New Location for Manufacturing/Applications Laboratory

I am pleased to announce that MicroSolv’s laboratory division, previously located in upstate New York, has now relocated to Wilmington, NC. Situated in the CREST Research Park of UNCW, this new facility will allow for collaboration between MicroSolv and those at the university. In addition, the new lab is very well-equipped and up-to-date, which will allow our operations to be very streamlined and productive.

                I have spent seven years at the previous laboratory and watched it develop over the years, ever since its very inception. It has been very gratifying to have seen the laboratory acquire new instruments and capabilities while at that location, and I think our new facility here in Wilmington will allow for even more of this kind of growth in the future.

                I look forward to seeing how the new location will enhance our processes in both manufacturing and research applications. Stay tuned for more updates on our progress with the new laboratory!

Eastern Analytical Symposium 2017

Recently I had the privilege of giving a lecture at the annual Eastern Analytical Symposium in Plainsboro Township, NJ. This year, the symposium was held in an elegant new venue, the Crowne Plaza Princeton-Conference Center. I have been attending EAS since 2011 and I think this has been my favorite one so far.  

                My presentation was at 11:00am on Wednesday the 15^th. The subject of my lecture was about a method my colleagues and I had been working on for chromatographic assay and organic impurities detection of chlorpheniramine maleate formulations. The current USP method calls for a time-consuming multistep liquid–liquid extraction procedure for the assay, and no method is provided for impurities analysis. Consequently, there was a lot of room for improvement in bringing these methods up-to-date.

As a basic compound, chlorpheniramine may be prone to prone to chromatographic tailing via undesirable electrostatic interaction with residual silanols present on the surface of some traditional silica-based stationary phases. Hence, the moderately hydrophobic Si–H-based surface of Cogent™ columns was postulated to be advantageous in terms of minimizing tailing contributions from this phenomenon.  

                We encountered a few obstacles in method development. The first involved peak splitting for the first two peaks. I had previously selected a diluent with 50% acetonitrile in an effort to ensure solubility of the more hydrophobic analytes, but this choice had the undesired effect of peak splitting. Reducing the organic content in the diluent to 10% resolved this issue. Another encountered problem dealt with trying to use a high enough extract concentration such that we could detect impurity peaks at the 0.1% level while being careful not to overload the column so much that we lost resolution of the critical peak pair. In this case, adjusting the gradient and using a lower mobile phase concentration of trifluoracetic acid (0.05% instead of 0.1%) helped to improve selectivity enough that resolution was sufficient at the higher extract concentration.

                Despite these setbacks, we produced a nice method for subsequent validation studies and I think the audience was quite pleased with the end result. This was my second year as a speaker at EAS and look forward to the opportunity to possibly present again at future meetings! Hope to see you there in 2018!

Method Validation Tip: Do Robustness Studies First

After method development, I find that a lot of chemists will delve into method validation studies like accuracy, precision, etc. but leave robustness for the end. However, I think it is better to do robustness first. The reason for this is that the results can help you further refine the method in ways that you may not have considered during the method development stage.

                Let’s say for example that you do your method development have a perfectly satisfactory separation where your critical peak pair is baseline-resolved. You move on to validation and go through all the necessary studies (e.g. accuracy, linearity, repeatability, intermediate precision, LOD, LOQ, etc.). Then in the robustness studies you find that alteration of one of the method variables actually leads to a superior separation. This was what I found when I varied the TFA concentration. I chose a 0.1% concentration originally during the method development stage but hadn’t given much thought to it at the time. Then in robustness studies, I found that the critical peak pair, which had been just baseline-resolved with 0.1% TFA, became vastly better separated upon a decrease to 0.05%. Before proceeding with the rest of the validation, I could use the 0.05% concentration in my method.

                I hope this information helps you as much as it helped me. I am always learning new tips and tricks to make chromatographic analytical techniques more streamlined and efficient.

UV Trace Drift in Gradients

Gradient elution mode is an invaluable tool for us as analytical chemists, but the slope of the UV trace that results can present issues. You can understand why this happens by considering how each of the solvent systems has its own UV absorption profile. Some of the drift is due to differences in the refractive indices between the two solvents as well. Because of these differences, the UV readout will change continuously over the course of the gradient, producing the slope that we observe. If it is too steep, it can obscure eluting peaks, reducing sensitivity. It would be ideal to have the slope as shallow as possible if we can do so.

