Recent advances in photodegradation of antibiotic residues in water

Graphical abstract

Abstract

Antibiotics are widely present in the environment due to their extensive and long-term use in modern medicine. The presence and dispersal of these compounds in the environment lead to the dissemination of antibiotic residues, thereby seriously threatening human and ecosystem health. Thus, the effective management of antibiotic residues in water and the practical applications of the management methods are long-term matters of contention among academics. Particularly, photocatalysis has attracted extensive interest as it enables the treatment of antibiotic residues in an eco-friendly manner. Considerable progress has been achieved in the implementation of photocatalytic treatment of antibiotic residues in the past few years. Therefore, this review provides a comprehensive overview of the recent developments on this important topic. This review primarily focuses on the application of photocatalysis as a promising solution for the efficient decomposition of antibiotic residues in water. Particular emphasis was laid on improvement and modification strategies, such as augmented light harvesting, improved charge separation, and strengthened interface interaction, all of which enable the design of powerful photocatalysts to enhance the photocatalytic removal of antibiotics.

1. Introduction

In the last few decades, antibiotics have been broadly utilized not only for human therapies, such as curing infectious diseases like COVID-19, transplants, chemotherapy, and surgical interventions, but also for therapeutic and non-therapeutic purposes in aquaculture and animal husbandry and for enhancing crop production [1], [2]. However, the use and overuse as well as the delayed metabolism of antibiotics have led to the inevitable discharge of their residues into aquatic environments, resulting in refractory pollution sources [3]. Conventional wastewater treatment plants (WWTPs) cannot efficiently remove antibiotic residues [4], [5]. This leads to their distribution into ecological systems [6], [7] and eventually into the human body through the food chain or drinking water [8], [9], [10]. Moreover, the long residence times of antibiotic residues in aquatic environments, even at low concentrations, may lead to the propagation of bacteria with antibiotic resistance and even multiple drug resistance, which may cause life-threatening infections [3], [11], [12]. Therefore, developing an effective approach to degrade or remove antibiotic residues from aquatic environments is crucial.

Different methods have been developed to treat antibiotic residues in water and wastewater before their final release into the environment. Table 1 summarizes the current main approaches, including traditional techniques and newly developed methods.

Table 1.

Methods for antibiotic degradation or removal in water.

Treatment	Materials	Antibiotic	Operating conditions	Disadvantages	Ref.
Adsorption processes	Anionic surfactant sodium dodecyl sulfate (SDS)	Amoxicillin Ampicillin	Contact time: 40 min Agitation speed: 350 rpm Temperature: 40 °C pH: 4	Low removal capacities; difficult separation; secondary environmental pollution; unsatisfactory recycling ability	[13]
	KOH-modified biochar	Norfloxacin Sulfamerazine Oxytetracycline	Temperature: 15–35 °C pH 3–9		[14]
	Graphene oxide/cellulose nanofibril hybrid aerogel	Doxycycline Chlortetracycline Oxytetracycline Tetracycline	Temperature: 25 °C		[15]
	Pyrogenic carbonaceous materials	Ciprofloxacin	Contact time: 72 h Temperature: 25 °C pH 7.5 and 9.5		[16]
	Clean and dried Kigelia pinnata modified with ionic liquid	Ibuprofen Ketoprofen Ampicillin Diclofenac	pH 2.5 or 5		[17]
	Manure-derived biochars	Lincomycin	pH 6 or 10		[18]
	Lanthanum modified diatomite	Tetracycline antibiotics	Contact time: 24 h pH 3–10		[19]
Adsorption processes	Grape stalk	Ofloxacin Chrysoidine	pH 4, 7, and 9	Low removal capacities; difficult separation; secondary environmental pollution; unsatisfactory recycling capacity	[20]
	Cleaned and dried Pachydictyon coriaceum and Sargassum hemiphyllum	Tetracycline	Temperature: 15–35 °C pH 3–9 Salinity: 0–100 mM NaCl		[21]
	Spent mushroom substrate	Sulfamethyldiazine Sulfamethazine Sulfathiazole Sulfamethoxazole	Temperature: 15 °C pH 3–11		[22]
Coagulation /flocculation /sedimentation	Amino-acid-modified-chitosan flocculants	Norfloxacin Sulfadiazine Tylosin	Temperature: 25 °C pH 6, 7, 8	Antibiotics cannot be completely removed and secondary pollution occurs readily	[23]
Ozonation	Ozone	Amoxicillin Doxycycline Ciprofloxacin Sulfadiazine	Neutral pH	Demands high equipment and energy costs	[24]
	Ozone/zero-valent iron	Flumequine	Contact time: 1h Initial pH 2.5 Fe (0) dosage: 60 g/L Ozone flow rate: 0.25 L/min Temperature: 30 °C		[25]
Ozonation	Medium-high frequency ultrasound and ozone	Amoxicillin	Medium-high ultrasonic frequency waves: 575, 861, 1141 kHz pH 7, 10	Demands high equipment and energy costs	[26]
	Ozone	Flumequine	Contact time: 6 min Temperature: 25 °C pH 3, 5, 7, 9, 11		[27]
	Ozone	Ofloxacin	Temperature: 25 °C pH: 2, 7, 12		[28]
	Multistage ozone and biological treatment system	Amoxicillin	Temperature: 25 °C pH: 10	Complex; high operating costs; continuous use is impractical	[29]
	Chemical coagulation and microfiltration	Ibuprofen Ephedrine Propranolol	Different doses of ZnO nanoparticles: 0.5, 0.7, 1.0, 1.3, 1.5, 1.7 g/L pH 7, 9		[30]
	Electric coagulation and photo-electro-Fenton process	Metronidazole	pH: 1, 3, 5, 7, 9 Metronidazole concentration: 50 mg/L Voltage: 5–30 V H2O2: 0–0.02 mol/L Temperature: ~20 °C UV at 230 nm Number of UV lamps: 1–4		[31]
Combined processes	A membrane bioreactor (MBR) integrated with solar Fenton oxidation	Sulfamethoxazole Erythromycin Clarithromycin	H2O2: 20–100 mg/L pH: 2.8 [Fe2+]: 5 mg/L	Complex; high operating costs; impracticability in continuous use	[32]
	Integrated adsorption-membrane filtration process	Norfloxacin Ofloxacin	pH: 7 Temperature: 25 °C UV at 276 nm and 293 nm pressure range: 34.47–172.36 kPa		[33]
	Ultraviolet, chlorination, ozone disinfection	Antibiotic resistance genes	Chlorine concentrations: 2–32 mg/L Ozone concentration: 2–10 mg/L UV (mJ/cm2): 10–160		[34]
	Adsorptive magnetic ion exchange resin	Sulfamethoxazole Tetracycline Amoxicillin	Contact time: 30 min Temperature: 25 °C MIEX resin dosage: 5 mL/LAntibiotic: 1000 μg/L		[30]
	Nanofiltration and chlorination	Sulfanilamide Sulfadiazine Sulfamethoxazole Sulfadimethoxine	Membrane effective area: 40.92 cm2 Operating pressure: 5 bar Temperature: 25 °C Flow rate: 120–150 mL/min Initial pH: 5.6, 7.2, 10 Initial concentration: 2.0 × 10−5 M		[35]

These current methods can be loosely classified into three types according to the treatment mechanisms or principles involved: physical removal, biological treatment, and chemical degradation. As shown in Table 1, physical methods such as adsorption, sedimentation, flocculation, and filtration only separate the antibiotic residues from the water and generate problematic products such as brine and contaminated adsorbents. Alternatively, biological approaches have recently emerged, and most antibiotic residues in the environment can be removed through this route [36], [37], [38]. However, the artificial introduction of active organisms into aquatic environments may disrupt the ecological balance of their biomes, which may cause irreversible ecosystem damage. Moreover, biological processes suffer from some key disadvantages such as being time-consuming and often unreliable. Therefore, the chemical degradation of antibiotics has gained strong interests.

Different chemical approaches such as ozonation, chlorination, and Fenton’s oxidation have been developed for the treatment of antibiotic residues in water, as shown in Table 1. Unfortunately, complete mineralization is difficult to achieve or may otherwise be prohibitively lengthy. In some cases, these methods may kill non-target organisms due to their low selectivity, which causes unintended damages [39], [40]. Additionally, this method incurs high capital and high operating costs. Combinations of physical and chemical degradation processes can significantly reduce the toxicity of treated effluents during the removal of antibiotic residues from water. Nevertheless, these methods are complex and costly [41].

Alternatively, photocatalysis has broad application prospects for environmental remediation due to its unique advantages, including (1) easily-attainable reaction conditions (i.e., near ambient temperature and mostly ambient pressure), the use of oxygen in the air to produce a powerful oxidant, and the use of solar radiation as an energy source; (2) potentially complete decomposition of organic pollutants into innocuous inorganic molecules such as carbon dioxide and water; (3) strong redox ability, low cost, no adsorption saturation, and long durability. Therefore, photocatalysis has increasingly garnered worldwide interest and has been broadly implemented in novel energy extraction and environmental control strategies.

Photocatalysis is an advanced oxidation process that has previously been applied in the treatment of antibiotic residues. As shown in Fig. 1 , less than 40 studies had been published on the photocatalytic treatment of antibiotic residues, prior to 2014. However, a sharply increasing number of publications on this subject have been released thereafter. Substantial progress has been made in recent years; however, photocatalysis still suffers from some key imperfections, including insufficient visible light utilization, rapid annihilation of photogenerated carriers, and incomplete mineralization, all of which strongly restrict its commercial application. A thorough overview concerning the fundamentals, improvement, and modification of strategies and challenges of the photocatalytic treatment of antibiotic residues was yet to be compiled, and therefore, these topics have been comprehensively addressed herein.

Fig. 1.

Number of search results of recent publications addressing the photocatalytic treatment of antibiotic residues using “photocatalytic” and “antibiotic treatment” as keywords (collected from the Web of Science Core Database: March 9, 2020).

