Anal. Chem. 2004, 76, 1690-1695
Apparatus and Initial Results for Pressurized Planar Electrochromatography David Nurok,* James M. Koers, Allyson L. Novotny, Matthew A. Carmichael, and Justin J. Kosiba
Department of Chemistry, Indiana University Purdue University Indianapolis, 402 N. Blackford Street, Indianapolis, Indiana 46202 Robert E. Santini,*,† Gregory L. Hawkins, and Randall W. Replogle
The Amy Facility for Chemical Instrumentation, Department of Chemistry, Purdue University, West Lafayette, Indiana 47907
Pressurized planar electrochromatography (PPEC) is a new planar chromatographic technique in which the mobile phase is driven by electroosmotic flow, while the sorbent layer is pressurized in a manner that allows heat to flow from the layer through an electrically insulating, thermally conducting, sheet of aluminum nitride ceramic. A prototype apparatus for performing PPEC is described. Separation by PPEC is faster than by conventional TLC, and an example is presented of a 24-fold enhancement in the speed of separation. PPEC was performed on both regular and high-performance C18 layers, and the latter yield substantially faster separation. The sorbent layer requires conditioning at elevated temperature before use, and solute migration velocity increases with this temperature. The flow rate increases in a linear manner with increasing voltage and diminishes in a nonlinear manner with increasing pressure. Both electrical current and Joule heating diminish with increasing pressure, and the diminution of flow at high pressure can be compensated by an increase in voltage. PPEC is more efficient than classical TLC. Theoretical plate heights diminish with increasing Rf and are in the range 29-21 and 55-27 µm for the high-performance and regular plates, respectively. PPEC retains the advantages of classical TLC but has the ability to separate a substantially higher number of samples simultaneously. An example is presented on the separation of nine samples in 1 min on a 2.5 cm × 10 cm sorbent layer. Planar chromatographysalso called thin-layer chromatography (TLC)sis a very well established technique that has both attractive and unattractive features. The attractive features include the ability to separate multiple samples simultaneously, the fact that there is no need to transport the separated components to a detector, the ease of performing postseparation visualization reactions, the large number of such reactions that are available, and the ability to perform true two-dimensional separations (as compared to the virtual two-dimensional separations obtained by column chromatography). Despite these features, classical TLC is not widely used for quantitative analyses, even though quality scanners are * Corresponding authors. E-mail:nurok@chem.iupui.edu. † To whom questions on instrument design should be directed.
1690 Analytical Chemistry, Vol. 76, No. 6, March 15, 2004
commercially available. The major drawback of TLC is the poor flow profile, with the linear flow rate of the mobile phase given by
u ) κ/2Zf
(1)
where κ is the solvent velocity constant and Zf is the distance migrated by the solvent front. The inverse relationship between flow rate and the distance migrated by the solvent front results in a progressively slower rate of migration of the mobile phase. With the exception of separations that require only a short migration distance (typically on high-performance plates), the flow profile results in separations that can be lengthy (especially in the reversed-phase mode) and that are of low chromatographic efficiency. Forced-flow techniques have been introduced to overcome the problems related to the poor flow profile.1-3 The first forced-flow technique to be introduced was rotational planar chromatography,1 in which the TLC plate is rapidly rotated and the flow of mobile phase is driven by centrifugal force. The flow profile is better than in classical TLC, but suffers from the disadvantage that for geometric reasons the rate of mobile-phase flow diminishes as it moves from the center of the plate to the circumference. All the advantages of planar chromatography noted in the introduction are retained, apart from the ability to perform two-dimensional separations. Overpressured-layer chromatography (OPLC) is a method whereby the TLC layer is covered by a thin, flexible membrane that is subjected to high pressure.2 This allows mobile