S. Chen et al. / Applied Catalysis A: General 498 (2015) 63–68
67
Fig. 5. The time courses of H2 evolution from water over Pt/g-C3N4, PTCDIs/Pt/g-
C3N4 and Pt/PTCDIs nanofibers.
Fig. 6. Stable H2 evolution from water at PTCDI-1/Pt/g-C3N4. The reaction was con-
[25]. Despite the different morphologies, both PTCDI nanostruc-
tures form extended 1D charge carrier pathways (enabled by the
– stacking) and have large surface area, making them appropri-
ate for comparative photocatalytic investigation.
are quite consistent with the photocatalytic H2 evolution results.
Fig. 6 shows the stability of the photocatalytic activity of PTCDI-
1/Pt/g-C3N4 for H2 generation under visible-light illumination. As
both PTCDI-1 and g-C3N4 are stable under light irradiation in
aqueous triethanolamine solutions [37], the photocatalytic reac-
tion results in a stable H2 evolution rate, although there is still
about 20% decrease during 72 h probably due to separation between
PTCDI and g-C3N4 from stirring. Furthermore, while many dyes suf-
fer from photobleaching [14]. PTCDI-1 and PTCDI-2 are stable. They
can be regenerated by dissolving the reacted composite powders in
CHCl3. Indeed, reused PTCDI-1/Pt/g-C3N4 displayed an equivalent
photocatalytic H2 evolution rate (Fig. S5).
3.4. Photocatalytic H2 production
Photocatalytic H2 evolution reactions were performed in aque-
ous solutions with 10% (by volume) triethanolamine as the
sacrificial electron donor under visible-light (ꢀ ≥ 420 nm) illu-
mination. As shown in Fig. 5, steady H2 evolution from the
to 0.375 mol h−1 obtained for PTCDI-1/Pt/g-C3N4, which is about
ten times higher than that of Pt/g-C3N4. Furthermore, Pt/PTCDI
nanofibers without the g-C3N4 showed much lower activity (ca.
0.0225–0.0375 mol h−1, Fig. 5) under the same conditions, owing
to the intrinsic inefficient photoactivity of organic dyes without
a semiconductor matrix to act as an efficient charge separator
In principle, the mesoporous morphology of g-C3N4 facilitates
the surface deposition of PTCDI (and thus the charge separa-
tion), and mass diffusion during the photocatalytic H2 evolution
reactions [7]. Nevertheless, in both cases, composites contain-
ing symmetric dimethylamino)benzyl substituted PTCDI-1 give
rise to improved photocatalytic activity arising from enhanced
intramolecular charge separation (due to two electron donor
groups) and higher LUMO orbital than that of asymmetric PTCDI-
sion at white or UV light, demonstrating efficient photoinduced
intermolecular charge transfer within the – assemblies, which is
consistent with the H2 evolution capabilities of Pt/PTCDI nanofibers
[25]. Moreover, as summarized in Table 2, under the irradiation
of monochromatic light with the wavelength of 420 nm, the AQEs
of different photocatalysts are ranked as PTCDI/Pt/g-C3N4 > Pt/g-
C3N4 > Pt/PTCDI nanofiber. PTCDI-1/Pt/g-C3N4 shows the highest
activity, which is about sevenfold higher than that of Pt/g-C3N4 and
A “cooperative excitation process” is proposed to explain the
photo-induced charge transfer process that enhances H2 evolution
over PTCDI/Pt/g-C3N4 compared to bare Pt/g-C3N4 or Pt/PTCDIs. As
shown in Fig. 7, under visible-light (ꢀ ≥ 420 nm) illumination, both
PTCDI and g-C3N4 can be excited by absorbing photons. On one
hand, an ultra-fast intramolecular charge transfer occurred; that
is, after the visible-light-driven excitation of PTCDI, electrons are
transferred from the (dimethylamino)benzyl moiety (donor part)
to the HOMO orbital of the PTCDI core (acceptor part). Due to the
presence of large excess of triethanolamine (sacrificial reagent),
the (dimethylamino)benzyl moiety under the oxidation state can
be easily reduced to its ground state, thus preventing the charge
recombination, leaving excited electrons at the LUMO of PTCDI core.
The anionic radical of PTCDI thus formed acts a strong electron
donor. On the other hand, the g-C3N4 is also excited to promote
photo-excited electrons to its conduction band (CB), leaving holes
on its valance band (VB). The latter can be quenched by the elec-
trons transferred from the PTCDI’s LUMO due to their large energy
level difference (driving force). This process inhibits the recom-
bination of photogenerated electron–hole pairs within g-C3N4. It
should be noted that although the electrons transfer from the LUMO
of the PTCDI to the CB of g-C3N4 is possible, the process is quite
weaken due to the small energy level difference (∼0.1 eV). Thus,
the collected electrons in the CB of g-C3N4 are captured by the
loaded cocatalyst Pt nanoparticles, which provide active sites for
the reduction of water to H2 and suppress electron–hole recom-
bination [10]. The electrons are consumed by H2O to complete
the reduction reaction process for H2 production. Consequently,
efficient photoinduced charge separation is achieved both within
PTCDI aggregates (via intramolecular electron transfer) and at
the interface between PTCDI and g-C3N4 (via interfacial electron
Table 2
The data of the apparent quantum efficiency of various photocatalysts.
Photocatalysts
H2 evolution rate/mol h−1
AQE/%
Pt/PTCDI-1 nanofiber
Pt/PTCDI-2 nanofiber
Pt/g-C3N4
PTCDI-1/Pt/g-C3N4
PTCDI-2/Pt/g-C3N4
0.3
0.25
0.6
3.8
2.2
0.024
0.020
0.048
0.31
0.17