138
Anal. Chem. 1982, 5 4 , 138-140
it would be better to call the value of Walthall et al. (3) as “literature value” instead of “most probable value”. The sample decomposition procedure used was that of Rantala et al. (4). In Table I1 the value shown in the % recovery column after f is that of the accumulated indeterminate error (expressed as one standard deviation) in the computed value of % recovery. In the 95% confidence level, the ranges of recoveries are as follows: for P b 81.1% to 115.9%; for Ni 79.9% to 118.7%; for Mn 87.7% to 104.5%. Abbey’s correspondence wrongly assumes that the % recovery values quoted by us and the other results cited by him are true values without any uncertainity in them. This is certainly not true. In Table 11, the f values in both the “most probable value” and “recovered value” are estimates of uncertainity. Abbey has presented results of analysis of USGS Marine Mud MAG-1 by other techniques without showing any estimates of uncertainty associated with these results. All of the results quoted by Abbey were obtained by instrumental methods of analysis and were based on analytical calibration curves. Hence, they are relative values based on comparison with standards and are not necessarily true values. Errors involved in analytical calibration curve methods depend on how closely the composition of samples matches with that of standards used for the analytical calibration curve, also on the particular instrumental technique, the element of interest, etc. Agreement among the results of various instrumental techniques does not necessarily indicate that the results are true values-they can be true values only in the absence of any determinate errors. It is also worth reproducing here a relevant quotation (5): “Bruce Kowalski of the University of
Washington makes a careful distinction between the concentration determined from a calibration curve and the truth. If you just want to do calibrations to determine concentrations, there’s not much chemometrics can do for you, according to Kowalski. But if you’re interested in the truth-how much of an analyte is really in the sample-then information science has lots to offer.” The paper under discussion describes our attempt made to do away with calibration curves in an instrumental analytical technique-The Capacitive Discharge Technique in Graphite Furnace Atomic Absorption Spectrometry. The future alone can tell whether or not we have succeeded in this attempt. LITERATURE C I T E D Chakrabarti, C. L.; Wan, C. C.; Harned, H. A.; Bertels, P. C. Anal. Chem. 1081, 53, 444. Chakrabartl, C. L.; Wan, C. C.; Harned, H. A,; Bertels, P. C. Nature (London) 1080, 288, 246. Walthail, F. G.; Dorrzapf, A. F., Jr.; Fianagan, F. J. Geol. Surv. Prof. Pap. ( U . S . ) 1078, No. 840, 99. Rantala, R.T.T.; Loring, D.H. At. Abs. Newsl. 1075, 74, 117. New Dlrections in Analytical Chernlstry” Anal. Cbern. 1081, 53, 703 A.
Department of Chemistry Carleton University Ottawa, Ontario, Canada K1S 5B6
C . L. Chakrabarti* C . C . Wan H. A. Hamed P. C . Bertels
RECEIVED for review July 8, 1981. Accepted October 2, 1981.
Rotationally Cooled Laser Induced Fluorescence Determination of Polycyclic Aromatic Hydrocarbons Sir: In recent years the development of new highly selective and sensitive methods for the characterization and determination of polycyclic aromatic hydrocarbons (PAHs) and their derivatives in complex mixtures has received considerable attention. High selectivity is associated here with the ability to distinguish between substitutional isomers of PAHs. Attainment of this selectivity with capillary column-gas chromatography-mass spectrometry for complex mixtures is very difficult and time-consuming. Alternative approaches are, therefore, required. Given that the majority of PAHs fluoresce with reasonable quantum yields and that high sensitivities are afforded by fluorescence detection, the possibility of developing highresolution fluorescence based techniques is attractive. This is all the more so if the technique’s selectivity does not rely on physical separation, e.g., chromatography. In this paper discussion is limited to such techniques. They must, therefore, afford fluorescence vibronic bandwidths of only a few cm-l. When low-temperature solids are employed as host matrices for PAHs or other analytes of interest the most general approach to achieving such line widths is laser excited fluorescence line narrowing (1-10). Given that fluorescence line narrowing spectrometry affords quasi-line fluorescence spectra for PAHs (as well as other
species, e.g., rare earth ions (11)),its selectivity will often be limited by the vibronic absorption bandwidths of the PAHs. When organic glasses are employed these bandwidths are typically ~ 3 0 cm-l 0 and are site inhomogeneously broadened. For Shpol’skii or rare gas matrices the bandwidths can approach -100 cm-’. In this paper we present data and discussion which show that rotationally cooled laser induced fluorescence (RC-LIF), a gas-phase technique, possesses a selectivity which is significantly superior to the solid-state methods discussed above. The enhanced selectivity results from the fact that absorption bands of rotationally cooled molecules seeded in supersonic jets (12) exhibit widths of about 1 cm-’. The availability of detailed spectroscopic information of rotationally cooled species under the collision-free conditions of these supersonic beams has led to a number of applications of RC-LIF (13). Molecules as large as ovalene (14) and free-base phthalocyanin (15-18) have been observed to exhibit absorption (obtained by photoexcitation) and emission spectra with line widths which are typically instrument limited and with only minor contributions from vibrational hot bands. It therefore seems reasonable to expect that RC-LIF can be of use in the analysis of complex mixtures of large organic molecules (6).
