Polymethylmethacrylate Open Tubular Ion Exchange Columns

Jul 22, 2013 - ... The University of Texas at Arlington, Arlington, Texas 76019-0065, United States. ‡School of Pharmacy, East China University of S...
0 downloads 0 Views 739KB Size
Article pubs.acs.org/ac

Polymethylmethacrylate Open Tubular Ion Exchange Columns: Nondestructive Measurement of Very Small Ion Exchange Capacities Min Zhang,† Bingcheng Yang,†,‡ and Purnendu K. Dasgupta*,† †

Department of Chemistry and Biochemistry, The University of Texas at Arlington, Arlington, Texas 76019-0065, United States School of Pharmacy, East China University of Science and Technology, Shanghai 200237, China



S Supporting Information *

ABSTRACT: We describe an approach to prepare an open tubular ion exchange (OTIE) column by coating a monolayer of anion exchange nanoparticle to a 16−20 μm bore polymethylmethacrylate (PMMA) capillary. The latex nanoparticle was electrostatically attached to carboxylate groups on the inner wall of capillary, pretreated with strong base for hydrolyzing the ester. Several approaches to nondestructively measure ion exchange capacities (IEC) of the columns were examined: (a) adsorption−desorption of an intensely fluorescent ion, e.g. fluorescein, and off-line fluorometry, (b) loading a weakly retained ion (e.g., IO3−), frontal displacement by a strongly bound ion (e.g., Cl−), and online optical or conductometric boundary detection, and (c) similar to the above except displacement being accompanied by reaction (e.g., acid−base titration). To our knowledge, this is the first time on-column titration has been used to measure capacities. By using different pH displacer solutions, we demonstrate for the first time the possibility of pKa-differentiated ion exchange capacity measurements. The cation exchange capacity of bare PMMA capillaries was on the order of 1 pequiv/mm2 with little dependence on time and temperature of hydrolysis conditions. After AS18 latex coating, the strong base anion exchange capacity was on the order of 10 pequiv/mm2, very close to what would be estimated on the basis of monolayer coverage of the surface by individual latex particles. The latex used contained a significant, additional amount of weak base character, about the same as the strong base ion exchange capacity.

M

commercially introduced, predates packed capillary IC. Ishii and Takeuchi6 made OTIE columns bearing sulfonate groups by reacting phenyltriethoxysilane to 30−60 μm glass capillaries followed by sulfonation or by reacting mercaptoethyltriethoxysilane followed by oxidation. Between 1983 and 1991, Simon et al. used open tubes of 2.3−25 μm in inner diameter and demonstrated ion exchange separations using bare silica or silica capillaries variously coated with poly(butadienesulfonic acid), poly(butadienemaleic acid), or reacted the inner wall with sulfopropylhydroxysilane and aminoalkylalkoxysilane.7−9 These approaches have not proven very practical. A dendrimerlike layered polymer structure, containing >20 successive layers, has also been demonstrated for OTIE columns.10,11 As column diameters get smaller, however, procedures using polymerization reactions to create ion exchange sites become increasingly difficult. The probability that the capillary will get blocked increases exponentially. Mere surface functionalization, e.g., by reacting with a monomeric silane, may not block the capillary but typically does not provide sufficient capacity. Electrostatic binding of an oppositely charged subμm latex particle provides an attractive solution. A silica capillary is

any, if not most, materials have some ion exchange properties on the exposed surface. Organic polymers are often surface oxidized to form −OH or −COOH groups that display cation exchange properties. Lucy has persuasively argued that desalination of water through natural, biologically based ion exchangers has been explicitly described at least since the time of Moses.1 Ion exchange groups are deliberately introduced in many processes of analytical significance, ion chromatography (IC) being a prime example. Since its introduction in 1975, IC has emerged as the dominant technique for ion analysis.2 Although pellicular packings, including ion exchanger stationary phases, had been introduced even before,3 the commercial appearance of IC also introduced electrostatically agglomerated packings in which an oppositely charged latex nanomaterial is electrostatically attached to a charged surface. This technique is still in use today for making particular IC stationary phases. Functionally, they share the same advantage of better mass transfer kinetics as the presently popular superficially porous HPLC packings. As a trend, miniaturization has affected all technologies; IC is no exception. Suppressed conductometric IC using a 190 μm packed capillary was demonstrated in 1983;4 a portable capillary IC was introduced in 1998.5 A commercial version has appeared only recently. Interestingly, open tubular ion exchange (OTIE) chromatography, which has not yet been © 2013 American Chemical Society

