Journal of Chromatography A, 1030 (2004) 201–208
`
`Characterization of some physical and chromatographic properties of
`monolithic poly(styrene–co-divinylbenzene) columns
`Herbert Oberacher a,b,1, Andreas Premstaller a,2, Christian G. Huber b,∗
`
`a Institute of Analytical Chemistry and Radiochemistry, Leopold-Franzens-University, 6020 Innsbruck, Austria
`b Instrumental Analysis and Bioanalysis, Saarland University, Building 9.2, 66123 Saarbrücken, Germany
`
`Abstract
`
`Monolithic capillary columns were prepared by copolymerization of styrene and divinylbenzene inside a 200 ␮m i.d. fused silica capillary
`using a mixture of tetrahydrofuran and decanol as porogen. Important chromatographic features of the synthesized columns were charac-
`terized and critically compared to the properties of columns packed with micropellicular, octadecylated poly(styrene–co-divinylbenzene)
`(PS–DVB–C18) particles. The permeability of a 60 mm long monolithic column was slightly higher than that of an equally dimensioned
`column packed with PS–DVB–C18 beads and was invariant up to at least 250 bar column inlet pressure, indicating the high-pressure stability
`of the monolithic columns. Interestingly, monolithic columns showed a 3.6 times better separation efficiency for oligonucleotides than granular
`columns. To study differences of the molecular diffusion processes between granular and monolithic columns, Van Deemter plots were mea-
`sured. Due to the favorable pore structure of monolithic columns all kind of diffusional band broadening was reduced two to five times. Using
`inverse size-exclusion chromatography a total porosity of 70% was determined, which consisted of internodule porosity (20%) and internal
`porosity (50%). The observed fast mass transfer and the resulting high separation efficiency suggested that the surface of the monolithic sta-
`tionary phase is rather rough and does not feature real pores accessible to macromolecular analytes such as polypeptides or oligonucleotides.
`The maximum analytical loading capacity of monolithic columns for oligonucleotides was found to be in the region of 500 fmol, which
`compared well to the loading capacity of the granular columns. Batch-to-batch reproducibility proved to be better with granular stationary
`phases compared to monolithic stationary phase, in which each column bed is the result of a unique column preparation process.
`© 2004 Elsevier B.V. All rights reserved.
`
`Keywords: Monolithic columns; Capillary columns; Stationary phases, LC; Poly(styrene–divinylbenzene); Peptides; Nucleic acids
`
`1. Introduction
`
`Stationary phases based on microparticles have been suc-
`cessfully utilized as separation media for high-performance
`liquid chromatography (HPLC) for almost four decades
`[1–6]. However, HPLC columns packed with micropar-
`ticulate, porous stationary phases have some limitations,
`namely the relatively large void volume between the packed
`particles and the slow diffusional mass transfer of solutes
`into and out of the stagnant mobile phase present in the
`pores of the separation medium, resulting in considerable
`band broadening particularly with high molecular analytes
`
`Corresponding author. Tel.: +49-681-302-2433;
`∗
`fax: +49-681-302-2963.
`E-mail address: christian.huber@mx.uni-saarland.de (C.G. Huber).
`1 Present address: Institute of Legal Medicine, Leopold-Franzens-
`University, 6020 Innsbruck, Austria.
`2 Present address: Sandoz GmbH, 6250 Kundl, Austria.
`
`0021-9673/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
`doi:10.1016/j.chroma.2004.01.009
`
`[7,8]. One possible route to enhance the mass transfer rep-
`resents the complete elimination of diffusive pores, which
`restricts the mass transfer to a thin, retentive layer at the
`outer surface of the stationary phase, resulting in so-called
`micropellicular stationary phases [9].
`An alternative approach to alleviate the problem of re-
`stricted mass transfer and intraparticular void volume is the
`concept of monolithic chromatographic beds, in which the
`separation medium consists of a continuous rod of a rigid,
`porous polymer which has no interstitial volume but only
`internal porosity consisting of micropores and macropores
`[10–13]. Because of the absence of intraparticular volume,
`all of the mobile phase is forced to flow through the pores
`of the separation medium [14]. According to theory, mass
`transport is enhanced by such convection [15–17], which
`has a positive effect on chromatographic efficiency. There-
`fore, monolithic stationary phases have become a rapidly
`burgeoning field in the preparation of chromatographic sta-
`tionary phases in recent years [18].
