Abstract

The desalination and purification of sea or brackish water by utilizing solar energy are considered to be the most feasible solutions to overcome the problems of water shortage and pollution. In this study, a bifunctional Cu2-xSe-decorated hierarchical TiO2 nanotube mesh (CTNM) was designed and synthesized successfully for both solar water evaporation and photodegradation. Cu2-xSe enhances solar light absorption and solar water evaporation performance as a low-cost absorber because of its localized surface plasmon resonance (LSPR) effect. Meanwhile, the formation of the p-Cu2-xSe/n-TiO2 heterojunction improves the photodegradation performance by increasing separation and transport of photogenerated charge carriers. Hence, CTNM has a relatively high solar water evaporation conversion efficiency of 83.06% and also can photodegrade 95% of methyl orange after 3 h under 2.5 kW m−2 simulated solar irradiation, which demonstrate the extremely high utilization ratio of solar energy of CTNM.

INTRODUCTION

Water is the foundation of human origin and continuation. Rapid industrial growth and the worldwide population explosion have brought about an extremely increased demand for fresh water. Moreover, the contamination of rivers and lakes by industrial wastes and the large amounts of sewage discharged aggravate the serious global problem of water scarcity (Kalogirou 2005).

Many efforts have been made to search for effective solutions to the problem of water shortage, and the desalination of sea or brackish water is the most feasible solution to overcome the problem because the only nearly inexhaustible water sources are the oceans. The traditional water desalination technologies create a series of problems such as energy consumption and environmental pollution caused by utilizing fossil fuels (Elminshawy et al. 2015). Therefore, nanomaterials that utilize solar energy to evaporate water become significant for clean water generation (Zhou et al. 2016; Ye et al. 2017; Guo et al. 2018; Ren & Yang 2018).

As is well known, some contaminants in seawater such as alcohol, aldehyde, ketone, phenol and dyestuff that can also form azeotropic mixtures with water can be evaporated during water evaporation (Janakey et al. 2018), which can be finally converted to CO2 and H2O by photodegradation. So, if the photothermal material possesses both solar water evaporation and photodegradation performance, the processes of water desalination and purification can obviously be simplified. Bifunctional membranes possessing both TiO2 nanoparticle-based photocatalytic function and Au nanoparticle-based solar water evaporation have been reported (Liu et al. 2016; Huang et al. 2017), but the noble metal nanostructures are unstable under long periods of irradiation and heating with high cost. In this paper, Cu2-xSe is chosen to decorate TiO2 nanotube mesh (TNTM) for photothermal conversion and photocatalytic degradation because of excellent solar absorptivity, good photo-stability, low cytotoxicity and low production cost (Dorfs et al. 2011; Hessel et al. 2011). The p–n heterojunction between p-Cu2-xSe (Eg = 1.35 eV) and n-TiO2 nanotube (NT) favors the fast separation and transport of photogenerated charge carriers along the tube wall, which improves the photodegradation performance of the hybrid material (Ratanatawanate et al. 2009; Zhou et al. 2011; Han et al. 2015). In addition, TNTM with many radial NTs around Ti wires also positively influences the performance because of its large specific surface area, flexibility and light absorption independent of the direction of the solar light (Yang & Chen 2016). Hence, the CTNM exhibits a relatively high solar water evaporation conversion efficiency of 83.06% and an excellent methyl orange (MO) photodegradation efficiency of 95% in 3 h under 2.5 kW m−2 simulated solar irradiation.

METHODS

Preparation of TNTM

A large piece of raw Ti mesh was cut into square pieces of 2.5 × 2.5 cm2, which were ultrasonically degreased in acetone, isopropanol and methanol for 15 min respectively, and then chemically etched in a mixture of HF and HNO3 aqueous solution (HF:HNO3:H2O = 1:4:10 by volume) for 10 s, afterwards rinsed with deionized water and finally dried in air. Electrochemical anodic oxidation was performed at 60 V direct current voltage for 24 h in diethylene glycol solution containing 1.5 vol% HF, using Ti mesh as the working electrode and Pt plate as counter-electrode. The as-prepared samples were ultrasonically rinsed with deionized water and dried in air. To convert samples from amorphous phase to anatase phase, thermal treatment was performed in air at 450 °C for 3 h.