0003-2700/82/0354-0138$01.25/00 1961 American Chemical Society
ANALYTICAL CHEMISTRY, VOL. 54, NO. 1, JANUARY 1982
139
Flgure 3. Dispersed fluorescence spectrum by excitation of the
band
Flgure 1. Diagram of the RC-LIF scheme.
i
I
31.6
(32 458
cml-I) of naphthalene. Instrument resolution =
8'
1.25 A.
The focused laser beam intercepts the rotationally cooled beam 5 cm (-30 nozzle diameters) downstream of the orifice. The
'Ool
31
a
32.0 0 I C d *
32.2
32.4
32.6
,m-3
Figure 2. RC-LIF photoexcitation spectrum from a mixed sample of
naphthalene, a-methylnaphthalene, and /i?-methylnaphthalene.N[O0], denote their SI origin bands, respectively. Monitoring wavelength = 3425 A. Wand-pass = 50 A. a[Oo], and p [ O o ]
In the present work, we report our observations of the photoexcitation and dispersed fluorescence spectra of naphthalene and the a and 0isomers of methylnaphthalene from pure and mixed samples in a seeded supersonic expansion of helium. The ability to distinguish between these spectroscopically similar PAIHs, in particular the geometric isomers of methylnaphthalene, provides a test of the usefulness of this technique in supplying information suitable for ana1yt:ical determinations. The ipresent limitations of the technique are also discussed.
EXPERIMENTAL SECTION A diagram of the RC-LIF scheme is shLown in Figure 1. The output from a Quantal-Ray Nd:YAG pumped dye laser is frequency doubled by a IKDP crystal mounted in an Inrad Autotracker. The doubled output is focused into the supersonic expansion chamber where it crosses the rotationally cooled beam. Light baffles are mounted along the laser entrance and exit arms of the chamber to reduce scattered light. The exiting laser beam is then deflected into solution of rhodamine B where the resulting fluorescence is detected with a phototube, providing a reference for the laser intensity. The supersonic expansion chamber is similar in design to that described elsewhere (13). The chamber was fabricated from 4-in. stainless-steel tubing. Internal pressures of 0.1-5 mtorr are maintained during operation by two 4-in. oil diffusion pumps operating at 300 L s-l with a 50 L s-l mechanical forepump. The nozzle was made by heating an open end of 8-mm Pyrex tubing in a flame until a bulb with a pinhole was formed. The excess glass was then ground away leaving an orifice with a diameter of 160 i 10 wm.
resulting fluorescence propagating perpendicular to the two beams is collected through a 2-in. quartz window and imaged onto the slits of a 0.32-m htruments SA spectrometerwhere it is dispersed. The outputs from the signal and reference phototubes (both RCA 1P28-A) are sent to a pair of charge sensitive, integrating preamplifiers aind then to a Molectron LP-20 laser photometer, which averages the outputs separately. The outputs are then ratioed and sent to a strip chart recorder to record the normalized spectrum. Samples were prepared from naphthalene (Aldrich, 98%), a-methylnaphthalene (Aldrich,97%),and 0-methylnaphthalene (Fluka, 97%) without further purification. The room-temperature vapor pressures of the samples were diluted with He, maintaining a constant total pressure of 1.0-2.0 atm.