`
`1
`
`MTX1020
`
`

`

`202
`
`H. Oberacher et al. / J. Chromatogr. A 1030 (2004) 201–208
`
`In general, monolithic columns can be divided into two
`categories. (i) Silica-based monolithic columns are gener-
`ally prepared using sol–gel technology. This technology
`can be applied to create a continuous sol–gel network
`throughout the column formed by the gelation of a sol solu-
`tion [13,19,20]. (ii) The second category is represented by
`rigid organic polymer-based monolithic columns including
`acrylamide-based [21,22], acrylate- or methacrylate-based
`[12,23–25], and styrene-based polymers [26–30]. Porous
`matrices are obtained when polymerization and crosslink-
`ing take place in the presence of inert porogens which lead
`to a phase separation during the ongoing polymerization
`reaction, effecting the formation of permanent pores in the
`material.
`The concept of monolithic stationary phases is espe-
`cially favorable for the fabrication of capillary columns
`[28,31–33]. Covalent immobilization of the monolith at
`the wall of a fused capillary eliminates the necessity to
`prepare a tiny retaining frit, which is one of the more te-
`dious and difficult to control steps during the manufacture
`of packed bed capillary columns [34]. Moreover due to the
`in situ polymerization of the monolithic chromatographic
`bed within the confines of a fused silica tube the laborious
`steps of particle synthesis and column packing could be
`overcome. On account of this, we introduced monolithic
`capillary columns prepared by copolymerization of styrene
`and divinylbenzene inside a 200 ␮m i.d. fused silica cap-
`illary using a mixture of tetrahydrofuran and decanol as
`porogen for the highly efficient separation of single- and
`double-stranded nucleic acids by ion-pair reversed-phase
`HPLC (RP-HPIPC) and of peptides and proteins by
`reversed-phase HPLC (RP-HPLC) [29,35–37]. Continu-
`ing our recent work, a discussion of important physical
`and chromatographic properties of the monolithic capillary
`columns is presented in this paper. Characteristics including
`reproducibility of fabrication, loading capacity, pore size
`distribution, and molecular diffusion processes within the
`chromatographic bed were compared to the features of cap-
`illary columns packed with micropellicular, octadecylated,
`2.1 ␮m poly(styrene–co-divinylbenzene) particles.
`
`2. Experimental
`
`2.1. Chemicals and samples
`
`Acetonitrile (HPLC gradient-grade), acetic acid (analyt-
`ical reagent grade), methanol (gradient grade), and water
`(HPLC grade) were obtained from Merck (Darmstadt, Ger-
`many). Trifluoroacetic acid (TFA, for protein sequence
`analysis), tetrahydrofuran (puriss.), toluene (puriss.), and
`triethylamine (analytical reagent grade) were purchased
`from Fluka (Buchs, Switzerland). A 1.0 M stock solution
`of triethylammonium acetate (TEAA), pH 7.0, was pre-
`pared by adding acetic acid to a 1.0 M aqueous solution of
`triethylamine until pH 7.0 was reached.
`
`Polystyrene standards for size-exclusion chromatogra-
`phy were obtained from Polymer Standards Service (PSS:
`Mainz, Germany) (mass, 94 650), and from Waters (Milford,
`MA, USA) (masses, 440, 2350, 3600, 6870, 15 000, 35 000,
`49 300, 200 000, 470 000, 803 000, 1 260 000, 2 700 000,
`3 150 000, 3 390 000, 4 110 000, 6 590 000) and had a poly-
`dispersity lower than 1.10 except for the standards with
`a mass above 3000000, which had a polydispersity lower
`than 1.3.
`The oligodeoxynucleotide standard [a mixture of (dT)12 to
`(dT)18] was purchased as sodium salt from Pharmacia (Up-
`psala, Sweden). The synthetic oligonucleotide (dT)16 was
`ordered from Microsynth (Balgach, Switzerland) and used
`without further purification. The peptide standard was ob-
`tained from Sigma (St. Louis, MO, USA).
