Beginner's Guide to Convergence Chromatography

This chapter first reviews the terminology used in both CC and then describes what types of samples and analytes are candidates for analysis by the ACQUITY UPC² System. We lay out the role of co-solvents, mobile phase additives, and sample diluents, as well as the effects of pressure and temperature on density and how this affects separations. Finally, we introduce a generic protocol for developing a method.

Terminology

As described earlier, CC is similar to RPLC, but substitutes compressed CO₂ for the weak solvent (mobile phase A) rather than water. Conventional terms such as solvent, co-solvent or modifier, all refer to the primary liquid component(s) of mobile phase B, the strong eluting solvent. Typically, in CC the co-solvent is methanol, but can also be other organic solvents such ethanol, isopropyl alcohol, acetonitrile, or combinations of these. An additive is a salt or liquid added to the co-solvent at a low concentration to improve peak shape and/or analyte solubility. The additive can also influence chromatographic selectivity. Typical additives include diethyl amine, ammonium hydroxide, formic acid, trifluoroacetic acid, ammonium formate, ammonium acetate, or small amounts of water. The appropriate concentration is additive dependent; for example, the addition of water at above 5% risks the formation of a biphasic mobile phase if the other method conditions are not appropriately chosen.

One of the first questions asked about any new analytical methodology is, “Can my sample be analyzed by this technique (convergence chromatography)?” The simplest answer is: if the sample can be dissolved in an organic solvent, it is a candidate for CC. There is no universal answer to this question and some experimental work is needed to confirm. This compatibility with organic injection solvents is very convenient since many sample preparation techniques produce samples in organic solvent (for example, liquid/liquid extraction, solid phase extraction, protein precipitation). The beauty of CC is that these samples, in an organic solvent, can be directly injected onto the ACQUITY UPC² System and do not require the laborious and time-consuming evaporation or reconstitution steps often required in RPLC. Chapter 5 deals with this topic in more detail. For the analytical scientist, it is always useful to learn as much about a sample as possible (Figure 23) – the more you know, the better the chance to develop a robust method. One useful piece of information regarding the solubility of compounds in various organic solvents relates to the partition coefficient (P) (usually referred to as log₁₀ P). The partition coefficient (P) is the ratio of concentrations of a compound in the two phases of a mixture of two immiscible solvents at equilibrium, usually water and 1-octanol (Figure 24). These coefficients are a measure of the differential solubility of the compound between these two solvents. The partition coefficient measures how hydrophilic or hydrophobic a compound is.

In terms of CC, the partition coefficient could help determine whether a target compound is a candidate for analysis by the ACQUITY UPC² System. As a rule of thumb, compounds with log P values between 2 and 9 make suitable candidates for CC.

Co-Solvents in Convergence Chromatography

The co-solvent plays two roles. First, it influences the solvating power of CO₂. Second, it influences the interaction between the analyte and the stationary-phase. Changing co-solvent (for example, methanol to acetonitrile) affects both retention and selectivity. The role of the co-solvent in CC is analogous to that of the strong solvent in reversed-phase LC; CO₂ alone has approximately the eluting strength of heptane.

Table 1, in Chapter 1, shows the eluotropic (eluting strength) series for a range of organic solvents and highlights the four most common co-solvents used in CC: acetonitrile, isopropyl alcohol, ethanol, and methanol. The solvents listed are all miscible with CO₂ resulting in a wide range of available retention and elution strengths.

Co-solvents added to the CO₂ mobile phase generally decrease an analyte’s retention time. When the co-solvent concentration increases, the polarity of the mobile phase changes, decreasing the retention time(s). Figure 25 illustrates the effect of changing co-solvent concentration on retention in an isocratic separation. As the concentration of the strong, eluting co-solvent (methanol) decreases, the retention of the analytes increases. This is the same phenomenon observed in RPLC.

Figure 26 demonstrates the way mobile phase strength changes with different co-solvents. Methanol is the strongest co-solvent and elutes the analytes the fastest. Isopropyl alcohol is weaker than methanol, but stronger than acetonitrile, while acetonitrile is the weakest of the three co-solvents in CC and retains the analytes longest. The same relative type of chromatographic behavior occurs in other modes of chromatography; stronger solvents lessen retention and elute analytes faster.

Different co-solvents can be mixed in CC, changing solvent strength and creating differences in retention. Figure 27 illustrates the effect of adding a weaker co-solvent (acetonitrile) to methanol, for a gradient separation of metoclopramide and related impurities. As the acetonitrile concentration increases, the methanol concentration and therefore solvent strength decreases, and longer retention times are observed. Slight changes in selectivity, improved resolution, and peaks sharpen with a different co-solvent for this separation.

