Isotopic Determination of Oxygen-18 in Gases by Neutron Time-of-FlightSpectrometry Max Peisach, Re& Pretorius, and Paul J. StrebelV Southern Universities Nuclear Institute, P.O.Box 17,Faure, C.P.,South Africa Oxygen-18 was determined in gases by time-of-flight spectrometry of the prompt neutrons from the reaction 180(p,n)18F roduced by a pulsed beam of protons of 5.0 MeV. h !e analysis is nondestructive and, with a beam current of 0.5 to 1.5 PA requires from 5 to 30 minutes per sample. The method is applicable to isotopic concentrations from 0.204 atom % (natural level) to 100%. The relative standard deviation was &3.8% and the sensitivity, 7.2 X 10-8 g/cm2.
MANYNUCLEAR METHODS for determining oxygen-18 have been developed, not only because of the use of this isotope as a tracer for oxygen, but also as a means of determining oxygen, the most abundant isotope of which has less favorable nuclear activation properties. Activation analysis has been carried out using the activities of fluorine-18, generated either by proton irradiation (1) or by neutron-induced “knock-on” protons (2, 3), of oxygen-19, produced by thermal neutron capture (4,of carbon-15, generated by fast neutrons (5), or of nitrogen-17, produced by secondary tritons in the presence of lithium-6 by the successive nuclear reactions 6Li(n,a)3H and 180(t,a)17N induced by thermal neutrons (6). Analytical methods, in which the emission of prompt nuclear reaction products was measured, were developed using alpha particle emitted in the (d,a) reaction (7)or neutrons emitted in the (a,n) reaction (8). These latter methods have the advantage that the conditions of measurement are not determined by the decay properties of radioactive products. The technical difficulties connected with the irradiation of gaseous samples, with few exceptions, limit the use of methods such as described above, to liquids and solids. An attempt to determine oxygen-18 in carbon dioxide was reported (4) but the procedure involved the conversion of the gas to solid ammonium carbamate, which was then irradiated. There is thus need for an analytical method which could extend the advantages of nuclear methods to the determination of oxygen-18 in the gaseous phase. In this work an attempt was made to determine oxygen-18 nondestructively in gases by time-of-flight spectrometry of prompt neutrons emitted from the reaction 180(p,n) 18F induced by a pulsed beam of protons. This method has already been applied to the determination of deuterium (9) and other elements (10)using a pulsed deuteron 1 Department of Chemistry, University of Cape Town, Rondebosch, C.P., South Africa.
Present address, Graduate College, Princeton University, Princeton, N. J. 08540.
Table I. Neutron Energies from (p,n) Reactions Ep = 5.0MeV
Target *C
Natural abundance, 98.89 1.11 99.63
Q-value (P,no) MeV (1.4
170
99.759 0.037
-18.390 -3.004 -5.931 -3.543 -16.431 -3.544
‘80
0.204
-2.450
lac
4N “N 160
0.37
Neutron energy, MeV
... ...
1.941 (no) 1.381 (no)
...
1.390 (no) 0.854 (nl) 2.495 (no) 1.555 (nl) 1.445 (nz) 1.406 (n3) 1.361 (na) 0.744 (n5) 0.288 (ne)
beam and isotopes of calcium (11) using a pulsed proton beam. Gaseous samples in which the isotopic concentration of oxygen-18 is to be determined frequently contain the elements carbon and nitrogen, in addition to oxygen. The nuclear properties of the stable isotopes of these elements under the conditions of proton irradiation are of considerable interest and are listed in Table I. The Q-values of the (p,no) reactions of the stable isotopes and the natural abundances are given in the table. Using proton beams with energies less than 5.5 MeV, the highest energy attainable in this investigation, neutron emission from the very abundant nuclides I2C, 14N,and I 6 0 is energetically impossible. In the same table, values are given for the energies of neutrons expected to be formed by a proton beam of 5 MeV from the heavier isotopes of these elements, the neutron group nt corresponding to the ith excited state in which the product nucleus is left. The neutron energies given in the last column of Table I were calculated from the kinematics of the nuclear reactions concerned, as were reported previously (9). Many neutron groups are obtainable from oxygen-18 and there is no a priori reason to prefer any one for use for analytical purposes, except that the no neutrons from carbon-1 3 and nitrogen-15 have energies relatively close to those of the n3and n4neutrons from oxygen18 and hence could be a possible source of interference.