Preparation of CTNM

All experiments were performed with a LK98BΠ electrochemical workstation using the TNTM as the working electrode, platinum plate as counter electrode and saturated calomel electrode as reference electrode, respectively. Cu3Se2 was electrochemically deposited into TiO2 NT channels in 30 mL fresh electrolyte containing 0.1 M CuSO4·5H2O and 0.05 M H2SeO3 aqueous solution. Pulse electro-deposition was adopted with a pulse voltage of −0.45 V, a pulse time of 1 s and an interval time of 4 s, which permitted depleted ion equilibrium in the interval time. Cu3Se2-decorated TNTM with 10, 30 and 50 electrochemical cycles was taken out from the electrolyte and rinsed with deionized water, then dried in air, respectively. To convert Cu3Se2 into Cu2-xSe, thermal treatment was performed in N2 at 300 °C for 3 h.

Solar water evaporation

The solar water evaporation performances of all samples were tested in a nested cubic container with a test room of a surface area of 2.7 × 2.7 cm2 and a depth of 3 cm. The container was placed on an electronic balance to measure the weight of evaporated water. For each run, the container was filled with deionized water, and the sample on supporting glass-fiber cotton floated on the water surface. A 500 W xenon lamp was used to simulate solar irradiation and the aperture diameter was adjusted to be the same on the sample each time. After certain time intervals, the weight of water in the container was recorded. A K-type thermocouple was used to measure the temperature of the surface water and the temperatures were recorded after certain time intervals (Figure S1, Supplementary Material, available with the online version of this paper).

Photodegradation activity test

Photodegradation activity of the samples was evaluated by decomposing 10−5M MO solution under the 500 W Xe lamp. Before photodegradation, MO adsorption equilibrium on the sample surface was established by mechanical stirring in the dark for 30 min. The MO concentration was recorded at every simulated solar light irradiation time interval of 60 min at 462 nm by UV1700 UV-Vis spectrophotometer.

RESULTS AND DISCUSSION

The X-ray diffraction (XRD) patterns in Figure 1(a) indicate that tetragonal Cu3Se2 (JCPDS card No. 47-1745) is successfully deposited by the electrochemical deposition method into TNTM prepared through the electrochemical anodic oxidation method (Gong et al. 2001) and converted into Cu2-xSe after thermal treatment because six new peaks appear at 26.75°, 44.60°, 52.91°, 64.98°, 71.59° and 82.24°, which can be assigned to the (111), (220), (311), (400), (331) and (422) crystal faces of Cu2-xSe (JCPDS card No. 06-0680). The samples with 10, 30 and 50 pulse cycles were denoted as O-10, O-30 and O-50, and S-10, S-30 and S-50 after thermal treatment, respectively.

Figure 1

(a) XRD patterns of samples TNTM (curve a), O-10 (curve b), O-30 (curve c), O-50 (curve d), S-10 (curve e), S-30 (curve f) and S-50 (curve g); (b) XPS survey spectrum of S-30; (c) experimental and fitted Cu 2p XPS spectrum of S-30; (d) UV-Vis-NIR absorption spectra of samples TNTM, S-10, S-30 and S-50.

Figure 1

(a) XRD patterns of samples TNTM (curve a), O-10 (curve b), O-30 (curve c), O-50 (curve d), S-10 (curve e), S-30 (curve f) and S-50 (curve g); (b) XPS survey spectrum of S-30; (c) experimental and fitted Cu 2p XPS spectrum of S-30; (d) UV-Vis-NIR absorption spectra of samples TNTM, S-10, S-30 and S-50.

The electrochemical reaction is described by Equation (1):  
formula
(1)
The phase transformation during thermal treatment can be depicted as:  
formula
(2)

The Cu2-xSe phase content increases with pulse cycles based on the intensity of the Cu2-xSe peaks. According to the Scherrer equation, the calculated Cu2-xSe grain size is 14 nm, 21 nm, and 39 nm for samples S-10, S-30 and S-50, respectively. Figure 1(b) and 1(c) show the X-ray photoelectron spectroscopy (XPS) spectra of the sample S-30. From the XPS survey spectrum in Figure 1(b) it turns out that the sample contains Cu, Ti, O and Se elements. The high resolution XPS spectrum of Cu in Figure 1(c) indicates the existence of Cu2+ ions in sample S-30 because of the partial oxidization of Cu+ ions at the surface.

The ultraviolet-visible-near infrared (UV-Vis-NIR) absorption spectra of TNTM, S-10, S-30 and S-50 (Figure 1(d)) indicate that TNTM mainly absorbs ultraviolet light with an absorption edge at about 387 nm corresponding to the band gap of anatase TiO2 (about 3.2 eV). The wide absorption around 1,000 nm can be attributed to oxygen vacancies and Ti3+ ions (Kagan et al. 1999). The decoration of Cu2-xSe extends the absorption range of TiO2 from 400 nm to 2,000 nm. Compared with S-10 with insufficient Cu2-xSe and S-50 with too much Cu2-xSe, S-30 has the highest absorbance due to the appropriate Cu2-xSe quantity that not only enhances light absorption but also does not obstruct the light absorption of the NT structure, which can effectively trap light in extremely long and narrow inter-NT channels such that the light can hardly escape. In addition, the localized surface plasmon resonance (LSPR) peaks at about 1,700 nm can be observed in spectra of S-10, S-30 and S-50, which have a slight red shift and widen with the decrease of grain size and nanoparticle size (Smithard 1973; Ganière et al. 1975).

Figure 2 shows scanning electron microscope (SEM) images and energy dispersive spectroscopy (EDS) spectra of TNTM, S-10, S-30 and S-50. Figure 2(a) indicates that vertically oriented TiO2 NTs separate from each other with a tube diameter of about 165 nm and a wall thickness of about 43 nm. The sideview image in Figure 2(b) indicates that the tube length is about 3.5 μm. Figure 2(d) and 2(f) indicate that a few Cu2-xSe are deposited on the tube openings and walls for ten cycles. After 30 cycles, the NT surface becomes extensively coated with relatively uniform Cu2-xSe (Figure 2(g) and 2(h)). With further increase of deposition to 50 cycles, obvious agglomerations of Cu2-xSe are observed on the top surface of the TiO2 NTs which partially block the NT channels (Figure 2(j) and 2(k)).

Figure 2

SEM images and EDS spectra of TNTM (a)–(c), S-10 (d)–(f), S-30 (g)–(i) and S-50 (j)–(l); the inset shows the optical image of TNTM.

Figure 2

SEM images and EDS spectra of TNTM (a)–(c), S-10 (d)–(f), S-30 (g)–(i) and S-50 (j)–(l); the inset shows the optical image of TNTM.

Figure 3 depicts the surface water temperature and evaporated water mass changes when 2.5 kW m−2 simulated solar light irradiates the samples floating on the surface of the deionized water over time. Cu2-xSe was deposited on Ti mesh for 30 cycles as a control sample denoted as TCS. The surface water temperature and the evaporated water mass increase with the irradiation time. Especially, S-30 has the highest final surface water temperature (57.7 °C), which is obtained by fitting the temperature values in the last 1 h and the largest evaporated water mass (16.141 kg m−2). Compared with S-30, S-10 and S-50 have lower surface water temperature and evaporated water mass values due to the low utilization efficiency for solar light. The sample TNTM and the sample TCS have similar surface water temperature and evaporated water mass values.

Figure 3

(a) Surface water temperature and (b) evaporated water mass changes under 2.5 kW m−2 simulated solar irradiation over time with water itself, and samples TNTM, TCS, S-10, S-30, and S-50.

Figure 3

(a) Surface water temperature and (b) evaporated water mass changes under 2.5 kW m−2 simulated solar irradiation over time with water itself, and samples TNTM, TCS, S-10, S-30, and S-50.

The solar water evaporation conversion efficiency (η) is an important index to evaluate a solar photothermal conversion material for water evaporation. Table S1 (Supplementary Material, available with the online version of this paper) summarizes the calculated η values indicating that S-30 possesses the highest η of 83.06%, and S-10 and S-50 have a little lower η values of 73.15% and 62.54%, respectively. TNTM has the lowest η of 46.14%. In addition, TCS has an η value of 50.37%, a little higher than TNTM.

Figure 4(a) shows the degradation efficiencies of samples TNTM, TCS, S-10, S-30 and S-50 under 2.5 kW m−2 simulated solar light irradiation. For comparison purposes, the autodecomposition of the MO aqueous solution under the same simulated solar light irradiation was studied, and its final photodegradation rate was 3%. After 3 h of simulated solar irradiation, 95% of MO was decomposed by S-30, and only 85%, 60%, 74% and 6% of MO was decomposed by S-10, S-50, TNTM and TCS, respectively. As shown in Figure 4(b), S-30 exhibits the highest photodegradation rate constant (k), which is about 3.3 times as high as that of TNTM. Furthermore, its k value increases over time because the concentrating of heat by photothermal conversion gradually improves the rate of photodegradation reactions. S-10 and S-50 had higher k than TNTM, by about 2.1 times and 1.5 times, respectively.

Figure 4

(a) Photodegradation of MO and (b) a plot of ln(C0/C) versus irradiation time with MO solution itself, and samples TCS, TNTM, S-10, S-30, and S-50; (c) the energy band structure diagrams of Cu2-xSe and TiO2 before and after contact. (CB: conduction band, VB: valence band, NHE: normal hydrogen electrode.)

Figure 4

(a) Photodegradation of MO and (b) a plot of ln(C0/C) versus irradiation time with MO solution itself, and samples TCS, TNTM, S-10, S-30, and S-50; (c) the energy band structure diagrams of Cu2-xSe and TiO2 before and after contact. (CB: conduction band, VB: valence band, NHE: normal hydrogen electrode.)

According to the experimental and theoretical analysis, the energy band structure diagram of the p-Cu2-xSe/n-TiO2 heterojunction is expounded schematically in Figure 4(c). It is well known that the photocatalytic activities of photocatalysts mainly depend on the separation and transport of photogenerated charge carriers. When the p-Cu2-xSe and n-TiO2 are combined to form the p–n heterojunction, an interfacial electric field is built in the interface between Cu2-xSe and TiO2, which leads to the upward band bending of n-TiO2 and downward band bending of p-Cu2-xSe (Zhang & Yates 2012; Han et al. 2015).

Under simulated solar light irradiation, Cu2-xSe acting as a photosensitizer can be easily activated by solar light and generated electrons and holes. The photogenerated electron–hole pairs are separated effectively in the p-Cu2-xSe/n-TiO2 heterojunction interface. The efficient charge separation can increase the lifetime of the charge carriers and give them enough time to react with the reactants adsorbed onto the photocatalyst surfaces so as to improve the photodegradation activity. In such a case, the photogenerated electrons can reduce the oxygen to generate the active species ·O2, which can further lead to oxidation of MO. The holes would be readily scavenged by H2O or OH, leading to ·OH radicals, and accelerating the MO degradation.

Figure S2 (Supplementary Material, available online) shows the schematic diagram of solar water evaporation and photodegradation processes for sample CTNM. When the simulated solar light irradiates samples, light is effectively trapped in inter-NT channels. Moreover, the radial NTs around Ti wires can absorb light from different directions so that the utilization efficiency of solar light is further improved. Meanwhile, CTNM absorbs photon energy by the two mechanisms of LSPR (Herzog et al. 2014) and electron transitions (intraband transition and interband transition) (Yang et al. 2009) that make CTNM sufficiently utilize the light in the solar spectrum. The photogenerated hot electrons on the CB of the TiO2 and Cu2-xSe participate in the photodegradation reaction. The other hot electrons relax initially by electron–electron scattering, which transfer their energy to lattices and cool via inelastic electron–phonon collisions. Finally, phonon–phonon interactions result in a complete relaxation of the initially absorbed photon energy to make TiO2 NTs and Cu2-xSe become hot. Therefore, CTNM can almost completely convert the absorbed light into thermal and chemical energy. The photodegradation process is closely associated with evaporation. Photodegradation takes place in the Cu2-xSe-decorated TiO2 NT channels when accompanied by water evaporation. The concentration of Cu2-xSe can also be increased in TiO2 NT channels rather than the Cu2-xSe dispersing into the bulk solution to improve the efficiency of photodegradation (Huang et al. 2017). Thus, the TiO2 NT channels play a key role in concentrating Cu2-xSe for photodegradation, while also concentrating heat for water evaporation that can increase the rate of the photodegradation reaction (Clavero 2014). Water evaporation also makes MO solution be transported to the NT channels and degraded. Additionally, the NT channels are able to absorb the MO through the water evaporation process. The more water is evaporated, the more MO the NT channels can absorb, leading to a decrease in the concentration of MO.

Previously, the loss of energy efficiency in photocatalytic water purification has been the result of two processes: the absorption process with the relatively narrow absorption bands of the photocatalysts and photothermal conversion with the absorbed photons (Liu et al. 2016). The integration of the photothermal-driven solar water evaporation function with the photodegradation function helps expand the range of useful solar light for clean water generation and also takes advantage of the heat generated by the catalysts for further efficiency enhancement. The bifunctional CTNM enables the generation of both pure and partially purified water in a single step. Pure water is generated through the solar water evaporation process and partially purified water is the result of the photodegradation process.

CONCLUSION

In summary, CTNM was successfully synthesized by annealing Cu3Se2-decorated hierarchical TiO2 NT mesh, which was fabricated by electrochemical methods. The LSPR peak has a slight red shift and widens with the decrease of grain size and nanoparticle size resulting from the decrease of electrochemical deposition cycles. Moderate amounts of Cu2-xSe dramatically improve the light absorption, solar water evaporation and photodegradation performances. S-30 has a relatively high solar water evaporation conversion efficiency of 83.06% and also can photodegrade 95% of MO after 3 h under 2.5 kW m−2 simulated solar irradiation. The high solar water evaporation efficiency with the high photodegradation rate demonstrates that CTNM has an extremely high utilization ratio of solar energy. Hence, utilizing CTNM as an efficient and low-cost nanomaterial combining solar water evaporation and photodegradation functions for generating pure and purified water is a promising approach to address the dual issues of fresh water shortage and water pollution.

REFERENCES

REFERENCES
Dorfs
D.
,
Härtling
T.
,
Miszta
K.
,
Bigall
N. C.
,
Kim
M. R.
,
Genovese
A.
,
Falqui
A.
,
Povia
M.
,
Manna
L.
2011
Reversible tunability of the near-infrared valence band plasmon resonance in Cu2-xSe nanocrystals
.
J. Am. Chem. Soc.
133
(
29
),
11175
11180
.
Elminshawy
N. A. S.
,
Siddiqui
F. R.
,
Sultan
G. I.
2015
Development of a desalination system driven by solar energy and low grade waste heat
.
Energ. Convers. Manage.
103
,
28
35
.
Ganière
J.-D.
,
Rechsteiner
R.
,
Smithard
M.-A.
1975
On the size dependence of the optical absorption due to small metal particles
.
Solid State Commun.
16
(
1
),
113
115
.
Gong
D.
,
Grimes
C. A.
,
Varghese
O. K.
,
Hu
W.
,
Singh
R. S.
,
Chen
Z.
,
Dickey
E. C.
2001
Titanium oxide nanotube arrays prepared by anodic oxidation
.
J. Mater. Res.
16
(
12
),
3331
3334
.
Guo
Z.
,
Ming
X.
,
Wang
G.
,
Hou
B.
,
Liu
X.
,
Mei
T.
,
Li
J.
,
Wang
J.
,
Wang
X.
2018
Super-hydrophilic copper sulfide films as light absorbers for efficient solar steam generation under one sun illumination
.
Semicond. Sci. Technol.
33
(
2
),
025008
.
Herzog
J. B.
,
Knight
M. W.
,
Natelson
D.
2014
Thermoplasmonics: quantifying plasmonic heating in single nanowires
.
Nano Lett.
14
(
2
),
499
503
.
Hessel
C. M.
,
Pattani
V. P.
,
Rasch
M.
,
Panthani
M. G.
,
Koo
B.
,
Tunnell
J. W.
,
Korgel
B. A.
2011
Copper selenide nanocrystals for photothermal therapy
.
Nano Lett.
11
(
6
),
2560
2566
.
Janakey Devi
V. K. P.
,
Sai
P. S. T.
,
Balakrishnan
A. R.
2018
Ammonium-based ionic liquid as an entrainer for the separation of n-propanol + water and isopropanol + water mixtures
.
J. Chem. Eng. Data
63
(
3
),
498
507
.
Kalogirou
S. A.
2005
Seawater desalination using renewable energy sources
.
Prog. Energ. Combust.
31
(
3
),
242
281
.
Liu
Y.
,
Lou
J.
,
Ni
M.
,
Song
C.
,
Wu
J.
,
Dasgupta
N. P.
,
Tao
P.
,
Shang
W.
,
Deng
T.
2016
Bioinspired bifunctional membrane for efficient clean water generation
.
ACS Appl. Mater. Interfaces
8
(
1
),
772
779
.
Ratanatawanate
C.
,
Tao
Y.
,
Balkus
K. J.
2009
Photocatalytic activity of PbS quantum dot/TiO2 nanotube composites
.
J. Phys. Chem. C
113
(
24
),
10755
10760
.
Ye
M.
,
Jia
J.
,
Wu
Z.
,
Qian
C.
,
Chen
R.
,
O'Brien
P. G.
,
Sun
W.
,
Dong
Y.
,
Ozin
G. A.
2017
Synthesis of black TiOx nanoparticles by Mg reduction of TiO2 nanocrystals and their application for solar water evaporation
.
Adv. Energy Mater.
7
(
4
),
1601811
.
Zhou
L.
,
Tan
Y.
,
Ji
D.
,
Zhu
B.
,
Zhang
P.
,
Xu
J.
,
Gan
Q.
,
Yu
Z.
,
Zhu
J.
2016
Self-assembly of highly efficient, broadband plasmonic absorbers for solar steam generation
.
Sci. Adv.
2
(
4
),
e1501227
.

Supplementary data