Chromatography Forum

Hello,

I have a quite sensitive substance (TBHP in toluene) and analyze it by GC-MS at different temperatures.
It should be stable below 80°C.
If I use the same method only changing temperature or heating rate, I get sometimes different signal areas
although I am below 80°C.
Sometimes there is an oxidation product (benzaldehyde), but instead of becoming more at higher temperatures
with decreasing TBHP-area, it is the other way round.
The TBHP signal gets higher and simultaneously does the area of benzaldehyde.
How is this possible, if the measurement is done with the same method by an autosampler?

May there be any temperature dependence of the MS detector signal?
Why? The TIC is measured and the number of ions do not depend on the temperature.

Any idea?

Thanks

Thomas

Where are we talkign about 80 degrees? GC inlet,GC oven temperature, transfer line or ion source? You have oportunities for decompsition in each of these thermal zones. What type of injection are you making? (PTV, split, splitless, SPME...) And what line are you using and what stationary phase in the GC?

Hello,

thanks for your answer. I mean an oven/column temperature of 80°C.
ion source is 200°C. I tried several inlet conditions (PTV: a) starting at 67°C, after 3 min rise to 82°C, b) 150°C constant).
I use a stabilowax column (30 m, 0.25 mm).
Still more strange is, that we had a quite complicated column construction before (spl injector-column, switch with (restrictor- MS, mole sieve column-WLD) ), but were able to get equivalent areas (measured by MS) in the chromatogram from 70 to 90°C.
Now we have only a simple construction (PTV-column-MS) but the area is not reproducible. It changes very much from 60 to 90°C.

I have to say that I don't use an internal standard which is usually used in GC. It seemed not to be necessary yet.
Perhaps I should use one, but I am not sure, if the one I would take is appropriated.
The standard I would use should be stable, but the peak is not near to the one of the TBHP.
I heard that this would be favorable or did I misunderstand?
Anyway I don't see why the standard peak should be near the peak of the TBHP.
This does only make sense if the signal changes with temperature which should not be the case...
So the question rises again...

regards

Thomas

The MS signal should not be dependent on column temperature, but I can imagine that column temperature can influence peak shape, and therefore the area seen by the integrator.

Is there a reason why you could not go back to the split/splitless injector? PTV inlets can be a bit tricky - and it can even get to the liners and the seals in the inlet.

Hello,

I was a little bit confused how it is possible that the area of the peak is so strongly fluctuating.
But now I am sure, there is a problem, even if I have no idea what problem it is.
The substance I normally investigate is unstable so there may be a lot of reasons for problems.
But even the same experiments with a stable substance yield very different areas up to 6% error.
This is as we would evaluate the area manually by cutting it out of a paper or estimating it otherwise.
So this is not normal. Changing column temperature also gives different areas.
Last year with another configuration - splitless injector - column - MS there was no problem.

It is a good hint that the PTV may be a little bit tricky. The problem is we need it because we have
unstable compounds and want to be sure to safely vaporize them.
I am sure that somewhere some of my substance does not reach the column because
a problem at the MS interface should also cause a bad tuning file due to a high nitrogen level, vacuum
problems etc. which are not the case.
Do you have any idea which special problems are linked to the PTV and if they may be the reason?

Thanks

Thommy7571

Your injection port and oven temperature are below your solvent boiling point so the results will be very dependent on inlet conditions. Thus Don's comment about PTV inlets being tricky.

Are you running the carrier gas in constant pressure or constant flow mode?

If constant pressure, then you will have different flows, decreasing with increasing column temperature, which will cause the MS to become more sensitive on the runs with higher oven temperature(which is what I believe you are seeing).

As the flow goes lower, the vacuum will become greater thus the sensitivity of the MS increases. If you are using a constant head pressure in the inlet for every run, then that could be why you have greater response at higher oven temperatures.

Review your PTV inlet flow and pressure programme - it sounds as if you are turning on a purge flow close to the time at which the sample finishes evaporating and transferring to the column. Small differences in temperature and evaporation rate means that you are purging different amoiuhts each time instead of transferring everything to the column.

Peter

Are you running the carrier gas in constant pressure or constant flow mode?

If constant pressure, then you will have different flows, decreasing with increasing column temperature, which will cause the MS to become more sensitive on the runs with higher oven temperature(which is what I believe you are seeing).

