Silver nanoparticles stabilized by a polyaminocyclodextrin as catalysts for the reduction of nitroaromatic compounds

Article abstract of DOI:10.1016/j.molcata.2015.07.031

Silver nanoparticles stabilized by means of poly-(6-N,N-dimethyl-propylenediamino)-(6-deoxy)-β-cyclodextrin were synthesized, characterized by different techniques (UV-vis spectroscopy, Dynamic Light Scattering, High Resolution Transmission Electron Microscopy, Fourier-transform IR Spectroscopy) and used as catalysts for the reduction of various nitrobenzene derivatives with sodium borohydride. The nanocomposites obtained appear to have an organized structure, with a metal core surrounded by a layer-structured coating shell. Kinetic data, rationalized in terms of a modified Langmuir-Hinshelwood model, evidenced a non-linear dependence of the reaction rate on the concentration of the catalyst. This was explained on the grounds of the catalytic activity of differently covered catalyst areas. Careful analysis of kinetic data, in particular the effect of the para substituent on the nitroarene structure and the trends of the induction period observed at the beginning of the reaction, provided with interesting insights on the reaction course, and brought us to critically reconsider several mechanistic ideas reported in previous literature.

Full text of DOI:10.1016/j.molcata.2015.07.031

Accepted Manuscript
Title: Silver nanoparticles stabilized by a
polyaminocyclodextrin as catalysts for the reduction of
nitroaromatic compounds
Author: Marco Russo Francesco Armetta Serena Riela Delia
Chillura Martino Paolo Lo Meo Renato Noto
PII:
S1381-1169(15)30042-X
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.molcata.2015.07.031
MOLCAA 9582
To appear in:
Journal of Molecular Catalysis A: Chemical
Received date:
Revised date:
Accepted date:
18-5-2015
16-7-2015
27-7-2015
Please cite this article as: Marco Russo, Francesco Armetta, Serena Riela,
Delia Chillura Martino, Paolo Lo Meo, Renato Noto, Silver nanoparticles
stabilized by a polyaminocyclodextrin as catalysts for the reduction of
nitroaromatic compounds, Journal of Molecular Catalysis A: Chemical
http://dx.doi.org/10.1016/j.molcata.2015.07.031
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

1
Silver nanoparticles stabilized by a polyaminocyclodextrin as catalysts for the reduction of
nitroaromatic compounds
Marco Russo, Francesco Armetta, Serena Riela, Delia Chillura Martino, Paolo Lo Meo,* Renato
Noto
Università degli Studi di Palermo, Dip. STEBICEF - V.le delle Scienze, Parco d’Orleans, Ed. 17.
90128 – Palermo (Italy)
e-mail: paolo.lomeo@unipa.it;
Graphical abstract
Highlights





Ag-polyaminocyclodextrin nanocomposites show a layer-organized structure
These systems were used as catalysts for nitroarene reduction with NaBH4
The reduction reaction showed peculiar kinetic features
Kinetics were rationalized on the by a modified Langmuir-Hinshelwood model
Results brought us to reconsider some literature mechanistic ideas
Abstract
Silver nanoparticles stabilized by means of poly-(6-N,N-dimethyl-propylenediamino)-(6-deoxy)--
cyclodextrin were synthesized, characterized by different techniques (UV-Vis spectroscopy,
Dynamic Light Scattering, High Resolution Transmission Electron Microscopy, Fourier-transform IR

2
Spectroscopy) and used as catalysts for the reduction of various nitrobenzene derivatives with
sodium borohydride. The nanocomposites obtained appear to have an organized structure, with a
metal core surrounded by a layer-structured coating shell. Kinetic data, rationalized in terms of a
modified Langmuir-Hinshelwood model, evidenced a non-linear dependence of the reaction rate
on the concentration of the catalyst. This was explained on the grounds of the catalytic activity of
differently covered catalyst areas. Careful analysis of kinetic data, in particular the effect of the
para substituent on the nitroarene structure and the trends of the induction period observed at
the beginning of the reaction, provided with interesting insights on the reaction course, and
brought us to critically reconsider several mechanistic ideas reported in previous literature.
Keywords
polyaminocyclodextrin
nitroarene reduction
silver nanocomposite
1. Introduction
The reduction of aromatic nitrocompounds is an important synthetic route leading, under the
proper reaction conditions, to different nitrogen-containing derivatives [1]. In particular, the
reduction to anilines (Scheme 1), which has also a great interest as an industrial process, is
generally carried out with strong reducing agents or under harsh conditions. The most common
protocols provide the use of LiAlH4, transition metals in strongly acidic media (Fe and conc. HCl is a
typical industrial method), or catalytic hydrogenation.
Unfortunately, these conditions may be unsuitable whenever particularly delicate functional
groups are present in the molecule. Therefore, the availability of milder alternative methods has
an undoubted appeal. In this context, the use of NaBH4 in the presence of noble metal
nanoparticles (NPs) has attracted a certain interest [2,3]. The reduction of 4-nitrophenol or 4-
nitroaniline (to 4-aminophenol and 1,4-diaminobenzene respectively) has been successfully
carried out using NPs stabilized by various capping agents, including biopolymers such as dextran
[4,5] or alginate [6-8], and PAMAM- or PPI-type dendrimers [3,9,10]. The latter class of auxiliaries
is particularly interesting, because dendrimers provide a strictly defined micro-environment for
the formation of the NP and for the actual nitroarene reduction process as well.

3
Cyclodextrins (CDs), both native or chemically modified with suitable donor groups, constitute a
class of alternative capping agents for the formation of metal NPs which have been occasionally
employed [11,12]. The most known example is probably constituted by the heptakis-(6-thiol)-(6-
deoxy)-CD [13], which has been successfully used for the synthesis of Au or Ag NPs. Recently,
some of us synthesized a new class of polyamino-cyclodextrin products [14], which have been
tested as effective auxiliary agents for the preparation of Ag nanoparticles. The CD derivatives
were obtained by a simple nucleophilic displacement reaction between the heptakis-(6-iodo)-(6-
deoxy)-CD and various polyamines (Scheme 2).
However, due to the unavoidable occurrence of poly-substitution, the reaction affords mixtures of
products bearing a different number of polyamino pendants [15]. As a matter of fact, the same
polyamine unit may substitute more than one iodine on the same CD scaffold. The obtained
materials, which are isolated as partial hydroiodides, have been fully characterized by means of
Nuclear Magnetic Resonance, Electrospray Ionization Mass Spectrometry and potentiometric
titration. The combined use of these techniques allowed to identify the different constituents of
the mixtures, to determine the average number of polyamine pendants (n) and formal HI
molecules (x) per CD unit, and to calculate the possible protonation or charge status as a function
of the pH.
Despite the nitroarene reduction catalyzed by metal NPs has been known for relatively long time,
the actual mechanism of catalysis has not been satisfactorily understood, and even the kinetic
features of the process are still debated. There is no general agreement about the kinetic order of
the reaction, which is usually reported as first order in the nitroarene [3,6,10,12,16-18], but which
has been also occasionally described as zeroth order [2,8]. Moreover, the possibility to apply this
reaction to further nitroarene derivatives has never been explored in detail. Therefore, a
systematic investigation and in general a critical reconsideration of the entire topic seemed
interesting. In this context, we reasoned that the use of a cyclodextrin-capped Ag-NP system as
the catalyst could be particularly intriguing. Indeed, CDs as auxiliary agents may provide both a
suitable micro-environment for the formation and the stabilization of the NP, and a sort of
receptor site on the catalyst surface, constituted by the CD cavity, able to bind and direct the
nitroarene substrate.

