Supramolecular Perylene Diimides for Photocatalytic Hydrogen Production

Abstract

Energy depletion and environmental pollution have emerged as pressing global concerns, demanding the urgent promotion of green and clean energy sources. As such, the efficient utilization of solar energy for hydrogen production has gained significant research attention, with semiconductor photocatalysis emerging as an effective strategy. However, harnessing the full potential of semiconductor photocatalysis still poses great challenges. Notably, the limited utilization of visible light and the substantial recombination of photogenerated electron–hole pairs adversely affect photocatalytic performance, ultimately impeding the further development and practical application of semiconductor photocatalysis. Perylene diimide (PDI), an n-type semiconductor distinguished by its conjugated π-π bonds, exhibits remarkable photoelectric properties. Its energy band gap falls within the absorption range of visible light, ensuring remarkable light absorption efficiency. Furthermore, the photogenerated charge can be efficiently conducted along the π-π stacking in its structural unit, significantly reducing electron–hole recombination. Consequently, PDI holds immense potential for achieving visible-light-driven photocatalytic hydrogen production. Yet, despite these attributes, the photocatalytic efficiency of pure PDI is still far from practical use, necessitating innovative modifications to elevate its catalytic performance. In this review, we begin with an in-depth exploration of the principles underlying photocatalytic hydrogen production and discuss various strategies aimed at enhancing photocatalytic performance. We also engage in a comprehensive discussion and summation of the challenges encountered and the future prospects of PDI-based materials. Our endeavor is to pave the way for groundbreaking advancements in the field of photocatalysis, ultimately contributing to a cleaner and more sustainable future.

1. Introduction

Global population growth and industrial development have significantly increased energy consumption and waste generation [1]. Current energy systems remain heavily dependent on fossil fuels, with petroleum constituting 32.9% of the global energy supply in 2015, followed by coal (29.2%) and natural gas (23.8%) [2]. This reliance on carbon-intensive resources necessitates the development of sustainable zero-emission alternatives [3]. Solar energy, the most abundant renewable source, delivering over 120,000 TW annually [4,5], presents a viable solution through its conversion into storable hydrogen energy.

While multiple hydrogen production methods exist—including fossil fuel reforming, photocatalytic systems, photovoltaic (PV) electrocatalysis, photoelectrocatalysis, and electrocatalysis [6,7,8,9]—current global hydrogen production (70 million tons/year) remains predominantly fossil-based (96% from natural gas, coal, and oil) [10]. Photocatalytic systems utilizing semiconductor catalysts, especially photocatalytic water decomposition, have garnered significant attention since their discovery [11,12]. Unlike conventional fossil-derived hydrogen processes that emit pollutants and greenhouse gases [13,14], this approach theoretically produces only hydrogen and oxygen through water splitting. Achieving large-scale, low-emission, and high-efficiency hydrogen production remains hindered by substantial technological bottlenecks.

To achieve low-emission and high-efficiency hydrogen production, success hinges on the cleanliness of production technologies, pathways, and raw material sources. Therefore, photocatalytic overall water splitting and organic reforming reactions constitute the principal modalities of the photocatalysis system. Photocatalytic overall water splitting for hydrogen production represents a highly prevalent and economically viable approach that integrates light harvesting with the redox reaction [15,16]. This process employs semiconductor photocatalysts as active agents to convert low-density solar energy into hydrogen energy [17]. The field of semiconductor photocatalytic materials is predominantly populated by three key categories: inorganic semiconductors (TiO₂ [18,19], ZnO [20,21], CdS [22,23], Bi₂WO₆ [24,25], PbS [26,27], etc.), organic semiconductors, and their composite forms. Inorganic semiconductors have held an early advantage in photocatalysis, but they are predominantly constrained by a narrow light absorption spectrum, low molar extinction coefficients, and challenges in tailoring their energy levels and structural configurations [28]. In contrast, organic semiconductors have emerged as a focal point of recent research, due to their chemically adjustable electrical and optical properties, cost-effectiveness, structural versatility, and diversity [29,30]. Notably, organic supramolecular semiconductors distinguish themselves through their extensive light absorption range, versatile assembly methodologies, and enhanced stability, offering substantial potential for diverse photocatalytic applications [31]. Notable examples of these supramolecular materials include polymeric graphitic carbon nitride (g-C₃N₄) [32,33], fullerenes (C₆₀) [34,35], porphyrins (Pors) [36,37], phthalocyanines (PCs) [38,39], pyridines [40,41], ruthenium polypyridine complexes [42,43], polythiophenes (PTHs) [44,45], polypyrroles (PPYs) [46,47], and perylene diimides (PDIs) [48,49], etc., alongside their numerous derivatives.

PDIs and their derivatives stand out as promising n-type photocatalysts due to their distinct advantages over other organic semiconductors. A suitable ~2 eV band gap enables efficient visible-light absorption (400–750 nm), narrower than that of g-C₃N₄ (2.7 eV) and broader than TiO₂’s UV-limited response [50]. Their high molar absorption coefficients enhance light harvesting [51]. Their exceptional photothermal stability (up to 500 °C in air) surpasses that of easily degraded porphyrins [52]. Chlorine-substituted PDIs exhibit superior electron mobility (0.081 cm²·V⁻¹·s⁻¹), outperforming fullerene-based systems [53]. Functional group modification, e.g., phosphoric acid substitution, boosts hydrogen evolution to 46 mmol·g⁻¹·h⁻¹ under visible light, outclassing many conjugated polymers and MOFs [54]. These properties make PDIs critical for advancing efficient, stable solar-to-hydrogen conversion. Supramolecular PDI photocatalysts are formed through the self-assembly of monomeric PDI molecules, utilizing non-covalent interactions, such as hydrogen bonding, π-π stacking, and van der Waals forces, for structural integrity [55]. Initially, researchers employed PDI supermolecules as photosensitizers or charge carrier transport media in conjunction with other photocatalysts. Chen and his team [56], for instance, devised a PTCDI/Pt/g-C₃N₄ photocatalyst system, wherein PTCDI (perylene tetracarboxylic diimide) acted as the antenna chromophore, g-C₃N₄ as the primary catalyst, and Pt nanoparticle as the cocatalyst. Building on this, Chen et al. [57] subsequently developed a PTCDI/Pt/TiO₂ composite catalyst system, with PTCDI functioning as the photosensitizer, P25 as the catalyst and Pt also as the cocatalyst. This system achieved stable hydrogen production from water splitting across a wide range of visible-light wavelengths. Later research revealed that PDI-based supermolecules could independently complete the entire photocatalytic reaction process. Liu et al. [58] for example, reported a visible-light photocatalytic system using self-assembled perylene-3,4,9,10-tetracarboxylic diimide (PDINH), where strong π-π stacking interactions facilitated the efficient migration and separation of photogenerated charges, leading to effective photocatalytic activity. Another example is the self-assembled supramolecular system developed by Kong et al. [50], based on phosphate-substituted PDI (P-PMPDI). The phosphate groups, with their potent electron-withdrawing effect, promoted the separation of photoinduced charge carriers, enabling highly efficient photocatalytic hydrogen production. These advancements highlight the significant strides made in harnessing the potential of PDI-based supramolecular semiconductors for photocatalytic applications. Table 1 presents a comparison of the photocatalytic hydrogen generation performance of PDI-based nanoscale semiconductor materials and their composites in recent years.

Recent advances in perylene diimide (PDI)-based materials have been extensively reviewed, with the existing literature focusing on molecular design (e.g., bay/imide substitution, symmetry modulation) [59,60], self-assembly mechanisms [29,59,61], electronic structure regulation [61,62], and broad photocatalytic applications, including pollutant degradation [29,63], CO₂ reduction [64], and organic synthesis [61]. Notably, several reviews have discussed PDI-based photocatalysts for hydrogen production as a subtopic within general photocatalysis [63,64], while Li et al. [65] recently outlined molecular design strategies for PDI derivatives in photocatalytic water splitting. However, no comprehensive review has specifically addressed the multifaceted modification strategies of PDI-based materials exclusively for photocatalytic hydrogen evolution. Our work fills this critical gap by systematically summarizing state-of-the-art approaches to enhance PDI performance in hydrogen generation, focusing on three key dimensions: (1) molecular-level modifications (substitution and doping), (2) nanostructuring (aggregate morphology, self-assembly, and dimensionality control), and (3) heterojunction engineering, thereby establishing a dedicated framework for hydrogen-centric optimization.

Table 1. Summary of photocatalytic properties of PDI-based photocatalysts.

