Analytical and Bioanalytical Chemistry

, Volume 391, Issue 7, pp 2551–2556

Anion exchange silica monolith for capillary liquid chromatography

Authors

    • Department of Chemistry, Faculty of ScienceUniversiti Teknologi Malaysia
  • Yuta Watanabe
    • Department of Biomolecular EngineeringKyoto Institute of Technology
  • Tohru Ikegami
    • Department of Biomolecular EngineeringKyoto Institute of Technology
  • Kosuke Miyamoto
    • Department of Biomolecular EngineeringKyoto Institute of Technology
  • Nobuo Tanaka
    • Department of Biomolecular EngineeringKyoto Institute of Technology
Technical Note

DOI: 10.1007/s00216-008-2063-3

Cite this article as:
Jaafar, J., Watanabe, Y., Ikegami, T. et al. Anal Bioanal Chem (2008) 391: 2551. doi:10.1007/s00216-008-2063-3

Abstract

An anion exchange monolithic silica capillary column was prepared by surface modification of a hybrid monolithic silica capillary column prepared from a mixture of tetramethoxysilane (TMOS) and methyltrimethoxysilane (MTMS). The surface modification was carried out by on-column copolymerization of N-[3-(dimethylamino)propyl]acrylamide methyl chloride-quaternary salt (DMAPAA-Q) with 3-methacryloxypropyl moieties bonded as an anchor to the silica surface to form a strong anion exchange stationary phase. The columns were examined for their performance in liquid chromatography (LC) and capillary electrochromatography (CEC) separations of common anions. The ions were separated using 50 mM phosphate buffer at pH 6.6. Evaluation by LC produced an average of 30,000 theoretical plates (33 cm column length) for the inorganic anions and nucleotides. Evaluation by CEC, using the same buffer, produced enhanced chromatographic performance of up to ca. 90,000 theoretical plates and a theoretical plate height of ca. 4 μm. Although reduced efficiency was observed for inorganic anions that were retained a long time, the results of this study highlight the potential utility of the DMAPAA-Q stationary phase for anion separations.

https://static-content.springer.com/image/art%3A10.1007%2Fs00216-008-2063-3/MediaObjects/216_2008_2063_Figa_HTML.gif
Figure

Micro-LC performance evaluation of a strong anion exchange silica monolith column, 100H-MOP-DMAPAA-Q, 33 cm in length, with a mobile phase of 50 mM phosphate buffer, pH 2.8; linear velocity: u = 1.8 mm/s; UV-Vis detection at 254 nm. Sample solution (5 mg/mL of each component, 4 mL) was injected in split flow injection mode at a split ratio of ca. 1:1900 with a pump flow rate of 1.5 mL/min

Keywords

Anion exchangeSilica monolithMicro liquid chromatographyCapillary electrochromatographyInorganic anions