`
`2.2. High-performance liquid chromatography
`
`The Ultimate fully integrated capillary HPLC system
`(LC Packings, Amsterdam, The Netherlands) was used
`for all chromatographic measurements except the inverse
`size-exclusion experiments where HPLC was carried out
`using a micro pump (model Rheos 2000, Flux Instruments,
`Karlskoga, Sweden) controlled by a personal computer
`with Janeiro II software (Flux Instruments), a microinjector
`(model C4-1004, Valco Instruments, Houston, TX, USA)
`with a 20 nl internal sample loop, a variable-wavelength
`detector (model UltiMate UV detector, LC Packings), and a
`Personnal Computer-based data system (Chromeleon 6.00,
`Dionex-Softron, Germering, Germany). The primary flow
`rate and hence the pressure was set at the micropump, the
`flow was split using a tee-piece and a restriction capillary
`(100 cm × 50 ␮m i.d.), and the resulting flow rate through
`the column was measured at the column exit. The use of
`tetrahydrofuran as mobile phase necessitated the use of
`stainless steel tubes for all high-pressure connections. The
`detection cell was in all cases a 3 nl ULT-UZ-N10 cell (LC
`Packings).
`Monolithic poly(styrene–co-divinylbenzene) (PS–DVB)
`capillary columns were prepared according to the published
`protocol [29] and have been commercialized as Mono-
`lith by LC Packings. Octadecylated PS–DVB particles
`(PS–DVB–C18) were synthesized as published in the lit-
`erature [38]. The PS–DVB–C18 stationary phase has been
`commercialized as DNASep by Transgenomic (Santa Clara,
`CA, USA). Granular capillary columns were prepared ac-
`cording to the procedure described in [34].
`
`3. Results and discussion
`
`3.1. Column permeability
`
`Porous polymeric stationary phases in contact with or-
`ganic solvents often lack sufficient mechanical strength and
`the polymer may be deformed under the pressure gradient
`
`2
`
`

`

`H. Oberacher et al. / J. Chromatogr. A 1030 (2004) 201–208
`
`203
`
`tive swelling of the polymer rod in contact with the organic
`solvent, the permeability decreased and the backpressure
`was higher than initially expected. However, this observa-
`tion is not really relevant for HPLC since tetrahydrofuran is
`hardly ever used as mobile phase component, and all other
`solvents (water, acetonitrile, methanol), which are common
`solvents for HPLC, do not cause any considerable swelling
`of the chromatographic bed.
`Finally, numerical values for the specific permeability of a
`monolithic capillary column were determined. Acetonitrile
`and water were passed through a 55 mm× 0.2 mm monolith
`◦
`at a pressure of 90 bar and a temperature of 20
`C. The linear
`flow velocity was 1.3 mm/s for acetonitrile and 0.58 mm/s
`for water. The specific permeability B0 of the column was
`2.9 × 10
`−15 m2 for acetonitrile and 3.5 × 10
`−15 m2 for wa-
`ter. The specific permeability with water is thus 20% higher
`than with acetonitrile. This indicates that some swelling of
`the stationary phase and restriction of the accessible pore
`volume occurs also with acetonitrile, yet to a much lesser
`degree than with tetrahydrofuran. The apparent particle di-
`ameter dp was calculated for acetonitrile as a solvent us-
`ing the Kozeny–Carman equation. A volumetric flow rate
`of 3.0 × 10
`−11 m3 s
`−1 and an elution time of the unretained
`compound of 43.3 s yielded a porosity of 0.75. Using the
`specific permeability given above, a value of 280 nm was
`calculated for the apparent particle diameter.
`The monolithic columns were synthesized to exhibit
`hydrodynamic properties comparable to that of packed
`columns. The back pressure in a 6 cm long monolithic
`column at a flow rate of 3 ␮l/min water was typically in
`the range of 90–120 bar, which compared well to a col-
`umn packed with PS–DVB–C18 beads of equal dimensions
`which exhibited a back pressure of 150 bar. The lower back
`pressure in monoliths is an indication of an increased total
`column porosity.
`
`3.2. Batch-to-batch reproducibility of column fabrication
`
`Since the identification of individual components by
`HPLC is usually based on comparison of retention time, a
`high reproducibility of retention is a prerequisite for any
`chromatographic method. Retention times should show
`only a slight fluctuation between runs on one and the same
`column and between runs on different column batches, re-
`spectively. In general, the run-to-run reproducibility is an
`indicator for the quality of the HPLC system and is typically
`in the range of better than 1.0%. The batch-to-batch repro-
`ducibility, on the other hand, reflects the reproducibility of
`column fabrication.