Photocatalyst	Preparation Method	Structure and Form	Photocatalytic Hydrogen Production Efficiency	Light Source	Wavelength	Electron Source for H2 Evolution	Reference
PTA and Cl-PTA	Hydrolysis-acidification reassembly strategy	Nanosheets	PTA: 15.2 mmol·g−1·h−1 Cl-PTA: 27.1 mmol·g−1·h−1	CEL-HXF300 300W xenon lamp 1	Full-spectrum irradiation (UV + visible light)	Ascorbic acid	[48]
PyBpDBSO-5	Suzuki coupling reaction	Flaky amorphous structures	48.5 mmol·g−1·h−1	Visible light	>420 nm	TEOA	[49]
PTCDIs/Pt/g-C3N4	Self-assembly method	Amorphous structure	0.015 mmol·g−1·h−1	300 W xenon lamp (CEL-S500, Beijing AULTT) with 420 nm cutoff optical filter attached 1	>420 nm	TEOA	[56]
PDI-phthalic	Self-assembly method	Nanosheets	1.1 mmol·g−1·h−1	300/500 W xenon lamp with a 420 nm cut-off filter 2	>420 nm	Ascorbic acid	[66]
PDI/Zn0.8Cd0.2S	Self-assembly method	Nanorods/nanospheres	71.98 μmol·g−1·h−1	300W xenon lamp (1000 mW·cm−2)	>420 nm	Water	[67]
PDI/Zn0.7Cd0.3S	Co-precipitation-hydrothermal method	Nanoparticles	5.166 mmol·g−1·h−1	Visible-light irradiation	Monochromatic light at 420 nm	Na2S and Na2SO3 aqueous solution	[68]
N-APDI	Self-assembly method	Nanosheets	61.49 mmol·g−1·h−1	300 W xenon lamp	>400 nm	Ascorbic acid	[69]
P-PMPDI	Self-assembly method	Multilayer nanobelts	11.7 mmol·g−1·h−1	Visible light	400 nm–780 nm	Ascorbic acid	[50]
P-PMPDI-Zr	Self-assembly method	Multilayer nanobelts	50.46 mmol·g−1·h−1	CEL-HXF300 300 W xenon lamp 1	400 nm–780 nm	Ascorbic acid	[54]
CBZ-PDCA-PT2	Heating + evaporating + filtration, Suzuki coupling	Nanoparticles	30 mmol·g−1·h−1	Sun simulator SCIENCETECH SF-300-A equipped with airmass filter AM1.5 G 3	Visible light	Ascorbic acid	[70]
TPPS/PDI	Self-assembly method	Nanowires	30.36 mmol·g−1·h−1	Full-spectrum light	Full-spectrum irradiation	Ascorbic acid	[71]
g-C3N4/rGO/PDIP	Wet-chemistry reduction + solvent evaporation + heat treatment	Nanosheets/nanorods (core/shell)	15.8 μmol·g−1·h−1	Visible light	≥420 nm	Water	[72]
PDI/g-C3N4	Self-assembly method	Nanorods/nanosheets	1649.93 μmol·g−1·h−1	Full-spectrum light	Full-spectrum irradiation	TEOA	[73]
COF (TATF-COF)/PUP	Imidazole solvent method + via in situ coupling	Nanosheets/nanorods	94.5 mmol·g−1·h−1	350 W xenon lamp	>420 nm	Ascorbic acid	[74]
PDI-TiO2	Solvent compounding method	Nanosheets/nanoparticle	238 mmol·g−1·h−1	300 W xenon lamp	Visible light	Water	[75]
ZIS/PDIIM	Hydrothermal method	Nanosheets/nanorods	13.04 mmol·g−1·h−1	PLS-SXE300 300 W xenon lamp (0.6 W/cm2)1	320 nm–780 nm	Na2S and Na2SO3 aqueous solution	[76]
PDINH/TiO2	Hydrothermal method + mixing	Nanosheets/nanorods	1.2 mmol·g−1·h−1	PLS-SXE 300 300 W Xe arc lamp 4	UV–VIS light irradiation	Methanol	[77]
GQDs/PDI-14%	Self-assembly method + mixing	Zero-dimensional quantum dots/nanofibers	1.6 mmol·g−1·h−1	Visible light	>420 nm	Ascorbic acid	[78]
PTA	Facile hydrolysis reassembly of PTCDA	Nanosheets	118.9 mmol·g−1·h−1	300 W Xe lamp (~530 mW cm−2)	≥300 nm	Ascorbic acid	[79]
CN-P-0.2%	Thermal condensing of cyanamide	Porous structure	17.7 mmol·g−1·h−1	LED	450 nm	TEOA	[80]

¹ CEL-HXF300/CEL-HXFS500/PLS-SXE300 300W xenon lamps: Beijing China Education Au-light Co., Ltd., located in Beijing, China. ² 300/500 W xenon lamp: Beijing PerfectLight, located in Beijing, China. ³ SCIENCETECH SF-300-A: SCIENCETECH, located in London, Canada. ⁴ PLS-SXE 300 300 W Xe arc lamp: Beijing Trusttech Co. Ltd., located in Beijing, China.

Table 1. Summary of photocatalytic properties of PDI-based photocatalysts.

Photocatalyst	Preparation Method	Structure and Form	Photocatalytic Hydrogen Production Efficiency	Light Source	Wavelength	Electron Source for H2 Evolution	Reference
PTA and Cl-PTA	Hydrolysis-acidification reassembly strategy	Nanosheets	PTA: 15.2 mmol·g−1·h−1 Cl-PTA: 27.1 mmol·g−1·h−1	CEL-HXF300 300W xenon lamp 1	Full-spectrum irradiation (UV + visible light)	Ascorbic acid	[48]
PyBpDBSO-5	Suzuki coupling reaction	Flaky amorphous structures	48.5 mmol·g−1·h−1	Visible light	>420 nm	TEOA	[49]
PTCDIs/Pt/g-C3N4	Self-assembly method	Amorphous structure	0.015 mmol·g−1·h−1	300 W xenon lamp (CEL-S500, Beijing AULTT) with 420 nm cutoff optical filter attached 1	>420 nm	TEOA	[56]
PDI-phthalic	Self-assembly method	Nanosheets	1.1 mmol·g−1·h−1	300/500 W xenon lamp with a 420 nm cut-off filter 2	>420 nm	Ascorbic acid	[66]
PDI/Zn0.8Cd0.2S	Self-assembly method	Nanorods/nanospheres	71.98 μmol·g−1·h−1	300W xenon lamp (1000 mW·cm−2)	>420 nm	Water	[67]
PDI/Zn0.7Cd0.3S	Co-precipitation-hydrothermal method	Nanoparticles	5.166 mmol·g−1·h−1	Visible-light irradiation	Monochromatic light at 420 nm	Na2S and Na2SO3 aqueous solution	[68]
N-APDI	Self-assembly method	Nanosheets	61.49 mmol·g−1·h−1	300 W xenon lamp	>400 nm	Ascorbic acid	[69]
P-PMPDI	Self-assembly method	Multilayer nanobelts	11.7 mmol·g−1·h−1	Visible light	400 nm–780 nm	Ascorbic acid	[50]
P-PMPDI-Zr	Self-assembly method	Multilayer nanobelts	50.46 mmol·g−1·h−1	CEL-HXF300 300 W xenon lamp 1	400 nm–780 nm	Ascorbic acid	[54]
CBZ-PDCA-PT2	Heating + evaporating + filtration, Suzuki coupling	Nanoparticles	30 mmol·g−1·h−1	Sun simulator SCIENCETECH SF-300-A equipped with airmass filter AM1.5 G 3	Visible light	Ascorbic acid	[70]
TPPS/PDI	Self-assembly method	Nanowires	30.36 mmol·g−1·h−1	Full-spectrum light	Full-spectrum irradiation	Ascorbic acid	[71]
g-C3N4/rGO/PDIP	Wet-chemistry reduction + solvent evaporation + heat treatment	Nanosheets/nanorods (core/shell)	15.8 μmol·g−1·h−1	Visible light	≥420 nm	Water	[72]
PDI/g-C3N4	Self-assembly method	Nanorods/nanosheets	1649.93 μmol·g−1·h−1	Full-spectrum light	Full-spectrum irradiation	TEOA	[73]
COF (TATF-COF)/PUP	Imidazole solvent method + via in situ coupling	Nanosheets/nanorods	94.5 mmol·g−1·h−1	350 W xenon lamp	>420 nm	Ascorbic acid	[74]
PDI-TiO2	Solvent compounding method	Nanosheets/nanoparticle	238 mmol·g−1·h−1	300 W xenon lamp	Visible light	Water	[75]
ZIS/PDIIM	Hydrothermal method	Nanosheets/nanorods	13.04 mmol·g−1·h−1	PLS-SXE300 300 W xenon lamp (0.6 W/cm2)1	320 nm–780 nm	Na2S and Na2SO3 aqueous solution	[76]
PDINH/TiO2	Hydrothermal method + mixing	Nanosheets/nanorods	1.2 mmol·g−1·h−1	PLS-SXE 300 300 W Xe arc lamp 4	UV–VIS light irradiation	Methanol	[77]
GQDs/PDI-14%	Self-assembly method + mixing	Zero-dimensional quantum dots/nanofibers	1.6 mmol·g−1·h−1	Visible light	>420 nm	Ascorbic acid	[78]
PTA	Facile hydrolysis reassembly of PTCDA	Nanosheets	118.9 mmol·g−1·h−1	300 W Xe lamp (~530 mW cm−2)	≥300 nm	Ascorbic acid	[79]
CN-P-0.2%	Thermal condensing of cyanamide	Porous structure	17.7 mmol·g−1·h−1	LED	450 nm	TEOA	[80]

¹ CEL-HXF300/CEL-HXFS500/PLS-SXE300 300W xenon lamps: Beijing China Education Au-light Co., Ltd., located in Beijing, China. ² 300/500 W xenon lamp: Beijing PerfectLight, located in Beijing, China. ³ SCIENCETECH SF-300-A: SCIENCETECH, located in London, Canada. ⁴ PLS-SXE 300 300 W Xe arc lamp: Beijing Trusttech Co. Ltd., located in Beijing, China.

