Colloids and Surfaces B: Biointerfaces 130 (2015) 101109

Contents lists available at ScienceDirect

Colloids and Surfaces B: Biointerfaces

journal homepage: www.elsevier.com/locate/colsurfb

Surface modied natural zeolite as a carrier for sustained diclofenac

release: A preliminary feasibility study
Bruno de Gennaro a, , Lilia Catalanotti a , Piergiulio Cappelletti b , Alessio Langella c ,
Mariano Mercurio c , Carla Serri d , Marco Biondi e,f , Laura Mayol e,f
a
Dipartimento di Ingegneria Chimica, dei Materiali e della Produzione Industriale, Universit di Napoli Federico II, Piazzale V. Tecchio 80, 80125 Napoli, Italy
b
DiSTAR, Universit di Napoli Federico II, Via Mezzocannone 8, 80134 Naples, Italy
c
Dipartimento di Scienze e Tecnologie, Universit del Sannio, Via dei Mulini 59/A, 82100 Benevento, Italy
d
Dipartimento di Scienze Biologiche ed Ambientali, Universit di Messina, Piazza Pugliatti 1, Messina, Italy
e
Dipartimento di Farmacia, Universit di Napoli Federico II, Via Domenico Montesano 49, 80131 Napoli, Italy
f
Interdisciplinary Research Centre on Biomaterials CRIB Universit di Napoli Federico II, P.le Tecchio 80, Napoli, Italy

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

Article history: In view of zeolite potentiality as a carrier for sustained drug release, a clinoptilolite-rich rock from
Received 23 December 2014 California (CLI CA) was supercially modied with cetylpyridinium chloride and loaded with diclofenac
Received in revised form 21 March 2015 sodium (DS). The obtained surface modied natural zeolites (SMNZ) were characterized by confocal
Accepted 24 March 2015
scanning laser microscopy (CLSM), powder X-ray diffraction (XRPD) and laser light scattering (LS). Their
Available online 3 April 2015
owability properties, drug adsorption and in vitro release kinetics in simulated intestinal uid (SIF) were
also investigated.
Keywords:
CLI CA is a Na- and K-rich clinoptilolite with a cationic exchange ability that ts well with its zeolite
Clinoptilolite-rich rock
Surfactant modied natural zeolite
content (clinoptilolite = 80 wt%); the external cationic exchange capacity is independent of the cationic
Cetylpyridinium surfactant used. LS and CLSM analyses have shown a wide distribution of volume diameters of SMNZ
Adsorption kinetics particles that, along with their irregular shape, make them cohesive with scarce ow properties. CLSM
Diclofenac release observation has revealed the localization of different molecules in/on SMNZ by virtue of their chemical
Kinetic models nature. In particular, cationic and polar probes prevalently localize in SMNZ bulk, whereas anionic probes
preferentially arrange themselves on SMNZ surface and the loading of a nonpolar molecule in/on SMNZ
is discouraged. The adsorption rate of DS onto SMNZ was shown by different kinetic models highlighting
the fact that DS adsorption is a pseudo-second order reaction and that the diffusion through the boundary
layer is the rate-controlling step of the process. DS release in an ionic medium, such as SIF, can be sustained
for about 5 h through a mechanism prevalently governed by anionic exchange with a rapid nal phase.
2015 Elsevier B.V. All rights reserved.

1. Introduction
A variety of different minerals have been used as excipients
in pharmaceutical preparations since they have certain desirable
physical and physicalchemical properties, such as high adsorp-
tion capacity, specic surface area, swelling ability, reactivity to
acids, water solubility, hygroscopicity and slightly alkaline reaction
Corresponding author. Tel.: +39 0817682551; fax: +39 0817682394.
E-mail addresses: bruno.degennaro@unina.it (B. de Gennaro),
lilia.catalanotti@unina.it (L. Catalanotti), piergiulio.cappelletti@unina.it
(P. Cappelletti), alessio.langella@unisannio.it (A. Langella),
mariano.mercurio@unisannio.it (M. Mercurio), carlaserri@virgilio.it (C. Serri),
marco.biondi@unina.it (M. Biondi), laura.mayol@unina.it (L. Mayol).
(pH). Clearly, such minerals must not be toxic to humans. Oxides,
carbonates, sulphates, chlorides, phosphate and phyllosilicates are
commonly used as excipients in pharmaceutical preparations [1].
More recently, some tectosilicates, such as zeolites, have also fea-
tured in pharmaceutical preparations due to their adsorptive and
ion exchange properties [24].
Clinoptilolite application in human medicine has proven use-
ful in the external treatment of skin wounds and athletes foot,
in kidney dialyses for the removal of ammonia ions from body
uids, as antidiarrheal remedies, or as an active ingredient in
antacid drugs [57]; furthermore clinoptilolite has also been
reported as having an antimicrobial, antitumor and antiviral action
[811].
Mineralorganic interaction can also be used to sustain
the release of active ingredients to improve their therapeutic

