Origin and Abatement of Heterogeneity at the Support Granule Scale of Silver on Silica Catalysts

Article abstract of DOI:10.1002/cctc.201700753

Incipient wetness impregnation is used commonly to form supported metal nanoparticle catalysts. Recently, it has been revealed that this approach may induce severe heterogeneity between catalyst granules of the same batch. At least a 10-fold variation in metal loading was observed, which affect the catalytic performance of individual catalyst granules severely. However, the origin of this heterogeneity is still unclear. Here we show that every elementary step in the preparation procedure of a Ag on silica catalyst has an effect on the resulting interparticle heterogeneity, but the influence of the drying step is the most important. This is because drying by capillary force results in a heterogeneous sample. Specifically, the position of a granule in the stagnant drying bed influences the resulting color and, thus, Ag loading significantly. This is further demonstrated by varying the drying conditions: freeze-drying and fluidized-bed drying led to a more homogeneous Ag loading. An investigation of the fluidized-bed-dried sample by using optical microscopy revealed a large fraction of transparent granules (94 %), which indicates that almost all the Ag nanoparticles in this sample are confined within the 6 nm pores. The optimized supported Ag on silica catalyst shows a good catalytic performance. This adaptation of the drying step can be implemented easily on a laboratory scale, is scalable, and does not require the use of expensive solvents or metal precursors.

Full text of DOI:10.1002/cctc.201700753

Accepted Article
Title: Origin and Abatement of Heterogeneity at the Support Granule
Scale of Silver on Silica Catalysts
Authors: Eva Plessers, Jeroen E. van den Reijen, Petra E. de Jongh,
Krijn P. de Jong, and Maarten B.J. Roeffaers
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.201700753
Link to VoR: http://dx.doi.org/10.1002/cctc.201700753
A Journal of
www.chemcatchem.org

10.1002/cctc.201700753
ChemCatChem
FULL PAPER
Origin and Abatement of Heterogeneity at the Support Granule
Scale of Silver on Silica Catalysts
Eva Plessers,[a] Jeroen E. van den Reijen,[b] Petra E. de Jongh,[b] Krijn P. de Jong,[b] Maarten B.J.
Roeffaers*[a]
Abstract: Incipient wetness impregnation is commonly used to form
supported metal nanoparticle catalysts. Recently, it has been
revealed that this approach may induce severe heterogeneity
between catalyst granules of the same batch. At least 10-fold
variations in metal loading were observed, severely affecting the
catalytic performance of individual catalyst granules. However, the
origin of this heterogeneity is still unclear. Here we show that every
elementary step in the preparation procedure of a silver on silica
catalyst has an effect on the resulting interparticle heterogeneity, but
the influence of the drying step is the most important; as a result of
techniques. The most frequently used characterization methods,
such as X-ray diffraction, chemisorption, EXAFS and UV/Visible-
Diffuse
Reflectance
Spectroscopy,[7–13]
yield
average
nanoparticle sizes. This value reveals little to no information
about nanoparticle distribution and/or dispersion within a single
and between different micrometer- - or millimeter-sized support
granules. Zooming in on the nanoparticles using transmission
electron microscopy (TEM) on the contrary, allows precise
determination of individual supported nanoparticle dimensions
and distribution given that a representative sample of the
powder has been investigated.[1,14,15] Besides electron
microscopy, a wide variety of X-ray based imaging techniques
are developed to study the heterogeneity of solid catalysts.[16–20]
For industrial scale millimeter-sized support bodies it is well
known that non-uniform metal loadings such as ‘egg-shell’, ‘egg-
yolk’ or ‘egg-white’ are more generally obtained than a ‘uniform'
homogeneous distribution.[3,6,21-23] Based on literature reports
and model calculations Morbidelli, Gavriilidis and Varma[21]
conclude that several experimental factors e.g. precursor
solution concentration, pH, ionic strength and temperature as
well as drying conditions cause this heterogeneity during
supported metal catalyst preparation. Also for micrometer-sized
granules, typically used for catalysts discovery at the laboratory
scale, intraparticle variations in nanoparticle size, dispersion and
distribution have been shown. Electron tomography for example
revealed that typical impregnation and drying results in a non-
uniform metal distribution and nanoparticle size whereas
optimization of drying and calcination resulted in more
capillary forces drying results in
a
heterogeneous sample.
Specifically, it is shown that the position of a granule in the stagnant
drying bed largely influences the resulting color and thus silver
loading. This is further demonstrated by varying the drying
conditions: freeze-drying as well as fluidized bed drying lead to a
more homogeneous silver loading. Optical microscopic investigation
of the fluidized bed dried sample revealed a large fraction of
transparent granules (94 %) indicating that almost all silver
nanoparticles in this sample are confined within the 6 nm pores. The
optimized supported silver on silica catalyst shows an improved
catalytic performance. This adaptation of the drying step can be
easily implemented at the laboratory scale, is scalable and does not
require the use of expensive solvents or metal precursors.
