HS/GC stops collecting data on vial 8 and i dont know why

[MestizoJoe]

I have this strange problem. Let me first state that I am using a G1888A headspace sampler with an Agilent 6890 GC.

So, I ran a sample set that had 13 vials twice. Both times the first seven vials are completed and the eighth vial showed 'missing data' in empower. Also, the sample set was:

vial 1 Air
vial 2 blank
vial 3-7 standard
vial 8 blank
vial 9 sample endogenous
vial 10 spiked sample
vial 11 blank
vial 12 standard
vial 13 blank

For vial 6 & 7 the chromatogram does not show any standard peaks.

The run time is 20min but the vial equilibration time in headspace is 40min.

GC oven: 40C (2min) ---(10C/min)---->220 (0min)

This exact situation occurred twice.

So far what I've tried to remedy the problem is changing the GC oven program to

40C (2min) ---(10C/min)---->220 (2min) --------(20C/min)----->40C

to see if the problem was that the GC was not ready when the headspace injected.

I'll post if this worked later. Any help would be appreciated.

[MestizoJoe]

I found that in the headspace parameters I had set the GC cycle time to 25min, with vial equilibration time at 40min as per the method (which is probably a mistake in the method).

The headspace injected every 25 minutes even though the GC was still analyzing an injection. I didn't see multiple injections in the column (i.e. if the main peak is at 3min i didn't see a new peak at 28 min). I think that's because even though the headspace sample vial was punctured the sample was not introduced into the GC inlet.

Anyway, it's working now. I just have to make sure GC cycle time is about 3-4 minutes longer than the time required to perform all other operations in the headspace (if the vial equil time is 40min, the GC cycle should be 43min or more).

[krickos]

Anyway, it's working now. I just have to make sure GC cycle time is about 3-4 minutes longer than the time required to perform all other operations in the headspace (if the vial equil time is 40min, the GC cycle should be 43min or more).

No, something else is wrong. Cycle time does not has to be longer than vial equilibration time. GC cycle should be: Total runtime of analysis+ ccooling time to start temp+set equlibration time at start temp+ "additional safty factor time".

I do not have empower but same instrument as you, essentially all of our methods has shorter GC cycle times than the vial equilibration time with exceptions of a few new very fast methods.

In fact yout initial GC cycle time of 25 mins seems to be the issue as the run time is 20mins so 25min is too short but do not necessary have to be longer than vial equilibration.

So if I guess a bit: 20 runtime+ 8 min cooling time +´2 equilibration at start temp+5min safty factor= 35mins. So yes, your headspace made dual needle punctures of vials in some runs.

Your headspace is technically competent enough to handle more than one vial in headspace oven otherwise why have more than one vial position in HS oven

Though at times one has to fiddle a bit with cycle times but usually due to some operations colliding time wise but then we talk about just a few minutes +/-.

[Peter Apps]

Two tips -

you can set the headspace sampler to wait until the GC signals ready before it injects - if I recall correctly this can mean that the sample sits in the valve loop so it is a measure of last resort.

If your GC cycle time and sample equilibrium time have common multiples then sooner or later a sample injection will clash with another GC operation; simple example - GC cycle of 15 min isothermal so no colling time, vial equilibrium 45 min, the third (or fourth) vial will inject while the GC is getting itself organised for the next run. Use times that do not divide into one another.

Peter

[MestizoJoe]

This is in response to krickos

Of course GC cycle time was the issue. That’s the parameter I changed to get the samples to run.

You seem to disagree with what I think the reason for that is.

When the GC cycle time was 25min and the vial equilibration time was 40min, I witnessed the first vial go into the headspace oven for 40 minutes and the injection proceeded with no problems.

However, after the first vial came out of the headspace oven and the next vial went into it, about 23 minutes went by and the vial was injected again; the vial did not equilibrate for 40min. So your explanation does not correlate to what I observed since the next injection didn’t happen in 25min cycle + cooling + etc = 35min. It happened in 23.4 or so minutes.