http://dx.doi.org/10.1016/j.colsurfb.2015.03.052
0927-7765/ 2015 Elsevier B.V. All rights reserved.
102 B. de Gennaro et al. / Colloids and Surfaces B: Biointerfaces 130 (2015) 101109
properties. Here, the minerals rst work by carrying and then
releasing the drug. Zeolites possess a great potential as drug carriers
due to their large specic surface area and high adsorption capacity
[12]. In order to enhance the adsorption of drug molecules, cationic
surfactants have been used to modify the surface properties of zeo-
lites. Surfactants can replace only the native inorganic cations at
the external surface of the zeolite (external cation exchange capac-
ity ECEC), while bilayers are formed with sufcient quantities
of surfactants. Thus, adsorption of cationic surfactants onto zeolitic
surface usually includes ion exchange (when the quantity of surfac-
tant is less than the zeolite ECEC) and, in addition to ion exchange,
long chain hydrophobic interaction (when the amount of surfac-
tant is greater than the zeolite ECEC). This bilayer formation results
in a charge reversal on the external zeolite surface, providing sites
where anions will be retained and cations repelled, while neutral
species can partition into the hydrophobic core [13]. Additional
cation exchange capacity remains on the internal surfaces of the
zeolite. Thus, modication of natural zeolites with an organic sur-
factant gives rise to a product that is able to sorb organic compounds
and exchange inorganic cations and anions, while retaining favor-
able hydraulic properties [14]. In this way, adsorption of surfactants
at the solidliquid interface results in an increased hydrophobicity
of the mineral surface providing a high afnity for drug molecules.
On this basis, the aim of this work is to verify the possible use of
a clinoptilolite-rich rock from California as a potential carrier for
molecules of pharmaceutical interest. Indeed, recent literature tells
us that clinoptilolite-based tuffs that have been supercially modi-
ed with surfactants, such as cetylpyridinium chloride (CP-Cl), have
proven able to load and sustain the release of a model anionic drug
such as diclofenac [1517]. Although these tuffs have undergone
an extensive technological characterization, a description of their
other important features in terms of pharmaceutical applications
is still lacking. This is why, in this work, we intend to investigate
zeolite powder properties, such as their uidity/compressibility,
which is a crucial feature for the fabrication of oral dosage forms.
Another aim is to shed light on the mechanisms of drug load-
ing on/release from surface modied zeolite materials by means
of model uorescent probes and subsequent confocal microscopy
analysis. It should also be emphasized that differences in zeolite
geographical origin and cationic content can behave differently in
terms of exchange mechanism and, therefore, the formation of the
bilayer.
For this purpose, modication of the zeolite surface, accord-
ing to external cation exchange capacity of the zeolite-rich
rock, was performed, using cetylpyridinium chloride (CP-Cl) as
cationic surfactants, while diclofenac sodium (DS) was used as
the model drug since it has been already used in previous
similar studies and since it possesses numerous pharmaceuti-
cal applications for both oral and topical route of administration
[1517]. In actual fact, DS is a non-steroidal anti-inammatory
drug (NSAID) used to reduce inammation and/or as an anal-
gesic, reducing pain in certain conditions. The name diclofenac
derives from its chemical name: 2-(2,6-dichloranilino) pheny-
lacetic acid, and it is supplied as, or contained in, medications
under a variety of trade names [18]. It is used to treat pain,
acute migraines and inammatory disorders, such as muscu-
loskeletal complaints, osteoarthritis, dental pain or dysmenorrhea
[1921]. DS has been found effective against all strains of mul-
tidrug resistance. It may therefore have the capacity to treat
uncomplicated urinary tract infections caused by Escherichia coli
[22].
Surface modied natural zeolites (SMNZ) were characterized by
confocal scanning laser microscopy (CLSM), powder X-ray diffrac-
tion and laser light scattering (LS). Their owability properties, drug
adsorption and in vitro release kinetics in simulated intestinal uid
(SIF) were also investigated.
2. Materials and methods
2.1. Materials and their characterization
The geomaterial used in this study for the surfactant modied
zeolite preparation was a clinoptilolite-rich rock from California
(United States) marketed by St. Cloud Mining Company (CA, USA).
Rhyolitic tuffs and sedimentary rocks of the Miocene and Pliocene
ages crop out in the southern part of the Ash Meadows (CA, USA).
These rocks, part of 16- to 10-Ma volcanic deposits of the south-
western Nevada volcanic eld, are widely zeolitized and virtually
pure deposits of clinoptilolite occurring in a vitric, non-welded, ash-
ow tuff [23].
Rock mineralogy was investigated by X-ray powder diffrac-
tion (XRPD), using a Panalytical XPert Pro diffractometer equipped
with RTMS detector (XCelerator), XPert High Score Plus 3.0 soft-
ware. Quantitative mineralogical characterization was carried out
using the RIR/Rietveld method with internal standard, by means
of TOPAS 4.2 software (BRUKER AXS Company). Microporosi-
metric characterization was also carried out by N2 adsorption
at 77 K, and the specic surface area was evaluated using the
BrunauerEmmettTeller (BET) method. A Micromeritics ASAP
2020 volumetric instrument was used for this purpose, and samples
were degassed at 473 K for 4 h prior to characterization.
The batch exchange method (BEM) was applied to assess the
cation exchange capacity (CEC) of the tuff sample by extracting
Na+ , K+ , Mg2+ and Ca2+ cations with a solution of ammonium
chloride (Aldrich, assay 99.5%, CAS [12125-02-9]) [24]. The exter-
nal cation exchange capacity (ECEC) was estimated by contacting,
under continuous stirring in batch conditions, using a suitable
amount of zeolite sample (2.5 g) in contact with 10 mL of 20-mM
surfactant solution in a 50-mL polyallomer centrifuge tube. Every
24 h, the exhausted solution was replaced with a fresh one and
this procedure was repeated for three days. The samples were
then centrifuged and the supernatants analyzed for cation con-
centrations via ICP-OES (Perkin Elmer Optima 2100 DV). All batch
experiments were performed in triplicate. The selected surfactant
was cetylpyridinium-chloride (CP-Cl) (Sigma, assay 99.0102.0%,
CAS [6004-24-6]), for which the sorption plateau, i.e. the max-
imum surface exchange amount, was previously evaluated. The
anion-exchange capacity (AEC) of the obtained surfactant mod-
ied clinoptilolite-rich tuff was estimated by contacting 0.5 g of
these samples for three days, at 25 C and under continuous stirring,
with 50 mL of solutions containing NO3 , having concentrations
equal to 100 mM. Three different replicates were prepared for each
exchange and the solutions were changed and recovered every 24 h.
The anion concentrations were analyzed via HPLC and the AEC of
the sample was determined.
DS (CAS [15307-79-6]), spray dried lactose, starch, microcrys-
talline cellulose (CMC) and colloidal silica were from Farmalabor
(Italy). Poly(vinyl alcohol) (PVA, Mowiol 4088) (CAS [9002-89-
5]), the salts used to prepare the phosphate buffered solutions (PBS)
and simulated intestinal uid (SIF), and the uorescent probes used
for CLSM observations, i.e. toluidine blue (TB) (dye content 80%,
CAS [92-31-9]), nonpolar Nile red (NR) (CAS [7385-67-3]), uo-
rescein isothiocyanate isomer ICelite (FITC) (CAS [3326-32-7]),