Introduction
homogeneous
loading.[14,24-26]
Studies
on
interparticle
Supported
metal
nanoparticles
are
widespread
in
heterogeneous catalysis since they combine both high activity
and stability. These metal nanoparticle catalysts are also
commonly used in catalyst discovery research at laboratory
scale, often prepared via incipient wetness impregnation (IWI). It
is generally known that the preparation method largely
influences the final catalytic performance, as metal nanoparticle
size and distribution inside the support depend on the
preparation method.[1-6] This was the driving force for the
development and application of various characterization
heterogeneities at the support granule level however have to the
best of our knowledge only been reported twice: (1) advanced
electron tomography on platinum loaded sub-micron USY
zeolites[1] (2) a correlative approach using optical microscopy on
silver loaded mesoporous silica granules.[2] In both cases at
least 10-fold variations in metal loading between different
support particles were found. Furthermore, these differences in
metal loading were shown to severely affect the catalytic
performance when measuring at the same granule scale.[2] In
contrast to millimeter-sized support bodies, where transport of
the metal precursor via the impregnation solution during drying
is responsible for the heterogeneous metal loading, the origin of
the heterogeneity during the preparation of supported
nanoparticles at micrometer-sized granules has not yet been
investigated.
[a]
[b]
Dr. E. Plessers, Prof. Dr. Maarten B.J. Roeffaers
Centre for Surface Chemistry and Catalysis
KU Leuven
Celestijnenlaan 200f, 3001 Heverlee, Belgium
E-mail: marten.roeffaers@kuleuven.be
In this work, optical microscopy was used to study the influence
of the different elementary steps on the loading heterogeneity
during the preparation of supported silver on micrometer-sized
silica granules. Typical impregnation procedures and a silver
nitrate precursor solution that displays limited chemical or
J.E. van den Reijen, Prof. Dr. P.E. de Jongh, Prof. Dr. K.P. de Jong
Inorganic Chemistry and Catalysis
Debye Institute for Nanomaterials Science, Utrecht University
Universiteitsweg 99, 3584 CG Utrecht, The Netherlands
Supporting information for this article is given via a link at the end of
the document.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700753
ChemCatChem
FULL PAPER
electrostatic interactions with the support are used. Variations in
silver loading at the granule-level were assessed via the
nanoparticles, whereas an increased coloration was associated
to an elevated silver concentration and the formation of large
Figure 1. Schematic representation of incipient wetness impregnation followed by drying and calcination.
coloration of individual support granules within the resulting
powder. From the conducted experiments it can be concluded
that each of the elementary steps namely impregnation, drying,
calcination and activation induces interparticle heterogeneity at
the granule level. For this particular support and metal
(precursor) system the drying step brings about most of the
heterogeneity. Specifically, this work also shows how fluidized-
bed drying effectively abates most of the heterogeneity in
loading; this optimized drying method can easily be implemented
into the preparation of supported metal catalysts. The optimized
supported silver on silica catalyst also shows an improved
catalytic performance.
silver nanoparticles on the outer surface of the granules.
Although the interparticle heterogeneity in silver loading at the
silica support granule level was reported before,[2] the origin of
this heterogeneity is still unclear. In order to identify the cause,
four possible origins of heterogeneity were evaluated in detail:
(1) variability between support granules in surface charge or
metal impurities; and heterogeneity induced during the (2)
impregnation, (3) drying and (4) calcination (Figure 1). The
influence of each of these parameters on the resulting
interparticle heterogeneity, was assessed based on the color
heterogeneity of the powder obtained after the last calcination
and activation step.
Results
Heterogeneity at the support granule level
First, 5 wt% silver supported on a silica gel (Sigma-Aldrich,
Fluka 60752) with 409 m2/g BET specific surface area and 6 nm
pore diameter was prepared via standard incipient wetness
impregnation (IWI) using a 3.6 mM silver nitrate solution.[3,5,6]
After calcination, the resulting silica supported silver
nanoparticles powder has a typical yellowish and homogeneous
appearance. However, as reported before,[2] optical microscopy
revealed an unexpected color variability with support granules
ranging from transparent over yellow to red. The color at the
support granule level is associated with the silver nanoparticles,
more specifically to the surface plasmon resonance. This
plasmon absorbance is largely determined by the nanoparticle
size, nanoparticle shape and refractive index of the
environment.[27] Increasing the silver nanoparticle diameter
inside a silica matrix from below 10 nm to about 100 nm results
in a red-shift of the surface plasmon resonance absorbance
peak from 420 nm to about 520 nm and changing coloration
from weak yellow to intense red.[8] However, not only size, but
also silver nanoparticle concentration will influence the overall
optical appearance of the support granule. As evidenced before,
via correlation of optical microscopy with high resolution SEM
and EDX, the coloration in our silver on silica sample is related
to both nanoparticle loading and size.[2] Transparent granules
contain a low concentration of pore confined 6 nm silver
Variability in surface charge density Since it is generally
known that the adsorption strength of the metal precursor onto
the
support
plays
an
important
role
during
the
impregnation,[6,21,28] variability in surface charge between support
granules might be the first possible origin of the observed
heterogeneity. For the studied system, silver cations show a
weak adsorption onto the silica support surface because of the
limited surface charge density at the prevailing pH.[21,29] Silica
immersed in neutral water acquires some negative surface
charges of about -0,02 C/m2,[30] predominantly through the
deprotonation of terminal silanol groups.[31] For 1 gram of silica
gel with
a
BET specific surface area of 409 m2/g, this
corresponds to 85 µmol negatively charged O-atoms on the
surface. To prepare a 5 wt% silver on silica material, a solution
with 460 µmol ionic silver is used, meaning that for this specific
system less than 20% of the total amount of silver ions can be
electrostatically adsorbed on the silica surface. To test the
hypothesis, the surface charge density of the support granules
was probed via the ion exchange of a cationic dye Rhodamine
6G. Based on the literature reported silica surface charge
density,[30] the silica gel powder was contacted with an aqueous
Rhodamine 6G solution containing sufficient cationic dye to
compensate all theoretical surface charges (see supporting
information, Figure S2).
