In order to gain useful information from the separation technique, it is necessary to detect and measure the analytes. Detection may be qualitative and/or quantitative. Most CE detection is done on-capillary; that is, a section of the capillary is linked to the detection device and the capillary itself is the detection cell. It is also possible to couple to detectors that are outside of the separation capillary although this does require a specialized interface.

1.4.1. Geometry and Path-Length

The most frequently used method of detection involves absorbance of energy as the analytes move through a focused beam of light. The scale of the detection apparatus and of the signal produced has created some unique challenges for the instrument developer.

On-capillary detection eliminates the problems of coupling the capillary and its power supply to flow cells or other devices. However, detection through the capillary is complicated by the curvature of the capillary itself. The capillary and the fluid it contains make up a complex cylindrical lens. The curvature of this lens must be accounted for in order to gather the maximum amount of light and thereby maximize signal-to-noise ratio. The effective length of the light path through the capillary is actually about 63.5% the stated i.d. of the capillary. Thus a 50-^m capillary has an effective path length of only 32 ^m. This can be compared to typical HPLC detectors that have detectors in the 5-10 mm range. Because of this very small light path, the absorbance signal obtained from a CE system is also correspondingly small. Therefore, a peak with an absorbance of 0.002 AU is a significant peak. Noise levels are correspondingly small and are usually measured in microabsorbance units. The maximum absorbance of typical CE detectors is 0.2 AU. The capillary lumen occupies only a small part of the diameter of the capillary; the remainder is transparent silica. This geometry allows for large amounts of stray light to enter the detector so that a fully opaque sample would not reduce the light passing through the capillary to zero. This factor must be considered in the design of effective CE detector hardware and software.

Several novel approaches have been taken to increase the path length (and the sensitivity) of CE detectors. One approach has been to use a specially constructed low-volume flow cell. These cells carry the analytes through two right-angle bends (30). The segment between the bends, which may be over 1000 ^m long, is thus at right angles to the direction of the capillary and parallel to the direction of the light beam. Properly designed, these cells can dramatically increase sensitivity but at the risk of some loss of resolution. Another approach has been to create a wide zone or "bubble" in the capillary at the window (31). A 50-^m i.d. capillary may have a window that is 150 ^m in diameter, giving a sensitivity increase of approximately threefold.

1.4.2. Absorbance

Absorbance detectors are the most commonly encountered types of detector in CE instrument systems. They rely on the absorbance of light energy by the analytes. This absorbance creates a shadow as the analytes pass between the light source and the light detector. The intensity of the shadow is proportional to amount of material present.

The simplest absorbance detector, shown in Fig. 11A uses only portion of the available energy. The broad-spectrum light from a source lamp is passed through a filter or diffracted by a grating so that a narrow range of the spectrum is used. In some cases lamps (such as hollow cathode types) are used, that produce light at only a few discrete wavelengths. Monochromatic

Fig. 11. Comparison of single wavelength (A) and photodiode array (B) absor-bance detectors.

absorbance detectors are relatively inexpensive and rugged. However, because only a part of the spectrum is used, information may be lost from complex samples that have components with differing absorbance maxima.

Another type of absorbance detector looks at changes over a wide range of wavelengths simultaneously. The photodiode array detector (PDA) shown in Fig. 11B delivers the entire spectrum of light available from the source


Fig. 12. Laser-induced fluorescence detector.


Fig. 12. Laser-induced fluorescence detector.

lamp to the capillary window. The light passing through the capillary is diffracted into a spectrum that is projected on a linear array of photodiodes. In this manner it is possible to record the entire absorbance spectrum of analytes as they pass by the detector window. PDA-type detectors are more expensive and less rugged than are monochromatic detectors. Because the spectrum can be divided into as many as 512 channels the amount of data acquired in a single run can be very large. The PDA type detector, however, is not suitable for all applications because nearly all the energy of the source lamp is focused onto a very small region of the capillary. Some capillary coatings and buffers will decompose under this onslaught of energy unless some of the energy is filtered out. Despite these limitations the information provided by the PDA detector can be valuable for confirming the identity of analytes. By comparing the change in spectral signature across a peak it is possible to estimate peak purity (32).

1.4.3. Fluorescence

Fluorescence detectors do not rely on the measurement of shadows. These systems use an external energy source to excite the analyte molecules to a higher energy state. When these excited molecules return to the normal state they emit energy of a lower wavelength, which can be detected and recorded as evidence of the passage of the analytes. Fluorescent detectors in CE systems often use lasers as the source of the excitation energy (LIF detection, Fig. 12). Lasers have the advantage of producing intense light at a single wavelength. The intensity of the light contributes to good excitation efficiency. In addition, the monochromatic nature of the laser beam makes it easy to filter out any stray laser light to keep it from interfering with the detection of analytes. Analytes will vary in their excitation and absorbance wavelengths so that a fluorescent detector will not see all the components that may be in a sample. For analytes that are fluorescent or can be made fluorescent by a chemical reaction, the sensitivity of this type of detector can be 10-1000 times better than an absorbance detector (33).