In our review, we initially introduce the fundamental principles underlying photocatalytic hydrogen production, emphasizing the key role of the band structure in determining the effectiveness of photocatalysts. We then proceed to explore the intricate performance conditions associated with organic photocatalysts within the context of photocatalytic hydrogen production. This involves a comprehensive examination of various factors, including the catalyst’s stability under reaction conditions, its light absorption capabilities, and the efficiency of charge separation and transfer processes. These conditions are crucial for maximizing both the yield and rate of hydrogen production. Subsequently, we shift our focus to the unique properties and intricate structure of PDI supramolecular materials. By gaining a deep understanding of the specific properties and structural features of these materials, we can tailor their design to enhance their performance in hydrogen production. The emphasis of our paper is on modification strategies aimed at optimizing PDI-based materials for photocatalytic hydrogen production. Finally, we summarize the current challenges faced in the research field and propose strategic directions for future development. These challenges include the need for more efficient and stable photocatalysts, the optimization of reaction conditions, and the scalability of the photocatalytic process for practical applications. By addressing these challenges and leveraging the insights gained from recent research, we aim to provide a design guideline for the preparation of high-performance PDI-based photocatalysts. Our ultimate goal is to significantly advance the field of photocatalytic hydrogen production.

2. Fundamentals

2.1. The Process of Photocatalytic Water Splitting

Photocatalytic water splitting into hydrogen can be divided into at least four steps: photon absorption, photoexcited charge separation, charge diffusion/transport, and catalytic reaction at the catalyst’s active sites.

In inorganic semiconductors, upon illumination, electrons within the valence band (VB) of the photocatalyst undergo excitation, transitioning to the conduction band (CB) while leaving holes in the VB, thereby generating electron–hole pairs (hereinafter referred to as Process I). Similarly, the energy band configuration of organic semiconductors exhibits an analogy to inorganic semiconductors, as illustrated on the right-hand side of Figure 1. Specifically, the “Highest Occupied Molecular Orbital” (HOMO) aligns with the VB, and the “Lowest Unoccupied Molecular Orbital” (LUMO) corresponds to the CB, with the energy separating these two levels designated as the band gap, E_g.

In the context of exciton excitation within organic semiconductors, electrons in their ground state are confined to the HOMO, with the LUMO remaining unoccupied. Upon absorption of energy by an electron in the HOMO, exceeding the threshold of Eg, there is a likelihood for this electron to cross the band gap and relocate to the LUMO, transitioning the molecule to an excited state. A defining characteristic of organic semiconductors is their low dielectric constant, which favors the storage of energy in the form of bound states known as excitons. These excitons can manifest as Frenkel excitons or as Coulomb-bound electron–hole pairs. To quantify these phenomena, Equation (1) provides a formulation for calculating the exciton binding energy [81], while Equation (2) offers an expression for determining the exciton mass [82]. These equations are crucial for understanding the fundamental physics governing exciton dynamics in organic semiconductors.

$$E_{X_{BE}}=E_H\ast m^{*}/{\epsilon }_r^2$$

(1)

$${\displaystyle \frac{1}{m^{*}}}={\displaystyle \frac{1}{m_e^{*}}}+{\displaystyle \frac{1}{m_h^{*}}}$$

(2)

The exciton binding energy, denoted as $E_{X_{BE}}$, is a crucial parameter that characterizes the stability of excitons in a material [83]. Meanwhile, $E_H$, with a value of −13.6 eV, represents the energy level of a hydrogen atom in its 1s orbital, serving as a reference point for energy comparisons. The reduced mass of the exciton, $m^{*}$, plays a significant role in determining the exciton’s dynamic properties. The relative permittivity, ${\epsilon }_r$, influences the electric field interactions within the material, affecting exciton stability. Additionally, the effective masses of the electron, $m_e^{*}$, and the hole, $m_h^{*}$, are essential in describing the motion of charge carriers within the exciton. A larger exciton binding energy corresponds to a smaller Bohr radius, indicating a stronger binding force between the electron and the hole within the exciton, which makes spontaneous dissociation less likely. Due to differences in relative permittivity, organic semiconductors exhibit exciton binding energies ranging from 0.2 to 0.5 eV, resulting in a compact exciton Bohr radius. In contrast, inorganic semiconductors possess exciton binding energies of less than 10 meV, leading to a more extended exciton Bohr radius. This fundamental difference has profound implications for exciton dynamics: excitons in organic semiconductors are relatively stable and difficult to dissociate spontaneously, whereas excitons in inorganic semiconductors are more susceptible to dissociation [84,85]. As a result, excitons in organic semiconductors exhibit shorter diffusion lengths, increasing the probability of exciton–exciton recombination and reducing the quantum efficiency of the material.

Subsequently, these electrons and holes, functioning as carriers, traverse to the catalyst’s surface (Process II), where they engage in redox reactions with water molecules (Process III). This crucial phase encompasses the hydrogen evolution reaction (HER) and the oxygen evolution reaction (OER) [86].

Photocatalytic water splitting is similar to electrolytic water splitting. When illuminated, water undergoes a reaction characterized by a Gibbs free energy of 237 kJ·mol⁻¹, clearly indicating its non-spontaneous nature. In the realm of electrochemistry, water possesses a standard reduction potential of 0 V (referenced against H₂O/H₂) and a standard oxidation potential of 1.23 V (referenced against H₂O/O₂). Consequently, the electrolysis of two water molecules to yield two hydrogen molecules and one oxygen molecule necessitates the input of two 1.23 eV of energy [87]. In the context of photocatalytic water splitting, the inherent energy losses encountered during the catalytic process necessitate a photocatalyst with a band gap exceeding 1.23 eV. This prerequisite ensures that the minimum of the conduction band (CBM) exceeds the reduction potential of H⁺/H₂, while the maximum of the valence band (VBM) falls below the oxidation potential of O₂/H₂O, as illustrated on the right side of Figure 1. Furthermore, the absorption wavelength of the incident light must be less than 1000 nm, marking the threshold energy requirement for photocatalytic water splitting.

The carrier transfer process inherent in Process II of photocatalytic water splitting introduces inevitable energy losses. Similarly, Process III necessitates kinetic overpotentials to surmount the energetic barriers associated with the HER and OER. Consequently, photocatalysts typically exhibit band gaps greater than 1.8 Ev [88]. To facilitate catalysis under visible light, the band gap must be further constrained to less than 3.0 eV, corresponding to an absorption wavelength exceeding 400 nm [89]. The band structure and position of the photocatalyst are thus determinant factors in the feasibility of photolytic water splitting and significantly influence the absorption range of the solar spectrum.

Beyond the fundamental considerations of band structure and position, the efficiency of these photocatalysts depends on several additional factors. These include the light absorption capabilities of photogenerated electrons and holes, the efficacy of charge separation and migration, and the HER and OER processes occurring on the photocatalyst’s surface. Regarding band structure and position, it is crucial to recognize that a larger band gap does not inherently equate to superior performance. A widened band gap prevents the absorption of photons with longer wavelengths, resulting in diminished light utilization efficiency and a consequent reduction in solar-to-hydrogen efficiency (STH). Therefore, a band gap ranging between 1.8 and 2.2 eV is deemed optimal. Inorganic photocatalysts often face this issue, with most only able to absorb ultraviolet light for catalytic hydrogen production [90]. The greatest advantage of organic photocatalysts is that their band structure can be adjusted through the design of molecular structure and arrangement to achieve broad/full-spectrum absorption. However, the inherently low dielectric constant (${\epsilon }_r$ = 3~5) of organic materials poses challenges, leading to sluggish carrier migration rates and limited migration distances. These limitations necessitate further refinement and modification to harness the full potential of organic photocatalysts.

