Synthesis of Ag nanoparticles on oxide and carbon supports from Ag diammine precursor

Article abstract of DOI:10.1016/j.jcat.2016.08.018

To determine whether the method of “strong electrostatic adsorption” (SEA) can be extended to the preparation of uniform and highly dispersed supported Ag catalysts, the adsorption of silver diammine, Ag(NH₃)₂⁺, also known as Tollen's reagent, has been examined over five supports (i.e., SiO₂, γ-Al₂O₃, ZrO₂, Nb₂O₅, and carbon) with different surface areas and PZCs. The speciation of Ag in solution was followed by UV–vis spectroscopy, Ag uptake as a function of the solution pH was determined by atomic absorption, and Ag particle size was determined by powder XRD and STEM. At the Ag concentrations, pH, and surface loadings employed, silver diammine complexes convert into soluble Ag(H₂O)₂⁺ aquo complexes below pH 11 and hydrolyze to insoluble Ag₂O above pH 11. The deposition of either Ag species over carbon at any pH appears to be reactive and results in large particles. The silver aquo complexes appear to adsorb via ion exchange near the PZC of alumina, zirconia, and niobia but not silica, while the silver diammine complexes appear to adsorb electrostatically at high pH over all the oxides tested - niobia, silica, zirconia, and alumina in the order of ascending PZC. Niobia, which has the lowest PZC, adsorbs the highest surface density of Ag via both mechanisms. The particles obtained via electrostatic adsorption of Ag(NH₃)₂⁺ at high pH are somewhat smaller than those formed from Ag(H₂O)₂⁺ ions at the lower pH values. In the absence of ammonia in solution at high pH, deposition of Ag aquo complexes occurs via surface precipitation and gives large particles. In sum, SEA of silver ammine is demonstrated to be a simple, reproducible way to synthesize small particles on all supports but carbon.

Full text of DOI:10.1016/j.jcat.2016.08.018

Journal of Catalysis xxx (2016) xxx–xxx
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Synthesis of Ag nanoparticles on oxide and carbon supports from Ag
diammine precursor
Eleni A. Kyriakidou, Oleg S. Alexeev, Andrew P. Wong, Christina Papadimitriou, Michael D. Amiridis,
⇑
John R. Regalbuto
Department of Chemical Engineering, University of South Carolina, Columbia, SC 29208, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
To determine whether the method of ‘‘strong electrostatic adsorption” (SEA) can be extended to the
Received 11 June 2016
Revised 25 August 2016
Accepted 27 August 2016
Available online xxxx
preparation of uniform and highly dispersed supported Ag catalysts, the adsorption of silver diammine,
+
Ag(NH
Nb
3
)
2
, also known as Tollen’s reagent, has been examined over five supports (i.e., SiO
2
,
c
-Al
2
O
3
, ZrO
2
,
2
O
5
, and carbon) with different surface areas and PZCs. The speciation of Ag in solution was followed
by UV–vis spectroscopy, Ag uptake as a function of the solution pH was determined by atomic absorption,
and Ag particle size was determined by powder XRD and STEM.
Keywords:
At the Ag concentrations, pH, and surface loadings employed, silver diammine complexes convert into
Strong electrostatic adsorption
Ag catalysts
Silver diammine
Silica
Alumina
Zirconia
+
soluble Ag(H
2
O)
2
aquo complexes below pH 11 and hydrolyze to insoluble Ag
2
O above pH 11. The depo-
sition of either Ag species over carbon at any pH appears to be reactive and results in large particles. The
silver aquo complexes appear to adsorb via ion exchange near the PZC of alumina, zirconia, and niobia but
not silica, while the silver diammine complexes appear to adsorb electrostatically at high pH over all the
oxides tested - niobia, silica, zirconia, and alumina in the order of ascending PZC. Niobia, which has the
Niobia
Carbon
lowest PZC, adsorbs the highest surface density of Ag via both mechanisms. The particles obtained via
+
electrostatic adsorption of Ag(NH
3
)
2
at high pH are somewhat smaller than those formed from
+
2 2
Ag(H O) ions at the lower pH values. In the absence of ammonia in solution at high pH, deposition of
Ag aquo complexes occurs via surface precipitation and gives large particles. In sum, SEA of silver ammine
is demonstrated to be a simple, reproducible way to synthesize small particles on all supports but carbon.
Ó 2016 Elsevier Inc. All rights reserved.
1
. Introduction
wetness (IW), wet impregnation (WI), and sol-gel methods. While
this survey is not exhaustive, a number of trends affecting the size
Supported silver nanoparticles are catalytically active for a vari-
ety of chemical reactions, including decomposition of ozone [1],
reduction of NO [2,3], epoxidation of alkenes [4,5], dehydrogena-
tion of alcohols [6–9], and reduction of nitrophenols [10,11]. Some
of these reactions require relatively large Ag particles, but some
of Ag particles in these materials can be identified. For example,
AgNO is used most frequently as a precursor for all these conven-
tional preparation techniques and high temperature treatments
are employed to form metal particles. The support nature appears
to be a crucial factor, as average dimensions of silver particles
3
x
(
i.e., dehydrogenation of alcohols and selective catalytic reduction
2 3 2
formed on Al O and SiO supports were found to be on the order
of NO ) proceed with higher rates over supported Ag nanoparticles
x
of 11 and 3 nm, respectively [12,13]. For each specific support, cal-
cination in air at approximately 500 °C typically leads to a higher
metal dispersion as compared to the case when the calcination
with relatively small dimensions, suggesting that the architecture
of supported Ag particles must be strictly controlled for each
individual chemical process.
step is followed by reduction in H
2
at elevated temperatures
Extensive efforts have been made to synthesize supported silver
nanoparticles with desired dimensions and shapes. Table S1
[14]. In the latter case, however, the size of Ag particles increases
with reduction temperature. For example, in the case of 8%
provides
a
short literature survey of selected properties of
-supported Ag samples prepared by incipient
Ag/SiO
Ag particles were found to be 4.5 and 16.7 nm after reduction with
at 200 and 500 °C, respectively [15]. Finally, Table S1 also shows
that when AgNO was used to prepare silver nanoparticles in solu-
2
prepared by IW and calcined at 500 °C, average sizes of
Al - and SiO
2
O
3
2
H
2
3
⇑
Corresponding author.