phase to be pumped through the layer, and the technique can be considered a form of high-performance liquid chromatography with a flat sorbent bed. It retains all the advantages of classical TLC, with the added advantage that the mobile-phase flow rate is controllable. OPLC can be performed in the “off-line” mode in (1) Nyiredy, Sz. In Planar Chromatography: A Retrospective View for the Third Millennium; Nyiredy, Sz., Ed.; Springer Scientific: Budapest, Hungary, 2001; pp 177-199. (2) Tyihak, E.; Mincsovics, E. In Planar Chromatography: A Retrospective View for the Third Millennium; Nyiredy, Sz., Ed.; Springer Scientific: Budapest, Hungary, 2001; pp 137-176. (3) Rozylo, J. K.; Malinowska, I. In Planar Chromatography: A Retrospective View for the Third Millennium; Nyiredy, Sz., Ed.; Springer Scientific; Budapest, Hungary, 2001; pp 200-219. 10.1021/ac0303362 CCC: $27.50
© 2004 American Chemical Society Published on Web 02/14/2004
which the separation is stopped while all analytes are still on the TLC plate or in the “on-line” mode in which the mobile phase flows through a detector after traversing the sorbent bed. Planar electrochromatography (PEC) is a separation technique that is performed on a TLC plate in the presence of a large electric field.3 Movement of the mobile phase is due to electroosmotic flow, but capillary mediated flow also occurs unless special precautions are taken. The technique was first described by Pretorius and co-workers4 in 1974, who reported a 15-fold increase in the speed of the technique as compared to conventional TLC. The report lacks sufficient experimental detail for duplicating the results, and this may have contributed to the absence of any further reports for 20 years. Since 1994, there have been several reports on PEC in both the normal-phase and the reversed-phase modes, and with initially dry5-7 or with prewetted layers.8-10 These reports were all for PEC performed at atmospheric pressure. Reports on the fastest separations,4,8 the longest migration distances,4,8 and those with the greatest peak resolution,9 were obtained with prewetted layers. There is only one recent report on PEC in the normal-phase mode with prewetted plates.8 The mobile phases used were either pure ethanol or pure acetonitrile, which are suitable for separating only highly polar solutes. PEC was ∼12 times faster than the corresponding separation by conventional TLC. PEC in the reversed-phase mode is not as fast but can be very efficient as demonstrated by the separation of seven compounds in less than 4 min.9 The performance of PEC at atmospheric pressure is limited by the nature of electroosmotic flow in an open bed. Under inappropriate conditions, mobile phase accumulates on the surface of the sorbent layer and this results in streaking of the solutes.9 This phenomenon has been ascribed to mobile phase traveling from the source along the surface of the plate10 or to liquid being driven to the surface by the electroosmotic force.9 The accumulation of liquid on the surface can be prevented by clipping a small glass plate, coated with paraffin oil, to the TLC layer close to the mobile-phase origin.10 The accumulation of liquid can also be prevented by increasing the amount of Joule heating (by increasing the ionic strength of the mobile phase or by increasing the applied voltage), and good results are obtained when there is a balance between liquid driven to the surface and liquid evaporating from the surface due to this heating.9 Under these conditions, solvent focusing due to evaporation of the mobile phase, causes some peak sharpening. If there is too much Joule heating, evaporation of the mobile phase is excessive, and the plate dries. Even under optimum conditions, the sharpest peaks have been reported only for short migration distances.9 (4) Pretorius, V.; Hopkins, B. J.; Schieke, J. D. J. Chromatogr. 1974, 99, 2330. (5) Pukl, M.; Prosek, M.; Kaiser, R. E. Chromatographia 1994, 38, 83-87. (6) Malinowska, I. J. Planar Chromatogr. 2000, 13, 307-313. (7) Malinowska, I.; Rozylo, J. K.; Krason, A. J. Planar Chromatogr. 2002, 15, 418-424. (8) Howard, A. G.; Shafik, T.; Moffatt, F.; Wilson, I. D. J. Chromatogr., A 1999, 844, 333-340. (9) Nurok, D.; Koers, J. M.; Carmichael, M. A. J. Chromatogr., A 2003, 983, 247-253. (10) Dzido, T. H.; Majewski, R.; Polak, B.; Golkiewicz, W.; Soczewinski, E. J. Planar Chromatogr. 2003, 16, 176-182.
