ture. To increase the signal-to-noise ratio, the integration times can be extended without adverse effects. An instrument designed to be a reflection spectrometer would use source optics to increase the signal. Also, it might only be necessary to use several preselected wavelengths for monitoring, rather than the entire spectrum. In summary, the principle of remote monitoring of oil films on water using an active source and infrared spectroradiometer has definite possibilities. No physical contact between the sensor and the sample is required, although one would certainly consider damping the effects of weather by using a network of baffles below the source and spectroradiometer. Additional studies of the differences to be found from one oil to another, the effects of aging on the infrared spectrum, and
the actual monitoring of naturally occurring oil slicks are being carried out at this time. Acknowledgment
One of us (H. B. M., Jr.) would like to acknowledge support from the FWQA, Grant no. 16020 ELH. Literature Cited Hansen, W. N., J. Opt. SOC.Amer. 58,380 (1968). Mattson, J. S., Mark, H. B., Jr., Kolpack, R. L., Schutt, C. E., Anal. Chem. 42,234 (1970). Sommerfeld, A., “Optics,” Academic Press, New York, N. Y., 1954. Received f o r review June 12,1970. Accepted October 28, 1970.
Fluorometric Determination of Bromine : Application to Measurement of Bromine Aerosols Herman D. Axelrod, Joseph E. Bonelli, and James P. Lodge, Jr. National Center for Atmospheric Research, Boulder, Colo. 80302
A simple analytical procedure is described for the measurement of trace Br with application to the determination of Br aerosols and particulates. The Br sample (0.5 ml.) is added to a 9.5-ml. solution made from 9 ml. of 1.1 X 10-6M fluorescein in glacial HOAc and 0.5 ml. of 30% H202.The sample is allowed to react for 45 min., and the fluorescence is measured at 440 nm. excitation and 470 nm. emission. The fluorescence is quenched by the Br, and the quenching amount can be related to the Br concentration through sample standards. Atmospheric samples can be obtained from high-volume air filters. The filters are washed with water, and the liquid is analyzed for Br. Greater effective sensitivity might be achieved by use of glacial HOAc as a filter washing solution. A large excess of C1- in the sample does not interfere with the analysis. This method can be used to determine as little as 2 ng./ml. of Br- in the final analysis solution.
M
easurement of Cl/Br ratios in aerosols and airborne particulates has recently been of interest to scientists. The ratios vary, depending upon the source of C1 and Br. For example, automobiles with ethylene di-bromide gasoline additives can potentially alter natural Br levels. Duce, Wasson, et a/. (1963), Duce and Winchester (1965), Duce, Woodcock, et al. (1967), Lininger, Duce, et al. (1966), and Winchester and Duce (1967) have made worldwide measurements for Cl/Br/I ratios. Their principal analysis technique was the neutron activation analysis, and their procedure (Duce and Winchester, 1965) was a modification of an earlier method by Cosgrove, Bastian, et al. (1958). Saltzman (1961) recommended the reaction of bromine with neutral iodide as a means of measuring atmospheric bromine; however, the Saltzman method lacks specificity. Hunter and Goldspink (1954) developed a Br method involving the bromination of rosaniline, but this method was complicated. 420 Environmental Science & Technology
The neutron activation analysis procedures are good, but they require costly equipment and, occasionally, long analysis times. Feigl(l958) showed that the bromination of fluorescein made a sensitive spot test. We used Feigl‘s reaction as the basis of a fluorometric quantitative technique, and found that 2 ng./ml. of Br in the final analysis solution can be determined in a 10,000-fold mole excess of C1. Experimental
Reagents. Only AR-grade chemicals were used, and laboratory water was first passed through a de-ionizing column and then distilled. Fluorescence Measurement. Fluorescence measurements were made with a Perkin-Elmer Model 203 spectrofluorometer equipped with a xenon lamp source (continuous spectrum). Quartz cells (1-cm.) were used with the excitation wavelength set at 440 nm., and the emission wavelength set at 470 nni. All measurements were made in an air-conditioned room, 23°C. Analysis Procedure. For Br- concentrations of 0 to 2 X M (0 to 160 ng./ml.) in the final solution, 1.1 x 10-6M fluorescein in glacial HOAc was prepared. (This range is for monobromides in solution.) To 9 ml. of the fluoresceinHOAc solution, Hz02and sample were added so that the H202 final concentration was above 1%, and the total volume of water added was no more than 1 ml. The sample was allowed to react at room temperature for 45 min. The fluorescence was measured and the Br value calculated from the analytical curve. The 100% blank was made from the fluorescein-glacial HOAc solution, Hz02,and water. The standards were prepared from NaBr and treated identically to the samples. Results and Discussion
Fluorescein Spectra. Figure 1 shows the excitation and emission spectra for 10-6M fluorescein. At this concentration, the excitation peak is at 440 nm. and the emission peak is at 470
