Integrated Sampling and Analytical Approach for Common


Integrated Sampling and Analytical Approach for Common...

0 downloads 94 Views 240KB Size

Environ. Sci. Technol. 2007, 41, 8388–8393

Integrated Sampling and Analytical Approach for Common Groundwater Dissolved Gases KIMBERLEY MCLEISH,† M . C A T H R Y N R Y A N , * ,‡ A N D A N G U S C H U † a Department of Civil Engineering, and Department of Geoscience, University of Calgary, Calgary, Alberta T2N1N4

Received June 30, 2007. Revised manuscript received October 14, 2007. Accepted October 15, 2007.

A novel passive gas diffusion sampler (PGDS) combines sampling, storage and direct injection into a single gas chromatograph (GC). The sampler has a 4.5 mL internal volume when deployed, is easy to operate, and eliminates samplepartitioning. The associated GC method analyzes for a large, dynamic sampling range from a single, small volume injection. Dissolved gases were separated on parallel Rt-Molsieve 5A and Rt-Q-PLOT columns and eluted solutes were quantified using a pulse discharge helium ionization detector (PD-HID). The combined sampling and analytical method appears to be less prone to systematic bias than conventional sampling and headspace partitioning and analysis. Total dissolved gas pressure used in tandem with the PGDS improved the accuracy of dissolved gas concentrations. The incorporation of routine measurements of dissolved biogeochemical and permanent gases into groundwater investigations will provide increased insight into chemical and biological processes in groundwater and improve chemical mass balance accuracy.

Introduction Dissolved gases (e.g., N2, O2, H2, CO, CH4, CO2, N2O, and H2S) are produced and consumed in many biogeochemical reactions in groundwater (1). Analysis of these and other naturally occurring permanent gases (e.g., Ar, Ne, Kr, Xe) are useful for groundwater age dating (2, 3), interpretation of subsurface biogeochemical processes and mass balance calculations (1), geochemical exploration (2), seismology, paleoclimatology (3), groundwater tracers (4, 5), and measurement of volatile organic compound (VOC) contamination (6–9). Dissolved gas sampling is under-utilized because routine sampling and analytical procedures are difficult (10). Typically, gas samples are analyzed from groundwater that is pumped to the surface into glass bottles that are filled and capped in the field with no visible headspace (11–13). Conventional analytical techniques for groundwater gas samples include purge and trap analysis (14), headspace extraction (11), or collection of free gas in the field by a “bubble stripping method” (15, 16). Dissolved gas concentrations can be altered by degassing (17) and air contamination during sampling (18) and partitioning, and injection processes during analysis (19). Also, few investigators have * Corresponding author phone: 403-220-2793; fax: 403-284-0074; e-mail: [email protected]. † Department of Civil Engineering. ‡ Department of Geoscience. 8388

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 24, 2007

used a total dissolved gas pressure (TDGP) meter in combination with gas sampling to achieve more accurate dissolved gas concentrations (20). Although measurement of gases from a single sample would be ideal, most analytical methods are incapable of separating a wide variety of gas species, and multiple separation and/or detection methods are used (21). One method is to analyze samples with two separate injections, or split the sample onto two columns with quantification on two different detectors. Cryogenic cooling can be used in combination with a single injection where gases are separated on two columns connected in series (with the first column at 45 °C and the second at -70 °C) and subsequent quantification by two different detectors (22). Although column switching (or flow reversal) can separate CO2, N2O, Ar-O2, and N2 (19), it does not provide baseline separation of Ar and O2. Molsieve 5A columns effectively separate permanent gases including critical pairs such as Ar and O2, but are highly retentive for CO2 and N2O (23). Q PLOT columns provide effective separation of gaseous and volatile hydrocarbons such as CO2 and CH4, but do not separate permanent gases (23). Dissolved gases are analyzed using thermal conductivity detectors (TCDs), flame ionization detectors (FIDs), electron capture detectors (ECDs), and pulse discharge helium ionization detectors (PD-HIDs). TCDs have a universal response to a wide variety of compounds (organic and inorganic), but are insensitive to small gas concentrations, require large sample volumes, and have a limited linear range (104 (24);). FIDs have lower detection limits (approximately 100 times less than TCDs) and greater linear ranges (107), but are unable to detect inorganic gases (24). ECDs have extremely low detection limits (5 fg/s), a linear range of 104, but are specific for electronegative compounds such as pesticides and CFCs (24). PD-HIDs are capable of detecting both organic and inorganic compounds, and have a linear response of 105 (25). We introduce an integrated sampling and analytical approach for accurate, inexpensive, and routine measurement of dissolved biogeochemical and permanent gases. Molsieve 5A and QPLOT columns are combined in parallel to analyze for CH4, CO2, N2O, Ar, O2, and N2 from a single injection using one conventional gas chromatograph (with no retrofitting required) and one PD-HID detector. The combination of a passive diffusion groundwater gas sampler and direct GC injection syringe has been recently published (26). This paper presents a similar approach that also (i) compares the passive gas diffusion sampler with conventional headspace partitioning and analysis (a common groundwater gas sampling method used in commercial and research studies (11) and recent publications 12, 13); (ii) analyses for a suite of common groundwater gases; (iii) incorporates the use of a TDGP probe in tandem with groundwater gas sampling to achieve more accurate dissolved gas concentrations (20); (iv) optimizes an gas chromatographic separation and detection approach for groundwater gas analysis; and (v) can be used in deep groundwater (i.e., under high hydrostatic head).