1.4.4. Amperometry and Conductivity

These two detection techniques potentially can offer a high degree of sensitivity and be applied to a wide variety of analytes, including those without appreciable UV absorption. Amperometry and conductivity are difficult to do in practice, and to date these devices are not commercially available. Both of these detection schemes require the use of sensing electrodes that are scaled to the dimensions of the electrophoresis capillary. In addition, these detectors place some limitations on the type of separation buffer employed.

In amperometric detection, an electroactive analyte undergoes an electrochemical reaction inside a detector cell (34). The CE separation is generally carried out at microampere currents and kilovolt potentials, whereas the detection cell must operate at picoampere currents and millivolt potentials. The two circuits must therefore be isolated, usually by connecting the capillary to the high voltage at a point prior to the end of the capillary where the detection cell is located. This system depends on EOF to carry the analyte past the high voltage electrode to the detection cell. Amperometric detectors have been used successfully for the detection of biogenic amines at levels as low as 10-8 m (35). With continued advances in microfabrication this type of detector should become routinely available.

There are two types of conductivity detector (36,37). Both require that two electrodes be placed within the separation capillary. In the first type a short distance down the length of the capillary, separates the electrodes. Because there is a voltage gradient down the length of the capillary a portion of that voltage gradient can be measured between the two sensing electrodes. The presence of analyte zones will change the potential drop in the area of the zone. This change in potential will be sensed as the zone passes the electrodes. In the other arrangement the electrodes are placed opposite one another across the diameter of the capillary. In this case there is no voltage drop between the two electrodes. A circuit is constructed that passes a sensing voltage across the capillary diameter. If the conductivity of the system changes as analytes pass between the electrodes, there will be a change in the current in the sensing circuit. Conductivity detectors work best when there is a substantial difference between the conductivity of the analyte zones and the background buffer (38). These conditions are not optimized for peak shape and lead to unacceptable peak asymmetry. Various schemes have been described for avoiding this problem. Conductivity detectors have been used for the measurement of inorganic ions and in isotachophoretic separations (Section 2.4.).

1.4.5. Capillary Electrophoresis-Mass Spectrometry

Hybrid systems such as CE-mass spectrometry (CE-MS) offer an additional dimension of analysis in addition to detection. CE data consists of migration time, quantity, and (using a PDA detector) spectral signature. MS adds the additional data of molecular weight and, using collision dissociation and MS-MS systems, structural information as well. This combination of techniques provides an orthogonal approach to analysis in a single analytical run.

The predominant form of MS that has been coupled to CE has been electrospray MS (39). A more in-depth discussion of CE-MS can be found in Chapter 15. This section will focus on that technique. The outlet end of the CE capillary is inserted into the electrospray interface. Because the volume of liquid emerging from the capillary is very small, a make-up liquid is pumped through an axial needle. This sheath flow also provides a return connection to the high-voltage power supply of the CE system. The liquid is mixed with a flowing gas stream and nebulized into a spray. The spray vaporizes and the ionized analyte particles are carried into the MS detector. The MS system is usually set to scan across an expected range of mass values. Because the width of peaks in CE can be very small, the MS instrument must be able to scan across the desired mass range very rapidly or peaks may be missed.

Most mass spectroscopists prefer to use buffer systems that are volatile, such as ammonium formate, in order to reduce the accumulation of buffer salts inside the MS instrument. These buffers may not be optimized for the separation of the analyte mixture, although an incomplete separation can be acceptable in CE-MS because the MS systems provide an additional dimension of separation.

Many of the existing commercial CE systems have been interfaced to commercial MS systems. The design of these systems allows the use of UV or other detectors prior to the MS interface, but usually require quite long and awkward reaches of capillary to connect the two systems. Unlocking the true potential of this method will require the development of a CE system that is fully integrated with the MS system.

1.4.6. Indirect Detection

In the previous discussions it has been assumed that detection will be direct, that is, the presence of the sample in the detector cell or window will cause an increase in the output signal. This is not always the case. There are certain applications where the decrease of the background signal provides evidence of the passage of analytes through a detector. In some cases a detector that is useful for indirect detection would not detect the analytes in direct mode. An example of this is the detection of inorganic ions such as sodium or sulfate with a UV detector.

