Journal of Colloid and Interface Science 419 (2014) 148–159

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

www.elsevier.com/locate/jcis

Removal of sulfamethoxazole sulfonamide antibiotic from water by high

silica zeolites: A study of the involved host–guest interactions
by a combined structural, spectroscopic, and computational approach
Sonia Blasioli a, Annalisa Martucci b, Geo Paul c,d, Lara Gigli b, Maurizio Cossi c,d, Cliff T. Johnston e,
Leonardo Marchese c,d, Ilaria Braschi a,d,⇑
a
Department of Agricultural Sciences, University of Bologna, Viale G. Fanin 44, 40127 Bologna, Italy
b
Department of Physics and Earth Sciences, University of Ferrara, Via G. Saragat, 1, 44100 Ferrara, Italy
c
Dipartimento di Scienze e Innovazione Tecnologica, Università del Piemonte Orientale A. Avogadro, Viale T. Michel 11, 15121 Alessandria, Italy
d
Centro Interdisciplinare Nano-SiSTeMI, Università del Piemonte Orientale A. Avogadro, Viale T. Michel 11, 15121 Alessandria, Italy
e
Department of Agronomy, Purdue University, 915 West State Street, West Lafayette, IN 47907-2053, USA

a r t i c l e i n f o a b s t r a c t

Article history: Sulfonamide antibiotics are persistent pollutants present in surface and subsurface waters in both agri-
Received 27 August 2013 cultural and urban environments. Sulfonamides are of particular concern in the environment because
Accepted 16 December 2013 they are known to induce high levels of bacterial resistance. Adsorption of sulfamethoxazole sulfonamide
Available online 24 December 2013
antibiotic into three high silica zeolites (Y, mordenite, and ZSM-5) with pore opening sizes comparable to
sulfamethoxazole dimensions is reported. Sulfamethoxazole was almost completely removed from water
Keywords: by zeolite Y and MOR in a few minutes. Adsorption onto ZSM-5 showed an increased kinetics with
Adsorption
increasing temperature. Antibiotic sorption was largely irreversible with little antibiotic desorbed. Sulfa-
Water depollution
Zeolite Y
methoxazole incorporation and localization into the pore of each zeolite system was defined along with
Mordenite medium-weak and cooperative host–guest interactions in which water molecules play a certain role only
ZSM-5 in zeolite Y and mordenite.
Ó 2013 Elsevier Inc. All rights reserved.

1. Introduction employed. Using biological treatment with activate sludge, SMX

Sulfamethoxazole (SMX) is a broad spectrum antibiotic used in
human and veterinary therapy which belongs to the sulfonamide
family (sulfa drugs). Like other sulfonamides, SMX is a competitor
of p-aminobenzoic acid in the biosynthesis of tetrahydrofolic acid.
In studies on humans it was found that 20% of SMX daily dose was
excreted via urine/feces in its original form, 50–60% was trans-
formed to acetylated derivative (N1-acetylsulfamethoxazole, see
Fig. 1) and 15–20% to glucuronide conjugate [30,31,13].
The inefficiency of sewage treatment plants to degrade sulfon-
amides and the practice to land apply sulfonamide-containing
manure for agronomic purpose contribute to the spreading of anti-
biotics in the environment. As reported by Michael et al. [36], SMX
concentrations in urban wastewater treatment plants depend on
its consumed amount and the type of wastewater treatment
⇑ Corresponding author at: Department of Agricultural Sciences, University of
Bologna, Viale G. Fanin 44, 40127 Bologna, Italy. Fax: +390512096203.
E-mail addresses: sonia.blasioli@unibo.it (S. Blasioli), annalisa.martucci@unife.it
(A. Martucci), geo.paul@mfn.unipmn.it (G. Paul), 90647@studenti.unimore.it
(L. Gigli), maurizio.cossi@mfn.unipmn.it (M. Cossi), cliffjohnston@purdue.edu
(C.T. Johnston), leonardo.marchese@mfn.unipmn.it (L. Marchese), ilaria.braschi@
unibo.it (I. Braschi).
removal efficiency ranged between 100% and <25% [36]. Two
reasons for the high variability could have been the reconversion
of the N1-acetyl derivate in the parent compound, which would
have underestimated SMX removal efficiency, and to its anionic
nature (pKa2 = 5.7 [26]) that limits sorption on activate sludge
resulting in lower efficiency [38]. As a result, wastewater treat-
ment plants represent potential reservoirs of sulfonamides that
may contribute to the evolution and spread of antibiotic resistance
[33]. Owing to its anionic nature, SMX is not strongly retained by
soils containing organic carbon and clays [15,16] and moves
through the soil profile accumulating in the aquatic environment
[7]. The most important consequence of antibiotic release in natu-
ral environments is the development of resistant bacteria [18].
Several techniques have been developed to degrade and remove
antibiotics from water and wastewater [34]. Among these, adsorp-
tion on activated carbon is one of the most efficient; however, the
high cost of the adsorbent, their susceptibility of pore blocking, and
the difficulties of regeneration are disadvantages [9]. As an alterna-
tive low-cost adsorbent, this study evaluated the mechanisms of
SMX removal efficiency from water using several commercial high
silica (HS) zeolites.

