Chromatography Forum

current HiVac 8.17e-06

Changing from constant pressure to constant flow might account for the differences - at constant pressure the flow when the oven is hot is slower than at the beginning of the programme, while with constant flow it stays the same throughout. The faster flow at high temperature with constant flow carries more column bleed per unit time to the detector, hence more baseline drift, and the increased flow means that the peaks are moving through the column at different temperatures, so you get a change in selectivity which moves the peaks relative to one another. The difference in flow will also make the peaks narrower, so they will be taller for a given quantity of analyte.

At a split ratio of 2:1 the gas controls will struggle to maintain flows and pressures because the total flow is too low. Either go to splitless (with a purge flow of at least 20 ml/min), or increase the split ratio to 10:1.

Peter

With your original question of how to produce results like those obtained from the previous method. The only concern I have in looking at the chormatogram is the lack of resolution between 20:0 and 20:5. There are a couple of possibilites on thise - well actually three.

1) Since these did separate better on the longer column and lower flow, you may be able to run the separation with a slightly lower column flow and get that resoluion back. The colulmn will work well with a linear velocity down to about 23 cm/sec. You are proabably not going to want to get this low because lower head pressure on the column allows more air to diffuse into the inlet and you see it in the background of the mass spectrometer.

2) Try a slightly slower ramp through that portion of the chromatogram. I would suggest slowing the whole ramp down by just a degree per minute or so, just to see if it has a positive effect.

3) Given the degree of resolution, you may have adequate separation for the work that you are doing.

With any approach, you need to answer some questions that take you away from running a standard mixture: How does the chromatogram look with real samples? Do you see two peaks where you have both compunds present at widely differing concentrations? Does overloading of the column generate peaks so large that they swallow up neighbors?

Adjust as you feel that you may need and then take some samples and run them with your revised method. Check to see if the results match up with those obtained previously on the old method. You need to do this across the range of sample types you see and with enough samples in each sample type to detect any difference you would see in results. The question of how many samples is answered with the maximum allowed bias from your application, the variability you see in a few injections, and the computations for a power table (statistics book).