Earth and Planetary Sciences ETDs

Publication Date



In recent years the preferred analytical procedure for the extraction of oxygen from silicates and oxides for oxygen isotope analysis has been the oxidation of the samples with either fluorine (F2) or bromine pentafluoride (BrF5) to produce oxygen. Several standards are available for monitoring the extraction procedure, and one in particular (NBS-28) is used in many laboratories. The oxygen-18 oxygen-16 ratios obtained for NBS-28 vary by as much as one permil between laboratories. The major cause of this variation lies in the extraction procedure used. Extraction of oxygen from iron bearing minerals and whole rocks (and minerals containing certain other multivalent transition metals) results in the accumulation of iron or other metals in the sample tubes. These metals can react with oxygen liberated in all later reactions, thus depleting that sample in oxygen-18. Hot fluorine gas can be used to remove these metals or at least to minimize their effects. To facilitate the use of fluorine gas in this application, a solid source of fluorine gas has been adapted. This fluorine source can be used to produce oxygen-free fluorine at low pressures, thereby reducing the danger in handling the fluorine, and eliminating the need for oxygen blank corrections. Modifications have been made to remove two other sources of contaminants, nitrogen and atmospheric moisture, that can affect oxygen isotope values. First, argon is used to bring the manifold and sample tubes to a pressure of one atmosphere before sample loading. Argon is used in place of nitrogen because nitrogen forms nitrogen-­oxygen and nitrogen-oxygen-fluorine compounds in the conversion process, and may fractionate the sample. Second, atmospheric moisture contamination, from loading samples outside a dry box, has been reduced to below a detectable level by thermal decomposition of the contaminant formed, NiF2•4H20. This technique seems to be effective over the range of humidities encountered in this laboratory (up to fifty percent relative humidity). The techniques described in this study minimize contaminants in the manifold and sample tubes so that regardless of the composition of the proceeding sample, a consistant value for NBS-28 can be obtained. This laboratory obtains a value for NBS-28 of 9.9 ± 0.17 (one sigma) permil using the modified procedures developed in this study. This average and standard deviation are calculated for twenty-three NBS-28 analysis, and are representative of values obtained over a period of nearly four years.

Degree Name

Earth and Planetary Sciences

Level of Degree


Department Name

Department of Earth and Planetary Sciences

First Committee Member (Chair)

Gary Perrin Landis

Second Committee Member

Douglas Gridley Brookins

Third Committee Member

Rodney Charles Ewing



Document Type


Included in

Geology Commons