CHEMICAL STATE OF TRITIUM IN THE ATMOSPHERE 147 If one assumes that sources other than those in the United States were releasing tritium during the period 1950 to 1955, the release curve_should turn upward. When similar release rates are assumed for such other tritium-producing laboratories, a value of 0.4 to 1.5 x 10° T.U. is obtained for a total of two or three of these laboratories. The tritium released in the early tests then is superimposed on a level that already has been made fairly high by 1953. Unknown factors in the estimation of the amount of tritium re- leased from all sources are the starting dates for the releases and the period of time over which they occurred. Also unknownis the relative rate of release from the laboratories outside the United States. The Significance of accidental releases of large amounts of tritium cannot be determined. Robinsonet al.'® refer to “accidents releasing appreciable quantities of gaseous tritium.” However, a given size release would have a greater significance for the T/H levels observed in hydrogen prior to 1953 than would the same amount if released at the present time. These uncertainties are probably no greater than the rather large errors in the present estimation. The quoted value of 15 kg of tritium released during the Ivy and Castle tests is based on an estimated value for a rate of release per megaton of nuclear energy.’ These release-rate values must depend, in a manner not well defined, on the composition of the nuclear devices exploded. It is entirely possible that the tritium level in 1953 and 1954 was due to releases by the production systems only. At this point at least three different conclusions are possible: (1) the Ivy and Castle tests did not release significant amounts of “un- burnt” tritium as HT, (2) the total tritium released in these series was actually much less than the estimated 15 kg (operating at maximum capacity, the thermal-diffusion column at Los Alamoscould have pro- duced 15 kg in the 5.5 years considered between 1949 and 1954); and (3) perhaps the production systems and the bombtest each might have released approximately equal amounts of HT. SOURCES OF ATMOSPHERIC TRITIATED METHANE In their description of the tritium-production system, Robinson et al.!° show a mass-spectrum curve for the impurities observed in tritium handled in a system containing stopcocks and vacuum grease. These impurities include OH*t, H,O*+, OT*, T,O+, and CTf. The state- ment “at somewhat lower isotopic purity a CHT, ion of mass 22 appears, etc.” is made. An all-metal system that employed Kel-F seals was constructed to avoid these impurities. Although this did eliminate the problem of labeled hydrocarbons appearing in the tritiumproduction system, it did not eliminate the release of tritium-labeled molecules, particularly methane, into the atmosphere.

Select target paragraph3