WATER QUALITY--Guidelines on Collection of Ground Water Samples for Analysis of Organic Compounds In Reply Refer To: March 21, 1983 EGS-Mail Stop 412 QUALITY OF WATER BRANCH TECHNICAL MEMORANDUM NO. 83.12. Subject: WATER QUALITY--Guidelines on Collection of Ground Water Samples for Analysis of Organic Compounds In recent years, the Water Resources Division has witnessed a significant increase in the level of interest and concern, both among our cooperators and internally, regarding contamination of ground waters with manmade organic substances. At the present time, there are no Geological Survey-approved standard methods for collecting ground-water samples where the constituents of interest are manmade organic compounds and, indeed, it appears that much additional research and practical field experience is needed before acceptable methods are established and described in approved Techniques of Water Resources Investigations (TWRI) format. The materials used in casings, sampling devices, and containers must be tested for contamination or adsorption effects; quality-assurance procedures need to be developed and implemented; for some constituents at some concentrations, new approaches to sampling may be needed and some are being studied and tested at the present time. It is anticipated that much of the research needed to develop or improve and test methods of sample collection will be performed at research sites supported by the Toxic Waste-- Ground Water Contamination thrust program, and it is hoped that each project leader will be conscious of the need to design into each project experiments which evaluate the integrity of the sample collection and handling procedures. Research performed elsewhere and reported in the open literature will also add to the technology of ground-water sampling in coming years. For the time being, in the absence of approved Survey methods, the Quality of Water Branch, with the assistance of the Toxic Waste Program Coordinator, will periodically distribute to the field offices publications, committee reports, or citations of literature articles dealing with this subject. These materials, we hope, will provide satisfactory interim guidance to personnel engaged in data collection and interpretive work 1n ground-water resources. In addition, the Quality of Water Branch and the Office of Hazardous Waste Hydrology will assist, if needed, in the design of ground-water sampling programs. In March 1982, the New Jersey District circulated a questionnaire to several persons within the Division with experience in ground-water sampling to obtain information that would assist them in design of sampling programs and in making the most appropriate choices for materials and methods to use, based on the current state of knowledge. The Organic Substances Task Group, an advisory committee of the Quality of Water Branch with membership from the laboratories, the research program, the operational program, and the Branch, also discussed the questionnaire at length and reported on those discussions on November 30, 1982. The portion of that report dealing with the questionnaire is attached for your information and guidance. In addition to providing preliminary guidance on materials and methods to use when various substances at various concentrations are sought or expected in ground waters, the report recommends the development of other equipment, analytical methods, and quality-assurance procedures. These recommendations are being considered at the present time. For additional guidance, your attention is directed to the following publications: Pettyjohn, R. A., Dunlap, W. J., Cosby, R. and Keeley, J. W., 1981, Sampling ground water for organic contaminants. Ground Water, v. 19, no. 2, p. 180-189. Scalf, M. R., McNabb, J. F., Dunlap, W. J. Cosby, R. L., and Fryberger, J., 1981, Manual of ground-water sampling procedures: National Water Well Association, Worthington, Ohio, 93 p. Gibbs, J. P., Schuller, R. M., and Griffin, R. A., 1981, Procedures for the collection of representative water quality data from monitoring wells: Cooperative Ground Water Report No. 7, Illinois State Water Survey, Champaign, Illinois. Dunlap, W. J., McNabb, J. F., Scalf, M. R., and Cosby, R. L., 1977, Sampling for organic chemicals and microorganisms in the subsurface: USEPA-60O/2-77-176, 26 p. Also, the following articles in the winter 1983 issue of Ground Water Monitoring Review are of particular value. Voytek, J. E., Jr., Considerations in the design and installation of monitoring wells, p. 70-71. Rinaldo-Lee, M. B., Small- vs. large-diameter monitoring wells, p. 72-75. Johnson, T. L., A comparison of well nests vs. single-well completions, p. 76-78. Cadwgan, R. M., Barvenik, M. J., Ehrenfried, A. D., and Ullinskey, G., Improving monitoring efficiency of deep wells, p 110-118. Wilson, L. G., Monitoring in the vadose zone: Part III, p. 155-166. R. J. Pickering Attachment WRD Distribution: A, B, FO, PO Key Words: Water quality, sampling, organic compounds, ground water, toxic waste. This memorandum does not supercede any previous memorandum. GUIDELINES ON SAMPLING GROUND WATER FOR ORGANIC COMPOUNDS The following discussion resulted from the deliberations of the Task Group over a questionaire prepared by the New Jersey District about March, 1982 and circulated to several researchers