The instrument is GC-MS with autosampler equipped with S B Zebron column (5% Phenyl-Arylene, 95% Dimethylpolysiloxane 30m*0.25mm*0.25micron), that should be very similar to the SPB 5 (5% Phenyl 95% dimethylpolysiloxane).
Target is evaluation of benzene concentration in water before and after adsorption tests: 250 ml of water and benzene (50-200ppm in water) are placed in vial with 750 ml of methanol and toluene (100 ppm in methanol) as internal standard. 1 microliter is injected in the column. Best program evaluated is: split ratio=10, flowrate (He) 1.5 ml/min, 1.8 min at 50C then ramp of 25C/min to 160. You can see the benzene peak well shaped at minute 3 and 1 minute later the toluene peak.
Ratios benzene/toluene of different consecutive analyses from same sample is sometimes very different (30%). Between each analysis every time an injection with methanol (2 min at 50 then up to 200 C) provides to clean the column.
Excluding leaks and any type of instrument problems, how could be explained this difference? I could also try to increase the split ratio, but I can't believe that this is the cause of a such strong difference.
Are the peaks being integrated the same way each time? Magnify up the base line area of the benzene and Toluene peaks and make sure they are being integrated in exactly the same manner. Try manual integration where you manually draw in the best base line and drop a perpendicular to this at the same place between the 2 peaks.
Are there sufficient data points over the peaks to accurately determine their areas? Each peak should have at least 10 data points along its integrated area for accurate determination. If not increase the data collection rate. Elution at 3 minutes on the column you mention should give a very sharp peak suggesting a high data collection rate is required. What are the mass spec parameters? If you are doing a full scan try only scanning from 75 to 95 amu and see if this improves the results.
Hi Andrea,
What about the benzene/toluene ratios in your standards, are they within a few %?
Did you do the duplo analysis of your samples from the same GC-vial or did you fill several GC-vials with the same sample?
Alex
Dear Bruce and Alex,
thank you. All about integration and data points is well setted. I both do double analyses on same vial (stirring before) and I also fill several vials with the same sample. By increasing equilibration time and shaking vials before every analyses I was able to have good repeatability for some consecutive analyses (ratios: 0.68,0.70,0.69,0.69,0.70) but then you can have sometimes strange results (0.90). What I saw is that you've sometimes too big difference in the total peak areas (sometimes they are 50%more then the previous analysis), with or without a good ratio between them. The autosampler take exactly one microliter, I changed septa and I check for leaks just close to the septa. I should check for leaks in other points and I'll do it, but I'm pretty sure that I will no find leaks. Do you have any suggestion?
Check the syringe in the autosampler for a worn plunger or worn plunger tip. Is the syringe barrel clean? If not it should be cleaned or replaced. Also check the septum and injector purge/vent lines for cleanliness and check if there is a chemical trap in the vent line(s) that might need to be changed.
Hi Andrea,
A possible cause of the variability may be due the injecting water samples. The expansion volume of water is large. An injection of 1 ul into a port at 250 C with a column head pressure of 15 psi will expand to about 1100 ul of vapor. A typical 4 mmid liner has about 1000 ul of volume. A 2mmid liner has about 250 ul of internal volume. A good rule of thumb is to keep the expansion volume to less than 70% of the liner volume. Most common split/split less port liners are 4mm id or less, so the sample vapor is expanding out of the liner into the outer parts of the injection port where a septum purge flow will carry sample out of the port, sample can condense in the unheated portions of the port, expand back into the carrier gas line and condense there. All this can cause reproducibility problems. Water can get a reputation of being unsuitable as a solvent for GC analysis because proper adjustments are not made for the large expansion volume.
Agilent has a series of free calculators on the www.Agilent.com website that are quite useful for this type of calculation.
Good luck
- Badges
- Science topic
- Similar topics
- Philosophy
- Esthetics
- Interpretation