Catalytic performance of bulk and colloidal Co/Al layered double hydroxide with Au nanoparticles in aerobic olefin oxidation

Article abstract of DOI:10.1016/j.apcata.2019.117155

A Co/Al layered double hydroxide material was synthesized in both bulk and exfoliated (colloidal) forms. Anion exchange with methionine allowed immobilization of Au nanoparticles previously prepared by a biomimetic method using an anti-oxidant tea aqueous extract to reduce the Au salt solution. The catalytic performance of bulk and exfoliated clays Au-hybrid materials was assessed in aerobic olefin epoxidation. Both catalysts were very active towards the epoxide products and with very interesting substrate conversion levels after 80 h reaction time. The Au-exfoliated material, where the nanosheets work as large ligands, yielded higher product stereoselectivity in the case of limonene epoxidation. This arises from a confined environment around the Au nanoparticles wrapped by the clay nanosheets modulating access to the catalytic active centres by reagents. Mechanistic assessment was also accomplished for styrene oxidation by DFT methods.

Full text of DOI:10.1016/j.apcata.2019.117155

Accepted Manuscript
Title: Catalytic performance of bulk and colloidal Co/Al
layered double hydroxide with Au nanoparticles in aerobic
olefin oxidation
´
Authors: Sonia R. Leandro, Cristina I. Fernandes, Ana S.
Viana, Ana C. Mourato, Pedro D. Vaz, Carla D. Nunes
PII:
DOI:
Article Number:
S0926-860X(19)30310-2
https://doi.org/10.1016/j.apcata.2019.117155
117155
Reference:
APCATA 117155
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
15 May 2019
16 July 2019
17 July 2019
Please cite this article as: Leandro SR, Fernandes CI, Viana AS, Mourato AC, Vaz PD,
Nunes CD, Catalytic performance of bulk and colloidal Co/Al layered double hydroxide
with Au nanoparticles in aerobic olefin oxidation, Applied Catalysis A, General (2019),
https://doi.org/10.1016/j.apcata.2019.117155
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.

Catalytic performance of bulk and colloidal Co/Al layered double
hydroxide with Au nanoparticles in aerobic olefin oxidation
Sónia R. Leandro,a Cristina I. Fernandes,a Ana S. Viana,a,b Ana C. Mourato,a Pedro D. Vaz,c
Carla D. Nunesa,b*cmnunes@ciencias.ulisboa.pt
a
Centro de Química e Bioquímica, Faculdade de Ciências da Universidade de Lisboa, 1749-016
Lisboa, Portugal
b
Centro de Química Estrutural, Faculdade de Ciências e Instituto Superior Técnico, Universidade
de Lisboa, 1049-001 Lisboa, Portugal
c
CICECO – Aveiro Institute of Materials, Department of Chemistry, University of Aveiro, 3810-
193 Aveiro, Portugal
*Correspondence to: Dr. Carla D. Nunes Centro de Química e Bioquímica, Departmento de
Química e Bioquímica, Faculdade de Ciências da Universidade de Lisboa, Campo Grande, Ed. C8,
1749-016 Lisboa, Portugal Centro de Química Estrutural, Faculdade de Ciências e Instituto
Superior Técnico, Universidade de Lisboa, 1049-001 Lisboa, Portugal Tel: (+351) 217 500 216
Fax: (+351) 217 500 088
Graphical abstract
Co/Al hydrotalcite-Au nanoparticle catalysts show versatility for aerobic oxidation with high
(stereo)selectivity.
1

Highlights
 Biomimetic Au nanoparticle synthesis yields active oxidation catalysts
 Exfoliated clays were very selective in aerobic olefin epoxidation
 Exfoliated catalysts showed higher stereoselectivity than bulk ones
 Exfoliated catalysts with higher catalytic potential than the bulk-based ones
2

Abstract
A Co/Al layered double hydroxide material was synthesized in both bulk and exfoliated (colloidal)
forms. Anion exchange with methionine allowed immobilization of Au nanoparticles previously
prepared by a biomimetic method using an anti-oxidant tea aqueous extract to reduce the Au salt
solution. The catalytic performance of bulk and exfoliated clays Au-hybrid materials was assessed
in aerobic olefin epoxidation. Both catalysts were very active towards the epoxide products and
with very interesting substrate conversion levels after 80 h reaction time. The Au-exfoliated
material, where the nanosheets work as large ligands, yielded higher product stereoselectivity in the
case of limonene epoxidation. This arises from a confined environment around the Au nanoparticles
wrapped by the clay nanosheets modulating access to the catalytic active centres by reagents.
Mechanistic assessment was also accomplished for styrene oxidation by DFT methods.
Keywords
Gold nanoparticles; Amino acids; Clays; Aerobic oxidation catalysis;
3

