Evaluation of HPLC columns A study on surface homogeneity by qov12652

VIEWS: 16 PAGES: 9

									J. Sep. Sci. 2003, 26, 313–321                                                                  Buszewski et al.         313


Bogusław Buszewskia),                Evaluation of HPLC columns: A study on surface
Katarzyna Krupczynskaa),
                   ´
Renata M. Gadzała-                   homogeneity of chemically bonded stationary
Kopciucha),                          phases
Gerhard Rychlickib),
Roman Kaliszanc)                     The aim of the current work is to study the heterogeneity of the adsorbent surface on
                                     the basis of physicochemical investigations and chromatographic tests. A series of
a)
   Department of Environmental       packing materials with octadecyl chains chemically bonded to a silica matrix were
Chemistry and Ecoanalytics,          prepared for this purpose. The surface and structural properties of bare silica and
Faculty of Chemistry, Nicholas       silica-based octadecyl phases were characterized by porosimetry, elemental analy-
Copernicus University, 7
                                     sis, 29Si CP/MAS NMR, etc. The most advanced characterization methods based on
Gagarin St., 87-100 Torun,´
Poland                               adsorption microcalorimetry (heat of wetting) measurements were employed to obtain
b)
   Department of General             information about the heterogeneity and topography of unmodified and modified silica
Chemistry, Faculty of                gel. For the chromatographic study, these phases were evaluated on the basis of the
Chemistry, Nicholas                  retention data under non-aqueous conditions. A test series of solutes with various
Copernicus University, 7             chemical properties, such as pK a values, was used. It was found that heterogeneity
Gagarin St., 87-100 Torun,´          of the packing surface results in low HPLC resolution. Use of a non-aqueous mobile
Poland                               phase (n-hexane) reduces analytical interference by eliminating hydrophobic interac-
c)
   Department of Biopharmacy         tions between alkyl ligands and the analyte.
and Pharmacodynamics,
                           ´
Medical University of Gdansk,        Key Words: Chemically bonded stationary phases; Surface characterization; Reten-
107 J. Haller St., 80-416            tion mechanism; Aqueous conditions; Non-aqueous conditions;
     ´
Gdansk, Poland
                                     Received: July 4, 2002; revised: October 16, 2002; accepted: October 17, 2002




1 Introduction                                                   graphic column packing and the stationary phase with
                                                                 components of mobile phase adsorbed on its surface [10].
The dynamic development of chromatography and related
                                                                 Octadecyl stationary phases contain two kinds of active
techniques has become possible thanks to a wide variety
                                                                 centers: bonded octadecyl ligands and residual silanols.
of chromatographic column packings [1, 2]. Silica is still
                                                                 Stationary phases with ligands having free electron pairs
the most commonly used adsorbent, and its surface is
                                                                 (e. g. alkylamide phases) are also employed in liquid chro-
subject to physical and chemical modifications. A quality




                                                                                                                                Original Paper
                                                                 matography [6, 11]. In such a case, several kinds of cen-
analysis of the retention properties of the stationary phase
                                                                 ters can be distinguished: residual silanols, residual
allows selection of columns appropriate for solving a given
                                                                 amine groups (which are non-blocked because of the
analytical problem. The stationary phase might be pre-
                                                                 steric effects), and hydrophobic ligands including N-acyl-
sented as a uniform layer of chemically bonded organic
                                                                 amine groups with free electron pairs. The solvation of
ligands. Due to steric effects it is impossible to block all
                                                                 ligands by solvent molecules from a hydro-organic mobile
the superficial hydroxyl groups of the silica support by
                                                                 phase probably creates a hydrolytic cushion on the adsor-
molecules of organic modifiers [3 – 6]. Thus, we obtain a
                                                                 bent surface [11].
packing with organic ligands located non-homoge-
neously. The remaining silanes can also react with the sol-
                                                                 n-Hexane was used in this study in order to avoid disad-
vent molecules by specific and non-specific interac-
                                                                 vantageous interactions between alkyl ligands and the
tions [7]. In the case of hydro-organic eluents, silica
                                                                 analyte, and to avoid the formation of a hydrolytic cushion
groups are strongly hydrated by water molecules and that
                                                                 by the support surface. A decrease in the mobile phase
is why retention encompasses both hydrophobic and sol-
                                                                 polarity (non-aqueous conditions) brings about changes
vophobic interactions [8, 9]. Hence, in liquid chromatogra-
                                                                 in the alkyl chain conformation. Relatively high mobility of
phy a distinction should be made between the chromato-
                                                                 chains enables the mobile phase components to pene-
Correspondence: Bogusław Buszewski, Department of En-            trate between them. The non-aqueous mobile phase
vironmental Chemistry and Ecoanalytics, Faculty of Chemis-       reduces the solvation of alkyl ligands and the occurrence
try, Nicholas Copernicus University, 7 Gagarin St., 87-100       of hydrophobic-dispersion interactions between chains.
     ´
Torun, Poland. Phone: +48 56 6114308.                            These interactions enhance the effect of steric interac-
Fax: +48 56 6114837. E-mail: bbusz@chem.uni.torun.pl.            tions on the chromatographic process [9].

