Independent Consultants in Environmental and Forensic Chemistry

Volume 3, Issue 3, Summer 1999

Are Contaminants in Groundwater Really Retarded or Is It Just Us?

The commentary by Ernesto Baca (AOn the Misuse of the Simplest Transport Model.@ Ground Water, Vol. 37, No. 4, p. 483, July-August 1999) reminds environmental professionals that age-dating chemical releases by dividing a contaminant=s plume length by its average velocity in groundwater is wrong and not technically supportable. The reason for this lies in understanding contaminant plumes and the concept of average velocity.

Contaminants dissolved in groundwater travel as fast as the groundwater but are retarded by adsorption/desorption on aquifer materials. The effect of the adsorption/desorption phenomenon is to spread the concentration of the contaminant over the distance of groundwater flow forming a bell-shaped curve. The average velocity is determined by dividing the distance from the source to the apex of the curve by the time required to traverse the distance.

The curve can be sharp or broad depending on the aquifer=s average effective porosity and/or the degree of retardation. Since an inorganic ion is assumed to be minimally affected by the adsorption/desorption phenomenon, most retardation experiments use an inorganic substance, such as bromide a ion, to trace and measure the average groundwater velocity. Organic contaminants are more susceptible to the adsorption/desorption phenomenon, and, as a result, the average flow velocity of an organic contaminant is less than the average groundwater velocity, and the contaminant=s flow seems retarded.

If the apex of a bromide curve is taken as the average groundwater velocity, half of the groundwater is moving faster and half is moving slower than the average. The portion of the plume moving faster than the average forms the leading edge of the plume. At a downgradient location, the leading edge of the bromide curve will be detected before the apex.

As a result of retardation, an organic contaminant=s curve will be broader than the groundwater (or bromide) curve. While the average velocity (the apex of the curve) of the organic compound may be slower than the average velocity of an inorganic ion, the leading edge of the organic plume will still keep pace with the leading edge of the inorganic plume. If detection of bromide is limited by a one ppm detection limit and an organic contaminant can be detected at 0.1-0.2 ppb (3 to 4 orders of magnitude less than for bromide), the organic compound will be detectable at a downgradient location before the bromide ion, and the organic contaminant will appear to move faster than the groundwater. Meanwhile, the apex of the organic curve will still lag behind the apex of the inorganic curve, and a retardation factor (average groundwater velocity divided by average contaminant velocity) will be calculated.

In a real life situation, the length of an organic plume is determined by the farthest downgradient observation of the compound. If this length is divided by the average velocity of the contaminant, the calculated time of groundwater transport would be wrong, as noted by Baca. The farthest downgradient observation would be the leading edge of the plume, not the apex of the contaminant curve. To obtain a more accurate estimate of the time of groundwater transport, the distance to the apex of the contaminant curve, not the leading edge, must be divided by the average contaminant velocity.

What=s in a Data Package?

During the course of planning a project, the conversation should turn to the laboratory deliverables requirements. What kind of analytical data package will be required for the project? The laboratory can provide several types to meet various needs, but all data packages are not created equal.

The contents of the data package will differ depending on the type of data package requested. An EPA Contract Laboratory Program (CLP) or CLP-type data package provides the most complete documentation and is sometimes referred to as a Level IV package. All raw data applicable to the samples analyzed are included. The raw data include chromatograms and quantitation reports for tuning (for GC/MS analyses), calibration standards, field samples, and quality control (QC) samples. The QC samples include method/preparation blanks, spikes (matrix and/or blank), duplicate analyses, and laboratory control samples. Mass spectra are also included, if applicable. Various forms which summarize the results of QC samples, calibration data, and reported sample results are also included.

A Level IV data package contains sufficient information to allow a detailed validation of the data.

The Reduced Data Deliverables package provides significantly less instrument printout data and is sometimes referred to as a Level III package. For organics analyzed pursuant to CLP methods, all laboratory deliverables required for Level III are the same as for the CLP except that chromatograms of calibration standards and chromatograms and spectra for matrix spikes and matrix spike duplicates are not required. Deliverables for inorganics consist of the reporting forms as specified in the Statement of Work (SOW). Reduced Data Deliverables for non-CLP methods consist of completed summary forms only. Since much of the raw data are not included, a detailed data validation cannot be performed on a Level III package; only an evaluation can be performed.

The least complete package is a results only package. In this case, no supporting data are provided to verify the sample results. Complete reliance is placed on the laboratory to have accurately and correctly performed the preparation procedures, instrumental calibration, sample analyses, data reduction, and transcription of the results for the samples. Since no raw data are available, a data validator cannot verify the results in a results only data package.

Regardless of the level of the supporting data documentation, a copy of the chain of custody record must be included in every data package.

Preparation and analysis procedures are the same for a results only package as for a Level IV package. The same raw data are generated during the analyses. Only the level of documentation included in the data package submitted to the client differs. While any data package (with the exception of a Level IV package) should be able to be upgraded at any time, the costs to upgrade a data package can be expected to be significantly higher at a later date. It is easier for the laboratory to produce a substantive data package at the time of analyses than to retrieve and compile the data later. Depending on the laboratory and the elapsed time since the analyses, costs to upgrade a data package after the fact may approach the cost of the original analyses. Also, it may not be until you request the data package upgrade that you find that the laboratory did not conduct the appropriate calibrations, quality control, and/or calculations.

The ability to retrieve any supporting data at a later date is also contingent on the history and/or future of the laboratory. The collapse of many environmental laboratories in the early 1990s and the recent trend towards acquisition of small laboratories by large corporations makes the fate of laboratory records uncertain. As with buying any product with a lifetime guarantee, you must ask yourself about whose lifetime you are talking.

