Boosting exciton dissociation and charge transfer in CsPbBr₃ QDs via ferrocene derivative ligation for CO₂ photoreduction

Significance

Perovskite quantum dots (QDs) have garnered significant interest in the field of photo-catalytic CO₂ reduction due to their superior extinction coefficient, eminent photoelectric energy conversion capability, and adjustable photoelectrochemical properties. However, obstacles such as coulomb interactions in excitons, surface capture states, and surface barrier energy deteriorate this process and are still challenging to be solved. In this work, we proposed a facile surface engineering strategy by directly grafting the ferrocene carboxylic acid (FCA) ligand to unlock the restricted multiple-exciton dissociation and inefficient charge transfer dynamic bottleneck of the photo-catalytic CO₂ reduction on emerging perovskite QDs and mark a significant nine-fold enhancement compared to pristine CsPbBr₃ QDs.

Abstract

Photo-catalytic CO₂ reduction with perovskite quantum dots (QDs) shows potential for solar energy storage, but it encounters challenges due to the intricate multi-electron photoreduction processes and thermodynamic and kinetic obstacles associated with them. This study aimed to improve photo-catalytic performance by addressing surface barriers and utilizing multiple-exciton generation in perovskite QDs. A facile surface engineering method was employed, involving the grafting of ferrocene carboxylic acid (FCA) onto CsPbBr₃ (CPB) QDs, to overcome limitations arising from restricted multiple-exciton dissociation and inefficient charge transfer dynamics. Kelvin Probe Force Microscopy and XPS spectral confirmed successfully creating an FCA-modulated microelectric field through the Cs active site, thus facilitating electron transfer, disrupting surface barrier energy, and promoting multi-exciton dissociations. Transient absorption spectroscopy showed enhanced charge transfer and reduced energy barriers, resulting in an impressive CO₂-to-CO conversion rate of 132.8 μmol g⁻¹ h⁻¹ with 96.5% selectivity. The CPB-FCA catalyst exhibited four-cycle reusability and 72 h of long-term stability, marking a significant nine-fold improvement compared to pristine CPB (14.4 μmol g⁻¹ h⁻¹). These results provide insights into the influential role of FCA in regulating intramolecular charge transfer, enhancing multi-exciton dissociation, and improving CO₂ photoreduction on CPB QDs. Furthermore, these findings offer valuable knowledge for controlling quantum-confined exciton dissociation to enhance CO₂ photocatalysis.

Converting anthropogenic CO₂ into valuable chemicals and fuels using renewable solar energy is a crucial step toward achieving intermittent solar energy storage and advancing toward a sustainable, environmentally friendly energy future (1–3). However, as an electron-consuming-dependent reaction, the complicated multi-electron photo-catalytic CO₂ reduction reaction (CRR) creates a significant barrier in both thermodynamics and kinetics, preventing CO₂ activation and conversion (3–6). Thus, developing advanced photocatalysts with robust photon-electron conversion capacity is crucial to ensure an adequate hot electron supply in this reaction (4, 6–9). Perovskite quantum dots (QDs) have garnered substantial interest as catalysts in the field of photo-catalytic CRR, owing to their remarkable quantum-confined effect. This effect enables the generation of multiple excitons, resulting in enhanced efficiency by simultaneously forming delocalized excitonic states via carrier multiplication or multiphoton absorption in the size-confined core region (10–12). Nevertheless, intrinsic kinetic competition between the multiple-exciton dissociation and exciton–exciton annihilation severely hinders the direct application of the multiple-excitons generation effect in perovskite QDs-based photocatalysis (13–15).

Efficient multiple-exciton dissociation and charge transfer to reactive site acceptors prior to the exciton–exciton annihilation process is a prerequisite for maximizing the multiple-excitons generation effect in perovskite QDs-based photocatalysis. However, as depicted in Fig. 1A, obstacles such as coulomb interactions in excitons (16, 17), surface capture states (18, 19), and surface barrier energy (20) deteriorate this process, which are still challenging to be solved. Encouragingly, recent research on exciton regulation in metal chalcogenide QDs has demonstrated the prominent impacts of surface ligands on these factors, providing relevant insights for overcoming these limitations through surface engineering. For instance, Frederick et al. employed phenyl dithiocarbamate, a hole-delocalizing ligand, and parasubstituted phenyl rings with electron-donating or electron-withdrawing groups to modulate the quantum confinement and energy of excitonic holes in CdSe QDs (21). Meanwhile, Jier Huang and colleagues demonstrated rapid exciton dissociation in CdSe QDs, facilitated by electron extraction and transfer to adsorbed Re-bipyridyl complexes with a swift electron transfer rate of 2.3 picoseconds, outpacing exciton annihilation in CdSe (11). These previous studies reveal that surface ligand-dependent quantum confinement regulation and electronic structure modulation are crucial in determining exciton dissociation and charge transfer in QDs. Thus, to overcome the limitation of multiple-exciton dissociation and charge delivery in perovskite QDs, surface engineering, involving the grafting of functionalized ligands onto QDs surface, is proposed in this work to mitigate excitonic coulomb interactions, decrease surface energy barriers, and occupy the capture state, thereby expectantly gaining superior photo-catalytic performance in perovskite QDs.