Received: June 20, 2013 Accepted: July 22, 2013 Published: July 22, 2013 7994

dx.doi.org/10.1021/ac4018583 | Anal. Chem. 2013, 85, 7994−8000

Analytical Chemistry

Article

greater concern is that the reported ion exchange capacities varied by large amounts depending on the choice of the probe ion: For the 25 μm diameter tube, for example, the observed capacity increased by an order of magnitude from iodate to thiocyanate as the probe ion. Using the same basic technique, by using a neutral salt at a neutral pH, cannot in principle result in such great differences in the measured ion exchange capacity. Herein, we demonstrate three complementary approaches to measure capacities of OTIE columns as detailed below.

negatively charged at all but very acidic pH, and positively charged latex easily binds to such a surface and provides much greater capacity than a silane monolayer; this strategy has been well-used.12−14 Much as the advent of silica optical fibers made available silica capillaries (virtually all capillary chromatography columns are still exclusively based on silica), the technology behind precise extrusion of polymeric optical fibers have now made available polymeric capillaries in a variety of materials. Unlike silica, they are not intrinsically fragile and do not need protective coatings for use. They can be oxidized/hydrolyzed/ sulfonated to create a negatively charged surface or otherwise functionalized, generally by chemistries no more difficult than those used with silica. When optical transparency is not an issue (e.g., in conductometric measurements), there are advantages beside cost and material diversity. The consequence of deliberately deformed channels on mass transfer has been considered for decades,15 but experimental studies with capillaries that only permit large bend radii have not been possible. It is well-known that a straight tube is the worst possible geometry for inducing axial dispersion.16 Considering that many different types of positively charged latex particles have already been made for use in commercial IC stationary phases, latex attachment becomes a particularly useful technique for fabricating anion exchange columns from polymeric capillaries with carboxylate surfaces. Latex attachment dramatically increases ion exchange capacity because a typical ion exchanged functionalized latex provides many more exposed ion exchange sites on its exterior compared to the number of oppositely charged wall surface sites it takes up. Although the ion exchange capacity is one of the most important characteristics of a chromatographic column, regardless of the chemical architecture, capacities of OTIE columns are hardly, if ever, reported because of the difficulties of measuring such low capacities. A review on capillary IC,17 for example, contains no column capacity data whatsoever. Early on, Ishii and Takeuchi18 sulfonated a phenylsilylated surface-etched soda-lime glass column (5.4 m × 60 μm i.d.) and determined its ion exchange capacity by (a) converting the column to H+-form with acid, (b) washing thoroughly with water, and (c) flushing with 0.1 M NaCl and collecting the effluent. The column was completely converted to the Na-form in step c. The liberated HCl in the effluent was titrated with 1 mM NaOH, requiring ∼80 μL for the titration. This approach is not practical for columns much smaller in bore and length, with 10−100× smaller absolute capacities. Although fused silica has been essentially the only material used in more recent OTIE columns, in our experience, the long-term stability of cationic latex on silica capillaries at either extremes of pH is very poor. This is true regardless of the nature of surface preparation of the capillary, short of creating other types of acidic functional groups. The only other extant capacity data of OTIE columns are due to Hutchinson et al.19 They determined the capacities of latex coated monoliths and a few OTIE fused silica columns by (a) loading the column with a very dilute (10 μM) solution of an UV-absorbing anion until breakthrough was indicated by an on-column UV detector, (b) washing thoroughly with water, and (c) flushing with 100 mM ClO4− and integrating the peak produced by the liberated UV absorbing anion. Such an approach is workable but often produces a somewhat distorted, difficult to integrate peak. The significant change in refractive index (RI) coincident with the peak elution seriously affects the measured absorbance, causing an unaccounted for error. A