2.2. The Structure/Property of PDI

PDI is a typical n-type semiconductor material, known for its excellent photochemical and thermal stability (light absorption range: 400–750 nm, E_g ≈ 2 eV, temperature resistance in air: 500 °C), tunable photoelectric properties (through changes in molecular stacking), and unique redox properties, making it widely used in visible-light catalysis. Structurally, the PDI molecule is composed of a central, rigid, and planar perylene ring flanked by two terminal imide groups. As depicted in Figure 2, the termini are designated as “imide positions”, whereas positions 1, 6, 7, and 12 are collectively termed “bay positions”. Consequently, two primary derivatization strategies emerge: one entails the introduction of diverse substituents onto the N atoms situated at the imide positions, and the other involves the modification of the bay positions (1, 6, 7, and 12) of the PDI with electron-withdrawing moieties, such as halogen atoms, cyanides, and alkynyl groups, or electron-donating entities, like phenolate, phenyl, and secondary amino groups [91].

The HOMO and LUMO orbitals of the PDI molecule are predominantly contributed by C and O atoms, with the N atoms in the amide bonds playing a negligible role in orbital formation [92,93,94]. Hence, modifications and substitutions at the bay positions can effectively modulate the intrinsic energy levels of the PDI molecule, whereas alterations at the imide positions remain inert to its intrinsic energy levels and redox potentials, thereby preserving its molecular spectral absorption and emission profiles. Vajiravelu et al. substituted PDI at the 1,6-bay positions with electron-donating groups such as 3,4,5-trimethoxyphenyl, thiophene, 6-methoxynaphthyl, 5-hexyl-2-thiophene, and thiophenyl. The presence of these electron-donating aromatic units reduced the band gap of PDI from 2.32 eV to 1.54 eV [95]. Zhang et al. [51] introduced four chlorine atoms at the 1,6,7,12-bay positions of PDI, resulting in an excellent mobility of 0.081 cm²·V⁻¹·s⁻¹. Jones et al. [53] achieved even higher electron mobility (0.64 cm²·V⁻¹·s⁻¹) by introducing cyano groups at the 1,7-bay positions. Further, Yi et al. [96] formed PDI dimers at the 1,7-bay positions. Due to their enhanced intermolecular and intramolecular interactions, the dimers still exhibited similar absorption in the 400–500 nm range, with a broadened absorption band.

The PDI molecule undergoes structural modifications to the perylene core, which, in turn, forms different intermolecular interactions that assemble into micro- and nanostructures, significantly affecting its optical and electrical properties. The perylene ring in the PDI molecule is a rigid polycyclic π-electron conjugated system, allowing PDI molecules to form supramolecular stacking structures through π-π interactions. The π-π stacking modes of PDI include two configurations: J-type and H-type aggregates [97]. H-aggregates primarily adopt a face-to-face parallel stacking arrangement, whereas J-aggregates typically exhibit an offset parallel stacking configuration [98]. This disparity in stacking modes dictates that PDI molecules with fewer spatial constraints (i.e., more linear structures) prefer to form H-aggregates, whereas those with greater spatial demands (i.e., more branched structures) tend to form J-aggregates or remain non-aggregated. Evidently, the coplanar H-aggregates of the PDI core skeleton exhibit a superior degree of π-π stacking and π-conjugation, manifesting semiconductor photoelectric properties [99]. Conversely, the partially offset J-aggregates of the PDI core skeleton display reduced π-π stacking, retaining photoelectric properties akin to those of monomeric PDI [100]. The π-delocalized channels facilitated by the π-π stacking structure expedite electron transfer, favoring the electron transfer mechanism (ET) [101], whereas J-aggregate PDI exhibits a propensity towards the energy transfer mechanism (EnT) [52]. The H-aggregate structure boasts enhanced photogenerated carrier mobility and separation efficiency, coupled with a heightened oxidation capacity of the active components, thereby promising its potential in photocatalytic reactions.

3. Modification Strategies

Despite the ability of PDI photocatalysts to independently execute the entire photocatalytic process, their performance is constrained by three primary factors: (1) high exciton binding energy and the propensity for exciton (electron–hole pair) recombination; (2) limited carrier mobility; and (3) a short exciton diffusion length. In recent decades, various strategies have been devised to address these challenges by modifying the nanostructures of PDIs. These strategies mainly include the following: (1) molecular-level design; (2) nanostructuring; and (3) heterostructuring.

3.1. Molecular-Level Modifications

PDI molecules feature two substitutable sites: the “imide positions”, positioned at both termini, and the “bay positions”, located at positions 1, 6, 7, and 12. Modifications at the imide positions primarily entail the incorporation of diverse substituents onto the N atoms. It is noteworthy that these N atoms, as integral parts of the amide bonds, contribute minimally to the formation of the HOMO and LUMO orbitals of PDI molecules [92]. Consequently, alterations at this site do not shift the intrinsic energy levels, redox potentials, spectral absorption, or emission characteristics of PDI molecules. However, the introduction of varying substituents at the imide positions of PDI molecules alters the electron distribution and molecular dipole moment, stemming from differences in electronegativity [102]. This, in turn, exerts an influence on the intensity of the internal electric field (IEF). A potent IEF intensity enhances the separation of photogenerated charges and boosts catalytic activity.

Pu et al. replaced the imide positions of PDI molecules with negatively charged functional groups, including methane, amino, and carboxyl. DFT calculations and characterization results confirmed [103] that these negatively charged groups effectively modulated the molecular dipoles, thereby altering the intensity of the internal electric field (IEF). Among the substituted PDIs, the carboxyl-substituted PDI exhibited the maximum absorption wavelength of 740 nm. Its photocurrent density reached 2.12 mA∙cm⁻², surpassing those of the methane- and amino-substituted PDIs, which were 0.86 mA∙cm⁻² and 1.47 mA∙cm⁻², respectively. Furthermore, the charge transfer internal resistance of the carboxyl-substituted PDI was lower, at 40.6 kΩ, compared to 90.5 kΩ and 116.5 kΩ for the methane- and amino-substituted PDIs, respectively. These findings correlated well with the IEF magnitudes derived from DFT calculations. Clearly, the combined influence of electron-accepting and electron-donating groups on the molecular dipole moment determines the IEF intensity. An increased molecular dipole moment leads to a higher IEF intensity, which favors the separation of photogenerated charges, and subsequently enhances photocatalytic activity [103].

In another study, Kong et al. [50] incorporated carboxyl and phosphoryl groups into the imide positions of PDI molecules. The carboxyl-substituted PDI (P-CMPDI) (Figure 3a) exhibited a higher molecular dipole moment compared to the phosphoryl-substituted PDI (P-PMPDI) (Figure 3b). This was particularly advantageous under long-wavelength monochromatic light radiation, where photocatalytic activity is typically less effective. The absorption edge spectrum of P-PMPDI extended up to 780 nm. The increased molecular dipole strengthened the internal electric field between the central portion and the terminal substituents, facilitating the separation of photogenerated electron–hole carriers. Consequently, P-PMPDI displayed a significantly reduced photoluminescence intensity compared to P-CMPDI, which exhibited a strong emission peak at 689 nm. When Pt nanoparticles were used as cocatalysts, the ultimate hydrogen production rate of P-PMPDI was 11.7 mmol∙g⁻¹∙h⁻¹, approximately 11 times higher than that of P-CMPDI (0.83 mmol∙g⁻¹∙h⁻¹) (Figure 3c). After continuous irradiation under visible light for 12 h, no significant deactivation of P-PMPDI was observed (Figure 3d), and the apparent quantum yield (AQY) at 550 nm was 2.96%. Due to the high electronegativity of P-PMPDI, it possessed a dipole moment of 5.8 Debye, enabling its spectral response to cover nearly the entire visible spectrum. In addition, the team further added the coordination metal Zr(IV) to the self-assembled solution of P-PMPDI to generate P-PMPDI-Zr (Figure 3e) [54]. The existence of phosphate group–metal coordination bonds with Zr resulted in a dipole moment of 5.94 Debye, leading to more compact stacking, with the characteristic stacking distance between adjacent perylene cores reduced from 0.331 nm to 0.316 nm (Figure 3f). Compared to P-PMPDI (1.66 eV), P-PMPDI-Zr has a narrower band gap of 1.65 eV, with maximum absorption in the near-infrared region. The photovoltage of P-PMPDI-Zr is also significantly higher than that of P-PMPDI (Figure 3g). Its photocurrent density is 1.62 mA·cm⁻², much higher than that of P-PMPDI (0.2 mA·cm⁻²). Ultimately, the hydrogen evolution rate of P-PMPDI-Zr reaches 46 mmol·g⁻¹·h⁻¹, with an apparent quantum yield of up to 11.7% at 630 nm in the near-infrared region (Figure 3h) [54].

