Increasing Speed and Decreasing Cost of PCB Analysis of Alternative Fuels and Raw Materials
David Gossman, Gossman Consulting, Inc.
Environmental Regulations in many countries control the use of waste-containing PCBs. It may therefore be necessary to test incoming alternative raw materials and/or fuels for PCBs based on regulatory requirements or a desire to limit
corporate liabilities. The concentration for which this control is needed may vary considerably state to state. In many areas of the world including the US the critical threshold is 50 ppm but it can be as low as 2 ppm and as high as
500 ppm depending on the specific circumstances. In any case it is desirable to have testing methods that can perform this testing in a minimal amount of time (generally less than 30-60 minutes) and at a minimal cost. There are three
steps to achieve this goal. First, implement a process that evaluates clearly the data quality objective (DQO) and thus limits or eliminates unnecessary testing. Second, develop a sample preparation process that targets the specific
sample matrix, and minimizes sample preparation time. Finally, develop analytical instrumentation specifications and operating conditions that meet the DQO while minimizing the analytical time required.
The DQO Process
The DQO process is a seven-step process summarized at http://gcisolutions.com/AWMADG99.htm. This process needs to be used to help set clear data quality objectives based on a variety of factors – all outlined in the process. By
performing this process first, the laboratory can be assured that the data that is produced will meet the end use needs and that more time and cost than is needed will not be invested. It is often the case that laboratories perform
testing blind to the final use of the data. For example, in most PCB testing of alternative raw materials and fuels all that is required is a threshold determination – is the concentration above or below a specific value?
Generally this type of determination does not require the multipoint calibration curves that may be present in standard methods. Instead the sample preparation, and method calibration needs to focus on meeting the need to quantitate at
the threshold level. Concentrations below the threshold level are of little concern and concentrations above the threshold level need only be noted as such.
Sample preparation for PCB analysis consists of three basic steps for most alternative fuels and raw materials. Sample extraction, the first step, may not be necessary for some samples such as waste oils or solvents. Solid samples will
require extraction with a solvent – generally the same solvent that is used for sample dilution, the second step. The precise nature of the sample extraction and the time required will depend on the specific sample matrix. Spiked
samples should occasionally be used to validate this step in the methodology.
Sample dilution is performed in a suitable nonpolar solvent. The most common solvent seen in standard methods is n-hexane. I strongly recommend against using n-hexane – it is both more toxic and more volatile than other suitable
solvents. Isooctane is preferred, reducing error due to solvent volatilization and potential exposure issues for laboratory personnel. The amount of sample dilution required will depend largely on the quantitation limit needed. For
samples needing a low quantitation limit, such as 1-2 ppm, a ten-fold dilution is recommended. Higher dilutions are recommended when higher quantitation limits will meet the DQOs. It is important to note that the dilution solvent can be
prespiked with internal standards. Said internal standards are not used to quantify the results but rather to verify that the third step in sample preparation process, sample cleanup, does not result in the loss of the analytes of
interest. I recommend that the dilution solvent contain 10-100 ppb of 2,4,5,6-tetrachloro-m-xylene and decachlorobiphenyl. These internal standards neatly bracket the chromatographic range over which various PCB Aroclors elute.
Once the sample has been extracted and diluted in the appropriate solvent, which has been prespiked with internal standards, it is now necessary to perform a chemical cleanup of the sample. This sample cleanup can vary considerably
depending on the degree of interferences present in the original sample and the proportion of them that extract into the solvent used for extraction. In order to minimize sample preparation time and cost it is generally necessary to
investigate a variety of sample cleanup options to determine the optimal threshold for cleaning up the sample but not going into overkill. For a typical waste oil sample there are two recommended sample cleanups that can be performed
fairly quickly. The first is the use of concentrated sulfuric acid to wash the diluted sample and chemically destroy interfering compounds such as oxygen containing compounds like esters and phthalates. Depending on the level of cleanup
needed this step can be repeated a number of times. An additional step that is often performed is to clean the sample with florosil. Most often this is done by passing the diluted sample through a preprepared florosil column. Another
faster method that can achieve the same result is to simply add a small amount of florosil to the diluted sample and agitate it. Interfering compounds are adsorbed onto the florosil.
Taking this series of steps into account and using waste oil needing a 2 ppm quantitation limit as an example, the following would be a typical resulting operating procedure:
1. Dispense 4.5 ml of solvent that has been prespiked with internal standards into a test tube.
2. Pipet 0.5 ml of waste oil into the solvent. Agitate the sample of oil just prior to removing the sample and then mix the diluted sample using a vortexer.
3. Add 2 ml of concentrated sulfuric acid to the test tube in 0.5 ml increments. With each incremental addition mix with the vortexer allowing any reaction to diminish prior to adding additional acid. Make certain that
the test tube is pointed toward the back of the fume hood during this procedure in case of a more violent reaction.
4. Centrifuge the sample to separate. If the organic (top) layer is reasonably clear go on to step 5, otherwise decant the organic layer to a new test tube and repeat steps 3 and 4 as necessary.
5. Decant most of the organic layer to a clean test tube and add ~1 g of florosil. Agitate the sample for at least 15 seconds with the vortexer. Allow the sample to separate and withdraw ~1 ml of solvent to be placed
in the gas chromatograph autosampler vial.
This entire procedure can be accomplished by a practiced technician within a matter of a just a few minutes.
Instrumentation and Analytical Conditions
There are numerous columns that can be used to efficiently separate and quantify PCB Aroclors. Many years ago the standard was a packed glass column with a mixed phase support. Today I prefer a wide bore capillary column with a thick
nonpolar phase. The exact configuration that will be chosen by the analyst may depend on available instrumentation and other potential uses such as chlorinated herbicide analysis. Electron capture detectors are generally used for PCB
determination although other detectors have been successfully used including mass selective detectors. Further, to minimize analytical times and maintain appropriate peak separation the use of aggressive gas flow levels are sometimes
needed. Modern instruments with electronic flow control can aid significantly in maintaining high gas flow through the column during the temperature-programmed run. With sample turn-around time of critical importance and the need to
push the quantitation limit down to <200 ppb in the diluted sample the following parameters would provide a good starting point for developing a standard operating procedure:
· Column – SPB-608, 30 m x 0.53mm ID, 1.0 µm film
· Carrier – helium ~10-12 ml/min
· Detector – ECD, 320°C, nitrogen makeup gas
· Injection – 250°C, splitless, 2.0 µl
· Oven – 120°C (2 min) to 300°C (15°C/min) hold 300°C (5 min)
If interferences were minimal a thinner coating of 0.5 µm film would speed the analysis and enhance the resolution a small amount. All of these parameters should be considered a starting point for optimizing an operating
Standards of 200 ppb of various Aroclors should be maintained in a computerized database for quick overlay and comparison with samples. Aroclors typically used are 1016, 1221, 1232, 1242, 1248, 1254, 1260 and 1262. A mixture of two of
these Aroclors, such as 1232 and 1260 can be used for routine checks on the retention time calibrations. Separately prepared check standards can be used to verify that the sensitivity of the ECD has not changed over time.
By using a system of data quality objectives and then developing sample preparation and instrumentation analysis procedures to optimize the determination of PCBs in alternative raw materials and/or fuels, a system can be put into place
that allows an individual sample to be analyzed within 30 minutes of receipt with a high degree of confidence that the results meet a predetermined threshold determination.