Noble Metal Free, Visible Light Driven Photocatalysis Using TiO2 Nanotube Arrays Sensitized by P‐Doped C3N4 Quantum Dots

1901275 (1 of 15) © 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
www.advopticalmat.de
FULL PAPER
Noble Metal Free, Visible Light Driven Photocatalysis
Using TiO2 Nanotube Arrays Sensitized by P-Doped
C3N4 Quantum Dots
Pawan Kumar,* Piyush Kar, Ajay P. Manuel, Sheng Zeng, Ujwal K. Thakur, Kazi M. Alam,
Yun Zhang, Ryan Kisslinger, Kai Cui, Guy M. Bernard, Vladimir K. Michaelis,
and Karthik Shankar*
DOI: 10.1002/adom.201901275
1. Introduction
Sunlight assisted photoelectrochemical
(PEC) water splitting is a powerful process
for storing light energy in the chemical
bonds of hydrogen, which is a portable fuel
that can be ignited to release clean energy.[1]
TiO2 is an important semiconductor used
as the photoanode for photoelectrochemical
water splitting because of its chemical sta-
bility and the high oxidative power of its
photogenerated holes. However, both the
anatase and rutile forms of TiO2 are wide
bandgap (WBG) semiconductors, due to
which TiO2-based photocatalysts harvest
<5% of the energy in the sun's spectrum,
since the bulk of the solar energy reaching
the earth's surface resides at visible and
near-infrared wavelengths. The limited light
harvesting abilities of stable, photocorro-
sion-resistant and earth-abundant metal
oxide semiconductors (e.g., TiO2, SrTiO3,
NaNbO3, etc.) has prompted a major
research effort into synthesizing new visible
light-responsive semiconductors. One family
of semiconductors that shows great promise
for photocatalysis is that of graphenic semi-
conductors, which are conjugated polymers
based on the substitution of one or more carbon atoms in the
sp2
-hybridized graphene framework by nitrogen and other heter-
oatoms, which results in the opening of a bandgap.[2]
Graphenic
semiconductors are made of earth-abundant elements (C, N, O,
etc.), have remarkable thermal and photochemical stability, ade-
quate charge transport, and band edges and bandgaps that are tun-
able through molecular structure.[3]
The most celebrated member
of this family is bulk graphitic carbon nitride (g-C3N4; Eg = 2.7 eV),
whose energy levels are well-matched with the requirements of
water-splitting, CO2 reduction and the degradation of numerous
organic compounds.[4] In recent years carbon nitride quantum
dots (CNQDs) are gaining popularity over conventional carbon/
graphene quantum dots in various applications including sensing,
bioimaging, chemical sensing, photocatalysis, etc. due to their
excellent photochemical and thermal stability, nontoxic nature,
and bright photoluminescence.[5]
Bulk g-C3N4 is an earth-abundant, easily synthesizable, and exception-
ally stable photocatalyst with an electronic bandgap of 2.7 eV. Herein, the
concepts of P-doping and size quantization are combined to synthesize
highly fluorescent P-doped carbon nitride quantum dots (CNPQDs) with a
bandgap of 2.1 eV. CNPQDs are hosted on anatase-phase and rutile-phase
TiO2 nanotube array scaffolds, and examined as photoanodes for sunlight-
driven water-splitting and as photocatalysts for surface catalytic reactions.
Square-shaped rutile phase TiO2 nanotube arrays (STNAs) decorated with
CNPQDs (CNPQD-STNA) generate 2.54 mA cm−2 photocurrent under
AM1.5 G simulated sunlight. A champion hydrogen evolution rate of
22 µmol h−1 corresponds to a Faradaic efficiency of 93.2%. In conjunction
with Ag nanoparticles (NPs), the CNPQD-STNA hybrid is also found to be
an excellent plexcitonic photocatalyst for the visible light-driven transfor-
mation of 4-nitrobenzenethiol (4-NBT) to dimercaptoazobenzene (DMAB),
producing reaction completion at a laser power of 1 mW (532 nm) while
Ag NP/TNA and Ag NP/STNA photocatalysts cannot complete this trans-
formation even at 10 mW laser power. The results point the way forward
for photochemically robust, noble metal free, visible light harvesting
photoacatalysts based on nanostructured heterojunctions of graphenic
frameworks with TiO2.
Dr. P. Kumar, Dr. P. Kar, A. P. Manuel, S. Zeng, U. K. Thakur,
Dr. K. M. Alam, Y. Zhang, R. Kisslinger, Prof. K. Shankar
Department of Electrical and Computer Engineering
University of Alberta
9211-116 St, Edmonton, Alberta T6G 1H9, Canada
E-mail: pawan@ualberta.ca; kshankar@ualberta.ca
Dr. K. Cui
Nanotechnology Research Centre
National Research Council of Canada
Edmonton, Alberta T6G 2M9, Canada
Dr. G. M. Bernard, Prof. V. K. Michaelis
Department of Chemistry
University of Alberta
Edmonton, Alberta T6G 2G2, Canada
The ORCID identification number(s) for the author(s) of this article
can be found under https://doi.org/10.1002/adom.201901275.
Adv. Optical Mater. 2019, 1901275

www.advancedsciencenews.com
© 2019 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1901275 (2 of 15)
Two major problems with graphitic carbon nitride semi-
conductor electrodes whether constituted in the bulk or in the
form of quantum dots or sheets, are limited light absorption
and fast carrier recombination resulting in efficiency losses. A
third problem is that some of the reactants in the aforemen-
tioned photocatalytic reactions adsorb poorly on the surfaces of
graphenic semiconductors. The doping of carbon nitride mate-
rials with heteroatoms such as P, F, S, B, etc. has been reported
to increase visible light absorption performance and photo-
stability.[6] Phosphorus doping is particularly appealing where
the P atoms substitute carbon in the tris-s-triazine (heptazine)
ring system and participate in conjugation by donating an extra
electron in the ring network leading to an electron rich, low
bandgap carbon nitride framework.[7] Consequently, carbon
nitride based semiconductors have been found to perform
best as photocatalysts, when they are part of heterojunctions
with other semiconductors such as TiO2, ZnO, SnO2, BiOBr,
Fe2O3, SrTiO3, etc.[8] In these heterojunctions, separation of
photogenerated charge carriers is superior to stand-alone elec-
trodes made of graphenic semiconductors, either due to facile
photoinduced electron transfer or due to the formation of a
Z-scheme. In addition, the strong adhesion of the graphenic
semiconductor to a metal oxide semiconducting scaffold can
also ameliorate the problem of leaching.
In this report, we present the synthesis of a graphenic semi-
conductor, namely phosphorus doped carbon nitride quantum
dots (CNPQDs) via a facile thermal condensation polymeriza-
tion reaction. We formed heterojunctions of CNPQDs with cir-
cular anatase and square-shaped rutile TiO2 nanotube arrays
(STNA) and investigated the resulting films as photoanodes
for water splitting. Among the many different types of micro-
and nanostructures of titanium dioxide used for photocatalytic
applications, TiO2 nanotube and nanowire arrays are uniquely
suited as photoanodes for photoelectrochemical water-splitting
due to their self-organized, highly ordered, vertically oriented
geometry which enables orthogonalization of light harvesting
and charge separation,[9] slow light and light trapping effects,[10]
and directed pathways for the transport of electrons.[11] As
mentioned in the previous paragraph, graphitic carbon nitride
(g-C3N4) and related materials are a relatively new class of
compounds, that have proven to be very effective in the
photocatalytic field.[8d] From the photocatalytic standpoint, gra-
phitic carbon nitride semiconductors are metal free and have
interesting properties, including unique electrical and optical
properties. Therefore, they can potentially eliminate the use of
precious as well as toxic metals from photocatalytic devices.
2. Results and Discussion
2.1. Synthesis, Morphological, and Structural Characterization
The phosphorus doped CNPQDs were synthesized by solid
state thermal condensation reaction of urea, citric acid, and
1-butyl-3-methylimidazolium hexafluorophosphate (BMIM-
PF6) at 200 °C in an autoclave by a slight modification of pre-
viously reported procedures (Figure 1a).[12] In this synthetic
process, urea and citric acid react together via condensation
polymerization to generate the basic carbon nitride framework
while BMIM-PF6 serves as the source of phosphorus doping.
Figure 1. Schematic diagram illustrating the synthesis of a) phosphorus doped carbon nitride quantum dots (CNPQDs). b) Steady state photolumi-
nescence (PL) spectrum of phosphorus doped carbon nitride quantum dots (CNPQDs) and carbon nitride quantum dots (CNQDs) in water (330 nm
excitation wavelength). Inset showing fluorescence of CNPQDs sample under UV light irradiation. c) 15N and 31P CPMAS NMR spectra of CNPQDs.
d) Flame annealed square shape TiO2 nanotube arrays (STNAs) from aqueous TiO2 nanotube arrays (TNAs). e) Hydrothermal treatment for decoration
of CNPQDs on STNA (CNPQD-STNA) and TNA (CNPQD-TNA).