rhodamine 6G (Rhod) (dye content 95%, CAS [989-38-8]) were pur-
chased from SigmaAldrich (USA).
2.2. Size and size distribution
Laser light scattering (Coulter LS 100Q, USA;  = 750 nm) on
a dispersion of bare zeolite (NZ) and surface-modied natural
zeolites (SMNZ) in 0.5% w/v aqueous PVA solution enabled the
determination of mean volume diameter and size distribution of
 B. de Gennaro et al. / Colloids and Surfaces B: Biointerfaces 130 (2015) 101109
the particles. Average diameter values were calculated on three
independent samples.
2.3. Confocal laser scanning microscopy (CLSM)
In order to investigate whether the active molecule localizes on
the SMNZ surface and/or into the bulk, model uorescent probes
were loaded with the same procedure used for DS and reported in
Section 2.6. CLSM analyses were carried out on an LSM 510 Zeiss
confocal inverted microscope equipped with a Zeiss 40 Apochro-
mat objective lens (Carl Zeiss, Germany). Excitation wavelengths
were 488 nm for FITC, 543 nm for Rhod and NR and 633 nm for TB.
Cross sections of SMNZs were obtained by merging uorescence
and transmission images.
2.4. Determination of Carrs index
In view of SMNZ formulation as oral dosage forms, it is crucial
to evaluate the cohesion, size uniformity and surface area of their
powders. The parameter that indirectly correlates to those prop-
erties is the Carrs index, CI(%) which is an indicator of powder
compressibility and therefore a measure of the relative signi-
cance of inter-particle interactions. CI is expressed as the quotient
between the apparent (A ) and tapped (T ) densities, as described
by the following equation [25]:
A T
CI (%) = 100
A
In Eq. (1), A is the mass of powder particles divided by the
total volume they occupy immediately after being poured into a
graduated cylinder. More specically, the total volume includes
particle volume, inter-particle void volume and internal pore vol-
ume. The tapped density T is obtained by dividing the powder
mass by the volume occupied after mechanically tapping the cylin-
der containing the sample until little further volume decrease (<2%)
is observed. This test was performed with an automatic tapping
apparatus described in the European Pharmacopoeia. CI(%) was
determined for bare natural zeolites (NZ), SMNZ and DS-loaded
SMNZ by triplicate experiments.
2.5. Surfactant modied natural zeolite (SMNZ) preparation
In order to avoid any release of cations that may adversely
interact with the human body (having already demonstrated a neg-
ligible release of heavy metal cations), the mainly Na+ form of the
clinoptilolite-rich sample was prepared by ion exchange with 1 M
NaCl (Baker, CAS [7647-14-5]) aqueous solution. The system was
stirred for 12 h at room temperature and the process was repeated
until the concentrations of all outgoing cations had reduced to
below 1 ppm, indicating that all the exchangeable cations had been
replaced by sodium. Finally, the sample was washed and air-dried.
The surface modication of CLI CA with CP-Cl (CLICP) was car-
ried out by mixing, with a high speed disperser at 10,000 rpm and
xed time (about 6 h), a suitable amount of zeolite rock sample
in a solution containing the ionic surfactant in a S/L ratio equal
to 1/40 [26]. In order to verify the maximum surfactant sorption
on the CLI CA surface, the sorption equilibrium study was per-
formed. For this purpose, CLI CA sample was treated with CP-Cl
solution by contacting 2.5 g of raw zeolite and 20 mL of a surfac-
tant aqueous solution, placed into 50 mL Nalgene centrifuge tubes
under continuous stirring and for 24 h, which was evaluated suf-
cient to achieve equilibrium. The initial surfactant concentrations
ranged from 25% to 400% of the sorption plateau as a function of
the ECEC. The mixture was then centrifuged to yield a clear super-
natant solution, which was analyzed via liquid-chromatography
for later analysis (Dionex DX120). From the surfactant sorption
103
isotherm, in order to prepare the surfactant modied CLI CA, the
initial surfactant concentration equal to 150% of the sample ECEC
has been xed. After dispersion, the suspensions were ltered and
the supernatants used to determine the cations released.
2.6. In vitro drug loading/release tests
The surfactant/zeolite composites were tested in order to inves-
tigate the DS sorption capacity in aqueous solutions. For this
purpose, sorption kinetic test and release one were performed.
Kinetic studies were carried out in batch experiments at room
temperature by shaking the reaction mixture containing 2.5 g
of CLICP in 500 mL of DS aqueous solution having concentra-
tion of 400 mg, corresponding to the sample ECEC, since ion
exchange is assumed to be the main loading mechanism, buffered
at pH 7.4 and at room temperature [15]. To check the effect of
ions in the loading medium on the adsorption of DS on SMNZ,
loading experiments were carried out by dissolving the drug in
distilled water buffered with PBS (DWPB) or distilled water con-
taining NaOH (DWNB), to adjust the pH to 7.4. At scheduled
time intervals, aliquots (8 mL) of the supernatant were withdrawn
and centrifuged; the unloaded drug was quantied by means
of reversed-phase high-performance liquid chromatography (RP-
HPLC). The chromatograph was equipped with an HPLC LC-10AD
pump (Shimadzu, Milan, Italy), a 7725i injection valve (Rheodyne),
an SPV-10A UV-vis detector (Shimadzu) set at the wavelength of
(1)
281 nm and a C-R6A integrator (Shimadzu). RP-HPLC experiments
were carried out using a Luna 5 m C18 column (Phenomenex,
Klwid, USA; 250 4.6 mm, 100 A porosity). The mobile phase was a
70/30 v/v mixture of acetonitrile and water at neutral pH. The ow
rate was set to 1 mL/min and run time was 10 min. The linearity of
the response was assessed in the 0.150 g/mL range (r2 > 0.99).
In vitro release experiments were carried out using a standard
sampling-separation method. DS loaded SMNZ were dispersed in
simulated intestinal uid (SIF), prepared as described in the United
States Pharmacopoeia 25 Ed. and later editions (68.05 g KH2 PO4 ,
8.96 g NaOH, diluted in distilled water to 10 L) [27]. A total of 15 mg
of drug loaded SMNZ was added into 10 mL of SIF and placed in
a thermostatic bath at 37 C under continuous shaking (100 rpm).
The samples were kept in the thermostatic bath until a plateau in
the drug released percentage was found. DS loaded SMNZ/SIF ratio
was chosen in order to ensure sink conditions. At scheduled time
points, 5 mL of supernatant were withdrawn after centrifugation at
9000 rpm (room temperature, 5 min) and replaced with the same
volume of fresh medium. DS was quantied by spectrophotometric
assay (UV-1800, Shimadzu Laboratory World, Japan) at 275 nm. The
linearity of the response was assessed in the 0.210 g/mL range
(r2 > 0.99). Experiments were run in triplicate.
3. Results and discussion
3.1. Material characterizations
Table 1 reports the quantitative mineralogical of the selected
geomaterial and chemical analyses of the clinoptilolite pure phase.
Based on quantitative XRPD, the clinoptilolite content was about
80 wt% along with cristobalite (3 wt%), K-feldspar (2 wt%), quartz
(traces) and a micaceous phase (2 wt%). Chemical analyses of
the pure zeolite samples, apart from the common high silica
content (66.75 wt%) can be distinguished on the basis of the extra-
framework cations. In actual fact, in contrast with the European
clinoptilolites, which are usually greatly enriched with Ca2+ and
K+ , the CLI CA is a Na- and K-rich phase (Table 1).
The CEC of the clinoptilolite-rich sample, evaluated using BEM,
resulted in 1.97 meq/g, coherent with the zeolite content. In actual
104 B. de Gennaro et al. / Colloids and Surfaces B: Biointerfaces 130 (2015) 101109