Fig. 1.

The band edge carrier transfer in (A) CPB and (B) CPB-FCA upon excitation. (C) Schematic representation of the exciton dissociation and electron transfer in CPB-FCA.

Ferrocene [Fe(C₅H₅)₂, labeled as Fc], discovered in 1951 by T. J. KEALY, is a sandwich-shaped molecule (22). This structure comprises two cyclopentadienyl (Cp) rings bound to an iron (Fe) atom, contributing to its robust isoelectronic 18-electron configuration (23). The interaction between the Fe atom and the Cp rings is facilitated by a “π-back bonding” mechanism, wherein the delocalized π-electrons in the Cp rings donate electron density to d-orbitals of the Fe atom, thereby stabilizing the molecule and rendering Fc electrons free to easily involved in electron transition (24). Due to its unique structure, Fc exhibits an electron-rich character in various chemical reactions (25–27). For instance, electron-donating Fc derivatives were frequently used as terminal building blocks in donor–acceptor array systems and achieved extremely long charge separation time. Imahori and coworkers achieved an extremely long-lived charge-separated state by using a ferrocene-zincporphyrin-freebaseporphyrin-fullerene tetrad which reveals a cascade of photo-induced energy transfer and multi-step electron transfer (28). Ferrocene carboxylic acid (labeled as FCA) shares a similar electronic structure with Fc, but it has a carboxylic acid (−COOH) group attached to one of its Cp rings. Due to the −COOH group, FCA can potentially replace long-chain oleic acid (OA) molecule as a microelectronic-modulating ligand to functionalize perovskite QDs for regulating the internal exciton dissociation and optimizing interface charge transfer path (schematically displayed in Fig. 1B). Regrettably, while previous studies have investigated various functional ligands associated with ferrocene derivatives, their integration into photo-catalytic processes remains limited.

In this study, a surface engineering strategy was proposed, as depicted in Fig. 1C, involving the direct grafting of the FCA ligand to unlock the restricted multiple-exciton dissociation and inefficient charge transfer dynamics of the photo-catalytic CRR on emerging perovskite QDs. The CsPbBr₃ QDs functionalized with the FCA ligand (CPB-FCA) exhibited remarkable and consistent CO yield rates of 132.8 μmol g⁻¹ h⁻¹ with an impressive selectivity of nearly 96.5%. This performance exhibited an approximate nine-fold enhancement compared to pristine CPB (14.4 μmol g⁻¹ h⁻¹) and even surpassed most halide perovskite photocatalysts. Ultrafast exciton dynamics and time-resolved photoluminescence decay were examined to investigate the exciton dissociation process. The CPB-FCA system exhibited a significant increase in negative ground state bleach (GBS) signal amplitude, along with recovery dynamics and the occurrence of an extra signal at 519 nm, indicating an additional electronic transformation pathway facilitated by FCA, which resulted in a lower exciton binding energy, as evidenced by a red shift of the photo-induced absorption (PIA) signal by 5 to 10 nm, indicating a reduction in surface barrier energy. It was demonstrated that FCA effectively operated as a dielectric screening agent while occupying the capture state on the surface of CPB QDs, reducing coulomb interactions in electrons, breaking the surface barrier energy, and triggering exciton dissociation. The decrease in surface barrier energy was further confirmed by conductive atomic force microscopy (CAFM) and Kelvin probe force microscopy (KPFM), which detected interfacial energetic and electronic dependent charge transfer properties in this unique perovskite QDs-FCA ligand complex system. The surface potential distribution, ranging from −215.8 to −181.0 mV (CPB) and −120.4 to −70.1 mV (CPB-FCA), combined with the resonance chemistry of Cp, vividly illustrated the reinforced interfacial charge delivery controlled by the surface electron distribution of FCA. Ultimately, this study revealed that manipulating quantum-confined multiple-exciton dissociation and interfacial charge delivery in semiconductor perovskite QDs, enabled by Fc-based functionalized ligands, could provide an effective strategy to enhance photo-catalytic CRR efficiencies.

Results

Synthesis and Structural Characterization.

With a few minor adjustments, the Protesescu et al. approach was used to synthesis CPB QDs of excellent quality (29). In a typical procedure, the Cs-oleate precursor and the PbBr₂ precursor were separately synthesized with OA and oleylamine (OAm). And the Cs-oleate precursor was injected into the PbBr₂ precursor at 150 °C for 5 s to obtain the as-prepared CPB QDs. The as-prepared CPB QDs were preserved in n-hexane under low-temperature conditions of 5 °C. CPB-FCA was prepared by a straightforward ligand exchange reaction with FCA, carried out under ambient air conditions. For detailed information regarding the synthesis of materials and protocols for ligand exchange, please refer to the experimental sections. The structural and morphological characteristics of pristine CPB and CPB-FCA were analyzed using transmission electron microscopy (TEM) and X-ray diffractometer (XRD), and the results are presented in Fig. 2 A–C. The TEM images in Fig. 2 A and B reveal a regular cubic structure with high crystal quality in both CPB and CPB-FCA. The high-resolution TEM (HRTEM) displays lattice fringes consistent with the d₂₀₀ spacing of CsPbBr₃ for both pristine CPB and FCA-grafted CPB QDs. No obvious change could be observed in the TEM results of the two samples, indicating the insignificant influence of FCA exchange in morphology altering. Furthermore, the XRD results in Fig. 2C demonstrate that the diffraction peaks match precisely with the cubic phase of CsPbBr₃ (JCPDS no.54-0752), proving that the perovskite crystal structure of the as-prepared samples remained intact following the FCA ligand exchange procedure.

