Introduction

Antibiotics are one of the most emergent pollutants in natural aqueous environments worldwide due to the increased antibiotic abuse, resulting from the increased demand of the pharmaceuticals and the rapid development of the drug and medical industry (Akhil et al. 2021; Kumar et al. 2021; Madhura et al. 2018; Shi et al. 2019; Wei et al. 2020; Yang et al. 2021b). Tetracycline, one of the highly effective broad-spectrum prescribed antibiotics, is extensively used in the therapeutic treatment, and livestock and poultry industry (Daghrir and Drogui 2013; Peng et al. 2021; Zheng et al. 2021). Nevertheless, tetracycline gets into the environment through urine and feces since it is incompletely absorbed in the body and has poor metabolic transformation (Barhoumi et al. 2017; Guo et al. 2019a). Moreover, tetracycline is not easily degradable because it is chemically stable, thus difficult to be eliminated under the natural environment (Yan et al. 2021a; Yang et al. 2018; Younis et al. 2020). Therefore, it is necessary to identify an effective and feasible technology for degrading tetracycline in water environments.

Light-activated semiconductor-based photocatalysis represents the most promising green and eco-friendly technique for wastewater treatment (Liu et al. 2019a; López et al. 2021; Saravanan et al. 2020). It has received a lot of attention due to its high efficiency, low-cost, and outstanding stability (Peng et al. 2021; Zhou et al. 2019). Electrons (e) and holes (h+) are generated under light irradiation and migrate to the surface of photocatalyst (López et al. 2021; Soni et al. 2021), inducing redox reactions and forming of reactive oxygen species, such as superoxide radical (O2 · −), singlet oxygen (1O2), hydroxyl radical (OH·), hydroperoxyl radical (HOO·), and hydroperoxide (H2O2). The reactive oxygen species with high oxidative and reductive activity can easily degrade organic and inorganic pollutant molecules.

Although many significant advances have been made in photocatalysis, two major bottlenecks limit the practical application of pristine photocatalysts under solar light. First, several pure photocatalysts with wide band gaps can only be activated using ultraviolet light (Li et al. 2020d), less than 5% of the sunlight spectrum. Second, photoproduced e and h+ pairs of single semiconductor photocatalysts generally likely to recombine, reducing quantum and photocatalytic efficiencies (Louangsouphom et al. 2018; Soltani et al. 2019). Various strategies, such as element doping (He et al. 2020b), semiconductor coupling (Guo et al. 2019b), and dye sensitization (Liu and Wang 2019) have been developed in the past decades to overcome the inherent drawbacks associated with pristine photocatalysts. However, semiconductor coupling is one of the most popular research hotspots for photocatalysis since it allows for the introduction of heterojunction systems, effectively improving the separation of e and h+ (Ifebajo et al. 2018; Trinh Duy et al. 2020).

Many studies have reported on tetracycline degradation after visible light irradiation using different heterostructure photocatalytic materials in recent years. For instance, Guo et al. (2019a) fabricated a 2D/2D Z-scheme heterojunction photocatalyst by coupling CuInS2 and g-C3N4 semiconductors to efficiently remove tetracycline from water under visible illumination. Acharya et al. (2020) also reported that a boron nitride/B-doped-g-C3N4 heterojunction composite could effectively inhibit the recombination of electron–hole pairs, thus degrading tetracycline antibiotics under visible light. Wang et al. (2018a) built a visible-light-driven BiOBr/Bi2SiO5 heterojunction to enhance the elimination of tetracycline in an aqueous solution. Although several articles have reviewed the development of heterojunction architectures and their application in the purification of different environmental pollutants (Li and Li 2020; Low et al. 2017; Nemiwal et al. 2021; Wang et al. 2014; Yu et al. 2014a), none has comprehensively summarized and emphasized on tetracycline photodegradation.

This study focused on photocatalytic degradation of tetracycline using various heterojunction photocatalytic composites under light irradiation. Moreover, the fundamental mechanisms and electron–hole separation of different heterojunction systems are summarized. The proposed degradation pathways of tetracycline during heterojunction photocatalysis also are assessed. Furthermore, the photocatalytic performance of heterogeneous heterojunction photocatalysts is objectively evaluated and compared using several quantified metrics including quantum efficiency, space time yields, and figures of merit. Finally, future research requirements for heterojunctions and their applications in water environmental remediation are discussed as well.