E-mail address: regalbuj@cec.sc.edu (J.R. Regalbuto).
tion and Ag colloids thus formed were further used to prepare solid
catalytic materials, large Ag particles with sizes on the order of
http://dx.doi.org/10.1016/j.jcat.2016.08.018
021-9517/Ó 2016 Elsevier Inc. All rights reserved.
0
Please cite this article in press as: E.A. Kyriakidou et al., Synthesis of Ag nanoparticles on oxide and carbon supports from Ag diammine precursor, J. Catal.
2016), http://dx.doi.org/10.1016/j.jcat.2016.08.018
(

2
E.A. Kyriakidou et al. / Journal of Catalysis xxx (2016) xxx–xxx
1
50–220 nm were formed regardless of the support nature and
platinate (and related hydrolysis products) over alumina [31],
cationic Pt tetraammines over silica [32–34], cationic and anionic
Pt complexes over carbon [35–37].
In this work, the uptake of Ag diammine complexes by supports
2 5 2 2 3 2
with low and mid-range PZCs (i.e., Nb O , SiO , Al O and ZrO ) is
examined and used to prepare highly dispersed supported Ag
nanoparticles. Unlike other ammine complexes, silver exhibits
two uptake maxima, the first of which (at the lower pH) appears
to involve the additional deposition mechanism of ion exchange
near the PZC of some of the materials.
activation conditions used [16]. Overall, these literature examples
clearly show that conventional preparation techniques based on
the impregnation of Ag salts and high temperature treatments pro-
vide only a limited control over Ag particle sizes in solid materials
because Ag easily aggregates at elevated temperatures.
Some literature reports also describe the synthesis of Ag/SiO
from AgNO using a modified IW impregnation method in which
the pH of the initial AgNO aqueous solution was adjusted to 11
with ammonium hydroxide. When 2% Ag/SiO prepared in this
way was calcined in air at 500 °C, the following reduction treat-
ment in H at 200 °C for 2 h yielded Ag nanoparticles on the SiO
surface with an average size of approximately 0.8 nm [17]. While
pH adjustments were used in this case mainly to improve the
interaction between Ag species and the support surface during
the impregnation step, one cannot ignore the formation of
Ag(NH
2
3
3
2
2
2
2
. Materials and methods
Commercial Nb
Mineração – CBMM), ZrO
Evonik) and oxidized carbon (Asbury) supports with BET surface
2
O
5
ꢀ(Companhia Brasileira de Metalurgia
e
2
(Toyota), -Al (Toyota), fumed SiO
c
2
O
3
2
(
+
3
)
2
complexes in solution under reported conditions.
2
areas of 186, 96, 167, 148, and 114 m /g, respectively, were used
as supplied. AgNO (99% purity, Aldrich), 1 N NH OH (Alfa Aesar),
and 5 N NaOH (Alfa Aesar) aqueous solutions were used as
Only few other reports, demonstrating the formation of sup-
3
4
ported silver nanostructures from diammine complexes, Ag
+
(
NH
3
)
2
, can be found to date [18–20]. For example, it has been
received. Deionized water purified to 18.2 M
was used to prepare all aqueous solutions.
X
ꢀcm (Milli-Q grade)
+
shown that the treatment of SBA-15-supported Ag(NH
plexes in O at 300 °C followed by reduction with H at the same
temperature leads to the formation of relatively small (4 nm) Ag
nanoparticles or long (50 nm) Ag nanowires, depending on the
Ag loading used [18]. The formation of Ag nanoparticle ‘‘necklaces”
has been observed upon treatment of colloidal silica with Ag(NH
solution followed by washing and ageing [19]. When Ag colloids
prepared via chemical reduction of Ag(NH
were used to fabricate Ag/SiO , it was possible to control sizes of
supported Ag nanoparticles in the 2–10 nm range using a chemical
etching method [20].
In addition to the synthesis of various supported Ag structures,
several literature reports illustrate the beneficial catalytic proper-
ties of nano-sized particles. For example, it has been shown that
Ag -exchanged zeolites are capable of activating CH
fore, catalyzing the reaction of CH
pene [21]. Hydrotalcite-supported Ag particles with an average
diameter of 3.3 nm were found to be active for dehydrogenation
3 2
) com-
2
2
The PZC of each support was determined by the method
described elsewhere [38]. In a typical measurement, appropriate
amounts of each specific support were added to aqueous solutions
adjusted to different initial pH values to achieve a surface loading
+
)
2
2
3
(
the total support area per liter of solution) of 10,000 m /L. Solu-
tions were allowed to stabilize for approximately 1 h and the pH
was measured again. The PZC of the solid corresponds to a plateau
on a plot of final pH vs the initial pH [38]. Based on these measure-
+
3 2
) by formaldehyde
2
ments, Al
values of 7.8 and 6.8, respectively, while SiO
have low PZC values of 4, 3, and 4.7, respectively.
