CRITICAL JOURNAL

REVIEW
QUALITATIVE AND
QUANTITATIVE
ANALYTICAL CHEMISTRY
STUDY PROGRAM S1
BILINGUAL CHEMISTRY
EDUCATION

SCORE:

CRITICAL JOURNAL REVIEW

GROUP III

NAME : 1. AHMAD RAMADANA (4182131001)

2. NAJLA RIZANI FAHLEVI (4181131012)

3. NATALIN PERTIWI SIAHAAN (4183332001)

PROGRAM : S1- BILINGUAL CHEMISTRY EDUCATION

LECTURER : Dr. Techn MARINI DAMANIK, S. Si., M. Si.

COURSES : QUALITATIVE AND QUANTITATIVE ANALYTICAL

CHEMISTRY

STUDY PROGRAM S1 BILINGUAL CHEMISTRY EDUCATION

FACULTY OF MATHEMATICS AND SCIENCE
STATE UNIVERSITY OF MEDAN
APRIL 2020
 PREFACE

Praise the presence of God Almighty for all His mercy so that this paper can be
arranged to completion. Not to forget we also express my gratitude for the assistance from
those who have contributed by contributing both their material and thoughts in this Critical
Journal Review (CJR).
The assignment was made to fulfill the Critical Journal Review (CJR) task in the
Qualitative and Quantitative Chemistry. We hope this paper is one of the references for
readers if they want to know the contents of this Journal. Thus is the introductory words
that we said, more and my lack of apology because someone did not escape mistakes. In
the end let me say thank you.

Medan, May 2020

Group III

i
 LIST OF CONTENTS

PREFACE ...............................................................................................................................I

LIST OF CONTENTS .......................................................................................................... II

CHAPTER I INTRODUCTION ........................................................................................... 1

1.1 Identities of Journal 1.................................................................................................. 1

1.2 Identities of Journal2................................................................................................... 1

CHAPTER II DESCRIPTION CONTENTS OF JOURNAL .............................................. 2

2.1 Description Contents of Journal 1............................................................................... 2

2.2 Description Contents of Journal 2............................................................................... 6

CHAPTER III ADVANTAGES ......................................................................................... 13

3.1. The Advantages of Journal I ..................................................................................... 13

3.2. The Advantages of Journal II .................................................................................... 13

CHAPTER IV DISADVANTAGES ................................................................................... 14

4.1. The Disadvantages of Journal I................................................................................. 14

4.2. The Disadvantages of Journal II ............................................................................... 14

CHAPTER V CLOSING ..................................................................................................... 15

5.1. Conclusion ................................................................................................................ 15

5.2. Suggestion ................................................................................................................. 15

REFERENCES .................................................................................................................... 16

ii
 CHAPTER I
INTRODUCTION

1.1 Identities of Journal 1

Journal Title Journal of Pharmaceutical Analysis
Article Title An analytical method for Fe(II) and Fe(III) determination
in pharmaceutical grade iron sucrose complex and sodium
ferric gluconate complex
Publication Year 2012
Author of The Daniele Merlia, Antonella Profumoa,n, Carlo Dossi

Journal
Publisher Production and hosting by Elsevier B.V.
The City Rises Xi’an Jiaotong University.
Doi 10.1016
1.2 Identities of Journal2
Journal Name Journal of Microporous and Mesoporous Materials
Article Title Iron within the erionite cavity and its potential role in
inducing itstoxicity: Evidence of Fe (III) segregation as
extra-framework cation
Publication Year 2017
Author of The Alessandro Pacella ,MarziaFantauzzi , Davide Atzei b,
Carlo Cremisini , Elisa Nardi ,Maria Rita Montereali ,
Journal
Antonella Rossi , Paolo Ballirano
Publisher www.elsevier.com/locate/micromeso
The City Rises Italy
Doi 10.1016

1
 CHAPTER II
DESCRIPTION CONTENTS OF JOURNAL

2.1 Description Contents of Journal 1

2.1.1. Introduction
Most iron deficiency anemic as respond well to treatment with oralor parenteral iron;
in the latter case, polymeric complexes of Fe(III) with sugars, such as iron sucrose
complex (ISC) andsodium ferric gluconate complex (SFGC) , are frequentlyused to
stabilize iron hydroxide nanoparticles in the colloidalsuspension .According to USP
(United State Pharmacopeia), they must contain less than 20% Fe(II) with respect to the
totaliron (5.3%–6.4% and 4.5%–7.0%, for ISC and SFGC respectively), and low molecular
weight (MW) complexes must beundetectable by gel permeation chromatography and
normalpulse polarography (NPP) . The standard method reported on Pharmacopoeia, and
still used, is, in fact, NPP in acetate ionic strength buffer.

2.1.2. Research Purpose

In this paper, a robust voltammetric method at hanging dropping mercury electrode
(HDME) for the determination of Fe(II) and Fe(III) in SFGC and ISC formulations at
about 1% (w/v) concentration, is described and validated against standard USP method.
This methodology combines greater sensitivity and lower limit of detection (LOD) with
reduced mercury consumptions. A redox titration method for the selective determination of
Fe(II) is also reported.

2.1.3. Experimental Method

2.1.3.1. Reagents and chemicals
Chemicals were obtained from Aldrich and used as received.Solutions were
prepared with Milli-Q grade water. Cericammonium sulfate was standardized against
As2O3 bypotentiometric titration. ISC and SFGC were prepared and
purifiedaccording to known procedures

2.1.3.2. Instrumentation
Electrochemical measurements were performed in a conventional three-
electrode cell (volume 10 mL) with HDME as the working electrode (BASi, PWR-3
interfaced to EF-1400controlled growth Hg electrode), Pt wire as the
auxiliaryelectrode, and Ag/AgCl (3 M NaCl) as the reference electrode.Potentiometric

2
 titrations were performed using a Pt working electrode with an Ag/AgCl (3 M KCl)
reference electrode on an Orion 520 potentiometer

2.1.3.3. Quantitative calculations

Fe(II) in the preparations is calculated by the Eq. (1) from the NPP or
differential pulse voltammetry (DPV) response ratios
2
[1 − ( )] × [𝐹𝑒] = %𝐹𝑒(𝐼𝐼)
𝑅
Where R is Fe(II)/Fe(0) (henceforth denoted R1) to Fe(III)/ Fe(II) (denoted R2) wave
response ratio; [Fe] is the total iron concentration in the vial, typically 20 g/L.
According to Eq. (1), in the absence of Fe (II) in the original solution, the theoretical R
value should approximate two, since Fe(III)/Fe(II) is one-electron reduction process
and Fe(II)/Fe(0) is bioelectronics, and all the Fe(II) reduced to Fe(0) derives from the
reduction of the Fe(III) complex in the sample.

2.1.4. Research Result and Discussion

2.1.4.1. Determination of Fe(II) by NPP
In the USP polar graphic method for the determination of Fe(II) in ISC
aqueous solutions, the half-wave potential, E1/2, isrelated to the MW of the complex
[8]. The two waves at750mV and 1400 mV correspond to Fe(III)/Fe(II) and
Fe(II)/Fe(0) reductions, respectively. The absence of additional peaksmay, therefore,
excludes the presence in the sample of low MW complexes at concentration higher
than 5% of the total ironcomplex. According to Eq. (1), R values greater than two (i.e.
thesecond wave height greater than twice the first), must beobserved whenever Fe(II)
is present in the initial solution.However, the waves often show poor reproducibility,
in termsof current intensity and E1/2 values. As a consequence, thecalculated
concentration of Fe(II) tends to be extremely high oreven negative, with no analytical
significance. Such large Rvalues are likely to be related to systematic overestimation
of Fe(II), which is, instead, unlikely to occur during the preparationof ISC and SFGC,
where the highly alkaline melt comprisingmixture of ferric oxide and sucrose or
glucose would result in afast oxidation of Fe(II) by air. This prompted us to undertake
aninvestigation of the electrochemical behavior of the two Fecomplexes with the final
aim to devise an alternative analyticalmethod that could ensure adequate
reproducibility and precision, as required in Quality Control and Quality Assurance
protocols, reducing, at the same time, Hg consumption during the analysis

3
2.1.4.2. Determination of Fe(II) by DPV
DPV was initially carried out using the instrumental parameters reported in the
USP methodology [8], and successivelyoptimized. The final working conditions were:
LiClO4 50 g/L,SFGC or ISC 330 mg/L, step 30 mV; pulse width 100 ms; pulse period
325 ms, pulse amplitude 100 mV; scan rate92 mV/s; Ei¼0 V; Ef¼1700 mV. A two-
peak voltammetric profile (Fig. 1) is observed for ISC, and peak height is shownto be
dependent on ISCconcentration. The peak around700 mV is due to [Fe(III)]/[Fe(II)]
reduction, whereas thesecond peak, around 1400 mV, is ascribed to [Fe(II)] reduction
to zerovalent iron. Analogousresults were observed withthe SFGC.A good linearity
between current and concentration wasfound in the range 0.05–50 mg/L Fe for both
irons.

Formulations, using lithium perchlorate as a non-complexing supporting

electrolyte (Table 1). The determination of Fe(II) and Fe(III) complexes in real
samples can be obtained, without any pre-treatment or separation, by standard addition
method using ISC and SFGC standard solutions. Noteworthy, this method can allow
the analysis of samples even if the standard solutions are not available. In fact, the
R1/R2 ratio can be used to quantify Fe(II) selectively and, after determination of the
total iron content by alternative methods, typically by spectroscopic techniques (e.g.

4
ICP-OES, AAS), Fe(III) is calculated from the difference between these two values.
With this simple approach, Fe(II) concentrations as low as 1% of total iron can be
easily quantified.

2.1.4.3. Influence of free Fe(III) and Fe(II) and weaker iron Complexes
In order to evaluate the effect of free Fe ions on the composition of the
complexes, ferric and ferrous ammonium sulfate were added to the solution of each
iron formulation and voltammograms were recorded subsequently. It was verified that
the addition of free Fe(II) and Fe(III) shifted the equilibria of SFGC and ISC toward
new complexes with intermediate MW . Interestingly, an analogous behavior was also
observed with labile Fe(II) complexes, e.g. the gluconate complex

2.1.4.4. Validation of the proposed DPV method

The analytical method was validated according to International Conference on
Harmonization (ICH) guidelines. The linearity, assessed by linear regression
determinations,
Complex DPV NPP
R1/R2 RSD(%) R1/R2 RSD(%)
ISC 2.03 3.8 1.86 12.4
SFGC 2.09 3.6 1.77 11.5
Expected value 2.0 2.0

was calculated by the least-square regression method. Thecalibration graphs were

obtained with 10 standard solutions inthe concentration range 0.05–50 mg/L as Fe,
both on ISC andSFGC. The correlation coefficient (r) value was found to be0.998. A
set of 6 samples containing 20 mg/mL, as Fe, of each pharmaceutical formulation was
analyzed to assess repeatability and precision. The repeatability was evaluated
byassaying samples during the same day, whereas the intermediate precision was
investigated by comparing results on twodifferent days. LOD and LOQ values for the
complexesevaluated from the linear regression were 15 mg/L and 50 mg/L,for ISC and
SFGC (as Fe), respectively. Recovery was evaluatedby addition of known amounts of
standard solutions of eachdrug to the commercial formulations. The spiked solutions
werethen analyzed by the proposed DPV method and the results werein the range
94%–117%. Similarly, the intraday and inter day precisions showed good results, with
percentage errors ranging between 3.2% and 4.5%.

5
 2.1.4.5. Potentiometric determination of Fe(II) in ISC and SFGC
In order to assess the total Fe(II) concentration in the samples, a potentiometric
determination using Ce(IV) and ferroine as indicator, was under taken after the
complete chemical decomposition, evidenced by disappearance of color, of the
complexes (200 mg) with 2 mL of a 1:1 mixture of commercial96% H2SO4/85%
H3PO4. Phosphoric acid stabilizes Fe(III)ions and leads to a sharp titration end point.
Hence, Fe(II) concentration as low as 1% with respect to the total iron, could easily be
quantified. This potentiometric method was compared with the DPV Fe(II) analysis,
using standard ISC and SFGC samples, as well as a ISC preparation contaminated
with Fe(II). Fe(II) concentration was found to be below LOD in the standard samples,
and 0.061 mmol/g ISC (DPV) and 0.066 mmol/g ISC (potentiometric) in the
contaminated sample. The good agreement of the results obtained by the two methods
confirmed the reliability and the accuracy of the DPV proposed method for Fe(II) and
Fe(III) quantification in iron sucrose and sodium ferric gluconate complexes

2.1.5. Conclusion
The proposed method was then validated against the USP NPP method. The same
standard samples, with an expected R1/R2 ratio of 2.0, have been analyzed by both (NPP
and DPV) methods. The comparative results are shown in Table 2. The proposed DPV
method has been found statistically more precise (RSD¼3.6% against 11.5% for SFGC
and RDS¼3.8% against 12.4% for ISC) and accurate than NPP, and, therefore, it can be
directly applied to the analysis of the pharmaceutical preparations. This work was
financially supported by FAR, Fondo Ateneo per la Ricerca Universita´ di Pavia, Italy. We
thank Prof. Nick Serpone for useful discussions.

