Xu
Hu‡
a,
Zhijie
Zhu‡
a,
Yuxuan
Zhou‡
a,
Shuang
Liu
a,
Chunpeng
Wu
a,
Jiaqi
Wang
a,
Yihao
Shen
a,
Tianran
Yan
a,
Liang
Zhang
ab,
Jinxing
Chen
ab,
Kai
Feng
ab,
Alexander
Genest
c,
Günther
Rupprechter
c,
Xingda
An
*ab,
Chaoran
Li
*ad and
Le
He
*ab
aInstitute of Functional Nano & Soft Materials (FUNSOM), Soochow University, Suzhou, 215123, Jiangsu, PR China. E-mail: xdan@suda.edu.cn; crli@suda.edu.cn; lehe@suda.edu.cn
bJiangsu Key Laboratory of Advanced Negative Carbon Technologies, Soochow University, Suzhou, 215123, Jiangsu, PR China
cInstitute of Materials Chemistry, Technische Universität Wien, Vienna 1060, Austria
dJiangsu Key Laboratory for Carbon-Based Functional Materials & Devices, Soochow University, Suzhou, 215123, Jiangsu, PR China
First published on 4th April 2024
Here we present an effective strategy to achieve strongly enhanced catalytic activity of platinum-copper bimetallic clusters through augmented plasmonic photochemical effects of an aggregated nanostructure. The excitation by light irradiation significantly amplified both the population and migration rate of hot carriers generated on the catalyst surface, thereby achieving a record-high reduction rate of p-nitrophenol at elevated concentration. This work provides important insights for the rational design of efficient metallic photocatalysts; it is also relevant for further improvement of the efficiencies of plasmonic photocatalysis and solar energy harvesting.
To further improve catalytic performance, reducing the catalyst particle size has become a commonly employed strategy in heterogeneous catalysis.23 While the reduction in size increases the specific surface area, enhances the exposure of reaction sites on the catalyst surface and lowers the energy barriers for reactions,24–26 it poses unique challenges in plasmonic photocatalysis. The decrease in nanoparticle size is typically accompanied by a simultaneous decrease in absorption cross-section, a notable attenuation of hot carriers and a blueshift of their resonance frequency.27,28 These factors impede effective sunlight capture, resulting in a substantial decline in photochemical effects. Therefore, it remains an urgent need but great challenge to develop metallic plasmonic photocatalysts with both high reactivity and strong photochemical effects.
Herein, we present an aggregated plasmonic Pt–Cu bimetallic cluster photocatalyst fabricated through a simple wet-chemical method. We employed the p-nitrophenol (p-NP) reduction reaction as a model reaction to probe the intrinsic catalytic properties and photochemical effects of Pt–Cu cluster aggregates CAs. Photoelectrochemical test methods and in situ characterization were utilized to deepen our understanding of the mechanism the photochemical effects. The photocatalysts exploit the inherent structure consisting of aggregates of small-sized particles to enhance the overall light-absorption properties of the catalysts, while simultaneously maintaining high intrinsic catalytic performance. This resolves the contradiction that arises when attempting to achieve both high intrinsic catalytic activity and strong photochemical response, paving the way for the construction of efficient plasmonic photocatalysts.
Ideal plasmonic photocatalysts should possess both high catalytic performance comparable to that of small-sized particles and strong photochemical effects similar to those of large-sized particles. According to the surface plasmon and molecular orbital theories, when many small particles are closely assembled to form a larger structure, the plasmonic hybridization effect broadens the overall light responsivity, thereby enhancing the photochemical efficiency of catalysts.29,30 Simultaneously, these small particles may retain their high intrinsic catalytic performance. On these grounds, we hypothesize that aggregated cluster catalysts possess great promises as excellent photocatalysts (Fig. 1a).
