Charge transfer and photocatalytic activity in CuO/TiO2 nanoparticle heterojunctions synthesised through a rapid, one-pot, microwave solvothermal route

Article abstract of DOI:10.1002/cctc.201500315

Rapid charge carrier recombination is a major limiting factor over efficiency in many semiconductor photocatalysts. To address this, copper(II) oxide/titanium dioxide (CuO/TiO₂) heterojunctions were synthesised through a novel, rapid solvothermal microwave procedure using a low-cost copper precursor and commercial P25 TiO₂, taking as little as five minutes to synthesise well-defined CuO nanoparticles onto the host TiO₂, achieving an intimate contact. The resultant composites encompass pure CuO particles of approximately 6-7 nm diameter, confirmed by means of high resolution transmission electron microscopy and X-ray photoelectron spectroscopy analysis. Photoelectrochemical water splitting was enhanced by nearly 2 times using the junction, whilst ≈1.6 times enhancement in the photocatalytic mineralisation of a model organic pollutant 2,4-dichlorophenoxyacetic acid (2,4-D) was observed. Furthermore, we studied the initial decomposition mechanism of 2,4-D by means of GC-MS analysis. The increase in catalytic activity, investigated by impedance analysis (Mott-Schottky plots) and photoluminescence spectra, is attributed to photoelectron transfer from the more negative conduction band (CB) of TiO₂ to CuO, leaving the photohole on TiO₂ to take part in oxidation reactions. This strategy allows for in situ charge separation which facilitates superior photocatalytic activity for both pollutant degradation and water splitting.

Full text of DOI:10.1002/cctc.201500315

DOI: 10.1002/cctc.201500315
Full Papers
Charge Transfer and Photocatalytic Activity in CuO/TiO2
Nanoparticle Heterojunctions Synthesised through
a Rapid, One-Pot, Microwave Solvothermal Route
Savio J. A. Moniz and Junwang Tang*[a]
Rapid charge carrier recombination is a major limiting factor
over efficiency in many semiconductor photocatalysts. To ad-
dress this, copper(II) oxide/titanium dioxide (CuO/TiO2) hetero-
junctions were synthesised through a novel, rapid solvother-
mal microwave procedure using a low-cost copper precursor
and commercial P25 TiO2, taking as little as five minutes to syn-
thesise well-defined CuO nanoparticles onto the host TiO2, ach-
ieving an intimate contact. The resultant composites encom-
pass pure CuO particles of approximately 6–7 nm diameter,
confirmed by means of high resolution transmission electron
microscopy and X-ray photoelectron spectroscopy analysis.
Photoelectrochemical water splitting was enhanced by nearly
2 times using the junction, whilst ꢀ1.6 times enhancement in
the photocatalytic mineralisation of a model organic pollutant
2,4-dichlorophenoxyacetic acid (2,4-D) was observed. Further-
more, we studied the initial decomposition mechanism of 2,4-
D by means of GC-MS analysis. The increase in catalytic activity,
investigated by impedance analysis (Mott–Schottky plots) and
photoluminescence spectra, is attributed to photoelectron
transfer from the more negative conduction band (CB) of TiO2
to CuO, leaving the photohole on TiO2 to take part in oxidation
reactions. This strategy allows for in situ charge separation
which facilitates superior photocatalytic activity for both pollu-
tant degradation and water splitting.
Introduction
In recent years, semiconductor photocatalysis has shown great
potential as a low-cost, environmental friendly and sustainable
technology to remove harmful contaminants from water sup-
plies and for renewable energy generation (water splitting into
hydrogen fuel).[1] However, inefficient carrier separation result-
ing in rapid recombination limits overall efficiencies for both
contaminant removal and renewable energy applications.[2]
Heterojunctions based on the combination of narrow band-
gap metal oxides and/or co-catalysts with titanium dioxide
(TiO2) have been widely used as composite photocatalysts to
alleviate recombination and speed up reaction kinetics, where-
by charge carriers are vectorially transferred from one semicon-
ductor to either another semiconductor or co-catalyst in inti-
mate contact.[3] Importantly, photocatalysis facilitates a cheap
and efficient route for pollutant removal from water supplies,
which is of utmost importance in developing countries. A fur-
ther advantage is that through using the environmentally
friendly oxidant O2, photocatalytic reactions can be performed
at room temperature through formation of superoxide radicals
(oxygen reduction reaction; ORR).[4] In aqueous solution, va-
lence band holes can oxidise hydroxyl ions to highly reactive
hydroxyl radicals, which have been shown to be crucial for
degradation of organic contaminants.[5] P25 TiO2 is one of the
most commonly studied photocatalysts, owing to its low cost,
high dispersibility (surface area), appreciable stability in solu-
tion and reasonable activity under full arc (visible and UV) light
irradiation. As P25 is comprised of ꢀ70% anatase and 30%
rutile TiO2 polymorphs, it can itself be regarded as a homojunc-
tion with a staggered energy band alignment at the anatase/
rutile interface, therefore, photogenerated electrons will prefer-
entially move to anatase owing to its lower conduction band
minimum energy, and holes will move to rutile owing to its
higher valence band maximum energy.[6] To improve the activi-
ty further, cheap, earth abundant, non-toxic co-catalysts can
be brought in close contact to the TiO2 surface to aid charge
separation.[7] It is preferable that the method employed to de-
posit co-catalysts is both cost effective and easy to scale up.
For example, photodeposition of FeIII species on TiO2 was dem-
onstrated to be significantly more effective for both water
splitting and pollutant degradation, caused by electron transfer
to Fe2O3 from TiO2.[8] Theoretical calculations revealed that this
junction also renders the TiO2 surface more reactive toward O2
molecules for the ORR, which is the rate-determining step in
photocatalytic environmental purification. However, the low ef-
ficiency of photodeposition (less than 50%) limits the overall
potential for scale-up, hence a mild, alternative efficient syn-
thetic strategy is required.