2.2 Description Contents of Journal 2

2.2.1. Introduction
Erionite is a zeolite with fibrous morphology that has been classified as a Group 1
Human-Carcinogen by the International Agency for Research on Cancer (IARC) . In fact,
its inhalation has been shown to cause malignant mesothelioma (MM), with tumorigenicity
500e800 times greater than that of chrysotile and crocidolite asbestos. Being prevalently
the product of diagenetic alteration by weathering processes of sediments in volcanic ash ,
it is relatively widespread worldwide. It commonly occurs associated with structurally
related offretite .From the structural point of view, erionite belongs to the socalled ABC-6

6
family . This family is built up from the stacking along the c-axis of layers of six-
membered rings of (Si,Al)O4 tetrahedra following an ABC scheme. Erionite is one of the
four known minerals characterized by a six-layers sequence: AABAAC, erionite ,
AABBCC, chabazite, ABBACC, bellbergite, and ABABAC, liottite. It is hexagonal, space
group P63/mmc, and has an average chemical formula K2(Na,Ca0.5)8[Al10Si26O72].30H2O.
Three different species, erionite-K, erionite-Ca, and erionite-Na are recognized on the basis
of the most abundant extra-framework (EF) cation . The framework of erionite is
composed of columns in which cancrinite (ε) cages and double-6 rings (D6R) regularly
alternate.
2.2.2. Experimental Method
2.2.2.1. Sample description
The erionite pristine sample is from Rome, Oregon, USA. The hand specimen
has relevant chemical variability resulting in extended patches of erionite-Na
alternating with prevailing erionite-K areas. Erionite-Na found in this specimen is Ca-
free and may be accordingly identified with ease by both SEM-EDX and, owing to
slightly different cell parameters, by XRPD preliminary analyses. Particular care was
paid to identify and separate large areas of the sample consisting exclusively of
erionite-K, which was used throughout the present investigation. Its crystal-chemical
formula, as obtained from SEM-EDX analysis, is (Na1.45K2.98-
Ca0.61Mg0.63)[Al7.42Si28.58O71.74]29.49H2O. The raw material was subsequentlenriched
following an especially tailored procedure.
2.2.2.2. Fe (III) binding to erionite
Erionite-K fibres (25 mg) were suspended in 25 mL of 25, 50, 100, 250 and
1000 mM FeCl3 solutions in a 50 mL Falcon™ polypropylene tube. Experiments were
performed in triplicate. In addition, a “blank” procedure was carried out suspending
erioniteK fibres (25 mg) in water solutions (without FeCl3) adjusted to the same pH of
the FeCl3 solutions with HCl (Ultrex II, ultrapure reagent, J.T. Baker) diluted in
ultrapure deionized water (Milli-Q - Millipore, Bedford, MA, USA). Experiments at
pH 2.7 and 3.0 were performed in triplicate, the others in duplicate. All the
polypropylene tubes, closed and sealed with Parafilm, were placed in an oscillating
thermostatic water bath at 37 C. After 1 h of treatment the supernatant solution was
sampled and filtered, using a 0.22 mm nitrocellulose membrane filter, for ICP-OES
analysis. The fibres were recovered from the tube on filter, rinsed with ultrapure
deionized water to remove any residues of iron not bound, and dried prior to the

7
experiments of “incubation” in ascorbic acid and SEM-EDX, XRPD, and XPS
measurements.
2.2.2.3. Incubation of the fibres in ascorbic acid (AA)
Upon completion of Fe (III) binding experiments the fibres were recovered
from the filters, dried and homogenized. The accurately weighted fibres (about 25 mg)
were suspended in 25 mL of a 6 mM AA solution in a 50 mL Falcon™ polypropylene
tube. All the tubes, sealed with Parafilm, were placed in an oscillating thermostatic
bath set at the temperature of 37 C. Fibres previously treated with 250 and 1000 mM
FeCl3 solutions were incubated for 1 h and 24 h, whereas those treated with 25, 50 and
100 mM FeCl3 solutions were incubated for 24 h only. A “blank” procedure was also
carried incubating for 1 h and for 24 h in the 6 mM AA solution the fibres previously
incubated for 1 h in water solutions (without FeCl3) adjusted to pH 3.0 and 2.7 (the
same pH of the 250 and 1000 mM FeCl3 solutions). All the tests were performed in
duplicate. After completion of the incubation in AA solution, the tubes were
transferred in a nitrogen-filled glove bag and the supernatant solution was sampled
and filtered, using a 0.22 mm nitrocellulose membrane filter, for ICP-OES analysis.
Finally, the fibres were recovered from the tube on filter, rinsed with ultrapure
deionizedwater to eliminate residues of the solution, dried and stored undernitrogen
prior to SEM-EDX, XRPD, and XPS measurements.
2.2.2.4. Samples characterization
2.2.2.4.1. Surface area
Surface area, BrunauereEmmetteTeller (BET) multipoint method and
textural analysis were obtained by N2 adsorption/ desorption measurements, at the
liquid nitrogen temperature (196 C), using a Micromeritics 3-Flex analyser. Owing
to the small amount of treated samples the analysis was limited to the pristine one.
Nevertheless, no significant variation of the measured parameters is expected upon
iron loading as a reduction of the surface area is attained exclusively by inserting
extra K cations that are segregated at the centre of the large openings of the erionite
cavity . The sample was pre-treated under vacuum at 65 C for 15 h. The pores
distribution was determined from the adsorption curve by the
BarreteJoynereHalenda (BJH) method and from the analysis of the micropore
isotherm by the t-test taking the curve of Harkins and Jura . The total pore volume
was determined by the rule of Gurvitsch .

8
2.2.2.4.2. Scanning electron microscopy e energy dispersive X-ray spectroscopy
The micro-chemical characterization of Fe (III) loaded and AA incubated
samples was performed using a Quanta 400 SEM equipped with an EDX Genesis
EDS system (FEI, Hillsboro, OR, USA). Operating conditions were: 15 kV
accelerating voltage, 11 mm working distance, 0 tilt angle. Chemical data were
collected at least at 12 analytical points. The final crystal chemical formulae were
calculated, after renormalization of the chemical analyses hypothesizing a water
content of 18.5 wt% (corresponding to ca. 30 atoms per formula unit, apfu), on the
basis of 36 (Si þ Al) apfu. Both the balance error formula E% and the K content test
were used for selecting the reliable analyses.

2.2.2.4.3. ICP-OES analysis

One mL of each filtered solution was diluted (1:10) with 1% nitric acid
solution and analysed by ICP-OES in order to measure the amount of Ca, Mg, Na,
K, Si and Al released by the erionite fibres and that of Fe persisting in solution after
their suspension. All measurements were performed using an Optima 2000 DV
ICP-OES spectrometer (Perkin-Elmer, Waltham, MA, USA) equipped with a cross
flow nebulizer placed inside a Scott spray chamber. ICP Aristar (BDH) standard
solutions in nitric acid for Ca, Mg, Na, K, Si, Al and Fe (1000 mg/L) were used to
prepare the calibrating solutions for ICPOES analyses. To ensure adequate quality
assurance, the analyses of the standard solutions were regularly repeated after the
measurement of a defined number of samples. ICP data of cation release and of Fe
bound to erionite fibres are listed in the present work under the form of nmol/mg of
erionite. The amount of Fe bound to erionite was calculated by difference between
the Fe concentration measured in the initial solution ([Fe]) and that persisting in
solution after 1 h of sample incubation ([Fe]1h). The net cation release was
calculated by difference between the values measured after suspension in the FeCl3
solution and those in the blank procedure (suspension of fibres in HCl solutions
with the same pH of the FeCl3 ones). In order to verify whether Fe (III) was
acquired by erionite through ion exchange or fixed at the surface, a charge balance
was calculated using the same approach detailed in ref. [42]. In particular, the
number of charges of each element was determined multiplying the nanomoles by
the valence of the corresponding element and finally, the net charges released into
solution were compared to the total charges acquired as Fe.

9
2.2.3. Results and Discussion
2.2.3.1. Solutions analysis
Erionite samples were suspended in 25 (pH ¼ 3.7), 50 (pH ¼ 3.5), 100 (pH ¼
3.2), 250 (pH ¼ 3.0) and 1000 mM (pH ¼ 2.7) FeCl3 solutions for 1 h and, for
comparison, in HCl solutions with the same pH of the ferric solutions. ICP-OES
analyses of the supernatants were performed to measure the amount of the released
cations from and the amount of the iron bound to erionite. The incubation of erionite
in FeCl3 solution resulted in a decrease of iron content in the FeCl3 solution. The
amount of iron bound by erionite was calculated starting from
[𝐹𝑒]𝑏 = [𝐹𝑒]0 − [𝐹𝑒]1ℎ
where [Fe]0 is the starting concentration of iron in the FeCl3 solutions and [Fe]1h is the
residual concentration after 1 h of incubation, determined by ICP analysis. The higher
was [Fe]0 the higher was iron bound by erionite [Fe]b (Fig. 1): the amount of Fe (III)
loaded by the fibres ranges from 23(1) to 376(40) nmol/mg after incubation in 25 and
1000 mM FeCl3 solutions, respectively.

2.2.3.2. Solutions analysis after incubation of Fe (III) loaded erionite in ascorbic acid
ICP data acquired on the FeCl3 treated sample after 24 h of incubation in AA
revealed the presence of the EF cations in the supernatant and 23nmol/mg for the
samples previously suspended in FeCl3 25- and 1000-mM solutions, respectively.
Together with the EF cations, small amounts of iron are released after 24 h of
incubation time.

2.2.3.3. Bulk analysis

N2 adsorption analysis of the pristine sample indicates a BET specific surface
area, measured in the conventional p/p0 0.05e0.3 range, of 252(5) m2g1 and an
external surface area of 10.1(5) m2g-1. The total volume of pores is of 0.151 cm3g1, of
which 0.130 cm3g-1 attributable to microspores. SEM-EDX chemical analysis revealed
a content of ca. 1.6(6) wt% of Fe2O3 loaded by the fibres incubated in a 250 mM FeCl3
solution (Table 2). Simultaneously, a marked reduction of the Na2O content (ca. 40%)
in the FeCl3 treated samples with respect to the pristine one is observed. The total site
scattering (s.s.) of EF cations suffers a moderate abatement as compared to that of the
pristine sample, passing from 92.3 to 81.1 e. Moreover, the analysis of the 250 m M
FeCl3 incubated fibres suspended in AA as well as those of the 1000 mM FeCl3 (not
shown) incubated fibres and those correspondingly suspended in AA revealed E%
10
 values markedly outside the admitted range of ±10% (up to 40%) so that it was
impossible to retrieve reliable crystal chemical formulae.

2.2.3.4. Surface analysis

The surface characterization of the erionite samples (pristine, treated with 250
and 1000 mM FeCl3 solutions and after incubation in AA for 1 and 24 h) was
performed by XPS. Together with the signals of the erionite (Si, O, Ca, Na, Al, K, Mg
and Fe) only small amounts of adventitious carbon were found in the survey spectra.
The absence of any chlorine signal in the Fe loaded samples ruled out the possibility
of the occurrence of residual FeCl3.
After 24 h of incubation in ascorbic acid the amount of released iron increased
with the concentration of the FeCl3 solutions. Regarding the effect of incubation time
it can be noted that a great part of iron is released during the first hour. The percentage
of iron bound to the fibres that was released into solution was around 6e7% rising to
ca. 10% in the 250 and 1000 mM FeCl3 treated samples. This result is consistent with
previous findings of Eborn and Aust showing that Fe located at the fibre surface is
difficult to be mobilized by ascorbate. The observed EF cation release in AA of the Fe
loaded samples is comparable or even lower than that of the pristine ones incubated in
AA. This result indicates that, after reduction of Fe (III) fixed at the fibre surface, the
possible incorporation of Fe (II) as EF cation through ion exchange can be considered
as non-existent or very limited. XPS analyses confirmed the release of iron by
detection of a lower total iron concentration on the surface. In addition it could be
shown that upon incubation in AA of the 250 mM FeCl3 suspended sample Fe (II)
oxide appears after 1 h of immersion, thus indicating the reduction of iron Fe (III) to
Fe (II).
2.2.4. Conclusion
In this work we investigated the chemical, structural and surface modifications of
fibrous erionite-K from Rome after incubation in FeCl3 solutions. ICP-OES results,
supported by XPS investigation, unambiguously show that Fe (III) is mainly fixed at the
fibre surface, in perfect agreement with previous results of Eborn and Aust . However, a
fraction of Fe (III), which is ca. 2/3 in the case of very diluted Fe (III) solutions and rapidly
decreasing as concentration rises, is segregated by an ion-exchange mechanism in the
erionite cavity at the Ca3 site, similarly to Fe (II) . The maximum amount of apfu Fe (III)
residing at Ca3 is of 0.17(7) and is observed in the case of the sample incubated in the 250

11
mM FeCl3 solution. This is a significantly smaller iron content as compared to that ion
exchanged under the form of Fe (II). Therefore, Fe (II) and Fe (III) are both fixed at the
same crystallographic site, albeit with a significantly different efficiency with respect to the
Fe available in solution. In addition, our results suggest that AA is able to reduce fairly
easily only Fe (III) residing at the fibres surface and characterized by low nuclearity,
whereas this reaction does not occur (or possibly occurs only very marginally) in the case
of the ion exchanged metal. Bronchoalveolar lavage of normal humans reported total iron
content as small as 0.21 mmol L1 (12(2) mg L1) strongly suggesting that under
physiological conditions iron tends to be prevalently ion exchanged. This is a very
important piece of information as it, unfortunately, represents the optimum condition for
iron to behave as a very active site in the generation of ROS. In fact, it has been shown that
in Fe loaded zeolites at rather low Fe content, the most active sites are those located in
well-defined crystallographic positions and possessing a very low nuclearity (single ions,
dimers or oligomers).

12
 CHAPTER III
ADVANTAGES

3.1. The Advantages of Journal I

When viewed from the year the journal was published, this journal was up to date
because the journal was published in 2012. This journal included graphs and tables that made
it easy for readers to understand the results of the research. The abstract presentation in this
journal is brief and clear. There is keyword section that makes it easy for readers to recognize
related topics in this journal. In the experiment section, this journal includes quantitative
calculations. This journal explains the contents systematically, densely and clearly. This
journal is good for adding insight from readers and can be used as a reference.

3.2. The Advantages of Journal II

When viewed from the year of publication of this journal, this journal is up to date
because it was published in 2016. This journal includes tables and graphs so as to make it
clear in describing the data. In terms of the structure of journal writing, there is a keyword
section added to make it easier for readers to recognize the topic being discussed, there is an
info section as clarifying journal identity and journal history. The abstract presentation in this
journal is solid and clear. In the experimental section, it is written systematically and
completely. Presentation of results section and discussion, explained in detail and clearly.
Formulas and calculations are well written. This journal has many reference sources, thus
increasing the completeness of the contents of the journal. This journal also includes
conclusions, acknowledgments, and appendix sections A. Supplementary data to complement
the contents of the journal. This journal is good to be used as reference material while adding
to the reader's insight

13
 CHAPTER IV
DISADVANTAGES

4.1. The Disadvantages of Journal I

This journal does not include a DOI and will be complete if it is included in the DOI. In
terms of structure, it will be better and more complete if the conclusions are included. This
journal cites a number of reference sources under 2000. When viewed from journal sources,
this makes the journal less sophisticated.

4.2. The Disadvantages of Journal II

In this journal, there are writing formulas that are too small so that readers find it difficult
to see. This journal does not include a DOI and will be complete if it is included in the DOI.
In terms of structure, it will be better and more complete if the conclusions are included. This
journal cites several reference sources under the year 2000. When viewed from journal
sources this makes the journal less sophisticated and in writing references it would be better if
it was written in alphabetical order so that it was easy to see the list of references.

14
 CHAPTER V
CLOSING

5.1. Conclusion
From the summary in each journal, we can see that the journal 2 have more explanation
than journal 2. Journal 2 have graph, and table than Journal 1. In the experimental section,
Journal 2 is written systematically and completely. Presentation of results section and
discussion, explained in detail and clearly. Formulas and calculations are well written.
To choose a better one, journal 2 is good in term of completeness of the material provide
and the journal 1 is good in term of explenation. But for the references, for the both of journal,
have several references sources under the year 2000.

5.2. Suggestion
Both of these journals have presented the material well, as well as a clear and good
writing structure. Each journal has advantages and disadvantages. In some flaws it will be
better upgraded and renewed. But, for choosing the best from the both journal, journal 2 have
more in term of completeness of the material. However, for choosing in the easy deliver one
from the both journsl, journal 1 is the best of it.

15
 REFERENCES

Merli, D., Profumo, A., & Dossi, C. (2012). An analytical method for Fe (II) and Fe (III)
determination in pharmaceutical grade iron sucrose complex and sodium ferric
gluconate complex. Journal of pharmaceutical analysis, 2(6), 450-453.

Pacella, A., Fantauzzi, M., Atzei, D., Cremisini, C., Nardi, E., Montereali, M. R., ... &
Ballirano, P. (2017). Iron within the erionite cavity and its potential role in inducing
its toxicity: Evidence of Fe (III) segregation as extra-framework cation. Microporous
and Mesoporous Materials, 237, 168-179.