To test this hypothesis, we prepared Pt–Cu bimetallic cluster aggregates (Pt–Cu CAs) using a wet-chemical synthesis method. Small particle aggregates with the similar morphology and structure but different Pt–Cu ratios were synthesized by controlling the ratio of metal precursors, denoted as Pt1–Cu2 CAs, Pt1–Cu8 CAs and Pt1–Cu12 CAs (Fig. S1†). The Pt–Cu atomic ratio of Pt1–Cu8 CAs, as determined by inductively coupled plasma mass spectroscopy (ICP-MS), was 12:88. Transmission electron microscopy (TEM) images clearly reveal the structure, consisting of nanocrystalline particles with a rounded shape around 3 nm, coupled with each other, resulting in an overall size of approximately 20 nm (Fig. 1b). The porous nature of the structure maintains a high specific surface area, while a large number of atoms at the edges/corners provide abundant active sites, suggesting high intrinsic surface reactivity. Elemental distribution in the bimetallic nanocrystalline particles was characterized using Energy-Dispersive Spectrometry (EDS) elemental mapping (Fig. 1d–g), confirming uniform distribution of Pt and Cu in each nanoparticle and cluster aggregate. Fig. 1h displays the X-ray diffraction (XRD) of Pt1–Cu8 CAs. It can be seen that the (111), (200) and (220) diffraction peaks were located between the diffraction peaks of the pure face-centered cubic (fcc) Pt (JCPDS: 65-2868) and Cu (JCPDS: 65-9743) crystalline phases, and no characteristic peaks of Cu or its oxides (e.g., Cu2O and CuO) have been detected, indicating the formation of Pt–Cu alloy nanocrystals. Representative high-resolution transmission electron micrographs (HRTEM) reveal lattice fringes with an interplanar spacing of 0.21 nm corresponding to the (111) planes of the fcc Pt–Cu alloy (Fig. 1c).31
The bonding state between the Pt–Cu bimetallic elements was investigated using X-ray Photoelectron Spectroscopy (XPS). The Pt valence electron spectrum in Fig. 1i shows three pairs of peaks attributed to metallic platinum (Pt0), Pt2+, and Pt4+, suggesting some surface oxidation due to the small size of alloy particles.32 In contrast, the Pt element in Pt–Cu alloys exhibited a stronger affinity towards metallic state in the Extended X-ray Absorption Fine Structure (EXAFS) map (Fig. S2†). Due to the difficulty of distinguishing between metallic copper (Cu0) and copper oxides (Cu2O and CuO) using XPS, the chemical state of the Cu element in Pt–Cu alloys is typically determined by Auger Electron Spectroscopy (AES). The AES spectrum of Cu in Fig. 1j can be accurately fitted with a curve consisting of three peaks corresponding to metallic copper (Cu0), Cu1+, and Cu2+. The position of the LMM peak for elemental Cu was significantly higher than that of metallic Cu, indicating that elemental Cu was more biased towards the oxidation state in the Pt–Cu alloy. These results confirm the steady-state electron migration from Cu to Pt in the Pt–Cu bimetallic cluster aggregates, which is caused by the higher Fermi energy level of Cu compared to Pt.
The favourable electronic and structural modulation in the Pt–Cu cluster aggregates hold great promise in catalysis. To verify their photocatalytic properties, p-NP degradation was used as a test reaction. As a highly toxic compound, p-NP can cause damage to the human liver, kidney, and nervous system. Current methods for removing p-NP can be categorized into physical, biological, and chemical approaches. Among them, physical and biological methods are gradually being phased out due to stability and secondary pollution issues. The catalytic reduction of p-NP into an important pharmaceutical and chemical intermediate, p-aminophenol (p-AP), aligns with the current mainstream concept of green environmental protection.33–36 In this study, we evaluate the catalytic performance of Pt–Cu CAs by reducing p-NP to p-AP with borohydride. According to the Beer–Lambert law, the concentration of p-NP is proportional to the intensity of the absorption peak at 400 nm. Notably, during the catalytic reduction by Pt–Cu CAs, the characteristic absorption peak of p-NP gradually disappears while an absorption peak centered at about 300 nm appears, which can be attributed to the absorption feature of the p-AP product.37,38 Synthesized cluster aggregates with the same morphology structure but different Pt–Cu ratios showed that Pt1–Cu8 CAs exhibited the highest catalytic activity for the reaction in dark (Fig. S3–S4†). Moreover, under the same light intensity (100 mW cm−2) the catalytic performance of Pt1–Cu8 CAs exhibited significant enhancement when compared to the other two counterparts (Fig. S5 and S6†). These results indicate that the appropriate ratio of bimetallic components is decisive on the maximization of the bimetallic synergistic effect.39–42
We next investigated the influence of different p-NP concentrations on the reaction rate using Pt1–Cu8 CAs. Typical catalysts for the p-NP reduction reaction shows rapid decrease in catalytic performance and even suspension of the reaction process as the reactant concentration increases, which limits practical industrial applications. Our results reveal that although higher reactant concentrations still led to much longer reaction times, yet at a commonly used p-NP concentration of 1 × 10−4 mol L−1, the reactivity of Pt1–Cu8 CAs reaches an excellent turnover frequency (TOF) of 91 mmol g−1 min−1, setting a new record in p-NP reduction reaction rate (Fig. 2a and b, Fig. S7 and Table S1†). The structure of the cluster aggregates appears to remain unchanged during the catalytic reaction. The presence of lattice structures observed within the cluster boundaries in the HRTEM images suggests that the clusters are not completely isolated within the aggregates. This observation may provide an explanation for the stability exhibited by the aggregated structure (Fig. S8†).