Fig. 2.

(A and B) TEM images and (C) XRD patterns of CPB and CPB-FCA. (D) The ¹H NMR spectra and (E) the UV-Vis spectra of CPB, CPB-FCA, and FCA, respectively. Insets in (A and B) are the corresponding high-resolution TEM images. Insets in (D) include the chemical structural formula of OA and OAm.

The grafting of the FCA ligand will significantly impact the chemical and electronic properties of the original interfacial microenvironment formed by OA and OAm around CPB QDs. Thus, we conducted ¹H NMR spectra to investigate the surface chemistry changes of CPB QDs before and after FCA grafting. As shown in Fig. 2D, the ¹H-NMR spectrum of pristine CPB QDs exhibits prominent resonances associated with OA and OAm molecules used during the synthesis process (30). Specifically, the characteristic protons resonances from ligands (e.g., 1 and β for OA and OAm, respectively), the vinylene protons (5 and ε), the methylene protons (4 and δ), the methylene groups (3), and the methyl resonance (6) can all be clearly distinguished in the pristine CPB QDs (31, 32). After FCA grafting, the ¹H-NMR spectrum features of CPB-FCA change drastically when compared to pristine CPB QDs. Specifically, the peculiar protons linked with OA were replaced by the Fc protons (δ = 4.2 to 4.6 ppm) connected with the FCA ligand in the case of CPB-FCA, while the peculiar protons relating to OAm were mostly unchanged. This ¹H-NMR spectroscopy confirmed the successful replacement of native OA ligands with FCA ligands while preserving the OAm ligands. Inductively coupled plasma optical emission spectroscopy experiments were performed using an Fe 2187 spectrometer to determine the content of Fe. As shown in SI Appendix, Table S1, the molar content of Fe in CPB-FCA is about 0.6%. We also employed ultraviolet and visible absorption (UV-Vis) spectroscopy to double-verify FCA grafting by examining CPB, CPB-FCA, and FCA absorption spectra. In Fig. 2E, the characteristic absorption bands at 247 nm were attributed to the π-π* transition of monosubstituted ferrocene in FCA (33), providing evidence for the successful FCA exchange. Moreover, the significant increase in absorption intensity confirms the efficient electronic modulation of FCA grafting on CPB QDs, which will benefit the following photo-catalytic process. And this remarkable electronic modulation behavior will be analyzed and discussed in detail in the following paragraphs.

Photo-induced Ultrafast Electron Transition.

The significant absorption enhancement after FCA grafting greatly stimulated our interest in further exploring the mechanism of electronic modulating on FCA-grafted CPB QDs. Thereout, we conducted fs-TAS to meticulously investigate the ultrafast exciton dynamics in CPB and CPB-FCA. The experiments were performed using a 400 nm excitation (3.1 eV) and a pump power of 25 μW. Fig. 3 A–F presents the 2D pseudocolor TA plots and TA spectral profiles with varying time delays for CPB and CPB-FCA, respectively. SI Appendix, Fig. S1 compares the GSB recovery kinetics between CPB and CPB-FCA. As shown in the 2D pseudocolor TA plots of CPB (Fig. 3A) and CPB-FCA (Fig. 3B), the spectral features observed in both samples are consistent with earlier research and include a negative GSB near the bandgap resulting from the band-filling effect (both electrons and holes contribute to the GSB signal) (34–37), a long-lived positive photo-induced absorption (L-PIA) at higher energies due to carrier absorption in the band-edge state (34, 37, 38), and a short-lived positive PIA (S-PIA) below the bandgap attributed to bandgap renormalization (37, 38). These signals are mainly influenced by state filling and Coulomb interactions (39). The state-filling-induced TA signals provide information about carrier populations in QDs, while the effects of Coulomb interactions on TA spectra can be explained by the Stark effect associated with local fields generated by photo-excited carriers (39). Generally, when a 25 μW power density pump-probe at 400 nm excited the CPB QDs, the initial time (0 fs in Fig. 3C) showed a GBS growth process with GSB feature at 502 nm and two PIA features at 445 ~ 470 nm (L-PIA) and 519 ~ 525 nm (S-PIA). The GSB and PIA signals can be explained as the state-filling-induced bleach and the red shift of the lowest excitonic state caused by hot excitons, respectively (36). Besides, Fig. 3 C and E show the growth and recovery process of the GBS signal for CPB, respectively. Hot carriers (HCs) relax to the lowest-energy band-edge states and finally recover at the end, causing a strong GSB signal to take the place of the S-PIA signal, which causes the initial intensity of the GSB signal to increase to its maximum (0.15 ~ 2 ps in Fig. 3C). The biexciton-induced stark effect, which is often responsible for the derivative-like spectral characteristic at early delay times, results from the Coulomb interaction between the hot exciton and band-edge exciton (35–37). The L-PIA appears due to the absorption of carriers from band edge states. The CPB-FCA shows analogous GSB and PIA features in Fig. 3D, while the performance parameter varies. We can observe that the negative amplitude of the GSB signal of CPB-FAC is about twice as high as that of CPB QDs at a 503 nm period. The two PIA signals displayed notable red-shift to 445 ~ 475 nm and 529 ~ 530 nm (0.15 ~ 2 ps in Fig. 3D), respectively. The leach amplitude of CPB-FCA QDs increases systematically, indicating the increase of photoexcited carriers in CPB-FCA QDs. SI Appendix, Fig. S1 demonstrates the overall GBS dynamic change process of CPB and CPB-FCA. It suggests that FCA could be a powerful dielectric screening leading to effective exciton dissociation, which is associated with the stronger GSB signal of CPB-FCA QDs (39). The comparable red-shift in the PIA band suggests the renormalization of the whole band structure of CPB-FCA QDs. This absorption peak shift originates from bandgap renormalization (BGR) mediated by the band-filling effect in CPB QDs through a photo-excited charge transfer from the FCA and pump-induced photoexcitation in CPB QDs (40, 41). The pump-induced photoexcitation should be the same because the pump fluence was restricted during the experiment. When combined with the stronger GSB signal, the red-shift of PIA can be attributed to the robust photo-excited charge transfer from FCA.