0021-9797/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcis.2013.12.039
 S. Blasioli et al. / Journal of Colloid and Interface Science 419 (2014) 148–159
Fig. 1. DFT optimized stick and ball (left) and electron density (right) structure of sulfamethoxazole. Geometry optimization with GGA functional BLYP. Plot of total electron
density with isosurface at 8  103 Å3.
Zeolites are microporous crystalline alumino-silicates exten-
sively used because of their high surface area, adsorption capacity
and cationic exchange capacity. Moreover, the sorption of
hydrophobic and hydrophilic solutes can be modulated by control
of zeolite SiO2/Al2O3 ratio. Of particular interest in this study is the
relation between the molecular dimensions of the zeolite pores
relative to the size of SMX.
Studies on the use of high silica (HS) zeolites (high SiO2/Al2O3
ratio) adsorbents for pharmaceutical compounds from water are
limited [37,4,12,19,20,6,24,25]. HS zeolite Y has been tested for
the adsorption of sulfa drugs from soil solution containing salts
and dissolved organic matter with excellent results (ca. 20% zeolite
dry weight – DW [4,5]). In the light of these recent findings, we be-
lieved interesting to investigate the adsorption of SMX from water
by a series of HS zeolites and the effect of water on the host–guest
interactions between zeolite and antibiotic.
2. Materials and methods
2.1. Chemicals
Sulfamethoxazole (4-amino-N-(5-methylisoxazol-3-yl)ben-
zenesulfonamide (DFT optimized structure in Fig. 1); MW
253.3 g mol1; Kow 0.4: [27]; pKa2 5.7), was purchased as analytical
standard by Dr. Ehrenstorfer GmbH (Germany) with a purity of
99%.
Because SMX contains ionizable functional groups, its solubility
depends on solution pH [23]. A SMX stock solution at the maximal
solubility was prepared by adding the antibiotic to distilled water
in equilibrium with the atmospheric CO2 (ca. pH 6.0), in amount
exceeding that required to saturate the solution. The suspension
was sonicated for 15 min, shaken at 50 °C for 30 min and, after
cooling at room temperature (RT), filtered through 0.45 lm Dura-
poreÒ membrane filters to eliminate the undissolved solute. SMX
solubility, measured by means of high performance liquid chroma-
tography (HPLC), was 202.3 ± 4.2 lM at RT.
Three different HS zeolites have been chosen: Y, mordenite
(MOR), and ZSM-5. Powders of zeolites Y and MOR (code
HSZ-390HUA and HSZ-690 HOA, respectively) were purchased in
their protonic form from Tosoh Corporation (Japan). ZSM-5 (code
TSP-3022) was supplied in its ammonium form (<0.04%) by Tricat
(USA).
2.2. Temperature effect on adsorption
Adsorption experiments were performed at both RT and 65 °C
by adding each zeolite to a 30 lM antibiotic solution (zeolite:
antibiotic solution ratio of 1 mg:2 mL) in polyallomer centrifuge
tubes (Nalgene, NY, USA). Suspensions were shaken for 24 h and
then centrifuged at 20,000g for 15 min. Finally, an aliquot of the
149
supernatant was withdrawn and analyzed by HPLC. The amount
of adsorbed antibiotics was calculated by the difference between
the initial and final concentration. Controls run in the absence of
zeolites allowed to verify the stability of SMX and any possible
adsorption of the drug onto the centrifuge tubes. The experiments
were carried out in triplicate.
2.3. Adsorption kinetics
Kinetics were performed at RT on zeolite Y and MOR and at
65 °C on ZSM-5 with a zeolite:antibiotic solution ratio of
1 mg:2 mL. SMX was weighted on a Mettler Toledo AT21 Compar-
ator balance (weigh limit = 1 lg). The concentration of adsorbed
antibiotic was obtained by HPLC as reported in Section 2.2. The
experiments were conducted in triplicate. Control experiments
were conducted to verify the SMX stability for the entire duration
of the kinetics.
2.4. Adsorption and desorption isotherms
Owing to the moderate SMX solubility and to the high adsorp-
tion capacity of the selected zeolites, the adsorption isotherms
were built by mean of subsequent adsorption cycles with zeo-
lite:antibiotic solution ratio of 1 mg:2 mL as follows: suspensions
of zeolites Y or MOR in the antibiotic solution at maximal solubility
were shaken for 30 min, whereas ZSM-5 for 24 h. Then the suspen-
sions were centrifuged and the supernatant analyzed by HPLC. The
supernatant was then completely removed and substituted by
fresh antibiotic solution. Several subsequent adsorption steps were
performed until the zeolites reached the maximum adsorption
capacity (from here on called exhausted/loaded zeolites). All
experiments were conducted in triplicate.
Adsorption data were expressed as Ce (lM) – the antibiotic con-
centration in aqueous phase at equilibrium – and Cs (lmol g1
adsorbent) – the amount of antibiotics adsorbed onto zeolite (i.e.
the sum of adsorbed antibiotic after each adsorption step) – calcu-
lated by the difference between the initial and final (Ce)
concentrations.
Desorption isotherms were performed on loaded zeolites ob-
tained after the last adsorption step by substituting one SMX solu-
tion half-volume with distilled water. The new suspension was
shaken for 24 h, centrifuged and the supernatant analyzed by
HPLC. Then, a supernatant second half-volume was again removed
and substituted with distilled water. The dilution step was re-
peated several times, until no decrease was detected in the SMX
equilibrium concentration.
2.5. Chromatographic analysis
SMX concentration was determined by a Jasco HPLC-Diodarray.
The system was assembled with the detector set at 267 nm, Jasco
150 S. Blasioli et al. / Journal of Colloid and Interface Science 419 (2014) 148–159
ChromNAV1.14.01 chromatography data software, a Jones Chro-
matography model 7971 column heater, and a 4.60 mm  150 mm
Synergi 4 lm Hydro-RP 80A analytical column (Phenomenex,
USA). The column was kept at 35 °C and eluted with acetoni-
trile:water eluant (23:77 by volume, pH 2.7 for H3PO4) at 1 mL/
min. Under these conditions, SMX retention time was 11.5 min.
All solvents were HPLC grade.
2.6. Thermogravimetric analysis
Thermogravimetric analyses (TGA) were carried out using a
TGA/SDTA851e, Mettler Toledo. About 10 mg of sample (SMX, bare
and loaded zeolites) were heated in a ceramic crucible under air
flux at 25–1000 °C temperature range and 10 °C/min.
2.7. Specific surface area and pore volume measurements
The specific surface area and pore size distribution were mea-
sured by nitrogen adsorption–desorption isotherms obtained at li-
quid nitrogen temperature (196 °C) in the 5  106–760 Torr
(1 Torr = 133.33 Pa) pressure range using a Micromeritics Tristar
3000 sorptometer (Norcross, GA, USA). Prior to analysis, samples
were outgassed at 150 °C and 106 Torr for a minimum of 12 h.
BET surface areas were calculated from the linear part of the BET
plot and BJH pore sizes were obtained from adsorption isotherms.
2.8. X-ray diffraction and structure determination
X-ray powder diffraction patterns of zeolites after SMX
adsorption (Y-SMX, MOR-SMX, and ZSM-5-SMX, respectively) were
measured on a Bruker D8 Advance Diffractometer equipped with a
Sol-X detector, using Cu Ka1, a2 radiation. The spectra were col-
lected in the 3–110° 2h range with a counting time of 12 s step1.
Rietveld profile fittings were performed using the GSAS [17]
computer program with the EXPGUI interface [29]. Full matrix
least-square refinements of SMX loaded zeolites were carried out
in the Fd-3 (for zeolite Y), Cmc21 (for MOR), and P212121 (for
ZSM-5) space groups, respectively. The crystallographic data and
refinement details are reported as Supporting information (SI) in
Table 1S. The final atomic positions and thermal parameters are gi-
ven in Tables 2S–4S.
2.9. Infrared analysis
Absorption spectra were collected on a Thermo Electron Corpo-
ration FT Nicolet 5700 Spectrometer with 4 cm1 resolution. DRIFT
and ATR-FTIR spectra were obtained on a Perkin–Elmer GX2000
Fourier Transform infrared spectrometer equipped with MCT and
DTGS detectors with 4 and 2 cm1 unapodized resolution, respec-
tively. Further details are available as SI.
2.10. Computational modeling
All the calculations were performed with GAUSSIAN03 code
[11], at the Density Functional Theory (DFT) level, using the hybrid
B3LYP functional [3] and the double-zeta polarized 6-31G(d,p)
[14,10] gaussian basis set. Solvent effects were included through
the C-PCM [8] implemented in GAUSSIAN03.
Two SMX conformations have been considered, since two tauto-
meric forms are possible, according to the protonation state of the
sulfonamide and heterocyclic nitrogen atoms: the amide structure
(Fig. 1) resulted the most stable, while the imide form relative en-
ergy was 8.00 kcal mol1.
Harmonic vibrational frequencies were computed for amide
conformation in its optimized geometry, and compared to experi-
mental IR spectra.
2.11. Solid state nuclear magnetic resonance spectroscopy
All solid state NMR (SS-NMR) spectra were acquired on a Bruker
Avance III 500 spectrometer and a wide bore 11.7 T magnet with
operational frequencies for 1H and 13C of 500.13 and
125.77 MHz, respectively. A 4 mm triple resonance probe with
MAS was employed in all the experiments. The samples were
packed on a Zirconia rotor and spun at a MAS rate between 10
and 15 kHz. For the 13C{1H} CPMAS experiments, the magnetic
fields tHrf of 55 and 28 kHz were used for initial excitation and
decoupling, respectively. During the CP period the 1H RF field tH rf
was ramped using 100 increments, whereas the 13C RF field tCrf
was maintained at a constant level. During the acquisition, the pro-
tons are decoupled from the carbons by using a TPPM decoupling
scheme. A moderate ramped RF field tH rf of 62 kHz was used for
spin locking, while the carbon RF field tCrf was matched to obtain
optimal signal and the CP contact time of 2 ms was used. The relax-
ation delay between accumulations was 0.5–75 s and all chemical
shifts are reported using d scale and are externally referenced to
TMS at 0 ppm. Dehydration of the adducts were carried out (sam-
ple packed on the NMR rotor) in a special hand-made quartz cell
that was connected to a high vacuum line (residual pressure
p < 104 mbar), allowing in situ heat treatment at 120 °C for 2 h.
Then the rotor was extracted from the cell, rapidly closed and
was submitted to SS-NMR analysis (both 1H and 13C CPMAS NMR
studies).
3. Results and discussion
The HS Y zeolite (Y) structure is characterized by eight super-
cages per unit cell with a diameter of approximately 12 Å joined
via circular windows and delimited by four 12-membered ring
(MR) pore openings (7.4 Å) per cage [2]. The MOR structure is char-
acterized by straight 12MR (6.5 Å  7.0 Å) and 8MR (5.7 Å  2.6 Å)
channels running along the c axis. The 12MR channels are inter-
connected along [0 1 0] through 8MR side pockets (3.9 Å). The
ZSM-5 structure consists of two intersecting sets of tubular chan-
nels, delimited by 10MR openings. The sinusoidal channel system
is parallel to the [1 0 0] direction and has a cross section of
5.1 Å  5.5 Å. The straight channel system runs parallel to the
[0 1 0] direction and has 5.3 Å  5.6 Å [0 1 0] openings. Structural
and physico-chemical characteristics of the three zeolites are given
in Table 1.
Like other sulfonamides, the molecular structure of SMX is not
linear, but shows a distorted ‘‘V-shaped’’ geometry. The molecular
dimensions (Fig. 1). Make possible the diffusion of the antibiotic
into the zeolites.
3.1. Adsorption
SMX adsorption onto the selected zeolites has been carried out
at RT and 65 °C for 24 h (see Fig. 1S in the Supplemental informa-
tion (SI) section) in order to evaluate the effect of temperature on
adsorption. The choice of 65 °C was also done because some prac-
tical applications of these sorbents might have working conditions
such a high temperature (e.g. the thermophilic phase of manure/
sewage composting often reaches this temperature and, as re-
ported in Section 1, these wastes are usually highly polluted with
sulfa drugs). Complete removal of SMX by zeolite Y was observed
at both temperatures, in agreement with prior work [4]. Decreased
SMX sorption on MOR occurred was observed at elevated temper-
ature where the sorbed amount was 65% of the initial concentra-
tion at RT but below 40% at 65 °C. A similar trend was observed
for sulfachloropyridazine sulfonamide sorption onto zeolite MOR
[20]. In contrast, the higher temperature favored SMX adsorption
 S. Blasioli et al. / Journal of Colloid and Interface Science 419 (2014) 148–159 151