This protocol is different from conventional synthesis involving
hydrothermal reaction of g-C3N4 QDs and phosphonitrilic
chloride trimer.[13] The afforded CNPQDs show bright tur-
quoise fluorescence under UV light irradiation. When excited
by 330 nm photons, CNPQDs in water gave an intense broad
photoluminescence (PL) peak centered at 462 nm as seen in
Figure 1b, demonstrating emission at a longer wavelength than
the previously reported PL for carbon nitride based quantum
dots.[14] The observed redshifted PL suggests phosphorus
doping increases the electronic charge density on the carbon
nitride framework leading to a reduction in the bandgap due
to the contribution of the extra electron from P to the heptazine
ring system.[15] The UV–vis absorption spectrum of CNPQDs
in water shows an intense peak at 300 nm corresponding to a
π→π* transition and another weak shoulder peak at 487 nm
was assigned to n→π* transition.[16] The presence of n→π*
transition at 487 nm demonstrates enhanced visible light
absorption in comparison to previously reported carbon and
carbon nitride quantum dots (Figure 4a).[12b,17] The synthesized
CNPQDS were instantaneously dispersible in water which was
attributed to their small size, electron rich surface (promoted
by phosphorus doping), and the presence of ample amounts
of amino and carboxylic moieties at the edge of the sheet-like
quantum dots.[6a] To probe this assumption, the average particle
size distribution and surface charge on CNPQDs were deter-
mined using dynamic light scattering (DLS) and zeta potential
measurements, respectively. The average diameter of CNPQDs
in water was found to be 28.12 nm (Figure S1, Supporting
Information) while the surface potential was measured to be
−20.1 mV (Figure S2, Supporting Information) which dem-
onstrates extremely small size quantum dots and an electron
rich surface decorated with COOH/NH2 groups.[12c,18] The
successful synthesis of carbon nitride quantum dots doped
with phosphorous atoms was confirmed by solid-state nuclear
magnetic resonance (NMR) spectroscopy acquired using the
cross-polarization magic-angle spinning (CPMAS) technique
(Figure 1c). The 15N CPMAS NMR spectrum of CNPQDs
reveals two NMR signals centered at −250 and −291 ppm. The
15N NMR resonance centered between −240 and −270 ppm is
attributed to secondary (NC2) and tertiary (NC3) nitrogen atoms
of heptazine units constituting the carbon nitride skeleton while
the peak located between −290 and −310 ppm is ascribed to
primary nitrogens of terminal NH2 groups.[19]
These observed
NMR resonances demonstrate the presence of a carbon nitride
framework decorated with terminal NH2 groups. A 31
P CPMAS
NMR spectrum shows an intense peak at 6 ppm confirming the
presence of doped phosphorous atoms in a carbon nitride scaf-
fold as reported earlier.[7a,20]
Additionally, a weak heptet signal
was also observed at −145 ppm (not shown) attributed to a PF6
phosphorous due to the presence of residual BMIM-PF6.
Anatase phase TiO2 nanotube arrays (TNAs) were prepared by
electrochemical anodization of Ti foil in water-based electrolyte
containing 4% v/v of hydrofluoric acid (HF) and 8% v/v of acetic
acid followed by annealing at 450 °C for 3 h.[21] The obtained
anatase phase TiO2 nanotubes were subjected to high tempera-
ture annealing in a propane flame (≈750–900 °C) to afford rutile
phase STNAs (Figure 1d).[21]
The change in the cross-section of
the nanotubes from circular to square, represents an interesting
morphological change with important implications for the
interaction of light with titania nanotubes.[22] Anatase phase
TNA and flame annealed square-shaped TiO2 nanotube array
samples were decorated with CNPQDs via hydrothermal treat-
ment at 120 °C for 12 h in an autoclave to obtain CNPQD-TNA
and CNPQD-STNA nanostructures (Figure 1e).
Field emission scanning electron microscopy (FESEM) was
used to determine the rough morphological attributes of TNAs
and STNAs (Figure S3, Supporting Information). A typical
FESEM image of TNA shows cylindrical nanotubes with an
average diameter of 77 nm while cross-sectional view shows the
nanotube to have an average length of 470 nm (Figure S3a,b,
Supporting Information). The top-view of TiO2 nanotubes
after high temperature flame annealing (STNA) clearly shows
the morphological transformation of cylindrical nanotubes
into square-shaped nanotubes with 60–70 nm edge length
(Figure S3c, Supporting Information) while cross-sectional view
shows the average length of STNAs was ≈400 nm (Figure S3d,
Supporting Information).
The ultrafine morphological attributes of phosphorus
doped carbon nitride quantum dots decorated STNA samples
(CNPQD-STNA) were investigated with high resolution trans-
mission microscopy (HRTEM) (Figure 2). The HRTEM images
of CNPQD-STNA at 10 nm scale bar show the presence of dark
spots and amorphous region of CNPQDs on STNA (Figure 2a).
The high magnification TEM images at 5 nm scale bar show
lattice fringes with 0.33 and 0.27 nm interplanar d spacing
corresponding to (110) and (101) planes of rutile phase TiO2
(Figure 2b,c).[23] Another d spacing of 0.32 nm in HRTEM
images was assigned to the (002) plane of carbonaceous mate-
rials demonstrating long range π–π stacking of CNPQDs on the
surface of STNA (Figure 2c).[24] Further, HRTEM images show
an unusually high d spacing of 0.63 nm, which were Moire
fringes that originate in layered materials due to mismatched
superimposition (dislocation or rotational) of two lattices with
almost identical d spacing (Figure 2b).[25] The Moire fringe
with almost double d spacing (0.63 nm) confirms presence of
stacked CNPQDs on STNA.
Elemental mapping of CNPQD-STNA from top shows an
even distribution of C, N, O, P, and Ti elements centered on
the wall of nanotubes thus confirming the decoration of TiO2
nanotubes with CNPQDs (Figure 2d). The presence of various
peaks corresponding to C, N, O, P, Ti in the energy-dispersive
X-ray (EDX) spectra of CNPQDs, further corroborate the deco-
ration of STNA with CNPQDs (Figure 2e). To discern the pres-
ence of CNPQDs on STNA, electron energy loss (EEL) spectra
of CNPQD-STNA samples were collected (Figure 2f). The EELS
spectra display peaks corresponding to P L-edge, C K-edge, N
K-edge, and Ti L-edge at 138.16, 287.14, 401.31, and 464.95 eV,
respectively and reveal a chemically stable P-doped carbon
nitride structure on the STNA.[26]
The EELS mapping of sample
shows well distributed concentration of C, N, and P over the
nanostructures (Figure 2g). Further, EELS line scan mapping
over nanotubes structures show intense signal of C, N, and
P on the wall of TiO2 nanotubes overlapped with Ti signals
verifying the decoration of STNA with CNPQDs (Figure 2h).
The phase structure and crystalline nature of TNA and
STNA samples with and without CNPQDs were investigated
with X-ray diffraction (XRD) (Figure S6a, Supporting Informa-
tion). The X-ray diffractogram of TNA sample was measured by

maintaining a 2° angle between substrate and incident X-ray,
and shows a diffraction peak at a 2θ value of 25.2°, which was
assigned to the (101) plane of tetragonal anatase phase TiO2 and
matched well with JCPDS card #21-1272. Other peaks of TiO2
were not distinguishable, likely due to the masking of signal
from intense Ti substrate peaks. After decoration of CNPQDs on
TNA, the peak intensity of (101) anatase peak decreased slightly
suggesting coverage of TiO2 surface with CNPQDs, which
shields some signals. For STNA, the XRD peaks were observed
at 27.3°, 36.0°, 41.2°, and 54.3° corresponding to (110) (101) and
(111) and (211) planes of tetragonal rutile phase TiO2 with a
slightly orthorhombic distortion of the TiO6 octahedron (JCPDS
card #21-1276).[27]
After hydrothermal treatment of STNA with
CNPQDs under aqueous condition the peak intensities were
Figure 2. HRTEM images of CNPQD-STNA a) at 10 nm scale bar showing amorphous layers of CNPQDs deposited on STNA, b) at 5 nm scale bar
showing lattice fringes and Moire fringes, insets shows rutile phase TiO2 (110) plane d spacing and TEM FFT diffraction pattern, c) at 5 nm scale bar
showing lattice fringes and insets shows rutile (110) and (101) planes d spacing and TEM FFT diffraction pattern. d) Bright field electron image, and
STEM elemental maps for C, N, O, P and Ti. e) STEM-EDX spectra of CNPQD-STNA. f) Inner shell ionization edge (core loss) electron energy loss
spectra (EELS) of CNPQD-STNA showing P-L edge, C-K edge, N-K edge, and Ti-L edge acquired on STEM. g) EELS elemental mapping of CNPQD-
STNA showing map area, C-K edge, P-L edge and N-K edge. h) EELS line scan mapping of CNPQD-STNA for P (pink), C (orange) and Ti (navy blue)
elements showing coinciding signal of P and N with Ti confirming decoration of STNA with CNPQDs.

decreased while a new peak was observed at ≈26° due to (002)
reflection of stacked carbon nitride sheets which demonstrated
the successful decoration of STNA with CNPQDs.[23]
To probe the presence of CNPQDs and measure the mode
specific IR vibrational features, Raman spectra of the syn-
thesized materials were collected (Figure S6b, Supporting
Information). Raman spectra of TNA displayed the character-
istic fingerprint bands of tetragonal anatase phase TiO2 at 142.0,
396.1, 511.9, and 630.3 cm−1 indexed to Eg, B1g, B1g+A1g, and Eg
active modes.[28]
After hydrothermal grafting of CNPQDs on
the TNA in CNPQD-TNA, the intensity of these bands grew
weaker and a few new bands at 450.1, 550.3, and 691.6 cm−1
appeared which were assigned to heptazine ring (tris-s-triazine,
C6N7) breathing modes.[29]
The presence of two signature
peaks at 1358 and 1585 cm−1
attributed to the D and G bands
of graphitic carbon based materials confirmed the presence
of CNPQDs on the surface of TNA.[30]
The D band signifying
defects in the aromatic conjugated network originated due to
out of plane vibration of sp3
carbons while the G band pecu-
liar to graphitic network originated due to in-plane vibrations
of sp2 carbons in the aromatic system.[31] The integral intensity
ratio of D and G bands (ID/IG) was calculated to be 1.92 dem-
onstrating plenty of defects present in the form of COOH/
CNH2 terminated functionalities on the heptazine nucleus.[32]
Raman spectra of STNA exhibited IR active vibration bands at
138.2, 444.6, and 635 cm−1
assigned to B1g, Eg, and A1g modes
of tetragonal rutile phase TiO2 respectively while another less
intense broad peak at 241.2 cm−1 was assigned to multiphonon
processes.[33] The observation of weak anatase peaks in the
Raman spectrum of STNA might be due to incomplete phase
transformation during the flame annealing step. Analogous to
CNPQD-TNA, the Raman spectra of CNPQD-STNA displayed
a heptazine nucleus-specific peak at lower frequency and the
characteristic D and G bands having an integral intensity
ratio (ID/IG) of 1.84, which verified the presence of defect rich
P-doped carbon nitride quantum dots over the surface of STNA.
The surface and sub-surface of chemical composition, and ele-
mental oxidation states in CNPQD-TNA and CNPQD-STNA were
probed using X-ray photoelectron spectroscopy (XPS) (Figure 3).
The elemental survey scan of CNPQD-TNA and CNPQD-STNA
exhibits C1s, N1s, O1s P2p, and Ti2p signal along with other inner-
core excitation peaks confirms presence of all the constituting ele-
ments of CNPQDs and TiO2 in material (Figure S8a, Supporting
Information). Core-level high resolution XPS (HR-XPS) spectra
of CNPQD-STNA in C1s region were deconvoluted into four
peak components centered at 284.8, 286.6, 287.8, and 288.9 eV
(Figure 3a). The peaks at 284.8 and 286.6 eV were assigned to
sp3 hybridized turbostratic carbons (CC) and CO carbons of
carboxylic group, respectively.[34] Two other peak components
located at BE ≈ 287.8 and 288.9 eV originated from sp2
carbons
in NCN2 bonds and sp3
carbon in CNC bonds which are
basic constituents of the graphitic carbon nitride framework.
Figure 3. Core-level high resolution XPS spectra of CNPQD-STNA in a) C1s, b) N1s, c) P2p, and d) Ti2p regions.