i    2003 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim                               1615-9306/2003/0303–0313$17.50+.50/0
314        Buszewski et al.                                                                      J. Sep. Sci. 2003, 26, 313–321

The separation process in high performance liquid chro-           Table 1. Physicochemical characteristics of bare Kromasilm
matography (HPLC) is based on specific and non-specific           100 (AT 0191).
interactions analyte E eluent E stationary phase.
                             e          e
                                                                  Parameter                   Abbreviation    Units       Value
Which type of interactions (adsorption, partition, ion
exchange, steric exclusion) will predominate depends on           Particle shape                  –             –      Spherical
the structural properties of the analyte, the composition of      Mean particle size              dp           lm          5
the mobile phase, and the dynamics of chemically bonded           Specific surface area          SBET         m2/g       295
phases [1 – 3, 6, 12 – 16]. The stationary phase selectivity      Pore volume                     Vp          cm3/g      0.92
as well as chromatographic data acquisition depend on: (i)        Mean pore diameter              D             A        113
the chemical nature of bonded ligands, (ii) coverage den-         Concentration of OH            aOH         lmol/m2      7.1
                                                                  groups
sity of the surface, (iii) the subsurface structure of a chemi-
                                                                  Trace amount of metals          CM          ppm         a 20
cally bonded film, (iv) the properties of the support surface
of the chemically bonded phase [6, 10, 13 – 19]. Despite a
large number of chromatographic analyses, many aspects
connected with the mechanism of separation have not yet           2.2 Equipment
been investigated, especially those related to the proper-        The porosimetry parameters such as the specific surface
ties of chromatographic packings. Hence, this paper               area, pore volume, and pore diameter of bare silica (Kro-
focuses on a study of homogeneity of surface coverage             masilm 100 AT 0191; Akzo Nobel, Bohus, Sweden) were
density of chemically bonded stationary phases. A series          determined by the low-temperature nitrogen adsorption –
of different home-made columns were tested for this pur-          desorption method using an ASAP 2010 MicroPore Sys-
pose. Compounds with various chemical properties (pK a)           tem instrument (Carbo Elba, Milan, Italy).
served as test analytes. A whole group of physicochem-
                                                                  The degree of coverage of silica support with bonded
ical techniques such as: porosimetry, elemental analysis,
                                                                  ligands in the home-made packings was calculated from
nuclear magnetic resonance spectroscopy, infrared spec-
                                                                  the carbon content (PC) determined by elemental analy-
troscopy, differential scanning calorimetry, etc. were used
                                                                  sis [16] with a CHN analyzer Model 240 (Perkin-Elmer,
to characterize chromatographic packings [20 – 24]. The
                                                                  Norwalk, CT, USA) (Table 2).
determination of the structure and physicochemical char-
acter of adsorbents allowed prediction of the pattern of          The 29Si solid-state NMR experiments were performed on
analyte molecule behavior during the chromatographic              an ASX Bruker spectrometer, model 300 (Rheinstetten,
process, and provided some information about its quality.         Germany) in the magic-angle spinning (MAS) mode.
Observations of phase changes and heats of wetting by             Details of these measurements were described in
organic phases during microcalorimetric measurements              Ref. [27].
made it possible to characterize the surface before and           Microcalorimetry measurements were made using a
after chemical modification [25, 26].                             home-made microcalorimeter of the Tiana-Calve isother-
                                                                  mal type. The temperature of measurement was 37 l 18C.
                                                                  Home-made stationary phases were packed into
2 Experimental                                                    125 mm64.6 mm ID stainless steel columns. The col-


