Alloying Gold with Copper Makes for a Highly Selective Visible-Light Photocatalyst for the Reduction of Nitroaromatics to Anilines

Article abstract of DOI:10.1021/acscatal.5b02643

Finely control of product selectivity is an essential issue in organic chemical production. In the synthesis of functionalized anilines via reduction of the corresponding nitroarenes, the challenge is to selectively reduce only the nitro group in the presence of other reducible functional groups in nitroarene molecules at a high reaction rate. Normally, the nitroarene is reduced stepwise through a series of intermediates that remain as byproducts, increasing the aniline synthesis cost. Here we report that alloying small amounts of copper into gold nanoparticles can alter the reaction pathway of the catalytic reduction under visible-light irradiation at ambient temperature, allowing nitroaromatics to be transformed directly to anilines in a highly selective manner. The reasons for the high efficiency of the photocatalytic reduction under these comparatively benign conditions as well as the light-excited reaction mechanisms are discussed. This photocatalytic process avoids byproducts, exhibits a high reaction rate and excellent substituent tolerance, and can be used for the synthesis of many useful functionalized anilines under environmentally benign conditions. Switching of the reaction pathway simply by tailoring the bimetallic alloy NPs of the photocatalysts is effective for engineering of product chemoselectivity.

Full text of DOI:10.1021/acscatal.5b02643

Research Article
pubs.acs.org/acscatalysis
Alloying Gold with Copper Makes for a Highly Selective Visible-Light
Photocatalyst for the Reduction of Nitroaromatics to Anilines
†
†
†
‡
†,§
∥
‡
Qi Xiao, Sarina Sarina, Eric R. Waclawik, Jianfeng Jia, Jin Chang, James D. Riches, Haishun Wu,
⊥
,†
Zhanfeng Zheng, and Huaiyong Zhu*
†School of Chemistry, Physics and Mechanical Engineering, Science and Engineering Faculty, Queensland University of Technology,
Brisbane, QLD 4001, Australia
‡
School of Chemical and Material Science, Shanxi Normal University, Linfen 041004, China
§
Institute of Advanced Materials (IAM), National Jiangsu Synergistic Innovation Center for Advanced Materials (SICAM), Nanjing
Tech University, Nanjing 211816, China
∥
Institute for Future Environments & School of Earth, Environmental and Biological Sciences, Queensland University of Technology,
Brisbane, QLD 4001, Australia
⊥
State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Taiyuan 030001, China
*
S Supporting Information
ABSTRACT: Finely control of product selectivity is an essential issue in organic
chemical production. In the synthesis of functionalized anilines via reduction of the
corresponding nitroarenes, the challenge is to selectively reduce only the nitro group in
the presence of other reducible functional groups in nitroarene molecules at a high
reaction rate. Normally, the nitroarene is reduced stepwise through a series of
intermediates that remain as byproducts, increasing the aniline synthesis cost. Here we
report that alloying small amounts of copper into gold nanoparticles can alter the
reaction pathway of the catalytic reduction under visible-light irradiation at ambient
temperature, allowing nitroaromatics to be transformed directly to anilines in a highly
selective manner. The reasons for the high efficiency of the photocatalytic reduction
under these comparatively benign conditions as well as the light-excited reaction
mechanisms are discussed. This photocatalytic process avoids byproducts, exhibits a high
reaction rate and excellent substituent tolerance, and can be used for the synthesis of
many useful functionalized anilines under environmentally benign conditions. Switching of the reaction pathway simply by
tailoring the bimetallic alloy NPs of the photocatalysts is effective for engineering of product chemoselectivity.
KEYWORDS: Au−Cu alloy, photocatalysis, reduction, reaction pathways, visible light
1
1
INTRODUCTION
compound to the aniline are slow steps. These steps involve
cleavage of the last O−N bond and addition of H to the azo
group, respectively, and high reaction temperatures and high
pressures of hydrogen are used to accelerate these processes,
even though excellent catalysts have been discovered. These
conditions can reduce other functional groups in the reactants.
We previously reported that Au nanoparticles (NPs)
■
Nitroaromatic compounds are among the most important
industrial chemicals in use today. For example, approximately
5% of nitroarenes are consumed in the production of anilines,
which are important precursors to rubber chemicals, pesticides,
dyes, explosives, and pharmaceuticals. An essential issue with
this reduction process is achieving high product selectivity.
1
9
2
3
−6
supported on catalytically inert zirconium oxide (ZrO ) can
exhibit high catalytic activity for the selective transformation of
nitroaromatics to the corresponding azo compounds in
2
Many useful anilines contain reducible functional groups. It is a
challenge to reduce only the nitro group at a high reaction rate
while maintaining any other reducible groups in the reactants
12
3
,5,6
isopropanol solution under visible-light irradiation at 40 °C.
unchanged.
The selective reduction process has been
The supported Au NPs strongly absorb visible light by means
of the localized surface plasmon resonance (LSPR) effect and
efficiently channel the light energy to the reactant, inducing
extensively studied using a variety of heterogeneous catalytic
3
−6
systems.
In those reported catalytic systems, the reduction
proceeds via the well-known pathways first proposed by Haber
13−15
7
reaction under moderate reaction conditions.
catalytic reduction using supported Au NPs also proceeds via
The photo-
in 1898 (Figure 1). The nitro group of the reactant is reduced
to the corresponding aniline in a stepwise manner through
nitroso and hydroxylamine intermediates.
8
−11
Condensation of
the reactive intermediates yields coupling derivatives such as
Received: November 22, 2015
Revised: January 18, 2016
10,11
azobenzene and azoxybenzene,
but the conversions from
the azoxy compound to the azo compound and from the azo
©
XXXX American Chemical Society
1744
DOI: 10.1021/acscatal.5b02643
ACS Catal. 2016, 6, 1744−1753

ACS Catalysis
Research Article
Figure 1. Reaction pathways for the reduction of aromatic nitro
compounds to the corresponding anilines proposed by Haber: (a)
direct route; (b) coupling route.
Haber’s pathways, as intermediate azoxy compound and trace
over-reduced product aniline were observed. When this
photocatalytic system is used to produce functionalized anilines
from nitroarenes, high selectivity for the aniline product is
difficult to achieve.
To address the issue of high selectivity for the aniline, the
discovery of catalytic systems that proceed by a different
reaction mechanism than Haber’s long established pathways
could prove useful. To develop a viable new route that
produces anilines as the sole product, solving the high
selectivity issue is a fundamentally important consideration.
In the present study, we have found one such photocatalytic
system. Gold−copper alloy NPs with a small Cu fraction were
prepared, and it was found that reduction catalyzed by these
Au−Cu alloy NP photocatalysts under visible-light irradiation
at ambient temperature does not follow Haber’s reaction
pathways but rather yields the corresponding anilines directly as
the sole product. The photocatalytic process can be used to
selectively reduce the nitro group in the presence of other
reducible groups in the reactant nitroarenes to yield function-
alized anilines, which has remained a challenge in controlling
this reduction process.