Materials and Methods Gas Chromatographic Analysis. Molsieve 5A and Q PLOT column performance for common groundwater gas analysis was tested using air, a mixed gas standard (5% CH4, 5% CO2, 0.43% N2O, 1% Ar, 5% O2, and 5% N2 with a balance of helium), and pure gas components (Praxair). Standards for CH4, CO2, O2, Ar, and N2 were prepared by filling a gastight syringe 10.1021/es0716094 CCC: $37.00

 2007 American Chemical Society

Published on Web 11/15/2007

FIGURE 1. External and cross-sectional views of the passive gas diffusion sampler. with selected volumes of gas and diluted with helium to prepare samples with differing concentrations. Samples were injected using a 6-port, two position sampling valve (Valco) and a 5 µL sample loop (1/16 inch stainless steel) that was loaded manually from a gastight syringe (Hamilton). Gas analysis was conducted on an HP-5890 (HewlettPackard) gas chromatograph fitted with a Valco model D-1 pulse discharge detector operated in the helium ionization mode with an electrometer (Valco). Data were collected using a HP 3396 Series II Integrator (Hewlett-Packard). Helium (UHP Praxair, 99.999%) was used for the carrier and discharge gas. A helium purifier (Valco), placed in-line immediately after the cylinder, removed water and impurities from the carrier gas. The sampling valve, injection line, and oven were maintained at ambient temperature (28 °C) and the detector at 200 °C. Flow rate through the detector was 50 mL/min, with 30 mL/min supplied from the columns and an additional 20 mL/min in helium makeup gas from the cylinder. Parallel Rt-QPLOT (Restek, 30 m × 0.53 mm) and RtMsieve 5A PLOT (Restek, 30 m × 0.32 mm) columns were used (23) (Supporting Information (SI) Figure S-1). The parallel columns were connected to the injection line and detector using 0.5 m precolumn Rt-QPLOT sections. Columns were connected with “press fit Y” connectors (Valco) and polyamide resin (Valco). Passive Gas Diffusion Sampler Construction and Deployment. The PGDS is a modified gastight syringe constructed of materials that are unlikely to alter the gas composition in the sample. The sampler was designed to have minimal internal volume and membrane thickness, and maximal diffusion surface area (to minimize equilibration times) within the limits of commercially available materials. The sampler consists of three major parts; the syringe, the diffusion barrel, and the plunger (Figure 1; Figure S-3 and more details in Supporting Information). A 1 mL Hamilton Sample Lock Gastight syringe comprised part of the body of the sampler (Figure 1). Samples are contained in the PGDS by closing the valve at the base of the needle during deployment. The diffusion barrel consists of a solid and sintered stainless steel skeleton covered with a gas-permeable silicone

TABLE 1. Calibration Statistics for Parallel Dual Column/pulse discharge-helium ionization detector (PD-HID) Analytical System gas species

R2 value

mean relative standard deviation (RSD)

slope (area count /% Vol)

detection limit (% Vol)