and district project personnel with experience in sampling ground waters from production and observation wells. A copy of the questionaire is attached for reference. The Task Group found that many of the questions could be answered only in the context of the project objectives or the local hydrologic conditions, but that several reflected a genuine lack of knowledge and techniques within the Division regarding materials and methods to use in obtaining samples of ground water for organic compound analysis. In particular, there is a need for l) well-designed devices to sample purgeable and non-purgeable organic compounds and 2) documentation of casing and sampler materials and sampling techniques specific for compounds of interest under different hydrologic conditions. The discussion begins with general guidelines based on the collective experience of the Task Group members, followed by answers to the questions presented by New Jersey and closes with recommendations for action or studies which should be considered to improve Division practices in this area. PURGEABLES o General Discussion Purgeable organic compounds are operationally defined as those having <2% solubility by weight in water and boiling points 100 are hydrophobic. The non-purgeables have molecular weights greater than phenol. Generally the toxicities of homologous series are positively correlated with molecular weights and octanol/water partition coefficients. The hydrophobic compounds are strongly but reversibly sorbed to surfaces so that cross-contamination through the sampling device is always possible and must be guarded against, and most will reside on suspended material if it is present. Since the non-purgeables have little tendency to degas, there is no need for the sampling device to be sealed unless the analytes of interest are present in low concentrations and degraded by aerobic bacteria. Many priority pollutants are both. But there are a multitude of other mechanisms by which samples may be altered by casing, pumping, sampler and sample container that more than make up for the sampling problems avoided by low volatility. o Casing Material The casing material of choice depends upon the concentration and corrosive action of the organic compounds in the aquifer. The most corrosive materials are the solvents (TCE, methylene chloride) and the acids (acetic, formic and citric). 1. High concentrations of non-corrosive materials! The term "high" implies elevated with respect to the potential for contamination from the casing. It is difficult to quantify by a single independent measure such as DOC since "high" will vary from compound-to-compound. In this situation, the casing material can be most anything, even PVC, so long as the joining compound is used sparingly. 2. Low concentrations of anything! It is absolutely essential that galvanized, copper or stainless be used. Casing and joints must be cleaned before being set with high-pressure hot water. 3. Corrosive materials! Stainless, galvanized and copper have limited lifetimes when exposed to corrosive organic materials, particularly organic acids, which complex with the metals causing dissolution of the casing and accumulation of precipitate at the bottom of the well. In the injection wells in Florida for example, stainless casing are replaced every 5 years. In extremely corrosive environments, a fiberglass casing may be needed. o Pipe and Pump Materials 1. High concentrations of anything! The pipe and pump materials can be stainless, galvanized, copper, Teflon or PVC. The joining compound should be used sparingly and the pumping system should be used on more than one well only with great caution. In cases of severe contamination, a pumping system should be dedicated to one well or, at most, wells showing comparable levels of contamination. Do not clean the system after use; no procedure is fully effective. But considerable pumping time should be allowed at each sampling to equilibrate the system with the water to be sampled. One obvious precaution when sampling a production well is to avoid oil-lubricated pumps wherever possible. Where one cannot be avoided, check on the oil consumption rate and check for oil standing on the water with a measuring tape. 2. Low concentrations of anything! Stainless and galvanized are the best materials followed by Teflon. The sorptive properties of Teflon make it less desirable, especially when small diameter tubing is used. The name Teflon applies to a number of chemically distinct products that vary in important chemical and physical properties. Some kinds of Teflon are more suitable than others. The high surface to volume ratio of smaller tubing (<1/2 inch) increases the probability of significant sorption of hydrophobic materials over that of larger tubing. PVC and polyethylene should not be used in any case. Any plastic will leach phthalates and an assortment of other materials, and act as strong hydrophobic sorbants. o Sample Collection, Sample Containers and Preservation The methods for pumping an observation or production well described for sampling the purgeables are satisfactory for sampling non-purgeables as well. The methods are the best we have seen to date at isolating water fresh from the bearing zone and minimizing the fraction of old or dead water sampled. Where hydrophobic