Groundwater Monitoring with the Waterra Inertial Pump
Abstract
The Waterra Inertial Pump has been available for use in narrow diameter monitoring wells for several years. This device offers many advantages over other types of sampling equipment. Since this pump operates via a principle not used by any other type of device and is commonly constructed of materials not widely used in field sampling, some questions exist regarding its suitability. A review of published literature concerning the effect of different polymer materials clearly shows that the materials used do not have a measurable impact on sample quality. Additionally, published equipment tests as well as recent field testing of the Waterra Inertial Pump had show it to be an excellent performer in narrow diameter monitoring wells for sampling volatile organics.
Introduction
Narrow diameter monitoring wells (2 inch) have been used in Canada for many years now. This type of well offers many advantages over the traditional 4' or 6' well. Narrow diameter wells are less expensive to install, and materials costs as well as the volume of cuttings produced are reduced. In addition, the purge volumes in 2' monitoring wells are significantly less than in larger wells. This can produce significant savings in labour and operating costs especially if the purge water must be containerized. Narrow diameter monitoring wells do create some problems as they limit the types of equipment suitable for use in these wells.
Waterra Pumps Limited has developed a line of pumps designed specifically for use in narrow diameter monitoring wells. This paper outlines the advantages of the Waterra Inertial Pump and discusses some of the concerns regarding groundwater sampling devices.
Sampling devices for narrow diameter monitoring wells
Prior to the availability of the Waterra Inertial Pump, several methods were used for developing, purging and sampling narrow diameter monitoring wells. These methods are characterized by the types of equipment involved, either simple devices or complex devices.
The most common and simple method of purging and sampling monitoring wells is bailing. Less commonly used and more complex are air lift devices and bladder & piston (positive displacement) pumps. The advantages and disadvantages of these types of equipment are listed here.
ADVANTAGES | DISADVANTAGES |
simple design | inefficient |
easy to operate | prone to contamination |
can be dedicated | cross contamination possible if not dedicated |
inexpensive | |
portable |
ADVANTAGES | DISADVANTAGES |
high purge rate | poor portability, requires compressed gas cylinder or compressor |
can be dedicated | continued operating cost due to consumption of nitrogen |
inexpensive | not suited to wells less than 1/3 submersed |
air-water mixing, not suitable for sampling |
ADVANTAGES | DISADVANTAGES |
good purge rate | poor portability, requires compressor and controller |
good sampling system | expensive to dedicate because of high unit cost |
inexpensive | continued operating cost due to reagent used in cleaning and rinseblank analysis if not dedicated |
Inertial pump
This selection of equipment left the groundwater geologist with a choice between inexpensive, inefficient equipment that may produce samples of dubious quality or expensive equipment which is difficult to transport and offers the potential of cross contamination if the equipment is not dedicated.
Waterra's goal was to produce a narrow diameter monitoring well pump that would bridge the gap between the existing methods. The task was to develop an affordable well purging and sampling device that was efficient, reliable and suitable for dedication. The result was the Inertial Pump.
ADVANTAGES | DISADVANTAGES |
simple design | pump efficiency decreases in larger diameter wells |
easy to operate | some automated equipment is heavy |
suitable for dedication | |
inexpensive | |
suitable for sampling | |
good purge rate |
Groundwater sampling and pump materials
A common concern in groundwater monitoring is the possible negative effect that polymer materials may have on sample chemistry. Numerous papers have been published discussing 'Sorption effects'. All polymer materials, Teflon included, will sorb certain organic chemicals from a water sample. The rate at which sorption occurs is dependent upon several parameters. The most important parameters are the type of organic chemicals present in the sample and the type of polymer material that these chemicals are in contact with. The typical sorption study procedure consists of sealing a piece of tubing in a container with an organic solution of a known concentration. Alternatively the organic solution may be sealed in a section of tubing. Samples are drawn from containers at various time intervals and the concentration of the organics are measured. Any loss in the concentration is deemed to be a result of tubing sorption.
In general, blended chemical species tend to sorb faster than singular species and more rigid (crystalline) polymers tends to sorb at a slower rate than more flexible tubes. A commonly referenced paper concerning the effect of tubing materials on sample quality is 'Sample Tubing Effect on Groundwater Samples' by M.J. Barcelona et al. In this study samples are drawn from sealed lengths of 0.25 inch ID tubing. Five different tubing materials were studied, Teflon (TFE), polypropylene (PP), polyethylene (PE), Tygon (flexible vinyl PVC) and silicon rubber (SIL). The first two samples were collected after 5 minutes of exposure and the balance at 10 minute intervals.