The HR-XPS of CNPQD-TNA displayed four peak compo-
nents at 284.8, 286.5, 287.9, and 289.0 eV corroborated to CC,
CO, NCN2, and CNC type carbons, respectively (Figure
S8b, Supporting Information). The HR-XPS of CNPQD-STNA
and CNPQD-TNA in N1s region after deconvolution gave three
peak components situated at BE ≈ 399.2, 400.0, and 400.8 eV
which arose from secondary CNC, tertiary N(C)3 and
primary HN(C)2 type nitrogens, respectively (Figure 3b
and Figure S8c, Supporting Information).[35] Further, decon-
volution of P2p signals of CNPQD-STNA and CNPQD-TNA
show two peak components at 132.4 and 133.4 eV due to
P2p3/2 and P2p1/2 peak components of doped phosphorus. The
presence of a single P2p peak with two orbital splitting compo-
nents demonstrates that P atoms share an identical chemical
environment in heptazine ring system and rules out the pos-
sibilities of PO type phosphorous (Figure 3c and Figure S8d,
Supporting Information).[36] The inner-core HR-XPS spectra of
both CNPQD-STNA and CNPQD-TNA in Ti2p region consisted
of two major peak components located at binding energies of
458.6 and 464.4 eV, which were assigned to Ti2p3/2 and Ti2p1/2
peak components of Ti4+ state in TiO2 nanocrystals (Figure 3d
and Figure S8e, Supporting Information).[37] The 5.8 eV peak
splitting between Ti2p3/2 and Ti2p1/2 peaks and their respec-
tive positions was in well agreement with Ti4+ coordinated to
O2− (Ti4+-O2−) in the TiO6 octahedron of tetragonal TiO2 crys-
tals.[38] Additionally, Ti2p XPS spectra of CNPQD-STNA and
CNPQD-TNA displayed two weak shoulder peaks at relatively
lower BE ≈ 457.7 and 463.2 eV originated from Ti2p3/2 and
Ti2p1/2 peak components of Ti3+ present in the form of lattice
defects.[39] The flame annealing of TiO2 samples tends to create
Ti3+ defects and oxygen vacancies due to abstraction of lattice
oxygens at high temperature which generates sub gap energy
levels just below the conduction band (shallow traps) resulting
in a slight narrowing of the bandgap.[40] In our previous reports
we have demonstrated that STNA possesses a higher density
of defects in comparison to TNA which promotes a better vis-
ible light response and enhanced photocatalytic performance
in the photoreduction to CO2 to methane.[41] The O1s peak of
CNPQD-STNA and CNPQD-TNA split into two peak compo-
nents positioned at 529.5 and 531.6 eV (Figures S7 and S8f,
Supporting Information). The peak at 529.5 eV was attributed
to oxygen atoms in the crystal lattice of TiO2 (Ti–O–Ti) while
the peak at 531.6 eV originated from the contribution of surface
adsorbed non-lattice OH in TiO2 and CO and OH of car-
boxylic groups on the surface of CNPQDs (Figures S7 and S8f,
Supporting Information). XPS results clearly demonstrate the
presence of intact phosphorus doped carbon nitride quantum
dots on the surface of tetragonal TiO2.
2.2. Optical Properties and Electronic Band Structure
The optical absorption performance of materials was determined
using UV–vis absorption spectra collected in diffuse reflectance
mode (Figure 4). The CNPQDs show π→π* and n→π* transi-
tion peaks centered at 300 and 487 nm, respectively revealing
visible light absorbance (Figure 4a).[16a]
It is noteworthy that the
optical absorption of CNPQDs reported in this work with a band
edge at ≈600 nm (Figure 4a) is significantly different from the
absorption behavior of N, P-codoped carbon dots which exhib-
ited a band edge at ≈450 nm.[42] The emission peak of CNPQDs
occurs at 462 nm (λexc = 330 nm) as shown in Figure 1b while
N, P-codoped carbon dots exhibit excitation-dependent emission
behavior with a PL emission maximum at 410 for 330 nm exci-
tation.[42] The DR UV–vis spectrum of TNA displayed the semi-
conductor band edge of anatase at ≈385 nm (Figure 4b). The
weak yet broad absorption feature centered at ≈450 nm arose
due to the presence of mid-gap states due to oxygen vacancies
creating deep level traps facilitating longer wavelength absorp-
tion.[43] The flame annealing of TNA at elevated temperature
transformed anatase phase TiO2 into rutile phase TiO2 in STNA
and UV–vis absorption edge was red-shifted by ≈25 to 410 nm
corresponding to the 3.0 eV bandgap of rutile. After decoration
of CNPQDs on TNA and STNA, the visible light absorption per-
formances of CNPQD-TNA and CNPQD-STNA were enhanced
with each type of sample showing maximum absorption in
between 460 to 550 nm.
The electronic structure and band edge energies of materials
with respect to vacuum level were calculated from ultraviolet
photoelectron spectroscopy (UPS) work function (WF) spectra
and valence band spectra acquired using UPS (Figure 4). The
value of work function was calculated by subtracting the energy
of emitted secondary electrons (Ecut-off) from the incident energy
of UV light source (He laser: 21.21 eV) using the following
expression WF (φ)  =  21.21−Ecut-off. The cut-off energy was deter-
mined from the point of intersection of the extrapolated linear
region of work function on X and Y scale. From the WF spectra,
the Ecut-off energies of TNA, STNA, CNPQD-TNA, and CNPQD-
STNA were estimated to be 16.91, 16.71, 16.89, and 16.63 eV,
respectively which correspond to WF values of 4.30, 4.50, 4.32,
and 4.58 eV, respectively (Figure 4c). From the work function
value, it was obvious that Fermi level of STNA was 0.20 eV below
the Fermi level of TNA which demonstrate transformation of
anatase phase TiO2 in TNA to rutile phase TiO2 in STNA during
high temperature annealing. The obtained Fermi level values
and 0.20 eV energy difference was in close agreement with lit-
erature values.[44] After decoration of CNPQDs on TNA and
STNA, the position of Fermi level was not changed significantly
which suggests that the formation of an electronic heterojunc-
tion between TiO2 and CNPQDs did not result in band bending.
Further, linear extrapolation of the leading edges in UPS valence
band spectra gave the position of valance band maxima (VBM)
(Figure 4d). The VBM positions for TNA and STNA were cal-
culated to be 3.21 and 3.11 eV, respectively. The CNPQD-TNA,
CNPQD-STNA displayed two valence band edges corresponding
to CNPQDs and TiO2 suggesting that in the CNPQD-TiO2 nano-
composite, each material retains its individual electronic identity.
The VBM edge values for CNFQDs-TNA were found to be 2.70
and 1.32 eV while CNPQD-STNA displayed VB position at 2.97
and 2.42 eV assigned to TiO2 and CNPQDs respectively.
2.3. Performance of CNPQD Decorated Titania Nanotubes in
Photoelectrochemical Water Splitting
Photoelectrochemical water splitting experiments were per-
formed in a three-electrode configuration wherein the bare/het-
erojunction TiO2 nanotube array scaffold on Ti foil was assigned