constituents are sought or anticipated and suspended sediment appears in the sample, it would be advisable to filter the sample through 0.45u silver membrane filter and submit the filter for analysis. The sample container should always be glass baked at 450-500 C. Baking forms a relatively non-reactive surface which minimizes sorption. Avoid plastic or acid-washed glass. Caps should be lined with aluminum foil, dull side in (the shiny side is coated with an organic compound). Teflon caps (as in the septum bottles) are satisfactory but contact with the cap should be minimized by standing the bottle upright during transit. No method of preservation is completely satisfactory but none has been found better than chilling. Where biodegradation is possible, the bottle should be topped. Samples should be shipped chilled by the fastest route and analyzed as quickly as possible. Some workers have observed two or more phases in water samples from highly contaminated wells. If multiple phases are present in the aquifer, sample collection and data interpretation are complicated. First, the pumping rates and the locations of screens will affect the mix of phases in the sample. Secondly, the analysis of mixed phases is custom and the analytical values have little meaning with respect to concentrations in the aquifer. For the present, when mixed phases appear in a sample, the phases would be separated in the field service units to avoid changes in transit, and the phases analysed separately. The interpretation of the data must be left to the observer. For example, an immiscible phase separated out in the field will probably contain compounds that would have been in the water or suspended material in the absence of the immiscible phase. Decisions regarding the treatment of such data must be left to the judgement of the researcher. However, guidelines on methods for handling immiscible fluids in the field should be developed and made available to the Division. Responses to Questions 1. Which type of well is preferable, production or observation? Neither is preferable either from a hydrologic or a chemical viewpoint. The objectives of the study may dictate the spatial representativeness of water from a well which, in turn, will constrain the pumping rate. Sample contamination depends largely on the casing, pipe and pump materials used. 2. Well design characteristics? The general rule to avoid multiply-screened wells wherever possible applies here. The mix of waters drawn from the various bearing zones of a multiplyscreened well depends on the rate of pumping and a number of other variables in a complex and undefinable way, sufficient to obscure the vertical definition of aquifer characteristics. 3. Sampling strategies in consolidated versus unconsolidated formations? The question must be addressed at the local level in an analysis of the hydrologic conditions. 4. Are submersible pumps preferred? Not necessarily but they are handy to use subject to the comments of the general discussion. 5. Experience with the Johnson-Keck submersible pump? Tennessee used this pump in a study of a contaminated aquifer. It is stainless with either Viton or EPDM stator and rotor, compact, cleanable with solvents and autoclavable. Teflon stators and rotors may be available in the near future. This should be a good pump for use where low levels of contaminants are being studied. The pump was purchased by HIF and is being fieldtested in Maine. The cleanup procedure where hydrophobic materials are encountered would include an alkaline detergent scrub, hot water rinse, methylene chloride rinse and an alcohol rinse. 6. Location of submersible pump? Surface skimming? See the procedure described for sampling purgeable organic compounds. Regarding sampling surface skims, see the general comments under nonpurgeable compounds, mixed phases. 7. Use of a thief sampler? The discussion has assumed use of submersible pumps because of their convenience. But a thief sampler may be the best device to use for low capacity observation wells. The casing should be evacuated with a bailer, air-lift or submersible. After the water level has recovered, drop the thief to the midpoint of the screen or below the top of the water, whichever is deeper. A thief is less desirable than a submersible for taking a sample for the purgeable compounds because the sampler must be opened to withdraw the sample. 8. Evidence for removal of "dead" water? None of the parameters that an observer may monitor is a reliable measure of the removal of dead water. The mechanistic approach described for sampling the purgeable compounds is the best. 9. Flow rate for sampling observation wells? There is no universally appropriate flow rate. The rate of choice depends upon the characteristics of the well and the bearing zone. See the general discussion. 9a. Sampling from oil-lubricated production wells? Obviously, sampling from water-lubricated wells is preferable. Where use of an oil lubricated pump cannot be avoided, the observer should be aware of the condition of the seals and the material used in them. Butyl rubber leaches PNA's and other hydrophobic materials. One may expect contamination from the seals and oil to include parafins, polyolefins, plasticizers and halogenated biphenyls. Contamination from purgeable materials is of less concern. 