The ability to separate and analyse complex samples is integral to the biological and medical sciences. Classic column chromatography has evolved over the years, with chromatographic innovations introduced at roughly decade intervals. These techniques offered major improvements in speed, resolving power, detection, quantification, convenience and applicability to new sample types. The most notable of these modifications was high performance liquid chromatography (HPLC). Modern HPLC techniques became available in 1969; however, they were not widely accepted in the pharmaceutical industry until several years later. Once HPLC systems capable of quantitative analysis became commercially available, their usefulness in pharmaceutical analysis was fully appreciated.
Sensitivity in chromatographic analysis is a measure of the smallest detectable level of a component in a chromatographic separation and is dependent on the signal–to–noise ratio in a given detector. Sensitivity can be increased by derivatisation of the compound of interest, optimisation of chromatographic system or miniaturisation of the system. The limit of detection is normally taken as three times the signal–to–noise ratio and the limit of quantification as ten times this ratio.
If we approximate peaks by symmetric triangles, then if R is equal to or more than 1, the components are completely separated. If R is less than 1, the components overlap.
The systems used in chromatography are often described as belonging to one of four mechanistic types: adsorption, partition, ion exchange and size exclusion. Adsorption chromatography arises from interactions between solutes and the surface of the solid stationary phase. Generally, the eluents used for adsorption chromatography are less polar than the stationary phases and such systems are described as ‘normal phase’. Partition chromatography involves a liquid stationary phase that is immiscible with the eluent and coated on an inert support. Partition systems can be normal phase (stationary phase more polar than eluent) or reversed–phase chromatography, referred to as RPC (stationary phase less polar than eluent).Ion–exchange chromatography involves a solid stationary phase with anionic or cationic groups on the surface to which solute molecules of opposite charge are attracted. Size–exclusion chromatography involves a solid stationary phase with controlled pore size. Solutes are separated according to their molecular size, with the large molecules unable to enter the pores elute first. However, this concept of four separation modes is an over–simplification. In reality, there are no distinct boundaries and several different mechanisms often operate simultaneously.
Other types of chromatographic separation have been described. Ion–pair chromatography is an alternative to ion–exchange chromatography. It involves the addition of an organic ionic substance to the mobile phase, which forms an ion pair with the sample component of opposite charge. This allows a reversed–phase system to be used to separate ionic compounds. Chiral chromatography is a method used to separate enantiomers, which can be achieved by various means. In one case, the mobile phase is chiral and the stationary phase is non–chiral. In another, the liquid stationary phase is chiral with the mobile phase non–chiral or, finally, the solid stationary phase may be chiral with a non–chiral mobile phase.
HPLC instrumentation includes a pump, injector, column, detector and recorder or data system (Fig. 1). The heart of the system is the column in which separation occurs. Since the stationary phase is composed of micrometer–size porous particles, a high–pressure pump is required to move the mobile phase through the column. The chromatographic process begins by injecting the solute onto the top of the column. Separation of components occurs as the analytes and mobile phase are pumped through the column. Eventually, each component elutes from the column and is registered as a peak on the recorder. Detection of the eluting components is important; this can be either selective or universal, depending upon the detector used. The response of the detector to each component is displayed on a chart recorder or computer screen and is known as a chromatogram. To collect, store and analyse the chromatographic data, computers, integrators and other data–processing equipment are used frequently.
Figure.1. Typical HPLC system
Mobile phase reservoir
The most common type of solvent reservoir is a glass bottle. Most of the manufacturers supply these bottles with special caps, Teflon tubing and filters to connect to the pump inlet and to the sparge gas (helium) used to remove dissolved air. When the mobile phase contains excessive gas that remains dissolved at the pressure produced by the column, the gas may come out of the solution at the column exit or in the detector, which results in sharp spikes. Spikes are created by microscopic bubbles that change the nature of the flowing stream to make it heterogeneous, while drift may occur as these microscopic bubbles gradually collect and combine in the detector cell. The main culprit is oxygen (from the air) that dissolves in polar solvents, particularly water. Degassing may be accomplished by one or a combination of the following methods: apply a vacuum to the liquid, boil the liquid, place the liquid in an ultrasonic bath, bubble a fine stream of helium through the liquid (sparging) or by commercial on–line degassing units.
High–pressure pumps are needed to force solvents through packed stationary phase beds. Smaller bed particles (e.g. 3 μm) require higher pressures. There are many advantages to using smaller particles, but they may not be essential for all separations. The most important advantages are higher resolution, faster analyses and increased sample load capacity. However, only the most demanding separations require these advances in significant amounts. Many separation problems can be resolved with larger particle packings (e.g. 5 μm) that require less pressure.
Flow–rate stability is another important pump feature that distinguishes pumps. Constant–flow systems are generally of two basic types: reciprocating piston and positive displacement (syringe) pumps. The basic advantage of both systems are their ability to repeat elution volume and peak area, regardless of viscosity changes or column blockage, up to the pressure limit of the pump. Although syringe–type pumps have a pressure capability of up to 540 000 kPa (78 000 psi), they have a limited ability to form gradients. Reciprocating piston pumps can maintain a liquid flow for an indefinite length of time, while a syringe pump needs to be refilled after the syringe volume has been displaced. Dual–headed reciprocating piston pumps provide more reproducible and pulse–free delivery of solvent, which reduces detector noise and enables more reliable integration of peak area. Reciprocating pumps now dominate the HPLC market and are even useful for micro-HPLC applications, as they can maintain a constant flow at flow rates in μL/min ranges.