1. Introduction
Current environmental concerns are focused on the development of cheap catalysts leading to
highly selective materials and therefore reducing by-products [1-3]. Such concerns aim at a lower
consumption of natural resources as well as benign and impactless reaction media.
A large fraction of catalytic processes makes use of precious metals (second- and third-row
transition metals). While still much unrivaled in terms of catalytic performance, these metals suffer
from many drawbacks impacting negatively both industry and society [4].
On the other hand, Nature sustains its activity and offers an incredible range of materials relying on
metals (Mg, Ca, Mn, Fe, Co, Ni, Cu, Zn, Al) that are typically cheaper, less toxic, and more benign
than their heavier counterparts [4]. Exploring this rapidly, expanding new frontier in a field that is
of importance for practical reasons (cost and availability, environmental impact, toxicity), will aim
at the discovery of fundamentally new reactivity that can offer solutions to a broader range of
catalytic problems. Understanding better the upgrading processes and developing viable alternatives
will contribute to achieve new frontiers in catalysis [5].
To address the above concerns, clays and particularly layered double hydroxides (LDHs) are among
the inorganic materials that have experienced a high development in recent years, owing to their
applications in many fields, namely heterogeneous catalysis [6]. These materials are known for their
ion exchange capacity, making possible the introduction of adequately functionalized guests. The
importance of these materials is based on their ability to retain chemical species with electrical
charges compatible to those of the layers. Concerning catalytic applications, the use of LDHs can
still go much further in this field. The versatility of these materials goes beyond their ion exchange
capacity, with tunable composition and form (powder, colloid, gel) [7-10]. The use of
delaminated/exfoliated LDHs has been reported with several advantages, especially at the level of
asymmetric synthesis [11]. This process makes possible to assess the LDHs as colloidal nanosheets
that transform such materials in nanostructured catalysts. More recently, LDHs have been
engineered as nanoparticles, as monoliths with hierarchical porosity, or as gelated nanoparticles
with high concentration rendering in both cases such materials further new applications [9,10,12].
Increasing research in recent years led to new catalysts capable of carrying out reactions at lower
temperatures and pressures with higher selectivity towards the desired products. Fostering new
frontiers in reactivity and product selectivity this work aimed at exploring the chemistry of LDHs
concerning its forms (bulk powder and delaminated) [7,10,11,13].
Concerning catalytic applications, the use of LDHs can still go much further in this field [14-22].
With this goal in mind we aim to develop new catalysts or catalyst forms towards more selective
and efficient processes. In this work we present a study on the versatility of LDH-based catalysts
4

with Au nanoparticles. Au is known for its catalytic versatility in a variety of reactions. Although
most of its applications is usually in aerobic oxidation of organic compounds (alkanes, olefins or
alcohols) [23] the truth is that Au nanoparticles have found much more applications, including
activation of oxygenated compounds, hydrogen transfer reduction or C–C coupling reactions [24].
In this work the versatility of these catalysts was explored under a multidimensional perspective.
While LDHs can be prepared under different forms (bulk, exfoliated or gelated) we also explored
their activity for their use as oxidation catalyst supports for Au nanoparticles, by tuning the reaction
conditions and substrates. This work reports the use of Co/Al LDH materials that builds up on
knowledge from previous research using Mg/Al LDH materials [13] to assess whether the
composition of the host material influences catalytic performance. The outcome of this research is
presented, discussed and benchmarked with other literature reports.
2. Experimental
2.1. General
All reagents and solvents were of analytical grade and bought from Sigma-Aldrich. In this work we
have used L-methionine as the intercalating ligand and will be denoted as met. HTCo/Al-CO3,
HTCo/Al-Cl, and HTCo/Al-met materials were synthesized by adopting a literature methodology [7].
ExHTCo/Al-met was exfoliated in formamide according to literature methods [7,13]. Au nanoparticles
(AuNP) were prepared adopting a biomimetic methodology described in the literature [25]. Detailed
procedures can be found in the SI material.
2.2. Catalysts Preparation
HTCo/Al-met-Au
AuNP (17 mg) were added to a suspension of HTCo/Al-met (700 mg) in deionized type II water. The
resulting slurry was stirred under a N2 atmosphere, at room temperature (303 K) for 12 h. The
resulting materials were then filtered off, washed with deionized water (3x20 mL), and dried in a
vacuum oven (313 K) for 48 h.
ExHTCo/Al-met-Au
The ExHTCo/Al-met exfoliated material was used to prepare hybrid-sheet nanostructures with AuNP.
A given amount of the colloidal nanosheet material (15 ml, 0.10 mmolmet) was mixed with 4 mg of
AuNP (0.02 mmolAu). The mixture was sonicated for 1 minute and then left stirring for 12 h at room
temperature (303 K). The resulting colloid, denoted ExHTCo/Al-met-Au, had a different color
(burgundy) due to the presence of the AuNP and was found to be stable for a long time as well.
5

2.3. Catalytic epoxidation of olefins
Catalytic tests were carried out in a Carousel 12 Plus Reaction Station from Radleys. A given
amount (1.6 mmol) of the olefin substrates – cis-cyclooctene, styrene, 1-octene, trans-2-hexen-1-ol,
R-(+)-limonene – was mixed with dibutylether (1.6 mmol) (internal standard) and 15 L (5 mol %)
of tert-butyl hydroperoxide (5.5 M solution in decane; tbhp). Then the catalyst was added to this
mixture. This was either 10 mL of exfoliated materials ExHTCo/Al-met-Au (0.016 mol Au) with an
extra 5 mL of acetonitrile or 100 mg of HTCo/Al-met-Au (0.016 mol Au) bulk catalyst with 8 mL of
toluene. The magnetically stirred mixture was heated to 353 K and kept for 80 h.
Substrate conversion and product yields were monitored by sampling at regular time intervals and
analyzing it using a Shimadzu QP2100-Plus GC/MS system fitted with a capillary column
(Teknokroma TRB-5MS or TRB-1MS).
All reactions started with tbhp acting as radical initiator with O2 from air at atmospheric pressure
being the oxidant. Control experiments were run in the presence of (a) N2 inert atmosphere, and (b)
in the presence of a radical scavenger (2,6-di-tert-butyl-4-methylphenol). This allowed confirming
that O2 was the oxidant and the existence of a radical reaction mechanism. The heterogeneous
nature of the HTCo/Al-met-Au catalyst was evaluated by analyzing its Au content by ICP-OES at the
end of an oxidation reaction (cis-cyclooctene as substrate) and comparing it with the Au loading of
the fresh catalyst.
3. Results and discussion
3.1. Catalysts characterization
Layered double hydroxides (LDH) are very versatile materials that can be prepared with organic
anions by three major approaches: (i) anion exchange, (ii) coprecipitation and (iii) reconstruction
[6]. In all these methods, when preparing LDHs with anions other than carbonate, contamination
from CO2 must be prevented, since the carbonate anion is readily incorporated and persistently held
in the interlayer [6]. Therefore, decarbonated and deionized water with an inert atmosphere must be
used to avoid exposure of the reacting material to air.
In this work we have used the coprecipitation method to prepare Co/Al LDH (denoted as HTCo/Al
clay-based hybrid materials with L-methionine (met) as evidenced in Scheme 1 [7,13].
)
The synthesis provides the carbonate clay (HTCo/Al-CO3), which is then ion-exchanged with
chloride anions to yield HTCo/Al-Cl. The latter is easier to undergo further ion exchange (due to
lower charge density of the Cl– anion) to prepare organic-LDH materials. Introduction of met
amino acid by ion exchanging the Cl– anions was carried out by a literature protocol [13], yielding
6