`In order to evaluate the batch-to-batch reproducibility
`of the two column types, the retention times of a mix-
`ture of seven homologous oligothymidylic acids ranging
`in size from 12 to 18 nucleotides eluting from 17 differ-
`ent 60 mm × 0.2 mm i.d. monolithic PS–DVB columns
`and from 10 columns of the same dimensions packed with
`PS–DVB–C18 particles were measured. The experimental
`
`tetrahydrofuran
`
`water
`
`methanol
`
`acetonitrile
`
`4
`2
`flow rate [µL/min]
`
`6
`
`300
`
`200
`
`100
`
`pressure drop [bar]
`
`0
`
`0
`
`Fig. 1. Graph illustrating plots of pressure drop vs. flow velocity of
`different liquids. Column, PS–DVB monolith, 60 mm × 0.2 mm; mobile
`phases, (䉬) tetrahydrofuran, (䊏) water, (䉱) methanol, (䊉) acetonitrile;
`◦
`C.
`temperature, 20
`
`normally encountered in HPLC columns. In order to eval-
`uate the mechanical stability of our column material, the
`pressure drop across the column was measured upon per-
`fusing it with various solvents in a wide range of flow rates.
`Fig. 1 shows the effect of flow rate on the back pressure in a
`monolithic capillary column for four different solvents. An
`excellent linear dependence of the column inlet pressure on
`the flow rate is indicated by a regression factor R better than
`0.9998 for all measured curves. Thus, with any given sol-
`vent the permeability of the 60 mm × 0.2 mm i.d. column
`was invariant up to at least 250 bar column inlet pressure
`and no impairment of the column integrity occurred. This
`confirms that the rod is not compressed even at high flow
`rates.
`Using water, a flow rate of 1.9 ␮l/min caused a back pres-
`sure of 98 bar. For methanol and acetonitrile a smaller pres-
`sure drop was registered and the order of permeability of
`the column agrees with that expected when comparing the
`viscosities of the utilized solvents (Table 1). However, the
`highest back pressure at equal flow rate was registered when
`using tetrahydrofuran as solvent (Table 1). According to the-
`ory, for a given porous structure with a given column per-
`meability, the pressure drop through a column at a given
`flow rate is only dependent on the viscosity of the solvent.
`Therefore the back pressure of tetrahydrofuran should be
`between that of methanol and acetonitrile. Yet, if the pore
`structure and hence the permeability of the column changes
`depending on the utilized solvent, this relationship is no
`longer valid. Specifically tetrahydrofuran caused a distinc-
`
`Table 1
`Viscosity of solvents and pressure drop per flow rate
`
`Solvent
`
`Water
`Methanol
`Acetonitrile
`Tetrahydrofuran
`
`Viscosity η (kg/(ms))
`1.002 × 10
`0.597 × 10
`0.360 × 10
`0.486 × 10
`
`−3
`−3
`−3
`−3
`
`Pressure drop/flow
`rate (bar/(␮l/min))
`45.6
`35.4
`25.0
`49.6
`
`3
`
`

`

`204
`
`H. Oberacher et al. / J. Chromatogr. A 1030 (2004) 201–208
`
`(a) granular column
`
`(b) monolithic column
`
`8.0
`
`7.0
`
`6.0
`
`5.0
`
`4.0
`
`retention time [min]
`
`12 13 14 15 16 17 18
`
`12 13 14 15 16 17 18
`
`length [nt]
`
`length [nt]
`
`Fig. 2. Comparison of the average retention times of (dT)12–18 on (䉱) 10 monolithic and (䊏) 17 granular capillary columns. Columns, (䉱) PS–DVB
`monolith, 60 mm× 0.2 mm i.d., (䊏) PS–DVB–C18, 2.1 ␮m, 60 mm× 0.2 mm i.d.; mobile phase: (A) 100 mM TEAA, pH 7.0; (B) 100 mM TEAA, pH 7.0,
`◦
`20% acetonitrile; linear gradient 25–60% B in 10.0 min; flow rate, 2.0–3.0 ␮l/min; temperature, 50
`C; detection, UV, 254 nm; sample, (dT)12–18, 1.25 ng.
`
`results are depicted in Fig. 2. The average standard de-
`viation of the retention times among various batches of
`granular and monolithic capillary columns was found to
`be 4.2 and 9.5%, respectively, which clearly demonstrates
`that the uniform packing of presynthesized particles into
`an empty tube to form a granular column is much easier
`to control than the complete de novo synthesis of a chro-
`matographic bed accomplished during the fabrication of the
`monoliths.