4
In the present work we investigated the reduction reaction of various nitroarene derivatives 1-8
(Figure 1) with NaBH4 in the presence of two Ag-NP catalyst systems, obtained by using the poly-
(6-N,N-dimethyl-propylenediamino)-(6-deoxy)-CD (AmCD, Figure 2) as the capping agent.
The structural and morphological characterization of the Ag-NP systems chosen as catalysts was
performed by means of combined UV-Vis spectroscopy, Dynamic Light Scattering (DLS), High
Resolution Transmission Electron Microscopy (HR-TEM) and FT-IR spectroscopy. These techniques
allowed to get information on the size and polidispersity of the NP suspensions, as well as to
investigate the morphology of the NPs. The chosen nitroarene substrates differ for the substituent
at the para position with respect to the reactant nitro group. They were suitably selected to
present significant variations in their molecular properties [19,20]. In particular, we considered the
electrondonating effect, the possible charge status and the steric hindrance of the para
substituent, as well as the intrinsic affinity of the substrate for the CD cavity.
2. Results and discussion
2.1. Synthesis of the NP catalysts
According to previous work [14], AmCD-capped Ag-NPs can be obtained by reacting in the dark a
solution of the [Ag(NH3)2]+ complex in the presence of a suitable amount of the auxiliary AmCD,
with an excess of formaldehyde at 40 °C for 90 minutes. Noticeably, on the grounds of the
analytical data reported elsewhere [14], AmCD bears on average 5.7 N,N-dimethyl-
propylenediamino pendants and 4.0 formal HI molecules per CD unit. The formation of the NPs is
evident by the appearance of the typical reddish color, due to the characteristic surface plasmon
resonance (SPR) absorption band, usually centered between 400 and 450 nm. In a series of
preliminary experiments we explored the effect on the NP formation process from changing the
concentrations of the reactants and their molar ratio, in particular: i) by fixing the [Ag(NH3)2]+
concentration (1.0 mM) and varying the AmCD (0.410.0 mN; the normality of the AmCD solution
is defined according to the average number of donor sites, namely 11.4, per CD unit); ii) by fixing

5
the AmCD (2 mN) and varying the [Ag(NH3)2]+ (0.25.0 mM); iii) by simultaneously varying both the
reactants, while keeping constant their equivalent ratio (2:1 eq/eq, i.e. the same as in the starting
[Ag(NH3)2]+ complex). The obtained systems were preliminarily characterized by UV-vis
spectroscopy; the results obtained are summarized in Table 1.
The outcome of the NP formation process strictly depends on the AmCD/[Ag(NH3)2]+ equivalent
ratio used. Nearly no reaction occurred with ratios larger than 4:1. In the other cases the systems
showed an absorption band centered at 401±2 nm, the intensity of which (expressed by the
relevant molar extinction coefficient, ) regularly increases on decreasing the equivalent ratio.
Moreover, the same ratio heavily affects the stability of the system. Samples obtained with a 2:1
ratio are stable for several weeks, if kept in the dark at room temperature. By contrast, systems
with a 0.4:1 ratio or prepared with the largest [Ag(NH3)2]+ concentrations, appear cloudy or turbid
and form reddish-brown precipitates within few days (lower ratios were also explored, indeed, but
resulted in the quick separation of part of the reduced Ag as the typical “silver mirror”). On the
whole, these preliminary findings suggest that the auxiliary AmCD protects the silver ion towards
its reduction under the conditions used. Data indicate that NP formation is hampered at the
highest AmCD/[Ag(NH3)2]+ ratios, is partial for the ideal 2:1 value, and becomes almost complete
at the lowest ratios. Of course, the presence of a minimum amount of AmCD is needed in order to
ensure the actual formation and stabilization of the NP system. Therefore, it seems reasonable to
hypothesize that NPs prepared with a 2:1 ratio are constituted by a metal core, surrounded by a
shell of AmCD units complexing a certain amount of Ag+ ions by means of their the amine donor
groups. The presence of a significant positive charge on the NP surface ensures the time stability
of the system. On the other hand, in systems prepared at low (namely 0.4:1) ratios the NPs formed
bear nearly no charge on their surface, and consequently may aggregate or coalesce.
In order to confirm these hypotheses, we thoroughly investigated two particular representative
NP systems, which were subsequently used as the catalysts for studying the nitroarene reduction
reaction. These systems, indicated hereinafter as Catalysts I and II, were prepared at a 1 mM
[Ag(NH3)2]+ concentration, and AmCD concentrations as large as 0.4 mN (Catalyst I) and 2 mN
(Catalyst II), which correspond to AmCD/Ag+ mole/mole ratios (as large as 0.035 and 0.175
respectively. Moreover, as it will be illustrated more in detail in section 2.3, in order to perform
the nitroarene reduction reaction Catalysts I and II are subjected to dilution (up to 1:750 v/v., i.e.

6
1.33 M) and subsequent interaction with NaBH4. This may cause deep modifications in the
structure of the nanocomposite, as it can be evidenced by simple observations. After dilution 1:1
or 1:9 v/v of the preformed Catalyst II with water, a significant enlargement and decrease in
intensity of the SPR band centered at 401 nm, and the appearance of a new band at 567 nm can
be noticed in the normalized UV-vis spectra (Figure 3). The corresponding color change is apparent
even at naked eye. The observed behavior might account in principle for either an increase in NP
average size, or concomitant changes in the polar character and/or the local refractive index of the
coordination shell immediately surrounding the metal core [21]. Furthermore, the residual Ag+
possibly present at the NP surface (in particular in the case of Catalyst II) can be easily reduced to
Ag0 by borohydride ion. As a matter of fact, addition of NaBH4 to a solution containing [Ag(NH3)2]+
and AmCD (at the
same concentrations as in Catalyst II) instantaneously causes the formation of a turbid brown
suspension, which precipitates down in few minutes. For these reasons, together with “as-
prepared” Catalysts I and II, we subjected to characterization three further control systems,
namely two dilutions 1:1 and 1:9 v/v of Catalyst II with water (indicated as IIdilA and IIdilB
respectively), and a mixture 1:1 v/v between Catalyst II and a freshly prepared NaBH4 48.0 mM
solution (indicated as IIred).
2.2. Structural and morphological investigation of the NP catalysts
The knowledge of NPs size and morphology is of outmost importance in order to rationalize their
activity as catalysts [22]. These information are typically achieved by means of direct methods
such as transmission electron microscopy (TEM), as well as by indirect scattering-based methods
such as Dynamic Light Scattering (DLS). The latter one affords an evaluation of particle size in
suspension, allowing to analyze a more statistically representative sample at the cost of lack in
high resolution details about the morphology of nanoparticles. For our purposes, it is worth
stressing here that TEM and DLS provide with complementary information about the structural
organization of our systems: the former focuses on the metal core alone, whereas the latter looks
at the entire nanocomposite, including its coating and hydration shell.
As representative examples, TEM micrographs for Catalysts I and for control systems IIdilB and
IIred are shown in Figure 4 (further images are provided in the Supporting). The pristine Catalysts I
and II show well defined, quite spherical, individual metal cores with a relatively low polydispersity
and scarce tendency to aggregation. Statistical analysis of TEM images indicates unimodal