Figure 2. Photocatalytic performance for the reduction of nitro-
benzene. (a) Au−Cu alloy NPs @ZrO
catalysts with varied
2
composition. (b) The dependence of the product selectivity to aniline
on the Au−Cu alloy NPs @ZrO2 catalysts with varied Au/Cu
composition under visible-light irradiation.
selectivity of the nitrobenzene reduction catalyzed by
Au Cu @ZrO was compared with that for the reaction
2.6
0.4
2
catalyzed by the Au @ZrO catalyst under the same
3
.0
2
conditions, as shown in Figure 3a,b. Control experiments in
the dark showed similar trends in the product selectivity but
lower conversion rates under identical reaction conditions
(Figure 3c,d).
RESULTS AND DISCUSSION
■
A series of catalysts consisting of Au−Cu alloy NPs supported
on ZrO were prepared as the photocatalysts (for experimental
2
details, see the Experimental Section). ZrO was chosen as the
It is evident that the direct reduction of nitrobenzene to
aniline using the Au−Cu alloy NP catalyst is distinctly different
from the reductive coupling of nitrobenzene catalyzed by the
Au NP catalyst. When Au Cu @ZrO catalyst was used
2
support because it is a catalytically inert material in the present
study and can be used to achieve a homogeneous distribution
12
of metal NPs on the surface. The total metal amount of the
catalyst was maintained at 3 wt %, which was an optimal metal
2.6
0.4
2
(Figure 3a), the selectivity for aniline approached 100% from
the initial reaction stage and remained at 100% until the
reaction was complete. In contrast, when the pure Au NP
photocatalyst was used, there was no aniline product in the
early reaction stage (Figure 3b). The nitrobenzene conversion
reached 100% within 3 h, which was faster than occurred with
Au−Cu alloy NPs, but the main product was azobenzene.
Azoxybenzene was generated in the initial stage of the reaction
(0.5 h). As the reaction proceeded, the selectivity for
azoxybenzene dropped substantially while the selectivity for
azobenzene significantly increased (0.5−2 h). The azobenzene
selectivity reached 100% within 2 h and then declined as aniline
emerged as the further-reduced product. The aniline selectivity
was also much lower at the end of the 6 h reaction compared
with that using Au−Cu alloy NPs as the photocatalyst. The
complex sequence of reactions and intermediates formed in this
12
loading on the basis of our previous work, and the Au:Cu
ratio was varied to obtain several alloy catalysts (labeled
Au3−xCu @ZrO ; e.g., the Au Cu @ZrO catalyst contained
x
2
2.6
0.4
2
2
.6 wt % Au and 0.4 wt % Cu). The Au Cu @ZrO catalyst
2.6 0.4 2
exhibited superior photocatalytic activity, and the aniline yield
under irradiation was approximately 3.6-fold that of the control
reaction in the dark (Figure 2a). The photocatalytic perform-
ance of the catalysts with varied Au:Cu ratios for the reduction
was also investigated, and the results are summarized in Figure
2
a. Importantly, the catalysts exhibited unique selectivity for
aniline, whereas reactions catalyzed by Au NPs have produced
mainly azobenzene. The selectivity behavior changed as a
function of Au/Cu composition: when the Cu content was
.4−0.6 wt % (and the corresponding Au content was 2.6−2.4
wt %), the product was solely aniline (Figure 2b). The product
0
1
745
DOI: 10.1021/acscatal.5b02643
ACS Catal. 2016, 6, 1744−1753

ACS Catalysis
Research Article
Figure 3. (a, b) Time−conversion plots for nitrobenzene reduction under visible-light irradiation using (a) Au Cu @ZrO and (b) Au @ZrO as
2.6
0.4
2
3.0
2
the catalyst. (c, d) Time−conversion plots for nitrobenzene reduction in the dark using (c) Au Cu @ZrO and (d) Au @ZrO as the catalyst.
2.6
0.4
2
3.0
2
The reactions were conducted in an argon atmosphere at 40 °C using 2 mL of isopropyl alcohol mixed with 0.025 mmol of KOH, 0.1 mmol of
2
nitrobenzene, and 50 mg of catalyst. The irradiation intensity was 0.5 W/cm , and the reaction time was 6 h. For the dark reactions, all the reaction
conditions were identical to those for the photocatalytic reactions except that the reaction tubes were kept out of the light.
process with the pure Au NP catalyst indicates that the
photocatalytic coupling follows the pathways proposed by
aniline yield. Both reactants exist in the liquid phase and have
adequate access to the catalyst dispersed in the liquid. This is
part of the reason why the reaction can proceed at ambient
pressure, which leads to significant savings in the cost of the
reactor. The catalyst can be readily recycled after reactions
without significant loss of activity (for details, see Figure S1 in
the Supporting Information (SI)).
7
,10,11
Haber.
alloy NPs does not.
In contrast, the reduction catalyzed by Au−Cu
The photocatalytic coupling of nitrobenzene on the pure Au
NP surface does not consume the reducing agent except in the
1
2
initial stage, while the direct reduction of nitrobenzene to
aniline should. Isopropyl alcohol was the reducing agent in the
photocatalytic reduction, which provided hydrogen and was
oxidized to acetone. The content of acetone in the reaction
system catalyzed by the Au−Cu alloy NPs increased gradually
as the reaction proceeded (Figure 4), in proportion with the
Figure 5a shows a typical transmission electron microscopy
(
TEM) image of the Au Cu @ZrO catalyst. The mean size
2.6 0.4 2
of the alloy NPs was 5 nm. The lattice fringes of the Au−Cu
alloy NPs can be observed in the high-resolution TEM (HR-
TEM) image in Figure 5b. The lattice fringe spacing of 0.22 nm
corresponds to the interplanar distance of Au−Cu alloy (111)
16
planes. Energy-dispersive X-ray spectroscopy (EDS) line scan
analysis of a Au−Cu alloy NP indicated that Au and Cu were
uniformly distributed in the alloy NP (Figure 5a inset). This
was also apparent in the EDS mapping and X-ray photoelectron
spectroscopy (XPS) analysis provided in Figures S2 and S3 in
the SI, which showed that the Au content was much higher
than the Cu content. The characterization results confirmed the
formation of Au Cu alloy NPs.
