The limits of HPLC

Martin et al. published a paper on the theoretical limits of HPLC which is well worth reading.They used relatively simple mathematics to calculate pressure-optimized columns for which the length L, particle size and flow rate u of the mobile phase were selected such that a minimum pressure Ap is required to solve a separation problem. It has been shown that these optimized colunms are operated at their van Deemter curve minima. Some astonishing facts have emerged from the study, provided that the chromatography is performed on well packed columns (reduced plate height h = 2-3 see Section 8.5). 

The relationship is given by AT = where p is pressure and C the heat capacity of the liquid at 

A pressure-optimized column has L = 2.2 cm, dp = 6.9 )Xm and A/ = 2.3 bar Shorter columns are preferred for simple problems with a separation factor of ca. 1.2. Both time and solvent are saved by using columns 3-5 cm in length (also available commercially), but the injection and extra-column volumes and the detector time constant must be kept small in order not to deteriorate the separation performance. 

Strictly, the nomograms apply only for columns with a very specific van Deemter curve and samples with defined diffusion coefficients. However, they can always be used as a means of establishing whether or not a specific system is suitable for solving a specific separation problem. Also, they provide a view of the HPLC limits for those who understand how to interpret them. 

Calculate the number of theoretical plates that can be achieved with a pressure of 10 bar and 10 pin silica under optimum conditions (low viscosity system). 

A pressure-optimized column has L = 10 cm and dp = 6.3 rm. The recommended pressure is 11.8 bar (for a mobile phase viscosity of 0.4 mPa, which is a typical value in adsorption chromatography the viscosity may be up to 4.5 times greater in reversed-phase chromatography, increasing the pressure required to around 50 bar). This pressure is much lower than usual in most separations. 

An example of an eight-component separation on a 5 cm column is given in Fig. 2.28. 

---

One of the limitations of HPLC has been its restriction to the analytical mode of operation. However, increasing attention has been focused upon the development of preparative systems. For example, the use of HPLC in the isolation and identification of the components of complex mixtures of pyrethrum extracts, progesterone, and cholesteryl phenylace-tate was described [265]. Various large diameter columns for preparative work have also been developed such as that available from Du Pont and described by Wolf [266] and by De Stefano [267] in a variety of preparative applications. 

Retinoic acid, an endogenous retinoid, is a potent inducer of cellular differentiation. Because cancer is fundamentally a loss of cellular differentiation, circulating levels of retinoic acid could play an important role in chemoprevention. However, physiological concentrations are typically below the limits of HPLC detection. Sensitive techniques, such as negative chemical ionization (NCI) GC/MS have been employed for quantification, but cause isomerization and also fail to resolve the cis and trans isomers of retinoic acid. Normal phase HPLC can resolve the cis and trans isomers of retinoic acid without isomerization, and mobile phase volatility makes it readily compatible with the mass spectrometer. Based on these considerations, a method combining microbore normal phase HPLC separation with NCI-MS detection was developed to quantify endogenous 13-cis and all-trans retinoic acid in human plasma. The limit of detection was 0.5 ng/ml, injecting only 8 pg of retinoic acid onto the column. The concentration of 13-cis retinoic acid in normal, fasted, human plasma (n=13) was 1.6 +/- 0.40 ng/ml. 

A comprehensive presentation of the limits of HPLC was published by G. Guichon as The limits of the separation power of unidimensional column liquid chromatography . ... 

Swartz, M., Murphy, B., and Sievers, D. 2004. UPLC Expanding the limits of HPLC. 

The limits of lifetime detection and resolution in on-the-flight fluorescence lifetime detection in hplc were evaluated for simple, binary systems of polycycHc hydrocarbons (70). Peak homogeneity owing to coelution was clearly indicated for two compounds having fluorescence lifetime ratios as small as 1.2 and the individual peaks could be recovered using predeterrnined lifetimes of the compounds. Limits of lifetime detection were deterrnined to be 6 and 0.3 pmol for benzo[b]fluoranthene and benzo[k]fluoranthene, respectively. 