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700753
ChemCatChem
FULL PAPER
Figure 2. Silver incipient wetness impregnation on various silica supports: (A) standard silica gel (Aldrich), (B) NaOH pre-activated silica gel, (C) tetraethyl
orthosilicate (TEOS) based self-synthesized silica gel and (D) high purity silica (CARiACT) (5-6 wt% Ag, vacuum drying at RT 16 h, calcination 16 h 500 °C).
Using confocal microscopy with a 405 nm laser, it can be
concluded that there is 5-fold variability in Rhodamine
studies described in this work were therefore always executed
with the commercial silica gel (obtained from Sigma-Aldrich) with
a pore diameter of 6 nm and granule size ranging between 40 -
75 µm.
a
concentration between different support granules. Since this
interparticle heterogeneity in Rhodamine loading is a direct
measure for the surface charge heterogeneity, this can only
partially explain the observed 10-fold heterogeneity in silver
loading between different silica support granules.[2] Although, as
mentioned above, only less than 20 % of the used silver ions at
this loading of 5 wt% can be electrostatically adsorbed on the
silica surface this factor will become important when preparing
low loaded (noble) metal catalysts.
Heterogeneity induced during impregnation
Co-impregnation Since variation in surface charge at the
granule level can partially explain the observed heterogeneity in
silver loading,
a co-impregnant can be introduced in the
The heterogeneity in surface charge could have various origins,
of which differences in silanol density amongst the different silica
gel particles is the most plausible. In an attempt to homogenize
the surface charge, the silica gel was pretreated with an
aqueous 0.1 M NaOH solution followed by extensive washing
with distilled water before impregnation (see supporting
information). This treatment should deprotonate all surface
hydroxyl groups to the same extent. When using the pre-
activated silica support in the regular IWI procedure, the bulk
powder showed a reduced yellow coloration (Figure 2b) which
was confirmed using optical microscopy. Silanol activation
clearly decreases surface charge heterogeneity. Unfortunately,
the base pretreatment results in partial silica dissolution and
fragmentation making a detailed comparison at the single
support particle level, for example via Rhodamine 6G staining,
difficult. However, the experiment does support the idea of
particle-to-particle variations in silanol density.
impregnation solution to compete with the cationic silver for the
negatively charged adsorption sites. This strategy is for example
often used for platinum impregnation of Al2O3 pellets.[4,21] The
effect of co-impregnation with an excess of ammonium nitrate
was investigated. This procedure however did not lead to
complete homogeneity (see supporting information, Figure S3).
Pore filling An important question when performing the
conventional incipient wetness impregnation procedure is if all
pores are effectively filled with the precursor solution. An uneven
pore filling could also cause a difference in metal loading at the
support granule level. A positive indication of complete pore
filling, often used in the laboratory, is the sudden transition of a
seemingly ‘dry’ powder to a moist state upon addition of the last
drop of the precursor solution.[5] Since the correct determination
of the support mesopore volume is an important parameter to
reach this situation, nitrogen physisorption was performed on the
silica gel supports under study (see supporting information).
Using an impregnation solution with a volume that equals the
pore volume determined by N2-physisorption precludes that up
until adding the last drop of the precursor solution, capillary
forces drive the uptake of the solution into the support. However,
careful execution and macroscopic observation of the
impregnation step indicated that vigorous stirring while adding
the precursor solution is very important to maintain the 'dry'
powder state until the very last drop. Macroscopic examination
thus indicated that complete pore filling was reached, since only
at the very last drop all pores are filled and the system changes
to a moist state, and pore filling during impregnation is thus not
the origin of heterogeneity. Furthermore, different techniques are
reported in literature that also confirm that pore filling was
homogeneous after impregnation.[25,32]
Metal impurities Besides intrinsic heterogeneities in surface
charge density, metal impurities (Na, Al, Ti, ...) present in the
silica gel before impregnation also severely influence the overall
surface charge. As silica gel is typically made starting from less
pure water glass solutions,[4] the IWI procedure was repeated
using
a self-synthesized silica gel made from tetraethyl
orthosilicate (TEOS) and a commercially obtained ultra-pure
silica (CARiACT) (see supporting information). The thus
obtained powders showed an even darker coloration (Figure 2c-
d). Zooming in at the support granule level, revealed an even
larger heterogeneity than for the standard silica gel support
showing that the presence of impurities is not at the origin of the
observed heterogeneity. As these different silica materials have
a different surface area, granulate size and pore size, a detailed
one-to-one comparison is not possible. Further impregnation
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700753
ChemCatChem
FULL PAPER
Figure 3. Schematic representation of colour distribution at the catalyst granule level as the result of varying preparation conditions (5-6 wt% Ag, calcination
2 h (top) and 16 h (bottom) 500 °C). Colour heterogeneity (n=250) illustrated with pie diagram, colours represent colour index of individual Ag/SiO2 granules
(determination see Supporting Information and reference 2).
