Improved Indeno[123-cd]pyrene-Dibenz[ah]anthracene Separation on the New Rxi-SVOCms Column

When we started development of the Rxi-SVOCms column, we focused on maximizing the resolution of PAH isobars while targeting a 16-minute elution time for benzo[ghi]perylene (BghiP) on a 30 m x 0.25 mm x 0.25 µm column using our optimized split injection SVOA acquisition parameters [GC_EV1604]. The benzo[b]fluoranthene - benzo[k]fluoranthene (BbF - BkF) separation draws the most attention, but we also put significant resources into maximizing the indeno[123-cd]pyrene – dibenz[ah]anthracene (IcdP - DahA) separation in the allotted time.

Figure 1 - Indeno[123-cd]pyrene (left) extracted ion chromatogram showing good resolution from the dibenz[ah]anthracene interferences (right). Peak areas for [276+] are marking the apices.

So why do these two PAHs, which have different nominal masses, need to be chromatographically separated? Because DahA (nominal mass 278) has a very strong signal at m/z = 276 (the nominal mass of IcdP) under electron ionization at 70eV. Coelution of these two compounds causes several problems with the analysis:

• Calibration bias
• Mischaracterization of sample toxicity
• Misidentification of the DahA [276+] signal as IcdP (this is especially devastating during proficiency test studies)

Calibration Bias

Let’s look at an extracted ion chromatogram for IcdP where IcdP and DahA are well separated on the Rxi-SVOCms cat#16623 (Figure 1 Were IcdP and DahA to coelute, the quant ion signal ([276+]) for IcdP would be inflated 38.2% by DahA ions. Running a calibration under these conditions would cause a significant negative bias in the calculated value for IcdP concentration in a sample extract.

Mischaracterization of Sample Toxicity

A PAH mixture’s toxicological profile is characterized using Potency Equivalency Factors (PEFs). Carcinogenicity is expressed as Benzo[a]pyrene (BaP) equivalent:

• Indeno[123-cd]pyrene PEF = 0.1
• Dibenz[ah]anthracene PEF = 2.4
• Benzo[b]fluoranthene PEF = 0.1
• Benzo[k]fluoranthene PEF = 0.1
• Benzo[a]pyrene PEF = 1.0

IcdP’s PEF is only 0.1 while DahA’s is 2.4, so reporting IcdP’s concentration in samples with a negative bias of nearly 40% shouldn’t be that bad, right? The answer really depends on the relative abundance of each PAH in real samples. Let’s look at the NIST SRM “PAHs in Coal Tar” analysis (Figure 2). In this case, IcdP contamination exceeds DahA by a large amount. Were this a food commodity being tested for PAHs, a near 40% low bias in the calculated value of a PAH with a high signal could be the difference between properly characterizing the food as unsafe for humans and a false negative allowing the commodity to be distributed.

Figure 2 - Indeno[123-cd]pyrene and dibenz[ah]anthracene in NIST SRM 1597a (part of GC_EV1606)

Misidentification of the [276+] Ion Signal from Dibenz[ah]anthracene

This is the simplest situation, but it can lead to disastrous results. If IcdP and DahA coelute in your chromatographic run, and you receive a PT sample that contains DahA only at a moderately high level, you will be unable to tell if the [276+] is from DahA or IcdP. Misreport those compounds on consecutive PT rounds and the local regulating body will likely pull the certificate authorizing your laboratory to perform the analysis.

Recommended Analytical Conditions for the Rxi-SVOCms Column

Figure 3 shows the continuing calibration verification level (20 ng/µL) from our 8-point calibration ranging from 1 ng/µL to 120 ng/µL, or 100 pg to 12 ng column loading for each of the 86 target compounds and surrogates on cat#16623 – a 30 m x 0.25 mm x 0.25 µm Rxi-SVOCms Column (GC_EV1604).

Figure 3 - GC_EV1604 - Example total ion chromatogram of a 94 component semivolatile standard injection with a 10:1 split injection and 2 ng column loading for each component

Optimized for a 16-minute run, we get excellent resolution of both critical PAH pairs. 86% valley as reported by MSD ChemStation for BbF and BkF and 87% valley for IcdP and DahA. If 16 minutes isn’t fast enough for your needs, tune in next time when I explain how to cut 5 minutes from the analysis time without sacrificing the critical separations.