Mobile Phase Buffers in Liquid Chromatography: The Buffer–Detector Battle

Buffered solutions that resist changes in pH are very important to various aspects of the practice of liquid chromatography. Unfortunately, the buffering additives that are best for chromatographic performance sometimes cause problems with detection, such as baseline drift and noise, or decreases in sensitivity. In this installment, I primarily discuss considerations to keep in mind when choosing buffering additives that will be used for LC methods involving UV absorbance detection. Keeping these ideas in mind at the start of the method development process can help prevent problems from developing during the lifetime of a new method, as well as help troubleshoot problems with existing methods that are buffer-related.

In last month’s installment of “LC Troubleshooting” (1), we reviewed several essential ideas related to the use of mobile phase buffers in liquid chromatography (LC). We touched lightly on several topics ranging from how to choose a buffering additive to how to prepare the buffer solution (and whether to filter it before use), and then provided some representative references in each of these areas for readers who want to dig a little deeper into any of the subtopics. These are topics that we think are often underappreciated in the practice of LC, and an occasional refresher on why we use buffers at all, as well as best practices for their preparation and use, is warranted. In this month’s installment, we take a closer look at the relationship between choices we make related to buffers (for example, the buffering additive itself or its concentration) and the different detectors we use (for example, mass spectrometric [MS] and spectrophotometric detectors) for different applications. Unfortunately, very often a buffer that is a good choice for one application can be a terrible choice for a different one, so knowing how different variables affect the compatibility of a buffer with a given detector is very important.

Buffers and Ultraviolet (UV) Absorption Detectors

The biggest buffer-detector conflict we face when using UV absorbance detection is related to the light-absorbing character of the buffering additive. In the absolute worst case, where the mobile phase buffer itself absorbs a large fraction of the photons emitted by the detector light source, there are few photons left to be absorbed by the analyte moving through the flow cell that we are trying to detect. This can lead to noisy signals and a limited linear dynamic range. It is not hard to find guidance about which buffering additives are or are not suitable for UV absorbance detection with this idea in mind—that we want to avoid situations where the buffer itself absorbs much of the available light. For example, a recent post in the LCGC Blog provides “UV cutoff” values for several common buffers (2). The concept of the UV cutoff is that it is the wavelength at which the absorbance of the buffer solution containing a particular buffering additive is 1.0 absorbance unit. For example, the blog post indicates that the cutoff is 210 nm for formic acid. Typically, the absorbance of the solution increases at wavelengths shorter than the UV cutoff; thus, the cutoff value is a kind of marker that we should stay away from when developing a method. However, the UV cutoff just indicates one point on the wavelength axis, and when these cutoffs are reported, there is usually no sense for the role that the concentration of the buffering additive plays in determining which detection wavelengths can be used. The cutoff value does not give us any idea how steep the slope of absorbance is versus wavelength as we approach the cutoff from the high wavelength side. Therefore, it doesn’t give us any idea how close to the cutoff we can work before running into problems. With this perspective in mind, we can learn some things by looking at the actual absorption spectra of different buffer solutions.

Phosphate Buffers

Figure 1 shows the UV absorption spectra for a handful of common buffering additives in water at concentrations that are typical for LC use. The range of these spectra is pretty dramatic, from the nearly transparent 10 mM potassium phosphate (pH 7; purple trace), to the highly absorptive 25 mM ammonium formate (pH 3.2; short-dashed black trace). The very low absorptivity of the potassium phosphate solution, even down to 210 nm, show why it is so attractive when using UV detection. Not only does the transparency leave a lot of photons to be absorbed by the analytes we are actually trying to detect, but it also allows the use of solvent gradients involving these buffers without having to worry about a shifting baseline due to changes in the fraction of the mobile phase containing the buffering agent. For example, Figure 2a shows the UV absorbance baselines at wavelengths of 210 nm and 254 nm obtained when running a gradient from 5 to 70% acetonitrile (ACN) (95 to 30% 10 mM phosphoric acid in water) over 5 min. We see that the absorbance values change less than 2 mAU from one end of the gradient to the other at both wavelengths, again because the buffer solution is similarly transparent at these two wavelengths. However, buffers containing phosphates are nonstarters for mass spectroscopic (MS) detection, as well as evaporative light scattering detectors (ELSD) and charged aerosol detectors (CAD). The phosphates are not at all volatile, which immediately leads to problems if we try to use them with these other detectors.