An additional pump feature found on the more elaborate pumps is external electronic control. Although it adds to the expense of the pump, external electronic control is a very desirable feature when automation or electronically controlled gradients are to be run. Alternatively, this becomes unnecessary when using isocratic methods. The degree of flow control also varies with pump expense. More expensive pumps include such state–of–the–art technology as electronic feedback and multiheaded configurations.
Modern pumps have the following parameters:
- Flow–rate range, 0.01 to 10 mL/min.
- Flow–rate stability, not more than 1% (short term).
- For size exclusion chromatography (SEC), flow–rate stability should be <0.2%.
- Maximum pressure, up to 34 500 kPa (5000 psi).
An injector for an HPLC system should provide injection of the liquid sample within the range of 0.1 to 100 mL of volume with high reproducibility and under high pressure (up to 27 600 kPa). The injector should also minimise disturbances to the flow of the mobile phase and produce minimum band broadening. Sample introduction can be accomplished in various ways. The injection valve has, in most cases, replaced syringe injection. Valve injection offers rapid, reproducible and essentially operator–independent delivery of a wide range of sample volumes. The most common valve is a six–port Rheodyne valve in which the sample fills an external stainless steel loop. A clockwise turn of the valve rotor places the sample–filled loop into the mobile–phase stream, which deposits the sample onto the top of the column. These valves can be operated manually or actuated via computer–automated systems. One minor disadvantage of valve injection is that the sample loop must be changed to obtain various sample volumes. However, this is a simple procedure that requires a few minutes only. In more sophisticated HPLC systems, automatic sampling devices are incorporated. These autosamplers have a piston–metering syringe–type pump to suck the preset sample volume into a line and transfer it to a sample loop of adequate size in a standard six–port valve. Most autosamplers are computer controlled and can serve as the master controller for the whole system.
In HPLC, liquid samples may be injected directly and solid samples need only be dissolved in an appropriate solvent. The solvent need not be the mobile phase, but frequently it is wise to choose the mobile phase to avoid detector interference, column–component interference, loss in efficiency or all of these. It is always best to remove particles from the sample by filtration or centrifugation, since continuous injections of particulate material eventually cause blockage of injection devices or columns.
Sample sizes may vary widely. The availability of highly sensitive detectors frequently allows the use of small samples that yield the highest column performance.
It often is advantageous to run ion exchange, size–exclusion and reversed–phase columns above room temperature and to control precisely the temperature of liquid–liquid columns. Therefore, column thermostats are a desirable feature in modern HPLC instruments. Temperature variation within the HPLC column should generally be held within ±0.2°. To maintain a constant temperature is especially important in quantitative analysis, since changes in temperature can seriously affect peak–size measurement. It is often important to be able to work at higher temperatures for size–exclusion chromatography of some synthetic polymers because of solubility problems. High–velocity circulating air baths, which usually consist of high–velocity air blowers plus electronically controlled thermostats, are the most convenient for HPLC. Alternatively, HPLC columns can be jacketted and the temperature controlled by contact heaters or by circulating fluid from a constant–temperature bath. This latter approach is practical for routine analyses, but is less convenient when columns must be changed frequently.
These valve devices are used to divert the flow from one column to another within a single HPLC system. Column–switching techniques can be used during method development when several columns are to be evaluated for their efficiency, retention, etc. More recently, the use of column switching has been employed in the on–line analysis of biological matrices. Raw plasma or other sample matrix is injected directly onto the first column. Chromatographic conditions are optimised such that interfering substances are eluted from the column while the analytes of interest are retained. The column switch then diverts the eluent that contains the analytes of interest from the ‘clean–up column’ onto the analytical column, which then separates the analytes of interest for quantification or characterisation. Another use of column switches is in gradient chromatography for which high throughput is essential. The first column is switched off–line to re–equilibrate to initial conditions, while the second column is brought on–line for the next injection. This conserves valuable analysis time that would otherwise be wasted waiting for the column to re–equilibrate. The most up–to–date information on the use of column switching can be found by searching the current literature.
Today, optical detectors are used most frequently in HPLC systems. These detectors pass a beam of light through the flowing column effluent as it passes through a flow–cell. Flow–cells are available in preparative, analytical and micro–analytical sizes The variations in light intensity, caused by ultraviolet (UV) absorption, fluorescence emission or change in refractive index (depending on the type of detector used) from the sample components that pass through the cell, are monitored as changes in the output voltage. These voltage changes are recorded on a strip–chart recorder and frequently are fed into an integrator or computer to provide retention time and peak–area data.
Most applications in drug analysis use detectors that respond to the absorption of UV radiation (or visible light) by the solute as it passes through the flow–cell. Absorption changes are proportional to concentration, following the Beer–Lambert Law. Flow–cells generally have path–lengths of 5 to 10 mm with volumes between 5 and 10 μL. These detectors give good sensitivities with many compounds, are not affected by slight fluctuations in flow rate and temperature, and are non–destructive, which allows solutes to be collected and further analysed if desired.
The simplest detectors are of the fixed–wavelength type and usually contain low–pressure mercury lamps that have an intense emission line at 254 nm. Some instruments offer conversion kits that allow the energy at 254 nm to excite a suitable phosphor to give a new detection wavelength (e.g. 280 nm). Variable–wavelength detectors have a deuterium lamp with a continuous emission from 180 to 400 nm and use a manually operated diffraction grating to select the required wavelength. Tungsten lamps (400 to 700 nm) are used for the visible region.