Mechanisms of semiconductor heterojunction photocatalysis and the separation of electron–hole pairs

Several researches have reported the basic photophysical and photochemical mechanisms underlying heterojunction photocatalysis (Low et al. 2017; Wang et al. 2014; Xu et al. 2019; Yu et al. 2014a). An appropriate heterojunction interface can effectively inhibit the recombination of photoinduced e and h+ during photocatalysis, thereby enhancing the overall photo-activity of catalysts (Li and Li 2020). Heterojunction typically could be grouped into six categories based on the heterojunction structure: traditional (e.g., type I, type II, and type III), p–n, Z-scheme, step-scheme, Schottky, and surface heterojunctions (He et al. 2020b; Yang et al. 2021c). Various heterojunction architectures generally show diverse photocatalysis procedures and mechanisms for the separation of photoexcited carriers.

Traditional heterojunction

Both photogenerated electrons and holes in a conventional type I heterojunction photocatalyst can flow from the semiconductor with the higher conduction and lower valence bands to another semiconductor (Fig. 1a) (Marschall 2014). However, type I heterojunction cannot spatially separate charge carriers, leading to the accumulation of electron–hole pairs in the same semiconductor, thus accelerating their recombination. Besides, photocatalysis reactions occur in the semiconductor with relatively lower oxidation and reduction potential (Low et al. 2017), reducing photocatalysis performance under light irradiation. In type II heterojunctions (Fig. 1b), the electrons transfer to the semiconductor with a lower conduction band level. Meanwhile, holes transfer to the semiconductor with a higher valence band level, decreasing the contact and recombination of electron–hole pairs (Zhou et al. 2012). Although the structure of type III heterojunction (Fig. 1c) is similar to that of type II, the bandgaps of both semiconductors do not overlap in the interface since the levels of band is set so extreme (Lu et al. 2019). In terms of traditional heterojunction system, only type II heterojunction can be an ideal system to improve the separation of electron and holes.

Fig. 1
figure 1

Schematic illustration of band structure and charge carries transfer of traditional heterojunction systems. a Type I: both e and h+ migrate to the same semiconductor photocatalyst. b Type II: e and h+ transfer to the photocatalyst with a lower conduction band and the photocatalyst with a higher valence band, respectively. c Type III: migration of e and h+ between the two semiconductors cannot occur (Reprinted with permission of Wiley from Low et al. 2017). e, electron; h+, hole; CB, conduction band; VB, valence band

Yuan et al. (2019b) prepared a novel type II In2S3/InVO4 heterojunction by loading Na2S with pre-synthesized InVO4 microspheres in a reaction solution (Fig. 2a). The as-prepared In2S3/InVO4 composite showed outstanding photocatalytic performance for tetracycline degradation under visible light. The photo-activity of In2S3/InVO4 was 11.71 times and 2.26 times higher than that of pristine InVO4 and In2S3, respectively. The photosensitization of InVO4 by In2S3 and enhanced charge separation through the compact interface junction between semiconductors significantly promoted photocatalytic activity of In2S3/InVO4. Huang et al. (2019) also synthesized a series of different hierarchical heterojunctions, including g-C3N4/BiOI (type I), g-C3N4/Bi4O5I2 (type I), and g-C3N4/Bi5O7I (type II) using direct precipitation and in situ calcination transformation (Fig. 2b–d). g-C3N4/Bi5O7I had the highest tetracycline photodegradation activity due to its increased specific surface area, leading to highly efficient charge separation in type II heterojunction.

Fig. 2
figure 2

a The possible charge-transfer mechanism of In2S3/InVO4: type II heterojunction system built between In2S3 and InVO4 (Reprinted with permission of Elsevier from Yuan et al. 2019b). b The synthesis process, c Mott–Schottky curves, and d transient photocurrent responses of g-C3N4/BiOI, g-C3N4/Bi4O5I2, and g-C3N4/Bi5O7I under visible irradiation (Reprinted with permission of Elsevier from Huang et al. 2019). e, electron; h+, hole; ·O2 , superoxide radical; TC, tetracycline

pn heterojunction

Internal electric field occurs on the interface between p- and n-type semiconductors in the p–n heterostructure system due to the diffusion of charge carriers between two types of semiconductors (Fig. 3) (He et al. 2021a). This internal electric field quickly drives electrons to the conduction band of n-type semiconductor and holes to the valence band of p-type semiconductor. Therefore, the presence of the internal electric field in p–n heterojunction system ultimately effectively separates electrons and holes.

