N-Functionalised Imidazoles as Stabilisers for Metal Nanoparticles in Catalysis and Anion Binding

Article abstract of DOI:10.1002/open.202000145

Metal nanoparticles (NPs) have physicochemical properties which are distinct from both the bulk and molecular metal species, and provide opportunities in fields such as catalysis and sensing. NPs typically require protection of their surface to impede aggregation, but these coatings can also block access to the surface which would be required to take advantage of their unusual properties. Here, we show that alkyl imidazoles can stabilise Pd, Pt, Au, and Ag NPs, and delineate the limits of their synthesis. These ligands provide an intermediate level of surface protection, for which we demonstrate proof-of-principle in catalysis and anion binding.

Full text of DOI:10.1002/open.202000145

Full Papers
doi.org/10.1002/open.202000145
ChemistryOpen
1
2
3
4
5
N-Functionalised Imidazoles as Stabilisers for Metal
Nanoparticles in Catalysis and Anion Binding
6
7
Christopher J. Serpell,*[a] James Cookson,[b] and Paul D. Beer*[c]
8
9
Metal nanoparticles (NPs) have physicochemical properties
which are distinct from both the bulk and molecular metal
species, and provide opportunities in fields such as catalysis
and sensing. NPs typically require protection of their surface to
impede aggregation, but these coatings can also block access
to the surface which would be required to take advantage of
their unusual properties. Here, we show that alkyl imidazoles
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
can stabilise Pd, Pt, Au, and Ag NPs, and delineate the limits of
their synthesis. These ligands provide an intermediate level of
surface protection, for which we demonstrate proof-of-principle
in catalysis and anion binding.
1. Introduction
lability, allowing exchange with incoming ligands,[6,7] and fluid-
ity on the surface.[8] Such dynamic potential may promote
catalytic activity. Thiols are most commonly employed since
these exhibit strong interactions with noble metal surfaces;
they have become the ‘default’ ligands, especially for AuNPs,
since the introduction of the Brust method[9] for synthesis,
despite sulfur being well known as a catalyst poison. However,
donation of electrons from other elements is entirely feasible.
Nanoparticle stabilisation has been reported through
nitrogen,[10] oxygen,[11] and phosphorous[12] donors, and more
recently carbon.[13]
Due to their exceptionally high surface area to volume ratio,
metal nanoparticles (NPs) are natural candidates for new
catalysts and sensors. The benefits of nanoparticulate systems
exceed mere metal economy: the confinement of electrons to
nanoscale domains can improve activity and selectivity, and
offer surface-sensitive optical properties; and a shell of stabilis-
ing ligands, polymers, or surfactants can be tailored to fine tune
the interaction of the substrate with the surface.[1,2] NPs require
this stabilising shell to prevent irreversible agglomeration which
is catastrophic for nanoelectronic properties and surface area.
Stabilisers operate either sterically (using ligands or polymers)
or electrostatically, preventing the close approach of separate
NPs,[3] but this occlusion of the surface may also apply to target
substrates for catalysis or sensing, resulting in reduced
activity.[4,5] Conversely, a functionally active NP must have a
surface which is accessible to the substrate - but less robust
surface coverage may also promote agglomeration. The level of
stability afforded varies according to factors such as the ratio of
capping agent to metal, and the magnitude of the interaction
with the surface. The use of lone-pair donating ligands which
form a coordinate bond with the surface is particularly versatile
because the strength of the attachment can be correlated with
molecular analogues, and judicious choice of ligand can impart
The strength of the ligand-metal interaction will play a key
role in determining access of catalytic or analytic substrates to
the surface. For the production of new nanoparticles for
catalysis or sensing, the ideal ligand should strike a balance
between strength of coordination and lability, be easily
synthesised from inexpensive commercial materials, exhibit
broad scope for modification, and be non-protic. We believe
that the hard/soft mismatch between nitrogen donors and
noble metals could be used to achieve this, and in particular,
alkyl imidazoles comply with all these requirements. To date,
use of the basic imidazole nitrogen to stabilise NPs has been
sparse. The first observation of NP stabilisation by alkyl
imidazoles was the discovery that such ligands were present as
impurities in ionic liquids, and when NPs were produced in
those media, the neutral N-donors were the true NP ligands.