Table 1
Structural and physico-chemical characteristics of zeolites under investigation.

Zeolite Cation type SiO2/Al2O3 (mol/mol) Unit cell Chemical formula

+
Y H 200 [Si192O384]

MOR H+ 200 [Si48O96]

ZSM-5 NHþ
4
500 [Si96O192]

onto ZSM-5 which increased from 40% at RT to 75% at 65 °C. Given
that diffusivity of species inside zeolite porosities is favored at in-
creased temperature and that adsorption is unfavored (being an
exhothermic process), the amount retained by zeolite is a result
of a counterbalance between the two processes. Among the three
zeolites studied, ZSM-5 has the smallest channel system, and the
reduced diffusivity of drug molecules at RT into its pore system
is the limiting factor. Consequently, the SMX adsorption kinetics
and isotherms have been collected at RT in zeolite Y and MOR
and at 65 °C in ZSM-5.
SMX adsorption kinetics are shown in Fig. 2. SMX was quickly
adsorbed on zeolite Y and MOR with <25% of the initial SMX
concentration remaining in solution after <1 min of contact. In
contrast, adsorption equilibrium was not achieved even after
2 weeks on ZSM-5. In the case of sulfachloropyridazine, 4 h was
required to reach the adsorption equilibrium onto MOR [20]. This
is attributed to the presence of the larger heterocyclic ring (6
membered chloropyridazine ring) of sulfachloropyridazine
compared to the 5 membered methylisoxazole ring of SMX.
Taking into consideration the significant differences of SMX
sorption on Y and MOR compared to ZSM-5, subsequent adsorption
experiments on zeolite Y and MOR were performed using 30 min
contact time compared to 24 h for ZSM-5 (Fig. 3). At low SMX con-
centrations of <20 lM, SMX showed an affinity for the zeolites in
the order: Y  ZSM-5  MOR. At SMX concentrations >180 lM,
the affinity of this zeolite for the drug was undoubtedly the highest
being in the order: ZSM-5  Y  MOR. The adsorption isotherms
onto both Y and MOR zeolites were L-type (plateau calculated val-
ues = 1250 and 270 lmol g1, R2 = 0.991 and 0.985, respectively).
The adsorption isotherm of SMX onto ZSM-5 was S-shaped
showing higher affinity of SMX at higher SMX concentrations.
SMX sorption on ZSM-5 did not achieve a sorption plateau that

30 (a) Y MOR ZSM-5

30 (b) ZSM-5
SMX concentration

SMX concentration

25 25
20 20
(µM)

(µM)

15 15
10 10
5 5
0 0
0 20 40 60 0 100 200 300
Time (min) Time (h)

