350 MELHUISH AND CHAPMAN: DETERMINATION OF [Analyst, VOl. 91 The Determination of Helium-3 in Argon at L’evels of BY K. R. MELHUISH AND H. CHAPMAN (Dounreay Experimeflntal Reactor Establishment, Thztrso, Caithness) The helium in the argon is first concentrated by removing the argon on an activated charcoal trap a t - 190” C. The helium is then transferred to a modified A.E.1. Ltd., MS2 mass spectrometer equipped with a Vibron amplifier, in which the volume of the helium-3 (and helium-4) is measured. With 1-litre samples the limit of detection is approximately 2 x p.p.m. by volume. Current experience on establishing the level of helium-3 in the argon blanket gas of the Dounreay Fast Reactor is outlined. THE measurement of the level of helium-3 and its day-to-day variation in the argon “blanket” gas of the Dounreay Fast Reactor has been carried out. Helium-3 is produced within the reactor from the radioactive decay of tritium, which is produced by two nuclear reactions: (i),.$Li (n, a ) :H on the lithium present as an impurity in the sodium - potassium metal (primary coolant), and (ii), ternary fission in the uranium fuel of the driver charge. The driver charge fuel elements are vented so that fission-product gases escape into the argon gas blanket. Calculations showed that the volume of helium-3 produced from tritium decay should be from about 0.3 to 0.6 ml per year. Therefore, as the argon blanket gas has a volume of about 108 ml, in order to detect the day-to-day change in the helium-3 level it was necessary to measure about 1 to 2 x lop5 p.p.m.by volume. By using refined techniques with a conventional mass spectrometer, the detection limit for the direct determination of helium in argon can be lowered to about 5 p.p.m. by volume and by a simple impurity-concentration technique1 to 0-05 p.p.m. by volume. Therefore, in order to achieve the required 1 to 2 x p.p.m. by volume, a superior concentration procedure was required and, as the sample size was limited to 1 litre, arising from the associated radioactive fission-product gases (radiation levels up to 1 Roentgen per hour per litre), improved spectrometric sensitivity was also required. The determination of 10-s-ml amounts of rare gas is usually only undertaken with special instruments and techniques, e.g., as used on rare gases from meteorites.2 EXPERIMENTAL MASS SPECTROMETRIC DEVELOPMENT- A standard A.E.I.Ltd., MS2 mass spectrometer and a Vibron Mass Spectrometer Amplifier Type 51A were made available for the work. By using the standard d.c. amplifier equipped with a 4 x 10lO-ohm resistor the limit of detection for helium was about 5 x lW5ml. The I‘ibron amplifier was then fitted (the Vibron head being fitted with a 1012-ohm resistor and 3-pF condenser). This lowered the detection limit to about 2 x 10P ml, but the noise level was appreciable and the time constant of the detection system long (about 3 seconds) for normal peak-scanning techniques to be used. Measurements had to be made with the instrument controls already “tuned” to the required mass number, and the stability of various voltage and current supplies was of paramount importance.For maximum sensitivity the mass-spectrometer ion-source voltages were adjusted (tuned) to the first ion repellor maximum on mass number 3 for the helium-3 determinations and on mass number 4 for early work with “natural” helium-4. To increase further the sensitivity of the mass spectrometer a 2-litre reservoir of the double-inlet system of the MS2 was replaced with a small reservoir of about 50-ml capacity. Corrections, however, have to be applied to the recorded results to allow for the pumping away of the gas through the mass spectrometer leak valve. (The leak has molecular flow characteristics, and hence the flow is proportional to M-, where M is the molecular weight of the gas concerned.) With these modifications, the detection limit should now be approxi- mately 5 x ml of helium.June, 19661 