Step-by-Step Growth of HKUST-1 on Functionalized TiO₂ Surface: An Efficient Material for CO₂ Capture and Solar Photoreduction

Abstract

The present study reports on a simple preparation strategy of a hybrid catalyst, TiO₂/HKUST-1, containing TiO₂ anatase nanoparticles (NPs) with tailored morphology and photocatalytic activity coupled with a porous metal-organic framework (MOF), namely HKUST-1, as an advanced material for the CO₂ photocatalytic reduction. In detail, TiO₂/HKUST-1 catalyst was prepared via an easy slow-diffusion method combined with a step-by-step self-assembly at room temperature. The growth of crystalline HKUST-1 onto titania surface was achieved by functionalizing TiO₂ nanocrystals, with phosphoesanoic acid (PHA), namely TiO₂-PHA, which provides an intimate contact between MOF and TiO₂. The presence of a crystalline and porous shell of HKUST-1 on the TiO₂ surfaces was assessed by a combination of analytical and spectroscopic techniques. TiO₂/HKUST-1 nanocomposite showed a significant efficiency in reducing CO₂ to CH₄ under solar light irradiation, much higher than those of the single components. The role of MOF to improve the photoreduction process under visible light was evidenced and attributed either to the relevant amount of CO₂ captured into the HKUST-1 porous architecture or to the hybrid structure of the material, which affords enhanced visible light absorption and allows an effective electron injection from TiO₂-PHA to HKUST-1, responsible for the photochemical reduction of CO₂.

1. Introduction

The significant increase in the CO₂ level in the past several decades is a matter of great concern [1,2]. While there are considerable investments in developing methods to reduce CO₂ emissions, it is apparent that its atmospheric concentration will continue to monotonically increase for the foreseeable future, due to fossil fuel and industry consumption, land-use changes, such as deforestation and fires, and so on [3,4,5]. Suggestions have been made to sequester carbon dioxide, such as sorption into new functional materials [6]. However, the possibility to recycle CO₂ via conversion into high-energy content fuels appears an attractive option, but the process is energy demanding and useful only if a renewable energy source can be used. A possible sustainable approach lies in the TiO₂-assisted photocatalytic reduction of CO₂ to renewable fuels (i.e., CO, CH₃OH, HCOOH, CH₄) using solar light as photon source [7,8,9].

Although TiO₂ has been demonstrated to be a promising photocatalyst for CO₂ reduction [10,11,12,13], its activity is limited owing to the poor adsorption capability, the low quantum yield and solar energy conversion efficiency. Hence, the combination of TiO₂ with materials with remarkable adsorption capacity, like zeolites or other mesoporous materials [14,15,16] has been found to be a promising method for enhancing its photocatalytic performance. Besides, many efforts have been devoted to extending the light absorption of TiO₂ to the visible region, but the activation of the TiO₂ photocatalytic response in the visible region still remains a challenge.

To overcome these issues, several recent studies support the idea of using Metal Organic Frameworks (MOF) as porous coating of TiO₂ [17]. MOFs materials, made by assembling metal ions with organic ligands as linkers, possess exceptionally high pore volume and surface area, due to uniform and continuous structural cavities, and different active sites on the pores surface [18,19]. These characteristics depend both on the kind of metal and linker, and provide cage-like cavities with tunable properties, such as charge, polarity, redox potential, hydrophobicity/hydrophilicity, and aromatic/lipophilic character [20,21]. In addition, it is important to point out that MOFs could act also as a photosensitizer of TiO₂, extending its absorption toward the visible range and inhibiting the electron-hole recombination process, thanks to the charge transfer process between semiconductor and MOF [22,23,24,25].

In the last decade, some examples of TiO₂ hybrid systems containing Zr [26], Fe [27], Al [28], Cd [29], Cu [30], based MOFs have been reported to reduce CO₂ mainly to CO, formate and methane. In any case, the interfacial interaction between TiO₂ and MOFs was identified as the crucial point for the improved photocatalytic conversion [23,24].

Among MOFs, a copper-based one, [Cu₃(BTC)₂(H₂O)₃]_n, (BTC = benzene-1,3,5-tricarboxylate), namely HKUST-1 [31], has been recognized as an excellent material for CO₂ storage [32,33] due to the presence of unsaturated metal sites, available for interaction with CO₂, H₂ and other gases [34,35]. However, only a few TiO₂ materials containing HKUST-1 for CO₂ photoreduction have been reported so far [36,37]. Li et al. synthesized a HKUST/TiO₂ core/shell structure by coating a nanocrystalline TiO₂ shell onto the HKUST-1 core, with HKUST-1/TiO₂ ratio of 0.5 [36]. This MOF/semiconductor material was able to convert CO₂ into CH₄ under UV illumination with improved performance compared to bare TiO₂. The authors concluded that in HKUST/TiO₂, UV photogenerated electrons can be effectively transferred from TiO₂ to the MOF, avoiding charge recombination, and, in turn, to the gas molecules adsorbed by the porous structure. However, the suggested electron transfer was correlated to the photocatalytic reduction performed exclusively under UV irradiation.