Although impregnation in vacuum conditions is known to be
beneficial for the resulting homogeneity, this is practically difficult
when using micrometer-sized granules.
might occur. To minimize redistribution of metal precursor within
individual support granules and to optimize metal dispersion
alternative drying methods such as freeze-drying or vacuum
drying have been explored.[14,33] Based on these earlier findings,
the influence of vacuum drying and freeze-drying on the
resulting color heterogeneity at the granule scale for our silver
on silica powder was evaluated. From the color distribution
shown in Figure 3c it is clear that vacuum drying has a positive
effect. Not only does it increase the fraction of light yellow
colored granules (from 24 % to 29 %), the fraction of darkly
colored orange to red granules decreases similarly (from 0,5 %
to 0,4 %). An even more homogeneous color distribution at the
granule scale was obtained after freeze-drying (Figure 3d).
Drying bed Clearly, the drying conditions of the freshly
impregnated powder play an important role. Within the static
drying bed temperature gradients as well as concentration
gradients might develop. In order to examine the influence of the
position of a support granule inside the drying bed on the
resulting color and thus silver loading, the powdered bed was
sliced in half after drying and calcined while carefully maintaining
this same powder configuration (Figure 4a-d). The overview
picture taken of the powdered bed after calcination, Figure 4 (e-
h), clearly shows that granules that were present in the interior
of the bed during drying are markedly less colored than granules
that resided closer to the exterior of the bed. This observation
conflicts with the general assumption that during drying metal
precursor precipitation takes place within or at the outside of the
original granule and no migration between granules exist; note
that within one support granule precursor migration has been
observed before.[14,24–26]
Wet impregnation Besides IWI, another frequently used
impregnation method is ‘wet impregnation’ or ‘diffusional
impregnation’ in which the support is wetted with the pure
solvent before impregnation with the precursor solution.[4,6] Wet
impregnation was executed to test whether interparticle
heterogeneity could be decreased. Visual inspection of the
powder obtained via wet-impregnation after the calcination step
however, suggests an even larger heterogeneity because of the
darker yellow color of the obtained bulk powder. As mentioned
above, the intensity of the plasmon absorbance of silver
nanoparticles decreases upon decreasing their size, while an
increasing size results in
a more intense and red-shifted
absorbance. In addition, also silver loading determines the color
of the resulting powder. Close inspection using optical
microscopy indeed reveals the fraction of dark colored orange to
red granules containing excessive amounts of silver and larger
silver nanoparticles is strongly increased (from 0,5 % to 2,6 %).
The complete color distribution among the individual supported
silver granules as observed via optical microscopy is shown in
Figure 3b.
Heterogeneity induced during drying
Alternative
drying
methods Typically, drying after
impregnation is performed in a static oven, however metal
precursor redistribution within and between support granules
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700753
ChemCatChem
FULL PAPER
Fluidized bed drying To prevent directional transport of the
metal precursor between support granules as is observed during
static drying, an alternative to static drying was investigated. de
Jong and co-workers have explored fluidized bed drying
approaches to improve the cobalt distribution within mesoporous
silica when impregnating with Co(NO3)2.[26] The impregnated
powder is loaded into a tubular reactor through which a gas is
flowed from bottom to top, creating a fluidized bed. The effect of
this drying method on the resulting color heterogeneity was
tested by drying freshly prepared silver impregnated silica under
air flow (30 mL/min, 80 % N2 20 % O2) at 100 °C for 8 hours.
Macroscopically, the obtained powder appears more lightly
colored indicating a more homogeneous silver distribution which
was also confirmed by optical transmission microscopy. Figure
5a shows the color distribution in the sample, the fraction of
transparent granules is strongly increased from 74 % to 94 %,
while the average silver loading is maintained.
Both fluidized bed drying and fluidized bed calcination can be at
the origin of the improved homogeneity of the powder.
Subsequently a traditionally dried silver on silica powder (100 °C
static, 16 h) was calcined in similar fluidized bed conditions (air
flow 30 mL/min, 80 % N2 20 % O2, 1 °C/min to 500 °C for 2 h).
The obtained powder strongly resembles that obtained by
following the standard impregnation procedure (Figure 5b and
Figure 2a). The remarkable increase in homogeneity can hence
be attributed to the fluidized bed drying. Note that also under
these conditions, the drying temperature plays a crucial role.[26,33]
For this particular system a drying temperature of 100 °C was
found to give the best results.