2.1 Reagents and materials                                        Table 2. Surface characterization of modified silica.

The solid support of home-made phases was Kromasilm               Packing           Type of       PC [%]     Coverage Percent
100 AT 0191 (Akzo Nobel, Bohus, Sweden). Table 1                                    phase                     density of suface
shows the physicochemical characteristics of bare silica                                                      [lm/m2] coverage
gel. All the parameters are close to those expected for the         L
                                                                  MC18             monomer        10.66        1.75        23.3
“theoretically optimal” adsorbent [27, 28].                         L
                                                                  MC18 + EC        monomer        12.68        2.13        28.5
The following reagents were used for chemical modifica-           MCH18            monomer        17.10        3.10        41.5
tion of the silica support material: octadecyldimethyl-           MCH + EC
                                                                     18            monomer        17.45        3.17        42.4
                                                                    L
chlorosilane (Johnson Matthey ALFA Products, Karls-               DC18             polymer        16.02        3.27        43.8
                                                                    L
                                                                  DC18 + EC        polymer        16.84        3.40        45.5
ruhe, Germany), octadecylmethyldichlorosilane (Petrarch
                                                                  DCH
                                                                    18             polymer        17.35        3.64        48.7
Systems Inc. Levittown, Pennsylvania, USA), trimethyl-            DCH + EC
                                                                    18             polymer        17.38        3.71        49.6
chlorosilane (Sigma-Aldrich Chemie, Steinheim, Ger-
many), morpholine (Reachim, Moscow, Russia). Organic              where EC = end-capping, H = high coverage density, L = low
solvents were of HPLC grade (POCh, Gliwice, Poland).              coverage density, M = monomer phases, D = polymeric
The test solutes were of various origins.                         phases.
J. Sep. Sci. 2003, 26, 313–321                                                      Evaluation of HPLC columns           315