The limitations of PEC at atmospheric pressure can be overcome by pressurizing the sorbent layer. This will prevent liquid from accumulating on the surface of the layer, in addition to other advantages discussed below. The pressurizing medium needs to be electrically insulated from the chromatographic system, and this can be conveniently done by covering the sorbent layer with a sheet of aluminum nitride ceramic, which is both an electrical insulator and a thermal conductor. The presence of the aluminum nitride ceramic sheet allows the pressurizing medium to act as a heat sink and thus maintain the sorbent layer at a constant temperature. The pressurizing medium can also be actively cooled by a circulating liquid. There are additional advantages to working at elevated pressure. The capillary channels between sorbent particles become narrower, and this should result in enhanced efficiency. Moreover, the narrowing of the channels will result in a lower electrical current and reduced Joule heating for any given voltage. Thus, PEC can be performed at a higher voltage for a given amount of Joule heating, and this should result in both faster and more efficient separations. Electroosmotic flow is directly proportional to the electric field,11 which is the applied voltage divided by the relevant length of the chromatographic system. This assumes that the resistance across the system is constant. It is not clear that this assumption can be made for the PPEC apparatus described below, and for this reason, voltage rather than electric field is used in this report. The report below describes a prototype apparatus for pressurized planar electrochromatography (PPEC). The initial results indicate that the technique can yield separations that are both rapid and of high efficiency. EXPERIMENTAL SECTION Materials Used. The hydraulic pump, gauge, tubing, and ram were purchased from EnerPac (Milwaukee, WI). Merck LiChrospher RP-18 WF254s plates (Catalog No. 1.05646) were supplied as a gift by Merck KgaA (Darmstadt, Germany), and Merck RP18 F254s plates (Catalog No. 15389-7) were supplied as a gift by EMD Chemicals, Inc. (Gibbstown, NJ). The latter are referred to as regular plates in the text. Acetonitrile, acetic acid, sodium acetate, and methanol were purchased from Fisher Scientific (Pittsburgh, PA). Benzanilide, 17R-acetoxyprogesterone, 2′-acetonaphthone, and o-nitroaniline were a gift from Don Risley of Eli Lilly and Co. 4-Cholesten-3-one was purchased from Sigma-Aldrich (St. Louis, MO). Analyte mixtures were prepared in methanol. The buffer solution was prepared by mixing equal volumes of 1 M acetic acid and 1 M sodium acetate. The concentration of the solution was then adjusted such that, when mixed (on a volume/volume basis) with acetonitrile and additional water, the mobile phase was at the desired molarity. The water was filtered with a Milli-Q system. The reported pH of 4.7 for the mobile phase is a nominal value and refers to the value before the addition of acetonitrile. The silicone rubber sealant was a gift from Dr. Emil Mincsovics. A commercially available sealant that can be used is ethylene (55%)/vinyl acetate (45%) copolymer beads obtainable (11) Baker, D. R., Capillary Electrophoresis; Wiley-Interscience: New York, 1995; p 23.