Indeed, intentional addition of methyl imidazole improved the
uniformity and stability of AuNPs, and in PdAu alloy NPs the
catalytic activity was also increased.[14] Through incorporation of
an anion binding moiety in the alkyl chain, we have employed
imidazole ligands in the creation of core@shell bimetallic
particles for enhancement of catalytic selectivity.[15] The
imidazole motif has been used to pre-coordinate gold ions to
the reducing biopolymer chitosan to produce polymer-coated
luminescent AuNPs which were active for catalysing hydride
reductions.[16] In a similar vein, polysiloxane microspheres were
decorated with imidazole groups, and used to bind Pd which
was then reduced into nanoparticles on the surface to create
catalysts for hydrogenation.[17] These studies indicate that there
is further potential for exploration of catalysis by imidazole-
[a] Dr. C. J. Serpell
School of Physical Sciences, Ingram Building, University of Kent, Canterbury,
CT2 7NH (UK)
E-mail: c.j.serpell@kent.ac.uk
[b] Dr. J. Cookson
Johnson Matthey Technology Centre, Reading, RG4 9NH, (UK)
[c] Prof. P. D. Beer
Chemistry Research Laboratory, Department of Chemistry, University of
Oxford, Mansfield Road, Oxford, OX1 3TA, (UK)
E-mail: paul.beer@chem.ox.ac.uk
Supporting information for this article is available on the WWW under
https://doi.org/10.1002/open.202000145
Functional Supramolecular Systems
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA. This
is an open access article under the terms of the Creative Commons Attri-
bution License, which permits use, distribution and reproduction in any
medium, provided the original work is properly cited.
ChemistryOpen 2020, 9, 683–690
683
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Full Papers
doi.org/10.1002/open.202000145
ChemistryOpen
stabilised NPs. Beyond catalysis, there are otherwise unrelated 2. Results and Discussion
1
2
3
4
5
6
7
8
9
reports of imidazole-NPs: copper NPs produced by pulsed-laser
ablation in the presence of DNA were found to be stabilised in
part by the imidazole N of purine DNA bases;[18] replacement of
one of the phenyl rings in azobenzene with an imidazole
enabled binding of the photoswitching unit to AgNPs, where its
switching capacity was retained;[19] and the lability of the
imidazole NÀ NP bond has been used in the development of
sensors for sulfur mustard, a chemical warfare agent, by ligand
displacement.[20] As indicated above, in its bis-alkylated imidazo-
lium form, the heterocycle may provide NP stabilisation either
noncovalently,[21] or as an N-heterocyclic carbene,[13] (a stronger
interaction[22]) while other imidazole-containing ligands have
been produced which provide stabilisation through a motif
other than the basic nitrogen, e.g. 2-mercapto-1-meth-
ylimidazole.[23,24] The related triazole heterocycle has also been
used as an N-donor to stabilise AuNPs for sensing and
catalysis,[25–27] and tetrazoles have also been explored.[28]
Supported by these positive but disparate reports, there is
much unexplored potential for alkyl imidazoles as ligands for
metal nanoparticles in catalysis and beyond.
Here, we report the synthesis and characterisation of gold,
palladium, platinum, and silver NPs stabilised by simple alkyl
imidazole ligands, defining the limiting requirements for NP
stabilisation in terms of alkyl chain length, molar equivalents,
and metal used. We then provide two case studies for how
these NPs may be used, showing proof-of-principle for
enhanced catalytic activity despite the presence of the ligands,
and for anion coordination using functional ligands.
2.1. Nanoparticle Synthesis
To make use of alkyl imidazole capping of metal NPs,
stabilisation requirements and synthetic protocols needed to be
established. This involved discovering the minimum required
alkyl chain length and alkyl imidazole ligand:metal ratio for
stabilisation, testing reaction conditions, and experimenting
with a range of metals. Alkyl imidazoles were readily synthes-
ised in excellent yield by deprotonation of imidazole with
sodium hydride in THF, followed by reaction with bromoalkanes
(Scheme 1a). Alkyl imidazole ligands 1–3 were isolated as
increasingly viscous oils, while 4 was a waxy solid. NPs were
then synthesised by an adaptation of the Brust method.[9]
Starting with hydrogen tetrachloroaurate, potassium tetrachlor-
opalladate, and potassium tetrachloroplatinate, the metallate
anions were extracted into toluene from water using the phase
transfer agent Aliquat 336. The layers were separated, and the
alkyl imidazole ligand (1–3) was added to the organic layer in
molar ligand:metal ratios of 1 and 10. After cooling in an ice
bath, dropwise addition of aqueous sodium borohydride
followed. In case of Au, a colour change from yellow to deep
purple, typical of AuNPs, was observed, whereas for Pd a black
solution was obtained. When using Pt, no nanoparticle
formation was observed, despite allowing the reaction to warm
to room temperature, or addition of hydrochloric acid to
promote the evolution of dihydrogen from the borohydride
salt. Work-up was achieved by washing with acid (to remove
excess alkyl imidazoles) and water (to remove inorganic side
products), followed by evaporation of the solvent, giving
products MNP.Lx where M is the metal used, L is the ligand
(compounds 1–3), and x is the number of equivalents of ligand
to metal used in the reaction (not necessarily that of the
product). The AuNP samples were thick oils which displayed a
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Scheme 1. Synthetic routes towards imidazole-stabilised NPs. (a) Brust-style two-phase chemical reduction. (b) Thermal degradation of platinum complexes.
