Two-Electrode Configuration for Separations Based on Electrosorption in Electrochemically Modulated Liquid Chromatography (EMLC)

Abstract

This invention discloses a design a capillary column for use with electrochemically modulated liquid chromatography (EMLC). The capillary design, which results in a marked reduction in the flow of current through the column, enables the use of a two-electrode column construction that overcomes the mechanical and electrical shortfalls of the conventional standard bore design.

Claims

1. An electrochemically modulated liquid chromatography (EMLC) column for the separation of a sample, the EMLC column comprising: a conductive stationary phase packed in a capillary fluidic channel which also functions as a working electrode in an electrochemical cell, wherein the working electrode has a low level of background current due to the relatively small surface area of the capillary fluidic channel and packed stationary phase; and a dual-purpose reference or pseudo reference electrode which provides a stable reference potential that is not appreciably perturbed by the background current of the working electrode, thereby eliminating the need to a current-carrying counter electrode.

2. The EMLC column of claim 1, wherein the capillary fluidic channel has an inner diameter of less than 250 m.

3. The EMLC column of claim 1, wherein the working electrode has a background current of less than 1000 nA.

4. The EMLC column of claim 1, wherein fixed values of potentials or time-dependent potential waveforms are applied to the working electrode to manipulate the separation of the sample.

5. The EMLC column of claim 1, wherein fixed values of currents or time-dependent current waveforms are applied to the working electrode to manipulate the separation of the sample.

6. The EMLC column of claim 1, wherein the separation of the sample is manipulated by changes in the solvent and/or electrolyte composition and concentrations of the mobile phase.

7. The EMLC column of claim 1, wherein the separation of the sample is manipulated by inducing an electrochemical transformation in one or more components in the sample.

8. The EMLC column of claim 1, wherein the stationary phase consists of a conductive material which includes but is not limited to reticulated vitreous carbon, boron-doped diamond, glassy carbon, porous graphitic carbon (PGC), stainless steels and other metals, semiconductors, and conductive and redox-active polymeric and inorganic materials.

9. The EMLC column of claim 1, wherein the cross-section of the capillary fluidic channel is circular, square, rectangular, or triangular shaped.

10. The EMLC column of claim 1, wherein the length of the capillary fluidic channel is from hundreds of microns to hundreds of centimeters.

Description

BRIEF DESCRIPTION OF THE FIGURES

[0007] The accompanying figures, when linked with the detailed descriptions that follow, serve to illustrate various embodiments of the invention, which aid in framing the operational principles and associated advantages of the invention.

[0008] FIG. 1 is a schematic illustration of a standard bore EMLC column;

[0009] FIG. 2 is a configurational diagram for capillary EMLC system;

[0010] FIG. 3 presents the chemical structures of the aromatic sulfonates use to access the improvements in chromatographic performance;

[0011] FIG. 4A shows a chromatogram collected on a 250 m ID EMLC column at open circuit potential (50 mV vs. Ag|AgCl pseudo-reference electrode) for the sample mixture prepared with the compounds listed FIG. 3;

[0012] FIG. 4B shows a zoomed-in profile for the elution band for the early-eluting analyte BS from FIG. 4A;

[0013] FIG. 4C shows a zoomed-in profile for the elution band for the moderately-eluting analyte NTS-2 from FIG. 4A;

[0014] FIG. 4D shows a zoomed-in profile for the elution band for the late-eluting analyte CBS from FIG. 4A;

[0015] FIG. 5 shows chromatograms collected on a 250 m ID EMLC column at different E.sub.app (vs. Ag|AgCl pseudo-reference electrode);

[0016] FIG. 6 plots of log(k) vs. E.sub.app for separations of [A] the monovalent (1) analytes: CBS, PTS, and BS; [B] the trivalent (3) analytes: NTS-1, NTS-2, and N*; and [C] analytes with different coulombic charges: NTS-2 (3), NDS (2), CBS (1) and AHBS; and

[0017] FIG. 7 shows a cyclic voltammogram for AHBS (235 M) in a degassed EMLC mobile phase (0.1 M LiClO.sub.4, 0.15 M LiCl, 4% acetonitrile.

DETAILED DESCRIPTION

[0018] By way of context, the embodiments of the present invention are described within the framework of analytical- and preparative-scale separations by electrochemically modulated liquid chromatography (EMLC) It should, however, be readily recognized by practitioners skilled in the art that these embodiments apply well beyond this illustrative example to include the use of this invention across all areas of investigative and applied measurement science and technology.

