Quadrupole Scan Speed and the 8270 Instrument Checkout Mix – Part II - The DFTPP Evaluation

Yesterday’s blog gave an overview of how choosing the wrong scan speed could be detrimental to the tailing factor evaluation. Before someone asks, I thought we'd spend today looking at the impact of scan speed on the decafluorotriphenylphosphine (DFTPP) tune evaluation.

I know I’ve covered this before, but here are the tuning criteria from EPA 8270 D (rev 5) intended for scanning quadrupole instruments:

11.3.1.1 - In the absence of specific recommendations on how to acquire the mass spectrum of DFTPP from the instrument manufacturer, the following approach should be used: Three scans (the peak apex scan and the scans immediately preceding and following the apex) are acquired and averaged. Background subtraction is required, and must be accomplished using a single scan acquired within 20 scans of the elution of DFTPP. The background subtraction should be designed only to eliminate column bleed or instrument background ions. Do not subtract part of the DFTPP peak or any other discrete peak that does not coelute with DFTPP.
EPA Method 8270 "Table 3 - DFTPP Key Ions and Ion Abundance Criteria"

To summarize, the tune evaluation consists of a background subtracted average of the 3 apex scans evaluated against the criteria in Table 3. Using the same instrument settings used in the tailing factor evaluation blog, let’s evaluate DFTPP at 3.1 and 5.9 Hz.

First, there is the qualitative chromatographic evaluation.  Figure 1 is an extracted ion chromatogram showing the six major ion fragments of DFTPP, collected at 3.1 Hz. Notice that the green trace of m/z = 442 is out of sync with the other mass peaks. This is also the case, but to a lesser degree, in Figure 2, the extracted ion chromatogram of DFTPP collected at 5.9 Hz. This offset of the high mass peak is due to spectral tilting, which is one of the issues experienced when scanning too slow.
Figure 1 - Extracted major DFTPP ion chromatogram collected at 3.1 Hz
Figure 2 - Extracted major DFTPP ion chromatogram collected at 3.1 Hz

Spectral tilt occurs when the amount of an analyte entering the detector increases during an individual scan (mass/second). Temperature programing and constant carrier gas flow combined with high efficiency GC columns yields chromatographic peaks with very steep slopes. The 5975C used here scans from high mass to low mass because it takes less energy to reset from low to high. Using DFTPP as an example, it would stand to reason that if you slowly scan from m/z = 500 to m/z 35 while the amount of DFTPP entering the detector is increasing, the resulting value for m/z = 442 would be biased low, and the m/z = 51 value would be biased high. The opposite would be true for the tailing side of the peak, where the amount of DFTPP entering the detector is dropping over time. 3.1 Hz is slow enough to see these effects when the three scans specified by EPA 8270 (the peak apex scan and the scans immediately preceding and following the apex) are inspected individually (Figure 3, Figure 4 & Figure 5)
Figure 3 - 3.1 Hz DFTPP scan 1463 (pre-apex)

Figure 4 - 3.1 Hz DFTPP scan 1464 (apex)

Figure 5 - 3.1 Hz DFTPP scan 1465 (post-apex)

Consider that the mass spectrum for each “scan” is just a bar graph showing the response for each ion. In the absence of spectral tilt, there would be no change in the slope of a line drawn from the top of the bar for m/z = 51 to the top of the bar for m/z 442, as shown in Figure 6, when you compared each scan for a given compound.
Figure 6 - 3.1 Hz DFTPP scan 1463 (pre-apex) with trace for visualizing low to high slope shifts

To perform this tilt evaluation for the 3.1 Hz scan rate, we created an X,Y scatter plot, with the responses for m/z = 51 and m/z = 442 entered for each of the three scans used for the DFTPP evaluation (Figure 7). The slopes are very different.
Figure 7 - Spectral Tilt evaluation (m/z = 51 and m/z = 442) at 3.1 Hz

Nearly doubling the scan rate to 5.9 Hz does not eliminate the spectral tilt, but it does minimize the effects. Looking back at Figure 2, it is clear the 3 scans used for the DFTPP evaluation are all clustered near the actual peak apex, when the 2 adjacent scans from the 3.1 Hz data were a ways down the front and back of the DFTPP peak (Figure 1). This clustering near the apex is reflected in the ion ratios of the 3 individual scans used for the tune evaluation (Figure 8, Figure 9, and Figure 10)
Figure 8 - 5.9 Hz DFTPP scan 2638 (pre-apex)

Figure 9 - 5.9 Hz DFTPP scan 2639 (apex)

Figure 10 - 5.9 Hz DFTPP scan 2640 (post-apex)

Evaluating the spectral tilt using the same X,Y scatter plot setup used for the 3.1 Hz evaluation shows less variability for the for m/z = 51 and m/z = 442 ions acquired at 5.9 Hz (Figure 11) and a much smaller range of slopes.
Figure 11 - Spectral Tilt evaluation (m/z = 51 and m/z = 442) at 5.9 Hz

A box plot of responses for the averaged scans for the 3.1 Hz (Figure 3, Figure 4 and Figure 5) and 5.9 Hz (Figure 8, Figure 9 and Figure 10) evaluations highlights the reduced response variability for virtually all the ions of significance (Figure 12). This is because the faster scan rate allows repeated sampling in a zone where the concentration of analyte in the detector is fairly stable (the pseudo-plateau at the peak apex).
Figure 12 - Ion Responses for the 3 DFTPP scans averaged at 3.1 and 5.9 Hz

Over the course of two days, I evaluated 40 DFTPP tune verification runs – 20 at 3.1 Hz, and 20 at 5.9 Hz. One of the 20 tune evaluations collected at 3.1 Hz failed because the ratio of m/z = 68 to m/z = 69 exceeded 2.0. All 20 DFTPP runs collected at 5.9 Hz met all tune criteria. The 68:69 ion ratio was interesting because even though the variance for the 5.9 Hz data was larger than that of the 3.1 Hz data, the median was approximately 0.5 for the 5.9 Hz data though greater than 1.5 for the 3.1 Hz (Figure 13).
Figure 13 - Boxplot and Individual Value plot of the m/z = 68 to m/z = 69 ratio for 3.1 and 5.9 Hz DFTPP evaluations (n=20)

Overall, average tune performance between the slow and fast acquisition populations was similar (Figure 14 and Figure 15), even though there were zero failures in the 5.9 Hz data set. I suspect the lower scan to scan variability of the faster scan rate highlighted in Figure 12 makes the occurrence of an evaluation failing outlier less likely.
Figure 14 - Boxplot of 3.1 and 5.9 Hz DFTPP Tune Evaluation Ion Ratios (n=20)
Figure 15 - Variances for the 3.1 and 5.9 Hz DFTPP Tune Evaluation Ion Ratios (n=20)