(c) One-phase chemical reduction.
ChemistryOpen 2020, 9, 683–690
www.chemistryopen.org
684
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Full Papers
doi.org/10.1002/open.202000145
ChemistryOpen
gold sheen despite being redispersible in organic solvents,
whereas black oils were obtained for the PdNPs. Stable nano-
particles were obtained for both gold and palladium using alkyl
imidazole ligands 2 and 3 in both 1:1 and 1:10 metal:ligand
ratios, whereas using ligand 1, the only stable product was
AuNP.110, indicating the smallest n-propyl group was not
capable of sterically stabilising PdNP formation. In an effort to
obtain PtNPs, an alternative method was pursued (Scheme 1b),
in which platinum acetoacetonate was dissolved in 2 and
thermally decomposed, giving a colour change from orange to
dark brown, characteristic of PtNPs (PtNP.2n). However, it was
not possible to separate PtNP.2n from the excess 2. Synthesis of
alkyl imidazole stabilised AgNPs (starting from AgNO3, Sche-
me 1c) was attempted using ligands 2 and 3, but gave only
agglomerated bulk metal. Stable AgNP.410 could be generated
using ten equivalents of hexadecylimidazole (4), but after
reaction could not be separated from the free ligand, giving a
pale pink waxy material.
All the NP samples were characterised by TEM (Figure 1,
Table 1, Figure S1–12, Supporting Information). Large and
uneven particles were seen for AuNP.110, indicating the paucity
of the stabilisation given by the short n-propyl chain imidazole.
Much smaller sizes and improved uniformity were seen across
the board for the other AuNP samples. High magnification of
AuNP.21 revealed the lattice fringes, designating the planes of
metal atoms within the NP and indicating a crystalline nature.
Small and uniform spherical NPs were observed for the Pd
samples PdNP.21, PdNP.210, and PdNP.310 but irregular mor-
phology found for PdNP.31. Use of Pd gave smaller NPs than
gold by about 1 nm in diameter, and with better monodisper-
sity. Using either metal, the particle size was lower when ten
equivalents of the ligand to metal were used, rather than
equimolar ratios. This corresponds to the greater stabilised
surface area that can be achieved by providing more capping
agents. PtNP.2n were extremely small (<1 nm), which explains
why separation from excess ligand was challenging. AgNP.410
displayed very large NPs with high dispersity. Since stable
solutions were obtained of the NPs, their clustering in the TEM
images is most likely a drying effect; while it is possible that
agglomeration would also occur upon drying, since the NP
samples could be repeatedly dried and redissolved we do not
believe that this is significant. The AuNPs displayed plasmon
resonance bands (PRB) in the visible region typical of the
material (Table 1), whereas no PRB was observed for Pd or Pt, as
is usual.[29]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
These studies show that stable and well-formed NPs can be
obtained using Au and Pd using alkyl imidazole ligands via the
Brust method. For these, a propyl chain on the imidazole is
insufficient, presumably since it does not provide sufficient
steric stabilisation, but a hexyl chain is effective, as is a dodecyl
chain. Ligand:metal ratios of both 1:1 and 10:1 are suitable
with hexyl and dodecyl imidazoles; the higher ratio gives
smaller NPs. PtNPs and AgNPs are accessible but require
adapted synthetic methods and do not display useful size
distributions.
2.2. Assessing Surface Accessibility Through Catalysis
We then proceeded to examine the degree of surface
accessibility by conducting catalytic tests on the nanoparticu-
late metals – if the ligands prevent substrates reaching the
surface, then the catalytic activity should be significantly lower
than commercial catalysts, or our own samples following ligand
removal. Preliminary evaluation showed that only the palla-
dium-based systems were catalytically active in commercially
relevant hydrogenation conditions, therefore these were tar-
geted for further investigation and comparison with ligand-free
controls.
Heterogeneous catalysts were prepared by adsorbing sol-
utions of PdNP.21, PdNP.210, PdNP.31 and PdNP.310 onto
activated carbon, each in projected metal loadings of both 1 wt
% and 5 wt%. This was achieved by stirring a chloroform
solution of NPs with carbon until the solution lost all colour
(typically one hour). Filtration, washing with chloroform, and
Figure 1. TEM images of selected alkyl imidazole-stabilised NPs. Further
images can be found in the Supporting Information.