I haven’t figured out how to make this work without making the GC cycle time longer than the vial equilibration time. Maybe the method is a “new fast methodâ€

[GaryR]

How do you figure that the GC cycle time was the issue, when by your own observation vial 2 only thermostatted for 23 minutes?
Just because you changed the parameter and it ran does not mean that was the problem.
Krickos’ logic is sound and backed up by experience, and not just his own.
If you’re going to ask for help show a bit of gratitude when help is offered. Just because it’s not what you want to hear doesn’t mean that it is not correct.
I don't believe there was any intent at being condescending, english is not the first language of all users here so simplicity and clarity can do a lot to overcome language issues.

[MestizoJoe]

You're right GaryR, I should show more gratitude. I apologize to krickos and anyone else I may have offended.

I now understand what krickos was trying to say about vials in the oven. I didn't see more than one vial go into the oven. It only went in one at a time.

So, the first one went in for 40 min and injected. The next one only goes in for 23 or so minutes because the next injection is only 20 min away. So, my problem seems to be that the vials are not being introduced into the oven when they ought to be.

And now I feel like a big bonehead. Once again, I apologize.

[Peter Apps]

You need to check all the settings on the headspacer - as well as the wait for GC ready there is also a setting that selects for constant equilibrium time if I recall correctly.

The underlying issue might be an incompatability between the Empower and the headspacer - even Chemstation had some issues with the instrument that I had. The G1888 is an Agilent clone of a Dani instrument and communication is not always what it might be. Can you run the headspacer off its front panel and just have the GC send it start and ready signals ?

Peter

[krickos]

You're right GaryR, I should show more gratitude. I apologize to krickos and anyone else I may have offended.
.

Apology accepted, myself I could also have been a bit clearer and left out the smiles.

Gonna pop over to another of our sites today where the QC department have started to shift over to Empower on their Agilent GCs and headspaces, with luck I will find some time to check the communication set up in Empower or if there is some specific config file that has to be taken into account.

[GaryR]

I’ve been giving it some thought....
And just reading the above responses:
It does sound as if you have a hardware communication issue.
The GC should only start an injection when it is triggered by the headspace.
But if the HS injects while the GC is acquiring, the sample goes straight out the vent, as you see by no mid-run spikes.

And your HS should be smart enough to be able to take the vials into the oven at the correct intervals to be pressurised and ready to inject after the GC has had time to come back to initial temp (cycle time).
And it should also wait for the GC 'Ready' signal before it injects. As Peter pointed out the headspace will wait with the sample in the injector loop, pressurised, until the GC is ready.

So, it sounds to me as if your headspace is not waiting for the GC to come ready before it injects. This is why extending the GC cycle time setting for the headspace has solved (masked) your problem. As Krickos said, the cycle time should be around 38 minutes (with 5 min safety margin).

- So you're right that the method is incorrect, but not that the thermostatting time is incorrect, but that the cycle time is too short.

(And to be realistic, going from 220C to 40C in 5 minutes is do-able, so a cycle time of 25 min will probably be OK in that the sample wouldn't be sitting in the loop all that long before being injected)

So, your injection should have waited for the GC 'Ready' with the sample in the loop.

But, it didn't.
Hang on.... Just noticed it also didn’t thermostat your subsequent vials for the correct time, so maybe you also have a firmware issue with your HS as well.

All signs point to your HS.

Reboot everthing.
Double check you set up. Try again.
Oh, and a bit more info could be helpful.
How long have you been running this method, is it something new, or new hardware or software?
Does the system work for other analysis?
Oh, and not dissin’ . Just trying to help with the troubleshooting!
Let us know how it goes.

[krickos]

Hi again

Well sadly the guy I hoped to met had sick kids at were at home.

But there are some basic communication settings that you should check, if you use chemstation newer softwares with the above instrumentation my experiance is that the communication through the method is "seemless".

But when you use empower one has to step back in development it seems.

I found the following settings on my collegues lab under the headspacepart:

-Feedback signal: Set to "ignore" (other possible settings are wait & abort)
-Handshake mode: 6890/6850

It is possible I think that the feedback signal can be set wrong. If using ignore the HS should ignore the ready signal from the GC after colling down and continue to equilibrate vial and inject when it is ready.
If you have set wait it is possible that the ready signal from the GC triggers an injection prematurely.
Sorry a bit speculative but best guess I could come up with.

Have a nice week end everyone.