HTCo/Al-met material. Based on the S-content from elemental analysis, the amino acid loading was
1.82 mmol·g–1. According to the empirical formula expected for the HTCo/Al material –
[Co6Al2(OH)16](Cl)2·4H2O – the intercalated HTCo/Al-met clay can be rationalized as
[Co6Al2(OH)16](met)1.11(Cl)0.89·4.6H2O with the water content being estimated from
thermogravimetric analysis, as will be discussed later in this work.
Reaction of HTCo/Al-met with gold nanoparticles (AuNP) afforded the new HTCo/Al-met-Au hybrid
material and the Au loading was found to be 1.16 wt.-% corresponding to 5.9×10-2 mmolAu·g–1 for
this material.
The amino acid containing clay HTCo/Al-met synthesized previously was also subjected to an
exfoliation procedure to prepare 2D mixed-metal inorganic nanosheet materials modified with the
L-containing amino-acid. This step yielded a deep pink colloidal formamide solution of the clay
denoted as ExHTCo/Al-met. The choice of this amino acid to conduct exfoliation was found to be
quite adequate given that the colloids were found to be stable for at least 24 months. This matches
other reports found in the literature addressing the exfoliation of LDH materials with larger anions
(such as sodium dodecyl sulfate) [6,7].
[Scheme 1 here]
Thermogravimetric analysis (TGA) also confirmed the intercalation of met in the interlayer spacing
of the HT LDH material. According to Figure 1, met abruptly decomposes at temperatures in the
500 – 650 K range. After this range, decomposition carries till ca. 3.2% at 1073 K, but at slower
pace.
In the case of HTCo/Al-met its thermal profile evidences a mass loss of 12% till ca. 493 K, which
may be due to water (both adsorbed and intercalated). The material then further decomposes in a
two-step way till 1073K, accounting for a total of 27% of the initial weight. Assuming this to arise
exclusively from decomposition of methionine, this yields a loading of 1.82 mmol·g–1 in the
HTCo/Al-met material. This shows an excellent agreement with results from elemental analysis
(discussed above). Caution should however prevail, as at higher temperature some dehydroxilation
is likely to occur and therefore the real content of methionine may be slightly lower [13]. The
decomposition profile should be commented as well; in this case the profile is different from that
observed for neat met, which can be a proof that met is indeed intercalated within the interlayer
spacing of the HT host and is therefore protected against thermal decomposition due to specific
host-guest interactions [26].
7

Electronic spectroscopy (using both absorption and diffuse reflectance techniques) was used to
assess formation on the Au-containing hybrid materials. For both materials containing the AuNP,
HTCo/Al-met-Au and ExHTCo/Al-met-Au, the measured spectra are shown in Figure 2.
The measured spectra show that both HTCo/Al-met-Au and ExHTCo/Al-met-Au materials displayed
the surface plasmon resonance band at wavelengths similar to those observed for the as-synthesized
gold nanoparticles (AuNP), indicating the successful formation of the hybrid materials [13]. In
addition, changes at the measured absorbance maxima of the hybrid materials also evidence that
there may be interaction between the AuNP and the host HTCo/Al-met and ExHTCo/Al-met materials
[13]. This is most evident for the absorbance maxima of the exfoliated ExHTCo/Al-met-Au material,
where not only the surface plasmon resonance band shifted (to 558 nm) but also the methionine
bands experienced shifts. In this case the bands from methionine (631 and 671 nm for the neat
compound) shifted to 610 and 656 nm, respectively, indicating that the interaction between AuNP
and methionine is stronger than that found for the bulk HTCo/Al-met-Au material, where the shifts
were smaller.
The morphology and size of the exfoliated nanosheets in ExHTCo/Al-met were examined by TEM, as
shown in Figure 3.
As shown in Figure 3a for ExHTCo/Al-met, the TEM image of this material shows an assembly of
thin sheets with homogeneous contrast, reflecting their ultrathin nature and uniform thickness
[7,13], also supported by AFM as discussed in the following lines. The diffraction pattern (inset in
Figure 3a) shows a hexagonal arrangement agreeing with that expected for these materials [7]. On
the other hand, Figure 3b shows the interaction between AuNP and the exfoliated sheets the success
of the assembly in ExHTCo/Al-met-Au. This image shows clearly the presence of the LDH exfoliated
thin sheets of ExHTCo/Al-met (grey areas) along with the geometric shapes (dark spots with pyramid,
hexagon, etc… shapes) typical from AuNP structures.
AFM imaging of exfoliated nanosheets was also carried out, as illustrated in Figure 4, to have
additional information on their morphology and thickness before and after interaction with AuNP.
Typically, larger EXHTCo/Al-met nanosheets exhibited a thickness of ca. 4 nm (Figure 4a pointing to
the presence of multilamellar nanosheets [7]. Upon the formation of the hybrid materials ExHTCo/Al
-
met-Au (Figure 4b), individual gold nanoparticles and small nanoclusters are clearly depicted over
the clay nanosheets, conferring a more heterogenous morphology. Nevertheless, the ExHTCo/Al-met-
Au nanosheets profile reveals that overall thickness, comparing profile data from both ExHTCo/Al
-
8

