Iniferter-mediated grafting of molecularly imprinted polymers on porous silica beads for the enantiomeric resolution of drugs

Article abstract of DOI:10.1002/jmr.2443

A surface-imprinted chiral stationary phase for the enantiomeric resolution of the antidepressant drug, citalopram, is presented in this work. N, N′-diethylaminodithiocarbamoylpropyl(trimethoxy)silane has been used as silane iniferter for the surface functionalization of the solid silica support. A molecularly imprinted polymer thin film, in the nm scale, was then grafted on the silanized silica using itaconic acid as the functional monomer and ethylene glycol dimethacrylate as the cross-linker in the presence of the template S-citalopram. The total monomer amount was calculated to obtain the desired thickness. Non-imprinted stationary phases were prepared similarly in the absence of S-citalopram. Characterization of the materials was carried out by scanning electron microscopy, thermogravimetric analysis, elemental analysis and Fourier transform infrared spectroscopy. Stationary phases have been applied to the chromatographic separation of the target. Conditions for best chromatographic resolution of the enantiomers were optimized, and it was found that a mobile phase consisting of a mixture of formate buffer (40 mM, pH 3) and acetonitrile (30:70 v/v) at 40 °C provided best results. Binding behaviour of the developed material was finally assessed by batch rebinding experiments. The obtained binding isotherm was fitted to different binding models being the Freundlich-Langmuir model, the one that best fitted the experimental data. The developed material has shown high selectivity for the target enantiomer, and the stationary phase could be undoubtedly exploited for chiral separation of the drug.

Full text of DOI:10.1002/jmr.2443

Special issue article
Received: 2 June 2014,
Revised: 23 September 2014,
Accepted: 27 October 2014,
Published online in Wiley Online Library: 16 February 2015
(wileyonlinelibrary.com) DOI: 10.1002/jmr.2443
Iniferter-mediated grafting of molecularly
imprinted polymers on porous silica beads for
†
the enantiomeric resolution of drugs
a
a
b
Raquel Gutiérrez-Climente , Alberto Gómez-Caballero , Mahadeo Halhalli ,
Börje Sellergren , M. Aránzazu Goicolea and Ramón J. Barrio *
c
a
a
A surface-imprinted chiral stationary phase for the enantiomeric resolution of the antidepressant drug, citalopram, is
presented in this work. N, N′-diethylaminodithiocarbamoylpropyl(trimethoxy)silane has been used as silane iniferter
for the surface functionalization of the solid silica support. A molecularly imprinted polymer thin film, in the nm scale,
was then grafted on the silanized silica using itaconic acid as the functional monomer and ethylene glycol dimethacrylate
as the cross-linker in the presence of the template S-citalopram. The total monomer amount was calculated to obtain the
desired thickness. Non-imprinted stationary phases were prepared similarly in the absence of S-citalopram. Characteriza-
tion of the materials was carried out by scanning electron microscopy, thermogravimetric analysis, elemental analysis
and Fourier transform infrared spectroscopy. Stationary phases have been applied to the chromatographic separation
of the target. Conditions for best chromatographic resolution of the enantiomers were optimized, and it was found that
a mobile phase consisting of a mixture of formate buffer (40mM, pH 3) and acetonitrile (30:70 v/v) at 40 °C provided best
results. Binding behaviour of the developed material was finally assessed by batch rebinding experiments. The obtained
binding isotherm was fitted to different binding models being the Freundlich–Langmuir model, the one that best fitted
the experimental data. The developed material has shown high selectivity for the target enantiomer, and the stationary
phase could be undoubtedly exploited for chiral separation of the drug. Copyright © 2015 John Wiley & Sons, Ltd.
Additional supporting information may be found in the online version of this article at the publisher’s web site.
Keywords: iniferter; MIP; citalopram; chiral stationary phase
is time consuming and often leads to highly irregular, variable
dimensions and poor yield of useful MIP particles (Baggiani, 2005).
INTRODUCTION
Molecular imprinting is a technique where a tailor-made polymer
is synthesized under the guidance of a template molecule. The
presence of this molecule during the polymer formation, in
combination with functional and cross-linking monomers, leads
to the generation of molecular imprints with specific recognition
capability. The subsequent removal of the template is crucial to
leave the imprinted sites free, which are chemically and sterically
complementary in shape, size and functional groups to the
template itself or structural analogues (Piletsky and Turner, 2006).
Molecularly imprinted polymers (MIPs) have been employed
over the last years (Whitcombe et al., 2014) as a useful tool for dif-
ferent analytical applications including chromatography (Sulitzky
et al., 2002; Chen et al., 2012), capillary electrochromatography
In order to have a better control of MIP particle morphology and
to provide materials having improved chromatographic characteris-
tics, different strategies such as suspension polymerization (Kempe
and Kempe, 2004), multi-swelling polymerization (Haginaka and
Kagawa, 2002), precipitation polymerization (Wang et al., 2003)
and imprinting core-shell particles (Perez-Moral and Mayes, 2004)
have been employed. However, the conditions required to generate
high-affinity binding sites are in most cases incompatible with the
conditions needed to obtain the desired particle characteristics
*
Correspondence to: Ramón J. Barrio, Department of Analytical Chemistry,
Faculty of Pharmacy, University of the Basque Country, 01006 Vitoria-Gasteiz
(Álava), Spain.
