Journal of Chromatographic Science, Vol.

41, March 2003

Gas Chromatography Problem Solving and Troubleshooting

Question

The results of a N,N-dimethylformamide (DMF) purity analysis performed in my lab is different than the results obtained
in another lab. We are using identical capillary GC methods and techniques, but the DMF purity differs by about 5% for
the same DMF sample. Each lab could consistently reproduce their results. In addition, the DMF peak is not symmetrical
and all attempts to correct this problem have failed. What is the cause of the purity difference and asymmetric DMF peak?
Which purity result is correct?

Downloaded from https://academic.oup.com/chromsci/article/41/3/164/287003 by guest on 07 August 2022

Answer

The capillary GC analysis of high-purity solvents requires specific GC conditions and a few compromises. In order to
detect very low level impurities in a high-purity sample, a sufficient quantity of the sample has to be injected into the
column. This results in a very large amount of the primary compound entering the column. Column, injector, and detector
constraints occur because of the presence of both extremely high and low level compounds in the same sample.
Probably the most overlooked GC parameter for high-purity analysis is detector range. With a few exceptions, detector
response or output is proportional to the amount of compound eluting from the column. The detector output is plotted
versus time, which results in the familiar chromatogram. The area or height of each peak is calculated by the data
system. Most detectors have a linear response over an amount range. For example, if the compound amount increases by
two, the peak area or height also increases by a factor of two, as long as the range of the detector is not exceeded.
Detectors have a maximum output signal
that is not surpassed even if the
compound is present in extremely high A
amounts. The output signal will remain at
this sample maximum value until enough
of the compound has eluted from the
column. Exceeding a detector's output
range results in flat-top or square peaks.
An example is shown in Figure 1A. Peak
areas or heights for out-of-range peaks can
not be accurately calculated. While
increasing the amount of an out-of-range B
compound results in larger peak area or
height, they are no longer proportional
with the sample amount. Relative peak
areas or heights become skewed when
one or more of the peaks are out of range.
Sample purity can no longer be accurately
calculated by comparing individual peak
areas with the total peak area.
The chromatogram in Figure 1 was Figure 1. DMF chromatogram using a detector range of 0 at full scale (A) and reduced scale (B).
obtained using a detector range of 0. A The chromatographic conditions were a DB-WAX column (30-m × 0.53-mm i.d., 1.0 µm), split in-
range of 0 is the default or typical value for jector at 250°C and split ratio of 1:5, FID detector at 320°C, helium as the carrier gas at 31 cm/s,
most GC systems, and this range is suitable and a column temperature program of 70°C for 1 min (70–250°C at 20°/min).
for most capillary GC analyses. A flat-top

The purpose of Chromatography Problem Solving and Troubleshooting is to have selected experts answer chromatographic
questions in any of the various separation fields (GC, GC–MS, HPLC, TLC, SFC, HPTLC, open column, etc.). If you have
questions or problems that you would like answered, please forward these to the Journal editorial office with all pertinent
details: instrument operating conditions, temperatures, pressures, columns, support materials, liquid phases, carrier gas,
mobile phases, detectors, example chromatograms, etc. In addition, if you would like to share your expertise or experience in
the form of a particular question accompanied by the answer, please forward it to: JCS Associate Editor, Chromatography
Problem Solving and Troubleshooting, P.O. Box 48312, Niles, IL 60714. All questions/answers are reviewed to ensure
completeness. The Journal reserves the right not to publish submitted questions/answers.
Dean Rood
Associate Editor

164
Journal of Chromatographic Science, Vol. 41, March 2003

peak is obtained when the detector range is exceeded. Injecting a smaller sample amount, using a very high split ratio, or
diluting the sample with another solvent are not viable options because the low level impurities will be below the
detection limit of the method (these options also create other problems). The detector’s maximum output signal can be
increased by increasing the detector range value. Figure 2 shows the chromatogram obtained using a range of 6 for the
exact same sample and GC conditions as in Figure 1. The peak using a range of 6 is now onscale without a flat top. An
accurate peak area can now be obtained. Table I shows the data obtained using a detector range of 0 and 6. The purity
results (peak %) for the two analyses are noticeably different. A new bottle of DMF used in this example reported purity of
99.9%. The purity obtained using a detector range of 6 is much closer to the value reported by the DMF supplier.
Assuming the DMF sample is not contaminated, the +99.9% result is probably the more accurate analysis value.
Increasing detector range also decreases detector sensitivity. This is evident for the chromatograms in Figures 1B and
2B. The peaks in Figure 1B (range = 0) are larger, but the background is also increased. The relative peak percent of the

Downloaded from https://academic.oup.com/chromsci/article/41/3/164/287003 by guest on 07 August 2022

peak at 7.911 min is calculated to be 3.798% for the detector range of 0. This value appears to be too high when visually
comparing the relative peak sizes in the chromatogram. The
sloping or noisy baseline is probably contributing to an Table I. Integration Data for Figures 1 and 2
integration error, which affects the peak-area and relative-
area percent accuracy. Even though the sensitivity is lower at Range 0
a range of 6 (Figure 2B), the small peaks are large enough to
be easily detected and properly integrated. The more Retention time Area Symmetry Area %
reasonable result at a detector range of 6 is caused by the 1.844 5281.97 1.508 0.080
onscale DMF peak and the more accurate integration of the 2.188 4507.76 1.385 0.069
small peaks. 2.502 49,864.96 1.017 0.760
The DMF is present at a very high level (> 99.9%). Even 3.223 5701.30 1.037 0.087
with a very small injection volume, the amount entering the 6.197 6,228,946.50 5.810 94.887
capillary column significantly exceeds the column capacity 6.570 20,976.26 1.800 0.320
and an overloaded peak is obtained. Overloaded peaks have a 7.911 249,311.27 22.467 3.798
pronounced slope or leading edge. The peak in Figure 2A is a
good example of an overloaded peak. Trying to obtain a Range 6
symmetry value between 0.9 and 1.1 for an overloaded peak is
a futile exercise. To minimize the severity of the peak overload, Retention time Area Symmetry Area %
wide-diameter and thick-film capillary columns are often used. 1.842 41.50 0.829 0.002
Sample capacity increases as capillary column diameter and 2.186 29.53 0.847 0.001
film thickness increase. Though a smaller diameter and thinner 2.499 785.53 0.840 0.032
film column can be used, it will become overloaded at a much 3.222 40.36 0.901 0.002
lower sample level. One benefit of using the higher detector 6.166 2,482,055.25 2.031 99.961
range is the better symmetry value (Table I). 6.570 49.66 0.954 0.002
It is assumed that the detector has equal response for 7.910 23.08 1.031 0.001
every sample compound when calculating sample purity
using relative peak areas. Although this
assumption is incorrect, the unequal A
detector response is usually ignored
because correction factors can be difficult
to calculate and often do not significantly
change the results. The amount of error is
dependent on the specific type of
detector. Because the same sample
and method was used in both labs,
it is very unlikely that the purity analysis Time (min)
difference is caused by unequal response-
factor issues. B
The different purity results obtained
between the two labs are probably related
to the detector ranges being used. If one of
the labs is using a detector range that is
too low and out-of-range peaks are
obtained, incorrect peak areas are
probably being reported, which negatively
affects the purity results. Obtaining ideal Time (min)
symmetry for an overloaded peak is an Figure 2. DMF chromatogram using a detector range of 6 and at full scale (A) and reduced scale
unrealistic expectation, and attempts to (B). The chromatographic conditions were the same as for Figure 1.
improve the symmetry will not work.

165