Analytical Chemistry, Vol. 76, No. 6, March 15, 2004
1691
from Sigma-Aldrich. The Teflon sheet was purchased from the Mc Master-Carr Supply Co. (Chicago, IL). The aluminum nitride ceramic sheet was a gift of Stellar Industries Corp. (Leominster, MA). TLC Plate and Holding Frame. The TLC plate is cut to 3.3 cm × 12 cm and ∼0.5 cm of the sorbent layer is removed from each of the long sides using an Analtech (Newark, DE) scraper. Each of these sides is then coated with the silicone rubber sealant, which is allowed to set overnight. An alternative procedure is to wash the plates before the scraping and coating with the sealant. This procedure, which results in a lower current during PPEC, was found only at the end of the study and was used only for the separation shown in Figure 9. The plates are first baked at 150 °C for 20 min and then soaked for 5 min at room temperature in a solution consisting of 70 parts ethanol and 30 parts 1.0 M hydrochloric acid. The plates are then soaked in deionized water for 5 min and dried at 100 °C for 2 min. The plate is conditioned in an oven at 150 °C for 20 min and allowed to cool in a desiccator. Sample spots are applied 4 cm from the bottom of the plate (i.e., 3 cm from the bottom of the pressurized region) by means of a Camag Nanomat II fitted with a Nano Applicator (Wilmington, NC). After the spots have dried, the plate is briefly dipped in the run buffer, and the excess liquid is allowed to drain along the long axis of the plate for ∼5 s. If the plate is not positioned correctly during the draining, the spots will deviate from the vertical axis during PPEC. The plate is then inserted in the plate holder as described below. An alternative procedure is to dip each end of the plate in the run buffer, such that the sorbent layer is wet to within ∼3 mm of the sample spot, which remains in an initially dry section of the plate. This procedure was not used in the current report, but preliminary experiments indicate that it gives satisfactory results. Figure 1 shows an exploded view of the Delrin frame in which the TLC plate is held. The plate is shown facedown, such that the platinum electrode will contact the sorbent layer at the top end of the TLC plate. A filter paper wick (shown in contact with the electrode) prevents liquid from accumulating at the top of the plate. A 0.25-mm-thick sheet of Teflon (shown as a rectangle with a lip) covers the sorbent layer. The lip covers a 1-cm section of the TLC plate that extends from the bottom of the frame into a pool of the run buffer. A rubber strip (not shown) in the frame presses the electrode (the cathode) against the sorbent layer. The cathode is not under high pressure but is in proximity to the pressurized region of the plate. The assembled frame is placed between the two die blocks shown in Figure 2. Pressure is applied to the layer by the two die blocks using the apparatus described below. PPEC is commenced as soon as the apparatus is pressurized. The run buffer is contained in a modified pipet that has been bent into an “L” shape, with a slot cut in the bulb (see Figure 2). A wire anode is threaded through a hole in the side of the pipet and is in contact with the run buffer. After PPEC, the plates are allowed to dry and are then viewed in a light box, at λ ) 254 nm, and later scanned at the same wavelength with a Shimadzu (Kyoto, Japan) CS9000U dualwavelength flying-spot scanner. 1692 Analytical Chemistry, Vol. 76, No. 6, March 15, 2004
Figure 1. Exploded view of the Delrin frame and components. From bottom to top the diagram shows a section of the Delrin frame, the Teflon sheet, the filter paper wick, the electrode, the TLC plate facedown, and the second section of the Delrin frame.
Apparatus for Applying Pressure. Figure 3 is a photograph of the apparatus used for pressurizing the TLC layer. It is housed in a Plexiglas cabinet that is fitted with two safety switches. When the door is opened, one switch disables the power supply and the second activates a circuit that allows any residual charge to drain to earth. The pressure is supplied by an incompressible fluid, and thus, there is no hazard from this source if the apparatus fails. The hand-operated hydraulic press is in front of the apparatus and is connected with flexible tubing to a ram that presses against one of the die blocks discussed in the above paragraph. The base of the apparatus is a 35 in. × 8 in. × 1.5 in. block of solid aluminum. Cast iron angle brackets at each end of the base plate anchor the TLC plate against the pressure of the hydraulic ram. The four support rods between the angle brackets were retroadded to stabilize the apparatus, after it was found that the base would slightly bow when high pressure was applied. The die block assembly rests on four alignment rods. This assembly is shown in Figure 2 (only three rods are visible in the
Figure 2. Die blocks and alignment rods. Only three of the four alignment rods are visible, due to the camera’s orientation. The mobile-phase reservoir and alligator clip holding anode connection are also shown.