Table 1
Mineralogical and pure zeolite chemical composition of the selected CLI CA samples.
The errors in mineralogical assessment are reported in brackets.

CLI CA wt.% Clinoptilolite wt.% Sodium form

(wt.%)

Clinoptilolite 80 (4) SiO2 66.75 67.95

Cristobalite 3 (1) Al2 O3 11.32 11.51
K-feldspar 2 (1) TiO2 0.04 0.04
Quartz tr. Fe2 O3 0.10 0.1
Mica 2 (1) MgO 0.11 0.04
Amorphousa 13 (4) CaO 0.44 0.03
MnO 0.06 0.05
Na2 O 3.32 6.51
K2 O 4.80 0.81
H2 Ob 13.07 12.99
Si/Al 5.00 5.01
R 0.83 0.83
E% 4.11 2.41 Fig. 1. Representative distribution of mean volume diameters and micrograph of
CEC 2.30 2.30 SMNZ.
a
Amorphous matter also includes partially or totally disordered phases.
b
Calculated by difference. R = Si/(Si + Al); E%: according to Passaglia and Sheppard
[28]; CEC (Cation Exchange Capacity) expressed in meq/g. Errors in parentheses.
locates into the SMNZ bulk, while the anionic FITC prevalently
interacts with the surfactant double layer, thus locating on SMNZ
fact, the theoretical CEC calculated using the zeolite chemical for- surface. This suggests that surface modication with the surfactant
mula was 2.30 meq/g, while the theoretical CEC based on the zeolite does not alter the CLI CA capacity for cation exchange, conrming
quantity in the sample came to 1.84 meq/g, thereby indicating the previous conjecture [29]. With regards to the hydrophilicity of
cation release from other raw sample constituent. The ECEC eval- the loaded molecules, the images displayed in Fig. 3 showed that
uated by CP-Cl exchange was 0.49 meq/g, thus conrming that the non-polar NR has a strong tendency to self-aggregate outside the
ECEC would appear not to be inuenced by surfactant type, since SMNZ particles. This observation strongly suggests that the loading
this value is comparable with those obtained for the same zeolite of a non-polar molecule into the SMNZ is very much discouraged,
sample using a different surfactant molecule. Moreover, the AEC of at least with the loading procedure used here. By contrast, CLSM
the produced SMNZ was equal to about 0.18 meq/g, approximately observations have shown that polar molecules localize preferen-
corresponding to the half the ECEC. Finally, the surface area val- tially into SMNZ bulk (Fig. 3). In actual fact, the uorescent signal
ues (BET) was found to be equal to 17.67 m2 /g, which is certainly in this latter case was quite weak, suggesting a limited interaction
consistent with literature data [13]. of uncharged molecules with the SMNZ surface and/or bulk.

