5 shows first-derivative spectra obtained for a mercury penlamp with the unmodified and modified smoothing functions. Differences are quite apparent, with data smoothed with the unmodified function showing overlapping peaks where single peaks are expected and observed with the modified function. The smoothing functions were evaluated with intensity (0 data obtained for Nd(II1) solutions monitored with an image-dissector based instrument constructed in this laboratory.
Table I presents least-squares statistics for first-derivative signals vs. concentration for solutions of Nd(II1) without any smoothing and with smoothing by each of the functions discussed above. All statistical parameters are degraded for the data smoothed by the unmodified function relative to data without smoothing or smoothed by the modified method. Whereas the unmodified S-G functions degrade the signal, the modified function introduces little or no distortion relative to the data without any smoothing. More complete studies are needed to fully understand the advantages and limitations of these modified Savitsky-Golay functions. However, preliminary results presented here suggest that these modified functions may offer real advantages in some situations. LITERATURE CITED (1) Savitzky, A,; Golay, M. J. E. Anal. Chern. 1964, 36, 1627. (2) Kuo, F. F.; Kaiser, J. F. "System Analysis by Digltal Computer"; Wiiey: 1966; pp 218-285. (3) Hamming, R. W. "Dlgltal Filters": Prentice-Hall: Englewood Cliffs, NJ, 1983; pp 34-45.
Timothy A. Nevius H a r r y L. Pardue* Department of Chemistry Purdue University West Lafayette, Indiana 47907 RECEIVED for review February 27, 1984. Accepted May 29, 1984. This work was supported by Contract DE-ACOZ79EV10240 from the Department of Energy.
Secondary Ion Mass Spectrometry of Pyrene: Enhancement of Molecular Ion Emission by Antimony Trichloride Sir: Mass spectra of positive ions produced by sputtering (i.e., FABMS, SIMS) from condensed phase organic analytes are generally characterized by abundant even electron ions ( I ) . Secondary positive ion emission is enhanced if the analyte is basic, or contains a basic moiety such as -NH2, or exists as an organic salt. Secondary ion emission is particularly enhanced when the analyte has first been dissolved in a polar, protic matrix, such as glycerol. In this case, pseudomolecular ions such as [M H+], [M + Na+], [M + K+], etc., or intact cations from organic salts are apparent in the SIMS spectrum. Relative intensity of these species in the SIMS spectrum has been related to their concentration in the condensed phase solution (2). Radical cations (i.e., molecular ions) would not be expected to exist in significant concentrations in a glycerol matrix; in fact, no enhancement of secondary emission of positive, molecular ions is observed from such a matrix. If secondary emission of molecular ions can be enhanced by a mechanism similar to that effected on organic bases by glycerol, it is reasonable to study matrices which enhance the concentration of the radical ions in solution. Molecular ions have been sputtered from polynuclear aromatic hydrocarbons (PAH) deposited on solid, metal supports ( 3 , 4 ) . Enhancement of secondary molecular ion emission has been observed when PAH's were deposited on carbon (5) or liquid metal substrates (6). In none of these experiments has the molecular species been dissolved into the support. More recently, however, secondary molecular ion emission has been reported for N,N-tetramethylbenzenediamine (TMPD) in glycerol and MezSO solutions (7). In these solutions, the TMPD exists in significant concentrations as part of a
+
0003-2700/84/0356-225 1$01.50/0
charge-transfer complex with quinone. Consistentwith these results, we report enhanced secondary ion emission of pyrene molecular ions when pyrene is dissolved in a matrix of SbC13. Molten SbC13solutions are known to ionize many PAHs via a reversible, one-electronoxidation (eq 11,the resultant PAH+ being stable and dissolved in the melt (8). The system is an aprotic analogue to organic base/glycerol solutions in which the basic analyte can exist as a solvated, protonated ion prior to bombardment. 1/3SbC1, .t PAH
* 1/3Sb0+ C1- + PAH+
(1)
SbC13posesses properties which are desirable for an organic SIMS matrix. These include an acceptably low pressure in the mass spectrometer (vide infra), a low melting point, and the ability to conduct charge away from the bombarded surface (9). Significantly, PAHs are soluble in SbC13(8,lO). Performance of SbC13 as a SIMS matrix was evaluated by comparison of the intensity of secondary emission of pyrene molecular ions from pyrene/SbCl, and pyrene/glycerol mixtures as well as the secondary emission of the same ions from a neat pyrene sample dispersed on a solid metal probe tip. Because the concentration of dissolved pyrene molecular ions in SbC1, is known to increase when the mixture is melted (8), secondary emission of pyrene molecular ions was measured as a function of temperature. EXPERIMENTAL SECTION The secondary ion mass spectrometer used in this study has been described previously (11) and has been modified by the addition of a source heater and a thermocouple. The 5-keV Ar+ 0 1984 Arnerlcan Chemical Society
2252
SIMS
ANALYTICAL CHEMISTRY, VOL. 56,
OF
SIC,,
NO. 12, OCTOBER 1984 A
I
Flgure 1. SIMS spectrum of SbCI3 at 21 "C.