Additives In Convergence Chromatography

As in RPLC, additives in CC improve peak shape and/or the resolution of the separation. In Figure 27, the effects of adding ammonium formate to all mixtures of co-solvents for all four chromatograms is shown. Additives can modify the stationary phase surface or act as ion pairs, changing selectivity. Basic additives tend to improve peak shape for basic compounds and may slightly change the selectivity. Examples of basic additives include ammonium hydroxide, 2-propylamine, and triethylamine. Acidic additives can improve peak shape for acidic compounds, and may change the selectivity. Common acidic additives include trifluoroacetic acid, formic acid, and acetic acid. Figure 28 shows a separation of acidic analytes and, as this example shows, increasing the concentration of the acidic additive improves peak shape.

Changing between different additives can have a drastic effect upon peak shape and retention (Figure 29). For these basic analytes (beta blockers), a methanol co-solvent with no additive results in poorly shaped peaks. The addition of formic acid actually worsens the peak shape. Formic acid also absorbs at the 220 nm detection wavelength, resulting in a sloping baseline. For these strong bases, the addition of ammonium acetate (20 mM) dramatically improves peak shape, as does diethylamine, as basic compound peak shape often improves when using basic additives.

Sample Diluents In Convergence Chromatography

While a wide range sample diluents are compatible with CC, selecting the right diluent is sometimes necessary to achieve the best peak shape. Sample diluent strength can strongly affect peak shape and solubility in CC. As with other modes of chromatography, we recommend a weak sample diluent (as weak as possible), balancing analyte solubility and peak shape. With CC, that means that the sample should be dissolved in an organic solvent near the top of the eluotropic series (Table 1). Waters recommends heptane/2-propanol (90:10) as a good generic solvent, balancing solubility (isopropyl alcohol), and peak shape (heptane). Water content in the sample should be reduced, or eliminated if possible. Figure 30 shows seven overlaid chromatograms of peaks for the neutral compound butylparaben. As the injection volume increases, the effects of injection solvent strength on peak shape appear. In the strong co-solvent methanol, peak fronting occurs as injection volume increases. The weaker isopropyl alcohol shows less fronting and slightly taller peaks as compared to methanol. The recommended sample diluent isopropyl alcohol/hexane provides sharp, symmetrical peaks for all injection volumes.

Pressure, Temperature, and Density

The Automatic Back Pressure Regulator (ABPR) setting affects retention time by changing the density of compressed CO₂. As the ABPR pressure setting increases, the density increases while retention times decrease. Despite the fact that mobile phase composition has the greatest effect upon the separation, adjusting the pressure and tuning the density of the mobile phase can fine-tune a method. Figure 31 shows that increasing the ABPR setting (pressure) while leaving all other parameters unchanged results in shorter retention times.

 As in RPLC, column temperature affects both selectivity and retention in CC, and different analytes are affected to differing degrees. Increasing column temperature increases the energy of analyte molecules, which, like in LC or GC, should lead to decreased retention on stationary phases. But in CC, increasing temperature at constant pressure also decreases mobile phase density, which reduces its solvation power, and hence leads to increased retention. So in CC, varying temperature has a counter-acting effect. Mostly, as column temperature increases, mobile phase density decreases, and retention times increase (Figure 32). Also, note the presence of an additional (small) peak at 50 °C. This results from slight selectivity differences since temperature affects different analytes in different ways.

Peak shape, retention, and selectivity are manipulated by varying and understanding the roles of co-solvent, additive, sample diluent, pressure, temperature, and stationary phase. Combining everything shown in this chapter, Table 5 shows how CC separations are optimized using these tools. Note, however, that for a given separation, the importance of each parameter may vary. For example, sometimes the stationary phase plays a larger role than choice of co-solvent in manipulating selectivity. In other situations, switching from methanol to 50:50 methanol/acetonitrile plays a larger role than changing column chemistries.

Generic Method Development Protocol - Approach 1

Figure 33 describes a protocol for quickly determining whether a sample can be dissolved in an injection solvent compatible with CC. If the log P of the analyte is a value between -2 and 9, it can. If it is less than -2, the analyte is only soluble in an aqueous solvent and therefore may not be compatible with CC. If the log P value of the analyte is unknown, it must dissolve in a suitable organic solvent prior to injection. Remember that convergence chromatography behaves more like a normal-phase separation, where solvents such as methanol are a very strong solvent (the opposite of reversed-phase LC, where it is a fairly weak solvent).

Consequently, the sample needs to be dissolved (or diluted) in a weaker solvent such as heptane/isopropyl alcohol. A balance must be struck between sample solubility and peak shape; a discussion of the effect of sample diluent on peak shape in CC appears earlier in this chapter.