3.2. Size and size distribution

The mean volume diameters of NZ, SMNZ and DS-loaded SMNZ
are reported in Table 2. The addition of the surfactant and the drug
led to a slight increase in the mean size.
In Fig. 1 a representative SMNZ mean volume diameter distribu-
tion and a transmitted light micrograph of SMNZ are shown. Clearly,
the SMNZ has a very irregular shape and a wide size distribution.
3.3. Confocal laser scanning microscopy (CLSM)
CLSM observations have enabled the discovery of the localiza-
tion of different molecules in/on SMNZ according to their chemical
nature. In particular, cationic/anionic and polar/non-polar uores-
cent probes were chosen to simulate different active molecules
to be loaded into SMNZ. These probes were loaded into the SMNZ
using the same procedure as applied to load DS. Figs. 2 and 3 provide
representative images of SMNZ loaded with four uorescent dyes,
namely TB (cationic), FITC (anionic), Rhod (polar) and NR (non-
polar). Images clearly show that when a charged molecule is loaded
into the SMNZ (Fig. 2), electrostatic interactions control the local-
ization of the molecules. In particular, cationic TB preferentially
Table 2
Mean volume diameters of NZ, SMNZ and DS loaded SMNZ. Standard deviations
were calculated on three independent experiments.
Sample Mean diameter (m) DSa
NZ 5.63 1.24
SMNZ 6.37 1.50
DS-loaded SMNZ 6.97 0.59
a
DS were calculated on at least three independent experiments.
3.4. Powder owability
With a view to using SMNZ as an excipient for the oral delivery
of DS and envisaging a possible specic dosage form to be for-
mulated, powder owability must be preliminarily assessed. The
parameter that indirectly correlates to this property is the Carr
index, CI(%), which is the quotient between the apparent (A ) and
the tapped (T ) densities. It is known that, in a freely owing pow-
der, the bulk and tapped density are close in value and, therefore,
the CI(%) is small. By contrast, in a poorly owing powder there
are stronger inter-particle interactions and, therefore, the CI(%) is
higher. It is generally recognized that good powder owability cor-
responds to CI(%) values of less than 15, while above 25 means
that the powder is considered to be cohesive (as reported in Italian
Pharmacopoeia, XII ed.). The results of this analysis have shown
that CI(%) values for the NZ, SMNZ and DS loaded SMNZ were,
respectively, 42.2, 33.1 and 31.9. These ndings have shown that
the addition of the surfactant and the drug slightly enhances the
ow properties of the powders, probably by discouraging van der
Waals inter-particle interactions. However, the reduction in CI(%)
value with surfactant/drug addition does not sufce and we there-
fore attempted to improve the SMNZ powder owability by adding
selected lubricants such as spray dried lactose, starch, CMC (micro-
crystalline cellulose) and colloidal silica in different w/w ratios
with respect to SMNZ. Unfortunately, lubricant addition did not
lead to any signicant reduction in CI(%) and this can probably
be ascribed to the wide distribution of mean volume diameters of
the SMNZ particles, along with their irregular shape. These results
suggest that SMNZ powders need further processing, such as a
different grinding technique or granulation to improve their owa-
bility.
 B. de Gennaro et al. / Colloids and Surfaces B: Biointerfaces 130 (2015) 101109 105

Fig. 2. Representative CLSM images of SMNZ loaded with cationic (TB, blue)/anionic (FITC, orange) dyes. The gure reports the merge of transmission, uorescence images
and the Z stack. The bar is 20 m. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of the article.)

3.5. Surfactant modied natural zeolite (SMNZ) preparation
Fig. 4(left) reports the equilibrium isotherm of the sample
treated with the selected surfactant. It should be remarked that
the CP-Cl only leads to the formation of a partial or patchy bilayer,
as the highest surfactant concentration reaches only 150% of the
ECEC. These results are consistent with data reported by Li and
Bowman [30], which, although referring to another surfactant,
underlined the inuence of the surfactant leading anion in the
bilayer formation, and how the Cl did not enable it to form a com-
pact micelle such as Br does. Actually, this aspect cannot lead us to
suppose that the Br-SMNZ is better than the Cl-SMNZ one; Bow-
man himself noted that SMZ formed with HDTMA-Cl sometimes
use a higher portion of their anionic exchange capacity towards
polluting anions such as chromates and arseniates [30]. According
to the surfactant sorption results, the surfactant/zeolite composites
tested in order to investigate the DS sorption capacity were pre-
pared from a CP-Cl solution having a concentration equal to 150%
of the ECEC.
not only exchanged by anionic sites of the SMNZ, but also enters
the patchy bilayer, as previously suggested by Krajisnik et al. [17].
The adsorption rate of DS onto SMNZ can be described by dif-
ferent kinetic models. In particular, to clarify the mechanisms of
adsorption of DS on SMNZ, DS adsorption data were analyzed using
a pseudo-rst order, pseudo-second order, and intra-particle diffu-
sion models, as described below.
The pseudo-rst order Lagergen model is described by the fol-
lowing equation [31]:
k1
ln (qe qt ) = ln qe t (2)
2.303
where qt and qe (mg/g) are the amounts of drug adsorbed at time t
(min) and at equilibrium, respectively, and k1 (min1 ) is the kinetic
constant of the pseudo-rst order adsorption. These latter amounts
are calculated from the plot of ln(qe qt ) against t. In actual fact,
as shown in Table 3, the calculated and experimental values of qe
are very different, thereby indicating that the adsorption process
is not adequately described by a pseudo-rst order kinetic model.

Table 3
3.6. In vitro drug loading and release tests Pseudo-rst order, pseudo-second order kinetic parameters of adsorption process
and ion exchange parameters.

The adsorption kinetics have shown a rapid loading of the DS on DWPB DWNB
the SMNZ. Indeed, within 20 min, the phenomenon was at equilib-
Pseudo-rst order qe (mg/g) 40.3 65.0
rium, and the calculated DS loading, expressed as DS mass per unit k1 (1/min) 0.0787 0.134
mass of SMNZ, was 40.3 and 65.0 mg/g when drug loading was per- qe,calc (mg/g) 20.7 19.3
formed at pH 7.4 in DWPB or DWNB, respectively. This increase can r2 0.973 0.819
be explained if we consider that in DWPB there is greater compe- Pseudo-second order k2 (g/mg min) 0.00397 0.0134
qe,calc (mg/g) 41.8 62.9
tition between the drug and the phosphate buffer, while in DWNB, h = 1/k2 q2e (min g/mg) 0.144 0.0188
the lower ion concentration in the medium favors DS adsorption. r2 0.995 0.994
However, if we compare the results of the maximum DS loaded on Kinetic parameters kads (g/mg min) 0.225 1.962
the SMNZ surface with that of the AEC experimentally evaluated, it kdes 107 (1/min) 4.17 9.90
qe,calc (mg/g) 38.2 63.6
is clear that in the case of DWNB, the value of the maximum loading
r2 0.916 0.940
capacity (0.23 meq/g) is greater than the AEC, indicating that DS is
106 B. de Gennaro et al. / Colloids and Surfaces B: Biointerfaces 130 (2015) 101109

