ISTD setting in Data Analysis


We recently encountered a DA processing error for many of our batches when running for Fe, Al, Zn and Cd.  For some background, where are a nanotechnology/materials lab synthesizing core-shell nanoparticle structures, and use ICP to determine composition and concentration after processing. 

I found that often, the ISTD wasn't defined in the batch DA, stating <none> instead of mass 89 for Yttrium.  Mass 89 was acquired and properly defined as the ISTD in the prior DA tab, but not selected in the fully quant tab.  What exactly is the ISTD doing, in the context of post acquisition DA?  How and why are the results changed when the ISTD is defined in the fully quant tab?

When I selected mass 89 for the ISTD and re-processed the batch, many of the calculated sample ppb values increased by a few to 10% or more.  In some cases they decreased slightly, but mostly increased. Turns out this has been a problem with more than a few of our batches from these elements (EPA multi element standard) since we typically "Create from existing"...  At least we caught it.  Going forward, we will be aware of this setting and implement some control solutions for these lower throughput analytes for us.  We run tons of gold and silver and have good control over those. 

I understand the basics of IS correction for matrix, SIS or plasma effects in ICP-MS, but the corrections with this field weren't proportional to the ISTD recovery.... In some cases, samples with IS recovery within a couple percent would be hugely different after re-processing DA when the ISTD box was changed from <none> to 89 in the Fully Quant tab.  In other cases, not as much.

Can anyone explain this to me???

Many Thanks,

David G

David Garcia
Characterization Services Manager
nanoComposix inc.
office: (858)-565-4227 ext# 155
cell: (760)-822-0089

"Success usually comes to those who are too busy to be looking for it.”
-Henry David Thoreau

  • Hello , thanks for sharing this with the community.

    There are some questions I have before suggesting any action:

    1) What are your calibration ranges for Fe, Al, Zn and Cu?

    2) What is the Y concentration?

    3) Are you doing online addition through the "T" piece?

    4) Can you see the IS stability graph after choosing m/z 89 as ISTD? How does it look like (around 100% throughout your run)?

    Fact: Y should be added in the very same concentration throughout your analysis, and should not be present in your sample. As far as I could get from your case, results change when you apply the IS correction in DA. What happens is that the CPS obtained from the blank for Y is set up as a reference. Any other read of Y is divided by the CPS from the blank. For instance, if you're reading 10,000 CPS for Y in blank and 9,800 CPS in Standard 1, your IS recovery will be 98% (or, in terms of ratio, 0.98). Your CPS for Standard 1, for the elements you choose 89 as IS, will be divided by this ratio. Hence, your calibration curve will be (concentration) x Ratio, instead of (concentration) x CPS. Normally everything goes well in this step (calibration standards are controlled solutions, you know exactly what you added).

    This said, when you introduce your sample, the same comparison between Y CPS obtained in your sample and the calibration Y read will be done, and this factor (ratio) will be applied to it. In an ideal world, calibration standards and samples will give the same Y CPS, then the ratio will be 100% and your ratio will be 1.00, not changing the final result. In the real world, your samples contain interferents that could affect the liquid transportation (such as different viscosity, surface tension, etc) and plasma characteristics (electron density, temperature, etc) and the Y signal will not be the same as the obtained from the blank solution. This will give you a different ratio - can't say if higher or lower, it depends on the characteristics of your sample, and your result will be divided by this ratio, then you get a different result. Nonetheless, the IS correction in ICP-MS is quite frequent due to signal oscillation, space-charge effects, transportation effects, among others. One way to check if your IS correction is working fine is to use a CRM or spike sample. Even with the final concentration correction, you need to achieve a satisfactory percentage recovery (let's say +- 10%).

    Using the ISTD Stability Graph you can see if during the analysis of your samples the Y recovery goes down or up during your run. You could get some actions like matrix-matching the analytical curve, diluting more the sample or even running aerosol dilution to reduce the quantity of solvent plasma load to the plasma. Conditioning of the cones with a sample also helps getting better long-term stability.

    You could try using another IS. Scandium or rhodium are quite common.

    Let me know if this helps you out.

    Best regards,


Was this helpful?