28
M. Piao et al.
Electrochemical Measurement of Porphyrin
CuTHPP was modified on a glassy carbon electrode (GCE)
using acetone solution and dried at room temperature. A Pyrex
electrolytic cell was employed, filled with Ar-purged 0.5 M
Na2SO4 (20 mL). A Pt wire and Ag|AgCl electrode were
employed as the counter and reference electrodes, respectively.
Impedance measurements were recorded using a CHI660 elec-
trochemical workstation.
C2H4
CuTHPP
LUMO
eꢀ
CO2
eꢀ
eꢀ
eꢀ
ꢁ
Photocatalytic Measurement
H2O
A double-sided photocatalytic film with an appropriate area of
2 ꢁ 10 cm ꢁ 20 cm was introduced into a 1-L quartz chamber
equipped with valves for evacuation and gas feeding. Following
loading of the sample, the chamber was evacuated and sealed to
,10 mTorr using a mechanical pump. Carbon dioxide (CO2,
99.99 % pure) was forced to pass through a bubbler containing
deionized water before entering the reaction chamber. These
two processes of evacuation and CO2 pumping were repeated
thrice. The approximate excess pressure is less than 1.0 psi. All
CO2 conversion experiments were performed under irradiation
O2
Scheme 1. Transfer process of photogenerated electrons for CuTHPP/
graphene composite. Under illumination, the photoexcited electrons are
transferred from the porphyrin molecules passing through graphene sheets.
Due to the hydrophobicity of graphene, the transfer of electrons from
graphene to H2O molecules is difficult.
of a solar simulator. The light power density was 100 mW cmꢀ2
.
2.72 eV, as shown in Fig. 1b. To estimate the conduction band
potential (ECB), Mott–Schottky plot for CuTHPP was con-
structed under the condition of pH ¼ 7.0, as shown in Fig. 1a.
It can be seen that the slope of the plot was positive, thus
suggesting that CuTHPP is an n-type semiconductor. When
illuminated, the generated electrons are transferred from the
porphyrin molecules to graphene, while the holes remain in the
porphyrin because of the properties of these two materials
mentioned above. The ECB of CuTHPP was approximately
ꢀ0.74 V versus Ag|AgCl (ꢀ0.54 V versus NHE (normal hydro-
gen electrode)). According to the band gap and ECB of CuTHPP,
we can achieve the relationship between the energy bands of this
porphyrin and the energy levels of the redox couple, as clearly
seen in Fig. 1. In this case, theoretically, CO2 can be reduced to a
carbohydrate, and H2O can be oxidized to O2 using the present
photocatalyst.
Owing to the large number of defects in reduced graphite
oxide (RGO), the transfer of electrons will be greatly hindered,
and the defects will also lead to the introduction of some trap
levels in the band gap of graphene. These traps will result in a
difficulty in transferring electrons from the GS to the adsorbed
molecule for following reactions.[22] In consequence, the prep-
aration of graphene using conventional methods (graphite oxide
reduction) is not sufficiently effective. Therefore, the presence
of fewer defects and the higher conductivity of graphene are
desirable in this photocatalytic composite.
The equilibrium temperature of the samples was ,408C (ꢂ58C).
Although the experiments have also been conducted at lower
temperatures (,358C), no definite influence of temperature on
product formation rates has been found.
Results and Discussion
In order to construct an outstanding photocatalyst with high
selectivity, two types of special materials were used as shown in
Scheme 1. CuTHPP was chosen as the antenna (light exciter)
rather than inorganic materials because CuTHPP has a greater
ability to absorb light than inorganic materials. Compared with
inorganic catalyst materials, organic molecules possess rela-
tively simple light absorption suitable for investigating, with no
interference from the effect of surface state and impurity energy
levels. More importantly, organic molecules as catalysts do not
possess surface dangling bonds, which are otherwise abundant
on the surface of inorganic materials. Thus, the interaction
between porphyrin and reactants is simpler to control. All above
characteristics are beneficial to improve the selectivity of pho-
tocatalysts. Another important reason to choose such porphyrin
molecule as antenna is its excellent photostability.
Graphene was introduced into this photocatalyst system as a
charge transfer mediator for its high electrical conductivity and
tuneable band gap properties.[26–28] Moreover, graphene is an
outstandingCO2 absorberaccordingtoliteraturereports,ofwhich
the theoretical maximum uptake of CO2 is 37.93 wt-%.[29] After
combining these two components, under light irradiation, the
excited electrons would move from the antenna to the CO2
molecules while easily passing through the graphene component,
thus greatly enhancing the conversion rate effectively. Moreover,
the hydrophobic property of graphene allows the catalytic reac-
tion to proceed on the surface in a more controllable fashion
because H2O molecules are unable to accept photogenerated
chargesfromthegraphenesurface.Therefore, these twomaterials
werecombinedtorealize hydrocarbonphotogeneration, asshown
in Scheme 1.
After preparation, a series of characterizations were per-
formed to investigate the quality of the graphene sheets, as
shown in Fig. 2. The exfoliation state of the material was first
ascertained by scanning electron microscopy (SEM) measure-
ments (Fig. 2a). After exfoliation, the thickness of the flakes
decreased significantly when compared with that of graphite, as
clearly shown in the SEM image. Transmission electron micro-
scopy (TEM) was used to reveal the number of layers in the GS.
The high-resolution TEM image (Fig. 2b) shows that the
structure of GS consisted of a few layers.
The structure of graphene was further investigated by Raman
spectroscopy, as shown in Fig. 2c. The G band (,1580 cmꢀ1
)
In order to confirm whether this photocatalyst can reduce
CO2 into hydrocarbon, the energy level of this material was
investigated using the Mott–Schottky method and UV-visible
measurement technique (Fig. 1a, b). Through the UV-visible
measurements, the energy gap of CuTHPP was estimated as
and 2D band (,2700 cmꢀ1) were observed clearly in graphene,
and a weak peak for D band (,1350 cmꢀ1) was visible that is
attributed to edge effects.[30] However, the signal intensities of
the D peak are obviously different from that of the RGO reported