Fig. 3
figure 3

Schematic illustration of the internal electric field and charge-carrier transfer of a p–n heterojunction system: i before light irradiation, the diffusion of e and h+ between p- and n-type semiconductors forms an internal electric field. ii After light irradiation, photogenerated e and h+ are, respectively, driven to n- and p-type semiconductors under the influence of the internal electric field. (Reprinted with permission of Wiley from Low et al. 2017). e, electron; h+, hole; CB, conduction band; VB, valence band

Liao and co-workers (Liao et al. 2020) constructed a visible-light-driven BiFeO3/TiO2 p–n heterojunction composite via facile hydrolysis and precipitation method (Fig. 4a–c). BiFeO3/TiO2, BiFeO3, and TiO2 had tetracycline degradation efficiency of 72.2%, 64.9%, and 38.3%, respectively, after 180 min of visible illumination. BiFeO3/TiO2 had an enhanced photocatalytic performance due to the synergistic effect of the ferroelectric effect of BiFeO3 with the internal electric field of BiFeO3/TiO2 p–n heterojunction, which significantly inhibited charge-carrier recombination. Fiber-shaped Ag2O/Ta3N5 p–n heterojunctions are efficient photocatalysts designed by combining porous Ta3N5 nanofibers with Ag2O nanoparticles via in situ anchoring (Fig. 4d–f) (Li et al. 2017). Besides the internal electric field of the p–n junction, Ag2O/Ta3N5 had hierarchical pores and a high specific surface area, increasing the photocatalytic activity.

Fig. 4
figure 4

a The degradation mechanism and charge-transfer pathway of BiFeO3/TiO2: p-n heterojunction system occurred between BiFeO3 and TiO2. b and c Scanning electron microscopy images of BiFeO3/TiO2 p-n heterojunction (Reprinted with permission of Elsevier from Liao et al. 2020). d and e Scanning electron microscopy images and f preparation process of Ag2O/Ta3N5 p-n heterojunction (Reprinted with permission of Elsevier from Li et al. 2017). e, electron; h+, hole; CB, conduction band; VB, valence band; OH·, hydroxyl radical

Z-scheme heterojunction

Z-scheme heterojunction can be classified into traditional, all-solid-state, and direct Z-scheme photocatalysts. Their charge-transfer routes are similar but no intermediate is used in direct Z-scheme system (Low et al. 2019; Yu et al. 2018). Recombination in a direct Z-scheme photocatalyst occurs between the weak electrons from the semiconductor with less negative conduction band and holes from the semiconductor with less positive valence band (Fig. 5) (Ani et al. 2018; Qi et al. 2017). Therefore, the holes with high oxidation ability and the electrons with high reduction ability cannot recombine and be continuously maintained for further photocatalytic reactions, hence increasing the total redox potential of the Z-scheme catalyst system. Z-scheme heterojunction provides a new insight for enhancing the photocatalytic activity of photocatalysts: consuming the useless electrons and holes and remaining high redox potential of the whole system.

Fig. 5
figure 5

Schematic illustration of the energy band structure and charge-carrier transfer on a direct Z-scheme heterojunction system: e from PS II directly recombine with h+ from PS I. Thus, e with high reduction potential in PS I and h+ with high oxidation potential in PS II can be remained. (Reprinted with permission of Elsevier from Qi et al. 2017). e, electron; h+, hole; CB, conduction band; VB, valence band; PS, photosystem