Chiral separations have become of significant importance because the optical isomer of an active component can be considered an impurity. Optical isomers can have potentially different therapeutic or toxicological activities. The pharmaceutical Hterature is trying to address the issues pertaining to these compounds (155). Frequendy separations can be accompHshed by glc, hplc, or ce. For example, separation of R(+) and 5 (—) pindolol was accompHshed on a reversed-phase ceUulose-based chiral column with duorescence emission (156). The limits of detection were 1.2 ng/mL of R(+) and 4.3 ng/mL of 3 (—) pindolol in semm, and 21 and 76 ng/mL in urine, respectively. 

There have been also found the quantitative characteristics of the methods. They are as follows for HPLC method the linearity is 0.1 ng to 2 ng the detecting limit is 0.1 ng the limit of the quantitative estimation makes up 0.0004 mg/kg a coefficient of variation is 2.74% for the chromatodensitometry method the linearity is 2 ng to 10 ng the detecting limit is 0.6 ng the coefficient of variation is 2.37%. The data obtained have been treated using a regressive analysis. 

As a matter of fact, the main advantage in comparison with HPLC is the reduction of solvent consumption, which is limited to the organic modifiers, and that will be nonexistent when no modifier is used. Usually, one of the drawbacks of HPLC applied at large scale is that the product must be recovered from dilute solution and the solvent recycled in order to make the process less expensive. In that sense, SFC can be advantageous because it requires fewer manipulations of the sample after the chromatographic process. This facilitates recovery of the products after the separation. Although SFC is usually superior to HPLC with respect to enantioselectivity, efficiency and time of analysis [136], its use is limited to compounds which are soluble in nonpolar solvents (carbon dioxide, CO,). This represents a major drawback, as many of the chemical and pharmaceutical products of interest are relatively polar. 

In the first chapter, I have discussed the limitations of high performance liquid chromatography (HPLC) and mass spectrometry when used in isolation and how the combination of the two allows these to be overcome. In this chapter, the effect of combining the two techniques with regard to the individual performance characteristics are explored. 

The main one is the incompatibility of HPLC, utilizing flow rates of ml min of a liquid, and the mass spectrometer, which operates under conditions of high vacuum. Even if this can be overcome, attention must then be focussed on the ionization of the analyte, bearing in mind the limitations of El and Cl discussed earlier in Chapter 3, and the generation of analytically useful mass spectra. 

Some Chemical Considerations Relevant to the Mouse Bioassay. Net toxicity, determined by mouse bioassay, has served as a traditional measure of toxin quantity and, despite the development of HPLC and other detection methods for the saxi-toxins, continues to be used. In this assay, as in most others, the molar specific potencies of the various saxitoxins differ, thus, net toxicity of a toxin sample with an undefined mixture of the saxitoxins can provide only a rough approximation of the net molar concentration. Still, to the extent that limits can be placed on variation in toxin composition, the mouse assay can in principle provide useful data on trends in net toxin concentration. However, the somewhat protean chemistry of the saxitoxins makes it difficult to define conditions under which the composition of a mixture of toxins will remain constant thus, attaining a reproducible level of mouse bioassay toxicity is difficult. It is therefore useful to review briefly some of the chemical factors that should be considered when employing the mouse bioassay for the saxitoxins or when interpreting results. Similar concepts will apply to other assays. 

The use of HPLC for quantification of phenols is often limited to a single class of phenolics and then often only to low-molecular weight compounds that are available as standards. It is, therefore, often necessary to use colorimetric assays such as the Folin-Ciocalteau assay which rely on the reducing ability of phenols to quantify the amount of total phenolics in a sample (Waterman and Mole, 1994 Singleton et al, 1999 Schofield et al, 2001). The degree of condensation of polyphenols can be quantified by colorimetric assays such as the acid-butanol assay and the vanillin assay (Waterman and Mole, 1994 Schofield et al, 2001). 