as the photoanode (working electrode) while Pt was used as
photocathode (counter electrode) and the voltage bias was applied
with respect to an Ag/AgCl reference electrode. All three elec-
trodes were immersed in 1 m KOH electrolyte and the photo-
anode was irradiated with AM1.5 G one sun simulated sunlight.
The photocurrent response, expressed in the terms of the photo-
current density, was measured in linear sweep voltammetric
(LSV) mode by sweeping the applied bias from −0.8 to +1.0 V
versus Ag/AgCl. In the dark, almost zero photocurrent den-
sity was observed for each of the tested samples (Figure 5a–e).
Under one sun illumination, all samples showed appreciable
photocurrents at an applied bias of +0.6 V versus Ag/AgCl
(1.23 V vs RHE equivalent to water oxidation potential). Pristine
TNA and STNA samples achieved current densities of 0.04 and
0.30 mA cm−2
, respectively under simulated sunlight passed
through a 420 nm UV cut-off filter while the current densities
without the UV cut-off filter were found to be 0.17 and 0.75 mA
cm−2, respectively, at +0.6 V versus Ag/AgCl. The photocurrent
response of STNA was 7.5 and 4.4 times higher than TNA under
one sun irradiation with UV filter (>420 nm) and without UV
filter respectively. These results clearly demonstrate the superior
photoelectrochemical performance of STNA due to the pres-
ence of the lower bandgap (3.0 eV) rutile phase in comparison
to anatase phase TiO2 (3.2 eV) in TNA. Further, Ti3+ defects and
oxygen vacancies were responsible for the visible light response
of materials. The enhancement of water splitting performance
was also due to shortened carrier collection pathways, because of
the short (in length) and thin walled STNA formed in aqueous
electrolyte. In order to make the most of the finite carrier
mobility and charge carrier lifetime, short aqueous nanotubes
with thin walls are well suited. Recombination is minimized
before charge carriers reach the interface with the electrolyte,
and before the electrons reach the Ti foil on their way to the
external circuit. As we shall show in the next section (finite-dif-
ference time-domain (FDTD) simulations), the square-shaped
pore morphology also results in more effective absorption of
photons in both bare and CNPQD-decorated TiO2 nanotubes.[45]
The photocurrents generated by TNA and STNA samples
were tremendously increased after decoration of CNPQDs
(Figure 5a,b). The measured photocurrent densities for CNPQD-
TNA and CNPQD-STNA under solar simulated light with UV
filter (>420 nm) were calculated to be 0.24 and 0.45 mA cm−2
Figure 4. a) UV–vis spectra of CNPQDs in water showing π→π* and n→π* transition and UV–vis spectra of bulk g-C3N4. b) Diffuse reflectance UV–vis
spectra of TNA, CNPQD-TNA, STNA, and CNPQD-STNA. c) UPS work function spectra of TNA, CNPQD-TNA, STNA, CNPQD-STNA and inset shows
cut-off energy (Ecut-off) of secondary electrons. The work function (WF) was determined from the UPS work function spectra using the equation WF
(φ)  =  21.21 − Ecut-off, where 21.21 eV is energy of the incident, He discharge lamp used for UPS. d) UPS valence band spectra of TNA, CNPQD-TNA,
STNA, CNPQD-STNA showing the estimated values of the respective valence band maxima below the Fermi level. Color: TNA (blue), CNPQD-TNA
(green), STNA (red) and CNPQD-STNA (wine red).

while without UV filter, the photocurrent densities were found to
be 1.26 and 2.54 mA cm−2
, respectively. CNPQD-STNA show the
highest photocurrent density which was 1.9 times higher than
CNPQD-TNA under AM1.5 G irradiation (>420 nm) and 2.0
times without UV filter. The photocurrent response of materials
during light On–Off cycles exhibited evident peaks and troughs
and confirms the photoresponse of the materials to be real and
not an artifact of the scan rate (Figure 5c) The photoelectrochem-
ical performance of the samples under near-monochromatic
illumination at discrete longer wavelengths was tested using
450 and 500 nm light emitting diodes (LEDs) with power densi-
ties of 54.15 and 40.58 mW cm−2
at the surface of the samples
(Figure 5d). The relative photocurrent density of STNA based
samples was highest among all samples. The photocurrent den-
sities of STNA and CNPQD-STNA samples under 450 nm irra-
diation were found to be 0.20 and 0.59 mA cm−2 while under
500 nm irradiation the value of photocurrent densities were
measured to be 0.08 and 0.65 mA cm−2
, respectively (Figure 5d).
Under identical conditions TNA and CNPQD-TNA samples
displayed 0.04 and 0.24 mA cm−2
current density at 450 nm
Figure 5. Photocurrent density versus applied potential plot for a) CNPQD-STNA and CNPQD-TNA. b) STNA and TNA. c) Photocurrent density change
during light on-off cycle for CNPQD-STNA under dark conditions, under solar simulated AM1.5 G irradiation with 420 nm cut-off filter, under AM1.5 G
irradiation without UV cut-off filter. d) Photocurrent density versus applied potential plot for STNA and CNPQD-STNA under dark conditions, under
450 nm LEDs irradiation (power density 54.15 mW cm−2) and 500 nm LEDs irradiation (power density 40.58 mW cm−2). e) TNA, CNPQD-TNA under
dark conditions, under 450 nm LEDs irradiation (power density 54.15 mW cm−2). f) IPCE and APCE% of TNA, CNPQD-TNA, STNA, CNPQD-STNA
samples in 450 and 500 nm wavelength LED. Plot between ABPE% versus potential showing maximum ABPE% of TNA, STNA, CNPQD-TNA, CNPQD-
STNA g) under solar simulated AM1.5 G irradiation (>420 nm), under AM1.5 G irradiation without UV cut-off filter h) under 450 nm wavelength LEDs
irradiation, under 500 nm wavelength LEDs irradiation. All the measurements were carried out in 1.0 m KOH solution at a scan rate of 0.1 mV s−1.
i) Bar diagram showing experimentally observed H2 evolution rate and corresponding Faradaic efficiencies using TNA, CNPQD-TNA, STNA CNPQD-
STNA samples under AM1.5 G irradiation.

(Figure 5e). Under 500 nm wavelength illu-
mination, the current densities of TNA and
CNPQD-TNA were measured to be 0.05 and
0.32 mA cm−2. The calculated current den-
sity of CNPQD-STNA was 8.1 and 13.0 times
higher than bare STNA and TNA respectively
at 500 nm wavelength. Diagnostic efficien-
cies of materials were estimated to investigate
materials system/interface performance (Sec-
tion S2, Supporting Information).[1a,46] This is
a proof of concept study intended to showcase
the potential of CNPQDs decorated on nano-
structured n-type semiconducting scaffolds
as visible light harvesting photonodes for
sunlight-driven water splitting. Therefore, nei-
ther the nanotube length nor the total loading
of CNPQDs was optimized for maximum
performance. The applied bias photoconver-
sion efficiency (ABPE) measures the overall
conversion of light into electricity, and ABPE
values approaching 1% were obtained for
the unoptimized systems presented in this
study under AM1.5 G one sun illumination
(Figure 5g). A particularly relevant metric
consists of the absorbed photon-to-electron
conversion efficiency (APCE) at visible wave-
lengths which measures the internal quantum
yield of conversion of incident photons into
electrons that generate current in the external
circuit. APCEs as high as 3.5% for 450 nm
photons and 4.7% for 500 nm photons, were
achieved using CNPQD-STNA samples
(Figure 5f). The results of hydrogen evolution
experiments using H-cells (Figure S14, Sup-
porting Information) indicated a champion H2
production rate of 22 µmol h−1 corresponding
to a Faradaic efficiency of 93.2% (Figure 5h
and Section S2, Supporting Information).
2.4. FDTD Simulations of TNAs, STNAs,
CNPQD-TNAs, and CNPQD-STNAs
Both cylindrical and square cross-sectioned
titania nanotubes can behave as symmetric
hollow core waveguides for incident photons
in sunlight due to the large refractive index
contrast between the high index TiO2
annulus and the surrounding air medium.[47]
In Figure 6a,
we see that for bare cylindrical anatase-phase TNAs, the elec-
tric field intensity at 500 nm does not decrease exponentially
in the air cladding. The fundamental mode is almost cut-off as
the electric field spreads into the air medium surrounding the
nanotubes. For 350 nm photons, the waveguiding effect of the
cylindrical nanotubes is strong in spite of the bandgap absorp-
tion of TiO2 as evidenced by a strong electric field that is fully
confined within the hollow core of the nanotube (Figure 6b).
On the other hand, 350 nm photons are almost fully attenuated
due to absorption by TiO2 in bare rutile-phase square nanotubes
(STNAs) with a small evanescent field at the outer wall of the
nanotubes (Figure 6d). STNAs also exhibit a strong confine-
ment of the electric field for 500 nm photons within the hollow
core with an exponential decrease in the field intensity outside
the nanotubes (Figure 6c). Following decoration of the STNAs
with CNPQDs, the electric field intensity at the top of the
square nanotubes is dramatically weaker and the optical energy
is confined to a much smaller volume within the nanotubes
due to the rapid attenuation of light because of absorption by
the CNPQDs. In contrast, the circular TNAs when decorated
with CNPQDs still show an appreciable electric field intensity
Figure 6. Results of FDTD simulations of discrete titania nanotubes showing electric field
intensities for a) TNAs at 500 nm. b) TNAs at 350 nm. c) STNAs at 500 nm. d) STNAs at
350 nm. e) CNPQD-STNA at 500 nm. f) CNPQD-TNA at 500 nm.