met and ExHTCo/Al-met-Au materials was maintained (ca. 4 nm in Figures 4a and 4b), highlighting
the stability of methionine interlayer.
Figure 5 shows the powder XRD patterns of the nanohybrid materials HTCo/Al-met and HTCo/Al
-
met-Au. These materials display a series of well-developed Bragg reflections in the 2 range 4-
70º, which can be indexed to the hexagonal layered LDH structure with rhombohedral symmetry
[27,28]. This also agrees with the electron diffraction pattern from the TEM measurements
(discussed above). No peaks of impurities were discerned, indicating the high purity of the product.
Several intense peaks due to basal reflections (00l) at low 2θ angles allowed us to estimate directly
the normal basal spacing to the plane (001) which is equal to the thickness of one lamella of the
brucite structure (ca. 4.8 Å) plus the thickness of the interlayer region [29]. To calculate this the
second 2θ value was used corresponding to plane (003). In the case of the HTCo/Al-met parent
material the 003 reflection was located at a 2 value of 5.8°, corresponding to an interlayer spacing
of 10.4 Å, after subtracting the brucite layer thickness (ca. 4.8 Å). This indicates that the met amino
acid adopted a tilted vertical arrangement between the lamella [30,31]. Caution should prevail in
this analysis as the interlayer arrangement depends on many other factors (e.g. hydration level and
temperature conditions), which were not accounted for in this estimation.
In addition, peaks due to the presence of the AuNP structures were identified at 2θ = 38.2°, 44.4°
and 64.6° (Figure 3, arrow markers). The average size of the AuNP structures according to the
Debye-Scherrer equation and based on data from the peak at 2θ = 38.2° was estimated to be 31 nm.
This value agrees well with the size found for those structures based on the AFM measurements, 40
nm, and TEM images (average size found was 29 nm). Such size magnitude would render the AuNP
structures as inactive for catalytic applications. However, according to Figure 3 (TEM images), a
closer look at the bottom image shows that there is a size distribution for the AuNP structures. Sizes
varied between ca. 4 and 70 nm with ca. 56% of the nanoparticles to fall between 4 and 24 nm,
according to the method we followed [32]. Such size distribution agrees with the observed color,
rendering potential catalytic activity (which is known to be favored for smaller nanoparticles) as
will be discussed later in this work. In this case, we observed that the rough control of the size and
shape of the nanoparticles is advantageous for the observed catalytic activity. Given that the scope
of the manuscript is also on taking advantage on the biomimetic synthesis method of the
nanoparticles, we think this is a feature of this approach, although we are aware that the Au
nanoparticles size is far from optimum.
9

Diffuse reflectance infrared spectroscopy (DRIFT) was used to characterize the hybrid materials,
shown in Figure 6.
The FTIR spectra of HTCo/Al-Cl, HTCo/Al-met and HTCo/Al-met-Au (Figure 6a) show similar
profiles. They present strong bands at 1360-1400 cm-1 and 810 cm-1, corresponding to the stretching
2-
and deformation modes of CO3 . All materials showed a band centered at ca. 3440 cm-1 being
assigned to the νO-H mode of hydroxyl groups, while the presence of a shoulder at ca. 3150 cm-1
suggests that water molecules are interacting with other species by hydrogen bonding. The
remaining bands observed below 800 cm-1 may be associated with bonds between metals and
oxygen atoms in the structure of the clay [13]. According to the spectrum of the material containing
the methionine amino acid, HTCo/Al-met, it was possible to assign bands to the νC–H, νC–O and
νC=O modes of the amino acid moiety (Figure 6b).
Bands corresponding to νN-H vibrational modes, expected in the 3300-3500 cm-1 range could not
be observed due to overlapping. By the same token, νC–N modes, expected between 1020-1220 cm-
1 were not observed due to overlap of the νC–O modes. The characteristic band of the binding of the
metals of the lamella hydroxyl, νM–O modes, appears in the form of a broad band around 600 cm-1.
3.2. Catalytic studies on epoxidation of olefins
The synthesized nanohybrid clay materials HTCo/Al-met-Au and ExHTCo/Al-met-Au were tested as
catalysts for epoxidation of olefins. The test-set comprised cis-cyclooctene, 1-octene, trans-hex-2-
en-1-ol, R-(+)-limonene and styrene as substrates. All reactions were carried out using tert-butyl
hypdroperoxide (tbhp in decane) as the initiator (5 mol%), under aerobic conditions (molecular O2
as the oxidant) and run at 353 K, with results gathered in Table 1. Control runs (without catalyst and
in the presence of tbhp and O2) gave virtually no conversion of starting material at 353 K. The
materials without Au, HTCo/Al-met and ExHTCo/Al-met, were also tested for their intrinsic catalytic
activity and to assess their influence in the results. The results showed that both materials had a
catalytic activity with average cyclooctene conversion below 5% after 80 h.
The oxidation of cis-cyclooctene using HTCo/Al-met-Au and ExHTCo/Al-met-Au materials as
catalysts (Table 1) displayed very good to excellent substrate conversion after 80 h of reaction
(lowest value was 74% for ExHTCo/Al-met-Au). The observed selectivity values towards the epoxide
followed the same trend (above 90%), but not “100%”, which is expected for radical-based
mechanisms due to side-reactions, as recently described by us [13]. The by-products found in the
reaction were both the enone and enol being detected with selectivity below 5% (Table 1, entries 1
10