nm. (spectra were not corrected for instrumental factors). The spectrum, however, is concentration-dependent. For 1O-W fluorescein in 90% HOAc, 1.5% HzOz, the excitation spectrum shows two peaks, at 405 and 465 nm., and the emission is somewhat flat and broad at 505 nm. The brominated product is nonfluorescent so that the subsequent reduction of a high fluorescein concentration will shift the emission peaks. However, this shift does not prevent the use of high fluorescein concentrations for the analysis of large amounts of Br. Analytical Curve and Stoichiometry. For 10-6M fluorescein in 90% HOAc, 1.5% HzO?,the fluorescence suppression is linear with increasing Br concentration down to 0% fluorescence. At this fluorescein concentration, the slope of the analytical curve indicates a stoichiometry of 2 Br/fluorescein. Whether a disubstituted molecule is required for total supression is not clear, since the fluorescein can be brominated in as many as four positions in the upper rings (Figure 2). Sensitivity and Precision. Using 10-6M fluorescein, 1 X 10-*M Br (1.6 ng./ml.) in the final analysis solution can be determined accurately. Higher fluorescein concentrations can be easily used for higher levels of Br. The HOAc concentration, however, must be 90% or above. Lowering the acidity lowered the sensitivity; other acids were tried but with little success. For maximum sensitivity, the H?Ozconcentration should be at least 1Z ; lower H 2 0 2concentrations cause incomplete reactions. The fluorescein concentration was lowered to lO-’M, but the reaction was incomplete with subsequent loss of sensitivity. The precision was determined by analyzing 10 samples each of 80 ng./ml. and 20 ng./ml. Br in the final solution. The standard deviation for each was 2.1 % and 2 . 0 z absolute, respectively, on an arbitrary 0 to 100% fluorescence scale.
I
t
I
-Excitation
>
t cn
z
W k
z
w V
z w
0
cn W
(z
0
3 LL I
3 1
400
t
500 WAVELENGTH nm
Figure 1. Excitation and emission spectra of 10+M fluorescein in 90% HOAc, l.5yOHzOz
Table I. Materials Investigated for Interferencea Substance Error, (100-fold mole excess) over Br so3*0 s04’+5 I-20
z
~
0
~
3
f5
-
NOzNos-
-15 0 0 0 0 0 0 -15 0 0 0
c1- *
K+ NH4f Mg2+ Ca *+ Fe z+ Fe 3+ Cut+ Pb 2+ HCOOH CHsCOCH3 HCHO CH3CHzOH Ethylenediamine
+5 -5
a 10-6Mfluorescein, 10-6M Br-, 90% HOAc, 1.5% *.10,000-foldmole excess over Br-.
0 0 0 HzOZ.
The percentage fluorescence suppressions for the above samples were at each end of the scale, indicating that the standard deviation is 2 % absolute suppression along the entire scale. Reaction Time. The reaction was continued to what was believed to be completion, 45 min. after sample addition, (After only 30 min., the reaction is 90% complete, but good results require an additional 15 min.) The samples will then show the same fluorescence reading for at least an additional 4 hr. Therefore, there is no need to measure a sample at a precise time. Interferences. The interferences investigated are listed in Table I. All of them were in 100-fold mole excess over Br-, except C1- (10,000-fold mole excess). Some ions did cause problems, but it is unlikely that such large excesses would be found in atmospheric samples with the C1 and Br. The Iinterference probably resulted from the solution being yellow from the formation of 12. An interesting effect was noted: Chloride in large excess did not interfere. Iron(I1) did interefere, but when Fe(I1) was present with an equal C1- concentration, the Fe(I1) interference effect was 10% greater. Analysis of Filters. High-volume air filters were taken at two selected sites in the Panama Canal Zone region (Lodge and Pate, 1966). A small section of glass-fiber filter was washed with refluxing water to dissolve any collected Br-. Two samples obtained at Ft. Sherman (near a harbor) showed the air to contain 7.0 and 5.4 ng./mn3Br. Albrook Forest samples (jungle) had Br levels of 2 and 0 ng./m.3 The Ft. Sherman values are about a factor of 3 lower than the Hawaiian air values reported by Winchester and Duce (1967). The difference could be attributed to the fact that Ft. Sherman is located near very calm water while the Hawaiian samples could have been taken near typical ocean turbulance. The use of water for filter extraction can present problems because of the need to keep HOAc concentrations above 90 %.
CYooH ocooH I
FLUORESCE IN
TETRABROMOFLUORESCEIN
Figure 2. Fluorescein and a possible brominated product
Volume 5, Number 5, May 1971 421
This restricts the sample size, resulting in a lowered effective sensitivity. An alternate technique could be the use of glacial HOAc as the extraction solvent. If the material were dissolved directly into the acid, the reagent concentration could be adjusted accordingly, and considerable sensitivity could be gained. Because the Br determination can be accomplished in the presence of large excesses of C1, this method, coupled with a C1 analysis technique, can be used for measuring Cl/Br ratios. Acknowledgment
The authors thank John B. Pate for providing the highvolume air filter samples. Literature Cited
Cosgrove, J. F., Bastian, R. P., Morrison, G . H., Anal. Chem. 30,1872 (1958).