Fig. 3.

TA spectra and kinetics of CPB and CPB-FCA QDs. 2D pseudocolor map of (A) CPB and (B) CPB-FCA. Comparative GSB growth and GSB recovery dynamics of (C and E) CPB and (D and F) CPB-FCA.

We further compared the GSB and PIA kinetics of CPB and CPB-FCA QDs to characterize electronic transportation (SI Appendix, Figs. S2 and S3). All the kinetics fitting parameters and errors are listed in SI Appendix, Table S2. The rise time of GSB feature in CPB QDs (scaled and inverted for comparison) was 0.6 ± 0.06 ps. This time corresponding to the hot carrier relaxation time into the lowest excitonic state. Furthermore, it exhibited decay characteristics with time constants and amplitudes of 4.774 ± 0.63 ns and 31.6 ± 1.4%, as well as 65.88 ± 4.5 ps and 68.4 ± 1.7%. Here, the radiative recombination of the electron and hole causes the longer component in this case, whereas the trap state-mediated radiative process causes the shorter component (37, 41, 42). In the GBS feature of CPB-FCA QDs, there is a faster rise time of 0.48 ± 0.04 ps and decayed time constants of 1.069 ± 0.16 ns (22.9 ± 1.8%) and 45.69 ± 2.6 ps (77.1 ± 1.8%). The shorter components (τ₁ and τ₂) become faster, and a significant amplitude enhancement is noticed after the treatment of FCA, indicating the more intense trap state-mediated electronic transportation. Additionally, as seen in Fig. 3F, the spectra of the CPB-FCA display a newly formed GBS feature at 519 nm. This feature is in good agreement with the UV-Vis spectum of FCA and can be attributed to the depletion of FCA molecules in their ground state during the GBS decay process. Hence, the surface trap state-mediated electronic transportation is also can be regulated by FCA. In SI Appendix, Table S2 the GSB feature of CPB QDs (scaled and inverted for comparison) formed with a rise time of 0.6 ± 0.06 ps, corresponding to the hot carrier relaxation time into the lowest excitonic state, and decay with time constants and amplitudes of 4.774 ± 0.63 ns and 31.6 ± 1.4%, 65.88 ± 4.5 ps and 68.4 ± 1.7%. Here, the longer component is due to the radiative recombination of the electron and hole, and the shorter component is due to the trap state-mediated radiative process (37, 41, 42). In the GBS feature of CPB-FCA QDs, the quicker rising time is 0.48 ± 0.04 ps and the decaying time constants are 1.069 ± 0.16 ns (22.9 ± 1.8%) and 45.69 ± 2.6 ps (77.1 ± 1.8%). The shorter components (τ₁ and τ₂) become faster, and a significant amplitude enhancement is noticed after the treatment of FCA, indicating the more intense trap state-mediated electronic transportation.

Interfacial Electron Transportation between CPB and FCA.