Fig. 3. Representative CLSM images of SMNZ loaded with polar (Rhod, red)/non-polar (NR, purple) dyes. The gure reports the merge of transmission, uorescence images
and the Z stack (Rhod). The bar is 20 m. (For interpretation of the references to color in this gure legend, the reader is referred to the web version of the article.)
 B. de Gennaro et al. / Colloids and Surfaces B: Biointerfaces 130 (2015) 101109
Fig. 4. Surfactant sorption isotherm between CLI CA and CP-Cl in concentration ranging from 25% to 400% of the ECEC (left). Adsorption rates of DS onto SMNZ from DWPB
and DWNB (right), with intra-particle diffusion model for DS loading in DWPB and DWNB at room temperature and pH 7.4 (inset). Solid lines represent numerical simulations.
The pseudo-second order model considers that adsorption kinetics
correlates in a linear fashion to the square of the number of avail-
able binding sites [32]. The linear form of the pseudo-second order
equation is the following [33]:
t 1 t in the SMNZ surface and in the loading medium is achieved. DS
= + , (3)
qt k2 q2e qe
where k2 (g/(mg min)) is the rate constant of pseudo-second order
adsorption, obtained from the plot of qt /t as a function of time. In
this case, the experimental and calculated values of qe are simi-
lar. Moreover, the comparison between the correlation coefcients
(r2 ) of these two models indicate that the pseudo-second order
model could more accurately predict adsorption behavior than the
pseudo-rst order model results, thereby supporting the conclu-
sion that adsorption kinetics is a pseudo-second order reaction.
In order to determine the rate-limiting step of the adsorption
process, the intra-particle diffusion model was adopted. In the case
of zeolites, pore diffusion may play a role in drug adsorption onto
SMNZ and, for this reason, adsorption proles were analyzed as
follows [34]:
0.5
qt = kp t +I
where kp (mg/(g min0.5 ))
is the rate constant for the intra-particle
diffusion and I is the intercept of qt against the square root of
time. Hence, if I = 0, the intra-particle diffusion is the only rate-
limiting stage while, for increasing values of I, the contribution
of the adsorption surface is increasingly important. As shown in
Fig. 4 (right-inset), the plots of qt vs t0.5 in DWPB and DWNB are
not linear and r2 values are very low, thus indicating that the intra-
particle diffusion is not a rate-controlling step of the adsorption
process. In particular, I value is higher in the case of DWNB than in
DWPB (I = 54.3 and 27.0 mg/g, respectively) and this indicates that
the boundary layer has the greatest inuence when DWNB is used.
Table 3 also reports the values of the kinetic parameters of
drug adsorption/desorption process. To quantify these parameters,
loading data were t by modeling DS adsorption by an ion exchange
equation. The latter was written considering a single average SMNZ
particle, having the mean volume diameter. Drug diffusion coef-
cient in the loading medium has been assumed to be independent
of drug concentration and, taking into account the results of Eq. (4)
tting, ion exchange has been considered as prevalently occurring
at SMNZ surface [35]. Since drug loading was carried out under
vigorous stirring, diffusion resistance in the loading medium can
be neglected and drug loading has been assumed to be driven by
adsorption/desorption kinetics. Consequently, drug loading can be
schematized as follows: (i) initially, each binding site is occupied
107
by anions present in the loading medium; (ii) once at SMNZ surface,
DS either occupies a free binding site or displaces counterions;
(iii) the loaded drug slowly diffuses through surfactant layer. Ion
exchange proceeds until equilibrium between drug concentration
loading can therefore be modeled based on a pseudo-chemical
equilibrium relation, and described by a stoichiometric exchange
between the drug and the counterions at SMNZ surface [36,37]:
kads
DS + Cat + DS Cat + (5)
kdes
Here Cat+ is the binding site and DS Cat+ the complex within the
surfactant layer, kads and kdes the kinetic constants for drug adsorp-
tion and desorption, respectively. The following kinetic equation
at the SMNZ frontier can be written on the drug participating to
the loading:
dqDS Cat + free free
= kads qDS qCat + kdes qDS Cat + (6)
dt
free
where qDS Cat + and qDS are, respectively, the time-dependent
concentrations of the drug bound to the surfactant and in the
(4)
free
loading medium, while qCat + is the concentration of the positive
charges in the surfactant, unoccupied by DS anions. At t = 0,
qDS Cat + = 0. That is:
free
dqDS mSMNZ dqDS Cat +
= , (7)
dt mL dt
where mSMNZ is the mass of SMNZ during the loading process, and
free
mL is the liquid mass of the loading medium. At t = 0, qDS is the
initial concentration of DS in the loading solution. The model is
free
completed considering that qCat + = qe qDS Cat + . All the concen-
trations have been expressed as mg/g. Fitting results (Fig. 4(right))
are satisfactory and show that desorption process is clearly
negligible and that adsorption kinetics are strongly encouraged
when a DWNB is employed for drug adsorption. In particular, the