HELIUM-3 I N ARGON AT LEVELS OF 351 To establish the detection limit of the mass spectrometer, samples of air containing 5.4 p.p.m.by volume of helium were used. Aliquots (0-2 ml) were dispensed into the 50-ml reservoir, and the mass number 4 peak was measured. Five aliquots gave an average reading of 67.2 i 1.4 (2a) divisions on a 10-inch recorder chart (100 divisions = full-scale deflection) with a noise level of approximately 2 divisions. ml of helium-4 could be measured to within k2 per cent. ; 2 x 10 per cent. and the absolute detection limit, i.e., twice noise level, is about 2 x ml. Use of a calibrated helium leak also confirmed the accuracy of these absolute measurements.Taken in conjunc- tion with a 1-litre sample this gave a detection limit of 2 x The major drawback to the detection of helium-3 was that at mass number 3, the peaks from the hydrogen deuteride ion HD+ and, if sufficient hydrogen is present, the association ion H,+ also occur. To resolve the peaks in the mass doublet requires a resolving power of at least 500. Thus 1 x ml is measurable to within p.p.m. by volume. 3He : 3.01699 a.m.u. HD : 3.02289 a.m.u. The standard MS2 has a resolving power of about 200, with the collector slit a t 0.020 inch; this resolving power was insufficient to distinguish the two peaks at mass number 3 in a mixture of helium-3, hydrogen and deuterium (Fig. 1). Decreasing the collector slit width to 0.005 inch gave only sufficient resolving power to distinguish the two peaks (Fig.2). Decreasing of the collector slit reduced the sensitivity by a factor of four. In view of the poor resolution and the lowered sensitivity a hydrogen-removal stage was added to the concentration procedure, the collector slit being opened out again to 0.020 inch. It was found that, with volumes of air in excess of 0.2 ml, there was a distinct suppression effect on the mass-4 peak, probably arising from space-charge effects that caused de-focusing of the ion beam. This set the limit for the volume of concentrated gas from the concentration apparatus. Ion accelerating voltage Fig. 1. Peak shape of mass 3 Fig. 2. Peak shape of mass 3 with O.02-inch collecter slit: graph .A, with 0.005-inch collecter slit ; mixture hydrogen deutcridc only; graph B, a of helium-3 and hydrogen deuteride mixture of helium-3 and hydrogen showing partial resolution of mass-3 deu teride doublct CONCENTKATING THE HELIUM- Helium and argon are both members of the "inert"-gas family, therefore they cannot be separated by chemical procedures, and use must be made of variations in their physical properties.The boiling-point of argon is -1857" C and that of helium is -268.9" C, so the argon can be condensed at liquid nitrogen temperature (-195.So C) leaving helium as a gas.352 MELHUISH AND CHAPMAN DETERMINATION OF [Analyst, VOl. 91 A few experiments were made with a simple apparatus, the argon was condensed in a cold-trap containing an absorbent and the helium was recovered by pumping. It was established that activated charcoal held the argon better than a molecular sieve, and an apparatus was constructed (Fig.3). T U C1 T" 2 To rotary pump A = Sample attachment point F = Concentrate sample tube B = B.T.S. reagent container 0 = 3-way taps C,, C, = Cold traps (- 1 9 6 O C) T = Taps D = Mercury diffusion pump + = Ground-glass joint E = Simple manometer Fig. 3. Schematic diagram of the helium conccntration apparatus A series of experiments was carried out to determine the recovery of helium. The helium was added by using known volumes of air (5-4 p.p.m. by volume of natural helium)- the main constituents of air, i . e . , oxygen and nitrogen, are retained on the liquid-nitrogen cooled charcoal, the helium being pumped from trap 1 to trap 2 and finally into the sample tube.These are results on concentrates from 50-ml aliquots of air that should contain 2.7 x ml of helium. The analyses were carried out on a standard MS2 gas mass spectrometer. The results