166 DETECTION 210 220 230 240 250 260 270 280 Wavelength (nm) Absorbance (b) X Y X (260 nm) Y (260 nm) X + Y (260 nm) (a) Figure 4.12 Illustration of spectral deconvolution of analytes. (a) Hypothetical chro- matograms for individual injections of X and Y at 260 nm shown with combined response X + Y at 260 nm; (b) spectra for X and Y. ratio across the peak. The same dataset collected at 240 and 280 nm could be used to determine peak purity by calculation of the 240/280 ratio at every point across the peak. If the peak were pure X or Y, the ratio would be constant, whereas if the mixture of Figure 4.12a were present, the ratio would be > 1when X was predominant and <1whenY was the major compound. The nonconstant nature of the ratio would indicate the presence of a peak mixture, even though the peaks overlapped chromatographically and appeared as a single peak at 260 nm. Peak-purity algorithms compare the consistency of the spectrum across the entire peak and in some cases can identify the presence of minor impurities (e.g., <1%) that are eluted under the tail of the major peak. For additional examples of the determination of peak purity by DAD, see [20, 21] or the detector manufacturer’s literature (e.g., [22]). 4.4.4 General UV-Detector Characteristics Table 4.5 summarizes the general characteristics of UV detectors. UV detectors are ideal for use with gradient elution; many common, UV-transmitting solvents are available in HPLC grade for use as mobile phases (Tables I.2 and I.3 of Appendix I). The UV detector is very useful for the trace analysis of UV-absorbing solutes, but its widely varying response for different solutes can be a disadvantage if the compound of interest does not absorb in the UV (or visible) region. UV detectors are reliable and easy to operate, and are particularly suitable for use by less-skilled operators. 4.5 FLUORESCENCE DETECTORS 167 Table 4.5 UV-Detector Characteristics Capable of very high sensitivity (for samples that absorb in the UV) Good linear range ( > 10 5 ) Can be made with small cell volumes to minimize extra-column band broadening Relatively insensitive to mobile-phase flow and temperature changes Very reliable Easy to operate Nondestructive of sample Widely varying response for different solutes Compatible with gradient elution Detection wavelength can be selected Internal troubleshooting and calibration checks are common Built-in test procedures that can be carried out at detector startup identify many potential detector problems and can provide automatic wavelength calibration. The background, or baseline absorbance, of UV detectors can increase with continued use. This usually indicates that the cell windows have become dirty and need cleaning or replacement. Regular detector-cell flushing (as when the column is flushed) and sample cleanup can make more thorough cell cleaning a rarity. Lamp life, a concern in the past, is seldom an issue today. Useful lifetimes of > 2000 hr are common, and internal circuitry monitors lamp performance and can alert the user when the lamp output has deteriorated. Although the linear response range of UV detectors may be > 2 AU, according to manufacturer’s specifications, most analysts try to operate the detectors at <1 AU for best results. Stabilizing the flow-cell temperature through thermostatting or use of a capillary-tubing heat exchanger helps to reduce noise and drift from flow rate or temperature changes. Figure 4.13a shows an example chromatogram for the determination of derivatized roxithromycin (ROX) in human plasma by UV detection at 220 nm [23]. An internal standard, erythromycin (IS), was added to 50 μL of plasma followed by solid-phase-extraction sample cleanup and derivatization with 9-fluorenylmethyl chloroformate (FMOC-Cl). With UV detection at 220 nm, the method could monitor plasma concentrations of ROX but was unable to reach the LLOQ of <1 μg/mL necessary for pharmacokinetic studies. (See discussion of Section 4.5 for comparison of the UV response of Fig. 4.13a for this sample to the fluorescence response of Fig. 4.13b.) 4.5 FLUORESCENCE DETECTORS Fluorescence detectors are very sensitive and selective for solutes that fluoresce when excited by UV radiation. Sample components that do not fluoresce do not produce a detector signal, so sample cleanup may be simplified. For example, a simple acetonitrile/buffer extraction allowed detection of as little as 30 pg of (naturally 168 DETECTION Time ( min ) IS (area = 951) ROX (area = 1901) IS (area = 961) ROX (area = 609) UV absorbance (220 nm)Fluorescence (255 nm ex/315 nm em) (a) (b) UV fluorescence 501015 501015 Figure 4.13 Chromatogram for the determination of roxithromycin (ROX) in human plasma by (a) UV detection at 220 nm, and (b) fluorescence detection (excitation 255 