[0019] Note that relational terms such as first and second, top and bottom, and the like may be used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying an actual relationship or order between such entities or actions. The terms comprises, comprising, or any variations thereof, are intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus that consists of a number of different and/or related elements is not limited to only those elements but may include other elements not expressly listed or inherent to such a process, method, article, or apparatus. An element preceded by comprises does not, without more constraints, preclude the existence of a number of additional identical elements in the process, method, article, or apparatus that comprises the element.

[0020] Electrochemically modulated liquid chromatography (EMLC) is a separation technique that merges liquid chromatography (LC) and electrochemistry. In EMLC, an electrically conductive stationary phase like porous graphitic carbon (PGC) also serves as the working electrode in a three-electrode electrochemical cell. This design enables the manipulation of the interfacial properties of the stationary phase (e.g., the structure of the electrical double layer) by the application of fixed electrical potentials (E.sub.app) or potential waveforms by means of an external potentiostat. These changes can alter analyte retention and, in some cases, the order of analyte elution. EMLC, therefore, differs from all other known forms of liquid chromatography (e.g., reversed-phase LC) in that the stationary phase may be altered prior to or during a separation via an externally controlling the value of E.sub.app.

[0021] Despite the potential utility of EMLC as a separation technique, there are two challenges that have limited its widespread adoption. Both are related to configuring an LC column to, at the same time, function as an electrochemical cell: (1) the availability of materials suitable as conductive stationary phases and (2) the engineering design and operation of the column itself. Stationary phases for EMLC must have the same characteristics shown to be important to conventional LC (i.e., homogeneous in size and surface chemistry, chemical and mechanical stability, and a high surface area). These stationary phases must also be electrically conductive. Carbonaceous materials (e.g., reticulated vitreous carbon, boron-doped diamond, glassy carbon, and PGC) and a number of other conductive materials meet these requirements, noting that the majority of the work to date on EMLC has used PGC as the stationary phase material.

[0022] In addition to the characteristics of the stationary phase, there are several challenges inherent to the functional operation of an EMLC column. These challenges reflect the need to effectively incorporate a number of electrochemical components (e.g., reference and counter electrodes, insulating membranes and fittings, and electrical contacts) into the design of an EMLC in a manner that does not diminish chromatographic performance.

[0023] FIG. 1 summarizes the common features of earlier EMLC column designs, all of which followed the layout of a conventional three-electrode (i.e., working, reference, and counter electrodes) electrochemical cell. In this example, a PGC stationary phase (101) is carefully packed inside a porous stainless-steel tube (102), which acts as a column housing and as a counter electrode, and a silver-silver chloride (Ag|AgCl) reference electrode (103) is placed in an external electrolyte bath (104) that contains an electrolyte solution in contact with the porous stainless-steel counter electrode. The counter electrode and PGC are separated by a perfluoro sulfonated ion-exchange membrane (105). The ion exchange tubing tightly conforms to the inner wall of the stainless-steel tube, acting to electrically isolate the working electrode from the counter electrode while also facilitating ion transport between electrodes. A stainless-steel retaining frit (106) makes conformal contact with both the PGC and a metallic end-fitting (107). The metallic end-fitting serves as the electrical connection to the working electrode and a connector at the liquid inlet (108) to the tubing from the chromatographic pumping system. The insulating polyether ether ketone union (PEEK) (109) electrically separates the stationary phase from the counter electrode (102). The stationary phase is further separated from the counter electrode just below the column inlet (110) by an insulating polymer frit (111), which is connected to the column by another metallic end-fitting (112). A section of tubing comprised of an ion-exchange polymer (105) serves to separate the stationary phase and column housing along the length of the column, thereby preventing electrical short-circuit.

[0024] While the design of the EMLC column has proven its utility in the past, the ion exchange tubing can deform under the pressures applied to drive fluid flow through the column, which can compromise run-to-run repeatability and broaden the profiles of the elution bands. The ion exchange tubing can also rupture to cause an electrical short circuit between the PGC and counter electrode and, consequently, column failure.