Many organic compounds absorb at 254 nm and hence a fixed–wavelength detector has many uses. However, a variable–wavelength detector can be invaluable to increase the sensitivity of detection by using the wavelength of maximum absorption. This is particularly useful when analysing proteins that absorb at 280 nm, or peptides that are detected commonly at 215 nm. Using a variable–wavelength detector can also increase the selectivity of detection by enhancing the peak of interest relative to interfering peaks.
Eluents must have sufficient transparency at the selected detection wavelength. Buffer salts can also limit transparency. The spectra of some drugs change with pH and the sensitivity and selectivity of an assay can sometimes be controlled by changing the eluent pH. The influence of such changes on the chromatography must also be considered.
Other detectors commonly used include diode array, refractive index (RI), fluorescence (FL), electrochemical (EC) and mass spectrometry (MS). Infra–red (IR) and nuclear magnetic resonance (NMR) spectrometers may also be used as detectors.
Photodiode array detectors
The photodiode array detector (DAD) is an advanced type of UV detector. Depending on the wavelength, a tungsten lamp and a deuterium lamp are used as light sources. The polychromatic light beam is focused on a flow–cell (volume 8 to 13 μL) and subsequently dispersed by a holographic grating or quartz prism. The spectral light then reaches a chip that contains 100 to 1000 light–sensitive diodes arranged side by side. Each diode only registers a well–defined fraction of the information and in this way all wavelengths are measured at the same time. Note that although having more diodes in an array increases the resolution of UV spectra, it lowers the absolute sensitivity since less radiation is absorbed by each individual diode. The wavelength resolution of up–to–date detectors is of the order of 1 nm per diode, with a wavelength accuracy of better than ±1 nm and a sensitivity below 10−4 absorbency units. All operations of the detector are controlled by a computer: correction of fluctuations of the lamp energy, collection of signals (Iλ) from all the diodes, storage of the data of the mobile phase (I0λ, measured at the start of the chromatogram) and calculation of the absorbance according to the Beer–Lambert Law from Iλ to I0λ. The number of spectra recorded per second can be chosen from between 0.1 and 10; usually one spectrum/sec is optimum with respect to chromatographic resolution and noise. At the end of the run, a three–dimensional spectrochromatogram (absorbance as a function of wavelength and time) is stored on the computer and can be evaluated qualitatively and quantitatively. A detailed description of the DAD operation is given in Huber and George (1993).
Diode array detection offers several advantages. Knowledge of the spectra of compounds of interest enables interfering peaks to be eliminated such that an accurate quantification of peaks of interest can be achieved despite less than optimal resolution. Simultaneous detection at two wavelengths allows calculation of an absorbance ratio. If this ratio is not constant across a peak, the peak is not pure, regardless of its appearance. An additional advantage of diode array detection is the subtraction of a reference wavelength. This reduces baseline drift during gradient elution. HPLC–DAD systems linked to libraries of UV spectra are particularly useful in clinical and forensic toxicology in screening for drugs in biological samples and its use in this context is described in detail in later (Pragst and Herzler, personal communication).
Refractive index detector
The RI detector is a universal detector, in that changes in RI (either positive or negative) that arise from the presence of a compound in the eluent are recorded. However, it is also the least–sensitive detector (as much as 100 times less sensitive than UV detection). RI detectors may be used for excipients such as sugars in pharmaceuticals. Many factors influence RI and must be controlled during separation, such as temperature, eluent composition and pressure. The chromatography is best facilitated using a thermostatically controlled cabinet and high–quality pump to minimise pressure fluctuations.
In FL detectors, the solute is excited with UV radiation and emits radiation at a longer wavelength. Most detectors allow the selection of both excitation and emission wavelengths. There are only a few drugs and natural compounds that have strong natural fluorescence (e.g. ergot alkaloids), however, many drug derivatives are fluorescent compounds. FL detection can offer great selectivity, since excitation and emission wavelengths as well as retention time can be used to identify drugs. It is necessary to choose eluents carefully when using FL detection. The eluent must neither fluoresce nor absorb at the chosen wavelengths. It is also necessary to consider the pH of the system, in that some drugs only show fluorescence in certain ionic forms.
EC detectors measure the current that results from the electrolytic oxidation or reduction of analytes at the surface of an electrode. These detectors are quite sensitive (down to 10–15 mole) and also quite selective. Two types of detector are available. The coulometric detector has a large electrode surface at which the electrochemical reaction is taken to completion. The amperometric detector has a small electrode with a low degree of conversion. Despite the difference in conversion rate, in practice these two types have approximately the same sensitivity. Eluents for EC detection must be electrically conductive. This is accomplished by the addition of inert electrolytes. EC detection is most easily used in the oxidative mode, as use in the reductive mode requires the removal of dissolved oxygen from the eluent.
The recent development of the so–called hyphenated techniques has improved the ability to separate and identify multiple entities within a mixture. These techniques include HPLC–MS, HPLC–MS–MS, HPLC–IR and HPLC–NMR. These techniques usually involve chromatographic separation followed by peak identification with a traditional detector such as UV, combined with further identification of the compound with the MS, IR or NMR spectrometer.