7
diameters distributions centered at ca. 27 and 19 nm respectively. Therefore, larger NPs are
formed in the presence of a lesser amount of capping agent. Dilution does not seem to have
significant effects on the dimensions of the pristine metal cores; however, a fair tendency to
aggregation can be noticed, together with the appearance of a new population of small faint
objects having an average diameter of ca. 3.6 nm. Considerable modifications are apparent after
treatment with NaBH4. TEM images show a large tendency to aggregation, together with
modification of both the size and the morphology of the pristine metal cores and, most
interestingly, the formation of new small sized cores. Statistical analysis of images reveals a
completely different diameter distribution, with a maximum frequency centered at ca. 9.7 nm.
Different information are provided by DLS determinations. Data for “as-prepared” systems
indicate the presence of smaller aggregates for Catalyst I than for Catalyst II, with average
diameters as large as 86 and 149 nm respectively. Sizes were evaluated by applying the CONTIN
method (see Experimental for details; representative DLS curves are reported in Supporting), and
again show unimodal distributions. For the control systems IIdilA and IIdilB average diameters of
149 and 59 nm respectively are found, indicating that significant dilution (namely 1:9 v/v) causes a
noticeable decrease of the aggregate size. By contrast, once Catalyst II is treated with NaBH4, for
the system IIred the aggregates in suspension grow dramatically in size, with an average value of
the diameter of 464 nm, and show high polydispersity. The latter result is in agreement with TEM
evidence of NPs aggregation.
It is interesting to notice that, on the grounds of TEM results, metal cores for Catalysts I and II may
be estimated to possess on average 6·105 and 2·105 silver atoms respectively, and average surface
areas as large as 2300 and 1100 nm2 respectively. On the other hand, simple molecular models
suggest that a single AmCD unit may cover an area on the NP surface, which can be roughly
estimated as comprised between 3 and 7 nm2 [23]. Keeping into account the actual composition of
the systems, in particular the values of the aforementioned AmCD/Ag+ mole ratios  (section 2.1),
it can be calculated that only a minor fraction (less than 5%) of the AmCD present might actually
come in direct contact with, and fully cover, the metal core. It must be also considered that the
height of a single CD torus is reported to be 0.79 nm [24]; therefore, a single AmCD layer cannot
be much thicker than 1 nm. Thus, everything considered, the whole of the TEM and DLS results
can be explained as follows (Scheme 3).

8
Pristine NP cores, formed after the slow formaldehyde reduction of silver ions embedded in a sort
of Ag+·AmCD macroaggregate, are initially surrounded by a thick layer-structured coating,
constituted by shells of AmCD units held together by various supramolecular interactions,
including the mutual inclusion of the polyamine tails into the CD cavities and the complexation of
residual Ag+ ions with the polyamines. This coating offers both a steric and an electrostatic barrier
against the aggregation/coalescence of the NPs and, of course, is thicker for Catalyst II, prepared
with the largest amount of AmCD. Dilution of the catalyst causes the progressive dissolution of the
coating shells (as accounted for by the decreased size estimated by DLS), with formation of small
Ag+·AmCD microaggregates (the faint objects observed in IIdilB), which act as seeds for the
formation of new small Ag cores after treatment with NaBH4. The latter process may cause the
deposition of further metal on the pristine cores, modifying their size and shape (as observed in
IIred). The concomitant exhaustive reduction of Ag+ removes at the same time the electrostatic
barrier, causing extensive aggregation.
Further confirmation for the composite nature of our AgNP systems was achieved by means of FT-
IR evidence. By subjecting a large amount of Catalyst I to centrifugation, a sample of Ag NPs was
isolated and its IR spectrum (KBr) was recorded. Noticeably, owing to the strong intrinsic
absorption due the nanosized Ag component, the sample disc results fairly opaque to IR radiation.
After suitable correction, however, the resulting spectrum shows very intriguing features (Figure
5).
Comparison with the spectrum of free AmCD, reveals the presence in the composite of the signals
expected for the organic capping agent. More in detail, we observe that the pristine large AmCD
band at 3310 cm-1, due to superimposition (because of extensive hydrogen bonding) of the -OH
and -NH- groups stretching, splits in two bands centered at 3470 and 3136 cm-1 respectively. This
indicates a neat discrimination between the two different groups, reasonably owing to the
occurrence of the N-Ag interactions at the NP surface. The AmCD fingerprints at 1153, 1086, 1041
(C-O stretching) and 943 cm-1 (-OH bending) attributable to the CD scaffold, are present in the
spectrum of the composite with minor shifts, together with new bands at 998 and 839 cm-1. The
latter one, in particular, replaces the corresponding AmCD band centered at 756 cm-1 which can
be attributed to N-H wagging. Therefore, this large shift confirms the occurrence of strong N-Ag

9
interactions at the primary CD rim. Finally, The spectrum of the composite shows the trivial C-H
stretching bands (30002800 cm-1), and two strong bands at 1456 and 1385 cm-1 due to carbonate
ion (superimposed to the expected C-H bending bands of the AmCD at 1466 and 1368 cm-1), the
presence of which can be easily explained as due to CO2 uptake by the quite alkaline NP solution.
2.3. Kinetic study of the nitroarene reduction
In order to study the kinetics of the reduction reaction of nitroarenes 1-8 mediated by Ag-NPs,
sample systems were prepared by mixing the proper amounts of stock solutions of the reactants
and the catalyst (see Experimental), in such a way to have initial concentrations of the substrate
and NaBH4 as large as 0.1 mM and 4.0 mM respectively, and a catalyst concentration ranging
between 1.33 and 83.3 M. The latter concentration is expressed as the equivalent concentration
of silver ion (CAg) needed to obtain the amount of NPs introduced in the sample. The reducing
agent, which is the last component introduced in the system immediately before starting the
registration of the kinetic trace, is always used in large excess with respect to the nitroarene
substrate. Kinetic experiments were performed at 313 K, following by UV-vis spectrophotometry
the disappearance of the absorption band of the substrate, at its max value. A typical kinetic trace
is shown in Figure 6.
In agreement with literature [2,17,25,26,27], a preliminary induction period is present.
Subsequent data points monotonically decrease, till a plateau is almost abruptly reached. In
striking contrast with previous literature, the decreasing branch of the kinetic trace cannot be
satisfactorily fitted neither as the exponential decay expected for a first order process, nor as a
simple straight line accounting for a zeroth-order process. Rather, experimental points appear
tentatively adaptable at a first approximation by a parabolic trend. In a preliminary set of control
experiments performed with 4-nitroaniline 1, we fixed the amount of catalyst (50 M) and suitably
varied the borohydride concentration in the range 1.08.0 mM. We verified that the initial
reaction rate, as expressed by the first order coefficient (c1) of the fitting parabolic curve, linearly
increases with the reducing agent concentration within the range considered (Figure 7, see data in
Supporting). According to previous reports [5,23,28], a curvilinear dependence of the initial rate
vs. borohydride concentration would have been expected at much larger concentrations (> 10
mM).

10
The kinetics of NPs catalyzed reactions has been frequently approached by means of the
Langmuir-Hinshelwood model [7,18,25,29-31], providing the reversible association of the
reactants on different sites of the catalyst surface. This, in particular, explains the curvilinear
dependence of the initial reaction rate [5,28], which has been rationalized admitting a fast
-
association pre-equilibrium between the catalyst and the BH4 ion, followed by the transfer of a
hydride ion onto the NP surface [7,25,32]. However, we preferred to develop a modified version of
the model. In particular, we considered the Ag-NP catalyst surface divided in effective area cells a,
each as large as the average surface covered by a single AmCD unit. Then, we hypothesized that
the reaction product is formed in a rate-limiting step from a surface-substrate-borohydride ternary
complex, deriving through a set of fast pre-equilibria (Scheme 4).
It can be assumed that a cells have an actual concentration (Ca) proportional to CAg according to
the relationship Ca = ·CAg , where  is the ratio between the total surface for the NPs generated
by one mole of silver ion, and the average surface which can be covered by one mole of AmCD
units. From Scheme 4 the following differential kinetic expression can be written:
d | ArNO2 |
kKBRCa | BH4 || ArNO2 |
1 KB | BH4 | KR | ArNO2 | KBR | BH4 || ArNO2 |
v  
 k | a  BH4  ArNO2 |
dt
(1)
-
where KB, KR and KBR are the stability constants for the binary (a·ArNO2) and (a·BH4 ) complexes
-
and the overall stability constant for the ternary (a·BH4 ·ArNO2) complex respectively; k is the
kinetic constant for the rate determining step. Equation (1) can be simplified by considering that
-
the observed linear dependence of the reaction rate on |BH4 | implies the algebraic condition:
-
-
1+KR|ArNO2| >> KB|BH4 |+KBR|BH4 ||ArNO2| , then:
kKBRCAg | BH4 || ArNO2 |
kKBRCa | BH4 || ArNO2 |
1 KR | ArNO2 |
d | ArNO2 |



dt
1 KR | ArNO2 |
(2)
which is equivalent to the expression expected on the grounds of the original Langmuir-
Hinshelwood model. It has to be pointed out [31] that an expression such as equation (2) is never
approached as such, because it can be reduced to either a first order or a zeroth order expression,