2.6
0.4
A challenge for Cu-based NP catalysts is that O in air can
2
oxidize surface Cu atoms of the alloy NPs, which results in a
17
rapid loss of activity during reactions exposed to air. It is
known that the Cu oxidation rate depends on the alloy NP
composition and that increasing amounts of Au can improve
1
8
the stability of the catalyst. In bulk Au−Cu alloys, Au can
Figure 4. Time−conversion plots for acetone formation during the
reduction of nitrobenzene under visible-light irradiation using the
Au Cu @ZrO and Au @ZrO catalysts.
protect Cu from oxidation by limiting Cu O surface island
growth. Thus, we anticipated that Au−Cu alloy NPs with a
2
19
2.6
0.4
2
3.0
2
1
746
DOI: 10.1021/acscatal.5b02643
ACS Catal. 2016, 6, 1744−1753

ACS Catalysis
Research Article
Table 1. Reaction Pathway Study of the Catalytic
Performance Using Au Cu @ZrO and Monometallic
2
.6
0.4
2
Au @ZrO Catalysts with Various Intermediates as
3
.0
2
a
Substrates
selectivities (%)
entry
substrate
catalyst
conv. (%)
3
4
5
Figure 5. Characterization of the photocatalyst. (a) Typical TEM
image of Au Cu @ZrO . The inset shows a line profile analysis
b
2.6
0.4
2
1
1
Au−Cu
Au
100
100
100
100
100
100
70
5
3
4
85
83
75
0
providing information about the elemental composition and Au/Cu
distribution of the NP. (b) HR-TEM image of a typical Au Cu alloy
b
2
1
4
c
2.6
0.4
3
2
Au−Cu
Au
19
15
4
6
NP. (c) XPS profiles of Cu and Au species in the Au Cu alloy NPs.
c
2.6
0.4
4
2
85
3
(
d) UV/vis spectra of the Au Cu @ZrO and monometallic Au @
c
2.6
0.4
2
3.0
5
1 + 2
1 + 2
4
Au−Cu
Au
93
0
ZrO catalysts.
2
c
6
8
92
0
d
7
Au−Cu
Au
100
10
0
low Cu content would be stable in air. The XPS spectra in
Figure 5c confirmed this. The binding energies of Cu 2p1/2 at
around 952.0 eV and Cu 2p at 932.5 eV can be mainly
d
e
8
4
55
−
0
a
The reactions were conducted under an argon atmosphere at 40 °C
using 2 mL of isopropyl alcohol mixed with 0.025 mmol of KOH, 0.1
mmol of substrate (for entries 5 and 6, 0.05 mmol of substrate 1 and
.05 mmol of substrate 2 were used), and 50 mg of catalyst at an
irradiation intensity of 0.5 W/cm for the indicated reaction times. The
conversions and selectivities were calculated from the amounts of
products formed and reactants converted as measured by gas
chromatography (GC). Reaction time = 3 h. Reaction time = 1 h.
Reaction time = 16 h. The other product selectivity was 90%
3
/2
0
20,21
attributed to the Cu state.
loading of Cu and low signal intensity, analysis of the Cu Auger
23
However, because of the low
0
0
+
22
2
lines to assign Cu and Cu was not possible. Sugano et al.
suggested that in the Au−Cu alloy NP photocatalyst the
partially oxidized surface Cu atoms could be successfully
reduced by plasmon-activated Au atoms to maintain the Au−
Cu alloying effect. Thus, even if a very small amount of copper
oxide exists on the surface, it could be reduced to the metallic
b
c
d
e
hydroazobenzene.
2
3,24
state under irradiation.
The binding energies for Au are
25
identical to those in bulk gold metal.
azoxybenzene was observed (entry 4). The reaction proceeded
rapidly and was complete within 1 h, indicating that
phenylhydroxylamine was highly reactive. The small amounts
of aniline in the product (19% and 15%) could have formed via
a disproportionation pathway simultaneously with the coupling
route as proposed in Figure 1, which is based on the
The light absorption property is an important characteristic
of a photocatalyst. The monometallic Au NP catalyst exhibits a
distinctive light absorption band at 525 nm due to the LSPR
effect (Figure 5d). In comparison, a red shift of the LSPR band
was observed for the Au−Cu alloy NPs, which is also evidence
26
10
of alloying (Figure 5d and Figure S4 in the SI). Character-
izations of the photocatalysts, however, could not provide a
direct explanation for the dramatic change in product
selectivity. To gain insight into the reaction pathway of the
process with the alloy photocatalyst, reductions of the possible
intermediates (nitrosobenzene, phenylhydroxylamine, and their
mixture) were conducted using Au Cu @ZrO or mono-
literature.
Further experiments using a mixture of nitrosobenzene and
phenylhydroxylamine as the reactant gave similar results (Table
1, entries 5 and 6). The reaction between nitrosobenzene and
phenylhydroxylamine occurred spontaneously (without cata-
lyst) and rapidly as soon as they were mixed, yielding mainly
azoxybenzene. Thus, we can exclude the formation of
intermediates such as nitrosobenzene and phenylhydroxylamine
in the reduction using Au Cu @ZrO catalyst, because if they
were formed in the reaction, either of them would result in the
coupling products. Comparison of the results of entry 2 with
those of entries 4 and 6 indicates that phenylhydroxylamine is
the key intermediate to afford the main product azobenzene in
the reaction catalyzed by monometallic Au NPs. Moreover, the
Au−Cu alloy NPs can also catalyze the reduction of
azobenzene to the sole product aniline with 100% selectivity
within 16 h under visible-light irradiation, whereas the pure Au
NPs are slow to yield the aniline product even after prolonged
reaction time (entries 7 and 8). This suggests that the
2
.6
0.4
2
metallic Au @ZrO as the catalyst, with the other
3
.0
2
experimental conditions maintained identical to those for the
nitrobenzene reduction. The results are presented in Table 1.
As shown in entries 1 and 2, when nitrosobenzene was the
reactant, azoxybenzene was the main product no matter which
catalyst was used, and the product selectivities were very
similar. However, when phenylhydroxylamine was the reactant
2.6
0.4
2
(
entries 3 and 4), a striking difference in the product
selectivities was observed: the main product with the alloy
NP catalyst was azoxybenzene, while only a small fraction of
azobenzene was detected (entry 3). The main product with the
monometallic Au NP catalyst was azobenzene, and no
1
747
DOI: 10.1021/acscatal.5b02643
ACS Catal. 2016, 6, 1744−1753

ACS Catalysis
Research Article
formation of aniline on the Au−Cu alloy surface is highly
favored, while pure Au NPs exhibit a much weaker ability to
catalyze the hydrogenation of azobenzene to aniline, causing
the Au NP surface to exhibit selectivity for azobenzene in the
reduction process.
We propose a reaction pathway that would explain this
difference, as shown in Figure 6. Analysis confirmed that the
Figure 7. DFT-simulated (a) nitrobenzene and (b) aniline on the Au
and Au−Cu alloy surfaces. The nitrobenzene molecule has much
stronger interaction with the Au−Cu alloy surface than with the pure
Au surface, as shown in panel (a). When an aniline molecule is
adsorbed on the Au−Cu alloy surface, the N atom in the aniline
molecule is attracted to the Cu atom, and the length of the N−Cu
bond is 2.23 Å. However, when an aniline molecule is close to the
surface of the Au NP, the N atom is repelled, as shown in panel (b).