MS as a detector for an HPLC system has gained wide popularity over the past several years. Advances in data systems and the simplification of the user interface have facilitated the ease of use of a mass spectrometer as an HPLC detector. The most common types of mass spectrometers used in HPLC are quadrupoles and ion traps. Tandem mass spectrometers (also called triple quadrupoles) are also commonly available and are widely used in the pharmaceutical industry for the quantitative analysis of trace concentrations of drug molecules.
The process of mass analysis is essentially the same as in any other mass spectrometric analyses that utilise quadrupole or ion–trap technology. The unique challenge to interfacing an HPLC to a mass spectrometer is the need to convert a liquid–phase eluent into a gas phase suitable for mass spectral analysis. Modern mass spectrometers commonly utilise a technique known as atmospheric pressure ionisation (API) to accomplish this. API can be subdivided into electrospray (ionspray) ionisation (ESI) and atmospheric pressure chemical ionisation (APCI). Each technique has its own advantages. ESI is particularly useful for the analysis of a wide variety of compounds, especially proteins and peptides. APCI is also very well suited for the analysis of a large variety of compounds, particularly the less polar organic molecules. Both techniques are very rugged and well suited to pharmaceutical analysis.
An important consideration when using API is the need for volatile mobile–phase modifiers in the chromatographic separation. Acetic acid, formic acid, etc., are commonly used as acidic modifiers. Ammonium formate and ammonium acetate salts can also be used when more pH control is required for the separation. Organic modifiers are most often methanol or acetonitrile. One very important issue that must be considered when developing a method using API (electrospray, in particular) is the phenomenon of ion suppression. Co–eluting contaminants compete with the analyte of interest for ionisation, which results in a loss of signal for the analyte of interest. This can be very problematic if extremely small quantities of analyte are to be measured (as is often the case when MS is being used). Additional sample cleanup or adjustment of the chromatography to prevent coelution of the contaminant is often necessary to correct this problem.
HPLC–MS–MS is commonly used in the pharmaceutical industry and in forensic science to analyse trace concentrations of drug and/or metabolite. MS–MS offers the advantage of increased signal–to–noise ratio, which in turn lowers the limits of detection and quantification easily into the sub ng/mL range. MS–MS is also a very useful technique in the qualitative identification of previously unidentified metabolites of drugs, which thus makes MS–MS a very powerful technique in research laboratories. Several recently published studies have utilised MS–MS as a high–throughput analytical technique in the pharmaceutical industry.
HPLC–IR has proved to be an effective method to detect degradation products in pharmaceuticals. IR provides spectral information that can be used for compound identification or structural analysis. The IR spectra obtained after HPLC separation and IR analysis can be compared to the thousands of spectra available in spectral libraries to identify compounds, metabolites and degradation products. An advantage of IR spectroscopy is its ability to identify different isomeric forms of a compound based on the different spectra that result from alternative locations of a functional group on the compound. Unlike MS, IR is a non–destructive technique in which the original compound is deposited on a plate as pure, dry crystals and can be collected afterwards if desired.
HPLC–NMR is also growing in popularity for the identification of various components in natural products and other disciplines. Although a relatively new hyphenated system, HPLC–NMR has several applications on the horizon. The miniaturisation of the system and the possibility of measuring picomole amounts of material are both areas currently attracting a large amount of attention. Also, in the future HPLC–NMR systems will be interfaced with other detectors, such as Fourier transform IR and mass spectrometers. This will provide a wide range of possibilities for further applications, which could include the analysis of mixtures of polymer additives and the ability to identify unknowns without first having to isolate them in a pure form.
Since the detector signal is electronic, use of modern data–acquisition techniques can aid in the signal analysis. In addition, some systems can store data in a retrievable form for highly sophisticated computer analysis at a later time.
The main goal in using electronic data systems is to increase analysis accuracy and precision, while reducing operator attention. There are several types of data systems, each of which differ in terms of available features. In routine analysis, where no automation (in terms of data management or process control) is needed, a pre–programmed computing integrator may be sufficient. If higher control levels are desired, a more intelligent device is necessary, such as a data station or minicomputer. The advantages of intelligent processors in chromatographs are found in several areas. Firstly, additional automation options become easier to implement. Secondly, complex data analysis becomes more feasible. These analysis options include such features as run–parameter optimisation and deconvolution (i.e. resolution) of overlapping peaks. Finally, software safeguards can be designed to reduce accidental misuse of the system. For example, the controller can be set to limit the rate of solvent switching. This acts to extend column life by reducing thermal and chemical shocks. In general, these stand–alone, user–programmable systems are becoming less expensive and increasingly practical.
Other more advanced features can also be applied to a chromatographic system. These include computer–controlled automatic injectors, multi–pump gradient controllers and sample fraction collectors. These added features are not found on many systems, but they do exist, and can save much time and effort for the chromatographer.
Typical HPLC columns are 10, 15 and 25 cm in length and are fitted with extremely small diameter (3, 5 or 10 μm) particles. The columns may be made of stainless steel, glass–lined stainless steel or polyetheretherketone (PEEK). The internal diameter of the columns is usually 4.0 or 4.6 mm for traditional detection systems (UV, FL, etc.); this is considered the best compromise between sample capacity, mobile phase consumption, speed and resolution. However, if pure substances are to be collected (preparative scale), larger diameter columns may be needed. Smaller diameter columns (2.1 mm or less) are often used when HPLC is coupled with MS. The smaller diameter columns also have the advantage of consuming less solvent because of their lower optimal flow rates. HPLC systems sold today can often be plumbed with narrower tubing diameters to take advantage of the benefits of these smaller column diameters.