10. Pipe, pump and casing materials? See general comments. 11. Is PVC acceptable? PVC can be used only where concentrations of the compounds of interest are high, well above the levels potentially leached from the PVC. See general comments. 12. Contamination from a PVC pumping system? Generally, PVC is not acceptible for use on the pumping system. The hydrophobic materials strongly but reversibly sorb and there is always a question about the effectiveness of cleanup between samplings. 13. Line pressure affecting adsorption by PVC? Fast flow rates minimize contact but one cannot be certain that hydrophobic compounds do not sorb and desorb under even the highest flow rates. 14. Solvent attack of PVC? An article in a recent issue of Ground Water discussed materials to use for sampling. The authors recommended in order of preference; stainless> Teflon> glass> PVC. We think galvanized is nearly as good as stainless. They also did some experiments on leaching of PVC and other polyvinyl plastics by solvents. Generally, stainless and galvanized are resistant to attack by solvents but PVC is not. When there is any leaching at all, the level of contamination is always uncertain. 15. Replacement of PVC tubing? See general comments. 16. Sampling for purgeable compounds? See general comments. 17. Temperature effects on the septum bottles? Gas losses in a reaeration study are small unless an air bubble is introduced. But it is felt that the Fisher Tube is more reliably sealed than the septum bottle. 18. Permissible storage times for purgeables? The presence of a gas bubble indicates loss of a purgeable compound to the gas phase, which cannot be sampled with the resent system. The absence of an air bubble does not guarantee against losses by other mechanisms. Presently the storage guideline is 14 days under refridgeration but samples should be analysed as soon as possible. i9. Preparation of EPA reference samples? The Task Group has no experience with the EPA vials of reference materials. But the following guidelines seem appropriate in lieu of specific instructions from EPA. The reference material comes in vials dissolved in methanol. Use a 25uL chromatographic syringe. Inject an appropriate volume at the bottom of the septum bottle (insert needle through the septum) or deep into the Fisher Tube from one end. 20. Rate of check sample flow to the lab? One blank, one spike and one duplicate per shipment at least and additional combinations not to exceed 20% of the samples. Blanks, spikes and duplicates sample provide needed assessment of the usefulness of any particular set of analyses but because of the rapidly changing capabilities and needs, and the likelihood of detecting substances not anticipated for which the sampling program was optimised, workers may frequently discover that the results are not good enough for a specific purpose. Although the general guidelines for spiked, blank and duplicate samples may be satisfactory for some purposes, the decision as to what is good enough ultimately rests with the data user. RECOMMENDATIONS 1. The Central Laboratories should acquire Grob Closed Loop stripping devices and modify them as appropriate to mate with a Fisher Tube (also modified) and another device recommended below for development, to maintain a sealed sample from collection through analysis for the purgeable compounds. 2. A system for collecting and analysing water samples for purgeable organic compounds should be developed having the following features; a. Can be lowered down the well. b. Will open, collect and seal while in the well. c. Can be shipped without opening to the laboratory. d. Has a removable septum through which a spike or internal standard may be injected. e. The preferred material is glass; the sampler should be capable of being baked at 450-500 C. f. Will mate with a Grob Closed Loop stripping device in the laboratory. 3. In addition to developing a down-hole sampler, the Fisher Tubes should be modified as follows: a. Fitted with a removable septum through which a spike or internal standard may be injected. b. Modified to mate with a Grob Closed Loop stripping device in the laboratory such that the tube need not be opened to the air to remove the sample. 4. The downhole sampler and modified Fisher Tubes should be distributed by the Central Laboratories or HIF as part of a sampling kit that would include standards for injection into the sample in the field. The labs or HIF should also be responsible for cleaning and repairing the devices for reuse in the field. 5. HIF and the Central Laboratories should develop a program to evaluate pumping systems and procedures for their use, including cleanup procedures. The more expensive pumping systems should be purchased by HIF and rented to the Districts. 6. An assortment of mixed standards and surrogate internal standards for injection into samples in the field, and procedures for their use, should be developed by the Division. These materials would permit quality control procedures to be carried on from the point of sample collection. 7. Procedures for sampling dissolved, solid and immiscible phases should be incorporated into a TWRI. NEW JERSEY DISTRICT QUESTIONAIRE ON GROUNDWATER QUALITY SAMPLING Questions pertaining to pumping a well: 1. Which type of well (production or observation) is considered preferable in a trace organics sampling program for VOC's, phenols, phthalate esters, etc. ? 