Silver diammine nitrate (Ag(NH NO ), also known as a Tol-
len’s reagent, was prepared according to a two-step procedure
2
O
3
and ZrO
2
supports were found to have medium PZC
2
, Nb , and carbon
2 5
O
3
)
2
3
+
and, there-
3
4
reported elsewhere [20]. In short, a 3.15 ꢀ 10 ppm AgNO
3
aqueous
in
00 mL of deionized water. The resulting solution was transferred
4
and C
2
H
4
coupling to form pro-
solution was initially prepared by dissolving 0.315 g of AgNO
3
1
to a 1 L volumetric flask and 450 lL NaOH solution (5 N) was
added dropwise to the flask, leading to the formation of a brown
3
of alcohols [22]. It has been shown that Ag clusters deposited from
a molecular beam on amorphous alumina films and 3.5 nm Ag
nanoparticles formed from them at elevated temperatures can cat-
alyze propene epoxidation with the latter species being more
selective toward propylene oxide at temperatures up to 120 °C
precipitate (Ag O).
2
2AgNO
þ 2NaOHðaqÞ ! Ag OðsÞ þ 2NaNO3ðaqÞ þ H
2
O
ð1Þ
3ðaqÞ
2
4
The subsequent addition of 12 mL NH OH (1 N) to the mixture
[
23]. Finally, unique catalytic properties of subnano- and nano-
sized Al -supported Ag clusters in selective catalytic reduction
of NO by hydrocarbons and a variety of environmentally friendly
+
dissolves the precipitate due to the formation of Ag(NH
complexes.
3 2
)
2 3
O
x
organic reactions have been reported elsewhere [24–27].
Ag O þ 4NH3ðaqÞ þ 2NaNO3ðaqÞ þ H O
2
ðsÞ
2
ðlÞ
Although the findings mentioned above indicate that supported
Ag nanostructures tend to have higher activities and selectivities
for specific reactions, a controllable and repeatable method to syn-
thesize small Ag nanoparticles does not appear to have been
demonstrated. The purpose of this paper was to demonstrate that
!
2AgðNH
3
Þ2NO3ðaqÞ þ 2NaOHðaqÞ
ð2Þ
The 1 L volumetric flask was filled with deionized water to yield
a 200 ppm Ag concentration in the final solution with a basic pH of
11.3. The solution was vigorously stirred during the whole
preparation procedure and an amber flask was used due to the
light sensitivity of silver.
‘
‘Strong Electrostatic Adsorption” (SEA) [28–30] is one such
method.
SEA exploits the protonation-deprotonation chemistry of oxide
3 2 3
Approximately 40 mL aliquots of the stock Ag(NH ) NO solu-
and carbon surfaces in solutions. Every oxide support has a point of
zero charge (PZC) and by definition, the PZC is the pH at which the
support surface is neutrally charged. When the solution pH is
below the PZC value, the hydroxyl groups are protonated and
become positively charged. Under these conditions, the surface
can adsorb anionic metal complexes. The same groups become
deprotonated and negatively charged when the solution pH is
above the PZC of the oxide support, and thus positively charged
species can be adsorbed by the support surface [28,29]. To date,
the electrostatic adsorption mechanism has been demonstrated
mainly for noble metal complexes, including anionic hexachloro-
tion were used to determine the nature of Ag species existing in
solution at different pH values. In each case, the pH of the solution
was adjusted to a desire value with 1.2 M HNO or 1 M NaOH solu-
3
tions, as it was appropriate, the acquired solution was stirred until
no changes in pH were observed, and it was examined by UV–vis.
Adsorption experiments were performed at a constant surface
2
loading of 1000 m /L. In these experiments, appropriate amounts
of each support were added into 40 mL aliquots of the stock Ag
+
(NH
3
)
2
solutions the pH of which was adjusted in the 1–13 range
or 1 M NaOH solutions, as appropriate.
Suspensions thus formed were vigorously stirred and sampled for
(pHInitial) with 1.2 M HNO
3
Please cite this article in press as: E.A. Kyriakidou et al., Synthesis of Ag nanoparticles on oxide and carbon supports from Ag diammine precursor, J. Catal.
2016), http://dx.doi.org/10.1016/j.jcat.2016.08.018
(

E.A. Kyriakidou et al. / Journal of Catalysis xxx (2016) xxx–xxx
3
pH after 1 h (pHFinal). Then, 3–4 mL portions of these suspensions
were removed, the solid was filtered, and the liquid phase was ana-
lyzed by Atomic Absorption (AA) for the Ag content. The adsorp-
the range of pH values above 11.3, the concentration of Ag in solu-
+
tion sharply declined. This result was expected, as Ag(NH
3 2
) com-
plexes are not stable in strongly basic solutions and their
tion density of Ag (
C
Ag), expressed as micromoles Ag adsorbed
2
decomposition leads to the formation of insoluble Ag O [40].
per square meter of the support surface, was calculated as follows:
þ
þ 2OHꢁ ! Ag OðsÞ þ 4NH
2
AgðNH
3
Þ
3
þ H
2
O
ð3Þ
2
2
ðCAg;Initial ꢁ CAg;FinalÞ ð
l
mol=LÞ
2
+
C
Ag
¼
¼ uptake ð
l
mol=m Þ
At the low end of the pH scale, Ag(NH ) can undergo ligand
exchange to yield Ag(H
following:
2
3 2
SL ðm =LÞ
+
2
O)
2
aquo complexes [41] according to the
where CAg,Initial and CAg,Final are initial and final concentrations of Ag
in solution, respectively, and SL is the surface loading.