[0025] This invention discloses a column design, exemplified in FIG. 2, which is aimed at improving the instrumental capabilities of EMLC by the construction of a miniaturized, two-electrode capillary system. In this example, the inner diameter (ID) of the column from 3.3 mm, an ID reflective of standard-bore HPLC columns, to 250 m ID by using fused silica tubing as the column housing and a conductive stationary phase for chromatographic separations. Note that the conductive stationary phase is packed directly into the silica capillary column and acts as both the working electrode in a two-electrode electrochemical cell and as the stationary phase for chromatographic separations. Due, in part, to the lower surface area of the stationary phase, the reference electrode can carry the current that flows in the cell and maintain a stable (i.e., usable) reference potential. As a result, the counter electrode and ion exchange membrane are eliminated from the column design.

[0026] It is well known that capillary columns can maintain the efficiency of their larger-bore counterparts and provide additional benefits such as increased sensitivity, reduced amounts of stationary phase needed to pack columns, significantly decreased solvent consumption, and lower required sample volumes. There are two more uniquely distinct benefits to the adaptation of EMLC to capillary columns: (1) the design and placement of the individual electrodes are simplified, and (2) the ion exchange tubing is eliminated from the column design. Collectively, the design and chromatographic performance benefits of the new capillary EMLC hardware developed herein may encourage the more widespread study and use of this technique.

[0027] The design of the new column is exemplified in FIG. 2. It is constructed by cutting a polyimide-coated fused silica tubing (201) packed with a conductive stationary phase like PGC [360 m outer diameter (OD), with either a 100 or 250 m inner diameter (ID)] to a length of 25-30 cm. A glass fiber/silica composite retaining frit (202) was installed in a separate 5 cm piece of 250 m ID tubing (203). After installation of the frit, both the column and fritted tubing were connected via a stainless-steel union (204) with a 250 m diameter thru-hole. By orienting the two pieces of tubing in this way, a stationary phase such as PGC (205) is then slurry packed into the thru-hole (250 m diameter) of the union, thereby establishing conformal and electrical contact between the PGC packing and stainless-steel union. Note that electrical contact is maintained under the high fluid pressures operative in LC, meaning the union can be used as the electrical contact between the working electrode and potentiostat without increasing column backpressure or degrading chromatographic efficiency.

[0028] The column was connected to a conventional HPLC instrument by way of a 6-port injection valve that used a 90 nL loop (not shown) prepared from a length of 50 m ID polymer tubing and a standard 500 nL flow cell for detection by UV-Vis absorbance measurements. The flow path of liquid from the LC system was modified to accommodate the low rate of fluid flow through the capillary EMLC column by using a pre-column flow splitter (206) connected to a piece of polymer tubing (207) that was used to house the pseudo-reference electrode (208). There are three key elements of this design. It (1) places the pseudo-reference electrode in close proximity to the PGC stationary phase, which minimizes the uncompensated resistance in the electrochemical cell; (2) can withstand pressures up to 400 bar operative in the LC experiments; and (3) prevents the mobile phase from contacting wire leads used to make electrical contact with the pseudo-reference electrode.

[0029] As an example, the flow splitter and pseudo-reference electrode housing can be assembled by soldering one end of a 10 cm long, 0.4 mm diameter chloridized silver wire (Ag|AgCl electrode; 208) to a 50 m diameter tungsten wire (209). The tungsten wire functions as an electrical lead for connection between the reference electrode and potentiostat. The solder junction is then inserted into a 5 cm piece of 1 mm ID polyether ether ketone (PEEK) tubing and sealed with a heat-curable epoxied glue (not shown). The PEEK tubing was connected to one free thread of a PEEK tee-union (210) with the Ag|AgCl wire inserted through the thru-hole of the tee-union. The remaining length of the Ag|AgCl wire (7 cm) that protruded from the tee-union was sheathed with a piece of 0.5 mm ID PEEK tubing. The next step connects the PEEK tubing to the tee-union (i.e., the thread where the Ag|AgCl wire exits the union) on one end and to the pre-column flow splitter union on the other end by means of a 1/16 PEEK LC fittings (211). Finally, the free thread of the tee-union is connected to a 60 cm length of 50 m ID PEEK tubing (212), which places the 50 m ID tubing in line with the Ag|AgCl wire to provide the necessary flow resistance to achieve the desired mobile phase flow rate through the column.