E-mail: r.barrio@ehu.es
(Turiel and Martin-Esteban, 2005), sensor technology (Piletsky
et al., 2012) or even in sample treatment procedures such as
solid-phase extraction (Sellergren and Esteban, 2010), solid phase
microextraction (Prasad et al., 2008) or stir bar sorptive extraction
†
This article is published in Journal of Molecular Recognition as part of the Special
Issue ‘MIP2014: The Eighth International Symposium on Molecular Imprinting’.
a R. Gutiérrez-Climente, A. Gómez-Caballero, M. A. Goicolea, R. J. Barrio
Department of Analytical Chemistry, Faculty of Pharmacy, University of the
Basque Country, Vitoria-Gasteiz (Álava), Spain
(Gomez-Caballero et al., 2011). In what concerns chromatography,
MIP-based stationary phases, traditionally, have been almost
exclusively synthesized by bulk polymerization, obtaining irregular
particles because of grinding and sieving processes after radical
polymerization. This approach has been widely employed for
imprinting, not only because of its economic viability, but also
because of the fact that resulting polymers are very stable and are
able to withstand high pressures without cracking or collapsing.
However, this procedure is far long from ideal because the process
b M. Halhalli
INFU, Faculty of Chemistry, Technical University of Dortmund, Dortmund,
Germany
c B. Sellergren
Department of Biomedical Sciences, Faculty of Health and Society, Malmö
University, Malmö, Sweden
J. Mol. Recognit. 2016; 29: 106–114
Copyright © 2015 John Wiley & Sons, Ltd.

INIFERTER-MEDIATED GRAFTING OF MIP ON POROUS SILICA BEADS
(
size, porosity, pore volume and surface area) (Sulitzky et al., 2002).
In this work, the ‘grafting from’ approach has been employed to
develop a surface-imprinted chiral stationary phase for the chro-
matographic separation of the enantiomers of the antidepressant
drug, citalopram (Unceta et al., 2011). The surface-imprinted
stationary phase was synthesized via initiator, transfer agent,
terminator (iniferter) technique (Tamayo et al., 2005; Barahona
et al., 2010), what led to the formation of a chiral stationary phase
(CSP) based on a MIP. This application of MIP as CSP provides
benefits, such as selectivity and elution predictability, which are
not often found in commercial CSPs (Ansell et al., 2012) where
organic compounds with chiral functionalities are immobilized.
To overcome these problems, surface-imprinting technique (Lu
et al., 2007) emerged. Surface imprinting can be carried out by
different types of reversible-deactivation radical polymerizations
such as iniferter polymerization (Tamayo et al., 2005; Su et al.,
2
008; Barahona et al., 2010), atom transfer radical polymerization
and reversible addition fragmentation chain transfer polymerization.
End-functionalized monomers or polymeric chains can be
inserted on the surface of solid supports by ‘grafting to’ or ‘grafting
from’ approaches. These grafting methods allow the growth of
MIPs on preformed support materials of known morphology
(Arnold et al., 1995; Piletsky et al., 2000). Traditionally, the
preparation of bonded stationary phases based on silica has been
carried out almost exclusively by the ‘grafting to’ approach. Despite
its potential benefits improving kinetic characteristics and the sim-
plicity of the procedure, the synthesis of imprinted thin layers is
not a simple goal because of the difficulties in controlling film
thickness (Giovannoli et al., 2012) because of the presence of an
initiator in solution and the molecular weight of the polymers.
Moreover, grafting densities are limited because of kinetic and
steric factors (Edmondson et al., 2004). On the other hand, in the
MATERIALS AND METHODS
Materials
(S)-citalopram (SCIT) and (R)-citalopram (RCIT) oxalate were
purchased from Trademax (Shanghai, China). The monomers
itaconic acid (IA) ≥99% and ethylene glycol dimethacrylate
(EDMA) 98% were supplied from Sigma-Aldrich (Madrid, Spain).
Porous silica particles (SiliCycle®) were from Teknokroma (Barce-
lona, Spain) whose average diameter, pore size and surface area
‘grafting from’ approach, the polymerization can start from the
2
ꢀ1
surface itself if reactive groups are previously attached (Quaglia
et al., 2003). The greater control of the polymerization process, in
terms of length and density of surface polymer chains above all,
has favoured the recent increase of the use of this technique
were 10 μm, 10 nm and 432 m g respectively. The iniferter N, N
′-diethylaminodithiocarbamoylpropyl(trimethoxy)silane (DDCPTS)
was synthesized as reported elsewhere (Bossi et al., 2010). Reagents
required for its synthesis, that is sodium diethyl dithiocarbamate
trihydrate and chloropropyl trimethoxysilane ≥97%, were obtained
(Giovannoli et al., 2012).
Figure 1. (a) Silica surface activation and (b) functionalization.
J. Mol. Recognit. 2016; 29: 106–114
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jmr

R. GUTIÉRREZ-CLIMENTE ET AL.
from Sigma-Aldrich (Madrid, Spain). Acetone, acetonitrile (ACN) and
methanol (MeOH), were acquired from Scharlab (Barcelona, Spain)
and dry toluene from Panreac (Barcelona, Spain), all of them were
analytical or high-performance liquid chromatography (HPLC)
grade and used as received. Mass spectrometry grade ammonium
formate >99% and formic acid ≥98% were acquired from Sigma-
Aldrich (Madrid, Spain). Every buffer solution was prepared with
doubly deionized water Milli-Ro and Milli-Q water purifications
system (Millipore, Bedford, MA, USA).