A Perkin Elmer (AAnalyst 400) atomic absorption spectrometer
was used to quantify the concentration of Ag in solutions. Calibra-
tion standards were prepared from a 1000 ppm Ag stock solution
þ
þ
þ
OÞ2 þ 2NHþ4
AgðNH
3
Þ
þ 2H
3
O
! AgðH
2
ð4Þ
2
To study the speciation of Ag at lower pH values, UV–vis spec-
troscopy was used. Fig. 2 shows UV–vis spectra of Ag(H O) and Ag
(NH3) aqueous solutions containing 1500 ppm of Ag. This rela-
tively high concentration of Ag was chosen to obtain spectra with
a reasonable intensity of absorption bands. The UV–vis spectrum of
Ag(H O) + aquo complexes (Fig. 2A, spectrum (1) is characterized
by a single relatively strong absorption band at 300 nm, while the
+
2
2
+
(
J. T. Baker) and the calibration curve obtained was verified with
an independent second standard solution of 2 ppm Ag in 5% (v/v)
HNO (Ultra Scientific). All solutions were stored in amber glass
volumetric flasks for light protection. Before measurements, all
2
3
2
2
solutions were diluted with 5% HNO
of Ag concentrations.
3
to reach the 0–5 mg/L range
+
spectrum collected for the aqueous solution of Ag(NH3)2 (Fig. 2A,
UV–vis spectra were collected at room temperature using a
Shimadzu UV-2010PC spectrophotometer and quartz cells with a
path length of 10 mm. All samples were scanned in the
spectrum (2) features a weaker absorption band at approximately
400 nm [42,43]. Since these bands are well resolved, it is possible
to estimate the speciation of Ag in the solution during conversion
+
+
1
90–800 nm range at 0.5 nm steps. Deionized water was used as
2 2 2
of Ag(NH3) to Ag(H O) complexes with decreasing the solution
the reference for all measurements.
To synthesize catalytic materials in sufficient quantities, the
adsorption experiments were simply scaled up as has been previ-
pH by monitoring their intensity, as shown in Fig. 2B. The band
+
at 400 nm assigned to Ag(NH ) continuously declines in intensity
3
2
as pH lowers, while a new band simultaneously develops at
300 nm and increases in intensity, signifying transformation of
ously reported [37,39]. Approximately 1 g of each specific support
+
+
+
was mixed with an appropriate amount of a 200 ppm Ag(NH
3
)
2
2 2 2
Ag(NH3) into Ag(H O) complexes (reaction 4). When the results
stock solution the pH of which was adjusted to a desired value to
of the UV–vis experiments are quantified as shown in Fig. 3, it
2
+
yield a suspension with surface loading of 1000 m /L. The suspen-
becomes evident that Ag(H O) aquo complexes dominate in solu-
2
2
sion was stirred for 1 h before the solid was separated by filtration
and dried at room temperature for 48 h. The elemental analysis of
solutions before and after interaction with the support was used to
determine the Ag loadings in solid materials. Solid samples thus
tion at pH values below 9.3, while Ag ammines are the dominant
species in more basic solutions. However, there is a relatively wide
range of pH values (approximately 8.3–10.3) when both types of
Ag species coexist in solution in various concentrations. The forma-
+
+
formed were treated with a 10% H
2
/He mixture while the temper-
f 3 2 2 2
tion constant (K ) of Ag(NH ) species from Ag(H O) in water (Ag
+
+
7
ature was ramped at 2 °C/min to 100 °C and held at this tempera-
ture for 2 h; TPR profiles (not shown) confirmed the reduction of
Ag.
(H O) + 2 NH M Ag(NH ) + 2 H O) at 25 °C is 1.7 ꢀ 10 [44] and
2
2
3
3 2
2
equilibrium speciation computed based on this constant is reason-
ably consistent with the experimental data in Fig. 3.
Powder X-ray diffraction (XRD) measurements were performed
using a Rigaku MiniFlex II bench top system with Cu Ka radiation
3.2. Silver uptake surveys
(k = 1.5406 Å) operated at 30 kV and 15 mA. This instrument
features a high sensitivity solid state (D/tex Ultra) detector with
a detection size limit of about 1.5 nm. Scans were made in the
The uptakes of silver as a function of pHFinal and the correspond-
ing shifts in pH values over all examined materials are shown in
Fig. 4. Two trends are observed in the uptake surveys (left hand
side graphs). First, the maximum Ag uptake increases as the PZC
drops, and second, for all supports (except silica), two maxima
appear in the Ag uptake profiles. The pH shifts for the metal-free
control and metal-containing experiments (right hand side graphs)
also exhibit two trends. First, in the high pH region, the pH values
of the metal containing solutions are above those of the control
experiments, indicating the consumption of protons during
1
0
0–70° 2h range, with a step size of 0.02°, and a scan rate of
.5°/min. The XRD particle size was calculated using the Scherrer
equation with a shape factor of 0.94.
BET surface area measurements and TPR were performed using
a ChemBET-3000 analyzer (Quantachrome Instruments). All sup-
ports were outgassed in He at 120 °C for 3 h prior to measure-
ments. The surface area of each support was then determined by
2
N adsorption using a multipoint BET method.
adsorption. The consumption of protons is caused by the speciation
+
3
. Results and discussion
of Ag(NH
3
)
2
complexes, as the ammine ligands (NH
3
) released from
+
the metal complex are converted into ammonium ions (NH
tion 4). This process is nearly complete at a pH of approximately 6
Fig. 3). Speciation-driven proton consumption is especially evi-
4
) (reac-
3.1. Ag solution stability and speciation
(
The stability of the 200 ppm Ag diammine solution at different
dent in Fig. 4J, where larger pH shifts are observed for the higher
metal concentration. Second, for the alumina, zirconia, and niobia
cases, the pH of the metal containing solutions falls below the con-
trol experiment at low pH, indicating the generation of protons
during metal deposition.
pH values was examined first and results obtained are shown in
Fig. 1, where the concentration of Ag (determined by AA) for repli-
cated sets of solutions aged for 1 h is plotted as a function of the
solution pH. The data of Fig. 1 clearly show that the concentration
of Ag remains nearly unchanged in solution when 1.2 M HNO
3
in
The maximum uptake of cationic precursors via electrostatic
adsorption is expected to increase with lower PZC materials
[28,29] and this can explain the higher surface densities of the
high-pH maxima as substrate PZC drops (left hand side graphs of
appropriate amounts was used to adjust pH values in the 3–11
range, suggesting that this range of pH is the most suitable for
SEA measurements. In contrast, when 1 M NaOH was used to cover
Please cite this article in press as: E.A. Kyriakidou et al., Synthesis of Ag nanoparticles on oxide and carbon supports from Ag diammine precursor, J. Catal.