16
Journal of Pharmaceutical Analysis 2012;2(6):450–453

Contents lists available at ScienceDirect

Journal of Pharmaceutical Analysis

www.elsevier.com/locate/jpa
www.sciencedirect.com

SHORT COMMUNICATION

An analytical method for Fe(II) and Fe(III) determination

in pharmaceutical grade iron sucrose complex and sodium
ferric gluconate complex
Daniele Merlia, Antonella Profumoa,n, Carlo Dossib

aDipartimento di Chimica, Universita` degli Studi di Pavia, Via Taramelli 12/14/16, 27100 Pavia Italy
bDipartimento di Scienze Teoriche ed Applicate (DSTA), Universita` dell’Insubria, Via Dunant, 7, 21100 Varese Italy

Received 31 January 2012; accepted 18 May 2012

Available online 26 May 2012

KEYWORDS Abstract A robust voltammetric method has been developed and validated for the determination
Sodium ferric gluconate
of Fe(II) and Fe(III) in pharmaceutical iron polysaccharidic complexes. Undesirable low molecular
complex; weight iron complexes, at concentration about 3% in the pharmaceutical formulation, can be easily
Iron sucrose complex; determined with good accuracy and precision. This methodology can be proposed as a viable,
Iron polysaccharides; environmentally sustainable substitute for the conventional Normal Pulse Polarographic method in
Voltammetry US Pharmacopeia, with better analytical figures of merit, and reduced Hg consumption. A deeper
insight in Fe(II) and Fe(III) composition can be gained by the combined use of a new
potentiometric technique after chemical decomposition of the complex.
& 2012 Xi’an Jiaotong University. Production and hosting by Elsevier B.V. All rights reserved.

1. Introduction used to stabilize iron hydroxide nanoparticles in the colloidal

Most iron deficiency anemias respond well to treatment with oral
or parenteral iron; in the latter case, polymeric complexes of
Fe(III) with sugars, such as iron sucrose complex (ISC) and
sodium ferric gluconate complex (SFGC) [1–5], are frequently
n
Corresponding author. Tel.: þ39 0382 987581; fax: þ39 0382 528544.
E-mail address: antonella.profumo@unipv.it (A. Profumo)
Peer review under responsibility of Institute of Materia Medica, Chinese
Academy of Medical Sciences and Chinese Pharmaceutical Association.
suspension [6,7]. According to USP (United State Pharmacopeia),
they must contain less than 20% Fe(II) with respect to the total
iron (5.3%–6.4% and 4.5%–7.0%, for ISC and SFGC respec-
tively), and low molecular weight (MW) complexes must be
undetectable by gel permeation chromatography and normal
pulse polarography (NPP) [8]. The standard method reported on
Pharmacopoeia, and still used, is, in fact, NPP in acetate ionic
strength buffer.
In this paper, a robust voltammetric method at hanging
dropping mercury electrode (HDME) for the determination of
Fe(II) and Fe(III) in SFGC and ISC formulations at about 1%
(w/v) concentration, is described and validated against standard
USP method. This methodology combines greater sensitivity
and lower limit of detection (LOD) with reduced mercury

2095-1779 & 2012 Xi’an Jiaotong University. Production and hosting by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.jpha.2012.05.003
An analytical method for Fe(II) and Fe(III) determination in iron complexes by DPV
consumptions. A redox titration method for the selective determi-
nation of Fe(II) is also reported.
2. Experimental
Reagents and chemicals
Chemicals were obtained from Aldrich and used as received.
Solutions were prepared with Milli-Q grade water. Ceric
ammonium sulfate was standardized against As 2O3 by poten-
tiometric titration. ISC and SFGC were prepared and purified
according to known procedures [9,10].
Instrumentation
Electrochemical measurements were performed in a conven-
tional three-electrode cell (volume 10 mL) with HDME as the
working electrode (BASi, PWR-3 interfaced to EF-1400
controlled growth Hg electrode), Pt wire as the auxiliary
electrode, and Ag/AgCl (3 M NaCl) as the reference electrode.
Potentiometric titrations were performed using a Pt work- ing
electrode with an Ag/AgCl (3 M KCl) reference electrode
on an Orion 520 potentiometer.
Quantitative calculations
Fe(II) in the preparations is calculated by the Eq. (1) from the
Σ
NPP or differential pulse voltammetry (DPV) response ratios 1—
. ΣΣ
2 R
~ ½Fe] ¼ %FeðII Þ ð1Þ
where R is Fe(II)/Fe(0) (henceforth denoted R1) to Fe(III)/
Fe(II) (denoted R2) wave response ratio; [Fe] is the total iron
concentration in the vial, typically 20 g/L.
According to Eq. (1), in the absence of Fe (II) in the original
solution, the theoretical R value should approximate two, since
Fe(III)/Fe(II) is one-electron reduction process and Fe(II)/Fe(0)
is bielectronic [11], and all the Fe(II) reduced to Fe(0) derives
from the reduction of the Fe(III) complex in the sample.
nFAD1=2c
ilim ¼ ð2Þ
p1=2t1=2
p
In Eq. (2), the symbols have their usual meaning, and ilim is
calculated from NPP wave or by integration of the DPV peak.
Figure 1 DPV profile of ISC. Experimental conditions described in the text.
451
3. Results and discussion
Determination of Fe(II) by NPP
In the USP polarographic method for the determination of
Fe(II) in ISC aqueous solutions, the half-wave potential, E1/2, is
related to the MW of the complex [8]. The two waves at—750
mV and 1400— mV correspond to Fe(III)/Fe(II) and Fe(II)/
Fe(0) reductions, respectively. The absence of additional peaks
may, therefore, exclude the presence in the sample of low MW
complexes at concentration higher than 5% of the total iron
complex. According to Eq. (1), R values greater than two (i.e. the
second wave height greater than twice the first), must be
observed whenever Fe(II) is present in the initial solution.
However, the waves often show poor reproducibility, in terms
of current intensity and E1/2 values. As a consequence, the
calculated concentration of Fe(II) tends to be extremely high or
even negative, with no analytical significance. Such large R
values are likely to be related to systematic overestimation of
Fe(II), which is, instead, unlikely to occur during the preparation
of ISC and SFGC, where the highly alkaline melt comprising
mixture of ferric oxide and sucrose or glucose would result in a
fast oxidation of Fe(II) by air. This prompted us to undertake an
investigation of the electrochemical behavior of the two Fe
complexes with the final aim to devise an alternative analytical
method that could ensure adequate reproducibility and pre-
cision, as required in Quality Control and Quality Assurance
protocols, reducing, at the same time, Hg consumption during
the analysis.
Determination of Fe(II) by DPV
DPV was initially carried out using the instrumental para-
meters reported in the USP methodology [8], and successively
optimized. The final working conditions were: LiClO 4 50 g/L,
SFGC or ISC 330 mg/L, step 30 mV; pulse width 100 ms;
pulse period 325 ms, pulse amplitude 100 mV; scan rate
92 mV/s; E¼ i 0 V; E¼— f 1700 mV. A two-peak voltammetric
profile (Fig. 1) is observed for ISC, and peak height is shown
to be dependent on ISC concentration. The peak around
—700 mV is due to [Fe(III)]/[Fe(II)] reduction, whereas the
—
second peak, around 1400 mV, is ascribed to [Fe(II)] reduc-
tion to zerovalent iron. Analogous results were observed with
the SFGC.
A good linearity between current and concentration was
found in the range 0.05–50 mg/L Fe for both iron
452
Table 1 Voltammetric results obtained in the determina-
tion of Fe(III) and Fe(II) by calibration curves.
Complex Fe(III) peak Fe(II) peak (total iron
quantification)
ISC I (mA)¼ 0.017(2) mg/ I (mA)¼ 0.035(2) mg/
LFe(III)—0.1(1) LFe—0.2(3)
R 2 ¼ 0.998 (10 data R 2 ¼ 0.997 (10 data
points) points)
SFGC I (mA)¼ 0.015(2) mg/ I (mA)¼ 0.032(1) mg/
LFe(III)—0.1(1) LFe—0.1(2)
R 2 ¼ 0.998 (10 data R 2 ¼ 0.996 (10 data
points) points)
formulations, using lithium perchlorate as a non-complexing
supporting electrolyte (Table 1).
The determination of Fe(II) and Fe(III) complexes in real
samples can be obtained, without any pre-treatment or
separation, by standard addition method using ISC and SFGC
standard solutions. Noteworthy, this method can allow the
analysis of samples even if the standard solutions are not
available. In fact, the R1/R2 ratio can be used to quantify
Fe(II) selectively and, after determination of the total iron
content by alternative methods, typically by spectroscopic
techniques (e.g. ICP-OES, AAS), Fe(III) is calculated from the
difference between these two values. With this simple
approach, Fe(II) concentrations as low as 1% of total iron
can be easily quantified.
The effect of the analytical parameters, in particular the
composition of the supporting electrolyte, was found to have a
deep influence on the electrochemical behavior, leading to
modifications of the shape and of the ratio of the two
voltammetric peaks. The best results were obtained with
lithium perchlorate at a concentration of 50 g/L. The effect
of pH on the electrochemical behavior was also investigated in
the lithium perchlorate supporting electrolyte. As the result,
no buffering of the electrolyte solution was required, provided
that the pH was greater than 4.
Influence of free Fe(III) and Fe(II) and weaker iron
complexes
In order to evaluate the effect of free Fe ions on the
composition of the complexes, ferric and ferrous ammonium
sulfate were added to the solution of each iron formulation
and voltammograms were recorded subsequently. It was
verified that the addition of free Fe(II) and Fe(III) shifted
the equilibria of SFGC and ISC toward new complexes with
intermediate MW [2]. Interestingly, an analogous behavior
was also observed with labile Fe(II) complexes, e.g. the
gluconate complex [12].
Validation of the proposed DPV method
The analytical method was validated according to Interna-
tional Conference on Harmonization (ICH) guidelines [13].
The linearity, assessed by linear regression determinations,
D. Merli et al.
Table 2 Comparison of R1/R2 values obtained by DPV
and NPP methods (n ¼ 3).
Complex DPV NPP
R1/R2 RSD (%) R1/R2 RSD (%)
ISC 2.03 3.8 1.86 12.4
SFGC 2.09 3.6 1.77 11.5
Expected value 2.0 2.0
was calculated by the least-square regression method. The
calibration graphs were obtained with 10 standard solutions in
the concentration range 0.05–50 mg/L as Fe, both on ISC and
SFGC. The correlation coefficient (r) value was found to be
0.998. A set of 6 samples containing 20 mg/mL, as Fe, of each
pharmaceutical formulation was analyzed to assess repeat-
ibility and precision. The repeatability was evaluated by
assaying samples during the same day, whereas the intermedi-
ate precision was investigated by comparing results on two
different days. LOD and LOQ values for the complexes
evaluated from the linear regression were 15 mg/L and 50 mg/L,
for ISC and SFGC (as Fe), respectively. Recovery was evaluated
by addition of known amounts of standard solutions of each
drug to the commercial formulations. The spiked solutions were
then analyzed by the proposed DPV method and the results were
in the range 94%–117%. Similarly, the intraday and interday
precisions showed good results, with percentage errors ranging
between 3.2% and 4.5%.
The proposed method was then validated against the USP
NPP method. The same standard samples, with an expected
R1/R2 ratio of 2.0, have been analyzed by both (NPP and
DPV) methods. The comparative results are shown in Table 2.
The proposed DPV method has been found statistically more
precise (RSD 3.6%¼ against 11.5% for SFGC and RDS
3.8%¼against 12.4% for ISC) and accurate than NPP, and,
therefore, it can be directly applied to the analysis
of the pharmaceutical preparations.
Potentiometric determination of Fe(II) in ISC and SFGC
In order to assess the total Fe(II) concentration in the samples,
a potentiometric determination using Ce(IV) and ferroine as
indicator, was undertaken after the complete chemical decom-
position, evidenced by disappearance of color, of the com-
plexes (200 mg) with 2 mL of a 1:1 mixture of commercial
96% H2SO4/85% H3PO4. Phosphoric acid stabilizes Fe(III)
ions and leads to a sharp titration end point. Hence, Fe(II)
concentration as low as 1% with respect to the total iron,
could easily be quantified. This potentiometric method was
compared with the DPV Fe(II) analysis, using standard ISC
and SFGC samples, as well as a ISC preparation contami-
nated with Fe(II). Fe(II) concentration was found to be below
LOD in the standard samples, and 0.061 mmol/g ISC (DPV)
and 0.066 mmol/g ISC (potentiometric) in the contaminated
sample. The good agreement of the results obtained by the two
methods confirmed the reliability and the accuracy of the DPV
proposed method for Fe(II) and Fe(III) quantification in iron
sucrose and sodium ferric gluconate complexes.
An analytical method for Fe(II) and Fe(III) determination in iron complexes by DPV
Acknowledgment
This work was financially supported by FAR, Fondo Ateneo
per la Ricerca Universita´di Pavia, Italy. We thank Prof. Nick
Serpone for useful discussions.
References
[1] L. Gyurcsik, B. Nagy, Metal complexes of carbohydrates and
their derivatives: coordination equilibrium and structure, Coord.
Chem. Rev. 203 (2000) 81–149.
[2] L. Nagy, A.J. Szorcsik, Equilibrium and structural studies on
metal complexes of carbohydrates and their derivatives, Inorg.
Biochem. 89 (2002) 1–12.
[3] F. Jones, H. Colfen, M. Antonietti, Iron oxyhydroxide colloids
stabilized with polysaccharides, Colloid Polym. Sci. 278 (2001)
491–501.
[4] L.L. Brunton, J.S. Lazo, K.L. Parker, in: Goodman & Gilman’s
The Pharmacological Basis of Therapeutics, 11th Edition,
McGraw-Hill, New York, 2006, pp. 1442–1450 (Italian edition).
[5] S. Sweetman, in: Martindale: The Complete Drug Reference, 35th
Edition, Pharmaceutical Press, New York, 2007, pp. 1366–1370.
View publication stats
453
[6] E.M. Coe, L.H. Bowen, J.A. Speer, et al., The recharacterization
of a polysaccharide iron complex (Niferex), J. Inorg. Biochem. 58
(1995) 269–278.
[7] D.S. Kudasheva, J. Lai, A. Ulman, et al., Structure of carbohy-
drate-bound polynuclear iron oxyhydroxide nanoparticles in
parenteral formulations, J. Inorg. Biochem. 98 (2004) 1757–1769.
[8] United States Pharmacopeia, 31th Edition, US Pharmacopeial
Convention, 2008, 2, pp. 2449–2450.
[9] Y.-Y. Zhang, J.-H. Liu, Optimization of process conditions for
preparing an iron-polysaccharide complex by response surface
methodology, Chem. Biochem. Eng. Q. 25 (1) (2011) 75–81.
[10] L. Nagy, T. Gajda, J. Ku¨rti, et al., Spectroscopic studies of
iron(III) complexes of D-saccharose and D-glucose in the solid
state and in solution, J. Radioanal. Nucl. Chem. 209 (1) (1996)
225–234.
[11] F. Scholz, Electroanalytical Methods: Guide to Experiments and
Applications, Springer Verlag, Berlin, 2002, pp. 105.
[12] R.L. Pecsok, J. Sandera, The gluconate complexes. II. The ferric-
gluconate system, J. Am. Chem. Soc. 77 (6) (1995) 1489–1494.
[13] (a) ICH Q2A, Validation of Analytical Procedures: Definitions
and Terminology, Geneva, 1995, in 2005 incorporated in Q2(R1)
(b) ICH Q2B, Validation of Analytical Procedures: Methodology,
adopted in 1995, Geneva Q2B, in 2005 incorporated in Q2(R1).
 Microporous and Mesoporous Materials 237 (2017) 168e179