After the activation step, the silica surface was functionalized
using DDCPTS as silane iniferter (Figure 1). For this, 4.3 g of iniferter
were dissolved in 250 ml of dry toluene where, subsequently, 4 g of
dry silica particles were dispersed. The suspension was stirred,
purged with nitrogen for 15 min and kept in the dark for 24 h. Then
silica beads were filtered out and rinsed successively with MeOH
and ACN. The minimum amount of DDCPTS to be employed for
functionalization was calculated based on the amount of silanols
on the silica surface (8 μmol m ) and considering a 3:1 silanol:
iniferter molar ratio. An excess of iniferter was employed to avoid
having uncoated silica particles. The final iniferter density on silica
particles was determined by EA.
ꢀ
2
For the MIP synthesis, SCIT was required to be used in its neu-
tral form. This procedure was carried out as reported elsewhere
(Gomez-Caballero et al., 2011).
Polymer grafting on the functionalized silica was finally carried
out by photo-initiated radical polymerization. Based on a previously
reported computational modelling (Gomez-Caballero et al., 2011), IA
was chosen as functional monomer because of its strong binding
energy with SCIT. Initially, depending on the desired polymer film
thickness (Table 1), different amounts of IA and SCIT were dissolved
in 15-ml ACN in a 50-ml flask, and it was degassed with high-purity
nitrogen for 3 min. Then, the cross-linker (EDMA) and the functional-
ized silica (2 g) suspended in 5 ml of ACN were added. The mixture
was degassed by the freeze–pump–thaw degassing method using
liquid nitrogen and a vacuum pump. Finally, ultraviolet irradiation
was applied to the mixture for 3 h, and polymerized silica particles
were collected, filtered and washed with ACN. The NIP-coated silica
particles were synthesized using the same procedure but in the ab-
sence of SCIT.
Instrumentation
A Philips UV model HP/3151/A (Amsterdam, the Netherlands)
fitted with 4 × 75 W lamps was used for photo-polymerizations.
Elemental analysis (EA) was performed with a Eurovector 3000
(Milan, Italy) elemental analyzer. Fourier transform infrared
spectroscopy analysis of polymers was performed using a FT-IR
spectrometer, model 6300 type A, from Jasco (Madrid, Spain).
Thermogravimetric analysis (TGA) was carried out in a Mettler
Toledo SDTA 851 analyzer (Mettler Toledo International Inc.,
ꢀ
1
Greifensee, Switzerland) at a heating rate of 10 °C min up to
00 °C under nitrogen atmosphere. Morphological studies were
carried out by scanning electron microscopy (SEM) in a JSM-
400 scanning microscope (JEOL Ltd., Tokyo, Japan), with an
9
6
accelerating voltage set to 20 kV. Samples were mounted on
conductor tape and were chromium coated (20 nm thickness).
EA, SEM and TGA analysis were performed by the Advanced
Research Facilities (SGIker) of the University of Basque Country.
Chromatographic experiments were carried out using stainless-
steel columns (4.6 × 100 mm) (Phenomenex-Micron Analítica, Madrid,
Spain) packed with surface-modified silica. The packing was per-
formed using the Pack in a Box column packing system from Restek
Column packing
Stainless steel analytical (100 mm L × 4.6 mm i.d) columns were
packed using a slurry packing procedure: 1.5 g of the chiral
molecularly imprinted material was suspended in 20 ml of ACN.
The slurry was packed in the column at constant pressure
(2000 PSI), using ACN as the pressurizing agent.
(
Bellefonte, USA) comprised of a dual piston pump and a 20-ml
Once the material packed, MIP and NIP columns were placed in
a chromatograph and a mobile phase that consisted of 200-mM
formic acid in MeOH was passed through at 1 mlmin for 5 h.
Finally, the columns were washed with a water:ACN 90:10 mixture
and then with 100% ACN.
reservoir. Chromatographic evaluation of MIP and non-imprinted
polymer (NIP) columns was performed in an Agilent 1100-series bi-
nary pump system (Agilent Technologies Inc., Palo Alto, CA, USA) at
a flow rate of 1 ml min and a temperature of 40 °C. The mobile
phase consisted of a mixture of formate buffer (40 mM, pH 4) and
ACN (30:70 v/v). Fluorescence detection was carried out at 240nn
ꢀ
1
ꢀ
1
Binding capacity and adsorption isotherms
(excitation) and 308 nm (emission).
Batch rebinding analyses were performed using the following
procedure: 20 mg of the surface-modified silica (imprinted or
non-imprinted) were weighted into a 5-ml glass vials followed
by the addition of 2 ml of different standard solutions of SCIT
Preparation of the chiral molecularly imprinted stationary phase
Initially, the surface of commercial silica particles was activated
by rehydroxylation (Figure 1). There are various methods for
rehydroxylation whose impact on the quality of bonded phases
has been thoroughly evaluated by Köhler and Kirkland (Koehler
et al., 1986; Koehler and Kirkland, 1987) and Unger et al. (1991)
who studied acidic/hydrothermal treatment of a series of
commercially available silicas. In this work, an acid treatment
was applied to commercial silica using HCl. 4 g of silica were
suspended in 150 ml of a 17% HCl solution, and the mixture
was kept under reflux during 24 h. Once the mixture had cooled
down, silica particles were filtered and rinsed several times with
water and then twice with 250 ml of MeOH to avoid clotting
during drying (du Fresne von Hohenesche et al., 2004). Finally,
silica particles were dried at 120 °C at least for 12 h. After this
Table 1. Composition of the different polymerization mixtures
depending on the estimated coating thickness
Thickness
nm)
SCIT
(mmol)
IA
(mmol)
EDMA
(mmol)
Silica
(g)
(
0.5
1
2
3
4
0.04
0.08
0.14
0.18
0.19
0.17
0.32
0.55
0.71
0.79
1.00
1.90
3.32
4.27
4.74
2
2
2
2
2
SCIT, (S)-citalopram; IA, itaconic acid; EDMA, ethylene glycol
dimethacrylate.