2016), http://dx.doi.org/10.1016/j.jcat.2016.08.018
(

4
E.A. Kyriakidou et al. / Journal of Catalysis xxx (2016) xxx–xxx
Fig. 1. Concentration of Ag in an aqueous silver diammine solution as a function of pH.
+
+
Fig. 3. Experimental mole fractions of Ag(H
2
O)
2
(1) and Ag(NH
3
)
2
(2) species
+
+
calculated from UV–vis data and theoretical Ag(H
2
O)
species formation equilibrium: Ag(H
2 3 2
and Ag(NH ) mole fractions
+
+
(
NH
(K
b
3) - calculated by using the Ag(NH
3
)
2
2
O)
2
+ 2
+
7
+
3
M Ag(NH
)
3 2
+ 2
H
2
O
(K
f
= 1.7 ꢀ 10 [39]) and the NH
3 4
/NH equilibrium
= 1.8 ꢀ 10ꢁ - as a function of the solution pH.
5
Fig. 4). The surface densities observed for all supports except niobia
2
2
(
i.e., in the range from 0.4
lmol/m for alumina to 1.6 lmol/m for
silica) are typical of cationic metal complexes retaining one or two
+
hydration sheaths [28,29,32–37]. The diameter of Ag(NH
3
)
2
is
3
.93 Å [45], such that the area of a hydrated (dH2O = 2.75 Å) silver
2
2
ammine complex is p(3.93/2 + 2.75) = 70. Å /complex, which con-
verts to 2.4
2
lmol/m . By the same calculation, if the diammine
complex retains two hydration sheaths, the maximum density
2
.
would be 0.95 lmol/m It would then appear that the diammine
complexes retain an average of between one or two sheaths. Surface
+
2 2
Fig. 2. UV–vis spectra of (A) 1500 ppm aqueous solutions of Ag(H O) (1) and Ag
+
+
densities over silica approaching this value have been seen with [Co
(
NH
3 2 3 2
) (2) and (B) 1500 ppm aqueous solution of Ag(NH ) at pH (1) 11.3, (2) 9.3,
3
+
3+
(
NH )
3 6
]
3 6
and Ru(NH ) [39,46].
and (3) 8.3.
Please cite this article in press as: E.A. Kyriakidou et al., Synthesis of Ag nanoparticles on oxide and carbon supports from Ag diammine precursor, J. Catal.
2016), http://dx.doi.org/10.1016/j.jcat.2016.08.018
(

E.A. Kyriakidou et al. / Journal of Catalysis xxx (2016) xxx–xxx
5
Fig. 4. Ag uptakes for a 200 ppm silver diammine solution as a function of pH and corresponding pH shifts observed over alumina (A and B), zirconia (C and D), oxidized
2
carbon (E and F), silica (G and H), and niobia (I and J), respectively. In all cases, the surface loading is 1000 m /L.
Please cite this article in press as: E.A. Kyriakidou et al., Synthesis of Ag nanoparticles on oxide and carbon supports from Ag diammine precursor, J. Catal.
(2016), http://dx.doi.org/10.1016/j.jcat.2016.08.018

6
E.A. Kyriakidou et al. / Journal of Catalysis xxx (2016) xxx–xxx
A notable result over niobia is the complete uptake of the
00 ppm solution over a very wide pH range; the niobia surface
3.3. Synthesis of Ag nanoparticles
2
could only be saturated with a 1500 ppm solution (Fig. 4I).
In order to synthesize the amount of catalyst needed for the
The maximum level of uptake over niobia was as high as
XRD characterization, the adsorption survey experiments were
2
+
1
4
lmol/m . High metal ammine uptakes over niobia have been
simply scaled up to 1.0 L of solution. Adsorption of Ag(NH
3
)
2
was
observed in the past with platinum ammines [47]. Niobia con-
performed at pH values where the two Ag adsorption maxima were
2
tains more than 12 OH/nm [48] on its surface, that is much more
observed (Fig. 4, corresponding to ion exchange and electrostatic
2
2
+
than that of alumina (8 OH/nm ) and silica (5 OH/nm ) [9,32]. The
significantly higher hydroxyl density of niobia may impart to the
deprotonated surface sufficient coulombic attraction to overcome
the solvation energy of removing the hydration sheaths of the
uptake), and deposition of Ag(H
2
O)
2
was conducted at a lower
pH corresponding to electrostatic adsorption and a higher pH of
precipitation (Fig. 5). The XRD patterns of Ag nanoparticles pre-
+
+
3 2 2 2
pared from Ag(NH ) versus Ag(H O) are shown in Fig. 6 on the
adsorbing complex. Calculated as unhydrated disks of radius
left and right hand sides, respectively, and the XRD particle sizes
are summarized in Table S1. The deposition of Ag ammines from
a 200 ppm solution over silica at intermediate (8.3) and optimal
(9.9) pH levels (Fig. 6a) yields samples with Ag loadings of 1.7
and 2.5 wt.%, respectively. After reduction these samples exhibit
no crystallinity in XRD, consistent with the formation of ultra-
small Ag nanoparticles below the limit of XRD detection (about
1.5 nm). With 200 ppm Ag aquo complexes, no Ag crystallites were
observed at the intermediate pH of 8.3 (Fig. 6b) but at pH 10.5, the
resulting Ag particles were on the order of 17 nm.