umns were packed under a pressure of 50 MPa a using             3.2 CP/MAS NMR investigations
home-made equipment based on a DSF-122 packing
                                                                CP/MAS NMR spectroscopy, especially the spectrum for
pump (Haskel Inc., Burbank, CA, USA).                           29
                                                                   Si, gives quantitative information about the density of
Chromatographic measurements were made using:                   silica gel coverage with ligands. For bare silica three char-
                                                                acteristic signals correspond to geminal (Q2) d = – 91 ppm
– an HP 1050 liquid chromatograph system (Hewlett
                                                                and free/bonded (Q3) d = – 100 ppm silanol groups and
  Packard, Waldbronn, Germany), equipped with a UV-
                                                                oxosilanes (Q4) d = – 108 ppm [29, 30]. An analysis of 29Si
  Vis detector and a HP ChemStation-2 for data collec-
                                                                CP/MAS NMR spectra for the materials from Table 2, pre-
  tion and control of the process;
                                                                sented in Figure 1a, b, c, d, shows that along with chemi-
– a Shimadzu liquid chromatograph system (Shimadzu,             cal modification of active centers (silanols) the intensity of
  Kyoto, Japan) equipped with binary pump (LC-                  signals of particular Q2 and Q3 bands decreases, whereas
  10ADVP) and diode array detector (SPD-M10AVP),                the intensity of signal Q4 increases (Figure 1.a, b). Band
  controlled by Class VP 5.0 software.                          M corresponds with the monomeric structure of the chemi-
                                                                cally bonded phase (one-point bonding ligand-support)
Solutes were injected using a Rheodyne Model 7125 (Ber-
                                                                and appears in the spectrum within a chemical shift
keley, CA, USA) sampling valve with a 20 lL sample loop.
                                                                d = 12.5 – 13.0 ppm. It is interesting that the intensity of
                                                                signals Q2 and Q3 decrease with an increase in the inten-
                                                                sity of signal M. Disadvantageous geminal groups (Q2)
3 Results and discussion                                        are not completely reduced for thin stationary phases
                                                                                                      L
                                                                (with low coverage density) (MC18). Even end-capping
3.1 Surface characterization of stationary phases               does not cause complete blocking of these centers. This
A series of packings was produced by a synthetic                means that some groups of geminal silanols are still
approach (in our laboratory). Their surface was subjected       unblocked, which can have a negative influence on the
to a thorough physicochemical assessment. The pack-             separation of polar analytes (especially on basic com-
ings, prepared in a substitution reaction with mono- and        pounds and irreversible sorption). The presence of Q2
difunctional octadecylchlorosilanes, permitted separa-          groups can also affect the penetration of quite big mole-
tions of different quality, depending on the reagent con-       cules (e. g biopolymers). Their conformation changes in
centration (Table 2). In order to avoid any detrimental         proportion to alkyl silica groups on the packing surface. In
effect of non-blocked silanols, secondary silanization          the case of thin coverage, 28% of silanols are effectively
(end-capping) by trimethylchlorosilane was perfor-              blocked by alkylsilyl ligands. As a result, penetration of
med (TMS). It should be noted that, in the exchange reac-       quite small trimethylochlorosilane molecules to residual
tion between the silica surface and the analyte, a predomi-     silanols is possible. This effect is not observed for densely
nant factor (determining homogeneity and ligand location        covered materials (Figure 1.a, b) [6, 17, 21, 22, 28 – 30].
on the surface) is the steric effect.                           The spectra for packings obtained using difunctional
Our investigations show that it is easier to control the for-   silane are presented in Figure 1.c, d. Figure 1.c shows
mation of chemically bonded phases in the case of mono-         that geminal groups are not blocked completely. However,
meric structures. A 30% decrease in the reaction sub-           end-capping effectively reduces their contribution. Bands
strates causes an increase in surface coverage density
                                                                                        L
                                                                (Figure 1.c for DC18) corresponding to one-point (D1;
(from thin to dense) by 45%. According to the results, we       d = –2.5 ppm and/or d = – 6 ppm) and multiple bonding
can expect a uniform arrangement of free silanol groups         (D2; d = – 10 ppm, D3; d = – 14 ppm, D4; d = – 16 ppm) to
on the support surface. This is also confirmed by small dif-    the silica support are observed on the spectrum and show
ferences in the carbon percentage and coverage density          small differences in the values of chemical shifts (signal
for packings after end-capping. The situation is more com-      D1 – 3) [6, 22, 23, 28 – 30].
plex in the case of polymeric packings, whose formation
can lead to different structures of chemically bonded
                                                                3.3 Microcalorimetric study
phases, i. e. single or double bonding between the station-
ary phase and the ligand. The differences between thin          Microcalorimetry was used to measure the heats of wet-
and dense coverage, despite control over modifier contri-       ting for n-hexane and methanol. The wettability (solvation)
bution in the reaction, are inconsiderable and concern          of non-modified and chemically modified materials may
about 5% of the surface coverage. Smaller differences           be indirectly determined by such measurements. It can
are obtained for packings after end-capping. A film formed      help to define the conformation of chemically bonded
on the surface prevents blocking of residual silanols,          phases and to predict the possibility of solvent penetration
which come from the support and modifier hydroxyl group.        through a film of stationary phase. The penetration and
The steric effect can again be observed.                        solvation depend on the organization of brush-type
316       Buszewski et al.                                                                 J. Sep. Sci. 2003, 26, 313–321




      Figure 1. 29Si CP/MAS NMR spectra for the bonded C18 packing.