This article comprehensively reviews the latest advances in the photocatalytic treatment of antibiotic residues in water. Specifically, the fundamentals of photocatalysis will be introduced. Different strategies will then be summarized and classified, followed by a discussion on the improvement and modification strategies to enhance the photocatalytic degradation of antibiotics. Finally, this review will conclude with a summary of major challenges and key perspectives.

2. Basic principles of photodegradation

When a semiconductor is exposed to irradiation with energy beyond its optical band gap, electrons are excited and shifted from its valence band (VB) towards the conduction band (CB), producing an equal number of positively charged holes in the VB. When the potential of VB vs NHE is more positive than ${\mathrm{H}}_2\mathrm{O}/\bullet \mathrm{O}\mathrm{H}(+2.72\mathrm{V}\mathrm{v}\mathrm{s}\mathrm{N}\mathrm{H}\mathrm{E})$ or ${\text{O}\mathrm{H}}^{-}/\bullet \mathrm{O}\mathrm{H}(+1.89\mathrm{V}\mathrm{v}\mathrm{s}\mathrm{N}\mathrm{H}\mathrm{E})$, and the potential of CB vs NHE is more negative than${\text{O}}_2/\bullet {\text{O}}_2^{-}($−0.33 V vs NHE), the semiconductor will be able to generate $\bullet \mathrm{O}\mathrm{H}$ and $\bullet {\text{O}}_2^{-}$. Thereafter, the photoinduced electrons and holes separate and migrate to the surface of the semiconductor, and redox reactions will occur at the reactive site on the semiconductor surface (Fig. 2 ) [42], [43].

Fig. 2.

Schematic representation of the semiconductor photocatalysis process [44]

The reaction mechanisms of semiconductor photocatalysis are typically expressed by the following equations [45], [46]:

Semiconductor $+$ Light Energy (λ$\ge E_g$)$\to$ Semiconductor ($e_{\mathit{cb}}^{-}$ $+$ $h_{\mathit{vb}}^{+}$) (1)

$h_{\mathit{vb}}^{+}+$${\mathrm{H}}_2\mathrm{O}\to$${\mathrm{H}}^{+}+\bullet$OH (${\mathrm{H}}_2\mathrm{O}/\bullet \mathrm{O}\mathrm{H}|+2.72\mathrm{V}\mathrm{v}\mathrm{s}\mathrm{N}\mathrm{H}\mathrm{E}$) (2)

$h_{\mathit{vb}}^{+}$$+$${\text{O}\mathrm{H}}^{-}$$\to \bullet$OH (${\text{O}\mathrm{H}}^{-}/\bullet \mathrm{O}\mathrm{H}|+1.89\mathrm{V}\mathrm{v}\mathrm{s}\mathrm{N}\mathrm{H}\mathrm{E}$) (3)

$e_{\mathit{cb}}^{-}$$+$${\text{O}}_2$$\to \bullet {\text{O}}_2^{-}$ (${\text{O}}_2/\bullet {\text{O}}_2^{-}|$ − 0.33 V vs NHE) (4)

Pollutant $+$ Active species $\left(h_{\mathit{vb}}^{+},e_{\mathit{cb}}^{-},\bullet \mathrm{O}\mathrm{H},\bullet {\text{O}}_2^{-}\right)\to$Degradation Products (5)

Through these chemical processes, solar energy can be directly converted and effectively utilized. For instance, Zhao et al. [47] successfully prepared ZnSe quantum dot (QD)/g-C₃N₄ composites with remarkable photocatalytic properties and utilized them for the photocatalytic degradation of ceftriaxone sodium. Based on the results, $\bullet$OH and h⁺ were found to be the main active materials. The potential reaction mechanisms are described below.

• ZnSe QDs/g-C₃N₄ $+$ hv $\to$ h⁺ $+$ e⁻

• h⁺ $+$ H₂O $\to$ H⁺ $+$ •OH

• e⁻ $+$ 1/2O₂ $+$ H⁺ $\to$ •OH

• h⁺ $+$ •OH $+$ Ceftriaxone sodium $\to$ CO₂ $+$ H₂O $+$ other small molecules

However, limited optical utilization and the quick annihilation of photoexcited electron-hole pairs diminish the effects of photocatalytic activity. Photocatalysts are able to overcome these deficits, provided that they meet the following criteria: (1) appropriate spectral absorption range, (2) proper band energy structure for sufficient separation and transport of electron–hole pairs, and (3) adequate active sites for adsorption or reaction [48], [49], [50]. It is essential to satisfy all three of the aforementioned prerequisites to improve photocatalytic efficiency, and abundant efforts have been made to systematically design photocatalysts and optimize photocatalytic dynamics.

3. Strategies for photocatalytic activity improvement

The photocatalytic treatment of antibiotic residues in aquatic environments has recently become the focus of much attention, and diverse strategies have been proposed to improve the photocatalytic efficiency (Scheme 1 ). Examples of such reported strategies are summarized in Table 2 .

Scheme 1.

Strategies for photocatalytic efficiency improvement.

Table 2.

Summary of previous studies using different strategies to improve photocatalytic degradation of antibiotics.