Abbreviations

CEC

Capillary electrochromatography

TMOS

Tetramethoxysilane

MTMS

Methyl trimethoxysilane

DMAPAA-Q

N-[3-(dimethylamino)propyl]acrylamide methyl chloride-quaternary salt

MOP

3-Methacryloxypropyl

Introduction

Anion exchange chromatography plays an important role in the separation of inorganic and organic ions. The benefits of monolithic phases have been explored for ion analysis [1] and progress made in the fabrication of capillary silica monoliths has contributed to the increased research activity in ion chromatography [2]. Capillary ion chromatography using monolithic columns provides high-performance or rapid and efficient ion analysis. Suzuki et al. [3] used monolithic silica capillary columns modified with cetyltrimethylammonium salts for the separation of five inorganic anions in seawater matrix in less than 1 min. A retention factor, k = 3.2, and a plate height, H = 87 μm, were reported for nitrate in 500 mmol/L sodium chloride with 0.1 mmol/L of cetyltrimethylammonium chloride (CTAC), but the capillary needed the addition of the CTAC modifier to stabilize the retention times. The separation of inorganic ions using a 250 μm ID quaternary ammonium latex-coated monolithic polymeric stationary phase was reported by Zakaria et al. [4]. Rapid separation of seven anions was achieved in 2 min using conductivity detection; however, the separation efficiencies were moderate, with H of 55 μm for iodate showing k = 0.30 in 20 mmol/L Tris/acetate buffer at pH 8.0. A reversed-phase monolithic silica capillary column coated with a surfactant, N-dodecyl-N,N-(dimethylammonio)undecanoate (DDMAU), was recently reported for the separation of inorganic ions [5]. A more sensitive detector for ion analysis, i.e., an on-column contactless conductivity detector, was used. The column efficiency of H = 29 μm for iodide (k = 5.35) in 0.5 mmol/L phthalate buffer, was better than for some other types of capillary ion-exchange columns that have been reported [2]. Glenn et al. [6] coated a conventionally sized monolithic silica column (Merck Chromolith) with Dionex AS9-SC latex nanoparticles. This column afforded a very fast separation of seven anions in 30 s, although larger H values were observed compared to a didodecyldimethylammonium bromide (DDAB)-coated silica monolithic column, which produced H values as low as 10 μm for iodate ion in 6 mmol/L o-cyanophenol buffer, pH 7.0 [7]. Such ion exchange stationary phases were prepared by modifying a reversed-phase column. Ikegami et al. [8] reported a novel method for the preparation of an anion and cation exchange microHPLC column. The column was used for the separation of nucleic acid, nucleotides and inorganic anions, and presented H = 12 μm for 5′CMP (k = 0.74) in 0.05 mol/L NH4H2PO4 (pH 3.0) buffer. Pelletier and Lucy [9] prepared a short silica-based monolithic column coated with DDAB that demonstrated an efficiency of H = 36–40 μm (k = 1.5–5.0) for \({\text{IO}}^{ - }_{{\text{3}}} \), Cl, \({\text{NO}}^{ - }_{{\text{2}}} \), Br and \({\text{NO}}^{ - }_{3} \) ions in 9 mmol/L 4-cyanophenol eluent, pH 7.3. However, the capacity of this column decreased with time, and periodic recoating of the column is needed.

Interest in monolithic silica capillary columns for CEC has also grown recently. Most of the capillaries used in CEC have been silica-based ODS-type stationary phases that are modified into anion-exchange or mixed-mode phases by using dynamic and adsorptive coatings. Scherer and Steiner [10] synthesized a porous encapsulated silica monolith with a mixed mode of strong anion exchange and reversed phases. The column exhibited a high efficiency, H = 7 μm for neutral solutes in 5 mmol/L phosphate, pH 4.0 water–acetonitrile eluent, while slightly higher H values were obtained for some carboxylate ions. A CEC ion exchange capillary using a latex-coated polymer monolith stationary phase was reported for the separation of inorganic anions with Tris/perchlorate electrolyte and UV detection at 195 nm [11]. A plate height of 13 μm was obtained for bromide ion in 20 mmol/L or higher perchlorate solution. An anion exchange column for CEC containing a propyl-N,N,N-trimethylammonium group on silicon alkoxide monolith, as reported by Lin et al. [12], was used for the separation of inorganic anions using 40 mM phosphate buffer (pH 3.5), with a plate height of 7 μm for \({\text{HCrO}}^{ - }_{{\text{4}}} \) ion. Overall, column efficiencies have improved tremendously for some new stationary phase materials reported in CEC mode.

Here, we describe preliminary results from the evaluation of a novel strong anion exchange silica monolith capillary column prepared by the in situ polymerization of a monomer, N-[3-(dimethylamino)propyl]acrylamide methyl chloride-quaternary salt (DMAPAA-Q), anchored to 3-methacryloxypropylsilyl moieties. Good column efficiency was expected with the monolithic silica capillary columns, because similar columns with ODS modification produced H = 7–10 μm at optimum linear velocity (u = 1.5–2 mm/s) in 80/20 acetonitrile–water for aromatic hydrocarbons. The anion exchange columns were evaluated in pressure-driven liquid chromatography and capillary electrochromatography for the separation of common nucleotides and inorganic anions.

