In the laboratory, we go through great efforts to ensure data quality. We must know our instruments inside and out. We must understand the chemistry on which our analytical instruments are based. We have many resources to help us understand our analysis, from published ASTM, UOP, EPA methods to the instrument manufacturer technical support. We go through training on our instruments, we grow our Subject Matter Expert Network. In fact, any one of us has at least 3 “phone a friend” options on our speed dial for just about any emergency. If you joined my Analytical Instrument Resources Group, there are many more SME’s just a post or message away.
The more we use these instruments, the better we are at understanding problems when they malfunction. If a result from our instrument is wrong, there are a plethora of engineers, operations supervisors, operators, managers, senior managers ready to point out the error, and chastise the laboratory for not doing their job right. The laboratory is the favorite scapegoat of any refinery. So, you can imagine it is in our best interest to learn our instruments and become experts as soon as possible. There really is no room for error, we simply must always be perfect. You should not be surprised to learn that lab personnel tend to suffer our own form of undiagnosed “PTSD” for certain analysis. So much that they might incorporate certain rituals like holding their breath, spinning around three and a half times, closing one eye, standing on one leg, before hitting the start button!
I have mentioned in a previous article “Why the Laboratory Matters”, that there is always some distrust between the lab and operations. Obviously, operations and engineering need to be just as much an SME as lab personnel. So, we each will posture and point fingers, they have their resources and consultants that they call as well. The difference is, for analytical instruments, we have considerably less points of failure than process units. The whole purpose of having a laboratory is to have a controlled environment, instrument manufacturers invest in stringent quality control measures to ensure that the instrument performs as expected. In the laboratory, we have more control over anything that could possibly go wrong than what operations does.
So why does the sample point matter?
The sample point matters because it is the very first step in analysis. The laboratory can only tell you what is in a small aliquot of the sample in the container that was delivered to the laboratory. In fact, consider a GC analysis. The typical injection amount for a GC is 1 microliter, or 0.001 milliliters or 0.000001 liters, that’s it, the entire process is being controlled off of results from the tiniest portion of the sample that was delivered to the laboratory, in fact a typical GC method will have a 100 to 1 split, which brings the portion of the sample analyzed to 0.00000001 liters.
Now imagine a contaminated container, what are the chances of getting accurate results? Or let’s consider the sample point, for a liquid sample, the source would have to be piped into the sampling station. If this pipe is lined up so that it is continuously flowing, there is a better chance of getting a representative sample, but if it is not continuously flowing, this would essentially be a dead leg, the entire volume of the pipe up to the source is stagnant. You can still get a representative sample, but you must flush the line out, discarding the entire contents of the dead leg and filling it with fresh sample before obtaining the sample that will be sent to the lab. This is an important detail that will influence the accuracy of the data.
I have never been an operator, so I cannot speak to the training or attention to detail that is emphasized to the operators catching the samples, but I do know that they have lots of responsibilities, not to mention all the hazards surrounding them in the unit. It is not beyond the realm of possibility that properly flushing the sample point is not performed.
The lab can only give you the results for the sample that is delivered to the laboratory. We have taken steps to ensure that the instrument is working correctly before approving and submitting the results. But the analysis started at the sample point before it got to the lab. To save everyone’s time, heartache, and frustration, look at your sample points, have a chemist review it before installing or upgrading. The reality is that the designers of the sample point are not as familiar with the consequences of a poor sampling design as the lab is, so there should be lab representation in any and all sample point discussions. The goal should always be to make it easy to do it right, and hard to do it wrong. Lab PTSD, the struggle is real!