Chromatography Forum

Not knowning how many injections you do each day, rule of thumb I go t from somewhere: no more than 100 injetions on a septum. And, if you are using a SPME fiber, I'd make it fewer as the end of the spme fiber aparatus is going to tear up the septum a bit faster.

Are the liquid injections done with an autosampler? If manual, are you using a fast injection with a cold needle or are you witdrawing the sample into the syringe barrel and allowing the needle to heat?

Hi Eva

The solvents that you are using could be causing some problems. Until you are familiar with the instruments and techniques, and are sure that the hardware is actually working properly I suggest that you make up solutions in a volatile non-polar solvent such as hexane or dichloromethane. This will give you clean evaporations in the inlet, without excessive vapour volumes. Run a set of replicates with 1 ul injections of solution with about 1 mg per ml of each component and a 10:1 split. If that looks OK dilute the standard solution by ten, and run splitless, with the split opening after 30s.

Are the unknowns that you are identifying in natural materials (that have to be extracted etc) or in solutions that are already made up ?

Peter

Not knowning how many injections you do each day, rule of thumb I go t from somewhere: no more than 100 injetions on a septum. And, if you are using a SPME fiber, I'd make it fewer as the end of the spme fiber aparatus is going to tear up the septum a bit faster.

Are the liquid injections done with an autosampler? If manual, are you using a fast injection with a cold needle or are you witdrawing the sample into the syringe barrel and allowing the needle to heat?

At most each week I do 40 injections at the least I do 6 injections/week.

All injections are done with the autosampler -a Gerstel MPS-2.

It seems that a look at a chromatogram would be helpful in the flavor mixtures. These types of mixtures can be troublesome, particularly with water in the mix. Peter's advice on trying some injections in a more nicely behaved solvent is a good idea. If you have reproducabiltiy problems wiht a solution of a few compunds in methylene chloride - it takes us in one direction. If the problem stays with the typical solvents for flavorings, it takes us in another.

Often there is a sample of somethign like an 8270 mix around the lab. This is a great mix of environmental contaminants and does a nice job exercising an insrument. Or you may have some grob mix - another good test mix for a column. If not, a few failry non-polar comounds in methylene chloride at about one part per thousand (plan on a split injction) should be good. I would stay away from phenols, organic acids, and amines for this mix, at least for now. (Yes they are in the 8270 and grob mixes - but we can ignore them as we start out, if we need to.)

Here are the images of the chromatogram, the FID data for that sample, the FID for a blank sample, and the fid compared against the TIC.

http://imgur.com/a/8RBHd

<iframe%20class="imgur-album"%20width="100%"%20height="550"%20frameborder="0"%20src="http://imgur.com/a/8RBHd/embed"></iframe>

Sorry, I couldn't get it to post any other way. I tried to put the link in between the BBCode tags and it never loaded up properly.

Anyway, does this help?

On the differences between FID and TIC - mass range can be part of the issue... The FID generally responds to all carbon atoms in the molecue - regardless of molecuar weight. The mass spectrometer will give a signal only for masses included in the scan range - and if the molecule generates only ions outside of that mass range, the mass spectrometer will see nothing.

So, in the chromatograms you posted, what mass range or selction of masses was being used for the mass spectrometer?

For variability - how does the integration look for peaks across several runs. Peak shape does not look too bad, but I can not see what the baseline and noise look like around the smaller peaks.

In addition to Don's comments. It looks as if you have very different flow rates through the two columns - this is easily caused by the vacuum in the MS, check your instrument setup, there will be a column outlet pressure selection somewhere, make sure that it is at atmospheric for the FID side and vacuum for the MS column.

Having checked that make an injection of the headspace of some solvent that the MS can detect, and check that you see a peak at the same retention on both detectors.

Peter

With Peter's suggestion: A disposable cigarette lighter is a good source of butane. just release some of the gas without lighing it and draw it into a syringe.

Are the columns the same diameter, stationary phase, and length?

The mass range or selection of masses used for the mass spectrometer is 25 amu to 500 amu.

re: variability - you can see a close up of the peak shapes here http://imgur.com/a/81kIx.

re: different flow rates through the two columns -I have set the column outlet pressure selection to atmospheric for the FID side and vacuum for the MS column.
The columns are the same diameter, stationary phase, but one is shorter by about four feet, because of column maintenance issues.

I have not been able to get the two columns to elute at the same retention times. I'm not sure how to. I tried changing flows that didn't work, I tried matching flows, I tried changing the split on the FID, but none of those helped. Do you guys have suggestions for how to fix that?

I did an injection of ethyl acetate on both detectors and did not see a peak at the same retention on both detectors. I will try butane today.

Agilent doesn't think it's column bleed, so I'm still feeling like I'm muddling around in the dark. Any thoughts or advice are greatly appreciated.

I assume the larger peaks which show clean baseline work well, but the small peaks that are almost lost in the grass have high RSD's? For what I see as the smaller peaks in the trace with the noisy baseline, I would expect high RSD's - even with careful placement of integration marks.

On flow rates: Do you have the colums set for the same linear velocity? This may help, but with different operating pressures - even with the same average linear velocity in each column, expect some differences in retention times between the columns. I would expect the differences to change with GC oven ramp rate.

And, the difference in column length will give a difference in retention time. Be sure that the actual column length is entered into the instrument. This is used in calculating appropriate head pressure for the column. If the coumn is signifiantly shorter than the value set in the isntrument, the head pressure will be set too high - making the column flow even faster.

Is 1 to 2 feet difference in length enough to make a big difference in terms of entering the column length?

The column is set to have the same total flow is that the same as the linear velocity?

Yes, the RSDs vary greatly for the small peaks. Sometimes the RSDs also vary greatly for the larger ones, though.

Is 1 to 2 feet difference in length enough to make a big difference in terms of entering the column length?probably not, but earlier the difference was 4 feet, and if it is really, say 8 feet, then it might be significant
The column is set to have the same total flow is that the same as the linear velocity?No, it is completely different, and this might be the cause of the problem. The total flow setting is the the sum in volume flow terms of the column flow, the split flow and the septum purge flow. It does not control the flow through the column. The flow through the column depends on the pressure in the inlet (and at the outlet), which is controlled by setting either the volume flow through the column, or the linear flow through the column. Set a constant linear flow for both columns of 35 cm/s and the electronic gas controls will take care of programming the pressure, and your retention times should then be very similar.

Yes, the RSDs vary greatly for the small peaks. Sometimes the RSDs also vary greatly for the larger ones, though. sample transfer from inlet to column has probably been disrupted by not having the carrier gas properly controlled

Peter

Is there a good way to figure out how long the column actually is?
I know my boss has cut it several times. I've cut it a few times. I don't know how much has been cut off in sum total.

Is there a good way to figure out how long the column actually is?
I know my boss has cut it several times. I've cut it a few times. I don't know how much has been cut off in sum total.

Tedious but accurate - measure the diameter of the coils and count how many coils there are, 2 pi r x coils = length.

Peter