Packed capillary microcolumns are also gaining wider use when interfacing the HPLC to a mass spectrometer and extremely low flow rates (nL/min) are needed to maximise sensitivity for the analysis of proteins and peptides.
Packing of the column tubing with small diameter particles requires high skill and specialised equipment. For this reason, it is generally recommended that all but the most experienced chromatographers purchase pre–packed columns, since it is difficult to match the high performance of professionally packed HPLC columns without a large investment in time and equipment.
In general, HPLC columns are fairly durable and one can expect a long service life unless they are used in some manner that is intrinsically destructive, such as with highly acidic or basic eluents, or with continual injections of ‘dirty’ biological or crude samples. It is wise to inject some test mixture (under fixed conditions) into a column when new and to retain the chromatogram. If questionable results are obtained later the test mixture can be injected again under specified conditions. The two chromatograms are compared to establish whether or not the column is still useful.
The description of column dimensions and assignment of a category to that size varies greatly depending on the reference cited. The following categories were suggested by Rozing et al. (2001), and may be more stratified than other categories.
Preparative columns generally are larger bore than analytical columns. Some have inner diameters as large as 100 mm and may have lengths up to 600 mm. These columns are usually packed with packing materials of larger particle size that may range from 10 to 50 μm particle size. The flow rate used with these columns normally exceeds 5 mL/min.
The normal bore for an analytical column can range from 3.9 mm to 5.0 mm inner diameter, but the most common is 4.6 mm. This diameter is the best compromise between sample capacity, mobile phase consumption, speed and resolution. The normal flow rate for this type of column is 1.5 to 5 mL/min.
A mini or narrow bore column has an inner diameter of 2.1 mm to 3.9 mm. The flow rate for this column size ranges from 500 to 1500 μL/min.
Microbore columns have a 1.0 mm to 2.1 mm inner diameter and have flow rates of 100 to 500 μL/min. These small columns save solvent, are popular when HPLC is interfaced with MS and provide increased sensitivity in situations of limited sample mass.
Capillary columns have inner diameters of 50 μm to 1.0 mm and have a typical flow rate of 0.2 to 100 μL/min. So–called ‘nanobore’ columns usually fall into the lower end of this size range. The inner surface of these very narrow columns must be extremely smooth. Since this is difficult to obtain with stainless steel columns, many of these columns are glass–lined stainless steel. Fused silica columns also fall into this category.
Silica–based packing materials
Silica (SiO2,xH2O) is the most widely used substance for the manufacture of packing materials. It consists of a network of siloxane linkages (Si–O–Si) in a rigid three–dimensional structure that contains interconnecting pores. The size of the pores and the concentration of silanol groups (Si–OH), which line the pores, can be controlled in the manufacturing process. Thus, a wide range of commercial products is available with surface areas that range from 100 to 800 m2/g and average pore sizes from 4 to 33 nm.
Spherical packing materials are now the only types being introduced for analytical HPLC. Irregular shaped materials are still being used to pack preparative columns. The silanol groups on the surface of silica give it a polar character, which is exploited in adsorption chromatography using organic eluents. Silanol groups are also slightly acidic and hence basic compounds are adsorbed particularly strongly. Unmodified silicas can thus be used with aqueous eluents for the chromatography of basic drugs.
Silica can be altered drastically by reaction with organochlorosilanes or organoalkoxysilanes to give Si–O–Si–R linkages with the surface. The attachment of hydrocarbon chains to silica produces a non–polar surface suitable for RPC in which mixtures of water and organic solvents are used as eluents. The most popular material is octadecylsilica (ODS), which contains C18 chains, but materials with C1, C2, C4, C6, C8 and C22 chains are also available. The latest silica–based bonded phase to be introduced is a long C30 phase, which has 24% carbon coverage to make it one of the most retentive phases available.
During manufacture, such materials may be reacted with a small monofunctional silane (e.g. trimethylchlorosilane) to reduce further the number of silanol groups that remain on the surface (endcapping). Recent advances in column technology include multiple reactant endcapping, use of Type B (high purity, low trace metal, low acidity) silica and encapsulating the surface with a polymeric phase. These silicas are often referred to as ‘base–deactivated’ and are especially useful in RPC in the pH range of 4 to 8 when many basic compounds are partially ionised. Variations in elution order on different commercial packing materials of the same type (e.g. ODS) are often attributed to differences in surface coverage and the presence of residual silanol groups. For this reason it must not be assumed that a method developed with one manufacturer’s ODS column can be transferred easily to another manufacturer’s ODS column.
A vast range of materials have intermediate surface polarities that arise from the bonding to silica of organic compounds that contain groups such as phenyl, cyano, nitro, amino, fluoro, sulfono and diols. There are also miscellaneous chemical moieties bound to silica, as well as polymeric packings, designed to purify specific compounds.
Propylphenylsilane ligands attached to the silica gel show weak dipole–induced dipole interactions with polar analytes. Usually this type of bonded phase is used for group separations of complex mixtures. Newer phases have phenyl backbones that allow π–π (stacking) interactions. These are recommended for peptide mapping applications. Amino–compounds show some specific interactions with phenyl–modified adsorbents.
A cyano–modified surface is very slightly polar. Columns with this phase are useful for fast separations of mixtures that consist of very different components. These mixtures may show a very broad range of retention times on the usual columns.
Cyano–columns can be used on both normal- and reversed–phase modes of HPLC.