11
by imposing the conditions 1>>KR|ArNO2| or 1<<KR|ArNO2| respectively. Nevertheless, it can be
integrated as:
t = ti + (1/kobs)(C0-|ArNO2|) + (1/kobsKR) ln(C0/|ArNO2|)
(3)
where ti is the length of the induction period, C0 the starting concentration of the substrate and
-
-
kobs = k(KBR/KR)CAg|BH4 |. The latter condition may be also written as kobs = kKB*CAg|BH4 |, where
KB* is the equilibrium constant for the association between the preformed (a·ArNO2) binary
-
complex and BH4 . Evidence for the occurrence of a strong association between nitroarenes and
Ag-NPs, by Raman spectroscopy, has been already reported in literature [16,33]. The integrated
equation (3) comprises both a zeroth-order (linear) and a first-order (logarithmic) term in the
substrate. So, it can be suitably used to accomplish an excellent fitting the experimental kinetic
traces, as shown in Figure 6 [34].
The expression for kobs predicts a linear dependence of the reaction rate on both the
concentrations of the reducing agent and the catalyst as well [18, 35]. Surprisingly, however,
experimental results showed for kobs a non-linear dependence on CAg (Figure 8, the complete set of
data is collected in the Supporting). In some cases, namely with Catalyst II and anionic substrates 2
(sodium N-4-nitrophenyl-glicinate), 3 (sodium N-methyl-N-4-nitrophenyl-glicinate) and 5 (sodium
4-nitrophenate), we even found non-monotonical trends. No reaction could be observed for
substrate 7 (sodium 4-nitrobenzoate) with both catalysts, and for 8 (nitrobenzene) with Catalyst II.
Keeping into account the characterization of the catalysts discussed in Section 2.2, the observed
dependence of kobs on CAg can be explained considering that, under the high dilution condition
used, the catalyst should have lost most of its coating. Thus, we can hypothesize that the effective
surface cells a may be present in the reaction system in three different conditions, namely naked,
covered with one AmCD unit only, or covered with two (or even more) AmCD layers. Of course,
differently covered a cells are each other in equilibrium and have different catalytic effectiveness.
In particular, we can assume that doubly-covered (or even multiply-covered) cells should be
catalytically inactive, because the reactants are likely to interact with the outer AmCD layer, so
they cannot approach the catalytically active NP surface. This has been verified in a set of control
experiments performed with 4-nitroaniline 1 and Catalyst I (50 M). As a matter of fact, by adding

12
increasing amounts of AmCD, we noticed a progressive decrease in the reaction rate, until
substantial inhibition was achieved for AmCD concentrations larger than 1.0 mM. Therefore, we
considered a more complex mechanistic scheme (Scheme 5), keeping also into account the
possible host-guest binding equilibrium between the free AmCD and the nitroarene substrate. A
similar inverse dependence of the reaction rate on the amount of capping agent has been
reported for the reduction of nitrophenol with Au NPs stabilized with natural extracts of Breyhnia
rhamnoides [16].
In principle, the reaction rate should depend on the concentrations of the possible active catalyst-
substrate-borohydride ternary complexes, according to the expression:
-
-
v = k(0)|a·ArNO2·BH4 | + k(1)|(a·AmCD)·ArNO2·BH4 | =
-
-
= k(0) KR(0)KB*(0)|a||ArNO2||BH4 | + k(1)K(1)KR(1)KB*(1)|a||AmCD||ArNO2||BH4 |
(4)
From Scheme 5, considering the mass balance on the surface cells and the equilibrium conditions,
and applying few suitable simplifications, we can obtain the following kinetic equation:
d | ArNO2 |
v  
 k(0) | a BH4  ArNO2 | k(1) | (a AmCD) BH4  ArNO2 |
dt
[(k0KR(0)KB*(0)  k(1) K1KR(1)KB*(1) | AmCD |) CAg | BH4 |] | ArNO2 |
(5)

(1 K1 | AmCD | K1K2 | AmCD |2 )  (KR(0)  K1KR(1) | AmCD| K1K2KR(2) | AmCD |2 ) | ArNO2 |
In equation (5) the concentration of free AmCD is accounted. The total amount of AmCD in the
system (CCD) is bound to CAg by the relationship: CCD = ·CAg, where  is the AmCD/[Ag(NH3)2]+
mole ratio defined above (Section 2.1). However, at a first approximation it may be assumed
|AmCD| ·CAg , i.e. the concentrations of the various complexes may be considered negligible
with respect to free AmCD [36]. Then, the following final kinetic expression can be obtained:
(k0 KR(0) KB*(0)  k(1) K1KR(1) KB*(1) CAg )CAg | BH4 || ArNO2 |
d | ArNO2 |
v  

(1 K1CAg  K1K2 2CA2g )  (KR(0)  K1KR(1) CAg  K1K2 KR(2) 2CA2g ) | ArNO2 |
dt
(6)
Despite its complicated appearance, equation (6) assumes to the same form as equation (2), and
can be integrated in the same way to afford the final expression:
t = ti + (1/kobs)(C0-|ArNO2|) + (1/kobsKapp) ln(C0/|ArNO2|)
(7)

13
where:
[k(0) KB*(0) ]  [k1K1 (KR(1) / KR(0) )KB*(1) ]CAg
1 K1 (K R(1) / K R(0) )CAg  K1K2 (K R(2) / KR(0) ) 2CA2g
kobs  CAg | BH 4 |
(8)
(9)
and
KR(0)  K1KR(1) CAg  K1K2 KR(2)  2CA2g
Kapp

1 K1CAg  K1K2  2CA2g
Equation (8) can be also written as:
kobs,0CAg  kobs,1Kapp,1CA2g
| BH4 | 1 Kapp,1CAg  Kapp,1Kapp,2  2CA2g
kobs

(10)
where kobs,0= k0KB*(0) , kobs,1= k1KB*(1) , Kapp,1 = K1(KR(1)/KR(0)) and Kapp,2 = K2(KR(2)/KR(1)). Noticeably,
kobs,0 and kobs,1 provide a quantitative estimation of the catalytic efficiency of the naked and the
singly-covered NP surface respectively. The catalytic efficiency depends, in turn, on the
morphological characteristics of the NPs (accounted for by ), on the ability of the NP-substrate
pre-complex to bind the reducing agent (accounted for by KB*(0) or KB*(1)) and on the intrinsic
reactivity of the substrate onto the particular catalyst cell (accounted for by k0 or k1); Kapp,1 and
Kapp,2 represent the conditional binding constants between the NP surface and the AmCD layers in
the presence of the different substrates.
Equation (10) can adequately describe the non-monotonical kobs vs. CAg trends observed with
Catalyst II and anionic substrates. However, as long as data relevant to Catalyst I are considered,
the use of equation (10) leads to unreliable Kapp,2 values, characterized by unacceptably large
2
indeterminations. From an algebraic viewpoint, being the term Kapp,1Kapp,2CAg negligible with
respect to 1+Kapp,1CAg , a simplified form of equation (10) can be more suitably used:
kobs,0CAg  kobs,1Kapp,1CA2g
kobs
| BH4 |