Fig. 2. Adsorption kinetics of sulfamethoxazole onto zeolites Y, MOR (at RT) and ZSM-5 (at 65 °C): (a) 1 h and (b) 14 d contact time.
152 S. Blasioli et al. / Journal of Colloid and Interface Science 419 (2014) 148–159
Y Mor ZSM-5
des Y des Mor des ZSM-5
1200
C s (µmol/g)
800
400
0
0 50 100 150 200
Ce (µM)
Fig. 3. Adsorption (colored marks) and desorption (white marks) isotherms of
sulfamethoxazole onto high silica zeolites. Contact time: 30 min at RT for zeolite Y
and MOR and 24 h at 65 °C for ZSM-5. Mean of three replicates, standard deviation
lower than 4%. (For interpretation of the references to color in this figure legend, the
reader is referred to the web version of this article.)
was evident for Y and MOR zeolites. At SMX concentrations
<180 lM, the initial curve slope was a C type (R2 = 0.998). At con-
centrations >180 lM, an isotherm branch with stiffer slope oc-
curred, still recalling a C type (R2 = 0.902). This trend can be
explained assuming that, in 0–180 lM concentration range, the
drug interacts with specific adsorption sites which are different
from those occupied in the adsorption branch at higher antibiotic
concentrations. Based on the crystal structure of ZSM-5 and the
molecular dimensions of SMX, two adsorption sites are available
[21,22]. The first and most accessible sorption site is localized at
the intersection of the linear and sinusoidal channels. Once these
sites are occupied, sorption can onto sites within the linear chan-
nel. In the case of SMX, this hypothesis is supported by the Rietveld
analysis (vide infra, Fig. 6) that localize SMX molecule at the inter-
section of the two channels. The data suggest that once SMX mol-
ecules are present at the intersection of the linear and sinusoidal
channels, additional sorption into the smaller, less-accessible lin-
ear channels can occur (>180 lM range). Adsorption can proceed
up to the highest concentration investigated where sites are still
available as shown by the almost unchanged slope of the isotherm
branch in the 180–200 lM concentration range. The adsorption
process of SMX onto each zeolite is irreversible with essentially
no SMX desorption was observed by water from any of the three
zeolites (Fig. 3).
According to HPLC analysis, the amount of SMX sorbed was
about 30%, 10%, and 5% (wt/wt) for Y, ZSM-5, and MOR, respec-
tively (Table 2). These findings were also confirmed by TGA
(Fig. 2S) and XRPD results. Since SMX has been detected in muni-
cipal sewage treatment plant effluents at 0.1–2 lg mL1 concen-
tration range [39], 1 g of zeolite Y could clean up more than 6 m3
water maximally contaminated with the drug.
3.2. Porosimetric analysis
Evidence of the entrapment of SMX molecules inside the zeolite
channels was obtained by comparing the specific surface area
(SSA) and pore volume of the ‘bare’ zeolites (no SMX) with those
containing the maximum surface loading of SMX (Table 3). The
Table 2
Maximal amount of sulfamethoxazole adsorbed onto zeolites. The unit cell composition of zeolite-SMX systems is also reported.
Zeolite SMX adsorbed onto zeolite (mg 100 mg1 DW zeolite) SD in parenthesis
HPLC TGA
Y 29.6 (0.6) 23.9 (3.1)
MOR 5.2 (0.3) 5.8 (0.3)
ZSM-5 9.6 (0.1) 8.0 (1.0)
SSA of zeolite decreased by 79% resulting from SMX sorption; the
corresponding decrease in pore volume was 49%. This sharp reduc-
tion in both SSA and pore volume is attributed to SMX molecules
residing in the pores and channels and is in agreement with the
SMX sorption results (Fig. 3).
In ZSM-5, a decrease in SSA and pore volume of 56% and 48%,
respectively, were found. SMX is larger than that the smallest pores
in ZSM-5 resulting in an SMX adsorption capacity that is below sat-
uration. These features are in good agreement with the fact that
the adsorption data point investigated (ca. 200 lM along Ce axis,
Fig. 3) follows in the isotherm branch at the highest concentration:
the corresponding adsorption isotherm data point is far away from
the saturation concentration, given that the slope of the corre-
sponding isotherm branch is linear. The loaded ZSM-5 sample
hence is seemingly able to adsorb an additional considerable
amount of antibiotic.
The observed decrease of both SSA and total pore volume in ex-
hausted MOR was more limited: 32% and 21%, respectively. Though
the remaining free surface and volume of the sorbent after antibi-
otic loading were found to be remarkably high, the investigated
adsorption data point lays in the plateau region of the adsorption
isotherm. Likely, the lowest antibiotic loading onto MOR seems
to be due to the limited diffusion of bulky SMX molecules inside
its ‘‘mono-dimensional’’ pore system.
3.3. Rietveld refinement
Changes in both the positions and intensities (Figs. 3–5S) of the
diffraction peaks on bare and loaded zeolites revealed structural
modifications induced by the presence of SMX.
3.3.1. Zeolite Y
The observed and difference Fourier maps generated using
GSAS program, revealed the presence of a number of extraframe-
work ions inside the supercages, which can be attributed to encap-
sulated organic molecules. The largest and intense peak in the
difference Fourier map was attributed to the sulfur atom. The re-
fined occupancy factor of sulfur atom allowed us to determine
the amount of sulfa drugs adsorbed in the zeolite channel systems.
SMX embedding caused a lowering in real Fd-3m symmetry in the
parent Y zeolite [4] to Fd-3 as well as a significant decrease in unit
cell volume (Table 1S) in comparison to those of the untreated
materials [4].
On the whole, Rietveld structure refinement revealed the pres-
ence of about 14 antibiotic molecules per unit cell, which are
hosted in the eight supercages of Y zeolite. As a consequence, about
2 molecules per unit cell were embedded inside each zeolite cage.
All these molecules are hosted in crystallographic sites partially
statistically occupied. The cubic Fd-3 symmetry imposes that some
of these are symmetrically equivalent (for instance, C1 crystallo-
graphic site describes C1, C2, C3, C4, C5, and C6 atoms of the aniline
ring at the same time, as reported in Table 2S). The geometry of
sulfa drug adsorbed in Y zeolite is consistent with that of pure
SMX, reported by Takasuka and Nakai [28]. Fig. 4 shows that the
aniline ring is almost centered on the supercage with its plane per-
pendicular to the threefold axes of the unit cell, whereas the
Unit cell composition TGA (XRPD)
XRPD
24.0 [Si192O384] 11.5 (14) SMX
5.0 [Si48O96] 0.6 (0.6) SMX
6.5 [Si96O192] 1.9 (1.5) SMX
 S. Blasioli et al. / Journal of Colloid and Interface Science 419 (2014) 148–159
Table 3
Specific surface area and total pore volume of high silica zeolites before and after sulfamethoxazole adsorption determined by N2 physisorption measurements and applying the
Brunauer–Emmet–Teller (BET) and non-local density functional theory (NLDFT) methods, respectively.
Sample Specific surface area (m2/g)
Bare zeolite Loaded zeolite
Y 648 135
MOR 441 301
ZSM-5 342 150
isoxazole ring is found with six different orientations. The aniline
ring (C1 site) is situated in the window that joins neighboring
supercages, whereas the positions occupied by isoxazole atoms
appear disordered and consequently are only partially localized
(C7 site in Table 2S). The isoxazole and aniline rings of adsorbed
SMX form a typical V configuration with the torsion angle
SAN2AC7 of about 126°, with respect to that (121.1°) of crystalline
SMX [28]. The refined distances of N2 (N2AO1 = 2.68(11) Å and
N2AO4 = 3.29(9) Å, respectively) indicate interactions with
framework oxygens (Fig. 4). The 12MR increased its Crystallo-
graphic Free Area (CFA) when compared with the bare material,
whereas the double 6MR (D6R) unit is contracted (Fig. 4). This
combined effect of widening/contraction of pore openings justifies
the variations in the unit cell parameters reported in Table 1S
[4,35].
3.3.2. MOR zeolite
No significant change of unit cell parameters occurred
(Table 1S) for MOR zeolite. However, the TAO11AT angle changed
from 156.4° (MOR-SMX) to 172.8° (MOR) which clearly revealed a
strong deviation from the topological Cmcm symmetry. A similar
lowering of symmetry in MOR after sulfachloropyridazine [20]
and toluene [1] adsorption has been reported.
Rietveld structure refinement revealed the presence of about
0.6 antibiotic molecules per unit cell, located in the large 12MR
channel, with the aniline ring (crystollographic sites C1, C2, and
C3, corresponding to C1/C6, C2/C5, C3/C4 atom numbering of
Fig. 1, respectively) oriented parallel to the [0 1 0] direction
(Fig. 5). The SMX position is similar to that recently found for sul-
fachloropyridazine loaded into the same high silica MOR [20]. Also
in the case of SMX, the drug molecules show two different orienta-
tions due to the screw axis parallel to [0 0 1], with the isoxazole
ring (crystallographic sites C7, C8, and C9 corresponding to C7,
C8/N3, C9/O3 atom numbering of Fig. 1, respectively) oriented
Fig. 4. Location of sulfamethoxazole (SMX) and oxygen–oxygen distance (Å) of the apertures in zeolite Y after drug adsorption, assuming an oxygen ionic radius of 1.35 Å.
Crystallographic Free Area (CFA) and ellipticity (e) are also reported.
153
Total pore volume at P/Po = 0.99 (cm3/g, STP)
Bare zeolite Loaded zeolite
0.57 0.29
0.28 0.22
0.21 0.11
towards the side pocket. The presence of the mirror plane orthog-
onal to [1 0 0] imposes that in isoxazole ring C8 crystallographic
site can host either C or N atoms whereas in C9 site O or C atoms
can be hosted: owing to the position of SMX molecule, the real
symmetry of MOR-SMX is lower than Cmc21.
The isoxazole and aniline rings form a typical V configuration
with the torsion angle SAN2AC7 of 123.8°, which is slightly larger
than that reported for crystalline SMX [28] (121.1°). The positions
of oxygen atoms bound to the sulfur and those of methyl group are
not localized, due to a strong disorder due to sorption.
The refined distances of SMX atoms with the framework oxy-
gens (C9AO9 = 2.80(4) Å, C8/N3AO4 = 2.89(4) Å, respectively)
indicate that these species strongly interact with the framework
thus altering the framework geometry. After SMX adsorption, the
12MR channel widens and becomes more circular thus increasing
its CFA (Fig. 5). The 12MR widening causes a narrowing of both
8MR channel and 8MR of the side pocket (see CFA values in
Fig. 5). This widening/contraction mechanism is responsible of
the small unit cell volume variation reported in Table 1S.
3.3.3. ZSM-5 zeolite
As reported in Table 1S, variations of the unit cell parameters in
comparison to those of the untreated material [20] indicated a
monoclinic (P21/n) orthorhombic (P212121) phase transition
depending on the nature and/or amount of encapsulated mole-
cules. Rietveld structure refinement revealed the incorporation of
about 1.5 antibiotic molecules per unit cell: the aniline ring being
at the intersection of straight and sinusoidal channels, and the
isoxazole ring protruding towards the sinusoidal channel.
The positions occupied by the oxygen atoms bound to the sulfur as
well as those of methyl group in ZSM-5-SMX, appear strongly disor-
dered and consequently are not localized. The torsion angle SAN2AC7
in SMX loaded into the zeolite was about 118.7°, which is smaller than
in pure crystalline antibiotic (121.1° [28]). The distances between N1
154 S. Blasioli et al. / Journal of Colloid and Interface Science 419 (2014) 148–159