adsorption kinetic constant is roughly 9-fold higher when DWNB
is employed as a loading medium, therefore strongly pointing out
to the need for ion-free liquid medium during adsorption process.
The release prole of DS from SMNZ is shown in Fig. 5. In actual
fact, DS loading was found to be reversible with the drug being
gradually released in the presence of a ionic medium, such as SIF.
In particular, SIF (pH 6.8) was chosen as the release medium
because of the insolubility of DS in a strongly acidic environment
such as gastric uids (pH 1.2) [38]. In actual fact, DS is a salt of a
weak acid (2-[(2,6-dichlorophenyl)amino]benzeneacetic acid) and
108 B. de Gennaro et al. / Colloids and Surfaces B: Biointerfaces 130 (2015) 101109
Fig. 5. In vitro release prole, in SIF at 37 C, of DS from SMNZ. Inset: DS released as
a function of the square root of time. Standard deviations were calculated on three
independent experiments.
its solubility therefore greatly depends on the ionization constant,
Ka , and on the pH of the dissolution medium. More specically,
when pH values are more than 1 unit below pKa of DS (4), the
active molecule is presented mostly in its free acid form, which
is even less soluble than salt [39]. As the pH value increases, the
solubility of the active ingredient increases due to the contribution
made by the ionized form, and the highest solubility of the ionized
form is in PBS at pH 8.0.
Results highlighted the ability of SMNZ to sustain DS release for
about 5 h with fairly reproducible kinetics. In detail, about 40% of
the drug was released in the rst hour, followed by a slower release
phase for up to 5 h and a nal rapid release phase. The release of DS
from SMNZ is predominantly triggered by ion exchange since the
drug for preference distributes on the particle surface, as suggested
by CLSM analysis. More specically, the ion exchange process can
be described as a series of two steps including the stoichiometric
exchange of ions with solute and mass transfer. Since ion exchange
is very fast, drug diffusion is the step controlling drug release kinet-
ics [35]. Indeed, the plot of the percentage of released DS as a
function of the square root of time is basically linear for up to 5 h
(inset of Fig. 5). This shows that, in this time frame, drug diffusion
does control the release process [40]. Afterwards, a nal phase of
rapid release occurs, suggesting that, at this stage of the process,
a different release mechanism prevails other than ion exchange,
presumably related to the partial absorption of molecules of DS
inside the SMNZ that can be released only when the supercially
exposed DS has been exchanged. This further conrms our previ-
ous observation that the higher SMNZ loading capacity compared
to AEC indicates that some drug molecules may cross the patchy
bilayer and be internalized within the SMNZ [17].
4. Conclusions
In this work we investigated the possibility of using a
clinoptilolite-rich rock from California (CLI CA) as a carrier for
sustained drug delivery. For this purpose, surface modied natu-
ral zeolite (SMNZ) was produced according to the external cation
exchange capacity of the zeolite-rich rock. The cationic surfac-
tant cetylpyridinium chloride was employed to this aim, while
diclofenac sodium (DS) was used as model drug.
CLI CA is a Na- and K-rich phase, while the zeolite content and
the external cationic exchange capacity were independent of the
nature of the cationic surfactant used. Powder owability, esti-
mated by means of Carrs index % (CI(%)), revealed that the powder
was cohesive with scarce ow properties. The addition of lubricants
did not lead to a signicant reduction in CI(%) thus suggesting that
a granulation process is most likely required to produce a dosage
form for oral administration. This can probably be ascribed to the
extensive distribution of volume diameters of the SMNZ particles
along with their irregular shape. Moreover, SMNZ was found to be
a very versatile drug carrier since molecules with different chemi-
cal properties can be potentially loaded in/on SMNZ. In particular,
cationic and polar probes prevalently localize in SMNZ bulk, anionic
probes preferentially arrange themselves on SMNZ surface, while
the loading of a nonpolar molecule in/on SMNZ is discouraged.
Numerical simulations have shown that DS adsorption is a pseudo-
second order reaction and that diffusion through the boundary
layer is the rate controlling step of the process. DS release in an
ionic medium, such as SIF, can be sustained for about 5 h through a
mechanism prevalently governed by anionic exchange with a rapid
nal phase probably ascribed to the release of DS fraction within
the patchy bilayer of SMNZ.
In conclusion, as pointed out previously [41], the surface modi-
cation of zeolitized cationic exchangers enables us to broaden the
range of ways in which these materials can be applied. The results
reported here pave the way for further investigation focused on the
pharmaceutical application of SMNZ, since our preliminary data
evidenced the versatility of their use.
Acknowledgements
The authors wish to thank INNOVA S.c.a.r.l. for the XRPD and ICP-