To elucidate the advantages of the cluster aggregates, which combine strong photochemical effects with high catalytic activity, we next investigated control groups including larger particles of the same aggregates size (Pt1–Cu8 LPs) and smaller particles of the same monomer size (Pt1–Cu8 SPs) for comparative analysis (Fig. S9†). In the dark environment, both small-sized particle samples and aggregates exhibited higher intrinsic catalytic activity than large-sized metal nanocrystalline particles (Fig. 2c). The Pt1–Cu8 CAs and Pt1–Cu8 SPs expose a greater number of active sites to the reaction environment, thereby ensuring high intrinsic catalytic performance. The catalytic reaction rate of Pt1–Cu8 LPs and Pt1–Cu8 CAs was enhanced nearly four-fold under a visible light intensity of 50 mW cm−2, whereas Pt1–Cu8 SPs exhibited no significant alteration. This demonstrates that the aggregates’ structure significantly improves the photoresponsivity of small-sized metal particles, achieving high catalytic activity (Fig. 2c and Fig. S10†). Current densities tested by cyclic voltammetry further validate the advantages of nanocrystalline particle aggregates with larger photocurrent densities and earlier onset potentials (Fig. S11 and S12†). Pt–Cu CAs are not only limited to the reduction reaction of p-NP, but also show remarkable photo-enhancement in the reduction of 2-nitrophenol and 3-methyl-4-nitrophenol, which further demonstrates the universality of our strategy (Fig. S13†).
To confirm that the enhancement in light-driven catalytic activity stems from the photochemical effect, we designed catalytic reaction tests under different light intensity irradiations from 0.5 suns to 3 suns (1 sun = 100 mW cm−2). As expected, the results showed a first order linear relation between external light intensity and catalytic activity (Fig. 2d and Fig. S14†), demonstrating that the catalytic activity of the p-NP reduction reaction catalysed by Pt1–Cu8 CAs is controlled by photochemical effects.43,44 We next repeated this test under the same light intensity (1 sun) but with different wavelengths of light emitting diode (LED) irradiation. It was found that the catalytic reaction rate was wavelength-dependent (Fig. 2e and Fig. S15†), further corroborating the above results. In comparison with the general thermochemical reaction pathways where heating activates molecules, the distribution of molecular energy in the system follows the Boltzmann distribution.45,46 However, photochemical reactions can still be catalysed at very low temperatures, provided the wavelength of light is appropriate and can be absorbed by the substance. Overall, Pt–Cu bimetallic nanocrystalline particle aggregates significantly improved the rate of p-nitrophenol reduction, accompanied by a strong photochemical effect which further enhances its catalytic activity in high concentration of p-NP (Fig. 2b).
The cluster aggregates contribute to an efficient plasmonic hybridization effect, enhancing the absorption and photoresponsivity of Pt1–Cu8 SPs. To investigate the influence of the plasmonic hybridization effect on light absorption properties, we dissolved an equal mass of Pt1–Cu8 CAs and its control catalysts in deionized water to simulate the actual catalytic environment and tested their absorption spectra (Fig. 3a). First, through the comparison of control particle samples with two different particle sizes, it is evident that the light absorption ability of small-sized metal particles is much weaker. Due to the unique structure of the cluster aggregates, the extinction area of the catalyst system matrix is significantly expanded. Moreover, the simulation of the extinction efficiency of Pt–Cu nanoparticles (normalized by the geometrical cross-section) shows that when two or more nanoparticles are adjacent to each other, a new peak appears at a longer wavelength due to plasmonic hybridization. The peak red shifts with the number of nanoparticles, greatly enhancing the absorption of visible and infrared light (Fig. 3b). The comparison of the absorption spectra of Pt1–Cu8 CAs with those of Pt1–Cu8 LPs also supports this idea.
To further corroborate the efficacy of plasmonic hybridization, the photo-induced electric field around the densely arranged nanocrystalline cluster structure and discrete controls were simulated using the finite-difference time-domain (FDTD) method (Fig. 3c and d).47 The electric field is strongly localized on the surface of the aggregation structure. The maximum electric field intensity on the surface of Pt–Cu CAs is much higher than that of the discrete arrangement of small particles. Additionally, the photo-induced electric field around Pt–Cu LPs with the same feature sizes as Pt–Cu CAs was simulated. The maximum electric field intensity on the surface of Pt–Cu CAs is much higher than that of Pt–Cu LPs under single-wavelength illumination in the infrared region (950 nm) (Fig. S16a†), further corroborating that the plasmonic hybridization effect dramatically expands the catalyst light response range. The arrangement of particles also influences the effectiveness of the plasmonic hybridization effect in enhancing the LSPR effect. When nanoparticles are arranged solely in a single row, the degree of coupling between particles is limited, resulting in a weaker local electromagnetic field strength (Fig. S16b†). However, when the nanocrystalline particle aggregates are coupled with each other again, the plasmonic hybridization effect is further enhanced to improve the overall light response property (Fig. S16c and d†).