We conducted a comprehensive analysis of the localized electronic behavior by employing CAFM. This detailed investigation aimed to delve deeper insights into the effects of incorporating FCA on the dynamics of interfacial electron transportation. Typically, the surface morphologies of all samples were measured by AFM in tapping mode with a Fe slice as the bottom electrode, applying a bias voltage of 1 V between the needle tip and the sample and recording the current value between the conducting probe and the sample. SI Appendix, Fig. S4A illustrates the 2D plot of the local current distributions of CPB, and the corresponding current distributions along the yellow line exhibit fluctuations around −0.05 to 0.05 nA (depicted as a blue line in SI Appendix, Fig. S4C). As shown in SI Appendix, Fig. S4 B and C, CPB-FCA demonstrates a significant increase in interfacial current compared to CPB, ranging from −0.08 to 0.1 nA. These results indicate that FCA can efficiently strengthen molecular conductance, promoting interparticle charge delivery in CPB-FCA through enhanced interfacial electron transfer. This charge delivery is further supported by the transient photocurrent responses, which exhibit a more robust photo-response current in CPB-FCA compared to pristine CPB (Fig. 4A). To provide additional confirmations of electron transfer between CPB and FCA, we compared the photoluminescence intensity and normalized time-resolved photoluminescence decay process of CPB and CPB-FCA. The Inset of Fig. 4B indicates that the photoluminescence intensity of CPB-FCA is lower than that of pure CPB due to the addition of FCA, which provides an alternative energy-transfer pathway alongside the intrinsic radiative channel for excited-state electron transfer. Furthermore, the normalized time-resolved photoluminescence decay graphs following pulsed laser stimulation at λ = 373.4 nm, as shown in Fig. 4B, corroborate the rapid separation of charges facilitated by FCA. All the decay traces were fitted well with bi-exponential functions, and the result is summarized in SI Appendix, Table S3. It is noteworthy that the photoluminescence of CPB-FCA decays noticeably faster than that of pure CPB, with average photoluminescence decay times (τ_average) for CPB-FCA and CPB being 0.35 and 4.34 ns, respectively. The photoluminescence quenching (SI Appendix, Fig. S5) and the shortening of decay time of CPB in the presence of FCA are definitely because of the energy or charge-transfer (electron or hole transfer) process (36, 43, 44). Therefore, the combined findings from TA, CAFM, transient photocurrent response, and photoluminescence characterizations collectively provide evidence of intense electron transfer between CPB and FCA. However, how do the electrons transfer and what is the reason behind FCA providing such incredible benefits for electron transfer?

Fig. 4.

(A) Transient photocurrent responses, (B) time-resolved fluorescence emission decay spectra, and (C) XPS spectra of CPB and CPB-FCA, respectively, for the Cs 3d core-level regions indicating the presence of the alumina shell on the QDs surface. (D) Band alignment of pure CPB and FCA indicating electron transfer from FCA to CPB and hole transfer from CPB to FCA. (E) Schematic representation of the surface electron transfer and the band edge carrier transfer from FCA molecules to CPB QDs upon excitation.

The utilization of KPFM for surface potential distribution analysis has provided further elucidation regarding the specific role of FCA in enhancing charge separation dynamics. The surface potential distribution images for CPB and CPB-FCA, depicted in SI Appendix, Fig. S6, show values of −215.8 ~ −181.0 mV and −120.4 ~ −70.1 mV, respectively. These data illustrate that introducing FCA leads to a two-fold increase in the potential CPB, breaking the surface barrier and facilitating electron transfer from FCA to CPB. Moreover, X-ray Photoelectron Spectroscopy (XPS) spectra have confirmed significant alterations in electron density within the Cs atoms of CPB subsequent to the inclusion of FCA. This modification is distinctly observed in both Fig. 4C and SI Appendix, Fig. S7, demonstrating a shift toward lower electron binding energies in the inner shell of Cs, Pb, and Br atoms post-FCA introduction. This shift strongly implies an electron transfer from FCA to CPB, facilitated by the coordination of FCA with CPB through the Cs atom. Additionally, the band alignment of the CPB was obtained from UV-Vis and Mott-Schottky analysis (SI Appendix, Fig. S8) to understand the exact nature of the charge transfer phenomenon. The CPB is an n-type semiconductor with a redox potential of the conduction band (CB) and valence band (VB) located at −3.95 eV and −6.2 eV, respectively, versus the absolute vacuum scale (45). The highest unoccupied molecular orbitals (HOMO) and the lowest occupied molecular orbitals (LUMO) energy levels were determined from previous research to be approximately −5.07 eV and −2.27 eV (46). Fig. 4D demonstrates the energy-level alignment of CPB with respect to FCA. The HOMO of FCA lies above the VB of the CPB, suggesting that the hole transfer process is thermodynamically feasible from CPB to FCA. The position of the LUMO of FCA lies below the CB of CPB, indicating that the electron transfer process from FCA to CPB is thermodynamically feasible as well (8, 34, 35). Based on the analysis of band structure, work functions, XPS and KPFM, the photogenerated electrons would spontaneously transfer from FCA to CPB until their Fermi levels reach equilibrium.

