high boilers in split/splitless

[WK]

Hi All,
I have samples which contain a range of mw/boiling points. I have some which are mw300-400 which elute late in split injection.
When I dissolve sample in hexane and do splitless injection they do not appear - which is good for me since less cleaning of MS etc.
This is a useful discrimination since high boilers dont go into MS - and I still get my lower mw interesting peaks.
Can anyone explain what's happening here?
WK

[CE Instruments]

All split/splitless injectors discriminate, i.e. they do not transfer the same proportion of each molecular weight onto the column. Many start to discriminate at about C20 with higher hydrocarbons dropping off. The reason is a failure of the injector to vapourise the sample sufficiently. In split injection this means they remain in the solvent and exit via the split vent. In Splitless mode the injector must provide more energy to to vapourise the whole sample, when it fails the unvapourised sample is purged from the injector as the split vent is opened. Splitless injection is still a split of the sample, you are transfering more of the low MW compounds and the higher MW compounds do not reach the phase on the column and are lost.

[Victor]

WK seems to be suggesting that s/he is perhaps seeing some discrimination of high boiling compounds in split injection, but s/he is seeing much more discrimation in splitless injection. I am not sure about the explanation given for this by CE instruments.

Firstly, did you change anything e.g. the liner in going from split to splitless injection? If you did then you have changed the conditions of the experiment.

Discrimation is generally attributed to incomplete vaporisation of the heavy sample components from the syringe needle. Such discrimation is reduced in rapid autoinjectors, where the needle hardly has time to heat up and thus the sample leaves the needle as a liquid plug rather than an aerosol. However, this discrimation mechanism is identical in split and splitless injection, as long as the same conditions are used. It then seems that some further mechanism is operating in the injector. It could be that there are cold spots in the injector on to which the heavier components gradually re-condense during the splitless time period giving further discrimination. Maybe this could also occur on pieces of septum which might gather on the bottom of the injector. It is possible that these processes do not have the time to occur in split injection where the portion of sample that is destined to enter the column does so very rapidly, rather than over a period of a minute or so during the splitless injection. However, you need to be sure you are comparing identical situations. Split injection is often performed with the initial column temperature much higher than in splitless injection- I could guess that this might also have some affect on the results.

[Peter Apps]

As is all too often the case, any answer has to be largely speculative because the query does not contain anywhere near enough information on operating conditions etc.

Nonetheless;

The simplest explanation is that the high molecular weight components are not soluble in hexane, so they never get injected.

It is also possible that the late eluting peaks are column bleed rather than sample components; with a silicone phase an injection of a sample with water in it can release a pulse of siloxanes that appear as peaks rather than baseline drift. Dissolving the sample in hexane will tend to partition out the water, hence less column bleed with your splitless injection.

Depending on the inlet temperature, the high boilers take time to evaporate into the carrier gas, if the split valve is opened too soon the high boilers do not get a chance to evaporate onto the column. It is also possible that they are evaporating so slowly that their peaks are too broad to be detected as such - this will depend on the inlet temperature, the column temperature at injection and whether or not you programme the column temperature. In this scenario the heavy compounds are still entering the MS, you just cannot see them as peaks.

On Agilent GCs run under Chemstation the default splitless time is 0.0 for some reason, and it is easy to forget to change it, resulting in a split injection with all sorts of pressure pulses as the valves open and close.

Depending again on temperatures, and also on the volume that is injected you may be recondensing solvent onto the column, which can generate all sorts of striking peak distortions, although I would be surprised if it would make peaks disappear completely.

You need to check (with a flow meter, not software !) what your split and septum purge flows really are - as the sample or solvent evaporates in the (presumably) hot inlet it generates a pressure pulse that can send gas flows in all sorts of unexpected and undesirable directions.

If you really need to discriminate against high boilers then some kind of backflush system will be far more reliable, unfortunately the commercial systems are absurdly costly.

Peter

[WK]

Interesting point from Peter Apps,
The "backflush effect" is interesting. This was what I likened it to.
I managed to detect my latest eluting peak of interest without eluting all the subsequent high boilers. Which is ideal for me in this particular analysis.