2.1 Morphology and Structure of PB2T-TEG-TiO_2−X

The structure and synthesis of PB2T-TEG-TiO_2−X are shown in Figure 1a. The polymer PB2T-TEG was synthesized by Stille polymerization from dibromobenzene attached with OEG side chains and 5,5′-bis(trimethylstannyl)-2,2′-bithiophene. The procedure of polymer synthesis is described in the Experimental Section. The polymer is soluble in chloroform and dispersible in ethanol. The ¹H NMR spectra showed characteristic signals of the OEG side chains at 3.3–7.6 ppm, indicating that hydrophilic non-conjugated groups were present in the polymer (Figure S1, Supporting Information). The number-average molecular weight (Mn) and dispersity (Đ) of the polymer were 41 kDa and 1.5, respectively, determined by size exclusion chromatography in dimethylformamide. The oxygen-doped semiconductor TiO_2−X mesoporous nanospheres that were obtained by high-temperature hydrogen reduction were dispersed in ethanol. Subsequently, the ethanol dispersion of TiO_2−X was mixed with the PB2T-TEG ethanol solution under ultrasonication and the mixture was heated while stirring to form composite materials. In order to optimize the performance of the heterojunction, the contents of PB2T-TEG were varied to be 1, 3, 5, and 7 wt.% relative to TiO_2−X. The prepared samples were denoted as TiO_2−X-PY (Y = 1,3,5,7), e.g. the PB2T-TEG content of 3 wt.% was denoted as TiO_2−X-P3. In the ethanol dispersion, the TiO_2−X and PB2T-TEG samples had opposite charges, allowing the construction of TiO_2−X-PY through the electrostatic interactions between TiO_2−X and PB2T-TEG (Figure 1b). The microscopic morphology of the heterojunction was observed by scanning electron microscopy (SEM), in which TiO_2−X is a mesoporous nanosphere with a rough surface (Figure S2, Supporting Information). As shown in Figure 1c, PB2T-TEG of graphene-like layered assemblies is randomly wrapping around the mesoporous TiO_2−X nanospheres, suggesting the existence of close interactions between PB2T-TEG and TiO_2−X. The disposition of PB2T-TEG over the TiO_2−X mesoporous sphere was unequivocally confirmed by transmission electron microscopy (TEM) (Figure 1d,e). The high-resolution transmission electron microscopy (HRTEM) images also clearly showed the existence of a distinct heterogeneous interface between PB2T-TEG and TiO_2−X (Figure 1f). The lattice strip spacing of TiO_2−X is 0.342 nm, corresponding to the lattice plane of anatase TiO₂ (101).^[47] However, due to its lower crystallinity, the lattice stripes of amorphous PB2T-TEG are indistinguishable. Nevertheless, pure TiO_2−X exhibits notably clear and uncovered grain surfaces (Figure S3, Supporting Information). The energy-dispersive spectroscopy (EDS) was further employed for elemental mapping. The results confirmed that C, S, O, and Ti elements are evenly distributed within the TiO_2−X-P3 nanostructure, and the S distribution indicates the successful encapsulation of PB2T-TEG on TiO_2−X (Figure 1g–j).

