Before conducting our experiments we go for validation of our experimental setup with some known data sets. As we are also providing instrument errors and uncertainty analysis as well with our work. Therefore, what is the main motive behind doing that? Is it just to satisfy the world community or it is for our satisfaction that our setup will provide good results. Please do suggest some good research in this field.
What you are describing is called a "benchmark experiment". You can search Google Scholar for this term. The purpose is, as you said, to satisfy yourself and the rest that the output from a machine is reliable. Not only the machine is reliable but the experimenter is also trustworthy. However this does not guarantee or prove that any new effect you report is 100% correct until and unless there is an explanation and someone else can also reproduce it. A lesson in science is "Paper never refuses an ink" which means anything which is published is to be believed with a grain of salt. A very good story is that of Micheal Kasha who was a very famous physicist in the field of photochemistry and a student of G.N. Lewis. He showed odd experimental his professor (the quote is from his own recollection). Kasha recalls the dialogue of his mentor.
Kasha writes "When Lewis saw that the Beckman Spectrophotometer gave these results on linear molecules as well, he said, "Take Me To Your Spectrophotometer!" Lewis asked me to explain the various components and their operation. He had me unscrew the phototube and cell compartments, exposing the exit slit of the spectrometer. Then he said,
"Set the monochromator dial for 550 nm," and poised himself in line with the exit beam.
He said,
"I see GREEN."
Then,
"Set the monochromator for 600."
He said,
"I see YELLOW."
Then,
"Set the monochromator for 700."
He said,
"I see RED."
Then,
"Now, set the monochromator for 1,000 nm."
With excitement,
"Aha! I am the first man to see the INFRARED. And it is GREEN!"
Lewis had suspected that the monochromator was misbehaving, and we quickly saw that the whole of the visible spectrum could be seen on this instrument by scanning the near infrared from 750 to 2,000 nm. Thus, any visible dye would show an illusory absorption spectrum on the background of this scattered light, apparently reflected off the instrument wall. There are probably some false curves in the earlier literature based on this error. Lewis taught me to be
wary of instrumental performance."
(Adapted from: Kasha's The Triplet State in J. Chem. Ed)
There are two types of errors as random errors and systematic errors.
Random errors are usually accounted by statistical treatment of data.
Systematic errors are due to instrumental errors, personal errors, and method errors.
Your question falls into the third category of systematic errors.
Methods errors are introduced due to nonideal behavior of a process. These methods are very hard to detect or correct. Therefore, methods should be validated. To validate a method you need to have standard materials. The analyte concentration of these standards must be known to a high degree of certainty. NIST provides range is standard reference materials (SRMs). According to your requirements, you can purchase the material and validate your instrument and then make the measurement of your sample. Interpolation or extrapolation (not recommended though), you can find the unknown.
So in simple, validations are done to avoid systematic method errors.
It is all about getting standardization and authantication of the experiment.
Calibration is important for any experimental set up. Take an example of a load cell, which is commonly used in many types of lab tests. It has to be calibrated in air. Then it has to be again calibrated when fitted on a model in different environment. While carrying out expetiments is a repeative task, prior validation is very important to ensure that all measurements make sense after a hard work.
Only if your research work will be standardized, your results would be reproducible.
- Badges
- Science topic
- Similar topics
- Instrumentation Engineering
- Instruments