USLC columns put the right tools in your LC method development toolbox

- By: ,

Courtesy of Restek Corporation

Understanding Selectivity in Reversed Phase Separations – A Simplified Approach to HPLC and UHPLC Column Selection

The most significant influence on chromatographic peak separation, or resolution, is column selectivity. Unfortunately, column selectivity is also the least understood and most underutilized parameter. To improve selectivity, method developers often concentrate on manually altering mobile phases, operational parameters, and instrumentation. But because stationary phases offer more significant selectivity differences, you can drastically speed up HPLC and UHPLC method development by instead focusing on column choice. In this article, we discuss column selection for reversed phase separations and, using the hydrophobic-subtraction model (H-S model), identify a set of just 4 stationary phases—Restek’s USLC column set—that encompasses the widest selectivity range available on the market.

The Role of Selectivity in Liquid Separations
When performing a liquid separation, we generally focus on choosing the right instrumentation—especially since the recent advent of UHPLC—and end up choosing columns rather hastily, either by proximity (using the column that is already on the instrument or in the closest drawer) or by habit (using a column that has offered problem-free service in the past). While never optimal, this practice should be particularly concerning for a method developer because improper column choice can lead to needlessly labor- and time-intensive method development. If we consider the impact of column selectivity on peak separation, or resolution, we can see why choosing the right column can be so advantageous.

Resolution is the result of 3 cumulative terms: efficiency (N), retention capacity (k), and selectivity (α). How well we resolve our analytes, and how quickly we do so, depends upon our ability to control these 3 factors. Of the 3, the selectivity term mathematically affects resolution to the greatest degree (Equation 1). Put another way, resolution is largely a function of selectivity.

Changing Columns to Create Significant Changes in Selectivity
Since resolution is largely a function of selectivity, any discussion about improving resolution should focus primarily on altering selectivity. It has often been taught in HPLC method development that one can effectively alter selectivity by adjusting mobile phases to reach a desired separation. This, of course, is true. However, mobile phase adjustment can be laborious—often involving many preparation adjustments and column equilibration times—and typically creates only marginal selectivity differences. In addition, some elution profiles are not practical with certain mobile phases and detection modes, including mass spectrometry (MS) and refractive index (RI).

On the other hand, changing stationary phases (i.e., columns) can be much easier and can also result in more significant selectivity differences because stationary phases can offer alternate and even orthogonal separations. These alternate separations can also be scouted very quickly using precise scouting gradients.

With the number of columns commercially available today, choosing the right one can be difficult, even overwhelming. By quantifying stationary phase selectivity, we can create new guidelines for effectively and easily choosing columns to help reduce method development time and increase method ruggedness.

Quantifying Column Selectivity Using the Hydrophobic-Subtraction Model (H-S Model)
Many models exist for choosing solvents and mobile phase additives, but not until recently has stationary phase characterization received much attention. Column selectivity has been largely overlooked due, in part, to its complexity, particularly for liquid separations. But now that Snyder et al. have proposed their popular hydrophobic-subtraction model (H-S model) [1], we can begin to compare and quantify stationary phase selectivity in reversed phase separations and determine (often through orthogonal separations) which stationary phases produce the greatest degree and range of selectivity differences. Only then can we identify a small set of columns that will form the contents of an efficient and effective method development toolbox.

The H-S model is a novel treatment that empirically defines reversed phase selectivity by analyzing a varied collection of solute test probes and then utilizing 5 established selectivity parameters—hydrophobicity (H), steric hindrance (S*), hydrogen bond acidity (A), hydrogen bond basicity (B), and cation exchange activity (C)—to identify the contributions of silica sorbents and stationary phases on selectivity. This model has been used by many organizations, including United States Pharmacopeia (USP), to find column equivalency.

The selectivity value (Fs) of the H-S model is normally used to find the similarity between columns, but it can conversely be used to find column dissimilarity, even orthogonality, to highlight selectivity differences and simplify column selection. Table I compares a variety of stationary phases and reveals which phases offer increased selectivity. (Because the H-S model evaluates the contributions of both stationary phase and silica support on selectivity, we intentionally kept the silica support constant throughout our experiments to isolate the effect of stationary phases on selectivity.) Each value was calculated relative to a C18 benchmark. The columns showing high Fs values—like the 4 Restek USLC phases shown in blue—exhibit the greatest dissimilarity in selectivity relative to the C18, so they are excellent choices when a C18 does not provide the selectivity needed.

Characterizing Selectivity at the Molecular Level
Often during method development, after we have made our initial column choice, we still find ourselves struggling to resolve compounds as we try to find a “better” column. This difficulty is often due to an inability to find a column with alternate selectivity. Quantifying stationary phase selectivity (Table I) is a very important step in identifying a small and effective column set for method development, but we must further define selectivity at a molecular level to ensure that the columns in our method development toolbox exhibit not just high selectivity, but also alternate selectivity based on potential analyte types.

Selectivity (α) is practically determined from the difference in retention factors (k) of 2 peaks. Therefore, to produce alternate selectivity, we must alter the retention of one peak relative to the other. (Increasing the retention of both peaks equally results in higher retention capacity, but no change in selectivity because the difference between the 2 peaks does not change.) If we focus column selection on intermolecular interactions, we can see how specific phases create selectivity by altering the retention profile of specific solutes in relation to others—true selectivity.

