Synthesis of selected aromatic aldehydes under UV-LED irradiation over a hybrid photocatalyst of carbon nanofibers and zinc oxide

Article abstract of DOI:10.1016/j.cattod.2018.10.061

Zinc oxide (ZnO) prepared by chemical vapor deposition was combined with different amounts of carbon nanofibers (CNF) to obtain hybrid materials, which were thoroughly characterized using several techniques. The photocatalytic performance was evaluated towards the photocatalytic synthesis of vanillin (VAD) starting from vanillyl alcohol (VA). The incorporation of the carbon phase in ZnO (from 5% to 20% wt.) was found to increase the surface area and the photocatalytic performance of the materials. The latter was attributed to the efficient separation of charge carriers generated on the optical semiconductor. With the best performing material, the one containing 10% of CNF, the selectivity of the synthesis towards vanillin generation was increased by a factor of 2.5 compared to previous studies, with the additional advantage of carrying the reaction in aqueous medium. The same photocatalyst was successfully applied to the selective synthesis of other aromatic aldehydes, namely anisaldehyde, piperonal, and benzaldehyde. A relationship between the efficiency of the photocatalytic oxidation of the alcohols and the activating nature of their aromatic ring substituents was proposed.

Full text of DOI:10.1016/j.cattod.2018.10.061

Accepted Manuscript
Title: Synthesis of selected aromatic aldehydes under
UV-LED irradiation over a hybrid photocatalyst of carbon
nanofibers and zinc oxide
Authors: Raquel A. Fernandes, Maria J. Sampaio, Eliana S. da
´
Silva, Philippe Serp, Joaquim L. Faria, Claudia G. Silva
PII:
S0920-5861(18)31117-9
DOI:
Reference:
https://doi.org/10.1016/j.cattod.2018.10.061
CATTOD 11724
To appear in:
Catalysis Today
Received date:
Revised date:
Accepted date:
15-7-2018
20-10-2018
24-10-2018
Please cite this article as: Fernandes RA, Sampaio MJ, da Silva ES, Serp P, Faria JL,
Silva CG, Synthesis of selected aromatic aldehydes under UV-LED irradiation over
a hybrid photocatalyst of carbon nanofibers and zinc oxide, Catalysis Today (2018),
https://doi.org/10.1016/j.cattod.2018.10.061
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.

Synthesis of selected aromatic aldehydes under UV-LED irradiation
over a hybrid photocatalyst of carbon nanofibers and zinc oxide
Raquel A. Fernandesa, Maria J. Sampaioa, Eliana S. da Silvaa, Philippe Serpb, Joaquim
L. Fariaa, Cláudia G. Silvaa,*
aLaboratory of Separation and Reaction Engineering - Laboratory of Catalysis and
Materials (LSRE-LCM), Faculdade de Engenharia, Universidade do Porto, Rua Dr.
Roberto Frias s/n, 4200-465, Porto, Portugal
bLaboratoire de Chimie de Coordination, UPR CNRS 8241, Composante ENSIACET,
Université de Toulouse, UPS-INP LCC, 4 Allée Emile Monso, BP 44362, 31030,
Toulouse Cedex 4, France
*Corresponding author e-mail address: cgsilva@fe.up.pt (C.G. Silva)

Graphical abstract
Highlights
 CNF/ZnO materials are efficient photocatalysts for selective alcohol oxidation.
 Composite materials showed higher photocatalytic performance than neat ZnO.
 The incorporation of CNF promotes charge separation and transportation.
 Inserting 10% weight of CNF into the ZnO matrix, leads to the best activity and
selectivity.
 Organic substrate electronic nature conditions the photocatalytic performance.
Abstract
Zinc oxide (ZnO) prepared by chemical vapor deposition was combined with different
amounts of carbon nanofibers (CNF) to obtain hybrid materials, which were thoroughly
characterized using several techniques. The photocatalytic performance was evaluated

towards the photocatalytic synthesis of vanillin (VAD) starting from vanillyl alcohol
(VA). The incorporation of the carbon phase in ZnO (from 5.0% to 20% wt.) was found
to increase the surface area and the photocatalytic performance of the materials. The latter
was attributed to the efficient separation of charge carriers generated on the optical
semiconductor. With the best performing material, the one containing 10% of CNF, the
selectivity of the synthesis towards vanillin generation was increased by a factor of 2.5
compared to previous studies, with the additional advantage of carrying the reaction in
aqueous medium. The same photocatalyst was successfully applied to the selective
synthesis of other aromatic aldehydes, namely anisaldehyde, piperonal, and
benzaldehyde. A relationship between the efficiency of the photocatalytic oxidation of
the alcohols and the activating nature of their aromatic ring substituents was proposed.
Keywords: Photocatalysis; Chemical synthesis; Aromatic Alcohols; Zinc oxide; Carbon
nanofibers
1
Introduction
Since the energetic and environmental issues have become a serious concern to the human
society, both governments and scientists were committed to find technologies more
environmentally-friendly to reduce the ecological impact of the conventional industrial
processes. The chemical industry currently represents one of the largest industrial sectors
in the world, with the synthesis of fine chemicals playing a major role in it. Harsh
conditions of temperature and pressure, use of toxic organic solvents and high energetic

