Both the accuracy and precision of ICP measurements is dependent, in part, upon the calibration technique used. This section is focused upon errors (both fixed and random) that can be introduced through the use of different calibration techniques using accurate calibration standards, samples that have been prepared accurately to within defined error limits, and an instrument that has been 'set-up' correctly using a procedure programmed where there are no spectral/mass interferences that include background correction. You may believe that if the above errors have been confined to within acceptable and known limits that there is nothing else to worry about. Unfortunately, this is not the case.
The most common calibration technique options for ICP measurements are calibration curve and standard additions. In addition, the option of using internal standardization is available for the calibration curve technique and the ability of matrix matching may also be available. ICP-MS has the added option of using an internal standard that is an enhanced isotope of the element being measured (i.e., isotope dilution ICP-MS). This discussion will be limited to the above approaches, to the introduction of the analyte as a nebulized solution, and to the use of Ar as the plasma gas.
Before reading ahead, it may be helpful to restate the assumptions made above and make some additional considerations:
- ICP is a 'Comparative Method' where the measurement of an unknown sample is based upon chemical standards i.e. the measurement is a comparison process.
- It is not assumed that the calibration standards and samples have identical matrices.
- It is assumed that the calibration is linear. This means that the standard and sample elemental concentrations give an instrumental response that is described by the equation for a straight line.
- It is assumed that the analyst has prepared the chemical standards accurately to within defined error limits (i.e., the uncertainty of the prepared standard solution is known and has been calculated).
- It is assumed that the stability of the standards, however and by whoever prepared, is known and are only used within these defined limits of time, matrix, concentration, temperature/humidity, and container material(s).
- The uncertainty of the measurement of an unknown can only be worse (greater) than the uncertainty of the calibration.
- It is assumed that there are no spectral/mass interferences. This of course is an area of great concern and effort on the part of the analyst. This assumption is made to allow us to focus completely upon the potential errors involved with the calibration process.
- It is assumed that the sample prepared for analysis involves no positive or negative contamination errors and no sampling errors. It is therefore assumed that the uncertainty in preparation can be described by the random and known sampling, weighing and volume dilution errors. Again, this is an assumption that is often not the case but is made to allow us to focus completely upon the potential errors involved with the calibration process.
ICP is a matrix-dependant technique. Based upon the above assumptions and the fact that ICP is a comparative method, the prime concern is the availability and use of appropriate calibration standards. The problem analysts face is that ICP (ICP-OES and ICP-MS) is extremely matrix-dependent. Therefore, the ideal situation is that the matrices of the standards and samples be identical.
This section lists several recommendations. Discussions relating to these recommendations are provided in the next section for the reader who would like more detail.
Recommendation (a) - Match the acid content of your calibration standards and samples in both the type of acid used and the concentration of the acid.
Recommendation (b) - Match the elemental matrix components of your calibration standards and samples to the greatest extent possible. In this situation, the analyst who knows the composition of the sample has this capability.
Recommendation (c) - With unknown sample matrices, matching is not possible and is most accurately dealt with using the technique of standard additions. However, this approach is slow as compared to the calibration curve technique with the use of internal standardization.
Recommendation (d) - The use of internal standardization is very effective in many cases but may introduce--or not correct for--all errors. This statement does not apply to isotope dilution ICP-MS that is considered to be a primary analytical technique.
Recommendation (e) - "Chemical calibration is an approximation at best. The analytical chemist must be constantly aware of the possibility of bias introduced by the nature of the standards used, which may be the major source of bias in the analytical data. Appropriate reference materials should be used to evaluate this and other aspects of the measurement process."1
Discussion (a and b) - The matrix will influence the nebulization efficiency, which is proportional to the signal intensity. Nebulization efficiency is the percent of solution that reaches the plasma. Therefore, if the nebulization efficiency is 1%, then 99% of the solution is going to waste and 1% is making it to the plasma. Typically, nebulized solution 'mist particles' that are greater in diameter than 8 microns will go to waste. If a matrix component changes the efficiency from 1.0% to 0.8%, then a relative drop of ~ 20% would be expected from this effect alone. The droplet size distribution of a pneumatic nebulizer is governed by the physical properties of the solution as well as the volume flow rates of liquid (influenced by peristaltic pump speed and tubing diameter) and gas (sample Ar flow rate). The physical properties claimed to influence the droplet size distribution are the surface tension, viscosity, and density. See Inductively Coupled Plasmas in Analytical Atomic Spectrometry; Montaser, A., Golighty, D. W., Eds.; VCH Publishers: New York, 1992 - page 703 for more detail and additional references on this topic.
For the ICP analyst, the most common matrix component that will alter the physical properties of a solution is the acid content. This is not to say that other differences such as the presence of trace organics (added intentionally or not) should not be considered. However, the identity and concentration(s) of one or more acids is an issue that virtually all ICP analysts have to decide upon. The ICP analyst is most commonly involved in the preparation of samples where one or more inorganic mineral acids are required to bring about dissolution of the sample and/or to maintain solution stability of the analyte(s) of interest. The acids most commonly used are HNO3, HCl, HF, HClO4, H2SO4, and H3PO4 and are listed in the order of best to worst.
The effect of acid matrix upon nebulization efficiency is such that a change in acid content from 5 to 10% v/v will cause a decrease in efficiency of 10 to 35% depending upon the acid used, nebulizer design and liquid and gas flow rates. Matching the matrix to within 1% relative is necessary for the most accurate (we use the term "assay") work (i.e., a 5% HNO3 acid solution would be made to 5.00 ± 0.05%.
The matrix will influence the plasma temperature, which is related to the signal intensity for ICP-OES. The other effect matrix components have on the ICP cannot be explained by a change in nebulization efficiency. The effect is one where the matrix components give the appearance of taking power away from the plasma (lowering the temperature of the plasma). It has been reported that this effect is related to the excitation potential of the line and that the effect increases as the excitation potential increases. A similar effect would be seen by decreasing the applied RF power or by increasing the sample (nebulizer) Ar flow rate since both result in a reduction of the plasma temperature. Therefore different lines of the same element would be affected differently according to their excitation potentials. In addition, when choosing an internal standard element it follows that the excitation potentials of the internal standard and analyte lines should be as close as possible, unless the calibration standards and samples are matrix matched. For more information and additional references, see:
Inductively Coupled Plasmas in Analytical Atomic Spectrometry; Montaser, A., Golighty, D. W., Eds.; VCH Publishers: New York, 1992 - pages 279-281
ICP-MS suffers from nonspectral matrix effects. The effect most commonly encountered is referred to as 'quenching' and is thought to be due to defocusing of the ion optics by space charge effects. Generally, as the concentration of the 'matrix element(s)' increases, the analyte signal will be suppressed. Quenching increases in effect as the matrix element absolute concentration increases, the matrix element mass increases and the analyte mass decreases. This effect is absolute in nature and not a function of the relative concentrations of the matrix elements and analyte elements. Therefore, when sensitivity allows, it can be diluted out. It is also greater in effect as the RF power is lowered. The effect is such that an element matrix concentration of 100 ppm can severely suppress a 'cool plasma' (RF power W ~ 800) and has little effect at normal power (an RF power of ~ 1300 W). For more information and additional references, see:
Inductively Coupled Plasma Mass Spectrometry; Mantaser, A., Ed.; Wiley-VCH: New York, 1998 - page 543
In the next section, recommendations (c) and (d) will be discussed along with some additional considerations.