~~
EXPERIMENTAL
(1) B. A. Thompson, ANAL.CHEM.,33, 583 (1961). (2) L. H. Hunt and W. W. Miller, Ibid., 37,1269 (1965). (3) D. C. Aumann and H. J. Born, Naturwissenschafren, 51, 159 (1964). (4) G. J. Fritz, I. Han, and W. H. Ellis, Intern. J . Appl. Radiation Isotopes, 16, 431 (1965). (5) V. P. Guinn, Trans. Amer. N u d . SOC.,9,83 (1966). (6) S. Amiel and M. Peisach, ANAL.CHEM.,35, 323 (1963). (7) G. Amsel and 0. Smukowski, Compt. Rend., 251,950 (1960). (8) S. Amiel and A. Nir, “Radiochemical Methods of Analysis Vol I,” I.A.E.A., Vienna, 1965, p 287. (9) M. Peisach and R. Pretorius, ANAL.CHEM.,39, 650 (1967). 850
ANALYTICAL CHEMISTRY
Isotopically enriched oxygen-18, as chemically pure carbon dioxide containing 97.45 atom l80and 0.27 atom 10 ’ was supplied by Yeda Research and Development Co., (IO) M. Peisach, Chem. Communications, 1966, 632.
(11) W. R. McMurray, M. Peisach, R. Pretorius, P. van der Menve, and I. J. van Heerden, ANAL.CHEM.40,266 (1968). (12) F. Everling, L. A. Koenig, J. H. E. Mattauch, and A. H. Wapstra, “Nuclear Data Tables, Part I,” National Academy of Sciences, Washington, 1961.
Q6
NEUTRON ENERGV (MeV) 1.0 2.0 I
I
I
4.0 I
97.65 atomY, oxygen-18 as C%
Ep = 5.0 t W 600
400
200
Y
8 0
Figure 1. Neutron time-of-flight spectra of enriched oxygen-18, nitrogen-15, and carbon-13 Proton energy ( E p ) = 5.0 MeV Flight path = 3.13 m
Q $lo Q
5
20
lo
0-
0
Rehovoth, Israel. From this stock, samples were prepared containing known concentrations of oxygen-18 in natural oxygen, carbon dioxide, and nitrogen. The irradiation cell, 3 cm long with a volume of about 7 cc, and the gas handling apparatus, have already been described (9). A circular tantalum collimator, 3.5 mm in diameter, defined the cross-sectional area of the beam. However, the beam area is irrelevant because the number of counts recorded during an irradiation with a predetermined integrated quantity of charge, is independent of the beam area, The beam entered the irradiation cell through a thin nickel window and was stopped in the thick tantalum rear wall of the cell. Neither the nickel nor the tantalum could yield neutrons at the proton energies used in this investigation. Pulsed proton beams were obtained from the 5.5 MV Van de Graaff accelerator at the Southern Universities Nuclear Institute. Pulses were of 4-nsecond duration and
loo
200 CHANNEL NUMBER
400 nseconds apart. Low average beam currents of between 0.5 and 1.5 A were used to prevent damage to the thin nickel window. Most measurements lasted between 5 and 30 minutes per sample. The electronic equipment was the same as was used previously (9). Low level electronic noise was rejected by energy discrimination and pulses from gamma-rays by pulse shape discrimination. The neutron detection threshold was somewhat below 700 keV so that neutrons below this energy could not be observed. NEUTRON SPECTRA
Enriched Samples. Typical neutron time-of-flight spectra obtained from enriched carbon-13 as l3COt, nitrogen-15 as 15NH3 and oxygen-18 as T O Z with 5.0-MeV protons are shown in Figure 1. The energy of each neutron group is VOL 40, NO. 4, MAY 1948
851
300 CHANNEL NUMBER Figure 2. Typical background spectrum E p = 5.0MeV
Flight path = 3.13 rn about 150 keV lower than the corresponding values given in Table I because of the energy lost by the proton beam in passing through the nickel window. The spectrum obtained from oxygen-18 consisted of an isolated small single peak corresponding to the no neutron group and a large composite peak which included counts from the neutron groups nl, n2, na, and n4that could not be resolved under the conditions of the experiment. For analytical purposes it is not necessary to resolve these neutron groups, because the relative number of counts obtained from each is constant at constant incident