at the outer edge of the concentric nanotube-CNPQD hybrid.
These results indicate the CNPQD-STNAs to be significantly
more efficient absorbers of visible light (particularly photons
with wavelengths of 350–500 nm) than CNPQD-TNAs and
demonstrate the importance of nanotube shape in determining
the interactions with light.
It is well-known that the optical spectra of sub-micrometer
long titania nanotubes exhibit pronounced optical interference
fringes due to reflections from the air-TNA and substrate-TNA
interfaces.[48] Indeed, the simulated cross-sectional electric field
profiles of both CNPQD-TNA and CNPQD-STNA show undu-
lating intensities due to constructive and destructive interfer-
ence (Figure S10a,b,d,e, Supporting Information). However, the
CNPQD-STNA samples (Figures S10d and S13e, Supporting
Information) are indicated to attenuate the electric field intensity
much more strongly in comparison to CNPQD-TNA samples
(Figure S10a,b, Supporting Information), thus demonstrating
the enhanced light harvesting ability of square-shaped mixed
phase titania nanotube samples decorated with P-doped g-C3N4
quantum dots. The simulated optical spectra indicate absorption
[1−(R+T)] peaks at 395 and 530 nm for CNPQD-TNA and cor-
responding absorption troughs at 365 and 440 nm (Figure S10c,
Supporting Information). In the experimentally measured UV–
vis spectrum of CNPQD-TNA (green curve in Figure 4b), we see
an absorption peak at 390 nm and a broad absorption feature
centered at ≈500 nm along with absorption troughs at 365 and
440 nm. Thus, the FDTD simulations capture an impressive
number of features seen in the measured UV–vis spectrum of
CNPQD-TNA while also confirming that the overall absorptance
of the CNPQD-STNA samples is significantly higher than that
of CNPQD-TNA samples in the 350–550 nm spectral range.
2.5. Plausible Mechanism of Enhanced Photoelectrochemical
Water Splitting Performance
To obtain greater insight into the superior photoelectrochem-
ical performance of CNPQD-STNA, we have proposed a plau-
sible mechanism (Figure 7) based on experimental findings on
electronic band structure and existing literature. The minimum
requirement for efficient water splitting without any sacrifi-
cial donor on a semiconductor surface is a more negative CB
position than the proton reduction potential (0.00 eV vs NHE
at pH-0; H+
/H2) and a more positive valence band position
than the water oxidation potential (+1.23 eV vs NHE at pH-0;
H2O/O2).[49]
This implies that the bandgap value of the semi-
conductor should be more than 1.23 eV with appropriate values
of the electronic band edges. Regular furnace annealed TNA
is composed of anatase TiO2 (Eg = 3.2 eV) which absorbs only
UV light (λ < 386 nm) and cannot drive water splitting reac-
tion in the visible regime. However, the presence of defects and
mid-gap states that originated in the anodization process and
were not completely healed during annealing, enables longer
wavelength absorption which is evident from the absorption
band at 450 nm in the UV–vis spectrum of TNA.[43a,50] How-
ever, these traps are detrimental to the water splitting process
due to the annihilation of photogenerated charge carriers via
trap assisted recombination.[51]
This explained the low observed
photocurrent density and low ABPE for TNA samples under
irradiation from the 450 nm LED (blue curve in Figure 5b,e).
The flame annealing of TNA samples at elevated temperature
results in the transformation of anatase TiO2 into rutile phase
TiO2 in STNA with a band-edge at 410 nm. The ≈25 nm red-
shift in the band-edge enables STNAs to harvest a significant
fraction of violet and blue photons, and results in the gen-
eration of photocurrent densities of ≈0.75 mA cm−2 at +0.6 V
versus Ag/AgCl by bare, undecorated STNA samples (cyan
curve in Figure 5b). Work function spectra indicate that the
Fermi level of rutile phase TiO2 in STNA lies 0.2 eV lower
than anatase phase TiO2 in good agreement with literature
values.[44] After decoration by phosphorus doped CNPQDs, the
photoelectrochemical performance was dramatically increased
to 2.54 mA cm−2 (red color in Figure 5a–e) which was due to
enhanced harvesting of visible light by CNPQDs coupled with
efficient injection of photogenerated electrons from the CB of
CNPQDs to the CB of TiO2.[8a] Therefore photosensitization
was indicated to be the leading mechanism in photocatalysis.
The resulting electrons in the conduction band of TiO2 moved
to the Pt cathode through the external circuit to reduce water
while the holes at photoanode facilitated water oxidation.
Despite the similar geometric dimensions of TNA and STNA,
the CNPQD-STNA samples exhibited photocurrents that were
100% larger than those obtained using CNPQD-TNA samples.
This large difference in performance was primarily attributable
to two reasons. The 0.2 eV lower CB position in rutile phase
STNA compared to anatase phase facilitated electron injection
from the CNPQDs into STNA and enabled higher internal
quantum yields. Secondly, the square-shaped morphology of
the STNAs played a role in enhancing light-matter interactions
(Figure 6 and Figure S10, Supporting Information). The overall
water splitting reaction can be expressed as follow
4OH 4h O 2H O at anode2 2+ → +− +
(1)
4H O 4e 2H 4OH at counter electrode2 2+ → +− −
(2)
High photoelectrochemical performance by a semiconductor
photoanode such as shown in Figure 5a,c is only possible
Figure 7. Mechanistic scheme of photoelectrochemical water splitting on
CNPQD-STNA nanohybrid.

when the photogenerated holes are efficiently transferred to
the hydroxyl ions in the aqueous electrolyte per Equation (1). In
carbon nitride quantum dots, secondary and tertiary N atoms
bonded to C constitute the highest occupied molecular orbital
(HOMO) (equivalent to valence band) while the π-conjugated
system behaves like the conduction band lowest unoccupied
molecular orbital (LUMO).[4d,52] Phosphorus doping takes place
by replacement of ring carbons with phosphorus which con-
tributes loosely held extra electron to the conjugated network
resulting in the uplifting of the VB and reduction of bandgap.[7a,37]
It is clear from this work that P-doped C3N4 quantum dots are
able to efficiently transfer holes to electrolyte species.
2.6. Photoreduction of 4-Nitrobenzenethiol (4-NBT)
to Dimercaptoazobenzene (DMAB)
Harvesting solar energy to drive surface catalytic reactions is
an enticing protocol due to their benign environment-friendly
nature and use of unexhaustive energy resource. Recently
plasmonic materials coupled with semiconductor or graphenic
materials has shown wide potential in plasmon-exciton (collec-
tively called “plexciton”) co-driven surface catalytic reaction due
to a synergistic improvement in plasmon-to-electron conversion
efficiency.[53] In this venture, plasmonic nanoparticles (NPs)
integrated with semiconductor photocatalysts such as Cu2O,[54]
MoS2,[55] TiO2,[50b,51] graphitic carbon nitride (g-C3N4),[56] doped
and undoped graphene,[53d,57] etc. have been explored for the
surface photocatalytic conversion of aromatic thiols into azo
compounds.[58] To probe the feasibility of CNPQDs decorated
TiO2 nanotubes to drive surface catalytic reactions, the photo-
catalytic reduction of 4-NBT into p,p′-dimercaptoazobenzene
was chosen as a model reaction and the progress of the reac-
tion was monitored by surface-enhanced Raman spectroscopy
(SERS) using a 532 nm laser. Prior to measurement, 10 nm
Ag was sputtered on to the samples followed by deposition
of 4-NBT and laser power was varied from 0.1 to 10.0 mW
to obtain the laser power dependent degradation of 4-NBT to
DMAB. Figure 8a shows the normal Raman spectra of pris-
tine 4-NBT and DMAB along with comparative SERS spectra
of 4-NBT adsorbed Ag sputtered glass, TNA, STNA, CNPQD-
TNA, and CNPQD-STNA samples at 1.0 mW incident laser
power. The Raman spectrum of 4-NBT contains three signa-
ture peaks at 1101, 1332, and 1576 cm−1
corroborated to SC
stretch, NO2 vibration and CC stretch respectively while the
Raman spectrum of DMAB has peaks at 1142 (CN stretch),
1389 and 1438 cm−1
(NN stretch).[53d,59]
After irradiation with
532 nm laser at a laser power 1.0 mW, bare Ag nanoparticles
on glass substrate show negligible change in Raman signal
demonstrating plasmonic silver alone was incapable to reduce
4-NBT (Figure 8b). Likewise, 4-NBT adsorbed Ag coupled bare
TNA samples showed a small decrease in peak intensity of
NO vibration at 1576 cm−1
and appearance of new peaks cor-
responding to DMAB A1g mode at 1142 (CN stretch), 1389
and 1438 (NN stretch) cm−1, respectively, revealing that a
very small fraction of 4-NBT was reduced to DMAB at 1 mW
laser power (Figure 8c). However, for Ag NP coupled STNA
samples a significant drop in NO peak intensity at 1576 cm−1
was observed along with a rise in the amplitude of the DMAB
peaks which confirmed the better photocatalytic performance
of STNA (Figure 8d). After decoration by CNPQDs, poorly
performing TNA exhibited excellent conversion of 4-NBT to
DMAB (Figure 8e). Using CNPQD-TNA and CNPQD-STNA,
almost complete conversion of 4-NBT to DMAB was attained
at a low laser power of 1 mW while bare Ag NPs, Ag NP/TNA,
and Ag NP/STNA were unable to complete the transformation
from 4-NBT to DMAB even at 10 mW laser power (Figure 8e).
From these observations it can be concluded that CNPQDs dec-
oration of TNA and STNA dramatically amplifies photocatalytic
performance. Prior reports show that graphene and graphenic
frameworks are highly effective at extracting hot electrons from
the plasmon decay in silver nanostructures.[3d,60] Therefore, a
higher plasmon-to-electron conversion efficiency on CNPQDs
leads to increased hot electron-density to facilitate high cata-
lytic conversion. In addition, the long interaction times of har-
vested hot electrons with phonons in graphenic frameworks[61]
improves the rate of phonon-driven catalysis.
3. Conclusion
Phosphorus doping of bulk graphitic carbon nitride (g-C3N4)
has been previously demonstrated to narrow the bandgap of
g-C3N4 from 2.7 to 1.7 eV. We synthesized P-doped g-C3N4
quantum dots (CNPQDs) with an average size of 28 nm using
a facile and scalable solid-state thermal condensation polym-
erization reaction method. The CNPQDs were found to have
a negatively charged surface with a zeta potential of −20 mV
and were easily dispersed in water. CNPQDs had an enlarged
bandgap of 2.1 eV and exhibited a more red-shifted photolu-
minescence (λmax = 478 nm) in comparison to other reported
carbon nitride/carbon quantum dots. The CNPQDs were
hosted on anodically formed titania nanotube array scaffolds to
create heterojunction photocatalysts. The titania nanotubes had
an average diameter of 70 nm and a tube-length of ≈0.5 µm.
Two distinct types of titania nanotube supports were used for
decoration with CNPQDs— anatase phase cylindrical TNA and
rutile phase STNA. The TNA and STNA samples significantly
differed in their interactions with light both prior to, and fol-
lowing decoration with CNPQDs. Optical FDTD simulations
indicated the STNAs to exhibit a lesser waveguiding effect
for UV photons than TNAs and provide a stronger confine-
ment of 500 nm photons than TNAs while displaying a higher
density of hot spots. These differences enabled CNPQD-
STNAs to perform enhanced light harvesting in comparison
to CNPQD-TNAs, as confirmed by experimentally measured
optical absorption spectra. When used as photoanodes for
sunlight-driven water splitting, CNPQD-STNAs generated
a photocurrent density as high as 2.54 mA cm−2
at +0.6 V
versus Ag/AgCl while achieving significant quantum yields
for wavelengths as long as 500 nm. The resulting hydrogen
generation rate was measured to be 22 µmol h−1
, which cor-
responded to a Faradaic efficiency as high as 93.2%. When
loaded with sputtered Ag NPs, the CNPQD-STNAs were trans-
formed into plexcitonic photocatalysts due to the interaction
of the plasmonic and excitonic resonances in the weak cou-
pling regime. Ag NP/CNPQD-STNAs were demonstrated to
be excellent visible light-driven photocatalysts by completing