Recently, He and coworkers [37] reported the synthesis of TiO₂ embedded in the HKUST-1 matrix via a rapid aerosol route with HKUST-1/TiO₂ ratio of 3.33. This significantly improves the CO₂ photoreduction to CO, in comparison with pristine TiO₂. The enhanced performance was attributed only to the enhanced reactant adsorption on the catalyst due to MOF presence, even if no electron transfer processes were highlighted.

Despite the promising performance of the abovementioned hybrid TiO₂-MOF catalysts, several critical points should be still clarified in order to define the activation role of MOF in the CO₂ photoreduction, specifically under visible light.

In this context, we report a simple synthesis of porous hybrid catalyst, based on HKUST-1 anchored to anatase TiO₂ nanoparticles (NPs) with controlled rhombic elongated (RE) morphology, by means of a phosphonic acid as coupling agent. TiO₂ NPs were chosen for their photoactivity [38], while HKUST-1 for its ability in favoring high CO₂ adsorption as well as controlled release kinetics [39], so that MOF pores may behave like “nano-reactors”, in which the substrate is confined near to the TiO₂ surface.

In detail, we have prepared, by a cheap and easy to scale-up soft-chemistry method, TiO₂ nanocrystals having tailored structural and morphological features, with exposed {101} and {010} facets which are known to be active in photo-reduction reactions [40,41]. TiO₂ NPs were functionalized with an asymmetric organic linker, namely 6-phosphohexanoic acid (PHA), suitable to covalently interact with the oxide surface by Ti-O-P bond, as well as to selectively bind copper metal ions by carboxylate functionality. The growth of HKUST-1 on PHA-modified TiO₂ NPs (TiO₂-PHA) was obtained using a slow-diffusion method combined with a step-by-step self-assembly approach [42]. After complexation of Cu²⁺ metal ions by the carboxylic groups, the metal-carboxylate centers react with the BTC ligand activating the crystalline growth of the HKUST-1 structure on the oxide surface.

This synthetic strategy, based on the TiO₂ functionalization with an organic linker, greatly (i) facilitates NPs dispersion, preventing undesirable agglomeration phenomena, (ii) promotes the crystalline growth of MOF layers on TiO₂ surface and (iii) favors an intimate contact between TiO₂ and MOF, which makes easier, in principle, the electron transfer at the hybrid interface.

TiO₂/HKUST-1 nanocomposite was characterized by a combination of analytical and spectroscopic techniques. After determining the structural and morphological features, the CO₂ adsorption aptitude and the photocatalytic performances of the developed hybrid material were examined. CO₂ photoreduction tests were performed under sunlight, in ambient conditions and in a heterogeneous gas/solid set-up, in order to simulate the conditions under which CO₂ capture and fixation proceed in a single step, while most of the existent studies describe gas/liquid systems. Based on the photocatalytic results, a model for explaining the possible charge transfer between HKUST-1 and TiO₂-PHA in TiO₂/HKUST-1 have been proposed, in connection with the peculiar structure and morphology of hybrid material.

2. Results and Discussion

TiO₂/HKUST-1 catalyst was prepared via a simple slow-diffusion method combined with a step-by-step self-assembly at room temperature. Firstly, the solvothermal synthesis of shape-controlled anatase nanocrystals was performed according to a previously reported procedure [43] (Scheme 1, STEP 1) by reaction of the titanium (IV) butoxide (TB), in the presence of oleic acid (OA) and oleylamine (OM). After removing the residual amounts of capping agents by using tetramethylammonium hydroxide (TMAH) [44] (STEP 2), the pre-synthetized TiO₂ NPs have been functionalized with PHA (STEP 3), able to covalently interact with the oxide surface by the phosphonic group, as well as to bind copper metal ions by carboxylic acid, promoting the grown of HKUST-1.

After that, TiO₂/HKUST-1 was prepared by a step-by-step self-assembly at low temperature (40 °C) (Scheme 2). In detail, the copper centers of the copper acetate (AcCu(II)) can interact with the carboxyl groups of TiO₂-PHA and react with the carboxyl groups of H₃BTC ligands.

2.1. Spectroscopic and Morphological Characterization of TiO₂-PHA and TiO₂/HKUST-1

The characterization of TiO₂-PHA and TiO₂/HKUST-1 was performed in order to demonstrate: (i) the effective functionalization of TiO₂ NPs by PHA phosphonic groups and consequently (ii) the crystalline growth of the HKUST-1 structure on the oxide surface.

The crystal structure of the materials was checked by powder X-ray diffraction (PXRD) analysis (Figure 1). PXRD patterns of TiO₂/HKUST-1 sample (Figure 1a) showed the superimposed pattern of HKUST-1 (Figure 1c) [31] and anatase TiO₂ diffraction peaks [45] (Figure 1b), corresponding to (101), (004), (200), (105), and (211) reflections highlighted in the figure. In particular, the pattern indicates that HKUST-1 shells in TiO₂/HKUST-1 are highly crystalline, as confirmed by transmission electron microscopy (TEM), while pure HKUST-1 contains a greater fraction of amorphous phase.