Figure 4. Migration distance versus time for the following compounds listed in order of increasing Rf: 17R-acetoxyprogesterone, 2′-acetonaphthone, benzanilide, and o-nitroaniline. PPEC is on a LiChrospher C18 plate at 118 atm and 6 kV. The mobile phase is 55% aqueous acetonitrile containing 50 mM acetate buffer at a nominal pH of 4.7.
Figure 3. Apparatus for PPEC.
photograph due to the camera’s orientation). The face of the movable section is fitted with a 1-mm-thick sheet of aluminum nitride ceramic, which allows the die block to act as a heat sink while pressurizing the layer. Pressure is applied to an area of 2.5 cm × 10 cm of the surface of the TLC plate when the hydraulic ram pushes the die block assembly against the Teflon sheet covering the TLC layer. RESULTS AND DISCUSSION The velocity of electroosmotic flow was found to be constant under moderate conditions. This was determined by performing PPEC on the following mixture of compounds for different lengths of time: 4-cholesten-3-one, 17R-acetoxyprogesterone, 2′-acetonaphthone, benzanilide, and o-nitroaniline. The results are shown in Figure 4. The figure includes only four compounds, because 4-cholesten-3-one does not migrate under the experimental conditions. The good linearity of these plots indicates that the temperature remains constant during a separation and that the die block acts as an efficient heat sink. A similar set of experiments was performed in which the run time was kept constant at 5 min, the pressure maintained at 59 atm, and the voltage varied from 3 to 7 kV. This demonstrated that there is a linear relationship between the distance migrated (or migration velocity because all runs were for the same time) and applied voltage. This is what is expected from theory.11 For this set of experiments, the buffer concentration was reduced to 15 mM to prevent excessive Joule heating. The mobile phase was 55% aqueous acetonitrile.
Figure 5. Migration distance versus oven temperature for LiChrospher plates conditioned for 20 min. Separation is for 3 min at 59 atm and 6 kV. The concentration of acetate buffer was 25 mM. Apart from this, the mobile phase and solute mixture was the same as used for Figure 4.
Both regular and high-performance (LiChrospher) bonded C18 TLC plates were used. Data from the manufacturer indicate that the sorbent layer of the regular plates consists of irregularly shaped particles with a size distribution of ∼5-20 µm, an average size of ∼11 µm, and a carbon load of ∼14%. The layer of the LiChrospher plates consists of spherical particles with a size distribution of ∼5-11 µm, an average particle size of ∼7 µm, and a carbon load of ∼6%. The Merck reversed-phase plates require activation in an oven before use, and the temperature at which the plates are conditioned is an important experimental variable for PPEC. For the LiChrospher plates, there is a surprisingly strong linear relationship between solute migration distance and oven temperature, as shown in Figure 5. A similar study for the regular plates, over a larger temperature range (100-180 °C), yielded plots that were slightly concave. Migration velocity increases also with the length of time that the plate is conditioned, and this increase is inversely related to solute Rf. As an example, when the regular plates are conditioned at 200 °C, an increase in heating time from 3 to 7 min results in an increase in the migration distances in the range 109-158% when PPEC is performed at 7 kV and 59 atm for 10 min, with a mobile phase that is 55% aqueous acetonitrile containing 25 mM acetate buffer. Analytical Chemistry, Vol. 76, No. 6, March 15, 2004
1693
Figure 8. A 1-min separation on a LiChrospher C18 plate at 9 kV and a pressure of (a) 11.8, (b) 19.7, or (c) 118 atm. The same mobile phase and solute mixture was used as for Figure 7. Figure 6. Separation of 4-cholesten-3-one, 17R-acetoxyprogesterone, 2′-acetonaphthone, benzanilide, and o-nitroaniline. by conventional TLC performed in a horizontal DS chamber using 55% aqueous acetonitrile on (a) a LiChrospher C18 plate and (b) a regular Merck C18 plate. The separation time for the two runs was 24 and 23 min, respectively.