The interaction of the N atom of the nitrobenzene or aniline molecule
with a Cu atom in the Au−Cu alloy NP is much stronger than that
with Au atoms on the Au NP surface.
Figure 6. Proposed reaction pathway for the reduction of nitro-
benzene in the present study.
monometallic Au NPs catalyzed the reaction via Haber’s
coupling route. By this pathway, the final product aniline was
obtained mainly after the coupling step (azoxy→ azo→
hydrazo→ aniline). In the reaction system of the present
study, H−Au species form on the surface of metal NPs as
isopropanol provides hydrogen while being oxidized to
1
2
acetone. The H−Au species can react with nitrobenzene to
yield nitrosobenzene. According to Haber’s mechanism,
nitrosobenzene is a reactive intermediate that is transformed
molecules is a prerequisite for the coupling reaction, so in
contrast to the Au−Cu alloy, the reactant molecules
(intermediate molecules with N atoms) on a pure Au surface
are labile and can move for coupling. The reactant molecules
immobilized on the Cu atoms are directly reduced to aniline.
The product aniline is also much more stable on the Au−Cu
alloy surface than on the pure Au surface. The simulation
results confirm that the formation of aniline on the Au−Cu
alloy surface is favored thermodynamically compared with that
on the Au NP surface. Moreover, in regard to the difference
between the surface electronic properties of the alloy NPs and
pure Au NPs, the electronegativity of Au (2.54) is much higher
10,11
to phenylhydroxylamine rapidly.
This is the reason why
nitrosobenzene was not detected during the reaction. The high
concentration of phenylhydroxylamine favors the formation of
azobenezne on the Au NPs (Table 1, entries 4 and 6). Blaser
proposed that Haber’s coupling route is favored under basic
1
0
conditions. This is certainly true in the case of the
monometallic Au catalyst. When the Au−Cu alloy NPs were
used as the photocatalyst, nitrobenzene was transformed
directly to aniline as the sole product, and we did not detect
any byproduct such as azobenzene or other intermediates
during the reaction. The detailed mechanistic understanding of
this phenomenon is still unclear and may need further future
work to explore in detail. Interestingly, even under basic
conditions (when KOH was added to the reaction system) the
reduction catalyzed by Au−Cu alloy NPs followed the direct
reduction route.
27
than that of Cu (1.90). There are electron-rich sites of Au and
slightly positively charged sites of Cu on the surface of the alloy
NPs. The interaction between aniline and the positively
charged sites is stronger than that between aniline and the
pure Au NP surface. Hence, thermodynamically the formation
of aniline on the alloy surface is more favorable than that on the
pure Au surface, and thus, alloying Au with Cu in the present
study always enhances the product selectivity for aniline.
The isopropyl alcohol was the reducing agent and was first
To further understand the reason why aniline was the sole
product of the reaction catalyzed by Au−Cu alloy NPs, we
performed a density functional theory (DFT) simulation study
of the adsorption of the reactant nitrobenzene and the product
aniline on the surfaces of Au and Au−Cu clusters, respectively
1
2,28
oxidized to acetone, yielding H−Au NP
or H−Cu NP
20
species. Given the high Au content in the alloy NPs, the H−
Au NP species are considered to play the major role in reacting
with the oxygen atoms of N−O bonds in the nitrobenzene to
induce the reduction on the surface of the alloy NPs.
(Figure 7 and Table S1 in the SI). The DFT simulation results
indicate that both the reactant nitrobenzene and the product
aniline adsorb more strongly to the Au−Cu alloy surface than
to the pure Au surface. The interaction of the N atom of the
nitrobenzene or aniline molecule with a Cu atom in the Au−Cu
alloy NPs is much stronger than that with Au atoms on the NP
surface. This immobilizes the reactant molecules adsorbed at
Cu atom sites on the alloy NP surface. Mobility of the reactant
The general applicability of the Au−Cu alloy NP-photo-
catalyzed reduction of nitroaromatics was studied. Table 2
shows the results for the reduction of nitroaromatics to anilines
using the Au Cu @ZrO catalyst under visible-light irradi-
2.6
0.4
2
ation and in the dark. Visible-light irradiation promoted each
reaction, as the yield of the light-irradiated reaction was much
1
748
DOI: 10.1021/acscatal.5b02643
ACS Catal. 2016, 6, 1744−1753

ACS Catalysis
Research Article
Table 2. Photocatalytic Reduction of Substituted
Nitroaromatics Using the Au Cu @ZrO Catalyst
to be able to reduce nitro groups in the presence of alkenes
(Table 2, entry 9) but could also reduce them in the presence
of alkynes without detectable concurrent reduction of the
unsaturated unit (entry 12). Nitrile and keto substituents and
heteroarene groups were unaffected as well. Paracetamol (N-
acetyl-p-aminophenol) is a major ingredient in numerous
a
2
.6
0.4
2
29
analgesic, cold, and flu remedies. Typically, it is produced by
reduction of 4-nitrophenol to 4-aminophenol followed by
29
acetylation with acetic anhydride. The reduction of 4-
nitrophenol is the key step, and H or NaBH are the widely
2
4
used reductants. The reaction can be easily monitored by UV/
30−32
vis spectroscopy.
In the present study, we found that 4-
nitrophenol can be selectively reduced to 4-aminophenol using
the Au−Cu alloy NP catalyst under irradiation (Figure S5 in
the SI).
Evidently, the Au−Cu alloy NP catalyst is an efficient
photocatalyst for selective reduction of nitroaromatics directly
to functionalized anilines using visible light under mild reaction
conditions compared with those in the reported elegant
3
−6
works. To achieve a higher catalytic rate, we simply increased
the reaction temperature slightly from 40 to 60 °C for several
substrates under typical reaction conditions (Table S2 in the
SI). The reactions achieved 100% conversion within 45 min
and still maintained the high chemoselectivity to reduce the
nitro groups. This is due to the continuum of metal electron
energy levels in the metal NP photocatalysts, which provides
the capacity to couple the stimuli of irradiation and heat to
33,34
drive reactions.
This is an apparent merit compared with
traditional heterogeneous catalysts and semiconductor photo-
catalysts.
In the photocatalytic process, the light drives the reduction
and reductive coupling by exciting metal electrons, and these
light-excited electrons (so-called hot electrons) provide the
activation energy required for cleavage of the N−O bond. This
conclusion is supported by the fact that the catalytic activity of
the Au Cu @ZrO photocatalyst for the reduction of
2
.6
0.4
2
nitrobenzene to aniline significantly depends on the irradiation
intensity (Figure 8a). When the irradiation intensity was raised,
the aniline yields increased, and at higher intensity the
contribution of irradiation to the overall reaction rate was
greater. A higher irradiation intensity excites more hot electrons
to populate levels above the Fermi level of the metal NPs,
a
The reactions were conducted under an argon atmosphere at 40 °C
using 2 mL of isopropyl alcohol mixed with 0.025 mmol of KOH, 0.1
mmol of reactant, and 50 mg of Au Cu @ZrO catalyst. The
irradiation intensity was 0.5 W/cm , and the reaction time was 6 h. In
all cases, the product selectivity for the aniline was 100%. The yields
were calculated from the amounts of products formed and reactants
33−39
2.6
0.4
2
which can facilitate the reactions, as reported for Au NPs.
