Temperature programme doesn't separate on new GC

[Seabiscuit]

I was routinely using a Varian 3400 FID to analyse FAME, using a temperature programme we have developed over the years. We use a 100 m CP-Sil 88 column for FAME, hydrogen as a carrier gas at 2 ml/min, injecting 2 ul but with a split of 1:50. We are always very interested in the C18:1/C18:2 region of the chromatogram, which elutes just over half way through a run (total run time 60 min) and contains many isomers in very small quantities.

We now have a new GC (Bruker 450) with EFC. When I applied the same GC conditions and temp programme, we do not achieve the same degree of resolution around the C18:1/C18:2 region. The GC engineer advised us to increase injection volume or reduce the split to increase peak size, but all this does is mask the smaller isomers which then co-elute. I have tried reducing the column flow rate to 1 ml/min but the peaks are still co-eluting. I then tried slowing the temp ramp at this stage, but this also had no effect.

I guess it is a case of trial and error with any new machine, but I am puzzled as to why exactly the same conditions, column etc will not produce the same degree of accuracy we had with the old machine. Any advice would be welcome. Thanks!

[chromatographer1]

People will often state they are getting different results while doing an analysis exactly the same way they did before, with the same column, but the only difference is the GC.

Well, it only means that something is different, but they don't realize it.

Were you using constant pressure for the carrier, but now are using constant flow rate, or the opposite?

Does the column you are using now still performing the separation on the older GC?

It is possible the injector inlet is different in temperature or design that it might affect the separation but that is unlikely, but possible.

best wishes,

Rod

[Peter Apps]

Rod beat me to it - my money is on your using constant (volume) flow with the new instrument, while you had constant pressure on the old one.

And as a general point, reducing the carrier flow rate will nearly always make resolution worse, not better.

Also, check that you have the right carrier gas selected in the GC setup.

Peter

[Seabiscuit]

Thanks for the replies...yes I guess that's the reason for the difference - we now have a constant flow rate, whereas before we had constant pressure. The column is exactly the same column we had in the old machine. There were liner issues - Bruker installed a bog standard liner in the new machine so I have since swapped it for a focusliner (what we used on the old machine) but it didn't make much difference.

Why would reducing the flow rate not improve resolution?

Any tips as to what I can try next? I'm no expert at GC operations so any help would be most welcome, thanks.

[Peter Apps]

Since having constant flow rate is very likely the cause of the problem your first step needs to be to change to constant pressure, only if that does not work do you need to think about the next step. On the front panel of the GC (or the software equivalent depending on how you are doing things), press the pressure gauge icon at the top right. This takes you to the carrier gas flow control screen. About half way down on the left there is a tick box for constant flow, unselect it. A set of values for pressure appear at the bottom of the screen, change them to what you had on the old instrument.

The resolution gets worse at slower than optimum flow rates because the peaks take longer to elute and band spreading by longtitudinal diffusion gets worse with time.

Peter

[Seabiscuit]

I did this and the resolution has slightly improved so thanks. However it is still no where near as clear as it was with the old machine. In addition, the peak heights are almost 10 times smaller than they were before so the detector response isn't as good.

If I tried to increase peak size (by for example reducing the split flow), I am worried that I will "lose" some of my less abundant isomers as larger peaks may mask them. Should I first concentrate on trying to improve resolution before trying to increase peak size?

Thanks for all your help on this.

[Johnny Rod]

As Peter says, it would be best to replicate your old method exactly on the new instrument, then you have a good starting point. After that I'd say, determine the optimum flow rate (linear flow speed) for your column, so you get the best resolution (greatest number of theoretical plates). For hydrogen this is about 30-55cm/s so is probably close to your original 2ml/min (depending on the column diameter). Using constant flow mode will mean you get more or less the same volume flow and so linear speed at all temperatures, so the later peaks should appear sharper and elute quicker. Thus they are taller so your sensitivity is increased. You would probably have to monkey around with the temperature ramp to optimise.

I also should ask, what settings are you using for the FID?

[Seabiscuit]

Thanks - FID is set at 255degC. Do you want to know gas flow rates to the FID?

The column we're using is 100 m, 0.25 mm i.d.

Everything is exactly the same as was set on the old machine - detector gas flow rates, temperature programme, and now I have set the column to constant pressure rather than constant flow as that's what the old machine used (30 psi, equates to around 2 ml/min at the starting column temp of 70degC). I'm using the same split ratio (50:1) and injecting the same volume of sample (2 ul). But, for the region of interest, the resolution of peaks is not as good as it was before and the peak heights are smaller.

Would type of liner have any big effect? As I mentioned I have put a focusliner in but are there different types of focusliner? The liners in the new injector are much larger than in the old injector - is this having an effect on peak height?

Is there any way I can post pictures on here? I wondered if it would be worth sharing example chromatograms. Thanks again for all your advice!

[Ultimate Science 101]

Make sure to also check the attenuation (range) you are using on the detector. At least from the Varian 3800 to the Varian (Bruker) 450 GC, the equivalent attenuation values are now switched in the control software so that the equivalent Range 1 setting on the 3800 is now Range 10 on the 450, and vice-versa. Another good idea at this point is to check column cuts and installation distances into both the injector and detector, ensuring they are to proper specifications for your new system.

