Ok, Baju, apparently, you cannot have a negative amount of chromium in your sample.

You state that a blank gave you a 192 reading, so I would expect this to be your intercept (at zero chromium concentration). But you state that your calibration curve determined it to be 3452. Here is where I suspect your problem is, not in your calculations. (Some could argue that your intercept should actually be zero, but in reality, you always have some background noise that will not make it zero most of the time).

Assuming you did not make a mistake determining your calibration curve (e.g., it is easy to forget a zero, when it comes to very low numbers), I can see two situations where this can happen:

First, your calibration curve is made up of samples that contain very high concentrations (for example, 1000 mg, 5000mg and 10000 mg), and you just extrapolated this all the way down to zero. In this case you should try to dilute your controls that made up your calibration curve and see whether you end up with a slightly different slope. If your sample just contains a few micrograms, for instance, you may just be out of the proper dynamic range.

Second, your calibration curve contains something that interferes with your measurement and your sample does not contain that same background. In this case you would have to find a control sample with a similar background to your sample.

Hope this helps you getting to the bottom of this.

Dear Bayu, if the data is not within the calibration curve, can not be extrapolated ... I recommend a method called: standard addition. In summary consists of taking a sample of known concentration, the answer is within the calibration curve, then he adds a ml or more of your sample, calculate the concentration of the second and subtracting the concentration of the first (solution of known concentration) and get the concentration of your sample.

Dear Baju, Generally Calibration has been plotted to know the relation ship between concentration and Intensity. Any how as per your problem you stated that there is good linearity between concentration of chromium in water with the intensity what you measured. one thing we have to remember that what ever the sample concentration that you had prepared, expected to be lies in the linear range of Standard plot of chromium. i think the problem may be with the concentration of sample that you have prepared and needed to increase the concentration to get positvie result.

Thank you for all the answer.

I made the the series of standard solution from 0 - 50 ppm (0, 0.1, 0.5, 5, 10, 20, 25, 50 ppm)

if i get the opportunity to use icp-oes again i will do the standard addition method.

But now i have to report this data,

is it okay to make the intercept to be zero?? In excel we can do that. If i do that, i still have a good calibration curve with correlation 0.999

Some authors agree not to introduce the "zero" calibration curve, however, others if ... by premise that such "zero" is a concentration. I do not introdusco the "zero" since I have no certainty of the analytical response between the lowest concentration solution and "zero"

What concentration do you get when you set the calibration curve through zero? And what do you get when you set it at 192, which is what you actually measured?

You have to check LOD LOQ issues and method validation in general.

Yes , should be check LOD and LOQ for method .

Limit of detection less than limit of quantitation (LOD

Hi Bayu

This tends to happen when the matrix of your sample does not match the matrix of your standards, or you've had carryover in your standards. Best bet would be to use standard additions. Cheers Frank