2
Moreover, when the irradiation intensity was maintained
constant, we found that the reaction rate was higher at the
wavelength absorbed more strongly by the catalyst for a
number of reductions, as shown in the action spectra of the
b
converted as measured by GC. 100 mg of Au Cu @ZrO catalyst
2.6
0.4
2
c
was used, and the reaction temperature was 60 °C. 0.05 mmol of
reactant, 2 mL of 1:1 isopropyl alcohol/tetrahydrofuran mixed solvent,
and 100 mg of Au Cu @ZrO catalyst were used. The reaction
temperature was 60 °C, and the reaction time was 48 h. The reaction
temperature was 60 °C, and the reaction time was 16 h.
20,35−37
reactions.
Such an action spectrum is shown in Figure
8b. In the action spectrum, the apparent quantum yield (AQY)
of the reduction is closely related to the irradiation wavelength.
The action spectrum for the reduction of nitrobenzene to
aniline using the Au Cu @ZrO photocatalyst shows that the
2.6
0.4
2
d
2.6
0.4
2
higher than that of the reaction in the dark in every case. The
products were the corresponding anilines, and no coupling
products such as azo or azoxy compounds were detected. For
the reduction of halogen-containing nitroaromatics, dehaloge-
nation could sometimes be inevitable. In our case, the
dehalogenation was suppressed, and quantitative conversion
of these substrates was realized (entries 3−5).
AQY for aniline matches well the light absorption of the alloy
NP catalyst. The highest yield is achieved under irradiation at
the wavelength where the alloy NPs have the most intense
absorption due to the LSPR effect.
The hot electrons activate the reaction by two mechanisms:
First, the hot electrons can be transferred to the molecular
orbital of the reactant molecule adsorbed on the metal NP
40
The most challenging substrates are those that bear other
easily reducible moieties. The reduction of these substrates
allows production of important products: functionalized
anilines. The Au Cu @ZrO catalyst was not only observed
surface that is associated with the chemical bonds to be
cleaved (here the N−O bonds). There is an energetic
requirement for hot electron transfer: the hot electrons must
have an energy above a threshold that depends on the energy
2.6
0.4
2
1
749
DOI: 10.1021/acscatal.5b02643
ACS Catal. 2016, 6, 1744−1753

ACS Catalysis
Research Article
Figure 8. Impact of light intensity and wavelength. (a) Dependence of the catalytic activity of the Au Cu @ZrO photocatalyst for the reduction
2.6
0.4
2
of nitrobenzene to aniline on the intensity of the light irradiation. The percentages inside the figure show the contributions of the light irradiation
effect. (b) Action spectrum for the photocatalytic reduction of nitrobenzene using the Au Cu @ZrO catalyst. (c) Action spectrum for the
2.6
0.4
2
photocatalytic reduction of 4-nitrobenzyl alcohol using the Au Cu @ZrO catalyst. (d) Hot electron distributions of metal NPs under irradiation
2.6
0.4
2
at different wavelengths (400 and 530 nm) and schemes showing how the hot electrons contribute to activation of nitrobenzene and 4-nitrobenzyl
alcohol reactant molecules. The hot electrons with energies above the energy of the LUMO of the molecules adsorbed on the alloy NP can be
transferred into those orbitals, thereby potentially inducing reaction. The calculated HOMOs and LUMOs for nitrobenzene and 4-nitrobenzyl
alcohol are shown in Figure S6 in the SI. The LUMO of 4-nitrobenzyl alcohol (LUMO 2) is higher than that of nitrobenzene (LUMO 1), resulting
in two thresholds (thresholds 1 and 2) for activation of the molecules. Thus, LUMO 2 requires light excitation of energetic electrons to higher
energy levels to activate the molecule. As the strong chemisorption of adsorbates involves resonant electronic transitions between hybridized metal
and adsorbate states, the relative positions of LUMO 1 for nitrobenzene and LUMO 2 for 4-nitrobenzyl alcohol with respect to the Fermi level are
only qualitative and schematically show the relative positions of the LUMOs of the chemisorbed adsorbates.
level of the molecular orbital associated with the N−O
although the strongest light absorption of the catalyst appears
at 560 nm. It is also noted that the AQY values at wavelengths
of 590 and 620 nm are approximately one-sixth of the values at
400 and 470 nm, while the light absorption at the longer
wavelengths is about 20% less than that at the shorter
wavelengths. Hence, the photothermal effect cannot have
induced such a large difference in the AQY values. Actually, the
AQY values at wavelengths of 590 and 620 nm provide an
approximate measure of the photothermal effect. Evidently for
this reaction, cleavage of the N−O bonds is predominantly
driven by hot electron transfer.
4
1
bonds. Hot electrons with sufficient energy can be transferred
to the reactant molecules and weaken the N−O bonds,
resulting in the reaction. Second, although low-intensity
illumination of plasmonic metal NPs most likely does not
play a role in inducing chemistry, particularly in continuous
4
2
flow, the hot electrons can dissipate their energy via
electron−phonon interactions to heat the lattice of the
4
1
NPs, especially when the hot electrons do not have sufficient
energy for the direct transfer to the adsorbed reactant
4
1
molecules.
If the light energy absorbed by the hot electrons dissipates
entirely as heat in the NPs, the AQY depends on the light
absorption at all wavelengths. The higher AQY is observed at
the wavelength more strongly absorbed by the catalyst.
Nonetheless, if the energy threshold for the hot electron
transfer is low, such a dependence can also be observed. The
reduction of nitrobenzene is likely to have a low energy
threshold (Figure 8b).
We noted that the yield of 4-nitrobenzyl alcohol reduction is
much lower than that of nitrobenzene reduction (Table 2).
This means that it is more difficult to cleave the N−O bonds in
4-nitrobenzyl alcohol with hot electrons excited by the LSPR
light absorption compared with the corresponding process in
nitrobenzene. The alcohol group substituted on the aromatic
ring raises the energy of the lowest unoccupied molecular
orbital (LUMO) associated with the N−O bonds (Figure 8d).