FIGURE 1: UV absorption spectra for dilute aqueous solutions of additives commonly used for liquid chromatography.

FIGURE 2: UV absorbance baselines (blue traces) observed during solvent gradients (black traces shows composition changes) from 5–70% ACN as is typical in gradient elution LC. (a) Buffer solution is 10 mM phosphoric acid in water; (b) Buffer solution is 0.1% trifluoroacetic acid in water. LC Conditions: Flow rate, 0.5 mL/min; Column, 30 mm x 2.1 mm i.d. C18; Flow cell path length, 10 mm.

Solutions of Carboxylic Acids

Next, consider the red and blue traces in Figure 1, which correspond to solutions of 0.1% formic acid (FA), and 0.1% trifluoroacetic acid (TFA) in water, respectively. Here, we see that the absorbance of the 0.1% formic acid solution rises to 1.0 AU around 210 nm, which is where the “cutoff” value of 210 nm for formic acid comes from. Interestingly, the absorptivity of trifluoroacetic acid is quite a bit weaker at around 0.5 AU at 210 nm. This is most likely due to the fact that the molecular weight of TFA is approximately double that of FA, and thus, for the same mass of additive in the mobile phase, there are about half as many moles of TFA in solution compared to FA. Figure 2b shows that when we run a solvent gradient with using trifluoroacetic acid as the buffering additive, we observe a very steep baseline slope when detecting at 210 nm. In practice, this can be largely mitigated by adding the acid to both the aqueous and organic solvent components of the mobile phase. Users interested in employing this trick are referred to reference (3) for a more detailed discussion of how to do so, as there are several details that deserve attention when trying to squeeze the most performance out of a system like this. We must also be careful when adding acids to alcohols if they are used as the organic solvent, as this can result in ester formation in the mobile phase over time (4). The spectra for the formic acid and trifluoroacetic acid solutions in Figure 2a show that they are pretty transparent at 254 nm, and even down to approximately 230 nm for trifluoroacetic acid, but as we approach 210 nm, we should expect absorption in excess of 100 mAU for both solutions.

Buffers Prepared from Ammonium Salts

The last four solutions represented in Figure 1 involve ammonium salts of organic bases. First, we see that the 25 mM ammonium formate solution (pH 3.2; short-dashed black trace) has the highest absorption of any of the solutions represented in the figure. This is because the formate concentration is much higher in this solution than in the 0.1% formic acid solution (the formal concentration of formate in this solution is about 22 mM if the acid is added on a w/w basis). Preparation of the ammonium formate solution begins with dissolving 25 millimoles of ammonium formate salt in about 900 mL of water. The pH of this solution is about 6.5. Decreasing the pH to 3.2 (without the addition of a strong, nonvolatile acid such as hydrochloric acid) requires 80 millimoles of additional formic acid, bringing the total number of millimoles of formate added to 105. This concentration of formate is much higher than the 22 mM in the 0.1% formic acid solution, which leads to much higher UV light absorption. This is a good example of the importance of understanding the details associated with the preparation of specific buffer solutions. However, the series of ammonium formate solutions at different concentrations (all black traces, but different line types) shows the importance of the additive concentration when considering UV cutoff values. For the 25 mM solution, the cutoff occurs around 230 nm, but for the 10 nm and 5 mM solutions, the cutoffs are around 220 nm and 210 nm, respectively.