To detect such analytes, the capillary is filled with a buffer that has mobility close to that of the analytes and also has a significant UV absor-bency. The buffer/chromophore must carry the same charge as the analyte. For the analysis of sulfate, a buffer containing sodium chromate is used (40). In the area of the capillary occupied by the sulfate band the chromate is displaced, creating a chromate depleted region. Because chromate absorbs light at 254 nm, by monitoring background electrolyte absorbance at this wavelength the presence of sulfate will be indicated by a negative peak, with an area corresponding to the amount of chromate displaced and hence to the amount of sulfate present. Most CE data analysis packages can invert these negative peaks, producing a normal-appearing electropherogram. 1.4.7. Data Analysis

The collection and analysis of CE data has many characteristics in common with other chromatography-like analyses. Many analysts have utilized data systems developed for HPLC and GC systems. These data systems may not be optimal for CE for three reasons:

1. The signals obtained from CE are usually very small. A typical HPLC peak may have a maximum absorbance of 0.2 AU, whereas a CE detector may have a full range of 0.2 AU. A typical peak in CE may have a maximum absorbance of 0.002 AU.

2. CE peaks can be quite narrow with a width of only a few seconds. Some older data systems (particularly HPLC systems) cannot respond sufficiently fast to deal with this data. If data is to be collected digitally, it must be collected at a high data rate to define adequately the peak shape.

3. CE peaks are often non-Gaussian. Because of the low band broad spreading in CE and the effects of electrofocusing CE peaks tend to be more triangular and less bell-shaped than peaks in HPLC and GC. At times CE peaks may have one edge that is nearly vertical. Some data systems have difficulty locating the peak start and stop times for these shapes.

Two types of peak-detection methods dominate analytical data analysis: slope sensitive algorithm and the moving median filter algorithm. Slope sensitive algorithms look at the slope of the baseline over some interval of time. When the slope exceeds a pre-determined value, a peak is said to have begun. The point at which the slope goes to zero identifies the peak apex, and the point at which the slope returns to the starting value defines the peak end.

The slope-sensitive method looks for peaks independently of the baseline shape. Most commercial software packages use this method.

Moving median filter algorithms take a different approach. Peaks are relatively high frequency (impulse) events when compared to baseline drift. These algorithms seek to define how the baseline would look in the absence of peaks by filtering out all impulse events. Whatever differs from the baseline is defined as a peak. In practice, the algorithm makes several passes through the data set (iterations) to determine the best fit. This type of algorithm, commercially available as "Caesar," is more successful at dealing with the abrupt slope changes in CE data than are slope-sensitive algorithms (41).

A data system for CE needs to include some calculations that can be deemed "CE specific." These include the calculation of mobility (a measure of the velocity of an analyte through the capillary) and corrected peak area. Corrected area is necessary because the peaks passing through a CE detector do not all pass through at the same velocity. Early eluting peaks move through more rapidly than do later eluting peaks. This is unlike the situation in HPLC where the velocity through the detector cell is dependent only on the flow rate and not on the retention time. Because later eluting peaks are moving more slowly, they appear to be larger relative to earlier eluting peaks. Corrected peak area normalizes peak area to unit migration time and allows accurate comparison of the components in a mixture.

1.4.8. Fraction Collection

Fraction collection is placed under the heading of detection under the assumption that fractions are being collected for analysis outside the CE instrument. At first glance, fraction collection in CE appears to be analogous to fraction collection in HPLC, but there are significant differences (42).

The velocity of peaks through the HPLC detector is constant if the flow rate is constant. This is not true in CE where earlier eluting peaks are moving more rapidly than later eluting peaks. The delay time between peak detection and elution (the time for transit from the window to the capillary outlet) will therefore vary, becoming longer with each successive peak. This time can be quite long. Consider a capillary 60-cm long with the window located 10-cm from the outlet end. If a peak passes the detector at t = 5 min it will not emerge from the end of the capillary until t = 6 min. A peak detected at t = 10 min will emerge at t = 12 min. Peaks that pass the detector between 5 and 10 min will all be in the outlet segment of the capillary between t = 6 and t = 12 min. A further complication is that collection requires that the outlet of the capillary be moved to a new collection vial for each peak that is to be collected. At each move the circuit will be broken. The time needed for the power supply to ramp down and ramp up at each move, and the diminished peak velocities during the ramping segments cannot be disregarded. Complex algorithms are needed to do this process smoothly.

A more serious limitation to the collection of fractions from CE is the very small amount of sample that the fraction will contain. Consider a capillary 100 ^m in diameter and 50 cm long, with the window located 10 cm from the end. A 5 psi-sec injection into this capillary will inject about 189 nL of sample (see Section 1.3.1.). If the analyte concentration is 1 mg/mL in the starting sample and the peak is collected with 100% efficiency, the recovered sample will be 189 ng. Repeating this run 10 times would yield less than a two micrograms of product. Although fraction collection in CE is possible, one must question whether it is worth the effort.

Drug Free Life

Drug Free Life

How To Beat Drugs And Be On Your Way To Full Recovery. In this book, you will learn all about: Background Info On Drugs, Psychological Treatments Statistics, Rehab, Hypnosis and Much MORE.

Get My Free Ebook

Post a comment