nm, emis- sion 315 nm). Retention: ROX (10.7 min), internal standard erythromycin (5.1 min), both cleaned up by solid-phase extraction and derivatized with 9-fluorenylmethyl chloroformate (FMOC-Cl). Adapted from data of [23]. fluorescing) riboflavin in food products by HPLC with fluorescence detection [24]. Fluorescent derivatives of many nonfluorescing analytes can also be prepared (e.g., [25]), and this approach can be attractive for the selective detection of compounds for which sensitive or selective detection methods are otherwise not available. A schematic of a fluorescence detector is shown in Figure 4.14. The light source usually is a broad-spectrum UV lamp, such as the deuterium lamp used in UV detectors, or a xenon flash lamp. The excitation wavelength is selected by a filter or monochromator, and it illuminates the sample as it passes through the flow cell. When a compound fluoresces, the desired emission wavelength is isolated with a filter or monochromator and directed to a photodetector, where it is monitored and converted to an electronic signal for data processing. Because fluorescence is emitted in all directions, it is common to monitor the emitted light at right angles to the incident light—this simplifies the optics and reduces background noise. The least 4.5 FLUORESCENCE DETECTORS 169 lamp filter or monochromator photocell sample in sample out Figure 4.14 Schematic of a fluorescence detector. Dashed lines show optical path. expensive fluorometers use filters to select both excitation and emission wavelengths, whereas the most expensive use two monochromators (allowing a wide choice for both excitation and emission wavelengths). Remember, the fluorescence process is not 100% efficient, so energy is lost. This means that the emission wavelength always must be at lower energy (higher wavelength) than the excitation wavelength. For many samples, the fluorescence detector is 100-fold more sensitive than UV absorption—and is one of the most sensitive HPLC detectors. In other cases the sensitivity advantage of fluorescence over UV detection may be smaller but adequate for the task at hand. A comparison of the detector response to roxithromycin (ROX) by fluorescence and UV is shown in the RPC separations of Figure 4.13 [23]. ROX does not fluoresce naturally, so derivatization (9-fluorenylmethyl chloroformate [FMOC-Cl]) of the sample and internal standard (IS) was used to enable detection by fluorescence. When comparing the UV response (Fig. 4.13a) to fluorescence (Fig. 4.13b), the fluorescence response for the derivatized IS is approximately the same as the UV response, but the derivatized ROX peak response tripled with fluorescence detection. The baseline noise was approximately the same for both UV and fluorescence. This increase in response by the fluorescence method was adequate to reduce the LLOQ to <1 μg/mL of ROX in human plasma, which was required for pharmacokinetic studies. Because of its high sensitivity the fluorescence detector is particularly useful for trace analysis, or when either the sample size is small or the solute concentration is extremely low. The linear dynamic range of the fluorescence detector usually is smaller than that of UV detectors, but it is more than adequate for most trace analysis applications. While the dynamic range (the range over which a change in sample concentration produces a change in the detector output) of fluorescence detectors can be fairly large (e.g., 10 4 ), the linear dynamic range may be restricted for certain solutes to relatively narrow concentration ranges (as low as 10-fold). For all quantitative analyses using the fluorescence detector (or any other detector, for that matter), the linear range should be determined through the use of appropriate calibration (Section 11.4.1). In comparison to other detection techniques, fluorescence generally offers greater sensitivity and fewer problems with instrument instability (e.g., from temper- ature and flow changes). If solvents and mobile-phase additives free of fluorescing materials are used, the fluorescence detector can be used with gradient elution. The major disadvantage of the fluorescence detector is that not all compounds fluoresce. 