Figure 7. PPEC for 3 min at 118 atm and 10 kV on (a) a LiChrospher C18 plate and (b) a regular Merck C18 plate. The mobile phase is 55% aqueous acetonitrile containing 25 mM acetate buffer at a nominal pH of 4.7. The solute mixture is the same as that used for Figure 6.
Figure 6 illustrates that both the regular and the LiChrospher plates yield similar results when separating the test mixture of solutes by conventional TLC. This is not the case for PPEC, where the LiChrospher plates yield dramatically better results than the regular plates, as shown in Figure 7. The faster performance on the LiChrospher plates is interpreted as being due to the lower carbon load of the particles in the sorbent layer. This should expose more of the silica surface and a larger concentration of silanol groups to the mobile phase. At a given pH this would result in a higher number of ionized silianols in the layer and an increase in the electroosmotic flow. The excellent peak shape may be due to the fact that the layer consists of spherical particles of a narrow size distribution. PPEC is a substantially faster technique than conventional TLC, as is illustrated by the separation of the five test solutes in 1 min, shown in Figure 8 at three different pressures. The separation was on LiChrospher plates, at 11.8, 19.7, and 118 atm, respectively. At 19.7 atm, the separation is of substantially better quality and is ∼24 times faster than the equivalent separation by conventional TLC, shown in Figure 6a. The quality of the sorbent bed is expected to increase with increasing applied pressure, due to a diminution of the interstitial spaces within the bed. This may 1694 Analytical Chemistry, Vol. 76, No. 6, March 15, 2004
possibly explain the difference in separation quality between the 11.8 and 19.7 atm runs. There is a diminution in migration velocity as pressure is increased, and at 118 atm, the migration velocity is slower, and the separation quality worse, than that obtained at the two lower pressures. The diminution in quality is due to inadequate separation time. A substantially better chromatogram (Figure 7a) is obtained in 3 min under identical conditions, apart from the voltage being 10 kV rather than 9 kV. The diminution in migration velocity at the highest pressure is most probably due to a decrease in the size of the capillary channels between the sorbent particles in the TLC layer. Wan12 suggested that, for column electrochromatography, such a diminution in velocity occurs when the capillary channels are sufficiently small to cause overlap of the electrical double layers on opposite channel walls. The size of the double layer shrinks with an increase in buffer concentration, and it is expected that this overlap can be prevented, and the migration velocity increased, by working with a more concentrated buffer solution. This is not practical in the current apparatus, which lacks external cooling. At high voltage, an increase in buffer concentration will result in a substantial increase in the temperature of the sorbent layer due to Joule heating. An added advantage of working at high pressure is that there is a diminution of current and, hence, the amount of Joule heating, as the pressure is increased while holding all other variables constant. This is illustrated in the following example. PPEC was performed for 15 min at 5 kV on regular C18 plates with the same mobile phase as used for the separations in Figure 7. At 19.7 atm, the current drops from about 3.5 to 2.0 mA, whereas at 118 atm the current drops from about 2.2 to 1.3 mA. Separation of Multiple Samples. PPEC is well suited to the simultaneous separation of multiple samples. In addition to spotting the samples along a line parallel to the mobile-phase origin, the samples can also be spotted along lines parallel to the direction of mobile-phase flow. This is possible because the plate is prewetted before PPEC, and the separation of all samples commences simultaneously. The approach is illustrated in Figure 9, which shows the separation of nine samples of the fivecomponent test mixture in 1 min. The conditions should not be compared to those used for other separations in this report, because prewashed plates were used. Reproducibility. The reproducibility of the method was tested by performing six replicate analyses on the test mixture of (12) Wan, Q.-H. J. Chromatogr., A 1997, 782, 181-189.