2
3
=
.93 Å, the uptake of Ag ammine complexes would be
12 Å /complex, which converts to 8.3 lmol/m . This is not too
p
(3.93) -
2
2
2
far from the observed maximum value of about 14 lmol/m ,
especially if it is considered that the linear, rod-shaped diammine
complexes would pack more densely than disks on which the cal-
culation is based. The dip in Ag uptake after the first maximum at
pH 7 can be explained by the transition of uptake mechanism
from ion exchange to electrostatic adsorption; as pH increases
the neutral surface becomes more negatively charged, as the Ag
+
+
(
OH)
2
species are converted to Ag(NH
3
)
2
species. Deposition by
Over niobia, two Ag diammine uptake maxima were observed
at pH 7.8 and 11.5 (Fig. 4I) and Ag loadings of Ag/Nb O prepared
2 5
+
ion exchange of Ag(OH)
2
over the AOH groups is apparently not
compensated by electrostatic adsorption of the diammine com-
plexes until the pH is significantly above the PZC of the niobia
surface.
at these pH values were found to be 14 and 15 wt.%, respectively.
Broad XRD peaks were observed that correspond to Ag particles
of approximately 3.5 nm in size (Fig. 6c). The relatively low area
of these peaks relative to the 15 wt% sample in Fig. 6d suggests that
further fractions of the ammine-derived samples exist as particles
below the limit of XRD detection. XRD patterns of samples pre-
pared from 200 ppm Ag aquo complex solutions at pH 5.25, where
Silica (Fig. 4G and H) exhibits neither a low pH maximum, nor a
low pH production of protons, implying that silver species do not
2
ion exchange over silica. The maximum uptake of 1.6 lmol/m
(Fig. 4G) is consistent with other metals adsorbing with one hydra-
tion sheath on silica [33]. On the other hand, protons are also not
released over carbon (Fig. 4E and F), and the Ag uptake occurs over
a very wide pH range. This can be attributed to the reactivity of the
silver complex to carbon, for which, in fact, Tollen’s reagent is
known [49]. Specifically, the carbon uptake curve has two maxima
+
at pH 8 and 11.8, where the majority of the Ag species are Ag(H
O)
2 2
+
3 2
and Ag(NH ) , respectively (Fig. 3). This indicates that depending
on the adsorption pH carbon is reactive towards different Ag spe-
cies. In summary, it appears that the deposition of silver over car-
bon is purely reactive, over silica purely electrostatic, and over
alumina, zirconia, and niobia involves ion exchange near the PZC
of these oxides and electrostatic at higher pH values.
+
In the absence of ammonia, Ag(H
pH leading to the formation of Ag
parison between Ag(NH
O)
2 2
can precipitate at high
2
O. Fig. 5A and B shows a com-
+
+
3 2 2 2
) and Ag(H O) adsorption over silica
and niobia supports, respectively. To prevent immediate precipi-
+
tation of Ag(H
2
O)
2
in generating its uptake curve the supports
aqueous solution at a pH of approx-
were added into the AgNO
3
imately 6.0 and the pH was adjusted upward with appropriate
amounts of NaOH. In this manner, typical of deposition-
+
precipitation [50], the precipitation of Ag(H
2
O)
2
is induced over
the support surfaces. The Ag uptake curves obtained with the
AgNO aqueous solution for these low PZC supports, silica in
Fig. 5A and niobia in Fig. 5B, are similar. In both cases, the depo-
3
+
sition of Ag(H
2
O)
2
is most likely electrostatic at pH 6 and below,
+
as Ag(H O) does not precipitate in this pH region. However, pre-
cipitation of Ag(H
2 2
+
2
2
O) takes place at pH higher than 6 and this
deposition mode supplements the SEA in some range of pH values
but most likely prevails as the solution becomes increasingly
basic. The Ag uptake curves reach maxima at a pH of approxi-
mately 9.5 and 7.5 for silica and niobia, respectively, and remain
unchanged thereafter, consistent with a complete precipitation of
+
Ag(H
2
O)
2
(surface or otherwise) under these conditions. The par-
ticle sizes resulting from these two deposition regimes, surface
+
+
2 2 3 2
Fig. 5. Comparison of Ag uptakes from 1500 ppm Ag(H O) and Ag(NH ) solutions
+
+
2 2 3 2
precipitation of Ag(H O) at high pH and SEA of Ag(NH ) , are
by silica (A) and niobia (B) as a function of final pH (In all cases, the surface loading
2
reported in the next section.
is 1000 m /L).
Please cite this article in press as: E.A. Kyriakidou et al., Synthesis of Ag nanoparticles on oxide and carbon supports from Ag diammine precursor, J. Catal.
2016), http://dx.doi.org/10.1016/j.jcat.2016.08.018
(

E.A. Kyriakidou et al. / Journal of Catalysis xxx (2016) xxx–xxx
7
1
7 nm
3
1
.5 wt%
.5 wt%
2
1
.5 wt%
.7 wt%
3
3
.5 nm
.5 nm
3
.7 nm/
1
2 nm
1
2
5 wt%
2 wt%
1
5 wt%
4 wt%
3
2
.4 nm/
0 nm
1
1
2
9 nm
2 nm
2.1 wt%
1.8 wt%
1
2
7 nm
2 nm
Fig. 6. XRD patterns of solid samples prepared over silica (a and b), niobia (c and d), and carbon (e and f) from Ag diammine or Ag aquo complex solutions, respectively.
the aquo and ammine complexes show the same uptake (Fig. 5b),
and pH 9, where the aquo uptake is much higher than the ammine,
are shown in Fig. 6d. In both cases, Ag peaks are sharpened and of
higher overall intensity than the peaks of the Ag ammine precur-
sors in Fig. 6c. The pH 5.25 pattern was deconvoluted into two
peaks of 20 and 3.4 nm, for an overall weighted average size of
On carbon, XRD-patterns shown in Fig. 5e and f for the ammine
and aquo complexes respectively include sharp Ag metal peaks and
the determined Ag particle sizes are on the order of 20 nm, regard-
less of the pH and the precursor used. This is consistent with the
reactive deposition mechanism of silver over carbon. It appears
that the aquo complexes react in a fashion similar to the ammine,
even though the lower XRD peak intensity of Ag/C prepared from
Ag diammine at pH 11.7 (Fig. 6e, upper pattern) suggests that at
3
.9 nm, while the pH 9 sample is comprised of 2.7 and 12 nm peaks
for an overall average of 4.9 nm.