So before we can confirm alternate selectivity, we first need to characterize the types of intermolecular interactions commonly encountered in reversed phase chromatography (RPC). In our experiments, we measured 4 major types of interactions—dispersion, polarizability, hydrogen bonding, and cation exchange. To further simplify things and more easily define a guideline, we can relate these measured interactions to chemical properties as noted below:

  • Dispersion is the term for the van der Waals interactions that exist to some extent in all organic molecules, including polar molecules. It is the major driver for RPC and is a major retention mechanism for alkyl phases (i.e., C1 through C18). Since the retention is proportionate to the hydrophobicity of the molecule, we can call these interactions hydrophobic retention.
  • Polarizability is the ability of a stationary phase to change its electron distribution in the presence of an analyte and induce a dipole interaction. It is commonly seen in phenyl-based columns and is the main reason we often switch from a C18 to a phenyl to find alternate selectivity. The Restek Biphenyl column has 2 phenyl rings to enhance polarizability. These interactions are most commonly seen in dipolar, unsaturated, or conjugated compounds and fused-ring compounds with electron withdrawing groups (like nitro groups). For our purposes, we will define these interactions simply as dipolar retention.
  • Hydrogen bonding is used in RPC when a solute and a stationary phase form a chemical bond in which a hydrogen atom of one molecule is attracted to an electronegative atom, especially a nitrogen, oxygen, or fluorine, of another molecule. Although hydrogen bonding results in retention of other solute types, we will focus on its ability to increase retention for acidic compounds and will call it acidic retention.
  • Cation exchange is an electrostatic interaction between a cationic solute and an anion within the stationary phase. Cation exchange, or electrostatic interaction, is most commonly employed in RPC for the retention of protonated bases. Therefore, for simplicity, we will call it basic retention.

Extending the H-S Model to Simplify Column Choice
To determine a simplified guideline for column selection, Restek has extended the H-S model by analyzing empirical selectivity data of our stationary phases (Table I) against the RPC molecular interactions described above (Table II). Through matching stationary phases to specific solute types based on these measured intermolecular attractions, we can aid method development in 2 significant ways: First, we can find a small set of columns with a wide range of alternate selectivity for use in method development. Second, we can define a process for selecting columns based on the chemical properties of our analytes when scouting column selectivity.

Extrapolating the retention data for the solute probes in the H-S model allows us to correlate the retention characteristics of specific solutes to stationary phase types. Ultimately, this correlation has enabled us to match column type to the selective retention of our analytes’ chemical properties, making column selection truly definable by the chemical composition of our analytes.

Figure 1 illustrates the retention profile of a C18 compared with the profiles of the 4 Restek Ultra Selective Liquid Chromatography (USLC) columns. We can see changes in selectivity across these columns as illustrated by the circled areas showing heightened retention for particular solute types. (Selectivity is the retention of one solute relative to another.) The 4 USLC columns exhibit varied retention profiles based upon solute type and, therefore, will exhibit alternate selectivity relative to one another. Because we have a small, quantified column set—4 Restek USLC phases—that is highly selective and exhibits significantly different retention profiles based on specific solute chemical properties, we can now match columns to specific analytes and, thus, simplify method development.

Figure 1: Stationary phase selectivity can be determined by looking for column types with varying retention profiles. When compared to a C18, the 4 Restek USLC phases offer diverse retention profiles—that is, a true range in selectivity.

Confirming the Alternate Selectivity of the USLC Column Set
To further confirm that each USLC column provides alternate selectivity—not only when compared to the C18 benchmark, but also when compared to the other columns in the set—we quantified the column set’s range of selectivity (S) as described by Neue et al. [2]. Looking at the retention characteristics of the H-S model solute probes, we can define selectivity as the degree of scatter along the regression line when comparing stationary phases to the conventional C18 benchmark (Figure 2).

Two very similar stationary phases will produce similar retention for the solute probes and, when graphed, will show high linearity and high correlation. Two very dissimilar, or alternately selective, stationary phases that differ in the retention of the solute probes will show a high degree of scatter around the regression line. More scatter reveals that columns are more different, or orthogonal, from one another because it shows larger differences in selectivity. To measure this difference and use it as a means of comparing stationary phases, we can calculate a selectivity (S) value for the columns in the USLC column set. Note that because silica and mobile phase contributions could also alter the retention of the test probes, it is important to use identical silica supports and mobile phase compositions as to not bias the results and to allow focus only on the stationary phase contributions to selectivity.

With a selectivity value (S) of 46.7, Restek USLC phases produce an incredible range of alternate selectivity —using only 4 columns.

Conclusion: The Right Tools for Maximum Selectivity
The H-S model offers the chromatographic method developer a practical approach to column selection. With a simplified model described above, we can now easily create predictable and alternate selectivity, effectively influencing the most significant factor contributing to resolution. Now that we have identified the small USLC column set with a wide range of quantified selectivity, we can quickly determine the best column for nearly any instrument platform and reversed phase or HILIC application by referencing predefined retention profiles. This column set can also be used to get the most out of column switching by providing a functional column set.

The Restek USLC column set, consisting of a balanced Aqueous C18, a Biphenyl, a fluorinated PFP Propyl, and a polar embedded IBD, has a profile that encompasses the widest range of reversed phase selectivity available today. Putting the right tools—like the USLC column set—in your method development toolbox means maximum alternate selectivity and peak separation with minimal effort.

The authors gratefully acknowledge the contributions of Dr. Lloyd Snyder from LC Resources and Dr. Frank Dorman from The Pennsylvania State University. The authors also wish to thank the contributing team of researchers Randy Romesberg, Bruce Albright, Mike Wittrig, Brian Jones, and Vernon Bartlett.

Customer comments

No comments were found for USLC columns put the right tools in your LC method development toolbox. Be the first to comment!