1 Kapp,1CAg
(11)

14
The results obtained in analyzing the apparent kinetic constants kobs for substrates 1-8 with either
equation (10) or (11) are summarized in Table 2.
We can notice that kobs,0 values spun over a much larger range for Catalyst II than for Catalyst I,
indicating larger variations in reactivity between the various substrates, and therefore a larger
selectivity for the catalyst. Moreover, larger Kapp,1 values are generally observed with Catalyst I
than with Catalyst II. These differences among the two catalysts may be ascribed to the different
amount of unreduced Ag+ ion present on the catalyst, which is likely to interact with the
-
nitroarene substrate before the addition of borohydride. Then, after BH4 is introduced to start the
title reaction, it also reduces the residual Ag+. Thus, the consequent modifications in the
morphology of the catalyst may be affected by the possible occurrence of significant Ag+-substrate
interactions. This seems particularly evident in the case of anionic substrates, for which non-
monotonical trends in kobs values vs. CAg are observed.
Analysis of data reported in Table 2 indicates that the presence of a strong electrondonating
substituent at the para position with respect to the reacting nitro group, is an important
requirement for the reaction to occur. As far as kobs,0 values are concerned, nitroanilines 1-4 and
nitrophenate 5 appear significantly more reactive than nitroanisole 6 and nitrobenzene 8, whereas
nitrobenzoate 7 is not reactive at all. This result might appear quite surprising, because the
reduction reaction should imply either an electron transfer process [16,37] or the nucleophilic
attack of a hydride species to the nitro group. The latter type of process is consistent with the
-
hypothesis reported in literature that BH4 transfers a hydride ion onto the NP surface [7,25,32]. In
both cases the reaction should be unfavorably affected by electrondonating substituents. By
contrast, the behavior observed indicates that the mechanistic course of the process provides the
interaction of the nitroarene with an electrophilic species, which must be identified with the
catalyst surface. Nitroanilines 1-4 give further insights on the role of the para substituent group.
N-substituted derivatives 2-4 do not appear more reactive than parent nitroaniline 1, despite alkyl
substituents on the N atom should enhance the overall electrondonating character of the group.
Therefore, the presence of a negative charge on the ancillary chain bound to the N atom
(substrates 2 and 3), or the occurrence of significant steric hindrances (substrate 4), play an
unfavorable role. This probably reflects a more difficult approach of the anionic reducing agent to
the nitroarene-catalyst pre-complex, and a lower stability of the reactive ternary complex. The

15
reactivity of nitrophenate 5 with Catalyst II appears exceptionally high. This is probably due to
peculiar modifications in the NP surface morphology, occurring as a consequence of particular Ag+-
substrate interactions, as discussed previously.
On passing to analyze kobs,1 values, we can immediately notice that the singly-covered surface cells
appear by far less catalytically active than uncovered ones. More in detail, kobs,0/kobs,1 ratios range
from 4.6 (substrate 4 with Catalyst II) up to 97 (substrate 5 with Catalyst I). A lesser activity for
singly-covered cells is not a trivial finding, because it indicates that inclusion into the receptor-like
cavity does not lead the nitroarene to approach more easily the catalytic surface. This, however, is
in contrast with the fact that the organic molecule is correctly included with the nitro group
directed towards the primary CD rim [19,20,38]. By considering data for Catalyst I, the lowest kobs,1
values are observed for substrates 2 (sodium N-(4-nitrophenyl)-glicinate) and 5 (sodium 4-
nitrophenate). The former one is known to form a very stiff inclusion complex with native CD
[20], due to the occurrence of multiple hydrogen bonding with the secondary host rim. The same
effect likely blocks this substrate quite rigidly within the AmCD cavity. By contrast, other
substrates unable to undergo multiple hydrogen bonding form more flexible complexes [20], and
consequently are able to slip down more easily onto the NP surface after inclusion. It is interesting
to notice that even the parent nitrobenzene 8 appears relatively reactive onto the singly-covered
catalyst. Finally, the scarce reactivity of 5 may be likely attributed to the occurrence of a strong
hydrogen bond interaction of the anionic centre with the secondary host rim, which forces its
average position far from the NP surface.
A few considerations are due to the apparent binding constants Kapp present in equations (7) and
(9) and to the conditional binding constants Kapp,1 and Kapp,2 present in equations (8), (10) and (11).
Kapp reflects the actual contribution from the first order term to the overall reaction kinetics. This
contribution increases in importance as Kapp decreases; consequently, larger deviations are
observed in the kinetic trace with respect to the ideal linear trend expected for a purely zeroth
order process. Unlike a true binding constant, Kapp depends on the concentration of the catalyst
according to equation (9). Observed values usually range between 4·104 up to 4·105 M-1, and
frequently show bell-shaped trends as a function of CAg, with few exceptions. However, these
values are affected by very large indeterminations (up to 40 %), so that a detailed analysis would
be a pointless exercise. On passing to Kapp,1 and Kapp,2, their values depend on both on the intrinsic
affinity of the AmCD towards the naked or singly covered NP surface (irrespective of the substrate,
as accounted for by K1 and K2 respectively), and on the different abilities of the naked, singly- or

16
doubly-covered cells to bind the nitroarene (as accounted for by the ratios (KR1/KR0) and (KR2/KR1)).
Therefore, large differences in Kapp,1 or Kapp,2 among different substrates have to reflect similar
differences in the latter ratios. For instance, nitrobenzene 8 shows with Catalyst I a much smaller
Kapp,1 value as compared to the other substrates. This clearly mirrors its much lesser affinity for the
host cavity, consequent to the smaller value of its binding constant with native CD [39] as
compared with nitroanilines. The same argumentation can be applied to nitroanisole 6 with
Catalyst II. However, it should be reminded that Kapp,1 values for Catalyst II are influenced by the
deep modifications of the catalyst upon introduction of the reducing agent, as discussed
previously. The outcome of the possible substrate-Ag+ interactions on such modifications is not
easily predictable. Regarding Kapp,2, its estimation was possible only in few cases, but values
appear quite lower as compared to Kapp,1. It is worth stressing that Kapp,2 accounts for the
interaction between superimposed layers of AmCD, which probably involves the inclusion of the
diamine pendants of the outer layer into the cavities of the inner one. Therefore, data suggest that
the latter interaction is significantly less effective.
Finally, regarding induction times ti, for each substrate-catalyst combination examined we
observed that values increase on decreasing the concentration of the catalyst (data for substrate 1
with Catalyst I are shown in Figure 9 as a representative example).
Induction times vary with both the substrate and the catalyst. Nitroanilines 1-4 show smaller
values than nitrophenate 5 and nitroanisole 6, whereas nitrobenzene shows the largest values. At
the largest NP concentrations, slightly shorter induction times are observed with catalyst I, with
the exception of nitrophenate 5. The occurrence of an induction period has been justified in
different ways. It was formerly admitted that the catalytic process can start only after a tiny layer
–
of silver oxide possibly present on the NP surface had been reduced by BH4 [2]. It has been
alternatively proposed that induction times (viewed as the inverse of a sort of kinetic constant
[25]) might be related to the diffusion of the reacting species onto the catalyst surface [17], or to
the reconstruction of the NP surface [25]. Our results, however, seem to rule out these
interpretations, because they cannot explain neither the observed dependence on the
concentration of the catalyst and on the structure of the substrate as well, nor the fact that
induction occurs also with Catalyst II, having a significant amount of Ag+ immediately reduced in
situ by borohydride. On the other hand, it is well known that induction is typical for mechanisms
involving free-radical species. Therefore, it might be tentatively interpreted as an indication that