2
step, an amount of 8 μmol of silanols per m of silica was
assumed as a physicochemical constant (Vansant et al., 1995).
wileyonlinelibrary.com/journal/jmr
Copyright © 2015 John Wiley & Sons, Ltd.
J. Mol. Recognit. 2016; 29: 106–114

INIFERTER-MEDIATED GRAFTING OF MIP ON POROUS SILICA BEADS
2
Table 2. Elemental analysis results
Where δ is the area density (in μmol of grafted group per m of native
silica), p is the carbon percentage (w/w %), S is the specific surface
2
ꢀ1
Elemental
area (m g ) of the native silica, C is the atomic weight of carbon,
n is the number of carbon atoms per grafted group and M is the
molecular weight of the anchored group.
As illustrated in Table 2, the elemental composition of the
iniferter-modified silica and the native one are significantly dif-
Support
composition
Area density Coverage
ꢀ
2
%
N
%C
%H (μmol m
)
(%)
Silica
Si-iniferter
Si-MIP
<0.1 <0.3 1.12
0.98 6.62 1.89
0.86 22.83 3.36
0.9 21.3 3.23
—
1.88
—
—
70.83
—
ꢀ
2
ferent. The estimated area density (1.88 μmol m ) was close to
the theoretical maximum value of 2.67 μmol m , calculated
ꢀ
2
ꢀ
2
considering a maximum silanol density of 8 μmol m and a stoi-
chiometric 3:1 silanol: iniferter ratio. From the estimated area
density, surface coverage was calculated based on Equation (2).
Calculated coverage (C) was found to be 70% of the silica
surface.
Si-NIP
—
—
MIP, molecularly imprinted polymer; NIP, non-imprinted polymer.
ꢀ
1
and RCIT independently (concentration range 3–90 mg L ). Next, the
vials were stirred at room temperature with a MovilROD tube shaker (J.
P. Selecta, Barcelona, Spain) at a rolling speed of 15 rpm. After 24 h, the
silica was left to decant, and the supernatant was removed carefully,
filtered and analyzed by HPLC in order to measure the amount of free
1
8=3
00δ
CSiꢀIniferter ¼
(2)
(
(
unbound) analyte in solution. To this end, a ZORBAX Eclipse XDB-C18
4.6 × 150 mm, 5 μm) column from Agilent Technologies (Palo Alto, CA,
USA) was used, and a mixture of tetramethylammonium chloride
0.4%, pH 4) and ACN (70:30 v/v) was employed as mobile phase at a
Concerning NIP and MIP modified materials, the carbon content
was significantly higher than the one observed in the iniferter-
modified silica. In contrast, the nitrogen percentage remained
constant. The only molecule containing nitrogen among the
ones employed (apart from the template) is the iniferter. In this
sense, carbon percentage increase can be attributed to the poly-
mer growth on silica surface.
(
ꢀ
1
flow rate of 1 ml min at room temperature.
RESULTS AND DISCUSSION
Infrared spectra of the bare silica support, the iniferter-modified
silica and the MIP-grafted or NIP-grafted silica were registered and
compared next (Figure 2). The infrared spectrum of the bare silica
Characterization of surface-modified silica
Characterization of the surface modified silica was performed by
Fourier transform infrared spectroscopy, TGA and EA. Morphological
analysis of the material was carried out by SEM.
Elemental analysis data for unmodified silica, iniferter-modified
silica and MIP-grafted or NIP-grafted silica are shown in Table 2. From
these results the surface coverage of the functionalized silica was
calculated using the Berendsen-de Galan equation, Equation (1)
ꢀ1
shows a broad and intense band at 1110 ꢀ 1000 cm that was
attributed to the Si-O stretching vibration. This band can be distin-
guished in each of the recorded spectra. Insets 1 and 2 (Figure 2)
depict magnifications of each of the registered spectra. In inset 1,
ꢀ1
bands around 2950 cm
are clearly distinguished that were
assigned to -C-H stretching vibrations of methyl and methylene
groups. In spectrum (b), the presence of these bands in compari-
son with spectrum (a) is clear, what is indicative of the iniferter
coupling. In (c) and (d) spectra, a moderately intense band is
(Sandoval, 1999).
1
06p
Sð100Cn ꢀ pMÞ
ꢀ1
noticeable at 1723 cm (inset 2, Figure 2), which was assigned
δ ¼
(1)
to C = O stretching vibration. Both the cross-linker (EDMA) and
Figure 2. Fourier trasnsform infrared spectroscopy spectra of (a) silica, (b) iniferter modified silica, (c) non-imprinted polymer-modified silica and (d)
molecularly imprinted polymer-modified silica.