Experimental

Reagents and chemicals

Sodium dihydrogen phosphate, NaH2PO4.2H2O, disodium hydrogen phosphate, Na2HPO4.12H2O and thiourea were obtained from Nacalai Tesque (Kyoto, Japan). (3-methacryloxypropyl)trimethoxysilane (MOP-silane) was supplied by ShinEtsu Silicon Chemicals (Tokyo, Japan); ammonium persulfate (APS) was purchased from Wako Pure Chemicals (Osaka, Japan), and N-[3-(dimethylamino)propyl]acrylamide methyl chloride-quaternary salt, DMAPAA-Q, was obtained as a 75% water solution from Kohjin Co. Ltd. (Tokyo, Japan). A 50 mM phosphate buffer at pH 6.6 was prepared, and the pH values were measured with a Horiba B-212 compact pH-meter (Kyoto, Japan). Water was distilled by a Milli-Q A10 system (Millipore, Billerica, MA, USA). All buffer solutions were degassed with helium and sonicated prior to use. The inorganic anion standards were prepared as a stock solution of 1000 mg/L from the following species: potassium iodide, potassium iodate, potassium bromate, potassium nitrate, potassium bromide, and sodium thiocyanate (Nacalai Tesque), and 1 mg/mL of individual standards of cytidine-5′-monophosphate (5′CMP, Wako), adenosine-5′-monophosphate (5′AMP, Wako), uridine-5′-monophosphate (5′UMP, Nacalai Tesque), guanosine-5′-monophosphate (5′GMP, Wako). All standard reagents were of analytical grade.

Instrumentation

The liquid chromatographic evaluation of the strong anion exchange monolithic silica capillary column was carried out using a LC-10AD VP pump (Shimadzu, Kyoto, Japan) and a CE1575 detector (JASCO, Hachioji, Japan). Samples were injected using a model 7725 Rheodyne injector (Park Court, CA, USA) used in split-flow injection mode [13]. The capillary columns were evaluated at ambient temperature. Data were collected and processed by EZChrom Elite (Scientific Software, Inc., Pleasanton, CA, USA).

CEC experiments (performed in the same mobile phase as used in LC) were performed on an HP3D CE instrument (Agilent Technologies, Waldbron, Germany) with a UV detector and ChemStation software. The column was thermostatted at 25 °C under 8 bar of nitrogen pressure applied to both the inlet and the outlet vials. Samples were injected electrokinetically, −3.0 kV for 3 s from the cathodic end, and a negative separation voltage was applied.

Off-column detection

The background absorbance of the silica monolith is relatively noisy if a direct detection window is made in the silica section of the capillary. Both LC and CEC used UV detection, which was performed off-column at a signal wavelength of 210 nm. For the HPLC evaluation, the 100H-MOP-DMAPAA-Q capillary, 33 cm in length, was connected to an empty 50 μm ID, 360 μm OD silica capillary (15 cm long) using a 1 cm union made of Teflon tubing of 0.33 mm ID (Part number 6010, GL Science Inc., Tokyo, Japan). A detection window was created in the empty capillary, 7 cm from the union.

For the CEC measurements, the 75H-MOP-DMAPAA-Q capillary was connected to an empty 50 μm ID, 360 μm OD silica capillary of length 10 cm using a 1 cm union made of Teflon tubing of 0.33 mm ID (Part number 6010, GL Science Inc.), for off-column detection. Both capillary ends must be smoothly cut to ensure dead-volume-free connection. A detection window was made 1.5 cm from the joint union, and the capillary was then inserted into the alignment interface of the CE cassette holder. The void volume between the column outlet and the detection window was approximately 0.03 μL. The use of the anion exchange silica monolithic capillary connected to the empty silica capillary provided a useful way to reduce the background absorbance from the monolith.

Synthesis of the anion exchange monolithic silica capillary columns

The synthesis of the monolithic silica is a two-step process, involving (i) the preparation of the hybrid silica monolith, and (ii) the chemical modification to obtain the anion exchange functionality. Monolithic silica capillary columns were prepared from a mixture of tetramethoxysilane and methyltrimethoxysilane, as reported [13]. The TMOS/MTMS hybrid capillaries were then converted to the 3-methacryloxypropylsilyl-bonded (MOP) phase by applying a 1:1 mixture of pyridine:MOP-silane solution through the column for 24 h at 80 °C using N2 at a pressure of 1 MPa. This was followed by a wash with toluene for 24 h via N2 pressure, and methanol for 48 h [14].

Surface modification of the monolithic silica was carried out by radical copolymerization of the bonded MOP anchor groups with the anion exchange monomer DMAPAA-Q. Two hundred microliters of DMAPAA-Q were mixed with 2 mL of a solution of ammonium persulfate (APS) (5 mg APS/mL water) that worked as the initiator of the polymerization reaction. The mixture was flushed into the capillary using a syringe pump for one hour. Polymerization was carried out at 60 °C by immersing the capillary in a water bath for 2 h. After the polymerization reaction, the capillary was flushed in the reverse direction with water, using a HPLC pump, for 3 h. Then the capillary was placed in a methanol flushing line for three days. Scanning electron micrographs were obtained using a SEM (S-3000N, Hitachi, Tokyo, Japan). The columns were labeled as 100H-MOP-DMAPAA-Q for the μLC separation and 75H-MOP-DMAPAA-Q for the CEC separation. The abbreviations 100H- or 75H- were used, where the numbers represent the column i.d. in microns and H represents the hybrid-type silica. The columns were first evaluated for the separation of nucleotides, and then tested for the separation of inorganic anions. Column efficiencies, in terms of plate numbers, N, and plate height, H, were based on measurements of the peak width at half height.

