U.S. Patents No. 6,481,300; No. 6,837,120; others pending
|
|
|
|
Report Cover Table of Contents Sec. 1 Sec. 2 Sec. 3 Sec. 4 Sec. 5 Sec. 6 Sec. 7 Sec. 8 SECTION 6 DISCUSSION As described in Sections 1.1 and 2.3, one of the objectives of this demonstration was to include one or more “baseline” sampling methods to provide data against which the results of the alternative diffusion and grab samplers can be compared. This was achieved by incorporating the conventional methods (three-volume purge and low-flow purge) into the demonstration. However, these methods utilize inherently different sampling concepts than both the diffusion and grab sampling methods. Most notably, the conventional sampling methods induce groundwater flow into the well by creating an increased hydraulic gradient around the well resulting from the purge. Conversely, the alternative sampling methods rely solely on the natural flow of groundwater through the well. These methods might be expected to provide differing results since they are monitoring different flow conditions and potentially also different volumes of the aquifer. Furthermore, even results from the two conventional methods are expected to vary given the differences in purge volume and rate and the fact that low-flow samples are considered by some to be representative of a more discrete sample interval than samples obtained using a three-volume purge. Accordingly, although they represent “baseline” data in the sense that they are the commonly-used sampling methods that are generally accepted by the regulatory community, they do not necessarily represent the correct answer (only a different answer). Because there are many different ways to evaluate a data set as large and robust as the one collected during this demonstration, it is difficult to derive sweeping conclusions about the relative performance of one device compared to all the others. If all methods were measuring the same thing, comparison of the performance of one method to another would be more straightforward. However, in this demonstration, the purge and no-purge sampling methods actually might have measured different things (as described above). Accordingly, the performance of one sampling method relative to another is more difficult to quantify. Nonetheless, the sampling results were compared as described in Section 4. Sampling method- and analyte-specific conclusions and observations are summarized in the following subsections. These conclusions and observations were derived entirely from the ‘holistic conclusions’ presented in Tables 4.2 through 4.6. The holistic conclusions are necessarily ‘broad-brush’ and generalized and were assigned varying degrees of confidence depending on whether all of the quantitative comparisons performed resulted in the same observation. For example, the holistic conclusion that low-flow anion concentrations are less than Snap Sampler™ anion concentrations does not mean that this is always the case. The results of all four comparative tests did not consistently indicate this conclusion; however, the weight of evidence indicated that this was true more often than not. The reader is encouraged to study the more detailed information presented in Section 4 (Tables 4.2 through 4.6) prior to making final decisions on use of the various samplers tested at McClellan. The summary data presented in Table 4.1 may be misleading when compared with the results for the individual analytes or analyte groups presented in Tables 4.2 through 4.6, and should not be used to evaluate a particular sampling method’s utility for a specific analyte or analyte group. It should be noted that sampling results were quantitatively compared using groupings of analytes rather than specific analytes (e.g., all metals rather than individual metals such as aluminum and zinc, and all VOCs rather than specific VOCs such as TCE). Pooling data for a multitude of analytes provides a general basis for comparison, but the comparison results may not be representative of how each of the individual analytes compared. For example, a more extensive study is being performed by URS that is comparing individual and pooled metals results obtained from more than 250 McClellan wells using both low-flow and three-volume purge methods, and the results of this comparison obtained to date do not agree with the results presented in this report (source: written communication from J. Rogalla [URS]). Therefore, the comparison results presented in this report may not definitively determine comparability among the different sampling methods. 6.1 SUMMARY OF RESULTS BY SAMPLING METHOD 6.1.1 LOW-FLOW PURGE Generally, results of the low-flow purge are equal to or lower in concentration than corresponding results from most of the other sampling methods. The only notable exception to this observation is with metals (not including hexavalent chromium), where the low-flow purge method typically produced higher concentrations than all of the other sampling methods (Table 4.5). Low-flow results for hexavalent chromium tended to be lower than results obtained using other methods (Table 4.4). Although it is not entirely evident why these trends occurred, the following explanations are proposed. The three-volume purge samples were collected from a bailer after the purge was complete, while the low-flow samples were collected directly from the pump discharge. As shown in Table 3.3, the final water temperature in the low-flow purge sample was usually higher than for the three-volume purge samples. The temperature differences ranged from 0.01 to 3.8 degrees Celsius (ºC) with a mean value of 1.7 ºC . This may have resulted from the heat generated by the pump motor and impeller, and could at least partially explain why the VOC concentrations in the low-flow purge samples are frequently lower than the concentrations of the same analytes derived using other sampling methods (i.e., a higher water temperature could result in a higher volatilization rate and correspondingly lower