CH4 CO2 Ar O2 N2

0.9957 0.9956 0.9894 0.9929 0.9927

( 0.17% ( 0.02% ( 0.05% ( 0.14% ( 0.34%

9.2 × 104 2.7 × 104 4.0 × 104 3.8 × 104 3.0 × 104

0.001 0.003 0.025 0.026 0.033

membrane. A continuous void volume of 4.5 mL is formed by threading the barrel and syringe together. A longer, stainless steel plunger (Figure 1) replaced the short, aluminum one provided by the syringe manufacturer. The Teflon plunger tip was a machined replica of the original, so that a gastight seal with the inside of the glass syringe was maintained. During deployment, the air-filled sampler is placed down hole with the plunger tip sitting within the diffusion barrel (as depicted in Figure 1), allowing a continuous gas phase between the silicon diffusion membrane and the syringe. After an appropriate deployment time (see subsequent sections), the sampler is retrieved from the well. Immediately after retrieval, the plunger is pushed into the syringe, trapping the sample for storage and transport. The sample is subsequently directly injected into a gas chromatograph using a luer adapter fitted to the Valco sampling valve. The sampler is then readily reassembled for redeployment. A metal cage provides protection for the sampler while down hole and the means by which to lower it into position (SI Figure S-3). Where necessary at individual sampling sites, an adjustable pressure valve (Swagelok, CA series) has also been incorporated into the sampler design to facilitate sampling at high hydrostatic and gas pressures (Supporting Information). Total dissolved gas pressure was estimated at the end of the gas sampler deployment period using a commercially available TDGP probe (Common Sensing Inc.). The TDGP VOL. 41, NO. 24, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8389

FIGURE 2. Percent of initial gas concentration vs time during laboratory diffusion experiments to determine the apparent diffusion coefficient for the silicon membrane (DSM) used in the passive gas diffusion sampler. The solid lines represent the analytical solution acquired using the DSM indicated, and the equation developed by Sanford et al. (1996). The dashed lines indicate the analytical solution for DSM values that are (25% greater or smaller than the best fit value. values are reported in percent relative to water equilibrated with the atmosphere (20). Laboratory Estimation of DSM Coefficient and Deployment Times. Equilibration time and thus the length of deployment is one of the most important considerations for in situ samplers (27, 28). Deployment times in purely advective environments (Darcian flow >6 m/yr; (27)) are limited by the time required for the solutes to diffuse through the membrane. Diffusion equilibration times for passive gas samplers in purely advective environments are estimated using the following equation, which is based on Fick’s law of diffusion (29) Cr ) HCC CW (1 - exp [

-DSMA t ]) VL

(1)

Where Cr is the gas phase concentration within the sampler (M L-3), Hcc is the dimensionless Henry’s Law constant, Cw is the dissolved gas concentration in the water (M L-3), DSM is the permeability coefficient in the silicone membrane (L2 T -1), A is area of membrane exposed to the aqueous solution (38 cm2), t is the time of immersion, V is the internal volume of sampler (4.5 cm3), and L is the silicone membrane thickness (0.16 cm). In diffusion controlled environments (Darcian flow 20% by vol.) and CH4 (>10% by vol.) required an increased split flow rate (200–400 mL min-1). Alternatively, increasing the temperature after 5 min from 28 to 120 °C greatly improved the shape of later eluting CO2 and CH4 peaks from the Molsieve 5A column and allowed for better quantification. Therefore, multiple samples could be processed without changing the split flow rate with the incorporation of a simple temperature program for the gas chromatograph. Comparison to Conventional approach. Average dissolved gas concentrations sampled using serum vials and headspace partitioning were more variable than those sampled by PGD. Conversely, PGDSs showed good reproducibility between four replicate samples in the wastewater treatment plant reactor (Table 3). Small differences in concentrations noted between the two sampling methods may be the result of the changes in the wastewater treatment plant reactor conditions at the time of collection (dissolved gas concentrations obtained with the PGDS are timeweighted toward the end of the period of subsurface deployment, whereas the conventional samples represent gas concentrations at a particular point in time) and not due to specific deficiencies in a particular sampling method. However, relatively large differences, such as that observed between the oxygen concentrations (Table 3) are likely the result of the introduction of air to, or loss of gasses from, the sample during collection or laboratory manipulation. Dissolved gas concentrations from PGD sampling at a number of shallow groundwater monitoring sites appear to be less prone to systematic bias that can occur during conventional sampling and manipulation during headspace partitioning and analysis (Figure 4). Lower PGDS O2 concentrations suggest atmospheric contamination occurred during conventional sampling and/or analysis. Argon and N2 concentrations (in the absence of denitrification at these sites) are found in similar concentration in groundwater and water equilibrated with the atmosphere, and have similar concentrations in the two sampling approaches. Methane and CO2 are typically present in higher concentrations in groundwater, and tend to have higher concentrations in the PGD samples, suggesting these gases are lost during conventional sampling and/or headspace extraction. Importance of TDGP in Gas Analysis. TDGP measured in the groundwater at the sampling sites ranged from 100.4 8392

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 41, NO. 24, 2007

FIGURE 4. Comparison of gas concentrations obtained from the passive gas diffusion sampler (PGDS) vs conventional sampling and headspace partitioning and analysis for samples collected from monitoring wells at oil and gas facilities, domestic water wells, and a wastewater treatment plant. Both types of samples were analysed using the gas chromatographic approach reported in this paper. Total dissolved gas pressure (TDGP) is not considered in any gas concentrations presented in this figure.