PRINCIPLES Stated ion exchange capacities depend on the measurement conditions. A large ion may not be able to access all the sites a small ion can. The capacity determined for an anion exchanger that contains both weak and strongly basic sites will depend on the experimental pH. Rather than a single measured value being “correct”, the range of ion exchange capacities measured by different means should be regarded as the spectrum of capacities that the exchanger can exhibit. The examples in this paper mainly pertain to an anion exchanger, but the principles are equally applicable to cation exchangers. The cation exchange capacity determination method for our hydrolyzed polymethylmethacrylate (PMMA) tubes is not described in detail, but it is the exact analog of the conductometric frontal reaction chromatography procedure by acid−base reaction carried out for the anion exchangers. Fluorescent Ion Displacement (FID). Of the three techniques discussed here, only this one uses off-line measurement. This is similar to the Hutchinson et al. displacement approach.19 Because of the low capacities involved, a fluorometric method, rather than absorbance measurement, is used and the measurement is made offline to avoid RI effects. An intensely fluorescent ion, e.g., a dilute solution of fluorescein, is used to load an anion exchange column, which is then washed thoroughly. A concentrated solution of a nonabsorbing ion like sulfate or perchlorate is then used to elute the fluorescent ion in a small volume. After making up the eluate to a known volume, the fluorescence is measured and compared with a calibration plot of the fluorescent ion in the same matrix. Frontal Displacement Chromatography (FDC). Frontal displacement chromatography is easy to perform in an open tubular capillary provided the flow velocity is slow enough for the solution in the lumen to equilibrate with the wall. Imagine that the column is equilibrated with an optically absorbing and relatively easily displaced ion such as iodate (IO3−) to saturate the sites with iodate. Following washing with water, the dilute solution of a displacing ion such as chloride or perchlorate that has a different absorptivity is pumped through the column. This much more strongly bound ion displaces the iodate immediately in a frontal displacement manner, the concentration of the iodate in the displacement band being equal to the concentration of the displacing ion. With the background being water, the first appearance of the displaced iodate at the UV detector located at the end of the column is registered by an abrupt upward step function. After a period t, as the iodate zone ends and the transparent displacing ion begins to elute, the observed absorbance decreases in a similarly abrupt step function, making it easy to determine the width of the iodate zone. The anion exchange capacity C in equivalents is thus: C = [displacer ion, equiv/cm 3] × Q (cm 3/s) × t (s) (1) 7995

dx.doi.org/10.1021/ac4018583 | Anal. Chem. 2013, 85, 7994−8000

Analytical Chemistry

Article

where Q is the flow rate of the displacer ion. It will be obvious that the counterion present with the displacer ion should not participate in the observation; i.e., preferably, it will be optically transparent, and this requirement is not difficult to meet with alkali metal salts. With iodate being a relatively low mobility ion, it is also possible to observe the progress of the frontal displacement by on-column conductometry, as both Cl− or ClO4− have significantly greater mobility than IO3− and the conductivity difference between the zones is easier to distinguish if a low mobility counterion, e.g., Li+, is chosen. In this case, a two-step staircase would be observed, the first excursion from the baseline corresponding to the elution of LiIO3 and the second further excursion corresponding to the appearance of LiCl/LiClO4. Frontal Reaction Chromatography (FRC). In FRC, displacement is accompanied by a reaction that takes place quantitatively. The most obvious such reaction is an acid−base titration: For example, an anion exchanger may be loaded with hydroxide and washed with water, and then, a strong acid, e.g., HNO3, is passed through it until breakthrough occurs, which can be followed either by UV absorbance or conductance measurement. The holdup volume (column interior volume) is often obvious from the data but, if not, needs to be independently known so that the time needed for the liquid in the capillary to be displaced can be computed. The end of acid consumption by the capillary is indicated by a step change in the detector output. Thus, in this case, the capacity C is computed from