Amino–phases are weak anion–exchangers. This type of column is mainly used in normal–phase mode, especially for protein separation and also the selective retention of aromatic compounds.
A newer type of silica packing has fluorinated surfaces. This phase is generally more hydrophilic than phases with hydrocarbons of similar chain length. It has increased retention and unique selectivity for halogenated organic compounds and lipophilic compounds.
Sulfonic functional groups separate compounds on the basis of hydrophobic interactions. These packing materials allow the isocratic separation of mixtures that normally require gradient elution.
Diols are slightly polar adsorbents for normal–phase separations. These are useful to separate complex mixtures of compounds with different polarities that usually have a strong retention on unmodified silica.
Cyclodextrins, amylose, avidin, ristocetin, nitrophenylethyl, carbamate, ester, diphenylethyldiamine and Pirkle–type functional groups are all bound to silica packing material to enable enantiomeric separations. These columns are often referred to as chiral columns. Strong ion–exchangers are also available, in which sulfonic acid groups or quaternary ammonium groups are bonded to silica. These packing materials are useful to separate proteins. There are also proprietary functional groups added to silica packing materials for a variety of uses. These include petrochemical analysis, environmental analysis, detection of deoxyribose nucleic acid (DNA) adducts, purification of double stranded DNA, separation of cationic polymers and separation of nitro–aromatic explosives.
For size–exclusion chromatography, a special type of silica is available that has a narrow range of pore diameters. Size–exclusion chromatography can be complicated by adsorption, but this can be reduced by treating the surface with trimethylchlorosilane.
The useful pH range for silica-based columns is 2 to 8, since siloxane linkages are cleaved below pH 2 while at pH values above 8 silica may dissolve. However, the pH range may be extended above 8 if a precolumn packed with microparticulate silica is included between the pump and injector to saturate the eluent before it enters the analytical column.
Zirconia packing materials
Zirconia is a metal oxide that is more chemically and thermally stable than silica. It can be used for separations conducted at temperatures as high as 200° and is unaffected by changes in ionic strength or organic content of the mobile phase. Zirconia packings have a wider pH range and are especially useful for basic separations at pH 10 or higher, where silica gel starts to dissolve. Zirconia can be used for RPC and is extremely stable and efficient through surface modification with polymer or carbon coatings. Other chemical modifications of zirconia produce packing materials suitable for normal–phase or ion–exchange chromatography.
Polymer–based packing materials
Several packing materials based on organic polymers are available. For example, unmodified styrene–divinylbenzene co–polymers have a hydrophobic character and can be used for RPC. Although they traditionally give lower column efficiencies than ODS-silica, this has improved greatly in the past few years. Polymeric materials are best when separation conditions require a mobile phase that can go beyond the upper pH limits of silica gel (usually pH 6.5 to 7), as they have the advantage of being stable over a wide pH range. Polymeric materials also provide different selectivity and retention characteristics to silica–based reversed phase packings. They also avoid problems associated with residual silanol groups (e.g. peak tailing). Ion–exchange materials of the styrene–divinylbenzene type are also available in which sulfonic acids, carboxylic acids or quaternary ammonium groups are incorporated in the polymeric matrix.
Monoliths are chromatographic columns that are cast as continuous homogenous phases rather than packed as individual particles, creating porous rods of polymerised silica that are mechanically stable. Monolithic phases have flow–through pores with macroporosity (approx. 2 μm) and mesopores, which are diffusive pores with an average pore diameter that can be controlled. To create the column, a silica gel polymer is formed, which, after ageing, is dried into the form of a straight rod of highly porous silica with the bimodal pore structure. The rod is then encased (or clad) in a PEEK cover, ensuring that there is absolutely no void space between the silica and PEEK material. The pore structure yields a very large internal surface area and ensures high–quality separations. In addition, the high porosity of the column means very high flow rates can be used with lower pressures. This enables separations in a fraction of the time needed when using a column with conventional packing materials.
Recently, a polymeric monolithic column was introduced. It contains a poly(glycidylmethacrylate–ethyleneglycol-dimethacrylate) co–polymer that has functional groups added to make various types of stationary phases.
An effective maintenance programme is essential to keep an HPLC system in proper working order. The maintenance programme should include preventative, periodical and necessary repairs of the HPLC system. This programme is essential to ensure that all of the components of the system are in proper working condition. In this section, the general maintenance of columns, pumps, injection valves and detectors is discussed. For information on the functions and uses of these components, refer to the earlier sections of this chapter.
It is always recommended that the maintenance guidelines provided with the system should be consulted to ensure compliance with the manufacturers’ suggestions. This guide should be utilised whenever maintenance is required.
The quality of solvents and inorganic salts is an important consideration. Soluble impurities can give noisy baselines and spurious peaks or can build up on the surface of the packing material, eventually changing chromatographic retention. Furthermore, the eluate may need to be collected for further experimentation and all contamination must be avoided. In addition, particulate matter should be removed, otherwise pump filters, frits and tubing can become blocked.
Now commercially available is a wide range of HPLC-grade solvents that are free from particulate matter, have low residues on evaporation and have guaranteed upper limits of UV-absorbing and fluorescent impurities. However, if a detector is not to be operated at its maximum sensitivity, analytical grade solvents may be used. A general rule of thumb is to use the highest purity of solvent that is available and practical depending on the particular application.