ligands. The heats of wetting (q) listed in Table 3 show        cases, an increase (on average by ca. 8%) in this value
that bare silica has the greatest energy value. Along with      can be noted, which may mean that when coverage den-
an increase in the carbon content on the surface (for           sity increases relatively long n-hexane chains better pene-
monomeric materials only) after end-capping with thin           trate into the surface and wet it (hydrophobic interactions
                                                                                                                L         L
coverage, a reduction in the heat of wetting is observed for    chain chain). The differences observed for MC18 and MC18
n-hexane (qSG = 2.413 cal/g, Table 3). In the remaining
               H
                                                                + EC are surely a result of better flexibility of chains in
J. Sep. Sci. 2003, 26, 313–321                                                         Evaluation of HPLC columns            317


Table 3. Heats of wetting of n-hexane and methanol on sta-       Table 4. Analytes used in chromatographic investigations.
tionary phases with different coverage density.
                                                                 Analyte            Functional      Molecular   l2 a)    log P b)
Packing                 Type of            q a)[cal/g]                              group            mass        [D]
                        phase
                                    n-hexane       methanol      Benzene            C6H6               78.11     0        2.13
                                                                 Toluene            1H3                92.14    0.37      2.73
Silica gel                 –          2.413           10.9       Aniline            1NH2               93.13    1.27      0.90
    L
MC18                    monomer       2.375          6.418       Phenol             1OH                94.11    1.22      1.47
    L
MC18 + EC                             2.170          2.856       Ethylbenzene       1C2H5             106.11    0.35      3.13
MCH  18                               2.088          3.754       Chlorobenzene      1Cl               112.56    1.69      2.89
MCH + EC
     18                               2.244          2.901       Propylbenzene      1C3H7             120.11    0.33      3.98
    L
DC18                    polymer       1.559          4.284       Nitrobenzene       1NO2              123.11    4.23      1.85
    L
DC18 + EC                             2.068          2.377       Butylbenzene       1C4H9             134.11    0.31      4.28
DCH 18                                1.776          2.103
DCH + EC
    18                                1.988          2.092       a)
                                                                      Dipole moment.
                                                                 b)
                                                                      Log P – n-octanol – water partition coefficient.
a)
     Heat of wetting.


changing the conformation (for both the mobile and the           Silanols play a significant role in chromatographic elution,
stationary phase). In this case, more trimethylsilyl groups      depending on their arrangement on the chemically
go inside and effectively block residual silanols (Table 2),     bonded stationary phase. Assuming that the support used
compared with the other stationary phases, where cav-            in the experiment is characterized by high purity (a low
ities prevent more effective chain mobility. In all cases, the   level of pollution by heteroatoms), a dominant influence of
well-known principle that like dissolves like (hydrocar-         silanols in the elution process could be expected. Taking
bon – hydrocarbon) can be observed. Due to long chains,          into consideration the fact that a more ordered structure is
bonded ligands show a preferred brush-type structure (a          obtained when n-hexane is used as a mobile phase, only
more ordered stationary phase) [6].                              an adsorptive separation mechanism is predicted.
The opposite situation is encountered with methanol as           Figure 2 presents the results of the so called methyl
wetting medium, where residual silanols predominate              selectivity of coverage and structures formed on the sup-
(from the perspective of unit processes).They interact           port surface [13, 16]. Selectivity (S) is a measure of differ-
with methanol via hydrogen bonds. Thus, high values of           ences in the retention of analyte containing various func-
heats of wetting for bare silica (q M = 10.9 cal/g) are
                                        SG                       tional groups, in relation to benzene.
obtained and a drastic decrease (on average by ca. 15%)
in the values measured is observed, depending on the
coverage density and structure of chemically bonded              S = ln a = ln k compound – ln k benzene                       (1)
phases (Table 3). A different orientation of bonded ligands
should be expected as a result of preferential solvation of      where S = selectivity and k = retention factor.
the residual silanols by methanol molecules. The following
regularity was observed for all materials: an increase in        The effect of end-capping can be noted in all of the cases
the coverage density of the stationary phase causes a            examined because both the selectivity of the phases pre-
decrease in heats of wetting. This means that penetration        pared and their potential for analytical applications are
of methanol molecules through a dense film is difficult and      improved. It is interesting that the biggest difference in
residual silanols are effectively screened by long chains of
the stationary phase. This is confirmed by the data
obtained for dense coverage (packings with H index,
Table 3). End-capping does not significantly influence
changes in heats of wetting; on the contrary, adsorption of
methanol molecules on the stationary phase takes place
by means of subtle interactions.



