Cosmic-ray produced atmospheric 39Ar activity (T1/2 = 269 yr) has been determined at 0.11 ±0.012 dpm/lt argon. Ice samples from two profiles in Greenland bore holes showed conclusively that 39Ar dating leads to correct ages. Corrections can be made for possible contamination of the samples with ambient air during field extraction and during laboratory processing by measuring 85Kr in the same samples.
The following isotopes: 14C, 39Ar, 85Kr, 3H, partly 32Si, 13C, and 18O were investigated in 20 groundwater samples. Unexpectedly large discrepancies between “14C ages” and “39Ar ages” were observed for many of these samples. For example, a horizontal profile of a confined sandstone aquifer in the Franconian Albvorland showed decreasing 39Ar and 14C activities from respectively, 100 percent to 17 percent and 80 percent to 0.3 percent of modern activity, corresponding to elapsed time periods of 700 and >20,000 years, respectively.
It seems unlikely that gas exchange through the aquiclude is the cause of this discrepancy. It can neither be explained by only assuming that the water represents a mixing of components with different ages. We detected the possibility of underground production of 39Ar in thermal spring water from Zurzach, Switzerland. Its argon showed higher specific activity than atmospheric argon. Elsewhere, however, samples were found containing no detectable 39Ar activity: two wells of a confined carstic aquifer in Ingolstadt, Germany, show less than 7 percent atmospheric activity.
In our opinion, the assumptions on which the 14C method in hydrology are based need to be critically re-examined. It is possible that for some aquifers carbonate exchange between solid and liquid phases in the aquifer changes 14C results to a larger degree than generally assumed.