Mary Snider, Chemist, Catalent Pharma Solutions, Somerset, NJ
Prof. Richmond Sarpong, Ph.D., Department of Chemistry, University of California, Berkeley
David Norris, President, David Norris Analytical, Kent, UK
Liang Zhao, Ph.D., Research Analyst, Johns Hopkins University, Baltimore, MD
Ravi Orugunty, Ph.D., Senior Scientist, Worldwide Clinical Trials, King of Prussia, PA
Barry E. Boyes, Ph.D., Director of Bioscience R&D, Advanced Materials Technologies, Wilmington, DE
Provide some general impressions about your LC instrumentation
Mary Snider: I like Agilent’s ease of changing lamps and the self diagnostics.
Richmond Sarpong: HPLC is indispensible for unearthing minor impurities that we cannot observe using other techniques such as NMR. We also employ it for conducting enantiomeric excess assays.
David Norris: HPLC is tough and reliable and it works for most of our samples. UHPLC is probably great for analysis of pesticides in drinking water, but for us it is lacking for difficult separations.
Ravi Orugunty: The Shimadzu Nexera pumps are robust; the Autosampler SIL 30 uses a unique method of sample delivery using two different valves – a low pressure valve that is part of the injection port, and a high pressure valve that contains the sample loop and the entire UHPLC setup. The system has extensive capability for washing the external surface and internal surface of the needle to remove carryover, as well as port washing. This system allows us to study carryover in an analytical method and take steps to eliminate it.
UHPLC provides faster sample run times, improved sensitivity, and lower solvent use. But there are tradeoffs in terms of the amount of material we can load onto the column, plus band broadening. Solving these issues requires deep knowledge of the system configuration, dead volumes, pump and mixing efficiencies, and the correct choice of column. In other words, there definitely is a learning curve with UPLC systems.
Barry E. Boyes: I like what works with high reliability and performance and low maintenance. Most modern HPLC systems show low extra-column band dispersion, which saves me the trouble of refitting them for our applications. This becomes critical for LC/MS work, as we are always fighting interface issues for post-column effects. A few microliters of bad design can turn a very useful separation into blobs!
I also like that current pumps, injectors and detectors come from the factory with appropriate fittings and tubing, and work pretty much without issues with the software and data handling capabilities.
Almost all newer HPLCs and UHPLC systems have the ability to monitor themselves for maintenance and repair. This is important, particularly with instruments that track by module identity, since components get moved around between instruments.
Q: How can LC workflows be improved?
A: Mary Snider: We have a lot of problems running Chemstation with Chemstore data management software. For example, Chemstation does not link the column to the method unless you are using the ID tags, and requires us to calculate signal to noise manually. I would also like Agilent to improve the autosamplers on its 1200 and 1260 models and improve reliability. Our well plate autosampler has required repair twice in the last year.
Richmond Sarpong: The improvements I envision are not necessarily with the instruments. I would like to see columns that are more robust and do not fail easily after a short period of use. Furthermore, instruments that are easier to clean and dismantle the parts easily (i.e., modular) to allow routine trouble-shooting would be great. I would also welcome more user-friendly manual processing.
David Norris: I would like to see software that makes manual data processing easy again. Auto-integration, which all the software packages seem to assume, is not for every user and method.
Liang Zhao: If any aspect needs improving it would be columns. I would like to see new column types with better selectivity for specific analytes, for example glycopeptides and other post-translational modifications of peptides. I would also like Internet connectivity so I can monitor instrument performance remotely.
Ravi Orugunty: In a CRO business model the ultimate driver is the number of samples that we analyze per unit time. I personally would like to see multiplexing UHPLC systems that are easily deployed for use in a GLP environment. In addition, almost all HPLC systems that I have seen in the market have very limited ability to store the pressure and other operational data that is provided by the various sensors on the HPLC systems. This data is critical as it provides valuable insight into column durability and catastrophic failure events. It also helps in trouble-shooting.