Herein, we demonstrate a highly efficient solvothermal mi-
crowave precipitation method to fabricate pure CuO/TiO2
nanocomposite junction photocatalysts, which is performed
rapidly under relatively mild conditions and derived from
cheap, soluble CuII salt precursors. Surprisingly, there are fewer
reports on the utilisation of CuO as either a photocatalyst or
[a] Dr. S. J. A. Moniz, Dr. J. Tang
Department of Chemical Engineering
University College London
Torrington Place, London, WC1E 7 JE (United Kingdom)
E-mail: Junwang.tang@ucl.ac.uk
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201500315.
ChemCatChem 2015, 7, 1659 – 1667
1659
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
centrifugation (3000 rpm, 30 min) and washed several times with
ethanol and dried at 1008C for 24 h. The exact amount (i.e. the effi-
ciency) of CuO loading was determined by energy-dispersive X-ray
spectroscopy (EDS) analysis.
co-catalyst despite possessing a band-gap in the visible region
and being composed of cheap, earth abundant elements. Ha-
shimoto has described the enhanced photocatalytic activity of
CuII-loaded TiO2 for organic oxidation, caused by electron
transfer to CuO forming CuI, then reduction of O2 which regen-
erates CuII again.[9,10] CuO as a co-catalyst has also been proven
to improve photocatalytic activity when combined with WO3
through a similar mechanism.[11]
Characterisation
X-ray diffraction (XRD) was performed using a Bruker D4 diffrac-
tometer in reflection geometry using Cu-Ka radiation (l=
1.54054 Š, 40 kV, 30 mA). UV/vis absorption spectra were collected
using a Shimadzu UV/vis 2550 spectrophotometer equipped with
an integrating sphere. Reflectance measurements were performed
on powdered samples, using a standard barium sulphate powder
as a reference. The reflection measurements were then converted
to absorption spectra using the Kubelka–Mulk transformation. The
morphologies of the products were characterised by HRTEM (JEOL-
2010F) coupled with an EDS detector (Oxford Instruments). High-
resolution XPS was performed using a Thermo Scientific K-alpha
photoelectron spectrometer using monochromatic Al-Ka radiation;
peak positions were calibrated to carbon (284.5 eV) and plotted
using the CasaXPS software. Specific surface area measurements
were taken using the BET method (N2 absorption, TriStar 3000, Mi-
cromeritics). Photoluminescence (PL) spectra were conducted
using a Renishaw 1000 Raman system using a l=325 nm laser at
room temperature.
Nevertheless, the ambiguity over the positions of the con-
duction and valence band positions in CuO has perhaps limit-
ed its investigation for photocatalytic applications. Further-
more, there are very few reports of the synthesis of robust
junctions using microwave precipitation, which is somewhat
surprising given the potential advantage of utilising this tech-
nique, which include high-speed, efficiency and excellent
yield.[12] Indeed, if the microwave-assisted reaction is per-
formed in a closed vessel under high pressure and stirring,
more energy input at the same temperature coupled with
a huge acceleration in reaction time can be achieved, and,
therefore, a reaction which under conventional hydrothermal
or solvothermal conditions would take several hours, can be
completed in as little as a few minutes. The reported CuO/TiO2
composites were characterised fully by means of XRD, UV/vis,
X-ray photoelectron spectroscopy (XPS) and high resolution
transmission electron microscopy (HRTEM), which altogether
revealed copper species present as pure CuII, that is, CuO in
the composites. The loading of CuO on TiO2 was optimised
and investigated for photocurrent generation and degradation
of a model herbicide 2,4-dichlorophenoxyacetic acid (2,4-D)
under full arc solar irradiation. 2,4-D is a member of the highly
toxic phenoxy family of herbicides and pesticides which were
widely used in many developing countries and applied
through spraying onto open crop fields for effective weed con-
trol, resulting in contaminated river and groundwater supplies.
Importantly, the initial photocatalytic degradation mechanism
of 2,4-D was elucidated by means of a gas chromatography-
mass spectrometry (GC-MS) study of the degradation inter-
mediates and the effect of different radicals on its degradation
were probed. The mechanism of improved photocatalytic ac-
tivity for CuO/TiO2 was probed using photoluminescence spec-
troscopy and Mott–Schottky analysis to determine band offsets
and, therefore, possible charge carrier migration pathways in
CuO/TiO2.
Photocatalysis measurements
The photocatalytic activity of the TiO2-based composites was firstly
investigated by measuring the photodegradation of 2,4-D, which
was monitored using UV/vis spectroscopy for decomposition and
total organic carbon (TOC) analysis for mineralisation, in an identi-
cal procedure to that published previously.[8] In a typical measure-
ment, 10 mg of photocatalyst was suspended in 100 mL of
a 50 ppm aqueous solution of 2,4-D in deionised water. The sus-
pension was sonicated then stirred in the dark for 2 h to achieve
the equilibrium adsorption. The suspension was then illuminated
with a full arc 300 W Xenon lamp. The concentration change of
2,4-D was monitored by measuring the UV/vis absorption spectra
of the suspension at regular 30 min intervals, using a Shimadzu
UV/visible 2550 spectrophotometer. Owing to the very small parti-
cle size of CuO, for each measurement the suspension was filtered
three times using a filter syringe to remove the insoluble photo-
catalyst before UV/vis measurements. The TOC was measured at
the same time intervals using a Shimadzu TOC-L analyser to con-
firm the degree of mineralisation. GC-MS analysis of the intermedi-
ates formed during the mineralisation of 2,4-D was obtained using
a Shimadzu GC-MS 2010 equipped with a silica capillary column by
injecting ꢀ0.1 ml of a 500 ppm initial concentration of 2,4-D in
water.