the photoreduction of 4-NBT to p,p′-dimercaptoazobenzene
at a laser power of 1 mW at 532 nm whilst Ag NP/STNAs
could not complete the photoreduction even at a laser power
of 10 mW at 532 nm, thus confirming the importance of the
coupling between the Ag NPs and CNPQDs in generating high
photocatalytic performance.
4. Experimental Section
Synthesis of Phosphorus Doped CNPQDs: The phosphorus doped
CNPQDs were synthesized by solid-phase thermal condensation of urea,
citric acid, and BMIM-PF6 by slight modification of previously reported
procedures.[12]
In this synthesis procedure, citric acid and urea work as
the source of carbon and nitrogen respectively to form the carbon nitride
Figure 8. Schematic showing plexcitonic surface catalytic reduction of 4-nitrobenzenethiol (4-NBT) to DMAB (p,p′-dimercaptoazobenzene) and
a) Raman spectra of pristine 4-NBT (black), DMAB (purple) and comparison of SERS spectra of plexcitonic surface catalytic transformation of 4-NBT
to DMAB on bare Ag (orange),10 nm Ag sputtered TNA (blue), STNA (red), CNPQD-TNA (green), CNPQD-STNA (wine red) under 532 nm laser
illumination at 1 mW power, and SERS spectra of plasmon-exciton co-induced surface catalytic degradation of 4-NBT to DMAB as function of laser
power on a) 10 nm Ag sputtered on glass, c) 10 nm Ag sputtered TNA, d) 10 nm Ag sputtered STNA, e) 10 nm Ag sputtered CNPQD-TNA, f) 10 nm
Ag sputtered CNPQD-STNA.

framework while BMIM-PF6 furnishes phosphorus for intentional doping.
Briefly, 1.94 g citric acid (10 mmol) and 3.60 g urea (60 mmol) were
mixed together in an agate mortar followed by addition of 1.0 g BMIM-PF6
(7.0 mmol). The afforded mixture was transferred into a 50 mL Teflon
lined stainless steel autoclave and heated at 200 °C for 2 h to achieve
condensation polymerization and phosphorus doping. After cooling to
room temperature, the obtained solid was washed with dichloromethane
to remove unreacted organics, followed by extraction in water. The obtained
solution was centrifuged at 12 000 rpm to sediment large particles.
Synthesis of TNA and STNAs: TNA samples were prepared by
electrochemical anodization of Ti foil using an aqueous electrolyte
recipe. The water-based electrolyte was composed of 4% v/v of HF
and 8% v/v of acetic acid. The anodization was carried out at room
temperature using a two-electrode system containing a graphite cathode
and Ti anode with an applied voltage of 15 V between the Ti anode and
graphite cathode. The obtained TiO2 nanotube samples were annealed
at 450 °C under air for three hours. The samples were denoted as TNA.
The TNA samples were subjected to flame annealing at an elevated
temperature (≈750−900 °C) using a propane torch for 2 min. The
obtained flame annealed samples due to their specific square shape was
denoted as flame annealed square-shaped TiO2 nanotubes array.
Synthesis of CNPQD-TNA and CNPQD-STNAs: For the synthesis of
CNPQD decorated TNA photocatalyst, 15 mL of CNPQDs (20 mg mL−1
)
was added to a 50 mL Teflon lined autoclave. To this solution Ti foil
having TiO2 nanotube arrays (TNA and STNA) were immersed and
keep standing vertically. Autoclave was put in an oven at 120 °C for
12 h for hydrothermal treatment. After cooling to room temperature,
the CNPQDs decorated TiO2 samples were withdrawn and washed with
deionized water and methanol. The obtained samples were dried at
60 °C and used for photoelectrochemical measurements.
Measurement of Photoelectrochemical Water Splitting Performance: To
perform photoelectrochemical water splitting experiment three electrode
system composed of bare and CNPQDs decorated TiO2 samples as
photoanode (working electrode), Pt cathode (counter electrode) and
Ag/AgCl reference electrode system, respectively, was immersed in 0.1
m KOH electrolyte. To stimulate a photoresponse, AM1.5 G simulated
sunlight from a Newport Oriel solar simulator (model 67 005) equipped
with a Class A filter was used and the power density on the surface
of samples was maintained at 100 mW m−2
. To probe the visible light
response of materials, the photocurrent was measured in response
to AM1.5 G simulated sunlight passed through a 420 nm cut-off
filter. Further, to determine the photoresponse behavior of samples
at discrete visible wavelengths and calculate quantum efficiencies,
experiments were carried out using 450 and 500 nm wavelength LEDs,
having 54.15 and 40.48 mW cm−2
power densities at the surface of the
sample, respectively. The photocurrent was measured by linear sweep
voltammetry by sweeping applied voltage from −0.8 to +1.0 V versus
Ag/AgCl. The intensities of illumination from the LEDs on the sample
surface were calibrated with a photodiode. Additionally, dark current
was also measured for comparing the photoresponse of samples. The
gaseous sample from the Pt electrode containing arm was withdrawn
after 30 min interval using a gas tight syringe and analyzed using a gas
chromatographs (GC)–pulse discharge detector (PDD) system. The
gaseous products of photoelectrochemical water splitting were identified
and quantified using gas chromatography on a custom-made Shimadzu
GC equipped with Porapak Q column and Mol Sieve column, and a PDD.
Conditions: He carrier flow rate: 0.5 mL min−1
, detector temperature:
160 °C, oven temp: started from 60 °C and raised up to 160 °C. For the
quantification of gaseous hydrogen, a standard gas mixture (Praxair gas,
Canada) was used as a calibration standard.
Measurement of the Activity for Surface Catalytic Reactions: Prior to
measurement, 10 nm Ag was sputtered on the samples. For comparison
10 nm Ag was also sputtered on glass. A methanolic solution of 4-NBT
(5 × 10−5
m) was drop-casted on the Ag sputtered samples followed
by drying in a vacuum oven at room temperature. The photocatalytic
performance of samples in plasmon-exciton co-promoted surface
photocatalytic reduction of 4-NBT to DMAB was determined using
surface-enhanced Raman scattering (SERS) spectroscopy on a Thermo
Scientific DXR2 Raman Microscope with a 532 nm laser by varying laser
power from 0.1 to 10 mW. The SERS spectra were acquired using a
50 µm pinhole aperture size, 2 µm spot size and 2 cm−1 per CCD pixel
element 900 lines mm−1 spectral dispersion grating. The spectra were
accumulated for 3 × 25 s.
Optical Simulations: Optical properties of the cylindrical and square
nanotube structures were investigated using Lumerical FDTD simulation
software. The nanotubes were simulated with circular and square cross-
sections with inner and outer radii of 40 and 70 nm, respectively, and a
height of 300 nm. The simulations were performed for both combinations
of individual nanotubes in a vacuum environment, as well as nanotubes
decorated by nitrogen-doped C3N4 quantum dots with radii of 12 nm.
Scattering and absorption cross-sections, electric field intensity profiles,
Poynting vector distributions, as well as reflection and transmission spectra
were captured using near and far-field profile and frequency monitors.
Material compositions of rutile (square) and anatase (cylindrical)
TiO2 nantoubes were simulated using appropriate refractive index data
as obtained via ellipsometry. An additional set of refractive index data
corresponding to anatase TiO2 was also obtained from Siefke et al.[62]
Refractive index data corresponding to N@C quantum dots was obtained
from Kumar et al.[63] with the resulting stand-alone simulations of N@C
quantum dot nanoparticles displaying absorption spectrum profiles
identical to the UV–vis absorption spectra of the N@C quantum dots
utilized in experiments. Lumerical's refractive index monitor confirmed
the structures were appropriately configured and identified by the software
throughout the course of the simulation. The structures were illuminated
by a light source incident at a normal angle from above with a bandwidth
of 350–800 nm. Snapshots of the electric field intensity profiles, and
Poynting vector distributions were taken at selective wavelengths between
300–600 nm, and as observed along different planes including the xy-,
xz-, and yz-planes. Appropriate symmetric and antisymmetric boundaries
were considered given the symmetry of the nanotube structures, and
allowed for shorter simulation times, and accurate data aggregation.
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
Acknowledgements
All authors thank the Natural Sciences and Engineering Research
Council (NSERC), the National Research Council (NRC), Future Energy
Systems, and CMC Microsystems for direct and indirect (equipment use)
financial support. Scholarship supports from Alberta Innovates to U.K.T.,
and from NSERC to R.K. and A.P.M. are acknowledged. Some device
fabrication and testing used research infrastructure made possible by
a Leaders Opportunity Fund grant to K.S. from the Canada Foundation
for Innovation and matched by the Alberta Small Equipment Grants
Program. The authors acknowledge use of the following user-fee based
characterization facilities: The Cell Imaging Facility, the Oil Sands and
Coal Interfacial Engineering Facility, and the University of Alberta nanoFab.
Conflict of Interest
The authors declare no conflict of interest.
Author Contributions
P.K. and P.K. contributed equally to this work. P.K. and P.K. synthesized
and characterized the materials and performed photoelectrochemical
water splitting experiments. A.P.M. performed FDTD simulation studies.