primary beam intensity is approximately 5 MA/cm2. Resulting secondary ions are accelerated to 8 keV for mass analysis using a tandem mass spectrometer of EBE design (12). Structural identity of secondary molecular ions was verified by using CAD/MS/MS with argon as collision gas. Sampleswere admitted to the mass spectrometer via solid probe. Detachable probe tips of gold- or chromium-plated brass were used because reaction of SbC13with stainless steel and aluminum probe tips resulted in intense emission of secondary metal and metal chloride cations. For the SIMS analysis of pyrene on a bare, Au probe tip (no liquid matrix), the probe tip was loaded with 0.39 mg of pyrene (Aldrich)by depositing 1mL of a 0.39 mg/mL solution of pyrene in toluene onto the probe tip, and evaporating the solvent. When the SbC13was used as a matrix, an inhomogeneous,4 mol % pyrene in SbC13 (Aldrich) mixture was prepared in a glovebox under positive pressure of anhydrous nitrogen. Approximately 10 mg of the mixture was placed on the probe tip and heated on a hot plate, whereupon the mixture quickly melted to yield a bright green solution. The probe tip was removed from the hot plate, allowed to cool, and then screwed onto the probe. The probe was then removed from the glovebox, placed in the mass spectrometer insertion lock, and evacuated as rapidly as possible. When glycerol was used as a matrix, a 4 mol % pyrene in glycerol (Kodak, ACS Reagent Grade) mixture was prepared by sonicating an inhomogeneous mixture of pyrene and glycerol for 2 h, and allowing the mixture to stand for several days. A 10-mg sample of the mixture, which was still inhomogeneous,was deposited onto the probe tip for analysis. Effects of sample temperature on secondary ion emission were evaluated by inserting the probe at 21 "C (ambient temperature) and heating the sample at the rate of 1.1OC/min. During heating, temperature was monitored via a thermocouple, previously calibrated using the same temperature program. We find this approach to yield the most reproducible results, since tuning of the instrument with each sample is required and since evaporation of both sample and matrix occurs.