Table 1. Characterisation data for NPs. a Toluene, 293 K. b Determined
using TEM, n=number of NPs measured.
Ligand/Au molar ratio
PRB/nma
Average diameter/nmb
AuNP.110
AuNP.21
AuNP.210
AuNP.31
AuNP.310
PdNP.21
PdNP.210
PdNP.31
PdNP.310
PtNP.2n
532
529
520
522
520
–
–
–
–
–
25.5�10.1, n=165
2.9�1.7, n=136
2.8�1.1, n=118
3.9�1.2, n=285
2.3�1.0, n=398
1.9�0.6, n=484
1.6�0.9, n=130
2.2�0.9, n=393
1.9�0.2, n=32
<1 nm
AgNP.410
–
25.6�36.7 nm, n=29
ChemistryOpen 2020, 9, 683–690
www.chemistryopen.org
685
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Full Papers
doi.org/10.1002/open.202000145
ChemistryOpen
vacuum drying of the carbon suspension gave samples ready
for analysis. TEM analysis of the supported particles displayed
some growth in particle diameter compared to the unsup-
ported particles; for example, PdNP.310 increased from 1.4 to
4.2 nm in average diameter – this may in part be due to a
flattening effect caused by the attraction of the NP to the
surface, or to partial agglomeration. The increase of particle size
of precious metal nanoparticles has previously been observed
and attributed to the formation of stronger bonds with defects
on the surface of various supports[30,31] To test the effect of the
presence or absence of ligands on the catalytic properties of
the NPs, a selection of the samples were calcined in air, burning
off the ligands. The appropriate temperature for this was found
using thermogravimetric analysis (TGA, Figure S15, Supporting
Information). For pure ligands 2 and 3, total decomposition had
degradation of a common pollutant.[32] The reduction of NB to
An was therefore taken as a preliminary test reaction for the
catalytic activity of PdNPs. Catalytic experiments using PdNPs,
varying the stabiliser chain length, metal to ligand ratio, catalyst
1
2
3
4
5
6
7
8
9
°
loading, and use of calcination, were undertaken at 50 C under
2 bar of hydrogen, the products being analysed by gas
chromatography (GC). Given a 15 min stirring time, none of the
reactions had reached completion, allowing a meaningful
comparison of catalytic activity based upon their % conversion
(Figure 2). The performance of the Pd NPs was comparable to
the Pd/C standard and very minimally affected by calcination –
the ratio of the % conversion for uncalcined to calcined ranged
from 0.93 to 1.05. This affirms that the imidazole ligand is
sufficiently labile that it does not impede access to the NP
surface for catalysis, consistent with our hypothesis that the
hard/soft mismatch of nitrogen donors with noble metals
would provide such behaviour, and with previous results.[15,20]
Although there were no clear activity trends with respect to
alkyl chain length (C6 vs. C12), or ligand to metal ratio, the
samples with nominally 5 wt% catalytic loading performed
consistently better than those with 1 wt% coverage, despite
reactions being controlled for total Pd content.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
°
°
occurred by 250 C and 350 C respectively. The main feature in
the TGA plots of the Pd samples adsorbed onto carbon was the
°
combustion of the support at 500 to 550 C. This is unsurprising
given that samples contained only 5 wt% imidazole-stabilised
NPs. Closer examination of the plots for NPs PdNP.21 and
PdNP.310 reveals downward kinks in the TGA traces correspond-
°
°
ing to the loss of ligand at 250 C and 350 C respectively. The
°
total loss of mass prior to reaching 400 C, corresponding to the
To test the Pd NPs on a more challenging reaction in terms
of selectivity, the reduction of 2,3-dichlorobenzonitrile was
investigated. This reaction can proceed by a number of routes
including dehalogenation and formation of secondary and
tertiary amines (Scheme 2). The target, (2,3-dichlorophenyl)
methanamine, is an amino aryl halide, a class of compounds
°
ligands, was compared with the residue at 600 C (pure metal),
to estimate the surface coverage. A ligand mass% of 32.7 and
41.0 were obtained for PdNP.21 and PdNP.310 respectively,
giving approximately 296 and 77 ligands per particle, or 11.2
and 6.7 ligands per nm2, taking into account average particle
size. An estimated theoretical maximum number of ligand/nm2
is 12.5, indicating that with the shorter alkyl chain the ligands
are closely packed, whereas with the longer chain there is
greater spacing. For catalytic tests, a representative selection of
°
NP samples were calcined at 450 C to ensure complete ligand
removal without combustion of the carbon support. Finally, the
Pd samples were analysed by Inductively Coupled Plasma
Emission Spectroscopy (ICP-ES), to ascertain the precise metal
loading, an important detail for preparing catalytic tests. The
results show that the level of loading is uniformly lower than
expected (Table 2). These losses of metal could occur through
the formation of small levels of agglomerated metal which do
not pass through filter paper, and through incomplete
adsorption onto the carbon support.