170 DETECTION 0 Time (min) He Heair Fluorescence Figure 4.15 Fluorescence quenching of naphthalene by dissolved oxygen in the mobile phase. Mobile phase sparged with helium (He) or air, as shown. Adapted from data of [25]. As with other fluorescence techniques, fluorescence detection can be compromised by background fluorescence of the mobile phase or sample matrix, and by quenching effects. An example of fluorescence quenching is shown in Figure 4.15 [25]. When the mobile phase is sparged with helium, a consistent signal is observed, but when air is bubbled through the mobile phase, the signal drops because oxygen quenches the fluorescence of the naphthalene peak (250-nm excitation, 340-nm emission). Sparging the oxygenated mobile phase with helium then displaces the oxygen and the signal returns to normal. The presence of oxygen in the mobile phase also shifts the baseline slightly, but this is of minor concern. The use of a laser (laser-induced fluorescence, LIF) as the excitation source is available in the LIF detector. The higher energy of the laser over the conventional deuterium or xenon lamp gives added sensitivity to this detector, but the excitation wavelength range is more limited (300–700 nm vs. 200–700 nm for conventional fluorescence). LIF detection is not widely used with conventional HPLC systems, but is more common with micro applications (micro-LC, capillary LC, capillary electrophoresis, etc.) where a small diameter (e.g., 100-μm i.d.) flow cell is required to limit dispersion. 4.6 ELECTROCHEMICAL (AMPEROMETRIC) DETECTORS Many compounds that can be oxidized or reduced in the presence of an electric potential can be detected at very low concentrations by selective electrochemical (EC) measurements. By this approach the current between polarizable and reference electrodes is measured as a function of applied voltage. Because a constant voltage normally is imposed between the electrodes, and only the current varies as a result of solute reaction, EC detectors are more accurately termed amperometric devices. EC detectors can be made sensitive to a relatively wide variety of compound types, as illustrated in Table 4.6. EC detection is common for the determination of catecholamine and other neurotransmitters. Many of the compounds in Table 4.6 also can be detected by UV absorption, but some compound types (e.g., aliphatic mercaptans, hydroperoxides) sensed by EC detection cannot be detected at all by UV absorption, or only with difficulty and low sensitivity at low wavelengths. 4.6 ELECTROCHEMICAL (AMPEROMETRIC) DETECTORS 171 Table 4.6 Some Compound Types Sensed by the EC Detector Oxidation Reduction Phenolics Ketones Oximes Aldehydes Mercaptans Oximes Peroxides Conjugated acids Hydroperoxides Conjugated esters Aromatic amines, diamines Conjugated nitriles Purines Conjugated unsaturation Heterocyclic rings a Activated halogens Aromatic halogens Nitro compounds Heterocyclic rings Note: Compound types generally not sensed include ethers, aliphatic hydrocarbons, alcohols, and car- boxylic acids. a Detected depending on structure. EC detectors can be used only under the condition that the mobile phase is electrically conductive, but this is a minor limitation, since most HPLC separations are done by reversed-phase with water or buffer in the mobile phase. By fine-tuning the detector potential, one can achieve great selectivity for electroactive compounds. The EC detector’s sensitivity makes it one of the most sensitive of all HPLC detectors, for example with detection limits to 50 fg on-column of dopamine. However, to operate under high sensitivity, extra care must be taken to use highly purified mobile phases to reduce background noise. In order to reduce the background noise, in some applications the mobile phase is routed through a high-potential pretreatment cell so as to oxidize or reduce background interferences before the mobile phase reaches the autosampler. A glassy carbon electrode is most commonly used in the electrochemical cell. In the configuration shown in Figure 4.16, the column effluent flows across a glassy carbon electrode, whereas in another popular configuration, the sample flows through a porous graphite electrode. Several electrode styles are available, for example, Figure 4.16c shows a dual-electrode configuration. The high susceptibility of the EC detector to background noise and electrode contamination has earned it a reputation as a difficult detector to use. However, newer units are much more trouble free and can provide excellent and reliable results in the hands of a reasonably careful operator. Figure 4.17 shows the electrochemical detection of acteoside, an active ingre- dient in many Chinese medicinal plants. Following intravenous administration of acteoside at 10 mg/kg, the analyte was detected in rat brain microdialysate at a concentration of ≈25 ng/mL (≈0.4 ng on-column) by reversed-phase HPLC [26]. More information about electrochemical detectors can be found in [27]. 