Fig. 4E presents a probable underlying mechanism regarding molecular electronic structure in our CPB-FCA system. It is widely recognized that the −COOH group demonstrates an electron-withdrawing resonance effect, wherein electrons from the carbon-oxygen double bond are delocalized toward the oxygen atom (47). In our system, the Cp ring allows a possible electron density shift because of the composition of five π electrons, which can move in conjugated systems (48). As a result, electrons in FCA are withdrawn from the Cp ring, leading to a redistribution of electron density (step I-III in Fig. 4E). Upon grafting FCA onto CPB, the coupling between the Cs atom and −COOH group of FCA establishes an electronic bridge. The electron-rich anchoring site exerts a Coulombic attraction on holes of CPB, which is corroborated by the position of the HOMO of FCA (SI Appendix, Fig. S9). Moreover, SI Appendix, Fig. S10 provides evidence for the direction and pathway of electron transfer by analyzing differential charge density using DFT calculations. The blue and yellow regions in this analysis indicate charge accumulation and depletion, respectively. Specifically, a distinct electron interaction was observed at the interface between FCA and CPB (SI Appendix, Fig S10B), transfer from FCA to CPA (step IV in Fig. 4E). However, the original OA ligand on pristine CPB significantly restricts this interaction (SI Appendix, Fig. S10C). Based on the discussions above, it can be concluded that intramolecular charge transfer regulated by FCA triggered multiexciton dissociation in CPB QDs, potentially leading to improved CO₂ reduction performance. And Cs atom is the potential active site for CO₂ reduction.

Simulated Solar Light-Driven CO₂ Reduction Properties and Mechanism.

Under AM 1.5G simulated solar illumination, the gassolid photo-catalytic CRR performance of CPB and CPB-FCA was evaluated. Prior to the performance, a series of preliminary control experiments were conducted. Fig. 5A shows that no products were detected when the reactants were exposed to the photocatalyst in the absence of light irradiation or under light irradiation without a photocatalyst, which confirms the essential role of both the photocatalyst and light irradiation in photo-catalytic CRR of this testing system. The CO and C₂H₄ production rate of CPB-FCA is 132.8 μmol g⁻¹ h⁻¹ and 1.6 μmol g⁻¹ h⁻¹ during the 5-h reaction period, as displayed by the blue and yellow bar in Fig. 5A. Furthermore, Quantum Efficiency (QE) calculations were performed for the catalysts. The QE% for CO production using CPB-FCA QDs amounts to 0.072%, demonstrating comparability to natural photosynthesis. Notably, our CPB-FCA QDs exhibit a superior QE% in a gas-solid system when compared to previously reported QDs functioning without cocatalysts or sacrificial reagents (49–53). Conversely, pure CPB displays a notably lower CO production rate of 14.4 μmol g⁻¹ h⁻¹ and a QE of 0.006%, indicating its inferior performance relative to CPB-FCA. Additionally, the photo-catalytic reaction does not manifest activity solely with FCA in catalyzing CRR, thereby positioning FCA as a charge transfer medium. The confirmation of water (H₂O) oxidation, resulting in the generation of oxygen gas (O₂), was achieved through isotope-tracing control experiments. These results are presented in SI Appendix, Fig. S11 A and B, where mass-to-charge ratio (m/z) values were used to distinguish between different combinations of oxygen isotopes: ¹⁶O, ¹⁸O, ¹⁶O¹⁶O, and ¹⁸O¹⁸O. SI Appendix, Fig. S11B, illustrates that following 10 h of continuous light exposure, there was a significant increase in the relative abundances of ¹⁸O, ¹⁶O¹⁸O, and ¹⁸O¹⁸O compared to their initial levels (SI Appendix, Fig. S11A). This marked increase provides strong evidence for the involvement of oxygen derived from water oxidation. Moreover, to trace the source of carbon monoxide (CO) produced during the photo-catalytic conversion of CO₂ on CPB-FCA, we conducted experiments using isotope-labeled ¹³CO₂, as demonstrated in Fig. 5C. The primary product resulting from the reduction of CO₂ was ¹³CO, with minimal detection of ¹²CO after 10 h of reaction. Furthermore, to verify their stability, long-term catalytic stability tests for CPB and CPB-FCA (Fig. 5B and SI Appendix, Fig. S12) and cycling experiments of CPB-FCA (Fig. 5D) were conducted. The CO production amount continues to increase up to 72 h of reaction time, reaching 1,623 μmol g⁻¹ in the first 20 h and 2,107 μmol g⁻¹ at 72 h. Furthermore, SI Appendix, Fig. S13 displays the XRD results for the catalyst after 10 h of irradiation, illustrating a reduction in the crystallinity while maintaining the fundamental phase. These results demonstrate that CPB-FCA exhibits effective and stable CO₂ photoreduction capabilities. Moreover, compared to other CPB-based photocatalysts and other perovskite QDs under similar experimental conditions, our CPB-FCA displays a record-high CO production rate, which were summarized in SI Appendix, Table S4 and shown in Fig. 5E. Therefore, it can be inferred that CPB-FCA possesses significant advantages for CO₂ photoreduction.

Fig. 5.