The introduction of various substituents at the bay positions of PDI molecules significantly alters their intrinsic energy levels, primarily due to the substantial contribution of carbon (C) and oxygen (O) elements to the HOMO and LUMO orbitals [92]. The electron-deficient nature of PDI molecules makes them prone to reduction, but resistant to oxidation, thereby enhancing the impact of bay substituents on their redox potentials. For example, the incorporation of ditetravinylene substituents at the 1,6-bay positions of PDI molecules induced a substantial red-shift in fluorescence, exceeding 120 nm [104]. Vajiravelu et al. substituted the 1,6-bay positions of PDI with electron-donating groups, including 3,4,5-trimethoxyphenyl, thiophene, 6-methoxynaphthyl, 5-hexyldithiophene, and thiophenyl. This led to notable red-shifts in both the absorption and emission spectra compared to unsubstituted PDI. These compounds exhibited broad absorption bands, indicating extended π-π conjugation facilitated by the aryl groups in the bay region of PDI. Cyclic voltammetry studies further revealed a significant reduction in the band gap, from 2.32 eV to 1.54 eV, upon the introduction of electron-donating aryl units [95].

Another study reported the introduction of four chlorine atoms at the 1,6,7, and 12 positions of PDI, which enhanced electron mobility to 0.081 cm²·V⁻¹·s⁻¹. This improvement was attributed to hydrogen bonding and two-dimensional π-stacking interactions [51]. In contrast, the introduction of cyano groups at the 1,7 positions of PDI lowered the LUMO energy level and increased intermolecular π-overlap, resulting in a higher electron mobility of 0.64 cm²·V⁻¹·s⁻¹ [53]. Furthermore, a PDI dimer with substitutions at the 1,7 positions was synthesized. Due to the simultaneous enhancement of intermolecular and intramolecular interactions, the dimer exhibited similar absorption in the 400–500 nm range, but with a broader absorption band. Notably, the electron mobility of the linear alkyl chain PDI dimer was nearly 20 times higher than that of the branched alkyl chain PDI monomer [96].

Doping PDI molecules with metal or nonmetal elements provides additional pathways to enhance their photocatalytic hydrogen production capabilities. Metal substitution or doping can facilitate charge transfer through mechanisms such as Metal–Metal Charge Transfer (MMCT) or Metal-to-Ligand Charge Transfer (MLCT), while the doped metal ions serve as electron mediators to promote charge transfer. Doping with nonmetal elements also contributes to improving PDI’s photocatalytic hydrogen production. Xu et al. developed three heteroatom-doped bay-annulated PDI supramolecular photocatalysts: N-APDI, S-APDI, and Se-APDI (where APDI represents a phosphate-substituted PDI supramolecule). According to Density Functional Theory (DFT) calculations, the introduction of heteroatoms significantly increased the dipole moments of the R-APDI supramolecules. Specifically, N-APDI exhibited a dipole moment of 2.57 Debye, which is 10.7 times greater than that of APDI (0.24 Debye). This enhanced dipole moment generated a stronger built-in electric field, which favored the separation and transfer of photogenerated carriers and consequently boosted the photocatalytic activity of the catalyst [69].

Molecular engineering of perylene diimide (PDI)-based photocatalysts centers on imide/bay position substitutions and metal/nonmetal doping, achieving precise band gap tailoring and enhanced charge carrier transport for optimized photocatalytic systems. Imide substitutions (e.g., carboxyl and phosphoryl groups) enhance the internal electric field (IEF) by modulating molecular dipole moments without altering intrinsic energy levels [103]. Bay-region modifications (e.g., chlorine, cyano, and aryl groups) directly modulate HOMO/LUMO orbitals, inducing substantial spectral red-shifts [50,54], narrowing band gaps [95], and improving electron mobility [53]. Elemental doping (e.g., Zr coordination and S/Se heteroatoms) further strengthens dipole moments [69] and enhances near-infrared response, enabling synergistic optimization of charge separation and photocatalytic activity [54]. The foremost advantage of molecular-level engineering lies in its theoretical guidance and precision design: DFT calculations provide fundamental insights for experimental optimization, while the high controllability of molecular engineering enables directional regulation of PDI photocatalytic performance through site-specific (1,6- vs. 1,7-positions) and functionality-tailored (electron-donating vs. electron-withdrawing) substituent design. However, limitations include the following: (1) Imide modifications can enhance the internal electric field (IEF), but demonstrate limited ability to significantly alter the energy levels or spectral characteristics of PDI [92], resulting in restricted light absorption range extension; (2) While bay-region substitutions effectively reduce band gaps and improve the light absorption range, they may increase carrier recombination probability [95] and induce structural stability concerns [51], coupled with the current scarcity of systematic investigations on bay-position modifications requiring further exploration; (3) Doping strategies face inherent challenges, including high synthesis costs, questionable long-term stability, and complex fabrication processes. Collectively, multidimensional molecular engineering through spatial site coordination and elemental functionalization systematically optimizes light absorption, charge dynamics, and stability.

3.2. Nanostructuring

The morphology of a catalyst directly influences its catalytic performance. By regulating the morphology, it is possible to increase the specific surface area, provide more active sites, alter energy band structures, and broaden the absorption spectrum range.

PDI exhibits two distinct π-π stacking configurations: H-type and J-type, which are governed by the interactions between adjacent molecules. In terms of functionality, H-aggregate structures outperform J-aggregates in terms of photogenerated carrier migration and separation capabilities, as well as the oxidation potency of their active components. Ghosh [105] conducted extensive research on the aggregates of perylene bisimide (PBI) dyes containing amide groups at imide positions, and discovered that straight-chain linear alkyl groups (with minimal steric hindrance) have the tendency to form H-aggregates, whereas branched alkyl chain substitutions give rise to slipped J-aggregates. J-aggregates demonstrate notable absorbance across the entire visible-light spectrum, with the lowest absorbance occurring at 550 nm. A few studies have shown that by adjusting the length of the methylene linker between the acyl (acylimide) position of the PDI derivative and the hydrophilic carboxylic acid head group, two supramolecular forms, H-PDI and J-PDI, with differing π-π stacking arrangements, can be synthesized in HCl aqueous solutions (Figure 4a) [52]. Short CH₂ chains (consisting of 2 units) yield H-type nanowires ranging from hundreds of nanometers to micrometers in length, while long CH₂ chains (consisting of 10 units) develop into J-type nanorods that are hundreds of nanometers long. The absorption edge spectrum of H-PDI is approximately 734 nm, which is red-shifted by 39 nm compared to J-PDI (695 nm) (Figure 4b). The current intensity of H-PDI aggregates is 7.7 times greater than that of J-PDI aggregates (Figure 4c). These findings suggest that H-aggregates exhibit superior light-harvesting efficiency and carrier mobility compared to J-aggregates. This is attributed to the higher π-electron conjugation and deeper valence band position of H-aggregates [52].

In terms of self-assembled morphology, variations in the hydrophobicity–hydrophilicity and steric hindrance of side chains influence the self-assembly process. Specifically, H-aggregates with short side chains in water exhibit a reduced nucleation rate and an extended crystal growth period, which favors unidirectional growth into nanowires, rather than the formation of the thick, block-like aggregates typical of J-type structures [92]. Researchers have thoroughly examined the impact of side chains on the molecular stacking conformations of two PTCDI molecules possessing distinct side chain architectures [92]. They synthesized two derivatives, namely N,N′-di(dodecyl)-perylene-3,4,9,10-tetracarboxydimide (DD-PTCDI) and N,N′-di(nonyl)-perylene-3,4,9,10-tetracarboxydimide (ND-PTCDI), with differing side chain modifications. Due to the combined effects of hydrophobic interactions between the linear side chains of DD-PTCDI molecules and π-π interactions between the perylene planes, along with the hindrance imposed by the steric bulk of ND-PTCDI side chains on molecular stacking, which disrupts one-dimensional self-assembly growth, these two molecules self-assemble into distinct aggregate morphologies: one-dimensional (1D) nanoribbons and zero-dimensional (0D) nanoparticles, respectively (Figure 4d). Wang et al. [106] developed organic nanofibers with exceptional catalytic performance under visible light, utilizing carboxyl-substituted PTCDI supramolecules (Figure 4e). These materials exhibit an H-type π-π stacking structure, where long-range electron delocalization within this stacking significantly boosts the migration efficiency of photoinduced charge carriers. Furthermore, the nanofibers provide an increased specific surface area, enhancing their catalytic capabilities (Figure 4f).

Structural alterations to both the terminal ends and the perylene ring of PDI molecules can tailor the morphology of their self-assembled structures. When it comes to modifications at the terminal ends, the morphology largely depends on the steric hindrance imparted by the side chains, as bulky side chain groups disrupt the periodic and orderly arrangement of the molecules. Experimental evidence suggests that PDI can self-assemble independently with ortho-, meta-, and para-phthalate substituents through solution-based methodologies [66]. Notably, the meta-phthalate and para-phthalate substituted structures arrange into one-dimensional nanorods, with lengths of 51.5 nm and 37 nm, respectively. Conversely, the ortho-phthalate substituted structure, due to pronounced steric effects and augmented hydrogen bonding, assembles into ultrathin nanosheets with a thickness of just 0.8 nm. This occurs because π-π stacking self-assembly is inhibited, preventing the formation of nanorod structures. It underscores the fact that variations in the steric effects and hydrogen bonding associated with different substituents lead to diverse assembly structures. Furthermore, the nanosheet structure, by providing abundant active sites and reducing the carrier transport distance, renders the ortho-phthalate substituted PDIs superior in catalytic activity. Under visible light, they exhibit a hydrogen production rate of 1.1 mmol∙g⁻¹∙h⁻¹, which is 9.2 times and 4.8 times greater than that of meta-phthalate-PDIs and para-phthalate-PDIs, respectively.