Once several target methods employing, e.g., LC/MS/MS techniques have been combined, a multi-residue method will evolve which includes the DEC S19 extraction procedures in combination with the generally applicable GPC cleanup and requires automatic multiple injections to circumvent the limitations of the limited HPLC peak capacity and the target-specific MS/MS methods. 

The limit of detection (LOD) is an important criterion of the efficiency of an analytical method. It is characterized by the smallest value of the concentration of a compound in the analytical sample. The detectable amount of anilide compounds is in the range 0.01-0.5 ng by GC and 0.1 ng by HPLC. The limit of quantitation (LOQ) ranges from 0.005 to 0.01 mg kg for vegetables, fruits and crops. The recoveries from untreated plant matrices with fortification levels between 10 and 50 times the LOD and the LOQ are 70-120%. The relative standard deviation (RSD) at 10-50 times the level of the LOD and LOQ are <10 % and <20%, respectively. 

Several determination methods such as GC, HPLC, gas chromatography/mass spectrometry (GC/MS) and liquid chromatography/mass spectrometry (LC/MS) are used for the analysis of neonicotinoid residues. The applications of GC/MS and LC/MS are of increasing importance. The application of HPLC to the determination of neonicotinoids residues is limited, especially when metabolites (such as acetamiprid and nitenpyram) can be easily determined by GC after derivatization. 

For the development of the LANA route, analytical techniques such as GC, TLC, FIPLC, NMR, and GC/MS were used. GC methods were developed to monitor formation of the Grignard reagent. Since all of the components of the LANA route are unstable to the elevated temperatures of GC, FIPLC and TLC techniques were chosen for qualitative and quantitative analysis of reaction samples, to monitor reaction progress, and to determine the purity of intermediates and final product. Because the process development time was limited and the LANA process was entirely dependent on HPLC analysis, we set criteria for the development of HPLC methods ... 

HPLC-PDA-MS) are already being used. Although HPLC-NMR-MS provides a very powerful approach for compositional and structural analysis, it by no means represents the limit of what is possible in terms of hyphenation. On-line extraction and the attachment of multiple detectors (e.g. IR, F) make the technique even more powerful. Other analytical laboratories such as TG-DTA-DSC-FTIR, TD-CT/Py/GC-MS/FTIR and HPLC-UV/NMR/IR/MS have been put to work, but do not represent practical solutions for routine polymer/additive analysis. 

Extracts of all matrices were analyzed by reversed-phase HPLC using ultraviolet detection at a wavelength of 266 nm (Soekhoe and Kerstens, 1995). The limit of detection (LOD) was 10 mg/L for all matrices. Recovery was > 90% and "between days" CV of the analytical chemical method was < 10%. 

Bergstrom et al. [63] used HPLC for determination of penicillamine in body fluids. Proteins were precipitated from plasma and hemolyzed blood with trichloroacetic acid and metaphosphoric acid, respectively, and, after centrifugation, the supernatant solution was injected into the HPLC system via a 20-pL loop valve. Urine samples were directly injected after dilution with 0.4 M citric acid. Two columns (5 cm x 0.41 cm and 30 cm x 0.41 cm) packed with Zipax SCX (30 pm) were used as the guard and analytical columns, respectively. The mobile phase (2.5 mL/min) was deoxygenated 0.03 M citric acid-0.01 M Na2HP04 buffer, and use was made of an electrochemical detector equipped with a three-electrode thin-layer cell. The method was selective and sensitive for mercapto-compounds. Recoveries of penicillamine averaged 101% from plasma and 107% from urine, with coefficients of variation equal to 3.68 and 4.25%, respectively. The limits of detection for penicillamine were 0.5 pm and 3 pm in plasma and in urine, respectively. This method is selective and sensitive for sulfhydryl compounds. 

---