demand required by the organic synthesis processes are factors that contribute to a highly
negative ecological impact.
The application of photocatalysis in the synthesis of organic compounds is not a recent
topic and has been the subject of several reviews [1–6]. Oxidation and reduction,
isomerization, C–H bond activation, C–C and C–heteroatom bond formation and
cycloaddition are examples of chemical transformations that have been successfully
carried out by photo-assisted processes[1]. Over the past decade, a boost in the interest in
photo-assisted synthetic organic chemistry, mostly concerning with semiconductor-based
photocatalytic processes, has been witnessed because it is considered a renewable, clean,
safe and economic alternative to conventional themal catalysis [7]. Semiconductor
materials may drive the oxidation/reduction of several species, eliminating the need for
hazardous oxidant and reductant substances. This type of heterogeneous photocatalytic
processes can be carried out in water solutions instead of using organic solvents reducing
the amount of generated toxic residues [8]. Moreover, the energetic costs of the processes
can be reduced by the use of either, natural energy sources such as sunlight, or highly-
efficient Light Emitting Diodes (LEDs) to activate the photocatalyst [9]. Heterogeneous
photocatalysis is currently applied in distinct fields such as wastewater treatment [10–
12], and the production of hydrogen [13,14], fuels and chemicals [15,16].
Vanillin is the major component of natural vanilla, which is one of the most important
flavoring agent used in food, cosmetic and pharmaceutical industries [17]. Vanillin
production is mainly carried out through a synthetical route, once the natural sources
(bean and pod of Vanilla orchid) are scarce and the extraction from these sources is not
economically profitable [18]. There are a few reports about the synthesis of vanillin by
heterogeneous photocatalysis. Augugliaro et al. [18] studied the synthesis of vanillin from

the oxidation of several precursors using titanium dioxide (TiO2) as catalyst, reaching 5
to 21% of selectivity and 1 to 6% of yield. In another report, Di Paola et al. [19] performed
the photocatalytic synthesis of vanillin from trans-ferulic acid by WO3-loaded TiO2,
obtaining 24.7% of selectivity and 1.1% of yield. Recently, some of us [20] reported the
photocatalytic synthesis of vanillin using a composite catalyst constituted of zinc oxide
(ZnO) and nitrogen-doped carbon nanotubes, reaching 15% of vanillyl alcohol
conversion and 27% of selectivity for vanillin formation. Yet, all these reports showed
that selectivity of the referred processes is still limited in terms of industrial application.
In the field of neat optical semiconductors, ZnO present electronical properties similar to
the benchmark TiO2. Moreover, its lower cost and morphological versatility may be
envisionaged as promising aspects to promote its use as an alternative to TiO2 [12,21].
Several modifications have been performed in neat ZnO towards increasing its
photoefficiency, including doping with metals (Ni, Cu, Mg, Ag) and non-metals (N, S),
formation of heterojunctions by combination with other semiconductors and coupling
with carbon materials [22]. Composites of ZnO and carbon materials proved to be
promising catalysts in the decomposition of several organic pollutants [11,23,24] and
already in organic synthesis [20]. In these examples, carbon nanofibers (CNF), can act as
photoexcited electron dam, storing the excited electrons generated during the irradiation
of ZnO, or alternatively as photosensitizers, improving light absorption [22].
Based in the promising results obtained using carbon/ZnO photocatalysts for the synthesis
of vanillin [20], and seeking for enhanced selectivity of the process, in this work we report
the development of CNF/ZnO hybrid materials and its application in the production of
aromatic aldehydes from the respective alcohols, under mild conditions of temperature
and pressure, and using energy-efficient light sources (UV-LEDs). Composite materials
of ZnO and different amounts of CNF (5 to 20% wt.) were prepared and characterized by

several techniques in order to assess their morphological, electronic, and optical
properties. The structure-activity effect of coupling CNF with ZnO was investigated in
the photocatalytic synthesis of vanillin. Furthermore, the effect of oxygen in the
photocatalytic pathway was also explored. The study was extended to the photocatalytic
synthesis of other aromatic aldehydes, using the best reaction conditions found in the
photocatalytic synthesis of vanillin.
2
Experimental
2.1 Materials
Zinc flake (325 mesh size, >99%, Alfa Aesar), ethylene and ternary oxide catalyst were
used for the synthesis of zinc oxide (ZnO) and carbon nanofibers (CNF). Tetrahydrofuran
(C4H8O, >99%, Sigma-Aldrich) was used in the synthesis of the composite catalysts.
Ultra-pure water was obtained from a Direct-Q Millipore system and it was used as
solvent in the photocatalytic reactions. Vanillyl alcohol (VA, 4-hidroxy-3-
methoxybenzyl alcohol, C8H10O3, >98%, Alfa Aesar), anisyl alcohol (AA, 4-
methoxybenzyl alcohol, C8H10O2, 98%, Alfa Aesar), benzyl alcohol (BA, C7H7OH,
>99%, Sigma-Aldrich) and piperonyl alcohol (PA, C8H8O3, 98%, Sigma-Aldrich),
vanillin (VAD, C8H8O3, 99%, Alfa Aesar), p-anisaldehyde (AAD, C8H8O2, 98%, Sigma-
Aldrich), benzaldehyde (BAD, C6H5CHO, >99%, Sigma-Aldrich), piperonal (PAD,
C8H6O3, 99%, Sigma-Aldrich) and ethylenediaminetetraacetic acid (EDTA,  98%,
Sigma-Aldrich) were used as received.
Formic acid (HCO2H, 98%, Fluka) and methanol (CH3OH, 99.6%, VWR) were used for
high performance liquid chromatography analysis.