[0030] After completing the column packing procedure, the column and connecting tubing (i.e., the segment containing the frit) were connected to the LC system and flushed with the mobile phase at a back pressure of 120 bar for several hours. The mobile phase was 4% acetonitrile:96% water containing 0.10 M LiClO.sub.4 and 0.15 M LiCl. Backpressures during the operation were typically 140 bar. After applying a potential to the PGC, the background current was allowed to decay to an approximately constant background prior to analyte injection. This process generally required 20-30 min; however, up to 40 min was required for the current to stabilize at more extreme potentials (i.e., +400 or 800 mV E.sub.app) due to the larger magnitude of the background current. Note that application of an electrical potential to the PGC (0.8 to +0.5 V vs Ag|AgCl, satd KCl) in the 3.3 mm standard-bore column generally results in a significant double-layer charging current of a few milliamps or more, which stabilizes after 20-60 min to nonzero current levels of 0.5-10 A. These significant background currents, which arise from faradaic reactions of O.sub.2 or electrolyte contaminants in the mobile phase, necessitate the use of a counter electrode to ensure the stability of the reference electrode potential. In comparison, background currents for the capillary configurations were typically on the order of a few nanoamperes.

[0031] The improvements in the chromatographic performance of the capillary design were demonstrated by using mixtures of the compounds shown in FIG. 3 by means of the chromatographic data presented in FIGS. 4-6. The concentrations were 4 (NDS), 7.5 (AHBS, BDS, BS, CBS, PTS), or 15 (NTS) g/mL, resulting in 0.36-1.35 ng injection masses for each analyte. In these experiments, the eluents were detected via UV-VIS absorbance spectroscopy with a diode array detector. Absorbance spectra were collected for every data point (sampling frequency of 1 Hz), which aided in peak identification based on characteristic absorption bands for each analyte. Unless specified otherwise, all chromatograms were generated by monitoring the absorbance at 220 nm. For peak identification, single analyte injectionsor, in some cases, injections of well-resolved pairs of analyteswere used in conjunction with UV-Vis absorbance spectra to identify the elution bands in the chromatogram for the eight-component mixture and for precise determination of retention times.

[0032] Eight aromatic sulfonates were chosen for column characterization because the retention behavior of these analytes on PGC with respect to E.sub.app has been well characterized..sup.5, 12 Five unsubstituted aromatic sulfonates (i.e., BS, BDS, NDS, NTS-1, and NTS-2) with different coulombic charges (1 to 3) and fused ring structures (benzene and naphthalene) were used to investigate the impact of electrostatic and interactions on retention as a function of E.sub.app. Benzenesulfonate derivatives with electron withdrawing (Cl, NH.sub.3.sup.+) or donating (CH.sub.3, OH) substituents were included in the test mixture to assess other noncovalent interactions (e.g., dispersive and dipole-dipole). Additionally, AHBS exhibits redox behavior on graphitic carbon at low potentials (70 to 0 mV vs Ag|AgCl, satd KCl) and was suspected of displaying unique retention behavior as a function of E.sub.app.

[0033] The performance of the capillary EMLC column was quantified by several different separation parameters, including the retention factor (k), the elution band asymmetry factor (A.sub.s), the number of theoretical plates (N), and the chromatographic resolution (R.sub.s). The retention factor, k, for each an analyte under a set of chromatographic conditions specified shortly, was calculated according to Eqn 1,

[00001]$\begin{array}{cc}{k}^{}=\frac{t_R-t_M}{t_M}& \left(1\right)\end{array}$

where t.sub.R is the unadjusted retention time of the analyte, and t.sub.M is the dead time of the column. The higher the value of k, the stronger the affinity of an analyte for the stationary phase.

[0034] The asymmetry factor, A.sub.s, which is an indicator of elution band tailing, was calculated according to Eqn 2,

[00002]$\begin{array}{cc}{A}_s=\frac{b}{a}& \left(2\right)\end{array}$

where a is the half widths of an analyte's elution band at 10% of its peak height toward the leading edge of the band and b is the half widths of an analyte's elution band at 10% of its peak height toward the trailing edge of the band. Values of A.sub.s greater than 1 indicate peak tailing, and values of A.sub.s less than 1 are diagnostic of fronting.

[0035] The number of theoretical plates, N, was used as a measure of column efficiency and was calculated using Eqn 3,

[00003]$\begin{array}{cc}N=5.5.Math.4.Math.\left(\frac{t_R^2}{W_{0.5}^2}\right)& \left(3\right)\end{array}$

where W.sub.0.5 is the full width of the elution band determined at half of its maximum peak height.