Experimental Section
A microwave solvothermal method was used to fabricate CuO and
CuO/TiO2 composite photocatalysts. In a typical experiment, 0.3 g
of commercial P25 (Degussa, composed of about 70 % anatase
and 30 % rutile) with a surface area of about 49 m2 gÀ1 was added
to 60 mL absolute ethanol (99.99%, VWR) under vigorous stirring
to form slurry. To this, the appropriate amount percentage weight
of copper(II) acetate monohydrate [(Cu(CO2CH3)2·H2O), (98%, VWR)]
(e.g. 1 wt %) was added under stirring. The solution was trans-
ferred to a Teflon hydrothermal reactor equipped with a magnetic
stirrer bar and temperature probe. The solution was subjected to
microwave heating (400 W) at 1508C for a hold time of 5 min at
a ramp rate of 108CminÀ1. The obtained slurry was filtered using
Photocurrent studies were performed to further test the photo-
catalytic activity of the composite junction. In a typical run, 15 mg
of photocatalyst and 10 mL of Nafion solution (5 wt %) were dis-
persed in a 1 mL water/isopropanol mixed solvent (3:1 v/v) fol-
lowed by 30 min ultrasound sonication to form a homogeneous
colloid suspension. For the measurements, 100 mL of the catalyst
colloid was deposited onto fluorine-doped tin oxide (FTO) conduc-
tive glass (TEC 15, 35 W, ꢀ1 cm2) by spin-coating at 500 rpm for
30 s to form a working electrode. A platinum mesh was used as
a counter electrode whilst an Ag/AgCl electrode was employed as
the reference electrode in a three electrode photoelectrochemical
cell. The electrolyte was 0.5 m Na2SO4 (pH 6.5) aqueous solution
ChemCatChem 2015, 7, 1659 – 1667
www.chemcatchem.org
1660
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
degassed with argon for 30 min. Mott–Schottky (impedance) plots
were obtained at a frequency of 1 kHz in the dark with an AC am-
plitude of 5 mV. Equation (1) was used to determine flat band po-
tential (Vfb).
20 wt% Cu loading all of the XRD peaks can be indexed to
both anatase and rutile forms of TiO2, with no obvious peaks
that could be ascribed to copper species, possibly due to the
low nominal loading or high dispersion over TiO2. In the ab-
sence of P25, copper(II) acetate was dissolved in ethanol and
heated to 1508C. The resultant dark brown solid was washed
and dried in an identical manner used for synthesis of the Cu/
TiO2 composites. The XRD pattern of this material (Figure 1a)
revealed the formation of pure CuO (Tenorite, JCPDS 48–1548),
indicating that the TiO2 does not catalyse the production of
copper(II) oxide species in this system. Thus microwave treat-
ment appears to be an effective preparative route to synthe-
sise CuO particles very rapidly.
ꢀ
ꢁ
1
2
kBT
e
¼
V À Vfb
À
ð1Þ
C2 e0ereNA
Here NA is the carrier density, e0 is the permittivity in vacuum; er is
the relative permittivity; V is the applied potential and T is the ab-
solute temperature. Value e is the electronic charge, and kB is the
Boltzmann constant. Hence a plot of 1/C2 versus potential (V) will
yield a line, which when extrapolated to the x-axis, will correspond
to the flat-band potential of the semiconductor.[13] Potentials were
referenced to the reversible hydrogen electrode (RHE) using the
Nernst equation [Eq. (2)]:
Raman spectroscopy was also utilised to identify the species
present in the composites. Figure S1 shows the Raman spectra
recorded at a laser wavelength of 514.5 nm for CuO/P25, pris-
tine P25 TiO2 and CuO powder synthesised by microwave sol-
vothermal treatment. As the anatase and rutile polymorphs
belong to different space groups (I41/amd, Z=4 for anatase,
P42/mnm, Z=2 for rutile), the peaks centred at 142 (Eg), 197
(Eg), 394(B1g), 517 (A1g) and 640 cmÀ1 (Eg) were attributed to
the anatase phase, whilst low intensity peak broadening at 445
and 610 cmÀ1 was attributed to the Eg and A1g vibrational
modes of the rutile phase.[14] The CuO particles exhibited the
characteristic Raman scattering pattern attributed to CuO in
the literature;[15] the broad peak corresponding to the Ag mode
was found at 282 cmÀ1, with the Bg vibration contributing to
the peaks observed at 335 and 620 cmÀ1. No other peaks were
identified from the Raman pattern which provides further evi-
dence that the sample is pure CuO. Likewise, for the 5% CuO/
P25 sample, broad peaks for CuO were observed at 280 and
618 cmÀ1, with the high intensity peak at 149 cmÀ1 originating
from the anatase phase of TiO2.
ERHE ¼ E0Ag=AgCl þ EAg=AgCl þ 0:059 pH
ð2Þ
E0
¼ 0:1976 at 25 ꢁC:
Ag=AgCl
Results and Discussion
Microwave solvothermal treatment of the preformed P25 TiO2
particles with copper(II) acetate at a temperature of 1508C in
ethanol solvent at a holding time of 5 min resulted in light-
brown coloured fine particles after filtration and washing in
ethanol. A typical X-ray diffraction pattern of the largest load-
ing of Cu (20 wt% Cu) on P25 TiO2 is shown in Figure 1b, to-
gether with the XRD pattern of untreated P25 powder. For
The UV/vis absorption spectra of pristine P25 and various
CuO/TiO2 with different CuO loadings (Figure 2) revealed the
expected band-gap TiO2 absorption at 3.2 eV, however, increas-
ing the CuO loading resulted in a slight red-shift and an in-
creased visible absorption was observed (l>400 nm), which
was not surprising given the dark brown appearance of pure
CuO. Indeed from the absorption spectrum of pure CuO (Sup-
porting Information Figure S2), a band-gap absorption of
ꢀ1.3 eV and strong visible light absorption was exhibited, as
expected for CuO according to literature reports.[16]
Figure 1. XRD patterns of a) CuO nanoparticles, b) 20% CuO/P25 and P25.
Figure 2. UV/vis absorption spectra of CuO/P25 composites.