2.2 Catalysts synthesis
CNF and ZnO tetrapods were prepared by catalytic chemical vapor deposition (CVD) as
previously described [25,26].
The CNF/ZnO composites were prepared by adding 0.200 g of ZnO to a suspension of a
given amount of CNF in tetrahydrofuran under vigorous stirring, followed by sonication
for 30 min. The obtained solution was stirred for 1 h and the so obtained solid was then
filtered, washed and dried at 60°C for 12 h.
The catalysts were denoted as XCNF/ZnO, where X is the weight percentage of CNF in
relation to the mass of ZnO the composite. For example, 10%CNF/ZnO means that the
composite catalyst was prepared using a suspension with 10 mg of CNF and 100 mg of
ZnO.
2.3 Photocatalyst characterization
The morphology of ZnO and composite catalysts was analyzed by scanning electron
microscopy (SEM), using a FEI Quanta 400 FEG ESEM/EDAX Genesis X4M (15 keV)
instrument. CNF was characterized by transmission electronic microscopy (TEM), using
a TEM-FEI Tecnai-G2-20-FEI 2006 microscope at 200 kV. The effective content of
carbon material in the catalysts was determined by the thermogravimetric (TG) analysis,
by heating samples from 50 °C to 900 °C at 10 °C min-1 rate under air flow using a STA
490 PC/4/H Luxx Netzsch thermal instrument. The diffuse-reflectance UV–Vis spectra
(DR UV-Vis) of the materials were obtained on a JASCO V-560 UV-Vis
spectrophotometer equipped with an integrating sphere attachment (JASCO ISV-469).
The bandgap of the catalysts was determined from the respective Tauc plot. The specific
surface area (SBET) was determined through the N2 adsorption isotherm at -196°C in the
relative pressure range 0.05–0.20, using a Quantachrome NOVA 4200e apparatus. Solid

state photoluminescence (PL) spectra were obtained at room temperature on a JASCO
(FP 82000) fluorescence spectrometer with a 150W Xenon lamp as light source, under
340 nm excitation, using bandwidths of 2.5 nm for emission and excitation.
2.4 Photocatalytic experiments
The photocatalytic synthesis of vanillin (VAD) by ZnO was carried out through the
oxidation of vanillyl alcohol (VA) in a cylindrical glass reactor filled with 50 mL of 0.50
mM alcohol aqueous solution. The experiments were carried out at room temperature
(20°C), without pH adjustment (c.a. 5.7). The effect of catalyst load in the catalyst
performance was studied by varying the ZnO load between 0.10 and 1.2 g L-1. The effect
of CNF content (from 0 to 20 wt%) and the presence of oxygen in the reaction medium
were also evaluated. A study without the presence of any catalyst was also performed to
evaluate the possibility of photolysis. The application of the photocatalytic system to the
synthesis of other aromatic aldehydes, namely anisaldehyde (AAD), benzaldehyde
(BAD) and piperonal (PAD) was also studied starting from the respective alcohols (anisyl
(AA), benzyl (BA) and piperonyl (PA) alcohols).
The reaction medium was continuously saturated with argon (20 mL min-1) and kept in
dark conditions for 15 min. The irradiation was performed under four UV-LED system
emitting at 370 nm located symmetrically from the outside at 4 cm from the reactor wall.
The average nominal irradiance of each UV-LED was 450 W m-2, determined using a
UV-vis spectroradiometer (OceanOptics USB2000 +). The suspension was irradiated for
8 hours and samples were periodically withdrawn to monitor alcohol and aldehyde
concentration by High Performance Liquid Chromatography (HPLC), using a Shimadzu
Corporation apparatus equipped with a Diode Array Detector (SPD M20A). A
KinetexTM F5 1.7 µm 100 Å column (100 mm x 2.1 mm) was used with a solvent

delivery pump (LC-30AD) using a flow rate fixed at 0.15 mL min-1. The temperature of
the column oven and autosampler were set at 35 °C and 4.0 °C, respectively. An
equilibrated mixture of 0.1% v/v of formic acid (A) and methanol (B) with volume ratio
of 30:70 (A:B) was isocratically eluted for 15 min. The concentration of VA, VAD, PA
and PAD were determined at 230 nm. The BA concentration was determined at 207 nm,
while BAD concentration was determined at 250 nm. The AA and AAD concentration
were determined at 226 nm and 282 nm, respectively.
The conversion (C) of alcohol, selectivity (S) and yield (Y) towards the aldehyde
production were determined as described elsewhere [20].
3
Results and discussion
3.1 Catalyst characterization
The morphology of the catalysts was observed by scanning and transmission electron
microscopy (SEM and TEM, respectively, Figure 1). The SEM micrographs of ZnO
(Figure 1a) reveal the presence of tetrapod-like structures (I) with needles growing from
a central particle, as previously reported [11,26]. The incorporation of CNF in ZnO phase
does not induce any morphological modification on ZnO tetrapods, but induced their
aggregation on the side-walls of CNF and the generation of interstices in the ZnO phase
(Figure 1b-e), as reported in literature for studies using composite catalysts constituted
from metal oxides and carbon materials [27,28]. The presence of carbon phase in
5%CNF/ZnO (Figure 1b) was difficult to detect due to the low content of CNF in
composite material, with few small fibers being observed. By increasing the CNF content
to 10%, a high amount of fibers can be observed (Fig. 1c). The SEM image of
10%CNF/ZnO (Figure 1c) shows a good distribution of CNF on the ZnO phase, providing

a better contact between the two phases, which is known to favor the photocatalytic
activity. Additionally, it is evident that the increase of the CNF content to 15% and 20%
(Figure 1d and 1e, respectively) leads to the formation of ZnO agglomerates (Fig. 1, zone
III) surrounding CNF. These features may indicate the exixtence of physical segregation
between both phases, reducing their interactions. This behavior has already been reported
in previous studies using carbon/metal oxide photocatalysts [20,29]. In the present work,
this phase segregation may be attributed to the use of pristine CNF to prepare the hybrid
materia (without any surface treatment for increasing their dispersibility in the solvent),
and their tendency to aggregate during the synthesis procedure, thus impeding the
dispersion of ZnO (see Supplementary Information, Fig. SI1). TEM analysis of CNF
(Figure 1f) shows a long tubular structure (II), with a diameter of c.a. 30 nm.