Strategies	Catalysts and concentration	Light source	Antibiotic and concentration	Degradation efficiency	Ref.
Vacancies	BiOCl with abundant oxygen vacancies 0.5 g/L	300-W Xe lamp (λ > 420 nm)	Tetracycline hydrochloride 10 mg/L	Approximately 87% optimum within 2 h	[51]
Vacancies	Oxygen vacancy-rich mesoporous ZrO2 1 g/L	300-W Xe lamp (λ > 420 nm)	Tetracycline hydrochloride 40 mg/L	Approximately 80% optimum within 150 h	[52]
Vacancies	BiOBr microspheres with oxygen vacancies 1 g/L	10-W LED lamp (0.4 mW‧cm−2)	Tetracycline (TC) 20 mg/L	Approximately 94% optimum within 90 min	[53]
Vacancies	ZnWO4-x nanorods with oxygen vacancy 0.2 g/L	Hg lamp 300-W UV or 300-W Xe lamp (UV–Vis-NIR)	Tetracycline 20 mg/L	Approximately 91% optimum within 90 min	[54]
Vacancies	Bi2MoO6 with oxygen vacancy 0.4 g/L	300-W Xe lamp (λ > 400 nm)	Ciprofloxacin 20 mg/L	Approximately 55% optimum within 120 min	[55]
Doping	Carbon-doped g-C3N4 0.5 g/L	Sunlight	Tetracycline 20 mg/L	Approximately 90% optimum within 90 min	[56]
Doping	P-O co-doped g-C3N4 1 g/L	350-W Xe lamp (λ>420 nm)	Enrofloxacin10 mg/L	Approximately 90% within 80 min	[57]
Doping	I and K co-doped g-C3N4 2 g/L	300-W Xe lamp (λ>420 nm)	Sulfamethoxazole 10 mg/L	Approximately 99% optimum within 45 min	[58]
Doping	Bi3+/g-C3N4 0.2 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 94% optimum within 30 min	[59]
Doping	Cr3+/SrTiO3 1 g/L	250-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 70% optimum within 60 min	[60]
Doping	Fe2+/Fe3+ immobilized on TiO2/fly-ash cenospheres 2 g/L	150-W tungsten halogen lamp (λ>420 nm)	Ciprofloxacin 10 mg/L	Approximately 80% optimum within 60 min	[61]
Doping	Ce3+ doped Bi2O3 0.6 g/L	300-W lamp (visible light)	Tetracycline 20 mg/L	Approximately 89% optimum within 180 min	[62]
Doping	Ti3+/N co-doped TiO2/diatomite granule 5 g/L	150-W Xenon lamp with a UV light filter	Tetracycline 10 mg/L	92% optimum within 150 min	[63]
Quantum dots	CQDs modified Bi2MoO6 1 g/L	300-W Xe lamp (λ>400 nm)	Ciprofloxacin 20 mg/L	88% optimum within 2h	[64]
Quantum dots	CQDs/BiOBr microspheres 1 g/L	Visible light irradiation	Ciprofloxacin 1 g/L	Approximately 65% optimum within 180 min	[65]
Quantum dots	TiO2/C-dots 0.1 g/L-0.5 g/L	Average intensity sunlight irradiation (72 klx)	Levofloxacin 10 mg/L	Approximately 99% optimum within 90 min	[66]
Quantum dots	ZnSe QDs/g-C3N4 Not provided	300-W Xe lamp (λ>400 nm)	Ceftriaxone sodium Concentration not provided	Approximately 80% optimum within 120 min	[47]
Quantum dots	Ag2O/TiO2 quantum dots 0.25 g/L	400-W halogen bulb (similar to sunlight)	Levofloxacin 10 mg/L	Approximately 81% optimum within 90 min	[67]
Quantum dots	CQDs/BiOI 0.5 g/L	300-W Xe lamp (λ>400 nm)	Tetracycline 20 mg/L	Approximately 70% optimum within 120 min	[68]
Quantum dots	CQDs/BiOBr 0.3 g/L	300-W Xe lamp (λ>400 nm)	Tetracycline 20 mg/L	Approximately 60% optimum within 120 min	[69]
Quantum dots	MoS2 modified Zn-AgIn5S8 quantum dots 0.1 g/L	250-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 74% optimum within 4 min	[70]
Quantum dots	3D ZnS-RGO nanospheres 1 g/L	300-W Hg vapor lamp	Norfloxacin 20 mg/L	92% optimum within 4h	[71]
Phase junction	Porous core–shell homojunction 0.5 g/L	UV lamp (254 nm, 90 W, Philips)	Tetracycline hydrochloride 50 mg/L	Approximately 81% within 300 min	[72]
Facet junction	CaCu3Ti4O12 Co-exposed (0 0 1) and (1 1 1) facets 0.4 g/L	300-W Xe lamp visible light	Tetracycline 0.24 g/L	Approximately 99% within 50 min	[73]
Facet junction	AgBr tetradecahedrons with co-exposed (1 0 0) and (1 1 1) facets 1 g/L	500-W halogen tungsten lamp (λ>420 nm)	Sulfadiazine 20 mg/L	Approximately 90% optimum within 90 min	[74]
Schottky heterojunction	Ag/Ag2MoO4 0.4 g/L	500-W Xe lamp (λ>420 nm)	Ciprofloxacin 20 mg/L	Approximately 99% optimum within 60 min	[75]
Schottky heterojunction	Ag/TiO2 (hollow nanosphere) 0.5 g/L	125-W high-pressure Hg lamp, (λ>435.8 nm)	Metronidazole 15 mg/L	Approximately 95% optimum within 120 min	[76]
Schottky heterojunction	Bi/BiOBr (nano-flowers) 0.8 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline hydrochloride Ciprofloxacin and Doxycycline 1 g/L	Approximately 100% optimum within 30 min	[77]
Schottky heterojunction	BiOCl-Ag (2D) 1.0 g/L	200-W Xe arc lamp (λ<420 nm)	Sulfonamides 10 mg/L	Approximately 80% optimum within 5h	[78]
Schottky heterojunction	Ag/Bi3O4Cl 0.5 g/L	250-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 94% optimum within 120 min	[79]
Schottky heterojunction	Ag/CCN 0.5 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	75% optimum within 15 min	[80]
Schottky heterojunction	Pt/g-C3N4 0.5 g/L	300-W Xe lamp (λ>400 nm)	Tetracycline hydrochloride 20 mg/L	Approximately 84% optimum within 40 min	[81]
Schottky heterojunction	Bi (Spheres)/g-C3N4 0.5 g/L	300-W Xe lamp (λ>420 nm)	Amoxicillin 100 mg/L	Approximately 5% optimum within 4 h	[82]
Schottky heterojunction	W-doped BaTiO3 0.2 g/L	Visible light irradiation	Tetracycline 20 mg/L	Approximately 80% optimum within 3h	[83]
Schottky heterojunction	Fe, Co, Ni, Fe-Co-, and Fe-Ni-doped ZnO 0.6 g/L	300-W Xe lamp (λ=365 nm)	Oxytetracycline 20 mg/L	Approximately 87% optimum within 2h	[84]
Schottky heterojunction	Pt/Bi/TiO2 1 g/L	300-W halogen-tungsten lamp (λ>420 nm)	Amoxicillin 10 mg/L	Approximately 87% optimum within 2h	[85]
Schottky heterojunction	Au/Pt/g-C3N4 1 g/L	500-W Xe lamp (λ>400 nm)	Tetracycline hydrochloride 20 g/L	Approximately 90% optimum within 3h	[86]
Schottky heterojunction	0D Bi nanodots/2D Bi3NbO7 nanosheets 0.5 g/L	300-W Xe lamp (λ>400 nm)	Ciprofloxacin 10 mg/L	Approximately 86% optimum within 120 min	[87]
Type Ⅱ heterojunction	AgI/BiVO4 0.3 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline (TC) 20 mg/L	Approximately 94% optimum within 1h	[88]
Type Ⅱ heterojunction	3D porous CdS/TiO2 1 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline and oxytetracycline (OTC) 40 mg/L	Approximately TC: 67% and OTC: 81% optimums within 50 min	[89]
Type Ⅱ heterojunction	MgFe2O4/MoS2	Radiation intensity: 47 mW/cm2	Tetracycline 10 mg/L	Approximately 92% optimum within 120 min	[90]
Type Ⅱ heterojunction	ZnWO4-CdS 1 g/L	300-W Xe lamp (λ>420 nm)	Ciprofloxacin 15 mg/L	Approximately 90% optimum within 1h	[91]
Type Ⅱ heterojunction	ZnO@ZnS nanorod 3×3 cm chip	500-W Xe lamp	Tetracycline 10 mg/mL	Approximately 80% within 140 min	[92]
Type Ⅱ heterojunction	MoS2/PbBiO2I 0.3 g/L	300-W Xe lamp (λ>400 nm)	Ciprofloxacin 10 mg/L	Approximately 80% optimum within 6h	[93]
Type Ⅱ heterojunction	Bi2SiO5/Bi12SiO20 1 g/L	100-W high-pressure Hg lamp	Tetracycline 10 mg/L	Approximately 79% optimum within 30 min	[94]
Type Ⅱ heterojunction	SrTiO3/Fe2O3 1 g/L	250-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 83% optimum within 140 min	[95]
Type Ⅱ heterojunction	p-C3N4/f-BiOBr 1 g/L	250-W Xe lamp (λ>400 nm)	Tetracycline 30 mg/L	Approximately 94% optimum within 300 min	[96]
Type Ⅱ heterojunction	Bi2O7Sn2-Bi7O9I3 1 g/L	Halogen lamp as simulated solar light	Tetracycline 35 mg/L	80% optimum within 90 min	[97]
Type Ⅱ heterojunction	NiFe2O4/Bi2O3 1 g/L	150-W xenon lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 91% optimum within 90 min	[98]
Type Ⅱ heterojunction	BiVO4/rGO 0.9 g/L	250-W Xe lamp (λ>420 nm)	Tetracycline 25 mg/L	99% optimum within 90 min	[99]
Type Ⅱ heterojunction	SrTiO3 nanocube coated CdS microsphere 1 g/L	250-W Xe lamp (λ>400 nm)	Ciprofloxacin 10 mg/L	Approximately 94% optimum within 120 min	[100]
Type Ⅱ heterojunction	g-C3N4/BiPO4 0.3 g/L	250-W high-pressure Hg lamp	Ciprofloxacin 10 mg/L	Approximately 97% optimum within 120 min	[101]
Type Ⅱ heterojunction	g-C3N4/Ag3PO4 0.5 g/L	300-W Xe lamp (λ>400 nm)	Ciprofloxacin Concentration not provided	Approximately 67% optimum within 15 min	[102]
Type Ⅱ heterojunction	In2S3/NaTaO3 0.5 g/L	300-W Xe lamp	Tetracycline hydrochloride 10 mg/L	Approximately 80% optimum within 180 min	[103]
Type Ⅱ heterojunction	Polyaniline/Bi4O5Br2 0.4 g/L	Visible light (λ＞420 nm)	Ciprofloxacin 10 mg/L Tetracycline 20 mg/L	CIP: 99% optimum within 50 min, TC: approximately 86% optimum within 240 min	[104]
Type Ⅱ heterojunction	CdS nanoparticles/porous carbon polyhedrons 1 g/L	300-W Xe lamp (λ>420 nm)	Cephalexin 20 mg/L	Approximately 90% optimum within 90 min	[105]
Type Ⅱ heterojunction	Microsphere-like In2S3/InVO4 0.5 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 71% optimum within 60 min.	[106]
Type Ⅱ heterojunction	g-C3N4/Bi4O5Br2 0.5 g/L	300-W Xe arc lamp	Ciprofloxacin 10 mg/L	Approximately 67% optimum within 150 min	[107]
Type Ⅱ heterojunction	Bi2WO6/g-C3N4 1 g/L	300-W Xe lamp UV light	Ceftriaxone sodium 10 mg/L	Approximately 94% optimum within 120 min	[108]
Type Ⅱ heterojunction	Flower-root shaped Bi2O3/Bi2MoO6 0.4 g/L	500-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 70% optimum within 190 min	[109]
Type Ⅱ heterojunction	Covalent triazine framework modified BiOBr nanoflake 0.2 g/L	500-W Xe lamp	Tetracycline 10 mg/L Ciprofloxacin 10 mg/L	Approximately TC: 90% and CIP: 60% optimums within 50 min	[110]
Type Ⅱ heterojunction	mpg-C3N4 and Bi2WO6 nest-like structure 0.5 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline hydrochloride 50 mg/L	Approximately 75% optimum within 120 min	[111]
Type Ⅱ heterojunction	TiO2 nanoparticle/SnNb2O6 nanosheet heterojunctions 1 g/L	500-W tungsten lamp	Tetracycline hydrochloride 35 mg/L	Approximately 76% optimum within 240 min	[112]
Type Ⅱ heterojunction	Bi4Ti3O12/BiOCl (2D/0D) composite 0.67 g/L	300-W Xe lamp	Tetracycline hydrochloride 20 mg/L	Approximately 84% optimum within 150 min	[113]
Type Ⅱ heterojunction	2D-2D g-C3N4/Bi4O5Br2 0.1 g/L	300-W Xe lamp (λ>400 nm)	Ciprofloxacin 10 mg/L	Approximately 50% optimum within 30 min	[114]
Type Ⅱ heterojunction	2D/2D Bi4Ti3O12/I-BiOCl 0.375 g/L	350-W Xe arc lamp (λ>420 nm)	Ciprofloxacin 10 mg/L Tetracycline hydrochloride 10 mg/L	Approximately 90% optimum within 120 min	[115]
Type Ⅱ heterojunction	Carbon-doped carbon nitride/Bi12O17Cl2 1 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 20 mg/L	94% optimum within 60 min	[116]
Type Ⅱ heterojunction	CuBi2O4/CuO 1 g/L	Visible light (λ＞400 nm)	Metronidazole 50 mg/L	36% optimum within 120 min	[46]
p-n heterojunction	p-n type BiOCl/titanium phosphate nanoplates 0.4 g/L	300-W Xe lamp	Ciprofloxacin 5 mg/L	Approximately 100% within 5 min	[117]
p-n heterojunction	p-n type CoO/g-C3N4 0.5 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 90% within 60 min	[118]
p-n heterojunction	p-n type Cu2O/SrTiO3 1 g/L	150-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 79% within 100 min	[119]
p-n heterojunction	p-n type flower-like BiOCl/BiOCOOH p-n 1 g/L	300-W Xe lamp simulated sunlight	Tetracycline 20 mg/L	Approximately 80% within 60 min	[120]
p-n heterojunction	p-n type Ag2O/g-C3N4 1 g/L	500-W Xe lamp (λ>400 nm)	Tetracycline hydrochloride 20 mg/L	Approximately 94% within 3h	[121]
p-n heterojunction	p-n type Co3O4-C3N4 0.5 g/L	Sunlight	Tetracycline 48 mg/L	Approximately 97% within 180 min	[122]
p-n heterojunction	p-n type 3D flower-like BiOBr/Bi2SiO5 1 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 20 mg/L	91% within 120 min	[123]
p-n heterojunction	n-p type SnO2 nanoparticles/BiOI 1 g/L	300-W Xe lamp (λ>420 nm)	Oxytetracycline hydrochloride 10 mg/L	Approximately 94% optimum within 90 min	[124]