Nitrobenzene (NB) is widely used in organic synthesis, and
as a solvent. Its reduction to aniline (An) is important both in
Figure 2. Conversion of nitrobenzene to aniline by PdNPs. 2 bar, 15 min,
°
terms of synthetic methodology and as
a step in the
50 C, MeOH, 1:1000 ratio of Pd:substrate.
Table 2. Projected and measured (ICP-ES) loading of Pd NPs onto carbon
supports.
Sample
Projected Loading
Measured Loading
PdNP.21
PdNP.21
PdNP.210
PdNP.210
PdNP.31
PdNP.31
PdNP.310
PdNP.310
1%
5%
1%
5%
1%
5%
1%
5%
0.44%
2.27%
0.66%
2.82%
0.81%
3.45%
0.31%
3.45%
Scheme 2. Hydrogenation pathways for 2,3-dichlorobenzonitrile.
ChemistryOpen 2020, 9, 683–690
www.chemistryopen.org
686
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Full Papers
doi.org/10.1002/open.202000145
ChemistryOpen
which are versatile synthons in organic chemistry, and for which
selective reduction catalysts are highly desirable.[33] PdNP.31 at
5 wt% loading (uncalcined, i.e. with ligands still present) were
selected to be compared with the commercial Pd/C standard.
The reaction was conducted in pure ethanol, and in acidic
media known to promote the desired selectivity.[34] While the
product distribution was dominated by the target compound
and the starting material, the presence of other products was
conspicuous, notably the 3-dechlorinated species. It is note-
worthy that when compared to the Pd/C standard, a significant
improvement in selectivity was observed for PdNP.31 in ethanol
– mild, neutral conditions suitable for reduction of functional-
since it would prevent formation of monometallic NPs during
the second reduction step.
1
2
3
4
5
6
7
8
9
Borohydride reductions were attempted in acetonitrile
(required for solubility of 5) using NaAuCl4, K2PdCl4, or AgNO3 as
metal salt precursors, with both 0.5 and 0.2 equivalents of the
ligand (Scheme 3). Agglomerated bulk metal was obtained from
the Au salt, while addition of the Pd salt caused precipitation of
a highly insoluble material (presumably 5.PdCl4) prior to
reduction. After dispersing that material as a suspension,
addition of methanolic sodium borohydride generated only
insoluble agglomerated products. The outcome was different
with the Ag salt, the solution of which turned yellow and then
dark brown, giving stable NPs as the sole product in the 1:2
ligand:metal ratio (AgNP.50.5), some agglomerated material at
1:5 ratio (AgNP.50.2). Filtration was performed and the solvent
was removed in vacuo, the residue being then triturated with
saturated aqueous NH4PF6, to give AgNP.50.5 and AgNP.50.2 as
their non-coordinating hexafluorophosphate salts. The samples
were characterised by UV-visible spectroscopy, powder X-ray
diffraction (PXRD), and TEM analysis. PXRD revealed that while
the major component of AgNP.50.2 was identical to bulk silver,
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
ised molecules (Table 3). Furthermore, the
% conversion
resulting from use of the alkylimidazole-coated NPs was again
close to that of the commercial Pd/C, again evidencing that
access of substrates to the metal surface was not impeded
(ratio of % conversion of PdNP.31 to commercial Pd/C between
0.95 and 1.16).
These catalytic tests provide proof-of-principle for the use of
alkylimidazole-stabilised NPs as new hydrogenation catalysts,
since the ligands due not diminish catalytic activity significantly,
but can aid product selectivity. We believe that this is due to
the hard/soft mismatch of the nitrogen donor with the noble
metal surface.
this metallic form was a much smaller part of AgNP.50.5
.
Scherrer analysis of the (220) reflection gave diameters of 10.3
and 8.5 nm for AgNP.50.5 and AgNP.50.2 respectively. UV-visible
spectroscopy revealed plasmon resonance bands at 419 and
413 nm for AgNP.50.5 and AgNP.50.2 respectively, both of which
are typical for Ag NPs of approximately 10 nm. This size analysis
was confirmed by TEM experiments. (Figure 3a, Fig S13–S14)
which gave diameters of 11.1 nm (�1.8, n=85) and 7.5 nm (�
1.6, n=154) for AgNP.50.5 and AgNP.50.2 respectively. Compared
to other NP samples, it can be seen from the TEM images that
the AgNPs form aggregates rather than being dispersed evenly.
