Production of 99.9997% 38Ar spike

Since I have nowhere properly described the main production batch (see Chimia 29,441(1975) for the two preliminary batches) of our highly enriched 38Ar spike, I'll briefly show and explain here the main strategy applied. This is not unique, but pretty much an optimum in investment, labor, time and quality of product. Our spike has been used for decades by dozens of laboratories all over the planet. But Potassium-Argon chronometry has become somewhat obsolete,
since the 40Ar/39Ar method has often advantages and requires much less demanding laboratory skill.- Perhaps, one day somebody needs such a 38Ar spike again. Then this short account might be helpful. There are still about 10 ampoules with 0.1 ml STP of this spike available. Dr. Roelant van der Lelij, Geological Survey of Norway, PO Box 6315 Sluppen, NO-7491 Trondheim, Norway; has taken over the 38Ar isotope distribution.
Fig.1 Preenrichment cascade: The feet of 7 identical 3 m columns in parallel are flooded (independently) with natural Argon. From the connected heads the argon isotope mixture is collected and fed to the bottom of one 3 m column. After reaching a dynamical steady state while slowly increasing the withdrawal at its top until the concentration of 40Ar equals that of 36Ar, it constantly produces a mixture with 2.34 % 38Ar (about 38 times the initial concentration in natAr) at a flowrate of about 9.6 % of the transport coefficient tau0 (AFAIK this methodology of dynamical preenrichment of a rare middle isotope has not been described before). 4000 ml of this mixture are collected and then fed into a 54 m column, Fig.2, consisting of 18 3 meter units, running in series, see the apparatus in Fig 6.
Fig.2 The mixture of 2.4% 38Ar and 48.8% 36Ar, 40Ar, each, produced by the small cascade 7+1 as in fig.1 needs several months to approximate the final distribution with a maximum of 24 % 38Ar in the middle. A cut between l0-lengths 16 and 20 containing 17.4% 38Ar and 41.3% 36Ar, 40Ar, each, is withdrawn and later fed into the center of the same column, now filled with HCl gas, Fig. 3, preenriched in H37Cl.
Fig.3 After several weeks the HCl and Ar isotopes give a complex distribution. The unwanted Ar isotopes collect preferentially at the ends. They are slowly removed from there, while the HCl carrying them is fed back in. This leads us to the stage in Fig. 4. Notice that the separation tube is now much less performant since HCl has a lower thermal diffusion coefficient than Ar and the characteristic tube length l0 has now grown from 1.5 m in pure Argon to 3.85 m in HCl. Hence the HCl isotope distribution is much broader than the Argon distribution in Fig. 2. The 38Ar distribution within HCl is not entirely accurate. The thermal diffusion factors are not well known. See the experimentally determined final distribution and the computed curve in Fig.5, which are, fortunately, not too different from each other. The effective mass difference between 36Ar and H35Cl is higher than that between 40Ar and H37Cl which promotes a better enrichment of 36Ar than 40Ar.
Fig.4 The flanking isotopes 36Ar and 40Ar have now been removed almost completely. At the bottom of the column the pressure in the column has been maintained by feeding in H37Cl, Fig. 5.
Fig.5 Experimental: This is the final distribution (of batch 1) using HnatCl as auxiliary gas. Only 38Ar is left in the tube with a peak at the HCl isotope mixture near the natural chlorine isotope ratio of 0.76/0.24! The 38Ar partial volume histogram sums up to 14 ml STP in a total volume of 4000 ml HCl isotope mixture. Each single histogram bar corresponds to the volume of 38Ar being left over after freezing out the HCl by liquid nitrogen from one thermal diffusion tube of 3 m length. The computed yellow 38Ar curve is about 12 times amplified (points, compared to the yellow line below). It reasonably represents the measured histogram. Of course, the total volume of 38Ar produced by the five steps shown, amounts to about 80 ml STP of 99.9997% isotopically pure 38Ar. The overall separation factor of 0.999997/0.000003/0.000629*0.999371 = 5.29E+08 from the natural abundance of 0.000629 is the highest ever achieved for a "middle" isotope. The top yellow points give the 38Ar molefraction relative to total Argon.
Fig.6 Laboratory setup of the 54 m thermal diffusion column used. Because of the auxiliary gas HCl the tubes have to be pyrex glas. The 18 3 meter units are connected in series. Isotope transport between the adjacent top-bottom is done with a gas-swing. The whole apparatus is glasblown vacuumtight and can be pumped by diffusion pumps. The heating wires are made from a Pt-Ir alloy and have to withstand several years at 700°C. The length changes by temperature variations are automatically adjusted by a magnetic clamp. Cooling by running tap water.
Udo Grünwald, my labtechnician, who helped with constructing and running the "machine", just handles a liquid nitrogen trap at the upper end of the column. We both loved to glasblow the whole thing together, some 80 m of tubing, while on a ladder 3 meters above floor, and were happy when it held the vacuum!
The product has been chemically cleaned, measured and handled in a ultrahighvacuum glas line.
The breakseal ampoules have finally been packed and shipped with 0.1 ml STP 38Ar, good for
some 10'000 Potassium-Argon age determinations (if handled competently!):