5.5. STEALTH PROJECT: AUTOMATED FOUR­CHANNEL FIELD GAS CHROMATOGRAPHS

5.5.1. OVERVIEW

The STEALTH GC was installed at ALT, HFM, ITN, and LEF with different custom configurations for each client. The four-channel STEALTH GC will replace the old HP5920 GCs at the Radiatively Important Trace Species (RITS) stations (BRW, NWR, MLO, SMO, and SPO) and is currently being constructed and laboratory tested.

The data acquisition software and operating system of the STEALTH GC computer is being upgraded. NOAH is currently cooperating with personnel at Harvard University in the development of new data acquisition software for the PC-based UNIX operating system, QNX. QNX is a multitasking and multi-user operating system that will facilitate data acquisition, data retrieval, data archival, and real-time display. The airborne GC, ACATS-IV, is currently being configured to test the QNX data acquisition software. In 1996, NOAH scientists will implement the software on an ACATS-IV deployment and on the new STEALTH station GCs.

The STEALTH GC is an ECD/GC system based on technology developed on ACATS-IV and LACE. The first channel encompasses a Shimadzu mini-2E ECD and a Porapak Q packed column (Figure 5.28). This channel allows for the measurement of N₂O and SF₆. Channel two of the instrument uses a Valco ECD along with a Unibeads 2S packed column. This configuration is capable of measuring N₂O, CFC-12, H-1211, CFC-11, and CFC-113. The third channel also uses a Valco ECD and an OV-101 packed column and is used to measure CFC-11, CFC-113, CHCl₃, CH₃CCl₃, CCl₄, and C₂Cl₄ (perchloro-ethylene, PCE). All of the aforementioned channels have been proven in other similar instruments that have been constructed and deployed to various sites.

Fig. 5.28. Chromatograms of all four channels of the STEALTH GC that will replace the old RITS HP5890 GCs.

Channel four, however, has just recently been developed and uses quite a different setup than the other three channels. This channel also uses a Valco ECD, but incorporates a GS Q capillary column rather than a packed column. This channel also uses a Neslab cryocooler which allows for the preconcentration trapping of the three trace gasses, HCFC-22, CH₃Cl, and CH₃Br being measured. With the current configuration of this channel, one is capable of measurements of better than 1% for HCFC-22, 0.5% for CH₃Cl, and 2% for CH₃Br. Figure 5.28d is a chromatogram of the newly developed fourth channel.

5.5.2. TOWER GC AT WITN IN COOPERATION WITH CCG

The GC and instruments that monitor CO₂ and ²²²Rn are housed in a building adjacent to a tall tower (WITN) in rural North Carolina. Diaphragm pumps located in the building continuously draw air from 51, 123, and 496 m above ground through 1 cm i.d. Dekabon tubing affixed to the tower. Detailed descriptions of sample handling and drying, and initial results of CO₂ measurements at WITN were published [Bakwin et al., 1995]. GC analyses of air from each sampling level and of two calibrated whole-air standards are performed hourly. Standards are stored at high pressure in "Aculife"-treated aluminum cylinders. Chromatographic and housekeeping data are logged by a 486SX PC and archived on 1.2 Gb optical disks that are sent to the NOAH laboratory each week for analysis. Instruments and gas supplies are maintained by a technician who visits the site weekly .

Monthly statistics of mixing ratios for several halocompounds and N₂O at the WITN tower are presented in Figure 5.29. Statistics for each sampling height (51, 123, and 496 m) are denoted by 1, 2, and 3, respectively, along the bottom of each plot . For each month, the mean (circle) and standard deviation (distance between circle and asterisk) of the mixing ratios of each species generally decrease with increasing sample height. Variability in trace gas mixing ratios within the continental boundary layer is determined by sources, sinks, boundary layer dynamics, and horizontal transport. Since each species plotted in Figure 5.29 has solely ground-based sources, it is expected that mixing ratios and variability should be greatest near the ground. This effect is inflated at night by the accumulation of emissions from local, ground-based sources in the shallow nocturnal stable layer.

Fig. 5.29. Monthly statistics of CFC-11, CFC-12, CFC-113, methyl chloroform, carbon tetrachloride, nitrous oxide, and sulfur hexafluoride mixing ratios at the WITN tower. Crosses represent medians (horizontal bars) and interquartile range (vertical bars). Circles and asterisks are means and means ±1 standard deviation, respectively. The numbers across the bottom of each plot are the sampling level (1, 2, and 3 refer to 51, 123, and 496 m, respectively).

Figure 5.30 gives statistics for the 51-496 m mixing ratio gradients, binned by hour of day, of N₂O, CH₃CCl₃, and SF₆ for November 1995. Significant vertical gradients of N₂O and CH₃CCl₃ were observed at night, indicating these compounds were emitted by local, ground-based sources. In contrast, the insignificant accumulation of SF₆ in the nocturnal stable layer suggests an absence of local, ground-based sources. During the late morning and afternoon, convection rapidly mixes air from the ground to >500 m, and vertical gradients approach zero.

Fig. 5.30. Statistics of 51 m - 496 m mixing ratio gradients for N₂O, CH₃CCl₃, and SF₆, binned by hour, for November 1995. Crosses indicate means (horizontal bars) ± the 95% confidence interval (vertical bars). Circles represent medians, and asterisks indicate upper and lower quartiles. The left panel gives statistics of 51 m - 496 m gradients for the entire month.

In studying regional emissions of trace gases, it is critical that the influences of local sources are minimized. At WITN, the boundary layer height during the night is typically <500 m. Hence, mixing ratio variability at 496 m during the nighttime is primarily driven by horizontal transport of polluted air to the site, and mixing ratios of long-lived species should reflect regional-scale emissions. Figure 5.31 shows the correlation of CH₃CCl₃ and C₂Cl₄ mixing ratios at 496 m between 2200-0900 EST during November 1995. An orthogonal distance regression through the data yields a slope of 0.62, which can be taken as the regional emission ratio of these two compounds. Using accurate (±5%) estimates of North American emissions of C₂Cl₄ [McCulloch and Midgely, 1996], CH₃CCl₃ emissions can be calculated. Using this methodology, emissions of halocompounds, especially those whose production and emissions are controlled by the Montreal Protocol, are monitored.

Fig, 5.31. Correlation between CH₃CCl₃ and C₂Cl₄ mixing ratios at 496 m between 2200 and 0900 (EST) during November 1995. The slope of an orthogonal distance regression (0.62) is taken as the regional emission ratio of these two compounds.

[back] [contents] [next]