This suggests that both TiO₂ NPs functionalization with specific carboxylate end-group of PHA and the use of step-by-step self-assembly synthesis promote the crystalline growth of MOF porous material, even under mild conditions. In detail, in the first step of HKUST-1 synthesis, PHA carboxylic groups allow the complexation of Cu²⁺ metal ions anchoring them onto the oxide surface. In the second step, the copper centres covalently bond to the BTC ligands enabling the growing of HKUST-1 on the TiO₂ surface and the control of its crystallographic orientations [46].

The observed HKUST-1 crystallinity represents a key feature to impart desired functionality to the final hybrid material, such as improved adsorption ability and consequent enhanced catalytic activity [47].

Attenuated Total Reflection Fourier Transform Infrared Spectroscopy (ATR-FTIR) spectroscopy was performed in order to investigate the TiO₂–MOF hybrid catalyst (Figure 2).

Specifically, the spectrum of PHA shows the following main bands (Figure 2a, grey line): the intense carboxyl stretching vibration at 1710 cm⁻¹; the P=O stretching vibration at 1220 cm⁻¹ and the methylene C–H bending at 1309 and 1213 cm⁻¹ (very weak).

After titania functionalization, the spectrum of TiO₂-PHA (Figure 2a,b, red line) still displays the characteristic absorption bands at 1715 cm⁻¹ deriving from COOH of the PHA, instead the PHA band at 1220 cm⁻¹, attributed to the P=O stretching vibration, disappears (red line in Figure 2b). TiO₂-PHA spectrum also shows the vibrations at 995 cm⁻¹ and at 1150 cm⁻¹, attributable to the symmetric stretching of P-O-Ti and P-CH₂ bonds, respectively [48]. These results suggest that phosphoryl groups of PHA interact with TiO₂ surfaces.

The spectra of TiO₂/HKUST-1 (Figure 2b, blue line) and pure HKUST-1 (Figure 2b, sky blue line) look rather similar and show characteristics analogous to those reported in the literature [49]. The bands at 1623 and 1550 cm⁻¹ and at 1440 and 1364 cm⁻¹, corresponding to the asymmetric and symmetric stretching vibrations of the BTC carboxylate groups, confirm that MOF coating TiO₂ in the composite maintains the structural features of HKUST-1.

In addition, the band centred at 1715 cm⁻¹, corresponding to the carbonyl stretching in pure TiO₂-PHA, shifts to lower energy in the TiO₂/HKUST-1 spectrum, evidencing between 1540 and 1650 cm⁻¹ the antisymmetric mode of the chelated carboxyl group. This supports the deprotonation of carboxylic acid group and coordination with copper, confirming the effective role of PHA as grafting group for HKUST-1.

The thermal behavior of the hybrid material was assessed by thermogravimetric analysis (TGA). Figure 3 shows TGA curves obtained for both as-prepared shape-controlled TiO₂ (black line) and TiO₂-PHA (red line) nanocrystals. In the as-prepared sample (black curve), a small weight loss (~3%) beginning at nearly 30 °C and continuing until 250 °C is detectable, which can be ascribed to physisorbed solvents removal. A more relevant weight loss is observable instead in TiO₂-PHA in a wide temperature range, from 230 to 400 °C, attributable to the thermal degradation of the PHA. The total amount of PHA grafted onto TiO₂ (11.2 wt. %) was evaluated by the net weight loss of TiO₂–PHA between 150 and 400 °C, i.e., considering the total weight loss with exclusion of that associated to TiO₂ (i.e., 3 wt. %).

TGA analysis was also performed on TiO₂/HKUST-1 (blue line), in order to determine the real amount of HKUST-1 in the hybrid material. The loading of HKUST-1 (73% wt. %) was estimated from the net weight loss of TiO₂/HKUST-1 between 150 and 400 °C, i.e., considering the total weight loss with the exclusion of that associated to pure PHA.

The morphological features of both TiO₂ NPs and hybrid TiO₂/HKUST-1 materials were investigated by TEM microscopy.

RE TiO₂ sample exhibits well-formed rhombic elongated nanocrystals with estimated size of 16.5 nm in width and 45–60 nm in length (Table S1 in Supplementary Materials), and with mainly exposed {101} and {010} crystal facets (Figure 4a,b), according to the results reported by Dihn et al. [50].

In TiO₂/HKUST-1 hybrid composite (Figure 4c,d), TiO₂ NPs maintain their peculiar anisotropic morphology and appear organized in nanometric aggregates (Figure 4c). At a higher magnification (Figure 4d), the presence of HKUST-1 shells uniformly grown onto the TiO₂ surfaces and intimately connecting the NPs can be observed (yellow arrows in d). The uniform distribution of HKUST-1 around TiO₂ crystals into the hybrid material is also relevant; this is expected from the use of functionalized TiO₂ NPs.