ChemCatChem 2015, 7, 1659 – 1667
www.chemcatchem.org
1661
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 4. TEM images of a) 5% CuO/P25 and b) 10% CuO/P25, c) HRTEM of
CuO nanoparticles synthesised by microwave solvothermal treatment,
d) HRTEM of 20% CuO/P25.
5–10 nm diameter CuO particles (Figure 5c). The particle size
distribution (see Figure S4, Supporting Information) indicated
that a mean particle size of between 6–7 nm could be synthes-
ised under these microwave conditions. The size and morphol-
ogy are in stark contrast to CuO particles synthesised under
microwave conditions using copper(II) acetate precursor with
and without urea, reported recently, which exhibited a flower-
like morphology consisting of ꢀ500 nm diameter structures.[19]
In the present work, it is proposed that CuO formation is most
likely attributable to the decomposition of in situ formed
Cu(OH)2 during hydrolysis or its reaction with ethanol solvent.
From this sample, the lattice spacing which corresponds to the
[111] 0.23 nm interplanar spacing of CuO was detected. The
HRTEM of the 20% CuO/P25 junction revealed lattice spacings
of 0.35 nm for anatase [101] and 0.32 nm for rutile [110]
planes respectively. Additionally, the CuO [110] lattice fringe
was observed at 0.27 nm for a single CuO particle which is
shown on the surface of a TiO2 particle. Thus a robust, intimate
contact is formed between the two individual materials, which
is beneficial for charge transfer. Notably, owing to the high
energy of HRTEM, some beam damage was observed which
led to distortion of the edges of the nanoparticles at this high
magnification. To check the distribution of CuO species on
TiO2, EDS mapping was performed on several samples (see
Supporting Information, Figure S5). Somewhat surprisingly, the
EDS map indicates that the loading efficiency of microwave
synthesis is close to 100% in these composites, for example,
the 10 wt% nominal copper loaded sample resulted in an
actual loading amount of Cu species as high as 9.7%. Similarly,
for the 5 wt% Cu loading, an actual loading of 4.4% Cu spe-
cies was observed. The distribution also appeared to be quite
Figure 3. XPS Spectra of a) Cu2p regions in CuO and CuO/P25 and b) Ti2p
region in CuO/P25.
X-ray photoelectron spectroscopy was utilised to ascertain
the exact chemical valence state of the copper species present
in the samples. Figure 3a shows the XPS spectra of both pure
CuO and 20 wt% CuO/P25 synthesised by microwave solvo-
thermal synthesis. In both cases, copper species is present as
CuII as both samples exhibit characteristic peaks for Cu2p3/2 at
934.2 eV and Cu2p1/2 at 954 eV.[17] The shake-up satellite peaks
characteristic of the presence of CuII are clearly observed be-
tween 940–945 eV and 963 eV. In comparison, but as expected,
the microwave solvothermal synthesis method did not alter
the chemical state of titanium in the CuO/P25 composite, evi-
denced by the characteristic Ti2p3/2 peak at 458.7 eV, associat-
ed with the presence of Ti4 + (Figure 3b).[18] XPS, therefore, pro-
vides additional evidence for the formation of pure CuII species
in the junctions.
The as-synthesised particles were analysed by TEM imaging.
The TEM image of pristine P25 TiO2 is shown in the Supporting
Information, Figure S3, which exhibits particles of approximate-
ly 20–60 nm in diameter. In contrast, the TEM images of both
5% and 20% CuO/P25 are shown in Figure 4a and b respective-
ly. One can still distinguish the larger P25 TiO2 as straight-
edged particles, however, there appears to be clusters of small-
er particles randomly positioned on the TiO2 facets. EDS analy-
sis revealed these particles to consist of copper species, whilst
the HRTEM image of the sample synthesised by microwave sol-
vothermal treatment in the absence of P25 resulted in
ChemCatChem 2015, 7, 1659 – 1667
www.chemcatchem.org
1662
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
however, the largest increase
was exhibited by the 5% and
10% CuO/P25 samples. In fact
the rate constant for the best
performing sample (5% CuO/
TiO2) was calculated to be
2.077 minÀ1
times that of pristine P25 whose
rate constant was 1.324 minÀ1
, approximately 1.6
,
whereas the rate constant for
the 10% CuO/TiO2 composite
was a little lower at 1.547 minÀ1
.
In all cases the degradation pro-
file represents first-order kinetics.
To ensure that 2,4-D was miner-
alised in this process instead of
being incompletely decomposed
to other products, the total or-
ganic carbon (TOC) analysis was
measured for all of the samples.
The results obtained were in
good agreement with the degra-
dation results observed in the
UV/vis spectra, with the 5%
CuO/TiO2 representing the best
performing sample. Furthermore,
as the TOC value reached zero
after a period of ꢀ120 min illu-
Figure 5. a) Typical degradation of 2,4-D over 1% CuO/P25 followed by UV/vis spectroscopy, b) degradation of
2,4-D over various CuO/P25 junction composites, c) TOC removal over CuO/P25 junction composites, d) degrada-
tion rates of the CuO/P25 composites, revealing 5% CuO/P25 to be the most active sample.
homogeneous over the sample, which indicates the advantage
of microwave synthesis for rapid and efficient nanoparticle syn-
thesis compared to other methods of co-catalyst deposition,
such as photodeposition or impregnation.
mination (for the 5% CuO/P25 sample), we can indeed con-
clude that 2,4-D was mineralised in this process. Hence the op-
timum CuO loading on P25 was found to be 5% CuO, closely
followed by 10% CuO, 1% CuO, 0.5% and lastly, 20% CuO, the
latter of which had a detrimental effect compared to pristine
P25.