`At this point it must be emphasized, that the variability in
`retention between different batches of monolithic columns
`most probably reflects slight differences in surface mor-
`phology. If the batch-to-batch reproducibility of the particle
`synthesis had been considered in this study, increased re-
`tention time deviations would have been observed also for
`the granular column format. Nevertheless, run-to-run repro-
`ducibility of retention times on a single monolithic column
`of 0.5–3% is well within the values that are characteristic
`for capillary chromatographic systems [39]. Consequently,
`monolithic column production requires a careful control of
`synthetic conditions and a rigorous selection of synthesized
`columns, if high reproducibility of analyte retention times
`between column batches is obligatory. Nevertheless, this is
`not a primary concern for routine proteomic and genomic
`applications, in which most of the important information is
`extracted from the mass spectral data that are not influenced
`by slight shifts in chromatographic retention.
`Another interesting difference between the two column
`types can be deduced from the average retention time val-
`ues (Fig. 2). On average the oligonucleotides eluted 24 s
`later from the granular column than from the monolithic col-
`umn. Since the PS–DVB particles were octadecylated, they
`showed a higher hydrophobicity than the untreated PS–DVB
`monolithic stationary phase. Therefore a higher amount of
`acetonitrile was necessary to elute the components of the test
`mixture from the granular column than from the monolithic
`column.
`
`3.3. Study of molecular diffusion processes within the
`chromatographic bed
`
`Recently, we have evaluated the chromatographic effi-
`ciency of monolithic columns by isocratic elution of an
`◦
`oligonucleotide at 50
`C column temperature. The number
`of theoretical plates exceeded 11 500 plates for a 60 mm col-
`umn, corresponding to 191 000 theoretical plates per meter
`clearly demonstrating the outstanding separation efficiency
`of the monolithic capillary columns [29]. Extending this
`communication we are presenting here the study of molec-
`ular diffusion processes within the chromatographic bed of
`capillary columns. For this purpose we measured the depen-
`dence of the height equivalent to a theoretical plate (HETP)
`from the linear flow rate by injecting the oligonucleotide
`(dT)16 as test substance. The plate height curves for a typ-
`ical monolithic and a typical granular capillary column are
`depicted in Fig. 3. Additionally, these plots were used to de-
`termine the optimum flow rate for the separation of nucleic
`acids by RP–HPIPC.
`For the monolithic column, a minimum plate height
`of 8.6 ␮m was determined at a linear flow velocity of
`0.71 mm/s, which corresponds to a volumetric flow rate of
`0.97 ␮l/min, whereas a minimum plate height of 30.8 ␮m
`was observed for the granular column at a linear flow veloc-
`ity of 0.51 mm/s, which corresponds to a volumetric flow
`rate of 0.59 ␮l/min. Obviously, using optimum flow rate
`conditions the monolithic column showed a 3.6 times better
`separation efficiency than the granular column. However it
`must be considered that for most applications the maximum
`column performance is not required. An increase in plate
`height by 40% on the monolithic column when using a flow
`rate of 2.03 instead of 0.97 ␮l/min is often acceptable, be-
`cause the analysis time is shorter by a factor of more than
`2. In fact, the total retention time of the analyte is 1.95 min
`at a flow rate of 2.03 ␮l/min and 4.23 min at a flow rate
`of 0.97 ␮l/min. But even at high flow rates the monolithic
`
`4
`
`

`

`H. Oberacher et al. / J. Chromatogr. A 1030 (2004) 201–208
`
`205
`
`4.5
`
`4.0
`
`3.5
`
`3.0
`
`b0.5 [s]
`
`2.5
`
`2.0
`
`0
`
`0.5 1.0
`
`1.5
`
`2.0 2.5
`
`3.0
`
`linear flow rate [mm/s]
`
`Fig. 4. Peak widths at half height of (dT)16 eluted in gradient mode from a
`monolithic capillary column at different flow velocities. Column, PS–DVB
`monolith, 60 mm×0.2 mm i.d. mobile phase: (A) 100 mM TEAA, pH 7.0;
`(B) 100 mM TEAA, pH 7.0, 20% acetonitrile; linear gradient 25–60% B
`◦
`in 10.0 min; flow rate, 0.43–2.63 ␮l/min; temperature, 50
`C; detection,
`UV, 254 nm; sample, (dT)12–18, 1.25 ng.