Barry E. Boyes: The sticker price of LC systems has gotten out of hand. LCs should not be $100,000 instruments! I recently helped out some students on “optimizing” their instruments, and the tubing, fittings and minor accessories costs were thousands of dollars. A small piece of tubing with fittings should not cost $150, nor should the cost of accessories, tubing, and fittings be punitive.
Reconfiguring all of the various late model LC instruments to move between applications is more complex than it needs to be. Changing injection sample volume ranges should not require pulling the manual every time, then following a 23-step process.
The software systems for automated data collection and analysis are still too convoluted and specialized for many normal humans. Across platforms, I have limited problems learning new menu structures, but teaching it to a novice is unreasonably painful. Most people are PC conversant, and I wonder how isolated from reality some of the software engineers have become.
By the way, I have had two very bad high pressure LC systems in my lab that were launched before their time. There is another very popular system that I won’t have in my lab as the plumbing and internal design is so unrealistically complex. My rule of thumb is that if I can’t fix 90 percent of the likely problems, then I don’t want it.
Q: Would you consider switching to or expanding your capabilities through acquiring UHPLC instrumentation, or switching more of your methods to UHPLC?
A: Mary Snider: Because we’re a contract lab and employ validated methods, we cannot use UHPLC. If we were given a validated method to transfer that required a UHPLC, we would purchase one.
Richmond Sarpong: We have not had any exposure to UHPLC, so it is an unknown quantity for us. We have been happy with the capabilities of HPLC for our purposes.
David Norris: We must use both UHPLC and HPLC as many of our methods cannot convert from HPLC to UHPLC. We got our first 1200 Series UHPLC three years ago. It does an acceptable job for easy assays, but for impurity assays we often must overload the column to see minor impurities, and peak shapes suffer badly for the on-scale impurities. In our hands, UHPLC lacks reproducibility. We have found storage stability studies to be quite difficult, as the instrument does not hold retention time over long periods if other assays are conducted in the interim. In other words, storage stability studies conducted at 3-6-9-12 and 24 months require dedicated columns, which is not the case with conventional reverse-phase columns. UHPLC columns have such low capacity that we are unlikely to see impurities we look for without using highly sensitive detectors such as MS-MS-MS.
Liang Zhao: If shorter run times and cost made sense for our lab, we would consider switching to UHPLC. HPLC makes more sense for large-volume injections, whereas for low-concentration samples (for example proteomics) one should probably select UHPLC.
Ravi Orugunty: This is a tricky question. The answer depends on the analyte, how it separates from the matrix, and the presence of interferences. If the analyte(s) show good behavior under classic reverse phase conditions, then it is reasonable to expect that the method would translate to a UHPLC system. Analytes that are chromatographed using ion exchange or HILIC I tend not to try on UHPLC, though in the future we may have UHPLC columns for these modes as well.
Q: For which samples/workflows do you use HPLC and UHPLC?
A: David Norris: We originally used UHPLC for storage stability and HPLC for chemical profiling, but we are beginning to convert our standard methods back to HPLC
Ravi Orugunty: Workflows that involve very clean extracts (such as liquid liquid extraction or solid phase extraction of analytes from a biological matrix that provide a reasonably clean sample can usually be incorporated into UHPLC methods. If the extracts for the analysis are very dirty, I would usually go with HPLC conditions. This usually avoids catastrophic column failures, unusually large retention time shifts due to secondary chromatography of matrix components, such as phospholipids, on the column.
Barry E. Boyes: UHLPC instruments are employed for higher flow rate separations, and when coupling several columns in series to take advantage of 100,000 plate-plus separations for very complex samples. I personally find the popularization of UHPLC as an acronym kind of irritating. I suppose that this is better than saying “low band dispersion contemporary HPLC instruments with extended pressure tolerance.” Oh well, that’s marketing in action! Since the instruments have gotten so pricey, I tend to reconfigure the instruments for various workflows (small molecules, large, complex, simple), rather than dedicate an instrument for a particular need.