Matrix effects and interferences.

ICP-MS is subject to significant matrix effects, there must be a good matrix match between the calibration standards and test samples, e.g. acid type and strength, total dissolved solids,   Dissolved carbon an be troublesome for analytes with high ionization potential, e.g. Zn, As.

As Ed pointed out, ICPMS is prone to matrix effect.  For a number of reasons, this can result in the “suppression” or the “enhancement” of the signal of the analyte compared to that in the standard solutions.  Assuming that there are no spectral interferences, the former which is the most common situation, would result in a lower result (a negative bias) than the expected concentration with external calibration, and vice versa.

Isotope dilution, at least theoretically, can eliminate the errors due to any signal suppression or enhancement (matrix effects), as the analyte in both the standard and the sample experiences the same matrix effect which would give us an accurate result.  Therefore, any signal enhancement by the matrix would result in a positive bias with external calibration as compared to isotope dilution, and vice versa.

The method of Standard addition, also can eliminate matrix effects (again, assuming no spectral interferences) which should behave similar to isotope dilution.

If the main matrix composition is somewhat known and consistent, a simple solution would be to use matrix-matched standards (e.g. for seawater).

Regards,

Nimal

To give an adequate answer you should provide more details; e.g. on sample type, matrix characteristics, acid used (HNO3 or HCl), which elements you are interested in. As pointed out by the colleagues there are many possibilities of cross-interferences and, therefore, more details are necessary for an answer. The standard addition method is always a good approach to solve matrix problems. Somehow less cumbersome is the use of internal standards as an internal reference. I used spikes of 10 ppb Rh and 10 ppb Ir for standards and samples. These spikes are similarly affected by matrix effects of the samples due to signal suppression and signal enhancement and so  you should be able to adjust your sample signals to the standard signals by recalculating the spike signals. (Of course, this is no help in cases of mass coincidences.)

Good luck and regards

D. Zachmann