Photocatalytic testing
No visible light driven activity (l> 420 nm) was observed
for any of the composites, which demonstrates that excitation
in CuO is not responsible for photocatalytic activity. Further-
more, it is likely that the poorer activity of 20% CuO/TiO2 is
owed to screening of photons by highly absorbing CuO, which
competes with TiO2 for light absorption. The active species in-
volved in the photocatalytic mineralisation of 2,4-D over CuO/
TiO2 were investigated. Isopropanol was utilised as an hydroxyl
The application of these CuO/TiO2 heterojunction composites
for photocatalytic removal of toxic herbicides was evaluated
by the photo-mineralisation of 2,4-D, which was monitored by
UV/vis absorption spectroscopy and TOC analysis. Similar to
our previous report, zero degradation of a 50 ppm solution of
2,4-D in deionised water was observed after irradiation with
a full-arc 300 W solar lamp at room temperature in the ab-
sence of photocatalyst.[8] To assess the activity of bare CuO
toward 2,4-D degradation, 10 mg of CuO was dispersed into
a 50 ppm 2,4-D aqueous solution and irradiated with a full arc
300 W Xe light source for 4 h, which resulted in zero degrada-
tion of the pollutant over this time frame, likely, in part, to be
attributable to the band positions of CuO, which are not suffi-
ciently aligned for oxidation reactions or the one electron re-
duction of O2.[20] However, the pristine P25 showed appreciable
activity, with an approximate 80% removal efficiency over the
four hour period according to the UV/vis spectrum. A typical
UV/vis spectrum observed during the degradation of 2,4-D is
shown in Figure 5a. Upon introducing CuO onto P25 TiO2, a no-
table enhancement in 2,4-D removal was observed for the ma-
jority of samples. For low CuO loadings, that is, 0.5% and 1%,
a little enhancement in degradation over P25 was observed;
C
radical (OH ) scavenger, EDTA as a hole scavenger, and potassi-
C
um iodide as both hole and OH scavengers. The system was
also purged with argon gas to investigate the absence of su-
peroxide radicals on the mineralisation mechanism. No degra-
dation of the organic was observed in the presence of the
hole scavenger, whereas a 25% removal was obtained in the
À
C
absence of O2 and ꢀ75% removal using OH scavenger, com-
pared to CuO/P25 (see Figure S6), somewhat similar to the re-
C
sults obtained by Pignatello using methanol as the OH scav-
enger.[21] Addition of both hole and hydroxyl radical scavengers
together resulted in no decomposition, indicating the decom-
position of 2,4-D is dominated by a hole mechanism. In fact, it
has been reported that the photocatalytic degradation of
C
2,4-D is 70% governed by OH and 30% by direct oxidation by
photoholes.[22]
ChemCatChem 2015, 7, 1659 – 1667
www.chemcatchem.org
1663
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
To clarify the reasons for the enhanced photoactivity of the
heterojunctions, the surface areas of the photo-deposited
CuO/TiO2 samples were measured using a surface area analyser
using the BET method (Table S1, Supporting information). A
typical N2 adsorption-desorption isotherm for 5% Cu/P25 is
shown in Figure S7, Supporting Information. For pure CuO par-
ticles, the surface area obtained was 77 m2gÀ1, much higher
than pristine P25 (49 m2gÀ1). However, there appears to be
almost negligible variation in surface area for the CuO/TiO2
composites indicating that microwave treatment did not ad-
versely affect the surface area of P25, nor increase the surface
area which could have resulted in better activity for 2,4-D deg-
radation. The stability of a photocatalyst is critically important
to lower overall manufacturing costs and would be more ben-
eficial to the environment. To assess the stability of these com-
posites for sequential photocatalytic mineralisation of 2,4-D, 5
cycles of repeated testing using the same photocatalyst
sample was conducted (see Supporting Information Figure S8).
It is notable that no significant decrease in activity was ob-
served during this repeated testing which is testament to the
appreciable stability of these robust composites, which could
allow for their utilisation in commercial devices for photocata-
lytic water treatment. Only a slight decrease in activity was ob-
served after the fifth cycle. Furthermore, post degradation XPS
analysis of the sample revealed very little difference according
to the Cu2p spectrum (Figure S9, Supporting Information),
apart from a slight broadening of the Cu2p3/2 signal at
934.2 eV.
ther degradation to water and CO2 eventually takes place.[23]
The initial decomposition pathway, therefore, involves cleavage
of the carboxylic acid moiety through its replacement with OH.
This is followed by addition of OH at the adjacent carbon, fol-
lowed by sequential substitution of the remaining chlorine
atoms by OH groups. Scheme 1 illustrates the major intermedi-
ates detected by GC-MS analysis.
Scheme 1. Intermediates detected using GC-MS during the photocatalytic
mineralisation of 2,4-D over P25 TiO2.
Photoelectrochemical testing of the composites was per-
formed to verify the mechanism of charge transfer in the het-
erojunctions. Photoelectrodes were synthesised by spin-coat-
ing the colloidal photocatalysts, which were suspended in
a Nafion/isopropanol solvent onto FTO-coated glass substrates.