U.K.T. conducted the electrochemical characterization of materials. K.M.A.
helped in collection of characterization data and also performed the 4-NBT
photoreduction experiments. Y.Z. helped in the synthesis of nanotubes
for the studies. R.K. helped in imaging and EDX analysis of samples. K.C.
performed HRTEM, elemental mapping and EELS on samples. S.Z. assisted
in calibration of LEDs and calculation of efficiencies. K.S. supervised the
research and edited the manuscript. G.M.B. and V.K.M. performed the
solid-state NMR experiments.
Keywords
carbon nitride quantum dots, hydrogen generation, metal-free catalysts,
nanotubes, semiconductor photoelectrocatalysis
Received: July 28, 2019
Revised: October 15, 2019
Published online:
[1] a) Z. Chen, H. N. Dinh, E. Miller, Photoelectrochemical Water Split-
ting, Springer, New York 2013; b) J. Stringer, L. Horton, Basic
Research Needs to Assure a Secure Energy Future, Oak Ridge, TN,
Oak Ridge National Laboratory, US Department of Energy 2003,
pp. A-97–A-101.
[2] a) H. Wang, T. Maiyalagan, X. Wang, ACS Catal. 2012, 2, 781;
b) A. Lherbier, X. Blase, Y.-M. Niquet, F. Triozon, S. Roche, Phys.
Rev. Lett. 2008, 101, 036808.
[3] a) J. Liu, Y. Liu, N. Liu, Y. Han, X. Zhang, H. Huang, Y. Lifshitz,
S.-T. Lee, J. Zhong, Z. Kang, Science 2015, 347, 970; b) X. Wang,
K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson,
K. Domen, M. Antonietti, Nat. Mater. 2009, 8, 76; c) D. Masih,
Y. Ma, S. Rohani, Appl. Catal., B 2017, 206, 556; d) P. Kumar,
E. Vahidzadeh, U. K. Thakur, P. Kar, K. M. Alam, A. Goswami,
N. Mahdi, K. Cui, G. M. Bernard, V. K. Michaelis, K. Shankar, J. Am.
Chem. Soc. 2019, 141, 5415.
[4] a) F. K. Kessler, Y. Zheng, D. Schwarz, C. Merschjann, W. Schnick,
X. Wang, M. J. Bojdys, Nat. Rev. Mater. 2017, 2, 17030; b) P. Kumar,
R. Boukherroub, K. Shankar, J. Mater. Chem. A 2018, 6, 12876;
c) M. Volokh, G. Peng, J. Barrio, M. Shalom, Angew. Chem., Int.
Ed. 2019, 58, 6138; d) W.-J. Ong, L.-L. Tan, Y. H. Ng, S.-T. Yong,
S.-P. Chai, Chem. Rev. 2016, 116, 7159.
[5] a) Y. Tang, Y. Su, N. Yang, L. Zhang, Y. Lv, Anal. Chem. 2014,
86, 4528; b) H. Li, F.-Q. Shao, H. Huang, J.-J. Feng, A.-J. Wang,
Sens. Actuators, B 2016, 226, 506; c) Q. Cui, J. Xu, X. Wang, L. Li,
M. Antonietti, M. Shalom, Angew. Chem., Int. Ed. 2016, 55, 3672.
[6] a) L. Jiang, X. Yuan, Y. Pan, J. Liang, G. Zeng, Z. Wu, H. Wang,
Appl. Catal., B 2017, 217, 388; b) Y. Wang, Y. Di, M. Antonietti,
H. Li, X. Chen, X. Wang, Chem. Mater. 2010, 22, 5119;
c) J. P. Paraknowitsch, A. Thomas, Energy Environ. Sci. 2013, 6,
2839; d) K. M. Alam, P. Kumar, P. Kar, U. K. Thakur, S. Zeng, K. Cui,
K. Shankar, Nanoscale Adv. 2019, 1, 1460.
[7] a) Y. Zhang, T. Mori, J. Ye, M. Antonietti, J. Am. Chem. Soc. 2010,
132, 6294; b) Y.-P. Zhu, T.-Z. Ren, Z.-Y. Yuan, ACS Appl. Mater. Inter-
faces 2015, 7, 16850.
[8] a) J. Su, L. Zhu, G. Chen, Appl. Catal., B 2016, 186, 127; b) Z. Zhao,
Y. Sun, F. Dong, Nanoscale 2015, 7, 15; c) C. Chang, L. Zhu,
S. Wang, X. Chu, L. Yue, ACS Appl. Mater. Interfaces 2014, 6, 5083;
d) J. Wen, J. Xie, X. Chen, X. Li, Appl. Surf. Sci. 2017, 391, 72;
e) G. Mamba, A. Mishra, Appl. Catal., B 2016, 198, 347; f) C. Ott,
F. Reiter, M. Baumgartner, M. Pielmeier, A. Vogel, P. Walke,
S. Burger, M. Ehrenreich, G. Kieslich, D. Daisenberger, Adv. Funct.
Mater. 2019, 29, 1900233; g) J. Xu, I. Herraiz-Cardona, X. Yang,
S. Gimenez, M. Antonietti, M. Shalom, Adv. Opt. Mater. 2015, 3,
1052; h) B. Chai, T. Peng, J. Mao, K. Li, L. Zan, Phys. Chem. Chem.
Phys. 2012, 14, 16745; i) L. Zhang, D. Jing, X. She, H. Liu, D. Yang,
Y. Lu, J. Li, Z. Zheng, L. Guo, J. Mater. Chem. A 2014, 2, 2071.
[9] a) K. Shankar, J. I. Basham, N. K. Allam, O. K. Varghese, G. K. Mor,
X. Feng, M. Paulose, J. A. Seabold, K.-S. Choi, C. A. Grimes, J. Phys.
Chem. C 2009, 113, 6327; b) K. R. Reyes-Gil, D. B. Robinson, ACS
Appl. Mater. Interfaces 2013, 5, 12400.
[10] a) X. Zhang, F. Han, B. Shi, S. Farsinezhad, G. P. Dechaine,
K. Shankar, Angew. Chem., Int. Ed. 2012, 51, 12732; b) G. L. Chiarello,
A. Zuliani, D. Ceresoli, R. Martinazzo, E. Selli, ACS Catal. 2016, 6,
1345.
[11] A. Mohammadpour, P. Kar, B. D. Wiltshire, A. M. Askar, K. Shankar,
Curr. Nanosci. 2015, 11, 593.
[12] a) P. Kumar, U. K. Thakur, K. Alam, P. Kar, R. Kisslinger, S. Zeng,
S. Patel, K. Shankar, Carbon 2018, 137, 174; b) J. Zhou, Y. Yang,
C.-y. Zhang, Chem. Commun. 2013, 49, 8605; c) Y.-C. Lu, J. Chen,
A.-J. Wang, N. Bao, J.-J. Feng, W. Wang, L. Shao, J. Mater. Chem. C
2015, 3, 73.
[13] J. Wu, S. Yang, J. Li, Y. Yang, G. Wang, X. Bu, P. He, J. Sun, J. Yang,
Y. Deng, G. Ding, X. Xie, Adv. Opt. Mater. 2016, 4, 2095.
[14] a) N. Wang, H. Fan, J. Sun, Z. Han, J. Dong, S. Ai, Carbon 2016,
109, 141; b) W. Wang, C. Y. Jimmy, Z. Shen, D. K. Chan, T. Gu,
Chem. Commun. 2014, 50, 10148; c) Q. Liang, W. Ma, Y. Shi, Z. Li,
X. Yang, Carbon 2013, 60, 421.
[15] B. Chai, J. Yan, C. Wang, Z. Ren, Y. Zhu, Appl. Surf. Sci. 2017, 391,
376.
[16] a) J. Zhang, X. Chen, K. Takanabe, K. Maeda, K. Domen, J. D. Epping,
X. Fu, M. Antonietti, X. Wang, Angew. Chem., Int. Ed. 2010, 49, 441;
b) A. Kumar, P. Kumar, C. Joshi, S. Ponnada, A. K. Pathak, A. Ali,
B. Sreedhar, S. L. Jain, Green Chem. 2016, 18, 2514.
[17] a) S. Barman, M. Sadhukhan, J. Mater. Chem. 2012, 22, 21832;
b) Y. Dong, H. Pang, H. B. Yang, C. Guo, J. Shao, Y. Chi, C. M. Li,
T. Yu, Angew. Chem. 2013, 125, 7954.
[18] T. T. Meiling, P. J. Cywin´ski, I. Bald, Sci. Rep. 2016, 6, 28557.
[19] a) B. V. Lotsch, M. Döblinger, J. Sehnert, L. Seyfarth, J. Senker,
O. Oeckler, W. Schnick, Chem. - Eur. J. 2007, 13, 4969; b) B. Jürgens,
E. Irran, J. Senker, P. Kroll, H. Müller, W. Schnick, J. Am. Chem.
Soc. 2003, 125, 10288; c) A. Schwarzer, U. Böhme, E. Kroke, Chem.
- Eur. J. 2012, 18, 12052; d) X. Li, I. V. Sergeyev, F. Aussenac,
A. F. Masters, T. Maschmeyer, J. M. Hook, Angew. Chem. 2018, 130,
6964; e) Y. Hu, Y. Shim, J. Oh, S. Park, S. Park, Y. Ishii, Chem. Mater.
2017, 29, 5080; f) Y. Wang, Y. Li, W. Ju, J. Wang, H. Yao, L. Zhang,
J. Wang, Z. Li, Carbon 2016, 102, 477.
[20] a) X.-X. Fang, L.-B. Ma, K. Liang, S.-J. Zhao, Y.-F. Jiang, C. Ling,
T. Zhao, T.-Y. Cheang, A.-W. Xu, J. Mater. Chem. A 2019, 7, 11506;
b) S. Guo, Z. Deng, M. Li, B. Jiang, C. Tian, Q. Pan, H. Fu, Angew.
Chem., Int. Ed. 2016, 55, 1830.
[21] a) P. Kar, Y. Zhang, S. Farsinezhad, A. Mohammadpour,
B. D. Wiltshire, H. Sharma, K. Shankar, Chem. Commun. 2015, 51,
7816; b) P. Kar, Y. Zhang, N. Mahdi, U. K. Thakur, B. D. Wiltshire,
R. Kisslinger, K. Shankar, Nanotechnology 2017, 29, 014002.
[22] P. Kar, S. Zeng, Y. Zhang, E. Vahidzadeh, A. Manuel, R. Kisslinger,
K. M. Alam, U. K. Thakur, N. Mahdi, P. Kumar, K. Shankar, Appl.
Catal., B 2019, 243, 522.
[23] S. Yang, Y. Gong, J. Zhang, L. Zhan, L. Ma, Z. Fang, R. Vajtai,
X. Wang, P. M. Ajayan, Adv. Mater. 2013, 25, 2452.
[24] Y. Zheng, Z. Zhang, C. Li, J. Photochem. Photobiol., A 2017, 332, 32.
[25] Y. Liao, W. Cao, J. W. Connell, Z. Chen, Y. Lin, Sci. Rep. 2016, 6,
26084.
[26] a) D. H. Park, K. S. Lakhi, K. Ramadass, M. K. Kim, S. N. Talapaneni,
S. Joseph, U. Ravon, K. Al-Bahily, A. Vinu, Chem. - Eur. J. 2017, 23,
10753; b) L. F. Chen, Z. H. Huang, H. W. Liang, H. L. Gao, S. H. Yu,
Adv. Funct. Mater. 2014, 24, 5104.
[27] a) H.-S. Kim, J.-W. Lee, N. Yantara, P. P. Boix, S. A. Kulkarni,
S. Mhaisalkar, M. Grätzel, N.-G. Park, Nano Lett. 2013, 13, 2412;