Selected area electron diffraction (SAED) analysis was in agreement with XRD and high-resolution TEM (HRTEM) data, confirming again the presence of TiO₂ anatase phase in both TiO₂-PHA and TiO₂/HKUST-1 samples and of crystalline HKUST-1 in the hybrid system, as revealed by lattice fringes analysis (Figure S1). Finally, STEM-EDS analysis reveal the presence of Ti and Cu with a ratio close to the theoretical 1:1 (Figure S2).

In summary, TEM investigation indicates the successful formation of the hybrid structure and supports close interaction between the RE NPs and HKUST-1.

As mentioned above, the coupling of HKUST-1 to TiO₂ NPs was aimed at: (i) favouring the CO₂ access to the catalytic sites by improving the porosity necessary to capture CO₂; (ii) activating TiO₂ photocatalytic response in the visible region.

Hence, the porosity of the TiO₂/HKUST-1 material was evaluated by nitrogen adsorption-desorption isotherms at 77 K (Figure 5a). The product exhibits a type IV isotherm with H1 hysteresis, which is typical of ordered mesoporous materials with uniform cylindrical pores [51]. The BET surface area resulted 349 m²/g and the BJH pore-size distribution (Harkins–Jura approximation) shows the presence of mesopores with pore widths ca. 4 nm. The total pores volume corresponds to 0.25 cm³/g. The specific surface area and total pore volumes of the bare TiO₂ NPs resulted instead 170 m²/g and 0.21 cm³/g respectively (Table S1).

In addition to a high specific surface area, TiO₂/HKUST-1 also showed a significant adsorption CO₂ capacity at room temperature (303 K) and atmospheric pressure (Figure 5b), reaching a percentage of CO₂ adsorbed of 10.16 wt. % (2.3 mmol/g). As expected, this amount is about three times lower than that reported for unsupported HKUST-1 under the same conditions (7.23 mmol/g) [52]. However, the CO₂ uptake is rather similar to that determined for other Cu-based MOFs with even higher BET surface area [53,54,55].

The diffuse reflectance ultraviolet-visible (DR-UV/Vis) spectra (Figure 6) were acquired for pure HKUST-1, TiO₂-PHA and TiO₂/HKUST-1, in order to study their optical absorption properties.

As expected, TiO₂-PHA (red line) absorbs mainly UV light while HKUST-1 (sky blue line) absorption extends over the visible range [49]. Notably, TiO₂/HKUST-1 (blue line) composite also shows a red shift of the optical absorption toward the visible range compared to TiO₂-PHA.

Based on the reflectance spectra (Figure S3), the band gap (E_g) of the samples (absorption edge in the case of nanocomposites) was determined by plotting the modified Kubelka-Munk function, [F(R_∞′)hν]^1/2, versus the energy of the excitation light. The calculated values of the TiO₂-PHA, HKUST-1 and TiO₂/HKUST-1 powders resulted 3.15 eV, 2.70 eV and 2.90 eV, respectively.

The enhanced absorption in the visible range of TiO₂/HKUST-1 compared to TiO₂, along with the smaller absorption edge value, provides a suitable platform for sensitizing TiO₂, making the obtained hybrid material an active catalyst under solar light.

In order to understand the electronic changes responsible for the absorption modification, the TiO₂/HKUST-1, TiO₂-PHA, and pristine TiO₂ and HKUST-1 samples were investigated X-ray photoelectron spectroscopy (XPS).

The survey spectra of TiO₂/HKUST-1 confirmed the presence of Cu, Ti, O, and C elements (Figure S4). The evolution of the Cu 2p, Ti 2p and O 1s core levels by XPS spectra (Figure 7 and Figure S5) allowed us to reveal the electronic structure of these levels in TiO₂/HKUST-1.

The Cu 2p_3/2 XPS spectra of HKUST-1 and TiO₂/HKUST-1 (Figure 7a) show peaks with similar lineshapes, suggesting that the copper chemical environment in the hybrid material is analogous to that of pristine HKUST-1. The main peak of Cu 2p_3/2 was observed at 935.0 eV and it was deconvoluted into two components at 935.0 eV and 932.8 eV, originating from Cu(II) (blue peak in Figure 7a) and Cu(I) (sky blue peak in Figure 7a). By supposing a homogeneous stoichiometry, XPS indicates that Cu(I) percentage with respect to the total copper amount is ~11% in HKUST-1. This amount increases up to 17% in the final hybrid material [56].

The Ti 2p core level spectra of TiO₂, TiO₂-PHA and TiO₂/HKUST-1 samples are shown in Figure 7b. The Ti 2p_3/2 binding energy (BE) of pure TiO₂ sample was found at 458.5 eV, corresponding to the Ti(IV) state (green peak in Figure 7b), with a small fraction of reduced Ti(III) at 457.9 (yellow peak in Figure 7b) [57], indicating the presence of Ti(III) defects originated during the solvothermal synthesis.