They were subsequently dried at 2008C to remove organic
species. Figure 6a shows the current-voltage curves of 5%
CuO/P25 and bare P25 in 0.5m Na2SO4 electrolyte (pH 6.5)
under both light and dark conditions. The photocurrent of 5%
CuO/P25 under full arc 150 W illumination is ꢀ30 mAcmÀ2 (at
1 V versus Ag/AgCl) which is over two times higher than that
of P25 (ꢀ12 mAcmÀ2), which is in reasonable agreement with
the increased activity (almost 1.6 times) observed for photo-
catalytic degradation of 2,4-D, compared to P25. Dark current
was negligible for both samples. The onset potential of CuO/
P25 appears slightly shifted toward more positive potentials,
indicating that the photo-generated electrons in TiO2 are rapid-
ly trapped by CuO, (see Figure 7a, inset ii). One can, therefore,
qualitatively reason that the conduction band of CuO is more
positive than that of anatase TiO2. Furthermore, the stability of
5% CuO/P25 was found to be excellent over the 1 hour test
period; the transient photocurrent was recorded at a fixed po-
tential of 0.5 V (versus Ag/AgCl) in the same electrolyte (Fig-
ure 7a, inset i). Initially the photocurrent transient exhibited
a sharp decrease, owing to fast recombination at the start of
the experiment, however, for the P25 sample, a gradual de-
crease in photocurrent was observed, which we attribute to re-
combination rather than photocorrosion. We used photolumi-
nescence spectroscopy (PL) to investigate the feasibility of
charge transfer in the CuO/P25 system. PL spectra has been
used extensively to reveal the efficiency of charge carrier trap-
ping, transfer, and separation and to investigate the fate of
photogenerated electrons and holes in semiconductors, be-
cause the PL emission results from the recombination of free
charge carriers.[24] Figure 6b shows the PL spectra of P25 and
To investigate the initial decomposition pathway of the
model contaminant 2,4-D, we utilised GC-MS to analyse the
degradation products at each stage of the photocatalytic min-
eralisation process. To improve the sensitivity of the GC-MS
technique, a 500 ppm aqueous solution of 2,4-D was used, to-
gether with 100 mg of P25 photocatalyst. Direct liquid injec-
tion of a 0.1 ml aqueous sample of 2,4-D in the absence of pho-
tocatalyst or light irradiation resulted in the expected peak at
m/z + =221, corresponding to the molecular ion of 2,4-D. After
30 min illumination, a peak at m/z + =164 was observed, indi-
cating production of 2,4-dichlorophenol species (Mw =
164 gmolÀ1). A further peak was observed after 60 min at
m/z+ =178, corresponding to either of the dihydroxy-substitut-
ed dichlorobenzenes; 3,5-dichlorocatechol or 4,6-dichlororesor-
cinol. After 120 min degradation, the only observed product
was dihydroxychlorobenzene, followed by trihydroxybenzene
after 150 min. This indicates that each chlorine atom is substi-
tuted for an OH moiety during the degradation process. It has
previously been shown using liquid chromatography, that the
first step in the photocatalytic degradation of 2,4-D is the one-
electron oxidation of the carboxylic acid group followed by hy-
drolysis to yield 2,4-dichlorophenol and glycoxylic acid, which
occurs most efficiently at pH 3.[21] Initial cleavage of the carbox-
ylic acid moiety of 2,4-D was also confirmed by Djebbar
et al.[22] using P25 TiO2 photocatalyst and GC-MS measure-
ments, however, the major final reaction product measured in
their system was 2-chloro-1,4-benzoquinone.
Further illumination of the solution did not yield conclusive
results from our GC-MS analysis; however, it is likely that fur-
ChemCatChem 2015, 7, 1659 – 1667
www.chemcatchem.org
1664
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Figure 6. a) I–V curves for P25 and 5% CuO/P25 in Na2SO4 electrolyte
(pH 6.5), insets show i) the transient photocurrent measured at 0.5 V versus
Ag/AgCl and ii) enlarged section of the I–V curve showing onset region;
b) Photoluminescence spectra of P25 and 5% CuO/P25 recorded at room
temperature using a 325 nm Raman PL laser.
5% CuO/P25 recorded at room temperature under 325 nm UV
excitation. A broad peak centred at 525 nm was observed for
pristine P25, whilst strikingly, a very broad emission at much
lower intensity was exhibited by CuO/P25. Furthermore, this
emission spectrum appears to have red-shifted and is centred
at approximately 575 nm. Owing to its low intensity we were
not able to accurately fit the spectrum to a model; however,
the CuO/P25 exhibits ꢀ85% reduction in luminescence, which
one can ascribe to efficient rectifying charge transfer between
the two materials despite possessing a similar band-gap ab-
sorption (from the UV/vis spectra).
Figure 7. Mott–Schottky plots for a) CuO, b) anatase TiO2 ; c) mechanism of
charge transfer in CuO/TiO2 heterojunctions.
approximation of the valence and conduction band levels in
both p- and n-type semiconductors if other techniques are not
available.
From the Mott–Schottky plots we can estimate the flat-band
position (Efb), which is generally considered to be close to the
valence band position for p-type semiconductors and located
approximately 0.3 V below the conduction band for n-type
semiconductors.[27,28] Consequently, using the same experimen-
tal procedure described earlier for the fabrication of P25 and
CuO/P25 electrodes, CuO nanoparticles, which were synthes-
ised by microwave solvothermal treatment in the absence of
P25, were formed into a working electrode after spin coating.
A pure anatase TiO2 electrode was formed in a similar manner.
From the Mott–Schottky plot (Figure 7a), CuO exhibited the
characteristic p-type semiconductor direction of slope, which
when extrapolated to the x-axis yielded a flat-band position of
+0.51 V versus Ag/AgCl. On the other hand, pure anatase TiO2
Further experiments were undertaken to clarify the most
likely direction for charge transfer in the CuO/P25 heterojunc-
tion. To do this, one must first rationalise the relative band po-
sitions of the two materials, CuO, and the majority phase in
P25, anatase TiO2. One method to experimentally determine
the conduction and valence band positions of semiconductors
is through the interpretation of impedance (Mott–Schottky)
plots. Notably, however, the absolute values obtained by this
technique to determine band positions of semiconductors can
be strongly affected by particle size, inhomogeneity, surface
defects, vacancies, and measurement temperature, to name
just a few factors,[25,26] but the method can be used to give an
ChemCatChem 2015, 7, 1659 – 1667
www.chemcatchem.org
1665
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
displayed typical n-type behaviour and the slope is in the op-
posing direction to that of CuO (Figure 7b). The value for Efb
for anatase was determined to be À0.52 V (versus Ag/AgCl).