p-n heterojunction	p-n type N-graphene QDs-BiOI/MnNb2O6 0.5 g/L	250-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 87% optimum within 60 min	[125]
p-n heterojunction	p-n type Fe3O4 quantum dots modified BiOCl/BiVO4 0.5 g/L	300-W Xe lamp (λ>420 nm)	Sulfamethoxazole (SMX, 5 mg/L), TC (20 mg/L), norfloxacin (NOR, 10 mg/L), and CIP (10 mg/L)	SMX: 91% within 90 min, TC: 87% within 30 min, NOR: 89% within 60 min, CIP: 87% within 90 min	[126]
Double heterojunction	CDs/MoS2/TiO2 nanobelt 0.5 g/L	250-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 82% within 3h	[127]
Double heterojunction	(g-C3N4)-ZnO/halloysite nanotubes (HNTs) 1 g/L	350-W Xe arc lamp	Tetracycline 20 mg/L	Approximately 87% optimum within 60 min	[128]
Double heterojunction	Ultrathin g-C3N4 nanosheets coupled with amorphous Cu doped FeOOH nanoclusters as 2D/0D heterogeneous catalysts 0.2 g/L	500-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 90% within 40 min	[129]
Double heterojunction	2D/2D/2D CoAl-LDH/g-C3N4/RGO ternary heterojunction 0.25 g/L	300-W halogen lamp visible-light irradiation	Tetracycline 20 mg/L	Approximately 100% optimum within 60 min	[130]
Double heterojunction	Ag-AgVO3/g-C3N4 0.2 g/L	300-W Xe lamp (λ>410 nm)	Tetracycline 30 mg/L	Approximately 84% optimum within 120 min	[131]
Double heterojunction	ZnFe2O4/Ag/Ag3VO4 1 g/L	Visible-light irradiation	Tetracycline 10 mg/L	Approximately 60% within 10 min	[132]
Double heterojunction	NiS and MoS2 nanosheet co-modified g-C3N4 ternary heterostructure 1 g/L	250-W metal halide lamp (λ>400 nm)	Tetracycline 10 mg/L Ciprofloxacin 10 mg/L	Approximately CIP: 71% and TC: 96% optimums, within 120 min	[133]
Double heterojunction	AgCl/Ag3PO4/ g-C3N4 1 mg/L	Visible-light irradiation (λ>400 nm)	Sulfamethoxazole 50 mg/L	Approximately 100% optimum within 90 min	[134]
Double heterojunction	3D Ag3PO4/TiO2@MoS2 0.5 g/L	800-W Xe arc lamp	OTC 5 mg/L ENR 5 mg/L	Approximately OTC: 75%, ENR: 92% within 10 min	[135]
Double heterojunction	Bi2O3/BiOCl supported on graphene sand (BO/BOC/GSC) composite (BO/BOC/CT) 0.5 g/L	Solar light intensity (35 × 103 ± 1000 lx)	Oxytetracycline 46 mg/L Ampicillin 37 mg/L	BO/BOC/GSC: approximately 90% for AMP and OTC BO/BOC/CT: approximately 80% for AMP and OTC	[136]
Double heterojunction	Core-shell structured Fe3O4@SiO2@CdS 0.2 g/L	1000-W tungsten-halide lamp (Philips) (λ>420 nm)	Tetracycline 100 mg/L	Approximately 80% optimum within 21 min	[137]
Double heterojunction	RGO-CdS/ZnS 0.5 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 15 mg/L	Approximately 90% optimum in 60 min	[138]
Double heterojunction	TiO2/Bi2WO6/carbon fibers 3 g/L	300-W Xe lamp (λ>400 nm)	Tetracycline hydrochloride 10 mg/L	Approximately 95% optimum within 60 min	[139]
Z-scheme heterojunction	Z-scheme beta-Bi2O3@g-C3N4 core/shell nanocomposite 0.5 g/L	250-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 80% optimum within 50 min	[140]
Z-scheme heterojunction	Z-scheme WO3-g-C3N4 0.5 g/L	300-W Xe arc lamp (1.5 AM solar simulator)	Sulfamethoxazole 10 mg/L	Approximately 92% optimum within 4h	[141]
Z-scheme heterojunction	Z-scheme AgI nanoparticle-sensitized Bi5O7I microspheres 0.5 g/L	300-W Xe lamp (λ>400 nm)	TC (20 mg/L), DTC (10 mg/L), OTC (10 mg/L), or CIP (10 mg/L)	Approximately TC: 95%, DTC: 90%, OTC: 80% and CIP: 90% optimums within 40 min	[142]
Z-scheme heterojunction	Z-scheme CdTe/TiO2 0.8 g/L	400-W halogen lamp (λ>400 nm)	Tetracycline hydrochloride 20 mg/L	Approximately 78% optimum within 30 min	[143]
Z-scheme heterojunction	Type II AgI/CuBi2O4 Z-scheme AgBr/CuBi2O4 0.5 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately Type II catalyst: 80% and Z-scheme catalyst: 90% optimums within 30 min	[144]
Z-scheme heterojunction	Z-scheme mesoporous Sn3O4 nanoclusters/g-C3N4 nanosheets 0.5 g/L	500-W Xe lamp (λ>420 nm)	Tetracycline hydrochloride 10 mg/L	Approximately 72% optimum within 120 min	[145]
Z-scheme heterojunction	Z-scheme Bi3TaO7 QDs/g-C3N4 nanosheets (NSs) 0.5 g/L	LED lamp (λ=420 nm, 86 W)	CIP and CPX 10 mg/L	Approximately CIP: 91%, CPX: 77% CPX within 120 min	[146]
Z-scheme heterojunction	Z-scheme WO3 nanosheet/K+Ca2Nb3O10− ultrathin nanosheet 1 g/L	250-W xenon lamp as simulated sunlight (no filters).	Tetracycline hydrochloride 35 mg/L	Approximately 86% optimum within 120 min	[147]
Z-scheme heterojunction	Z-scheme AgI/BiOBr 0.5 g/L	300-W Xe lamp (λ>420 nm)	Ciprofloxacin 10 mg/L	Approximately 91% optimum within 1h	[148]
Z-scheme heterojunction	Z-scheme Ag3PO4/g-C3N4 0.05 g/L	300-W Xe lamp (λ>400 nm)	Sulfamethoxazole 1 mg/L	Approximately 99% optimum within 90 min	[149]
Z-scheme heterojunction	Z-scheme CdS-Au-BiVO4 (0 1 0) 0.5 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	91% optimum within 90 min	[150]
Z-scheme heterojunction	Z-scheme BiVO4/Ag/Cu2O 0.4 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 20 mg/L	Approximately 91% optimum within 90 min	[151]
Z-scheme heterojunction	Z-scheme ZnFe2O4/Ag/PEDOT 0.2 g/L	250-W xenon lamp (1.8×105 lx)	Tetracycline 20 mg/L	Approximately 72% optimum within 120 min	[152]
Z-scheme heterojunction	Organic-inorganic Z-scheme PANI/Ag/Ag2MoO4 ~ 0.2 g/L	40-W UV tube (Phillips)	Ciprofloxacin 3 mg/L	Approximately 100% optimum within 40 min	[153]
Z-scheme heterojunction	Z-scheme (0 0 1) BiOCl-Au-CdS 1 g/L	300-W Xe lamp (AM 1.5)	Sulfadiazine 20 mg/L	Approximately 91% optimum within 4h	[154]
Z-scheme heterojunction	Z-scheme iodine vacancy-rich BiOI/Ag@AgI 0.3 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 20 mg/L	Approximately 86% optimum within 60 min	[155]
Z-scheme heterojunction	Z-scheme Ag2CO3/Ag/WO3 0.5 g/L	300-W Xe lamp (λ>420 nm)	CIP and TC 10 mg/L	Approximately CIP: 84% and TC: 81% optimums within 90 min	[156]
Z-scheme heterojunction	Z-scheme AgI/Ag/Bi3TaO7 0.5 g/L	300-W Xe lamp (visible light)	Sulfamethoxazole 5 mg/L	Approximately 98% optimum within 100 min	[157]
Z-scheme heterojunction	Z-scheme MIL-53(Fe)/Ag/g-C3N4 80 mg/L	Visible light	Clioquinol 10 mg/L	95% optimum within 100 min	[158]
Z-scheme heterojunction	Z-scheme CeVO4/3D RGO aerogel/BiVO4 0.5 g/L	500-W Xe lamp	Tetracycline 20 mg/L	Approximately 100% optimum within 60 min	[159]
Z-scheme heterojunction	TCPP/rGO/Bi2WO6 0.3 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 15 mg/L	Approximately 84% optimum within 60 min	[160]
Z-scheme heterojunction	Z-scheme Ag3PO4/Bi2S3/Bi2O3 1 g/L	300-W Xe lamp (λ>420 nm)	Sulfamethazine (SAZ) and cloxacillin (CLX) 10 mg/L	Approximately SAZ: 99% and CLX: 90% optimums within 90 min	[161]
Z-scheme heterojunction	RGO-Ag2O/TiO2 0.4 g/L	350-W Hg lamp (λless than356 nm), 300-W Xe arc lamp (visible light), 300-W infrared lamp, and 156-W APOLLO solar simulator	Tetracycline 10 mg/L	100% and approximately 100% optimums within 60 min, approximately 90% optimums within 120 min	[162]
Z-scheme heterojunction	Z-scheme g-C3N4/Ag2CO3/graphene oxide 0.8 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 20 mg/L	Approximately 82% optimum within 60 min	[163]
Z-scheme heterojunction	Z-scheme nitrogen-doped graphene QDs-BiVO4/g-C3N4 0.5 g/L	250-W Xe lamp (λ>420 nm)	Tetracycline 10 mg/L	Approximately 91% optimum within 30 min	[164]
Z-scheme heterojunction	Z-scheme graphitic carbon nitride (CN) and reduced graphene oxide (rGO) with AP 1 g/L	Both intense sunlight and weak indoor light irradiation	Norfloxacin 10 mg/L	Approximately 100% optimum within 30 min and 85% optimum within 2h	[165]
Z-scheme heterojunction	Z-scheme WO3/Fe3O4/g-C3N4 1 g/L	300-W Xe lamp (λ>400 nm)	Tetracycline 20 mg/L	89% optimum within 120 min	[166]
Z-scheme heterojunction	Z-scheme nitrogen-doped hollow mesoporous carbon spheres (N-HMCs) modified g-C3N4/Bi2O3 1 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline hydrochloride (TCH) and ciprofloxacin hydrochloride (CFH) 10 mg/L	Approximately 90% and 80% optimum within 60 min	[167]
Exposing active facets	Bi2O2(OH)(NO3) nanosheets with (0 0 1) active exposing facets 1 g/L	UV light irradiation (300-W Hg lamp)	Tetracycline hydrochloride 10 mg/L	Approximately 98% optimum within 25 min	[168]
Exposing active facets	Ultrathin Bi2O2(OH)xCl2-x solid solution with exposed (0 0 1) facets 0.5 g/L	Visible light (420-nm LED)	Ciprofloxacin 20 mg/L	Approximately 90% optimum within 150 min	[169]
Exposing active facets	Nanosheet BiVO4 with oxygen vacancies and exposed (0 0 1) facets 1 g/L	500-W Xe lamp without optical filters to simulate the sunlight	Oxytetracycline 20 mg/L	Approximately 96% optimum within 2h	[170]
Exposing active facets	Various well-defined Bi2WO6 crystals ~ 0.67 g/L	300-W Xe lamp (λ>420 nm)	Ciprofloxacin 10 mg/L	Approximately 70% optimum within 5h	[171]
Exposing active facets	Doped BiOCl nanoplates ~ 0.67 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline hydrochloride 30 mg/L	Approximately 90% optimum within 100 min	[172]
Exposing active facets	(0 0 1) Ag@NC-TiO2 square nanosheets 1 g/L	350-W Xe arc lamp (λ>420 nm)	Ciprofloxacin 10 mg/L	Approximately 97% optimum within 150 min	[173]
Exposing active facets	TiO2@g-C3N4 core–shell quantum 1 g/L	Xe lamp irradiation.	Tetracycline 20 mg/L	Approximately 100% optimum within less than 10 min	[174]
Porous materials	Intercalate structure g-C3N4@ATP 1 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline 20 mg/L	Approximately 90% optimum within 2h	[175]
Porous materials	3D hierarchical mesoporous BiOI Optimum 1.5 or 0.68 g/L	1000-W tungsten halogen lamp (λ>420 nm)	Tetracycline hydrochloride Optimum 2.1 or 2 mg/L	Approximately 100% optimum within 37.5 or 101.5 min	[176]
Porous materials	BiOI hollow microspheres 1 g/L	300-W Xe lamp (λ>400 nm)	Tetracycline 20 mg/L	80% optimum within 120 min	[177]
Porous materials	Ultra-thin Bi2MoO6 nanosheets 0.5 g/L-1.5 g/L	Sunlight	Ofloxacin 10 mg/L	Approximately 71% optimum within 90 min	[178]
Tailoring morphology	Rod-like SrV2O6 2–9 g/L	500-W tungsten lamp (λ>400 nm)	Metronidazole 20 mg/L	98% optimum within 60 min	[179]
Tailoring morphology	TiO2 nanobelts	Simulative solar light	Amikacin 5 mg/L	Approximately 70% optimum within 150 min	[180]
Tailoring morphology	ZnO Nanotubes 0.035 g/L	300-W Xe lamp (AM1.5 filter (1000 Wm−2))	Ciprofloxacin 6.6 mg/L	Approximately 12% optimum within 2h	[181]
Tailoring morphology	Bi5FeTi3O15 0.4 g/L	300-W Xe lamp (λ>420 nm)	Tetracycline hydrochloride 4.8 mg/L	Approximately 99% optimum within 1h	[182]
Tailoring morphology	Navel-like Bi2WO6 hierarchical microspheres	UV light irradiation (λ=365 nm)	Norfloxacin (Concentration not provided)	Approximately 67% optimum within 8h	[183]
Tailoring morphology	spearhead-like g-C3N4 1 g/L	Xe lamp (35 W, 6000 K)	Tetracycline 20 mg/L	Approximately 70% optimum within 180 min	[184]
3D aerogel	g-C3N4@CA/B-PET 5 g/L	Artificial solar light (125 mW/cm2)	Sulfaquinoxaline sodium 10 mg/L	Approximately 100% optimum within 60 min	[185]
3D aerogel	BiVO4/3D RGO aerogel/CeVO4 0.5 g/L	500-W Xe lamp (visible light)	Tetracycline 20 mg/L	Approximately 90% optimum within 120 min	[159]
3D aerogel	3D MoS2 nanosheets/graphene aerogel 0.8 mg/L	300-W Xe lamp (λ>420 nm)	Tetracycline hydrochloride 60 mg/L	Approximately 10% optimum within 75 min	[186]

