Challenge:

Environmental samples need to be analyzed for polynuclear aromatic hydrocarbons (PNAs or PAHs) at very low levels with accurate identification.  We were interested in analyzing PNAs at low detection limits required by the Illinois EPA.  From past experience, we knew that there were several problems associated with the HPLC method (SW-846 Method 8310) and also with Selected Ion Monitoring (SIM).  Specifically, complex matrices make identification of specific compounds difficult if the only information available is retention time data or response from a single selected ion.  A single PNA compound is hard to distinguish in a sample that contains a large amount of petroleum hydrocarbons.  We were interested in the selectivity offered by the mass spectrometry method and yet needed the sensitivity offered by the traditional HPLC or SIM method. This analysis provides the necessary sensitivity and selectivity without the limitations of a SIM or HPLC analysis.

Solution:
We found there were 3 factors that led to the successful use of GC/MS method 8270 for PNA determinations:

1.  Newer generation mass spectrometer systems have proven to be more sensitive.

2.  Reporting limits can also be decreased if more sample is introduced into the GC/MS system.

3.  Rapid analysis time can be accomplished with pressure programming.

1.  New Generation GC/MS System

Newer generation mass spectrometer systems have proven to be more sensitive.  The Agilent Technologies 5972 is more sensitive than older models.  This system will produce a signal to noise ratio of 127:1 for 10 picograms of hexachlorobenzene in a full-scan mode (Agilent MS Application Brief MS 92-11).  We have observed excellent response for 0.5 picogram of Naphthalene.

It is important to our application to obtain low reporting limits with full scan spectra.  Older systems can reach low detection limits if they are used in a selected ion mode (SIM) during acquisition.  The full scan spectra in this procedure will give more assured identification of target compounds and allow us to perform library search routines on non-target compounds.  This will enable us to distinguish target compounds in difficult matrices.

One of the disadvantages in using the HPLC methodology, in this application, is the total dependence on retention time for compound identification.  With GC/MS, unique ions, which have been isolated from the background matrix, can be distinguished and used for quantitation. Samples from underground storage tank locations are typically very complex, containing many petroleum hydrocarbon compounds. The potential for co-eluting non-target analytes is extremely high.

2.  More sample is introduced into the GC/MS system.

Reporting limits can also be decreased if more sample is introduced into the GC/MS system.  We can accomplish this by using a “pressure-pulse” injection.  The injection port column head pressure is electronically controlled and can be programmed to begin the analysis at an elevated pressure.  Increasing the column head pressure at the moment of injection forces more of the analyte onto the analytical column.  The initial pressure that we used was 20 psi, which was held for the 1-minute splitless injection time.  The 20 psi column head pressure produces a column flow rate of 6.5 mL/min.  After one minute the pressure is returned to a constant flow rate of approximately 1.0 mL/min.

Additional analyte can also be introduced onto the GC by using a 2.0 uL injection volume as opposed to the routine 1.0 uL injection volume.

We have found that the latest generation GC/MS system from Agilent Technologies is even more sensitive than the 5972 model.  The 5973 does not require the use of pressure-pulse injections to meet the sensitivity described in this article.

3.  The analysis time is about 30 minutes.

For this method application to be feasible, the analysis time must not be too long.  To expedite the analysis we have taken advantage of the column pressure programming feature and ramped the column pressure as the oven temperature increases.  The parameters are set so that a constant flow of 1.0 mL/min is maintained throughout the sample analysis.  The runtime is 30 minutes with these parameters.   In this study a 30 meter Supelco PTE-5 column was used.  Column types of 0.25 or 0.32 mm ID have been used.

GC Run conditions are as follows.

Initial temp = 50 deg C;  Initial Time = 3.0 min.; Rate(1) = 10 deg/min.; Final Temp(1) = 200 deg C; Rate(2) = 20 deg/min; Final Temp(2) = 290 deg C.

The results:

Linear Range 

We found that the working range of the analysis is from 0.25 to 30 ug/mL.  A linear relationship is shown in this recent calibration curve.  Curves have been very stable (one lasting about 4 months).  

Response Factor Report Initial Calibration

PNA Analysis by Method 8270 (Low Level)

Calibration Level: (ug/mL)

MDL Study 

We found that we were able to meet all the low-level method detection limit requirements. We have seen similar results for soil samples extracted by sonication or soxhlet and also for aqueous samples extracted by separatory funnel. Calculated MDL values range from 0.04 to 0.14 ug/mL.

Response 

We have found excellent peak shapes and recognizable mass spectra for the first point (lowest concentration) of our calibration curve. The amount injected is 0.5 pg. 

Because of the listed benefits of performing a full-scan mass spectral analysis, this is our preferred method of PNA determinations.