[0036] The efficiency of a separation can also be determined by the chromatographic resolution, R.sub.s, which is given in Eqn 4.

[00004]$\begin{array}{cc}{R}_s=\frac{2.Math.\left(t_{R.Math.2}-t_{R.Math.1}\right)}{w_{b.Math.1}+w_{b.Math.2}}& \left(4\right)\end{array}$

In Eqn 4, the subscripts 1 and 2 represent elution metrics for a pair of solutes that elute neighbor each other whereby the subscript 1 is for the more weakly retained solute, and the subscript 2 is for the more strongly interacting solutes. The parameter t.sub.Ri represents their corresponding unadjusted retention times, and w.sub.bi stands for the widths at the base of each of the elution bands. R.sub.s, therefore, take both the column efficiency and the difference in solute retention into account.

[0037] FIG. 4A presents a chromatogram for the separation of the eight-component mixture on a 1400.25 mm (lengthID) EMLC column at open circuit potential (50 mV vs. Ag|AgCl pseudo-reference electrode), whereas FIGS. 4B-D shows the elution bands for three of the components, BS, NTS-2, and CBS, in the mixture in order to more clearly demonstrate the improvements in performance. This separation was carried out at a flow rate of 10.60.2 L/min. As shown in FIG. 4, six of the eight components in the mixture were well resolved from each other, but AHBS and NTS-1 coeluted at with retention times spanning 3-4 min. An unknown ninth band eluted at 12.65 min. This band was subsequently identified as a naphthalene-based contaminant in the isomeric 1,3,6/7-naphthalenetrisulfonic acid stock and is hereafter abbreviated N*.

[0038] For the chromatogram shown in FIG. 4, the theoretical number of plates, N, for moderately retained analytes (2<k<10) and a dead-time marker (i.e., BDS) were calculated at 13,200-19,500 and 3,400 plates/m, respectively. Values for N up to 39,600 plates/m were achieved at lower, more optimal flow rates (e.g., 0.2 L/min; data not shown as the elution times for such separations span several hours). Importantly, these efficiencies meet or exceed the typical efficiencies (i.e., 2,000-5,000 plates/m) reported for the column design illustrated in FIG. 1, as well as the largest efficiency (14,000 plates/m) reported for the standard-bore configuration to date.

[0039] The chromatograms in FIG. 4 also demonstrate an improved symmetry in elution band profiles. As shown in FIGS. 4B-D, the elution bands for the three different solutes with the capillary EMLC column have little tailing and become increasingly symmetric for the analytes that are strongly retained on the column. Moreover, the asymmetry factor, A.sub.s, for BS, which has a retention time, t.sub.R, of 2.02 min, is 1.4 that is diagnostic of moderate tailing. As noted, however, the symmetry of the elution bands clearly improves as the value of A.sub.s decrease to 1.2 and 1.1 for NTS-2 (t.sub.R=6.33 min) and CBS (t.sub.R=10.71 min), respectively. For comparison, data from the standard bore EMLC columns yields A.sub.s values ranging from 2.1 to 4.4, which are large and effectively unusable levels of tailing.

[0040] The dependence of the retention of the mixture's components as a function of different fixed values of E.sub.app is presented in FIG. 5. The set of chromatograms was obtained with a 1200.25 mm PGC-packed capillary EMLC column at a flow rate of 8.4 (0.2) L/min (backpressure of 141 bar) for five different fixed values of E.sub.app: 400 (501), 100 (502), 0 (503), +200 (504), and +400 (505) mV (vs. Ag|AgCl pseudo-reference electrode). Elution band assignments are abbreviated for clarity according to the following assignments: A.sup.2 (BDS), B.sup. (BS), C (PTS), D.sup.3 (NTS-1), E.sup. (CBS), F.sup. (AHBS), G.sup.2 (NDS), H.sup.3 (NTS-2), I.sup.3 (N*).