The hot electrons generated in the metal by shorter-wavelength
irradiation have enough energy to effectively activate 4-
nitrobenzyl alcohol molecules for reaction (Figure 8d, left). It
Figure 8c displays a different action spectrum for the reaction
using 4-nitrobenzyl alcohol as the substrate: the reaction AQY
value is greater at shorter wavelengths (e.g., 400 and 470 nm)
1
750
DOI: 10.1021/acscatal.5b02643
ACS Catal. 2016, 6, 1744−1753

ACS Catalysis
Research Article
is therefore rational that the N−O bond-cleaving reaction is
predominately driven by hot electron transfer with a high
energy threshold in this case.
density distribution indicates electron transfer from the metal
NP to the LUMO of the adsorbed reactant molecules under
irradiation. These simulation results provide insight into the
dependence of the catalytic activity on the intensity and
wavelength of the incident light.
When the hot electrons are transferred into the LUMO of
3
5,37,39,42−44
the reactant molecules adsorbed on the metal NPs,
which is associated with the N−O bonds, the bonds are
weakened and the reduction is activated. Only the hot electrons
with energy above a threshold that depends on the energy of
the LUMO can be injected into the adsorbed reactant LUMO
to induce the reaction (see Figure 8d). The calculated energy
of the LUMO of 4-nitrobenzyl alcohol (−2.7 eV) is higher than
that of nitrobenzene (−2.9 eV), and thus, the Fermi level of Au,
which is −5.1 eV, lies about 2.4 and 2.2 eV below the LUMO
levels of 4-nitrobenzyl alcohol and nitrobenzene, respectively
CONCLUSIONS
■
This study has demonstrated that Au−Cu alloy NPs supported
on ZrO2 are an excellent photocatalyst for the selective
reduction of nitroaromatics to the corresponding functionalized
anilines as the sole products driven by visible-light irradiation
without requiring intensive heating and pressurized reagents.
The reasons for the high efficiency of the photocatalytic
reduction under these comparatively benign conditions have
been analyzed. During the photocatalytic reductions using this
catalyst, the nitro group is transformed to an amino group
directly rather than in a stepwise manner via a series of
intermediates as described by Haber. The interaction between
the nitro group and copper atoms is stronger than that between
the nitro group and gold atoms, immobilizing the reactant. This
effect facilitates direct reduction of the nitro group to an amino
group and impedes the coupling reaction between two
intermediate molecules. The change in reaction pathway and
mild reaction conditions are the essential reasons for the high
chemoselectivity for the anilines. The catalytic reaction rate can
be further optimized by a slight increase in the reaction
temperature, while the high chemoselectivity for reduction of
the nitro groups is maintained. This study reveals that product
chemoselectivity may be engineered simply by tailoring the
bimetallic alloy NPs of the photocatalysts. This approach
represents a promising new direction in the area of photo-
40
(
Figure S6 in the SI). Only the hot electrons with energies of at
least 2.4 eV above the Au Fermi level can induce the reaction of
-nitrobenzyl alcohol. The Fermi level of the alloy NPs is
4
slightly higher than that of Au, and thus, the energy required for
the injection is reduced (<2.4 eV). Shorter wavelengths more
effectively generate hot electrons with sufficient energy to be
transferred (Figure 8d, left). The light absorption due to the
LSPR effect is intense and can excite a large number of metal
electrons, but only to energy levels about 2 eV above the Fermi
level (Figure 8d, right). Hence, the LSPR absorption is not as
effective as the absorption of shorter wavelengths in driving the
reduction of 4-nitrobenzyl alcohol, while it effectively drives the
reduction of nitrobenzene because nitrobenzene possesses a
lower LUMO energy. We have to bear in mind that the above
discussion of the energetics of the reaction systems are
semiquantitative since the influence of the reactant adsorption
is not included. Another possible mechanism such as direct
charge transfer (direct excitation of an electron to the LUMO
without the formation of an excited electron distribution in the
45
catalytic chemical transformations. The knowledge acquired
45
in this study could be useful in the development of new
heterogeneous photocatalysts for the production of important
chemicals, functionalized anilines and others, and in under-
standing photocatalytic systems for organic reactions.
metal) may also be involved in the reaction. The present study
has limited capacity to distinguish between indirect and direct
charge transfer mechanisms for the reactions. Nonetheless, the
discussion reveals that the reduction reaction is mainly driven
by hot electron transfer from the alloy NP to the LUMO
associated with the N−O bonds.
EXPERIMENTAL SECTION
■
It is noted that high AQY values were found in the short-
wavelength range of the action spectra. In this range (400−450
nm), one photon can excite an electron (interband transition)
to a higher energy level than those excited by LSPR absorption
in Au−Cu NPs at the expense of LSPR excitation. These high-
energy hot electrons can subsequently generate more hot
electrons with lower energy by electron−electron collisions.
The higher AQY values suggest that the short-wavelength
photons can result in more effective adsorbate-specific energy
transfer to activate the N−O bonds on the NP surface. It has
been reported that size of the metal NP affects wavelength-
Chemicals. Zirconium(IV) oxide (ZrO , <100 nm particle
size, TEM), gold(III) chloride hydrate (HAuCl ·xH O,
99.999% trace metals basis), sodium borohydride (NaBH4,
2
4
2
≥98.0%), copper(II) nitrate hydrate (Cu(NO ) ·xH O,
3
2
2
≥99.9% trace metals basis) were used. All of the chemicals
employed in the experiments were purchased from Sigma-
Aldrich (unless otherwise noted) and used as received without
further purification. The water used in all of the experiments
was prepared by passage through an ultrapurification system.
Preparation of Catalysts. As a typical example, the
Au Cu @ZrO catalyst was prepared as follows: ZrO
2
2
.6
0.4
2
46
dependent photocatalytic reactions. Small NPs (approximately
.0 nm) have large fractions of surface atoms, making photon
powder (1.0 g) was dispersed in a mixture of HAuCl (13.2
mL, 0.01 M) and Cu(NO ) (6.25 mL, 0.01 M) aqueous
4
5
3
2
absorption likely to occur at surface atoms that are directly
bonded to adsorbates, which may facilitate the direct
photoexcitation of metal−adsorbate bonds to enhance the
solutions under magnetic stirring at room temperature. An
aqueous solution of lysine (3 mL, 0.1 M) was then added to the
mixture with vigorous stirring for 30 min, and the pH was 8−9.
To this suspension was added a freshly prepared aqueous
46
photocatalytic reaction on the NP surface under irradiation.