Duce, R. A,, Wasson, J. T., Winchester, J. W., Burns, F. J., J. Geophys. Res. 68, 3943 (1963). Duce, R. A,, Winchester, J. W., Radiochim. Acta 4, 100 (1965). Duce, R. A., Winchester, J. W., VanNahl, T. W., J. Geophys. Res. 70. 1775 (1965). Duce, R. A , , Woodcock, A. H., Moyers, J. L., Tellus 19, 369 (1967). Feigl, F., “Spot Tests in Inorganic Analysis,” 5th ed., Transl. R. E. Oespar, Elsevier, New York, 1958, p. 262. Hunter, G., Goldspink, A. A,, Analyst 79,467 (1954). Lininger, R. L., Duce, R. A., Winchester, J. W., Matson, W. R., J . Geophys. Res. 71, 2457 (1966). Lodge, J. P., Jr., Pate, J. B., Science 153,408 (1966). Saltzman, B. E., Anal. Chem. 33, 1100 (1961). Winchester, J. W., Duce, R. A., Naturwissenschaften 54, 110 (1967). Received for review June 23, 1970. Accepted September 25, 1970. The National Center for Atmospheric Research is sponsored by the National Science Foundation.
Nitric Acid and the Nitrogen Balance of Irradiated Hydrocarbons in the Presence of Oxides of Nitrogen Bruce W. Gay, Jr. and Joseph J. Bufalini Division of Chemistry and Physics, National Air Pollution Control Administration, Environmental Health Service, US. Public Health Service, US. Department of Health, Education, and Welfare, Cincinnati, Ohio 45226
In the photooxidation of systems containing hydrocarbons and nitrogen oxides in air, nitrogen balances have been poorLe., the amount of nitrogen consumed cannot be accounted for as products. Often, however, only gas-phase organic nitrogen products are considered; when surface-adsorbed products are analyzed, the nitrogen balances are greatly improved. Nitric acid is the principal surface product. It is believed to be formed primarily by hydrolysis of a nitrogen pentoxide intermediate on wall surfaces.
L
aboratory studies of the photooxidation of hydrocarbons in the presence of nitrogen oxides have satisfactorily explained many characteristics of photochemical smog. Many investigators agree concerning hydrocarbon reactivities, types of carbon-containing products, and amounts of oxidant formed in irradiations of particular hydrocarbon systems. Good carbon balances have been reported for irradiated systems in which the hydrocarbons were ethylene (Altshuller and Cohen, 1964), propylene (Altshuller et al., 1967), and 1-butene and trans-2-butene (Schuck and Doyle, 1959; Tuesday, 1961). In the past few years, greater interest has focused on nitrogen balance. Most determinations have not accounted for all of the nitrogen. In the photooxidation of propylene, Altshuller et al. (1967) reported that only 35 to 70% of the nitrogen consumed could be accounted for as products. With ethylene, only 13% of the nitrogen could be accounted for as methyl nitrate and nitrogen dioxide (Altshuller and Cohen, 1964). Similar results were observed with 422 Environmental Science & Technology
aromatics. With toluene, Altshuller et al. (1970) could explain only 10 to 20% of the nitrogen consumed as peroxyacetyl nitrate (PAN). With m-xylene, 10 to 75 % of the nitrogen originally present could be accounted for as PAN. In the work reported here, this lack of nitrogen balance has been resolved for various hydrocarbon-nitrogen dioxide systems. One earlier attempt to explain the nitrogen balance was in terms of a mechanism for the formation of molecular nitrogen. The data reported indicated that irradiation of nitrogen dioxide in a system with ethylene resulted in the formation of molecular nitrogen (Bufalini and Purcell, 1965). We performed studies to confirm these earlier findings and to extend the analyses to other hydrocarbon systems. The experimental results do not confirm the earlier findings concerning molecular nitrogen, but they do account for most of the nitrogen in the systems. Experimental
For the initial molecular nitrogen experiments, a 22-liter borosilicate flask, having a silvered outer surface and a double-walled borosilicate well, was used as a static reactor. A high-intensity Hanovia no. 679A36 mercury lamp fitted into the well and was cooled by flowing water between the double walls of the well. Molecular nitrogen and N20 were analyzed by vapor-phase chromatography with a helium photoionization detector. Other irradiations were made in the chamber described earlier (Altshuller and Cohen, 1964). This chamber is fitt$d with GE-F42-T6 black lamps with energy maximum at 3660 A. Temperature was maintained at 25 “C f 2 O during the irradiation. A plastic bag of 100 liters, fabricated from Teflon FEP