FIGURE 5. Gas concentrations and field-measured total dissolved gas pressures (TDGP) from a variety of groundwater monitoring locations (including monitoring locations at oil and gas facilities, domestic water wells, and a wastewater treatment plant). to 132% (Figure 5). Groundwaters with significant concentrations of biogeochemically produced gases (i.e., CO2 and CH4) tended to have higher TDGPs. Gases such as N2, O2 and Ar do not appear to be correlated with TDGP suggesting that they are not being biologically produced in substantial quantities at these groundwater sites. Failure to consider total dissolved gas pressure would thus underestimate gas concentrations by as much as one-third.

Acknowledgments Funds were provided by the Canadian Natural Science and Engineering Research Council, Imperial Oil, and Worley Parsons Komex Inc. Dr. E. A. Dixon provided much valued guidance and expertise. Three anonymous reviewers were very helpful.

Supporting Information Available Details on the gas chromatographic method, construction and deployment of the passive gas diffusion sampler, the

collection and preparation of conventional samples, and calculations used to estimate gas concentrations. This material is available free of charge via the Internet at http:// pubs.acs.org.

Literature Cited (1) Ryan, M. C.; MacQuarrie, K. T. B.; Harman, J.; McLellan, J. Field and modeling evidence for a “stagnant flow” zone in the upper meter of sandy phreatic aquifers. J. Hydrol. 2000, 233, 223–240. (2) Heaton, T. H. E. Dissolved gases: Some applications to groundwater research. Trans. Geol. Soc. S. Afr. 1981, 84, 91–97. (3) Stute, M.; Schlosser, P. Atmospheric noble gases. In Environmental Tracers in Subsurface Hydrology; Cook, P.; Herczeg, A., Eds.; Kluwar Academic Publishers: Boston, 2001. (4) Clark, J. F.; Hudson, G. B.; Davisson, M. L.; Woodside, G.; Herndon, R. Geochemical imaging of flow near an artificial recharge facility, Orange County, CA. Ground Water 2004, 42, 167–174. (5) Wilson, R. D.; MacKay, D. M. SF6 as a conservative tracer in saturated media with high intragranular porosity or high organic carbon content. Ground Water 1996, 34, 241–249. (6) Karp, K. E. A diffusive sampler for passive monitoring of volatile organic compounds in ground water. Ground Water 1993, 31, 735–739. (7) Vroblesky, D. A.; Rhodes, L. C.; Robertson, J. F.; Harrigan, J. A. Locating VOC contamination in a fractured-rock aquifer at the ground-water/surface-water interface using passive vapor collectors. Ground Water 1996, 34, 223–230. (8) Vroblesky, D. A.; Hyde, W. T. Diffusion samplers as an inexpensive way to monitor VOCs in groundwater. Ground Water Monit. Rem. 1997, (Summer), 177–184. (9) Interstate Technology Regulatory Council: Technology Overview of Passive Sampler Technologies 2006. http://www.itrcweb.org/ Documents/DSP_4.pdf. (10) Browne, B. A. Pumping-induced ebullution: A unified and simplified method for measuring multiple dissolved gases. Environ. Sci. Technol., 2004, 38, 5729–5736. (11) United States Geological Survey: Dissolved Gas Sampling Instructions 2006. Available at: http//water.usgs.gov/lab/ dissolved-gas/sampling/. (12) Blicher-Mathiesen, G.; McCarty, G. W.; Nielsen, L. P. Denitrification and degassing in groundwater estimated from dissolved dinitrogen and argon. J. Hydrol. 1998, 208, 16–24. (13) Wassenaar, L. I.; Hendry, M. J. Dynamics and stable isotope composition of gaseous and dissolved oxygen. Ground Water 2007, 45, 447–460. (14) Nadim, F.; Liu, S.; Hoag, G. E. The effect of air bubbles and headspace on the aqueous concentrations of volatile organic compounds in sampling vials. Ground Water Monit. Rem. 2001, 89–96. Winter. (15) Widemeier, T. H.; Swanson, M. A.; Moutoux, D. E.; Gordon, E. K.; Wilson, J. T.; Wilson, B. H.; Hansen, J. E.; Chapelle, F. H. Technical Protocol for Evaluating Natural Attenuation of Chlorinated Solvents in Groundwater; U.S. Environmental Protection Agency.: Washington, DC, 1998; http://www.epa.gov/ superfund/resources/gwdocs/protocol.htm. (16) Chappelle, F. H.; Vroblesky, D. A.; Woodward, J. C.; Lovley, D. R. Practical considerations for measuring hydrogen concentrations in groundwater. Environ. Sci. Technol. 1997, 31, 2873–2877. (17) Pearson, F. J., Jr.; Fisher, D. W.; Plummer, L. N. Correction of ground-water chemistry and carbon isotopic composition for effects of CO2 outgassing. Geochim. Cosmochim. Acta, 1978, 42, 1799–1807. (18) Dyck, W.; Da Silva, F. G. The use of ping-pong balls and latex tubing for sampling the helium content of lake sediments. J. Geochem. Explor. 1981, 14, 41–48.