and 5). The results obtained for cyclooctene are in line with those reported for the same substrate
using Au or Ag nanoparticles [33,34].Bawaked and coworkers, reported that cyclooctene oxide was
obtained with ca. 80% selectivity, while Chen described that cyclooctene was oxidized to its
epoxide with a selectivity of 97%, being in both cases similar to that reported in the present work
(above 90%).
To confirm the radical nature of the reaction we conducted cis-cyclooctene oxidation in the
presence of a radical trap (2,6-di-tert-butyl-4-methylphenol). In this case we observed very low
substrate conversion (< 5%) being consistent with the fact that free radicals are active species in the
catalytic cycle, agreeing with other reports for related systems [13,33,34]. The heterogeneous nature
of the HTCo/Al-met-Au catalyst concerning gold nanoparticle’s (AuNP) leaching was also evaluated.
After a cis-cyclooctene oxidation experiment finished, the catalyst was separated by filtration from
the reaction slurry and analyzed by ICP. Results showed that the Au loading was found to be 1.07
wt.-%. Compared to the Au loading in the fresh catalyst (1.16 wt.-%) showing a leaching of 8%,
confirming that the HTCo/Al-met-Au catalyst is truly heterogeneous.
Both catalysts were then further studied for their activity using other substrates such as 1-octene,
trans-hex-2-en-1-ol and R-(+)-limonene. This set of olefins yielded different results from the
previous one, with the oxidation of 1-octene being the lowest performing system. 1-octene is an
unbranched terminal olefin presenting low reactivity, similarly to other reports found in the
literature [13]. According to Table 1, HTCo/Al-met-Au catalyst made possible to achieve 32%
substrate conversion, while the exfoliated counterpart did not show any activity at all with this
substrate (Table 1, entry 6). For the case of the bulk catalyst the major product was found to be 2-
Hexyloxirane with 85% selectivity with octan-2-one being the by-product. Oxidation of the trans-
hex-2-en-1-ol allylic alcohol evidenced lower substrate conversion when the exfoliated catalyst was
used, compared to the performance of the bulk counterpart (Table 1, entries 3 and 7). In addition,
striking differences on product selectivity were found between both catalytic systems. While the
bulk HTCo/Al-met-Au catalyst showed an epoxide selectivity of 97% at the end of the reaction (80
h), the exfoliated one yielded the same product with a selectivity of only 64%. In both cases, the by-
product formed was the ,-unsaturated aldehyde (Table 1, entries 4 and 8). Formation of the ,-
unsaturated aldehyde has been reported in the literature for a set of allylic alcohols using AuNP
supported on CeO2 [35]. However, authors also noticed that when the catalyst changed to mixed
Au(Pd) nanoparticles supported on TiO2, the aldehyde was not the major product due to competing
catalytic processes. In the present case, we did not observe such high activity as in that case and the
product selectivity was different with the epoxide being the major product.
11

The oxidation of R-(+)-limonene followed a different trend from that observed for the allylic
alcohol described above. As shown in Figure 7, both catalysts achieved a very good level of
substrate conversion (above 74% after 80 h, according to Table 1). In addition, product selectivity
was found to follow the same trend for both catalysts, with carvone (the enone product) being the
major product (Table 2 and Figure 7). Moreover, for both catalysts the minor product was found to
be the endocyclic epoxide and carveol (the enol product) was never detected.
Despite this, the most striking feature was the cis/trans diastereomeric ratio evidenced by both
catalyst, which differed completely (Figure 7)
That was an outstanding feature distinguishing it from the bulk counterpart. According to Figure 7,
while the bulk catalyst yielded the epoxide without a clear preference for any of the diastereomers,
the exfoliated one originated mostly the cis diastereomer of the endocyclic epoxide (Table 1, entry
8) with a trend for slight increase with time (Figure 7c). That was a remarkable achievement
following the same trend as described in the literature [11,13]. As recently demonstrated by us, the
main reason for this, may arise from the fact that the LDH nanosheets with the chiral amino-acids
work as large ligands around the AuNP. This consequently constrains the way reactants interact with
active sites [11,13]. The higher cis/trans diasteriomeric ratio observed with the exfoliated catalyst
(Figure 7c) also demonstrates that in this case the same effect was observed concerning the
stereocontrol of the process by constraining how reactants reach the active sites [11,13].
This can be explained by the structural differences of both catalysts. According to Scheme 2, while
in the bulk catalyst the AuNP structures are outside the LDH particles, in the bulk they are wrapped
by the exfoliated LDH nanosheets. This results in the fact that the bulk catalyst provided better
access of the substrates to the active centers, whereas for the exfoliated one there must be some
diffusion process of substrates through the exfoliated nanosheets till the active centers are reached.
The outcome was the observation of higher substrate conversion (except for limonene) for the bulk
catalyst. Additionally, this diffusion feature of the exfoliated catalyst is also responsible for the
observed enhanced diastereaselectivity in limonene oxidation (as evidenced in Figure 7), since the
substrate approach is constrained as already discussed above.
As mentioned earlier in this work, the observed selectivity values for epoxide products were
generally high. It could however be anticipated, as it would be expected for radical-based
mechanisms due to side-reactions being supported by mechanistic proposals found in the literature.
12