3.1. Strategies to increase light harvesting

3.1.1. Defect engineering to increase light harvesting

Defects play two distinct roles in the photocatalytic process: (1) they act as trapping sites for electrons or holes. Depending on the nature of the defects (deep trap or shallow trap), it may improve the diffusion length of charge carriers (shallow trap) or promote the charge recombination (deep trap) thus reducing the photocatalytic efficiency; (2) they may extend the light absorption by modulating the electronic structure of the photocatalyst (e.g. creation of midgap impurities level), thereby influencing the photocatalytic activity. However, elaborative designs such as vacancy construction and elemental doping are necessary to take full advantage of the improvements and avoid or minimize the disadvantages [187], [188], [189], [190].

3.1.1.1. Vacancies for the improvement of light absorption

Vacancies are types of point defects that not only manipulate the composition of catalysts without introducing impurities but also allow for the customization of suitable band structures by modulating their types and concentrations. The concentration of the vacancies should be controlled within a delicately balanced range. Excessive vacancies would provide too many recombination centers for photoinduced electrons and holes, whereas insufficient vacancies would fail to achieve the desired photocatalytic performance. Vacancies may be categorized into three types: 1) anion vacancies (such as sulfur [191], [192], [193], [194], [195], halogen [196], [197], [198], [199], nitrogen [200], [201], [202] and oxygen vacancies (OV) [203], [204], [205], [206], [207]), 2) cationic vacancies (such as bismuth [208], [209], [210], carbon [211] and titanium vacancies [212], [213], [214], [215]), and 3) mixed vacancies, which combine anion and cation vacancies [216], [217]. For example, mesoporous ZrO₂ containing abundant oxygen vacancies was successfully synthesized by Zhang et al. [52] via a template-less solvothermal approach, following calcination. The absorption edge of the obtained sample was extended into the visible range, and the modified surface characteristics promoted adsorption ability towards organic compounds, leading to an enhancement of the visible-light-driven photodegradation of organic contaminants. Similarly, brown ZnWO_4-x nanorods containing oxygen vacancies exhibited a broadened photo-absorption capacity from the ultraviolet (UV) to the near-infrared (NIR) region, whereas conventional white ZnWO₄ nanorods without OV exhibited a UV absorption onset at 365 nm [54].

Fig. 3 shows the existence of an OV-mid-gap state and three possibilities for electron excitation: (i) UV-driven direct excitation from the VB towards the CB, (ii) visible-light-induced jump from VB to OV defect states, and (iii) near-infrared-light-driven transition from the OV-derived defect state to the CB [54]. The induced defect, or the inter-band energy level, enables electron jumps with less energetic illumination, thereby improving the light harvest capacity of ZnWO_4-x nanorods with abundant OV [54]. Moreover, Liu et al. [218] reported that anion vacancy could be produced in a TiO₂ lattice when Ti³⁺ was present. Excess energy levels below the CB would be induced in the presence of these anion vacancies [219], [220], [221], [222]. Overlapping of these levels with CB occurs by increasing the concentration of these vacancies, resulting in a decreased bandgap of TiO₂. Therefore, the visible-light-driven transition can take place from intrinsic VB to Ti³⁺ states. This behavior can extend the optical response to an onset wavelength of longer than 400 nm [223].

Fig. 3.

Schematic of the OV-induced photocatalytic process on ZnWO_4-x[54]

3.1.1.2. Heteroatoms or ion doping for light response improvement

The introduction of heteroatoms or ions into the photocatalyst, in the form of dopant, is an effective light response management method. Non-metal dopants (e.g., C [224], [225], B [226], [227], [228], [229], P [57], [230], S [231], [232], N [233], [234], [235], [236], and halogen [237], [238]) and metallic dopants (e.g. Fe³⁺ [239], [240], Ce³⁺ [62], Cr³⁺ [241], W⁶⁺ [242], Ti⁴⁺ [243], Cu²⁺ [241], Co²⁺ [244], [245] and Er³⁺ [246]) were most widely studied in semiconductor photocatalysts, especially metal oxides. Generally, a method can either introduce an impurity energy level lower than the CB minimum or higher than the VB maximum to narrow down the bandgap, enhancing light absorption.