Air dissolved in the mobile phase can lead to problems. The formation of a bubble in a pump head usually reduces or stops eluent flow, while bubbles formed in the detector can give spurious peaks. One commonly used remedy is to degas the eluent using an in–line vacuum chamber. HPLC solvents are pumped from the reservoirs into a vacuum chamber in–line with the HPLC eluent flow. This method ensures continuous and efficient degassing of the mobile phase. Vacuum degassing can also be performed off–line by applying a weak vacuum to the mobile phase reservoir while sonicating. Off–line techniques do not offer the advantage of continuous degassing throughout the analysis. Eluents can also be degassed by purging with helium, which has a very low solubility and drives the air out. This technique can be performed on–line and be controlled by the HPLC system, or off–line. Care must always be taken when degassing eluents that contain volatile components to avoid changing the composition.
It is convenient to prepare eluents as volume plus volume mixtures of solvents (i.e. the volume of each solvent is measured separately and then mixed). Volume changes can occur when solvents are mixed (e.g. methanol and water show a contraction in volume), which must be remembered if the volume of only one solvent is measured and the second solvent added to make up to volume (v/v).
True pH values can only be measured in aqueous solutions and any measurements made with a pH meter in aqueous–organic solvents should be described as ‘apparent pH’. In general, the apparent pH of a buffer solution rises as the proportion of organic solvent in the aqueous mixture increases. When an eluent is prepared it is usually best to dissolve the required buffer salts in water at the appropriate concentrations, adjust the pH and then mix this solution (v/v) with the organic solvents.
When the mobile–phase composition does not change throughout the course of the run, it is said to be isocratic. A mixed mobile phase can be delivered at a constant ratio by the pumps themselves or the solvent mixture can be prepared prior to analysis and pumped through a single reservoir. This is the simplest technique and should be the method of first choice when developing a separation.
HPLC can be performed with changes in composition over time (gradient elution). The elution strength of the eluent is increased during the gradient run by changing polarity, pH or ionic strength. Gradient elution can be a powerful tool to separate mixtures of compounds with widely different retention. A direct comparison can be drawn with temperature programming in gas chromatography (GC).
Eluent gradients are usually generated by combining the pressurised flows from two pumps and changing their individual flow rates with an electronic controller or data system, while maintaining the overall flow rate constant. Alternatively, a single pump with a low sweep volume can be used in combination with a proportioning valve, which controls the ratio of two liquids that enter the pump from two liquid reservoirs. Equipment and data systems that allow the gradient to take almost any conceivable form (e.g. step gradients, concave and convex gradient curves) are commonly available. The gradient can be programmed to return the system to the original eluent composition for the next analysis.
While most, if not all, commercially available pumps are capable of performing reliable gradient elutions, there are some potential difficulties. The technique can be very time consuming, as the column must be reconditioned with the initial eluent between runs. This drawback can be overcome by utilising a column–switching apparatus (see elsewhere in this chapter). In addition, drifting of the detector response and the appearance of spurious peaks that arise from solvent impurities may occur. While isocratic elution is usually favoured over gradients for simplicity, gradient elution can be a very important and useful technique in the separation of complex mixtures.
Recently, the use of ‘fast gradient’ separation has enabled the implementation of high throughput analysis in laboratories with a high sample load.
Derivatisation involves a chemical reaction that alters the molecular structure of the analyte of interest to improve detection and/or chromatography. In HPLC, derivatisation of a drug is usually unnecessary to achieve satisfactory chromatography. This applies to compounds of all polarities and molecular weights and is an important advantage of HPLC over GC. Derivatisation is used to enhance the sensitivity and selectivity of detection when available detectors are not satisfactory for the underivatised compounds. Both UV-absorbing and fluorescent derivatives have been used widely. UV derivatisation reagents include N-succinimidyl-p-nitrophenylacetate (SNPA), phenylhydrazine and 3,5–dinitrobenzoyl chloride (DNBC), while fluorescent derivatives can be formed with reagents such as dansyl chloride (DNS-Cl), 4–bromomethyl–7–methoxycoumarin (BMC) and fluorescamine. The characteristics of a good derivative in HPLC are similar to those in GC (i.e. stability, low background, convenience, etc.).
Derivative formation can be carried out before the sample is injected on to the column (pre–column) or by on–line chemical reactions between the column outlet and the detector. Such post–column reactions generally involve the addition of reagents to the eluent. With pre–column derivatisation there are no restrictions on reaction conditions (e.g. solvent, temperature) and a large excess of reagent can be used, as this can be separated from the derivatives during the chromatography. The major drawback of pre–column reactions is the need to obtain reproducible yields for accurate quantification, which is best achieved when the reactions proceed to completion. Furthermore, it is important that the products of pre–column derivatisation reactions be characterised fully. With post–column derivatisation, the reaction is well controlled by the flow rates of eluate and reagents, temperature, etc. Hence, it is less necessary for the reaction to proceed to completion or even for the chemistry to be understood as the system is calibrated by the injection of known quantities of the reference standards. A much more detailed discussion can be found in Snyder et al. (1997).
Separation of compounds by chiral chromatography began in the early 1980s. At that time, the separation of enantiomeric compounds was one of the most challenging problems in chromatography. However, in recent years more than 100 chiral columns have been made available. These columns are based on several different approaches to solve the many enantiomeric separation problems. Chiral columns are used in a variety of different applications that range from pharmacokinetic and pharmacodynamic studies to measuring enantiomeric impurity of amino acids.