PDI nanostructures typically encompass nanoparticles, nanofibers, and nanosheets, with the carrier transport channels they form being well suited to the relatively short exciton diffusion length (LD, ranging from 5 to 20 nm) of organic semiconductors [107]. Notably, the ultrathin dimensions of these nanostructures result in transport channel distances that are shorter than the mean free path of photogenerated carriers. This distinctive characteristic plays a pivotal role in accelerating carrier transport and facilitating the efficient separation and migration of photogenerated carriers, thereby enhancing the photocatalytic activity of the materials.

In one study, Zhang and his colleagues [108] synthesized highly crystalline π-conjugated-PDI nanosheets, π-stacked-PDI nanorods, and anisotropic-PDI nanoparticles. Their comparative analysis of the photoelectric properties of these three morphological structures of PDI molecules unveiled intriguing findings. Specifically, π-conjugated (−112) planes on PDI nanosheets demonstrated an exceptionally high exposure ratio of 99.3%, far surpassing the 1.2% and 0% recorded for the other two structures. Moreover, these π-conjugated molecules exhibited a stronger internal electric field, which was 2.80 times and 6.01 times greater than that of the π-stacked and anisotropic ones, respectively. Further insights were gained through DFT calculation analysis and photoelectrochemical (PEC) and photoluminescence (PL) experiments. These studies revealed that π-conjugated planes possess a higher electron density, leading to a more robust internal electric field and a favorable valence band edge potential. These characteristics enable the effective separation and transfer of photogenerated charges. Additionally, the low in-plane electrostatic potential and high surface energy of π-conjugated PDI molecules facilitate charge transfer, further enhancing their photocatalytic activity. Remarkably, this enhanced activity is 8–17 times greater than that reported for previously studied PDI photocatalysts.

Another study indicated that the thinner the nanosheets were, the better their performance would be. Meng et al. [109] prepared PDI photocatalysts by replacing 3,4,9,10-Perylenetetracarboxylic acid dianhydride (PTCDA) with 3-amino-2-hydroxypropanoic acid, beta-Alanine acid, and n-propylamine, resulting in s-PDI-D1, s-PDI-P1, and s-PDI-B1, respectively. These photocatalysts exhibited nanosheet structures with thicknesses of 6.2, 4.7, and 5.4 nm. Notably, s-PDI-P1, the thinnest among them, displayed a finer nanofiber structure under transmission electron microscope observation. Fourier transform infrared spectroscopy and Raman spectroscopy revealed that hydrogen bonding among carboxyl groups and between hydroxyl groups and the perylene core reduced interlayer spacing and increased the degree of π-conjugation. Consequently, this led to a shorter charge transfer distance, facilitating the effective separation and migration of exciton pairs. Specifically, the photocatalytic performance of s-PDI-P1 was 18 times higher than that of pure PDI, 1.8 times higher than that of s-PDI-D1, and 1.5 times higher than that of s-PDI-B1.

The nanostructural modulation strategy effectively enhances photocatalytic performance by tailoring aggregation motifs, assembly pathways, and dimensional characteristics. Aggregation control prioritizes H-type stacking (short alkyl chains, e.g., C2), which strengthens π-conjugation (absorption edge: 734 nm, 39 nm redshift vs. J-type) and boosts carrier mobility (7.7 × higher current intensity than J-aggregates) via enhanced internal electric fields and deeper valence bands [52,105,108]. Assembly regulation employs side chain engineering: linear hydrophobic chains (e.g., DD-PTCDI) drive 1D nanoribbon growth, while ortho-substituents with steric hindrance and hydrogen bonding enable ultrathin 2D nanosheets (0.8 nm thickness), achieving 4.8–9.2 × higher H₂ production than meta/para-substituted analogs [66,92]. Dimensional optimization reveals 2D nanosheets (e.g., ortho-PDI) as optimal, with thicknesses <5 nm (<exciton diffusion length, 5–20 nm), exposing 99.3% active (−112) π-conjugated planes and amplifying built-in electric fields (6× stronger), thereby elevating charge separation and photocatalytic activity by 8–18× versus conventional structures [66,107,108,109]. Collectively, H-type 2D nanosheets synergistically integrate broad-spectrum absorption, efficient carrier dynamics, and abundant active sites, establishing a paradigm for high-performance organic photocatalysts [52,105,108,109]. The current limitations of nanostructural modification strategies in photocatalyst design are highlighted as follows: (1) The trade-off between light absorption and carrier dynamics in H- vs. J-aggregates: H-aggregates (e.g., short-chain H-PDI) exhibit superior carrier mobility and oxidation potential (7.7× higher photocurrent than J-aggregates), yet their light absorption range is restricted (absorption edge: 734 nm). In contrast, J-aggregates (e.g., long-chain J-PDI) achieve broad visible-light coverage (absorption edge: 695 nm) but suffer from inefficient carrier separation due to slipped π-π stacking [52,105]; (2) Dimensionality-dependent performance compromises: Zero-dimensional nanoparticles provide abundant active sites, but suffer from disordered stacking, exacerbating carrier recombination. One-dimensional nanoribbons enhance directional carrier transport, but limit active site density, due to reduced surface accessibility. Two-dimensional nanosheets balance shortened carrier transport distances and high active site exposure. However, ultrathin structures (e.g., s-PDI-P1, 4.7 nm thick) face challenges in precise thickness control during assembly and increased surface defects acting as recombination centers; (3) Limited exploration of complex nanostructures: Current studies predominantly focus on 1D (nanorods, nanofibers) or 2D (nanosheets) morphologies, while advanced architectures (e.g., nanotubes, hollow nanospheres) remain underexplored. These hierarchical structures could further optimize active site distribution and mass transfer, but require innovative synthetic methodologies to address scalability and stability issues.

3.3. Heterostructuring

Making heterostructures by coupling two compositionally different semiconductors represents a facile method to enable the effective spatial separation of electron–hole pairs. Currently, there exist three different types of heterojunctions: Type II, Z-scheme, and S-scheme configurations.

A Type II heterojunction is commonly recognized as a conventional configuration within the realm of heterojunctions. As depicted in Figure 5a, the band gaps of the two constituent semiconductors exhibit a staggered arrangement, where the LUMO and HOMO energy levels of semiconductor A are positioned above those of semiconductor B. This staggered alignment naturally creates a driving force for the photogenerated electrons to migrate spontaneously towards semiconductor B, while the photogenerated holes migrate in the opposite direction towards semiconductor A. This spatial separation of electron–hole pairs is critical in minimizing their recombination rates and thereby enhancing the overall photocatalytic activity [110,111].

Zhang et al. [112] demonstrated the successful fabrication of a two-dimensional nanocomposite system comprising perylenetetracarboxylic diimide (PTCDI) and tetracyanoquinodimethane (TCNQ), utilizing a solution-based self-assembly approach. The primary intermolecular interactions between PTCDI and TCNQ molecules involve π-π stacking and hydrogen bonding, which play a crucial role in the formation of molecular aggregates, as illustrated in Figure 6a. These interactions were proved by X-ray diffraction (XRD) analysis, which exhibited broadening of the crystalline peaks and an augmentation in interlayer spacing from 0.34 nm to 0.38 nm. Within the PTCDI crystallite domain of the composite, the average free path of the photogenerated carriers was determined to be approximately 10 nm, surpassing the interlayer spacing. This configuration favors the migration of photogenerated carriers towards the surface and substrate. Impedance spectroscopy analysis further revealed a notable decrease in the charge transfer resistance of the photogenerated carriers within the TCNQ-PTCDI composite, in comparison to pristine PTCDI, as depicted in Figure 6b. The presence of intermolecular charge transfer interfaces within the nanocomposite facilitates the efficient transfer of photogenerated electrons from the HOMO of PTCDI to the LUMO of TCNQ. Consequently, the photocurrent density of the composite was observed to be tenfold higher than that of pure PTCDI, as shown in Figure 6c. Ultimately, the composite exhibited a photocatalytic efficiency that was 10.4 times greater than that of pure PTCDI.

