-
- Posts: 72
- Joined: Sat Mar 28, 2009 12:01 am
Apologies for the long message
We are determining ethanol in wine using GC.
We have done this lab for many years but I have never found the peak shapes to be great so decided to try to improve things.
The standards are made up in water to be between 0.1% to 0.5% (v/v) ethanol with 0.5% 1-propanol as internal standard.
I inject 0.2 ul (to avoid the well-known water expansions issues) using a 10 ul syringe. Would like to inject more but that's the joy of water. Also don't want to switch from water to another solvent.
Conditions are
Inlet 250C, split 100:1. pressure 14 psi
Column is HP-Innowax 30m x 320 um x 0.25 um.
Flow is 2.9 ml/min (46 cm/sec)
Oven is 75C isothermal
but is left overnight at 150C to flush column and possibly prevent build up of water on the column.
Detector is FID
Agilent Autosampler.
Analyte peaks elute within 2 minutes
A standard containing (0.3% ethanol and 0.5% 1-propanol) was used throughout these tests.
Results
Using a deactivated liner with glass wool such as the Agilent 5183-4647.
Precision of peak areas expressed (as %CV) around 1% for 10 repeats
Peak shapes show a lot of tailing with USP tailing factors (at 5%) > 1.5 and get worse as more injections are done.
Using a liner without glass wool
Precision of peak areas (expressed as %CV) around 25% for 10 repeats. Area ratio precision is good though - %CV is around 3% for 10 repeats
Peak shapes are excellent with USP tailing factor (at 5%) consistently close to 1.
- this was also the case when I used a Agilent 5183-4647 liner that I had removed the glass wool from.
In conclusion if I have a liner without glass wool I get great peak shape but low peak precision while if I have a liner with glass wool I get tailing peaks but much better peak precision.
Any suggestions what I might be able to try to obtain the Holly Grail of good peak shape and good peak area precision
Am I missing anything obvious?
To try to solve this issue, I replaced the gold seal, and split vent trap recently and also have trimmed the column. It seems though that the liner is the issue.
Thanks
Kevin