of the first series of experiments are summarised in Table I. TABLE I RECOVERIES OF HELIUM FROM 50ml OF AIR Total volume Volume of Aliquot of concentrated helium found, Helium added, Recovery, No. gas, ml ml ml per cent. 1 - 2-7 x 10-4 2.7 x 10-4 100 2 3.6 x 10-4 2.6 x 10-4 2.7 x 10-4 96 3 3.7 x 10-4 2.8 x 10-4 2.7 x 10-4 103 4 4.1 x 10-4 2.8 x 10-4 2.7 x 10-4 103 A second series of experiments was then undertaken, in which the samples consisted of Results are 1 litre of helium-free argon and 25 ml of air, i.e., 1-35 x 10-4ml of helium. summarised in Table 11.TABLE I1 RECOVERIES OF HELIUM FROM 25ml OF AIR I N 1 LITRE OF ARGON Sample No. 1 2 3 4 5 6 7 Total volume 3f concentrated gas, ml 2.9 x 10-4 2.3 x 10-4 2.8 x 10-4 4.0 x 10-4 2.5 x 10-4 3.8 x 10-4 2.5 x 10-4 Volume of helium found, ml 1.6 x 10-4 1.3 x 10-4 1.4 x 10-4 1.4 x 10-4 1.4 x 10-4 1.3 x 10-4 1.4 x 10-4 Helium added, ml 1-35 x 10-4 1.35 x 10-4 1-35 x 10-4 1-35 x 10-4 1-35 x 10-4 1-35 x 10-4 1.35 x 10-4 Recovery, per cent. 118 96 104 104 104 96 104June, 19661 HELIUM-3 IN ARGON AT LEVELS OF 353 As good recoveries of helium at this level from 1 litre of argon had been established, a third series of experiments was carried out with a very much lower level of helium. This time the helium-4 in the concentrated gas was measured with the modified mass spectro- meter.It should be noted that the concentration apparatus was made of Pyrex glass and may therefore be slightly porous to natural helium. Blank values on the apparatus were equivalent to 0.2 to 0.3 x Synthetic samples were prepared by using 1 litre of helium-free argon and 0.65 ml of air, i.e., 3.5 x 10-6 ml of helium-4. Results are given in Table 111. ml of helium. TABLE I11 RECOVERIES OF HELIUM FROM 0-65ml OF AIR IN 1 LITRE OF ARGON Volume of Sample helium, No. ml 1 3.8 x 2 3.5 x 10-6 3 3.9 x 10-6 4 7.0 x loF6 5 6.1 x 10-6 6 3-8 x Corrected for blank, ml 3.5 x 10-6 3.2 x low6 3.6 x 6-7 x 5.8 x 3-5 x 10-6 Volume of helium added, ml 3.5 x 10-6 3.5 x 10-6 3.5 x 10-6 3.5 x 10-6 3.5 x 10-6 3.5 x 10-6 Recovery, per cent. 100 92 103 191 166 100 The analysis showed that the hydrogen content of the concentrated gas was undesirably high, and in view of the resolution problem it was decided that it must be reduced to a minimal value.Experiments with helium-3 and varying amounts of hydrogen showed that a 10 per cent. increase in the mass-3 peak arising from H3+ (from the hydrogen) required more than 300 times as much hydrogen as helium. Hydrogen is not removed by activated charcoal a t -196” C ; however, several chemical methods are available and experiments were carried out with palladised asbestos and “B.T.S. reagent” (finely divided copper made up into pellets with an organic binder), as supplied by B.A.S.F. of Germany. Both methods appeared to be equally effective in reducing the hydrogen content, but the B.T.S. reagent was chosen because the addition of oxygen gas to effect the removal of hydrogen was not necessary. Blends, consisting of 1 litre of argon, 3.5 ml of air and 1 ml of hydrogen, were used for the experiments with the two reagents, and the results obtained are given in Table IV.TABLE IV RESIDUAL HYDROGEN FROM 1 ml OF HYDROGEN IN 1 LITRE OF ARGON Reagent and Sample No. Palladised asbestos 1 Palladised asbestos 2 Palladised asbestos 3 Palladised asbestos 4 B.T.S. 1 B.T.S. 2 Total volume 3f concentrated gas, ml 9.4 x 10-4 2.1 x 10-3 1.7 x 10-3 3.7 x 10-3 1.6 x 10-3 1.9 x 10-3 Volume of hydrogen, ml 8.1 x 10-4 1-94 x 10-3 1-62 x 10-3 3-58 x 10-3 1.8 x 10-3 1.51 x lop3 Hydrogen removal, per cent. 99-92 99.81 99.84 99.64 99.85 99-82 The hydrogen remaining was then 2 to 3 x ml, i.e., the hydrogen removal was 99.7 The hydrogen content of the concentrated gas has to be measured to 99.8 per cent.effective. to make sure that it does not exceed 300 times the helium-3 content. OUTLINE OF FINAL METHOD Samples of blanket gas are taken from a sample point in 500-ml stainless-steel, lead- shielded