Quality Control Samples
by Tim Loftus
James Dux, Ph.D. describes quality control in his book, Handbook of
Quality Assurance for the Analytical Chemical Laboratory, as
“those operations undertaken in the laboratory to ensure that the data
are generated within known probability limits of accuracy and
precision.” In other words, it is the stuff that you do that shows
whether your lab results are most likely right. Unfortunately, many
laboratories do not have the quality control they should have to
produce unquestionable results. Reasons can vary, but most people find
it hard to add 10% to 20% more analyses from QC samples while their
normal sample output remains the same. It seems like a waste of time
for a technician who knows that he is performing a test method
properly.
There may be more to it than that,
though. The technician may be performing the test properly, but is the
test equipment working properly? Are the reagents OK? Is there
interference in the sample? QC samples, run along with “real” samples,
are to help determine just these types of errors. Any problem with
equipment, reagents, and yes, technician error, will produce bad
results. Bad results may show an NPDES permit violation when, in fact,
there is no violation. QC samples, when they fall within acceptable
levels, will help to validate the results of the “real” samples.
QC samples should be run with every
batch of samples analyzed. Generally 10% to 20% of your analysis run
should be QC samples. Common QC samples include:
Method blank
– This is reagent water that has gone through the same procedure and
has all the reagents as the other samples. Method blanks are useful in
determining contamination. Ideally, the method blank should not be
greater than the method detection limit.
Known
– This is a sample that has a known amount of analyte in it. The
sample can be made in-house or it can be from an outside source. If it
is made in-house, it should be made from a different source of
chemicals than what the test standards are made from. The value of the
known sample should fall within the result range of the other samples.
Knowns help determine the accuracy of an analysis.
Spikes
– A spike sample is a “real” sample that has a known amount of analyte
added to it. The difference between the sample and the spiked sample
is determined and the percent recovery is calculated. Spikes help to
determine if interferences are present. Generally, 85% to 115%
recovery is acceptable.
Duplicate
– Two samples are taken side by side using the same procedures and
tested separately. Duplicate samples used to evaluate analytical or
measurement precision.
Replicate
– A single sample that is tested twice within the batch. The results
indicate precision.
Quality control also includes
control charts, standardization of equipment, cleaning glassware, and
a number of other issues, some of which will be the subject of future
articles. However, keep in mind that not all of these QC suggestions
can be used with all test methods. What may be used with phosphorus
analysis may not necessarily be appropriate with pH analysis or
coliform analysis.
While the above suggestions are only meant as a rough guideline,
there may be other requirements for QC samples in your NPDES permit or
state certification program. For those of you who do not have a formal
QC plan, consider incorporating some of these samples in your analysis
runs. It will take more of your time, but with good quality control,
you will not have to defend your results. Your results will defend
themselves.
If you have any questions,
suggestions, or comments, please contact LPC Chair Paul Fitzgibbons at
(401) 222-6780 ext 118 (lab@narrabay.com)
or Tim Loftus at (508) 949-3865. You can also visit our website at
newea.org. Once on the website, press the Lab Practices button.