As an illustrative example, a donor–acceptor (D-A) composite, designated as PDI-TPPS, was synthesized via π-π stacking interactions between the PDI electron donor and the tetra(4-sulfonatophenyl)porphyrin (TPPS) acceptor (Figure 6d) [71]. The absorption spectrum of the composite exhibited a full-spectrum response, presenting a theoretical spectral efficiency as high as 72%. Furthermore, it generated a significantly enhanced internal electric field (Figure 6e), characterized by a surface potential of E = 70.16 mV, which substantially surpassed the surface potentials of the individual components, namely bare PDI (E = 40.41 mV) and TPPS (E = 13.07 mV). This robust built-in electric field within the composite effectively promoted charge separation dynamics, resulting in an elongated excited-state lifetime for PDI-TPPS compared to its constituent molecules. Specifically, the carrier lifetime at 495 nm was prolonged to 2.02 µs, exceeding that of TPPS (1.72 µs). The synergistic combination of high spectral efficiency and a boosted internal electric field endowed the PDI-TPPS composite with exceptional photocatalytic hydrogen production activity, allowing it to achieve a hydrogen evolution rate of approximately 30.36 mmol∙g⁻¹∙h⁻¹. This rate was approximately 9.95 times and 9.41 times higher than that of pure TPPS and PDI, respectively (Figure 6f), highlighting the superior performance of the composite system.

In contrast to Type II heterojunctions, where photogenerated electrons typically migrate to the lower-lying LUMO, thereby decreasing their redox potential [113], the construction of Z-scheme heterojunctions represents an exceptionally efficient strategy to fully utilize the redox capabilities of photogenerated charge carriers. Bard first proposed the concept of a traditional Z-scheme photocatalyst in 1979 [114]. This photocatalyst comprises three integral parts: photocatalyst I (PS I), photocatalyst II (PS II), and the acceptor/donor (A/D) system, also referred to as the redox mediator or redox ion pair (Figure 5b). In this configuration, PS I and PS II remain non-contacting, yet under illumination, they independently generate electron–hole pairs. Their staggered band structures allow for the efficient transfer of electrons from the LUMO of PS II to HOMO of PS I. This process entails the following redox reactions, described in Equations (3) and (4) [113], involving the photogenerated charges:

$$A+e^{-}\to D$$

(3)

$$D+h^{+}\to A$$

(4)

However, traditional Z-scheme systems have typically been confined to liquid phases due to mobility requirements, necessitating the incorporation of solid-state mediators for electron transfer in order to realize fully solid-state Z-scheme photocatalysts. In 2006, the Tada group [115] pioneeringly designed a solid electron mediator to replace redox ion pairs, thereby enabling its application across liquid, gas, and solid phases. As depicted in Figure 7a, upon light excitation, Photosystem II (PS II) and Photosystem I (PS I) generate electron–hole pairs. The photogenerated electrons in the LUMO of PS II are transferred to the HOMO of PS I via an electron mediator, such as Pt, Ag, Au, or carbon materials, and are subsequently re-excited to the LUMO of PS I. Solid-state mediators significantly reduce charge transfer distances compared to liquid-phase ion mediators, thereby enhancing carrier separation and redox capabilities.

For instance, Chen et al. [72] introduced reduced graphene oxide (rGO) between a perylene tetracarboxylic diimide polymer (PDIP) and graphitic carbon nitride (g-C₃N₄), as illustrated in Figure 7b. This established a highly efficient Z-scheme interface for electron transfer, enabling photocatalytic overall water splitting. Experimental results revealed that the N = C-N peak of g-C₃N₄ shifted negatively by 0.25 eV, while the N-(C = O)3 peak of PDIP shifted to a higher position by 0.27 eV. This increase in electron cloud density on the g-C₃N₄ side and decrease on the PDIP side confirmed Z-scheme electron coupling interactions among the ternary components. Specifically, a high-quality interface heterojunction contact was formed between PDIP and g-C₃N₄, with rGO serving as the electron transfer mediator. Furthermore, the average fluorescence lifetime of the g-C₃N₄/rGO/PDIP composite was 5.04 ns, lower than that of g-C₃N₄ alone (6.98 ns). The charge transfer resistance (Rct) of the composite was only 1.2 kΩ, significantly lower than that of PDIP (68.7 kΩ) and g-C₃N₄ (167.5 kΩ), as shown in Figure 7c. Consequently, this Z-scheme heterojunction exhibited efficient and stable photocatalytic overall water splitting activity, with H₂ and O₂ evolution rates of 15.80 and 7.80 µmol∙g⁻¹∙h⁻¹, respectively. These rates were approximately 12.1 times higher than those of pure g-C₃N₄, as illustrated in Figure 7d. Additionally, a quantum efficiency of 4.94% at 420 nm and a solar-to-hydrogen conversion efficiency of 0.30% were achieved.

However, fully solid-state Z-scheme systems still fail to mitigate the adverse effects of side reactions and intermediate mediators. Therefore, direct Z-scheme photocatalysts were designed, as depicted in Figure 7e. Similarly to traditional Type II photocatalysts, only two photocatalysts are involved, but the distinction lies in the electron transfer mechanism. In direct Z-scheme systems, photogenerated electrons migrate from the LUMO of PS I to the electron-rich HOMO of PS II. In contrast, in traditional systems, they migrate from the LUMO of PS II to the hole-rich HOMO of PS I. This conversion transforms the electrostatic repulsion present in electron transfer in traditional Type II photocatalysts into electrostatic attraction, facilitating photogenerated electron migration while retaining potent redox capabilities.

For instance, an all-organic wide-responsive-responsive (254–700 nm) supramolecular 1D/2D heterojunction was fabricated by loading self-assembled perylene tetramethyl diimide nanowires onto the surface of g-C₃N₄ nanosheets, termed g-C₃N₄/PDI (Figure 7f) [73]. Upon exposure to full-spectrum light, photogenerated electrons at the HOMO energy level of perylene tetramethyl diimide could transition to the LUMO energy level of g-C₃N₄ to recombine with holes, thereby inhibiting the recombination of photogenerated carriers. The electrons and holes involved in redox reactions resided at the highest possible HOMO energy level of g-C₃N₄ and the lowest possible LUMO energy level of perylene tetramethyl diimide, respectively. Electron paramagnetic resonance (EPR) measurements affirmed the presence of robustly formed -O radicals, providing compelling evidence for the Z-scheme electron transfer mechanism operating between the two constituents. This Z-scheme electron transfer interface resulted in a charge transfer resistance of 5.81 kΩ for the composite, substantially reduced compared to the resistances of PDI (1695 kΩ) and g-C₃N₄ (7.62 kΩ) individually. Furthermore, the photovoltage intensity of the composite was nearly 400 times greater than that of pure g-C₃N₄, marking a significant enhancement in carrier separation efficiency. Time-resolved fluorescence decay spectra revealed that the composite exhibited a fluorescence lifetime of 0.752 ns, shorter than the 2.242 ns observed for pristine g-C₃N₄ nanosheets. The absorption spectrum of the composite integrated the responsive ranges of both components, achieving a nearly comprehensive spectral response (ranging from 360 nm to 690 nm) (Figure 7g). Consequently, the hydrogen production rate for the PDI/g-C₃N₄ composite was 1649.93 µmol∙g⁻¹∙h⁻¹, representing a 2.03-fold increase compared to the rate of 814.03 µmol∙g⁻¹∙h⁻¹ achieved with pure g-C₃N₄ (Figure 7h).

S-scheme photocatalysts represent a burgeoning category of heterojunctions constructed through coupling of a reduction photocatalyst (RP) and an oxidation photocatalyst (OP), facilitated by a sacrificial agent to eliminate photogenerated holes from the RP. The two components exhibit staggered energy band structures: specifically, the RP possesses a lower work function and a correspondingly higher Fermi level, whereas the OP exhibits a higher work function and a lower Fermi level. Upon contact, electrons spontaneously migrate from the RP to the OP to equilibrate their Fermi levels, ultimately achieving a common Fermi level. This electron transfer results in the formation of a negatively charged interface on the OP side and a positively charged interface on the RP side, generating an internal electric field, as depicted in Figure 5c. In terms of energy bands, due to alterations in the Fermi level, the band edge of the RP bends upwards upon the loss of electrons, whereas the band edge of the OP bends downwards upon gaining electrons. When illuminated, both the RP and OP are excited, generating electron–hole pairs. The band bending facilitates the efficient recombination of photogenerated electrons in the LUMO of the OP with photogenerated holes in the HOMO of the RP. The internal electric field accelerates carrier migration, thereby effectively inhibiting the recombination of photogenerated electrons in the CB of the RP with photogenerated holes in the valence band (VB) of the OP. Furthermore, the electrostatic attraction between photogenerated holes and electrons promotes their recombination at the interface, effectively eliminating unwanted carriers. This unique configuration enables the RP to supply electrons for reduction reactions and the OP to provide holes for oxidation reactions, thereby achieving the separation of photogenerated electron–hole pairs, while maintaining robust redox capabilities [116].