Figure 1. SEM micrographs of a) ZnO, b) 5%CNF/ZnO, c) 10%CNF/ZnO, d)
15%CNF/ZnO and e) 20%CNF/ZnO. TEM image of f) CNF. Highlighted zones: I) ZnO
tetrapods, II) long tubular CNF and III) ZnO agglomerates.

The composition and surface thermodynamics of the composite materials was assessed
by TG analysis (Table 1 and SI). Bare ZnO, CNF and composites were submitted to a
thermal process until 900°C under air flow and the content of CNF was determined (Fig.
SI2). No significant mass loss was verified in ZnO, as expected, since the material is
stable in this range of temperature [20]. The TG profile of CNF shows about 5% of
residue, which may be indicative of the catalyst metal used to grown the CNF [25]. The
composite materials show a loss of weight around 550°C, which also occurs in the case
of CNF, due to the carbon gasification [30]. The theoretical carbon content (%) in
composite materials was determined (Table 1) considering the weight of CNF used in the
preparation of the composite material in relation to its total weight. The values indicate a
weight loss between 4-15% for the different composite materials, which is in agreement
with their nominal carbon content.
The surface area of bare ZnO, CNF and composite materials (SBET) are listed in Table 1.
The experimental SBET of bare ZnO was found to be very low (11 m2 g-1) comparing with
that of CNF (81 m2 g-1). As expected, an enhance in SBET of the composites was verified
with the incorporation of CNF in the metal oxide phase, which is related with the increase
of CNF content from 5% to 20%. Additionally, the experimental SBET of composite
materials is slightly lower than the theoretical one, obtained from the relative weight
fraction of the CNF and ZnO phases in each composite (Table 1). This deviation may be
attributed to the surface contact between CNF and ZnO tetrapods and the agglomeration
of ZnO tetrapods over CNF, as observed in SEM micrographs [11]. Bare ZnO, CNF and
the composite materials produce isotherms type II, according to the IUPAC classification
(Fig. SI3), which are characteristics of a non-porous material. However, it is visible a type
H3 hysteresis loop, which could be explained in terms of the interstices formed between
the aggregates of ZnO and CNF [31,32].

Table 1. Carbon content (%) and specific surface area (SBET) of ZnO and of
the composite catalysts.
Carbon content (%)
SBET (m2 g-1)
Material
Experimental
Theoretical
-
Experimental Theoretical a
ZnO
-
11
12
15
15
18
81
-
5%CNF/ZnO
10%CNF/ZnO
15%CNF/ZnO
20%CNF/ZnO
CNF
4.04
7.75
11.3
14.5
94.7
4.76
9.09
13.0
16.7
100
14
18
21
25
-
a The theoretical surface area was determined considering the experimental weight
content of ZnO and CNF in the catalyst
The optical properties of the photocatalysts were studied by diffuse reflectance (DR) UV-
Vis and photoluminescence (PL) spectroscopies. The DR UV-Vis spectrum of bare ZnO
expressed in terms of Kubelka-Munk equivalent absorption units (Fig. SI4) shows a
characteristic intense absorption band in the UV spectral region (λ<400 nm) due to its
bandgap transition [20]. The incorporation of CNF led to higher absorption in the visible
region (λ>400 nm). This feature might be related to the inherent ability of CNF to absorb
light (Fig. SI4), as well to the creation of an electronic interphase interaction between the
CNF and ZnO phases [13,20]. The bandgap energy (Eg) from neat ZnO, taken from the
Tauc plot, accounts for a value of 3.28 eV (Fig. SI4). In a qualitative way the Tauc
analysis of the composite suggests that, in small amounts (i.e. for 5%CNF/ZnO ), the
carbon phase does not affect the Eg of ZnO. Increasing the amount of CNF to 15 and 20%
results in a small decrease in the Eg of the ZnO. This is possibly due to the presence of
ZnO aggregates (SEM analysis), as previously reported [33]. The determination of Eg of
CNF through a similar Tauc plot, was not possible due to their high absorption in whole
UV-Vis range [34].

To evaluate the degree of charge separation and transportation in the prepared materials,
photoluminescence (PL) measurements were performed (Fig. 2). The PL emission spectra
of ZnO reveals an intense UV peak (λmax = 379.6 nm, Fig. 2a) due to the exciton
recombination on the surface of ZnO tetrapods [11,35]. Additionally, several peaks are
observed in the visible range, including a broad green peak at 540 nm (inset Fig. 2a),
usually associated with the presence of defects on the catalyst surface [11,35]. With the
incorporation of CNF on the metal oxide phase, the PL intensity of ZnO is significantly
quenched in both emission regions (Fig. 2a), indicating that the recombination of the
photogenerated charge carriers was greatly inhibited in the CNF/ZnO nanostructures. It
follows, that the CNF serves as an efficient electron sink, and scavenges away the
electrons, hindering recombination [11,20], therefore improving the separation of charge
carriers, resulting in higher photocatalytic activity [36]. As clearly observed in Fig. 2b,
the composite containing 10% of CNF leads to the highest quenching of the fluorescence
emission. Among all the composites, this indicates that the catalyst 10%CNF/ZnO
presents the most efficient separation of charge carriers, which is beneficial to enhance
the interfacial charge transfer to adsorbed substrates, thereby improving the
photocatalytic performance [36]. Furthermore, the inhibition of fluorescence is
accompanied by a small blue shift of the maximum emission of ZnO, resultant of the
interaction between the metal oxide and the carbon phase (Fig. 2I). On the other hand, a
further increase in the content of CNF to 15 and 20% causes a small increase in the
fluorescence emission intensity in both emission regions (Fig. 2b), in relation to the
composite with 10% of CNF. This increase might be related with ZnO agglomeration
evident in SEM micrographs (Figs. 1d and 1e), which hinders the efficient interfacial
charge transfer between ZnO and CNF, enhancing the recombination of electron-hole.