Powder X-ray diffraction (PXRD) confirmed the crystal structure of PB2T-TEG-TiO_2−X, as shown in Figure 2a. The diffraction peaks at 25.4°, 38.2°, 48.1°, 54.1°, 55.0°, 63.5°, 69.1°, 70.4°, and 75.3° can be indexed to (101), (004), (112), (200), (105), (211), (204), (116), (220) and (215) crystal planes of anatase TiO₂ (JCPDS card No. 21–1272), respectively.^[48] TiO_2−X MS has a similar crystal structure with pristine TiO₂ MS (Figure S4, Supporting Information), indicating that TiO_2−X has good structural stability during high-temperature hydrogen reduction. Compared with TiO₂ MS, the (101) diffraction peak intensity of TiO_2−X MS decreases, and the full width at half maximum (FWHM) broadens due to oxygen vacancies and Ti³⁺ doping.^[49] The presence of a diffraction peak at 7.3° is attributed to the polymer PB2T-TEG. The peaks of TiO_2−X and PB2T-TEG are clearly observed in all prepared composite materials. The samples were observed with Fourier transform infrared spectrometer (FTIR) spectra, as shown in Figure 2b. In the FTIR spectrum of TiO_2−X-P3, there were strong peaks at 1085 and 1157 cm⁻¹, attributed to C─O stretching vibrations resulting from PB2T-TEG molecules. The peaks at 1632 cm⁻¹ and broad IR bands ≈3600–3000 cm⁻¹ originate from the absorption of water or ─OH groups.^[50] The peaks at 2932 cm⁻¹ are attributed to the C─O bond.^[51] The absorption peak at 1380 cm⁻¹ demonstrates the vibration of Ti─O.^[52] FTIR spectra showed the characteristic bands of the TiO_2−X-P3 contain the PB2T-TEG and TiO_2−X cores, thus confirming the successful synthesis of hybrid heterojunctions. In Figure 2c, the incorporation of thiophene units resulted in the detection of S 2p peaks in the TiO_2−X-P3 heterostructure with a notable shift toward higher energy compared to the S 2p peak of PB2T-TEG alone by XPS at a detection depth of ≈10 nm, suggesting that the organic PB2T-TEG is grafted onto TiO_2−X.^[53] As shown in Figure S5 (Supporting Information), the high-resolution O 1s spectrum of TiO_2−X-P3 can be fitted into three peaks of 529.5, 531.0, and 533.5 eV, which could be attributed to lattice oxygen, ─OH and C─O bonds,^{[54, 55]} respectively. Furthermore, additional small peaks associated with defect-induced Ti³⁺ can be observed in the high-resolution spectrum of TiO_2−X-P3 for Ti 2p (Figure 2d). Similarly, electron paramagnetic resonance (EPR) tests were carried out to investigate the defects in the photocatalysts, and it can be seen that after high-temperature hydrogen reduction, characteristic peaks attributed to oxygen defects appear in the black TiO_2−X (e.g., surface Ti^IV─O centers, with g values >2.0023) compared to TiO₂ MS (Figure 2e).^[56] While in TiO_2−X-P3 the EPR response is stronger due to the presence of unpaired electrons on the aromatic ring carbon atoms of PB2T-TEG.^[57] The X-ray absorption near edge structure (XANES) spectrum for as-prepared TiO_2−X-P3 was compared with those for authentic Ti(0), Ti(III), and Ti(IV) models in Figure 2f. Although the detailed structure of the XANES spectrum of TiO_2−X-P3 is basically the same as that of the Ti(IV) model compound, the slight edge shift to lower energy for TiO_2−X-P3 is consistent with Ti(III), suggesting that the oxidation state of Ti in the as-prepared TiO_2−X-P3 is a mixture of Ti⁴⁺ and Ti³⁺.^[58]

2.2 Optical Properties and Charge Transfer

The optical absorption and electronic structure of photocatalysts are important factors affecting hydrogen production. The optical absorption of the samples was characterized by UV–vis diffuse reflectance spectroscopy. The spectra are illustrated in Figure 3a and Figure S6 (Supporting Information). Pure TiO₂ exhibits strong absorption in the ultraviolet region, with negligible absorption in the visible region. However, defective doping and the incorporation of PB2T-TEG led to a significant enhancement of absorption in the visible region for TiO_2−X-P3. This enhancement implies that the heterojunction may generate more charge carriers under visible light irradiation. The inset of Figure 3a also shows Tauc diagrams with estimated band gaps of 2.91 and 2.75 eV for TiO_2−X and TiO_2−X-P3, respectively.^[59]