Given the dense nature of the aggregates and their presence
2.3. Anion Sensing Studies Using Silver Nanoparticles
Stabilised by a Bis-Imidazolium-Imidazole Ligand
We then decided to examine the possibility of non-covalently
binding a second metal to the surface prior to reduction, to
generate core@shell nanoparticles. This templation-via-anion-
coordination strategy has been used to produce Au/Pd core@-
shell NPs which showed improved reduction of chloronitroani-
line without dehalogenation compared to monometallic or
unstructured bimetallic species,[15] but the system was not
universally applicable. Here, we decided to investigate the use
of a more potent anion-host based upon charge-assisted
hydrogen bonding through use of a pre-macrocycle 5[35] which
binds chloride in the highly competitive DMSO-d6 (K=164
(6) molÀ 1 dm3), while the cyclised version is stronger yet at K=
420 (23) molÀ 1 dm3; coordination to the NP surface could be
considered a type of cyclisation and thus result in similar
strength. An increase in anion binding affinity could improve
the synthetic efficiency of the core@shell structure preparation
Table 3. Levels of conversion and selectivity (% of target found in product
mixture) for the reduction of 2,3-dichlorobenzonitrile to (2,3-dichlorophen-
yl)methanamine catalysed by Pd NPs and Pd/C 87 L in different solvents.
°
4 bar, 4 hr (target 75% conversion), 50 C, 1:1000 ratio of Pd:substrate.
Conversion
Selectivity
PdNP.31 EtOH
Pd/C EtOH
PdNP.31 EtOH/HCl
Pd/C EtOH/HCl
PdNP.31 AcOH
Pd/C AcOH
83.1%
82. 8%
74.9%
64.5%
73.1%
77.1%
50.7%
38.7%
82.2%
86.0%
39.1%
34.4%
Figure 3. Characterisation of AgNPs. (a) TEM images (more can be found in
the SI). (b) Change in UV-visible spectrum of AgNP.50.5 upon addition of 1–
10 equivalents of TBA chloride in 1:1 MeCN:H2O (298 K). (c) Change in
intensity of UV-visible spectrum at 413 nm according to apparent equiv-
alents of chloride added.
ChemistryOpen 2020, 9, 683–690
www.chemistryopen.org
687
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Full Papers
doi.org/10.1002/open.202000145
ChemistryOpen
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Scheme 3. Synthesis of AgNPs using bidentate imidazolium-imidazole ligands, and proposed structure of aggregates.
throughout the TEM sample, we do not believe that this is
related to drying. However, unlike most aggregated NPs, we
found that the AgNPs were readily redispersible. This is likely to
be due to the bidentate nature of 5, and its lack of
preorganisation – the imidazole nitrogens may equally coor-
dinate one or two NPs, producing an interlinked network as
postulated in Scheme 3.
acetoitrile, addition of one equivalent of TBA Cl to 5 caused
precipitation, we titrated AgNP.50.5 with TBA Cl in same
conditions. The point at which precipitation was observed was
then calibrated to the one equivalent mark. To prevent
precipitation in deeper analysis of the anion binding properties
of AgNP.50.5, titrations were subsequently performed in 1:1
acetonitrile/water. Addition of TBA Cl caused an overall blue
shift from 413 nm to ca. 405 nm for the PRB, which was
accompanied by non-linear changes in absorbance intensity –
an initial decrease was followed at the one equivalent mark by
a large increase, which levelled off rapidly (Figure 3b,c). The
retention of the PRB indicates that the chloride ions are not
dissolving the metal, as can occur in weakly stabilised AgNPs
due to the formation of an Ag2O layer.[36] Titrations curves of
this shape are difficult to interpret, but two important facts can
be extracted. Firstly, the clearly non-linear change in intensity
Anion binding of the AgNP network was examined via the
effect of the addition of anions upon the plasmon resonance
band (PRB) of AgNP.50.5 using UV-visible spectroscopy (Fig-
ure 3b,c). Experiments were undertaken titrating AgNP.50.5 with
tetrabutylammonium (TBA) salts with the aim of establishing
surface binding before creating core@shell NP systems. A
particular difficulty encountered when performing titrations
using NPs is that the exact receptor concentration is difficult to
ascertain. To determine this, with the prior knowledge that in
ChemistryOpen 2020, 9, 683–690
www.chemistryopen.org
688
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Full Papers
doi.org/10.1002/open.202000145
ChemistryOpen
means that the halide anions are interacting with the NPs as Experimental Section
1
2
3
4
5
6
7
8
9
well as the ligands. Secondly, since there is a strong response at
All experimental details, synthetic procedures, further character-
isation of NPs (TEM, TGA, PXRD), and anion binding studies can be
found in the Supporting Information.