J. Mol. Recognit. 2016; 29: 106–114
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jmr

R. GUTIÉRREZ-CLIMENTE ET AL.
comparison with the NIP. The total monomer addition was ad-
justed to get films of 0.5, 1, 2, 3 and 4 nm in average thickness, as
described elsewhere (Halhalli et al., 2012). The binding capacity of
each of the prepared materials was determined next by classical
frontal chromatography. Table 3 summarizes the binding capacities
of MIP and NIP modified silicas. The highest MIP capacity was ap-
preciated when 4-nm coatings were employed. Higher coating
thicknesses were not employed to avoid collapsing silica pores
and the subsequent surface area decrease. Sulitzky et al. (2002)
had previously reported that a 10-nm-pore-sized silica contain-
ing ultrathin grafted films (1 nm approximately) exhibited the
highest chromatographic efficiency. However, particles with
thinner films (<1 nm) were unable to separate the enantiomers.
In the present work, it was observed that thicknesses below 4 nm
were unable to resolve the racemate. Moreover, NIP retention
capacity was higher than the one observed for the MIP. This fact
could be attributed to the lower number of binding sites when
coatings were thinner.
Figure 3. Thermogravimetric analysis measurements of the (a) iniferter-
modified silica (SI-INF), (b) molecularly imprinted polymer-modified silica
and (c) non-imprinted polymer-modified silica.
the functional monomer (IA) are the only molecules among
the employed ones that contain C = O groups, so the presence
of this band in the infrared spectrum can be indicative of the
polymer formation.
Mobile phase optimization
The effect of the mobile phase composition was investigated
next in order to know the solvent mixture that led to best enan-
tiomeric resolution. The pH effect, buffer concentration and or-
ganic modifier percentage on the chromatographic separation
of citalopram enantiomers was investigated. Separation effi-
ciency was assessed based on the separation factor (α) and the
resolution (R), whereas imprinting effect was evaluated through
the imprinting factor. Figure S1 depicts the effect of each of
the tested variables on these three parameters. Best
enantiomeric resolution was observed working at pH:4, at a
buffer percentage of 30% and a buffer concentration of 40 mM
Thermal stability of native silica, iniferter-modified silica and its
derivatives, Si-NIP and Si-MIP, were investigated by TGA. As
depicted in Figure 3, in all cases, an initial mass loss can be ob-
served at 100 °C because of the solvent content of the material.
As regards the iniferter modified-silica (Si-INF), weight loss at
this point is higher because of the fact that the material
was not dried in order to prevent it from thermal premature
initiation. The weight loss observed around 400 °C is due to
the organic matter on the silica. In the iniferter modified silica,
the weight loss was lower than the one observed in the Si-
MIP or Si-NIP, because of the polymeric absence on the silica
surface. As could be expected, the polymer mass in the NIP
and the MIP is almost equal; so, the polymer coating in both
cases could be assumed to be similar.
A MIP-based stationary phase should have different properties
that make it a ‘true’ HPLC material. First of all, the column should
be composed of monodispersed and very regular beads, with an
average diameter from 2 to 10 μm (Baggiani, 2005). Because of
that, the performance of the MIPs is not exclusively dictated by
the interaction of the template with the functional monomer
pre-polymerization and post-polymerization at a molecular level
but also by their overall morphological features (Simoes et al.,
(Supporting information).
Figure 5 depicts overlaid chromatograms registered with the MIP
and the NIP columns at the optimized mobile phase composition. It
is noticeable that the NIP is not capable of resolving the racemate it
constrast to what can be observed with the MIP.
The effect of column temperature
Molecularly imprinted polymer selectivity is based on the differ-
ence of binding free energies of different substrates in an
imprinted site (Spivak, 2005), and these differences are affected
by temperature. Temperature has, at least, two different effects
on the resolution of chromatographic separations: the thermo-
dynamic and the kinetic effect (Peter et al., 1998). The former is
the one that influences the separation factor (α). Temperature
dependence of thermodynamic parameters is described in
Equations (3) and (4) (Berthod et al., 2004).
2
014). Optical SEM images showed that the silica surface, both
in the MIP (Figure 4c) and the NIP (Figure 4b) modified silicas,
had been coated with the polymer without any appreciable
difference concerning particle morphology. No particle agglom-
eration was observed, and based on the morphological similari-
ties between the MIP-silica, NIP-silica and the bare silica, it was
concluded that polymerization had occurred in the silica pores.
ΔðΔGÞ ¼ ꢀR T lnα ¼ ΔðΔHÞ ꢀ TΔðΔSÞ
ΔðΔHÞ ΔðΔSÞ
(3)
(4)
ln α ¼ ꢀ
þ
RT
R
Polymer thickness optimization
Being the chain growth mainly confined to the support surface,
the ‘grafting from’ approach allows to have a better control of
the thickness of the grafted polymer (Prucker and Ruehe, 1998).
It could be expected that mass transfer problems associated with
most of MIPs would be largely overcome by thin grafted films
Where Δ(ΔG), Δ(ΔH) and Δ(ΔS) represent respectively, the differ-
ence of the Gibbs free energy change of the enantiomer-selector
phase transfer and the differences in the enthalpy and entropy
for a given pair of enantiomers (Berthod et al., 2004). R is the
ꢀ
1
ꢀ1
gas constant (8.314 J mol
and α the separation factor.
K ), T is the absolute temperature
(Sulitzky et al., 2002).
In this work, different coating thicknesses were tested in order
to know the one that led to the best MIP performance in
Temperature effect was evaluated through the injection of
RCIT and SCIT independently in the chromatograph. Column
wileyonlinelibrary.com/journal/jmr
Copyright © 2015 John Wiley & Sons, Ltd.