To examine the impact of TiO_2−X-P3 heterojunction on the photo-induced electron separation and transfer, steady-state photoluminescence (PL), time-resolved photoluminescence (TRPL) spectroscopy measurements, and electrochemical tests were conducted. The PL emission peak of black TiO_2−X under excitation at 325 nm is attributed to band-to-band recombination, which weakens upon the formation of a heterojunction with PB2T-TEG (Figure 3b). Corresponding to the UV/Vis absorption, PB2T-TEG does not exhibit a PL emission peak due to its lack of optical absorption in the ultraviolet region. Under 420 nm excitation, the PL intensity of TiO_2−X-P3 decreases dramatically compared to pure PB2T-TEG, indicating that the construction of the heterojunction significantly improves charge/energy transfer.^[60] To gain in-depth insights into the charge transfer dynamics, TRPL decay measurements were performed (Figure 3c; Table S2, Supporting Information). The TiO_2−X-P3 heterojunction shows an increased average emission lifetime (τ = 5.31 ns) compared to black TiO_2−X (τ = 4.32 ns), indicating that faster electron transfer and slower charge recombination in TiO_2−X-P3. The same conclusion can be obtained through electrochemical impedance spectroscopy (EIS) measurement as well. It reveals photoinduced electron–hole separation and charge transfer resistance within the TiO_2−X-P3 heterojunction, as shown in Figure S7 (Supporting Information). Typically, the charge transfer resistance is reflected through the Nyquist curve, where a smaller diameter of the semicircular arc indicates higher efficiency in electron–hole separation and transfer.^[61] Apparently, the TiO_2−X-P3 heterostructure presents smaller arcs than that of TiO_2−X and PB2T-TEG, indicating that TiO_2−X-P3 has weaker electronic impedance and enhanced charge separation and transfer efficiency. The photocurrents of PB2T-TEG, TiO_2−X, and TiO_2−X-P3 were measured via several on–off cycles under intermittent UV light irradiation, correlating with the separation efficiency of the photogenerated charge carriers. As shown in Figure 3d, the photocurrent responses of all samples are stable and reproducible in the time scale of 360 s, and the transient photocurrent of TiO_2−X-P3 turns out to be ≈2.4 times and 4.3 times higher than those of TiO_2−X and PB2T-TEG, respectively. All these results illustrate that the formation of heterojunction promotes efficient electron–hole separation. As expected, the linear sweep voltammetry (LSV) of these samples also displayed that the TiO_2−X-P3 heterojunction yielded the highest photocurrent density (Figure S8, Supporting Information). The enhanced photoelectric properties of TiO_2−X-P3 can be attributed to increased charge carrier mobility in the heterojunction structure.

In addition, the relative positions of the energy bands of TiO_2−X and PB2T-TEG play an important role in determining electron transfer in photocatalysis. Thus, the band positions are also studied in detail. The electrochemical impedance measurements were conducted on the pristine TiO_2−X to evaluate their flat band potentials. As demonstrated in Figure 3e, the positive slopes of the extension lines at 500, 1000, and 1500 Hz frequencies exhibited the n-type semiconductor features of as-prepared TiO_2−X, and the flat-band potentials of the TiO_2−X is estimated to be −0.77 V (vs Ag/AgCl), suggesting Fermi levels locate at −0.57 V (vs NHE).^[62] The valence band (VB) positions were inferred from valence band X-ray photoelectron spectroscopy (VB-XPS) for the inorganic semiconductor TiO_2−X. As shown in Figure 3f, the distance between VB and Fermi level of TiO_2−X is reckoned to be 2.57 eV. Naturally, the TiO_2−X samples have the VB position of 2.00 eV. Furthermore, based on the energy gaps (E_g) of TiO_2−X from UV–vis absorption and the VB from the VB-XPS spectrum, the conduction band (CB) for the sample was calculated using the formula: E_CB = E_VB + E_g,^[63] which is −0.91 V for electrode potential (V vs NHE). Regarding the electronic structure of the hybrids, the relative energy levels of the highest occupied molecular orbital (HOMO) and the lowest unoccupied molecular orbital (LUMO) of the polymers were estimated by the electrochemical determination of their ionization potential (IP) and electron affinity (EA). For the PB2T-TEG, HOMO was calculated to be −5.00 eV, and LUMO was −3.22 eV versus Vacuum, according to its oxidation and reduction potentials (0.5 and −1.28 V vs NHE, respectively) from cyclic voltammetry (CV) measurements (Figure 3g). PB2T-TEG exhibited a bandgap of 1.78 eV, according to its HOMO and LUMO energies. Based on the above-mentioned results, the band structure diagrams of TiO_2−X and PB2T-TEG are illustrated in Figure 3h, which indicates a Type II energy level alignment and fulfills the thermodynamic requirements for driving H₂ evolution.