RESULTS AND DISCUSSION A portion of the photoexcitation spectrum obtained from a mixture of naphthalene, a-methylnaphthalene, and fl-methylnaphthalene is shown in Figure 2. The sample consisted of an equimolar liquid mixture of a- and &methylnaphthalene and a neat sample of solid naphthalene. On the basis of similar RC-LIF spectra obtained from pure samples, and from previous investigations (6,19),the features in the spectrum can be readily identified. With the exception of the weak, pressure-sensitive hot band structure which appears at the low excitation energy side of the major bands, all of the bands in the spectrum can be directly assigned to transitions originating from the vibrationless level of the ground states to specific vibronic levels in the SI states of the three species. Specifically, the features occurring at 31 702 cm-', 31 772 cm-l, and 32 020 cm-l were identified as the S1origins of @methylnaphthalene, a-methylnaphthalene, and naphthalene, respectively. Indeed, the spectrum appears directly as a weighted sum of the photoexcitation spectra from pure samples of the three constituents, showing no component interactions. No evidence of the formation of van der Waals molecules or polymers was observed (12,13). The absence of saturation broadening and other nonlinear laser intensity phenomena was confirmed by rerunning this and other spectra at reduced laser powers, with no observable changes. In Figure 2, the line widths are laser limited (laser bandwidth = 2 cm-'). Previously (6),using a laser bandwidth of 0.7 cm-l, we determined that the width of the band of naphthalene, occurring at 32 458 cm-l, is still laser limited although some structure is resolvable. Such line widths are typical, even at these low carrier gas pressures. Thus, Figure 2 represents a relatively low resolution photoexcitation
Anal. Chem. 1982, 5 4 , 140-142
140
spectrum. Despite this, the bands in the spectrum are well resolved and allow for easy identification of the three components of the sample. The identification of impurities in Upurensamples can be attained in this manner. In photoexcitation spectra of fresh samples of naphthalene, we have observed bands attributable to @-methylnaphthalene,indicating its presence as an impurity. Over a period of -7 h of operation, these bands disappear apparently due to degassing of the more volatile @-methylnaphthalenefrom the sample. In other samples, for instance, P-methylnaphthalene was doped with 250 ppm naphthalene. Photoexcitation spectra of such samples clearly show the presence of the dopant. The presence of impurities can also be determined from fluorescence spectra. Figure 3 shows the dispersed fluorescence spectrum from excitation of the 32 458 cm-l g1 band of naphthalene. This spectrum is unchanged whether obtained from pure naphthalene or from a trace amount of naphthalene in @-methylnaphthalene. If this spectrum represented an unknown impurity in ,f3-methylnaphthalene, it could be identified as such because it is impossible to interpret the spectrum in terms of &methylnaphthalene vibrations. Together with the photoexcitation spectrum of the sample, an unambiguous identification of the impurity can be made. The data presented show that even in the photoexcitation mode rotationally cooled spectrometry of a simple mixture can readily distinguish between substitutional isomers of PAHs. This will usually not be possible with solid-state techniques. On the basis of the linewidths in Figure 2 and the earlier discussion of inhomogeneously broadened vibronic bandwidths of PAHs imbedded in solids, it is clear that the selectivity of RC-LIF is significantly superior (10- to 100-fold) to that of fluorescence line narrowing spectrometry in organic glasses (4-6) or laser-excited Shpol'skii spectrometry (9, IO). We hasten to point out, however, that the selectivity of the latter two techniques is high and has allowed for the direct analysis of nonpolar PAHs in real samples (5,9).Recent work on polar derivatives of PAHs, e.g., amino-PAHs, has shown, however, that application of solid-state fluorescence techniques presents difficulties due to strong analyte-matrix interactions (20). This suggests that RC-LIF spectrometry will be important for the characterization of species prone to strong matrix interactions which obviate fluoresence line narrowing spectrometry. In the assessment of the analytical utility of RC-LIF spectrometry, its superior selectivity, adequate sensitivity, and absence of matrix effects must be weighed against the fact N
that quantitation presents a serious problem for real samples. We are now initiating studies to determine whether RC-LIF, as a stand alone technique, can be made semiquantitative. For RC-LIF to be quantitative it will be necessary to couple it with gas chromatography as has been discussed in earlier work (6).