Table 1. Reproducibility of Migration Distance and Theoretical Plate Height in PPECa regular
LiChrospher plate
migrn dist, theor plate migrn dist, theor plate mm ht,b µm mm ht,b µm solute 17R-acetoxy progesterone 2′-acetonaphthone benzanilide o-nitroaniline
av
SD
av
SD
10.39 0.95
55
19.24 1.45 30.97 1.88 42.36 1.71
43 36 27
av
SD
av
SD
10
14.38 0.48
29
4
7 4 5
25.35 1.16 33.38 1.24 41.75 1.12
24 22 21
3 3 2
a PPEC was for 10 min at 7000 V and 59 atm for the regular plates, and for 4 min at 6000 V and 59 atm for the LiChrospher plates. The mobile phase was 55% aqueous acetonitrile containing 25 mM acetate buffer at a nominal pH of 4.7. b The number of theoretical plates, n, were calculated using the formula, n ) 5.54(MD/w1/2)2, where MD is the solute migration distance and w1/2 is the peak width at half-height. The theoretical plate height was obtained by dividing MD by n.
Figure 9. Separation of nine samples by PPEC in 1 min on a LiChrospher C18 plate at 59 atm and 7 kV. The same mobile phase and solute mixture was used as for Figure 7.
compounds, on both the regular and the LiChrospher plates. The metal die block assembly warms very slightly during a run under the conditions used, and for this reason it was placed in a water bath at room temperature (typically at 21 °C) for 5 min and then dried before commencing each separation. The results are listed in Table 1. The LiChrospher plates yield better reproducibility as measured by a somewhat lower standard deviation in retention (except for the solute of lowest migration) and a considerably lower standard deviation in theoretical plate height. Theoretical plate height decreases with increasing migration distance and varies from 55 to 27 µm for the regular plates and from 29 to 21 µm for the LiChrospher plates. Better temperature control of the sorbent layer should improve the reproducibility of the technique. This can be achieved by incorporating channels in the metal die block for circulating a cooling liquid. This will enable the layer to be maintained at a constant temperature, which in turn will allow the use of higher voltages and higher buffer concentrations. The circulating liquid can also be used to maintain the layer at subambient or elevated temperature. There are other design features that will improve the reproducibility of the technique. The hydraulic pump should be replaced, as there is a 3-10% diminution of pressure during a separation. A device other than an alligator clip should be used to attach the power source to the separation system. The cathode
should be in a fixed position relative to the pressurized layer. In the present design, the exact placement of the cathode depends on the operator. Another method of improving the reproducibility will be to use an oven with forced air circulation and excellent temperature control for conditioning the plates. The oven used in the study was not in this category. TLC plates need to be conditioned at an identical temperature in order to obtain reproducible results (see Figure 5). Reproducibility will also be improved by using an automatic device for prewetting the sorbent layer. Some Conclusions. This report demonstrates that PPEC is substantially faster and more efficient than conventional TLC. PPEC retains all the attractive features of conventional TLC, with the additional feature that a larger number of samples can be separated in a single run. The apparatus described is a prototype and has several shortcomings that are discussed above. An improved version of the apparatus is currently being designed. ACKNOWLEDGMENT This research was supported by a grant from Pfizer, Inc. Dr. Heinz-Emil Hauck is thanked for suggesting the use of the LiChrospher plates and for providing technical information on the two types of TLC plates used in the study. Merck KGaA is thanked for the gift of the LiChrospher plates, and EMD Chemicals, Inc. is thanked for the gift of the regular plates. Dr. Emil Mincsovics is thanked for the gift of the silicone rubber that was used for sealing the edges of the TLC plates. Stellar Industries Corp. is thanked for the gift of the aluminum nitride ceramic. Steve White of CAD & Graphics Design is thanked for preparing Figure 1. Nabilah Rontu is thanked for valuable technical assistance. Received for review September 22, 2003. Accepted December 30, 2003. AC0303362
Analytical Chemistry, Vol. 76, No. 6, March 15, 2004
1695