Supporting documentation is not required for sample results, but you should require it if you want or expect to use the results.

Is MTBE in your future?

Methyl tertiary butyl ether, MTBE, was produced as a gasoline additive to increase the gasoline octane rating and reduce carbon monoxide in automobile exhaust. Initially used in the late 1970s at a few locations, MTBE became common in many gasolines when allowable lead was reduced to a maximum of 0.1 g/gallon in 1985. By the end of the 1980s, almost all gasolines contained MTBE. The 1990 Clean Air Act Amendments for reformulated gasolines continued the trend in the 1990s.

The human toxicity of MTBE is not known; however, it is a reported carcinogen in animal studies. At approximately 20 ppb, it gives drinking water a smell and/or taste. MTBE is easily washed out of gasoline by water and released to the environment. It dissolves in groundwater in measurable concentrations faster than benzene, toluene, ethyl benzene, and the xylenes (BTEX) and appears to travel faster than the BTEX compounds in a groundwater plume. MTBE does not readily biodegrade in soil and groundwater. It does not readily adsorb on soil particles and is difficult to purge from water. It is also difficult to remove from water by carbon adsorption. Therefore, MTBE not only dissolves in groundwater rapidly but also tends to remain in groundwater.

MTBE seems to be everywhere. It is one of the most volatile gasoline components and, therefore, readily evaporates into air spaces. Exhaust from internal combustion engines in automobiles and boats contains MTBE. Exhaust from two cycle engines (outboard motors, jet skis, lawn mowers, etc.) have added significant quantities of MTBE to air and surface waters. Low ppb quantities of MTBE have been reported in precipitation runoff that is reaching shallow aquifers. MTBE has also been reported in home heating fuel oil and in jet fuel. The observation of MTBE in environmental samples has become so common that some states have initiated actions to ban the use of MTBE.

Some environmental forensic Aexperts@ have used the observations of MTBE in groundwater, even in trace quantities, for age-dating gasoline releases. No MTBE found means that the release was prior to 1979 or 1985 or 1989 or 1992, etc. No MTBE found means that the gasoline released was leaded. MTBE found means the release of gasoline was post 1979, etc. In and of themselves, these conclusions are not totally accurate. With all of the other potential sources for MTBE available, observations (or lack of observations) of MTBE in the environment must be carefully interpreted, particularly if they are present in trace quantities. Different refiners initiated the use of MTBE at different times and in different grades. Since MTBE dissolves and moves rapidly in the groundwater, residual gasoline at a source may not contain MTBE. The simplistic conclusions offered by many Aexperts@ regarding the age of a gasoline release based on the observation or lack of observation of MTBE can be totally erroneous and misleading.

What is a Spike and What Does it Do?

A spike is a sample into which a known quantity of an analyte, or analytes, is added. That sample is then processed in a manner identical to field samples. A spike is used to determine if the preparation and analytical systems are in control and if the laboratory has the capability of accurately measuring the amount of an analyte in a sample.

There are two types of spikes - matrix spikes and blank spikes. Blank spikes, also referred to as laboratory fortified blanks, are prepared by adding the analyte to a pristine matrix in the laboratory. This matrix must be similar to the field samples being analyzed - deionized water for aqueous field samples and a solid matrix for soil samples. Since the matrix is pristine, the laboratory should be able to accurately measure the amount of the analyte in the blank spike. A matrix spike is prepared by adding a known quantity of an analyte to a field sample and is processed in a manner identical to the field samples. The matrix spike measures the ability of the laboratory and the analytical method to accurately quantify the analyte in an actual field sample. For the matrix spike results to be of value to you, the spike should be conducted on a sample from your site.

The amount of the added spike should be in the concentration range of interest. Ideally, the amount of an analyte added to a matrix spike should be comparable to the amount already present in the field sample so as not to overwhelm or be overwhelmed by the original amount. No meaningful information can be obtained from a spike which is significantly greater or significantly less than the concentration of the analyte in the sample. If the spike overwhelms the original sample concentration, the spike will not give adequate information regarding the ability to measure the analyte at the original concentration. If the spike is overwhelmed by the original concentration, the spike will not be distinguishable from the original concentration. Unfortunately, in most analytical sequences, the concentration of the analyte in the field sample is unknown prior to the preparation of the samples and matrix spike. Therefore, it is not known prior to analysis if the concentration of the spike is comparable to the concentration of the target analyte in the field sample.

The result from a spiked sample is reported as a percent recovery which is the amount measured divided by the amount added and multiplied by 100. Since a sample used for a matrix spike may contain some of the analyte in addition to the spiked amount, this original amount must be subtracted from the measured amount prior to calculating the percent recovery. Matrix spikes analyzed in accordance with CLP and other EPA methods have assigned recovery acceptance criteria. Some analytical methods provide for the laboratory to establish their own criteria. Meeting method acceptance criteria may suggest that the laboratory is performing the analyses in a reasonable manner, but the accuracy of the analyses must still satisfy the needs of the project and may require more stringent criteria.

Many laboratories report poor matrix spike recoveries as being due to a matrix effect. However, other than the poor recoveries, seldom is any evidence offered to demonstrate a matrix effect as opposed to a breakdown in the preparation and/or analytical procedures. Since the blank spike matrix is pristine, matrix effects should not be a factor. If matrix spike results are out of criteria but blank spike results are acceptable, a matrix effect may be present. Poor spike recoveries alone do not indicate a matrix effect.

When planned, implemented, and used properly, spikes can provide significant information about the accuracy of the analytical systems.