b) E. Hosono, S. Fujihara, K. Kakiuchi, H. Imai, J. Am. Chem. Soc.
2004, 126, 7790.
[28] a) C. Ratanatawanate, C. Xiong, K. J. Balkus Jr., ACS Nano 2008, 2,
1682; b) T. Ohsaka, F. Izumi, Y. Fujiki, J. Raman Spectrosc. 1978, 7,
321.
[29] Y. Kang, Y. Yang, L. C. Yin, X. Kang, G. Liu, H. M. Cheng, Adv. Mater.
2015, 27, 4572.
[30] M. Y. Chen, D. Li, X. Lin, V. Dravid, Y. W. Chung, M. S. Wong,
W. D. Sproul, J. Vac. Sci. Technol., A 1993, 11, 521.
[31] P. Cheng, Z. Yang, H. Wang, W. Cheng, M. Chen, W. Shangguan,
G. Ding, Int. J. Hydrogen Energy 2012, 37, 2224.
[32] S. Kumar, P. Kumar, S. L. Jain, J. Mater. Chem. A 2014, 2, 18861.
[33] D. Eder, A. H. Windle, Adv. Mater. 2008, 20, 1787.
[34] a) Q. Liang, Z. Li, X. Yu, Z. H. Huang, F. Kang, Q. H. Yang, Adv.
Mater. 2015, 27, 4634; b) Z. Song, T. Lin, L. Lin, S. Lin, F. Fu,
X. Wang, L. Guo, Angew. Chem., Int. Ed. 2016, 55, 2773.
[35] a) A. Kumar, P. Kumar, R. Borkar, A. Bansiwal, N. Labhsetwar,
S. L. Jain, Carbon 2017, 123, 371; b) N. Hellgren, R. T. Haasch,
S. Schmidt, L. Hultman, I. Petrov, Carbon 2016, 108, 242.
[36] J. Ran, T. Y. Ma, G. Gao, X.-W. Du, S. Z. Qiao, Energy Environ. Sci.
2015, 8, 3708.
[37] X. Chen, Q. Liu, Q. Wu, P. Du, J. Zhu, S. Dai, S. Yang, Adv. Funct.
Mater. 2016, 26, 1719.
[38] J. Park, T. Back, W. C. Mitchel, S. S. Kim, S. Elhamri, J. Boeckl,
S. B. Fairchild, R. Naik, A. A. Voevodin, Sci. Rep. 2015, 5, 14374.
[39] J. Chen, W. Song, H. Hou, Y. Zhang, M. Jing, X. Jia, X. Ji, Adv. Funct.
Mater. 2015, 25, 6793.
[40] a) M. S. Hamdy, R. Amrollahi, G. Mul, ACS Catal. 2012, 2, 2641;
b) S. G. Ullattil, P. Periyat, J. Mater. Chem. A 2016, 4, 5854.
[41] P. Kar, S. Zheng, Y. Zhang, E. Vahidzadeh, A. Manuel, R. Kisslinger,
K. Alam, U. Thakur, N. Mahdi, P. Kumar, Appl. Catal., B 2019, 243,
522.
[42] H. Li, F.-Q. Shao, S.-Y. Zou, Q.-J. Yang, H. Huang, J.-J. Feng,
A.-J. Wang, Microchim. Acta 2016, 183, 821.
[43] a) C. Di Valentin, G. Pacchioni, A. Selloni, J. Phys. Chem. C 2009,
113, 20543; b) Z. Lian, W. Wang, G. Li, F. Tian, K. S. Schanze, H. Li,
ACS Appl. Mater. Interfaces 2017, 9, 16959.
[44] a) S. N. Habisreutinger, L. Schmidt-Mende, J. K. Stolarczyk, Angew.
Chem., Int. Ed. 2013, 52, 7372; b) L. Jing, S. Li, S. Song, L. Xue,
H. Fu, Sol. Energy Mater. Sol. Cells 2008, 92, 1030.
[45] F. E. Osterloh, Chem. Soc. Rev. 2013, 42, 2294.
[46] O. K. Varghese, C. A. Grimes, Sol. Energy Mater. Sol. Cells 2008, 92,
374.
[47] a) X.-Y. Zhang, A. Hu, T. Zhang, X.-J. Xue, J. Z. Wen, W. W. Duley,
Appl. Phys. Lett. 2010, 96, 043109; b) A. Mohammadpour,
K. Shankar, J. Mater. Chem. 2010, 20, 8474.
[48] S. Farsinezhad, A. Mohammadpour, A. N. Dalrymple, J. Geisinger,
P. Kar, M. J. Brett, K. Shankar, J. Nanosci. Nanotechnol. 2013, 13, 2885.
[49] M. G. Walter, E. L. Warren, J. R. McKone, S. W. Boettcher, Q. Mi,
E. A. Santori, N. S. Lewis, Chem. Rev. 2010, 110, 6446.
[50] a) C. Mao, F. Zuo, Y. Hou, X. Bu, P. Feng, Angew. Chem. 2014, 126,
10653; b) Z. Wang, B. Wen, Q. Hao, L.-M. Liu, C. Zhou, X. Mao,
X. Lang, W.-J. Yin, D. Dai, A. Selloni, J. Am. Chem. Soc. 2015, 137,
9146.
[51] S. K. Cushing, F. Meng, J. Zhang, B. Ding, C. K. Chen, C.-J. Chen,
R.-S. Liu, A. D. Bristow, J. Bright, P. Zheng, ACS Catal. 2017, 7, 1742.
[52] J. Su, L. Zhu, P. Geng, G. Chen, J. Hazard. Mater. 2016, 316, 159.
[53] a) A. G. da Silva, T. S. Rodrigues, V. G. Correia, T. V. Alves,
R. S. Alves, R. A. Ando, F. R. Ornellas, J. Wang, L. H. Andrade,
P. H. Camargo, Angew. Chem., Int. Ed. 2016, 55, 7111; b) J. Wang,
W. Qiao, X. Mu, Appl. Sci. 2018, 8, 2026; c) A. P. Manuel, A. Kirkey,
N. Mahdi, K. Shankar, J. Mater. Chem. C 2019, 7, 1821; d) X. Ren,
E. Cao, W. Lin, Y. Song, W. Liang, J. Wang, RSC Adv. 2017, 7, 31189.
[54] T. You, L. Jiang, P. Yin, Y. Shang, D. Zhang, L. Guo, S. Yang,
J. Raman Spectrosc. 2014, 45, 7.
[55] X. Yang, H. Yu, X. Guo, Q. Ding, T. Pullerits, R. Wang, G. Zhang,
W. Liang, M. Sun, Mater. Today Energy 2017, 5, 72.
[56] P. Kumar, E. Vahidzadeh, U. K. Thakur, P. Kar, K. M. Alam,
A. Goswami, N. Mahdi, K. Cui, G. M. Bernard, V. K. Michaelis,
J. Am. Chem. Soc. 2019, 141, 5415.
[57] a) Q. Ding, Y. Shi, M. Chen, H. Li, X. Yang, Y. Qu, W. Liang, M. Sun,
Sci. Rep. 2016, 6, 32724; b) E. Cao, M. Sun, Y. Song, W. Liang, Nano-
technology 2018, 29, 372001; c) L. Kang, J. Chu, H. Zhao, P. Xu,
M. Sun, J. Mater. Chem. C 2015, 3, 9024.
[58] W. Lin, Y. Cao, P. Wang, M. Sun, Langmuir 2017, 33, 12102.
[59] a) B. Dong, Y. Fang, L. Xia, H. Xu, M. Sun, J. Raman Spectrosc. 2011,
42, 1205; b) S. Wu, Y. Liu, C. Ma, J. Wang, Y. Zhang, P. Song, L. Xia,
Langmuir 2018, 34, 7240.
[60] J. Wang, W. Lin, X. Xu, F. Ma, M. Sun, Chem. Rec. 2018, 18, 481.
[61] Q. Ding, Y. Shi, M. Chen, H. Li, X. Yang, Y. Qu, W. Liang, M. Sun,
Sci. Rep. 2016, 6, 32724.
[62] T. Siefke, S. Kroker, K. Pfeiffer, O. Puffky, K. Dietrich, D. Franta,
I. Ohlídal, A. Szeghalmi, E.-B. Kley, A. Tünnermann, Adv. Opt.
Mater. 2016, 4, 1780.
[63] V. B. Kumar, A. K. Sahu, A. S. M. Mohsin, X. Li, A. Gedanken, ACS
Appl. Mater. Interfaces 2017, 9, 28930.

Noble Metal Free, Visible Light Driven Photocatalysis Using TiO2 Nanotube Arrays Sensitized by P‐Doped C3N4 Quantum Dots

More Related Content

What's hot (20)

Similar to Noble Metal Free, Visible Light Driven Photocatalysis Using TiO2 Nanotube Arrays Sensitized by P‐Doped C3N4 Quantum Dots (20)

More from Pawan Kumar (20)

Recently uploaded (20)

Noble Metal Free, Visible Light Driven Photocatalysis Using TiO2 Nanotube Arrays Sensitized by P‐Doped C3N4 Quantum Dots