Figure 2. a) Solid state PL spectra of ZnO and of the composites obtained upon excitation
at 340 nm. Inset shows a zoom of the emission bands in the visible region. b) ZnO
fluorescence emission intensity (measured at maximum of emission) in function of CNF
percentage.
3.2 Photocatalytic selective synthesis of vanillin
The effect of CNF incorporation in the photocatalytic performance of ZnO was evaluated
in the synthesis of vanillin (VAD) through oxidation of vanillyl alcohol (VA) under UV-
LED irradiation (370 nm) and argon atmosphere, using water as solvent. In absence of
catalyst (Fig. SI5), one but negligible decrease in VA concentration was observed at the
end of 4 h reaction, proving the photo-inertness of the parent alcohol molecule under the
present irradiation conditions.
The amount of catalyst used is an important parameter that needs to be optimized for
avoiding the useless excess of it, as well as to ensure an efficient absorption of the
irradiating photons. An excess of catalyst concentration may produce screening effects,
which can lead to a decrease of the photocatalytic efficiency of the process. Therefore,
the optimization of the ZnO load was performed by varying the amount suspended
between 0.10 and 1.2 g L-1. The activity and selectivity of the catalysts was evaluated in
terms of C, S and Y, determined after 1 h of reaction (Figure 3a).

Figure 3. a) Effect of ZnO load on C, S and Y at 1 h of reaction. b) Effect of CNF
content on C, S and Y in long-term (8 h) reactions (0% CNF content corresponds to
bare ZnO).
From the conversion values there is a linear dependency with the amount of catalyst up
to a certain amount, confirming the heterogeneous nature of the catalytic regime.
However, above 0.75 g L-1 the reaction becomes independent of the catalyst load. This
behavior corresponds to the maximum amount of catalyst at which all the particles are
totally irradiated, as well to the reaction conditions such as reactor geometry and radiation
configuration [37]. For ZnO loads between 0.75 g L-1 and 1.0 g L-1, a maximum of 14%
for VA conversion, 25% of selectivity and 3% of yield towards VAD production was
found. The efficiency decreased for higher ZnO load (1.2 g L-1), with a decrease in C, S
and Y (Figure 3a). As already mentioned above, these results may be explained by the
increase in opacity of the reaction medium, triggering a significant amount of light
scattering [37]. Additionally, agglomeration (particle-particle interaction) is favored for
higher catalyst load, which may lead to the reduction of active sites in the surface of the
catalyst for light irradiation and substrate adsorption, decreasing the efficiency of the
photocatalytic system [38]. Considering the obtained results, 0.75 g L-1 of catalyst was

chosen as the optimal catalyst load for the subsequent studies using both ZnO and
CNF/ZnO hybrids.
The effect of carbon phase on ZnO performance was therefore studied in the
photocatalytic synthesis of VAD over 8 h reaction, using different contents of CNF (5,
10, 15 and 20 wt.%). The photocatalytic results presented in Figure 3b reveal that the
incorporation of CNF promotes an increase of the efficiency in terms of yield towards
VAD production when compared with bare ZnO, after 8 h of reaction. The higher
photocatalytic performance of CNF/ZnO nanostructures was already reported in literature
[20,22,36]. With the introduction of CNF into the matrix, excited electrons in the
conduction band of ZnO are prone to be transferred to the surface of CNF, promoting an
efficient separation of electron-hole compared to bare ZnO. The holes left the valence
band of ZnO may act as active sites for oxidation of the adsorbed molecules, as in the
case of VA, adding to the catalytic activity [36,39].
An experiment using bare CNF was performed (Fig. SI6), and the results showed that it
lacks photoactivity for VAD production. Thus, the photocefficiency of the hybrid
CNF/ZnO materials cannot be rationalized in terms of an additive catalytic action of CNF
and ZnO, but attributed to an electronic synergetic effect between the carbon and metal
oxide phases, as revealed by PL spectroscopy (Fig. 2).
When the CNF content was raised up to 10 wt.%, a continuous increase in selectivity and
yield is observed (Figure 3II). However, for higher CNF content (15 and 20 wt.%), the
suspensions are darkened, which might promote light absorption and scattering, thus the
decrease in catalytic performance, leading to lower values of S and Y (Figure 3b).
Although these two materials present the highest surface areas, the formation of
aggregates (Fig. 1) and the higher recombination of charge carriers demonstrated by PL
studies might be accountable for the lower activity of the composites.

Among all the composites tested, the 10%CNF/ZnO showed the best photocatalytic
performance with c.a. C = 50%, S= 40% and Y = 20% at the end of 8 h of reaction (Figure
3II). The higher photocatalytic performance of this material might be ascribed to a more
efficient separation of the photogenerated electron-hole and increased surface area. SEM
image reveals also an efficient interface contact between ZnO and CNF phases, which
contributes to the efficient charge carrier separation, as observed by PL results.
Furthermore, the increase surface area contributes to create more active sites for
adsorption of VA, which are beneficial to promote the photocatalytic activity.
Generally, heterogeneous photocatalytic oxidation processes are carried out under an
aerobic atmosphere. Therefore, in order to assess the role of oxygen in the photocatalytic
synthesis of VAD, reactions were performed using ZnO and the best performing
composite (10%CNF/ZnO) in the presence of oxygen (by saturating the suspensions with
air, Fig. SI7). The saturation of the reaction medium with argon promotes a continuous
synthesis of VAD using both materials over 2 h of reaction. On the other hand, in the
presence of oxygen, it is possible to observe a maximum in the VAD production at 5 min
of reaction and its total disappearance after 30 min using both ZnO and 10%CNF/ZnO.
Although VA oxidation was enhanced in presence of air, S was negatively affected,
reaching merely 7% contrasting with 40% for the reactions using argon. The presence of
oxygen acts as electron acceptor, generating highly oxidant reactive oxygen species
(ROS) [38,40–42], contributing to the increase in VA conversion, but negatively affecting
selectivity for VAD production, due to the highly unselective nature of these chemical
species. Under these conditions, the overoxidation of the aldehydes is expected with the
its conversion to short chain carboxylic acids, CO2 and H2O [18]. When air is replaced
by an inert gas such as argon, the formation of ROS is inhibited, resulting in an increase
in the selectivity of the reaction towards the formation of the aromatic aldehyde. In this

case, the main reactive species present in the photocatalytic reactions are expected to be
mostly photogenerated holes (h+) and electrons (e-) [43].
A summary of the results reported in the literature for the photocatalytic synthesis of
vanillin from vanillyl alcohol both in the presence and absence of oxygen in depicted as
Supplementary Information (Table SI1). Augugliaro et al. [18] achieved 21% of S in the
VAD synthesis after 90 min of reaction, using a home prepared TiO2 (HPC3), under air.
In a recent work by our group [20] we were able to achieve selectivity up to 27% in only
20 min of irradiation using a ZnO/N-doped carbon nanotube composite catalyst. In the
present work, although increasing the reaction time from 20 to 60 min, we were able to
improve the selectivity for VAD by 2.5 times, reaching 67% of S using the best
performing composite catalyst, 10%CNF/ZnO (Table SI1).
To investigate the reaction pathway of photocatalytic synthesis of VAD, aqueous
solutions in the presence of EDTA (hole scavenger) and the catalysts ZnO and
10%CNF/ZnO were irradiated under UV-LED light (data not shown). Under these
conditions, EDTA inhibited the photocatalytic synthesis of VAD, implying that the
photocatalytic production of VAD occurs mainly through oxidation of VA by
photogenerated holes by H+ abstraction [20]. On the other hand, electrons at the
conduction band of the photocatalyst drive the reduction of H+ to H2, considered a
valuable by product of the process.
3.3 Photocatalytic synthesis of selected aromatic aldehydes
Taking into account the photocatalytic performance of ZnO and 10%CNF/ZnO in the
synthesis of VAD, the synthesis of other aromatic aldehydes was evaluated under similar
reaction conditions. Anisyl alcohol (AA), piperonyl alcohol (PA) and benzyl alcohol
(BA) were used as substrate molecules for the production of their respective aldehydes.

The reactions were performed using the best operational conditions found for VAD
production, i.e., a catalyst load of 0.75 g L-1 and under anaerobic conditions.
As observed in Fig. 4, the photoproduction of vanillin (VAD), anisaldehyde (AAD) and
piperonal (PAD) continuously increase over the time using both catalysts. On the other
hand, the profile of benzaldehyde (BAD) production does not follow the same tendency,
reaching a maximum of concentration at 90 min, after which the concentration starts to
decrease, due to its overoxidation to benzoic acid, as confirmed by HPLC analysis (see
Supplementary Information, Fig. SI8). This behavior has already been reported in
previous studies being attributed to the attack of hydroxyl radicals, formed from the
reaction of HO– with positive holes, to BAD molecules [44,45].

Figure 4. Concentration profile of anisaldehyde (AAD), benzaldehyde (BAD), piperonal
(PAD) and vanillin (VAD) over 480 min of reaction, using (a) ZnO and (b) 10%CNF/ZnO
as catalysts
The conversion and selectivity obtained for the different aldehydes after 8 h of reaction
is summarized in Table SI3. In terms of selectivity for aldehyde production using both
ZnO and 10%CNF/ZnO materials, a higher efficiency for the photocatalytic synthesis of
VAD through the oxidation of VA was found, reaching 47% and 39% of S, respectively.
However, in relation to the alcohol conversion, the results revealed different behavior,
being AA and PA the substrates that present higher C for both catalysts. According to
Yurdakal et al. [46,47], electronic properties of substituent groups affect the rate of
oxidation of aromatic alcohols and this phenomena can be interpreted through the
Hammett relationship.
The substituent groups can act as electron-donor or electron-withdrawing groups,
depending on their Hammett constant in para- and meta- positions (σp and σm,
respectively, Table SI2) [46,47]. VA, AA and PA are benzyl alcohol derivatives, with
different substituent groups in para- and meta- positions (Table SI2). The alcohols
studied in this work present mostly negative Hammett constants (σ < 0), i.e. electron-
donor groups, resulting in an activating tendency of these groups for alcohol oxidation.
At variance, for positive Hammett constants (σ>0), i.e. electron-withdrawing groups,
substituent groups have a deactivating effect over the reactivity of aromatic ring [48].
Generally, heterogeneous photocatalytic reactions are known as electrophilic processes,
thus the presence of electron-donating groups in aromatic rings accelerate the chemical
reactions, while the presence of electron-withdrawing substituents retards the reaction
[48]. AA and PA molecules exhibits a methoxy (─OCH3) and methylenedioxy (─CH2O2)
groups, respectively, which are electron-donors, thus the highest conversions observed

for these alcohols may be attributed to the activating nature of their substituent groups. In
the case of BA, the presence of para-hydrogen, which may act also as electron-
withdrawing (low activating) showed negligible efficiency for BAD production when
compared with attached substituents groups of the previous alcohols.
In the case of VA, the aromatic ring has attached two substituent groups (─OH and
─OCH3 on para- and meta-position, respectively). Considering the Hammett constants
for both substituents and their respective positions (Table SI2), one can infer that the
combination of a very strong (─OH) and a less strong (─OCH3) activating group may
decrease the positive impact of ─OH (in para-position over the reactivity of aromatic
ring), leading to a decrease of the expected conversion of VA when compared with other
studied aromatic alcohols.
Overall, the results reported here reported indicate that the photocatalytic oxidation of
aromatic alcohols under anaerobic conditions follows a characteristic and unique pathway
depending on the nature and position of the substituents groups on the aromatic ring. On
the other hand, the properties of the photocatalysts used on this type of reaction assume
an important role for achieving highly selectivity for aldehydes production.
4
Conclusion
Neat ZnO and composites made of ZnO and CNF prepared with different amounts of
carbon phase (5 to 20 wt%) possess a significant photocatalytic activity and selectivity
towards the synthesis of vanillin under 370 nm UV-LED irradiation.
The better performing catalyst was the 10%CNF/ZnO, in which the presence of this
amount of carbon phase was reflected in an effective increase of the surface area and
efficient charge separation and transportation. The optimal load for this particular catalyst

was found to be 0.75 g L-1 reflecting the optimal balance between photon absorption,
suspension light scattering and substrate adsorption.
Removing the presence of oxygen, as done by replacing the air saturation condition by
argon increases the selectivity towards vanillin, either for ZnO or 10%CNF/ZnO.
The role of CNF is to act as electron acceptor, decreasing the recombination rate of the
photogenerated e-/h+ pairs. Mechanistically, photogenerated holes at the surface of ZnO
are responsible for the oxidation of vanillyl alcohol, while electrons at the CNF phase are
available for reducing H+ (from water and/or from alcohol dehydrogenation) to H2.
The photocatalytic system used proved to be about 2.5 times more selective than those
reported on literature for the photocatalytic synthesis of vanillin.
Moreover, the operation with mild conditions of pressure and temperature, using water
as solvent and an energy-efficient source of radiation, reduces the environmental footprint
of the conventional processes used for vanillin production, turning it environmentally
sustainable.
The modification of the substrate nature by including different substituent groups in the
aromatic ring of the substrate allowed to test the performance of the developed catalyst
in similar organic synthesis. The photocatalysts ZnO and 10%CNF/ZnO were also
equally successfully in the synthesis of other selected aromatic aldehydes, paving the way
for the developing the field of photoredox heterogeneous catalysis in synthetic organic
chemistry.
Acknowledgements
This work was financially supported by Project POCI-01-0145-FEDER-006984 –
Associate Laboratory LSRE-LCM funded by FEDER through COMPETE2020 -
Programa Operacional Competitividade e Internacionalização (POCI) – and by national
funds through FCT - Fundação para a Ciência e a Tecnologia and by the project

“AIProcMat@N2020 − Advanced Industrial Processes and Materials for a Sustainable
Northern Region of Portugal 2020”, with the reference NORTE-01-0145-FEDER-
000006, supported by Norte Portugal Regional Operational Programme (NORTE 2020),
under the Portugal 2020 Partnership Agreement, through the European Regional
Development Fund (ERDF ). R.A.F. acknowledges the PhD fellowship funded by Project
NORTE-08-5369-FSE-000028, supported by Norte Portugal Regional Operational
Programme (NORTE 2020), under the PORTUGAL 2020 Partnership Agreement,
through the European Social Fund (ESF). C.G.S. acknowledge the FCT Investigator
Programme (IF/00514/2014) with financing from the European Social Fund and the
Human Potential Operational Programme.

References
[1] B. König, European J. Org. Chem. 2017 (2017) 1979–1981.
[2] C. Michelin, N. Hoffmann, Curr. Opin. Green Sustain. Chem. 10 (2018) 40–45.
[3] A. Maldotti, A. Molinari, R. Amadelli, Chem. Rev. 102 (2002) 3811–3836.
[4] M. Fagnoni, D. Dondi, D. Ravelli, A. Albini, Chem. Rev. 107 (2007) 2725–2756.
[5] Y. Shiraishi, T. Hirai, J. Photochem. Photobiol. C Photochem. Rev. 9 (2008) 157–
170.
[6] S. Protti, M. Fagnoni, Photochem. Photobiol. Sci. 8 (2009) 1499–1516.
[7] D. Friedmann, A. Hakki, H. Kim, W. Choi, D. Bahnemann, Green Chem. 18
(2016) 5391–5411.
[8] G. Palmisano, V. Augugliaro, M. Pagliaro, L. Palmisano, Chem. Commun. (2007)
3425–3437.
[9] O. Tokode, R. Prabhu, L.A. Lawton, P.K.J. Robertson, UV LED Sources for
Heterogeneous Photocatalysis, in: D.W. Bahnemann, P.K.J. Robertson (Eds.),
Environ. Photochem. Part III, Springer Berlin Heidelberg, Berlin, Heidelberg,
2015: pp. 159–179.
[10] M.J. Sampaio, M.J. Lima, D.L. Baptista, A.M.T. Silva, C.G. Silva, J.L. Faria,
Chem. Eng. J. 318 (2017) 95–102.
[11] M.J. Sampaio, R.R. Bacsa, A. Benyounes, R. Axet, P. Serp, C.G. Silva, A.M.T.
Silva, J.L. Faria, J. Catal. 331 (2015) 172–180.
[12] C.G. Silva, M.J. Sampaio, S.A.C. Carabineiro, J.W.L. Oliveira, D.L. Baptista, R.
Bacsa, B.F. Machado, P. Serp, J.L. Figueiredo, A.M.T. Silva, J.L. Faria, J. Catal.

