1758
Anal. Chem. 1980, 52, 1758-1761
again and additional layers of adsorbate are being formed on the surface. One milliliter of air a t room temperature containing 14 ppm of nicotine vapor represents a vapor pressure of 0.01 mmHg of nicotine and a weight of 90 ng. Saturated vapor pressure of nicotine a t room temperature is about 0.2 mmHg. Thus, the relative pressure of 1 4 ppm nicotine in air is 0.01/0.2 or 0.05, a value sufficiently large t o indicate that most of the monolayer is completed. Therefore, a 14-ppm gas sample of nicotine may be considerably depleted by adsorption of nicotine on the walls of the container (e.g., a gas syringe). I t will be clear that this situation becomes more severe with decreases in saturated vapor pressure of the compound a t room temperature. On the other hand, once the nicotine has been injected into the gas chromatograph, losses due t o physical adsorption are negligible because of the very small value of t h e relative pressure of nicotine (e.g., the vapor pressure of nicotine a t 200 "C is almost 400 mmHg making the relative pressure of 14 ppm 0.000025). This is usually of no importance with respect to adsorption, although other effects such as absorption, chemisorption, and so on may be operative. For example, the large quantities of nicotine found on poly(viny1 chloride) Tygon tubing (Table 11) indicate that nicotine is absorbed (dissolved) by poly(viny1chloride) rather than merely adsorbed on its surface. CONCLUSION It was found that the major source of our inability to achieve quantitative recovery was adsorption of nicotine on surfaces located predominantly outside the gas chromatographic instrument and, also, that this adsorption and resulting loss of nicotine could be suppressed and eliminated by addition of NHIOH t o all nicotine solutions. Reliable determinations of nicotine in liquids and gases in concentrations as low as 2-5
pg injected onto the column can be made by gas chromatographic analysis when precautions are taken to minimize adsorption during sample preparation. The interferences caused by the adsorption phenomena described in this paper are probably of a general nature and operative for many other chemical compounds. This means that commonly used vapor sampling techniques, particularly those used for gas chromatographic purposes, may be inadequate in low concentration regions. ACKNOWLEDGMENT The authors gratefully express their appreciation to Robert Weker and Frederick Laidlaw for their technical assistance. LITERATURE CITED (1) Burrows, I. E.; Corp, P. J.; Jackson, G. C.; Page, 8. F. J. Analyst (London) 1971, 96, 81-84. (2) Beckett, A. H.; Triggs, E. J. Nature(London) 1988, 211, 1415. (3) Isaac, P. F.; Rand. M. J. Nature (London) 1972, 236, 307-31 1. (4) Feyerabend, C.; Levitt, T.: Russell, M. A. H. J. Pharm. Pharmcol. 1975, 27, 434-436. (5) Hinds, W. C.; First, M. W. New Engl. J . Med. 1975, 292, 844-845. (6)Russel, M. A. H.; Feyerabend, C.; Cole, P. V. Br. Med. J . 1978. 1 , 1043-1046. (7) Hengen, N.; Hengen, M. Clin. Chem. ( Winston-Salem, N.C.) 1978, 24. 50-53. (8) Bluthgen, A.; Nijhuis. H.; Hamrnan, J.; Heeschen, W. Milchwissenschaft 1978, 33, 137-141. (9) Dow, J.; Hall, K. J. Chromatogr. 1978, 153, 521-525. (10) Weast, R. C., Ed., "Handbook of Chemistry and Physics", 56th ed.; Chemical Rubber Co. Press, Cleveland, OH, 1975, p D-202. (11) Maskarinec, M. P.; Harvey. R. W.; Caton, J. E. J. Anal. Toxicol. 1978, 2, 124-126.
RECEIVED for review December 26, 1976. Accepted June 26, 1980. This research was supported in part by a contract from the National Cancer Institute and Enviro Control, Inc. (RFP-SHP-77-131).
