Dear All, Pier asked me whether I can have some comments on this problem. See also the message to isogeochem by Torsten Vennemann on 9 October 1996 for the hydrogen-zinc problem. As far as I remember the H2-glass exchange was shown to be important by the papers mentioned at high T, i.e. if the reaction tubes are kept in the furnace too long. When I started the dD business I used the stopcock-sealed tubes, but the N2 background was always unconveniently high (see Heaton and Chenery 1993, Chem. Geol. Isot. Geosci. Lett., 106:485-487.). So I use flame-sealed 6 mm pyrex tubes now. This way there is no N2 in the sample gas. However, I don't like storing the tubes containing Zn+ZnO and H2 for more than a couple of days. Even if there is no obvious shift in dD values after a week, the reproducibility gets worse. Usually I prepare and reduce the water samples one day and measure the dD values on the next day. This way for pure H2O the average reproducibility is about 0.5-0.7 permil. Why measuring next day. I prepared a set of standards and started measuring them immediately after the reduction. The dD values stabilized after ~2 hours, maybe due to the hydrogen-zinc reaction that needs time for equilibration, so anyway I would wait for hours after the reaction. Once I measured hydrogen gas stored together with the reacted zinc for ~2 years in flame-sealed tubes. There was no H2 in the tubes. If this is due to the formation of zinc hydroxide, a solution for long-term storage would be absorption of H2 on activated charcoal in flame-sealed tubes. Also heating the "zinc" to 350 oC before the sample is let into the MS could eliminate the problem. Best regards, Attila Dr. Attila Demeny Laboratory for Geochemical Research Hungarian Academy of Sciences Budapest, Budaorsi ut 45, H-1112, Hungary tel/fax: 36-1-319-3145 e-mail: [log in to unmask]