Fig. 5. Location of sulfamethoxazole (SMX) and oxygen–oxygen distance (Å) of the apertures in mordenite (MOR) after drug adsorption, assuming an oxygen ionic radius of
1.35 Å. Crystallographic Free Areas (CFA) and ellipticity (e) are also reported.

Fig. 6. Location of sulfamethoxazole (SMX) and oxygen–oxygen distance (Å) of the apertures in ZSM-5 after drug adsorption, assuming an oxygen ionic radius of 1.35 Å.
Crystallographic Free Areas (CFA) and ellipticity (e) are also reported.
 S. Blasioli et al. / Journal of Colloid and Interface Science 419 (2014) 148–159 155

and N3 with respect to framework oxygen atoms (N1AO2 = 2.90(8) Å,
N1AO20 = 2.76(9) Å, N3AO25 = 2.60(11) Å) suggested interactions of
SMX with the framework.
The incorporation of SMX molecules in the ZSM-5 structure is
confirmed by a change in the dimension of 10MRs, when compared
to the parent zeolite (Fig. 6). After adsorption both sinusoidal and
straight 10MR channels become more expanded and their CFA in-
crease (Fig. 6), thus indicating a framework flexibility of ZSM-5.
3.4. FTIR
FTIR spectra of the three zeolites and SMX bound to Y, MOR and
ZSM-5 zeolites are compared in Fig. 7. Absorbance FTIR spectra re-
corded in vacuum (without water), DRIFT spectra (with a moderate
water amount present), and ATR-FTIR spectra (with a large water
amount present) are reported to evaluate whether the host–guest
zeolite-SMX interactions were affected by water.
The interpretation of IR spectrum of SMX in CH2Cl2 and ad-
sorbed onto the zeolites was facilitated by comparing the observed
absorptions with the theoretical frequencies computed for the
amidic form of the isolated molecule as shown in Fig. 7. Sulfa drugs
can exist as amide (ASO2ANHA) or imide (ASO2AN@) form of
monomeric or dimeric species [5,6,20]. In the case of SMX, the
experimental IR in CH2Cl2 matches the amide computed spectrum
in good agreement with the stabilization energy of the amide spe-
cies determined by DFT calculations which was 8.00 kcal/mol low-
er than the imide form. The most important SMX vibrational
frequencies are listed in Table 4.
The experimental bands of the antibiotic in CH2Cl2 at 3503 and
3406 cm1 were assigned to the asymmetric and symmetric
stretching of ANH2 group (masNH2 and msNH2, respectively),
whereas the absorption at 3346 cm1 to the strectching of the sul-
fonamide ANH group (mNH). The absorption at 1624 cm1 is as-
signed to the bending of ANH2 (dNH2), and the features at lower
wavenumbers are due to phenyl or heterocycle ring vibrations cou-
pled with the bending modes of amino and amide groups, along
with deformation modes of the methyl group. The strong OAH
stretching bands in the 3700–3000 cm1 range of water adsorbed
onto both bare and loaded zeolites obfuscated observation of the
SMX bands. In order to minimize the interference from H2O, spec-
tral interpretations of host–guest interactions were based upon
spectra obtained under vacuum.
3.4.1. Zeolite Y-sulfamethoxazole system
Considering the absorption spectrum of loaded zeolite Y, the
bands corresponding to the stretching and bending modes of the
aniline amino group of the adsorbed antibiotic are slightly shifted
with respect to the drug in CH2Cl2 (Table 4) indicating that the NH2
group has weak interactions with the zeolite, in agreement with
the localization by Rietveld analysis of this ring in the center of
the cage and far away from zeolite walls.
A more relevant downshift observed for the sulfonamide
stretching mode (DmNH: 106 cm1) is indicative of either weak
H-bond involving the sulfonamide group or a geometry modifica-
tion of the adsorbed molecule. Although the distance between
the sulfonamide nitrogen and the cage wall oxygens is short en-
ough to give rise to strong H-bonding, the poor directionality of
the NH group with framework oxygens likely leads to a weak H-
bond. The IR features are consistent with a zeolite-induced change
in SMX geometry.

Fig. 7. DRIFT and in vacuum absorbance FTIR spectra of zeolite Y, MOR, and ZSM-5 before (dotted curve) and after (solid curve) sulfamethoxazole (SMX) adsorption. The
difference spectrum (gray curve) of the drug adsorbed into MOR obtained by subtracting the bare outgassed sample is reported. ATR spectrum of Y-SMX adduct in water is
also presented. Reference spectra of SMX in CH2Cl2 (black curve) and calculated (red curve) are also reported. Asterisk indicates unreliable band due to solvent subtraction.
(For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
156 S. Blasioli et al. / Journal of Colloid and Interface Science 419 (2014) 148–159

Table 4
Calculated and experimental IR frequencies of sulfamethoxazole before and after adsorption onto high silica zeolites (n.d.: not detectable; –: absent).

Assignment IR wavenumber (cm1)

Computed In CH2Cl2 Adsorbed on
Y MOR ZSM-5
masNH2 3707 3503 3494, 3478 3516 3520
msNH2 3594 3406 3392 3422 3424
mNH 3565 3346 3240 n.d. 3356
dNH2 1676 1624 1622 1629 n.d.
mNC/CChet ring + dNH 1664 1614(sh) 1635–1610 1616 1635–1610
mPhquad + dNH2 1652 1597 1596 1597 1598
mPhsext 1543 1504 1503 1502 1508
mNC/CChet ring + dNH + mCACH3 1542 – – – –
mCChet ring + dasCH3 1508 1461 1464 1468 1470
mCChet ring + dasCH3 1485 – – – –
mPhsext + masSO2 + dNH 1389 1384 1384 – –
mPhsext + masSO2 + dNH 1371 1375 1373 – –
dCCHin-plane + mCANH2 1341 – – – 1340
mCOhet ring + masSO2 + dNH 1318 1311 1333, 1323 n.d. 1340, 1321
mCN/COhet ring + dNH + dasCH3 1304 – 1308 – –