To further investigate the hot electron transfer process on
the Au−Cu alloy NP surface, we simulated the molecular
orbitals involved in the hot electron transfer processes for
nitrobenzene and 4-nitrobenzyl alcohol reactant molecules
adsorbed on the Au−Cu cluster surface (see Figure S7 in the
SI). The details of the time-dependent DFT (TD-DFT) results
are listed in Tables S3−S6 in the SI. The change in charge
solution of NaBH (2 mL, 0.35 M) dropwise. The mixture was
4
aged for 24 h, and then the solid was separated by
centrifugation, washed with water (three times) and ethanol
(once), and dried at 60 °C in a vacuum oven for 24 h. The
dried powder was subjected to thermal treatment in a mixture
of H (5 vol %) and Ar at 450 °C for 0.5 h. The obtained
2
powder was used directly as the Au Cu @ZrO catalyst. All
2.6
0.4
2
1
751
DOI: 10.1021/acscatal.5b02643
ACS Catal. 2016, 6, 1744−1753

ACS Catalysis
Research Article
of the other catalysts were prepared via the same methods with
different quantities of HAuCl4 and Cu(NO3)2 aqueous
solutions.
the catalytic performance at different irradiation wavelengths
(action spectrum experiments). The light intensity of the LED
light sources used for the wavelength-dependent experiments
2
Characterization of Catalysts. The sizes, morphologies,
and compositions of the catalyst samples were characterized by
TEM using a JEOL 2100 transmission electron microscope
equipped with a Gatan Orius SC1000 CCD camera and an
Oxford X-Max EDS instrument. The Au and Cu contents of the
prepared catalysts were determined by EDS technology using
an EDS attachment to an FEI Quanta 200 environmental
scanning electron microscope. Diffuse-reflectance UV/vis (DR-
UV/vis) spectra of the sample powders were examined using a
was 0.2 W/cm , and the light intensity was maintained constant
for each wavelength-dependent experiment. The other reaction
conditions were kept identical with the halogen lamp
photocatalytic reactions. The apparent quantum yield was
calculated as AQY = [(Ylight − Ydark)/(number of incident
photons)] × 100%, where Ylight and Ydark are the number of
products formed under light irradiation and in the dark,
respectively. The number of products formed, Y, was calculated
using the equation Y = y × (moles of reactant) × Avogadro
constant, where y is the product aniline yield for the reaction.
Varian Cary 5000 spectrometer with BaSO as a reference. The
4
light absorption data were normalized with OriginPro 8
software using the normaliziation method “normalize to [0,
ASSOCIATED CONTENT
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acscatal.5b02643.
■
1]”. The normalized data were plotted against wavelength to
*
S
get the light absorption spectrum of the catalyst. XPS data were
acquired using a Kratos Axis ULTRA X-ray photoelectron
spectrometer incorporating a 165 mm hemispherical electron
energy analyzer. The incident radiation was monochromatic Al
Kα X-rays (1486.6 eV) at 225 W (15 kV, 15 mA). Narrow high-
resolution scans were run with 0.05 eV steps and a 250 ms
dwell time. The base pressure in the analysis chamber was 1.0 ×
Figures S1−S8 and Tables S1−S7 (PDF)
AUTHOR INFORMATION
Corresponding Author
*E-mail: hy.zhu@qut.edu.au.
Notes
■
−
9
8
1
1
0
0
Torr, and during sample analysis the pressure was 1.0 ×
Torr. Peak fitting of the high-resolution data was carried
−
The authors declare no competing financial interest.
out using the CasaXPS software.
Photocatalytic Reactions. A 20 mL Pyrex glass tube was
used as the reaction container, and after the reactants and
catalyst had been added, the tube was sealed with a rubber
septum cap. The reaction mixture was stirred magnetically and
irradiated using a Nelson halogen lamp with a wavelength in
the range 400−800 nm (see Figure S8 in the SI) as the visible-
light source, and the light intensity was measured to be 0.5 W/
ACKNOWLEDGMENTS
■
The authors gratefully acknowledge financial support from the
Australian Research Council (ARC DP110104990 and
DP150102110).
REFERENCES
■
2
cm . The temperature of the reaction system was carefully
(1) Blaser, H. U.; Siegrist, U.; Steiner, H.; Studer, M. In Fine
Chemicals through Heterogeneous Catalysis; Sheldon, R. A., van Bekkum,
H., Eds.; Wiley-VCH: Weinheim, Germany, 2001; pp 389−406.
controlled with an air conditioner attached to the reaction
chamber. The control reaction system in the dark was
maintained at the same temperature to ensure that the
comparison was meaningful. All of the reactions in the dark
were conducted using a water bath placed above a magnetic
stirrer to control the reaction temperature; the reaction tube
was wrapped with aluminum foil to avoid exposure of the
reaction mixture to light. At given irradiation time intervals, 0.5
mL aliquots were collected and then filtered through a
Millipore filter (pore size 0.45 μm) to remove the catalyst
particulates. The liquid-phase products were analyzed by gas
chromatography (GC) using an Agilent 7820A gas chromato-
graph with an HP-5 column to measure the changes in the
concentrations of reactants and products. An Agilent HP5977A
mass spectrometer attached to an Agilent 7890B gas chromato-
graph with an HP-5MS column was used to identify the
products. The acetone concentration was tested using an
Agilent 6890 gas chromatograph with a DB-Wax column.
Typical reaction conditions were as follows: 2 mL of a 0.05
M solution of nitrobenzene in isopropyl alcohol (IPA) (0.1
mmol), 0.25 mL of a 0.1 M solution of KOH in IPA (0.025
mmol), and 50 mg of catalyst were added to the reaction tube,
and the reaction was run at a temperature of 40 °C under a 1
atm argon atmosphere for a reaction time of 6 h. The reactions
with other substrates were conducted under similar methods
with some conditions changed slightly as noted in Table 2.
Light-emitting diode (LED) lamps (Tongyifang, Shenzhen,
China) with wavelengths of 400 ± 5, 470 ± 5, 530 ± 5, 590 ±
(
2) Rosenblatt, D. H.; Burrows, E. P. The Chemistry of Amino Nitroso
and Nitro Compounds and their Derivatives; Wiley: Chichester, U.K.,
982; p 1085.
1
(
(
1
3) Corma, A.; Serna, P. Science 2006, 313, 332−334.
4) Grirrane, A.; Corma, A.; Garcia, H. Science 2008, 322, 1661−
664.
(
5) Jagadeesh, R. V.; Surkus, A. E.; Junge, H.; Pohl, M. M.; Radnik, J.;
Rabeah, J.; Huan, H. M.; Schunemann, V.; Bruckner, A.; Beller, M.
Science 2013, 342, 1073−1076.
̈
(6) Westerhaus, F. A.; Jagadeesh, R. V.; Wienhofer, G.; Pohl, M.-M.;
Radnik, J.; Surkus, A.-E.; Rabeah, J.; Junge, K.; Junge, H.; Nielsen, M.;
Bru
̈
ckner, A.; Beller, M. Nat. Chem. 2013, 5, 537−543.
7) Haber, F. Z. Elektrochem. Angew. Phys. Chem. 1898, 4, 506−514.
8) Makosch, M.; Sa, J.; Kartusch, C.; Richner, G.; van Bokhoven, J.
(
(
A.; Hungerbuhler, K. ChemCatChem 2012, 4, 59−63.
9) Liu, X.; Li, H. Q.; Ye, S.; Liu, Y. M.; He, H. Y.; Cao, Y. Angew.
Chem., Int. Ed. 2014, 53, 7624−7628.