(somewhat longer for larger capillaries) to allow the latex sufficient time to diffuse to the wall and bind to it. The column was then thoroughly washed with water before further testing. We did not dialyze the latex suspension for removal of surfactants, monomers, or other small molecules,19 as preliminary experiments did not indicate that such prepurification had any effect on the final results. Test Procedures. All flow was attained by dipping the capillary into the desired solution (often in a disposable microvial for convenience) that is put inside a small pressurizable Plexiglas reservoir. The container was pressurized with He or N2 gas controlled by an analog input controllable miniature pressure controller (www.parker.com, P/N 990005123-100). Flow rates were initially estimated from the Hagen−Poiseuille equation and later measured with nanoflow meters from www.Idex-hs.com or www.Bronkhorstusa.com (Bronkhorst μ-Flow L01 mass flow meter). The flow rates were further confirmed with gravimetric measurement of the collected effluent. The flow rates ranged from 2 to 100 nL/min. A capacitively coupled contactless conductivity detector (C4D, TraceDec, http://www.istech.at) was used for oncolumn monitoring of conductance. This detector provides a voltage output rather than absolute conductance readings; it was used without specific calibration as only relative changes in conductance was important. An ultraviolet absorbance detector (Linear PHD 206) and a spectrofluorometer (LS-50, www. perkinelmer.com) were also used as indicated. Fluorescent Ion Displacement. The column was flushed with 2 mM disodium fluorescein (pH 8.2) for 2 h, followed by pure water for 1 h to remove unbound fluorescein. Next, the bound fluorescein was eluted with 0.20 M Na2SO4 for 3 h at a slow flow rate. The effluent was collected and diluted to 5 mL, and the fluorescence was measured (λex 490 nm, λem 510 nm, 2 nm slits). The data were interpreted by a calibration plot of fluorescein in the same 0.20 M Na2SO4 medium. As fluorescein has two anionic sites, the ion exchange capacity was taken to be twice the number of moles of fluorescein eluted from the column. Frontal Displacement Chromatography. The column was flushed with 10 mM of KIO3 for 2 h and rinsed with ultrapure water for 40 min followed by 1 mM KCl while the conductivity or absorbance is detected by on-column conductivity or UV absorbance detectors. UV detection experiments were also conducted with chloride (as NaCl) as the initially loaded ion and nitrate (as KNO3) as the displacing ion. The column capacity was calculated as previously stated. Frontal Reaction Chromatography. FRC was carried out in two different ways. For undifferentiated total anion exchange capacity measurement, the column was flushed with 50 mM NaOH for 30 min for complete conversion to the hydroxide form. Next, pure water wash for 30 min was followed by 1 mM HNO3 at a carefully measured flow rate. Either C4D or UV detection (at 210 nm) was used. However, strong base, e.g., NaOH, at ≥10 mM or higher concentrations was found to result in dislodgment of the latex from the PMMA wall and subsequent column blockage with many latex types. This was especially true of latexes exhibiting a significant amount of weak anion exchange capacity. Differentiated measurements of weakly and strongly basic anion exchange capacities were carried out as follows. The column was washed with 10 mM NaOAc for 1 h to convert the column to the acetate form. Three different displacement solutions, of 1 mM HNO3, NH4NO3, or NaNO3, were each used in separate

C = [HNO3 , equiv/cm 3] × Q (cm 3/s) × (tb − th) (s) (2)

where tb and th are the observed breakthrough times for the acid and the time corresponding to the independently determined column holdup volume. Note that, unlike the previous approaches that only measure strongly basic exchange sites, because of the use of a highly acidic pH, this approach measures weakly basic sites as well. Indeed, it is possible to differentiate between different types of exchange sites by using loading or displacing solutions of different pH values. In effect, this is the easiest to use and most versatile and powerful approach to ion exchange capacity measurement.



EXPERIMENTAL SECTION Materials and Methods. Polymethylmethacrylate (PMMA) capillaries, 16−20 μm i.d., 340 μm o.d., were custom extruded by www.paradigmoptics.com. AS18 latex (d = 65 nm) suspensions were supplied by Thermo Scientific (Dionex). The particles have proprietary composition, but retention behavior of AS18 columns can be found in the literature from www. dionex.com. AS18 has a diameter of 65 nm. Various other latex coatings were also investigated to ensure that the measurement principle is equally applicable. However, systematic studies, as were done for the AS18, were not carried, out and no data are presented. Capillary Coating Procedure. The PMMA capillary was washed with ethanol for 30 min to remove any organic residues from the extrusion process, followed by thorough rinsing with 18.2 MΩ·cm Milli-Q water. Unless otherwise stated, the capillary was next flushed by 10% (w/v) NaOH for 90 min to hydrolyze the ester and create free carboxylate groups and then washed again with water. The latex solution as received was diluted 10× before passing through the capillaries back and forth, for at least 2 h 7996