Determination of Carbon in Atmospheric Aerosols by Deuteron Activation Analysis Mark Clemenson, Tihomir Novakov, and Samuel S. Markowitz" Department of Chemistry and Lawrence Berkeley Laboratory, University of California, Berkeley, California 94 720
Nuclear reactions induced by 7.6-MeV deuterons are used to determine total carbon in atmospheric aerosols. The "C(d,n)"N reaction produces the radionuclide 13N, a 10.0-mln posltron emitter, which is detected by its 0.51 1-MeV annihllation radiatlon. The detection system Is a Ge(LI) y-ray spectrometer. The method is nondestructive of the sample, permlttlng the sample to be studied by additional methods. Comparison of carbon found by deuteron actlvation analysis with that found by independent but destructive combustion methods shows a standard deviatlon of 10 YO for 15 samples analyzed over a wide range of carbon contents. The detection limit is estimated to be 0.5 pg/Cm2, corresponding to a carbon concentration of 0.2% in a sample of total thickness 250 pg/cm2.
Carbon has been found to be a major elemental constituent of urban particulate pollution ( 1 , 2 ) . The determination of the origin of the carbon in terms of primary and secondary 0003-2700/80/0352-1758$01 .OO/O
components is a difficult problem, to which a solution is necessary if meaningful control strategies are to be implemented. Some work has been done, but much more work is required (3-6). Of important analytical concern is the nondestructive determination of the total carbon content of atmospheric aerosol samples. The common method of carbon determination in atmospheric aerosol samples is by combustion analysis. This method, however, is destructive of the sample and thus does not allow other analyses to be performed on the same sample. Various nondestructive methods can be used for the determination of elemental concentrations in atmospheric aerosols. The most commonly used methods are X-ray fluorescence analysis and neutron activation analysis; neither is suitable for elements of low atomic number. The X-ray fluorescence method is of great importance for sensitive and nondestructive determination of elements as light as sulfur, but for elements lighter than sulfur (2= 16) two effects limit its usefulness: (a) the fluorescence yield drops to the range of a few percent for these elements and (b ) there are large 0 1980 American Chemical Society
ANALYTICAL CHEMISTRY, VOL 52, NO 11, SEPTEMBER 1980
X-ray absorptive effects because of the very low energy of the X-rays (KO.5 keV). Neutron activation analysis for low-2 elements is limited by unfavorable nuclear properties of the important nuclides in this area. The thermal-neutron absorption cross sections are very small for the important reactions. T h e induced radioactivities are also unsuitable for counting because of very long or very short half-lives. A new method for the nondestructive determination of carbon and other elements of low atomic number in atmospheric aerosols has been developed by Macias and co-workers ( 7 , 8 ) . Their method involves the in-beam measurement of y-rays from the inelastic scattering of 7-MeV protons accelerated in a cyclotron. This type of analysis, however, does require lengthy use of accelerator time. Other methods for nondestructive determination of carbon have been developed by Gordon and co-workers (9) and by Gladney and co-workers (IO). These methods involve the measurement of prompt y-rays following neutron capture. A nuclear reactor is the source of neutrons. Relatively large samples (-1 g), however, a n d long irradiation periods are required depending on the circumstances. An excellent literature survey dealing with chemical analysis by thermal-neutron-induced capture y-ray spectrometry has been prepared by Gladney ( 1 1 ) . A new activation-analysis method for the determination of carbon in atmospheric aerosols which uses only a short amount of beam time (2 min) will be described here. This method has already been used to determine nitrogen in aerosols (12), and future work will center on the development of a method for t h e determination of oxygen in aerosols. The use of activation analysis for elemental determinations has reached into almost every field where sensitive and nondestructive analyses are required. Charged-particle activation analysis is one method that may be strongly suited for the determination of carbon and other elements of low atomic number in samples of atmospheric aerosols. These samples are usually lightly loaded (low in mass) because of the low concentration of particulate matter in the atmosphere; therefore methods with excellent sensitivity such as activation analysis are