Internal vs. External Standard

[vxfx]

Hello all.
I'm trying to calculate the volume % of my samples (CH4, C2H6 and C3H8) and I just use the volume/peak area ratio between my external std and my sample. I also did a plot Ave Peak Area vs. Published Vol %. I used the linear equation from it to find my samples' vol. %
It looks good, so far.

However, my teacher told me I could double check these methods by using this method on my internal standard cause I know the amount of each gas I injected. For example, 10uL methane, 10uL CO2, 10uL C2H6 and 10uL C3H8.
So, technically...their vol.% relative to the external is 25% percent each. I'm puzzled because I do not get 25%.
Rather, I got about 15% CH4, 16% CO2, 31% C2H6 and 37% C3H8.
Even from the peak height, ethane and propane have higher peaks than CH4 and CO2. So, that makes sense if they have higher vol%, but they are supposed to have 25%. Baffling situation....

Or is it because I cannot relate the internal and external standard at all?
Well, I did use the internal standard for the calibration of external std. Somehow I think I cannot relate them both through equations or volume/peak area ratio as I did for my unknown samples. Let me know if I'm wrong or right.

Thanks.

[Peter Apps]

Your detector is giving a different signal for a unit volume of each gas, as expected.

Peter

[larkl]

As Peter indicates, you cannot assume the same detector response for each component. If you were using FID (I'm guessing you're not since you're measuring CO2), the methane peak would always be about half as big as the ethane since the FID is a carbon counter and there are half as many carbons in a mol of methane compared to ethane.

[vxfx]

I'm using GC-IRMS to test my samples.
That clears things out. I did so many different ways to calculate my internal relative to external, just 'cause I want to see if my methods make sense

Anyways, is it possible to answer one more question for me?
For my external, I have published vol.% of
CH4 81%
CO2 0.999%
C2H6 2.8%
C3H8 0.4%

I did 3 runs and I have 3 sets of Peak Area, thus I average them up
I did a linear plot of Average Peak Area vs. Vol% above.
I have a good linear plot and good linear equation. I'm using this equation.
But, somebody dispute it by saying I cannot plot a graph using the vol% of these 4 gases in one line, instead it should be Vol Injected (uL) vs. Peak Area of each run. Then use this equation to find the vol for my samples or internal std (to test the method). I tried it, but the numbers are crazy.

- I do think I can use Average Peak Area vs. Vol% of the four gases in one linear line, rather than doing one line for each gas.
IDK, Help Please...

[Peter Apps]

What is IRMS - isotope ratio mass spectrometry ? If so I suspect that you are measuring carbon 12 content, and then plotting that against mole fraction (which is what you call volume). First problem; three of the four points are in the bottom left corner of the graph, and one is in the top right corner, effectively giving two points, which is bound to give a nice straight line irrespective of the validity of the approach.

If your method (and by that I mean the injection, separation and detection as well as the calibration) is working right, then a given quantity (whether measured as volume, mole fraction or mass) of a given analyte must give the same detector signal under all conditions, in other words independent of the mixture (e.g. internal standard, external standard or sample) that it is in, or how the sample was injected. Your method apparently does not fulfill that condition, why, given the information at our disposal, is anybody's guess.

Peter

[Suresh Seethapathy]

Hello vxfx,

Forgive me if you already know the principle of GC-IRMS and this is just a repetition. GC-IRMS involves three steps

1. Separation by GC
2. Combustion of everything that elutes from gc column into CO2
3. Separation of C(12)O2 and C(13)O2 (and all other combination of isotopes) by a magnetic sector MS.
4. Quantitation of each isotopic species (m/z of 44 and 45 depending on whether the carbon dioxide contains C12 or C13) using separate faraday cups.

the total response you see for each analyte should be proportional to the number of carbon atoms they have. Accordingly, you are seeing almost the same percentage for CH4 and CO2, double the response for C2H6 and you should ideally have had around 45% for C3H8. A low of 37% for C3H8 could have resulted from incomplete combustion.

The volume % you have in this thread corresponds to C(12)02 peak areas or C(13)O2 peak areas or the sum of them?

Cheers,
Suresh
www.science.uwaterloo.ca/~sseethap

[Suresh Seethapathy]

vxfx,

If you can email me an excel sheet with the raw data, i can have a look at it and guess at what is happening.

BTW, what make/model GC-IRMS are you using?

Cheers,
Suresh.
www.science.uwaterloo.ca/~sseethap

[vxfx]

Thank you Suresh and Peter. I understand it now.
Sorry for the limited information that I gave, but to answer your question Peter, my experiment was conducted in the same conditions for sample,internal and external. So, that's great!

Suresh, the GC-IRMS I'm using is Hewlett Packard 6890 GC with Poraplot Q column. It's okay, I don't want to trouble you with my data.

Thanks alot.

FZ