The Ti 2p_3/2 level of TiO₂-PHA presents a broader line shape (458.2 eV) compared to pure TiO₂, indicating a higher amount of Ti(III) defects possibly generated after TiO₂ functionalization with PHA. It is well known that the concentration of Ti(III) strongly depends on the chemistry at the TiO₂ surface [58], where the presence of capping molecules, such as carboxylic or phosphonic acid [59,60], may cause the atomic oxygen diffusion, away from the lattice sites, reducing Ti(IV) to Ti(III) at the vacancy sites. This suggest that in TiO₂-PHA, phosphoesanoic acid induces a higher amount of Ti(III) sites at the TiO₂ surface, as confirmed by XPS analysis (Figure 7b).

In the case of TiO₂/HKUST-1, the Ti 2p peaks become even broader and shifts to higher BE compared to TiO₂-PHA. The fitting of the major band indicates the presence of three different components: Ti(IV) (green peak), Ti(III) species (yellow peak) and an additional Ti(IV) 2p peak at 459.3 eV (red peak). The appearance of additional Ti(IV) centers and the increase of Cu(I) atomic percentage in the final hybrid material (Figure 7a) suggests a partial oxidation of the Ti(III) species of TiO₂-PHA by the Cu(II) centers of HKUST-1. This hypothesis envisages the existence of an electron transfer between TiO₂ and HKUST-1 favoured by the intimate contact between titania and MOF anchored on its surface.

2.2. Photocatalytic Activity

The photocatalytic activity of TiO₂/HKUST-1 was evaluated in the CO₂ degradation. In accordance with previous studies reporting the gas phase CO₂ photoreduction [61,62,63,64], the main products were CH₄ and traces of CO [65,66,67,68]. We assume the detected hydrocarbons are unambiguously formed from CO₂ as carbon source, since the photocatalytic processes were performed after a peculiar cleaning procedure, in accordance with Mei et al. [69].

CO₂ photoreduction in the presence of TiO₂-PHA NPs (Table S2) produced a very low concentration of CH₄, equal to 0.69 μM (corresponding to 0.28 μmol g⁻¹) after 6 h of irradiation. The tests performed in the presence of bare HKUST-1 sample (Table S3) yielded also very low amounts of CH₄ (0.42 μmol g⁻¹). In the presence of TiO₂/HKUST-1 hybrid material, a significant generation of CH₄ was observed (Table S4), indicating that methane derived from CO₂ reduction and the catalytic ability of TiO₂/HKUST-1 is by the photocatalytic process affected during photocatalysis. A total amount of 2.63 μM CH₄, corresponding to 1.05 μmol g⁻¹, was produced after 6 h of irradiation without any significant morphological and structural changes of TiO₂/HKUST-1 (see Figure S6) after the second photocatalytic run.

As H₂ could be obtained as by-product during the CO₂ photocatalytic reduction, its presence was checked during the runs, but it was not detected.

Figure 8 reports the comparison among the evolution of CH₄ concentration along irradiation time for TiO₂-PHA, HKUST-1 sample and TiO₂/HKUST-1 hybrid system (during the first and the second test). While a small amount of CH₄ was formed by using the bare samples, a relevant CH₄ production was observed for TiO₂/HKUST-1. For this latter sample, the amount of products obtained during the run carried out after the cleaning procedure, was comparable to those measured during the first run. The higher activity of the hybrid sample suggests the occurrence of a synergistic effect between RE NPs, which are known to be active in photoreduction reactions, and the MOF grafted onto TiO₂ surface. This finding is in accordance with the XPS, FTIR and TEM results.

Finally, both the TiO₂ + HKUST-1 sample, obtained by mixing un-functionalized TiO₂ NPs and HKUST-1, and the commercial TiO₂ sample P25, as reference materials, were tested under the same experimental conditions. The TiO₂ + HKUST-1 sample displayed a negligible photoactivity, producing low amounts of CH₄ and CO, in line with the results obtained by using the single components as photocatalysts. This is probably due to: (i) the presence of bare TiO₂ NPs, which are not reactive under solar light; and (ii) the high agglomeration degree of TiO₂ NPs in TiO₂/HKUST-1 may hinder the contact between the catalytic active sites and CO₂ target molecules. In addition, in the presence of commercial TiO₂ P25, we observed only a negligible CH₄ production.

These results support the efficacy of our synthetic approach and highlight the important role of PHA functionalization in order to guarantee a close contact between TiO₂ NPs and MOF, which delivers to TiO₂/HKUST-1 remarkable photoreduction properties.

2.3. Proposed Photocatalytic Pathway

Given the analysis above, a possible mechanism for the enhanced photocatalytic CO₂ reduction over the synthetized TiO₂/HKUST-1 hybrid material was proposed.

To better describe the electronic band structures of TiO₂-PHA NPs and HKUST-1 at the hybrid interface in TiO₂/HKUST-1, the relative energy of the conduction band (CB) and valence band (VB) versus normal hydrogen electrode (NHE) of both pristine HKUST-1 and TiO₂-PHA were calculated, according to Schoonen et al. [70,71], by the empirical equations:

E_CB = χ(S) − E^e − 0.5 E_g

(1)

E_VB = E_CB + E_g

(2)

where E_VB and E_CB are the CB and VB potentials, respectively. Moreover, E_g is the band gap of the semiconductor and E^e is the energy of free electrons vs. hydrogen (~4.5 eV) [72]. Finally, χ is the electronegativity of semiconductor and it was calculated by the following equation:

$$\chi (\mathrm{S})=\sqrt[N]{{\chi }_1^r{\chi }_2^s\dots \dots {\chi }_{n-1}^p{\chi }_n^q}$$

(3)

where ${\chi }_n$, n, and N are the electronegativity of the constituent atom, the number of species, and the total number of atoms in the compound, respectively [73]. The superscripts r, s, p and q refer to the numbers of the atoms 1, 2, n⁻¹, and n, respectively, in the molecule where (r + s + … + p + q = N).

Although this method cannot give absolute values because the structural factors are neglected, it may provide a rough estimation of the relative energy of CB and VB versus normal hydrogen electrode (NHE).

For HKUST-1, the values of E_g and χ values were 2.7 and 6.17 eV and consequently, E_CB and E_VB result were 3.02 and 0.32 eV while for TiO₂-PHA E_g and χ were 3.15 and 5.84 eV, and its E_CB and E_VB as −0.24 and 2.92 eV, respectively, versus NHE, in line with the reported experimental data [74].

With regard to TiO₂-PHA, the presence of the electron trapping sites, associated to Ti(III) sites, as indicated by XPS analysis, was also considered. The energy levels associated to those electron trapping sites could range between 0.1 and 1 eV lower than the anatase CB [75,76].

By considering the values calculated for E_CB and E_VB of TiO₂-PHA and HKUST-1, a scheme of the energy levels at the TiO₂-PHA/HKUST-1 interface in TiO₂/HKUST-1 can be proposed which suggests a pathway for the CO₂ photoreduction process (Figure 9).

It can be observed that the energy levels of Ti(III) are located at values higher than the CB of HKUST-1. Thus, upon UV-Vis irradiation of TiO₂/HKUST-1, electrons photogenerated from TiO₂-PHA VB can be trapped into Ti(III) centers and easily injected into HKUST-1 CB, thanks to the intimate contact between titania and HKUST-1. This seems to indicate the key role of Ti(III) defective sites of TiO₂-PHA as donor of electrons [77], which can be transferred to HKUST-1 and used in the photocatalytic CO₂ reduction.

Besides, photogenerated holes can migrate toward the VB of TiO₂ and contribute to the water decomposition to OH• and H⁺ [78]. The whole process highly hinders the electron–hole recombination [79,80], boosting the photoreduction activity of TiO₂/HKUST-1.

Finally, considering the significant amount of copper in TiO₂/HKUST-1, we cannot exclude its involvement in the photocatalytic reaction. Indeed, the beneficial effect of Cu toward the CO₂ photoreduction is well-known [61,64]. In the present case, the improved TiO₂ photoactivity may be connected to the simultaneous presence in TiO₂/HKUST-1 of Cu(II) and Cu(I) species, as revealed by XPS analysis. According to Slamet et al. [81], Cu(II) can be easily reduced by photoexcited electrons to Cu(I), while this latter species, in the presence of H⁺ or O₂, can be re-oxidized giving Cu(II) centers and electrons. This redox cycle may also contribute to enhance the photoreduction performance of TiO₂/HKUST-1 hybrid material.

In order to provide a further experimental probe of the suggested mechanism, Electron Spin Resonance (EPR) investigation was tentatively performed on both TiO₂/HKUST-1 hybrid material and TiO₂ + HKUST-1 mixture. The results evidenced the presence of a broad and intense signal related to Cu²⁺ species, which are very abundant in HKUST-1. This feature appears almost unaffected by the irradiation and, due to its intensity and broadness, did not allow to discriminate the possible presence of other resonances (e.g., those related to Ti³⁺ or O⁻/O₂⁻ species of titania).

3. Materials and Methods

3.1. Materials

Ti(OBu)₄ or TB, 97%, C₁₈H₃₃CO₂H or OA, 90%, C₁₈H₃₅NH₂ or OM 70%, TMAH solution 25 wt. % in H₂O, PHA, copper(II) acetate AcCu(II), benzene-1,3,5-tricarboxylic acid (trimesic acid, H₃BTC), ethanol (EtOH) were all purchased from Aldrich and used without further purification.

3.2. Synthesis of Shape-Controlled TiO₂ NPs and Functionalization with PHA

In a typical experiment, TB (44 mmol, 15.0 g) was added to a mixture containing 88 mmol of OA (25.0 g) and 132 mmol (35.3 g) of OM in 25 mL of absolute EtOH. The obtained mixture was stirred for 15 min and then transferred into a 400 mL Teflon-lined stainless-steel autoclave containing 85 mL of absolute EtOH and 3.5 mL of Milli-Q water. The system was then heated at 180 °C and kept at this temperature for 18 h. After decantation, the TiO₂ powder was recovered from the autoclave, washed with EtOH several times, filtered and finally dried in vacuum (p < 10⁻² mbar) at room temperature. Then, the TiO₂ NPs were dispersed by ultrasound in EtOH solution and then stirred at room temperature for 3 days in the presence of a suitable amount of TMAH (molar ratio TMAH/TiO₂ = 25/1). After decantation, the TiO₂ powder was washed with EtOH several times, filtered and finally dried at 80 °C for 12 h (up to 87% yield).