[Peter Apps]

Instructions for posting chromatograms are in a sticky near the top of the GC page.

What model of inlet do you have on the Bruker 450, and what did you have on the Varian ? I am puzzled that the liners you now use are much bigger than those that you used in the Varian.

An off the wall thought - when you moved the colum to the new instrument did you connect it the same way around - inlet end to inlet, detector end to detector ?

Are you actually measuring resolution (separation between peak maxima divided by peak width) or just eyeballing the look of the chromatogram ?

Peter

[Seabiscuit]

I'll check the attenuation range...apologies for my ignorance but what exactly does this specify? The column (which was the same one we used on the old machine) was installed by the engineer so I know this will have been installed correctly.

Peter - bit confused as to what you mean by model of inlet? I have checked the liners - our old liners for our 1077 injector were 3.4 mm diameter, whereas the new liners (1177 injector) are 4 mm. they are visibly longer and wider, but I just assumed this was because the injector was different.

And yes when I say resolution I mean resolution by eye - peaks that used to separate with the old machine are now not separating as well....you can see where the peak apexes are but the troughs are nowhere near as distinct as they were.

[Seabiscuit]

The first chromatogram below is a section containing the area of interest, from the old machine. You can see the peak definition. It is a chromatographical challenge as some of these molecules are present at very small quantities. They are all also isomers with the same molecular mass....ignore the peak label for C18:1t13-14 as it is incorrect in this instance.



The next chromatogram is from the new machine. Note the area I have circled - these peaks we never had a problem with with the old machine but they aren't as distinct here which will lead to integration problems..



Not forgetting the peak size issue...but I will check the attenuation range this morning. thanks

[Peter Apps]

I'll check the attenuation range...apologies for my ignorance but what exactly does this specify? The column (which was the same one we used on the old machine) was installed by the engineer so I know this will have been installed correctly. You have a lot more faith in service engineers than I have and you need to check the column insertion depth in both inlet and detector, and make sure that the end of the column is clean, with no bits of ferrule inside. If the column was installed by the service engineer I presume that you do not know whether it got turned around in the process. Although in theory the direction of the column should not make any difference, in practice the inlet end suffers more damage than the detector end, and swapping end for end would put the damage near the detector where it might have more impact

Peter - bit confused as to what you mean by model of inlet? I have checked the liners - our old liners for our 1077 injector were 3.4 mm diameter, whereas the new liners (1177 injector) are 4 mm. they are visibly longer and wider, but I just assumed this was because the injector was different.The model of inlet is the numbers that you have given me. I presume that the liner diameters you are giving are internal diameters, since they do not correspond to any external diameters that Varian or Bruker uses. I very much doubt that the difference in internal diameter will degrade the separation to the extent that your chromatogram shows. However you say that you are using the same focus liner is you had in the Varian, so how can it be longer and fatter ?

And yes when I say resolution I mean resolution by eye - peaks that used to separate with the old machine are now not separating as well....you can see where the peak apexes are but the troughs are nowhere near as distinct as they were. Having seen the chromatogram I see what you mean - nasty ! You have about a 20 % increase in peak width by the looks of it.

Have you checked and calibrated the septum purge flow ?

Small bits of debris in the inlet can cause peak broadening - check that there are no fragments of septum on top of the wool, and no bits of setum or liner ferrule in the bottom of the inelt body

Peter

[Ultimate Science 101]

As Peter mentioned, never assume your installation distances are correct just because a service technician installed it for you. Columns can slip down down as you are tightening the ferrule, thus positioning your column too far from the detector. Same for the injector. This can happen to anyone. Always good to quickly check your installation distance visually or with a measurement tool after the ferrule has tightened.

I wouldn't think the change in injector type is responsible for your observations but it's possible. Your old system had a PTV injector, which has more capabilities than the standard split/splitless injector you have now; however, for hot splitless injection your current liner has more internal volume and thus can accept a higher injection volume without backflash, at the same temperature and pressure, than your previous liner. Ensure the installation distance into your 1177 injector is 37.5mm (double check this in the manual, but it's what I'm recalling) measured from the back of the nut.

Considering you have more internal volume in the liner than before, you might consider a pressure pulse (perhaps 5-10 psi above head pressure observed at the initial oven temperature) for say 0.25 minutes to help drive components out of the injector faster.

Good luck! Method development is fun... Although you're not quite to the performance you were on your older system, you're really not all that far off. Hopefully soon you'll be there.

[Ultimate Science 101]

After reading your posts again, I see you're running with a 50:1 split and not splitless. Not so sure the pressure pulse would make a huge difference with your fairly high split value, but I suppose no harm to try 1 injection.

I would definitely consider trying several types of liners designed for split injection to see if this improves performance. Make sure the wool in your new focus liner has the same deactivation chemistry as used in your former 1078 injector. You might also consider a carbofrit liner.

Is the autosampler also programmed in a similar way as done with your former system? Air gaps? Needle residence time? Plunger speed?

Just wanting to ensure even small details are identical to what was used before.