(A) Performance of photocatalytic CO₂ reduction for 5 h under various reaction conditions. (B) CO plots over CPB and CPB-FCA for the 18-h stability test. (C) Mass spectra of ¹³CO₂ isotope experiments in the presence of CPB-FCA. (D) Photocatalytic cycling experiments on CPB-FCA. (E) Comparison of the photocatalytic CO₂ reduction activity of various perovskite QDs-based photocatalysts.

Reaction mechanism analysis reveals that the FCA ligand demonstrates dual functionality by acting as a critical modulator for exciton dissociation and interfacial charge delivery, while also playing a crucial role in the absorption and activation of CO₂ in the photoreduction process. As shown in SI Appendix, Fig. S14, CPB-FCA exhibits a higher CO₂ absorption capacity compared to pristine CPB. Besides, the observed pore size distribution in SI Appendix, Fig. S15 demonstrates that the diameter of the most concentrated micropores in the adsorbent, referred to as the DA, is several micropore diameters for CPB (6.020 nm) and CPB-FCA (1.060 nm) separately, suggesting that the sandwich-like structure of FCA may play a significant role in facilitating CO₂ capture by leveraging the steric hindrance effect and maximizing the utilization of surface area and basic active sites. In situ DRFTS was conducted on pristine CPB and CPB-FCA to confirm the specific CO₂ photoreduction process by detecting time-resolved intermediates. Fig. 6 A and B show an overall evolution of peaks with increasing CO₂ adsorption time of CPB and CPB-FCA, respectively, where the CPB-FCA exhibit a more intensified CO₂ activation ability than that of CPB while with the same adsorption configuration. Specifically, the increased peaks of HCO₃⁻ (1,476 cm⁻¹), ·CO₂⁻ (1,676 ~ 1,620, 1,252 cm⁻¹) are attributed to the enhanced CO₂ adsorption facilitated by the introduction of FCA (54, 55). In Fig. 6 C–F, the reaction intermediates during the CO₂ reduction process of CPB-FCA and CPB were observed as the irradiation time extended. Specifically, COOH* and *CO were detected at 1,620 cm⁻¹ and 2,022 cm⁻¹, which are commonly recognized as critical intermediates in the conversion of CO₂ to CO (5, 56). Notably, these intermediates exhibited more pronounced changes in the CPB-FCA system as well. Furthermore, DFT calculations validate the preferential binding of CO₂ molecules to the Cs atom in CPB-FCA, exhibiting a significantly lower adsorption energy of −2.5256 eV, as illustrated in SI Appendix, Fig. S16. This adsorption energy is approximately half that of CPB and CPB-OA, underscoring the heightened CO₂ adsorption and activation capabilities of CPB-FCA. SI Appendix, Fig. S17 shows the electronic location function results of different CO₂ adsorptions on CPB, OA-capped CPB, and FCA-capped CPB. It demonstrates that the sandwich construction creates natural CO₂ capture sites between FCA and the Cs atom of CPB increasing CO₂ absorption and activation by trapping CO₂ between QDs and FCA. As concluded above, the significant improvement in photo-catalytic CO₂ conversion efficiency on CPB-FCA can be ascribed to the unique adsorption activation capability and efficient exciton dissociation facilitated by FCA and Cs atom is the active site for both CO₂ adsorption and electron transform.

Fig. 6.

In situ DRIFTS spectra for (A and B) adsorption and (C–F) reaction during CO₂ photoreduction on CPB-FCA and CPB.

Discussion

In summary, we report CPB QDs functionalized with rarely reported FCA ligands. Our CPB-FCA demonstrated more robust exciton dissociation and charge delivery ability compared to CPB during the electron kinetics. The direct observation of characteristic GSB and PIA features in both CPB and CPB-FCA, upon excitation of the sample with a pump-probe at 400 nm and a power density of 25 μW, provides solid evidence for the FCA-modulated electronic behaviors of state filling and Coulomb interactions in CPB QDs. And the overall GBS dynamic change process of CPB and CPB-FCA suggests that FCA could be a robust dielectric screening leading to effective exciton dissociation, which is indicated by the twice stronger GSB signal of CPB-FCA QDs. When combined with the stronger GSB signal, the red-shift of PIA can be attributed to the robust photo-excited charge transfer facilitated by electron-rich FCA. Moreover, the new additional GBS feature at 519 nm in the CPB-FCA spectra agrees well with the UV-Vis spectra of FCA. It can directly confirm that the surface trap state-mediated electronic transportation is also can be regulated by FCA. Combining the results of the TA, time-resolved photoluminescence decay, AFM (CAFM and KPFM), and XPS spectral, FCA serves to regulate the electronic transportation that is mediated by the surface trap state, while in the meantime, electron-rich FCA could produce an effective dielectric screening that triggers the exciton dissociation. In addition, the electron-rich FCA helped break the surface tension barrier for electronic transport, which improved hot electron utilization. As a result, the CPB-FCA exhibits an impressive and consistent photo-catalytic CRR performance of 132.8 μmol g⁻¹ h⁻¹. The observed diversity of exciton dynamics in CPB resulting from charge transfer regulated by FCA offers more possibilities for efficiently using the multiple excitons generated by perovskite QDs for better performance in photocatalysis. This FCA-related surface engineering also allows for manipulating quantum confinement and modulation of electronic structure in perovskite QDs, which plays a critical role in controlling exciton dissociation and charge transfer, thus effectively overcoming the limitations associated with multiple-exciton annihilation and charge delivery frustration and providing guidance in broader catalytic applications of perovskite and/or other QDs.