There are two problems with the procedure used by Barcelona et al. First, rarely is sample water in contact with the sampling device, including bailers, for any time near these durations. Secondly the use of 0.25 inch ID tubing maximizes the surface to volume ratio, producing an example that overstates the significant of sorption. Conventional monitoring well pumps never use tube of this size and bailers are certainly never constructed with 0.25 inch ID pipe.
Nevertheless, the date produced in this study illustrates two important facts. The first is that with relatively short duration's of exposure, 5 to 10 minutes, the performance differences between Teflon, polypropylene and polyethylene are insignificant. Secondly, desorption, the release of sorbed organics is irrelevant as none of the tubing materials released any significant portion of the organic compound they sorbed.
The insignificance of sorption is further illustrated in table 4 from 'Sampling Tubing Effects on Groundwater Samples', reproduced below. The sample residence time has been added to this table. This table compares the performance of the various tubing materials studied in a hypothetical model. The sorbative loss is calculated as a result of passing a 40 ppb mixture of chloroform, trichloromethylene, tetrachloroethane and tetrachloroethylene through a 15 meter length of tubing at a rate of 100 ml/minute. Three different tubing IDs are evaluated.
TUBE ID | TFE | PP | PE | PVC | SIL | RESIDENCE TIME |
1/4' | 4% | 6% | 10% | 14% | 15% | 4.5 minutes |
3/8' | 1% | 2% | 3% | 4% | 4% | 12.0 minutes |
1/2' | 1% | 1% | 1% | 2% | 2% | 19.5 minutes |
The two main parameters that determine sorbative loss with a given organic and polymer are the surface area of contact and the duration of the contact.
The residence time of the sample in the tubing in the above example is a function of the tubing ID. Since the larger diameter tubes have a larger storage volume and the flow rate is fixed to 100 ml/minute the sample requires more time to move through the tube. Initially one may think that this would result in greater sorbative loss however increasing the diameter of the tube significantly increases the volume of fluid within the tube relative to the surface area of the tubing available for sorption to take place. This reduces the impact that sorption can have on a sample.
This comparison is a worst case example because of the extended contact time between the solution and the tubing material. In a typical 100' installation using 1/2' ID tubing, flows are commonly 1 gpm. This flow rate results in a contact time between the sample and the tubing of 1 minute. Nevertheless, the percentage loss predicted by the model illustrates even in a worst case situation that the loss due to sorption in insignificant.
There is no question that polymer tubing materials sorb organics. What is questioned is the belief that the use of Teflon, instead of less expensive materials, results in a measurable improvement in sample quality. Most studies now conclude (Devlin, J.F., 1986, Schalla et al., 1988 and Thomey et al., 1991) that this is not the case. Field conditions and the method of sample collection are far more significant in their ability to bias sample analysis.
Sample bias due to volatization loss –The Barker & Dickhout Study, 1988
Volatilization loss is recognized as the most significant cause of negative sample bias in subsequent concentration determinations. Laboratory evaluations of sampling bias due to volatilization (Barcelona et al., 1984, Schalla et al., 1988) have predicted little to no variation of volatile halocarbons in samples collected with different sampling devices. The samples used in these studies contained low concentrations of volatile organics, less than 100 ppb and the lifts required for sampling were shallow, usually less than 20 feet.
Gas-charged groundwater is more likely to produce conditions conducive to significant volatile losses even with low levels of VOCs present. Barker and Dickhout (1988) evaluated the performance of various sampling systems for sampling gas-charged groundwater. The relative recoveries of volatile aromatic hydrocarbons with a peristaltic pump, a positive displacement bladder pump and an inertial lift 'Waterra Pump' were compared when sampling methane gas-charged groundwater. The suction lift peristaltic pump produced 9 to 13 percent negative bias relative to other methods. The inertial lift pump and positive displacement bladder pump produced similar results.
The performance of these sampling systems was further evaluated in an artificial well. Samples were drawn from a port near the bottom of the artificial well, as a control, and from the sampling devices simultaneously. The inertial lift pump consistently produced the highest concentrations of volatile hydrocarbons, the bladder pump concentrations were 13 to 20 percent lower and the peristaltic pump concentrations were 23 to 33 percent lower. This experiment was repeated with the Waterra Pump with non-degassing water from the artificial well. Under these conditions the Waterra pump suffered virtually on loss of volatile organics. Barcelona et al. (1984) also reported that the bladder pump and bailer produced no serious bias (<2 percent) under non-degassing conditions when sampling trihalomethanes.