[0041] Application of 400 mV to the PGC packing resulted in the elution of eight of the mixture's nine components in 4 min. The last component of the mixture, CBS, eluted with a t.sub.r of 5.65 min and was baseline resolved from all of the other components. At 100 and 0 mV values of E.sub.app close to the open circuit potential of the column (50 mV)there are significant changes in the retention, k, of each analyte; there were also changes in the order of elution. Comparing the elution of the separations performed at 400 and 0 mV, the overall separation time increased from 6.5 to 8.5 min. At 0 mV, five analytes (BDS, BS, N*, NDS, and CBS) are either moderately or effectively resolved (1.0<R.sub.s<1.5) from all of the other components of the mixture, with AHBS and NTS-1 coeluting together and PTS and NTS-2 coeluting together. Relative to 400 mV, the k values for the separations at 0 mV are larger, increase that was also accompanied with a few changes in elution order. NDS, for example, is more strongly retained than PTS at 0 mV. Increasing E.sub.app to more positive values further increases the retention of all of the components in the mixture. At the most positive E.sub.app tested (+400 mV), it takes 33 min for all nine analytes to elute. For comparative purposes, the retention factors for all analytes increase, relative to their respective values at 400 mV, by a factor of at least 2 and up to 38. Furthermore, five analytes (BDS, BS, PTS, N*, and NTS-1) are now baseline separated (R.sub.s>1.5) from all of the other components in the mixture. On the other hand, NDS and NTS-2 are poorly separated (R.sub.s=0.6), and CBS/AHBS coelute (R.sub.s<0.5).

[0042] FIG. 6 summarizes the results of the retention data collected over fixed values of E.sub.app ranging from +400 mV to 800 mV by plots of log(k) versus E.sub.app. As was evident from the chromatograms in FIG. 5, the retention for each of the 9 analytes decreases as the value of E.sub.app moved progressively negative. This dependence can be partially accounted for by electrostatic repulsion between the sulfonate moieties on the solutes and the negatively charged PGC surface. At a more detailed level, perspectives on how changes in E.sub.app alter retention can be developed by examining linear fits to the log(k) vs. E.sub.app plots in FIG. 6, which are summarized in Table 1. Note that BDS is not retainedit elutes with the injection peakat all values of E.sub.app except for +400 mV and is not included in this analysis. In most instances, the retention behavior of the different aromatic sulfonates can be attributed to the magnitude of their electrostatic charge, meaning that the greater the charge on the analyte, the greater the dependence of its retention to changes in E.sub.app. This perspective qualitatively matches the trends in the data listed in Table 1. More specifically, the slopes for the three monovalent anions (BS, PTS, and CBS) fall within the range of 0.42-0.52 V.sup.1, whereas those for NDS (2 charge) and NTS-1/2 (3 charge) are 1.31 and 1.90 V.sup.1, respectively. The slope for N* (1.95) is also consistent with that of a trivalent anion, meaning it is likely to be an isomer of NTS-1/2.

TABLE-US-00001 TABLE 1 Slopes (S) of the log(k) vs. E.sub.app plots for the aromatic sulfonates. S at E.sub.app > S at E.sub.app < 0.2 Analyte.sup.a 0 V (V.sup.1) V (V.sup.1) BS (1) 0.42 0.04 (0.999).sup.c 0.18 0.03 (0.932) AHBS(0) 2.42 0.42 (0.923) 1.05 0.14 (0.965) PTS (1) 0.46 0.09 (0.889) 0.19 0.03 (0.957) CBS (1) 0.52 0.06 (0.954) 0.23 0.02 (0.987) NDS (2) 1.31 0.11 (0.991) 0.46 0.04 (0.986) NTS-1 (3) 1.87 0.15 (0.990) 0.58 0.09 (0.956) NTS-2 (3) 1.90 0.15 (0.993) 0.58 0.06 (0.978) N*(3).sup.b 1.95 0.13 (0.992) 0.60 0.07 (0.976) .sup.aBDS eluted with the solvent front at most values of E.sub.app and was thus omitted from the Table. .sup.bThe coulombic charge for N* was inferred based on the slope of the log(k) vs. E.sub.app plot and the fact that N* is an impurity in the NTS-1/2 stock. .sup.cR.sup.2 values for the linear fits are in parenthesis.