660 mg of the washed TiO₂ NPs were dispersed by ultrasound in EtOH/water solution (4/1 v/v) and then functionalized with 540 mg of PHA. The solution was stirred for 24 h at reflux (Scheme 1, STEP 3), then the TiO₂-PHA NPs were collected by centrifugation and the powders were washed several times with EtOH and dried in an oven at 80 °C for 12 h.

3.3. Synthesis of TiO₂/HKUST-1 Hybrid Catalyst

In a typical reaction, a fixed amount of TiO₂-PHA (300 mg) was firstly dispersed and sonicated for 10 min in an EtOH/H₂O 1:1 solution. Simultaneously, two solutions with the precursors AcCu(II) and H₃BTC have been prepared. In the first, 355 mg AcCu(II) were dissolved in 20 mL of EtOH/H₂O 1:1 solution at 45 °C, while 222 mg of H₃BTC were dispersed in 10 mL of EtOH/H₂O to obtain the second solution. After that, the two solutions were mixed with the TiO₂-PHA suspension prepared before. The suspension was stirred for 30 min at 45 °C and then the product was collected by centrifugation and the powders were re-dispersed in the EtOH/H₂O solution. This suspension was mixed again with the AcCu(II) and H₃BTC solutions. The process was repeated three times in order to obtain a material with a higher percentage of MOF. Finally, the TiO₂/HKUST-1 powders obtained were washed 3 times with EtOH/H₂O 1:1 and dried at 80 °C in an oven for one night.

Notably, the photocatalyst was carefully treated to remove possible organic contaminants. To eliminate the solvent and other organic agents from the pores of the TiO₂/HKUST-1 hybrid materials, the powders were treated in vacuum at 150 °C for one night in a Büchi.

In order to highlight the photocatalytic performance of TiO₂/HKUST-1, the following reference materials were also prepared: (i) pristine HKUST-1 and (ii) a composite material constituted of not-functionalized TiO₂ and HKUST-1, obtained by the same procedure described before and labeled as TiO₂+HKUST-1.

3.4. Characterization of TiO₂ NPs, HKUST-1 and TiO₂/HKUST-1

Analyses of the crystalline materials for phase identification were performed by PXRD. PXRD data were collected on a Rigaku Miniflex 600 diffractometer in reflectance Bragg–Brentano geometry with graphite monochromatized Cu-Kα radiation (λ = 1.5406 Å) at 600 W (40 kV, 15 mA) power. Samples were mounted on a zero-background silicon sample holder by dropping powders from a spatula and then gently leveling the sample surface with a razor blade. Samples were not ground before PXRD measurements. Scan rates were 1°/min with 0.02° angular steps in the 2θ range 2–65°. Data were analysed with PDXL2 software (Rigaku, Tokyo, Japan) and Qualx2 [82].

To preliminarily check the TiO₂ washing and PHA functionalization and, successively, the formation of the HKUST-1 structure, ATR-FTIR measurements were performed on a Perkin Elmer Spectrum 100 instrument (1 cm⁻¹ resolution spectra, 650–4000 cm⁻¹ region, 16 scans).

To quantitatively assess the PHA and MOF grafting onto TiO₂ surfaces, thermogravimetric analysis (TGA) measurements were carried out. TGA thermograms were collected by a Mettler Toledo TGA/DSC1 STARe System, at a constant gas flow (50 cm³ min⁻¹). The sample powders were heated in air from 30 to 1000 °C. The thermal profile was the following: 30–150 °C at 2 °C min⁻¹; dwell at 150 °C for 120 min; 150–1000 °C at 5 °C min⁻¹.

Morphological characterization by HRTEM and SAED of bare TiO₂ NPs and TiO₂/HKUST-1 powders were performed on a ZEISS LIBRA200FE EFTEM. Elemental composition was evaluated by STEM-EDS analysis (Scanning Transmission Electron Microscopy-Energy Dispersive X-ray Spectrometry Oxford INCA Energy TEM 200, Oxford Instruments, Abingdon-on-Thames, U.K.) of representative grains. The powders were suspended in isopropyl alcohol, sonicated and deposited onto a holey carbon film supported TEM grids. For STEM-EDS analysis holey-carbon molybdenum TEM grids were used. Samples were analyzed after overnight drying.

Low-pressure N₂ adsorption isotherm on HKUST-1 NPs were recorded by a Micromeritics ASAP2020 apparatus. The specific surface area (SSABET, BET method) was measured after evacuation of the samples at 120 °C for 12 h. A liquid N₂ bath was used for measurements at 77 K.

The CO₂ sorption was evaluated at room temperature (293 K) and controlled CO₂ pressure (p_CO2 from 5 to 920 mmHg) by Micromeritics ASAP 2020 after evacuation at 120 °C for 12 h.