We have tested another substrate – styrene – which differs from the previously described since it
lacks H-atoms in -position to the olefin moiety and, therefore, it is likely to follow a slightly
different mechanism [13,33,36].
Oxidation of styrene usually yields a set of economically valuable products: the epoxide,
benzaldehyde and benzoic acid. While the latter was never observed in the tests carried out using
any of the catalysts, the remaining two were found to be produced in different ratios depending on
the catalyst used. According to Table 2, both catalysts performed quite well, with the bulk one
being more active concerning substrate conversion.
The reaction kinetics on the substrate conversion was found to be overlapping for the first 8 h of
reaction without distinguishing features for any of the catalysts, as evidenced in Figure 8. Another
interesting feature was the fact that when the reactions were run with catalyst but under N2
atmosphere slower kinetics were observed confirming O2 as a key-factor to feed the process, as
evidenced in Figure 8a. Concerning product selectivity, benzaldehyde was found to be the major
product when the HTCo/Al-met-Au bulk catalyst was used. On the other hand, when ExHTCo/Al-met-
Au was the catalyst, styrene oxide was the main product formed. Another interesting feature was
the fact that acetophenone was detected (in very low yield) when the bulk catalyst was used, but
surprisingly it was not detected when using the exfoliated catalyst. All these facts are shown in
Figure 8. This product selectivity tuning between both catalysts may arise from different surface
characteristics. According to Scheme 2, there are expected differences for the AuNP structures,
which are more exposed to the reaction medium in the HTCo/Al-met-Au bulk catalyst and not so
much in the ExHTCo/Al-met-Au one. This will result in different interaction levels between the
reactive species and surfaces, which may result in product selectivity tuning [37].
The presence of phenylacetaldehyde was reported in the literature, but not observed in this study
[13,33,34]. Acetophenone (Acph, Scheme 3) was detected instead (Table 2, entry 2). According to
the literature, phenylacetaldehyde may arise from an isomerization of the epoxide after opening the
oxyrane ring. In the present work, we believe that acetophenone may be formed by the same
process with the oxyrane ring opening in the “opposite direction” to that required to form
phenylacetaldehyde (Phac), as shown in Scheme 3. Oxidation of styrene to acetophenone has been
reported in the literature [38]. When using Au nanoparticles, benzaldehyde was the major product,
but when Pd was used in combination with Au, acetophenone was the major product [38].
13

The above proposal demonstrates how the radical mechanism may progress as styrene lacks the
presence of H-atoms in -position to the olefin moiety, which would make impossible to rationalize
H-atom abstraction [34,36]. Alternatively, styrene recombines directly with tbhp radicals yielding a
couple of intermediaries (1 and 2 in Scheme 3). These will then decompose to yield either styrene
oxide or benzaldehyde, depending on the path followed (Scheme 3). The t-butyl derivative radicals
will recombine (or not) with O2 closing and feeding the catalytic cycle [36].
This mechanistic proposal prompted us to further explore this deeper making use of Density
Functional Theory (DFT) calculations applied on the intermediaries mentioned above to assess the
associated thermodynamics. All calculations were performed exclusively for the energetics of the
reagents and products without accounting for the catalyst’s influence.
In this way, the very first step should be the initiation step with decomposition of tbhp to yield the
initiator radicals TB1 and TB2, according to Scheme 3. Following this, radical TB1 will react with
styrene (S) yielding intermediary 1 (Figure 9). The reaction is endergonic with a calculated G of
23.2 kcal mol-1 in acetonitrile (23.5 kcal mol-1 in toluene). Intermediary 1 will consequently
decompose straightforward towards formation of the epoxide (E) and radical TB. The reaction in
this case is exergonic with a calculated G of -24.7 kcal mol-1 in acetonitrile (-24.6 kcal mol-1 in
toluene).
On the other hand, concomitantly, radical TB2 can also react with styrene (S) to yield intermediary
2 (Figure 10). This reaction is endergonic (similarly to the formation of intermediary 1) with a
calculated G of 19.5 kcal mol-1 in acetonitrile (19.8 kcal mol-1 in toluene). According to Scheme 3,
decomposition of intermediary 2 is more challenging than that of 1. Peroxide 2 decomposes through
two possible pathways (Figure 10).
It can yield the epoxide E and radical TB1 (which will feed the cycle, Scheme 3) or di-radical 3 by
loss of TB radical. According to DFT results both paths are exergonic. Decomposition of 2 towards
E is predicted to have a calculated G of -13.6 kcal mol-1 in acetonitrile (-13.7 kcal mol-1 in
toluene), while decomposition towards 3 will have a calculated G of -35.8 kcal mol-1 in
acetonitrile (-36.0 kcal mol-1 in toluene). Further decomposition of intermediary 3 yielding
formaldehyde and benzaldehyde is predicted to be slightly endergonic with a G of only 3.2 kcal
mol-1 in acetonitrile (3.1 kcal mol-1 in toluene), according to Figure 11a.
This confirms that, although likely to occur, the isomerization step is not a favored pathway with
these catalysts, being a relevant differing feature of these catalysts compared to others found in the
literature [34,36].
14

On the other hand, the likely formation of phenylacetaldehyde (Phac) and/or acetophenone (Acph)
arising from an isomerization step on the epoxide E, is also predicted to be slightly exergonic as
evidenced in Figure 11b, which accounts for the low yield of these products. It is worth
remembering that in this work only acetophenone was obtained and exclusively when the catalyst
was HTCo/Al-met-Au (Figure 8).
Regeneration of TB radical by reaction with O2 (yielding TB2) is an endergonic process with a
calculated G of 37.5 kcal mol-1 in acetonitrile (32.1 kcal mol-1 in toluene) as shown in Figure 11c.
From the results discussed above, the steps involving the reaction of radicals TB1 and TB2 with
styrene (S), as well as the regeneration reaction of radical TB with O2 may account for an induction
period observed experimentally at early stages of the reaction, given that these reactions have free
energy barriers in the 20-40 kcal mol-1 range at 353 K.
4. Conclusions
Continuing our efforts to prepare sustainable and environmentally friendly catalysts we reported in
this work a set of clay-based catalysts with Au nanoparticles. Their activity was tested in olefin
oxidation catalysis. The set of clay materials covered both bulk and exfoliated (colloid) versions,
while the Au nanoparticles were synthesized using tea extract with a high load of antioxidants. This
was also a sustainable water-based biomimetic procedure.
The catalysts were highly selective towards the epoxide products. Substrate conversion was also
found to be generally very high (above 74%, Table 1) with the exception being the oxidation of 1-
octene. In this case, the bulk catalyst, achieved 32% substrate conversion, while the exfoliated
counterpart did not show any activity at all. This could be expected for this compound as it is an
unbranched terminal olefin.
By the same token the bulk catalyst evidenced higher activity than the colloidal catalyst for cis-
cyclooctene and trans-hex-2-en-1-ol. For cis-cyclooctene, both catalysts yielded the corresponding
epoxide almost exclusively, but for the allylic alcohol epoxide selectivity was always high, since it
was found to be dependent on the catalyst form.
Stereoselectivity increased when using the exfoliated catalysts, as demonstrated in the epoxidation
of limonene, where the synergy of a chiral ligand together with the clay nanosheets seems to play a
definitive role. (see Figures S1 and S2 in SI material).
Styrene oxidation has shown similar differences. While the bulk catalyst evidenced higher substrate
conversion, but lower epoxide selectivity (and a third product). The exfoliated one showed lower
15