RESULTS AND DISCUSSION Insertion of SbC13into the source produces a rise in pressure of approximately 1.5 x torr above a background of 5 X torr. SbC13is comparable to glycerol in our instrument, in spite of the fact that SbC13 has a higher vapor pressure: 1 torr a t 49 "C (13),compared to 4 X torr at 49 "C for glycetol (14). The secondary ion mass spectrum of SbC13 shows the ions expected from SbC1, (Figure 1). Lower mass ions corresponding to C1+ are not shown in Figure 1. SbC13 has a much lower mass spectral background than does glycerol, principally because it lacks the hydrocarbon backbone of glycerol. A SIMS spectrum of a 10-mg sample of a 4 mol % pyrene/SbC13 mixture acquired at ambient temperature (21 "C) has an intense ion at m / z 202 (Figure 2A). The intensity of this ion is very stable and is observed to be nearly constant for up to 2 h. Identity of this species as pyrene molecular ion was confirmed by MS/MS. The intensity of the ion at mlz 203 was 16-18% of the intensity of m / z 202, consistent with the 13C isotopic abundance expected from the pyrene molecular ion (C16Hlo+.). Other organic ions are observed which correspond to losses of H- and H, from the pyrene radical
1
h $
B
2
,
MI2
202
I
Figure 2. (A) SIMS spectrum of a 4 mol % solution of pyrene in SbCI, at 21 "C. (B) SIMS spectrum of a 4 mol % solution of pyrene in S K I 3 at 44 O C .
cation. No chlorinated organic species are observed. Molten SbCl3 (mp 75 "C) is known to oxidize many polynuclear aromatic hydrocarbons (8). To explore this effect, the source temperature was increased. Heating a t a rate of approximately 1"C/min resulted in an exponential increase in the intensity of the pyrene molecular ion signal (mlz 202), until a temperature of approximately 40 "C was achieved (Figure 3). The mlz 202 signal intensity reached maximum intensity at 60 "C, with a 50-fold increase over the intensity recorded at 21 "C. At temperatures above 65 "C the intensity decreased, presumably due to depletion of the pyrene sample. Intensity vs. temperature data similar to that presented in Figure 3 cannot be obtained by cooling the source from 65 "C to 21 "C, because both pyrene and the SbC13matrix evaporated rapidly at 65 "C. In contrast to the mlz 202 signal, the intensities of ions originating from SbC13 (represented by the SbC12+isotopic ion mlz 195 in Figure 3) initially underwent a 5-fold decrease between 20 "C and 30 "C. The intensities of these ions remained relatively constant until the temperature exceeded 65 "C, where they began to decrease. The relative abundance of Sb+, SbCl+, SbC12+,and SbC13+ions remained fairly constant throughout the experiment. Enhancement of the m / z 202 signal relative to the SbC13background ion signal a t an intermediate temperature is shown in Figure 2B. For comparison purposes, secondary ion intensity vs. temperature measurements were made with a 0.4-mg sample of neat pyrene deposited onto a clean, gold probe tip. This sample size was the same amount of pyrene present in the 10-mgsample of the SbC13solution. Under identical operating conditions (Le., in back-to-back runs with the pyrene/SbC13 experiments), pyrene molecular secondary ion intensity was
2253
Anal. Chem. 1984, 56.2253-2256 Pyrene
and
and 02,do catalyze the oxidation of pyrene by SbC13 These species were undoubtedly present in the reagent grade SbCl, used here. Enhancement cannot be solely ascribed to oxidation of pyrene by SbC13 prior to bombardment, because maximum secondary ion emission occurs at temperatures below the melting point of SbC13,i.e., below temperatures at which oxidation is known to occur. This suggests that other processes may be contributing to enhancement. Evaporation of SbC13,for example, would result in increased concentration of pyrene in the solution. However, the intensity of the SbC12' ion (m/z 195) remains relatively constant between 30 O C and 65 "C. This may indicate that evaporation of SbC13does not significantly affect secondary ion intensity over the duration of the experiment. Further experiments are under way to develop the capabilities of SbC13as a SIMS matrix and to identify the nature of the enhancement effect. These include measurements similar to those reported here, with samples prepared and SIMS analysis done under completely anhydrous conditions using zone-refined metal halides.
SbCI, Secondary I o n
I n t e n s i t y versus T e m p e r a t u r e
Pyrene,
m/z 2 0 2
0
ACKNOWLEDGMENT We thank Chris P. Leibman, Gleb Mamantov, and A. C. Buchanan I11 for helpful discussion. Registry No. SbC13, 10025-91-9;pyrene, 129-00-0.