Recently, N-doped TiO₂ photocatalysts with tunable doping contents were prepared by calcining sol–gel TiO₂ powder in the presence of NH₃ flow at 450–800 °C [247]. The light absorption capacity of samples prepared at a temperature region of 450–600 °C increased gradually with temperature, and their colors varied from pale yellow to emerald green as a result of the production of N 2p (localized states) beyond the VB of TiO₂ and the generation of OV (Fig. 4 (a)). When NH₃ treatment reached T > 600 °C, anatase TiO₂ was transformed into rutile, leading to a remarkable reduction in its Brunauer-Emmett-Teller (BET) area and the production of a TiN layer on the TiO₂ surface. The materials treated at 700 and 800 °C exhibited a conspicuous response to a wide range of spectra spanning from the visible to the NIR range, which originated from the production of OV and Ti³⁺ defects. Similarly, Wang et al. [248] synthesized lantern-shaped g-C₃N₄ with a Eu³⁺ dopant and found that a doping level was developed below the CB upon the introduction of Eu³⁺, resulting in a narrower bandgap and lowering the electron position compared to that of raw CN. This improved both the optical light response and photocatalytic performance.

Fig. 4.

(a) UV–vis diffuse reflectance spectra of a) 1) raw TiO₂ (5 0 0), and N-TiO₂ (T) samples prepared at 2) 450, 3) 500, 4) 550, 5) 600, 6) 700 and 7) 800 °C [247], (b) diffuse reflection spectra (DRS) curves of undoped and doped TDHG [63], Inset: Photographs of TDHG, N-TDHG, b -TDHG and b/N-TDHG photocatalysts.

In some cases, co-doping (e.g., the addition of two non-metals [57], [231], two metal ions [246], or a metal ion paired with a non-metal ion [63]) has been used to avoid imbalances of charges during aliovalent doping [187]. For example, b/N-TDHG (black Ti³⁺/N co-doped TiO₂/diatomite hybrid granule) could be obtained from a sol–gel process [63]. The light absorption of b/N-TDHG (Fig. 4(b)) exhibited a more pronounced redshift than other TDHG counterparts. Particularly, it had the strongest visible-light response among all obtained photocatalysts, which resulted from the synergistic action of co-doped N and Ti³⁺.

3.1.2. Addition of light sensitizers to enhance light absorption

The addition of photosensitizers is another simple method of enhancing the photoactivity of wide bandgap semiconductors in visible light. In recent years, a variety of QDs (e.g., ZnSe QDs [47], carbon QDs (CQDs) [64], [65], [66], [68], [69], [249], [250], [251], [252], [253], graphene [254], CdS [255], [256], Fe₃O₄ [257], g-C₃N₄ [258], CsPbBr₃ [259] and MoS₂ [260]) have been combined into semiconductors to enhance light absorption. Among them, CQDs have become the most widely used QDs to upgrade wide-bandgap semiconductors owing to their low cost, non-toxicity, and high stability [250], [261]. For example, up-converting CQDs were immobilized on the surface of La₂Ti₂O₇ (LTO) nanosheets to promote visible-light-induced photocatalysis [252]. The pristine (i.e., unmodified) LTO nanosheets had no optical response from 400 to 800 nm (Fig. 5 (a)). In contrast, all CQD/La₂Ti₂O₇ (C-LTO) composites exhibited distinct responses in the above range, and the response could be modulated accordingly by changing the amount of CQD.

Fig. 5.

(a) Light absorption curves [252] and (b) photocatalytic mechanisms of CQD/TNT photocatalyst [253]

Zhao et al. [253] successfully synthesized TiO₂ nanotube (CQD/TNTs) composites with an outstanding photocatalytic performance. These CQD co-catalysts had remarkable up-converting photoluminescence (PL) features. Longwave infrared rays (LWIR, >600 nm) absorption could be converted into optical light of shorter than 600 nm, enabling the production of IR-induced electrons and holes on TNTs (Fig. 5(b)). Additionally, the introduction of CQDs can accelerate the harvesting of photoexcited electrons and prolong the lifetime of photoinduced carriers.

3.2. Interfacial engineering strategies for carrier migration improvement

The construction of a junction interface is a typical approach to facilitate photocatalysis by modulating the separation/migration of the photoinduced electrons and holes at the interface. Therefore, in-depth studies have been conducted to design photocatalyst junction interfaces to monitor carrier migration capabilities. According to their phase composition, junction interfaces can be divided into homogenous junction interfaces and heterojunction interfaces.

3.2.1. Homogenous junction interfaces

Homogenous junction interfaces are built by identical compounds in the absence of additional components and have been the focus of significant interest. The construction of homogenous junction interfaces is primarily based on the established phase junction or facet junction, creating an effective migration pathway for the photoinduced charges.

3.2.2. Phase junction

Most crystals (e.g., TiO₂ [262], [263], [264], [265], [266], [267], [268], [269], [270], CdS [271], Bi₂O₃ [272], MoS₂ [273] and ZnIn₂S₄ [274]) have many different natural or artificial phases. The construction of the phase junction is a significant approach to improve the photocatalytic performance through accelerating the migration of electrons and holes. Many studies have been conducted on the design of photocatalyst phase junctions to boost carrier migration characteristics and capabilities [275], [276]. Jia et al. [277] studied the theoretical basis of the phase changes and interfacial characteristics of Cu₂ZnSnS₄ with a hetero-phase junction. A type-II band structure could be formed at the hetero-phase junction of (1 0 1), (1 1 0), and (1 0 0) facets, thereby facilitating the photoelectric capacity. However, type-I heterojunctions may be aligned at the phase junction of (1 1 2) and (0 0 1) facets, hindering the generation of charges and improving optoelectronic properties owing to a high potential barrier. Similarly, according to the band engineering design theory, a bonding-region-width-controlled phase junction surrounded by cubic CdS and a hexagonal counterpart was designed by Ai et al. [271]. The electron–hole separation was optimized at an appropriate bonding region width (0.76 nm), which exhibited the highest photocatalytic efficiency (45% quantum efficiency) among all the examined samples, in addition to good photocorrosion resistance.

Artificial phases caused by vacancies or self-doping have been used to construct a phase junction to control the separation of electrons and holes. Recently, Cao et al. [201] constructed a g-C₃N₄ p-n homojunction through introducing nitrogen vacancies (NVs). The differences in Mott-Schottky plots between bulk g-C₃N₄ (CN) and PN-2 indicated that the conductivity characteristics of the homojunction were consistent with a p-n model (Fig. 6 (a) and (b)). Additionally, the PN-2 homojunction exhibited a distinct improvement in its photocatalytic performance to treat Rhodamine B (RhB), compared to CN, which was attributed to the efficient separation of the photoexcited charges at the PN-2 homojunction (Fig. 6 (c) and (d)). Chade Lv et al. [278] synthesized Bi⁵⁺-BVO (Bi⁵⁺-self-doped Bi₄V₂O₁₁) nanotubes and formed p-n homojunctions by implementing the oxygen vacancy strategy. Density functional theory (DFT) simulations and laboratory findings reveal that Bi⁵⁺ auto-doping reduces the bandgap of Bi₄V₂O₁₁ (BVO), thereby enhancing light capture. Furthermore, Bi⁵⁺ auto-doping confers both n- and p-type semiconductor characteristics to BVO, which enables the construction of p-n homojunctions to delay the quick annihilation of electrons and holes.

Fig. 6.

Mott-Schottky curves on (a) CN and (b) PN-2; (c and d) schematic of carrier migration at the p-n homojunction [201]

3.2.3. Facet engineering

Previous studies have reported that, unlike isolated surfaces, the coexistence of multiple crystal planes in single particles could be recognized as interactive surfaces with synergistic effects [279], [280]. This means that redox reactions occur locally at the separate planes, resulting in spatial charge separation in semiconductors. For example, multi-morphologic silver bromide (AgBr) crystals containing distinct exposed planes were prepared by synchronously injecting silver nitrate and potassium bromide precursors [74]. The shape and exposed planes of the target sample could be easily modulated by varying the density of Br⁻ ions, which can significantly reduce the surface barrier of (1 0 0) and (1 1 1) planes and affect their growth rates. C-AgBr ((1 0 0) facet exposed cubes), T-AgBr ((1 0 0) and (1 1 1) facet exposed tetradecahedrons), and O-AgBr ((1 1 1) facet exposed octahedrons) could be obtained at specific Br⁻ concentrations of 10⁰, 10^0.5, and 10¹ mM. The obtained T-AgBr exhibited improved photocatalytic performance for the degradation of methyl orange (MO) and sulfadiazine (SD) owing to the formed facet heterojunction structures between the (1 1 1) and (1 0 0) facets. When irradiated with optical light, the electrons and holes are separately transported to the (1 0 0) and (1 1 1) facets. Therefore, electrons and holes could not merely be efficiently separated, but back-reaction was effectively prevented owing to the isolation of active redox sites.

3.2.4. Heterojunction interfaces

The formation of homojunctions is considerably challenging owing to its complexity and high cost. Therefore, the creation of heterojunction interfaces has been proposed as an alternative to improve charge separation efficiency and has been widely investigated in the past decades [281]. Typically, heterojunctions are divided into five categories: (1) Schottky heterojunctions, (2) type I heterojunctions, (3) type II heterojunctions, (4) p-n heterojunctions, and (5) Z-scheme heterojunctions.

3.2.5. Schottky heterojunctions

At the semiconductor–metal interface, photoinduced electrons may typically flow from the former to the latter to match their Fermi energies and form a Schottky barrier. The formed Schottky junction can efficiently capture the electron, which promotes electron-hole separation. To further comprehend the mechanisms by which the Schottky junction improves photocatalysis, He et al. [282] prepared a series of Schottky-junction Ag-loaded carbon nitride fibers (Ag/CNFs) through loading NaBH₄-reduced Ag on preformed CNFs. The Schottky-type Ag/CNFs exhibited outstanding photodegradation performance for tetracycline (TC) under visible irradiation. The enhanced photodegradation activity resulted from the synergistic effect between surface plasmon resonance (SPR) on silver nanoparticles (Ag NPs) and the efficient isolation of the photoinduced carriers at the constructed Schottky heterojunction. Moreover, on account of the SPR absorption of specific metals, the light response could even be extended to the entire sunlight range. Therefore, in addition to promoting charge separation, semiconductors and plasmonic metals could also jointly improve light harvesting.