`
`ules were three to four times smaller than the PS–DVB–C18
`particles (compare Fig. 3 in [29]), the improved Eddy dis-
`persion properties of the monolithic column type was rea-
`sonable. The improvement in longitudinal diffusion has to
`be explained by a decrease in the labyrinth factor, in which
`longitudinal diffusion is hindered by the walls of the pores
`present in the monolithic structure.
`At a later stage, the chromatographic performance was
`evaluated at different flow rates during gradient elution. A
`mixture of the seven oligothymidylic acids (dT)12 to (dT)18
`was separated at various flow rates on a monolithic capil-
`lary column using a gradient from 5 to 12% acetonitrile in
`100 mM TEAA in 10 min. Fig. 4 shows the dependence of
`the peak width at half height (b0.5) from the volumetric flow
`rate for (dT)16. A minimum b0.5 of 2.46 s was determined
`at a linear flow velocity of 1.33 mm/s, which corresponds
`to a flow rate of 2.0 ␮l/min. For flow rates from 0.85 to
`2.27 mm/s, the peak width showed little variation and ranged
`from 2.46 to 2.75 s using constant gradient conditions. Thus,
`in gradient elution mode the optimum in separation perfor-
`mance is found at a higher flow rates than in the isocratic
`mode. This finding is favorable since it enables the rapid,
`high-performance separation of biopolymers using high flow
`rates.
`
`3.4. Porosity
`
`Inverse size-exclusion chromatography (ISEC), as first re-
`ported by Halász and Martin [41], was utilized to reveal
`differences in porosity between different column configu-
`rations. The porosities and pore-size distributions of two
`monolithic capillary columns and one granular capillary
`column were determined by using tetrahydrofuran as sol-
`vent and polystyrene (PS) standards of different molecular
`mass. However, due to the fact that it is based on several
`assumptions, this method is not absolutely undisputed, yet
`presents an appropriate way to perform such measurements
`
`granular column
`
`monolithic column
`
`50
`
`40
`
`30
`
`20
`
`10
`
`0
`
`HETP [µm]
`
`0.0
`
`0.5
`
`1.0
`
`1.5
`
`2.0
`
`linear flow rate [mm/s]
`
`Fig. 3. Van Deemter plots for (dT)16 characterizing (䉱) a monolithic and
`(䊏) a granular capillary column. HETP values are not corrected for extra
`column dispersion. Columns, (䉱) PS–DVB monolith, 55 mm × 0.2 mm
`i.d., (䊏) PS–DVB–C18, 2.1 ␮m, 55 mm × 0.2 mm i.d.; mobile phase,
`100 mM TEAA, pH 7.0, 8% acetonitrile; flow rate, 0.17–1.89 mm/s; tem-
`◦
`C; detection, UV, 254 nm; sample, (dT)16, 500 fmol.
`perature, 50
`
`column yielded a 2.5 times lower plate height compared the
`granular column operated at optimum flow rate conditions.
`To study the differences of the molecular diffusion pro-
`cesses in granular and monolithic columns, we used the
`simplified Van Deemter equation [40] for characterizing the
`axial dispersion. To determine the portions of the individ-
`ual band broadening processes to the overall band broad-
`ening within the chromatographic beds of the two column
`types, Van Deemter functions were fitted to the measured
`plate height curves yielding the three parameters A, B, and
`C, which characterize Eddy dispersion, longitudinal diffu-
`sion, and mass transfer, respectively. According to theory,
`a major difference between the mass transfer characteristics
`of monolithic and granular columns was expected, whereas
`only a little change in Eddy dispersion and no variation in
`longitudinal diffusion was anticipated.