3.4 Chromatographic measurements
Test analytes of various chemical characters were used
for evaluating the influence of stationary phase surface         Figure 2. Correlation between selectivity vs. ln k. Mobile
homogeneity on the chromatographic process (Table 4).            phase: n-hexane, flow rate: 0.7 mL/min.
318   Buszewski et al.                                                                  J. Sep. Sci. 2003, 26, 313–321




           Figure 3. Correlation of Hansch parameters vs. s values (mobile phase: n-hexane, flow rate:
           0.7 mL/min).
J. Sep. Sci. 2003, 26, 313–321                                                     Evaluation of HPLC columns          319

selectivity is observed for thin monomeric adsorbents
                                       L           L
before and after end-capping (MC18 and MC18 + EC),
which was confirmed by 29Si CP/MAS NMR and microca-
lorimetric investigations. Practically, no differences were
observed for polymeric phases (Figure 2). The smallest
differences in retention were found for dense monomeric
             L
phases (MC18 and MCH + EC). This means that in the pre-
                       18
sent case the most homogeneous arrangement of ligands
on the silica surface was obtained [26, 28]. This obser-
vation is in agreement with the microcalorimetric data
(Table 3).
This relationship is better (quantitatively) presented by the
QSRR (Quantitative Structure-Retention Relationships)
method, employed for interpreting retention taking into
account characteristic descriptors for selected ana-
lytes [19, 20]. The theoretical background for further
examination of the molecular mechanism of RP HPLC
separation is Abraham’s equation [Eq. (2)] [36, 37].
                              H       H       H
      log k w = c + rR 2 + s p2 + aSa 2 + bSb 2 + vVx     (2)
             H    H    H
where R 2, p 2 , a2 , b2 , and Vx are solute descriptors. The
constants c, r, s ,a , b reflect the corresponding properties
of the HPLC system considered.
This method provides very good results regarding not only
the separation process, but also an evaluation of column
quality [12, 39]. However, it is complex and time-consum-
ing. Similar information can be obtained by the Hansch
correlations [33], where a functional group contribution (s
values) together with Hansch substituent constants (p) for
a series monosubstituted benzene derivatives are applied
to understand the mechanism of analyte retention on an
octadecyl stationary phases.
The contribution of functional groups was determined
according to Tomlinson [31] and Smith [32], using the
parameter s x. This parameter was calculated on the basis
of differences in retention of two compounds with different
functional groups:

                 sx = log k R-X – log k R-H               (3)

Values sx relate to Hansch partition coefficients [33]
[Eq. (4)], whose theoretical explanation is based on the        Figure 4. Correlation between ln a vs. number of carbon
solvophobic theory [18]:                                        atoms in substituent for alkylbenzenes mixture (stationary
                                                                phase: monomer (a) and polymer (b), mobile phase: n-hex-
                 px = log P R-X – log P R-H               (4)   ane, flow rate: 0.7 mL/min; stationary phase: monomer (c)
                                                                and polymer (d), mobile phase: MeOH/H2O v/v 70/30, flow
where X = substituent in an aromatic ring, R1H = non-           rate: 1 mL/min).
substituted compound, R1X = substituted compound.
The relationship between retention (s) and the Hansch           An investigation of hydrophobic interactions (chain chain)
parameters (Figure 3) shows that aniline is irreversibly        is a very important aspect of an evaluation of chromato-
adsorbed on the packings after the first modification step      graphic packings. In this study the lowest values of selec-
(no signal). A quality effect is observed for the materials     tivity were obtained for dense monomeric adsorbents.
after end-capping, where blocked silanols do not play an        This can mean that bonded chains are uniformly arranged
active part in the elution process. This is in agreement        on the support surface and perpendicularly oriented to
with literature data [34, 35].                                  this surface (Figure 4.a,b). In the case of polymeric
320       Buszewski et al.                                                                 J. Sep. Sci. 2003, 26, 313–321




         Figure 5. Correlation between Hansch parameters vs. s values (mobile phase: MeOH/H2O v/v 70/30, flow
         rate: 1 mL/min).


adsorbents selectivity is low, which means that the sur-        governed by hydrophobicity and that the principal reten-
face examined is well screened [27, 28].                        tion mechanism is a partition process [32, 38].