Leaching
Eliminating possible sources of contamination of groundwater samples has resulted in the evaluation of most of the materials used in pump and well construction as possible sources. The main source of contamination of samples by sampling equipment is from the use of flexible PVC tubing or bottles. Flexible PVC tubing contains a high percentage of plasticizers, usually a phthalate ester. These compounds rapidly leach into samples and can appear in subsequent analysis.
The potential of virgin polyethylene contaminating sample water by leaching has been evaluated (Devlin, J.F. 1986) This material did not substantially contaminate organic free deionized water, Polyethylene is considered to be leach free as it does not contain any plasticizers. However polyethylene was found to be difficult to decontaminate and should consequently not be used in non-dedicated pumping systems in order to avoid cross contamination.
Unpublished work
Several unpublished field studies have been completed that compare the performance of the Waterra Inertial Pump to other sampling devices. These studies have evaluated the relative performance of the various sampling devices for VOC sampling.
The new jersey study
The first study was completed by CEH, Inc. of Portsmouth, NH in 1990. The New Jersey Department of Environment (DEP) requested that CEH, Inc. collected duplicate sample sets from two wells with Teflon bailers and Standard Waterra Inertial pumps (Delrin D-25 valve and polyethylene tubing). The state regulators requested this study in order to support manufacturers claims that the Inertial pump is suitable for use as a sampling device. These duplicate sets were submitted to the same laboratory for analysis. The results, given below, are reported in ppbs.
WELL # MW-9 | TEFLON BAILER | WATERRA INERTIAL PUMP |
1,2 - Dichloroethyene | 1100 | 1100 |
1,1,1-Trichloroethane | 420 | 420 |
Trichloroethane | 11000 | 11000 |
Tetrachloroethene | 2100 | 2000 |
WELL # MW-10 | TEFLON BAILER | WATERRA INERTIAL PUMP |
1,2 - Dichloroethyene | 3100 | 1100 |
1,1,1-Trichloroethane | 990 | 960 |
Trichloroethane | 18000 | 17000 |
Tetrachloroethene | 3700 | 3300 |
The minor difference between the sample pairs were not considered significant and the company was permitted to use the pump as the sampling device on the project. It should also be mentioned that CEH, Inc. did not use the recommended technique for collecting volatile samples from the Waterra Inertial pump.
The oakridge national labs study
A second study in 1990 was carried out by Oak Ridge National Labs of Oak Ridge, TN. This study compared samples collected with Teflon bailers, Teflon and stainless steel bladder pumps and the Standard Waterra Inertial pump. Seven monitoring wells were sampled, twice with the bladder pump and once each with the Waterra pump and bailer. Trichloroethylene (TCE) was detected in all of the wells at very low levels. The results reported below are in ppbs.
WELL # | BLADDER #1 | BLADDER #2 | BAILER | WATERRA |
UG02-U | 12.4 | 16.0 | 13.6 | 12.6 |
MW072-U | 1.9 | 1.8 | 1.5 | 1.0 |
MW072-L | 2.5 | 2.6 | 2.2 | 2.6 |
UG03-L | 8.5 | 8.1 | 3.1 | 6.8 |
UG06-U | 11.4 | 10.0 | 11.2 | 10.2 |
UG06-L | 15.2 | 15.0 | 8.9 | 9.7 |
MW074-L | 18.9 | 16.0 | 1.3 | 16.7 |
From this sample set only one sample, from the bailers in well MW074-L, showed a dramatic variation from those collected by other devices. All other samples show goods agreement between all sampling devices.
Oak Ridge National Labs proceeded to further evaluate the Waterra Inertial Pump at a site in Kansas City, KS. This site was more severely contaminated and three different contaminants were found in the analysis, dichloroethylene (DCE), TCE and vinyl chloride. Eighteen monitoring wells were sampled with Teflon bailers and Standard Waterra Inertial pumps. Four bailer duplicates and three Waterra duplicates were collected.