the nominal one equivalent mark, it validates the stoichiometric
designation from the acetonitrile titration. A distinctly different
effect was seen upon the addition of TBA AuCl4. The PRB was
quenched almost immediately, indicating a significant interac-
tion or dissolution of the Ag, after which the spectrum of the
added salt dominated, giving a linear increase in absorbance. Acknowledgements
The superimposition of spectra also occurred when TBA2 Pd2Cl6
was used. In this case, the absorption spectrum of the salt
coincided with the PRB, so it was impossible to determine if the
PRB was similarly quenched. Attempts were then made to
synthesise Ag@Au and Ag@Pd NPs by combining AgNP.50.5
with ‘one equivalent’ of TBA salts of AuCl4 or Pd2Cl6 in
acetonitrile, followed by addition of methanolic sodium borohy-
dride. In both cases this produced a dark precipitate which
We thank S. P. D. Spratt, G. Goodlet, and S. J. P. Longworth (JMTC)
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
for TEM analysis, G. Nnorom-Junior, J. McNaught, and I. Briggs
(JMTC) to ICP-ES analysis, and J. McNaught (JMTC) for PXRD
analysis. C.J.S. thanks Johnson Matthey and the Engineering and
Physical Sciences Research Council for a CASE Studentship.
À
2À
could not be redissolved, indicative of agglomerated metal. The Conflict of Interest
formation of core@shell nanoparticles has therefore not been
successful in this case, either due to the nature of the
nanoparticle network, or the reduced stabilisation power of the
ligand after reduction of the second metal. Nonetheless, since
the spectral changes due to chloride anion addition are in the
visible region, the network material obtained could act as a
potential optical sensor for chloride or other anions which do
not absorb light in the range of the PRB.
The authors declare no conflict of interest.
Keywords: nanoparticles
· ligands · anion coordination ·
functionalised imidazoles · sensing applications
[1] M. H. G. Prechtl, J. D. Scholten, J. Dupont, J. Mol. Catal. A 2009, 313, 74–
78.
[2] I. Kanelidis, T. Kraus, Beilstein J. Nanotechnol. 2017, 8, 2625–2639.
[3] S. Ozkar, R. G. Finke, S. Özkar, R. G. Finke, J. Am. Chem. Soc. 2002, 124,
5796–5810.
3. Conclusion
We have outlined routes to obtain Au, Pd, Pt, and Ag NPs
stabilised through the basic nitrogen atoms of alkyl imidazole
molecules. The Brust method can be readily adapted for Au and
Pd systems, but Ag and Pt NPs were better produced by more
dramatic adaptations. An n-propyl chain on the imidazole
provides insufficient stabilisation for high-quality NPs, but n-
hexyl yields robust protection of the NPs. All alkyl imidazole
systems produced NPs which were readily redispersible in a
range of organic solvents (introduction of polar groups in place
of alkyl ligands is also possible, permitting tuning of
solubility[15,20]). The ligands appear to offer minimal barrier to
[4] P. Migowski, J. Dupont, Chem. Eur. J. 2007, 13, 32–39.
[5] W. Wu, E. V. Shevchenko, J. Nanopart. Res. 2018, 20, 1–14.
[6] V. Chechik, J. Am. Chem. Soc. 2004, 126, 7780–7781.
[7] S. Rucareanu, V. J. Gandubert, R. B. Lennox, Chem. Mater. 2006, 18,
4674–4680.
[8] A. K. Boal, V. M. Rotello, J. Am. Chem. Soc. 1999, 121, 4914–4915.
[9] M. Brust, M. Walker, D. Bethel, P. J. Schriffen, R. Whyman, J. Chem. Soc.
Chem. Commun. 1994, 801.
[10] G. Aragay, J. Pons, J. Ros, A. Merkoci, Langmuir 2010, 26, 10165–10170.
[11] M. Chen, W. Ding, Y. Kong, G. Diao, Langmuir 2008, 24, 3471–3478.
[12] M. Montalti, N. Zaccheroni, L. Prodi, N. O’Reilly, S. L. James, J. Am. Chem.
Soc. 2007, 129, 2418–2419.
[13] C. A. Smith, M. R. Narouz, P. A. Lummis, I. Singh, A. Nazemi, C.-H. Li,
C. M. Crudden, Chem. Rev. 2019, 119, 4986–5056.
[14] P. Dash, R. W. J. Scott, Chem. Commun. 2009, 812–814.
[15] C. J. Serpell, J. Cookson, D. Ozkaya, P. D. Beer, Nat. Chem. 2011, 3, 478–
483.