Please cite this article in press as: E.A. Kyriakidou et al., Synthesis of Ag nanoparticles on oxide and carbon supports from Ag diammine precursor, J. Catal.
2016), http://dx.doi.org/10.1016/j.jcat.2016.08.018
(

8
E.A. Kyriakidou et al. / Journal of Catalysis xxx (2016) xxx–xxx
least some of the Ag particles are below the XRD detection limit;
there may be an electrostatic component to a predominantly reac-
tive deposition for the ammine precursor.
[3] N. Aoyama, K. Yoshida, A. Abe, T. Miyadera, Catal. Lett. 43 (1997) 249.
[
[
[
4] J.G. Serafin, A.C. Liu, S.R. Seyedmonir, J. Mol. Catal. A Chem. 131 (1998) 157.
5] G. Lu, X. Zuo, Catal. Lett. 58 (1999) 67.
6] A.N. Pestryakov, Catal. Today 28 (1996) 239.
For the sake of comparison, two other Ag nanoparticles synthe-
sis methods – thermal spreading of Ag O and incipient wetness
2
[7] A. Nagy, G. Mestl, Appl. Catal. A Gen. 188 (1999) 337.
[8] K. Shimizu, R. Sato, A. Satsuma, Angew. Chem. Int. Ed. 48 (2009) 3982.
9] K. Shimizu, K. Sugino, A. Satsuma, Chem. Eur. J. 15 (2009) 2341.
[
impregnation – were explored and are detailed in the supplemen-
tary information. Thermal spreading was explored in light of liter-
ature reports on the ability to disperse Ag onto supports by
calcination over alumina [26] and silica [17]. Briefly, it was found
that Ag can be induced to wet the surfaces of alumina and niobia
to a limited extent, but not silica or zirconia. Second, incipient wet-
ness impregnation (IW) was conducted with Ag ammine at weight
loadings corresponding to the SEA experiments. Large particles
appeared on niobia and alumina, and these could be reduced in
size by calcination as seen in the thermal wetting experiments,
and particles over carbon were again large. Smaller particles
formed over silica and zirconia as determined by XRD, but upon
closer inspection with electron microscopy, small numbers of very
large particles did appear over silica.
[
10] N. Pradhan, A. Pal, T. Pal, Colloid Surface A 196 (2002) 247.
[11] K. Esumi, R. Isono, T. Yoshimura, Langmuir 20 (2004) 237.
[
12] K. Arve, K. Svennerberg, F. Klingstedt, K. Eränen, L.R. Wallenberg, J.O. Bovin, L.
Cˇ apek, D.Y. Murzin, J. Phys. Chem. B 110 (2006) 420.
[
13] M.J. Beier, T.W. Hansen, J.D. Grunwaldt, J. Catal. 266 (2009) 320.
[14] E. Sayah, D. Brouri, Y. Wu, A. Musi, P. Da Costa, P. Massiani, Appl. Catal. A Gen.
06 (2011) 94.
15] X. Zhang, Z. Qu, X. Li, M. Wen, X. Quan, D. Ma, J. Wu, Sep. Purif. Technol. 72
2010) 395.
[16] R. Vadakkekara, M. Chakraborty, P.A. Parikh, Colloid Surface A 399 (2012) 11.
4
[
(
[
17] H. Wei, C. Gomez, J. Liu, N. Guo, T. Wu, R. Lobo-Lapidus, C.L. Marshall, J.T.
Miller, R.J. Meyer, J. Catal. 298 (2013) 18.
[
18] D.C. Lim, I. Lopez-Salido, Y.D. Kim, Appl. Surf. Sci. 253 (2006) 959.
[19] M. Hanisch, M. Mackovic, N. Taccardi, E. Spiecker, R.N.K. Taylor, Chem.
Commun. 48 (2012) 4287.
[
[
[
[
20] M. Zienkiewicz-Strazałka, S. Pasieczna-Patkowska, M. Kozak, S. Pikus, Appl.
Surf. Sci. 266 (2013) 337.
21] T. Baba, Y. Iwase, K. Inazu, D. Masih, A. Matsumoto, Micropor. Mesopor. Mater.
101 (2007) 142.
22] T. Mitsudome, S. Arita, H. Mori, T. Mizugaki, K. Jitsukawa, K. Kaneda, Angew.
Chem. Int. Ed. 47 (2008) 138.
23] Y. Lei, F. Mehmood, S. Lee, J. Greeley, B. Lee, S. Seifert, R.E. Winans, J.W. Elam, R.
J. Meyer, P.C. Redfern, D. Teschner, R. Schlögl, M.J. Pellin, L.A. Curtiss, S. Vajda,
Science 328 (2010) 224.