At lower frequencies, the most significant spectroscopic change
is due to the unresolved band of the antibiotic in CH2Cl2 at
ca.1614 cm1, assigned to the heterocycle ring CAN vibrations
coupled with the sulfonamide dNH mode, which is increased in
intensity, upshifted and overlapped to the dNH2 band at
1622 cm1 after adsorption into zeolite. A similar feature has been
observed for other sulfa drugs loaded into zeolite Y [5,6] and is
consistent with the occurrence of van der Waals interactions
between SMX heterocycle ring and the zeolite framework.
In the 1350–1250 cm1 region a complex IR band is found for
the adsorbed SMX that is absent in the spectrum in of SMX in
CH2Cl2 solution. Likely, this difference can be mainly ascribed to
modifications of the geometry of the sulfonamide moiety as a
consequence of the heterocycle ring interaction with the zeolite
framework.
Insight of the effect of water on the drug adsorption into zeolite
Y can be obtained observing the ATR-FTIR spectrum of loaded zeo-
lite Y recorded in water solution (Fig. 7). Interestingly, the band of
the drug in CH2Cl2 at 1384 cm1 due to vibrations of the phenyl-
sulfonamide moiety (Table 4) is observed upshifted at 1402 cm1
in the ATR spectrum maybe due to the solvatation effect of the sul-
fonamide group. Moreover, the broad band of SMX in CH2Cl2 at
1311 cm1, which is assigned to deformations of the heterocycle
ring coupled with the SO2 asymmetric stretching and the NH bend-
ing, is found split into two new bands at 1339 and 1327 cm1 in
the ATR spectrum of the loaded zeolite. Likely, the solvation effect
of water modifies the electron distribution of the drug in adsorbed
form with respect to the molecule in CH2Cl2.
band in the 3400–3000 cm1 range with maximum at ca.
3300 cm1). This is a clear-cut evidence of H-bond interactions of
medium strength, which occur between silanols and SMX.
In the silanol-drug adduct, only the msNH2 at 3422 cm1 is
clearly visible. The weak band of masNH2 can be found in the differ-
ence spectrum (red dotted curve) at 3516 cm1. The small upshift
of the NH2 group when the antibiotic is in adsorbed form, indicates
that this group is far apart from the silica walls in agreement with
the fact that the aniline ring protrudes toward the MOR channel
center (Fig. 5). Unfortunately, the band of the stretching mode of
sulfonamide NH is not clearly measurable for the adsorbed drug
because it is overlapped to the intense and broad band of MOR
silanols.
At lower wavenumber, the band at 1614 cm1 of SMX in CH2Cl2
related to the stretching of the heterocycle ring coupled with sul-
fonamide dNH was strongly perturbed upon adsorption. Moreover,
the two bands at 1384 and 1375 cm1 in CH2Cl2 almost completely
disappear in loaded MOR, likely because the heterocycle ring is H-
bonded to zeolite silanols and the collective vibrations of this part
of the molecule are strongly perturbed. This feature is consistent
with the localization of the antibiotic heterocycle ring partly inside
the MOR side pocket (Fig. 5) with both the isoxazole ring nitrogen
and oxygen atoms at H-bond distance with the oxygens of the MOR
side pocket entrance. A similar scenario was already observed for
sulfachloropyridazine adsorbed into the same MOR zeolite [20].
The similarity of the absorbance (outgassed) and DRIFT spectra
rule out a significant water involvement in the stabilization of
the host–guest interactions observed.

3.4.2. MOR-sulfamethoxazole system 3.4.3. ZSM-5-sulfamethoxazole system

The bare MOR shows a broad and intense band in the region be-
tween 3750 and 3000 cm1 where silanol groups and bound H2O
absorb (Fig. 7, dotted red curve). This high intensity is attributable
to the large amount of silanol defects produced by the dealumina-
tion process. A defined and weak band due to isolated silanols in-
side microporosities is found at 3728 cm1 whereas the most
intense and broad band in the range 3700–2900 cm1 is assigned
to H-bonded donor silanols [32].
After antibiotic adsorption (Fig. 7, solid red curve), the band at
3728 cm1 almost completely disappears whereas the intense
band of H-bonded silanols is downshifted and enlarged at the same
time. These findings indicate that a fraction of silanols are bound to
the antibiotic. Clearly the band of silanols is eroded after antibiotic
adsorption (negative band at 3750–3400 cm1 in the difference
spectrum) and a new band is formed at lower frequencies (positive
In the absorbance spectrum of the loaded zeolite, the bands as-
signed to the stretching modes of the drug NH2 group are slightly
upshifted with respect to the SMX in CH2Cl2 (Table 4): similarly to
the MOR, these small perturbations reveal that some H-bondings
occur in CH2Cl2 and these are turned off in zeolite. Also the band
related to the stretching mode of the sulfonamide NH is upshifted
after adsorption with respect to the free molecule, ruling out the
involvement of this group in H-bonds with the zeolite network.
At lower frequency, the two bands at 1384 and 1375 cm1,
which are active in CH2Cl2, almost completely disappear in the
spectrum of the adsorbed drug similarly to the MOR-drug case:
also in the ZSM-5-SMX system the heterocycle ring is the drug
moiety more involved in the adsorption process, being localized
into the zeolite sinusoidal channel. Moreover, two bands of the
heterocycle ring vibrations at 1340 and 1321 cm1, which are not
 S. Blasioli et al. / Journal of Colloid and Interface Science 419 (2014) 148–159 157

resolved in CH2Cl2, become active and well resolved in the spectra
of the adsorbed antibiotic. These bands are related to vibrations of
the heterocycle ring which is localized into the smaller sinusoidal
channel and therefore they are found strongly perturbed with re-
spect to the molecule in CH2Cl2. The water amount occurring in
DRIFT samples does not seem to affect the nature of the host–guest
ZSM-5-SMX interactions based on similarity between the absor-
bance (water absent) and DRIFT (some water present) spectra.
3.5. Solid state NMR
Solid state NMR spectroscopy is a powerful method for the
study of host–guest adducts, since it provides information about
the structure and dynamics of both guest and host over a wide-
range of time scales, length scales and temperature scales. In the
1
H MAS NMR spectrum of pure sulfamethoxazole (SMX, Fig. 8(I)),
two broad signals in the range of 0.5–2.5 and 5.5–9.5 ppm due to
methyl and aromatic protons are found. The broadness is due to
the overlapping of resonances arising from different polymorphs
and/or due to the non-averaging of the 1HA1H dipolar interactions.
On the other hand, the 1H MAS NMR spectra of the 3 adducts
show a good resolution for a solid sample and consist of resonances
from SMX, physisorbed H2O as well as the host zeolites. Fig. 8(I)
show the 1H MAS NMR spectra of adducts both before and after
dehydration in vacuum at 120 °C for 2 h. For adduct Y-SMX, the
peaks from the aromatic and heterocyclic rings are well resolved
compared to that of pure SMX and are associated to the averaging
of the 1HA1H dipolar interactions. Resonance due to methyl group
appears at 2.3 ppm. Y-SMX exhibits similar spectra before
(Fig. 8(I)a) and after dehydration (Fig. 8(I)b), except the fact that
there is a difference in the region where water resonate. Before
dehydration, a broad peak at 3.9 ppm is clearly visible and is due
contributions from physisorbed water, ANH2 and ANH. In addi-
tion, isolated silanols appear as a shoulder of the main peak at
1.7 ppm after dehydration.
The behavior of MOR-SMX adduct is entirely different from
that of Y-SMX. Before dehydration, the resonance due to protons
of water, ANH2 and ANH is dominant in this spectrum (Fig. 8(I)c)
and the amount of adsorbed water is significantly higher than in
Y-SMX. After dehydration (Fig. 8(I)d), resonances due to ANH2
and isolated silanols are visible at 4.2 and 1.7 ppm, respectively.
Furthermore, a dramatic increase in intensity for the methyl res-
onance after dehydration is evident and can be due to the reso-
nance contributions from ANH as well as isolated silanol
resonances being in the same chemical shift range. This demon-
strates that both silanol and ANH groups are H-bonded to water
molecules in the hydrated sample. In the aromatic region, after
dehydration, three resonances can be clearly distinguished; how-
ever, the chemical shifts are different from the hydrated sample.
In addition, a very broad low intense peak in the range 12–
18 ppm is also visible in dehydrated MOR-SMX (Fig. 8(I)d, inset)
and are associated to H-bonded protons (highlighted by an ar-
row). It is clearly evident that water plays a dramatic role in
the arrangement of SMX molecules inside the mordenite cages
and drug molecules reorganize after dehydration as water leaves
the channels/cages.