[0043] It should be noted, however, that while coulombic charge can have a strong influence on retention upon changes in E.sub.app, there are a number of other contributing factors that play a role in the interaction of these analytes and the PGC stationary phase. For example, the ratio of the slopes for NDS and BS, 3.120.39, is larger than that predicted based solely on the basis of electrical charge (i.e., 2). The same observation can be made for the trivalent analytes (e.g., S.sub.NTS-2:S.sub.BS=4.520.56). This results indicate that interactions between the -system of the naphthalene ring and positively charged PGC surface also have an impact on retention, which is consistent with the high affinity of PGC for polarizable functional groups (i.e., phenyl and fused aromatic ring systems). In other words, the stronger dependence of retention at values above the potential of zero charge (PGC) for PGC (150 mV) reflects the attractive contributions of the electrostatics and - interactions of this system. On the other hand, the weaker dependencies of retention at values of E.sub.app below PGC's PZC are qualitatively consistent with the tradeoffs between the electrostatic repulsion of the negative charges analytes and the buildup of negative charge on the PGC surface and the attractive interactions of the two - systems.

[0044] There is one more point to draw out regarding the dependencies of analyte retention in FIG. 6D that warrants further discussion. FIG. 6D highlights the shift in retention behavior between 200 and 0 mV for NDS with the linear fits to the plots at two different potential ranges superimposed as dotted lines. AHBS, which being a zwitterion ion that is neutrally charged under the conditions used for these separations, has two dependencies not accounted for the above mechanistic descriptions. First, the retention dependence of AHBS to changes in E.sub.app was stronger than those found for any of the analytes tested (S=2.42 V.sup.1 at E.sub.app>0 mV). Second, there was a slightly more abrupt sudden change in the slope of the log(k) vs. E.sub.app plot for AHBS at between 0 and +100 mV not observed in FIG. 6C for any of the other analytes. As a test for the possible electrochemical transformation of the analytes, the measured absorbance for each solute upon elution was compared to that to standards analyzed by spectrophotometry.

[0045] Importantly, the neutral (zwitterionic) AHBS exhibited two features not accounted for in the above mechanistic discussion. First, AHBS was more responsive to E.sub.app than any of the analytes tested (S=2.42 V.sup.1 at E.sub.app>0 mV). Second, there was a sudden change in the slope of the log(k) vs. E.sub.app plot for AHBS at between 0 and +100 mV that was not observed for any of the other analytes (FIG. 6C). Because aromatic amines are known to undergo electrooxidation reactions at carbon electrodes, the possibility that such a process may account for the unique retention behavior of AHBS, a series of cyclic voltammetry experiments were carried out. These experiments using a glassy carbon electrode as a surrogate for the PGC stationary phase with AHBS dissolved in the EMLC mobile phase. The results of these experiments are presented in FIG. 7 [reference electrode: Ag|AgCl (sat'd KCl), counter electrode: Pt coil, scan rate: 100 mV/s]. Note that the arrow depicts the potential scan direction and that anodic currents are plotted as positively signed values.

[0046] The anodic trace in FIG. 7 (scan from 600 to +400 mV) has two waves with peak-current potentials (E.sub.p) at 40 and 0 mV. The reverse, cathodic trace, however, exhibits a single wave at E.sub.p=68 mV. Correcting for the pseudo-reference electrode potential (+89 mV) relative to Ag|AgCl (sat'd KCl), the potential over which AHBS undergoes faradaic reactions (E.sub.P+20 to +90 mV) overlaps the region in FIG. 6C where AHBS retention is least sensitive to changes in E.sub.app. These data indicate that AHBS, at these potentials, coexists in at least two interchanging chemical forms as it elutes from the EMLC column, which mitigates the responsiveness of k to E.sub.app. These observations underscore another attribute of EMLC as a separation tool: the use of electrochemical transformations as a means to further tune analyte retention.

[0047] In the foregoing specifications, specific embodiments of the present invention have been described. However, various modifications and changes, such as the chromatographic detector, the absolute position of the reference electrode (i.e., in a flow splitter or in-line with the stationary phase), or the presentation of the column itself (e.g., silica capillary, polymer tubing, channel fabricated in a glass substrate, etc.) can be made without departing from the scope of the present invention as set forth in the claims below. Accordingly, the specifications and figures are to be regarded in an illustrative rather than a restrictive sense, and all such modifications are intended to be included within the scope of the present invention. The benefits, advantages, solutions to problems, and any element(s) that may cause any benefit, advantage, or solution to occur or become more pronounced are not to be construed as a critical, required, or essential features or elements of any or all the claims. The invention is defined solely by the appended claims, including any amendments made during the pendency of this application and all equivalents of those claims as issued.