2. How does the design of the well (casing construction, location of screened interval, multiple screens) af#ect the suitability of that well for sampling metals and trace organics (compounds with varying degrees of solubility)? 3. How does the sampling strategy differ in a consolidated (rock) system vs. an unconsolidated (coastal plain) system? 4. Is a submersible pump the pre#erred type of pump to be used whenever possible in sampling an observation well? Should the use of gas - reciprocal pumps be limited to small diameter observation wells? 5. Have you had any experience with the Johnson-Keck (1.75" O.D.) submersible pump? 6. Should the submersible pump be located near the pumping head during sampling, or should it be located nearer to the screened interval? Is there ever any value in skimming the surface water in a well to collect certain organic substances? 7. Should a thief sampler be located at the screened interval and an appropriate sample be taken from it instead of at land surface? If so, what type of sampler would be suitable, and when should it be used? Should such a sampler be dedicated to only one site? 8. What is considered sufficient evidence of the evacuation of "dead water" from the casing of a pumped observation well? Should we monitor temperature, pH, and specific conductance during the pumping process until stability is achieved? Are there any other parameters that should be included in this monitoring scheme? Is it.valid to simply evacuate a certain number of casing volumes prior to collection regardless of field parameter stability? 9. What is considered a minimum flow rate necessary to correctly sample an abandoned production well or observation well? Could this rate be dependent on a minimum velocity of flow within the casing? What kind of effect does deteriorated or seldom used casing have on sample quality? Should we question the use of seldom used production wells in an organics sampling network? 9a. Is there a significant differencc between watcr and oil lubed production wells in terms of organic contamination from oil? Questions pertaining to material construction: 10. What type of material is prefered in the construction of pumps and piping that transport sample water to land surface for the collection of metals and various trace organics samples? Is PVC tubing (flexibe) or PVC pipe (glue joint or threaded) considered acceptable? How about stainless steel or tin or teflon lined pipe? 11. If PVC is acceptable, would there be any preference in minimizing the length of tubing, where possible, that the sample would have to travel through to the point of collection? 12. How can we effectively demonstrate that our present pumping system (consisting of a Gould's submersible pump 1/3 H.P., 200' of 3/4" I.D. flexible PVC (clear) tubing, and a rigid PVC pipe delivery system with pressure regulator) does not cross-contaminate samples? Should the system be flushed or cleansed with de-ionized water (or soap) between uses? 13. Does the line pressure in the PVC tubing have any effect on the adsorption of hydropholic organics onto the tubing? Do fluid velocity, residence time within the tubing, or the state of reactivity (pH) of the sample water have an effect on trace organics being transported through the system? 14. Sample water that contains organic solvents probably will attack PVC tubing, but under what conditions and at what concentrations? How can we know that such an attack has taken place? What types of compounds might be leached from the tubing? Would such damage to the tubing be permanent? 15. Should tubing be replaced on a regular basis? How would you determine the time interval? Would "food grade" PVC tubing be preferrable to "non-food grade", and why? Other related questions: 16. Are VOC vials the most effective collection and storage vessel for a VOC sample? We suspect that manufacturing defects which cause irregularities at the opening of the vial can cause an imperfect seal with the teflon membrane. To what extent is this a problem? Can significant defects be visually observed? Do the resultant air bubbles invalidate the sample? 17. What about the effects of temperature change (particularly in the 1! - 4 C range, where the density of water changes dramatically) on the integrity or the VOC sample? Can such temperature changes induce air/water flow around or possibly through the membrane? 18. How long can a refrigerated VOC sample be stored prior to analysis? Is the absence of an air bubble an adequate indicator of sample integrity? 19. What is the proper analytical proceedure and equipment needed to prepare EPA standard rcference samples (from vials of trihalomethanes) for VOC's in the field? Do we mix the sample in an open Cantainer before pouring into vials? Is the system even useful in developing check samples in the field? Should we use micro-syringeg for measuring the buffer stock and Type 1 reagent grade water for dilution? How would we, on the district level, quality assure this preceedure? 20. How often should VOC check samples (spiked and blank) be sent to the central laboratory? How would these samples, and the results obtained from this effort be merged into the existing quality assurance program at the central laboratory?