required. T h e general principles of chargedparticle activation analysis have been discussed by Markowitz and Mahony (13). Advances in the field have been reviewed by Lyon a n d Ross ( 1 4 ) . EXPERIMENTAL SECTION Targets. The targets used to determine the %(d,r~)’~N excitation function were 2-mg/cm2polystyrene foils, (C8H8),. The targets were 2.2 cm in diameter, and the deuteron beam was collimated into a 1.3-cm diameter central spot. Polystyrene was selected for its purity and high carbon content of 92.3% C. Irradiations. The Lawrence Berkeley Laboratory 88-in. cyclotron facility was used for the irradiations. An initial deuteron beam energy of 15.0 MeV was selected. Aluminum foils were used to degrade the beam from the initial energy to the desired energy. The range-energy tables of Williamson et al. (15) were used to calculate the required aluminum thickness. The targets were irradiated at different energies by the stacked-foil technique. The targets were typically irradiated for 2 min at an average beam intensity of 0.5 FA. The total charge received by the Faraday cup was measured by an integrating electrometer. Counting. The irradiated samples were analyzed by detecting the 0.511-MeV positron-annihilation radiation of 13N. The counting system consisted of a high-resolution Ge(Li) detector and fast electronics coupled to a 4096-channel computer-controlled analyzer. The data were recorded on magnetic tape for later analysis. The Ge(Li) detector used for this work has an active volume of 60 cm3. The resolution of the detector is 2.0 keV (full-width at half-maximum) at the 1.3325-MeV y-ray of 6oCo. The use of this system allowed the collection of y-ray information up to 2.0 MeV with excellent resolution. The information obtained permitted a more complete indentification of other radionuclides produced during the irradiation and also protected against y-ray interferences that might not be detected with a low-resolution
1759
-
03--_-
“C(d,n) 3N
b
-I 1301
:
,
I
I
I
,
,
, 14
Deu‘ero- eiergy \MeV’
Figure 1.
Excitation function for the 12C(d,n)13N reaction.
system such as a NaI spectrometer (6-7% resolution). In routine use a simple counting system consisting of a NaI detector and a single-channel analyzer could be used after the high-resolution system has demonstrated that there are no y-ray interferences with the 0.511-MeV annihilation radiation. The decay of the 0.511-MeV annihilation-radiation activity of an activated (Ed= 8.6 MeV) polystyrene target is a single component with a half-life of 10.0 min corresponding to the decay of 13N in the target. The samples were counted approximately 5 cm from the face of the Ge(Li) crystal. In this geometry, the 0.511-MeV y-ray overall detection coefficient (ODC) was 0.97%. ODC = ( c / m )in the photopeak/(d/rn) from the standard source. It was determined with a 22Nacalibrated standard, obtained from the International Atomic Energy Authority, Vienna. The decay curves were analyzed with the CLSQ computer code (16). RESULTS E x c i t a t i o n F u n c t i o n . T h e absolute cross sections at several energies were determined for the 12C(d,n)13Nreaction, with the polystyrene foils (Figure 1). The excitation function measured in 1959 by Brill’ and Sumin ( I 7) agrees well with our result. The estimated accuracy of the excitation function is approximately 5%. The production of 13Nfrom 13C (1.11% abundance) was neglected. T h e uncertainties in the measurement include: (a) statistical fluctuations in the counting rates, (b) decay-curve component resolution, (c) determination of the integrated charge, (d) measurement of the weight of the polystyrene foils, and (e) the determination of the overall detection coefficient. I n t e r f e r e n c e s . Two types of interferences that must be considered in charged-particle activation analysis using y-ray spectroscopy to detect the 0.511-MeV annihilation radiation are (a) the production of positron-emitting activities of similar half-life to the activity of interest and (b) the production of the activity of interest from elements other than the one under analysis. Because the annihilation radiation is only characteristic of a positron decay event, all positron-emitting nuclides will contribute to the 0.511-MeV radiation. T h e activities must be separated by their half-lives in a decay curve analysis. The possible interfering radionuclides must be identified and their importance carefully investigated. The production of positron-emitting activities of similar half-life to 13N ( t l j z= 10.0 min) is not a serious problem with the carbon determination because the 13N activity is by far the dominant positron-emitting activity; it is easily resolved in the decay-curve analysis. T h e production of the radionuclide of interest from a n element other than the one under analysis is the other type of interference that must be carefully investigated. T h e reactions of the incident particle on neighboring elements are the most common interference of this type. T h e interfering reactions can sometimes be avoided by selection of the incident