conversion but almost twice epoxide selectivity, showing that the choice of the catalyst form is
relevant.
The mechanistic proposal for styrene oxidation based on DFT calculations, also showed that the
regeneration of the tert-butyl radical (TB yielding TB2) may be responsible for the rate determining
step, in addition to formation of intermediaries 1 and 2 (Scheme 3).
This renders excellent perspectives for future developments of this system, which are being
endeavoured by us.
Acknowledgments
The authors would like to acknowledge FCT for financial support (project
UID/MULTI/00612/2013). CIF (SFRH/BD/81029/2011) and AM (SFRH/BPD/35036/2007) also
acknowledge FCT for financial support. G09 calculations were made possible due to the computing
resources provided by STFC Scientific Computing Department's SCARF cluster. We also thank the
reviewers for raising pertinent questions that contributed largely to improve this manuscript.
References
1 D. Astruc, Nanoparticles and catalysis, Wiley‐ VCH Verlag GmbH & Co. KGaA, 2008.
2 R.J. White, R. Luque, V.L. Budarin, J.H. Clark, D.J. Macquarrie, Chem. Soc. Rev. 38 (2009)
481–494.
3 L. Mai, F. Yang, Y. Zhao, X. Xu, L. Xu, B. Hu, Y. Luo, H. Liu, Mater. Today 14 (2011) 346–
353.
4 European Cluster on Catalysis, www.catalysiscluster.eu, (accessed July 2019).
5 SPIRE, www.spire2030.eu/sites/default/files/pressoffice/spire-roadmap.pdf, (accessed July 2019).
6 C.I. Fernandes, C.D. Nunes, P.D. Vaz, Curr. Org. Synth. 9 (2012) 670–694.
7 Z. Liu, R. Ma, M. Osada, N. Iyi, Y. Ebina, K. Takada, T. Sasaki, J. Am. Chem. Soc. 128 (2006)
4872–4880.
8 Y. Tokudome, N. Tarutani, K. Nakanishib, M. Takahashia, J. Mater. Chem. A 1 (2013) 7702–
7708.
9 N. Tarutani, Y. Tokudome, M. Fukui, K. Nakanishi, M. Takahashi, RSC Adv., 2015, 5, 57187–
57192.
10 Y. Tokudome, T. Morimoto, N. Tarutani, P.D. Vaz, C.D. Nunes, V. Prevot, G.B.G. Stenning, M.
Takahashi, ACS Nano 10 (2016) 5550–5559.
11 J. Wang, L. Zhao, H. Shi, J. He, Angew. Chem. Int. Ed. 50 (2011) 9171–9176.
16

12 D. Kino, Y. Tokudome, P.D. Vaz, C.D. Nunes, M. Takahashi, J. Asian Ceramic Socs. 5 (2017)
466–471.
13 S.R. Leandro, A. Mourato, U. Łapińska, O.C. Monteiro, C.I. Fernandes, C.D. Nunes, P.D. Vaz,
J. Catal. 358 (2018) 187–198.
14 L. Li, L. Dou, H. Zhang, Nanoscale 6 (2014) 3753–3763.
15 L. Wang, J. Zhang, X. Meng, D. Zheng, F. Xiao, Catal. Today 175 (2011) 404–410.
16 J. Wang, X. Lang, B. Zhaorigetu, M. Jia, J. Wang, X. Guo, J. Zhao, ChemCatChem 6 (2014)
1737–1747.
17 X. Wang, G. Wu, F. Wang, K. Ding, F. Zhang, X. Liu, Y. Xue, Catal. Commun. 28 (2012) 73–
76.
18 S. Nakagaki, M. Halma, A. Bail, G.G.C. Arizaga, F. Wypych, J. Colloid Interf. Sci. 281 (2005)
417–423.
19 X. Bai, Y. Gao, H. Liu, L. Zheng, J. Phys. Chem. C 113 (2009) 17730–17736.
20 F. Lin, R. Doong, J. Phys. Chem. C 115 (2011) 6591–6598.
21 K. Layek, M.L. Kantam, M. Shirai, D. Nishio-Hamane, T. Sasaki, H. Maheswaran, Green Chem.
14 (2012) 3164–3174.
22 H.K. Kadam, S G. Tilve, RSC Adv. 5 (2015) 83391–83407.
23 A. Corma, H. Garcia, Chem. Soc. Rev. 37 (2008) 2096–2126.
24 M. Stratakis, H. Garcia, Chem. Rev. 112 (2012) 4469−4506.
25 D. Cruz, P.L. Falé, A. Mourato, P.D. Vaz, M.L. Serralheiro, A.R.L. Lino, Colloids and Surf. B:
Biointerf. 81 (2010) 67–73.
26 I.J. Marques, P.D. Vaz, A.C. Fernandes, C.D. Nunes, Microporous Mesoporous Mater. 183
(2014) 192–200.
27 C. Taviot-Gueho, F. Leroux, C. Payen, J.P. Besse, Appl. Clay Sci. 28 (2005) 111–120.
28 H.-J. Kim, K. Ryu, J.-H. Kang, A.-J. Choi, T. Kim, J.-M. Oh, Sci World J. (2013) 421967.
29 J.C. Dupin, H. Martinez, C. Guimon, E. Dumitriu, I. Fechete, Appl. Clay Sci. 27 (2004) 95–106.
30 P.D. Vaz, C.D. Nunes, New J. Chem. 34 (2010) 541-546.
31 M. Vasconcellos-Dias, C.D. Nunes, P.D. Vaz, P. Ferreira, M.J. Calhorda, Eur. J. Inorg. Chem.
(2007) 2917-2925.
32 A.A. Giro Dos Santos, Biossíntese, caracterização e citotoxicidade de nanopartículas de ouro
usando extracto de lúcia-lima e erva de São Roberto (http://hdl.handle.net/10451/4136), MSc thesis,
Universidade de Lisboa, Lisboa, 2010.
33 S. Bawaked, N.F. Dummer, D. Betchell, D.W. Knight, G.J. Hutchings, Green Chem. 13 (2011)
127–134.
17

