Analytical Procedures for SF6
SF6 is determined in the laboratory using a purge-and-trap gas chromatography procedure with an ECD detector. This section describes details of the measurements, including the apparatus for measuring SF6 concentrations, the procedure for introducing the water into the stripping cell, the vacuum extraction procedure of SF6, the measurement procedure, blanks, standardization and calibration procedures, and instrument stability.
Extraction and Measurement of SF6
Figure A shows the apparatus used for the vacuum extraction of SF6 from ground water and is similar to the system described by Law et al (1994) and Wanninkhof and Ledwell (1991). The apparatus consisted of a 1-L glass stripping vessel and various valves that control the flow of gases, water, and the vacuum. First valves V-1, V-A and V-C are turned to the closed positions, then V-B to the vacuum and V-D are opened. A high vacuum is pulled in the 1-L gas stripper. The stripper is isolated from the vacuum by closing of valves V-B and V-D. The water sampling tube, a 3.2 mm OD copper tube is placed in the bottle and V-A is turned from the off position to the water-in position. The water intake tube is slowly lowered into the bottom of the bottle. The vacuum in the stripping cell pulls the water in and sprays into the vessel through 6 nozzles. After about 100 mL of water is added into the vessel, the V-A is turned into the N2 position and the stripper is pressurized with SF6-free N2. The N2 pressure expels the water to waste from the stripper when V-B is turned from the off to the waste position. This procedure cleans the intake tube and valves and prevents carry-over of SF6 from the previous sample. The stripper is then re-evacuated. One L of water is sprayed into the cell, vacuum extracting about 90 percent of the SF6 from the water. N2 is introduced at the bottom of the stripper through a stainless steel gas-dispersion tube when V-C is slowly opened. After the stripper was pressurized with N2, V-1 is turned allowing the N2 and stripped SF6 to pass the Ascarite-magnesium perchlorate drier and then the SF6 is retained on a 1 m 3.2 mm OD Porapak-Q trap (figure B). Two CO2, H2S and H2O traps are on the water stripping system, V-2 is used to switch from one trap to the other without interrupting the analysis to replace traps. The SF6 trap is pre-cooled to a temperature of –70 to 79°C in a dry ice-isopropyl alcohol dewar flask. After exactly 5 min of stripping at precisely 250 mL/min, the trap is switched from the stripping system to the analytical system by turning valve V-3. The stripping cell is emptied and prepared for the next sample. A large volume of N2 is needed to efficiently strip the SF6 from the 1L of water, but break-though of SF6 occurs on short traps. To significantly improve the chromatography of SF6, the sample is transferred from the 1 m trap by placing the trap in 95°C water to the 0.1 m pre-cooled trap 2. The transfer is completed in 1 min, trap 2 is isolated from the carrier flow by turning V-7, and then trap 1 is switched from the analytical system to the water stripping system by turning V-3.
The analytical system bears similarities to other analytical instruments used to measure SF6 (Law et al, 1994; Wanninkhof and Ledwell, 1991; Maiss and others, 1994), however, many improvements were made to increase flow and pressure stability, and increase sensitivity. It soon became obvious that the analytical systems described in the literature were inadequate for the measurement of very low concentrations of SF6 normally found in ground waters and required greater flow and pressure regulation. The instrument (figure B) was modified to allow the stripping of much larger volumes of water. The design of the gas distribution system was critical for obtaining good chromatograms from very small signals that are obtained from natural levels of SF6.
Ultra-pure carrier grade (UPC) N2 is used through out the system. The carrier gas is purified with a charcoal and a hydrocarbon-O2 trap. The pressure of the gas is controlled with four ultra-clean pressure regulators and the flow adjusted by three needle valves. Additional pressure and flow regulation is provided by dummy columns and a capillary restrictor. A flow controller after the ECD maintains constant carrier flow through the detector. All these measures prevent flow and pressure fluctuation during the switching of valves, which cause background drift and noise.
The sample introduction system consists of the 4-position selector valve V-4. This valve selects between two gas standards, air, and carrier gas. Gas samples can also be introduced into the sampling loops from glass ampoules by attaching the flasks to port 1 of valve V-4. The gas sampling loops are evacuated by first turning valve V-F to the vacuum position and then to the off position. The sample is released by breaking the pre-scored tip of the ampoule. The samples can be injected by valves V-5 or V-6 into the analytical system at ambient, sub-ambient, or greater than atmospheric pressures. All volumes in the gas-introduction system have been precisely measured and the pressures determined with a pressure transducer. The gas injection system consists of two valves and four sampling loops with volumes ranging from 0.1 to 15 mL. The loops can be directly injected into the analytical column or the SF6 can be trapped at about –75°C on the short Porapak Q trap. Alternatively, SF6 concentrations can be measured from large volumes of gas by attaching the sample vessel to port 5 of V-3 and a vacuum pump to port 4 of V-3. First, the lines are evacuated, then the tip pre-scored glass vessel broken, and the sample gently pulled through the large pre-cooled trap by the vacuum pump. The large trap is transferred from the water stripping system to the analytical through V-3 and then the SF6 moved from the large to the small trap.
After the SF6 is trapped, the trap 1 is isolated by valve V-7, then the trap is heated to 95°C, valve V-8 is switched from the back-flush position to the run position, and the trap injected into the analytical column. The SF6 enters the analytical column and the chromatogaphy phase begins. After the SF6 is measured, the valve V-8 is switched to the back-flush position preventing the O2 and other highly retentive compounds from entering the ECD. This procedure greatly reduces the analytical time.
Calibration of Analytical System
The instrument was calibrated using a blank, and 0.1, 0.3, 0.5, 0.6 cc of a 104 ppt Scott gas standard. The gas was directly injected into the analytical columns or was trapped on the Porapak-Q trap and then injected onto the column. Both procedures yielded identical results indicating 100-percent efficiency in the trapping of SF6. The system also was calibrated using a blank, and 5, 10, 15, 30, and 45 mL of a NOAA air standard. SF6 in all air samples and standards are trapped prior to injection into the analytical column.
Precision and Accuracy of Measurements
Standard deviations of better than 3 percent were routinely obtained for repeated measurements of standards. The calibration was linear through the entire measuring range. Standards were prepared by gravimetric procedures, accuracy is about 1 percent. For water samples, precision was about 50 percent at the detection limit of less than 0.01 fmol/L and about 5 percent for concentrations greater than 0.1 fmol/L.
Analysis of SF6 in water references
Busenberg, E., and PlummerL.N., 2000, Dating young ground water with sulfur hexafluoride: Natural and anthropogenic sources of sulfur hexafluoride. Water Resources Research, 36, 3011-3030.
Law, C. S., Watson, A. J., and Liddicoat, M. I., 1994, Automated vacuum analysis of sulfur hexafluoride in seawater: derivation of the atmospheric trend (1979-1993) and potential as a transient tracer, Marine Chem., 48, 57-69.
Maiss, M., J. Ilmberger, A. Zenger, and K. O. Munnich, A SF6 tracer study of horizontal mixing in Lake Constance, Aquat. Sci., 56, 307-328, 1994.
Sliwka, I., and Lasa, J., 2000, Optimisation of the head-space method in measuring SF6 concentration in water. Chem. Anal. (Warsaw), 45, 59-72.
Wanninkhof, R., and Ledwell, J. R., 1991, Analysis of sulfur hexafluoride in seawater, Jour. Geophys. Res., 96C, 8733-8740.