10) Blaser, H. U. Science 2006, 313, 312−313.
(11) Corma, A.; Concepcion, P.; Serna, P. Angew. Chem., Int. Ed.
2007, 46, 7266−7269.
12) Zhu, H. Y.; Ke, X. B.; Yang, X. Z.; Sarina, S.; Liu, H. W. Angew.
Chem., Int. Ed. 2010, 49, 9657−9661.
13) Chen, X.; Zhu, H. Y.; Zhao, J. C.; Zheng, Z. F.; Gao, X. P.
Angew. Chem., Int. Ed. 2008, 47, 5353−5356.
14) Sarina, S.; Waclawik, E. R.; Zhu, H. Y. Green Chem. 2013, 15,
(
(
́
(
(
(
1
(
(
814−1833.
15) Wang, C.; Astruc, D. Chem. Soc. Rev. 2014, 43, 7188−7216.
16) Khanal, S.; Bhattarai, N.; McMaster, D.; Bahena, D.; Velazquez-
Salazar, J. J.; Jose-Yacaman, M. Phys. Chem. Chem. Phys. 2014, 16,
16278−16283.
5, and 620 ± 5 nm were used as the light sources to investigate
1
752
DOI: 10.1021/acscatal.5b02643
ACS Catal. 2016, 6, 1744−1753

ACS Catalysis
Research Article
(
17) Liu, X.; Wang, A.; Li, L.; Zhang, T.; Mou, C. Y.; Lee, J. F. J.
Catal. 2011, 278, 288−296.
18) Xu, Z.; Lai, E.; Shao-Horn, Y.; Hamad-Schifferli, K. Chem.
(
Commun. 2012, 48, 5626−5628.
(
(
19) Wang, L.; Yang, J. C. J. Mater. Res. 2005, 20, 1902−1909.
20) Guo, X. N.; Hao, C. H.; Jin, G. Q.; Zhu, H. Y.; Guo, X. Y.
Angew. Chem., Int. Ed. 2014, 53, 1973−1977.
(
21) Goncalves, R. V.; Wojcieszak, R.; Wender, H.; Dias, C. S. B.;
̧
Vono, L. L. R.; Eberhardt, D.; Teixeira, S. R.; Rossi, L. M. ACS Appl.
Mater. Interfaces 2015, 7, 7987−7994.
(
22) Schu
7, 7743−7750.
23) Sugano, Y.; Shiraishi, Y.; Tsukamoto, D.; Ichikawa, S.; Tanaka,
S.; Hirai, T. Angew. Chem., Int. Ed. 2013, 52, 5295−5299.
24) Marimuthu, A.; Zhang, J.; Linic, S. Science 2013, 339, 1590−
593.
25) Cano, I.; Huertos, M. A.; Chapman, A. M.; Buntkowsky, G.;
̈ ̈ ̈
nemann, S.; Dodekatos, G.; Tuysuz, H. Chem. Mater. 2015,
2
(
(
1
(
Gutmann, T.; Groszewicz, P. B.; van Leeuwen, P. W. N. M. J. Am.
Chem. Soc. 2015, 137, 7718−7727.
(
26) Pramanik, S.; Mishra, M. K.; De, G. CrystEngComm 2014, 16,
5
(
6−63.
27) Section 9, Molecular Structure and Spectroscopy; Electro-
negativity. In CRC Handbook of Chemistry and Physics, 84th ed.; Lide,
D. R., Ed.; CRC Press: Boca Raton, FL, 2003; pp 9−74.
(
28) Abad, A.; Concepcion
Int. Ed. 2005, 44, 4066−4069.
29) Ellis, F. Paracetamol: A Curriculum Resource; Royal Society of
Chemistry: Cambridge, U.K., 2002; pp 3−12.
30) Saha, A.; Adamcik, J.; Bolisetty, S.; Handschin, S.; Mezzenga, R.
Angew. Chem., Int. Ed. 2015, 54, 5408−5412.
31) Li, A.; Luo, Q.; Park, S.-J.; Cooks, R. G. Angew. Chem., Int. Ed.
014, 53, 3147−3150.
32) Wang, X.; Liu, D.; Song, S.; Zhang, H. J. Am. Chem. Soc. 2013,
35, 15864−15872.
33) Linic, S.; Christopher, P.; Ingram, D. B. Nat. Mater. 2011, 10,
11−921.
34) Christopher, P.; Xin, H. L.; Marimuthu, A.; Linic, S. Nat. Mater.
012, 11, 1044−1050.
35) Xiao, Q.; Jaatinen, E.; Zhu, H. Y. Chem. - Asian J. 2014, 9,
046−3064.
36) Xiao, Q.; Liu, Z.; Bo, A.; Zavahir, S.; Sarina, S.; Bottle, S.;
Riches, J. D.; Zhu, H. Y. J. Am. Chem. Soc. 2015, 137, 1956−1966.
37) Sarina, S.; Zhu, H. Y.; Xiao, Q.; Jaatinen, E.; Jia, J.; Huang, Y.;
Zheng, Z.; Wu, H. Angew. Chem., Int. Ed. 2014, 53, 2935−2940.
38) Sarina, S.; Zhu, H.; Jaatinen, E.; Xiao, Q.; Liu, H.; Jia, J.; Chen,
́
, P.; Corma, A.; García, H. Angew. Chem.,
(
(
(
2
(
1
(
9
(
2
(
3
(
(
(
C.; Zhao, J. J. Am. Chem. Soc. 2013, 135, 5793−5801.
(
(
(
39) Long, R.; Li, Y.; Song, L.; Xiong, Y. Small 2015, 11, 3873−3889.
40) Brus, L. Acc. Chem. Res. 2008, 41, 1742−1749.
41) Smith, J. G.; Faucheaux, J. A.; Jain, P. K. Nano Today 2015, 10,
6
7−80.
42) Kale, M. J.; Avanesian, T.; Christopher, P. ACS Catal. 2014, 4,
16−128.
43) Zhao, L. B.; Huang, Y. F.; Liu, X. M.; Anema, J. R.; Wu, D. Y.;
Ren, B.; Tian, Z. Q. Phys. Chem. Chem. Phys. 2012, 14, 12919−12929.
44) Zhao, L. B.; Zhang, M.; Huang, Y. F.; Williams, C. T.; Wu, D.
Y.; Ren, B.; Tian, Z. Q. J. Phys. Chem. Lett. 2014, 5, 1259−1266.
45) Linic, S.; Aslam, U.; Boerigter, C.; Morabito, M. Nat. Mater.
015, 14, 567−576.
46) Kale, M. J.; Avanesian, T.; Xin, H.; Yan, J.; Christopher, P. Nano
Lett. 2014, 14, 5405−5412.
(
1
(
(
(
2
(
1
753
DOI: 10.1021/acscatal.5b02643
ACS Catal. 2016, 6, 1744−1753