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials

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

Iron within the erionite cavity and its potential role in inducing its
toxicity: Evidence of Fe (III) segregation as extra-framework cation
Alessandro Pacella a, Marzia Fantauzzi b, Davide Atzei b, Carlo Cremisini c, Elisa Nardi c,
Maria Rita Montereali c, Antonella Rossi b, Paolo Ballirano a, d, *
a
Dipartimento di Scienze della Terra, Sapienza Universita
di Roma, Piazzale Aldo Moro 5, I-00185 Roma, Italy
b
Dipartimento di Scienze Chimiche e Geologiche, INSTM Research Unit, Centro Grandi Strumenti Universita
di Cagliari, I-09042, Monserrato, Cagliari, Italy
c
ENEA, C.R. Casaccia, via Anguillarese 301, I-00123 S. Maria di Galeria, Roma, Italy
d
Rectorial Laboratory Fibres and Inorganic Particulate, Sapienza Universita
di Roma, Piazzale Aldo Moro 5, I-00185 Roma, Italy

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

Article history: In this work we report the results of the crystal chemical, structural and surface characterization of
Received 11 May 2016 erionite-K fibres from Rome (Oregon, USA) after interaction with Fe (III) chloride solutions at different
Received in revised form concentrations. In addition, Fe (III) loaded samples were investigated after incubation in ascorbic acid in
7 September 2016
order to monitor the mobility of reduced Fe (II) and to highlight its possible incorporation as EF cation
Accepted 13 September 2016
through ion exchange. Comparison between released and acquired charges under the form of Fe con-
Available online 15 September 2016
firms, in perfect agreement with previous studies that Fe (III) is mainly fixed at the fibre surface.
Nevertheless, in very diluted Fe (III) solutions (below 50 mM FeCl3) a significant fraction of Fe (III) is
Keywords:
Erionite
segregated by an ion-exchange mechanism in the erionite cavity at the Ca3 site, albeit with a significantly
Malignant mesothelioma lower efficiency with respect to Fe (II). It is worth mentioning that, as a result of the catalytic properties
Fe (III) binding of zeolites, the location of iron in well-defined crystallographic positions is the prerequisite for behaving
Ion exchange as a very active site in the generation of reactive oxygen species. Incubation in ascorbic acid revealed that
Surface characterization only Fe (III) residing at the fibres surface and characterized by low nuclearity is significantly reduced,
whereas this reaction does not occur (or possibly occurs very marginally) in the case of the ion-
exchanged metal. Considering that the total iron in lung fluids occurs at very low concentration (ca.
0.21 mM), our results strongly suggest that the physiological environment unfortunately represents the
optimum condition for iron to behave as a very active site.
© 2016 Elsevier Inc. All rights reserved.

1. Introduction Cappadocia, Turkey [8e10]. However, in recent years a growing

Erionite is a zeolite with fibrous morphology that has been
classified as a Group 1 Human-Carcinogen by the International
Agency for Research on Cancer (IARC) [1,2]. In fact, its inhalation has
been shown to cause malignant mesothelioma (MM), with
tumorigenicity 500e800 times greater than that of chrysotile and
crocidolite asbestos [3]. Being prevalently the product of diagenetic
alteration by weathering processes of sediments in volcanic ash
[4,5], it is relatively widespread worldwide [6]. It commonly occurs
associated with structurally related offretite [7]. Erionite-related
MM cases have been reported since the 70's in several villages of
* Corresponding author. Dipartimento di Scienze della Terra, Sapienza Universit

a
di Roma, Piazzale Aldo Moro 5, I-00185 Roma, Italy.
E-mail address: paolo.ballirano@uniroma1.it (P. Ballirano).
concern about the carcinogenetic potential of erionite has spread
out in the public opinion in the USA because of the reported
ubiquitous use in several areas of the country of building materials
containing erionite [11e15]. Recently, it has been reported the
possible occurrence of erionite-related pulmonary diseases in Iran
[16].
From the structural point of view, erionite belongs to the so-
called ABC-6 family [17]. This family is built up from the stacking
along the c-axis of layers of six-membered rings of (Si,Al)O4 tetra-
hedra following an ABC scheme. Erionite is one of the four known
minerals characterized by a six-layers sequence: AABAAC, erionite
[18], AABBCC, chabazite [19], ABBACC, bellbergite [20], and ABA-
BAC, liottite [21]. It is hexagonal, space group P63/mmc, and has an
average chemical formula K2(Na,Ca0.5)8[Al10Si26O72]$30H2O. Three
different species, erionite-K, erionite-Ca, and erionite-Na are
recognized on the basis of the most abundant extra-framework (EF)

http://dx.doi.org/10.1016/j.micromeso.2016.09.021
1387-1811/© 2016 Elsevier Inc. All rights reserved.
 A. Pacella et al. / Microporous and Mesoporous Materials 237 (2017) 168e179
cation [22]. The framework of erionite is composed of columns in
which cancrinite ( ) cages and double-6 rings (D6R) regularly
alternate [18,23]. Adjacent columns are fastened by a single six-
membered ring, which connects cancrinite cages at the same
level, thus forming the larger erionite cavities. Cancrinite and
erionite cages host the EF cations: the former cage contains a K ion
placed at its centre (site K1) and several EF cation sites have been
located along the axis of the erionite cavity [23e28]. Finally, an
additional K2 site, residing at the centre of the larger openings of
the erionite cage formed by the 8-member rings (8 MR), has been
found in erionite-K and related to the presence of extra-K ions
[27,28].
The molecular mechanism/s by which erionite induces cell
damage is/are still unclear. By analogy with asbestos it has been
proposed that the presence and structural coordination of Fe,
especially Fe (II), play a primary role in the genotoxic effect of the
fibres. However, in spite of its high carcinogenicity, erionite is a
nominally iron-free phase. Nevertheless, some authors reported
the presence of iron in several erionite samples [29,30] that has
been recently shown to be located at the zeolite surface under the
form of oxy-hydroxides [27,31], sulphates [31] or thin coating of
iron-bearing phyllosilicates [32]. Owing to the ion-exchange ca-
pabilities of zeolites, the biological activity of erionite has been
related to ion-exchanged and/or surface-deposited Fe participating
in Fenton chemistry [33e37]. In particular, Fe (III) is supposed to be
mainly fixed at the surface, Fe (II) being ion-exchanged [34].
However, the samples used in the chemical reactivity and cell
toxicity tests [34,36e38] were incompletely characterized from the
crystal chemical point of view and their surface characterization
was missing. Recently, by combining chemical (scanning electron
microscopy energy dispersive X-ray spectroscopy, SEM-EDX),
structural (X-ray powder diffraction, XRPD) and surface (X-ray
photoelectron spectroscopy, XPS) data, it was provided a detailed
picture of the crystal chemical modifications occurring into erionite
fibres whenever kept in contact with a source of Fe (II) [39]. Results
univocally indicate that Fe (II) is segregated within the erionite
cavity, at the Ca3 site with a six-fold coordination to water
molecules.
Here we report the results of the crystal chemical, structural and
surface characterization of erionite-K fibres from Rome, performed
after interaction with Fe (III)-containing solutions. We have used
FeCl3 as a source of Fe (III) following the same successful procedure
used by Eborn and Aust [34] for comparison purposes. The choice of
those authors is justified by the will of employing a simple solution,
from the chemical point of view, owing to the ion-exchange
property of the zeolite. In fact, due to the combined large ionic
radius and negative charge of chlorine, HCl was used for blank
measurements, in order to obtain the same pH conditions of FeCl3
solutions following the same procedure of reference [34]. Due to
the complexity of the task and to reduce the number of variables to
be controlled during experiments we have chosen, as in the case of
Fe (II) [39] a simplified system exclusively mimicking pH conditions
compatible with those found within the lung. The next step will be
to perform similar experiments in a significantly more complex
environment as Gamble's solution or ALF.
Modifications of the fibres were investigated by using the multi-
analytical approach adopted by the present research group in ref.
[39] coupling SEM-EDX, XRPD, ICP-OES (inductively-coupled
plasma optical emission spectrometry) and XPS. In addition, Fe
(III)-loaded samples were incubated in ascorbic acid (AA) in order
to monitor the mobility of reduced Fe (II) and to highlight its
possible incorporation as EF cation through ion exchange. AA is the
compound commonly used by biochemists to mimic reducing
conditions that may occur in vivo.
169
Therefore, the aim of this investigation is to complete the
atomistic description of the behaviour of erionite fibres whenever
kept in contact with Fe sources. This piece of information is of
relevant importance toward a detailed comprehension of the
mechanism/s inducing carcinogenicity and for developing possible
inactivation routes.
2. Experimental
2.1. Sample description
The erionite pristine sample is from Rome, Oregon, USA. The
hand specimen has relevant chemical variability resulting in
extended patches of erionite-Na alternating with prevailing
erionite-K areas. Erionite-Na found in this specimen is Ca-free and
may be accordingly identified with ease by both SEM-EDX and,
owing to slightly different cell parameters, by XRPD preliminary
analyses. Particular care was paid to identify and separate large
areas of the sample consisting exclusively of erionite-K, which was
used throughout the present investigation. Its crystal-chemical
formula, as obtained from SEM-EDX analysis, is (Na1.45K2.98-
Ca0.61Mg0.63)[Al7.42Si28.58O71.74]$29.49H2O [40]. The raw material
was subsequently enriched following an especially tailored pro-
cedure [41].
2.2. Fe (III) binding to erionite
Erionite-K fibres (25 mg) were suspended in 25 mL of 25, 50,
100, 250 and 1000 mM FeCl3 solutions in a 50 mL Falcon poly-
propylene tube. Experiments were performed in triplicate. In
addition, a “blank” procedure was carried out suspending erionite-
K fibres (25 mg) in water solutions (without FeCl3) adjusted to the
same pH of the FeCl3 solutions with HCl (Ultrex II, ultrapure re-
agent, J.T. Baker) diluted in ultrapure deionized water (Milli-Q -
Millipore, Bedford, MA, USA). Experiments at pH 2.7 and 3.0 were
performed in triplicate, the others in duplicate. All the poly-
propylene tubes, closed and sealed with Parafilm®, were placed in
an oscillating thermostatic water bath at 37 C. After 1 h of treat-
ment the supernatant solution was sampled and filtered, using a
0.22 mm nitrocellulose membrane filter, for ICP-OES analysis. The
fibres were recovered from the tube on filter, rinsed with ultrapure
deionized water to remove any residues of iron not bound, and
dried prior to the experiments of “incubation” in ascorbic acid and
SEM-EDX, XRPD, and XPS measurements.
2.3. Incubation of the bres in ascorbic acid (AA)
Upon completion of Fe (III) binding experiments the fibres were
recovered from the filters, dried and homogenized. The accurately
weighted fibres (about 25 mg) were suspended in 25 mL of a 6 mM
AA solution in a 50 mL Falcon polypropylene tube. All the tubes,
sealed with Parafilm®, were placed in an oscillating thermostatic
bath set at the temperature of 37 C. Fibres previously treated with
250 and 1000 mM FeCl3 solutions were incubated for 1 h and 24 h,
whereas those treated with 25, 50 and 100 mM FeCl3 solutions were
incubated for 24 h only. A “blank” procedure was also carried
incubating for 1 h and for 24 h in the 6 mM AA solution the fibres
previously incubated for 1 h in water solutions (without FeCl3)
adjusted to pH 3.0 and 2.7 (the same pH of the 250 and 1000 mM
FeCl3 solutions). All the tests were performed in duplicate.
After completion of the incubation in AA solution, the tubes
were transferred in a nitrogen-filled glove bag and the supernatant
solution was sampled and filtered, using a 0.22 mm nitrocellulose
membrane filter, for ICP-OES analysis. Finally, the fibres were
recovered from the tube on filter, rinsed with ultrapure deionized
170 A. Pacella et al. / Microporous and Mesoporous Materials 237 (2017) 168e179
water to eliminate residues of the solution, dried and stored under
nitrogen prior to SEM-EDX, XRPD, and XPS measurements.
2.4. Samples characterization
2.4.1. Surface area
Surface area, BrunauereEmmetteTeller (BET) multipoint
method [42] and textural analysis were obtained by N2 adsorption/
desorption measurements, at the liquid nitrogen temperature
( 196 C), using a Micromeritics 3-Flex analyser. Owing to the
small amount of treated samples the analysis was limited to the
pristine one. Nevertheless, no significant variation of the measured
parameters is expected upon iron loading as a reduction of the
surface area is attained exclusively by inserting extra K cations that
are segregated at the centre of the large openings of the erionite
cavity [43]. The sample was pre-treated under vacuum at 65 C for
15 h. The pores distribution was determined from the adsorption
curve by the BarreteJoynereHalenda (BJH) method [44] and from
the analysis of the micropore isotherm by the t-test [45] taking the
curve of Harkins and Jura [46]. The total pore volume was deter-
mined by the rule of Gurvitsch [47].
2.4.2. Scanning electron microscopy e energy dispersive X-ray
spectroscopy (SEM-EDX)
The micro-chemical characterization of Fe (III) loaded and AA
incubated samples was performed using a Quanta 400 SEM
equipped with an EDX Genesis EDS system (FEI, Hillsboro, OR, USA).
Operating conditions were: 15 kV accelerating voltage, 11 mm
working distance, 0 tilt angle. Chemical data were collected at
least at 12 analytical points. The final crystal chemical formulae
were calculated, after renormalization of the chemical analyses
hypothesizing a water content of 18.5 wt% (corresponding to ca. 30
atoms per formula unit, apfu), on the basis of 36 (Si Al) apfu. Both
the balance error formula E% [48] and the K content test [49] were
used for selecting the reliable analyses.
2.4.3. ICP-OES analysis
One mL of each filtered solution was diluted (1:10) with 1% nitric
acid solution and analysed by ICP-OES in order to measure the
amount of Ca, Mg, Na, K, Si and Al released by the erionite fibres and
that of Fe persisting in solution after their suspension. All mea-
surements were performed using an Optima 2000 DV ICP-OES
spectrometer (Perkin-Elmer, Waltham, MA, USA) equipped with a
cross flow nebulizer placed inside a Scott spray chamber. ICP Aristar
(BDH) standard solutions in nitric acid for Ca, Mg, Na, K, Si, Al and Fe
(1000 mg/L) were used to prepare the calibrating solutions for ICP-
OES analyses. To ensure adequate quality assurance, the analyses of
the standard solutions were regularly repeated after the measure-
ment of a defined number of samples.
ICP data of cation release and of Fe bound to erionite fibres are
listed in the present work under the form of nmol/mg of erionite.
The amount of Fe bound to erionite was calculated by difference
between the Fe concentration measured in the initial solution
([Fe]0) and that persisting in solution after 1 h of sample incubation
([Fe]1h). The net cation release was calculated by difference be-
tween the values measured after suspension in the FeCl3 solution
and those in the blank procedure (suspension of fibres in HCl so-
lutions with the same pH of the FeCl3 ones). In order to verify
whether Fe (III) was acquired by erionite through ion exchange or
fixed at the surface, a charge balance was calculated using the same
approach detailed in ref. [42]. In particular, the number of charges
of each element was determined multiplying the nanomoles by the
valence of the corresponding element and finally, the net charges
released into solution were compared to the total charges acquired
as Fe.
2.4.4. X-ray photoelectron spectroscopy investigation
X-ray photoelectron spectra were acquired using a Thetaprobe
(Thermo Fisher Scientific, Whaltman MA, USA). The fibres were
mounted on a standard sample holder using a bi-adhesive graphite
tape. Data were collected using a monochromatic AlKa source
(energy: 1486.6 eV) with a spot size of 400 mm at 6.7 mA and 15 kV.
The base pressure was set to 6 10 8 Pa. Survey spectra were ac-
quired in fixed analyser transmission (FAT) mode setting the pass
energy (PE) to 200 eV, while the high-resolution spectra of C1s,
O1s, Si2p, Mg1s, Ca2p, Na1s, K2p, Fe2p and Cl2p were collected
with a PE of 100 eV in the standard lens (SL) mode. In the latter
condition the full-width at half-maximum (FWHM) of the silver
Ag3d5/2 peak was of 0.83 eV. Step size of 1 eV and 0.05 eV were
selected for the survey and the narrow scans, respectively. The
linearity of the binding energy (BE) scale was checked by periodic
calibration according to ISO 15472:2010 [50]. BE values are referred
to the aliphatic C1s signal at 285.0 eV.
Spectra were processed using CASAXPS software [51]. Before
applying the curve-fitting procedure, the background was sub-
tracted according to the Shirley-Sherwood method [52]. The
product of Gaussian and Lorentzian functions was used for curve
fitting. Quantitative chemical analysis of fibre surfaces was per-
formed using the first-principle method [53] assuming sample
homogeneity. Peak areas were corrected for the sensitivity factors.
Further details about areas correction parameters are given in ref.
[54]. The accuracy of the calculated atomic concentrations is esti-
mated to be ±10%. Binding energy values and atomic percentages
are reported in this work as means on three independent mea-
surements with their corresponding standard deviations.
2.4.5. X-ray powder diffraction (XRPD)
Powders of the various samples were loaded into 0.5 mm
diameter SiO2-glass capillaries. XRPD data were collected in
transmission mode using a D8 Advance diffractometer (Bruker AXS,
Karlsruhe, Germany) operating in q/q geometry. The instrument is
fitted with focusing Go €bel mirrors on the incident beam, Soller slits
on both incident and (radial) diffracted beams, and a PSD VÅNTEC-1
detector. A preliminary diffraction pattern indicated the occurrence
of minor chabazite (ABC-6 family zeolite: ideally (Ca0.5,N-
a,K)4[Al4Si8O24]$12H2O) and nontronite (phyllosilicate: ideally
(Ca0.5,Na)0.33Fe3 2(Al0.33Si3.67)O10(OH)2$nH2O) and traces of quartz.
Therefore, a mixed Rietveld/Pawley method was adopted to take
into account the small contributions of impurities [41]. Structure
refinements were carried out using TOPAS v.4.2 [55] and modelling
the peak shape by FPA (Fundamental Parameters Approach). The
selected starting structural model of erionite was that of ref. [49].
Structural data of chabazite were taken from Yakubovich et al. [56]
allowing for refinement of cell parameters only. Structural data of
a-quartz were taken from ref. [57] and kept fixed. Finally, the
nontronite lattice used for the Pawley refinement was taken from
ref. [58]. The well-established Rietveld refinement procedure
adopted by the present research group in several structure re-
finements of erionite fibres was followed in the data analysis (see
for example [41]). Absorption correction was performed following
the formalism of Sabine [59] and the occurrence of preferred
orientation was modelled by spherical harmonics (six refinable
parameters up to the 8th order). The choice of the number of terms
to be used has been performed following the procedure described
in ref. [60]. Experimental details and miscellaneous data of the
refinements are listed in Table S1. A representative example of
Rietveld plots of the pristine sample is reported in Fig. S1. As in the
case of the investigation of the Fe (II) loading process of erionite
[39], a second set of diffraction patterns were collected on the same
samples using a parallel-beam D8 Focus diffractometer (Bruker
AXS, Karlsruhe, Germany) equipped with a Si(Li) solid state SolX
 A. Pacella et al. / Microporous and Mesoporous Materials 237 (2017) 168e179
Fig. 1. Fe (III) bound to erionite as function of the FeCl3 concentration. Data of Eborn
and Aust [42] are reported for comparison. Where no error bars are visible, the
standard deviation is within the symbol.
detector and operating in q/2q transmission geometry. This addi-
tional data collection was carried out for monitoring the intensity of
the FeKa and the FeKb fluorescence lines on the various samples, by
exploiting the capabilities of the solid-state detector [61].
3. Results
3.1. Solutions analysis
Erionite samples were suspended in 25 (pH 3.7), 50
(pH 3.5), 100 (pH 3.2), 250 (pH 3.0) and 1000 mM (pH 2.7)
FeCl3 solutions for 1 h and, for comparison, in HCl solutions with
the same pH of the ferric solutions.
ICP-OES analyses of the supernatants were performed to mea-
sure the amount of the released cations from and the amount of the
iron bound to erionite.
The incubation of erionite in FeCl3 solution resulted in a
decrease of iron content in the FeCl3 solution. The amount of iron
bound by erionite was calculated starting from
[Fe]b [Fe]0 e [Fe]1h
where [Fe]0 is the starting concentration of iron in the FeCl3 solu-
tions and [Fe]1h is the residual concentration after 1 h of incubation,
determined by ICP analysis. The higher was [Fe]0 the higher was
iron bound by erionite [Fe]b (Fig. 1): the amount of Fe (III) loaded by
the fibres ranges from 23(1) to 376(40) nmol/mg after incubation in
25 and 1000 mM FeCl3 solutions, respectively.
As far as the extra framework (EF) cations, ICP analysis of su-
pernatants highlights a significant release of Na and K during the
suspension of erionite in both the FeCl3 and the corresponding HCl
solutions (Table 1e1.A and 1.B). In FeCl3 solutions with concentra-
tion up to 250 mM, the amount of dissolved EF cations (especially
Na) is higher than in the HCl solutions with the same pH, whereas
no significant differences were found between the samples incu-
bated in the 1000 mM FeCl3 solution and the HCl solution at pH 2.7.
3.2. Solutions analysis after incubation of Fe (III) loaded erionite in
ascorbic acid
ICP data acquired on the FeCl3 treated sample after 24 h of in-
cubation in AA revealed the presence of the EF cations in the su-
pernatant (Table 1.C) with Na, for example, ranging between 329
171
and 23 nmol/mg for the samples previously suspended in FeCl3 25
and 1000 mM solutions, respectively. Together with the EF cations,
small amounts of iron are released after 24 h of incubation time;
the released iron increased with the concentration of the FeCl3
solutions, passing from 2 to 52 nmol/mg for the samples previously
suspended in FeCl3 25 and 1000 mM solutions, respectively. A
moderate growth of the Fe release with incubation time was also
observed considering results on 1 h “incubation” in ascorbic acid of
the samples previously suspended in FeCl3 25 and 1000 mM solu-
tions, respectively (last 2 columns of Table 1.C).
3.3. Bulk analysis
N2 adsorption analysis of the pristine sample indicates a BET
specific surface area, measured in the conventional p/p0 0.05e0.3
range, of 252(5) m2 g 1 and an external surface area of
10.1(5) m2 g 1. The total volume of pores is of 0.151 cm3 g 1, of
which 0.130 cm3 g 1 attributable to micropores. SEM-EDX chemical
analysis revealed a content of ca. 1.6(6) wt% of Fe2O3 loaded by the
fibres incubated in a 250 mM FeCl3 solution (Table 2). Simulta-
neously, a marked reduction of the Na2O content (ca. 40%) in the
FeCl3 treated samples with respect to the pristine one is observed.
The total site scattering (s.s.) of EF cations suffers a moderate
abatement as compared to that of the pristine sample, passing from
92.3 to 81.1 e . Moreover, the analysis of the 250 mM FeCl3 incu-
bated fibres suspended in AA as well as those of the 1000 mM FeCl3
(not shown) incubated fibres and those correspondingly suspended
in AA revealed E% values markedly outside the admitted range of
±10% (up to 40%) so that it was impossible to retrieve reliable
crystal chemical formulae.
For the fibres incubated in the 250 mM FeCl3 solution SEM-EDX
chemical analysis does not show any significant difference between
pre- and post-AA suspension, except for a slight depletion of Na.
Once again, it was not possible to retrieve a reliable crystal chemical
formula since the analyses were characterized by E% values well
outside the admitted range. A significant decrease of the apparent
reduction of the total s.s. of EF cations after incubation at increasing
acidic conditions is observed (Table 2: 1 h at pH 2.9 vs. 24 h at pH
2.7, which corresponds to 81.2 vs. 79.1 e ).
SEM images did not show any morphological difference among
pristine, iron-loaded and AA treated samples (Fig. S2).
3.4. Surface analysis
The surface characterization of the erionite samples (pristine,
treated with 250 and 1000 mM FeCl3 solutions and after incubation
in AA for 1 and 24 h) was performed by XPS. Together with the
signals of the erionite (Si, O, Ca, Na, Al, K, Mg and Fe) only small
amounts of adventitious carbon were found in the survey spectra.
The absence of any chlorine signal in the Fe loaded samples ruled
out the possibility of the occurrence of residual FeCl3.
The quantitative analysis of the fibre surface was performed
taking into account the peak areas of each element corrected for the
respective sensitivity factor (Table 3) [54]. A very good agreement
was found between the chemical composition of the pristine
erionite used for performing the incubation in the 250 mM FeCl3
solution and the corresponding AA-reduction procedure and that of
ref. [39]. A second batch of pristine erionite was enriched from the
same hand specimen, after preliminary check by SEM-EDX and
XRPD, for performing the incubation in the 1000 mM FeCl3 solution
and the corresponding AA-reduction. As the enrichment procedure
was repeated more times in order to obtain a more accurate
removal of accessory phases, fibres persisted for a longer time
within the distilled water (pH ca. 5.5). This extra time produced a
surface modification that was evidenced by lower Si and Al and
172 A. Pacella et al. / Microporous and Mesoporous Materials 237 (2017) 168e179