(19) Kampbell, D. H.; Wilson, J. T.; Vandegrift, S. A. Dissolved oxygen and methane in water by a GC headspace equilibration technique. Int. J. Environ. Anal. Chem. 1989, 36, 249–257. (20) Manning, A. H.; Solomom, D. K. Application of a total dissolved gas pressure probe in groundwater studies. Ground Water 2003, 41, 440–448. (21) McCarty, G. W.; Blicher-Mathiesen, G. Automated chromatographic analysis of atmospheric gases in environmental samples. Soil Sci. Soc. Am. J. 1996, 60, 1439–1442. (22) Sarradin, P. M.; Caprais, J. C. Analysis of dissolved gases by headspace sampling gas chromatography with column and detector switching. Preliminary results. Anal. Commun. 1996, 3, 371–373. (23) Restek Applications Note no. 59540 Restek Petrochemical. 1999. www.restek.com/restek/images/external/59540.pdf. (24) Harris, D. C. Quantitative Chemical Analysis (5th Edition); WH Freeman and Company, NY, 1999. (25) Wentworth, W. E.; Vasnin, S. V.; Stearns, S. D.; Meyer, C. J. Pulsed discharge helium ionization detector. Chromatographia 1992, 34. (26) Spalding, B. P.; Watson, D. B. Measurement of dissolved H2, O2, and CO2 in groundwater using passive samplers for gas chromatographic analyses. Environ. Sci. Technol. 2006, 40, 7861– 7867. (27) Harrington, G. A.; Cook, P. G.; Robinson, N. I. Equilibration times of gas-filled diffusion samplers in slow-moving ground water systems. Ground Water Monit. Rem. 2000, 20, 60–65. (28) Vroblesky, D. A.; Campbell, T. R. Equilibration times, compound selectivity, and stability of diffusion samplers for collection of ground-water VOC concentrations. Adv. Environ. Res. 2001, 5, 1–12. (29) Sanford, W. E.; Shropshire, R. G.; Solomon, D. K. Dissolved gas tracers in groundwater: Simplified injection, sampling, and analysis. Water Resour. Res. 1996, 32, 1635–1642. (30) Dasika, R.; Atwater, J. Groundwater Nitrate Profiling by Passive Sampling Over Extended Depth Beneath the Water-Table - A Demonstration. Water Res. 1995, 29, 2609–2612. (31) Ronen, D.; Magaritz, M; Itzbak, L. An in situ multilevel sampler for preventative monitoring and study of hydrochemical profiles in aquifers. Ground Water Monit. Rem. 1987, 7, 69–74. (32) Martin, H.; Patterson, B. M.; Davis, G. B. Field trial of contaminant groundwater monitoring: comparing time-integrating Ceramic dosimeter and conventional water sampling. Environ. Sci. Technol. 2003, 37, 1360–1364. (33) Kampbell, D. H.; Vandegrift, S. A. Analysis of dissolved methane, ethane and ethylene in groundwater by a standard gas chromatographic technique. J. Chromatogr. Sci. 1998, 36, 253– 256. (34) Walsh, K. P.; McLaughlan, R. G. Bubble Extraction of Dissolved Gases from Groundwater Samples. Water Air Soil Pollut. 1999, 115, 525–534. (35) Manahan, S. E. Environmental Chemistry; Lewis Publisher: NY, 1994. (36) Ettouney, H; Majeed, U. Permeability functions for pure and mixture gases in silicone rubber and polysulfone membranes: Dependence on pressure and composition. J. Membr. Sci. 1997, 135, 251–261. (37) Colt, J. Computation of Dissolved Gas Concentrations in Water as Functions of Temperature, Salinity, and Pressure. Am. Fish. Soc., 1984, Special Publication No. 14. (38) Hayduk, W.; Laudie, H. Prediction of Diffusion Coefficients for Non-electrolytes in Dilute Aqueous Solutions. AIChE J, 1974, 20, 611–615.

ES0716094

VOL. 41, NO. 24, 2007 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

8393