Li et al. (2018) fabricated a mesoporous Z-scheme heterojunction system with increased specific surface area (124.09 m2/g) via anchoring γ-Fe2O3 nanoparticles on the surface of g-C3N4 nanosheet to further enhance tetracycline photodegradation under visible light. γ-Fe2O3/g-C3N4 presented the best photocatalytic performance in the case of 5% γ-Fe2O3 loaded, with a rate constant tetracycline degradation value of 0.0134 min−1 (higher than that of pure g-C3N4, 0.0020 min−1). The more superior photocatalytic activity of γ-Fe2O3/g-C3N4 is due to the extended response range of visible light and improved separation of photogenerated charge carriers in Z-scheme heterostructure. Chen et al. (2016) synthesized a direct Z-scheme heterostructured photocatalyst AgI/BiVO4 via an in situ precipitation approach (Fig. 6). The AgI/BiVO4 (1:4 mass ratio) heterojunction demonstrated higher tetracycline degradation efficiency (94.91%) than the original BiVO4 (62.68%) and AgI (75.43%). The emergence of Ag during photocatalysis process transforms the composite structure from AgI/BiVO4 into AgI/Ag/BiVO4, developing a sandwich-like Z-scheme photocatalyst in AgI, Ag, and BiVO4, thus efficiently enhancing the separation of electron−hole pairs.

Fig. 6
figure 6

a Photocurrent responses of AgI, BiVO4, and AgI/BiVO4 b Photocatalysis mechanism of tetracycline degradation under visible light illumination using AgI/BiVO4 photocatalyst: Z-scheme heterojunction system. 3D EEMs after c 30 and d 60 min irradiation: fluorescence signals increased with the irradiation time, suggesting the decomposition of tetracycline under light over AgI/BiVO4 (Reprinted with permission of American Chemical Society from Chen et al. 2016). 3D, three-dimensional; EEMs, excitation—emission matrix fluorescence spectroscopy

Step-scheme heterojunction

Step-scheme systems are exhibited between two n-type semiconductors with staggered band structure, closely resembling that of type II except for the greatly different path of charge migration (Fig. 7) (Fu et al. 2019; He et al. 2020a; Xu et al. 2020a). “Step-scheme” is a new concept that can be used to describe the photocatalytic mechanisms clearly. An electric field is formed from the semiconductor with greater work function and lower Fermi level (semiconductor A) to that with smaller work function and higher Fermi level (semiconductor B) due to the spontaneous diffusion of electrons from semiconductor A to B across their interface until the Fermi energy levels of two semiconductors are equal (Hasija et al. 2021; He et al. 2020b). In the same time, the energy band gap of semiconductor A bends upward due to electron depletion, and that of semiconductor B bends downward due to electron accumulation (Hasija et al. 2021; He et al. 2020a). Therefore, the weak reductive electrons from the conduction band of semiconductor B will recombine with the weak oxidative holes from the valence band of semiconductor A under the effect of built-in electric field, band edge bending, and Coulombic interactions, similar to Z-scheme heterostructure. (Fan et al. 2020). Photocatalytic reduction and oxidation reactions occur on the surface of the high-potential semiconductor B and A, respectively (Li et al. 2021b). The concept of step-scheme is the first to propose the transfer of charge carriers under the effect of band bending, which is similar to how water flows downhill.

Fig. 7
figure 7

Schematic illustration of the internal electric field and charge-carrier migration on a step-scheme heterojunction system: i An internal electric field is built due to the Femi energy equilibrium; ii Femi levels of semiconductor A and B bend upward and downward, respectively; iii recombination occurs between eCB of semiconductor B and hVB+ of semiconductor A (Reprinted with permission of Elsevier from He et al. 2020b). eCB, electron; hVB+, hole; CB, conduction band; VB, valence band; Eg, energy gap

Step-scheme heterostructure has been widely designed and used to decompose tetracycline antibiotics in an aqueous solution. For instance, a flower-like step-scheme BiOBr/BiO(CH3COO)1−xBrx was successfully constructed by Jia et al. (2020) through a simple co-precipitation procedure at room temperature (Fig. 8). The pre-synthesized step-scheme junction exhibited outstanding visible-light-driven photocatalytic properties especially when the molar ratio of Br/Bi was 0.8 with tetracycline degradation efficiency of 99.2% after 2 h of illumination. The enhanced photocatalytic activity of BiOBr/BiO(CH3COO)1−xBrx photocatalyst was due to: (i) flower-like contour greatly improved the dispersion, thereby accelerating the migration of charge carrier at the interfaces and photocatalyst surface; (ii) the visible light response of the whole composite was enlarged, developing a step-scheme junction owing to the formation of the solid solution BiOBr/BiO(CH3COO)1−xBrx. Fan et al. (2020) fabricated a face-to-face hierarchical In2S3/Bi2O2CO3 photocatalyst. The pre-prepared composite had a two-dimension 2D/2D step-scheme structure and excellent photocatalytic performance and cycling ability for tetracycline. This optimized photodegradation capability of the In2S3/Bi2O2CO3 heterostructure was ascribed to the increased visible response region and restrained electron–hole recombination in step-scheme heterojunction.