17
the nitroarene substrate undergoes single electron transfer processes leading to short-lived radical
species, which are scavenged by the molecular O2 present in solution [40]. There is literature
evidence that induction times for 4-nitrophenol reduction are suppressed by exclusion of O2 from
the reaction system [27]. As a consequence, the target reaction cannot properly start unless all the
O2 has been consumed by reaction with the excess borohydride. The latter process will be
catalyzed by NPs and, as a consequence, its duration will depend on the concentration of the
catalyst. On the other hand, it can be expected that the NP-substrate interaction will favor this
process, with an increasing effect on increasing the electron-donating character of the substrate
itself.
3. Conclusions
Polyaminocyclodextrin-capped Ag NPs were synthesized and characterized by combined UV-vis,
DLS and TEM techniques. Our systems appear constituted by a metal core surrounded by a layered
coating shell, made by the AmCD units held together by supramolecular interactions such as tail-
cavity mutual inclusion or complexation of the residual unreduced Ag+ ions. This shell offers both a
steric and an electrostatic barrier to aggregation, but it is progressively dissolved upon dilution on
the system. Then, possible treatment with NaBH4 causes both a modification of the size and shape
of the pristine metal cores and the formation of new small cores.
Two Ag-AmCD NP systems were used as catalysts for the reduction of nitroarene derivatives with
sodium borohydride. The reaction shows a peculiar kinetic trend and an unexpected dependence
of the apparent kinetic constants kobs on the concentration of the catalyst. On the grounds of a
modified Langmuir-Hinselwood model, experimental data were rationalized by means of a
complex mechanistic scheme, providing the possibility that the catalyst surface may be present in
the reaction system partly naked, and partly covered with one or more AmCD layers. Data show
that the reaction course is favorably affected by the presence of electron-donating groups on the
substrate, indicating that the NP surface acts as an electrophile towards the nitro group.
Nevertheless, the presence of negatively charged or bulky groups on the substrate has a slightly
unfavorable outcome, probably due to a more difficult approach of the reducing agent. The
AmCD-covered catalyst surface seems less active, because of the intrinsic stability of the
substrate-AmCD inclusion complex. In the case of Catalyst II, which contains a significant amount

18
-
of unreduced Ag+ ion, its morphology changes as BH4 is introduced into the reaction system. The
latter process may be significantly affected by the presence of the substrate. Finally, the trends for
the induction period seem to provide evidence for the occurrence of free-radical pathways in the
intimate mechanism of reduction of the nitro group.
4. Experimental
4.1. Materials
All commercial (Aldrich, Fluka) reactants and materials were used as purchased, with no further
purification. The nitroaniline derivatives 2-4 were prepared according to literature [19,20], by a
nucleophilic aromatic displacement reaction between 4-nitrofluorobenzene and the proper amine.
The poly-(6-N,N-dimethyl-propylenediamino)-(6-deoxy)-CD AmCD [14] was synthesized by a
nucleophilic displacement reaction between the heptakis-(6-iodo)-(6-deoxy)-CD and a 20-fold
excess of 3-(N,N-dimethylamino)-1-aminopropane (which functions as the solvent too), at 60 °C
for 48 h. Its high-resolution ESI-MS spectrum gave the following signals: 990.4720 (C77H154N14O28
2HI·2H)++,
(C72H140N12O28·H2O·2H)++, 811.5024 (C72H140N12O28·2H)++, 778.4575 (C67H126N10O28·2H2O· 2H)++,
769.4513 760.4447 582.7033
(C67H126N10O28·H2O·2H)++, (C67H126N10O28·2H)++,
873.5507 (C77H154N14O28HNa)++, 862.5601 (C77H154N14O282H)++, 820.5090
(C77H154N14O282H·Na)+++, 575.3765 (C77H154N14O283H)+++, 547.3417 (C72H140N12O28·H2O ·3H)+++
,
1
541.3379 (C72H140N12O28·3H)+++. All other physical and spectroscopic data (m.p., IR, H NMR, 13C
NMR) were consistent with literature.
4.2. Instrumentation
ESI-MS mass spectra were recorded in positive mode on an AGILENT Technologies 6540 UHD
Accurate Mass Q-TOF LC-MS apparatus (1 kV nozzle voltage, 250 V fragmentor voltage). UV-vis
spectra were recorded on a Beckmann DU 800 spectrophotometer, equipped with a Peltier
thermostatic apparatus, which was used also to perform the kinetic experiments. FT-IR spectra
(2% in KBr) were acquired on a Bruker VERTEX 70 apparatus.
TEM micrographs were acquired using a JEM-2100 (JEOL, Japan) electron microscope operating at
200 kV accelerating voltage. A drop of each suspension was put onto a 3 mm Cu grid “lacey
carbon” for analysis and let the solvent to complete evaporation.

19
DLS measurements have been carried out by using a 90 Plus Particle Size Analyzer (Brookhaven
Instrument Corp., Holtsville, NY, USA) equipped with a BI-9000AT Digital Autocorrelator. The
source was a He-Ne 632.8 nm laser and the detector was a photomultiplier tube, both mounted
on a turntable; measurements were performed at 90°. By using the software supplied by
Brookhaven Institute, data were obtained directly as electric field self-correlation function, G(τ),
versus decay time, τ. In order to check data reproducibility and the stability of the suspension,
measurements have been repeated in triplicate on different freshly prepared samples. Electric
field self-correlation functions have been analyzed to obtain diffusion coefficients, D. The
scattering particle diameters, d, have been evaluated by means of the Stokes-Einstein equation: D
=KBT/3πηd once provided the absolute temperature T, and the viscosity of the solution η. Owing
to the low concentration of the systems, the latter can be assumed coincident with the one of the
pure solvent, within the limits of experimental errors. The self-correlation function computed by
using the Cumulant analysis [41], did not satisfactorily reproduce the experimental data, thus
suggesting that the particles in suspension were polydisperse. Therefore, a more appropriate
analysis was performed by means of the CONTIN method [42].
4.3. Synthesis of Ag-NPs
Ag-NP systems were prepared with a slight modification of reported procedure [14]: mother
solutions of [Ag(NH3)2]+ 15 mM and AmCD 6 mN were prepared in double-distilled water,
previously degassed with a fine stream of Ar for 15 minutes. Then, samples were prepared by
mixing in a dark vial the proper amounts of water, [Ag(NH3)2]+ and AmCD solutions up to 14 mL.
The samples, kept under Ar atmosphere, were placed in an oil bath thermostated at 40 °C,
inoculated with 1 mL of formaldehyde solution 33.4 mM, and allowed to react for 90 minutes. UV-
vis spectra of the samples were recorded immediately after the reaction time. Catalysts I and II
were carefully kept and stored in the dark before use. In order to collect a sample for FT-IR, 100
mL of Catalyst I were subjected to centrifugation at 5000 rpm for 15 minutes, and the resulting
solid residue was isolated and lyophilized (yield 14.3 mg).
4.4. Kinetics
Kinetic experiments were performed at 313 K, according to the following procedure. In a quartz
cuvette the proper amount of a stock solution of catalyst (1.0 mM) was diluted in the proper
amount of freshly double-distilled water (degassed with a fine stream of Ar for 15 minutes), in