Chiral stationary phases (CSPs) are designed to separate optical isomers. The use of these columns provides an efficient and economical way to separate optical isomers by HPLC. CSPs are used for both resolving optical isomers to determine enantiomeric purity and for isolating enantiomerically pure compounds. Fig. 29.2 shows the separation of enantiomers of flurbiprofen.
The columns can be classified according to two categories, class or origin. The class category is based on the structural properties of the chiral selector. The category is made up of five different column types (macrocyclic, polymeric, π–π associations, ligand exchange, miscellaneous) and hybrids. The macrocyclic chiral columns have had the largest impact on analytical enantiomeric separations. The origin category separates columns according to their source and classifies them into three types (naturally occurring, semisynthetic and synthetic chiral selectors).
The speed of a chromatographic method directly affects the economy and operating cost of the separation. High–speed HPLC is accomplished by using short microbore columns packed with small particles (3 μm). In addition, the use of higher temperatures increases the speed of HPLC separations through the 5- to 10–fold decrease in eluent viscosity upon an increase of the eluent’s temperature from 25 to 200°. High–temperature/high–speed HPLC is not universally useful because of several limitations. Silica–based stationary phases are unstable in aqueous media at temperatures above 50 to 60°. Some detectors are also not able to tolerate hot temperatures.
The quantification methods incorporated in HPLC derive mostly from GC methods. The basic theory for quantification involves the measurement of peak height or peak area. To determine the concentration of a compound, the peak area or height is plotted versus the concentration of the substance (Fig.3). For peaks that are well resolved, both peak height and area are proportional to the concentration. Three different calibration methods, each with its own benefits and limitations, can be utilised in quantitative analysis, external standard, internal standard and the standard addition method.
Fig.3. Example of a calibration curve for pseudohypericin.
The external standard method is the simplest of the three methods. The accuracy of this method is dependent on the reproducibility of the injection of the sample volume. To perform this method, a standard solution of known concentration of the compound of interest is prepared. A fixed amount, which should be similar in concentration to the unknown, is injected. Peak height or area is plotted versus the concentration for each compound. The plot should be linear and go through the origin. The concentration of the unknown is then determined according to Equation (29.3),
The calibrator concentrations should cover the range of the likely concentration in the unknown sample. Only concentrations read within the highest and lowest calibration levels are acceptable. Concentrations read from an extrapolated regression line may not be accurate. This applies to all of the quantification methods.
Although each method is effective, the internal standard method tends to yield the most accurate and precise results. In this method, an equal amount of an internal standard, a component that is not present in the sample, is added to both the sample and standard solutions. The internal standard selected should be chemically similar to the analyte, have a retention time close to that of the analyte and derivatise in a similar way to the analyte. For biological samples, the internal standard should extract similarly to the analyte without significant bias toward the internal standard or the analyte. Additionally, it is important to ensure that the internal standard is stable and that it does not interfere with any of the sample components. The internal standard should be added before any preparation of the sample so that extraction efficiency can be evaluated. Quantification is achieved by using ratios of peak height or area of the component to the internal standard, Equation (29.4):
The third method for quantification is the standard addition approach. This is especially useful when there is a problem with interference from the sample matrix, since it cancels out these effects. To perform this quantification, the sample is divided into two portions, so that a known amount of the analyte (a spike) can be added to one portion. These two samples, the original and the original–plus–spike, are then analysed. The sample with the spike shows a larger analytical response than the original sample because of the additional amount of analyte added to it. The difference in analytical response between the spiked and unspiked samples results from the amount of analyte in the spike. This provides a calibration point to determine the analyte concentration in the original sample. The method has a drawback if only a small volume of sample is available. Equation (29.5) is used for this method:
It is important to use a validated HPLC method when carrying out analyses. Typical analytical characteristics evaluated in an HPLC validation may include precision, accuracy, specificity, limit of detection, limit of quantification, linearity and range. Some appropriate suggestions for LC validation for postmortem and body fluids samples are published in the SOFT/AAFS Forensic Toxicology Laboratory Guidelines (http://www.soft-tox.org). It is important to consider the US Food and Drug Administration (FDA; http://www.fda.gov/cder/guidance) and US Pharmacopoeia (USP; http://www.usp.org) guidelines when validating HPLC methods used for pharmaceutical samples. USP 24 section <1225> provides guidance on the validation of compendial methods including definitions and determination. International Conference on Harmonisation (ICH) guidelines (http://www.ich.org) provide suggestions concerning the validation of pharmaceuticals. Valuable sources of information providing regulatory guidance may be found in the FDA website at http://www.fda.gov/cder/guidance.
System suitability tests evaluate the function of the overall HPLC system. This includes all parts that make up a system, such as the instrument, reagents, packing material, details of the procedure and even the analyst. These tests imply that the all the components of a system constitute a single system in which the overall function can be tested. These tests are very valuable and have been accepted in general application because reliable and reproducible chromatographic results are based on a wide range of specific parameters.
Most laboratories have a standard operating procedure that outlines the specifications of running a systems suitability test. For example, in pharmaceutical analysis at least five replicate injections should be made of a single solution that contains 100% of the expected active and excipient ingredients level. The peak response is measured and the standard deviation of that response should not exceed the limit set by the testing monograph or 2%, whichever of the two is the lowest. Using the USP method, the tailing factors of the analytes should be determined. The values should not exceed 2.0. Peak–to–peak resolutions are also determined by using the USP calculations and the value should not be lower than 1.5. The system test should be used to ensure the quality of the data and of the analysis.