LITERATURE CITED Personov, R. I.; AI'Shits, E. 1.; Bykovskaya, L. A. Opf. Commun. 1972, 6 , 189. Szabo, A. Phys. Rev. Lett. 1970, 25, 924. Szabo, A. Phys. Rev. Lett. 1971, 27, 323. Brown, J. C.; Edelson, M. C.; Small, G. J. Anal. Chem. 1978, 50, 1394. Brown, J. C.; Duncanson, J. A.; Small, G. J. Anal. Chem. 1980, 52, 1711. Brown, J. C.; Hayes, J. M.; Warren, J. A,; Small, G. J. I n "Lasers in Chemical Analysis"; Helftje, G. M., Lytle, F. E., Travis, J. C., Eds.; Humana Press: Cllfton, NJ, 1981; Chapter 12. Bykovskaya, L. A.; Personov, R. I.; Romanovskii, Y. V. Anal. Chlm. Acfa 1981, 125, 1. Shpolskil, E. V.; Bolotnikova, T. N. Pure Appl. Chem. 1974, 37, 183. Yang, Y.; D'Sllva, A. P.; Fassel, V. A,; Iles, M. Anal. Chem. 1980, 52, 1350. Yang, Y.; D'Silva, A. P.; Fassel, V. A. Anal. Chem. 1981, 53, 894. Gustafson, F. J.; Wright, J. C. Anal. Chem. 1979, 51, 1762. Levy, D. H.; Wharton, L.; Smalley, R. E. in "Chemical and Biochemical Applications of Lasers"; Academlc Press: New York, 1977; Vol. 2, Chapter 1. Smalley, R. E.;Levy, D. H.; Wharton, L. J. Chem. Phys. 1976, 64, 3266 and references thereln. Amirav, A.; Even, U.; Jortner, J. Chem. Phys. Lett. 1980, 69, 14. Fitch, P. H. S.;Wharton, L.; Levy, D. H. J. Chem. Phys. 1978, 69, 3246. Fitch, P. H. S.; Wharton, L.; Levy, D. H. J. Chem. Phys. 1979, 70, 2018. Fitch, P. H. S.; Haynam, C.; Levy, D. H. J. Chem. Phys. 1980, 73, 1064. Fitch, P. H. S.; Haynam, C.; Levy, D. H. J. Chem. Phys. 1981, 74, 6812. Stockburger, M.; Oatterman, H.; Klusman, W. J. Chem. Phys. 1975, 63, 4529. Chiang, I. M.S. Thesis, Iowa State Universlty, Ames, IA, 1981.
Ames Laboratory-USDOE Chemistry Iowa State University Ames, Iowa 50011
J o n a t h a n A. Warren J o h n M. Hayes Gerald J. Small* and Department of
RECEIVED for review August 14, 1981. Accepted September 28, 1981. Ames Laboratory is operated for the U.S. Department of Energy by Iowa State University under Contract No. W-7405-Eng-82. This research was supported by the Assistant Secretary for Environment, Office of Health and Environmental Research.
AIDS FOR ANALYTICAL CHEMISTS Generation of Dry Formaldehyde at Trace Levels by the Vapor-Phase Depolymerization of Trioxane Krlstln L. Gelsllng" and Robert R. Miksch Building Ventilation and Indoor Air Quality Program, Energy and Environment Division, Lawrence Berkeley Laboratory, University of California, Berkeley, California 94 720
Stephen M. Rappaport Department of Biomedical and Environmental Health Sciences, School of Public Health, University of California, Berkeley, California 94 720
Recent findings that low-level exposures to formaldehyde (HCHO) may produce allergenic effects (1,2)and that HCHO 0003-2700/82/0354-0140$01.25/0
may be carcinogenic ( I , 2) have led laboratories to undertake the measurement of parts-per-billion to parts-per-million 1981 American Chemical Socie!y