172 DETECTION sample inlet sample outlet locking collar reference electrode o-ring auxiliary electrode block gasket working electrode block quick- release mechanism (a) (b) (c) electrode flow in flow out Figure 4.16 Schematic of an electrochemical detector. (a) Top view of assembled flow cell; (b) exploded diagram of cell; (c) detail of dual electrode cell. Courtesy of Bioanalytical Sys- tems, Inc. 4.7 RADIOACTIVITY DETECTORS Radioactivity detectors are used to monitor radio-labeled solutes as they elute from the HPLC column. Detection is based on the emission of light in the flow cell as a result of radioactive decay of the solute, followed by emission of α-, β-, or γ -radiation. The continuous-flow monitoring of β-radiation in the eluent ordinarily involves the use of a scintillation technique, where the original radiation is converted to light. Depending on the method of combining the eluent and the scintillator, this can be classified as either a homogeneous or heterogeneous system. In homogeneous operation, a liquid-scintillation cocktail is mixed with the column effluent prior to entering the detection cell, where emitted light is monitored. Under heterogeneous conditions, the column outlet is routed directly into the detector cell, which is packed with beads of a solid scintillant. When adsorption of the analyte on the beads is a problem, the scintillant may be coated onto the walls of the detector cell. Homogeneous detectors are best used with analytical procedures where recov- ery of the sample is unimportant. The technique also can be applied to preparative HPLC, when a portion of the sample stream is split off to the detector. Hetero- geneous detectors are less sensitive, and therefore better suited for samples with 4.7 RADIOACTIVITY DETECTORS 173 0 5 15 2510 20 0 0.4 0.8 1.2 Time (min) Detector response (nA) acteoside Figure 4.17 Determination of acteoside (t R ≈ 15 min) in rat brain microdialysate with elec- trochemical detection. Adapted from data of [26]. higher levels of radioactivity (or for larger solute concentrations, as in preparative separations). Heterogeneous systems also are relatively free of chemical quenching effects, and solutes can be recovered easily. However, this detector exhibits relatively low counting efficiency for low-energy β-emitters, such as 35 S, 14 C, 3 H, and 32 P, and is better suited for stronger α-, β-, and γ -emitters (e.g., 131 I, 210 Po, and 125 Sb). One application of the radioactivity monitor is to determine the complete distribution and mass balance of a radio-labeled pharmaceutical dosed in an experimental animal. Such determinations are difficult, if not impossible, without the aid of radio-labeled drugs. Radiochemical detectors have a wide response range and are insensitive to solvent change, making them useful with gradient elution. With radioactivity detec- tors, it may be necessary to compromise sensitivity to improve chromatographic resolution and speed of analysis. Detection sensitivity is proportional to the number of radioactive decays that are detected, and this number is proportional to the volume of the flow cell and inversely proportional to the flow rate (proportional to residence time, which allows more atoms to decay during passage of a peak through the flow cell). Larger flow-cell volumes increase extra-column peak broadening and can diminish resolution, while slower flow rates mean an increase in separation time. Because detection sensitivity is often marginal, larger flow cells are generally preferred for radioactivity detection. In practice, peak tailing and peak broadening in a radiometric flow cell can be minimized by working with columns of larger volume (assuming that sufficient sample is available for larger mass injections to compensate for sample dilution).With radioactivity detection, a compromise between chromatographic resolution and detector sensitivity must be reached, the exact nature of which depends on the analytical requirements. 174 DETECTION 4.8 CONDUCTIVITY DETECTORS Conductivity detectors use low-volume detector cells to measure a change in the conductivity of the column effluent as it passes through the cell. Conductivity detectors are most popular for ion chromatography and ion exchange applications in which the analyte does not have a UV chromophore. Analysis of inorganic ions (e.g., lithium, sodium, ammonium, potassium) in water samples, plating baths, power plant cooling fluids, and the like, is an ideal use of the conductivity detector. Organic acids, such as acetate, formate, and citrate are also conveniently detected by conductivity. Conductivity detection can be compromised by the presence of a conductive mobile phase; for example, the mobile-phase buffer. Thus the presence of the buffer greatly increases the conductance of the mobile phase, which is only slightly increased by the presence of the solute. One way to minimize this problem is to use a suitable buffer in combination with a suppressor column (ion exchanger), in order to reduce the background conductivity of the mobile phase. For example, consider the need to detect one or more anionic solutes. The use of a Na 2 CO 3 -NaHCO 3 buffer with a cation-exchange suppressor column (termed an anion suppressor in ion chromatography terms) in the H + form will eliminate Na + and other cations from the mobile phase, and convert carbonate to weakly acidic carbonic acid. This reduces the conductivity of the mobile phase and allows an easier detection or small concentrations of anionic solutes. The application of a suppressor column is illustrated in Figure 4.18 for the dramatic improvement in conductivity detector response to F − ,Cl − ,andSO 2− 4 . 