response factor

[Noodles]

Can someone please explain if the following calculations is correct.

I ran a standard of p-xylene and o-xylene, 33% and 67%. My recovery was 31% and 68%.

then we ran samples, next we figured out the response factor for each analyte in this manner p-xylene=33/31=1.05 and 0-xylene=67/68=.98.
Then we multiplied the sample results with the factors.

My question is why would one use this instead of running an external standard, calibration the instrument and then running the samples.

This might be remedial to most members but any explantions will be helpful.

[Peter Apps]

"My question is why would one use this instead of running an external standard, calibration the instrument and then running the samples. "

It looks to me as if that is what you just did - ran a standard of known composition and compared the results from the samples with the results from the standards. You can do the arithmatic in any number of ways, but the underlying analytical process is the same.

What worries me is that you do not mention running replicates - I would not be unduly surprised to find a 2% discepancy between replicate injections if conditions are not set up just right, so you might be applying a response factor correction when the cause of the discrepancy was random error.

I presume that what you mean by "recovery" was the actual result from running a standard, and that it was based on % peak areas of the two components ?

Peter

[chromatographer1]

I suspect that Noodles had an internal standard in both samples, the standard to determine relative response by weight and a sample containing the aromatics and applied a correction factor to their results.

I agree with Peter unless you are injecting directly, not using a split, your results can vary by 2% from injection to injection. If your results must be well within these limits then you should be doing a direct injection only.

I assume that you would prefer to have confidence in the numbers you are generating to the nearest tenth of a percent perhaps, instead of to the nearest percent.

Discrimination of components in a sample seems to be a factor forgotten by many modern analysts. Reproducibility of sample replicates is an important factor in the quality of your results and must not be ignored.

best wishes,

Rod