Periodic groundwater sampling is often required as part of a long-term monitoring program. Traditional sampling and analytical techniques require shipping multiple liters of water to fixed laboratories that perform regulatory-approved analytical methods. The typical analysis and data reporting time can be up to 45 days, which delays vital information on contaminant concentrations being reported to the customer. Additionally, most sample holding times have been tested for a small set of environmental matrices where the assumption has been made that analyte concentrations will not change significantly if analyzed within this window, typically 7 to 40 days. The use of a field-portable gas chromatograph-mass spectrometer (GC-MS) alleviates these concerns.

The objectives of this ESTCP demonstration were to (1) demonstrate the suitability of field analysis for a suite of contaminants of concern (semivolatile munitions constituents) and (2) demonstrate the utility, comparability, and cost savings of groundwater analysis using the Griffin 450 GC-MS. This effort was designed to demonstrate the advantages and limitations of field-portable analytical instrumentation for the detection and quantification of munitions constituents in groundwater, which eliminates the need to ship water samples overnight, under chain of custody, to a fixed analytical laboratory. Specifically, this project tested the Griffin 450 GC-MS and compared the in-field results to traditional munitions constituent analysis using laboratory-based high performance liquid chromatography (HPLC) with ultraviolet absorbance detection following U.S. Environmental Protection Agency (USEPA) Method 8330.

Mass spectrometry (MS) analysis systems can provide valuable chemical information on almost any type of sample. Traditionally, MS has been confined to fixed-site laboratory analysis due to the size and fragility of the instruments typically employed for this application. Griffin has made efforts toward miniaturization, enabling this technology to be brought to the field to perform analysis. The Griffin instruments use a cylindrical ion trap (CIT) as the mass analyzer; this device is a simplified geometry of the classic hyperbolic ion trap and therefore more easily miniaturized. The Griffin instruments also use a low thermal mass (LTM) gas chromatograph (GC) as the GC. With a smaller ion trap, the vacuum manifold becomes smaller, and the resulting pumping and power requirements are reduced. The LTM GC column eliminates the need for a convective oven, greatly reducing the size and power consumption compared to standard GC systems. These modifications to the instrument design all serve to decrease the size and weight of the instrument. Griffin has also worked to ruggedize the instrument, enabling transport into the field for on-site analysis. The improved electronic stability and sensitivity of the Griffin 450 provided higher quality data, especially in humid environments, compared to the previous Griffin 400 model GC-MS.

The instrumentation was tested on 28 groundwater samples from two distinct field sites for a variety of analytes with concentrations ranging over three orders of magnitude. The compounds evaluated were nitrobenzene (NB), 1,3-dinitrobenzene (1,3-DNB), 2,4-dinitrotoluene (2,4-DNT), 1,3,5-trinitrobenzene (TNB), 2,4,6-trinitrotoluene (TNT), and hexahydro-1,3,5-trinitro-1,3,5-triazine (RDX). Split groundwater samples were collected and analyzed for these compounds to compare the results from a field-portable GC/MS method to the results from a conventional fixed laboratory method. Detection limits for the field-portable instrumentation are sufficient to meet regulatory threshold levels, generally around 0.002 mg/L. Linear regression comparison of the in-field results to traditional laboratory-based analysis suggests comparability between the techniques, with the slope of the regression for all analytes being between 0.8 and 1.2, except for TNB and RDX. However, the slope of the regression for RDX is between 0.8 and 1.2 for all concentrations below 10 mg/L.

As all of the paired results for NB were non-detects, only a limited evaluation was possible. However, the NB results were consistent in that both the field and laboratory methods reported non-detects for NB for all split sample analyses. The field method for RDX possessed a negative bias relative to the fixed laboratory method and exhibited relatively large variability across all concentration ranges evaluated. The field results were about 70% of the laboratory results on the average. Therefore, it is recommended that the field method be used to obtain only screening-level data for RDX. The field and laboratory results were essentially equivalent for concentrations less than or equal to 0.3 and 0.2 mg/L for 1,3-DNB and 2,4-DNT, respectively. The comparison was limited by the relatively small data set owing to several non-detects and the relatively small concentration range evaluated (about 0.01 - 0.1 mg/L). Results for TNT were reliable for screening only below a concentration of 0.05 mg/L; however, between 0.05 and 10 mg/L, results from the field and laboratory were equivalent. The field method consistently exhibits a significant positive bias for TNB (F=1.5 L). There was a strong correlation between the laboratory and field methods for concentrations greater than about 0.05 mg/L to the highest reported concentration, but the performance of the field method was relatively poor at smaller concentrations. The TNB field results >0.05 mg/L would need to be adjusted for bias prior to being reported.

The results indicate that similar reporting limits can be obtained using the field-portable instrument when coupled to solid phase extraction (SPE) sample preparation, although instrument stability at the low concentration range can be an issue. Furthermore, the linear dynamic range is somewhat limited, as compared to HPLC analysis, for samples with high analyte concentrations. The cost savings of the field method were found to be $29,600 a year, based on 12 week-long field trips, with a breakeven point of 3.54 years.

Field-portable GC-MS appears at this point to be suitable only for screening RDX, due to significant scatter in the comparison to laboratory results across the concentration range tested. The regression line data demonstrate that the slope is within the 0.8 to 1.2 limit except for TNB and RDX. The TNB data is skewed somewhat by two samples with high concentrations. A similar effect is observed for RDX with one high concentration sample skewing the results. These samples reflect the linear dynamic range limitations of the current instrument when large sample pre-concentration factors result from the SPE procedure. Additionally, deployment of the technology requires skilled labor at this point. Deployment of the technology to field sites is feasible for any site that has sufficient space and access for traditional groundwater collection activities.