non linear injector

[ym3142]

Hi

I need some real expert's help.

I have Agilent GC 7890A. Water Quattro GC MS and CTC sampler.

Under the same conditions when I inject 1 uL the peak area was 3000 and when I inject 1.5 uL the peak area 15000(yes, 15000). I have a splitless liner(900uL). I tried both 265 C and 310 C inlet temperature; and 250 and 310 source and interface temperature. I am using SIR (209.2 for this trichloroanisole)

Sample is trichloroanisole 100pg/uL in acetone.

If you know possible reason please let me know.

[sav]

Hi,
Check:
1. Linearity of ms response.
2. Syringe and AS.
3. Soft (if you use AS): syringe size etc.
4. Split (if you use split).
5. Splitless (time, flow of eluent etc.).

Best regards, Alex.

[Peter Apps]

There are all sorts of possible reasons but the most likely is that vaporizing inlets are terribly sensitive to flow disruption and pressure pulses as the volatile solvent evaporates at high temperatures. I would not expect the peak area to be linear against injection volume. The solution to the problem is simple - always inject the same volume.

Peter

[ym3142]

Peter, Thanks for your suggestion, which is nice to know.

Alex, thanks for your suggestions too. but could you please instruct on what to check about syringe and AS? And What is AS?

I am using splitless and having the following:
Inlet mode splitless
inlet type split/splitless
temp 265C
pulse time 0.00 min
split ratio 0.1
split flow 0.1 ml/min
pulse pressure 0.0 psi
purge flow 60 ml/min
purge time 0.2 min
gas saver disabled
vent pressure 0.0 psi
vent flow 0.00 ml/min
vent time 0.00 min
septum purge flow 3 ml/min
sampling end time 0.00 min
initial time 0.00 min
.....

Should I try pulse-splitless?
do I need more purge or vent?

Thanks and happy Thanksgiving!
Jim

[Peter Apps]

Hi Jim

Increasing the purge time to at least 30s might help, but I still do not think that you really have a problem - why would you change the injection volume from 1 ul to 1.5 ul instead of always using one or the other ?

Peter

[ym3142]

Thanks, Pete.

I agree with you if my system is normal.

However, as you may guess, I just took over the GC MS system . I am not sure if the system works as designed/intended. I am worried there may be some hardware installation mistake or parameter setting defects.

Again, your opinion is that in the GC MS world 1ul gave signal 5000 and 1.5 uL gave 15000 is "normal" and "OK". Is this right?

By the way, thanks a lot for your expertise helping many of us here over the years in this Forum.

Happy Thanksgiving.
Jim

[sav]

AS - AutoSampler.
Check quality of syringe (partly obstruction of needle etc.).
If column dimensions is 30m*0.25mm and He flow is 1ml/min, then try this cond.:
Split: 1:20-1:50
Splitless time: 1-2min
Temp. of inj.: 230-250oC
Liner without wool.
Oven temp. (initial): 30-40oC

Best wishes, Alex

P.S.: I hope that you don't want to make calibration with different inj. volumes.

[ym3142]

Alex

Thanks for your advice.

I will check the syringe.

I am using splitless. Do you suggest me to change to split mode? And, any reason that the conditions you suggested solve my problem of non-linear?

Happy ThanksGiving!

Jim

[sav]

Do you suggest me to change to split mode?

No. It's splitLESS (split is open>split is closed>injection>delay time>split is opened).

And, any reason that the conditions you suggested solve my problem of non-linear?

Just try it (when all points will be checked). See theory of split/splitless injection in GC (K. Grob etc.).

Are you check linearity of ms response (0.02-1ppm of TCA in SIR mode)?

[Peter Apps]

Thanks, Pete.

I agree with you if my system is normal. With the evidence so far, the only thing that is not normal is your comparison of peak areas with different injection volumes - this is not usual practice in GC, for reasons that your tests make glaringly obvious - flash vapourizing inlets are not robust to changes in sample volume

However, as you may guess, I just took over the GC MS system . I am not sure if the system works as designed/intended. I am worried there may be some hardware installation mistake or parameter setting defects. There may be something wrong with the hardware or the setup, but what you are doing is not the kind of test that will throw much light on it

Again, your opinion is that in the GC MS world 1ul gave signal 5000 and 1.5 uL gave 15000 is "normal" and "OK". Is this right? I would guess that it might well be normal (linearity of peak area vs injection volume is not a performance parameter that is usually tested in GC - in fact I cannot recall ever coming across it), but it is not OK - it points to a serious deficiency in the design of the inlet that is on nealry every GC in the world !

Is this effect consistent - how many injections have you done at each volume ?

By the way, thanks a lot for your expertise helping many of us here over the years in this Forum.no problem

Happy Thanksgiving.
Jim

Peter

[ym3142]

Thanks Pete and Alex.

Here is what I found: when I extend the purge time the response was increased dramatically for 1uL injection, and so that the response difference between 1 and 1.5 uL is much smaller, the best relative response is about 1.5(1uL 5000 and 1.5uL 11250).

[Peter Apps]

Thanks for the feedback.

Peter

[Bigbear]

The signal is the signal. You are comparing apples to oranges.
If you are worried about the injector setup do a simple experiment. Shoot 10 or so reps ( one injection/vial) and check the %RSD of what ever area you get. If it is < 10% I would doubt you have a problem.

[mckrause]

1. You have the wrong liner. 900 uL internal volume is too small.

2. You are running splitless - you will get huge (HUGE) area count swings in splitless. If you're trying to develop a relationship between mass injected and area by varying the injection volume go to a low level split injection - use 3-4 mL split flow.

3. The HP (Agilent) injector has had problems since its inception. Welcome to a really poor splitless injector. If you're insistent in using splitless injections (I'd recommend highly against it, if you can avoid it, since low level split works so much better) then you probably should seriously think about either incorporating the Restek modifications into your system or switching out to something like the SGE injector.

[ym3142]

Thanks very much, McKrause.

Merry Christmas!!!

Could you please further elaborate the following?

1. You have the wrong liner. 900 uL internal volume is too small.
If it is true how does this rationalize that 2uL provides disproportionally larger signal than 1 uL. When I image 2uL will be less than double of 1uL signal, right?

2. You are running splitless - you will get huge (HUGE) area count swings in splitless. why? If you're trying to develop a relationship between mass injected and area by varying the injection volume go to a low level split injection - use 3-4 mL split flow.

3. The HP (Agilent) injector has had problems since its inception. Welcome to a really poor splitless injector. If you're insistent in using splitless injections (I'd recommend highly against it, if you can avoid it, since low level split works so much better) then you probably should seriously think about either incorporating the Restek modifications into your system Before we seriously consider to buy different inlet system we have to know the reason. Would you mind providing some literature? or switching out to something like the SGE injector. So why the injector here solve the inlet problem?