LITERATURE CITED (1) Barber, M.; Bordoll, R. S.; Elliot, 0. J.; Sedgwick, R. D.; Tyler, A. N. Anal. Chem. 1982, 54 645A. (2) Caprloll, R. M. Anal. Chem. 1983, 55, 2387-2391. (3) Day, R. J.; Unger, S. E.; Cooks, R. G. Anal. Chem. 1980, 52. 557A. (4) Grade, H.; Cooks, R. G. J. Am. Chem. SOC.1978, 100, 5615-5621. (5) Ross, M. M.; Coiton, R. J. Anal. Chem. 1983, 55, 150-153. (6) Ross, M. M.; Colton, R. J. Anal. Chem. 1983, 55, 1170-1171. (7) DePauw, E. Anal. Chem. 1983, 55, 2195-2196. (8) Buchanan, A. C., 111; Livingston, R.; Dworkin, A. S.; Smith, G. P. J . Phys. Chem. 1980, 84, 423-427. (9) Sorlie, M.; Smith, G. P.; Norvell, V. E.; Mamantov, G.; Klatt, L. N. J . Electrochem. SOC. 1981, 128, 333-338. (IO) Perkampus, H.-H.; Schonberger, E. 2.Nafurfofsch., 8 1978, 31, 475. (11) Todd, P. J.; Gllsh, G. L.; Christie, W. H. I n t . J . Mass Spectrom. Ion Phys ., in press. Russell, D. H.; Smith, D. H.; Warmack, R. J.; Bertram, L. K. I n t . J . Mass Spectrom. Ion Phys. 1980, 35, 381-391. Weast, R. C., Ed. "Handbook of Chemistry and Physics," 64th ed.; Chemical Rubber Company Press: Boca Raton, FL, 1983. Miner, C. S.;Dalton, N. N. "Glycerol"; Reinhold: New York, 1953; p 266. I
\ ,
20
I
30
40
50
60
70
80
90
Temper a t u re
Figure 3. IT/125evs. temperature, where pyrene radlcal cation.
I is the Intensky of
the
20% of that generated from the pyrene/SbC13 sample a t 21 "C. Upon heating, the intensity of the pyrene molecular secondary ion was observed to increase by a factor of 3.5; however, the intensity of molecular ions at any temperature never equaled that produced from the pyrene/SbC13sample at room temperature. As before, at higher temperatures and longer time after introduction, signal decreased due to sample depletion. Furthermore, the probe tip appeared charred upon completion of the experiment, indicating significant sample damage. Attempts to generate pyrene molecular ions from a 4 mol % pyrene/glycerol mixture failed; we attribute this to component insolubility. SbC1, clearly enhances secondary emission of molecular pyrene ions. Proximity of SbC1, to pyrene molecules in solution may assist ionization by reducing the energy necessary to form the pyrene molecular ion, e.g., via formation of a charge-transfer complex. Such a mechanism is suggested by reports indicating that pure (i.e., zone refined) SbC13cannot oxidize pyrene (8). However, minor impurities, such as SbCl,
Gary S. Groenewold Peter J. Todd* Michelle V. Buchanan Analytical Chemistry Division Oak Ridge National Laboratory Oak Ridge, Tennessee 37831
RECEIVED for review April 2,1984. Accepted May 25, 1984. Research sponsored by the US. Department of Energy, Office of Basic Energy Sciences, under Contract No. DE-AC05840R-21400 with Martin Marietta Energy Systems, Inc., and by the US. Department of Energy Postgraduate Research Training Programs administered by Oak Ridge Associated Universities.
Reverse-Phase Polystyrene Column for Purification and Analysis of DNA Oligomers Sir: The rapid synthesis of DNA oligomers of defined sequence has been successively achieved by phosphotriester and phosphite triester methods on solid support (1,2).Now
the main effort in this area should be directed toward a rapid process for isolating the pure oligomer of desired sequence. There are, at present, two kinds of purification methods
0003-2700/84/0356-2253$01.50/00 1984 American Chemical Society