Fig. 8. 1H solid state MAS-NMR spectra of pure sulfamethoxazole (SMX), Y-SMX, MOR-SMX and ZSM5-SMX at a MAS rate of 15 kHz (I) before (a, c, e) and after dehydration (b,
d, f). Inset show the zoomed 1H MAS spectra of MOR-SMX before (red) and after dehydration (black). 13C solid state CPMAS-NMR spectra of SMX, Y-SMX, MOR-SMX and
ZSM5-SMX (II) before (a, c, e) and after dehydration (b, d, f). A cross polarization contact time of 2 ms and a MAS rate of 10 kHz was used in all the experiments. Inset show the
zoomed methyl region of 13C CPMAS spectra of ZSM5-SMX before (red) and after dehydration (black).  Denotes spinning sidebands. (For interpretation of the references to
color in this figure legend, the reader is referred to the web version of this article.)
158 S. Blasioli et al. / Journal of Colloid and Interface Science 419 (2014) 148–159
ZSM5-SMX adduct shows a similar behavior to that of MOR-
SMX, however, the line widths are narrower. The spectra are iden-
tical before (Fig. 8(I)e) and after dehydration (Fig. 8(I)f) apart from
the fact that there is a change in the region where physisorbed
water, ANH2 and ANH resonate. After dehydration, only resonance
due to ANH2 is visible as a broad peak centered at 4.2 ppm in that
region. Additionally, methyl resonance appears as a doublet, sepa-
rated by 1 ppm apart, and confirms the presence on two distinct
environments for this group. The existence of multiple resonances
at different chemical shifts, before and after dehydration, (aromatic
protons in MOR-SMX) as well as multiple resonances for identical
protons, before and after dehydration (methyl protons in ZSM5-
SMX), suggests that SMX molecules exist in a number of different
chemical environments within the host–guest interfaces. This
may arise either from the differences in the proximity between
heterocyclic ring (and/or aniline ring) and zeolite walls or from dif-
ferences in the packing order. Further evidences are available from
the 13C CPMAS NMR data.
Standard CPMAS conditions at ambient temperature were used
to acquire the 13C[1H] CPMAS NMR spectra (before and after dehy-
dration) with short contact time of 2 ms and Fig. 8(II) shows spec-
tra acquired on pure SMX and the three different adducts. The
assignments are based on theoretical calculations as well as on li-
quid NMR data (not shown). The sharp and narrow resonances are
consistent with the highly ordered crystalline nature of the pure
SMX sample [Fig. 8(II), topmost]. However, the behavior of the
three adducts are different to that of pure crystalline SMX.
The 13C CPMAS spectrum of Y-SMX before dehydration
(Fig. 8(II)a) shows sharp resonances for each carbon particularly
in the aromatic region, however, peaks broaden after dehydration
(Fig. 8(II)b). The peak positions are different from that of the crys-
talline pure SMX, with the number of resonances as well as chem-
ical shifts being different in the aromatic region. This observation
shows that there is a fast flipping of rings in the cages of zeolite
Y in the hydrated state, probably assisted by water, leading to an
averaging of the resonance peaks from the aromatic carbons,
whereas in the crystalline solid, the particular packing of mole-
cules implies a crystallographic unequivalence of the various 13C
sites. The distinct behaviors in terms of observed line widths in
Y-SMX adduct, before and after dehydration can be explained by
steric considerations. It is important to note here that the SMX
loading in zeolite Y is around 24 wt% (TGA results, Table 2). At such
high loading, it is expected that the SMX molecules are engaged in
the zeolite cages probably through both host–guest and guest–
guest interactions. Molecular motion (e.g., flipping of rings and
methyl group rotation) of SMX inside the cages at RT would disrupt
the molecular orientation and the motional rate is fast enough to
average the chemical shift anisotropy (CSA) of the aromatic
carbons as it is seen by sharp resonances in the hydrated state.
However, such CSA averaging does not take place when the SMX
molecules are physisorbed on the cage surface (as water molecules
leave the cages) and broad peaks are considered to be due either to
a restriction of mobility or to a distribution of environments.
The key difference in 13C CPMAS spectrum of MOR-SMX with
that of Y-SMX is that the resonances are generally broad even in
the hydrated state (Fig. 8(II)c). After dehydration (Fig. 8(II)d),
changes in chemical shifts as well as line widths are observed. As
was observed in the 1H MAS NMR spectra of MOR-SMX adduct,
water plays a cooperative role in the guest–host interactions be-
tween SMX and mordenite cages. As the water molecules leave,
SMX reorganizes inside the cages with heterocyclic ring being
interacting with the cage surface through hydrogen bonding
(Fig. 8(I)d, inset).
As for the remaining ZSM5-SMX adduct, the spectra (Fig. 8(II)e
and f) are very different from the other two adducts but similar to
13
C CPMAS spectrum of pure crystalline SMX with matching peak
positions. However, for the aromatic 13C pairs, namely C3/C5 and
C2/C6, only one resonance each is observed in this adduct and
can be attributed to the fast flipping of aromatic ring. While such
good resolution with sharp lines can be attributed either to a great-
er mobility of SMX inside the cages or to a more regular packing
with less equivalent orientations as in crystalline samples,
although long range ordering is highly unlikely in ZSM5 due to
the restriction of space. However, we have observed a splitting in
the methyl carbon resonance which confirms the distinct environ-
ments this group experiences (Fig. 8(II)f, inset). It is important to
note here that, in ZSM5-SMX, there are no visible changes observed
in the spectra before and after dehydration. As was pointed out in
the XRD data recorded at RT, the methyl group position in ZSM5-
SMX was highly disordered and could not be localized. Further-
more, XRD data show that the aniline ring is at the intersection
of straight and sinusoidal channels, while the heterocyclic ring pro-
trudes towards the sinusoidal channel. In fact, the methyl group
can experience different environments in such arrangement. Thus,
we have performed a low temperature 13C CPMAS experiment
(data not shown), and the data recorded at 193 K show that the
two resonances coalesce into one signal. It can be concluded that,
based on the combined 1H MAS and 13C CPMAS NMR as well as
XRD data, SMX molecules occupy crystallographically unique posi-
tion in framework channel intersections which resemble a crystal-
line state for the drug in ZSM5 host matrix. Furthermore, water
plays no role in the arrangement of SMX molecules in ZSM5
channels.
In summary, the observation of variable adsorption rates and
capacities of SMX on different host zeolite framework is of rele-
vance, and suggests that non-covalent guest–host and guest–guest
interactions play significant role in the stability of these adduct
systems. Water might play a role in the adsorption process of anti-
biotics in zeolites, however, necessarily not control the structure
and dynamics of these molecules in the final adduct as was con-
firmed by the solid state NMR study. The exact location of the drug
molecules inside the host system depends on different factors
including, dimension of the drug molecules, the size of the cages/
channels, the presence and concentration of other guest molecules
like water, as well the non-covalent interactions between and
among guests and host systems. Clearly, this multi-technique ap-
proach allows the investigation of the adsorption of wide range
of sulfa drug families on different host zeolite systems and to
understand the structural and dynamic details of the adsorbed spe-
cies. Such unprecedented detail understanding can help to apply
this principle to clean up polluted water bodies.
4. Conclusions
The efficacy of sulfamethoxazole (SMX) removal by high silica
zeolites was evaluated based on the interaction of SMX with three
representative high silica zeolites (Y, MOR, and ZSM-5) each with a
different pore size and architecture. Zeolite Y and MOR adsorbed
SMX with a favorable kinetics (a few minutes) up to 24% and 6%
zeolite DW, respectively. No equilibrium adsorption was reached
onto ZSM-5 after two weeks, thus revealing a slow adsorption
kinetics. However, a surprisingly positive temperature effect was
highlighted which can be relevant for applications in manure/sew-
age polluted with sulfa drugs, given that the thermophilic phase in
manure/sewage composting can reach high temperatures. The
adsorption of SMX into ZSM-5 after 24 h contact at 65 °C reached
8% zeolite DW. SMX desorption revealed the irreversibility of the
adsorption process for each zeolite.
The location of SMX inside the zeolite porosities was defined by
XRPD Rietveld analysis. The close vicinity of some drug moieties to
the oxygen atoms of the zeolite voids suggests the occurrence of
 S. Blasioli et al. / Journal of Colloid and Interface Science 419 (2014) 148–159 159