The same chromatographic investigations were con-
ducted under hydro-organic conditions (MeOH/H2O v/v
70/30). In this system a study of chain E chain interac-
                                          e                     4 Conclusions
tions is also very important. Figure 4.c and Figure 4.d
show that the smallest differences in selectivity can be        Chemical modification of the silica surface by mono- and
observed for polymeric packings with high coverage den-         difunctional silanes allows preparation of well-defined
sity (DCH and DCH + EC). This means that organic chains
         18        18
                                                                packings. It is possible to control the process of chemical
are uniformly arranged on the surface of this adsorbent.        modification by changing the composition of the reaction
                                                                mixture. This is easier to accomplish in the case of mono-
A comparison of mean s values obtained on the basis of          meric phases. 29Si CP/MAS NMR spectra, as a source of
the capacity factors determined for binary hydroorganic         information about adsorbents, enable the characterization
(water-methanol) mobile phase compositions with the             of the surface design of new phases, as well as control
Hansch constants indicates a very good correlation (Fig-        over the modification process. Measurements of heats of
ure 5). The highest correlation, observed for the MCH +  18     wetting make it possible to predict molecule penetration
EC packing (r = 0.9896), suggests that it is possible to pre-   through chemically bonded ligands. Chromatographic
dict retention on the chemically bonded phases tested. A        investigations under non-aqueous conditions confirm an
good correlation achieved for the packings studied indi-        ordered arrangement of the underlying surface structure
cates that retention on these RP HPLC phases is primarily       of a chemically bonded film. The biggest differences in
J. Sep. Sci. 2003, 26, 313–321                                                      Evaluation of HPLC columns             321

selectivity are obtained for thin monomeric phases (before      [19] R. Kaliszan, Quantitative Structure – Retention Rela-
and after end-capping). The smallest differences in reten-           tionships and Chromatographic Determination of Hydro-
                                                                     phobicity in: Handbook of Advanced Materials Testing
tion noted for dense monomeric phases suggest their
                                                                     (N.P. Cheremisinoff, P.N. Cheremisinoff, eds.), Marcel
homogeneity. The relationships between retention and                 Dekker Inc., New York 1995.
the Hansch parameters allow interesting conclusions to
                                                                [20] U.D. Neue, HPLC Columns. Theory, technology and
be drawn. After end-capping blocked silanols do not play             practice. Wiley-VCH, New York 1997.
an active part in the elution process.
                                                                [21] E.F Vansant, P. van der Voort, K.C. Vrancken, Charac-
                                                                     terization and chemical modification of the silica surface.
                                                                     Elsevier, New York 1995.
References                                                      [22] E. Bayer, A. Paulus, B. Peters, G. Laupp, J. Reiners, K.
                                                                     Albert, J. Chromatogr. 1986, 364, 25 – 34.
 [1] L.C. Sander, S.A. Wise, CRC Crit. Rev. Anal. Chem.         [23] H.A.M. Verhulst, L.J.M. van de Ven, J.W. de Haan, H.A.
     1987, 18, 299 – 315.                                            Claessens, F. Eisenbeiss, C.A. Cramers, J. Chromatogr
 [2] K.K. Unger, Packings and Stationary Phases in Chroma-           A 1994, 687, 213 – 223.
     tographic Techniques. Marcel Dekker, New York 1990.        [24] M. Pursch, L.C. Sander, K. Albert, Anal. Chem. 1996,
 [3] R.K. Gilpin, J. Chromatogr. Sci. 1984, 22, 371 – 379.           68, 4107 – 4114.
 [4] K. Karch, I. Sebastian, I. Halasz, J. Chromatogr. 1976,    [25] B. Buszewski, Chemia Stosowana 1988, 32, 203 – 211.
     122, 3 – 12.
                                                                [26] P. Staszczuk, B. Buszewski, Chromatographia 1988,
 [5] G.E. Berendsen, L. de Galan, J. Liq. Chromatogr. 1978,          25, 881 – 894.
     1, 561 – 569.
                                                                [27] B. Buszewski, M. Jezierska, B. Ostrowska-Gumkowska,
 [6] B. Buszewski, Preparation, Properties and Application           Mater. Chem. Phys. 2001, 72, 30 – 41.
     of Chemically Bonded Phase in Chromatographic Analy-
                                                                [28] B. Buszewski, M. Jezierska, M. Welniak, D. Berek, J.
     sis, D.Sc. Thesis, Slovak Technical University, Brati-
                                                                     High Resol. Chromatogr. 1998, 21, 267 – 271.
     slava, 1992.
 [7] S.G. Weber, W.G. Tramposch, Anal. Chem. 1983, 55,          [29] G.E. Maciel, D.W. Maciel, D.W. Sindorf, J. Am. Chem.
     1771 – 1776.                                                    Soc. 1982, 102, 7606 – 7611.

