NJDEP-SRP Low Level TO-15 Series: Part 3 – How to calibrate your system Joisey style… but first I shall digress on ISTDs

Last time, I showed you how well the Markes CIA Advantage and the Rtx-VMS concentrate and separate the volatile organic compounds (VOCs) of interest for NJ LL TO-15. Next on the agenda would be calibrating our system (i.e., preconcentrator-GC-MS). NJ tells us to calibrate our systems with a minimum of 5 points (more allowed) and that the range must be from 0.2 to 40 ppbv for the majority of compounds (exceptions to be discussed in the next blog).

So… the single biggest complaint I have heard from laboratories running (or attempting to run) NJ LL TO-15 is the 0.2 to 40 ppbv calibration criterion. In fact, I have had laboratories (who will remain anonymous) admit to running two different curves to span the required range. I do not think this is legit, but bottom line I can now attest to the fact that achieving linear calibration across this range can be challenging, especially for all components. It is a fine balancing act between injection volume, breakthrough volume, detection limits, saturation limits, etc... However, now I am writing this blog to hopefully save you some of the headaches I have acquired. The other stipulation NJ puts out there is that we must use internal standards (ISTDs) and that their concentration needs to be 10 ppbv. So here comes the digression…

Photo courtesy of nj.com

Since we are required to use ISTDs and our calibration is based on RRF, we obviously need to have some tight RSDs on our ISTD injections. Now I will spare you the details on how many different iterations I ran on the “Pre Sampling” ISTD loop fill parameters to optimize my system (or any of the other parameters for that matter), but those are the parameters I found to work best on my system. It is important to note that no two preconcentration systems are identical… they all have slightly different personalities (and yes this holds true for all manufacturers preconcentrators as well). With that being said, the parameters previously laid out should get you in the ballpark. Note: Markes was super responsive and got me custom software to load the ISTDs onto the CIA trap before the sample is loaded, hence the “MFC Sampling IS First” method, because I wanted my ISTDs to go along for the entire sample ride from start to finish and act more as surrogate standards. This is now standard in the Markes control software. FYI – This also improved my calibration RSDs.

Ok moving on. Markes shipped me my CIA with a 1.0 mL sampling loop, which is responsible for delivering the ISTDs to the CIA’s trap. In order to obtain the very best ISTD RSDs and consequently premium RSDs on the calibration RRFs I replaced the 1.0 mL sampling loop with a 5.0 mL sampling loop. Note: I would imagine Markes would be kind enough to ship you your CIA with a 5.0 mL sample loop already installed, as this modification is not a trivial task recommended for everyone. However, this modification accomplished the following:

1. With my standard injection volume of 250 mL, which we will talk about more next time, one would have to inject 2.5 ppmv of ISTDs on the 1.0 mL sampling loop to attain the required 10 ppbv concentration. Well this is simply just a pain, because the highest non-custom ISTD mix we offer is at 1.0 ppmv and obviously you only buy your calibration gases from us. For the record, we can easily provide you with the 2.5 ppmv ISTD mix, so you can utilize the 1.0 mL sampling loop. However, I prefer to take my 1.0 ppmv ISTD mix and dilute it into a 6 L canister down to 500 ppbv, which gives me 10 ppbv with my 5.0 mL sampling loop. This also helps to ensure that I never lose an entire expensive bottle of ISTDs by directly hooking the ISTDs mix to the CIA and having a leak (not that this ever happened to me folks… a friend of mine told me about this). AND… this is important, hence the CAPITALS, I can add a regulator to the 6 L canister so I can confidently run my ISTD mix at 4 psig. I do not have the time or space to explain why the pressure matters, but long story short you cannot have your sample loop under pressure at the time of injection (unless super consistent, which is difficult to attain). So take my word and understand that the aforementioned system works very well.

2. As with any analytical instrument, there is some variability. So let us say we have a ±10 µL swing from injection to injection on a 1.0 mL sample loop (remember the pressure effects I alluded to earlier). Well that 10 µL variability would represent 1 %; however, the same 10 µL fluctuation on a 5.0 mL sample loop would only represent 0.2% inconsistency. So in theory, the 5.0 mL sample loop minimizes the RSDs across ISTD injections, and in practice it does. Obviously, I acknowledge we are not talking about orders of magnitude here, so you will be perfectly fine with a 1.0 mL loop. I am not going to show you the data, but again you will have to take my word for it.

So in the end the “MFC Sampling IS First”, the 5.0 mL sample loop, with 500 ppbv of ISTDs regulated at 4 psig, and the parameters laid out before get you ISTD RSDs of 1% or less across 9 injections. Now with our ISTD precision locked down we can move onto the calibrations, which we will cover next time, so stay tuned…