4.9 CHEMILUMINESCENT NITROGEN DETECTOR One advantage that gas chromatography has over HPLC is the availability of several element-specific detectors, allowing selective detection of compounds containing nitrogen, sulfur, or phosphorus. In the 1970s much effort was given to developing element-specific detectors for HPLC, but for the most part the results have been discouraging. One exception is the chemiluminescent nitrogen detector (CLND), which was reported as early as 1975 [28]. Several commercial implementations and refinements have resulted in today’s CLND. The HPLC column effluent is nebulized with oxygen and a carrier gas of argon or helium and pyrolyzed at 1050 ◦ C. Nitrogen-containing compounds (except N 2 ) are oxidized to nitric oxide (NO), which is then mixed with ozone to form nitrogen dioxide in the excited state (NO 2 *). NO 2 * decays to the ground state releasing a photon, which is detected by a photometer. The signal is directly proportional to the amount of nitrogen in the original sample, so calibrants of known nitrogen content can be used to quantify the nitrogen content of unknown analytes. This is illustrated in Figure 4.19a [29], where the injection of 50-ng nitrogen equivalents of 7 different compounds give detector responses that are constant within ±10%. Care must be taken to maintain a nitrogen-free mobile phase, so the use of acetonitrile is ruled out. Many solvents are compatible with the CLND, as is shown in Figure 4.19b for the response of the injection of 1 mg each of 6 nitrogen-free solvents, compared to an injection of 1 ng nitrogen-equivalent of a standard. 4.10 CHIRAL DETECTORS 175 F – Cl – mobile phase (Na 2 CO 3 ) sample (F – , Cl – , SO 4 2– ) analytical column anion suppressor NaF, NaCl, Na 2 SO 4 in Na 2 CO 3 waste H 2 OH 2 O waste +− HF, HCl, H 2 SO 4 in H 2 CO 3 Time μS Time μS Without Suppression With Suppression counter ions F – Cl – SO 4 2– SO 4 2 – (a) (c) (b) conductivity detector Figure 4.18 Use of an anion suppressor column to enhance conductivity detector response to anionic analytes. (a) Schematic of instrumentation; (b) conductivity detector output without suppressor column; (c) chromatogram with suppressor column in use. Courtesy of Dionex. One detector manufacturer claims detection limits equivalent to 0.1 ng of nitrogen. A practical example is seen in Figure 4.20a [30] for the detection of 13 underivatized amino acids by ion-pair chromatography and CLND. The response per nitrogen atom is within 6% RSD, with detection limits of ≈0.3to0.5μg/mL for the amino acids. Figure 4.20b shows the chromatogram for an injection of 10 μLof wine filtered through a 1000-Da filter (note overloaded proline peak shows shorter retention and strong tailing compared to a; see Section 2.6 for further discussion of overload). 4.10 CHIRAL DETECTORS Chiral drug candidates often are encountered in the development of new pharmaceu- tical compounds. Different enantiomers can possess different efficacy, toxicology, or other pharmacological characteristics, and the final product generally is a single enantiomer or a known mixture of enantiomeric forms. Chromatographic separation of the enantiomers (Chapter 14) is vital to the analysis of such mixtures. Detection and identification can be further aided by the use of detectors that respond selectively to specific chiral forms. Chiral detectors come in three different formats; each of these uses the same principles as stand-alone instrumentation, but in a flow-cell format. Polarimeters (PL) measure the degree of rotation of polarized light (typically in the 400–700 nm range) as it passes through the sample. The degree of rotation is dependent on . visible) region. UV detectors are reliable and easy to operate, and are particularly suitable for use by less-skilled operators. 4.5 FLUORESCENCE DETECTORS 167 Table 4.5 UV-Detector Characteristics Capable. directed to a photodetector, where it is monitored and converted to an electronic signal for data processing. Because fluorescence is emitted in all directions, it is common to monitor the emitted. 21] or the detector manufacturer’s literature (e.g., [22] ). 4.4.4 General UV-Detector Characteristics Table 4.5 summarizes the general characteristics of UV detectors. UV detectors are ideal for