20
such a way to have 2.65 mL of solution; then, 100 L of a stock solution 3.0 mM of the substrate
were added, and the mixture was thermostated. A 48.0 mM solution of NaBH4 was prepared just
before use by dissolving a weighed amount in a volumetric flask; then 250 L were rapidly
introduced in the cuvette, and the registration of the kinetic trace immediately started, following
by UV-vis spectrophotometry the disappearance of the nitroarene substrate.
Acknowledgments
TEM micrographs were acquired at the Centro Grandi Apparecchiature - UniNetLab - Università di
Palermo funded by P.O.R. Sicilia 2000-2006, Misura 3.15 Quota Regionale; Prof. Eugenio Caponetti
and Dr. G. Nasillo are gratefully acknowledged for collaboration and useful discussion.
References and notes
[1] G.W. Kabalka, R.S. Varma, “Reduction of Nitro and Nitroso compounds” in Comprehensive
Organic Synthesis, Vol. 8, pag. 363-379, B.M. Trost and I. Fleming, Eds., Pergamon Press, Oxford
(UK), 1991.
[2] N. Pradhan, A. Pal, T. Pal, Colloids Surf. A 196 (2002) 247-257.
[3] a) K. Esumi, K. Miyamoto, T. Yoshimura, J. Colloid Interf. Sci. 254 (2002) 402-405. b) K. Esumi, R.
Isono, T. Yoshimura, Langmuir 20 (2004) 237-243. c) T. Endo, T, Yoshimura, K. Esumi, J. Colloid
Interf. Sci. 286 (2005) 602-609.
[4] a) K.P. Bankura, D. Maity, M.M.R. Mollick, D. Mondal, B. Bhowmick, M.K. Bain, A. Chakraborty,
J. Sarkar, K. Acharya, D. Chattopadhyay, Carbohydr. Polym. 89 (2012) 1159–1165. b) G. Yang, Q.
Lin, C. Wang, J. Li, J. Zhou, Y. Wang, C. Wang, J. Nanosci. Nanotechnol. 12 (2012) 3766-3774.
[5] c) R. Eising, A.M. Signori, S. Fort, J.B. Domingos, Langmuir 27 (2011) 11860-11866.
[6] L. Ai, J. Jang, Biores. Technol. 132 (2013) 374–377.
[7] S. Pyne, S. Samanta, A. Misra, Solid State Sci. 26 (2013) 1-8.
[8] S. Saha, A. Pal, S. Kundu, S. Basu, T. Pal, Langmuir 26 (2010) 2885-2893.

21
[9] a) H. Ye, R.W.J. Scott, R.M. Crooks, Langmuir, 20 (2004) 2915-2920. b) J. Zheng, M. S.
Stevenson, R.S. Hikida, P.G. Van Patten, J. Phys. Chem. B 106 (2002) 1252-1255.
[10] K. Hayakawa, T. Yoshimura, K. Esumi, Langmuir 19 (2003) 5517-5521.
[11] a) S. Pande, S.K. Ghosh, S. Praharaj, S. Painigrahi, S. Basu, S. Jana, A. Pal, T. Tsukuda, T. Pal, J.
Phys. Chem. C 111 (2007) 10806-10813. b) Y. Huang, D. Li, J. Li, Chem. Phys. Lett. 389 (2004) 14-18.
c) P.R. Gopalan, Int. J. Nanosci. 9 (2010) 487-494. d) P.F. Andrade, A. Fonseca de Faria, D. Soares
da Silva, J. A. Bonacin, M.d.C. Gonçavales, Colloid. Surf. B 118 (2014) 289-297.
[12] B.D. Loganathan, A.B. Mandal, RSC Adv. 3 (2013) 5238-5253.
[13] a) J. Liu, S. Mendoza, E. Román, M.J. Lynn, R. Xu, A.E. Kaifer, J. Am. Chem. Soc. 121 (1999)
4304-4305. b) J. Liu, W. Ong, E. Román, M.J. Lynn, A.E. Kaifer, Langmuir 16 (2000) 3000-3002. c) J.
Liu, W. Ong, A.E. Kaifer, C. Peinador, Langmuir 18 (2002) 5981-5983. d) Y. Guo, Y. Zhao, H. Wu, M.
Fan, Y. Wei, S. Shuang, C. Dong, J. Incl. Phenom. Macrocycl. Chem. 78 (2014) 275-286. d) X. Li, Z.
Qi, K. Liang, X. Bai, J. Xu, J. Liu, J. Shen, Catal. Lett, 124 (2008) 413-417.
[14] P. Lo Meo, F. D’Anna, M. Gruttadauria, S. Riela, R. Noto, Carbohyd. Res. 347 (2012) 32-39.
[15] It is worth stressing here that the synthesis of AmCD from native CD requires two steps only,
and therefore is much more straightforward and cheaper than the syntheses reported in literature
of other polyaminocyclodextrin derivatives, obtained as pure individual products and mainly
employed for polynucleotide complexation and gene transfection. See for instance: a) C. Ortiz
Mellet, J.M. Benito, J.M. García Fernández, Chem. Eur. J. 16 (2010) 6728-6742. b) A. Méndez-
Ardoi, M. Gómez-García, C. Ortiz Mellet, N. Sevillano, M.D. Girón, R. Salto, F. Santoyo-Gonzáles,
J.M. García Fernández, Org. Biomol. Chem. 7 (2009) 2681-2684. c) S. Srinivasachari, K.M. Fichter,
T.M. Reineke, J. Am. Chem. Soc. 130 (2008) 4618-4627. d) J. Deng, N. Li, K. Mai, C. Yang, L. Yan, L.-
M. Zhang, J. Mater. Chem. 21 (2011) 5273-5281.
[16] A. Gangula, R. Podila, M. Ramakrishna, L. Karanam, C. Janardhana, A.M. Rao, Langmuir 27
(2011) 15268-15274.
[17] J. Zeng, Q. Zhang, J. Chen, Y. Xia, Nano Lett. 10 (2010) 30-35.
[18] N. Bingwa, R. Meijboom, J. Mol. Catal. A 396 (2015) 1-7.
[19] a) P. Lo Meo, F. D’Anna, S. Riela, M. Gruttadauria, R. Noto, Org. Biomol. Chem. 1 (2003) 1584-
1590. b) P. Lo Meo, F. D’Anna, S. Riela, M. Gruttadauria, R. Noto, Tetrahedron 63 (2007) 9163-
9171.
[20] P. Lo Meo, F. D’Anna, M. Gruttadauria, S. Riela, R. Noto, Tetrahedron 60 (2004) 9099-9111.

22
[21] K.G. Thomas, “Surface Plasmon Resonances in Nanostructured Materials” in Nanomaterials
Chemistry, pag. 185-218, C.N.R. Rao, A. Müller and A.K. Cheetham Eds., Wiley-VCH, Weinheim
(Germany), 2007. It is worth stressing that polarity or refractive index variations may in turn be
the outcome of a ligand exchange at the NP surface. For instance, it has been recently shown that
the position of the SPR is significantly shifted in presence of halide ions. See: a) B.-H. Kim, J.-S. Lee,
Mater. Chem. Phys. 149-150 (2015) 678-685. b) X.C. Jiang, A.B. Yu, Langmuir 24 (2008) 4300-4309.
[22] a) Y. Zhu, M. Shen, Y. Xia, M. Lu, Catal. Commun. (2015) 64, 37–43. b) F. Valentini, D. Roscioli,
M. Carbone, V. Conte, B. Floris, E.M. Bauer, E. Caponetti, D. Chillura-Martino, Sens. Actuators B
212 (2015) 248–255. c) M.L. Saladino, D. Chillura Martino, E. Kraleva, E. Caponetti Catal. Commun.
36 (2013) 10–15. d) E. Caponetti, L. Pedone, M.L. Saladino, D. Chillura Martino, G. Nasillo,
Microporous Mesoporous Mater. 128 (2010) 101-107.
[23] This estimation is based on the fact that the CD torus alone is reported to have an average
diameter of 1.54 nm (see ref. [20]), and that the length of a stretched diamine pendant chain can
be estimated (MM2) as ca. 0.8 nm.
[24] J. Szejtli, Chem. Rev. 98 (1998) 1743-1753.
[25] S. Wunder, F. Polzer, Y. Lu, Y. Mei, M. Ballauff, J. Phys. Chem. C 114 (2010) 8814-8820.
[26] Y. Gao, X. Ding, Z. Zheng, X. Cheng, Y. Peng, Chem. Comm. (2007) 3720-3722.
[27] M. Blosi, S. Albionetti, S. Ortelli, A.L. Costa, M. Ortolani, L. Dondi, New J. Chem. 38 (2014)
1401-1409.
[28] A.M. Signori, K. de O. Santos, R. Eising, B.L. Albuquerque, F. C. Giacomelli, J.B. Domingos,
Langmuir 26(2010) 17772-17779.
[29] K. Vasenth Kumar, K. Porkodi, F. Roche, Catal. Comm. 9 (2008) 82-84.
[30] a) J.S. Zhang, W.N. Delgass, T.S. Fisher, J.P. Gore, J. Power Sources 164 (2007) 772-781. b) F.
Lin, R. Doong, Appl. Catal. A 486 (2014) 32-41.
[31] J. W. Rodriguez-Acosta, M. A. Mueses, F. Machuca-Martínez, Int. J. Photoenergy (2014) DOI:
10.1155/2014/817538.
[32] G. Guella, B. Patton, A. Miotello J. Phys. Chem C 111 (2007) 18744-18750.
[33] Q. Zhou, G. Qian, Y. Li, G. Zhao, Y. Chao, J. Zheng, Thin Solid Films 516 (2008) 953-956.
[34] Of course, the fitting analysis is carried out directly on spectrophotometric data. For this
purpose, equation (3) is trivially modified by posing: C0 = (Abs0-Abs)/ArNO2 and |ArNO2| = (Abst-
Abs)/ArNO2 ; where Abs0 and Abs are the initial and final absorbances respectively, Abst is the
absorbance at a given time t, andArNO2 is the molar extinction coefficient of the substrate at the

