I think the co-solvent which you are using for ethanol calibration is water and that is creating problem of getting non linear chromatogram. Because water boils at 100deg C and your inlet temperature is 175deg C. When any solvent is injected into a hot inlet it is converted to a gas and the resultant gas volume is much larger than the initial volume. Resultant gas volume of water + ethanol is varied at the inlet so you might be getting non linear chromatograms.

If that is not the case then check your sample preparation. Do gravimetric preparation of the calibration standards and then check for linearity.

Please see this discussion: https://www.researchgate.net/post/How_is_it_that_when_monitoring_a_reaction_using_GC_the_product_peak_area_increases_as_expected_while_the_reactant_peak_area_fluctuates

my first question will be whether you are using autosampler or its a manual injection system. If the area of the response is not reproducible in GC system, first thing to be checked is the injection volume..u can do one thing , inject same volume of sample five times or more and chek the std dev in the response area. If it is fine then we can go for other reasons..

Oven temperature program should be change. As mentioned by Shweta boiling point of water is 100 degree celcius but oven started lower than that point. Hence, water vapor return to liquid state and then changed to gas state with increase of temperature. I recommend start oven temperature higher than 100 degree celcius and increase slower rate or maintain constant temperature.

Bytheway, does there no problems in column? Water is extremely polar and could make damage on column coating matters.

At 175 C the vapor volume of a 1 ul injection of water is around 1000 ul which is more than many injection liners can tolerate, so you will get flashback into the carrier gas lines to the injector. This can cause a lot of problems, including carryover and imprecision. As Kae Kwon said, you need a higher oven temperature, but also reduce the injection volume. A lot of analysts are afraid of injecting water, but most modern bonded phase columns can tolerate water without degradation. However, you do not mention what type of capillary column you are using. It should be a polar phase to give good retention and peak shape, such as a bonded Carbowax. It could also be done on a PLOT column such as Rt-Q-BOND. Also, the ethanol must elute well separated from the water. If they co-elute then the disturbance of the FID will yield very poor precision. Check the literature of your column supplier for the application.

I think the water is the origin of your problems, due to reasons well explained in the other comments. By the way you're injectings solutions from 500 to 1.500 ppm of MetOH. Could these concentrations be too high for that capilar column and a FID detector?. You don't mention the split ratio (in case you're working in split mode). You could try with a 1 microlitre injection and a split ratio of 1:40, in order to minimize the volumen of water vapour injected, as well as use an internal standard (specially if you use manual injection). But as the other comments indicated, the main problem is the water itself....it's not a proper solvent in classic GC due to the vapor volume, and the extreme polarity it presents.

As Salvador commented, "water itself....it's not a proper solvent in classic GC". It definitely is not the ideal GC solvent, however, we don't always get to choose our sample matrix! I have analyzed samples such as these in the past, as well as other samples containing high water content, with very good results. It just has to be done properly, and with all the comments above, Golla should have enough information to be able to establish an accurate analysis.

My first suggestion is that check your solvent peak is coming in the same retention time and also area of the solvent peak. If area of the solvent peak is high then other peaks also separates in different retention time. use always split injection usually 1:40 or 1:100 (if concentrated).

As Robert and Salvador commented water is not a proper solvent for GC injection. As mentioned by Shweta boiling point of water is 100 degree Celsius but oven started at 70 degree C. Hence, water vapor return to liquid state and then changed to gas state with increase of temperature. I recommend start oven temperature higher than 110 degree celcius with 5 degree per minute. With the manual injection the retention time always changes, since you have to inject the sample and then giving start. Check your polarity of the column, since water is a highly polar solvents. the concentration of sample has to reduce since GC is sensitive than HPLC and it detects in ppm.

As mentioned above injecting water into GC is probably a bad idea and could very well be the cause of your problems as well as quick degradation of your column. I would suggest a simple organic extract as was described in the attached paper. Since published we have tested the method on Agilent 5890 GC-FID and it works very well.

Let me know if you need more details.

Yoram

First of all, water is not a good GC solvent. However if your research requires that no point of talking about the solvent. I taught a similar undergrad lab on 'determination of ethanol in beer" which is similar to your case. If you are only looking for EtOH, I don't understand the temperature program you use. Injection temperature of 175 is ok. But since the bpt of EtOH is ~78, you can keep the oven at ~65 until your peak is observed and then increase. It come quick. Very high temperatures (higher than bpt) can give you that kind of problems time to time. Since your injection volume is 1 uL, water will not have a very bad effect. Bur for your problem, (1) I strongly recommend to use an internal standard (1-propanol will be the prime candidate) to overcome errors due to injection volumes. (2) Are you doing split or splitless, if you do splitless, may be your concentrations are too high and initial dilutions will be necessary. If you use a narrow inlet liner, high amount of water vapor at 175 will also create problems. If you are doing split, just play with split ratios. (And also, if you are doing a manual injection, make sure you inject in a flash. Slow injections cause bad chromatograms as your solvent band is wider at start). Hope this will help. Good luck !

Your concentrations are quite high. You should use a headspace injection rather than a direct aqueous injection. Split the injection 20:1 if you make a gas phase injection, inject approximately 100 uL of the headspace. You will get a nice, clean ethanol peak.