dx.doi.org/10.1021/ac4018583 | Anal. Chem. 2013, 85, 7994−8000

Analytical Chemistry

Article

measured capacity. Also, it is significantly larger than a typical ion for which analyses are conducted by IC, it may not be able to access all the sites a typical ion can, thus causing a negative bias. Fluorescein binds strongly; on the basis of online visible absorbance measurements, a 3 h elution with 0.20 M Na2SO4 was used to ensure complete elution. Off all three procedures described herein, this is the most time-consuming. Note also the measured capacities per column are in the range of tens to hundreds of pequiv, ultimately resulting in fluorescein concentrations in the measured solution at the nM level, necessitating careful manipulations. Despite its limitations, this ion exchange capacity (perhaps most aptly named f luorescein binding capacity) is a parameter that can be measured without any special equipment. Frontal Displacement Chromatography. Figure 2 illustrates the case where iodate is the initially loaded ion, and

experiments to put nitrate on the column. Note that these solutions, respectively, have strongly acidic, barely acidic, and neutral pH, thus respectively converting all anion exchange sites, all but the most weakly basic sites and finally only the strongly basic sites to the nitrate form. In principle, it would have been possible to simply determine the time for the breakthrough of the displacement solution, but we found more reproducible results with a short water wash followed by 1 mM KClO4 to elute the nitrate that was monitored by optical absorption at 210 nm.



RESULTS AND DISCUSSION Effect of Hydrolysis Time and Temperature on PMMA Cation Exchange Capacities. The relevant data are shown in Figure 1. With the exception of some benefit of at least etching

Figure 1. Cation exchange capacities per unit area of 16−20 μm i.d. PMMA columns after different alkaline hydrolysis time (circles) and anion exchange capacities after coating the same with 65 nm AS18 anionic latex particles (squares). The error bars reflect three separate columns, not triplicate measurement on the same column. The anion exchange capacities determined by the frontal reaction chromatography-acid base titration method reflect total anion exchange capacity; of this, the strong base type capacity is approximately half.

Figure 2. Representative elution curve for open tubular ion exchange capacity measurement using frontal displacement chromatography with conductometric detection. Column AS18 coated 0.0195 × 560 mm PMMA, 1 mM LiCl displacer @ 41 nL/min, calculated total capacity 302.7 pequiv or 8.8 pequiv/mm2.

chloride is the displacing ion on an AS18 coated column. Both are fully ionized strong acid ions, and chloride is far more strongly retained than iodate on AS18, essentially resulting in frontal displacement. With a conductivity detector, the equivalent conductance of iodate (40.5 S cm2 equiv−1) and chloride (76.4 S cm2 equiv−1) also differ markedly. Thus, an obvious “step” zone of displaced iodate is readily distinguished. It will be obvious to the reader that if we used a UV detector in this experiment at any wavelength where chloride does not have significant absorption but iodate does, the iodate zone would have appeared as a rectangular “pulse”, rather than a step. In this case, because the beginning and the displaced ion step are accurately known, one does not need to know the column void volume. It is important to ensure that the column is fully in iodate form as iodate is so weakly held. Initial removal of other ions with 10 mM NaOH or NaOAc, prior to loading iodate for a prolonged period (10 mM, 2 h) yielded reproducible results. It is necessary to avoid extensive exposure to CO2 to prevent formation of HCO3−.

for some time, the alkaline hydrolysis step, regardless of time or temperature, had little or no influence on the measured cation exchange capacity, on the order of 1 pequiv/mm2. Obviously, surface roughening and generation of greater surface area does not occur by this procedure. This cation exchange capacity, after coating with AS18 latex, results in slightly more than an order of magnitude increase in capacity as an anion exchanger. This capacity is relatively uniform and shows no dependence on the previous etching conditions (r2 = 0.22). Determination of Anion Exchange Capacities. Fluorescent Ion Displacement. Choice of fluorescein was based on its well characterized chemical behavior, known pKa values, and its high quantum efficiency that allows it to be measured at very low levels using instrumentation available in most laboratories. The use of off-line measurement avoids RI and matrix effects. The downside of using fluorescein, however, is that it may show some hydrophobic binding (e.g., through π−π interactions with an aromatic latex particle) thus causing a positive error in the 7997

dx.doi.org/10.1021/ac4018583 | Anal. Chem. 2013, 85, 7994−8000

Analytical Chemistry

Article

Frontal Reaction Chromatography. Total ion exchange capacity measurement involves passing a strong acid through a base form of the exchanger and monitoring emergence of excess acid. Note that the type of capacity measured really depends on the pH of the final displacer. With a low pH, the total capacity is measured. The choice of HNO3 allows either optical or conductometric detection. The acid concentration is chosen to get a reasonable tb − th value. Figure 4 shows a

If UV detection is used, a second possibility is that an optically transparent loading ion (e.g., chloride) is chosen along with a more strongly held optically absorbing displacing ion (e.g., nitrate). Many of the precautions needed to completely convert a column to iodate do not apply while loading chloride. However, the void volume of the column must be accurately known. Figure 3 shows a typical trace from such an experiment.