                The other day, I was working on a method development project for the USP that called for low pH. We observed tailing for some of our peaks and speculated that the lower pH was needed for this reason. To this end, we tried an additive of 0.5% formic acid in the mobile phase solvents but found a rise in the UV trace of 200 mAU over the course of our 30 min gradient. In an impurity method, where low detection levels are of critical importance, this degree of noise would clearly not be practical. By using TFA instead, however, we were able to reduce the additive concentration by a factor of five and still obtain the good peak shapes that we sought. With the lower additive concentration, the trace slope had risen by about 20 mAU, a mere tenth of the slope obtained previously! This clearly represented a much more viable option for a prospective USP method.

                You just have to remember a few things with TFA, though. It can be prone to oxidation from the atmosphere, so extraneous peaks or a steep UV trace can result if it degrades. In that case, you would have the same problem as before! Your best bet is to try single-use ampules of TFA, since these will have minimal contact time with air. If you use a resealable bulk bottle, you can try adding a blanket of argon after every time it’s opened to ensure it is kept free of air. This oxidation is a slow process, so it only becomes an issue after storage of the opened bottle. Then of course, another thing to keep in mind about TFA is its MS-incompatibility, as it contributes significantly to ion suppression. So if you develop a nice method for UV-based analyses with TFA, just be aware that its applicability will be more limited than with something like formic acid. Even so, TFA can do wonders for some tailing peaks of basic compounds, so it is a good tool for the analytical chemist to keep at the ready.

                I hope these are some helpful tips for you to try in your own method development process. I was always a fan of TFA for those situations where symmetrical peak shapes may be difficult to obtain otherwise. The effect of the trace slope can be one of the more tricky aspects of a UV-based gradient method. And if you can, maybe also consider using a more gradual solvent gradient. This will have a direct effect on how steep the resulting UV trace is.

HPLC Troubleshooting Tutorial: Problems with the UV Baseline

I was using a Cogent Bidentate C8 2.o™ column in my HPLC method recently when I observed a strange phenomenon. Instead of a normal, flat baseline in the UV monitoring readout, there was a highly erratic pattern; the signal would descend and jump back to its starting point in a recurring “sawtooth” like waveform. Meanwhile, the pressure was stable and all other instrument parameters indicated normal operation. My isocratic mobile phase was fairly simple (50% DI water/ 50% acetonitrile/ 0.1% formic acid) and other columns I tried did not have the issue. No amount of equilibrating seemed to remedy the situation.

                I really wasn’t sure if the column had somehow been damaged but had no way of running a standard QC test if I could not observe analyte peaks amongst the unstable UV baseline. Next I thought maybe there was some contaminant on the column and it needed to be washed out. Initial trials using solvent systems such as 1:1 methanol: DI water proved unsuccessful, and then I had the idea that perhaps there was some trace immiscible solvent in the column and use of an aqueous-based mobile phase would not be appropriate.

                So I switched to pure isopropanol and gave it a try. Miscible with both reversed phase and normal phase solvents, isopropanol can often be an effective choice in general column cleaning. Soon after introducing the new solvent to the column, the UV baseline resolved into a typical, flat signal. Now the real test would be to see how my original solvent system behaved. Upon reintroducing the original mobile phase of 50% DI water/ 50% acetonitrile/ 0.1% formic acid, I was pleased to observe that the baseline had returned to normal.

                How can we account for this behavior? The outcome of the experiment seemed to indicate to me that the column did indeed have some minor amounts of a nonpolar solvent in it, which would create immiscibility issues with the reversed phase solvents I was using. These could have been residual packing solvents and will pose no problems once completely removed.

This just shows you if you think a column may be defective, try to investigate all your possibilities before prematurely concluding that it has been damaged. You can save your laboratory money by getting the most out of each valuable HPLC column instead of replacing it. I find issues like this can crop up in one form or another quite often; a column is suspected to be defective when in reality, there is a simple solution to resolve the issue. So the next time your HPLC column exhibits some unexpected behavior, try to ask yourself what might be causing it and whether it can be fixed.