[16] A. Nazirov, A. Pestov, Y. Privar, A. Ustinov, E. Modin, S. Bratskaya,
Carbohydr. Polym. 2016, 151, 649–655.
[17] P. Pospiech, J. Chojnowski, U. Mizerska, T. Makowski, K. Strzelec, N.
Sienkiewicz, Appl. Organomet. Chem. 2016, 30, 399–407.
[18] F. H. Haghighi, H. Hadadzadeh, H. Farrokhpour, RSC Adv. 2016, 6,
109885–109896.
catalysis,
a feature which we attribute to the hard-soft
mismatch of nitrogen ligands with the metals chosen here. In
particular, selectivity was improved, with lower levels of
dehalogenation compared to commercial Pd/C catalysts. We
have also shown the use of a functional imidazole ligand can
produce nanoparticle networks which report chloride binding
in aqueous solvents through a non-linear response of their
plasmon resonance band. Functionalised imidazoles are versa-
tile ligands for metal nanoparticles and their finely tuned
surface interaction have potential to lead to new advances in
catalysis and sensing.
[19] W. Hossain, C. Sen, C. Sinha, U. K. Sarkar, J. Nanosci. Nanotechnol. 2019,
19, 3583–3590.
[20] R. C. Knighton, M. R. Sambrook, J. C. Vincent, S. A. Smith, C. J. Serpell, J.
Cookson, M. S. Vickers, P. D. Beer, Chem. Commun. 2013, 49, 2293–2295.
[21] P. Migowski, J. Dupont, Chem. Eur. J. 2007, 13, 32–39.
[22] K. R. Geethalakshmi, X. Yang, Q. Sun, T. Y. Ng, D. Wang, RSC Adv. 2015,
5, 88625–88635.
[23] B. Lu, F. Lu, L. Ran, K. Yu, Y. Xiao, Z. Li, F. Dai, D. Wu, G. Lan, Int. J. Biol.
Macromol. 2018, 119, 505–516.
[24] R. Imperatore, G. Carotenuto, M. A. Di Grazia, I. Ferrandino, L. Palomba,
R. Mariotti, E. Vitale, S. De Nicola, A. Longo, L. Cristino, J. Biomed. Mater.
Res. Part A 2015, 103, 1436–1446.
[25] N. Li, P. Zhao, M. E. Igartua, A. Rapakousiou, L. Salmon, S. Moya, J. Ruiz,
D. Astruc, Inorg. Chem. 2014, 53, 11802–11808.
ChemistryOpen 2020, 9, 683–690
www.chemistryopen.org
689
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA

Full Papers
doi.org/10.1002/open.202000145
ChemistryOpen
[26] X. Liu, S. Mu, Y. Long, G. Qiu, Q. Ling, H. Gu, W. Lin, Catal. Lett. 2019,
149, 544–551.
[27] J. Cookson, Platinum Met. Rev. 2012, 56, 83–98.
[28] S. V. Voitekhovich, V. Lesnyak, N. Gaponik, A. Eychmüller, Small 2015,
11, 5728–5739.
[33] T. Mitsudome, Y. Mikami, H. Mori, S. Arita, T. Mizugaki, K. Jitsukawaa, K.
Kaneda, Chem. Commun. 2009, 3258–3260.
[34] A. David, M. A. Vannice, J. Catal. 2006, 237, 349–358.
[35] C. J. Serpell, J. Cookson, A. L. Thompson, P. D. Beer, Chem. Sci. 2011, 2,
494–500.
[36] C. Levard, S. Mitra, T. Yang, A. D. Jew, A. R. Badireddy, G. V. Lowry, G. E.
Brown, Environ. Sci. Technol. 2013, 47, 5738–5745.
1
2
3
4
5
[29] J. A. Creighton, D. G. Eadon, J. Chem. Soc. Faraday Trans. 1991, 87,
3881–3891.
[30] M. Y. Smirnov, A. V. Kalinkin, A. V. Bukhtiyarov, I. P. Prosvirin, V. I.
Bukhtiyarov, J. Phys. Chem. C 2016, 120, 10419–10426.
[31] F. Sanchez, L. Bocelli, D. Motta, A. Villa, S. Albonetti, N. Dimitratos, Appl.
Sci. 2020, 10, 1752.
6
7
8
Manuscript received: May 14, 2020
[32] W.-Y. Xu, T.-Y. Gao, J.-H. Fan, J. Hazard. Mater. 2005, 123, 232–241.
Revised manuscript received: May 18, 2020
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
ChemistryOpen 2020, 9, 683–690
www.chemistryopen.org
690
© 2020 The Authors. Published by Wiley-VCH Verlag GmbH & Co. KGaA