34 Z. Chen, R.L. Luck, Green Chem. 18 (2016) 3354–3359.
35 A. Abad, A. Corma, H. García, Pure Appl. Chem. 79 (2007) 1847–1854.
36 P. Cancino, V. Paredes-García, P. Aguirre, E. Spodine, Catal. Sci. Technol. 4 (2014) 2599–2607.
37 A.S. Alshammari, Catalysts 9 (2019) 402.
38 X. Wang, N.S. Venkataramanan, H. Kawanami, Y. Ikushima, Green Chem. 9 (2007) 1352–1355.
18

Figures
Figure 1. TGA profiles obtained under N2 atmosphere at 10 K/min of met, HTCo/Al-met and
HTCo/Al-met-Au.
19

Figure 2. Electronic spectra of gold nanoparticles (AuNP), methionine (met), and of the hybrid clay
materials HTCo/Al-met and HTCo/Al-met-Au. The spectra of the hybrid materials were obtained as
diffuse reflectance (DRUV) while the remaining were obtained from aqueous solutions in
transmission.
20

Figure 3. TEM images of ExHTCo/Al-met (a) and ExHTCo/Al-met-Au (b). The inset in the top image
displays the diffraction pattern of the sample and evidences a hexagonal symmetry expected for
these samples. In the middle image (b) one can see the AuNP wrapped with the sheets of ExHTCo/Al
-
met, while image (c) is a zoom of the area highlighted in image (b).
21

Figure 4. 3D AFM images of ExHTCo/Al-met (a) and ExHTCo/Al-met-Au (b) and corresponding
profiles.
22

Figure 5. XRD powder patterns of HTCo/Al-met and HTCo/Al-met-Au. The peak at. 2 = 5.74° is
assigned to the 003 peak corresponding to the interlayer space containing the methionine guests.
Diffraction peaks arising from AuNP in HT-met-Au are highlighted with an arrow at 2θ = 38.2°,
44.4° and 64.6° while the dotted lines next to it are eye-guides to mark the position of the
diffraction peaks arising from the clay. Diffraction peak indexation was based on literature data
[28].
23

Figure 6. DRIFT spectra of HTCo/Al-met and HTCo/Al-met-Au materials. Spectra in (a) the full
spectra are shown along with the spectrum of the parent HTCo/Al-Cl material. The CH bands of
met in the spectra of HTCo/Al-met and HTCo/Al-met-Au materials were marked for clearer
identification. Spectra in (b) display the fingerprint region (modes below 2000 cm-1) of the hybrid
materials; the spectrum of met was included for comparison.
24

Figure 7. Reaction kinetics (a) for catalytic R-(+)-limonene epoxidation at 328 K with HTCo/Al-met-
Au and ExHTCo/Al-met-Au catalysts; at the beginning there is evidence for a slight induction period
compatible with a radical mechanism. Also shown are the product selectivity evolution for the bulk
(b) and exfoliated (c) catalysts. It is worth noticing the difference observed for the cis/trans
diastereomeric ration between the bulk and the exfoliated catalysts.
25

Figure 8. Reaction kinetics (a) for catalytic styrene epoxidation at 328 K under air atmosphere with
HTCo/Al-met-Au and ExHTCo/Al-met-Au catalysts; at the beginning there is evidence for a slight
induction period compatible with a radical mechanism. The inset shows the reaction kinetics
performed under N2 atmosphere, showing that O2 is required. Also shown are the product selectivity
evolution for the bulk (b) and exfoliated (c) catalysts. It is worth noticing that both catalysts
displayed different major products.
26

Figure 9. Free energy profile (kcal mol-1) for the reaction of radical TB1 with styrene substrate (S),
yielding intermediary 1 and its subsequent decomposition towards the epoxide product (E) and
radical TB. The top values correspond to those calculated in acetonitrile (blue) while the bottom
ones were obtained for toluene (red) and are reported for the experimental reaction temperature
(353 K). The spin density of the radicals was plotted in blue with a 0.05 cut-off.
27

Figure 10. Free energy profile (kcal mol-1) for the reaction of radical TB2 with styrene substrate
(S), yielding intermediary 2 and its subsequent decomposition towards the epoxide product (E) and
radical TB1 and the alternative pathway yielding diradical 3 and radical TB. The top values
correspond to those calculated in acetonitrile (blue) while the bottom ones were obtained for toluene
(red) and are reported for the experimental reaction temperature (353 K). The spin density of the
radicals was plotted in blue with a 0.05 cut-off.
28

Figure 11. Free energy profile (kcal mol-1) for: a) decomposition of diradical 3 towards
benzaldehyde (B) and formaldehyde (F); b) isomerization of styrene epoxide (E) into acetophenone
(Acph) or phenylacetaldehyde (Phac); c) regeneration reaction of radical TB with dioxygen (O2) to
yield radical TB2. The top values correspond to those calculated in acetonitrile (blue) while the
29