FID Baseline Drift and Noise

[deanmc]

I have observed strange baseline behaviour (drift and noise) recently. Full info is below if you have the time to read it.

For your information, the specs of the GC and analysis method are:

GC: GC17-A
Injector: AOC-20i autoinjector
Injection volume: 0.6uL
Injection System: Split injection at 380 degC
Precolumn: Restek MXT Biodiesel TG precolumn, 2m x 0.53mmID x 0.16um
Column: Restek MXT Biodiesel TG column, 15m x 0.32mmID x 10um
Column temperature program: 50-380degC
Column pressure program: 20-65kPa
Detector: FID at 400degC

Normally, baseline is flat with a slight upward drift towards the end of the program as the temperature rises. Yesterday, I firstly noticed the baseline climbing at a considerable rate throughout the course of each sample and the overall analysis. Then ,after about six samples, the baseline passed through a gentle crest mid-analysis, and began to drift downwards. This downward drift continued throughout the remaining six samples. The starting point for the baseline was at the same level as the end point for the previous sample. Despite the drift, there was no noise.

In case of buildup in the column, I tried baking at 410C for two, one hour periods. Column pressure constant at 20kPa. Both times, the chromatogram had a small hump at the start, drifted downward as temp increased, before going through a trough, rising slightly and levelling out at constant temp. No peaks appeared. I retried regular analysis with solvent only, and saw the same trend.

I then cleaned the glass insert of the injector, replaced the septum, silica wool, and the graphite ferrules at the column ends. I don't have replacement ferrules for the injector or connector between the two columns so they are unchanged.

Since then, the baseline has been very erratic and noisy. Ran at 50C for 2h, baseline initially noise free but considerable noise appeared after 1/2-1h. Increased temp gradually to 250C, saw a very sharp, large increase in baseline (up to 13mV, if that is of use). After temp ramping stopped and column steady at 250C, saw a very sharp decrease in baseline, back to about the same level as before ramping, and it continued to drift downward for the next 1/2h that I ran it for. All the while, a very noisy signal.

Do you know what could be the cause/s of this change in baseline behaviour?

Thanks for your time,
Dean

[Peter Apps]

Best guesses - the huge hump that gradually pbaked out was some very heavy contaminant the slowly migrated through the system. The new noise is a bit of graphite ferrule in the detector from when you re-did the connection.

And a tip - put in a clean liner before you bake the column, otherwise you are just throwing more muck onto the column during the bakeout.

Peter

[Alexandre]

Dear Dean,

Looks like it is from your samples. I operated this Shimadzu model many years back. From my memory you need to put septum sweep flow right. I used to set it a little bit more than recommended and pls. check the actual flow with meter.

Keep feeding pressures of injector on higher side. After sample was deposited on column (2-3 min), increase split to wash all rubbish away from injector.

If your problem persists you may need to wash all lines that connected to injector port, no matter in or out.

Can you reduce the amount of sample on column by analysis of biodiesel solution?

[Bigbear]

Don't over look the split vent trap. You may also want to consider replacing the Cu tubing connecting the trap and inlet. Cut the new one as close to the old one in length ( somewhat critical). Cleaning it is not very effective.

[Johnny Rod]

Probably worth cleaning the FID (dismantle and sonicate in methanol) and maybe replacing the jet, as some of this dirt could well be on there. Check any insulators around the electrometer/collector for dirt/scoring as well.

[deanmc]

It appears the problem was firstly a buildup of heavy components working their way out, and then with a leaking column to column connector, as far as I could tell. There was some residue on the fastening screw of the connector. Removed pre-column and now using main column only.

Performed a test (according to Shimadzu manual) to see if split trap was blocked but not a problem.

On the point about FID, although I did not clean or replace any of these components, I noticed the FID signal is very sensitive to length of column inserted into FID. I had to cut column 1mm shorter than Shimadzu jig to avoid noise.

Thanks for your advice.
Dean

[Peter Apps]

Hi Dean

Thanks for the feedback - interesting point about column insertion depth.

Peter

[Johnny Rod]

Could be a chunk of something in the FID jet in that case.