concentrations of VOCs in the sample). In this particular case study, the final turbidity in the low-flow purge water was generally higher than it was in the three-volume purge water (Table 3.3). This indicates that more particulates were present in the low-flow water than in the three-volume water, and could explain why metals concentrations were usually higher in the low-flow purge samples. While the reasons for this are unclear, it may be due at least in part to the fact that the low-flow purge was performed prior to the three-volume purge and shortly after the sampling pump had been introduced into the well (disturbing the water column and potentially increasing turbidity levels in the well). However, in almost every instance conventional and grab samples collected for metals analysis were field filtered with a new 0.45-micron filter to remove particulates. One hypothesis is that elevated quantities of colloidal metals were present in the low-flow samples that passed through the filters. The use of dedicated pumps, rather than newly decontaminated submersible equipment, may result in lower turbidity results during purging, possibly eliminating the need for filtration to remove particulates. From a cost perspective, the low-flow purge method was the second most expensive method demonstrated (Section 5). 6.1.2 THREE-VOLUME PURGE From a “performance” perspective, concentrations in samples collected using the three-volume purge technique were generally equal to or greater than corresponding concentrations in other sampling devices. Four exceptions to this trend were noted, including VOC concentrations in the Snap Sampler™, metals concentrations in the lowflow samples, and hexavalent chromium concentrations in the Hydrasleeve® and RPPS samples. In each of these instances, the three-volume purge concentrations tended to be lower. Overall, of the two conventional sampling methods demonstrated, the threevolume purge method produced results that were the most similar to the results for the diffusion and grab sampling devices. Based on the cost analysis (Section 5), the three-volume purge method was the most expensive method demonstrated. It should be noted that if the cost analysis had assumed use of dedicated pumps for the three-volume purge (similar to what was assumed for the low-flow purge method), the estimated costs of the three-volume and low-flow purge methods would have been more similar. 6.1.3 HYDRASLEEVE® For VOC concentrations, the HydraSleeve® was most comparable to the three-volume purge and the PDBS. Samples obtained using this device usually had higher concentrations of VOCs relative to the low-flow purge, PsMS, and RPPS methods. For metals, it was comparable to the three-volume purge, PsMS, and RPPS. HydraSleeve® samples typically contained higher concentrations of metals than the RCS and lower concentrations of metals than the low-flow purge samples. For anions, the HydraSleeve® was comparable to all other sampling methods against which it was compared. For hexavalent chromium, the HydraSleeve® was most comparable to the PsMS and RPPS, and was greater than both of the conventional methods and the RCS. For 1,4 dioxane, the HydraSleeve® was most comparable to both conventional methods, and was greater than both the PsMS and the RPPS. The conclusions involving hexavalent chromium and 1,4 dioxane are tentative due to the limited number of comparisons and resulting low statistical power of the tests performed. The HydraSleeve® and Snap Sampler™ were not tested in the same wells; therefore, analytical results for these two samplers were not compared with each other. The HydraSleeve® was the least expensive method demonstrated according to the cost analysis (Section 5). 6.1.4 SNAP SAMPLER™ For the majority of comparisons, the concentrations in the Snap Sampler™ samples were higher than corresponding concentrations in samples from all other sampling methods. This was true for all comparisons involving VOCs, and for all comparisons of anions and 1,4 dioxane except for the three-volume purge samples, which were roughly comparable to the Snap Sampler™ concentrations for those constituents. It should be noted that the 1,4 dioxane comparisons included few data points. The Snap Sampler™ was not used to sample for metals or hexavalent chromium. The observed ‘high bias’ in the Snap Sampler™ concentrations suggests that they may be more representative of the actual concentrations in the well at the time of sample collection, particularly for VOCs as described below. The fact that water for VOC analysis does not have to be transferred from the Snap Sampler™ into separate sample containers appears to be the most reasonable explanation for the relatively higher VOC concentrations obtained using this method. The lack of sample transfer eliminates the potential for VOC loss as a result of sample transfer. The developer of this sampler reports that results of other tests also exhibit the same higherconcentration trends for VOCs as seen in this study (Britt et al., 2005). For anions and 1,4 dioxane, it is not clear why the Snap Sampler™ concentrations were typically higher than those in samples collected using other methods. It should be noted that the relatively high VOC concentrations in the Snap Sampler™ may also be due, at least in part, to differences in how these samples were treated at the laboratory. As stated in Section 3.3, most of the VOC samples submitted to the analytical laboratory in 20-ml vials (i.e., samples collected using the PDBS, RPPS, RCS, PsMS, and HydraSleeve®) were composited at the laboratory into one 40-ml vial for analysis. Therefore, most of the VOC samples collected using no-purge techniques, except for the Snap™ samples, underwent two episodes of sample transfer (one in the field and one at the laboratory). These transfers may have resulted in some VOC loss and increased the “noise” or variability in these no-purge VOC data sets. The VOC data indicate that minimizing VOC sample transfer can result