Figure 4. Influence of position in the drying bed. After impregnation and
drying (A-B) the powder is sliced (C), during calcination the sliced
configuration is maintained (D), optical examination of the calcined powder
reveals granules present in the exterior of the drying bed (G-H) are
predominantly yellow to orange coloured, while the granules present in the
interior of the bed are mostly transparent (E-F).
between fluidized bed and static calcination of the standard
dried catalysts. However, changing calcination time might affect
the samples’ heterogeneity, since at this stage the final
nanoparticles are formed. To test this hypothesis the calcination
time was increased from 2 h to 16 h. The increase in coloration
of the sample however is significant (Figure 3a-e vs f-j), top
versus bottom row). The prolonged calcination induces thus a
transition from non-colored
6
nm pore confined silver
nanoparticles to larger (colored) silver nanoparticles formed on
the outer surface of the granules. Even though the calcination
temperature, 500 °C, is well below the melting temperature of
silver (962 °C),[34] metal mobility is known to start already at half
the melting point, the so-called Tamman temperature. However,
nanoparticles do generally not migrate but rather diffuse around
their initial position. Since in our sample relatively low loadings
are used resulting in relatively large inter nanoparticle distances,
nanoparticles movement and coalescence is most likely not the
reason for the particle growth. On the other hand, during
calcination in air silver is partially oxidized to silver ions.[35]
These silver ions are highly mobile and can migrate between
silver nanoparticles, leading to particle growth.
Heterogeneity induced during calcination
Calcination time From the experiments discussed above, the
calcination seems not to play a decisive role in the final
coloration of the silica supported silver powder i.e. comparison
Calcination bed The particle growth phenomenon, similar to the
heterogeneities as a consequence during the drying step, might
also be related to the relative position of the granule inside the
powdered bed during the calcination (see supporting
information). To test this hypothesis, well dispersed individual
granules were calcined in an in situ heating stage with minimal
temperature gradients. Optical investigation indicates still color
heterogeneities can be observed (see supporting information,
Figure S4). The color heterogeneities that develop under such
conditions are due to the variability induced by the previous
steps in the IWI procedure (vide supra).
Figure 5. Color distribution in Ag/SiO2 after fluidized bed drying (A) and
static drying (B), followed by calcination in fluidized bed (drying 100 °C 8 –
16 h; calcination 1 °C/min 500 °C – 2 h).
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700753
ChemCatChem
FULL PAPER
Catalysis
support has been extensively studied at the millimeter-sized
support body level.[21] It has been found that the two factors that
determine the final metal loading are (1) convective flow from
the inside of the body towards the outer surface induced by
evaporation and (2) diffusive transport induced by concentration
gradients. Fast drying minimizes such effects preserving
maximally the metal precursor distribution inside the support
body.[21] However, reports have indicated that for weakly
adsorbing metal precursors accumulation of the precursor at the
external surface of the support body is significant even at fast
drying.[36] This results in the so-called egg-shell metal distribu-
tion.[3,6] The final metal distribution in support bodies has thus
been shown to be determined by the relative strength of
convection, diffusion, and adsorption and it can be influenced by
a variety of parameters like viscosity and pH of the solution,
solubility of the precursor, adsorption strength and size of the
support.[3,6,36] To rule out the latter, laboratory scale nanoparticle
catalysts are often supported on micrometer-sized granules in
which it is believed that these phenomena are not important in
contrast to millimeter-sized industrial pellets.[3] The coloration in
function of the position in the drying bed (Figure 4) clearly
refutes this general assumption. Although the convective and
diffusional transport of silver between different granules seems
unlikely, these experiments show that silver nitrate transport
during drying towards the ‘evaporative zone’ happens.
In order to test the impact of the improved homogeneity after
fluidized bed drying on catalysis, the catalytic performance of
this sample was compared to that prepared via the standard IWI
procedure (Table 1). Single support particle catalytic
experiments already revealed
a
large particle-to-particle
variation in catalytic performance in the chemoselective
hydrogenation of 4-nitrostyrene.[2] This reaction was chosen for
the bulk catalytic experiments. The standard and fluidized bed
prepared Ag/SiO2 samples were crushed before reaction to
eliminate mass transport effects, put in the high pressure reactor
together with the substrate (4-nitrostyrene, 2 mol% Ag), internal
standard (tetradecane), solvent (EtOH) and pressurized to 20
bar with H2 after flushing the reactor with N2. After 30 minutes at
140 °C the reaction was stopped and the products were
analyzed with GC. The samples prepared via fluidized bed
drying outperform the standard IWI material.
Table 1. Results of the catalytic hydrogenation of 4-nitrostyrene.
Reaction[a]
t [h]
X [%]
S4-Van [%] [b]
S4-EAn [%] [c]
Standard preparation
Fluidized bed
0,5
0,5
41
53
95
93
2
4
Since granule color is related to silver loading and size of the
silver particles on the granule’s outer surface, a homogeneous
color distribution originates from both an equal silver loading
inside all the individual granules and the formation of only pore
[a] Reaction conditions: 100 °C, 20 bar of H2, 2 mol% Ag, 0,13 M 4-
nitrostyrene in EtOH, 500 rpm. [b] 4-Van = 4-vinylaniline. [c] 4-EAn = 4-
ethylaniline.
confined
6
nm sized silver nanoparticles. The more
homogeneous color distribution as the result of freeze-drying
can be attributed to fixation of the solvent to the granules,
restricting the exchange of solution between two adjacent
granules and is thus a good indication of equal loading.
