Ternary gradient

FIGURE 5.7 Effects of binary and ternary gradient elution with methanol and acetonitrile on separation selectivity in RP HPLC. Column LiChrosorb RP-C18, 5 pm, 300x4.0mm i.d. flow rate ImL/min UV detection, 254nm. Sample 4-cyanophenol (1), 2-methoxyphenol (2), 4-fluorophenol (3), 3-fluorophenol (4), 3-methylphenol (5), 4-chlorophenol (6), 4-iodophenol (7), 2-phenylphenol (8), and 3-ferf-butylphenol (9). 

Green tea Dilution Dionex AminoPac PAIO (250 x 2 mm) Ternary gradient eluent consisting of water/0.25M NaOH/l.OM NaOAc No ED (equipped with a thin layer type amperometric cell) [248]... 

Solvent systems encompass a dizzying array of permutations of organic solvents, buffers, and other mobile-phase additives. However, the most commonly employed solvent systems involve acetonitrile, methanol, and/or tetrahydrofuran. Buffers are typically acetate (pKa 4.8) or phosphate (pKa 1.3 and 6.7) at approximately 100 mM. For the analysis of a small number of free amino acids, isocratic elution is often possible. For the determination of an overall amino acid profile from a hydrolysate sample, complicated ternary gradients are often necessary. 

Moreau used a ternary gradient including 0.04% triethylamine in water 23 different classes of both nonpolar and polar lipids, including glycolipids and phospholipids, were resolved within 1 hour (57). Besides NL, FFA, DPG, PE, PG, PI, PS, and PC, three major hopanoid classes were... 

Two predominant phenolic compounds (neochlorogenic and chlorogenic acids) in prunes and prune juice can be analyzed by reversed-phase HPLC with diode array detection along with other phenolic compounds (65). Phenolic compounds were extracted from prunes with methanol and aqueous 80% methanol and analyzed by HPLC. Ternary-gradient elution (a) 50 mM NaH4H2P04, pH 2.6, (b) 80% acetonitrile/20% (a), and (c) 200 mM phosphoric acid, pH 1.5, was employed for an 80-min run time. Four wavelengths were monitored for quantitation 280 nm for catechins and benzoic acids, 316 nm for hydroxycinnamates, 365 nm for flavonols, and 520 nm for anthocyanins. Phenolic analysis of pitted prune extract is presented in an HPLC chromatogram in Fig. 9, which is based on work done by Donovan and Waterhouse (65). 

However, due to the artifacts resulting from oxidation, hydrolysis of esters or ethers, or isomerization of phenolics during pretreatment of wines, as well as due to the low recovery rates of some phenolics, analysis of wine phenolics via direct injection of the filtered wine into the chromatographic column is often selected (80,82-84). For the red wine and musts (80), which were injected directly into the HPLC without sample preparation, a ternary-gradient system was often employed for phenolic compounds. Twenty-two phenolic compounds, including 10 anthocyanins, were analyzed from red wine. The separation of cinnamic acid derivatives (313 nm),... 

For the red wines (82-84), which were injected directly into the HPLC without sample preparation, a ternary-gradient system using aqueous acetic acid (1% and 5% or 6%), and acidified acetonitrile (acetonitrile-acetic acid-water, 30 5 6) was used for cinnamic acid derivatives, catechins, flavonols, flavonol glycosides, and proanthocyanidins. Due to the large number of peaks, the gradient was extended to 150 min for the resolution of many peaks of important phenolics. This direct injection method was able to separate phenolic acids and esters, catechins, proanthocyanidins, flavonols, flavonol glycosides, and other compounds (such as tyrosol, and rrans-resveratrol) in wine in a single analysis. However, use of acetic acid did not permit the detector (PDA) to be used to record the UV spectra of phenolics below 240 nm (84). 

Ternary gradient elution, 0.01 M ammonium acetate (pH 5.0), acetonitrile and methanol... 

Another complication may arise if we choose to vary the selectivity of a gradient program in LC by varying more than one parameter at the same time. For example, the concentration of two organic modifiers may be varied independently in RPLC (so-called ternary gradients) or both the mobile phase composition and the temperature may be programmed. 

In figure 6.9 we take a closer look at some ternary gradients in which the composition ([Pg.264]

Figure 6.9 Examples of linear ternary gradients in which the concentration of two modifiers (B and C) is varied simultaneously. The concentration of the base solvent (A) is not indicated in the figure. 

In figure 6.9c a ternary gradient is shown in which a small concentration of the second modifier C is present throughout the elution. Figure 6.9d shows a gradient that runs from 100% A to 100% B and subsequently from 100% B to 100% C. This may be a sensible program if C is a considerably stronger solvent than B. 

Compositions of the solvents in pseudo-binary gradients (A — or A —1) which are equivalent to the ternary gradients of figure 6.7. 

From the above we may conclude that many of the ternary gradients which may be used in LC can be seen as special forms of binary gradients. Of course, this conclusion is no longer correct if we do not restrict the discussion to linear gradients and allow the shape of the gradient for one solvent to be different from that for another. However, it may be difficult to find applications for which such complicated ternary gradients can be proved to yield better results than the simpler (pseudo-) binary ones. 

The selectivity in programmed solvent LC may be varied by varying the solvents used or by the application of ternary or even more complicated gradients. However, most ternary gradients can in fact be reduced to binary ones using mixed (pseudo-) solvents. 

The most useful secondary parameter for the optimization of the selectivity in programmed solvent LC is the nature of the modifter(s) in the mobile phase. The selectivity can be varied by selecting various solvents (pure solvents for binary or ternary gradients mixed solvents for pseudo-binary gradients). Analogous to the situation in isocratic LC, it is possible to use different modifiers (and hence to obtain different selectivity) while optimum retention conditions are maintained for all solutes. This possibility to optimize the selectivity in programmed solvent LC will be discussed below. 

The application of the Simplex procedure for the optimization of the selectivity in programmed solvent LC (e.g. for the application of ternary gradients) has not yet been reported. However, there is no apparent obstacle to the applicability of the Simplex procedure for this purpose. 

Figure 6.16 Illustration of the predictive optimization method for ternary gradients in RPLC of Jandera et ai [628]. All figures were recorded with linear gradients from 100% solvent A to 100% solvent B in 60 min. Stationary phase Lichrosorb Cl8. Flow rate 1.0 ml/min.

The predictive method of Jandera et al. [628] requires knowledge of the isocratic retention data of all solute components in binary and (preferably) ternary mobile phase mixtures. Once these data are available, the method may be very helpful in obtaining an adequate (but not an optimum) separation with a ternary gradient. Unfortunately, the data required for the application of this predictive method are almost never available, and hence a large number of experiments need to be performed before any predictions can take place. When this is the case the method is of very little practical use. 

Interpretive methods will generally arrive at the global optimum after a limited number of experiments. However, (by definition) the recognition of the individual solutes is required in each experimental chromatogram. Also, the computational requirements are relatively high, especially if the simultaneous optimization of several parameters is considered. For example, (linear) ternary gradients (one parameter) will be much easier to optimize than quaternary gradients (two parameters). 

---