Repoeting of LOQ

[Juan2011]

Dear All,
Here I come with a typical question of mine. We have a method which use calibration curve to quantify X. The range of calibration curve is from 1ppm(LOQ) to 25ppm. The sample preparation is as such
a) Take 1ml of sample
b) Add 10ml of reagent A to extract X.
c) Centrifuge it
d) Take 2ml of extract and add 2ml of colour developing reagent-C
The standard curve is prepared as follows
a) Make a series of standard solution from the stock
b) Take 2ml of standard and add 2ml of colour developing reagent-C
The calculation is as follows
Amount of X from Calibration curve =(Area of standard +Slope)/Slope= D
Amount of substance in the sample={ D x Dilution factor which is 11} / Volume of sample in ml
The amount obtained from calibration curve for one of my sample is 0.5ppm which is below the 1PPM (LOQ) of the method calibration range. Shall I report this sample as below 1ppm (LOQ)? If we calculate the amount of X is--( 0.5 x 11)/1=5.5ppm . I was wondering that we should report the result as below LOQ. Any body wants to say something

[sepscientologist]

What is the absorbance for your 0.5 ppm sample? The linear range ought to be around 3 orders of magnitude so if
the 25 ppm standard is up near 1 AU then you ought to be able to measure and report down to about 0.01 AU easily
which would be ~ 0.25 ppm.

[unmgvar]

i think the best thing is to declare indeed that you have a less then LOQ more then LOD result and in case of stability then you are following up on it
i would not declare an exact amount since you are below LOQ. you cannot make such a statement.
what are your specs for X? since LOQ is 1ppm then LOQ X should be 11ppm based on what you are saying
LOQ is stating the limit of your reporting capability
now if your required sepcifications of work need to be better then what you can do right now with the method, then you need to find a better method.
otherwise, less than LOQ

[lmh]

There's a slight fuzziness here. When you define LOQ as 1ppm, is this because your calibration curve stops at 1ppm, or does your calibration curve stop at 1ppm because you have previously found that the RSD on measurements below 1ppm is inadequate for the purpose?

If the signal is still substantial at 1ppm, and the calibration curve is linear, and extrapolated it passes close to the origin, then it is likely that your estimate of 0.5ppm is good. The danger in this case is that you are extrapolating and liable to systematic error. It is absolutely right to flag that it's an extrapolated estimate, and in many contexts extrapolation is just not permitted.

If, however, the signal at 1ppm is already at the border of what can be quantified with acceptable RSD (i.e. you are lower than the genuine LOQ on the method, not just lower than the lowest point you've so far taken a calibration) then the estimate is almost certainly bad. In this case the danger is that you are liable to random error (as well as systematic error). The only thing to do is report that the sample contained more than LOD and less than LOQ without further quantification.