Have you check the RSD% at your low level? Maybe is more than 5%, i.e. 10% (only a guess)

Thanks for you reply, if it be 8% then?

I mean what is the role of RSD%?

for what range of concentrations is your instrument calibrated? are both measurements within that range?

Did you check the chromatogram? is there overlap of the Ca and Mg peaks for the high concentration sample?

Is 6.6 ppm a single value or is it the mean concentration? In my opinion, these results are not that bad. However, did you check the chromatograms for overlaps and correct integration of peaks?

Peak integration methods are the right method to optmize along with standards.

I agree with Pieter.  What is your calibration range with each of your cations?  What is your resolutions?  You need to be within your calibration range.

Hi Taher,

As the people before me asked; did you do a multiple injection of the same sample or did you just injected the sample once and got this result? The other option is as stated above is that your result (the response from the sample) was not in your calibration range, in other words your result is not an calibrated and accurate result but an estimated and possibly inaccurate result..

Do you know the stability of the solutions? And did you analysed the solutions on the same day as you made/diluted them? 

One more question: What software do you use on your IC? If you're using Chromeleon (like I do), although I'm not entirely sure if other software has it too.. An I'd like you to think about is - which feeds into the suggestion about the peak integration - have a look at the QNT function of the software, which lets you 'choose' on which heights/areas the peaks will be integrated. So the peak integration will stay automatically, instead of manually integration that in order involves the human error which you want to avoid.

To end this, I wanted to answer your question about the %RSD. The lower your %RSD is, the more reproducible your data is. In other words, if you'd like to do the same analysis some other time you're likely to get the same results when having a low percentage relative standard deviation rather than when you're having a high RSD on your data.

Hope this is helping you!

Thank you very much for your kind help and complete explanation.

I use Chromeleon software and I just injected once. I think there is problem with my calibration curve. My reaction solution is 1 litre and every 10minutes I take 1ml and filter and dilute it (by weighing 20-40 times) I mean 2-3 drops in 4ml milliq water. then I store the solutions in fridge and after 3 weeks when all of the samples are ready I analyse them.

I have 1000ppm Ca and 1000ppm Mg standard solution which I bought from VWR company and are certified. I want to make calibration curve between 5ppb and 80ppm for Ca and Mg would you please guide me how to do it? it seem when I make standard solutions at lower ( less than 3ppm) concentrations I get less accurate results.

Depending on the concentration of your solutions the storage time differs.. also some compounds have shorter storage times due to certain properties, so you have to take that into account (lower concentration solutions tend to be unstable after a certain period of time) - you might want to check some literature references for it.

I always calculate my dilutions with the formula: C1 * V1 = C2 * V2 (check the website http://www.tocris.com/dilutioncalculator.php#.VUuXlflVhBd if you're not an expert in formulas :p). I made solutions of 10 ppb (0.01 ppm) concentration and had no problems with inaccurate concentrations.

Thanks for you reply. Just to check: Is it correct and good way I weight a vial and add 5 ml of 1000ppm Ca standard solution and weight it again and add 5ml of 1000ppm Mg standard solution and weight it again to have 10 ml of 500ppm Ca and 500ppm Mg solution and use it.

Then I get 5ml of this new solution and add 5ml DI water to have 250ppm Ca and Mg and continue this way (STEP BY STEP AND 10 steps)  to get about 5-10ppb Ca and Mg?

Sorry if replying my messages takes your time.

I'm not sure why you include the weighing step.. Can you clarify that for me?

The other steps you described sounds correct to me!

Because in each step I calculate how much I have diluted by weighting. I mean I weight  the vial (w1) then I add water and weight the vial again (w2) and add some drops of sample and weight the vial again(w3)

the amount of dilution can be calculated by: (w3-w2)/(w3-w1) by this way I want to reduce the pipetting error. I mean when I pipette 5ml I might pipette 4.99 or 5.01 but if I know the weight of the solutions I can calculate the amount of solution added.