proton energy and constant measuring angle. The small peak due to n6neutrons is just detectable, but the detection efficiency for such low energy neutrons is much smaller than for neutrons of about 1 MeV or higher, so that the height of the peak is not a true reflection of the relative reaction cross section. Background. A typical background spectrum as obtained from the irradiation of an empty gas cell with a beam of 5-MeV protons is given in Figure 2. This spectrum consisted of two relatively prominent peaks corresponding to 1.752and 1.328-MeV neutrons from carbon-13 and oxygen-18, respectively, probably because of deposits formed from residual vacuum oil vapors in the beam tube which decomposed at heated points of incidence of the irradiation beam on the cell window. In addition there was a low continuum of counts due to gamma-rays not entirely eliminated by pulse shape discrimination and neutrons scattered into the detector. Unlike the case for deuteron irradiation where all nuclidic components of carbon-, nitrogen-, or oxygen-containing gases generate neutrons at relatively low incident beam energies (23)) proton irradiation at comparable energies does not generate neutrons in such large numbers because of the highly endoergic nature of the corresponding proton-induced reactions on the more abundant lighter isotopes. Accordingly, the total number of neutrons produced in the gas cell from sources other than the nuclide under investigation, does not vary appreciably with the composition of the gas unless the gas contains other enriched isotopes of carbon, nitrogen, or (13) W. J. NaudB, M. Peisach, R. Pretorius, and P. J. Strebel, J . Radioanal. Chem. (in press).
852
ANALYTICAL CHEMISTRY
oxygen. Hence, the magnitude and spectral distribution of the background can be considered to be independent of the pressure and composition of the gas under analysis. CALIBRATION
The noneutrons from oxygen-18 are readily resolvable from all other neutron groups that can be generated in gases containing only carbon, nitrogen, and oxygen and would therefore appear to be most suitable for analytical purposes. However, the neutron yield from the (p,no) reaction induced by 5 MeV protons is so much lower than the yield from the combined (p,n) reactions leading to the first, second, third, and fourth excited states, that the integrated neutron count over the energy region covered by this combined group provides a better index of the oxygen-18 content of the gas. With 5 MeV protons, the appropriate energy region for integration lies between about 1.0 and 1.58 MeV. The integrated count in this region was used as a measure of the oxygen-18 content despite the fact that neutrons from nitrogen-15 and oxygen-17 might add to the total count. As the neutron count is proportional to the number of target nuclei in the path of the beam, it was possible to utilize the integrated neutron count obtained from a single gas sample, but measured over a range of pressures, as a calibration curve. The linear calibration curve obtained with a 5-MeV proton beam and approximately 10 atom oxygen-18 as W 0 2 over pressures ranging from 0.7 to 63.3 mm had a slope of 1080 counts per millicoulomb for 1 pg per cm* cross-sectional area of the beam. RESULTS OF ANALYSES
Table I1 lists the results of some determinations of oxygen18 in oxygen gas and in the presence of natural carbon dioxide and natural nitrogen with a proton beam of 5.0 MeV. The mean neutron count was 1091 counts per millicoulomb for 1 pg per cm2, which agrees with the calibration vahe, quoted above, within the precision of the method. The mean error, -0.045 pg per cm2, provides a measure of the accuracy of the method and the statistical analysis of the results shows that there is no bias. The relative standard deviation was
*3.82z.