The ammonium carbonate solution at pH 9.2 is an interesting one. It is quite transparent to UV light, at least down to 220 nm. Even at 210 nm, its absorbance is about six times less than that of 0.1% trifluoroacetic acid. In addition to this good transparency, it has other attractive attributes as a buffering additive. The pK_as of ammonium ion and hydrogen carbonate ion are approximately 9.3 and 10.3, which means that ammonium carbonate solutions provide good buffer capacity in the pH range of 8–11. Finally, ammonium carbonate is sufficiently volatile that it can be used with MS detectors, although a decision to do so should be made with care because the carbonate has been shown to affect the behavior of some analytes, such as proteins (5).

Buffers and Other Detectors

Evaporative Light Scattering (ELSD) and Charged Aerosol Detection (CAD)

The main guiding principle when choosing buffer solutions for use with ELSD and CAD detectors is to focus on buffering agents that are volatile, and tend toward lower concentrations (for example, below 100 mM). Typical lists of suitably volatile buffering additives include organic acids and bases (such as formic acid and ammonia) and salts of organic acids and bases (such as ammonium acetate and ammonium carbonate). The book by Gamache on CAD includes an extensive discussion of the impact of mobile phase additive composition (for example, just formic acid versus formic acid and ammonia) and concentration on the detection sensitivity observed for different compounds with CAD (6).

Electrospray Mass Spectrometric Detection

As with ELSD and CAD, the main guiding principle when choosing buffering additives for use with electrospray MS is also focused on their volatility. However, the situation is frankly more complicated for MS, due to the importance of factors such as adduct formation in the case of MS that simply are not relevant for ELSD and CAD. Given these additional factors, I will discuss details related to additive choices for MS detection in a future installment.

Summary

In this installment, I have discussed the light-absorbing characteristics of some additives that are commonly used for LC. The concept of the “UV cutoff” for mobile phase solvents and additives is useful, but we can gain more insight about the limitations and useful ranges of different additives by looking at full UV absorption spectra and looking at the dependence of the spectra on additive concentration. Adding these insights to our knowledge base can help us prevent problems from developing during the lifetime of a new method, while also troubleshooting problems with existing methods that are buffer-related.

References

(1) Stoll, D. R. Mobile Phase Buffers in Liquid Chromatography: A Review of Essential Ideas. LCGC International 2024, 1 (10), 6–11.

(2) Taylor, T. The LCGC Blog: UV Detection for HPLC – Fundamental Principle, Practical Implications. LCGC International 2018. https://www.chromatographyonline.com/view/lcgc-blog-uv-detection-hplc-fundamental-principle-practical-implications (accessed 2024-12-10)

(3) Choikhet, K.; Glatz, B.; Rozing, G. The Physicochemical Causes of Baseline Disturbances in HPLC, Part I – TFA-Containing Eluents. LCGC Europe 2003, 2, 2–9.

(4) Metzger, B.; Shoykhet, K.; Buckenmaier, S. Chemistry in a Bottle: Ester Formation in Acidified Mobile-Phase Solvents. LCGC N. Am. 2022, 411–416. DOI: 10.56530/lcgc.na.ns9481m3

(5) Hedges, J. B.; Vahidi, S.; Yue, X.; Konermann, L. Effects of Ammonium Bicarbonate on the Electrospray Mass Spectra of Proteins: Evidence for Bubble-Induced Unfolding. Anal. Chem. 2013, 85 (13), 6469–6476. DOI: 10.1021/ac401020s

(6) Charged Aerosol Detection for Liquid Chromatography and Related Separation Techniques; Gamache, P. H., Ed.; John Wiley & Sons, 2017. DOI: 10.1002/9781119390725

About the Authors

Dwight R. Stoll is the editor of “LC Troubleshooting.” Stoll is a professor and the co-chair of chemistry at Gustavus Adolphus College in St. Peter, Minnesota. His primary research focus is on the development of 2D-LC for both targeted and untargeted analyses. He has authored or coauthored more than 75 peer-reviewed publications and four book chapters in separation science and more than 100 conference presentations. He is also a member of LCGC’s editorial advisory board.

Ella Sontowski is a third-year undergraduate researcher working with Dr. Stoll’s research group at Gustavus Adolphus College.