GC/FID Ghost Peak Troubleshooting

[dpatlak]

My lab has recently purchased a Perkin Elmer Clarus 580 GC/FID system with a TurboMatrix 110 headspace autosampler for the analysis of ethanol and other volatiles in whole blood for DUI cases. It's a dual-channel system, running PE Elite BAC 1 & 2 Advantage columns.

For the past few months, I have been seeing some sort of noise or contamination peaks appearing intermittently on both columns. The issue is not consistently reproducible, i.e. it does not occur in every sample, or in any specific pattern (that I can see). I have changed carrier gas bottles (UHP He from Airgas), standards (ethanol controls and calibrants from Cerilliant), and run various blanks (air, milli-Q H2O) in an attempt to determine the source. The blanks were always perfectly clean, and the gas and standards had no effect. I have also tried incremental changes to nearly all of the headspace and GC parameters to see if anything has an effect; it seemed as though the HS needle withdrawal time and vial venting option did change the ghost peaks somewhat, but did not eliminate them entirely.

Through our service contract with Perkin Elmer, a technician has been out multiple times to address the issue, with no luck. The first visit he replaced the HS needle, the injection tower assembly, the gas vent trap, glass-lined tube connector, deactivated silica transfer line, injection liner, and injection septum; chromatography initially looked better, but the ghost peaks soon returned. The second visit, he replaced both solenoid valves, the PPC control board, and the PPC assembly itself. Again, initial chromatography looked promising but as soon as I ran a test batch through the system, the peaks were back.

My hope is that someone here has seen this before, or dealt with a similar situation and can shed some light on the situation. I am out of my depth with this instrument, and the issue has been run up the chain of command at Perkin Elmer with no results. Thanks in advance!

HS Parameters:
Equilibration time: 15.00 min
Thermostat temp: 70˚C
Pressurization time: 1.0 min
Vial pressure: 21 psi
Injection duration: 0.02 min
Needle temp: 100˚C
Transfer line temp: 150˚C
Transfer line: 0.32 mm ID fused silica tubing
Withdrawal time: 0.10 min
Vial vent: Off

GC Parameters:
Injection temp: 200˚C
Split flow: 10 ml/min
Oven program: 2.6 min hold at 40˚C; +45˚C/min to 130˚C; Hold at 130˚C
Carrier gas: He
Carrier program: Hold at 16 psi
Total acquisition time: 6.85 min
FID air flow: 400 ml/min
FID H2 flow: 40 ml/min
FID temp: 200˚C
Makeup gas: He
Makeup flow: 45 ml/min

Some examples from the latest batch:

Blank sample, t-butanol surrogate and n-propanol internal standard (this is how it should look!):


Aqueous 0.08 g/100 ml control, replicate 1
(Column A):


(Column B):


The second replicate vial of the same standard \
(Column A):


(Column B):


Close-ups of the ghost peaks:


Ethanol-free whole blood sample:


Close-up of a 0.20 g/100 ml ethanol standard. Note the repeating pattern in this one; makes me think it's some sort of multiple-injection, rather than true contamination:


A high-ethanol whole blood sample, with no extraneous peaks:


Apologies for the long post; please let me know if you need more information. Thanks for the help!

[Peter Apps]

Tricky one ! and that repeating pattern is odd.

Carryover, contamination or multiple injections ?

If it is carryover you will see smaller peaks after the whole setup has been left hot with all gasses flowing overnight - run a blank with water first, then some standards and samples and see whether the first run is clean and what precedes the reappearance of the peaks.

Contamination can come from the carrier gas - do you have a scrubber just upstream of the GC ? Changing cylinders will not eliminate dirt that is coming from the gas lines. Dirty vials and septa are also a possibility (and are the most likely to generate an intermittent problem). Bake some vials overnight at 200C in a GC oven, try a new batch of septa, make sure that finger dirt is not getting onto anything that the samples touch.

I have never used a Perkin Elmer headspacer, so I do not know if they are vulnerable to sending multiple pulses of headspace to the GC. If that is happening I would expect it to change if you change pressures and flows.

Peter

[dpatlak]

Thank you for the reply!

My thought is, not carryover because some blanks run immediately after high-ethanol samples look fine, and vice-versa.

Contamination is definitely still a possibility. We do not have a scrubber on the He line. It is split to supply the HS and GC with carrier gas; the main carrier path travels through the HS and transfer line to the GC, and the line supplying the GC provides makeup gas for the FID.

I have tried various lots of vials and septa, as well as cutting up a septum and running it inside a vial to ascertain septum bleed. There were a few small peaks, but nothing like what is seen above.

The odd part about this is that the system was completely fine for ~6 months during the validation phase. Great chromatography, perfect quantitation, etc. The noise started appearing in December or January and has persisted ever since, which makes me thing it is something in the system getting contaminated or worn out.

[Peter Apps]

Invest in a gas scrubber and put it just upstream of where the line divides to GC and headspacer - I suspect that you had a cylinder of dirty gas, or dirt got into the line during a cylinder change, and the lines are now contaminated. It will take a while for your GC and headspacer plumbing to bleed clean. Why this does not affect every run is a puzzle, but it is the most likely problem in current evidence.

Has anyone been using bubbly liquids to check for leaks anywhere ?

Helium makes a very poor make-up gas - it is not viscous enough to slow diffusion in the flame. I doubt you will see a difference in performance if you shut off the makeup gas.

Peter

[GOM]

Hi

Whilst I can't help you in this case - please don't apologise for your long post.

It is a model example to others of a post that gives as much information as you can, together with example chromatograms - thank you



I hope that you solve your problem

Kind regards

Ralph

[uzman]

It seems the noise stops after the last peak - n-propanol !