Why Sample Quality Matters

In my previous article Why the Sample Point Matters, I pointed out that the analysis for any sample starts before it gets to the lab, it starts when it is caught.? What you need to keep in mind is that the extra efforts needed to catch a sample are to ensure you have the best sample quality possible. We all have a job to do, and we all have our expertise. My expertise resides in the lab, I know nothing about controlling a process unit, I do know how to get good data. However, my job is not to run the samples, just like an engineer’s job is not to sit at the control board adjusting the process, it might be useful to think of us as coaches, calling plays behind the scenes.? As coaches, we are not in the field. In the lab, us coaches do not see every sample, we are not monitoring the data generated from every sample, quite frankly we do not care about the data for your sample.

I do not care about the sample data, because it’s not my job, I was not involved in the decisions made to produce the product that is in the sample, I cannot make the product better (directly), I do not control the process (directly).

This might seem harsh, and you may think less of me for revealing this but let me explain.? My job is to ensure the highest quality data is generated from my instruments, and I have many instruments. I typically know about problems with the instruments before the data is reported. How do I know this because we run Quality Control Samples.? This is how we ensure that our instrumentation is fit for use, in fact, it is also how we ensure that the technicians are properly trained. To have a good Quality Control Sample, we need to have sufficient volume and stability for months at a time. These QC’s are run at least once per shift.? If the QC fails, our technicians are trained in how to respond, if their response is insufficient, they contact me.? I can tell what the issue is by a simple question, did it fail high or low??

The chromatogram reveals everything, if evaporation took place, volatile compounds will be less concentrated while nonvolatile compounds will be more concentrated. This is where analytical skills come into play, I look at all the data and draw a conclusion. I cannot base my conclusion on a single data point.? The problem is almost always sample introduction, followed by bad sample.

So, before data is submitted, the technicians have confirmed that they have done their job properly, remember how I mentioned Lab PTSD, in my last article?? Many of our technicians suffer this, as I said previously, the lab must be perfect, no room for error. If they report a bad number, they will hear it from the board operator, they will hear it from the unit supervisor, and then our lab manager is getting calls, and I am getting calls, suddenly we are on an email chain where every manager, director, VP and president is copied, to alert that the people in the lab are incompetent.? If you are someone who authors these emails and copies everyone, I would encourage you to call your version of me first so that we can investigate it. Those emails are a double-edged sword, and I sir have the high ground, I will not hesitate to reply all and present all the evidence to demonstrate that incompetence does not lie with me or my team, but I digress.

So why does sample quality matter?

Well, despite what you have seen on your favorite CSI television show, we do not have time to improve the sample quality.? If you send in a water sample that is half hydrocarbon for Total Organic Carbons, the answer is yes, there are total organic carbons in your sample, and it is saturated. This sample does not need to be injected into our instrument. If you send in a sample that is to be run through a spectrophotometer, where you are shining a light through the sample to detect how much is absorbed, and the sample is black, we can give you a number, but it is meaningless, and if you act on it, it will waste your time.? Everybody wants more data, more is better, right? Wrong, too much data can be confusing, you might see trends where there are none. And bad data will suppress the real trends. You should not want more data; you should want quality data. And to get quality data, you need quality samples. Talk to your chemists, we are here to help, and we love to learn, it is highly likely we will help each other be better!

Bonus content:

I want to elaborate on my statement earlier about not making the product better or controlling the process directly. This is a secret, so make sure you do not under any circumstance tell anyone. The lab is actually very powerful and indirectly controls the entire process.? The lab is the heart of any chemical manufacturing facility, ultimately, operations and engineering will act on the data that is being generated from the lab.? Why? Because to sell the product, they will require a Certificate of Analysis, and this can only be obtained from the laboratory. But as Uncle Ben said, “With great power comes great responsibility.”? The lab must do everything to ensure that it is a source of truth, and sometimes we must withstand the pressures of upper management attempting to make us believe that we made a mistake, when we did not.? Do yourself a favor, become an expert on your instruments, keep your receipts and learn how to communicate calmly, you must have confidence before anyone will have confidence in you. If you do make a mistake, admit it openly and honestly, explain what happened and how you will prevent it from happening again. This is how you gain credibility, and how you become a valuable resource.

?