linearity test - GC method

[kskishore1973]

Hi

I calibrated GC for Ethylene impurities with calibration blend of hydrocarbon impurities in Ethylene balance. I used two standards(with concentration other than that used for calibration) to check its linearity and calculated R2 including Zero as third point.
Is two level linearity check is acceptable for method validation.
Please suggest most accurate method to dilute hydrocarbon gas blend.

Thanks in advance.

[rb6banjo]

It seems to me that this always comes down to the question "how much risk of being wrong are you willing to take?". We use a 1-point calibration standard to ensure the quality of a gas we use. Mostly it's because as long as the analyte concentration is less than a certain amount, the gas is acceptable for use. First, I think you have to answer that question for yourself for your situation. "What risk will be minimized or addressed by adding more calibration standards". Also, how high is the risk of nonlinearity of your detector? Flame detectors are quite linear over a very large range.......

Dilution of gases can be pretty easy. I use the canisters by Restek:

http://www.restek.com/catalog/view/43681

If you fill the canister and then vent it to atmospheric pressure, you repressurize the canister with analyte-free gas to a known pressure (need a canister with a gauge on it) to dilute it. Use the gas law to calculate your dilution factors. I've had good success with this approach.

[LabProARW]

I normally use 3 standards and certainly ignore the origin. Using only 2 standards is very risky as if 1 of those standards was not made exact - then the line drawn from the 2 points (one of which is "off") would tilt and the farther away from the second "good" standard the result was the farther "off" and in error the result would be. With 3 standards also, if one standard was not made exact then it shows very obvious by the line drawn through best fit of the 3 standards. I have found this a great help and can look at the line and considering the origin can spot which standard was not made well.