in more accurate detection of VOC concentrations present in the well water. The data also indicate that caution is advised when scoping the use of 20-ml VOA vials for VOC sample collection. The ability of the laboratory to analyze VOC samples contained in 20-ml vials without sample transfer should be confirmed, and use of 40-ml vials wherever possible is recommended. The Snap Sampler™ was more expensive than the other no-purge sampling methods based on the cost analysis described in Section 5, but as described above, it produced the most conservative sample results from a protectiveness standpoint. The cost per sample could be reduced by approximately $16 if two sample vials per sample are used instead of three, as was used at McClellan AFB and assumed in the cost analysis. It should be noted that the volume of water that can be collected using this device is relatively small compared to most of the other methods (Table 3.5). The vendor has developed a 125-ml sample bottle to accommodate somewhat larger sample volume needs. 6.1.5 PDBS The PDBS was only used to monitor VOCs. This diffusion sampler tended to return higher concentrations of VOCs than the low-flow method, the RCS, and the RPPS. It was most comparable to the three-volume purge, PsMS, and HydraSleeve® methods, and typically returned lower VOC concentrations than the Snap Sampler™. The PDBS was the second least expensive of the non-conventional samplers and was the least expensive diffusion sampler evaluated, according to the cost analysis described in Section 5. It has been shown in several other studies (e.g., Parsons, 2003b and 2004b) to be a reliable and inexpensive method of monitoring for most commonly-occurring VOCs in groundwater. 6.1.6 RPPS For VOCs and 1,4 dioxane, the results obtained using the RPPS were generally comparable (i.e., similar concentrations) to those obtained using the low-flow purge, PsMS, and the RCS methods. Conversely, the RPPS results for these analytes were usually less than (i.e., lower concentrations) obtained with the three-volume purge, HydraSleeve®, PDBS, and Snap Sampler™ methods. For anions, the RPPS was generally comparable with all other methods except the Snap Sampler™, which typically yielded higher concentrations. The RPPS results for hexavalent chromium tended to be higher than the low-flow purge, three-volume purge, and RCS results, but were similar to the HydraSleeve® and PsMS results for that analyte. For metals, the RPPS results were most comparable to the three-volume purge, HydraSleeve®, and PsMS; usually less than results obtained with the low-flow purge; and greater than results obtained with the RCS. In summary, the RPPS seems to have performed well at monitoring for anions, metals, and hexavalent chromium, but not as well at monitoring for VOCs and 1,4 dioxane. It is perhaps noteworthy that the RPPS appears to have performed best for inorganic, nonvolatile parameters, and less well for organic parameters (i.e., VOCs and 1,4 dioxane). Considering the relatively large pore size of the membrane used for this sampler relative to some of the other devices (Section 2.1.2), there may be a higher potential for volatilization when using this sampling device. The RPPS was the most expensive diffusion-based sampler according to the cost analysis (Section 5). 6.1.7 RCS Based on the statistical analysis results presented in Section 4, the RCS had lower concentrations of metals (not including hexavalent chromium) than all other methods. In contrast, this sampler returned hexavalent chromium concentrations that were similar to or higher than obtained using conventional methods. With the exception of the Snap Sampler™, it was generally comparable to all other methods for anions. The RCS also usually had lower concentrations of VOCs than the other methods except for the RPPS and low-flow purge methods, which produced results that were more similar to the RCS results. Based on limited data, the RCS performed acceptably for 1,4 dioxane. In general, the demonstration results indicate that this sampling device was in a lower bracket in terms of its performance at McClellan AFB. It usually produced concentrations that are comparable to or lower than the other devices. It was relatively inexpensive according to the cost analysis, being the third least expensive device evaluated (Section 5). In addition, the cellulose membrane of the RCS was occasionally observed to become brittle or was more easily torn upon retrieval of the sampler (see also Section 2.1.4). 6.1.8 PsMS The PsMS results for anions are comparable to anions results obtained using all other methods. Although the data set for 1,4 dioxane is small, the comparative analyses performed indicated that the PsMS is generally comparable to both the conventional methods and the RPPS, and usually produced lower concentrations than the HydraSleeve® and the Snap Sampler™. For hexavalent chromium and metals, the PsMS was comparable to all methods except low-flow purge, which generally returned lowermagnitude concentrations than the PsMS. For VOCs, the PsMS was comparable to the other diffusion samplers against which it was compared (PDBS and RPPS) and the lowflow purge, but returned lower-magnitude concentrations than the HydraSleeve® and the three-volume purge. The PsMS was not installed in the same wells as the RCS and Snap Sampler™ except for well MW-453, which was sampled using both the PsMS and Snap Sampler™. Therefore, the PsMS was not compared to these other devices. This method was relatively expensive compared to the most of the other diffusionbased devices, but it still was significantly less expensive than the conventional methods. 6.2 SUMMARY OF RESULTS BY ANALYTE 6.2.1 1,4-DIOXANE As indicated in Table 4.2, a low degree of confidence is assigned to the statistical results for this analyte due to the relatively small number of comparisons that could be made (ranging from 3 to 9). Therefore, the following conclusions and observations regarding 1,4 dioxane are tentative.