sample vessels at 35 p.s.i., and are allowed to stand for a day to reduce the radio- activity; they are then analysed. Between 500 and 1000 ml of gas are introduced into the helium-concentration apparatus (Fig. 3) and the volume is measured accurately. After passing the sample over the B.T.S. catalyst the argon is condensed in trap 1, the concentrated helium (and neon) is pumped into the interspace above the second Topler pump and then into trap 2, where any remaining traces of condensable gases are removed.The Topler pump is used to transfer the concen- trated gas into the sample tube where a pressure - volume measurement is made.354 MELHUISH AND CHAPMAN y l z a l y s t , Vol. 91 The sample tube is then transferred to the double-inlet system of the MS2 mass spectro- meter where the gas is expanded into the 50-ml reservoir. Measurements are made of the peaks at mass numbers 2, 3 and 4. To save the small amounts of gas from pumping steadily away via the leak valve to the mass spectrometer, the instrument controls are set to the desired mass number with monitor gas in the 2-litre reservoir. The leak valve to the 2-litre reservoir is closed and the leak valve to the 50-ml reservoir is opened for the sample peak height to be measured or recorded.The leak valve is then closed and the procedure repeated at the next mass number. A check is made on the time at which the leak valve was originally opened, To, the length of time for which it is open and the elapsed open-time at which each peak is measured; peak heights are then corrected for “pump-out” rates to give peak heights at To. The corrected peak height is then directly converted to the volume of helium-3 (or helium-4). The mass spectrometer is calibrated immediately before or after each sample by intro- ducing a known volume of an accurately prepared standard gas. RESULTS Sampling from the Dounreay Fast Reactor gas circuit was instituted in late January, 1965. Helium levels encountered were much higher than had been envisaged, being 1 to 2 x p.p.m. by volume of helium-3 and 20 to 50 p.p.m.by volume of helium-4. At these levels the size of sample taken was considerably reduced; 200 to 300 ml was usually quite sufficient. There was a considerable day-to-day variation of the helium-3 (and helium-4) content probably associated with sampling problems in a non-circulatory gas blanket system. Reproducibility has been checked by regular samples from a synthetic blend of helium-3 in argon at approximately 3 x p.p.m. by volume, and by taking several aliquots from one of the reactor gas samples. Results on the synthetic blend are as follows- helium-3 content: 0.0030 * 0.0001 p.p.m. by volume. Results on 3 aliquots from one reactor gas sample- helium-3: 0-8 x lod3 p.p.m. by volume, 0.8 x helium-4 : 31 p.p.m. by volume, 34 p.p.m. by volume, 34 p.p.m. by volume. p.p.m. by volume, 0-8 x p.p.m. by volume; DISCUSSION The limit of the method as used with the Dounreay Fast Reactor argon samples is 2 x 10-5 p.p.m. by volume, but for less radioactive samples the method should be capable of coping with at least 5 litres of argon, i.e., a limit of 4 x Other instrument modifications were considered, notably those by C ~ t h b e r t , ~ but were unnecessary in view of the helium levels encountered. Vse of the Cuthbert modifications should lower the limit to 1 x Recovery of helium from the argon is good and is certainly better than 40 per cent. Reproducibility is good, better than 10 per cent. The method has been proved in use over some 12 months. A method has been devised for measuring very low levels of helium-3 in argon. p.p.m. by volume. p.p.m. by volume. REFEREKCES 1. 2. 3. Parkinson, K. T., and Toft, L., AnaZyyst, 1965, 90, 220. Nier, A. O., in Waldron, J. D., Ediior, “Advances in Mass Spectrometry,” Pergamon Press, London, New York, Paris, Las Angeles, 1959, p. 507. Cuthbert, J., J . Scient. Instrum., 1964, 41, 431. Received December 221.ad, 1965