Critical Evaluation of HPLC Methods

The ability to properly develop, specify, and implement high performance liquid chromatography (HPLC) methods is critical for a successful separation, yet so many badly designed, badly specified, and poorly implemented methods exist. There are many parameters and variables that remain untouched between methods (the "lock and leave" syndrome) and are poorly understood in terms of their influence on a separation. Here, we describe important variables within a typical method to aid in understanding and highlight some of the commonly overlooked items when developing or implementing HPLC methods. For more information see the CHROMacademy webcast and Essential Guide Tutorial accompanying this article at www.chromacademy.com.

Acetonitrile is often chosen as the organic modifier for reversed-phase HPLC methods because of its low UV cut-off (190 nm) and relatively low viscosity (0.37 mPa·s). However, it is important to note that methanol, although not being so favorable in respect of UV cutoff (205 nm) or viscosity (0.6 mPa·s), has different physicochemical (solovophobic) properties that offer an alternative separation selectivity and can be particular useful when analyzing bases or less polar analytes. Isoeluograms can be used to obtain "equivalent" eutropic strengths when changing between modifiers, so that separations occur in a similar time frame but with an altered selectivity. This can be particularly useful when a suitable stationary phase for the separation cannot be easily identified; however, one should pay careful attention to the formation of azeotropic mixtures of methanol and water that lead to higher system back pressures, especially when using ultrahigh-pressure liquid chromatography (UHPLC) equipment. One should adjust the system pumping equipment to compensate for the viscosity differences of the various organic modifiers when performing on-line mixing.

When dealing with ionizable compounds, knowledge of the analyte pK_a values will allow the selectivity of separation to be "tuned," because retention in reversed-phase HPLC is related to analyte polarity and, therefore, the extent of ionization of functional groups. Acidic analytes have maximum retention in eluent systems at two pH units lower than the functional group pK_a and basic analytes in eluent systems at two pH units higher than the pK_a value. When dealing with analyte mixtures, the selectivity of the separation is altered by changing eluent pH until a suitable separation is obtained; however, one should note that separation robustness decreases with proximity to the analyte pK_a, where small changes in eluent pH give rise to larger changes in analyte retention. Recently, it has been fashionable to use eluent systems containing pH modifiers such as trifluoroacetic acid to adjust eluent pH well away from analyte pK_a values. For example, at pH 2.1, most acidic analytes will be fully neutral, most bases will be fully protonated (ionized), and modern stationary-phase chemistries can be chosen that will retain the more polar basic analytes. While this approach gives a robust method (eluent pH well away from the analyte pK_a values), we lose the ability to adjust the selectivity of the separation using eluent pH and rely solely on the nature of the organic modifier and the stationary-phase chemistry to optimize the band spacing.

One should note that trifluoroacetic acid is not a buffer and, as such, pH changes in the local environment as the sample is injected and enters the column may cause peak shape problems or retention time reproducibility issues. One should choose a true buffer (weak acid or base and its salt in cosolution such as trifluoroacetic acid and ammonium formate) whose pK_a is within one unit of the desired eluent pH value and is present at a high enough concentration to have the buffering capacity required (25 mM or less is typically sufficient with most common reversed-phase buffers).

Most modern HPLC stationary phases are made from spherical silica particles that are either fully porous or have a porous outer layer with a solid core. Smaller diameter particles are more efficient, but they will cause increased system back pressure in equivalent column dimensions. This increased efficiency can be used to effect high-resolution separations in a similar time frame to traditional particles or to obtain equivalent resolution in a reduced separation time frame. The larger the pore size within the stationary phase particle the lower the surface area, but the stronger the particle in terms of resistance to higher back pressures, and larger pore particles (300 Å) are typically used for biomolecule analysis to avoid exclusion of the large analyte from the pore system and hence poor retention. Some deactivation of the silica surface is usually necessary to obtain Gaussian peak shapes when analyzing polar or ionizable analytes to negate the interaction between these analytes and polar, lone (acidic) residual silanol groups on the silica surface. Analyses undertaken at low pH will help to reduce these interactions, which ultimately manifest themselves as tailing peaks, by operating at low eluent pH to help suppress the ionization of the acidic surface silanol species.

In general, short (circa 50 mm), narrow (2.1 or 3.0 mm i.d.) columns are used with smaller particles to generate high efficiencies in modern HPLC systems; however, these columns also highlight any problems with large extracolumn volumes in the HPLC system and all tubing and flow cell volumes need to be minimized to enjoy the benefits of the columns' increased efficiency.

When using UV detection, the data sampling rate and slit width must be optimized to realize the optimum sensitivity, acquisition wavelength, and bandwidth. Also, the reference wavelength and bandwidth need to be optimized to reduce baseline noise and drift in diode-array UV detection.