Reliable Measurements 15. ICP-OES Measurement By Paul Gaines, Ph.D. • Edited by Brian Brolin
This
section introduces sample measurement procedures, with part 15 covering
procedures for ICP-OES and later parts covering measurement for ICP-MS.
This section is for advanced chemists who are already familiar with
using the instrumentation. Most of the information and tips I supply
here will not be found in a standard textbook; rather they reflect the
major considerations of sample measurement based on my own personal
experience.
When developing an ICP-OES measurement procedure, the following concerns should be addressed:
Line selection
• Sensitivity requirements
• Precision requirements
• Spectral interferences and background correction
Matrix effects
Method validation (to be addressed later in a later part of this section)
The
first step in line selection is to simply choose the line(s) that meet
the sensitivity requirements for your measurement. More than one line
may be necessary due to spectral interferences.
Sensitivity:
Make sure that the line is not too sensitive
(i.e. - going outside the linear working range). Dilutions are easy and
hopefully you'll have more than one line that meets your lower limit of
detection and quantitative measurement requirements. If you do not have
line tables, we have listed the three most popular lines for each
element (also included are detection limits for radial and axial view,
as well as major interferences) in our Analytical Periodic Table.
Precision:
The
precision of an ICP measurement is a function of many factors that are
beyond our control. However, if you require an improvement in
precision, consider the following:
Try to keep
the analyte concentration well within the linear working range. When
the concentration falls < 100 times the detection limit, the
precision begins to fall.
Avoid lines requiring spectral correction.
Avoid lines that are in spectrally complex regions requiring sophisticated background correction algorithms.
Increasing the integration time to as high as 5 seconds should improve the precision. It should not be increased any higher.
Use
an all glass introduction system including a glass concentric
nebulizer. Your washout times will improve along with your precision.
Be
weary of salting out effects when analyzing high salt containing
solutions, but do not use high solids nebulizers unless necessary.
If your samples and standards have the same background then you should skip background correction.
If the matrices of the samples and standards can be matched
then eliminate the peristaltic pump and go 'free flow'. The pump
introduces pulsing and the pump tubing stretches, causing the
introduction rate to gradually change. If you can avoid the pump then
make sure the liquid levels of your samples and standards are the same. NOTE: Slight differences in hydrostatic pressure will make a difference.
Do
not begin the measurement unless the instrument has been allowed to
warm up for at least an hour. Additionally, working in a
temperature-controlled atmosphere is very advantageous.
Clean
your introduction system and replace the tubing when a) you notice an
increase in precision when using glass introduction systems, b) you
notice a 'bend or crink' in the Teflon tubing, c) the torch begins to
build up residue.
Spectral Interferences:
The
spectral tables are very useful, but they are no replacement for a
spectral interference study. These studies should be performed on every
instrument when it is installed and at least annually thereafter. These
time-consuming and tedious studies are very necessary. They are
performed by aspirating 1000 �g/mL solutions of the potential
interfering element and looking at the spectral regions around possible
analyte lines for unwanted spectral interferences. The types of
spectral interferences encountered are direct spectral overlap, wing
overlap, and near neighbors that may cause background correction
problems. The spectral region for the analyte when a blank solution or
the possible interfering element solution are analyzed are hopefully
the same, indicating no spectral interference.
A trace
metals analysis of the 1000 �g/mL solutions helps distinguish between a
true direct spectral overlap and the presence of the analyte in the
interfering element solution as an impurity. The absence of impurity
data requires confirmation of the suspected direct spectral overlap
using an alternate analyte line or another technique such as flame
atomic absorption. Therefore, it is important to use standards with
accurate trace metals impurity data.
Fig. 15.1
Spectrum
of a high Ca containing matrix (red line) causing a sloping background
for the Cu 219.959 or Ge 219.871 lines are used. Background correction
is difficult at best in these situations.
Fig. 15.2
Spectrum
of a high concentration of Fe (red line) showing a direct spectral
overlap upon the B 208.892 line and a wing overlap on a B 208.959.
Fig. 15.3
The following shows a wing overlap interference of Fe (red line) upon the Ba 233.527 nm line.
Fig. 15.4
Have
you ever wondered how you can get negative numbers? It is likely due to
a spectral situation similar to the one shown below. A nearby Fe line
interferes with the background correction upon the Al 396.152 nm line.
Example of a Method:
The
following method (MEB) covers a wide range of elements. This method is
currently under development and is presented for line selection
purposes - it is not a recommended method. The lines shown were chosen
on the basis of sensitivity and relative freedom from spectral
interference. It is important to point out that this method is intended
for an axial view instrument. Cs is used as an internal standard /
buffer (double duty) in this method. Data collected over the past 8
months indicates the method is giving accurate and reliable data.
However, at least another 12 months of work is planned before releasing
this method due to the unorthodox use of Cs as an internal standard. We
believe this is because the level of Cs used 'overwhelms' the matrix.
Matrix
effects are arguably the most subtle danger to the ICP-OES analyst.
Slight differences in the matrix can cause a considerable systematic
error having a definite bias. Internal standardization has been
the most common approach toward correcting this problem and this
approach does work well in many cases. The choice of the internal
standard is critical.
Please consider the following questions before using the internal standard approach for matrix correction:
Is the internal standard (IS) element compatible with your matrix? (Avoid using rare earths in fluoride matrices).
Are there any possible spectral interferences upon the IS line?
Is the concentration of the IS sufficient to give a good signal to noise ratio?
Can your sample possibly contain the IS element as a natural component?
Is the IS clean? Are the trace impurities reported on the certificate of analysis?
Is your method of addition of the IS very precise? Is the same amount added precisely to all standards, blanks, and samples?
Do you always use the same lot of IS for the standards and samples? (Using the same lot is very important).
If
your plasma temperature were to go up or down is the IS likely to
follow the same pattern of intensity change as the analyte? This is
where many IS problems occur (i.e. - an IS with the same plasma /
temperature behavior as the analyte is difficult [at best] to find for
each analyte while avoiding other issues listed above).
The technique of standard additions
is a tedious but more reliable approach for matrix correction. For
unknown matrices it may well be the fastest approach. When using
standard additions on unknown matrices, it is possible to have severe
spectral and background correction problems. It is cautioned here that
at least two spectral lines should be used and the spectral region
carefully scanned and studied. Attempt to make the additions
approximately 2x and 3x, where x is the analyte concentration.