multiple host–guest interactions between each sorbent and SMX. [6] I. Braschi, G. Paul, G. Gatti, M. Cossi, L. Marchese, RSC Adv. 3 (2013) 7427–7437.
[7] M. Burkhardt, C. Stamm, J. Environ. Qual. 36 (2007) 588–596.
The type and strength of the interactions, which are responsible
[8] M. Cossi, N. Rega, G. Scalmani, V. Barone, J. Comput. Chem. 24 (2003) 669–681.
for the irreversible extraction of the drug from water, were eluci- [9] R. Crisafully, M.A.L. Milhome, R.M. Cavalcante, E.R. Silveira, D. De Keukeleire,
dated by FTIR and SS-NMR spectroscopy. Medium-weak and coop- R.F. Nascimento, Biores. Technol. 99 (2008) 4515–4519.
erative H-bonds and van der Waals forces occur between zeolite [10] T.H. Dunning, J. Chem. Phys. 90 (1989) 1007–1023.
[11] M.J. Frisch, et al., Gaussian 03, Revision C.02, Gaussian Inc., Wallingford, CT,
porosity walls and the antibiotic: water seems to play a key role 2004.
in the stabilization of the host–guest high silica zeolite-SMX inter- [12] S. Fukahori, T. Fujiwara, R. Ito, N. Funamizu, Desalination 275 (2011) 237–242.
actions only for zeolite Y and mordenite. [13] H.J. Gill, J.L. Maggs, S. Madden, M. Pirmohamed, B.K. Park, Brit. J. Clin. Pharm.
42 (1996) 347–353.
In conclusion, favorable adsorption kinetics along with effective [14] P.C. Hariharan, J.A. Pople, Theor. Chim. Acta 28 (1973) 213–222.
and highly irreversible adsorption for SMX in zeolite Y pores make [15] M. Kahle, C. Stamm, Environ. Sci. Technol. 41 (2007) 132–138.
this environmental friendly material interesting for sulfa drugs re- [16] M. Kahle, C. Stamm, Chemosphere 68 (2007) 1224–1231.
[17] A.C. Larson, R.B. von Dreele, Los Alamos Natl. Lab. Rep. LAUR (2000) 86–748.
moval from water bodies, whereas the positive temperature effect [18] J.L. Martinez, Environ. Pollut. 157 (2009) 2893–2902.
revealed by ZSM-5 towards the antibiotic adsorption makes this [19] A. Martucci, L. Pasti, N. Marchetti, A. Cavazzini, F. Dondi, A. Alberti,
sorbent interesting for manure treatments also at the high temper- Microporous Mesoporous Mater. 148 (2011) 1742–1783.
[20] A. Martucci, M.A. Cremonini, S. Blasioli, L. Gigli, G. Gatti, L. Marchese, I. Braschi,
atures reached during thermophilic phase. Based on the sorption, Microporous Mesoporous Mater. 170 (2013) 274–286.
structural and spectroscopic data presented, zeolite Y showed the [21] B.F. Mentzen, Mater. Res. Bull. 27 (1992) 831–838.
highest affinity for SMX and can be thus considered optimal for [22] K. Nishi, A. Hidaka, Y. Yokomori, Acta Cryst. B 61 (2005) 160–163.
[23] C.F. McDonald, J. Parenteral Sci. Technol. 45 (3) (1991) 147–151.
sulfonamide removal from impacted waters.
[24] L. Pasti, A. Martucci, M. Nassi, A. Cavazzini, A. Alberti, R. Bagatin, Microporous
Mesoporous Mater. 160 (2012) 182–193.
Acknowledgments [25] L. Pasti, E. Sarti, A. Cavazzini, N. Marchetti, F. Dondi, A. Martucci, J. Sep. Sci. 00
(2013) 1–8.
[26] C.P. Price, A.L. Grzesiak, A.J. Matzger, J. Am. Chem. Soc. 127 (2005) 5512–5517.
Research co-funded by the Italian Ministry of Education, Uni- [27] P. Srinivasan, A.K. Sarmah, M. Manley-Harris, A.L. Wilkins, Sorption of
versity, and Research (Project: Zeolites as nano-reactors for the sulfamethoxazole, sulfachloropyridazine and sulfamethazine onto six New
environment: efficiency, selectivity and stability in the adsorption Zealand dairy farm soils, in: 2010, 19th World Congress of Soil Science, Soil
Solutions for a Changing World 1–6 August 2010, Brisbane, Australia.
of drugs from contaminated waters). G.S. Premachandra is [28] M. Takasuka, H. Nakai, Vib. Spectrosc. 25 (2) (2001) 197–204.
acknowledged for laboratory assistance. [29] B.H. Toby, J. Appl. Cryst. 34 (2001) 210–213.
[30] A.J. van der Ven, M.A. Mantel, T.B. Vree, P.P. Koopmans, J.W. van der Meer, Brit.
J. Clin. Pharm. 38 (1994) 147–150.
Appendix A. Supplementary material [31] A.J. van der Ven, T.B. Vree, E.W. van Ewijk-Benken Kolmer, P.P. Koopmans, J.W.
van der Meer, Brit. J. Clin. Pharm. 39 (1995) 621–625.
[32] I. Braschi, G. Gatti, C. Bisio, G. Berlier, V. Sacchetto, M. Cossi, L. Marchese, Phys.
Supplementary data associated with this article can be found, in
Chem. C 116 (2012) 6943–6952.
the online version, at http://dx.doi.org/10.1016/j.jcis.2013.12.039. [33] J. Gao, F.-Q. Yu, L.-P. Luo, J.-Z. He, R.-G. Hou, H.-Q. Zhang, S.-M. Li, J.-L. Su, B.
Han, Veterinary J. 194 (3) (2012) 423–424.
References [34] V. Homem, L. Santos, J. Environ. Manag. 92 (10) (2011) 2304–2347.
[35] M. Martucci, A. Alberti, L. Pasti, M. Nassi, R. Bagatin, J. Solid State Chem. 194
(2012) 135–142.
[1] R. Arletti, A. Martucci, A. Alberti, L. Pasti, M. Nassi, R. Bagatin, J. Solid State [36] I. Michael, L. Rizzo, C.S. McArdell, C.M. Manaia, C. Merlin, T. Schwartz, C. Dagot,
Chem. 194 (2012) 135–142. Water Res. 47 (3) (2013) 957–995.
[2] Ch. Baerlocher, L.B. McCusker, D.H. Olson, Atlas of Zeolite Framework Types, [37] A. Rossner, S.A. Snyder, D.R.U. Knappe, Water Research 43 (15) (2009) 3787–
6th revised ed., Elsevier, Amsterdam, 2007. 3796.
[3] A.D. Becke, J. Chem. Phys. 98 (7) (1993) 5648–5652. [38] T.A. Ternes, Water Res. 32 (1998) 3245–3260.
[4] I. Braschi, S. Blasioli, L. Gigli, C.E. Gessa, A. Alberti, A. Martucci, J. Hazard. Mater. [39] A.G. Trovò, R.F.P. Nogueira, A. Aguera, C. Sirtori, A.R. Fernandez-Alba,
17 (2010) 218–225. Chemosphere 77 (10) (2009) 1292–1298.
[5] I. Braschi, G. Gatti, G. Paul, C.E. Gessa, M. Cossi, L. Marchese, Langmuir 26
(2010) 9524–9532.