DR-UV/Vis spectra of carefully ground powders were recorded in the 800–200 nm range with a UV Lambda 900 PerkinElmer spectrophotometer (PerkinElmer, Waltham, MA, USA), equipped with a diffuse reflectance accessory Praying Mantis sampling kit (Harrick Scientific Products, Pleasantville, NY, USA). A Spectralon disk was used as reference material.

The surface chemical composition of the TiO_2, TiO₂-PHA NPs, HKUST-1 and TiO₂/HKUST-1 powders was investigated by XPS. Analysis was performed on the as-prepared powders samples, fixing them on the sample holder using carbon tape. The XPS spectra were acquired in ultrahigh vacuum (base pressure: ~4 × 10⁻¹⁰ mbar) at room temperature in normal emission geometry using a conventional Mg X-ray source (hν = 1253.6 eV) and a hemispherical electron energy analyzer (total energy resolution ~0.8 eV). Due to charging effects, all BEs are calibrated by fixing the C 1s BE of atmospheric contamination at 284.6 eV [83]. The standard deviation for the BEs values was ~±0.2 eV. Survey scans were obtained in the 0–1100 eV range. Detailed scans were recorded for the O 1s, Ti 2p, and Cu 2p regions. To individuate all the possible differences between the samples, XPS spectra were reproduced by fitting the experimental data using a Shirley background and several Doniach-Sunjich components [84], corresponding to different oxidation states and chemical environments [85]. The fitting parameters have been fixed following Kaushik [86] and taking into account the energy resolution used in the measurements (~0.8 eV).

The EPR investigation was performed by a Bruker EMX spectrometer operating at the X-band frequency and equipped with an Oxford cryostat. The spectra of TiO₂/HKUST-1 hybrid material and TiO₂ + HKUST-1 mixture were carried out at 130 °C in vacuum conditions (p < 10⁻⁵ mbar), before and after UV-Vis irradiation, directly inside the EPR cavity.

3.5. Photocatalytic CO₂ Reduction

The photocatalytic CO₂ reduction was carried out in a batch cylindrical gas–solid reactor (V = 120 mL) containing 0.3 g of powder distributed as a thin layer. The system was irradiated from the top with a solar light simulating lamp (1500 W high pressure Xe lamp) inside a SOLARBOX (CO.FO.ME.GRA.). The reaction temperature was 60 °C.

The possible presence of products deriving from C impurities was checked by means of the standardized procedure reported by Mei et al. [69]. In the first step, after prolonged purging with water-saturated He, the photocatalyst was irradiated for about 2 h in contact with Helium in water vapor, (i.e., in the absence of CO₂). Then the system was completely purged again with Helium, to remove all carbon-containing species in the gas phase. The cleaning procedure was repeated until carbon contaminations were removed from the sample surface. After this so-called cleaning step, the photocatalytic tests were performed. In detail, the system was saturated with wet CO₂. 500 μL of the gaseous mixture were withdrawn from the reactor for analyses at fixed irradiation times by using a gas-tight microsyringe. The evolution of CH₄ and CO was followed by a HP 6890 Series GC equipped with a packed column GC 60/80 Carboxen-1000 and a TCD detector, whilst the concentration of the organic species was measured by a GC-2010 Shimadzu gas chromatograph equipped with a Phenomenex Zebron Wax-plus column (30 m × 0.32 μm × 0.53 μm) and a flame ionization detector, using He as the carrier gas. Each photocatalyst was tested three times, to check the reproducibility of the photocatalytic runs, in terms of CO₂ (in the cleaning step), CH₄ and CO production.

4. Conclusions

In summary, a novel hybrid TiO₂/HKUST-1 photocatalyst was successfully obtained via a slow-diffusion method combined with step-by-step self-assembly approach at room temperature, starting from TiO₂ NPs functionalized with an asymmetric organic linker. Our synthetic approach promotes the crystalline growth of MOF layers on titania surface and favors an intimate contact between HKUST-1 and TiO₂-PHA.

The close contact between two hybrid components is a key point for expressing the remarkable performance of the hybrid material. In fact, this allows not only a high CO₂ uptake into the porous MOF structure but also, under solar light irradiation, favors an improved photoreduction activity of TiO₂/HKUST-1 compared to that of pure TiO₂-PHA and HKUST-1 samples.

The remarkable photoreduction ability of the material was related to the improved visible light absorption and to an effective electron injection from TiO₂-PHA to HKUST-1 involving both photogenerated electrons and those trapped in Ti(III) centers of TiO₂-PHA. This indicates that in the hybrid TiO₂/HKUST-1 photocatalyst, also the presence of electrons deriving from a redox active role of TiO₂-PHA can be exploited in the photocatalytic reduction of CO₂. Moreover, the performance of the material was confirmed by subsequent photocatalytic runs carried out with the same sample.

We expect that our synthetic approach will enable the possibility to fabricate a wider range of hybrid porous photocatalytic materials, namely TiO₂/MOF, with a designable MOF shell, suitable for a variety of energy and environmental applications. Particularly, the structure, composition, and function of the MOF shell could also be judiciously tailored by choosing different framework building blocks, i.e., metal ion and polyfunctional organic likers.