Figure 3. Representative elution curve for open tubular ion exchange capacity measurement using frontal displacement chromatography with UV absorbance detection. Column AS18 coated 0.0180 × 1000 mm PMMA, 1 mM KNO3 displacer @ 36 nL/min, calculated total capacity 218.5 pequiv or 3.9 pequiv/mm2.

Figure 4. Representative elution curve for open tubular total anion exchange capacity measurement using frontal reaction chromatography with conductometric detection. Column AS18 coated 0.0180 × 860 mm PMMA, 1 mM HNO3 titrant @ 28.5 nL/min, calculated total capacity 378 pequiv or 7.8 pequiv/mm2.

At the detection wavelength, there was essentially no absorption from chloride. The zero time in the figure corresponds to the time that nitrate introduction is begun. Breakthrough occurs at the indicated time of tb, and the time corresponding to the holdup volume th at the operating flow rate is also indicated in the figure. There is a transient RI disturbance accompanying the emergence of nitrate. However, this is reproducible, and the tb identification (midpoint of steeply rising trace) uncertainty is only ∼2% relative to tb − th. During the review of this manuscript, an anonymous reviewer pointed out that judicious use of an absorbing counterion (e.g., imidazolium) or even a neutral solute (e.g., thiourea, methyl isobutyl ketone added at a concentration to obtain a desired level of absorbance) in the displacing solution will serve to indicate the void volume of the column. From the time the pumping of the displacing solution begins on the water-filled column, the emergence of the first absorption step will be due to the emergence of this counterion/neutral solute and indicate the void time. The absorption due to the displacing ion (e.g., nitrate) will then appear later atop this counterion absorption. It will be fortuitous if the absorption from either the counterion/neutral solute or the displacing ion can be measured independently of the other without crosstalk using a two-wavelength detector. If a single wavelength measurement must be made, the counterion must be chosen such that it has about the same absorptivity as the displacing ion. This will ensure that both steps are easily discernible. A simpler alternative is to use a neutral solute whose concentration can be adjusted so that it will have about the same absorbance as the displacing ion at the measurement wavelength.

representative trace for such an experiment. Obviously, such an approach can also be used for the measurement of cation exchange capacities by reversing the roles of the loaded ion and the titrant. pH-Differentiated FRC: Subclassification of Ion Exchange Sites. As previously stated, the displacing solution pH essentially governs whether the particular site will be ionized. An AS-18 coated 0.016 × 420 mm column, with loading solutions of 1 mM HNO3, NH4NO3, or NaNO3 (approximate respective pH 3, 6.1, and 7) resulted in total measured column capacities of 288.4 ± 8.6, 116.2 ± 15.1, and 131.0 ± 35.7 (n = 3) pequiv (or 13.7 ± 0.4, 5.5 ± 0.7, and 6.2 ± 1.7 pequiv/ mm2). The ion exchange capacities at pH 6 and 7 are indistinguishable within experimental uncertainty, but the low pH experiment indicates that the additional weak base type capacity of the column is actually somewhat greater than the strong base capacity. While only the latter might be effective in traditional ion chromatography with a strongly alkaline eluent, such columns can conceivably be used for other types of separations in which the additional weak base character that presumably results from incomplete quaternization can play a significant role. In fact, the manufacturer was unaware of the extent weak base sites are present in this latex. The fact that this latex contains a significant amount of weak base character is also readily revealed by acid−base macroscale titration of the latex suspension; a titration plot and its first derivative is presented in Figure 5. Comparative Values. As the methods do not strictly measure an absolute value but are operationally defined, only a few columns were measured by more than one method. An 7998

dx.doi.org/10.1021/ac4018583 | Anal. Chem. 2013, 85, 7994−8000

Analytical Chemistry

Article

also produce attractive chromatographic performance; in a shortly forthcoming article, we will show that such columns can also produce attractive separations (