Consequently, the position of the valence band of CuO was
calculated to be +1.4 V (versus RHE) and its conduction band
located at +0.1 V (versus RHE); the latter value derived from
the band-gap absorption of CuO, which is approximately
1.3 eV, and is in good agreement with reported values.[29,30] For
TiO2, the conduction band is located at approximately À0.24 V
(versus RHE) and the valence band at +2.96 V (versus RHE), in
good agreement with the literature.[31] The difference in con-
duction band energies of the two materials suggests that elec-
tron transfer from TiO2 to CuO is feasible given the ꢀ0.34 V
offset which could provide sufficient driving force for charge
separation. This is further evidenced by the fact that pure CuO
exhibits zero activity for the organic decomposition, and thus
photocatalytic oxidation by CuO does not occur. Hashimoto
et al.[10] have proposed that loading of CuII on TiO2 results in
the in situ formation of CuI species, caused by electron transfer
from TiO2. In air, CuI species can reduce O2, whilst becoming
oxidised itself back to CuII. This multi-step reduction process is
initiated with a two electron reduction to peroxide [Eq. (3)]
and the following possible four-electron reductions [Eqs. (4)
and (5)].
in TiO2, followed by in situ rectifying electron transfer to the
conduction band of CuO. Photo-excited holes in TiO2 may then
migrate to the surface for the expected oxidation reactions. In
fact, as P25 is composed of anatase and rutile, the complete
CuO/P25 composite could be considered a triple junction,
however, the question over the direction of charge transfer be-
tween anatase and rutile phases is currently a topic of much
debate.[6] Furthermore, we noticed there is no photocatalytic
oxidation performed by CuO, which we will now discuss. First-
ly, as its valence band position is not well aligned for oxidation
reactions and thus exhibits no overpotential (driving force),
photocatalytic oxidation is not kinetically nor thermodynami-
cally feasible on the CuO surface. Secondly, and in contrast,
the surface of TiO2 has been demonstrated to be extremely
active for both photocatalytic oxidation reactions and surface
induced hydrophilicity, whereas crucially, CuO is inactive for
both processes.[36] This hydrophilicity is considered to be in-
duced by structural changes at the surface, that is, through the
diffusion of photogenerated holes to the surface of metal
oxides (TiO2) before being trapped at surface lattice oxygen
atoms, followed by the dissociative adsorption of a water mol-
ecule at these defect sites. It is more likely, therefore, that the
photohole diffuses to the TiO2 surface rather than into CuO
upon UV excitation, which rationalises the charge transfer
pathway depicted in Figure 7c and the lack of photocatalytic
activity exhibited by CuO alone. This also agrees with our ob-
servation upon which the photocatalytic decomposition of 2,4-
D is governed by a hole-dominated process.
2 CuI þ O2 þ 2 Hþ ! 2CuII þ H2O2
3 CuI þ O2 þ 4 Hþ ! 2 CuII þ CuIII þ 2 H2O
ð3Þ
ð4Þ
ð5Þ
4 CuI þ O2 þ 4 Hþ ! 4 CuII þ 2 H2O
Conclusions
As the reduction potential of O2/H2O2 is 0.7 V versus NHE,
this pathway could be viable for our CuO/TiO2 junction.[32]
Recently, a slightly different mechanism was reported for
CuO/TiO2 junctions tested for H2 evolution from water whereby
the authors observed significantly higher activity from 1.7%
CuO loaded TiO2 compared to the unmodified material, owing
to the in situ reduction of CuO to Cu species, which acts as
a co-catalyst for H2 production.[33] However, in the absence of
a strong hole scavenger such a mechanism would be unlikely
to occur in our junctions. Previously, CuI and Cu0 species were
photo-deposited onto TiO2 to yield Cu2O/TiO2/Cu triple junc-
tions that exhibited improved removal of the toxic organic pol-
lutant 2,4,5-Trichlorophenoxyacetic acid (2,4,5-T) from water.[34]
However, using this photo-reduction method it is extremely
difficult to control the selectivity and dispersibility of Cu and
Cu2O over the host photocatalyst. A slightly different mecha-
nism for increased photocatalytic activity for CuO/TiO2 hetero-
junctions studied for methylene blue decolourisation has been
put forward. Here, EPR spectroscopy revealed the presence of
highly dispersed CuO clusters and substitutional Cu2+ sites in
low weight percentage loadings of CuO in TiO2 (TiÀOÀCu link-
ages), whilst larger loadings resulted in CuO acting as a recom-
bination centre.[35] However, our mechanism is different. We
propose that the mechanism for the enhanced activity for
CuO/TiO2 junctions is a multi-step process: upon full-arc illumi-
nation there is light absorption and charge-carrier generation
To the best of our knowledge, we demonstrate for the first
time that mild modification of high surface area TiO2 photo-
catalysts with robust metal oxide redox co-catalysts using mi-
crowave co-precipitation is a promising strategy, which facili-
tates high loading efficiencies and results in the significant en-
hancement in photocatalytic activity for water splitting and
degradation of toxic organic pollutants. This technique could
be easily scaled up, for example, by using continuous flow
technology to rapidly synthesise large amounts of material for
commercial applications instead of the current batch method
described here. It was found that 5% CuO/P25 exhibited
a near 1.6 times enhancement in mineralisation of a model or-
ganic pollutant, 2,4-D, compared to the benchmark photocata-
lyst P25 TiO2. Furthermore, the photocurrent was found to be
enhanced by nearly 2 times and the CuO/P25 junction exhibit-
ed appreciable stability during both long-term water splitting
and organic degradation reactions. As our junctions do not in-
crease the visible light absorption of P25, their function is
purely for charge separation and transfer. We have also moni-
tored the photocatalytic mineralisation of 2,4-D and elucidated
the intermediates formed using GC-MS, which is important for
studies investigating the removal of persistent organic waste
matter from water supplies. Finally, the mechanism for this en-
hancement in efficiency was determined and discussed based
upon Mott–Schottky plots, photocurrent measurement and
photoluminescence spectroscopy, which revealed that the
ChemCatChem 2015, 7, 1659 – 1667
www.chemcatchem.org
1666
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[11] T. Arai, M. Horiguchi, M. Yanagida, T. Gunji, H. Sugihara, K. Sayama, J.
Phys. Chem. C 2009, 113, 6602.
[12] L. Pan, X. Liu, Z. Sun, C. Q. Sun, J. Mater. Chem. A 2013, 1, 8299.
[13] K. Gelderman, L. Lee, S. Donne, J. Chem. Educ. 2007, 84, 685.