To give an example, Li et al. [74] successfully developed an S-scheme all-organic heterojunction through the in situ coupling of a two-dimensional triazine-based imine-linked covalent organic framework (TATF-COF) with a perylene diimide urea polymer (PUP). As illustrated in Figure 8a, TATF-COF exhibits a crystalline nanosheet morphology with a size of approximately 600 nm, whereas PUP crystallizes into nanorods with a length of about 300 nm. DFT calculations reveal that PUP possesses a higher work function (5.23 eV) compared to TATF-COF (4.97 eV), with TATF-COF exhibiting a higher Fermi level. Upon contact between TATF-COF and PUP, electrons spontaneously migrate from TATF-COF to PUP until their Fermi levels equilibrate. Consequently, TATF-COF becomes positively charged due to the loss of electrons, while PUP becomes negatively charged upon gaining electrons. This results in the formation of an S-scheme internal electric field at the interface, directed from TATF-COF to PUP. The band edges of the two components undergo bending: the band edge of TATF-COF bends upwards, and the band edge of PUP bends downwards. Electron spin resonance (ESR) spectroscopy analysis, as shown in Figure 8b, indicates a significantly higher signal intensity of superoxide radicals in the TATF-COF/PUP composite compared to pure TATF-COF and PUP. This confirms the electron transfer mechanism of the S-scheme heterojunction. Under visible-light irradiation, with ascorbic acid as the sacrificial agent and Pt as the cocatalyst, the photocatalyst system containing 3 mg of PUP achieved an optimal hydrogen production rate of 94.5 mmol∙g⁻¹∙h⁻¹, 3.5 times higher than that of pure TATF-COF (Figure 8c). Additionally, it demonstrated an exceptionally high apparent quantum efficiency (AQE) of 19.7% at 420 nm.

In another work, Liu et al. synthesized an S-scheme TiO₂/PDI heterojunction through a straightforward solvent composite method [75]. Uniform and dense TiO₂ nanoparticles were loaded onto sheet-like PDI (Figure 8d), with TiO₂ serving as the RP and PDI as the OP. Due to their differing work functions, electrons spontaneously migrated from TiO₂ to PDI. Under solar irradiation, electrons in both PDI and TiO₂ were excited, with electrons in PDI transitioning from the HOMO to the LUMO and electrons in TiO₂ transitioning from the VB to the CB. Under the influence of the internal electric field, electron transfer from the CB to the LUMO was hindered, while electrons in the HOMO and holes in the VB were more prone to recombination due to the electric field (Figure 8e). Under the S-scheme charge transfer mechanism, this heterojunction achieved hydrogen evolution rates of 238 mmol∙g⁻¹∙h⁻¹ and oxygen evolution rates of 114.18 mmol∙g⁻¹∙h⁻¹ for visible-light photocatalytic overall water splitting, with a stoichiometric ratio of H₂ to O₂ close to 2:1 (Figure 8f).

Heterojunctions enhance photocatalytic performance by achieving efficient charge separation and extended spectral response through interface engineering (constructing intimate heterointerfaces), charge transport layer design (incorporating “electron highways” to accelerate carrier migration), and optimized energy level alignment. However, heterojunction modification strategies still confront common drawbacks and type-specific limitations. Common drawbacks include the following: (1) complex interface engineering requiring precise band alignment and molecular interactions, susceptible to lattice mismatch or surface defects; (2) interfacial charge transfer resistance and recombination risks induced by band bending or impurities; (3) stability and scalability issues, particularly for organic materials prone to photocorrosion and inhomogeneous self-assembled interfaces. Type-specific limitations include the following: (1) the Type II configuration suffers from weakened redox capability due to electron migration to the lower-energy LUMO; (2) the Z-scheme configuration relies on liquid/solid mediators, increasing costs and side reactions, while direct Z-scheme mechanisms require complex verification [73]; (3) the S-scheme configuration necessitates sacrificial agents (e.g., ascorbic acid) and constrained built-in electric fields governed by work function differences [74]. Future breakthroughs demand theoretical designs (e.g., DFT-guided interface optimization) and process innovations (e.g., defect passivation, sacrificial agent-free systems).

4. Conclusions and Perspectives

In summary, this article thoroughly explores the unique properties and complex structure of PDI molecules, and systematically reviews various modification techniques aimed at enhancing the photocatalytic efficiency of PDI-based materials. Perylene diimide (PDI) has emerged as a promising organic photocatalyst, due to its tunable π-conjugated architecture, controllable self-assembly, and strong electron-accepting capability. Advances in molecular engineering (e.g., site-specific substitutions and elemental doping), multidimensional nanostructuring (ultrathin 2D nanosheets with optimized H-aggregation), and heterojunction design (Z/S-scheme interfaces) have synergistically enhanced light absorption, charge separation, and catalytic activity, achieving spectral extension to 734 nm and 8–18-fold efficiency improvements. Figure 9 shows the existing engineering challenges and scientific challenges in the prospects of organic nanomaterials for photocatalytic hydrogen evolution. Correspondingly, challenges persist in scalable synthesis, long-term stability, interfacial degradation mechanisms, and the lack of computational databases for precise nanostructure control and reaction pathway elucidation. Addressing these engineering and scientific bottlenecks will drive the development of robust, sustainable PDI-based photocatalytic systems.

Engineering Challenges: (i) Complex Synthesis Protocols and Scalability Limitations: The molecular engineering and nanostructural modulation of PDI-based photocatalysts require precise synthetic control (e.g., low-temperature self-assembly, solvothermal conditions), leading to intricate processes that hinder large-scale production. For instance, precise thickness control (<5 nm) of ultrathin nanosheets necessitates multi-step assembly and purification, while doping strategies (e.g., Zr coordination) involve costly precursors and inert atmospheres. Furthermore, the absence of standardized experimental protocols and parameter databases (e.g., temperature gradients, precursor concentrations, post-treatment procedures) impedes reproducible synthesis of high-performance catalysts and efficient exploration of novel materials. Scalable fabrication of complex architectures remains underexplored. Potential solutions include the following: establish standardized synthesis protocols with quantitative correlations between critical variables (e.g., alkyl chain length, solvent polarity) and material properties; and develop template-assisted one-step synthesis (e.g., microfluidic-controlled self-assembly) and continuous processes (e.g., roll-to-roll coating), aided by machine learning for parameter optimization. (ii) Insufficient Long-Term Stability and Interfacial Degradation: Structural degradation (e.g., photo-oxidation of aromatic cores) and activity loss (e.g., leaching of doped metal sites) occur during prolonged photocatalytic cycles. Heterojunction interfaces suffer from lattice mismatch or carrier-induced stress cracks, resulting in interfacial delamination (e.g., >30% efficiency drop after 100 h irradiation in S-scheme systems). Potential solutions include the following: enhance stability via interfacial passivation (e.g., atomic layer deposition of Al₂O₃), covalent crosslinking (e.g., epoxy-functionalized PDI), or stress-buffering interlayers (e.g., flexible carbon matrices). (iii) Research on Economical Sacrificial Agent Replacements: Developing cost-effective sacrificial agent alternatives and sacrificial-free systems is critical to lowering operational costs and eliminating mechanistic ambiguities in photocatalytic hydrogen production. While advances have been made in broader photocatalytic systems [117,118,119], PDI-based photocatalysis remains underexplored, with only three donor-free systems reported to date [67,72,75]. Further efforts should prioritize synergistic strategies: (1) rational design of PDI derivatives with optimized charge kinetics to minimize sacrificial agent dependence; (2) innovative system engineering leveraging dual-functional redox mediators or self-sustaining cycles.

Scientific Research Directions: (i) Expanding the Variety of PDI Nanostructures: Although PDI has shown promising performance in the forms of nanosheets and nanorods/fibers, there is still a need for continued research into more diversified PDI nanostructures, such as quantum dots, three-dimensional porous structures, and hollow structures. These structures offer higher specific surface areas and more active sites. These new structures have the potential to significantly enhance PDI’s photocatalytic efficiency by providing additional reaction interfaces and active sites. (ii) Reaction Mechanism Research: In-depth research into the photocatalytic hydrogen production mechanism is essential to provide theoretical guidance for catalyst design and reaction process optimization. Techniques such as deep learning can be used to uncover potential correlations within data sets; in situ characterization can be employed to track the reaction process in real time; free radical capture methods can clarify the reaction mechanisms; and nano-electrochemical technologies can be utilized to study the electrochemical behavior of materials at the microscopic level. (iii) Enhancing and Deepening Theoretical Research Techniques: To better understand the mechanisms of photocatalytic hydrogen production and design superior material systems, advanced research techniques must be fully leveraged. However, our current understanding of the relationship between the structure, performance, and function of polymer photocatalysts is based largely on empirical correlations. The discovery of these highly active conjugated polymers is characterized by unpredictability and randomness, leading to many unproductive experiments. By exploring these materials from multiple levels and perspectives and fully utilizing computational simulations to predict material properties, we will be able to design photocatalytic materials with outstanding performance and clearer mechanisms, laying a solid foundation for the future development of photocatalytic hydrogen production technology.