Table 1
ICP analyses of the supernatants for cation release (calculated as nmol/mg of sample). Standard deviation reported for experiments performed at least in triplicate.

A: Incubation of erionite-K fibres in FeCl3 solutions (25e1000 mM)

25 mM (pH 3.7) 50 m (pH 3.5) 100 m (pH 3.2) 250 m (pH 3.0) 1000 mM (pH 2.7)

Si 1.6(5) 2.9(7) 2.9(6) 1.9 (0.2) 2.4(5)

Al <1 <1 <1 1.9 (0.4) 8.5(16)

Na 84(1) 150(2) 267(6) 367 (10) 465(11)

K 4.6(1) 9.6(3) 21(1) 63 (2) 169(12)

Mg ND ND 0.51(3) 3.0(1) 2.8(1)

Ca ND ND 0.95(7) 5.3(5) 44(8)

B: Incubation of erionite-K fibres in HCl solutions

pH 3.7 pH 3.5 pH 3.2 pH 3.0 pH 2.7

Si 3 3 3 9(2) 5.2(4)
Al 1 <1 <1 1.3(4) 20(1)

Na 45(11) 62(6) 110(11) 193(4) 428(23)

K 2 2.5(1) 16(2) 14(1) 136(6)

Mg ND ND ND 0.8(4) 5.2(4)
Ca ND ND ND ND 32(2)

C: Incubation in ascorbic acid (AA) of Fe(III) loaded erionite-K fibres

25 mM FeCl3 50 mM FeCl3 100 mM FeCl3 250 mM FeCl3 1000 mM FeCl3 250 mM FeCl3 1000 mM FeCl3
24 h in AA 24 h in AA 24 h in AA 24 h in AA 24 h in AA 1 h in AA 1 h in AA

Si 16(1) 17(1) 20.87(4) 21(3) 21.8(1) 7.12(4) 8(2)

Al 11.68(4) 14.0(2) 19(2) 17(1) 16(1) 3.0(1) 3.2(4)

Na 335(2) 296(6) 226.3(3) 117(10) 23(3) 118(3) 22(1)

K 77(2) 80.5(5) 80(2) 79(1) 48(2) 64.8(3) 41(3)

Mg 8.4(1) 10.00(0.04) 11(2) 11(1) 11(1) 2.72(1) 2.4(5)

Ca 21(1) 25.6(0.1) 32(1) 34(2) 29(3) 14.0(1) 12(1)

Fe 1.58(4) 3.19(7) 6(1) 25(5) 52(5) 17(5) 30(7)

% rel. Fe 7 6 7 10 11 7 7

D: Incubation in ascorbic acid of erionite-K fibres previously incubated in HCl solutions

pH 3.0 1 h in AA pH 3.0 24 h in AA pH 2.7 1 h in AA pH 2.7 24 h in AA

Si 6(1) 17(1) 4.5(6) 18(1)

Al 3 10(1) 2.0(4) 9.2(3)

Na 262(8) 264(18) 31(3) 33(1)

K 71.1(3) 88.5(2) 46(2) 58(2)

Mg 2.23(2) 5.5(2) 0.75(5) 3.9(2)

Ca 8 14(1) 14(4) 15(1)

Fe ND ND ND ND

higher O content than in the pristine surface of the former one
(Table 3). However, the Si/Al ratio between the two samples did not
change.
High-resolution spectra of the most intense signals were curve
fitted and the binding energies (BE) of the main peaks are listed in
Table S2. A weak Mg2p signal, very close to the limit of detection,
was only observed in the case of the pristine samples. The peak
fitting of Si2p was carried out using a doublet owing to spin-orbit
coupling. The energy separation between Si2p3/2 and Si2p1/2 was
kept constant and equal to 0.8 eV, the area ratio was constrained to
2:1: this fit agrees with those already reported for other silicon
based compounds [62,63]. The BE value of Si2p3/2 was found to
range between 102.7(0.2) eV and 103.0(0.1) eV. Al2p was curve
fitted with a single peak due to the small separation between Al2p3/
2 and Al2p1/2. The centre of gravity of the signal was found at a BE
ranging from 74.6(0.1) eV to 75.0(0.1). The Na1s signal consists of a
single peak at a BE of 1072.8(0.1) eV. The BE of Ca2p3/2 was found to
be of ca. 349 eV. Ca content was almost constant upon loading and
reduction, being the differences within the experimental uncer-
tainty. K2p consists of a doublet K2p3/2 - K2p1/2 separated by 3.0 eV
and with an area ratio equal to 2:1. The BE of the main component
was found at about 294 eV O1s signal (Fig. S3), that was
multicomponent, showed the contribution assigned to the bridging
oxygen in a silicate at about 532.5(0.2) eV [64] and that of eOH at
about 531.4(0.2) eV [65]; the signal attributed to crystallization
water was found at about 533.5(0.2) eV [66]. An additional peak
was found at ca. 530 eV after incubation in both the 250 and
1000 mM FeCl3 solutions and persisting upon reduction with AA: it
might be tentatively assigned to an oxide signal. It is worth noting
that the ratio between the O1s component of the signals ascribed to
eOH and to bridging oxygen is significantly different for the two
pristine samples being of 0.30(5) and 0.6(1) for the two sample
used for the 250 and 1000 mM FeCl3 incubation and AA-reduction
experiments, respectively. The different ratio supports the hy-
pothesis of different degrees of protonation/surface damaging of
the two pristine samples. The binding energy values of Si2p, Al2p
and O1s peak are in agreement with those reported for zeolite Na-Y
[67], which is classified, like erionite, as an intermediate zeolite (Si/
Al ratios between 1.5 and 5). Fe2p3/2 peak (Fig. S4) was resolved in
its components according to results reported in previous papers on
erionite and amphiboles [39,54,65]. The weak peak observed in the
pristine sample is due to the contribution of minor nontronite [39].
Consistently with the spectra recorded in the case of Fe (II) loading
[39] three main signals were observed at ca. 709, 710.5 and 712 eV
 A. Pacella et al. / Microporous and Mesoporous Materials 237 (2017) 168e179 173

Table 2
Chemical analyses, by SEM-EDX, of both pristine [40] and FeCl3 treated fibres. In addition, data of the fibres incubated in AA for 1 and 24 h are reported.* refers to data
recalculated after removal of Fe (III) lying a the surface (see text for explanation). E% of the data of the 250 mM FeCl3 treated sample modified after removal of surface Fe (III) was
2
Ca2 Sr2 Ba2 Fe3
calculated using the formula E% Al NaNa K K 2 2MgMg 2 Ca2 Sr2 Ba2 Fe3
%. Chemical formulae were not calculated in the case of analyses characterized by E% significantly
exceeding 10%. Standard deviations are reported in parenthesis.