[14] X. Chen, Y.-B. Lou, A. C. S. Samia, C. Burda, J. L. Gole, Adv. Funct. Mater.
2005, 15, 41.
most likely path of charge transfer in the CuO/P25 junctions is
attributed to electron transfer from TiO2 to CuO, followed by
hole transfer to the TiO2 surface. Further theoretical work is
currently underway to investigate the effect of CuO loading on
the surface characteristics of TiO2.
[15] W. Wang, Q. Zhou, X. Fei, Y. He, P. Zhang, G. Zhang, L. Peng, W. Xie,
CrystEngComm 2010, 12, 2232.
[16] H. S. Park, C. Lee, E. Reisner, Phys. Chem. Chem. Phys. 2014, 16, 22462.
[17] S. Poulston, P. M. Parlett, P. Stone, M. Bowker, Surf. Interface Anal. 1996,
24, 811.
[18] B. Erdem, R. Hunsicker, G. Simmons, Langmuir 2001, 17, 2664.
[19] G. Qiu, S. Dharmarathna, Y. Zhang, N. Opembe, H. Huang, S. L. Suib, J.
Phys. Chem. C 2012, 116, 468.
Acknowledgements
S.J.A.M and J.T. acknowledge funding from the European Com-
munity Seventh Framework Programme (4G-PHOTOCAT 309636).
[20] H. Chen, Y. Xu, RSC Adv. 2015, 5, 8108.
Keywords: copper
semiconductor · titania
·
oxidation
·
photocatalyst
·
[21] Y. Sun, J. J. Pignatello, Environ. Sci. Technol. 1995, 29, 2065.
[22] K. Djebbar, A. Zertal, T. Sehili, Environ. Technol. 2006, 27, 1191.
[23] Y. Yu, J. Ma, Y. Hou, J. Environ. Sci. 2006, 18, 1043.
[24] Q. Xiang, J. Yu, M. Jaroniec, J. Phys. Chem. C 2011, 115, 7355.
[25] J. W. Ondersma, T. W. Hamann, Energy Environ. Sci. 2012, 5, 9476.
[26] R. Beranek, Adv. Phys. Chem. 2011, 786759.
[27] W. Chun, A. Ishikawa, H. Fujisawa, T. Takata, J. Kondo, M. Hara, M. Kawai,
Y. Matsumoto, K. Domen, J. Phys. Chem. B 2003, 107, 1798.
[28] S. J. A. Moniz, J. Zhu, J. Tang, Adv. Energy Mater. 2014, 4, 201301590.
[29] Q. Huang, F. Kang, H. Liu, Q. Li, X. Xiao, J. Mater. Chem. A 2013, 1, 2418.
[30] F. P. Koffyberg, F. A. Benko, J. Appl. Phys. 1982, 53, 1173.
[31] S. Ho-Kimura, S. J. A. Moniz, A. D. Handoko, J. Tang, J. Mater. Chem. A
2014, 2, 3948.
[1] S. J. A. Moniz, R. Quesada-Cabrera, C. S. Blackman, J. Tang, P. Southern,
P. M. Weaver, C. J. Carmalt, J. Mater. Chem. A 2014, 2, 2922.
[2] H. Wang, L. Zhang, Z. Chen, J. Hu, S. Li, Z. Wang, J. Liu, X. Wang, Chem.
Soc. Rev. 2014, 43, 5234.
[3] S. J. A. Moniz, S. A. Shevlin, D. Martin, Z. Guo, J. Tang, Energy Environ.
Sci. 2015, 8, 731–759.
[4] J. Tang, Z. Zou, J. Ye, Angew. Chem. Int. Ed. 2004, 43, 4463; Angew.
Chem. 2004, 116, 4563.
[5] M. N. Chong, B. Jin, C. W. K. Chow, C. Saint, Water Res. 2010, 44, 2997.
[6] D. O. Scanlon, C. W. Dunnill, J. Buckeridge, S. A. Shevlin, A. J. Logsdail,
S. M. Woodley, C. R. A. Catlow, M. J. Powell, R. G. Palgrave, I. P. Parkin,
G. W. Watson, T. W. Keal, P. Sherwood, A. Walsh, A. A. Sokol, Nat. Mater.
2013, 12, 798.
[7] S. Neubert, P. Pulisova, C. Wiktor, P. Weide, B. Mei, D. A. Guschin, R. A.
Fischer, M. Muhler, R. Beranek, Catal. Today 2014, 230, 97.
[8] S. J. A. Moniz, S. A. Shevlin, X. An, Z.-X. Guo, J. Tang, Chem. Eur. J. 2014,
20, 15571.
[32] H. Yu, H. Irie, Y. Shimodaira, Y. Hosogi, Y. Kuroda, M. Miyauchi, K. Hashi-
moto, J. Phys. Chem. C 2010, 114, 16481.
[33] G. Moon, J. Joo, I. Lee, Y. Yin, Nanoscale 2014, 6, 12002.
[34] X. An, H. Liu, J. Qu, S. J. A. Moniz, J. Tang, New J. Chem. 2015, 39, 314.
[35] G. Li, N. Dimitrijevic, L. Chen, J. Phys. Chem. C 2008, 112, 19040.
[36] M. Miyauchi, A. Nakajima, T. Watanabe, K. Hashimoto, Chem. Mater.
2002, 14, 2812.
[9] X. Qiu, M. Miyauchi, K. Sunada, M. Minoshima, Y. L. Min Liu, D. Li, Y. Shi-
modaira, Y. Hosogi, Y. Kuroda, K. Hashimoto, ACS Nano 2012, 6, 1609.
[10] H. Irie, K. Kamiya, T. Shibanuma, S. Miura, D. A. Tryk, T. Yokoyama, K. Ha-
shimoto, J. Phys. Chem. C 2009, 113, 10761.
Received: March 26, 2015
Published online on May 12, 2015
ChemCatChem 2015, 7, 1659 – 1667
www.chemcatchem.org
1667
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim