Low level analysis of tritium by gas proportional counting
All numbers of quantities, etc., are typical only and vary from sample to sample.
A. Distillation
The water sample is distilled with continuous reflux to dryness or near dryness. During the procedure, the still is vented to the ambient air through a drying agent to avoid contamination of the sample by atmospheric water vapor.
B. Electrolytic Enrichment
The object of this procedure is to reduce the volume of the sample while preserving a large fraction of the tritium. A portion of the sample is charged into the electrolytic enrichment cell. To that portion, 2 mL of concentrated sodium hydroxide solution (made from dead water and sodium peroxide or oxide) is added, and the remainder of the sample is transferred to a container on top of the cell.
The sample is electrolyzed in a cold water bath. Once a day the solution in the cell is topped up and the procedure is continued. This procedure takes 10 - 14 days, and the remaining amount of water typically contains 80% of the original amount of tritium. The enriched water sample is then vacuum distilled from the sodium hydroxide, and the yield is weighed to – 2 mg.
C. Preparation of Sampling Gas
About 3 mL of the enriched water sample is injected into a vacuum system. The water evaporates, and the vapor is reduced by hot magnesium metal to hydrogen gas which is absorbed on activated charcoal at liquid nitrogen temperature in a stainless steel pressure cylinder. Approximately 4 L atm of hydrogen is obtained this way.
D. Low-Level Counting
The low-level gas proportional counters have an active volume of 1 L and are shielded by 2.5 cm of selected lead, a ring of anti-coincidence Geiger counters, 10 cm of paraffin wax, boric acid and/or borated polyethylene, and at least 20 cm of iron, plus the walls and ceiling of the building. The counter is first filled with 10 psi (67 kPa) of propane. Thereafter, the sample hydrogen gas, under pressure in its cylinder, is added to the counter for a total pressure of 40 psi (300 kPa). The counter is then sealed off, and the gas amplification is set to specifications by adjusting working voltage. After that, counting proceeds until criteria for accuracy or sensitivity have been met. The pulses are sorted into several channels, of which some are used for continuous control of the gas amplification, as shown in the cosmic radiation spectrum, etc. Counting times are 6 to 20 hours. A 1 TU original sample enriched from 275 to 6 mL typically shows 0.6 cpm in the tritium channel above a background of 0.40 cpm, known to – 0.02 cpm. The control of filling and counting procedures and calculation of results, as well as numerous checks on the performance of the machinery, are computerized.
E. Backgrounds and Standard
At least once weekly each counter counts dead hydrogen gas (from petroleum). In addition, water from the deep Floridan Aquifer (more than 10,000 years old water) is reduced to occasionally check on the tank hydrogen gas. This procedure sets the background count of the counting equipment. A further check on process blanks is that at least once a week a sample of dead water (from the Floridan Aquifer) goes through all the same procedures, including enrichment, as the unknown samples. In order to check on the efficiency of the enrichment procedure, at least once a week a sample of known activity is processed through the entire system of enrichment, reduction, and counting. The efficiency of each counter is determined by counting hydrogen gas made by reduction of standard water in our regular preparation system. This standard water is prepared from NIST (formerly United States National Bureau of Standards) SRM #4926 by dilution by weighing. The dependence of background, efficiency, etc., on pressure, gas composition, gas amplification, etc., is known, and the appropriate corrections are applied via the software of the computing system.
F. Update
Periodically, usually about every six weeks, all measurements in all counters for the preceding time period are recomputed, applying statistical tests, and scrutinized for flaws in quality. This includes all measurements of unknowns, backgrounds, blanks, enrichment factors, efficiencies, etc. Only after this step is the result considered final. The results are then reported in Data Releases, one for each project or job.
Further Technical Information
The laboratory has successfully participated in all international low-level cross-check projects run by the International Atomic Energy Agency in Vienna. These projects are run approximately every five years. The procedures of tritium analysis are described in the following publications from this Laboratory, reprints furnished on request:
Östlund, H.G., H.G. Dorsey and C.G. Rooth, 1974. GEOSECS North Atlantic radiocarbon and tritium results, Earth Planet. Sci. Lett., 23, 69-86.
Östlund, H.G., M.O. Rinkel and C. Rooth, 1969. Tritium in the equatorial Atlantic current system, J. Geophys. Res., 74(18), 4535-4543.
Östlund H.G., Tritium, in GEOSECS Atlantic, Pacific, and Indian Ocean Expeditions, Vol. 7, Shorebased Data and Graphics, 7-10, 1987. (Describes procedures and examples of reliability tests. Reprints available from our laboratory.)
G. Sample Identification and Flow of Information
Water samples for tritium analysis are received and inventoried using the accompanying packing list or chain of custody supplied by the client. A computer worksheet listing sample name, volume, sample collection date, and date of arrival into lab, as well as client information, is generated. At this time, each order is given a unique job number, and each sample decimal numbered within that job. For example, the job-sample number (JB#), 123.05 indicates the fifth sample in the listing for job 123. The computer input is proofread, and the worksheet and labels are printed. An abbreviated copy of the worksheet listing is given to the administrative personnel to be filed with the client's records. The worksheet is used by the preparation technician to keep track of the progress of the samples. Preliminary results are recorded on this sheet as they become available through the computer. From the time the worksheet is printed, the sample is referred to by its JB#. Triple copy labels are attached to each sample bottle.
When processing begins, one of the labels bearing the JB# is attached to the processing vessel; the label then "follows" the sample through the preparation steps; i.e., it is physically transferred to the next process container. The same label is used from the beginning distillation through electrolytic enrichment, vacuum distillation and reduction to hydrogen gas and is eventually attached to the face of the pressure gauge of the counter in which the sample is counted. During each step of the associated with the sample, a yellow preparation card reports the distillation date and unit #; electrolysis starting date; volume and cell #; and freeze-out date and unit #. Recorded also are the order number through the reduction system, and the cylinder # used to store the gas sample prior to counting.
When a sample is ready for activity measurement, the sample I.D. is merged with all the process information in the computer. Together, the data are entered into the particular sample file along with counter-fill data, sample pressure and temperature, a unique run number, and the time and date. Upon completion of the counting, temperature and pressure of the sample within the counter are again recorded for comparison and checked for computer input error. All records of the sample preparation information and counting results are stored in computer files. A listing of all samples prepared and counted the week before is printed every Monday to study for possible problems with the preparations and/or counting equipment.
Using these procedures, every sample can be easily traced from the moment it arrives in the lab to the final result.
