The LCGC Blog: From HPLC to LC-MS: Mobile-Phase Composition is the Main Consideration
I recently delivered a couple of lectures in a high performance liquid chromatography (HPLC) and liquid chromatography–mass spectrometry (LC–MS) workshop at U.T. Arlington. For my first talk, I discussed practical considerations for coupling HPLC to MS detection. It is the type of talk I have given many times before, but of course it is always good to refresh critical talking points. I remember that this was a really crucial topic when I was in graduate school (circa 2000), at a time when commercial electrospray ionization (ESI) MS systems from a wide variety of manufacturers were being promoted. Everyone was concerned with transferring their HPLC methods featuring UV detection to LC–MS. Overall, I think the rules for selecting mobile-phase compositions that are LC–MS compatible are pretty simple. I also think that based on the possible choices of mobile-phase additives, there is a lot less variation in mobile-phase composition for LC-MS methods, and this may have diminished some of the elegance associated with HPLC method development.
I am definitely not the best person to speak about all of the possibilities that are associated with HPLC–UV method development. I never quite understood all of the subtle choices in seemingly complicated HPLC–UV mobile-phase compositions that featured phosphate buffers and other ingredients that are quite foreign to LC–MS practitioners. Even today, one who wants to develop an HPLC–UV method probably wants it to be eventually compatible with LC–MS use, so they will restrain themselves to those relatively few ingredients that are MS compatible. As I sit and write this, I am inclined to ask readers if they would submit their comments associated with two related questions that I have: What are some of the most intricate non-MS-compatible mobile-phase choices (and applications) that you can recall having used to achieve a particular separation by HPLC? Do you think that some of art associated with HPLC method development has been lost based on the requirements for MS detection, or has this need been largely offset by the current availability of more efficient stationary phases?
Let us step back for a minute and bring everyone up to speed. Researchers spent a great deal of effort over many years trying to effectively couple HPLC to MS. Gas chromatography–mass spectrometry (GC–MS) advanced much more smoothly because of the availability of capillary-scale separations and the fact that analytes were already in a gaseous state when they reached the ionization source. For LC–MS to work, the main challenge was to segregate ionized (and typically nonvolatile) analytes from a large amount of solvent. Vacuum technology simply could not handle the huge volume of gas produced from the vaporization of all of the mobile-phase solvent. We need a vacuum in the mass spectrometer to achieve a significant mean free path for an ion. Ions have to be able to travel from source to detector without running into other stuff that might deactivate them. A simple calculation can tell you that even if you have water flowing at 0.1 mL/min, at atmospheric pressure, when vaporized, this is equivalent to 136 mL/min of gas flow. When reduced to the operating pressures of a mass spectrometer, this volumetric rate increases to somewhere in the range of 23,000 to 230,000 L/s! Typical MS vacuum pumps can remove gas at a rate of 50–1000 L/s, which is not directly compatible with typical HPLC flow rates. Various techniques were developed and investigated, including moving belt and particle beam interfaces, to try to remove solvent before or during ionization (1). The atmospheric pressure ionization techniques, which rely upon the volatilization and removal of solvent before ion sampling, were eventually able to solve the problem (2).
For the practitioner, development of modern LC–MS methods requires a choice of appropriate volatile solvents and additives in the mobile phase. Let’s confine ourselves at this point to coupling reversed-phase HPLC separations to ESI-MS. One can find a variety of resources on the web listing ESI-compatible solvents and solvent modifiers. Mainstream methods are typically limited to aqueous mixtures of methanol or acetonitrile, where water represents the weak solvent (a higher amount promotes retention) and methanol or acetonitrile represents the strong solvent (a higher amount promotes elution). These solvent mixtures provide adequate volatility and conductance to facilitate ESI. Modifiers used are generally acetic acid, formic acid, ammonium acetate, and ammonium formate. The former two alone can be used to suppress the ionization of acidic compounds, promoting chromatographic retention, whereas the incorporation of the latter two can create buffer systems to better stabilize pH. For ESI, presence of the acids can help facilitate formation of positive ions, whereas the presence of a buffer stabilizes ion production and can improve reproducibility. All are volatile and can be effectively removed in the ion source. The problem with using nonvolatile buffers and salts (for example, phosphates) is that they will be deposited at the inlet of the mass spectrometer and prevent entry of the ions.
There are other mobile-phase modifiers that are also MS compatible. If basic conditions are needed, then ammonium hydroxide can be used; however, one should be careful about the pH compatibility of the stationary phase (some silica begins to dissolve at pH 8–9). Volatile ion pair reagents, such as trichloroacetic acid, trifluoroacetic acid, triethylamine, and trimethylamine, can also be used in limited quantity (< 0.05%). These can sometimes aid the separation process for various analytes, but if they are present in too high of quantity, they can suppress ionization. Note that in LC–MS methods strict avoidance of trifluoroacetic acid, which is pretty useful for attaining low pH values in the mobile phase, is a common misconception. You just want to keep the concentrations modest.
I do not wish to oversimplify a science that requires significant thought and practice. Many times, there are compromises between those modifier choices or amounts that are better for separations, or are better for ion generation. For a given application, these effects must be carefully weighed. For reversed-phase LC, there are significant differences in types of stationary phases, and this is also a consideration, but it is not nearly as crucial as in hydrophilic interaction liquid chromatography (HILIC), for example (a topic for another time). I suspect that readers who comment, and even those who do not, will hold the opinion that today’s stationary phases for reversed-phase LC solve a lot of the difficulty in finding appropriate mobile phases to achieve efficient separations; not as many variations are needed to achieve excellent performance compared to a couple of decades ago. Optimizing choices of modifiers in LC–MS methods then is governed more by the ion response generated. Even so, I wonder if some of the art associated with HPLC–UV method development in past has been lost (did I just miss the boat?). However, maybe this is also a good thing. Developing an effective method is much more accessible to a much wider base of practitioners these days than ever before.
(1) J. Abian, J. Mass Spectrom. 34, 157–168 (1999).
Previous blog entries from Kevin Schug: