Hybrid Cu x O-TiO2 Heterostructured Composites for Photocatalytic CO2 Reduction into Methane Using Solar Irradiation: Sunlight into Fuel.

Photocatalytic CO2 conversion to fuel offers an exciting prospect for solar energy storage and transportation thereof. Several photocatalysts have been employed for CO2 photoreduction; the challenge of realizing a low-cost, readily synthesized photocorrosion-stable photocatalytic material that absorbs and successfully utilizes a broad portion of the solar spectrum energy is as yet unmet. Herein, a mesoporous p-type/n-type heterojunction material, Cu x O-TiO2 (x = 1, 2), is synthesized via annealing of Cu/Cu2O nanocomposites mixed with a TiO2 precursor (TiCl4). Such an experimental approach in which two materials of diverse bandgaps are coupled provides a simultaneous opportunity for greater light absorption and rapid charge separation because of the intrinsic p-n heterojunction nature of the material. As detailed herein, this heterostructured photocatalyst demonstrates an improved photocatalytic activity. With the CO2 reduction of our optimal sample (augmented light absorption, efficacious charge separation, and mesoporosity) that utilizes no metal cocatalysts, a remarkable methane yield of 221.63 ppm·g-1·h-1 is achieved.

Introduction

The continuous increase in atmospheric CO2 concentration is considered to be a key driver that induces climate change.[1] The prospect of unwanted climate change or climate heating–desertification has compelled investigations into the means to normalize atmospheric CO2 concentrations. As is well known, CO2 is a highly stable molecule, requiring significant energy input for its reduction.[2] In this regard, sunlight is considered to be the most useful energy source for promoting CO2 conversion into useful hydrocarbon products, such as methane or ethane, offering the possibility of turning sunlight into fuels compatible with the current energy infrastructure.

During the past several decades, titanium dioxide (TiO2) has received much attention as a photocatalyst because of its abundant availability, nontoxicity, photocorrosion resistance, and excellent charge transport properties.[3] However, the TiO2 bandgap of 3.2 eV limits its absorption to the ultraviolet region, making it utilize only 4% of the entire solar spectrum. To reduce the TiO2 bandgap such that it absorbs a greater portion of the solar spectrum energy while maintaining its commendable properties, various strategies have been investigated, such as anion doping,[4] loading of noble metals like Pt, Pd, Au, and so forth to act as cocatalysts,[5,6] and coupling of TiO2 with low-bandgap semiconductors.[7] Composites of CuO and TiO2 have been utilized for hydrogen evolution,[8] CO2 photoreduction,[9] and photodegradation of volatile organic compounds[10] and as visible light-responding photocathodes.[11] Therefore, we believe that coupling TiO2 with a suitable low-bandgap semiconductor, our interest herein, can result in a broad-spectrum visible light-absorbing material with improved photocatalytic activities. The present hybrid photocatalyst based upon CuO–TiO2 showcases a performance better than those of the previously reported studies on CuO–TiO2 for CO2 photoreduction into methane.[9,12]

Herein, we report the synthesis and the photocatalytic application of a mesoporous p–n heterojunction composite material without metal cocatalyst sensitization, as commonly used. Our material is composed of p-type CuO (CuO and Cu2O) with a bandgap of ∼1.35–1.7 eV,[13] coupled with n-type TiO2. The CuO–TiO2 photocatalyst is prepared using a facile and inexpensive two-step process: briefly, Cu/Cu2O nanocomposites are synthesized via thermal decomposition, which are then vigorously mixed with titanium tetrachloride (TiCl4) under an argon atmosphere, forming a gelatinous solution. The calcination under air of this gelatinous solution results in a mesoporous CuO–TiO2 composite.

During the ambient-atmosphere calcination process, (1) TiCl4 oxidizes to form TiO2; (2) Cu/Cu2O nanocomposites are oxidized, forming CuO (CuO and Cu2O); and (3) organic ligands used in the synthesis of Cu/Cu2O nanocomposites are removed, providing a firm heterojunction formation between CuO and TiO2 with a well-defined mesoporous morphology. A schematic diagram showing our experimental approach is shown in Figure .

Figure 1

Schematic view of our experimental approach for the synthesis of the mesoporous CuO–TiO2 photocatalyst.

We test the resulting photocatalyst for its ability to promote the ambient-temperature photoconversion of CO2 and water vapor to hydrocarbons and to discover a high rate conversion of CO2 to, almost exclusively, methane without the use of noble metal codopants. For the purpose of optimization, various samples of CuO–TiO2 are prepared with various amounts of TiCl4 (TiO2 precursor), namely, CT03, CT05, CT07, and CT09 corresponding to 0.3, 0.5, 0.7, and 0.9 mL of TiCl4, respectively, mixed with 20 mL of Cu/Cu2O nanocomposites dispersed in toluene. To the best of our knowledge, for the first time, hybrid mesoporous p-type CuO (CuO and Cu2O) coupled with n-type TiO2 for photocatalytic conversion of CO2 into hydrocarbon fuels without using metal cocatalysts is introduced.

Results and Discussion

Characterization of the CuO–TiO2 Composites

The X-ray powder diffraction (XRD) patterns of pure CuO, pure TiO2 (synthesized from TiCl4), and CuO–TiO2 samples are shown in Figure .[10] The XRD patterns of the CuO–TiO2 samples mainly show, an intense peak at 2θ = 26.2°, corresponding to d101 of anatase TiO2, confirming the presence of anatase TiO2 as well as a relatively small amount of Cu/Cu2O nanocomposites in the CuO–TiO2 samples. This result, we assume, is likely due to the low crystallinity and high dispersion of CuO species on TiO2 surfaces.[8,10] From the inductively coupled plasma-atomic emission spectroscopy (ICP-AES) analysis of the representative sample CT07, 2 wt % Cu is determined, which corresponds to approximately 80% of the Cu content obtained using the data from the field emission scanning electron microscopy (FE-SEM) energy-dispersive spectroscopy (EDS) technique (Figure S5).

Figure 2

XRD patterns of pure CuO, pure TiO2 and CuO−TiO2 samples.

Transmission electron microscopy (TEM) and high-resolution TEM (HR-TEM) images of the Cu/Cu2O nanocomposites and a representative CuO–TiO2 sample (CT07) are shown in Figure . The as-synthesized Cu/Cu2O nanocomposites show excellent size uniformity (Figure A) with the selected area electron diffraction (SAED) pattern showing d111 planes for both Cu2O and Cu. The HR-TEM image (Figure B) shows respective lattice fringes for Cu2O and Cu metal, which are well-matched with the SAED pattern.[10] To check the mesoporosity of the CuO–TiO2 sample, FE-SEM and HR-TEM images of the sample CT07 can be seen in Figures S2 and S3, respectively. Furthermore, the STEM image of the sample CT07 (Figure C) confirms the readily discernible mesoporous nature of the CuO–TiO2 samples. The STEM elemental mapping of the CT07 sample further (Figure S4) confirms the presence of Cu and Ti. Figure D exhibits the HR-TEM image of the sample CT07 showing well-defined lattice fringes of 0.28 and 0.35 nm, corresponding to the d110 plane of CuO and the d101 plane of TiO2, respectively. The SAED pattern of an interface region (inset of Figure D) further confirms the formation of CuO and TiO2 regions within the CuO–TiO2 matrix with rings indexed to the d110 plane of CuO and the d101 plane of anatase TiO2.[10]

Figure 3

(A) TEM image and SAED pattern (inset) of Cu/Cu2O nanocomposites and (B) HR-TEM image of single Cu/Cu2O nanocomposites, (C) STEM image of the sample CT07, by which the mesoporous nature of the photocatalyst can be clearly discerned, and (D) HR-TEM and SAED patterns (inset) confirming the presence of CuO and TiO2 within the mesoporous CuO−TiO2 composite.

The UV–vis diffuse reflectance spectra (UV–vis DRS) of pure TiO2 and CuO–TiO2 samples (Figure A) exhibit two prominent changes: (1) pure TiO2 is unable to absorb visible light, whereas the UV−vis DRS spectra of the modified samples exhibit a shift in the absorption edge to longer wavelengths (400–600 nm), which is attributed to the formation of a heterojunction between CuO and TiO2;[14,15] and (2) the absorption band in the near infrared (600–1000 nm) is observed because of the Cu loading and thus is attributed to a 2Eg → 2T2g interband transition in the CuII clusters deposited over TiO2.[16]

Figure 4

(A) UV–vis DRS of all samples showing red shift in the CuO–TiO2 samples attributed to the formation of a CuO–TiO2 heterojunction and absorption in the 600–1000 nm range because of CuO intrinsic absorptions. (B) Corresponding Tauc plots for the bandgap estimation of the CuO–TiO2 samples.

The bandgap values of all samples are estimated using Tauc plots (Figure B) and are listed in Table . The variation in the bandgap can be attributed to the amount of Cu/Cu2O present. The bandgap of the composites gradually increases with an increase in the TiCl4 content; the lowest value was 3.09 eV for sample CT03 containing 0.3 mL of TiCl4 mixed in a 20 mL Cu/Cu2O solution.

Table 1

Photocatalytic and Textural Properties of the CuO–TiO2 Samples

sample	SBETa (m2·g–1)	pore volumeb (cm3·g–1)	bandgapc (eV)	amount of CH4d (ppm·g–1·h–1)
CT03	11.05	0.115	3.09	23.33
CT05	16.49	0.102	3.12	55.07
CT07	26.95	0.094	3.15	221.63
CT09	22.95	0.164	3.2	168.03

The surface areas of the samples are determined by using the BET equation to a relative pressure (P/P0) range of 0.05−0.35 of the adsorption isotherm.

Barrett–Joyner–Halenda (BJH) equation using the desorption isotherm is used to calculate the pore volume.

Tauc plots are made for bandgap energy estimation.

Amount of CH4 evolved is calculated using eq .

The photoluminescence (PL) spectra of TiO2 (synthesized from the TiCl4 precursor), Cu/Cu2O nanocomposites, and the sample CT07 can be seen in Figure . Pure TiO2 depicts a sharp peak around 385 nm because of emission from band-to-band recombination with other peaks appearing in the range of 400–600 nm, attributed to the electron transitions from the inter-bandgap defect levels.[17,18] Cu/Cu2O nanocomposites exhibit a UV emission peak at 380 nm and a visible emission peak at 520 nm, attributed to the recombination of electron–hole pairs in free excitons or deep-level defects and photogenerated electrons in CuO, respectively.[19,20] For the sample CT07, the visible emission peak for CuO at 520 nm is quenched as compared to that of Cu/Cu2O nanocomposites, thereby suggesting the formation of a p–n junction at the interface, resulting in a reduced rate of recombination.

Figure 5

PL emission spectra of pure TiO2, Cu/Cu2O nanocomposites, and the sample CT07.

The Brunauer–Emmett–Teller (BET) surface areas and pore-size distributions are measured for all CuO–TiO2 samples (Figure S6), with their values displayed in Table . CT07 shows the largest surface area (26.95 m2·g–1), approximately 2.4 times higher than that of CT03 (11.05 m2·g–1). It is observed that on increasing the amount of TiCl4, the surface area decreases, possibly because of the increased aggregation induced by the higher amounts of the Ti precursor.

Further evidence for CuO–TiO2 formation comes from an X-ray photoelectron spectroscopy (XPS) analysis of the sample CT07. Figure A shows the Ti 2p region with two intense peaks at 457.6 and 464.0 eV, corresponding to Ti 2p3/2 and Ti 2p1/2 states, respectively, ensuring the presence of Ti4+ ions.[21] The satellite shoulder peaks appearing at higher binding energies are assumed to be because of the Ti4+ state in the Ti–O–Cu structure.[22] The Cu 2p region (Figure B) exhibits Cu 2p3/2 and Cu 2p1/2 main peaks appearing around 933.0 and 953.0 eV, with satellite peaks at 942.5 and 962.5 eV, respectively. The main Cu 2p peaks are further deconvoluted into four peaks, where peak 1 and peak 3 are assigned to the Cu+ species (Cu2O), whereas peak 2 and peak 4 are associated with the Cu2+ species (CuO).[14,23] The satellite peaks are attributed to the ligand-to-metal charge transfer, an important indicator of the presence of Cu2+ species[24] by an open 3d shell,[9] which is not observed for Cu+ or metallic Cu0 species because of their completely filled 3d shell.[10] Furthermore, the energy gap between peak 2 and peak 4 is 20.0 eV, which matches well with the standard value of 20.0 eV for CuO.[23] The Cu 2p XPS data suggest the presence of two phases, that is, CuO and Cu2O in the as-prepared samples. The O 1s region (Figure C) reveals the presence of three peaks located at 529.7, 531.2, and 532.3 eV corresponding to Ti–O, O–H, and C–O bonds, respectively.[8] The XPS spectra of all other synthesized samples, namely, CT03, CT05, and CT09, are given in Figures S7, S8, and S9, respectively. All samples show characteristic peak positions similar to that of the representative sample CT07, confirming the successful formation of the CuO–TiO2 heterojunction.

Figure 6

XPS of the sample CT07 showing (A) the Ti 2p region with characteristic Ti peaks assigned to Ti 2p3/2 and Ti 2p1/2, (B) characteristic Cu 2p peaks and satellite peaks assuring the presence of Cu2O and CuO, and (C) the O 1s region exhibiting three peaks located at 529.7, 531.2, and 532.3 eV corresponding to Ti–O, O–H, and C–O bonds, respectively.

Photocatalytic Conversion of CO2 into Methane and Its Proposed Mechanism

Photocatalytic CO2 conversion into hydrocarbon fuels is used to test the CuO–TiO2 sample. Pure TiO2 (synthesized from the TiCl4 precursor) and pure bulk CuO are used as reference samples. Analysis of the products obtained from all samples using gas chromatography (GC) predominantly shows methane as the main hydrocarbon product. Figure shows that the methane production rate after each sample is illuminated for 1 h of solar illumination. When CuO is coupled with TiO2, all samples show an increase in the CH4 yield. With an increasing amount of TiCl4 for the synthesis of CuO–TiO2, the amount of CH4 evolution increases, reaching a maximum value with CT07. A further increase in TiCl4 beyond CT07 decreases the CH4 yield, which we believe is due to the low surface area of CT09 limiting active sites to interact with the CO2 molecules. CT07 exhibits a methane evolution of 221.63 ppm·g−1·h−1, a value which is 11.1 times and 22 times higher than for pure CuO (20.01 ppm·g−1·h−1) and TiO2 (9.94 ppm·g−1·h−1), respectively. It is noted that this yield is better than those of our previously reported photocatalysts utilized for CO2 conversion into methane.[9,25−28]

Figure 7

Rates of CH4 evolution measured under simulated solar irradiation for pure TiO2, pure CuO, and all CuO−TiO2 samples. The sample CT07 shows the highest methane evolution rate of 221.63 ppm·g−1·h−1. The control test of the CT07 sample in an Ar/H2O(g) mixture exhibits negligible CH4 evolution.

The rate of CH4 production is calculated for all CuO–TiO2 samples using eq , as listed in Table . Control tests performed by illuminating CT07 in an Ar/H2O(g) atmosphere under similar irradiation conditions show a negligible amount of CH4 evolution. Thus, it can be inferred from the control experiment that CH4 evolved during the normal experiments is due to the photoreduction of CO2 and not because of the oxidation of surface-bound organics.

The stability of the representative sample CT07 is measured by its repeated testing for five cycles (Figure S11). The representative sample CT07 shows good stability without any acute decrease in the methane production rate: hence, one can see that the rate of CO2 reduction on the fifth test is approximately 88% that of the first.

An elucidation based upon experimental results and on the literature that reported the energy levels and the suggested process for the conversion of CO2 and water vapor into methane is shown in Figure . As revealed by XPS data, CuO–TiO2 contains both Cu2O and CuO, the conduction band edges of which are both more negative than TiO2.[29−31] Thus, upon illumination within the CuO–TiO2 photocatalyst, generation of electrons (e–) and holes (h+) takes place, whereby the electrons in p-type CuO can easily flow to the conduction band of TiO2 to contribute to the photoreduction of adsorbed CO2, whereas the photogenerated holes migrate in the opposite direction to oxidize the adsorbed H2O, releasing H+ and O2.

Figure 8

Schematic illustration of the photocatalytic reduction of CO2 into CH4 using CuO–TiO2 heterojunction samples.

The photoreduction of CO2 is a complex process. Initially, the adsorption of CO2 takes place over the semiconductor surface, leading to an activation of the CO2 molecule for reduction. Although CO2 is a linear molecule, its adsorption on a photocatalyst surface transforms it into a bent structure, with a decrease in the lowest unoccupied molecular orbital (LUMO) level of CO2, thus offering a lower barrier for accepting electrons under illumination.[32] When light is illuminated upon the CO2-adsorbed photocatalytic material, the photoexcited electrons generated are injected to the adsorbed CO2 to proceed with the reduction reactions, with the formation of various intermediate free radicals and products.[33] Among the various CO2 photoreduction mechanisms proposed, the carbene pathway is the most widely accepted pathway for yielding CH4 and/or CH3OH as the main products.[33] The literature for the carbene pathway mechanism is well-defined and is considered reliable based on the investigations made using electron spin resonance (ESR) and electron paramagnetic resonance (EPR) experimental techniques.[34−36] The carbene pathway begins with the injection of a single electron into the adsorbed CO2, forming an anion radical CO2•–.[37] Such a single electron reduction of CO2 to an anion radical CO2•– possesses a strong negative electrochemical potential of −1.9 V versus a normal hydrogen electrode (NHE).[36] Hence, with such a high potential required for this step, it seems highly improbable for the semiconductors to proceed with the reduction. Therefore, at this stage, it is considered that as soon as the CO2•– radical is formed, it reacts with the protons H+ (provided by water oxidation via filling holes) and photogenerated electrons to produce intermediate radicals and products. Such a process is known as “proton-assisted multielectron reduction” and is generally acceptable for the CO2 photoreduction process. The radicals and products produced at the intermediate stages further undergo a series of proton-assisted multielectron reductions, finally yielding CH4 as the main product. Thus considering the proton-assisted multielectron reduction via the carbene pathway, we propose a possible route for CO2 photoreduction, a schematic view of which is shown in Figure . Photogenerated electrons (e–) and holes (h+) are generated at the active sites of the CuO–TiO2 photocatalyst (eq ). The holes (h+) react with the adsorbed H2O to produce hydroxyl radicals (OH•) and protons (H+) (eq ). The surface-adsorbed CO2•– radical generated by the injection of a single electron (eq ) reacts with e– and H+, producing CO (eq ), which undergoes a further reduction process forming the surface-adsorbed C (eq ). This surface-adsorbed C reacts with 4e– and 4H+ to yield CH4 (eq ) as a main product. The proposed reactions involved in the CO2 photoreduction to CH4 can be described by eqs –6.

Experimental Section

Synthesis of Mesoporous CuO–TiO2 Heterostructured Composites

The synthesis of mesoporous CuO–TiO2 heterostructured composites was carried out using a simple impregnation step. An already prepared Cu/Cu2O nanocomposite solution (6 mL) (details in Supporting Information) was dispersed in 20 mL of anhydrous toluene in a rubber-capped vial and degassed for 30 min under vacuum to remove any air dissolved in the solution and filled with an inert gas (Ar). The degassed solution was transferred into the glove box, followed by a dropwise addition of a certain amount of 1 M titanium(IV) chloride solution (0.3, 0.5, 0.7, and 0.9 mL of TiCl4) to the degassed solution. The color of the solution changes from dark green to dark yellow. After 1 h of reaction, a well-dispersed solution mixture of dark yellow Cu/Cu2O nanocomposite and TiCl4 was formed and is allowed to oxidize spontaneously in air for 30 min under stirring. A change in the color of the mixture from dark yellow to reddish brown is observed. The gel-type mixture of Cu/Cu2O nanocomposite and TiCl4 was dried under air at 70 °C and then subsequently calcined in a tubular furnace at 400 °C, at a ramping rate of 6 °C/min under air flow (20 cc/min) for 3 h. The annealing process oxidizes both the Cu/Cu2O nanocomposite and TiCl4 forming CuO–TiO2 (where x = 1 or 2), a well-known oxidation process that has been reported earlier.[38−40] The CuO–TiO2 samples obtained with 0.3, 0.5, 0.7, and 0.9 mL of TiCl4 were labeled as CT03, CT05, CT07, and CT09, respectively.[41]

Photocatalyst Characterization

XRD studies were performed using a Panalytical, Empyrean diffractometer with Cu Kα radiation (λ = 1.54 Å) in the range of 2θ = 10°–90° at 1°/min. Surface morphologies and composition were observed using a field emission scanning electron microscope (Hitachi S-4800) equipped with an EDS attachment. High-resolution images were obtained using a field emission transmission electron microscope (FE-TEM, Hitachi HF-3300) operating at 300 kV, where the samples were prepared on a Ni grid.

The surface areas of the products were analyzed using the BET method (Micromeritics ASAP 2000 apparatus) at −196 °C. XPS (Thermo VG, K-alpha) with Al Kα line operating at 148 606 eV as the X-ray source was used to study the surface composition and oxidation states of CuO–TiO2. The optical properties of the samples were studied using UV–vis DRS Cary series (Agilent Technologies) with an attached diffuse reflectance accessory. PL was measured using a Cary Eclipse fluorescence spectrophotometer (Agilent Technologies), λexc = 300 nm for all samples. The copper content in the photocatalyst was deduced using the ICP-AES analysis using a Thermo Scientific iCAP 7400 duo ICP-AES instrument.

Photocatalytic CO2 Conversion

In the photocatalytic CO2 conversion experiment, an empty photoreactor (stainless steel; volume = 15.4 cm3) was purged with CO2 gas (1000 ppm in He) and vacuum simultaneously to remove any air or other impurities before and after the loading of the photocatalyst.[41] The photocatalyst (50 mg) was loaded into the photoreactor; moist CO2 gas (1000 ppm in He) was passed through a water bubbler, which then enters the photoreactor. The photocatalyst-loaded photoreactor, filled with a mixture of CO2 and H2O vapors mixture, was then illuminated by a 100 W Xenon solar simulator (Oriel, LCS-100) with an AM1.5 filter for 1 h, and the reaction products (500 μL) were analyzed using a Shimadzu GC-2014 gas chromatograph (Restek Rt-Q Bond column, ID = 0.53 mm, and length = 30 m) equipped with flame ionization (FID) and thermal conductivity (TCD) detectors. Figure S1 shows the schematic of the experimental setup for the photocatalytic CO2 reduction. The hour-normalized photocatalytic CH4 evolution rate is calculated using eq .Five cycles of CO2 photoreduction were performed to test the stability of the same sample; after every test, the photoreactor was purged with Ar gas and vacuum, then re-filled with CO2 gas (1000 ppm in He), followed by a 1 h illumination for the next testing cycle.

Conclusions

In summary, hybrid CuO–TiO2 photocatalysts are prepared via a facile experimental approach comprising two steps, that is, synthesis of Cu/Cu2O nanocomposites followed by mixing with TiCl4 and subsequent oxidation. The as-prepared samples are characterized using analytical techniques including XRD, TEM, UV–vis DRS, PL, BET, and XPS. A red shift in the light absorption is observed for the CuO–TiO2 samples, mainly attributed to the formation of nanoscale heterojunctions between CuO and anatase TiO2, providing a better charge separation and an increase in the optical absorption. Among the CuO–TiO2 samples, the sample CT07 produces the highest CH4 yield with production rates, 11.1 and 22 times higher than pure CuO and TiO2, respectively. The improved photocatalytic activity can be attributed to (1) the improved light absorption with a significant red shift in the absorption wavelength; (2) formation of p–n heterojunctions with suitable band edge positions for the improved separation of the photogenerated charge; and (3) large surface areas to promote interfacial reactions. A further increase in the TiCl4 amount, sample CT09, resulted in a decrease in the CH4 evolution rate, which we believe is due to a reduced surface area and a wider bandgap.

Our material synthesis strategy of hybrid CuO–TiO2 recommends coupling of low-bandgap materials with large-bandgap materials as an efficient approach for the design of high-performance photocatalysts.