Fig. 8
figure 8

a Photocatalysis mechanism of step-scheme BiOBr/BiO(CH3COO)1−xBrx systems: band gap of BiO(CH3COO)1−xBrx decreased to 2.91 eV and both BiOBr and BiO(CH3COO)1−xBrx photocatalysts could be excited by visible light. b Photocatalysis mechanism of type I BiOBr/BiO(CH3COO): only BiOBr could be excited by visible light. c Transmission electron microscopy and d high-resolution transmission electron microscopy images of BiOBr/BiO(CH3COO)1−xBrx (Reprinted with permission of Elsevier from Jia et al. 2020). e, electron; h+, hole; ·O2 , superoxide radical; S-0.4, BiOBr/BiO(CH3COO)0.6Br0.4

Schottky heterojunction

Schottky junctions are formed by contact between a semiconductor photocatalyst and a metal (Fig. 9) (Di Bartolomeo 2016). Photogenerated electron can transfer from metal to a semiconductor due to their unequal Fermi energies, thus building the Schottky barrier (Yang et al. 2021c), which suppresses the recombination of charge carrier.

Fig. 9
figure 9

Schematic illustration of the energy band structure and charge-carrier transfer on a Schottky junction system: e of the semiconductor tends to transfer to the metal rather than recombined with h+. e, electron; h+, hole; CB, conduction band; VB, valence band

2D/2D ultrathin interfacial Schottky Ti3C2 Mxene/SnNb2O6 nanosheets were synthesized by Wang and co-workers (Wang et al. 2021a) through an ultrasonication-assisted hydrothermal procedure (Fig. 10a–d). The 2D/2D Ti3C2/SnNb2O6 had broadened visible response and larger interfacial contact area. The internal electric field effectively separated the charge in Ti3C2/SnNb2O6 system, leading to a higher photocatalytic degradation efficiency of tetracycline with a high optimal reaction rate, which was 2.53 times that of pure SnNb2O6. Luo et al. (2015) prepare a Ag/Bi3TaO7 plasmonic photocatalyst with a Schottky junction using a facile photoreduction method (Fig. 10e). The synergistic effect of the surface plasmon resonance of Ag particle on the surface of Bi3TaO7 improved the separation process of carriers, thus heightening the photocatalytic activity of the whole composite. The best performance of tetracycline degradation of 85.42% by Ag/Bi3TaO7 was exhibited when 1 wt % Ag loaded on the surface of Bi3TaO7. Ag/Bi3TaO7 also showed excellent recyclability and only a slight degradation efficiency loss after five successive cycles.

Fig. 10
figure 10

a Ultraviolet photoelectron spectra edges of Ti3C2 and SnNb2O6. Theoretical calculations of work function values of b Ti3C2 and c SnNb2O6. d Charge density difference diagram of Ti3C2/SnNb2O6 (d) (Reprinted with permission of Elsevier from Wang et al. 2021a). e Proposed photocatalytic mechanism of Ag/Bi3TaO7 for degradation of tetracycline under visible irradiation: SPR absorbance of metallic Ag nanoparticles and charge carriers transfer in Schottky heterojunction of Ag/Bi3TaO7 (Reprinted with permission of American Chemical Society from Luo et al. 2015). TC, tetracycline; O2 · −, superoxide radical; ·OH hydroxyl radical; SPR, surface plasmon resonance

Surface heterojunction

Surface heterojunction mainly consists of two different crystal facets of a single semiconductor (Fig. 11) (Sajan et al. 2016; Yu et al. 2014b). The proposed mechanism underlying the separation of photosensitized carriers in surface heterojunction is similar to that of type II (Zhou and Li 2012). Nevertheless, the redox potential of surface heterojunction is lower than that of type II heterojunction due to the insignificantly different band structures of two co-exposed facets of a single semiconductor (He et al. 2020b).