Additionally, during freeze-drying there is no liquid-gas interface
and hence no capillary effects. Under normal drying conditions a
meniscus is formed, which ensures that upon drying larger pores
are emptied first, while smaller pores remain fully filled with
solution, and the meniscus retracts to the smaller pores, causing
redistribution of the active material. Freezing thus prevents
migration of the silver nitrate during drying, so more of the silver
nanoparticles are located inside the support granule and less
large silver particles are formed on the outer surface. The
beneficial effect of restricting the exchange of solution between
two adjacent granules is further evidenced by the large fraction
of transparent granules (94 %) after fluidized bed drying. The
transparent color does not mean that all these granules have a
very low silver loading, since the same initial silver salt
concentration is used in all impregnated samples. The silver
nanoparticles inside the transparent granules in the fluidized bed
sample are thus all present as pore confined 6 nm sized
nanoparticles. In the wet impregnated sample on the contrary, it
is more likely that larger silver nanoparticles form on the outer
surface giving rise to a higher fraction of orange and red colored
granules (2,6 %). Furthermore, fluidized bed and freeze-drying
randomizes the granules within the drying bed. This granule
randomization also counters the effect of temperature gradients
Discussion
The Rhodamine 6G experiments clearly demonstrate the
presence of interparticle heterogeneity in surface charge
between the individual silica support granules even before
impregnation. This interparticle heterogeneity originates from the
calcination step in the preparation procedure of the dry silica gel.
Temperature gradients from the outside towards the inside of
the wet gel develop; in the outer regions more severe heating
results in a more drastic dehydration and hence loss of surface
hydroxyls. After calcination the solidified gel is crushed into the
micrometer-sized granules used as support for metal
nanoparticles. As the whole individual micrometer-sized support
granule originates from one position within the calcined gel, only
limited intraparticle heterogeneities in surface hydroxyls can be
expected
within
the
granule. Moreover, interparticle
heterogeneities are expected because of the same reason,
since these support granules might originate from different
regions of the original wet and calcined gel. This is supported by
the Rhodamine 6G experiment showing a 5-fold variation in
surface hydroxyl density between individual granules and no
severe variations within the particles.
The experiments focusing on the drying step after impregnation
reveal that drying has a large influence on the resulting color
heterogeneity. The influence of the drying step onto to the
resulting metal dispersion and distribution inside and onto the
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700753
ChemCatChem
FULL PAPER
that develop during traditional drying. The fact that the position
inside the drying bed plays such an important role during the IWI
procedure probably has been overlooked because of
randomization of the granules position between the different
steps i.e. the dried powder is typically transferred to a heat
resistant crucible for calcination, removing the direct link
between final properties and the location. The optimized silver
distribution obtained via the fluidized bed drying also resulted in
an improved catalytic performance due to more efficient use of
the impregnated silver.
techniques.[2,37,38] Furthermore, optical screening could become
a general technique to study catalyst preparation procedures as
it can equally be used for a variety of other metals and metal
oxides that also develop colors like Au, Cu, Pt, Pd and Mn.
Experimental Section
Catalyst Preparation Ag/SiO2 was prepared via standard incipient
wetness impregnation. To 950 mg of dried silica gel (Aldrich 60752) an
aqueous solution of AgNO3 (78 mg AgNO3, 750 µL dist. water) was
added under vigorous stirring. After equilibration at RT (0.5 h), the silver
impregnated silica was dried in static air at 100 °C (16 h) and calcined in
static air (1-5 °C/min 500 °C, 2 h).
Conclusions
Optical microscopy was successfully used as a characterization
tool to improve the incipient wetness impregnation of silver
nanoparticles supported on silica gel. Variations in silver loading
were assessed via the coloration of individual support granules
within the resulting powder. The use of other silica support
materials, wet-impregnation and prolonged calcination in air
induced an even larger interparticle color heterogeneity at the
support granule scale. By changing drying conditions to fluidized
bed drying on the contrary, a nearly homogeneous sample was
obtained. Comparing the resulting granule color and its position
in the static drying and calcination bed revealed the large impact
of temperature gradients and silver nitrate transport across the
bed on the heterogeneity. In fluidized bed drying, the position of
the individual granules is continuously randomized eliminating
this impact. During freeze-drying the exchange of solution
Modifications to the preparation procedure: (1) NaOH pre-treatment: 0.1
M NaOH aqueous solution (5 mL/g silica), 2 h stirred, supernatant
removed via centrifugation (1000 rpm), washed with dist. H2O (7 times,
1000 rpm), dried (80 °C) and calcined; (2) wet impregnation: 1.5 mL dist.
H2O added to silica gel before impregnation, equilibrated 1 h at RT; (3)
vacuum drying: RT, 16 h, 10 mbar; (4) freeze drying: frozen at -20 °C (4
h) and liquid N2 cooled (1 h); freeze drying 4 h 0.3 mbar; (5) fluidized bed
drying: 200 mg impregnated powder is transferred to a tubular quartz
reactor (Ø
1 cm), 2 quartz wool plugs were used to create a
homogeneous flow through the bed and to prevent the powder being
blown out, a controlled flow of 30 mL/min 80% N2 20% O2 was used,
oven was heated to 100 °C (10 °C/min) for 8 h; (6) fluidized bed
calcination: 200 mg dry powder, tubular quartz reactor, 30 mL/min 80%
N2 20% O2, 1 °C/min 500 °C, 2 h.