MONITORING WELL | DCE BAILER/WATERRA |
TCE BAILER/WATERRA |
VINYL CHLORIDE BAILER/WATERRA |
113-U | 72/68 | 27/21 | 53/45 |
113-L | 62/65 | ND/ND | 28/28 |
114-U | 109/111 | ND/ND | 53/58 |
114-L | 308/420 | 62/66 | 67/73 |
115-U | 1000(1000*)/1500 | 62(ND*)/60 | 84(83*)/86 |
115-L | 990/1200(1300*) | 76/114(120*) | 47/51(46*) |
66-L | 74/78 | 8/8 | 12/12 |
40-L | 14/14 | ND/ND | ND/ND |
106 | 62(76*)/75(76*) | 13(16*)/12(12*) | ND(ND*)/ND(11*) |
24 | 17(18*)/19 | ND(ND*)/ND | 13(18*)/16 |
83-L | 18/14 | ND/ND | ND/ND |
95-L | 570/2000(1500*) | 48/100(ND*) | 69/280(250*) |
32-M | 126/240 | ND/ND | 6/ND |
32-L | 35/29 | 5/ND | ND/ND |
33-L | 40(36*)/40 | 54(50*)/61 | ND(ND*)/12 |
139-L | ND/ND | ND/ND | ND/ND |
38-L | ND/ND | ND/ND | ND/ND |
(*) duplicate sample |
The sample set from Kansas City produced some interesting results. Firstly it seems that the laboratory occasionally had problems detecting vinyl chloride and TCE because these compounds appeared erratically in the results. The TCE analysis from some of the bailer and Waterra samples in 95-L and 115-U as well as the vinyl chloride results in 106 were inconsistent.
Other than the periodic problems in the TCE and vinyl chloride analysis, these sample results show good relative correlation between the sampling devices except in monitoring wells 115-L, 95-L and 32-M. In these wells there are marked differences between the Waterra pump and the bailer, with the Waterra pump consistently producing the higher result. These samples are also from the wells that had the highest level of contamination.
Conclusions
The analytical results from the Kansas City and Tucson sites are consistent with observations of Schalla et al. (1988) in that different sampling devices usually give good agreement in duplicate samples when dealing with low levels of contamination. Although the data sets are not of sufficient size to make firm conclusions, it seems that the Inertial pump may be better able to provide a less disturbed sample of more severely contaminated groundwater than some other devices.
The review of published information clearly demonstrates that the use of HDPE in the place of Teflon does not produce any detectable difference in the sample chemistry. The impact of materials is further reduced if the routine of standard field procedures are considered. Normally the device used to purge the monitoring well is the same as that used to collect the sample. This prolonged flushing during well purging essentially eliminates any difference between all polymer materials.
References
Barcelona, M.J., Helfrich, J.A., Garske, E.E.
1985 'Sample Tubing Effects on Ground Water Samples' Anal. Chem., V. 57. No. 2, pp 460-464
Barcelona, M.J., Helfrich, J.A., Garske, E.E., Gibb, J.P.
1984 'A Laboratory Evaluation of Ground Water Sampling Mechanisms' Ground Water Monitoring Review, V. 4, No. 2, pp. 32-41
Barker, J.F., Dickhout, R.
1988 'An Evaluation of Some Systems for Sampling Gas-Charged Ground Water Water Monitoring Wells' Ground Water Monitoring Review, V.8, No. 4, pp
Devlin, J.F.
1986 'Evaluation of Some Materials and Pumping Systems Available for Use in Sampling Ground Water Contaminated with Volatile Organics' Proceedings of the Third Annual Eastern Regional Ground Water Conference, July 28-20, Springfield, IL pp. 503-526
Gillham, R.W., O'Hannesin, S.F.
1988 'Sorption of Aromatic Hydrocarbons by Material Used in the Construction of Groundwater Sampling Wells' ASTM Symposium: Standard development for groundwater and vadose zone monitoring investigation., Jan 27-29, Albuquerque, NM
Nielsen, D.M., Yeates, G.L.
1985 'A Comparison of Sampling Mechanisms Available for Small-Diameter Ground Water Monitoring Wells' Ground Water Monitoring Review, V. 5, No. 2, pp.83-99
Reynolds, G.W., Gillham, R.W.
1985 'Adsorption of Halogenated Organic Compounds by Polymer Materials Commonly used in Ground Water Monitors' Presented at the Second Annual Canadian/American Conference on Hydrogeology, June 25-29, Banff, Alberta, 19 pp.
Rannie, E.H., Nadon, R.L.
1985 'An Inexpensive, Multi Use, Dedicated Pump for Ground Water Monitoring Wells' Ground Water Monitoring Review, V. 8, No. 4, pp.. 100-107
Schalla, R., Myers, D.A., Simmons, M.A., Thomas, J.M., Toste A.P.
1988 'The Sensitivity of Four Monitoring Well Sampling Systems to Low Concentrations of Three Volatile Organics' Ground Water Monitoring Review, V.9, No. 3., pp. 83-99
Thomey, N., Ogle, R., Jackson, J.
1991 'A Comparison of Results for Samples Collected with Bailers Constructed of Different Materials' Proceedings of the 1991 Outdoor Action Conference, May 13-16, Las Vegas, NV, pp. 577-582
Customer comments
No comments were found for Groundwater Monitoring with the Waterra Inertial Pump. Be the first to comment!