2.3 Photocatalytic Activity

The photocatalytic H₂ evolution was implemented under simulated sunlight AM 1.5G with Pt as a co-catalyst and triethanolamine as a sacrificial reagent. Figure 4a depicts the cumulative hydrogen production versus time for PB2T-TEG, TiO_2−X, and TiO_2−X-P3 catalysts. The pristine polymer PB2T-TEG exhibits poor HER performance due to the rapid recombination of photo-induced electrons and holes. Meanwhile, in a series of hybrid TiO_2−X-PY (Y = 1, 3, 5, 7 wt.%), the photocatalytic activity shows a trend of initially increasing and then decreasing with the amount of PB2T-TEG. The initial increase can be explained by the strong absorption of the polymer PB2T-TEG and the subsequent decrease is mainly due to the excessive coverage of the active sites by the excess of PB2T-TEG. The HER of the TiO_2−X-P3 could reach 35.7 mmol g⁻¹ h⁻¹ at the dosage of 3 wt.% PB2T-TEG, which is 2.8 and ≈221 times higher than those of TiO_2−X (12.9 mmol g⁻¹ h⁻¹) and PB2T-TEG (0.161 mmol g⁻¹ h⁻¹), respectively. The results show that there is a synergistic effect between organic and inorganic semiconductors in heterostructures.

It is well known that one of the main limitations of polymer photocatalytic applications is their limited photochemical stability. In fact, linear polymers tend to degrade under prolonged light exposure. Here, an extended 16-h hydrogen production was conducted to assess the stability of the photocatalyst. Throughout the entire duration, the hydrogen evolution remained consistently stable, demonstrating total H₂ production of 495.85 mmol g⁻¹ in 16 h (Figure 4b). Even without the co-catalyst Pt, the HER of TiO_2−X-P3 can still reach 8.6 mmol g⁻¹ h⁻¹ (Figure S9, Supporting Information). To further highlight the contribution of PB2T-TEG to hydrogen production in the heterojunction, hydrogen production was investigated under UV light only, outside of the PB2T-TEG absorption range, and the HER of TiO_2−X-P3 is only twice as high as that of TiO_2−X. Furthermore, the photostability of TiO_2−X-P3 was characterized by XPS analysis of the recovered TiO_2−X-P3 after the photocatalytic reaction for 16 h, and no significant change is observed in comparison with the fresh TiO_2−X-P3 catalysts (Figure 2c; Figure S10, Supporting Information). Especially, Ti³⁺ has strong stability before and after photocatalytic reaction (Figure 2d; Figure S10c, Supporting Information). Both PXRD and TEM demonstrated the structural, compositional, and morphological stability of the TiO_2−X-P3 heterojunction during the photocatalytic process (Figures S11 and S12, Supporting Information).

Meanwhile, the high-resolution Pt 4f XPS spectrum of TiO_2−X-P3 after the photocatalytic reaction revealed that the in situ anchored Pt remained on TiO_2−X-P3 (Figure S12, Supporting Information), confirming the high stability of metallic Pt on the TiO_2−X-P3.^[64] AQY was obtained under the illumination of a 300 W xenon lamp with different wavelength band-pass filters (λ = 365, 410, 460, 500, 600 nm).^[65] TiO_2−X-P3 exhibits AQYs of 53.3% and 11.1% at 365 and 500 nm, respectively, with a co-catalyst Pt content of 0.62 wt.% (Figure 4c; Table S3, Supporting Information). The trend in AQY as a function of wavelength matches with the UV–vis absorption spectrum of TiO_2−X-P3, indicating that photocatalytic hydrogen evolution indeed occurred upon light harvesting. To the best of our knowledge, TiO_2−X-P3 presents one of the highest activities reported for photocatalytic H₂ evolution and could provide a pathway for efficient solar-to-chemical energy conversion by significantly improving the solar-to-hydrogen conversion efficiency of the current hybrid heterojunction (Figure 4d; Table S1, Supporting Information).^[66]