Oxides (wt%) Pristine 250 mM FeCl3 250 mM FeCl3* 250 mM FeCl3 AA 250 mM FeCl3 AA 250 mM FeCl3 AA 250 mM FeCl3 AA
1 h pH 2.9 1 h pH 2.9* 24 h pH 2.7 24 h pH 2.7*

SiO2 59.80(32) 59.71(101) 60.50(92) 59.66(46) 60.42(47) 59.70(46) 60.57(46)

Al2O3 13.18(35) 12.78(54) 12.91(53) 13.24(54) 13.41(55) 13.22(41) 13.41(42)
Na2O 1.56(31) 0.68(25) 0.69(25) 0.50(16) 0.51(17) 0.35(9) 0.35(9)
K2O 4.89(35) 4.23(21) 4.29(21) 4.01(54) 4.06(54) 3.79(49) 3.84(50)
MgO 0.88(17) 0.90(11) 0.96(8) 0.86(13) 0.87(13) 0.87(15) 0.88(15)
CaO 1.19(16) 1.65(23) 1.68(23) 1.78(13) 1.80(14) 1.93(20) 1.96(20)
Fe2O3 e 1.55(61) 0.47(18) 1.45(42) 0.43(13) 1.65(35) 0.49(11)
H2O 18.50 18.50 18.50 18.50 18.50 18.50 18.50
Total 100.00 100.00 100.00 100.00 100.00 100.00 100.00

Si 28.58(18) 28.76(32)
Al 7.42(18) 7.24(32)
Na 1.45(29) 0.64(24)
K 2.98(22) 2.60(14)
Mg 0.63(12) 0.68(6)
Ca 0.61(8) 0.86(12)
Fe3 e 0.17(7)
O 71.74(23) 71.79(16)
H2O 29.49(14) 29.34(35)

R 0.794 0.799 0.793 0.797

M/(M D) 0.782 0.678 0.653 0.62
E% 7.6 8.8 19.1 20.3
Tot. s.s. EF cat. (e ) 92.3 86.1 81.2 79.1

(Table S2) and assigned to Fe (II) bound to oxygen, to Fe (III) bound
to oxygen and to ferric oxide and hydroxides respectively.
The total iron content of the FeCl3 treated samples was higher
than in the pristine erionite, while it decreases after reduction in AA
(Table 3). Upon iron reduction in AA of the 250 mM FeCl3 incubated
sample, the peak at 709.2(0.1) eV starts to increase its intensity
confirming the attribution to Fe (II) variously bound to oxygen
atoms (Fe (II)-O). For the 1000 mM FeCl3-incubated sample no iron
reduction in AA was observed even after 24 h (Table 3).
4. Discussion
4.1. Dissolution of erionite bres
Concerning the erionite framework dissolution, very small
amounts of Si and even lower amounts of Al were released in the
25e250 mM FeCl3 solutions (with the exception of Si in 250 mM
FeCl3 solution) and in the HCl solutions at the corresponding pH
(Table 1.A and 1.B, respectively).
In the solution of 1000 mM FeCl3 and in the corresponding HCl
solution (pH 2.7), the framework experiences a much higher
dissolution of Al. Fe (III) e silica interactions can explain this dif-
ference that might lead to the formation of iron silicate aqueous
complexes and precipitates that limit further framework dissolu-
tion [68]. This higher concentration of Al in the solutions confirms
the preferential release of Al by zeolites incubated in solutions at
low pH [69]. Notably, zeolites characterized by Si/Al 3 (present
case: erionite-K 3.84) are little soluble and subjected to a non-
stoichiometric dissolution i.e. selective removal of Al, which pre-
vents further dissolution [70]. The contribution of admixed cha-
bazite (ca. 3 wt%) has to be considered as negligible as it is even less
soluble than erionite under the present acidic conditions [71].
ICP analysis of supernatants highlights a significant release of Na
and K during the suspension of erionite both in FeCl3 solutions and

Table 3
Surface quantitative analysis in at.% of the erionite samples and distribution of iron components, Fe (II)-O, Fe (III)-O e and Fe (III)-OOH in pristine, loaded and reduced erionite
samples (% of Fe2p3/2 peak area); results are the mean value over three independent measurements and the standard deviations are in parentheses. The anhydrous bulk
composition of the pristine sample, as recalculated from SEM-EDX data is also reported for comparison purpose. Standard deviations are reported in parenthesis.

Si Al O Na Ca K Mg Fe

Pristine bulk 25.2(2) 6.5(2) 63.3(2) 1.3(2) 0.54(7) 2.6(1) 0.6(1) e

Pristine 26.4(4) 6.0(4) 61.6(5) 1.8(3) 0.9(1) 2.1(2) 0.30(4) 0.9(2)
250 mM FeCl3 25.5(5) 6.0(3) 63.4(2) 0.4(1) 0.7(3) 1.58(4) e 2.3(2)
250 mM FeCl3 1 h AA 25.9(5) 5.4(4) 64.6(4) e 0.6(2) 1.3(1) e 2.2(2)
250 mM FeCl3 24 h AA 27.6(4) 5.4(7) 63.8(3) e e 1.3(1) e 2.0(3)
Pristine 22(1) 5.3(2) 66(2) 2.9(5) 0.9(1) 1.8(4) 0.28(3) 0.6(1)
1000 mM FeCl3 19.5(6) 4.2(2) 70.8(8) 0.2(1) 0.6(1) 1.4(2) e 3.3(1)
1000 mM FeCl3 1 h AA 19.6(2) 4.6(2) 71.6(2) e 0.6(1) 1.1(1) e 2.6(1)
1000 mM FeCl3 24 h AA 20.9(6) 3.7(6) 71(1) e 0.7(1) 0.95(3) e 2.5(3)

Iron components e area % of Fe2p3/2

Pristine 250 mM FeCl3 1 h AA 24 h AA 1000 mM FeCl3 1 h AA 24 h AA

Fe (II)-O 3(1) e 33(1) 44(3) e e e

Fe (III)-O 42(2) 34(3) 30(2) 30(3) 48(3) 51(3) 53(5)
Fe (III)-OOH 55(3) 66(2) 37(2) 26(1) 52(2) 49(3) 47(5)
174 A. Pacella et al. / Microporous and Mesoporous Materials 237 (2017) 168e179

Table 4
Charge balance involved in the process of Fe (III) loading of the erionite fibres. For each element, the net charge released was calculated by multiplying the net cation release by
the valence of the cation. In addition, the acquired charge was calculated by multiplying values of bound Fe (III) by 3. Standard deviations are reported in parenthesis. Total wt%
Fe2O3 of samples (TF) has been calculated by considering that 723 acquired charges (Fe) were equivalent to 1.55 wt% Fe2O3 as determined by SEM-EDX for the 250 mm sample.
The wt% Fe2O3 as EF cation (ETF) was calculated by multiplying TF by % Fe (III) as EF cation (%EF).

Net charge (e ) incubation in FeCl3 solutions

25 mM 50 m 100 m 250 m 1000 mM

Total released EF charge 41(9) 96(4) 162(7) 238(7) 94(30)

Acquired charges (Fe) 69(3) 150(0) 285(0) 723(6) 1127(121)
% Fe (III) as EF cation (%EF) 60 64 57 33 8
Total wt% Fe2O3 (TF) 0.15 0.32 0.61 1.55 2.42
wt% EF Fe2O3 (ETF) 0.09 0.20 0.35 0.47 0.19
apfu Fe as EF cation 0.03 0.07 0.12 0.17 0.06

Incubation in ascorbic acid of Fe(III) doped samples

25 mM FeCl3 50 mM FeCl3 100 mM FeCl3 250 mM FeCl3 250 mM FeCl3 1000 mM FeCl3 1000 mM FeCl3
24 h AA 24 h AA 24 h AA 1 h AA 24 h AA 1 h AA 24 h AA

Total released EF charge 472(5) 448(6) 393(8) 216(3) 284(18) 92(7) 151(12)

Incubation in ascorbic acid of pristine samples

1 h AA pH 2.9 24 h AA pH 2.9 1 h AA pH 2.7 24 h AA pH2.7

a
Total released EF charge 353(8) 391(21) 105(13) 129(5)
a
Calculated from the total cation release measured for the pristine sample after incubation in water (at the same pH of the FeCl3 solutions) and the following incubation in
AA.

HCl solutions (Table 1.A and 1.B). In particular, the release of Na
(ranging from ca. 83 to ca. 465 nmol/mg) is higher than that of any
other EF cation, confirming its smallest selectivity among all the
alkali and alkali earth metals elements in erionite [39,43,72].
Moreover, the EF cation release increases significantly as the pH of
the solution decreases (from 3.7 to 2.7), confirming the strong ef-
fect of pH on the (albeit limited) alumino-silicate dissolution, which
has been attributed to the replacement of cations with H [69,73].
The dissolution of cations during the subsequent incubation in
AA shows a clear trend (Table 1.C): the lower the pH of the solution
during loading (higher ferric chloride concentration), the lower was
the overall cation release during incubation. The sum of the
released cations (Na K Mg Ca) decreased from 435 to
111 nmol/mg passing from the 25 to the 1000 mM FeCl3 solution
(Table 1.C). Samples showing a significant Na release during the
Fe(III) loading process are able to release only a minor amount of Na
during incubation in AA and opposite. Note that the total Na release
(sum of loading and subsequent incubation in AA) is quite similar
for all samples (e.g. 415 with respect to 488 nmol/mg for the 25 and
1000 mM FeCl3 treated samples, respectively). These results could
reflect the existence of a maximum accessible volume of the fibre
limiting the possibility of ion exchange.
The charges released (dissolved cations) and acquired (loaded
iron) by erionite were calculated (see Section 2.4). During Fe (III)
loading in the 250 mM FeCl3 solution, the released charges [238(7)]
were only 33% of those bound as Fe (III) [723(6)], Table 4. This
percentage decreased to 8% in the more concentrated 1000 mM
FeCl3 solution (Table 4). Such behaviour is in qualitative agreement
with the results of Eborn and Aust [42], which show that charges
released after incubation in 100 and 500 mM FeCl3 solutions are,
respectively, 1/3 and 1/5 of those acquired by erionite as Fe (III). Our
results unequivocally indicate that Fe (III) is mainly fixed at the
fibres surface, consistently with the hypothesis reported in ref. [34].
The marked decrease of the difference between released and bound
charges with increasing FeCl3 concentration (Fig. 2) is clear evi-
dence that, for more diluted ferric solutions, the ion exchange
mechanism between EF cations and Fe (III) is highly facilitated. In
particular, decreasing the concentrations of FeCl3 solutions from