As the flow goes lower, the vacuum will become greater thus the sensitivity of the MS increases. If you are using a constant head pressure in the inlet for every run, then that could be why you have greater response at higher oven temperatures.

I am afraid I don't use a constant pressure program but a constant velocity so that can't be the reason,
but I know the problem with a constant pressure of former configurations.

Thanks for your answer

Thommy7571

Review your PTV inlet flow and pressure programme - it sounds as if you are turning on a purge flow close to the time at which the sample finishes evaporating and transferring to the column. Small differences in temperature and evaporation rate means that you are purging different amoiuhts each time instead of transferring everything to the column.

Peter

I already varied several conditions - including the injection temperature and program and the column program - before, so I don't think it is the problem, I also used just a constant temperature for the injector. After discussing the subject finally with a specialist of SHIMADZU I got some other hints:
1) a problem with the syringe, which might not be set in correctly or with the plunger inside which may not go down completely.
I fixed the plunge at the position where it it completely down in the autosampler.
2) A problem with the liner. In fact we used the PTV liner - included in our configuration from the beginning but never used - as it was without knowing
anything. We just connected the column to the PTV injector - 2 weeks ago - and used it. Now we found out, that there is a quartz glass liner without
glass wool (which seems to be normal). Worse would have been a steel liner.
We use 1 µl sample and toluene as a solvent. I heard that 0.5 µl sample may be better in some cases but I don't exactly know when it will help us. I
am just trying with 0.5 µl because the liner is smaller than the splitless injector liner.
I wonder if there will be a problem with a 10 µl syringe injecting only 0.5 µl.
I also wonder where I can set the sample size, only at the GC?
As the PTV seems to work better with a low boiling solvent I want to change to pentane but I wonder how to prevent vaporization losses of the
sample and calibration problems due to this. Will an internal standard be sufficient?

Thanks for any help for the first trials with an PTV- injector...
What are the optimal conditions and parameters for the PTV injector to use it?

Thommy7571

It is usual in PTV to use a solvent with a much lower boiling point than your analyte's - but this is mainly so that you can vent the solvent before transferring the analyte to the column. In your case you use the PTV to avoid thermal stress on your analyte, so the solvent BP is not an issue.

Certainly using the wrong liner can cause the problems you see, but 1 ul is already a small volume for a PTV injection and I would not have expected a problem with sample running down the walls of the liner and being lost at the bottom.

What can happen though is that sample sprays out of the needle and straight through the liner. A needle with a side hole is a big help with this, or try varying the injection speed.

Peter

Hello,

thanks for your hint. I don't know if I mentioned it before, but we just built up the dual system
with two injectors columns and detectors about 2 weeks ago.
We have a good tuning spectrum but if there is no problem with the sample amount (1 µl) or the method or the smaller liner, I would
suspect that there is a problem with the connection between column - PTV-injector (or column - MS) although we used
a "jig" to have the right length of the column in the MS and in the injector.
We already use the highest injection speed and the lowest suction speed.
Although we have a good base line and a good tuning spectrum, may there be any problem of this type (reproducibility)
due to some installation error?

By the way even the specialists of SHIMADZU are running out of ideas by now and we need to solve this problem urgently.

Thanks

Thomas Willms

Hello,

We already use the highest injection speed


Thomas Willms

Hi Thomas

The highest speed is the one that is most likely to spray liquid sample down to the bottom of the injector. Try slowing it down.

A baffled liner will help a little bit, one with a very small amount of inert packing will also help.

Peter

Hello,

I finally got the message that there are not only quartz liners but also
others deactivated liners with glass wool.
Off course I will try the other liners now.
Not only because it has some glass wool and might work better in general but especially because it
is deactivated as the liners we used before. This is essential for our peroxides.
On the other hand - if everybody uses not activated quartz liners, I wonder if this can be the reason.
I mean the normal liner should work as well, at least for normal stable compounds, whether or not
they are polar as alcohols or not.
What do you think?

I try now 0.5 µl because I found out that in some experiments, the value was not accepted and modified.

According to the postings etc. that shouldn't be the problem but I will give it a try.

Any other idea concerning any installation error?

Thanks

Thomas Willms