23
operational wavelength. Noticeably, Abs keeps into account the possible contribution to the
overall absorbance from the Ag-NP catalyst (the contribution from the aniline product is
negligible).
[35] a) S. Jana, S.K. Gosh, S. Nath, S. Pande, S. Paharaj, S. Panigrahi, S. Basu, T. Endo, T. Pal, Appl.
Catal. A 313 (2006) 41-48. b) K.B. Narayanan, N. Sakthivel, Biores. Technol. 102 (2011) 10737-
10740.
[36] The latter statement needs to be briefly justified. As a matter of fact, it has to be reminded
that most of the AmCD units present in the system cannot interact with the NPs surface and are
delivered into solution, as discussed previously. Delivered AmCD can form host-guest inclusion
complexes with the nitroarene substrate. The molar fraction of units engaged in this process
depends, in turn, on the binding constant KCD and on the analytical concentration of the substrate,
and it can be calculated algebraically. Now, by assuming that the KCD values for nitroarenes 1-8
towards AmCD are similar to those found in the case of native CD (see ref. [18] and [19]), it
follows that the concentration of a possible (ArNO2·AmCD) inclusion complex is actually negligible
with respect to free AmCD.
[37] X. Li, J. Wang, Y. Zhang, M. Li, J. Liu, Eur. J. Org. Chem. (2010) 1806-1812.
[38] a) T.-X. Lü, D.-B. Zhang, S.-J. Dong, J. Chem. Soc., Faraday Trans. 85 (1989) 1439-1445. b) H.-J.
Schneider, F. Hacket, V. Rüdiger, H. Ikeda, Chem. Rev. 98 (1998) 1755-1785. c) G.M. Bonora, R.
Fornasier, P. Scrimin, U. Tonellato, J. Chem. Soc., Perkin Trans. 2 (1985) 367-369. d) O.S. Tee, C.
Mazza, X.-X. Du, J. Org. Chem. 55 (1990) 3603-3609. e) P. Lo Meo, F. D’Anna, S. Riela, M.
Gruttadauria, R. Noto, Tetrahedron 58 (2002) 6039-6045. f) P. Lo Meo, F. D’Anna, S. Riela, M.
Gruttadauria, R. Noto, Tetrahedron 65 (2009) 2037-2042. g) P. Lo Meo, F. D’Anna, M.
Gruttadauria, S. Riela, R. Noto, Tetrahedron 65 (2009) 10413-10417.
[39] F. D’Anna, P. Lo Meo, S. Riela, M. Gruttadauria, R. Noto, Tetrahedron 57 (2001) 6823-6827.
[40] It is worth noting that in our particular system, unfortunately, it is very hard to verify
otherwise the occurrence of radical-involving mechanistic pathways. As a matter of fact, classical
radical scavengers such as phenols or thiols cannot be employed because, being acidic species,
-
they immediately decompose the BH4 ion.
[41] D. E. J. Koppel, Chem. Phys. 57 (1972) 4814-4820
[42] a) S.W. Provencher, Macromol. Chem. 180 (1979) 201-209. b) S.W. Provencher, Comput. Phys.
Commun. 27 (1982) 213-227.

24
NO2
NH2
[H]
[H] = LiAlH4 or Fe/HCl or H2/Pd
Scheme 1
HO
OH
HO
OH
Am-H
O
O
-H+
O
O
7
I
Am
7
. x HI
n
Scheme 2
+.AmCD
Ag
layers
Ag+.AmCD
macroaggregate
CH2O
dilution
Ag0
slow reduction
partial coating loss
.
+ AmCD
Ag
microaggregates
Ag0
new
NaBH4
f ast reduction
Ag0
modif ied
Ag0
pristine
Scheme 3
-
(a·BH4 ) + ArNO2
KB
KBR
k
-
-
a + BH4 + ArNO2
(a·BH4 ·ArNO2)
ArNH2
KR
KB*
-
(a·ArNO2) + BH4

25
Scheme 4
KCD
ArNO2 + AmCD
(ArNO2·AmCD)
-
(a·BH4 ) + ArNO2
KB(0)
KBR(0)
k(0)
-
-
a + BH4 + ArNO2
(a·BH4 ·ArNO2)
KB*(0)
KR(0)
-
(a·ArNO2) + BH4
K1
+ AmCD
ArNH2
-
AmCD
[(a·
)·BH4 ] + ArNO2
KB(1)
KBR(1)
k(1)
-
-
AmCD
(a·
AmCD
[(a·
) + BH4 + ArNO2
)·BH4 ·ArNO2]
KB*(1)
KR(1)
-
AmCD
[(a·
)·ArNO2] + BH4
K2
+ AmCD
-
AmCD
[(a·
2)·BH4 ] + ArNO2
KB(2)
KBR(2)
-
-
AmCD
(a·
2) + BH4 +ArNO2
AmCD
[(a·
2)·BH4 ·ArNO2]
X
KR(2)
-
AmCD
[(a·
2)·ArNO2] + BH4
Scheme 5

26
R : 1 -NH2
2 -NH-CH2-COO-Na+
-N(CH3)-CH2-COO-Na+
NO2
3
N
4
CH2OH
-O-Na+
-OCH3
-COO-Na+
5
6
7
R
8 -H
Figure 1. Nitroarene substrates
HO
OH
O
O
7
NH
N(CH3)2
. 4.0 HI
5.7
AmCD
Figure 2. Structure of AmCD
Figure 3. Normalized UV-vis spectra for pristine Catalyst II and its dilutions

27
a)
c)
b)
d)
e)
f)
Catalyst I
IIdilB
IIred
Figure 4. TEM micrographs of Catalyst I (a,d), IIdilB (b,e) and IIred (c,f). Micrographs a, b and c
acquired at a 50000X magnification; micrographs d, e and f acquired at a 150000X magnification.
Figure 5. FT-IR spectra of AmCD (….) and Ag NPs precipitate from Catalyst I (___).

28
Figure 6. Kinetic trace (squares) for the reaction of 1 (0.1 mM) with NaBH4 (4.0 mM) in the
presence of catalyst II (50 M).
-
Figure 7. Dependence of the initial reaction rate (c1) vs. BH4 concentration for the reaction of 1
(0.1 mM) in the presence of catalyst II (50 M).

29
-
Figure 8. Dependence of kobs/|BH4 | on CAg for Catalyst I (a) and Catalyst II (b).
Figure 9. Induction times ti for the reaction of 1 with Catalyst I at different concentrations.
Table 1.  values for the AmCD/Ag NP systems at the relevant max
.