Chromatography Forum

I work with a GC/FID instrument and I work in an industry which make adhesive cement formulations . As a part of our QC, we check for the concentration of the solvents present in the formulation.
I wanted some guidance on how to make the calibaration curves to be able to tell the unknown concentration.
for example acetone - lowest is around 4-5 highest is 95.
THF- 12 - 80.
MEK- 2- 40
CYH - 2- 30
and then there are solvent concentration which might be more than this which would be the unknown.
All these 4 solvents are usually present in a formulation.
I would be looking forward to some help on this.
Thank You
Prithi Rao.

If those numbers are concentration values, appropriate units MUST accompany them.

sorry, the units are in %.

Do I understand correctly ? You analyse mixture of 4 solvents.
Concentration is expressed in % (mass or volume ?)
How do you introduce this mixture to the inlet, in neat form or diluted in another solvent ?
Is you injection split ?

I guess you have already chromatography method with good resolution of components, and you need help with calibration curve ?

I donot dilute it just introduce it as is. Injection is split.I weigh the concentration in grams.

I would dissolve this solvent mixture in another solvent (eg. carbon disulfide) and then analyse it.

Before quantity analyses calibration for all 4 components have to be done.
Simply prepare series of standard solution of all 4 components in CS2 (known amounts in mg/ml).
Analyse them and calibrate.

This aproach seems the simplest (doesn't mean the best).

Results in mg/ml can be easily calculated to % (m/m) knowing dissolution factor in CS2.

-----------

Or the simplest way when one point calibration is sufficient:
Inject your neat substances one at a time. Those will be your highest points of calibration (expressed in g/mL, actually it is density). Force calibration curve through origin.

Keep split ratio high to avoid column overloading.

PrithiRao,

Dblux's comments are spot on. Split so you don't overload, use the neat as one data point in your curve and then make another mix yourself with all four components at some level (25% each) and put a second point in your curve. As long as your split ratio is high enough, the FID should be linear over this range of components so that a "3" point curve is sufficient.

This may be a place where normalized results make more sense as well.

Best regards,

AICMM

PrithiRao,
... and then make another mix yourself with all four components at some level (25% each) and put a second point in your curve. ...

Great idea to include this middle point. Quick and efficient.

So what if I do like for example
solvent a-25,50 ,100
b-25,50,100
c-25,50,100
d-25,50,100

first formulation will be 25% of solvent a, 25% of solvent B, 25% of solvent c and then 25% of solvent d.
and do the same for the next 2 formulas with 50% and 100%.
The thing is I need to cover a broad range to obtain the max concentration(some solvents is as low as 2% to as high as 90%). I also use IPA as the dilution solvent.

...
The thing is I need to cover a broad range to obtain the max concentration(some solvents is as low as 2% to as high as 90%). I also use IPA as the dilution solvent.

Calibration that was suggested covers the range from 0% to 100%.
You need to calibrate for all substances you deal with.
Rest is calculation, calculation of concentrations, for which you will need to use specific densities of substances.
Two days ago you were not diluting your analytes, now you do. But calculation of diluted samples won't be more difficult.
I would recommend you to read about calibration in one of GC manuals.
The most important stage is to achieve good chromatography resolution.