Results and discussion

SEM observation

Scanning electron micrographs (SEM photograph) of the 100H-MOP-DMAPAA-Q used for the μLC separation are shown in Fig. 1. The photographs showed that polymer agglomerates were not formed by the modification method involving copolymerization of the monomer with the anchor group.
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-008-2063-3/MediaObjects/216_2008_2063_Fig1_HTML.gif
Fig. 1

SEM photographs of 100 μm monolithic silica columns, MOP-DMAPAA-Q. a 800× magnification; b 2500× magnification

LC mode evaluation

The efficiency of the strong anion exchange column with quaternary ammonium functional groups was tested using a mixture of nucleotides as a sample in 50 mM sodium phosphate buffer at pH 2.8 in pressure-driven mode. The four nucleotides, 5′-CMP, 5′-UMP, 5′-GMP and 5′-AMP, were separated in less than 6 min, as shown in Fig. 2. The linear velocity, u = 1.8 mm/s, was determined by the elution time of an unretained marker, uracil, traversing the column length. This column (33 cm) was highly efficient, producing an average of 35,000 theoretical plates, H = 9.5 μm for the four nucleotides.
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-008-2063-3/MediaObjects/216_2008_2063_Fig2_HTML.gif
Fig. 2

Micro-LC performance evaluation of the anion exchange silica monolith. Conditions: column: 100H-MOP-DMAPAA-Q, 33 cm length, 50 mM phosphate buffer, pH 2.8; linear velocity: u = 1.8 mm/s; UV-Vis detection at 254 nm: peaks: (1) 5′-CMP, (2) 5′-AMP, (3) 5′-UMP and (4) 5′-GMP. Sample solution (5 mg/mL of each component, 4 μL) was injected in a split flow and injection mode at a split ratio of ca. 1:1900 with a pump flow rate of 1.5 mL/min

The separation of the six anions at a linear velocity of 1.4 mm/s is shown in Fig. 3. The anions were separated using a mobile phase of 50 mM phosphate at pH 6.6 and eluted in the order of iodate, bromate, nitrate, bromide, iodide, and thiocyanate. Thiourea, which was used as the dead time marker, was the first peak eluted. The performance of the DMAPAA-Q-modified column was much better than that of the DMAEA-Q column with 2-(triethylammonium)ethyl methacrylate functionality bonded to monolithic silica reported by Ikegami et al. [8]. The number of theoretical plates, N, and plate heights, H, of each peak are shown in Table 1. The early-eluting peaks are symmetrical and efficient with an average plate height H = 10 μm, although the later two peaks of iodide and thiocyanate showed significant tailing and broadening, presumably due to the slow equilibration of these highly polarizable ions. Better peak shapes can be expected at higher electrolyte concentrations [11].
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-008-2063-3/MediaObjects/216_2008_2063_Fig3_HTML.gif
Fig. 3

Micro-LC separation of anions mixture composed of (1) thiourea; (2) iodate; (3) bromate; (4) nitrate; (5) bromide; (6) iodide; (7) thiocyanate. The sample solution (2 mg/mL of each component, 4 μL) was injected in a split flow and injection mode with a pump flow rate of 0.40 mL/min. Column: 100H-MOP-DMAPAA-Q, 33 cm length; mobile phase: 50 mM sodium phosphate (pH 6.6), linear velocity: u = 1.4 mm/s, λ = 210 nm

Table 1

Column efficiency obtained for the LC separation of inorganic ions

Anion

N

H, μm

Thiourea

42300

8

\({\text{IO}}^{ - }_{3} \)

29900

11

\({\text{BrO}}^{ - }_{3} \)

27100

12

\({\text{NO}}^{ - }_{3} \)

35200

9

Br

31100

11

I

(23000 tailing)

14

SCN

(5680 tailing)