1760
ANALYTICAL CHEMISTRY, VOL. 52, NO. 11, SEPTEMBER 1980
Table 1. Nuclear Reaction Thresholds for the Reaction of Interest and the Principal Interfering Reactionsa
reaction I2C(d,n)l3N I4N(d,t)''N
threshold, MeV 0.33
4.91
reaction 14N(d,dn)"N 160(d,an)'3N
threshold, MeV 12.06 8.37
a The low atomic abundance (1.1%) of 13Cand the threshold of 7.19 MeV preclude interference from the reaction I3C(d,2n)I3N. Chmrel r m : e
particle energy to be below the reaction threshold. The possible interfering reactions of importance are listed in Table I. The reaction used for the determination of carbon is listed first. If the sample is irradiated with deuterons of energy less t h a n 8 MeV, the 14N(d,dn)13Nand 160(d,an)13Ninterfering reactions will be energetically forbidden. The 14N(d,t)13N reaction cross section was measured (with GaN as target) a t a deuteron energy of 7.6 MeV and found to be 1.3 mb. This cross section is much smaller than the 12C(d,n)13Ncross section at this energy (a = 70 mb). Furthermore, typical atmospheric aerosols contain 1-8 times as much carbon as nitrogen. Nitrogen, therefore, does not constitute an interference. The aerosol samples were irradiated a t a deuteron energy of 7.6 MeV. T h e use of this region of the excitation function minimizes possible cross-section fluctuations which might occur from small changes in the particle energy. Aerosol Sample Analyses. The atmospheric aerosol was collected on a silver-membrane filter with the use of a vacuum pump. Silver filters are the best commercially available filter for the analysis of aerosol samples using deuteron activation analysis. Because silver has a high atomic number, the Coulomb barrier can be used to minimize deuteron reactions on t h e filter material. T h e relatively massive filter (40 mg of Ag/cm2) is only slightly activated. This permits the most sensitive detection of the y radiation from the activation of the aerosol itself. T h e deuteron energy of 7.6 MeV that was used for the irradiation of the filter sample is below the Coulomb barrier of 8.4 MeV for deuterons plus silver. Carbon-containing filters such as Teflon and Millipore are unsuitable, of course, for the collection of the particulate because carbon is being determined in the aerosol. Oxygen-containing filters such as quartz are unsuitable because large amounts of I8Fare produced from the 180(d,2n)18Freaction, notwithstanding t h e low (0.2%) abundance of l80. T h e production of 109.8-min 18F in large amounts would interfere with the detection of 10.0-min 13N produced from carbon in the relatively small mass of aerosol deposited on the filter. Unfortunately, silver is relatively expensive. T h e typical loading of particulate on the silver filter was approximately 250 kg/cm2. A stacked-foil arrangement was used for the irradiations. In the stack was a polystyrene foil t h a t was used as the carbon standard, a filter sample, and aluminum foils. T h e aluminum foils were used to degrade t h e beam energy t o the desired value. The polystyrene standard was irradiated at a deuteron energy of 8.6 MeV. The filter sample was irradiated at a deuteron energy of 7.6 MeV. T h e stack was typically irradiated for 2 min a t a beam intensity of 0.5 PA. T h e decay of the 0.511-MeV annihilation radiation was followed for 2 to 3 h. A typical y-ray spectrum of an aerosol sample is shown in two parts in Figures 2 and 3. Figure 2 shows the region between 0 and 1 MeV. Besides the annihilation radiation, several y-rays are present. Most of these y-rays are a result of the activation of the silver filter. The y-ray at 844 keV, however, is due to the production of 27Mg by the 27Al(d,2p)27Mg reaction. The 27Mgactivity is a result of deuteron reactions involving the aluminum in the particulate matter and in the aluminum foil used to degrade the
Figure 2. The region from 0 to 1 MeV in the y-ray spectrum of an aerosol sample following deuteron irradiation. 134.
Aerosol gamma roy spectwm
Channei lumber
Figure 3. The region from 1 to 2 M e V in t h e y-ray spectrum of an aerosol sample following deuteron activation. ~. .
.
~
-
-
-
.
4erosol decay curve 511 keV
, \y 104-
5E
f
y o
m,!.
>
-
T
2
\.
- ' :1
$ompasite ,,2: 'IZrnin (liomin IF' + 6 5 h r "Cd)
.
---
"2
0
k>
-y