J. Mol. Recognit. 2016; 29: 106–114

INIFERTER-MEDIATED GRAFTING OF MIP ON POROUS SILICA BEADS
Figure 4. Scanning electron microscopy micrographs at different magnifications of (a) bare silica, (b) non-imprinted polymer-modified silica and (c)
molecularly imprinted polymer-modified silica.
heater was fixed to the desired temperature and kept constant at
Table 3. Binding capacities of the surface-modified silicas at
least for 20 min before injection. From the retention values of
different thicknesses
each compound, the retention factor (k) and the separation
factor (α) were calculated. Van’t Hoff plot was constructed next
Capacity
mg SCIT g material)
ꢀ
1
ꢀ
1
plotting the experimental ln α values versus 1/T (K ). As
illustrated in Figure 6, two linear sections can be distinguished.
The nonlinearity of the plot is an indicative of the differences
of the thermodynamic processes, and it happens as a result of
conformational changes of the analyte, the stationary phase or
as a consequence of their interaction (Lao, 2013).
(
Thickness
nm)
(
MIP
NIP
4
3
2
1
2.29
1.65
1.13
1.16
1.06
1.37
1.69
2.06
Experimental data of each section were fitted by linear regression
and from the intercept, and the slope of each equation Δ(ΔS) and
Δ(ΔH) were obtained. In Section A, which corresponds to temperatures
SCIT, (S)-citalopram; MIP, molecularly imprinted polymer; NIP,
non-imprinted polymer.
ꢀ
1
ranging from 50–70 °C, Δ(ΔH) was found to be ꢀ1.77 KJ mol and
J. Mol. Recognit. 2016; 29: 106–114
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jmr

R. GUTIÉRREZ-CLIMENTE ET AL.
Figure 5. Chromatographic profile of citalopram at (a) a non-imprinted stationary phase and (b) an imprinted stationary phase. Mobile phase was
ꢀ1
4
0 mM formate buffer:ACN (30:70 v/v) at a flow rate of 1 ml min and a temperature of 40 °C.
Adsorption isotherms were fitted using different mathematical
models including Langmuir, Bi–Langmuir, Freundlich and
Freundlich–Langmuir isotherm models (Supporting information).
Freundlich–Langmuir model was found to be the best fitting
one, which is described by Equation (5). Where N is the binding
t
site number, a is the average affinity binding and m is the het-
erogeneity factor. Fitting parameters are summarized in Table S1
for both the MIP and the NIP.
NT aFm
B ¼ 1
(5)
m
þ aF
The average binding affinity of the MIP (3.16 10 ± 0.26
ꢀ
2
ꢀ
2
ꢀ1
Figure 6. Van’t Hoff plot where ln α versus 1/T is represented for each of
the tested temperatures.
10 mM ) was noticeably higher than the one observed for
ꢀ
3
ꢀ3
ꢀ1
the NIP (2.80 10 ± 1.74 10 mM ). The heterogeneity factor
of the imprinted stationary phase was 0.82, whereas the m value
of the NIP column was found to be 0.59, what clearly denotes a
more homogeneous binding site distribution of the MIP material.
In what concerns the number of binding sites, the value ob-
served for the NIP material was around 10 times higher than
Δ(ΔS) was equal to ꢀ2.57 J molꢀ1. Negative ΔΔH values indicate an
exothermic transfer of the preferentially adsorbed enantiomer from
the mobile phase to the stationary phase and therefore chiral recog-
nition is favoured. The more negative the value, the stronger the in-
teractions with the stationary phase (Oberleitner et al., 2002).
However, too negative Δ(ΔS) can counteract the chiral recognition
spontaneity making ΔG increase (Equation (3)). In this case, because
ꢀ
1
the one observed for the MIP, 206.4 ± 133.5 μmol g versus
4.97 ± 3.09 μmol g . This could be attributed to a higher
ꢀ
1
2
amount of non-specific low-affinity binding sites related with
randomly distributed carboxylic functional groups in the poly-
mer. In other words, as could be expected, a higher number of
binding sites with a more heterogeneous distribution and with
much lower binding affinity was observed for non-imprinted ma-
terials in comparison with imprinted ones.
|ΔΔH| > |TΔΔS|, the process in Section (a) is enthalpy controlled and
temperature increase worsens enantioselectivity.
In Section (b), the opposite effect occurs. Here, ΔΔH and ΔΔS
ꢀ
1
ꢀ1
values are 0.86 KJ mol and 5.58 J mol respectively, so |ΔΔH| <
TΔΔS|, and the process is entropy controlled. In consequence, from
0 to 50 °C, as temperature increases, enantioselectivity is enhanced.
|
3
CONCLUSIONS
Binding capacity and adsorption isotherms
Iniferter-mediated polymerization of an imprinted polymer has
been demonstrated here to be a useful tool to coat silica based
solid support. The MIP chiral selector has been successfully
grafted on porous silica particles leading to a homogeneous
material that recognizes preferably the S enantiomer of the
Experimental adsorption isotherms were obtained by batch
rebinding experiments that are one of the best methods to
characterize binding behaviour of MIPs (Spivak, 2005). Results
are described in supporting information (Figure S3 and Table S1).
wileyonlinelibrary.com/journal/jmr
Copyright © 2015 John Wiley & Sons, Ltd.