Jiang and co-workers synthesized a novel Ag/Bi₃O₄Cl photocatalyst with different amounts of Ag via a facile photo-deposition process [79]. The TC degradation on 1.0 wt% Ag/Bi₃O₄Cl was drastically more pronounced, compared to pristine Bi₃O₄Cl, and 94.2% TC could be degraded in 120 min (Fig. 7 (a)). The improved photocatalytic properties resulted from the synergism of the photoinduced electrons on Bi₃O₄Cl and SPR from Ag NPs, which improved visible-light harvesting and facilitated the isolation of photoinduced electrons and holes (Fig. 7(b)).

Fig. 7.

(a) Photodegradation of TC by the as-synthesized plasmonic Ag/Bi₃O₄Cl under visible irradiation, (b) possible photocatalytic mechanisms on the Ag/Bi₃O₄Cl samples [79], (c) photocatalytic kinetics of prepared g-C₃N₄, Pt/g-C₃N₄, Au/g-C₃N₄, and Au/Pt/g-C₃N₄ nanocomposites [86] (d) schematic of the g-C₃N₄ photoinduced charge transport [284]

In other studies, bimetallic-supported semiconductor catalysts (Pt/Au/TiO₂ [283], [284], Au/Ag/TiO₂ [285], Pd-Cu/TiO₂ [286] and Ag-Cu/TiO₂ [287]) exhibited super activity and high selectivity, which cannot be found in single metals owing to the synergistic effect of bimetals. For example, Xue et al. [86] prepared plasmonic Au/Pt/g-C₃N₄ via a simple calcination-photo-deposition process. The prepared heterostructure photocatalyst exhibited an optimized photodegradation performance for antibiotic tetracycline hydrochloride (TC-HCl) treatment, and the visible-light-driven degradation was 3.4 times higher than that of pristine g-C₃N₄ (Fig. 7(c)). This enhanced photodegradation performance was attributed to the synergism between the SPR absorption on Au and the electron-trap effect of Pt NPs, which facilitated the light-harvesting capability and isolation of photoinduced charges on g-C₃N₄, thereby jointly boosting the photocatalytic properties. Similarly, TiO₂ nanotube arrays (Au-Pt/TNTAs) with small quantities of Au-Pt were successfully prepared. Notably, these structures exhibited remarkably enhanced visible-light harvesting and carrier separation capacity. These enhancements in photodegradation activity derived from the synergism from interfacial Schottky junctions and the synergistic effect between ternary Au, Pt, and TNTAs (Fig. 7(d)) [284].

3.2.6. Type I or II heterojunctions

Compared to single semiconductors, the formation of type I semiconductor heterojunctions can promote photocatalytic efficiency. However, electrons and holes are enriched on the same semiconductor, which cannot effectively restrain the recombination of photoinduced electrons and holes. This makes type I semiconductor heterojunctions questionable prospects for photocatalytic property improvement [288], [289]. To date, large numbers of traditional type II heterojunctions have been intensively investigated, some of which have rendered drastically optimized photocatalysis performances. For example, compared to pristine MgFe₂O₄, newly synthesized MgFe₂O₄/MoS₂ heterojunctions can accelerate electron–hole dissociation, albeit without extending the light-harvesting scope [90]. The as-synthesized MgFe₂O₄/MoS₂ sample delivered a prominent photoelectrochemical performance in TC treatment and H₂ evolution owing to its unique bandgap structure. In another study, a novel 2D-2D thin-layered g-C₃N₄/Bi₄O₅Br₂ type II heterojunction was synthesized by Ji’s group [114] through a simple ionic-liquid-assisted solvothermal process (Fig. 8 ). This 2D g-C₃N₄ component had a layered microstructure, providing enough potential to successfully build intense interface interactions with exfoliated Bi₄O₅Br₂ nanosheets. Implementing the 2D layered g-C₃N₄ component provided several clear advantages, including an enlarged specific surface area (SBET), broadened optical light harvesting range, optimized electron–hole separation, and accelerated transport of the interfacial electrons and holes, all of which enhanced the photocatalytic performance of the Bi₄O₅Br₂ system within a given light range.

Fig. 8.

Preparation process for g-C₃N₄/Bi₄O₅Br₂ nanocomposites [114]

3.2.7. p-n heterojunction

Schottky and type II heterojunctions may successfully split electron–hole pairs. However, it is still necessary to restrain the superfast electron–hole annihilation at the interface. Therefore, p-n heterojunctions have been implemented and widely used to manage photocatalytic properties. For instance, Kang et al. [118] constructed a high-performance CoO/g-C₃N₄ p-n junction through a simple solvothermal process. The prepared p-n junction exhibited excellent visible-light-induced photocatalytic properties and consistency for TC treatment. These excellent photocatalytic properties were attributed to the production of an interfacial electric field around the junction, which effectively prevented the annihilation of photo-induced charges. Furthermore, a SnO₂/BiOI n-p heterojunction was successfully achieved by Wen et al. [124] by depositing SnO₂ NPs onto the outer surface of BiOI nanosheets. The obtained n-p heterojunction exhibited a strong visible-light-driven photocatalysis capacity and durability for the degradation of MO and oxytetracycline HCl. This excellent photocatalytic activity resulted from the constructed p-n heterojunction, which significantly limited the annihilation of electrons and holes.

Similarly, a novel three-dimensional (3D) layered BiOBr/Bi₂SiO₅ p-n heterostructure was successfully synthesized to efficiently degrade TC [123]. The prepared BiOBr/Bi₂SiO₅ p-n heterostructure exhibited an optimized visible-light-driven photocatalytic degradation of TC, which was 3.6-fold higher than that of BiOBr. Moreover, this high TC photodegradation performance remained constant for 5 cycles. This optimized photodegradation property of the BiOBr/Bi₂SiO₅ heterostructure was ascribed to the following factors: (1) the introduced BiOBr facilitated light harvesting within a broad optical range, and (2) the formed p-n heterojunction restrained the annihilation of photoinduced electrons and holes. These features prominently promoted the generation of ·OH, ·O₂ ^–, and H⁺, leading to a promoted photocatalytic activity.

3.2.8. Z-scheme heterojunction

The development of Z-scheme photocatalysis was inspired by photosynthetic systems in nature and has attracted tremendous interests since the first report of a conventional Z-scheme photosystem in 1979 (Fig. 9 (a)) [290], [291]. Depending on the charge transport route involved, Z-scheme photosystems are classified as conventional, all-solid-state, and direct, which are distinguished by the specific transport medium used to facilitate the charge transport of reversible redox ion pairs, electronic conductors, and direct interfacial contacts, respectively [291], [292], [293], [294], [295]. For example, Ma et al. prepared a Z-scheme WO₃/K⁺Ca₂Nb₃O₁₀ ⁻ binary 2D-2D heterojunction photocatalyst from a facile hydrothermal co-assembly method at ambient conditions [147]. The as-prepared WO₃/K⁺Ca₂Nb₃O₁₀ ⁻ photocatalyst exhibited a remarkable improvement in the photodegradation of TC-HCl under simulated sunlight irradiation, compared to pristine WO₃ and K⁺Ca₂Nb₃O₁₀ ⁻. 20%-WO₃/K⁺Ca₂Nb₃O₁₀ ⁻ exhibited the maximum activity, which was 5.1 and 2 times higher than those of WO₃ and K⁺Ca₂Nb₃O₁₀ ⁻, respectively. These enhanced properties and sustainability were attributed to the tightly-coupled heterointerfaces, increased SBET area, optimized optical-capturing capability, and improved carrier dynamics of the semiconductor in the Z-scheme procedure.

Fig. 9.

(a) Roadmap of Z-scheme photocatalytic system evolution [293]; (b) suggested photoinduced charge transport on g-C₃N₄(60)/TNTAs [296]

Yang et al. [155] designed a new Z-scheme heterostructure Ag@AgI/VI-BOI (iodine-vacancy-rich BiOI), whereby Ag@AgI NPs were loaded locally on the outer surface of defect-rich BiOI nanosheets. This Z-scheme heterostructure possessed a remarkable visible-light-induced photocatalysis capacity for TC treatment, which was significantly higher than those of pristine BiOI, VI-BOI, and Ag@AgI. The enhanced photocatalytic properties of this Z-scheme heterostructure originated from the synergistic action between VI (iodine vacancies), interface junctions between AgI and VI-BOI, and Ag⁰. A direct visible-light-induced solid-state C₃N₄ NS/TNTAs (g-C₃N₄ nanosheet/TiO₂ nanotube array) Z-scheme heterostructure (Fig. 9(b)) was successfully fabricated in-situ by adding preformed g-C₃N₄ nanosheets into the anodizing bath solution [296]. Enhanced photocatalytic performances and good stability were observed on nanosheet-supplemented photocatalysts, and g-C₃N₄(60)/TNTAs exhibited an optimum activity for the degradation of rhodamine B (RhB) and colorless TC-HCl. The optimized photodegradation performance was derived from an enhanced light harvesting capacity, a suppressed carrier recombination, and an extended carrier lifetime.

3.2.9. Double heterojunction

In addition to heterojunctions, double heterojunctions such as QDs Schottky heterojunctions combined with type II, Z-scheme heterojunction [127], [130], [131], [132], type I combined with Z-scheme heterojunctions, and double Z-scheme heterojunctions have been applied to treat antibiotics (e.g., tetracycline hydrochloride, ciprofloxacin, oxytetracycline, ciprofloxacin, etc.) in water [135]. Tang et al. [297] prepared ternary Ag/CuNb₂O₆/CuFe₂O₄ heterostructures. The prepared multicomponent heterostructures exhibited stronger visible-light-induced photodegradation properties towards methylene blue (MB) and the pesticide imidacloprid than their single (CuNb₂O₆) and binary (Ag/CuNb₂O₆ or CuNb₂O₆/CuFe₂O₄) counterparts. The excellent photodegradation performance of the obtained ternary heterostructure was attributed to an enhanced visible-light harvesting capacity and rapid charge recombination restraint.