4.2. Fe (III) binding and release e ion exchange process

Upon suspension of erionite in FeCl3 solutions, both ICP-OES

(Table 1.A and Fig. 1) and XPS results (Table 2) evidenced that
some EF cations, mainly Na, were released and Fe (III) was acquired
by erionite.
The amount of loaded iron determined by solution analysis is in
good agreement with the results obtained by Eborn and Aust [34]
for incubation in low-concentration solutions (Fig. 1). It is worth
noting that, at concentrations exceeding 250 mM FeCl3, there is a
significant decrease of the Fe (III) binding efficiency (Fig. 1). The
amount of Fe (III) bound to erionite at 1000 mM FeCl3 was consis-
tently higher than that of Fe (II) fixed by a sample of erionite-Na at
1000 mM FeCl2 in our previous experiments [39], in agreement with
the results obtained in ref. [42] for incubation in 500 mM FeCl3 and Fig. 2. % Of Fe (III) ion-exchanged and segregated at Ca3 as a function of Fe (III) fixed
FeCl2. These results suggest a limited effect of the EF content on the per mg of erionite fibres. Where no error bars are visible, the standard deviation is
Fe-loading process. within the symbol.
 A. Pacella et al. / Microporous and Mesoporous Materials 237 (2017) 168e179
1000 to 25 mM, the percentage of ion exchanged Fe (III) rises from
ca. 8% to ca. 60% of the total iron acquired by the fibres (Table 4 and
Fig. 2). This might be due to the higher degree of Fe (III) polymer-
ization occurring in solutions with a higher Fe concentration, which
promotes the precipitation of oxy-hydroxides at the fibre surface
[74].
The XPS surface analyses confirm these results: together with a
cation (mainly Na) depletion, the surface of the samples after Fe (III)
loading resulted to be iron e enriched (Table 3) compared to the
pristine samples, arising from small amounts of adhering iron
bearing-nontronite [39], which is known to have a very small
dissolution rate. Nontronite thus contributed to the area of the
Fe2p3/2 signal also following the suspension in FeCl3 and incubation
in AA. To estimate the amount of loaded iron, the difference be-
tween the total iron in the FeCl3 treated samples and the pristine
ones was calculated and, from the fitted Fe2p3/2 spectra (Fig. S4) the
chemical state of iron present on the surface (Fe (II)-oxide, Fe (III)-
oxide, Fe (III)-oxy-hydroxide) and its quantity can be determined.
The percentage of each component in the Fe2p3/2 signal was
calculated (Table 3). It can be noted that after Fe (III) loading only Fe
(III) oxide and oxy-hydroxides are present. On the surface loaded in
the solution of 250 mM FeCl3 about 33% Fe (III)-oxide and 66% Fe(III)
oxy-hydroxide are found. If we attribute Fe (III)-oxide (BE 710.5 eV)
to iron bonded to oxygen atoms of the framework and/or residual
water molecules located within the erionite cavity, there is a
remarkably good agreement with both ICP data indicating that ca.
30% of Fe (III) is ion-exchanged.
Coherently, the peak at 712.2(0.2) eV of the sample incubated in
the 250 mM FeCl3 may be assigned to Fe (III) residing at the surface
(Fe (III)-OOH) [75]. Such value agrees with that reported for lep-
idocrocite (g-FeOOH) [76]. The surface deposition of Fe (III) oxy-
hydroxides is confirmed by the appearance of an additional O1s
signal at 530.3(0.2) eV (see above).
Upon incubation in the 1000 mM FeCl3 solution, the total amount
of iron significantly increases (Table 3) and Fe (III) oxide and oxy-
hydroxide are present in about the same amount. Owing to the
fact that ICP data suggest that only 8% of the acquired Fe by erionite
after incubation in the 1000 mM FeCl3 solution is ion-exchanged, a
significant amount of Fe (III)-O is expected to reside at the surface
of this sample, in addition to that occurring under the form of Fe
(III)-OOH. The significant presence of Fe (III)-O at the surface could
be due to the aggregation of the oxy-hydroxide clusters on the
sample surface leading to the formation of oxo-bridges (Fe-O-Fe) by
dehydration [77].
The higher total Fe (III) content in the sample incubated in the
1000 mM FeCl3 solution compared to that incubated in the 250 mM
solution is in reasonable agreement with the ICP results (376 vs
241 nmol mg 1).
SEM-EDX chemical analysis confirms the presence of iron on the
FeCl3 suspended erionite and the decrease of sodium content
(Table 2) but, as already said (Results section) it was impossible to
retrieve reliable crystal chemical formulae because of the E% higher
than the admitted one (±10%). Such systematic increase of E% could
be due to relevant alkali metal migration (in particular Na) occur-
ring in the external layer of fibres possibly subjected to protonation
and/or amorphisation induced by prolonged incubation at acidic
pH. This damaging is in principle able to magnify the severe elec-
tron beam sensitivity of erionite [78]. This hypothesis is substan-
tiated by the fact that alkali silicate glasses are severely affected by
Na loss [79]. It is presumable that this effect has played a significant
role in the already mentioned moderate abatement of the total s.s.
of EF cations observed in the fibres incubated in a 250 mM FeCl3
solution (from 92.3 to 81.1 e ) as compared to that of the pristine
sample.
175
According to ICP results (Table 1), only ca. 30% of the total Fe2O3
(corresponding to 0.5(2) wt%) of the 250 mM FeCl3 incubated fibres
was considered as loaded within the erionite cage, the remaining
iron being fixed at their surface. The chemical analysis was thus
recalculated and used to retrieve the crystal chemical formula of
the sample. Results shows that 0.17(7) apfu Fe (III) are loaded
within the erionite structure, a value comparable to ca. 0.2 apfu Fe
(II) estimated for the sample suspended in the 100 mM FeCl2 solu-
tion [39]. After recalculation the total s.s. of EF cations rises to 86.1
e , a value which is clearly still underestimated (Table 2). A rough
assessment of the Fe (III) apfu segregated within the framework
may be attempted by rescaling the ICP data of the various samples
with respect to the SEM-EDX data of the fibres incubated in the
250 mM FeCl3 solution i.e. considering the 723 acquired charges (Fe)
(Table 4) as corresponding to 1.55 wt% Fe2O3 resulting from SEM-
EDX (Table 2). As can be seen, the maximum amount of ion-
exchanged Fe (III) occurs in the case of the fibres incubated in the
250 mM FeCl3 solution, whereas in the remaining cases the Fe (III)
apfu are comprised between 0.03(2) and 0.12(5).
4.3. Incubation in ascorbic acid
After 24 h of incubation in ascorbic acid the amount of released
iron increased with the concentration of the FeCl3 solutions
(Table 1.C). Regarding the effect of incubation time it can be noted
that a great part of iron is released during the first hour The per-
centage of iron bound to the fibres that was released into solution
was around 6e7% rising to ca. 10% in the 250 and 1000 mM FeCl3
treated samples. This result is consistent with previous findings of
Eborn and Aust [34] showing that Fe located at the fibre surface is
difficult to be mobilized by ascorbate. The observed EF cation
release in AA of the Fe loaded samples is comparable or even lower
than that of the pristine ones incubated in AA (Tables 1.D and 4).
This result indicates that, after reduction of Fe (III) fixed at the
fibre surface, the possible incorporation of Fe (II) as EF cation
through ion exchange can be considered as non-existent or very
limited. XPS analyses confirmed the release of iron by detection of a
lower total iron concentration on the surface (Table 3). In addition it
could be shown that upon incubation in AA of the 250 mM FeCl3
suspended sample Fe (II) oxide appears after 1 h of immersion
(Fig. S4, Table 3, Fig. 3), thus indicating the reduction of iron Fe (III)
to Fe (II). The appearance of Fe (II), whose concentration increases
by prolonging incubation, is coupled to a marked decrease of the
Fig. 3. Variation of the three components of the Fe2p3/2 spectra as a function of the
treatment.
176 A. Pacella et al. / Microporous and Mesoporous Materials 237 (2017) 168e179
surface loaded Fe (III) and to a minor decrease of the ion-exchanged
Fe (III) compounds (Fig. 3). Those features indicate that reduction is
almost completely undertaken by iron residing at the fibres surface.
Conversely, for the 1000 mM FeCl3-incubated sample no iron
reduction in AA was observed even after 24 h. This result evidences
that the higher degree of Fe (III) polymerization in the solution
promotes the precipitation of both oxides and oxy-hydroxides at
the fibre surface characterized by increased iron nuclearity that
inhibits the interaction between the iron centres and AA.
SEM-EDX chemical analysis of the fibres incubated in the
250 mM FeCl3 solution does not show any significant difference
between pre- and post-AA suspension, except for a slight depletion
of Na in agreement with ICP-OES results. The rate of the apparent
reduction of the total s.s. of EF cations after incubation at increasing
acidic conditions decreases significantly (Table 2) confirming that
the structure damaging effects are restricted to the first few unit
cells and that the external layer may possibly act as a protecting
agent against the deepening of the process.
4.4. Structural modi cations of the Fe loaded erionite-K and after
incubation in AA
Comparison of the FeKa and FeKb fluorescence lines of the
pristine sample and of the samples incubated in the 1000 mM FeCl3
solution pre- and post-AA suspension is reported, as example, in
Fig. 4. Present data confirm the occurrence within the pristine
sample of minor Fe arising from admixed iron-bearing clay min-
erals [32,39,41]. The blank sample, kept in H2O at pH 2.7, shows a
very minor reduction of the Fe content arising from dissolution of
nontronite [80]. The process of Fe (III) loading significantly in-
creases the intensity of the FeKa and FeKb fluorescence lines that
only slightly reduces after incubation in AA (Fig. 4). Those data are
consistent with ICP-OES and XPS results.
Detailed scrutiny of the diffraction patterns and corresponding
Rietveld plots of Fe (III) loaded samples does not provide any clue
about the occurrence of detectable amounts of nano-Fe (III) oxides
and oxy-hydroxides. Rietveld refinements indicate a compression
of the c-parameter of the iron-loaded samples as compared to the
pristine one, which is observed in AA treated samples as well (see
below) (Table 5), whereas the variation of the a-parameter is
Fig. 4. Intensity of the FeK fluorescence lines of the pristine and 1000 mM FeCl3 treated
samples as collected at the multichannel analyser of the solid-state detector. Data were
rescaled on the basis of the common CuKa fluorescence line. The absence of the CuKb
fluorescence line is due to its removal by the incident beam multilayer (Go€bel) mirrors.
limited and irregular (Table 5). Analysis of the microstructural pa-
rameters indicates a moderate reduction of the 0 microstrain [81]
occurring after leaching. Modifications of the framework are almost
irrelevant. Table 6 reports the Si, Al partition at T1 and T2 sites
following Jones' determinative curve [82]. The populations of the
T1 and T2 sites of the pristine sample were Al3.52Si20.48 and
Al4.08Si7.92, respectively, consistent with present and reference
chemical data. As can be seen, albeit the T1-O and T2-O bond
distances of the leached and ascorbic acid treated samples are
remarkably constant, they are possibly shorter (1.6224(16) Å
instead of 1.627(2) Å) and longer (1.659(3) Å instead of 1.654(2) Å),
respectively, than the corresponding ones of the pristine sample
(Table 6). Numbers bracketed for the treated samples reflect the
actual data dispersion of the average values. The reported variation
of the T-O bond distances, which could be apparently interpreted
as arising from a Si Al migration occurring within the frame-
work, is in effect related to the occurrence of minor protonation at
some oxygen sites. A detailed scrutiny of individual T-O bond dis-
tances possibly points out to a prevailing segregation at neigh-
bouring O1 and O3 sites (T1-O1: pristine 1.607(3) Å;
treated 1.598(5) Å; T2-O1: pristine 1.673(2) Å;
treated 1.684(5) Å; T1-O3: pristine 1.628(1) Å;
treated 1.618(3) Å), that become Brønsted sites, permitting the
occurrence of the expected local intra-site hopping mechanism
[83,84] (Fig. 5). A rough estimate of the extension of protonation
may be attempted by comparing the total EF cation release in the
case of both Fe (II) and Fe (III) loading processes even considering
the slightly different chemistry of the corresponding pristine
samples. The two values are comparable (Fe
(II) 551(11) nmol mg 1; Fe (III) 465(11) nmol mg 1) and in the
case of Fe (II) leaching corresponds to the removal of ca. 1.5 apfu Na
from the structure i.e. 1.5 e . Therefore, we may hypothesize that
approximately the same number of apfu of H are required to
compensate the removed monovalent Na cations. This value cor-
responds to a protonation of ca. 2% of the oxygen atoms of the
framework or ca. 4% of those residing exclusively at the O1 O3
sites. The refined s.s. of the EF cation and water molecules sites of
the pristine sample agrees with reference structure refinements of
fibres of erionite-Na from the same locality (Table S3). It has to be
noticed that the total calculated s.s. of the EF cation sites is signif-
icantly higher than that arising from SEM-EDX analysis. This is due
to the alkali-metal migration occurring during EDX analyses. The
leaching process induces an EF cations mobilization, which results
in a progressive migration from both Ca1 and Ca2 sites toward Ca3.
Such process does not involve relevant modification of the total s.s.
of the EF cation sites and is coupled to rearrangements of the s.s. of
the bound water molecules, that occur without any significant
displacement and modification of the bonding network. In partic-
ular, there is an increase of s.s. at OW8 and OW9 and a decrease at
OW10. This is a pattern that qualitatively follows that observed
during the Fe (II) loading process [39] suggesting that Fe (III) is
similarly segregated at Ca3. This hypothesis is supported by the
smaller s.s. at Ca3 of the Fe (III) loaded sample with respect to the Fe
(II) loaded one, which positively correlates with a lower Fe popu-
lation devised for the 250 mM FeCl3 incubated sample (0.17 with
respect to 0.90 apfu). The coordination of Fe (III) at Ca3 is identical
to that of Fe (II) as it is linked to 3 OW7 at 2.20(4) Å and to
3 OW10 at 2.19(4) Å (Fig. 6). Combined ICP, SEM-EDX and XPS
data support the absence of reduction or possibly the occurrence of
a very minor reduction of Fe (III) segregated at Ca3. Structural data
provide further support to this hypothesis as no significant modi-
fication of the s.s. at Ca3, nor changes of its bonding neighbour, are
observed in AA treated samples as compared to those Fe (III)
loaded. However, it is worth noting that iron contributes with less
than 20% of the s.s. at Ca3.
 A. Pacella et al. / Microporous and Mesoporous Materials 237 (2017) 168e179 177

Table 5
Comparison of cell parameters and volume, 0 microstrain, and crystallite size of pristine, blank, Fe (III) loaded and AA reduced samples.

Sample a (Å) c (Å) c/a Volume (Å3) 0 Crystallite size (nm)

Pristine 13.23097(9) 15.06475(11) 1.1386 2283.89(3) 0.0618(15) 141(4)

Blank 250 mM FeCl3 13.23040(9) 15.05988(11) 1.1383 2282.96(3) 0.0617(15) 140(4)
250 mM FeCl3 13.22347(9) 15.05209(11) 1.1383 2279.39(4) 0.0538(16) 127(3)
250 mM FeCl3 1 h AA 13.22746(9) 15.04956(11) 1.1378 2280.38(3) 0.0565(15) 137(4)
250 mM FeCl3 24 h AA 13.22267(9) 15.04897(11) 1.1381 2278.64(4) 0.0540(16) 149(5)
Blank 1000 mM FeCl3 13.22872(9) 15.05219(11) 1.1378 2281.22(4) 0.0546(16) 140(4)
1000 mM FeCl3 13.22438(9) 15.05059(11) 1.1381 2279.48(3) 0.0560(15) 162(5)
1000 mM FeCl3 1 h AA 13.22952(9) 15.05212(11) 1.1378 2281.48(4) 0.0561(17) 132(4)
1000 mM FeCl3 24 h AA 13.22681(9) 15.05201(11) 1.1380 2280.53(4) 0.0573(16) 168(6)
1000 mM FeCl2a 13.22136(10) 15.05383(12) 1.1386 2278.92(4) e e
a
Refers to data taken, for comparison purposes, from reference [39].

Table 6
Si, Al partition at T1 and T2 sites and calculation of the R ratio following Jones [82].

T1-O T2-O T1 T2 T R

Pristine 1.6265 1.6539 Si20.48Al3.52 Si8.13Al3.87 Si28.61Al7.39 0.795

Blank 250 mM FeCl3 1.6246 1.6555 Si20.78Al3.22 Si8.01Al3.99 Si28.79Al7.21 0.800
250 mM FeCl3 1.6247 1.6568 Si20.77Al3.23 Si7.91Al4.09 Si28.68Al7.32 0.797
250 mM FeCl3 1 h AA 1.6226 1.6570 Si21.08Al2.92 Si7.90Al4.10 Si28.98Al7.02 0.805
250 mM FeCl3 24 h AA 1.6222 1.6563 Si21.15Al2.85 Si7.95Al4.05 Si29.10Al6.90 0.808
Blank 1000 mM FeCl3 1.6220 1.6587 Si21.17Al2.83 Si7.76Al4.24 Si28.93Al7.07 0.804
1000 mM FeCl3 1.6204 1.6627 Si21.43Al2.57 Si7.45Al4.55 Si28.88Al7.12 0.802
1000 mM FeCl3 1 h AA 1.6217 1.6634 Si21.22Al2.78 Si7.41Al4.59 Si28.63Al7.37 0.795
1000 mM FeCl3 24 h AA 1.6208 1.6610 Si21.36Al2.64 Si7.59Al4.41 Si28.95Al7.05 0.804
Average treated 1.6224(16) 1.659(3) Si21.12(24)Al2.88 Si7.75(24)Al4.25 Si28.87(16)Al7.13 0.802(4)

We may hypothesize that the EF cation rearrangement is
responsible for the marked contraction of the c-parameter and the
0 microstrain release. The absence of any relevant reduction of the
total EF cation s.s. sum implies that the H EF cation protonation
process, invoked for explaining their release to the incubation so-
lutions [69,73], should be confined to the external shell of fibres.
5. Conclusions

In this work we investigated the chemical, structural and surface

Fig. 5. ORTEP-3 [88] plot of the possible scheme of protonation.
Fig. 6. ORTEP-3 [88] plot of the six-fold coordination of Fe (III) within the erionite
structure after fibre incubation in the 250 mM FeCl3 solution.
modifications of fibrous erionite-K from Rome after incubation in
FeCl3 solutions. ICP-OES results, supported by XPS investigation,
unambiguously show that Fe (III) is mainly fixed at the fibre surface,
in perfect agreement with previous results of Eborn and Aust [34].
However, a fraction of Fe (III), which is ca. 2/3 in the case of very
diluted Fe (III) solutions and rapidly decreasing as concentration
rises, is segregated by an ion-exchange mechanism in the erionite
cavity at the Ca3 site, similarly to Fe (II) [39]. The maximum amount
of apfu Fe (III) residing at Ca3 is of 0.17(7) and is observed in the case
of the sample incubated in the 250 mM FeCl3 solution. This is a
significantly smaller iron content as compared to that ion-
exchanged under the form of Fe (II). Therefore, Fe (II) and Fe (III)
are both fixed at the same crystallographic site, albeit with a
significantly different efficiency with respect to the Fe available in
solution. In addition, our results suggest that AA is able to reduce
fairly easily only Fe (III) residing at the fibres surface and charac-
terized by low nuclearity, whereas this reaction does not occur (or
possibly occurs only very marginally) in the case of the ion-
exchanged metal. Bronchoalveolar lavage of normal humans [85]
reported total iron content as small as
0.21(4) mmol L 1(12(2) mg L 1) strongly suggesting that under
physiological conditions iron tends to be prevalently ion-
exchanged. This is a very important piece of information as it, un-
fortunately, represents the optimum condition for iron to behave as
a very active site in the generation of ROS. In fact, it has been shown
that in Fe loaded zeolites at rather low Fe content, the most active
sites are those located in well-defined crystallographic positions
and possessing a very low nuclearity (single ions, dimers or oligo-
mers) [86,87].
178 A. Pacella et al. / Microporous and Mesoporous Materials 237 (2017) 168e179