-

A van Deemter plot was constructed for thiourea, iodate, and bromate with linear velocities ranging from 0.8 mm/s to 2.1 mm/s. As shown in Fig. 4, optimum efficiency was observed for the ions at the lowest linear velocity in this range. The column efficiency of this anion exchange monolith is considerably higher than those of other types of ion exchange capillary columns reported [4]. These results highlight the potential usefulness of the DMAPAA-Q modified monolithic silica columns.
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-008-2063-3/MediaObjects/216_2008_2063_Fig4_HTML.gif
Fig. 4

Van Deemter plots for the anion exchange silica monolith. Diamonds, thiourea; squares, iodate; triangles, bromate. The same conditions were employed as in Fig. 3

CEC mode

The performance of the column was also evaluated with CEC. Figure 5 shows the electropherogram of four ions on a 75 μm ID capillary using a 50 mM phosphate buffer at pH 6.6. The EOF was reversed in the anion exchange monolithic capillary due to the positively charged surface, and so a negative separation voltage was applied to enable the migration of both the anions and the EOF marker toward the anode. Four anions were selected for the CEC evaluation, with an elution order of iodate, bromate, nitrate, and bromide. The elution order of the anions is similar to that observed in pressure-driven separation, while the contribution from the electrophoretic mobility is shown by the difference in selectivity observed with the nitrate ion. The separation was completed in less than 5 min. Thiourea, which is the EOF marker, eluted much later, at 17 min. This observation led to the conclusion that the anions migrate predominantly through electrophoresis under the influence of the electrostatic interaction with the quaternary ammonium groups on the ion exchange sites. Breadmore et al. reported a similar phenomenon for the late migration of an EOF marker on a silica monolith which was dynamically coated with a cationic polymer, poly(diallydimethylammonium chloride) [15].
https://static-content.springer.com/image/art%3A10.1007%2Fs00216-008-2063-3/MediaObjects/216_2008_2063_Fig5_HTML.gif
Fig. 5

CEC separation of a 0.20 mg/mL standard solution of four common anions. Experimental conditions: capillary 75H-MOP-DMAPAA-Q, 33.0 cm; BGS: 50 mM phosphate pH 6.6; −3 kV, 3 s injection; -15 kV, 8 bar separation; UV detection 210 nm; peaks: (1) iodate, (2) bromate, (3) nitrate, and (4) bromide. Thiourea showed a migration time of 17 min. The sample solution (0.20 mg/mL of each component) was used

The effect of applied voltage was investigated by monitoring the first two peaks of iodate and bromate. As shown in Table 2, the efficiency increased as the applied voltage was decreased. At −15 kV applied voltage, the capillary produced a plate height, H, of 4 μm. This plate height compares favorably with the results obtained using columns reported previously for inorganic ion separations [3, 12]. The CEC evaluation of the column also provided a better efficiency than μLC evaluation. Similar improvements in performance in electrodriven mode compared to pressure-driven LC have been reported for reversed-phase-mode monolithic silica columns [16].
Table 2

Effect of applied voltage on the column efficiency, N, and H

Analyte

−15 kV

−12 kV

−10 kV

−7 kV

Iodate

N

63000

79600

92500

93200

 

H, μm

5.0

4.0

3.5

3.4

Bromate

N

82400

84000

80800

90100

 

H, μm

3.9

3.8

4

3.6

The reproducibility of the anion exchange column was evaluated by deriving the relative standard deviations (RSD) of the migration times of the anions. Three replicates of migration time measurements for the four anions showed that the column presented good short-term stability, with RSDs in the range of 4.7–8.8%. The reproducibilities were not good but still acceptable for a CE method.

Conclusions

Polymerization of a DMAPAA-Q monomer in a MOP-modified hybrid monolithic silica capillary column afforded anion exchange columns. These MOP-DMAPAA-Q columns exhibited good performance for some ions, with large numbers of theoretical plates or small plate heights, although tailing was observed with late-eluting polarizable anions. The modified silica monolith columns achieved efficiencies of up to 40,000 theoretical plates in LC and 90,000 theoretical plates for the inorganic anions in CEC. CEC gave higher efficiency and faster separations than μLC. The results suggest that modifying a monolithic silica capillary column through the in situ polymerization of a monomer carrying a functional group can yield high-efficiency columns for ion-exchange-mode separations as well as for reversed-phase or HILIC-mode separations [14, 17].

Acknowledgement

JJ acknowledges the financial support from the JSPS-RONPAKU fellowship, VCC-10514, funded by the Japan Society for the Promotion of Science.

Copyright information

© Springer-Verlag 2008