Materials and Methods

Materials and Fabrication.

Cesium carbonate (Cs₂CO₃, 99.9% metals basis), lead bromide (PbBr₂, 99%), octadecene (ODE, 90%), OA (90%), OAm (80-90%), ferrocene carboxylic acid (C₁₁H₁₀FeO₂, 98%), ethyl acetate (99%), and n-Hexane (C₆H₁₄, 97%) were purchased from Aladdin Chemical Reagent Co., Ltd. Toluene (99%) and acetone (99%) were all obtained from Sinopharm Chemical Reagent Co., Ltd.

CPB QDs were synthesized using a modified hot-injection method. First, a mixture of Cs₂CO₃, OA, and ODE was heated to obtain the Cs-oleate precursor. The precursor was deaerated, annealed, and transformed into a light yellowish transparent colloidal Cs-oleate precursor. Second, PbBr₂ was reacted with OA and OAm to generate the PbBr₂ precursor. After purging with Ar, the temperature of the PbBr₂ precursor was increased. Then, 1.2 mL of the Cs-oleate precursor was injected into the PbBr₂ precursor, and the reaction lasted for 5 s at 150 °C. The synthesis was completed by cooling in an ice bath.

Ligand exchange was conducted under air. The crude CPB QDs (5 mL) were mixed with anhydrous n-hexane solution of hydrobromic acid (1 mL, 0.035 M). After stirring for 60 min, the QDs were washed with ethyl acetate, centrifuged at 10,000 rpm for 10 min, and re-dispersed in toluene.

Characterization.

QD size and morphology were characterized by TEM (JEM-2010). XRD spectra were obtained using a Rigaku D/max-2500 XRD. ¹H NMR spectra were recorded on a Bruker 400M. UV-Vis spectra were collected with a Shimadzu UV-Vis DRS (UV2550). PL spectra were measured using a Hitachi F-7000 spectrometer.

Evaluation of Photo-catalytic Activity.

The Labsolar-6A system (Perfect Light Co., China) was used to assess photo-catalytic CO₂ reduction. Gas chromatography (GC), a 300 W Xe lamp, and a quartzose reactor are all components of the system. The reactor temperature was kept at 20 °C by circulating water, and the Xe lamp had an AM 1.5G filter to replicate solar light with a 100 mW cm⁻² optical density. A glass fiber membrane was coated with 1.0 mL of toluene mixed with 3 milligrams of photocatalyst for 5 h reaction. The catalyst dosage employed for the extended 72-h experiment was 6 milligrams. Extra toluene was removed using vacuum drying. CO₂ (99.99%) and H₂O vapor were created by bubbling CO₂ through water (298 K). After being sealed at 101 kPa, the reactor was purged for 1 h with this combination. Using GC equipped with a flame ionization detector and a thermal conductivity detector, gaseous products were examined hourly. Selectivity for CO₂ reduction = 2R(CO)/Relectron × 100% [Relectron is the rate of electron consumption for the reduced product; Relectron = 2R(CO) + 8R(CH₄) + 12R(C₂H₄)]. The initial 5-h performance within the 72-h photocatalytic activity was utilized for Quantum Efficiency calculations.

Electrochemical Tests.

Three electrodes on a CHI660E workstation were used for the electrochemical experiments. Pt mesh functioned as the counter electrode, Ag/AgCl (saturated KCl) served as the reference electrode, and the sample-loaded photoelectrodes served as the working electrode. With 0.1 M TBAPF6, ethyl acetate served as the electrolyte.

AFM Tests.

AFM (KPFM and CAFM) was measured using a Bruker Dimension FastScan. In the measurement protocol, ensuring proper electrical conductivity between the sample and the substrate stands as a crucial aspect. For this study, iron sheets were utilized as the substrate material. The sample preparation procedure includes depositing the liquid sample onto the iron sheet and subsequently exposing it to infrared light dry treatment to facilitate the acquisition of data.

DFT Calculations.

The Vienna Ab Initio Simulation Package (VASP5.4), with a generalized gradient correlation functional, was used to carry out the spin-polarized DFT computations. 3 × 3 × 1 K points were used to sample the Brillouin zone. The plane wave cutoff energy was set at 400 eV. Except for atoms at the boundary, all atoms converged to 0.03 eV/Å. The Bader method was used to determine the volume of exchanged electrons (Δq). G = EDFT + EZPE − TS, where EDFT stands for the electronic energy computed with VASP, EZPE stands for zero-point energy, and TS stands for the contribution of entropy, was used to calculate the Gibbs free energies of adsorption (ΔGads) and desorption (ΔGdes) at 298.15 K.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix including related catalytic performance tests, materials characterization, Kelvin Probe Force Microscopy, Transient absorption spectroscopy, and mechanism investigation.