2.4 Photocatalytic Mechanism Analysis

To clarify the charge transfer process and understand the improved photocatalytic activity of the hybrids, fs-TA spectroscopy (Figure 5a–h) using the pump-probe technique under 315 nm laser excitation and high-resolution in situ XPS measurements for Ti 2p states of TiO_2−X-P3 were conducted (Figure 6a). As shown in Figure 5a,d, the TA spectrum of PB2T-TEG shows an obvious negative peak at ≈576 nm after excitation by a pump pulse with a wavelength of 315 nm, which belongs to ground state bleaching (GSB).^[67] Figure 5b,e displays the fs-TA spectroscopy for TiO_2−X, two GSB peaks can be clearly observed in the wide wavelength range of ≈340 nm and 512–640 nm on the micro-second-second time scale, which are in good agreement with the absorption edge of TiO_2−X. In addition to the GSB peaks, a positive peak with a wavelength of ≈423 nm is observed in the fs-TA spectrum of TiO_2−X, which is a spectral feature attributed to the excited state absorption (ESA) of the trapped holes on the defect state. As shown in Figure 5c,f, GSB signals dominated by PB2T-TEG appear at the TiO_2−X-P3 heterojunction at 584 nm. It is obvious that the GSB peak of PB2T-TEG in the TiO_2−X-P3 heterostructure is redshifted, which is probably caused by the introduction of strain at the interface of the heterostructure or the change of the built-in electrical field of the heterostructure.^[68] Thus, the decay times were further studied in the PB2T-TEG, TiO_2−X, and TiO_2−X-P3 composite. Figure 5g shows the kinetic profiles of TiO_2−X and TiO_2−X-P3 absorption at 340 nm. The decay traces were fitted to a three-exponential model, demonstrating that three pathways dominate the relaxation of photo-generated electrons (as shown in Table S4, Supporting Information). For the TiO_2−X-P3 composite, in addition to the very short diffusion lifetime of a few picoseconds (τ₁ = 4.8 ps) for the electrons on the lattice, two slower relaxation processes can be explained by the trapped states (τ₂ = 172 ps) and recombination of electrons in CB with holes in VB (τ₃ = 12 400 ps), respectively. Notably, TiO_2−X-P3 shows a longer electron diffusion lifetime than TiO_2−X, implying that the photogenerated charges can migrate further before decaying back to the ground state, leading to more efficient electron–hole separation.^[69] Further, the GSB decay signal at the peak corresponding to the PB2T-TEG bleaching peak was fitted using a four-exponential function since the interfacial electron-transfer process (Figure 5h; Table S4, Supporting Information). The newly fitted lifetime (τ₄) attributable to interfacial electron transfer is 19.2 ps in TiO_2−X-P3, in addition to the lifetimes attributed to electron diffusion on the lattices, carrier transfer to the trap states, conduction band electron, and valence band hole recombination.^[70] The electron transfer direction at the heterojunction interface can be further tested by in situ XPS.

During the XPS measurement process, three different lighting conditions were employed: dark, UV light, and visible light. Notably, as shown in Figure 6b, the binding energy of Ti 2p for TiO_2−X-P3 showed a noticeable negative shift upon visible light irradiation, in contrast to that under dark conditions, suggesting that more electrons are accumulated on TiO_2−X. Similarly, the Ti 2p binding energy of TiO_2−X-P3 exhibits a more pronounced negative shift under UV irradiation. ESR was performed to further explore the photocatalytic mechanism model by detecting the production of the reactive species in this system under dark and light irradiation. For the TiO_2−X and TiO_2−X-P3 systems, no ESR signal of reactive species occurred in the dark reaction. After 5 min of light irradiation, the characteristic peaks of DMPO-·O₂⁻ were observed (Figure 6c). The signal intensities of ·O₂⁻ radicals captured by TiO_2−X-P3 were notably higher than those captured by TiO_2−X after 10 min of light irradiation. This indicates that the heterojunction construction endows the photocatalyst with enhanced reduction capabilities. Based on the energy band structure of the photocatalyst, electron–hole separation behavior shown above, and photocatalytic performance, we propose a Type II-scheme model for the transfer of photo-generated electrons in the TiO_2−X-P3 heterojunction. Under UV light, the electron transfer takes place from the VB to CB of TiO_2−X corresponding to bandgap excitation. After that, the transfer of holes from the VB of TiO_2−X to the HOMO of PB2T-TEG occurs, leading to the separation of electrons and holes. Finally, protons are reduced by electrons at the CB of TiO_2−X (Figure 6d, left side). On the other hand, under visible light, both PB2T-TEG and TiO_2−X are excited, leading to electrons transfer from the LUMO of PB2T-TEG to the CB of TiO_2−X, and hole transfer from VB of TiO_2−X to HOMO of PB2T-TEG. Thereby, more electrons and holes are generated to facilitate the hydrogen production on the heterojunctions (Figure 6d, right-side). This can be the reason why the photocatalyst achieves high AQY both in the UV and visible regions. Overall, we can conclude that the OEG functionalized conjugated polymer can not only transfer more electrons in the visible region to TiO_2−X but also act as a hole trap in ultraviolet light to accelerate the process of photogenerated electrons in TiO_2−X, leading to the enhancement of photocatalytic reaction efficiency. The bifunctional polymer-modified inorganic photocatalyst with different optical bands has great potential for enhancing hydrogen production performance.