Characterization Optical microscopic images were obtained via the
eyepieces using an adapter from Micro-Tech-Lab (Austria) to connect a
Canon EOS5D color camera to an Olympus BX51 Upright microscope
between two adjacent granules is restricted leading to
a
homogeneous loading. Additionally, freezing prevents migration
of the silver nitrate to the surface during drying, resulting in more
silver nanoparticles located inside the support granule and
resulting in less large silver particles at the silica’s outer surface.
This is even more pronounced in fluidized bed drying, the large
fraction of transparent granules (94 %) indicates that almost all
the silver nanoparticles in this sample are pore confined 6 nm
sized nanoparticles. The optimized supported silver on silica
catalyst showed an improved catalytic performance supporting
earlier observations made at the single support particle level.[2]
Interestingly, this adaptation to the drying step of the IWI
procedure can be easily implemented at the laboratory scale, is
scalable and does not require the use of expensive solvents or
metal precursors.
with
a standard mercury lamp, equipped with infinity corrected air
objectives 4x (0.16 N.A.) and 20x (0.40 N.A.). Overview images of the
powder bed were obtained with a Leica (M165FC) stereomicroscope.
Color indexing is described in detail in reference 2. Shortly, images were
adjusted to obtain optimal contrast and brightness. The mean 8-bit value
in the red channel was used to increase the sensitivity towards
transparent, light yellow and yellow colored granules. Nitrogen adsorption
and desorption isotherms were measured using a Micromeritics 3Flex
3500 physisorption instrument at 77 K. The sample was degassed before
measurement at 423 K for 6 h under vacuum (10-2 mbar). The pore size
distribution was calculated using the BJH method (Harkins and Jura
thickness curve and Faas correction, 3Flex 3.00 software).
Catalysis Bulk hydrogenation reactions were performed in high pressure
15 mL TOP reactors (30 min., 100 °C, 20 bar of H2, 2 mol% of Ag, 0,13
Future work will involve determination of single particle silver
loading and size and distribution characterization both at the
outer surface and inside the granules requiring additional
characterization techniques like energy-dispersive X-ray
spectroscopy (EDX), scanning electron microscopy (SEM) and
transmission electron microscopy (TEM) combined with 3D-
representation or sectioning. This work demonstrates that such
detailed characterization are best repeated whenever changes
to the IWI procedure are implemented. It can for example be
expected that low loaded catalysts will suffer from the intrinsic
variation in surface charge at the granule level. The power of
optical microscopy is that simple and fast observation can be
easily combined with other more advanced characterization
M
4-nitrostyrene in EtOH, and 500 rpm). Analysis of the reaction
products was carried out via gas chromatography (Shimadzu, CP-Sil 5,
FID detector), and tetradecane was added as internal standard for
quantitative GC analysis. Identification of the compounds was carried out
using GC-MS (Agilent, HP-1, MS).
Acknowledgements
The authors thank the Research Foundation - Flanders (FWO)
(PhD scholarship to E.P.) and Belspo (IAP-VII/05). P. d J. and J.
v R. acknowledge funding from the Netherlands Organization for
Scientific Research (NOW-Vici 16.130.344) and K. P. d J. from
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700753
ChemCatChem
FULL PAPER
[19]
N. Ishiguro, T. Uruga, O. Sekizawa, T. Tsuji, M. Suzuki, N.
Kawamura, M. Mizumaki, K. Nitta, T. Yokoyama, M. Tada,
ChemPhysChem, 2014, 15, 1563–1568.
the European Research Council (EU FP7 ERC Advanced Grant
no. 338846).
[20]
[21]
W. Cha, Y. Liu, H. You, G. B. Stephenson, A. Ulvestad, Adv. Funct.
Mater. 2017, 27, 1700331.
Keywords: Impregnation • drying • catalyst preparation • optical
Morbidelli, A. Gavriilidis, A. Varma, Catalyst Design: Optimal
Distribution of Catalyst in Pellets, Reactors, and Membranes,
Cambridge University Press, 2001.
microscopy • supported metal catalyst
[1]
[2]
J. Zečević, A. M. J. van der Eerden, H. Friedrich, P. E. de Jongh, K.
P. de Jong, ACS Nano 2013, 7, 3698–3705.
[22]
[23]
[24]
T. Vergunst, F. Kapteijn, J. A. Moulijn, Appl. Catal. A Gen. 2001,
213, 179–187.
E. Plessers, I. Stassen, S. P. Sree, K. P. F. Janssen, H. Yuan, J.
Martens, J. Hofkens, D. De Vos, M. B. J. Roeffaers, ACS Catal.
2015, 5, 6690–6695.
X. Liu, J. G. Khinast, B. J. Glasser, Chem. Eng. Sci. 2012, 79, 187–
199.
J. R. Sietsma, H. Friedrich, A. Broersma, M. Versluijs-Helder, A. Jos
van Dillen, P. E. de Jongh, K. P. de Jong, J. Catal. 2008, 260, 227–
235.
[3]
[4]
P. Munnik, P. E. de Jongh, K. P. de Jong, Chem. Rev. 2015, 115,
6687–6718.