In addition to selecting a sample diluent, there are other considerations for developing a method. For example, what chromatographic conditions are appropriate for the analyte in question? Figure 34 shows a recommended generic set of starting conditions known to retain and separate a wide range of analytes. As with any mode of chromatography, these conditions are not appropriate in every case, and strategies for improving peak shape and changing selectivity and retention can be employed.

For such cases, Figures 35, 36, and 37 show the systematic approaches for improving peak shape, changing retention, and altering selectivity, and are not that different from those used in reversed-phase LC. Figure 35 provides a protocol for improving peak shape. The initial starting point is a recommended set of conditions based upon whether the analytes of interest are primarily basic or acidic in nature. Acidic compounds tend to have better peak shape under acidic conditions, and basic compounds tend to exhibit better peak shape under basic conditions. Strategies for improving peak shape include trying a different additive, changing the concentration of the additive, and changing the column chemistry.

Figure 36 gives a protocol for increasing retention. As with any form of liquid chromatography, the first option is to use a different (weaker) co-solvent. Methanol is the strongest co-solvent used in CC; using a weaker co-solvent, such as acetonitrile or another alcohol, increases retention time on the column. Flattening or reducing the slope of the gradient (lowering the final percentage of co-solvent or increasing the gradient duration) can be effective as well. Mixing co-solvents and reducing the concentration of methanol will weaken the overall co-solvent strength, thereby increasing retention. The ability to manipulate the density of mobile phase in the column is a property unique to SFC and CC. Doing so changes the overall retention, as lower density equals increased retention. This is done by decreasing the ABPR setting and/or increasing the temperature. Lastly, using a different column chemistry is another strategy for increasing retention.

Figure 37 illustrates a protocol for changing the selectivity (elution order, relative retention) of the separation. Trying different co-solvents, such as acetonitrile instead of methanol, is one approach. Changing from an alcohol-based, protic co-solvent to a non-alcohol, aprotic co-solvent such as acetonitrile has a much greater effect on selectivity than moving from methanol to another alcohol. Since a reduction in methanol concentration weakens overall co-solvent strength this can increase retention and change selectivity. Manipulating the density of the mobile phase can also change overall analyte retention, as these density effects can vary for specific analytes, which may be enough to optimize a given separation. Finally, if necessary, this protocol recommends a different column chemistry.

Column Family Method Development Strategy Protocol - Approach 2

The method development protocol described in the previous section focuses on generic starting conditions and the impact of varying mobile phase properties to fine-tune a separation. In this section two other alternatives - one for chiral separation and the other for achiral separation, are described. Both these alternatives call for combinations of a column and a method that statistically leads to the highest success rate for a large group of diverse test compounds. Generally speaking "brute-force" is applied during CC column selection – however the protocols suggested in this section propose a step-by-step strategy.

Method Development Strategy For Chiral Separations Using UPC² Trefoil Columns

The suggested screening protocol for chiral compounds is based on three Waters UPC² Trefoil chiral chemistries - AMY1, CEL1, and CEL2. The screening protocol demonstrated in Figure 38 recommends using this four-step, four-column optimal path screen and optimized co-solvent blends for the best chances of success.

As per the protocol, one should start screening with the AMY1, using a (1:1:1) blend of ethanol:isopropanol:acetonitrile with 20 mM ammonium acetate. If the method does not achieve the desired separation, the next method should employ CEL1 column with (1:1) blend of methanol and isopropanol with 0.2% TFA - and so on. Detailed method conditions are provided in the box of Figure 38.

Method Development Strategy For Achiral Separations Using UPC² Torus Columns

Screening protocol for an achiral separation follows a similar approach. It can be achieved in maximum three steps described below.

Step 1

Start with a rapid scouting step using a Torus 2-PIC (3.0 x 100 mm) column, and the following chromatographic conditions: 1.2 mL/min, 4–50% MeOH in 3 mins, 30 °C, 2,000 psi BPR

Based on the results obtained, note whether:

1. a) Selectivity and peak shape requirements are met, then further optimize method only if needed
2. b) Good selectivity was achieved, but peak shape needs improvement, then run a Torus 2-PIC column with an additive, either formic acid for acids or a basic additive for bases.
3. c) If the data shows that there is a need for different selectivity, then proceed to step 2.

Step 2

Based on the nature of the sample, choose the appropriate pathway outlined in the Defined Screening flow-path. Then continue down the flow-path using the suggested column and co-solvent combination until a suitable separation is achieved.

Step 3

To fine-tune the separation, one can optimize the separation by adjusting the method co-solvent composition, temperature, additive, and pressure. The figures in section 4.7 show you how to optimize a method in this way.