Hi everyone,
I have been experimenting with using H2 as a carrier gas in place of He on our GC system, in order to reduce analysis times. However, I've been having problems in getting good peak shapes with H2.
The following are the conditions I normally use with He: I use a 30m x 0.25 mm i.d. x 25 um film SGE BPX70 column (highly polar) to separate C10-C24 FAMES. Up to now, I have been successfully using He as the carrier gas , at 27 cm/s (optimum) and a split ratio of 30:1 at the injection point (250 degrees C injector temperature, 110 degrees C initial column temperature, 100 kPa head pressure, ramped to 220 degrees C over 25 min at constant pressure). I typically make 2 uL injections of FAMEs in hexane into a 3 mm straight deactivated glass liner. There is a 2.5 m retention gap (methyl deactivated)installed between the injector and the column.
When I use H2 as a carrier gas under the same conditions, FAMES up to C16 have excellent peak shapes, but late eluting FAMEs (C18-C24) exhibit bad fronting ("sawtooth" shape). This suggests to me that: (1) the column is overloaded, or (2) the higher BP FAMEs are having problems getting onto the column, or (3) the higher BP fames are having diffusional problems between the stationary phase and the gas phase in the presence of H2.
I calculated that at the inlet pressure I use, the H2 flow and velocity (60 cm/s) is about double that for He, and in the optimum range for H2. To compensate for the higher flow rate, I have tried increasing the split flow to maintain a 30:1 split ratio and so avoid loading more sample onto the column. However, I still get peak fronting under these conditions. I've also tried lowering the column temperatures in the temperature program so that the elution times approach those for He, and this doesn't solve the problem.
Can anyone suggest a fix for this problem?

thanks
Tony

Tony,

I wouldn't have expected the carrier itself to lead to overloading of the phase. I wonder if the sawtooth is due to an injection problem which is manifesting itself as an overloaded peak shape for later peaks?

Assuming that column etc is the same and only the carrier was changed try 64 kPa giving a flow of 51.5cm/s (1.2ml/min)which gives a speed gain of 1.4 and you can set your split back to 30:1
( calculated using the HP capillary column method translator program http://www.chem.agilent.com/cag/servsup/usersoft/main.html#mxlator)

Regards,

Ralph

Dear Tony,

In a temperature programmed analysis, the change alone from Helium to Hydrogen causes only a minor reduction in the analysis time, but, if such a change was made without accompanying change in a temperature program, it can cause a lot of troubles. It can cause unpredictable change or a total loss of resolution of some peak pairs. It can even cause a reversal in the elution order of some peaks.

From the point of view of the speed of a temperature-programmed analysis, the key benefit of switching from Helium to Hydrogen comes from the fact that the switching allows to speed up the temperature program WITHOUT CHANGES IN THE RESOLUTION OF ANY PEAK PAIR, and without causing an additional column overloading. The key to that outcome is in the preservation of the elution temperature of ANY solute while switching from one carrier gas to another. In order to achieve that outcome, you need to TRANSLATE the temperature program that you had with Helium to the one that is accordingly appropriate for Hydrogen. For example, if you switch to Hydrogen at 1.2 ml/min (as Ralph recommended) then you need to increase all heating rates by a factor of 1.443, and to reduce the durations of all temperature plateaus by the same factor. This will reduce the analysis time by 31% (1/1.44 is about 0.69).

By the way. If the reduction of the analysis time is important for you and you have some resolution to sacrifice, you can try to do something more than just the translation from Helium to Hydrogen. Try Hydrogen at 2.5 ml/min. You might lose in the peak resolutions by about 10% to 15%. However, the speed gain would be 2.52 (instead of 1.44 as in the translation alone). You will get 60% reduction in the analysis time. This, of course, will happen only if you increase all heating rates by a factor of 2.52, and reduce the durations of all temperature plateaus by the same factor. If you have more resolution to sacrifice, you can try other tricks (cut the column or further increase the flow rate or do both, but do not forget to always translate the temperature program).

In addition to all of the above, you can reduce the column ID. This is the surest way to reduce the analysis time and to improve the resolution if necessary. Try your options with the Method Translation software (see Ralph on Thursday, May 8, 2003 - 08:06 am).

Good luck,
Leon