Popular Reflections: ProEZGC – Modeling on the Go
In 2019’s “A GC in Your Laptop,” Jaap de Zeeuw discussed ProEZGC, a web program that simulates separations allowing trainees or potential users to “play” with gas chromatography optimization. Here, we catch up on the easy GC journey.
Are you still using the ProEZGC web program to help students and inexperienced users?
Yes – and it is still spreading. I’ve been presenting workshops at universities for the past four years. These are two 90 minute sessions, which include exercises. Session one is method translation and optimization. Using the same stationary phase, the goal is to get the exact same chromatogram (peak elution order), but faster (for example by trading in some efficiency for speed) or at lower operational cost (like using N2 instead of He or shorter columns). I offer them five different situations in which they have to use the method translation software and they must choose from multiple-choice answers. Session two is modeling using the ProEZGC. Here, the phase is not defined; we want to find the phase that offers the best separation. After the phase is found, students can experience the impact of flows, carrier gas type, temperatures, and column dimensions. Normally in the lab, a GC needs to be configured – you need to purchase different columns and you need someone on the job full time. If that’s all available, one can do 4–5 experiments in an afternoon. Now, with the modeling, one can do 400–500 experiments! In this session, they have to experience what happens with some peaks if you change program conditions, choose the best stationary phase for a set of 10 analytes and get the shortest possible run times by working with the options that the modeler provides.
We can measure the number of searches performed and we see a constant increase in usage. I also know that some universities have already adopted the program in their chromatography classes. I spoke with a university course director who wanted to pay the licence fee for using the program. He was amazed and happily surprised that there was no fee! In addition, Restek hosted a student from the latter university as part of his undergraduate thesis. He spent the summer performing experiments for the further development of ProEZGC at Restek’s headquarters in Bellefonte, Pennsylvania, USA.
Have there been any advances in separation simulation technology since you wrote the article in 2019?
The accuracy of the libraries have been improved and over 2000 compounds and half a dozen different new columns have been added. We also worked on the addition of PLOT columns. We did a poster recently on the method translation for alumina columns using H2, Ne and N2. The matches were very good, so it is likely that modeling is also possible.
In the article you listed the limitations of ProEZGC (for example, you can only model the components that are in the database, the program displays the components as Gaussian peaks, the program assumes ideal injection). Do you think the program could eventually simulate some of these non-ideal conditions and variables?
Yes – we believe that, just like retention time indexes, peak width indexes could be entered using empirical data to further expand the capabilities of ProEZGC.
With regard to training and education, has the COVID-19 pandemic changed how you view the importance of modeling technology?
For me, COVID-19 has opened a new world of teaching. I did hundreds of webinars and never could reach so many people in so many countries.
The cool thing is that, with ProEZGC, you can do modeling anywhere – indeed, you have a “GC in your laptop” so you can do a lot of work at home or in isolation. I had access to a prototype eight years ago when I was at a conference in Singapore. During a discussion, someone stated that “nitrogen is a dinosaur gas” and should not be used – and I felt challenged to prove the opposite! N2 has a low optimum velocity but it’s cheap! So, I used the modeler to simulate separations with N2 and found that you can replace a standard 30m x 0.25mm using helium for a 20m x 0.15mm using nitrogen and get exactly the same separation, in the same amount of time! I remember it was on a Saturday and I sent an email to my colleague Jack Cochran, who was in the US. He was curious and went in the lab on Sunday, took the most challenging column combination (using a Stabilwax column – when it works with a polar phase it will for sure work with a non-polar phase). He came back the next day sharing data that showed it was a perfect match. Later, this was tested with four other phases and four different component groups and it always worked.
Is there anything else you'd like to mention?
I am still excited about working with and presenting the program. I would like to see an option to show a T0 peak (methane), as well as a way to add some different peak heights for close eluting peaks. We can also consider different detectors and responses. If I can choose a solvent type, this peak could also be displayed. Also, the initial oven temperature depends on the solvent used and especially if I use the splitless injection technique.
Finally, I’d like to thank Chris Nelson and Chris English for their input.