OES analysis, and CISAG (Centro Interdipartimentale di Servizi per
Analisi Geomineralogiche) for the SEM-EDS analysis. The authors
also wish to thank Dr. Daniel T Eyde - CTO and President of St.
Cloud Mining (Tucson, Arizona - USA) for providing us with the
clinoptilolite-rich tuff used in this research. This work was carried
out with the nancial support of MIUR (Ministero dellIstruzione,
dellUniversit e della Ricerca) Progetti di Ricerca di Interesse
Nazionale (PRIN 2010).
References
[1] M.I. Carretero, M. Pozo, Appl. Clay Sci. 47 (2010) 171.
[2] G. Cerri, Appl. Clay Sci. 27 (2004) 141.
[3] M.C. Bonferoni, G. Cerri, Appl. Clay Sci. 36 (2007) 95.
[4] M.I. Carretero, M. Pozo, Appl. Clay Sci. 46 (2009) 73.
[5] F.A. Mumpton, Proc. Natl. Acad. Sci. U. S. A. 96 (1999) 3463.
[6] G. Rodrguez-Fuentes, M.A. Barrios, A. Iraizoz, I. Perdomo, B. Cedr, Zeolites 19
(1997) 441.
[7] G. Rodrguez-Fuentes, A.R. Denis, M.A. Barrios-lvarez, A.C. Iraizoiz, Micro-
porous Mesoporous Mater. 94 (2006) 200.
[8] K. Pavelic, M. Hadzija, L. Bedrica, J. Pavelic, I. Dikic, M. Katic, M. Kralj, M.H.
Bosnar, S. Kapitanovic, M. Poljak-Blazi, S. Krizanac, R. Stojkovic, M. Jurin, B.
Subotic, M. Colic, J. Mol. Med. 78 (2001) 708.
[9] K. Pavelic, M. Hadzija, in: S.M. Auerbach, K.A. Carrado, P.K. Dutta (Eds.), Hand-
book of Zeolite Science and Technology, Marcel Dekker, New York, 2003
(Chapter 24).
[10] M. Grce, K. Pavelic, Microporous Mesoporous Mater. 79 (2005) 165.
[11] M. Katic, B. Bosnjak, K. Gall-Troselj, I. Dikic, K. Pavelic, Front. Biosci. 11 (2006)
1722.
[12] A. Rivera, T. Faras, Microporous Mesoporous Mater. 80 (2005) 337.
[13] Z. Li, R.S. Bowman, Environ. Sci. Technol. 32 (1998) 2278.
[14] R.S. Bowman, G.M. Haggerty, R.G. Huddleston, D. Neel, M.M. Flynn, Am. Chem.
Soc. (1995) 54.
[15] D. Krajisnik, A. Dakovic, M. Milojevic, A. Malenovic, M. Kragovic, D.B. Bog-
danovic, V. Dondur, J. Milic, Colloids Surf. B: Biointerfaces 83 (2011) 165172.
[16] D. Krajisnik, A. Dakovic, A. Melanovic, M. Milojevic-Rkic, V. Dondur, Z.
Radulovic, J. Milic, Appl. Clay Sci. 8384 (2013) 322326.
[17] D. Krajisnik, A. Dakovic, A. Malenovic, L. Djekic, M. Kragovic, V. Dobricic, J. Milic,
Microporous Mesoporous Mater. 167 (2013) 94.
[18] http://www.novartis.com/about-novartis/company-history/index.shtml
[19] A.R. Salmann, Am. J. Med. 80 (4B) (1986) 29, http://dx.doi.org/10.1016/0002-
9343(86)90076-8
[20] Diclofenac Epolamine The American Society of Health-System Pharmacists.
Retrieved 3 April 2011.
[21] N.K. Dutta, S. Annadurai, K. Mazumdar, S.G. Dastidar, J.E. Kristiansen, J. Molnar,
M. Martins, L. Amaral, Int. J. Antimicrob. Agents 30 (3) (2007) 242, http://dx.
doi.org/10.1016/j.ijantimicag.2007.04.018, PMID 17644318.
 B. de Gennaro et al. / Colloids and Surfaces B: Biointerfaces 130 (2015) 101109
[22] K. Mazumdar, N.K. Dutta, S.G. Dastidar, N. Motohashi, Y. Shirataki, In Vivo 20
(5) (2006) 613, PMID 17091768.
[23] R.A. Sheppard, Clays and Zeolites, Los Angeles, California to Las Vegas, Int lst
Clay Conference Field Trip Guidebook, Nevada, 1985, pp. 51.
[24] G. Cerri, A. Langella, M. Pansini, P. Cappelletti, Clays Clay Miner. 50 (1) (2002)
127.
[25] J. Staniforth, in: M. Aulton (Ed.), Aultons Pharmaceutics: The Design and Man-
ufacture of Medicines, third ed., Elsevier, Canada, 2007, pp. 152, 197.
[26] P. Cappelletti, L. Catalanotti, A. Langella, M. Mercurio, B. de Gennaro, M. Rosella,
A. Dakovic, in: A. Dakovic, M. Trgo, A. Langella (Eds.), Book of Abstracts Zeo-
lite 2014, 9th International Conference on the Occurrence, Properties, and
Utilization of Natural Zeolites, Belgrade, Serbia, 813 June 2014, p. 37, ISBN:
978-86-82867-26-5.
[27] United States Pharmacopeia and National Formulary, United States Pharma-
copeial Convention Inc., Rockville, MD, USA, 25th Edition (2002) and 26th
Edition (2003).
[28] E. Passaglia, R.A. Sheppard, in: D.L. Bish, D.W. Ming (Eds.), Natural Zeolites:
Occurrence, Properties, Applications. Reviews in Mineralogy and Geochem-
istry, 45, Mineralogical Society of America, Washington, 2001, p. 69.
109
[29] R.S. Bowman, E.J. Sullivan, Z. Li, in: C. Colella, F.A. Mumpton (Eds.), Natural
Zeolites for the Third Millennium, De Frede Editore, Naples, Italy, 2000, p.
287.
[30] Z. Li, R.S. Bowman, Environ. Sci. Technol. 31 (1997) 2407.
[31] S. Lagergren, Kungliga Svenska Vetenskapsak ademien Handlingar 24 (1898) 1.
[32] Y. Gao, Y. Li, L. Zhang, H. Huang, J.J. Hu, J. Colloid Interface Sci. 368 (2012)
540546.
[33] Y.S. Ho, G. McKay, Process. Biochem. 34 (1999) 451.
[34] W.J. Weber, J.C. Morris, J. Sanit. Eng. Div. Am. Soc. Civ. Eng. 89 (1963) 31.
[35] M.J. Abdekhodaie, X.Y. Wu, Biomaterials 27 (2006) 3652.
[36] V. Boudy, N. Voute, D. Pradeau, J.C. Chaumeil, Int. J. Pharm. 239 (2002) 13.
[37] M. Biondi, S. Fusco, A.L. Lewis, P.A. Netti, J. Mater. Sci. Mater. Med. 24 (10) (2013)
2359.
[38] M. Kincl, F. Vrecer, M. Veber, Anal. Chim. Acta 502 (2004) 107.
[39] D. Hrter, J.B. Dressman, Adv. Drug Deliv. Rev. 46 (2001) 75.
[40] L. Mayol, M. Biondi, L. Russo, B.M. Malle, K. Schwach-Abdellaoui, A. Borzac-
chiello, Carbohydr. Polym. 102 (2014) 110.
[41] B. de Gennaro, L. Catalanotti, R.S. Bowman, M. Mercurio, J. Colloid Interface Sci.
430 (2014) 178.