In summary, if conventional (i.e., low-flow and three-volume purge) results are used as a baseline for comparison purposes, then the HydraSleeve®, Snap Sampler™, PsMS, and RCS appear to produce results that are similar to or higher than the conventional results, suggesting that they could be substituted for the conventional methods in at least some situations. Although similar to low-flow results, the RPPS results tended to be biased low relative to three-volume results; therefore, this sampler can not be wholeheartedly endorsed for use with 1,4 dioxane on the basis of the McClellan results. 6.2.2 ANIONS The holistic conclusions for anions summarized in Table 4.3 indicate a relatively high degree of parity among the samplers in terms of anion results. Fourteen of the 17 sampler-pair comparisons performed for anions (82 percent) generally yielded comparable results (indicated by an ‘equal’ sign in the ‘Holistic Conclusion’ column). In contrast, only 56 percent of 1,4 dioxane sampler-pair comparisons (Table 4.2), 36 percent of hexavalent chromium comparisons (Table 4.4), 43 percent of metals comparisons (Table 4.5), and 38 percent of VOC comparisons (Table 4.6) yielded equivalent results. The three sampler-pair comparisons for anions that did not yield equivalent results involved the Snap Sampler™, which tended to yield anion concentrations that were relatively elevated. In summary, if conventional (i.e., low-flow and three-volume purge) results are used as a baseline for comparison purposes, then all of the diffusion and grab samplers tested for anions (i.e., HydraSleeve®, Snap Sampler™, PsMS, RCS, and RPPS) appear to produce results that are either similar to or higher than the conventional results, suggesting that they could be substituted for the conventional methods in at least some situations. 6.2.3 HEXAVALENT CHROMIUM Similar to 1,4 dioxane, many of the holistic conclusions for hexavalent chromium listed in Table 4.4 are assigned a relatively low degree of confidence because they are based on a small number of comparisons. Therefore, the following conclusions and observations regarding hexavalent chromium are tentative.
6.2.4 METALS The holistic conclusions for TAL metals summarized in Table 4.5 indicate the following general conclusions and observations:
In summary, the McClellan data suggest that the low-flow results for TAL metals are anonymously high relative to results obtained using all of the other tested methods, and therefore they may not be a good indicator of ‘baseline’ results for comparison purposes. If the three-volume purge results are used as a baseline for comparison purposes, then the HydraSleeve®, PsMS, and RPPS appear to produce results that are similar to or higher than the baseline conventional results, suggesting that they could be substituted for the conventional methods in at least some situations. Of the three diffusion samplers tested, the RCS would rank the lowest in terms of monitoring for TAL metals based on the McClellan data. Comparison of metals results for filtered vs. unfiltered samples does not indicate a trend of low bias in the filtered samples relative to the unfiltered samples. 6.2.5 VOCs The holistic conclusions for VOCs summarized in Table 4.6 indicate the following general conclusions and observations:
In summary, if the conventional (i.e., low-flow and three-volume purge) results are used as a baseline for comparison purposes, then the Snap Sampler™, HydraSleeve®, and PDBS appear to produce results that are similar to or higher than the baseline conventional results, suggesting that they could be substituted for the conventional methods in at least some situations. Although similar to low-flow results, the RPPS results tended to be biased low relative to three-volume, HydraSleeve®, Snap Sampler™, and PDBS results. Similarly, the PsMS results tended to be biased low relative to the 3- volume and HydraSleeve® results, and the RCS tended to be biased low relative to the 3- volume, PDBS, and Snap Sampler™ results; therefore, these three diffusion samplers can not be confidently endorsed for use with VOCs on the basis of the McClellan results. Report Cover Table of Contents Sec. 1 Sec. 2 Sec. 3 Sec. 4 Sec. 5 Sec. 6 Sec. 7 Sec. 8
|