To gain a deeper understanding on the mechanism of plasmonic metal-based alloy catalysts in light-driven catalytic reactions, we first evaluated the electrochemical activity of Pt1–Cu8 CAs through cyclic voltammetry (CV) measurements in a mixed solution of 10−3 mol L−1p-NP and 0.1 mol L−1 sodium acetate, simulating the actual catalytic environment. Upon illumination, the Pt–Cu CAs exhibited a higher photocurrent density, proving that their charge separation efficiency was greatly enhanced (Fig. 4a), while their lower impedance indicated that their catalyst surface charge transfer efficiency was also significantly improved (Fig. 4b). The results of the Mott–Schottky test revealed a substantial increase in carrier concentration on the catalyst surface upon light irradiation, playing a crucial role in enhancing catalytic activity (Fig. 4c). Under chronoamperometry test, desirable stability was found for both illuminated and dark conditions, while the current densities with light was evidently higher, further evidencing the efficacy of photo-activation (Fig. 4d). Overall, it can be deduced from the photoelectrochemical measurements that the introduction of light significantly increased the carrier concentration and transfer rate on the catalyst surface, accelerating the desorption of the product p-AP and avoiding the poisoning of the active metal site. These aspects collectively led to the favourable catalytic turnover performance and enhanced catalyst stability.
To further confirm that the catalytic process was driven by the plasmonic photochemical effect of the Pt–Cu cluster aggregates rather than by additional mechanisms, such as photothermal activation, we next monitored the temperature change during the photoelectrochemical measurements of the reaction process. The results showed that the reaction temperature fluctuated by less than 3 °C around the room temperature of 18.6 °C, as measured by a thermocouple during a 12-hour test. To probe whether temperature change of this extent could have an effect on the reaction kinetics, we heated the reaction solution to different temperatures above RT and monitored the changes in photocurrent density. Notably, we observed only minimal temperature influence on the catalytic reaction. The effect of light on the heating of the catalytic system was essentially negligible over a range of temperature changes (Fig. 4e). The linear variation of electrochemical activity with light intensity, as assessed by cyclic voltammetry tests, further confirmed the photochemical effect of the catalyst (Fig. 4f).
In-depth analysis of the photocatalytic mechanism indicates that it is the hot electron transfer from Cu towards Pt that led to the favourable photochemical response.48 Metal Cu, with a higher Fermi energy level compared to metal Pt, tends to sustain a partially positive charge state after alloying, according to the results of local charge density calculations (Fig. S17†). During the catalysed p-NP reduction reaction, the strong electron-withdrawing nitro functional group is adsorbed to the metal Cu sites. Simultaneously, electrons are transferred to the metal Pt sites, which is more capable of hydrogenation, promoting the formation of hydrogen-active species. The hydrogen species then attack the nitro functional group and undergo an addition–substitution reaction to form an amino group (Fig. 5a). To gain insight into the photocatalytic mechanism of Pt–Cu CAs, in situ Fourier transform infrared diffuse reflectance (FTIR) tests were performed using CO as a gaseous probe molecule. Under light irradiation, the peaks of CO adsorbed at the Pt metal sites migrated to lower wavelengths, hot electrons were injected into the d orbitals of the Pt metal elements, and the center of the d band shifted upward, confirming that the direction of carrier transfer on the catalyst surface was from the Cu sites to the Pt sites.49,50 Moreover, the test under dark and light conditions revealed not only a change in the CO peak position but also a significant increase in peak height. These results evidence more favourable adsorption properties of the Pt–Cu cluster aggregates upon resonant illumination, which may be caused by the plasmonic charge transfer from Cu into the Pt species and a subsequent increase in d band center position (Fig. 5b). The results of in situ synchrotron radiation also verified this conclusion (Fig. 5c). Under light irradiation, the oscillation frequency of the electron cloud in the outer layer of the Cu plasmonic metal resonates with the frequency of the incident light, and the hot carriers are excited on the surface of the Cu metal particles, migrating to the Pt metal sites.51 This leads to a decrease in the valence state of the Pt metal element. Upon exposure to light, more hot electrons are generated by the excitation of the plasmonic metal Cu. These hot electrons are transferred to the Pt metal sites by electron–electron scattering, thereby enhancing hydrogen activation on electron-rich Pt sites. This process also promoted the adsorption of p-NP and desorption of the product p-AP, ultimately leading to an accelerated reaction rate.
Footnotes |
† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4gc00560k |
‡ These authors contributed equally to this work. |
This journal is © The Royal Society of Chemistry 2024 |