Jo et al. [130] developed a novel ternary 2D/2D/2D-configured LDH/CN/RGO heterostructure (CoAl-layered double hydroxide/g-C₃N₄/reduced graphene oxide) from a simple one-step hydrothermal synthesis. The prepared LDH/CN/RGO heterostructure exhibited a significantly superior visible-light-induced photodegradation activity for the treatment of Congo red (CR) and TC in water. Particularly, the LDH/CN/RGO heterostructure containing RGO (1 wt%) and LDH (15 wt%) featured optimum photodegradation properties in all the synthesized photocatalysts. The significantly enhanced photodegradation efficiency and outstanding consistency of the heterostructure LDH/CN/RGO were primarily attributed to the intense interface interaction between the CN, LDH, and RGO components (Fig. 10 ). This strong interfacial effect was further facilitated by the specifically arranged 2D/2D/2D microstructure, which ultimately accelerated the charge transport at the interface and efficiently separated photoinduced electrons and holes.

Fig. 10.

Photocatalytic degradation mechanisms of CR and TC by a ternary LDH/CN/RGO heterostructure [130]

3.3. Strategies to manage surface reactions

Although charges can be effectively separated, recombination in the process of electron transport remains problematic when the charges cannot be consumed promptly due to low activity or limited active sites on the semiconductor surface. Therefore, strategies have been proposed to promote intrinsic active sites on photocatalysts and the adsorption of pollutants.

3.3.1. Selective exposure of high energy facets

Typically, the redox reactions that occur on the outer surface or interface of semiconductors are highly susceptible to the exposed facets. Based on crystal anisotropy data, it is widely known that different crystal facets have different atomic arrangements, which result in different physical and chemical properties, including anisotropic surface electronic structures, tunable surface energy, and diverse molecule absorption abilities and reactivities [279].

Hao et al. [168] synthesized Bi₂O₂(OH)(NO₃) nanosheets (BON-NS) with a predominant exposure of the reactive (0 0 1) plane through a wet-chemical process using sodium dodecylbenzene sulfonate (SDBS) as a template. They found that the BON-NS samples consisted of loosely stacked nanosheets, which provided an exceptionally high specific surface area, as well as more efficient charge separation. This feature was likely derived from the dominant exposure of reactive (0 0 1) facet and the [Bi₂O₂(OH)]⁺ stacking layers down the c axis, which induces the orientation of an auto-built electric field. This allows the developed BON-NS to have an elevated capability for carrier separation and migration, leading to significantly superior UV-light photocatalysis performance for the treatment of antibiotics (tetracycline hydrochloride) than their bulk counterparts [168]. Additionally, hexagonal WO₃ (h-WO₃) was synthesized by employing sodium sulfate and ammonium sulfate as end-capping reagents through diverse hydrothermal methods [298]. Particularly, h-WO₃ NSs with exposed (0 0 2) planes exhibited the highest visible-light photodegradation performance owing to their large BET surface and interior electron–hole separation on high-energy (0 0 2) reactive planes.

3.3.2. Preparation of 2D porous materials

Porous materials have numerous notable advantages, including a high-density active core that facilitates photocatalytic reactions, high light absorption rates derived from the reflection and scattering of incident light within the pores, and high specific surface areas that improve the adsorption of contaminants and accelerate surface reactions.

Two-dimensional (2D) materials, including black phosphorus (BP), transition metal dichalcogenides (TMDs), metal oxides, metal carbides, metal nitrides, hexagonal boron nitride (h-BN), g-C₃N₄, and layered double hydroxides (LDH), have attracted substantial interest in the field of photocatalysis [299], [300], [301]. When performing as photocatalysts, the high BET surface and abundant surface reactive sites of 2D candidates may result in enhanced catalytic activity. To further improve photocatalytic properties, 2D porous materials with a more accessible BET surface and additional reactive sites are developed to completely expose the active sites to the contaminants and participate in the oxidation–reduction process, which degrades the contaminants and produces H₂/O₂.[302] Unfortunately, 2D porous materials, such as MoSe₂ [303], [304], MoS₂ [305], [306], [307], and g-C₃N₄ [308], [309], [310], [311], [312], [313], [314], [315], [316], [317], [318], [319], [320], [321], [322] are mostly used in photocatalytic hydrogen evolution or photodegradation of organic dyes and are rarely applied for the degradation of antibiotics. Nonetheless, 2D porous materials could be promising candidates for the photocatalytic treatment of antibiotics.

Kang et al. [311] prepared few-layer g-C₃N₄ NSs with foamy pores by ultrafast liquid-N₂-frozen exfoliation of its bulk counterpart within 10 s. The foamed porous super-thin g-C₃N₄ NSs exhibited interesting advances such as well-crystallized characteristics, a narrowed energy band, an additional unveiled edge, a slightly-augmented BET surface (135.6 m²/g) with an increased mesopore, and an enhanced charge transport capability. Unlike massive g-C₃N₄, the super-thin porous g-C₃N₄ exhibited a remarkably promoted capability to generate reactive oxygen species (ROS) and a 4-fold improvement in performance for visible-light-excited RhB decomposition. This primarily resulted from an increase in reactive sites and a shortened route for charge transport. These results suggest that this technology could be further developed to generate high-performance super-thin g-C₃N₄ NSs for contaminant treatment.

Template-free 2D porous super-thin g-C₃N₄ NSs with doped oxygen atoms were synthesized by She et al. [316]. The photodegradation performance (60.8%) of MO and the mean H₂ production speed (~189.3 μmol·h⁻¹) were nearly 71 and 5.2 times higher than those in the bulk phase, respectively. This enhanced photocatalysis derived from unique features such as an enriched adsorbable/reactive site, a reinforced oxidation–reduction capability and enhanced charge migration.

3.3.3. Preparing 3D aerogel materials

Aerogel photocatalysts with 3D continuous network structures can provide ultra-large accessible surface areas, rich porosities, and a large number of photocatalytic active sites, which can result in high absorption capacity and enhanced photocatalytic performance. Moreover, 3D aerogel materials can solve the problem of dispersion and recovery of photocatalytic materials in practical applications [186]. Liu et al. [159] recently fabricated ternary CeVO₄/3D RGO (reduced graphene oxide) aerogel/BiVO₄ photocatalysts (Fig. 11 ) and used them for visible-light-induced O₂ generation and TC removal. The synthesized aerogel composites displayed a well-defined 3D hierarchical microstructure built by the two vanadates and RGO. The optimum heterostructure exhibited a significantly facilitated visible-light-excited H₂O oxidation capacity as well as TC degradation. This optimized photocatalysis capacity resulted from the large BET surface and ternary Z-scheme heterojunction of the aerogel. These features enabled the 3D RGO aerogel to connect to the other semiconductors and act as distinct dual pathways to further enhance the charge transport.

Fig. 11.

Synthesis of BiVO₄/RGO/CeVO₄ heterostructures [159]

A freestanding MoS₂ nanosheets/graphene aerogel (3D MoS₂ NS/GA) heterostructure was recently created through a simple hydrothermal process [186]. The obtained 3D MoS₂ NS/GA exhibited an excellent visible-light photodegradation efficiency for TC-HCl disposal with superior stability and reusability. Chen et al. [185] prepared powerful cellulose aerogel/blended polyester fibers (3D g-C₃N₄@CA/B-PET) via a simple process by supporting g-C₃N₄ NSs on B-PET reinforced cellulose aerogel (CA). The prepared g-C₃N₄@CA/B-PET photocatalyst exhibited a superior photocatalytic performance to treat sulfaquinoxaline sodium and Cr (VI), compared to pristine g-C₃N₄. Moreover, heterostructure g-C₃N₄@CA/B-PET aerogel can be easily recycled and reused and is remarkably resistant to light and water currents.

4. Summary and outlook

The hazards of antibiotics and their treatment methods were briefly discussed herein, followed by a discussion on the current state of photocatalytic degradation of antibiotics in water. Photocatalysis plays an important role in treating antibiotic residues in water due to its superb features. However, efforts have to be made to further improve the efficiency for the wide application of this technology. The strategies for the improvement of degradation efficiency are the following:

• (1)increasing light-harvesting capacity via defect engineering
• (2)enhancing charge separation via interface engineering
• (3)accelerating surface reaction.

These strategies are promising based on previous studies. However, for further practical applications, many challenges remain to be addressed:

• a)It is difficult to identify the exact occurrence, formation, concentration, and types of defects (especially for vacancies) using current characterization techniques. This produces barriers for understanding the relationship between the structure and performance, which limits the further development of these technologies.
• b)The construction of traditional heterojunctions via interface engineering has been intensively studied. Nonetheless, only a few previous studies have focused on homojunctions and double heterojunctions, given that current preparation methods are complex, time-consuming, non-eco-friendly, and theoretically lacking.
• c)The BET surface and the number of active sites of a catalyst are key factors for accelerating surface reactions. In addition to increasing the specific surface area by controlling catalyst morphology, the selective exposure of highly active high-specific-surface-area crystal faces has recently become a popular research topic. However, most of these materials are used for hydrogen production, with few instances of antibiotic treatment applications.

Despite these challenges, this review provides valuable information on improvement and modification strategies for the design of high-performance photocatalysts to treat antibiotic residues in water. Besides, it is difficult to treat the antibiotic residues in practice by a sole technology. Innovating a complex system comprising photocatalysis and photoelectrocatalysis or electrocatalysis could be a promising alternative for the efficient removal of antibiotic residues in water. This review provides new insights into the design of high-efficiency photocatalysts for the degradation of antibiotic residues, thereby furthering the development of photocatalysis for water treatment and other fields.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.