J. Mol. Recognit. 2016; 29: 106–114

INIFERTER-MEDIATED GRAFTING OF MIP ON POROUS SILICA BEADS
drug. Grafted polymer thickness has been optimized, and best
MIP performance was observed using 4-nm-thick films. Lower
thicknesses did not provide a good resolution of the tested
racemate. The MIP-coated silica particles have demonstrated
their performance as CSP for the enantiomeric resolution of
the drug citalopram. The methodology presented here could
be easily extrapolated to other enantiomers of drugs for the
controlled development of thin film CSP that could separate
drug racemates.
Acknowledgements
The authors wish to acknowledge the Spanish Ministry of
Economy and Competitiveness (project CTQ2013-47921-P) and
the University of the Basque Country (project PPM12/08) for
the financial support. The Basque Government (Department of
education, language policy and culture) is also gratefully
acknowledged for a predoctoral BFI-2011-122 grant for R.
Gutiérrez-Climente. Technical and human support provided by
SGIker (UPV/EHU, MICINN, GV/EJ, ESF) is also acknowledged.
REFERENCES
Ansell RJ, Kuah JKL, Wang D, Jackson CE, Bartle KD, Clifford AA. 2012.
Imprinted polymers for chiral resolution of (±)-ephedrine, 4:
packed column supercritical fluid chromatography using molecu-
larly imprinted chiral stationary phases. J. Chromatogr. A 1264:
Lu C-H, Zhou W-H, Han B, Yang H-H, Chen X, Wang X-R. 2007. Surface-
imprinted core-shell nanoparticles for sorbent assays. Anal. Chem.
79: 5457–5461.
Oberleitner WR, Maier NM, Lindner W. 2002. Enantioseparation of various
amino acid derivatives on a quinine based chiral anion-exchange
selector at variable temperature conditions. Influence of structural
parameters of the analytes on the apparent retention and
enantioseparation characteristics. J. Chromatogr. A. 960: 97–108.
Perez-Moral N, Mayes AG. 2004. Noncovalent imprinting in the shell of
core-shell nanoparticles. Langmuir 20: 3775–3779.
Peter A, Torok G, Armstrong DW, Toth G, Tourwe D. 1998. Effect of
temperature on retention of enantiomers of β-methyl amino acids on
a teicoplanin chiral stationary phase. J. Chromatogr. A 828: 177–190.
Piletsky S, Turner A. 2006. Molecular imprinting of polymers. Landes
Bioscience: Texas.
1
17–123.
Arnold FH, Plunkett S, Dhal PK, Vidyasankar S. 1995. Surface modification
with molecularly-imprinted polymers for selective recognition.
Polym. Prepr. (Am. Chem. Soc., Div. Polym. Chem.). 36: 97–98.
Baggiani C. 2005. Chromatographic techniques. In Molecularly imprinted
materials. Science and Technology. Ramström O, Yan M (eds). Marcel
Dekker: New York; 517–734.
Barahona F, Turiel E, Cormack PAG, Martin-Esteban A. 2010. Chromato-
graphic performance of molecularly imprinted polymers: core-shell
microspheres by precipitation polymerization and grafted MIP films
via iniferter-modified silica beads. J. Polym. Sci., Part A: Polym. Chem.
4
8: 1058–1066.
Piletsky SA, Matuschewski H, Schedler U, Wilpert A, Piletska EV, Thiele TA,
Ulbricht M. 2000. Surface functionalization of porous polypropylene
membranes with molecularly imprinted polymers by photograft
copolymerization in water. Macromol. 33: 3092–3098.
Piletsky S, Piletsky S, Chianella I. 2012. MIP-based sensors. In Molec-
ularly imprinted sensors Li S et al. (eds). Elsevier: Amsterdam;
339–354.
Prasad BB, Tiwari K, Singh M, Sharma PS, Patel AK, Srivastava S. 2008. Mo-
lecularly imprinted polymer-based solid-phase microextraction fiber
coupled with molecularly imprinted polymer-based sensor for
ultratrace analysis of ascorbic acid. J. Chromatogr. A 1198-1199:
59–66.
Berthod A, He BL, Beesley TE. 2004. Temperature and enantioseparation by
macrocyclic glycopeptide chiral stationary phases. J. Chromatogr. A
1
Bossi A, Whitcombe MJ, Uludag Y, Fowler S, Chianella I, Subrahmanyam S,
Sanchez I, and Piletsky SA. 2010. Synthesis of controlled polymeric
cross-linked coatings via iniferter polymerization in the presence of
tetraethyl thiuram disulphide chain terminator. Biosens. Bioelectron.
060: 205–214.
2
5: 2149–2155.
Chen X, Yang W, Zhou Y, Jiao F. 2012. In situ synthesis of monolithic mo-
lecularly imprinted stationary phases for liquid chromatographic
enantioseparation of dibenzoyl tartaric acid enantiomers. J. Porous
Matter. 19: 587–595.
Edmondson S, Osborne VL, Huck WTS. 2004. Polymer brushes via surface-
initiated polymerizations. Chem. Soc. Rev. 33: 14–22.
Prucker O, Ruehe J. 1998. Mechanism of radical chain polymerizations
initiated by azo compounds covalently bound to the surface of
spherical particles. Macromol. 31: 602–613.
du Fresne von Hohenesche C, Ehwald V, Unger KK. 2004. Development of
standard operation procedures for the manufacture of n-octadecyl
bonded silicas as packing material in certified reference columns
for reversed-phase liquid chromatography. J. Chromatogr. A 1025:
Quaglia M, De Lorenzi E, Sulitzky C, Caccialanza G, Sellergren B. 2003.