4
. Conclusions
+
Silver diammine complexes are converted to Ag(H
2
O)
2
aquo
complexes below pH 11 and hydrolyze to insoluble Ag
2
O above
[
24] K. Shimizu, K. Sawabe, A. Satsuma, Cat. Sci. Tecnol. 1 (2011) 331.
pH 11. Deposition of either Ag complexes over carbon at any pH
[25] T. Chaieb, L. Delannoy, C. Louis, C. Thomas, Appl. Catal. B Environ. 142 (2013)
780.
[26] T. Chaieb, L. Delannoy, G. Costentin, C. Louis, S. Casale, R.L. Chantry, Z.Y. Li, C.
Thomas, Appl. Catal. B Environ. 156 (2014) 192.
appears to be reactive, resulting in large (ꢂ20 nm) metal particles.
+
The Ag(H
2
O)
2
aquo complexes appear to adsorb via ion exchange
near the PZC of alumina, zirconia, and niobia but not silica, but pre-
[
27] S. Kameoka, T. Chafik, Y. Ukisu, T. Miyadera, Catal. Lett. 51 (1998) 11.
+
cipitate to form large particles at high pH, while the Ag(NH
3
)
2
com-
[28] J.R. Regalbuto, in: K.P. de Jong (Ed.), Synthesis of Solid Catalysts, Wiley VCH
Verlag, 2009.
plexes appear to adsorb electrostatically at high pH over all the
oxides tested - niobia, silica, zirconia, and alumina. Niobia is able
to adsorb the maximum amount of silver via both mechanisms.
[
29] J.R. Regalbuto, in: J.R. Regalbuto (Ed.), Catalyst Preparation, Science and
Engineering, Taylor & Francis, 2007.
[30] A.P. Wong, E.A. Kyriakidou, T.J. Toops, J.R. Regalbuto, Catal. Today 267 (2016)
45.
31] J.R. Regalbuto, A. Navada, S. Shadid, M.L. Bricker, Q. Chen, J. Catal. 184 (1999)
35.
[32] J.T. Miller, M. Schreier, A.J. Kropf, J.R. Regalbuto, J. Catal. 225 (2004) 203.
1
The particles obtained via electrostatic adsorption at high pH
[
+
appear to be somewhat smaller than Ag(H
2
O)
2
ion exchanged at
3
the lower pH values. SEA with Ag diammines at high pH is demon-
strated to be a simple, reliable method to synthesize small Ag
nanoparticles over all oxide supports, but reacts to form large par-
ticles on carbon.
[
[
[
33] L. Jiao, J.R. Regalbuto, J. Catal. 260 (2008) 342.
34] M. Schreier, J.R. Regalbuto, J. Catal. 225 (2004) 190.
35] N. Job, S. Lambert, M. Chatenet, C.J. Gommes, F. Maillard, S. Berthon-Fabry, J.R.
Regalbuto, J.P. Pirard, Catal. Today 150 (2010) 119.
[
36] S. Lambert, N. Job, L. D΄Souza, M.F. Ribeiro Pereira, R. Pirard, B. Heinrichs, J.L.
Figueiredo, J.P. Pirard, J.R. Regalbuto, J. Catal. 261 (2009) 23.
37] X. Hao, S. Barnes, J.R. Regalbuto, J. Catal. 279 (2011) 48.
38] J. Park, J.R. Regalbuto, J. Colloid Interf. Sci. 175 (1995) 239.
Acknowledgments
[
[
The authors gratefully acknowledge Toyota Motor Engineering
Manufacturing North America, Inc., for the partial financial sup-
port of this work. JR gratefully acknowledges the support of NSF
grant CBET 1160023.
[39] L. Jiao, J.R. Regalbuto, J. Catal. 260 (2008) 329.
[
[
40] R.S. Liu, Controlled Nanofabrication, Advances and Application, Pan Stanford
Publishing, 2013.
41] B.W. Brian, Principles of Inorganic Chemistry, J. Wiley & Sons Inc, New Jersey,
2015.
&
[
[
42] H. Wang, X. Qiao, J. Chen, S. Ding, Colloid Surface A 256 (2005) 111.
43] I. Texier, S. Rémita, P. Archirel, M. Mostafavi, J. Phys. Chem. 100 (1996) 12472.
Appendix A. Supplementary material
[44] T.L. Brown, H.E. Iemay Jr., B.E. Bursten, C.J. Murphy, P.M. Woodward, Chemistry
the Central Science, Pearson Prentice Hall, USA, 2012.
[
45] T. Shoeib, R.K. Milburn, G.K. Koyanagi, V.V. Lavrov, D.K. Bohme, K.W.M. Siu, A.
C. Hopkinson, Int. J. Mass Spectrom. 201 (2000) 87.
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2016.08.018.
[46] L. D’Souza, L. Jiao, J.R. Regalbuto, J.T. Miller, A.J. Kropf, J. Catal. 248 (2007) 165.
[
[
[
[
47] Y. Zha, Ph.D. Dissertation, University of Illinois at Chicago, 2007.
48] I. Nowak, M. Ziolek, Chem. Rev. 99 (1999) 3603.
49] T. Textor, M.M.G. Fouda, B. Mahltig, Appl. Surf. Sci 256 (2010) 2337.
50] M. Skaf, S. Aouad, S. Hany, R. Cousin, E. Abi-Aad, A. Aboukaïs, J. Catal. 320
References
[
[
1] S. Imamura, M. Ikebata, T. Ito, T. Ogita, Ind. Eng. Chem. Res. 30 (1991) 217.
2] K. Shimizu, K. Sugino, K. Kato, S. Yokota, K. Okumura, A. Satsuma, J. Phys. Chem.
C 111 (2007) 6481.
(
2014) 137.
Please cite this article in press as: E.A. Kyriakidou et al., Synthesis of Ag nanoparticles on oxide and carbon supports from Ag diammine precursor, J. Catal.
2016), http://dx.doi.org/10.1016/j.jcat.2016.08.018
(