Published Chris Cook on May 19, 2020
The VGA-100 and VGA-101 use a deuterium lamp as the source for the high intensity ultraviolet radiation used by the instruments. The lamp signal varies with changes in temperature and pressure inside the bulb, and such changes occur when a cold lamp is turned on and begins to heat up. It takes approximately 15 minutes for the lamp to reach thermal equilibrium and lamp output to stabilize. We recommend having the lamp on for 30 minutes prior to running samples in the instrument to ensure enough time has passed. Running a solvent as the first sample in a sequence is a common procedure to ensure the lamp is ready.
But what does the data look like if the lamp is not given enough time to warm up prior to data acquisition? Let’s have a look.
Chromatograms from two runs of the VUV-CS PIONA Check Standard (using the ASTM D8071 PIONA analysis GC conditions) are shown below (Figure 1). The blue chromatogram represents a sample run just after the lamp had been turned on. As we can see, the baseline absorbance decreases as the run progresses, indicating that the lamp signal strengthens as the lamp warms up. The red chromatogram represents a sample run after the lamp was already warmed up, and far less change in the baseline is seen.
Summing retention regions at the end of the chromatograms with background regions from the beginning, we see the following absorbance patterns. In Figure 2 the red line again shows the run with the warmed-up lamp and the blue line shows the run with the cold lamp. The change in lamp signal with the warming lamp is noticeable, and it caused a greater signal increase in the lower wavelengths.
To see the effects of this when analyzing a peak in VUVision™ Software, we will zoom in on the 2-methylnapthalene peak at around 25.3 minutes (Figure 3). The peaks in the two chromatograms have slightly different baselines but are otherwise similar in shape and size.
Shown below in Figure 4, if we were to use background region from the beginning of the chromatogram from the cold lamp run, we see that the spectrum of the peak is distorted and VUVision has trouble matching anything at all.
When a background region close to the peak is chosen, as happens when using VUVision’s auto peak detection algorithm, the spectral match was much closer to 2-methylnaphthalene (Figure 5). A properly chosen background region compensates for gradual changes in the baseline and allows the spectra to be accurately identified.
The peak areas were slightly different, with a 0.8897 peak area for 2-methylnaphthalene in the cold lamp run and 0.9450 for the warm lamp run. This difference in chromatographic peak areas was consistent throughout the chromatograms, with the peak areas from the warm lamp run being slightly larger than the peak areas from the cold lamp run.
What if we did an ASTM D8071 PIONA analysis of the samples in VUV Analyze™ Software? To ensure the setup gives passing results with proper operation, we first look at the run done with the warmed-up lamp using the “Analyze Check Standard (CS)” function in the D8071 Setup Mode.
Next, we move on to an identical run but with a cold lamp (Table 2)
This is particularly interesting because the ethyl alcohol peak in the cold lamp run was smaller (14.5741 on the cold lamp run, 14.9749 for the warm lamp run), but because of changes in other analytes, the ethanol mass percent in the cold lamp run is the larger of the two when analyzed in VUV Analyze. The cold lamp start caused values to go far enough out of range to fail the VUV-CS quality control check.
Looking at the pdf reports of the two runs, we can see that many of the olefin peaks are quite small. The cold lamp run (Figure 7) misses many of the smaller peaks found in the run with the warmed-up lamp (Figure 6).
The decrease in background caused by the increasing lamp intensity causes problems in how VUV Analyze determines what is analyte and what is background. Checking the D8071 report to see the effects on area rejects, we find 0.324% area reject for the warm lamp and 0.267% reject for the cold lamp. This is also because of the loss of the small peaks. In the warm lamp run the small peaks were analyzed as peaks, but VUV Analyze could not accurately identify them. In the cold lamp run the peaks were not identified as peaks at all, so VUV Analyze did not include them in the area reject. Although the cold lamp affected D8071 results, it was not immediately obvious from looking at the D8071 report that anything was wrong.
In conclusion, making sure the lamp is warmed up prior to data collection is good practice. Failure to do so can affect results in unpredictable ways.