316 (2014) 182–190.
[13] C.G. Silva, M.J. Sampaio, R.R.N. Marques, L.A. Ferreira, P.B. Tavares, A.M.T.
Silva, J.L. Faria, Appl. Catal. B Environ. 178 (2015) 82–90.
[14] M.J. Sampaio, J.W.L. Oliveira, C.I.L. Sombrio, D.L. Baptista, S.R. Teixeira,
S.A.C. Carabineiro, C.G. Silva, J.L. Faria, Appl. Catal. A Gen. 518 (2016) 198–
205.
[15] J. Chen, J. Cen, X. Xu, X. Li, Catal. Sci. Technol. 6 (2016) 349–362.
[16] A. Corma, H. Garcia, J. Catal. 308 (2013) 168–175.
[17] G. Camera-Roda, V. Augugliaro, A. Cardillo, V. Loddo, G. Palmisano, L.
Palmisano, Chem. Eng. J. 224 (2013) 136–143.
[18] V. Augugliaro, G. Camera-Roda, V. Loddo, G. Palmisano, L. Palmisano, F.
Parrino, M.A. Puma, Appl. Catal. B Environ. 111–112 (2012) 555–561.
[19] A. Di Paola, M. Bellardita, B. Megna, F. Parrino, L. Palmisano, Catal. Today. 252
(2015) 195–200.
[20] M.J. Sampaio, A. Benyounes, P. Serp, J.L. Faria, C.G. Silva, Appl. Catal. A Gen.
551 (2018) 71–78.
[21] C. Tian, Q. Zhang, A. Wu, M. Jiang, Z. Liang, B. Jiang, H. Fu, Chem. Commun.
48 (2012) 2858–2860.
[22] K. Qi, B. Cheng, J. Yu, W. Ho, J. Alloys Compd. 727 (2017) 792–820.
[23] H. Yao, F. Li, J. Lutkenhaus, M. Kotaki, H.-J. Sue, AIMS Mater. Sci. 3 (2016)
1410–1425.
[24] L.-P. Zhu, G.-H. Liao, W.-Y. Huang, L.-L. Ma, Y. Yang, Y. Yu, S.-Y. Fu, Mater.

Sci. Eng. B. 163 (2009) 194–198.
[25] P. Serp, Carbon nanotubes and nanofibers in catalysis, in: Carbon Mater. Catal.,
John Wiley & Sons, Inc., 2008: pp. 309–372.
[26] R. Bacsa, Y. Kihn, M. Verelst, J. Dexpert, W. Bacsa, P. Serp, Surf. Coatings
Technol. 201 (2007) 9200–9204.
[27] M.I. Qureshi, F. Patel, N. Al-Baghli, B. Abussaud, B.S. Tawabini, T. Laoui,
Bioinorg. Chem. Appl. 2017 (2017) Article ID 1624243.
[28] M.J. Sampaio, R.R.N. Marques, P.B. Tavares, J.L. Faria, A.M.T. Silva, C.G. Silva,
J. Environ. Chem. Eng. 1 (2013) 945–953.
[29] L. Tian, L. Ye, K. Deng, L. Zan, J. Solid State Chem. 184 (2011) 1465–1471.
[30] L.M. Pastrana-Martínez, S. Morales-Torres, V. Likodimos, J.L. Figueiredo, J.L.
Faria, P. Falaras, A.M.T. Silva, Appl. Catal. B Environ. 123 (2012) 241–256.
[31] K.S.W. Sing, D.H. Everett, R.A.W. Haul, L. Moscou, R.A. Pierotti, J. Rouquerol,
T. Siemieniewska, Pure Appl. Chem. 57 (1985) 603–619.
[32] M. Thommes, Chem. Ing. Tech. 82 (2010) 1059–1073.
[33] B. Baruwati, D.K. Kumar, S. V Manorama, Sensors Actuators B Chem. 119 (2006)
676–682.
[34] S.S. Kumar, P. Venkateswarlu, V.R. Rao, G.N. Rao, Int. Nano Lett. 3 (2013) 30–
36.
[35] D. Thapa, J. Huso, J.L. Morrison, C.D. Corolewski, M.D. McCluskey, L. Bergman,
Opt. Mater. (Amst). 58 (2016) 382–389.
[36] J. Mu, C. Shao, Z. Guo, Z. Zhang, M. Zhang, P. Zhang, B. Chen, Y. Liu, ACS

Appl. Mater. Interfaces. 3 (2011) 590–596.
[37] C. Lu, Y. Wu, F. Mai, W. Chung, C. Wu, W. Lin, C. Chen, J. Mol. Catal. A Chem.
310 (2009) 159–165.
[38] K.M. Lee, C.W. Lai, K.S. Ngai, J.C. Juan, Water Res. 88 (2016) 428–448.
[39] K. Qi, H. Qi, Y. Xie, Y. Wang, Gen. Chem. 3 (2017) 113–117.
[40] M. Zhang, Q. Wang, C. Chen, L. Zang, W. Ma, J. Zhao, Angew. Chem. Int. Ed.
48 (2009) 6081–6084.
[41] A.O. Kondrakov, A.N. Ignatev, V. V Lunin, F.H. Frimmel, S. Bräse, H. Horn,
Appl. Catal. B Environ. 182 (2016) 424–430.
[42] Q. Wang, M. Zhang, C. Chen, W. Ma, J. Zhao, Angew. Chemie - Int. Ed. 49 (2010)
7976–7979.
[43] C. Ling, X. Ye, J.J. Zhang, J.J. Zhang, S. Zhang, S. Meng, X. Fu, S. Chen, Sci.
Rep. 7 (2017) 27–43.
[44] R. Marotta, I. Di Somma, D. Spasiano, R. Andreozzi, V. Caprio, Chem. Eng. J.
172 (2011) 243–249.
[45] M.J. Lima, P.B. Tavares, A.M.T. Silva, C.G. Silva, J.L. Faria, Catal. Today. 287
(2017) 70–77.
[46] S. Yurdakal, G. Palmisano, V. Loddo, O. Alagoz, V. Augugliaro, L. Palmisano,
Green Chem. 11 (2009) 510–516.
[47] L.P. Hammett, J. Am. Chem. Soc. 59 (1937) 96–103.
[48] C.G. Silva, J.L. Faria, ChemSusChem. 3 (2010) 609–618.