`The results of the curve fits, which are summarized in
`Table 2, showed that all three parameters were two to five
`times better on the monolithic column than on the gran-
`ular column. Surprisingly, the mass transfer term showed
`the smallest improvement of all three parameters. We be-
`lieve that due to the micropellicular configuration both of
`particles and monolithic beds, rapid mass transfer is possi-
`ble with both column types and therefore, the difference in
`the C term was relatively small. Since the monolithic nod-
`
`Table 2
`Parameters describing the molecular diffusion processes within a chro-
`matographic bed
`
`Column type
`
`Granular
`Monolithic
`
`Eddy
`dispersion, A
`(␮m)
`15.7
`3.0
`
`Longitudinal
`diffusion, B
`(␮m mm/s)
`3.6
`0.9
`
`Mass transfer, C
`(␮m/(mm/s))
`
`13.5
`6.1
`
`5
`
`

`

`206
`
`H. Oberacher et al. / J. Chromatogr. A 1030 (2004) 201–208
`
`monolithic column (b0.5 = 9.7 s)
`monolithic column (b0.5 = 2.7 s)
`granular column (b0.5 = 3.9 s)
`
`monolithic column (b0.5 = 9.7 s)
`monolithic column (b0.5 = 2.7 s)
`granular column (b0.5 = 3.9 s)
`
`80
`
`70
`
`60
`
`50
`
`40
`
`30
`
`20
`
`1.0
`
`0.8
`
`0.6
`
`0.4
`
`0.2
`
`0.0
`
`(a)
`
`Ve / Vk . 100 [%]
`
`(b)
`
`∆logΦ
`∆Ve
`
`.
`
`Vp
`
`1
`
`1
`
`1000
`
`10
`100
`pore size Φ [nm]
`Fig. 5. Pore size distribution for (×, 䊊) two monolithic and (䉫)
`one granular column determined by inverse size-exclusion chromatogra-
`phy. Columns, (×, 䊊) PS–DVB monolith, 60 mm × 0.2 mm i.d., (䉫)
`PS–DVB–C18, 2.1 ␮m, 60 mm × 0.2 mm i.d.; mobile phase, tetrahydro-
`◦
`furan; flow rate, 1.1 ␮l/min; temperature, 20
`C; detection, UV, 254 nm;
`samples, 20 nl polystyrene standards, 1 mg/ml each.
`
`clearly revealed that the surfaces of both materials exhibited
`surface roughness. The observed high separation efficiency
`of both materials for biopolymers suggested the presence a
`rough surface, which, on the one hand, offer fast mass trans-
`fer properties, and which, on the other hand, increase the
`overall surface area and consequently the loading capacities
`of the columns.
`The total porosity values of the examined monolithic
`columns were very similar and in the range of 70–71%
`(Table 3). This is not surprising since the ratio of monomers
`to porogens is identical for both monoliths tested in this
`analysis. However, these values exceeded the actual amount
`of porogenic solvent added to the polymerization mixture
`(60%, v/v) and reflect the contribution of volume shrink-
`age during polymerization. Although the packed column
`exhibited a much lower total porosity of only 47%, all
`three columns yielded nearly identical values for the internal
`porosity (Vp, 18–22%), which was calculated as the differ-
`ence between the maximum and minimum elution volumes
`of the PS standards.
`A plot of the relative pore size frequency versus the
`pore size (Fig. 5b) was used to determine the average pore
`size of the three stationary phases, which was 55 nm for
`the high-efficiency monolithic column and 89 nm for the
`medium-efficiency monolith. The average pore diameter
`calculated for the granular column was somewhat lower,
`namely 25 nm. Based on the measured average pore size
`values, the specific surface areas of the three columns were
`estimated. The high-efficiency monolithic column displayed
`a specific surface area in the range of 43 m2/g, while for the
`medium-efficiency monolith a value of 32 m2/g was calcu-
`lated. The specific surface for the particle packed-column
`finally was 96 m2/g. This result indicates that the specific
`
`as it works at least under conditions similar to those used in
`actual HPLC separations. Moreover, pore size distributions
`measured with tetrahydrofuran as solvent, which has been
`shown to swell the monolithic column bed (see Section 3.1),
`will certainly differ from the pore structure experienced by
`analytes chromatographed and non-swelling hydro-organic
`elution conditions common for protein, peptide and nucleic
`acid separations.
`To investigate the relationship between separation perfor-
`mance and column porosity, a monolith of very good and
`a monolith of moderate separation performance were char-
`acterized by ISEC and the data obtained were compared to
`those measured with a particle-packed column of a perfor-
`mance slightly lower than that of the good monolithic col-
`umn. For the evaluation of the performance of the three dif-
`ferent columns, the average peak widths at half height b0.5
`of a mixture of (dT)12 to (dT)18, which was separated using
`a gradient from 5 to 12% acetonitrile in 100 mM TEAA in
`10 min, were determined. The monolithic columns exhib-
`ited an average b0.5 of 2.7 s and 9.7 s, respectively, while the
`granular column yielded a value of 3.9 s for this parameter.