G. Ertl, H. Knözinger, F. Schüth, J. Weitkamp, Handbook of
Heterogeneous Catalysis, Wiley-VCH Verlag GmbH & Co. KGaA,
Weinheim, Germany, 2008.
[25]
[26]
T. M. Eggenhuisen, H. Friedrich, F. Nudelman, J. Zecevic, N. A. J.
M. Sommerdijk, P. E. de Jongh, K. P. de Jong, Chem. Mater. 2013,
25, 890–896.
[5]
[6]
[7]
R. J. White, R. Luque, V. L. Budarin, J. H. Clark, D. J. Macquarrie,
Chem. Soc. Rev. 2009, 38, 481–494.
P. Munnik, P. E. de Jongh, K. P. de Jong, J. Am. Chem. Soc. 2014,
136, 7333–7340.
K. P. de Jong, Synthesis of Solid Catalysts, Wiley-VCH, Weinheim,
Germany, 2009.
[27]
[28]
[29]
[30]
Y. Sun, Y. Xia, Analyst 2003, 128, 686.
S.-Y. Lee, R. Aris, Catal. Rev.-Sci.Eng. 1985, 27, 207–340.
J. Park, J. Regalbuto, J. Colloid Interface Sci. 1995, 175, 239–252.
R. Dinnebier, S. Billinge, Powder Diffraction: Theory and Practice,
The Royal Society Of Chemistry, Cambridge, 2008.
[8]
[9]
M. Quinten, Appl. Phys. B Lasers Opt. 2001, 73, 317–326.
M. Kosmulski, Chemical Properties of Material Surfaces, Marcel
Dekker, Inc, New York Basel, 2001.
J. Lyu, J. Wang, C. Lu, Q. Zhang, X. He, X. Li, J. Phys. Chem. C
2014, 118, 2594–2601.
[31]
[32]
S. H. Behrens, D. G. Grier, J. Chem. Phys. 2001, 115.
[10]
K. Arve, H. Kannisto, H. H. Ingelsten, K. Eränen, M. Skoglundh, D.
T. M. Eggenhuisen, M. J. Van Steenbergen, H. Talsma, P. E. de
Jongh, K. P. de Jong, J. Phys. Chem. C 2009, 113, 16785–16791.
Y. Murzin, Catal. Letters 2011, 141, 665–669.
[11]
[12]
G. Corro, U. Pal, E. Ayala, E. Vidal, Catal. Today 2013, 212, 63–69.
[33]
[34]
P. Munnik, N. Krans, P. De Jongh, K. De Jong, ACS Catal. 2014, 4,
3219–3226.
H. Wei, C. Gomez, J. Liu, N. Guo, T. Wu, R. Lobo, L. Christopher, R.
Lobo-Lapidus, C. L. Marshall, J. T. Miller, et al., J. Catal. 2013, 298,
18–26.
A. Brumby, P. Braumann, K. Zimmermann, F. Van Den Broeck, T.
Vandevelde, D. Goia, H. Renner, G. Schlamp, W. Weise, P. T.
Tews, Ullmann’s Encycl. Ind. Chem. 2012, vol. 33 p 15–94.
[13]
[14]
K. Shimizu, Y. Miyamoto, A. Satsuma, J. Catal. 2010, 270, 86–94.
[35]
[36]
[37]
J. Hu, W. Cai, H. Zeng, C. Li, F. Sun, J. Phys. Condens. Matter
2006, 18, 5415–5423.
G. Prieto, J. Zečević, H. Friedrich, K. P. de Jong, P. E. de Jongh,
Nat. Mater. 2013, 12, 34–9.
A. Lekhal, B. J. Glasser, J. G. Khinast, Chem. Eng. Sci. 2001, 56,
4473–4487.
[15]
J. Zecevic, G. Vanbutsele, K. P. de Jong, J. A. Martens, Nature
2015, 528, 245–248.
N. Liv, C. A. Zonnevylle, A. C. Narvaez, A. P. J. Effting, P. W.
Voorneveld, M. S. Lucas, J. C. Hardwick, R. a Wepf, P. Kruit, J. P.
Hoogenboom, PLoS One 2013, 8.
[16]
[17]
B. M. Weckhuysen, Angew. Chem. Int. Ed. Engl. 2009, 48, 4910–43.
S. Kalirai, U. Boesenberg, G. Falkenberg, F. Meirer, B. M.
Weckhuysen, ChemCatChem, 2015, 7, 3674–3682.
[38]
E. Debroye, J. Van Loon, X. Gu, T. Franklin, J. Hofkens, K. P. F.
Janssen, M. B. J. Roeffaers, Part. Part. Syst. Charact. 2016, 1–7.
[18]
H. Matsui, N. Ishiguro, K. Enomoto, O. Sekizawa, T. Uruga, M. Tada,
Angew. Chem. Int. Ed. 2016, 55, 12022–12025.
This article is protected by copyright. All rights reserved.

10.1002/cctc.201700753
ChemCatChem
FULL PAPER
FULL PAPER
Eva Plessers, Jeroen E. van den Reijen,
Petra E. de Jongh, Krijn P. de Jong,
Maarten B.J. Roeffaers*
Page No. – Page No.
Origin and Abatement of
Heterogeneity at the Support Granule
Scale of Silver on Silica Catalysts
This article is protected by copyright. All rights reserved.