Table 11. Some Determinations of Oxygen-18 Ep = 5.0MeV
Gas mixed with C W 2
Oxygen-18 content pg/cm2
Known A
Found B
Error (B-A)
0 2
4.44 9.95 15.09 20.69
4.08 9.71 15.14 21.40
-0.36 -0.24 $0.05 +O. 71
Nz
3.32 6.85 12.08 19.66 27.36 31.46
3.08 6.89 12.19 19.44 26.72 32.51
-0.24 +0.04 +o. 11 -0.22 -0.64 +1.05
coz
3.88 5.70 17.71 24.54 36.83
3.66 5.63 18.23 24.35 35.85
-0.22 -0.07 +0.52 -0.19 -0.98
Relative error, Z -8.11 -2.41 +O. 33 f3.44 -7.23 +O. 58 +0.91 -1.12 -2.34 +3.34 -5.67 -1.23 +2.94 -0.77 -2.66
Neutron counts per millicoulomb Observed N Per unit mass N/A 4639 10720 16584 23340
1045 1077 1099 1128
3555 7667 13399 21228 29088 35342 4187 6314 19921 26526 38944
1071 1119 1109 1080 1063 1123 1079 1107 1125 1081 1057
Mean error = -0.045 pg/cm2 Mean neutron count per pg/cm2 = 1091 for 1 mC Relative standard deviation = 3.82
x
PRECISION
From a knowledge of the background and the slope of the calibration line, the minimum amount of oxygen-I8 for which the neutron count could be determined with a precision of f3 % by irradiating with a total charge of 1 mC was found to be 1.76 fig per cm2 cross-sectional area of the beam. By increasing the total charge to 20 mC thereby increasing the number of counts and hence reducing its statistical uncertainty, about 280 ng/cm2 of oxygen-18 could be measured with the same precision. When a lower precision of & l o % was acceptable, this limit was reduced to 80 ng/cm2. SENSITIVITY
The background over the energy range of the nl, nz,n3,and n4neutrons from oxygen-18 amounted to about 700 counts per millicoulomb. Under these conditions the sensitivity limit for the detection of oxygen-18 with 1 mC total charge was 72 ng per cm2. The criterion selected for the calculation of the sensitivity was that weight of sample for which the net sample count was three times the standard deviation of the background count over the relevant energy interval. INTERFERENCES
When the determination has to be carried out in the presence of oxygen in which the oxygen-17 concentration is not enriched, interference from this nuclide is unlikely. However, it frequently happens that preparations enriched in oxygen-18
are also enriched in oxygen-17, albeit to a smaller extent. In such cases interference from oxygen-17 is likely. In the presence of carbon, oxygen-18 can readily be determined because none of the stable isotopes of this element can produce interfering neutrons. In the presence of nitrogen, interference from neutrons generated by nitrogen-1 5 may be expected when the concentration by weight of nitrogen-15 is about seven times that of oxygen-18. At the isotopic concentrations of oxygen-18 and nitrogen-15 found in nature, neutrons from nitrogen-15 would introduce errors in the determination of oxygen-18 when the nitrogen-to-oxygen atom ratio is greater than about 4:l. Other elements from which interference is energetically possible, when 5 MeV protons are used, are neon (from 22Ne) and sulfur (from 36S). The presence of neon in samples containing enriched oxygen-18 is improbable. Sulfur-36 occurs in nature in an isotopic concentration of 0.014 atom %; interference from this nuclide is thus expected to be small in unenriched samples. ACKNOWLEDGMENT
Acknowledgment is due to Bob NaudC who helped with the preliminary investigation. RECEIVED for review September 11, 1967. Accepted January 15,1968. This work was financially supported by the South African Atomic Energy Board. One of us (P.J.S.) was also supported by a bursary from the South African Council for Scientific and Industrial Research.
VOL 40, NO. 6, MAY 1960
853