Molecularly imprinted polymer films grafted from porous or nonpo-
rous silica: novel affinity stationary phases in capillary electro-
chromatography. Electrophoresis 24: 952–957.
Sandoval JE. 1999. Equation for calculating surface coverage from end-
capping of chromatographic bonded phases. J. Chromatogr. A 852:
375–381.
Sellergren B, Esteban AM. 2010. The use of molecularly imprinted
polymers for sampling and sample preparation. In Handbook of
sample preparation, Pawliszyn J, Lord HL (eds). John Wiley and Sons:
Hoboken; 445–473.
Simoes M, Martins N, Cabrita MJ, Burke AJ, Garcia R. 2014. Tailor-made
molecularly imprinted polymers for dimethoate and deltamethrin
recognition: synthesis, characterization and chromatographic evalu-
ation. J. Polym. Res. 21: 1–13.
177–187.
Giovannoli C, Passini C, Baravalle P, Anfossi L, Giraudi G, Baggiani C. 2012.
An innovative approach to molecularly imprinted capillaries for polar
templates by grafting polymerization. J. Mol. Recognit. 25: 377–382.
Gomez-Caballero A, Guerreiro A, Karim K, Piletsky S, Goicolea MA, Barrio
RJ. 2011. Chiral imprinted polymers as enantiospecific coatings of stir
bar sorptive extraction devices. Biosens. Bioelectron. 28: 25–32.
Haginaka J, Kagawa C. 2002. Uniformly sized molecularly imprinted poly-
mer for d-chlorpheniramine. Evaluation of retention and molecular
recognition properties in an aqueous mobile phase. J. Chromatogr.
A. 948: 77–84.
Halhalli MR, Aureliano CSA, Schillinger E, Sulitzky C, Titirici MM, Sellergren
B. 2012. An improved grafting technique for producing imprinted
thin film composite beads. Polym. Chem. 3: 1033–1042.
Kempe H, Kempe M. 2004. Novel method for the synthesis of molecularly
imprinted polymer bead libraries. Macromol. Rapid Commun. 25:
Spivak DA. 2005. Optimization, evaluation, and characterization of molec-
ularly imprinted polymers. Adv. Drug Delivery Rev. 57: 1779–1794.
Su S, Zhang M, Li B, Zhang H, Dong X. 2008. HPLC determination of sul-
famethazine in milk using surface-imprinted silica synthesized with
iniferter technique. Talanta 76: 1141–1146.
3
15–320.
Sulitzky C, Rueckert B, Hall AJ, Lanza F, Unger K, Sellergren B. 2002.
Grafting of molecularly imprinted polymer films on silica supports
containing surface-bound free radical initiators. Macromol. 35:
79–91.
Koehler J, Kirkland JJ. 1987. Improved silica-based column packings for
high-performance liquid chromatography. J. Chromatogr. 385:
1
25–150.
Koehler J, Chase DB, Farlee RD, Vega AJ, Kirkland JJ. 1986. Comprehensive
characterization of some silica-based stationary phases for high-
performance liquid chromatography. J. Chromatogr. 352: 275–305.
Lao W. 2013. Thermodynamic and extrathermodynamic studies of
enantioseparation of imidazolinone herbicides on chiralcel OJ
column. ISRN Chromatogr. 2013: 460787–460789.
Tamayo FG, Titirici MM, Martin-Esteban A, Sellergren B. 2005. Synthesis
and evaluation of new propazine-imprinted polymer formats for
use as stationary phases in liquid chromatography. Anal. Chim. Acta
542: 38–46.
Turiel E, Martin-Esteban A. 2005. Molecular imprinting technology in
capillary electrochromatography. J. Sep. Sci. 28: 719–728.
J. Mol. Recognit. 2016; 29: 106–114
Copyright © 2015 John Wiley & Sons, Ltd.
wileyonlinelibrary.com/journal/jmr

R. GUTIÉRREZ-CLIMENTE ET AL.
Unceta N, Goicolea MA, Barrio RJ. 2011. Analytical procedures for the deter-
mination of the selective serotonin reuptake inhibitor antidepressant
citalopram and its metabolites. Biomed. Chromatogr. 25: 238–257.
Unger KK, Lork KD, Pfleiderer B, Albert K, Bayer E. 1991. Impact of
acidic/hydrothermal treatment on pore structural and chromato-
graphic properties of porous silicas I. The conventional approach.
J. Chromatogr. 556: 395–406.
polymerization for affinity separation applications. Angew. Chem. Int.
Ed. 42: 5336–5338.
Whitcombe MJ, Kirsch N, Nicholls IA. 2014. Molecular imprinting science
and technology: a survey of the literature for the years 2004–2011.
J. Mol. Recognit. 27: 297–401.
Vansant EF, Van Der Voort P, Vrancken KC, and Editors. 1995. Characteri-
zation and chemical modification of the silica surface. Elsevier
Science B. V: Amsterdam.
SUPPORTING INFORMATION
Additional supporting information may be found in the online
Wang J, Cormack PAG, Sherrington DC, Khoshdel E. 2003. Monodisperse, mo-
lecularly imprinted polymer microspheres prepared by precipitation
version at the publisher’s web site.
wileyonlinelibrary.com/journal/jmr
Copyright © 2015 John Wiley & Sons, Ltd.
J. Mol. Recognit. 2016; 29: 106–114