 [8] A. Nahum, Cs. Horvath, J. Chromatogr. 1981, 203, 53 –      [30] G.E. Maciel, D.W. Maciel, D.W. Sindorf, J. Am. Chem.
     61.                                                             Soc. 1983, 105, 1487, 1848, 3767.

 [9] R.N. Nikolov, J. Chromatogr. 1984, 286, 147 – 154.         [31] E. Tomlinson, H. Poppe, J.C. Kraak, Int. J. Pharm. 1981,
                                                                     7, 225 – 230.
[10] L. Rohrschneider, J. Sep. Sci. 2001, 24, 3 – 9.
                                                                [32] R.M. Smith, J. Chromatogr. A 1993, 656, 381 – 385.
[11] B. Buszewski, R.M. Gadzała-Kopciuch, M. Jaroniec, J.
     Liq. Chromatogr. Rel. Technol. 1997, 20, 2313 – 2325.      [33] C. Hansch, A. Leo, Substituent Constants for Correlation
                                                                     Analysis in Chemistry and Biology, Wiley, New York,
[12] R. Kaliszan, M. van Straten, M. Markuszewski, C.C. Cra-         1979.
     mers, H.A Claessens, J. Chromatogr. A 1999, 835,
     455 – 486.                                                 [34] B. Buszewski, Chromatographia 1989, 28, 574 – 578.
[13] M. Jaroniec, J. Chromatogr. 1993, 656, 37 – 45.            [35] B. Buszewski, R.M. Gadzała-Kopciuch, M. Markus-
                                                                     zewski, R. Kaliszan, Anal. Chem. 1997, 69, 3277 – 3284.
[14] J.G. Dorsey, K.A. Dill, Chem. Rev. 1989, 89, 331 – 338.
                                                                [36] M.H. Abraham, University College London Data Base,
[15] C.F. Poole, S.K. Poole, Chromatography Today. Else-             1996.
     vier, Amsterdam 1991.
                                                                [37] M.H. Abraham, Chem. Soc. Rev. 1993, 22, 73 – 85.
[16] B. Buszewski, M. Jaroniec, R.K. Gilpin, J. Chromatogr.
     A 1994, 668, 293 – 304.                                    [38] B. Buszewski, M. Jezierska-Switala, R. Kaliszan, A.
                                                                     Wojtczak, K. Albert, S. Bachmann, M.T. Matyska, J.
[17] B. Buszewski, Z. Suprynowicz, P. Staszczuk, K. Albert,          Pesek, Chromatographia Suppl. 2001, 53, 204 – 212.
     B. Pfleiderer, E. Bayer, J. Chromatogr. 1990, 499, 305 –
     312.                                                       [39] M.A. Al.-Haj, R. Kaliszan, B. Buszewski, J. Chromatogr.
                                                                     Sci. 2001, 39, 29 – 37.
[18] R. Kaliszan, Quantitative Structure – Chromatographic
     Retention Relationships, Wiley & Sons, New York 1987.                                                         [JSS 1434]

								
To top