Under such aspect, the role of Fe-bearing, relatively unstable S. Quartieri, R. Rinaldi, M. Ross, R.A. Sheppard, E. Tillmanns, G. Vezzalini, Can.
Mineral 35 (1997) 1571e1606.
minerals under the appropriate physical-chemical conditions,
[23] J.A. Gard, J.M. Tait, in: Proceedings of the 3rd Intemational Conference on
which are mechanically driven within the lungs by the fibres, Molecular Sieves, University Press, Leuven, 1973, pp. 94e99.
should be investigated in detail. In particular, nontronite is a rela- [24] A. Kawahara, H. Curien, B. Soc. Fr. Mine ral. Cr. 92 (1969) 250e256.
tively widespread Fe-bearing clay mineral that has been found to [25] A. Alberti, A. Martucci, E. Galli, G. Vezzalini, Zeolites 19 (1997) 349e352.
[26] A. Gualtieri, G. Artioli, E. Passaglia, S. Bigi, A. Viani, J.C. Hanson, Am. Mineral. 83
engulf erionite fibres, modifying both their morphology and as- (1998) 590e606.
sembly [32]. Present results indicate a relatively small rate of [27] P. Ballirano, G.B. Andreozzi, M. Dogan, A.U. Dogan, Am. Mineral. 94 (2009)
dissolution of nontronite, as testified by the small amount of 1262e1270.
[28] P. Ballirano, G. Cametti, Microporous Mesoporous Mat. 163 (2012) 160e168.
released Fe (III) by the pristine sample. However, the incubation of [29] P.E. Eberly, Am. Mineral. 49 (1964) 30e40.
erionite fibres in a very diluted Fe (III) solution could result in a [30] A.U. Dogan, Y.I. Baris, M. Dogan, S. Emri, I. Steele, A.G. Elmishad, M. Carbone,
maximization of the segregation process of ferric iron at Ca3, which Cancer Res. 66 (2006) 5063e5068.
[31] A. Croce, M. Allegrina, C. Rinaudo, G. Gaudino, H. Yang, M. Carbone, Microsc.
is the prerequisite for behaving as a very active site in the gener- Microanal. 21 (2015) 1341e1347.
ation of ROS as a result of the catalytic properties of this zeolite. [32] R. Matassa, G. Famigliari, M. Relucenti, E. Battaglione, C. Downing, A. Pacella,
Further studies are in progress on the erionite samples from Rome, G. Cametti, P. Ballirano, Sci. Rep. 5 (2015) 16757.
[33] B. Fubini, L. Mollo, Toxicol. Lett. 82e83 (1995) 951e960.
concerning the biodurability and chemical reactivity tests. The [34] S.K. Eborn, A.E. Aust, Arch. Biochem. Biophys. 316 (1995) 507e514.
combined approach consisting in the full characterization of the [35] A. Carr, B. Frei, FASEB J. 13 (1999) 1007e1023.
physical-chemical features of the fibres coupled with the reactivity [36] E. Fach, W.J. Waldman, M. Williams, J. Long, R.K. Meister, P.K. Dutta, Environ.
Health Perspect. 110 (2002) 1087e1096.
studies will represent a fundamental contribution to the identifi-
[37] E. Fach, R. Kristovich, J. Long, W.J. Waldman, P.K. Dutta, M. Williams, Environ.
cation at the atomic level of the source of erionite toxicity. Int. 29 (2003) 451e458.
[38] J.F. Long, P.K. Dutta, B.D. Hogg, Environ. Health Perspect. 105 (1997) 706e711.
[39] P. Ballirano, A. Pacella, C. Cremisini, E. Nardi, M. Fantauzzi, D. Atzei, A. Rossi,
Acknowledgments G. Cametti, Microporous Mesoporous Mat. 211 (2015) 49e63.
[40] P. Ballirano, A. Pacella, Sci. Rep. 6 (2016) 22786.
[41] P. Ballirano, G. Cametti, Am. Mineral. 100 (2015) 1003e1012.
Erionite fibres from Rome, Oregon, were kindly provided by
[42] S.J. Gregg, K.S.W. Sing, Adsorption, Surface Area and Porosity, second ed.,
Prof. M. Gunter (University of Idaho, Moscow, Idaho, USA). Dr. Ida Academic Press, Inc, London, 1982, pp. 42e49.
Pettiti is acknowledged for BET measurements. Research was fun- [43] H.S. Sherry, Clays Clay Miner. 27 (1979) 231e237.
[44] E.P. Barrett, L.G. Joyner, P.P. Halenda, J. Am, Ceram. Soc. 73 (1951) 373e380.
ded by Sapienza University of Rome (P.R. Ateneo 2014, Prof. Bal-
[45] B.J. Lippens, J.H. de Boer, J. Catal. 4 (1965) 319.
lirano) and PRID 2014 (University of Cagliari). [46] W.D. Harkins, G. Jura, J. Am. Ceram. Soc. 66 (1944) 1366e1373.
[47] L. Gurvitsch, J. Phys. Chem. Soc. Russ. 47 (1915) 805e827.
[48] E. Passaglia, Am. Mineral. 55 (1970) 1278e1301.
Appendix A. Supplementary data [49] G. Cametti, A. Pacella, F. Mura, M. Rossi, P. Ballirano, Am. Mineral. 98 (2013)
2155e2163.
[50] M.P. Seah, Surf. Interface Anal. 31 (2001) 721e723.
Supplementary data related to this article can be found at http://
[51] N. Fairley, CasaXPS Version 2.3.15, Casa Software Ltd., Wilmslow, Cheshire,
dx.doi.org/10.1016/j.micromeso.2016.09.021. UK, 1999-2003.
[52] D.A. Shirley, Phys. Rev. B 5 (1972) 4709e4714.
[53] M.P. Seah, in: D. Briggs, J.T. Grant (Eds.), Surface Analysis by Auger and X-ray
References Photoelectron Spectroscopy, IM Publication Surface Science Spectra, Man-
chester, 2003.
[1] IARC, IARC Monographs on the evaluation of the carcinogenic risk to humans, [54] M. Fantauzzi, A. Pacella, G.B. Andreozzi, J. Fournier, A. Gianfagna, A. Rossi, 63,
in: Silica Some Silicates, vol. 42, 1987, pp. 225e239. Anal. Bioanal. Chem. 404 (2012) 821e833.
[2] IARC, IARC Monographs on the evaluation of the carcinogenic risk to humans, [55] A.X.S. Bruker, Topas V.4.2: General Profile and Structure Analysis Software for
in: Arsenic, Metals, Fibres Dusts, vol. 100 C, 2011, pp. 311e316. Powder Diffraction Data, Bruker AXS, Karlsruhe, Germany, 2009.
[3] D.L. Coffin, P.M. Cook, J.P. Creason, Inhal. Toxicol. 4 (1992) 273e300. [56] O.V. Yakubovich, W. Massa, P.G. Gavrilenko, I.V. Pekov, Crystallogr. Rep. 50
[4] R.A. Sheppard, A.I. Gude III, U.S. Geological Survey Professional Paper, 634, (2005) 544e553.
1973, p. 56. [57] Y. Le Page, G. Donnay, Acta Crystallogr. B32 (1976) 2456e2459.
[5] K.E. Bargar, T.E.C. Keith, in: D.W. Ming, F.S. Mumpton (Eds.), Natural Zeolites [58] R.A. Eggleton, Clay Miner. 12 (1977) 181e194.
‘93. International Committee on Natural Zeolites, Brockport, New York, 1995, [59] T.M. Sabine, B.A. Hunter, W.R. Sabine, C.J. Ball, J. Appl. Crystallogr. 31 (1998)
pp. 69e86. 47e51.
[6] R.W. Tschernich, Zeolites of the World, Geoscience Press, Phoenix, Arizona, [60] P. Ballirano, J. Appl. Crystallogr. 36 (2003) 1056e1061.
USA, 1992, p. 563. [61] G. Vignaroli, P. Ballirano, G. Belardi, F. Rossetti, Environ. Earth Sci. 72 (2014)
[7] M. Mattioli, M. Cenni, E. Passaglia, Period. Mineral. 85 (2016) 1e24. 3679e3698.
[8] Y.I. Bariş, A.A. Sahin, M. Ozesmi, I. Kerse, E. Ozen, B. Kolacan, M. Altino €rs, [62] A. Arcifa, A. Rossi, R.M. Espinosa-Marzal, N.D. Spencer, J. Phys. Chem. C.
A. Go€ktepeli, Thorax 33 (1978) 181e192. Nanomater. Interfaces 118 (50) (2014) 29389e29400.
[9] P. Dumortier, L. Coplü, I. Broucke, S. Emri, T. Selcuk, V. De Maertelaer, P. De [63] A. Penna, M. Careri, N.D. Spencer, A. Rossi, J. Am. Soc. Mass Spectrom. 26
Vuyst, I. Baris, Occup. Environ. Med. 58 (2001) 261e266. (2015) 1311e1319.
[10] M. Carbone, I. Baris, P. Bertino, B. Brass, S. Corertpay, A. Dogan, G. Gaudino, [64] K.N. Dalby, H.W. Nesbitt, V.P. Zakaznova-Herzog, P.L. King, Geochim. Cos-
S. Jube, S. Kanodia, C. Partridge, H. Pass, Z. Rivera, I. Steele, M. Tuncer, S. Way, mochim. Acta 71 (2007) 4297e4313.
H. Yang, A. Miller, Proc. Natl. Acad. Sci. U. S. A. 108 (33) (2011) 13618e13623. [65] M. Fantauzzi, A. Pacella, D. Atzei, A. Gianfagna, G.B. Andreozzi, A. Rossi, Anal.
[11] R.A. Sheppard, U.S. Geological Survey Open-file Report 96-018, 1996, p. 24. Bioanal. Chem. 396 (2010) 2889e2898.
[12] E.B. Ilgren, M. Ortega Brena, J. Castro Larragoitia, G. Loustaunau Navarrete, [66] J.F. Moulder, W.F. Stickle, P.E. Sobol, K.D. Bomben, in: J. Chastain (Ed.),
A. Fuentes Bren ~ a, E. Krauss, G. Fehe
r, Indoor Built Environ. 17 (2008) 496e515. Handbook of X-ray Photoelectron Spectroscopy, Perkin Elmer Corporation,
[13] C.R. Kliment, K. Clemens, T.D. Oury, Int. J. Clin. Exp. Pathol. 2 (2009) 407e410. 1992.
[14] P.H. Ryan, M. Dihle, S. Griffin, C. Partridge, T.J. Hilbert, R. Taylor, S. Adjei, [67] B. Herreros, H. He, T.L. Barr, J. Klinowski, J. Phys. Chem. 98 (1994) 1302e1305.
J.E. Lockey, J. Occup. Environ. Med. 53 (2011) 892e898. [68] G.S. Pokrovski, J. Schott, F. Farges, J.L. Hazemann, Geochim. Cosmochim. Acta
[15] B.S. Van Gosen, T.A. Blitz, G.S. Plumlee, G.P. Meeker, M.P. Pierson, Environ. 67 (2003) 3559e3573.
Geochem. Health 35 (2013) 419e430. [69] V.K. Ragnarsdottir, Geochim. Cosmochim. Acta 57 (1993) 2439e2449.
[16] E.B. Ilgren, H. Kazemian, J.A. Hoskins, EBPH 12, 2015 e10106-1-e10106-12. [70] R.L. Hartman, H.S. Fogler, Langmuir 23 (2007) 5477e5484.
[17] G. Gottardi, E. Galli, Natural Zeolites, Springer-Verlag, Heidelberg, 1985, p. [71] R.M. Carland, F.F. Aplan, Perspectives in Molecular sieve science, in:
409. W.H. Flank, T.E. Whyte Jr. (Eds.), ACS Symposium Series, vol. 368, 1988, pp.
[18] L.W. Staples, J.A. Gard, Mineral. Mag. 32 (1959) 261e281. 292e305 (Chapter 18).
[19] J.V. Smith, F. Rinaldi, L.S. Dent Glasser, Acta Crystallogr. 16 (1963) 45e53. [72] N.F. Chelishchev, V.F. Volodin, Dokl. Akad. Nauk. SSSR 237 (1977) 122e125.
[20] B. Rüdinger, E. Tillmanns, G. Hentschel, Mineral. Pet. 48 (1993) 147e152. [73] J. Schott, R.A. Berner, E.L. Sjo €berg, Geochim. Cosmochim. Acta 45 (1981)
[21] P. Ballirano, S. Merlino, E. Bonaccorsi, A. Maras, Can. Mineral 34 (1996) 2123e2135.
1021e1030. [74] S. Music, A. Ve rtes, G.W. Simmons, I. Czako-Nagy, H. Leidheiser Jr., J. Coll,
[22] D.S. Coombs, A. Alberti, T. Armbruster, G. Artioli, C. Colella, E. Galli, J.D. Grice, Interface Sci. 85 (1982) 256e266.
F. Liebau, J.A. Mandarino, H. Minato, E.H. Nickel, E. Passaglia, D.R. Peacor, [75] T.A. Ruda, P.K. Dutta, Environ. Sci. Technol. 39 (2005) 6147e6152.
 A. Pacella et al. / Microporous and Mesoporous Materials 237 (2017) 168e179
[76] F. Frau, D. Addari, D. Atzei, R. Biddau, R. Cidu, A. Rossi, Water Air Soil Poll. 205
(2010) 25e41.
[77] T. Misawa, K. Hashimoto, S. Shimodaira, Corros. Sci. 14 (1974) 131e149.
[78] A. Pacella, P. Ballirano, G. Cametti, Eur. J. Mineral. 28 (2016) 257e264.
[79] G.B. Morgan, D. London, Am. Mineral. 90 (2005) 1131e1138.
[80] S.R. Gainey, E.M. Hausrath, J.A. Hurowitz, R.E. Milliken, Geochim. Cosmochim.
Acta 126 (1) (2014) 92e211.
[81] P. Ballirano, C. Sadun, Struct. Chem. 20 (2009) 815e823.
[82] J.B. Jones, Acta Crystallogr. B24 (1968) 355e358.
179
[83] J.A. Ryder, A.K. Chakraborty, A.T. Bell, J. Phys. Chem. B 104 (2000) 6998e7011.
[84] J.A. Van Bokhoven, C. Lamberti, Coord. Chem. Rev. 277e278 (2014) 275e290.
[85] J.M.C. Gutteridge, S. Mumby, G.J. Quinlan, K.F. Chung, T.W. Evans, Biochem.
Biophys. Res. Commun. 220 (1996) 1024e1027.
[86] M. Schwidder, M.S. Kumar, K. Klementiev, M.M. Pohl, A. Bruckner, W. Grunert,
J. Catal. 231 (2005) 314e330.
[87] A. Zecchina, M. Rivallan, G. Berlier, C. Lamberti, D. Ricchiardi, Phys. Chem.
Chem. Phys. 9 (2007) 3483e3499.
[88] L.J. Farrugia, ORTEP-3 for Windows, University of Glasgow, Scotland, 1999.