`From the theory of SEC we learn that the smaller a
`molecule the more it can penetrate into the pores of a sta-
`tionary phase and the later it will elute from the column.
`Therefore the elution volume (Ve) of small molecules, such
`as toluene, represents the total volume of pores of all sizes
`in the rod, including the large channels. The polystyrene
`standards of a molecular mass between 440 and 6 590 000
`gradually eluted faster, consistent with the smaller accessi-
`ble volume in the column for samples of these sizes. It is
`important to note here that only pores with a diameter rang-
`ing from 2 to 650 nm could be probed with this approach
`and that the presence of pore sizes beyond this range cannot
`be excluded.
`Using the data sets obtained from inverse size-exclusion
`chromatography, the cumulative porosity was calculated as
`the portion of Ve in the total volume of the empty tube (VK).
`Fig. 5a shows a graph of the cumulative porosity against
`the pore size obtained for the three columns. The mono-
`lith that showed a good separation performance featured
`pores with a diameter as small as 5 nm. The upper exclusion
`limit and hence the maximum pore size was in the range
`of 200–300 nm. On the contrary, the monolith that exhib-
`ited poorer separation performance completely lacked pores
`with a diameter lower than 20 nm. The curve indicated that
`the average pore diameter was shifted to higher values. An
`upper exclusion limit could not be determined with ultimate
`certainty, because no polystyrene standards were available
`to probe the very large pores.
`The particle-packed capillary column was expected
`to show little size-exclusion effects due to the use of
`non-porous PS–DVB–C18 particles. However, the cumula-
`tive porosity curve indicates that pores with diameters in
`the range between 5 and 100 nm were present. A visual
`inspection of the surface morphology of the monolithic
`and particulate stationary phases (compare Fig. 3 in [29])
`
`6
`
`

`

`H. Oberacher et al. / J. Chromatogr. A 1030 (2004) 201–208
`
`207
`
`Table 3
`Porosity, average pore diameter and specific surface determined by inverse size-exclusion chromatography for two monolithic and one granular column
`
`Column
`
`Porosity
`
`Porea
`
`Internoduleb
`
`Totalc
`
`Pore diameter (nm)
`
`Specific surface (m2/g)
`
`Monolith 1 (b0.5 = 9.7 s)
`0.222
`0.483
`0.705
`89
`32
`Monolith 2 (b0.5 = 2.7 s)
`0.187
`0.521
`0.709
`55
`43
`Granular column (b0.5 = 3.9 s)
`0.185
`0.285
`0.470
`25
`96
`a εp = Vp/Vc: Vp, pore volume, obtained from the difference in elution volume of totally excluded and totally penetrating sizing standards; Vc, volume
`of the empty separation column.
`b εz = Vz/Vc: Vz, interstitial volume, elution volume of totally excluded sizing standard.
`c εT = εp + εz.
`
`ing capacity could be loaded onto the monolithic columns
`with only a minor increase in the peak widths at half height.
`Moreover, such overloading still yielded peak widths in the
`range of the analytical capacity of granular columns. This
`result is very important in the case of complex sample mix-
`tures that are commonly analyzed in proteomic or genomic
`applications and that tend to overload a column rather easily.
`
`3.6. Evaluation of the efficiency of monolithic columns for
`the separation of peptides and oligonucleotides
`
`As described in earlier publications, monolithic PS–DVB
`columns were successfully employed as separation media
`for peptide, protein, and nucleic acid analysis [29,35–37]. In
`this investigation the efficiencies of 21 monolithic columns
`for the separation of standard oligonucleotide and peptide
`mixtures were investigated. For comparison of column effi-
`ciency for peptide and oligonucleotide analysis, the average
`b0.5 of a mixture of (dT)12 to (dT)18 and of a mixture of
`
`4.5
`
`4
`
`3.5
`
`3
`
`2.5
`
`average b0.5, peptides [s]
`
`2
`
`1.5
`
`3.5
`3
`2.5
`2
`average b0.5, oligonucleotides [s]
`
`4
`
`Fig. 7. Average peak widths at half height of oligonucleotides vs. the
`average peak widths at half height of peptides eluted from 21 monolithic
`capillary columns. Columns, PS–DVB monoliths, 60 mm × 0.2 mm i.d.;
`