Fig. 11
figure 11

Charge-carrier migration of surface heterojunction system under light irradiation: different exposed facts of a single semiconductor have different band structures, leading to the formation of heterojunction system and migration of photogenerated eCB and hVB+. eCB, electron; hVB+, hole

The aforementioned heterojunctions are binary systems, and the behavior of electrons and holes has been discussed based on binary photocatalyst structures. Various ternary and quaternary heterojunction composites have been recently constructed to enhance photocatalytic activity (Jo and Tonda 2019; Yang et al. 2020).

Some studies have indicated that the separation mechanisms of charge carriers in ternary and quaternary heterojunction following that of binary heterojunction systems. Ni et al. (2020) successfully synthesized a ternary sandwich-like TiO2-x/ultrathin g-C3N4/TiO2-x for efficient degradation of tetracycline under visible light and found that direct Z-scheme system was introduced to this ternary heterojunction (Fig. 12). Liu et al. (2021) synthesized a visible-light-driven g-C3N4/TiO2/CdS ternary heterojunction nanocomposite to effectively remove tetracycline antibiotics from wastewater (Fig. 13a, b). The separation mechanism of electron–hole pair in g-C3N4/TiO2/CdS was similar to that of Z-scheme heterojunction. Yan et al. (2021a) developed an Ag/γ-AgI/Bi2O2CO3/Bi quaternary composite photocatalyst via a solvothermal-precipitation method and obtained a step-scheme structure (Fig. 13c, d). However, the transfer process of carriers of many other ternary and quaternary heterojunction systems is so complex and cannot be clearly and systematically explained. Heterogeneous heterojunction systems and their corresponding mechanisms regarding the migration of photogenerated carriers and degradation of tetracycline antibiotics are listed in Table 1.

Fig. 12
figure 12

a Formation route of TiO2-x/ultrathin g-C3N4/TiO2-x: planetary grinding method. b Possible photocatalytic mechanism diagram of TiO2-x/ultrathin g-C3N4/TiO2-x: direct Z-scheme heterojunction model (Reprinted with permission of Elsevier from Ni et al. 2020). TCH, tetracycline hydrochloride; O2 · −, superoxide radical; ·OH hydroxyl radical; UCN, ultrathin g-C3N4; S-TUCN, sandwich-like TiO2/ultrathin g-C3N4/TiO2; S-TUCNov, ov-mediated sandwich-like TiO2/ultrathin g-C3N4/TiO2

Fig. 13
figure 13

source and h+ acceptor. b First-order kinetics of g-C3N4, Cds, TiO2, KCNN, KCNNT, TiO2@CdS, KCNN@CdS and KCNNT@CdS for tetracycline removal under visible light (Reprinted with permission of Elsevier from Liu et al. 2021). c The effects of Bi metal and oxygen vacancies on the carrier transfer: oxygen vacancies acted as intermediate to transfer e to Bi metal. Then, e in Bi metal moved to the conduction band due to SPR effect. d Proposed photocatalytic mechanism of 30% Ag/γ-AgI/Bi2O2CO3/Bi: a novel S-scheme heterojunction mechanism. (Reprinted with permission of Elsevier from Yan et al. 2021a). e, electron; h+, hole; CB, conduction band; VB, valence band; O2 · −, superoxide radical; ·OH hydroxyl radical; KCNN, K doped g-C3N4 nanosheet; KCNNT, K doped g-C3N4 nanosheet-TiO2; Eg, energy gap; SPR, surface plasmon resonance

a Migration of charge carriers in g-C3N4/TiO2/CdS: a type II and Z-scheme synergistic heterojunction mechanism. CdS acts as an e

Table 1 Photocatalysis mechanisms of various heterojunctions photocatalysts for the removal of tetracycline under light irradiation
Table 2 Performance evaluation of various heterojunction photocatalysts against tetracycline

Proposed photodegradation pathways of tetracycline

Various heterojunction photocatalysts have been used for the photodestruction of tetracycline in aqueous solutions. The breakdown of tetracycline molecules involves the attack of some sites by various reactive oxygen species generated during the heterogeneous photocatalysis processes. Tetracycline degradation during photocatalytic reactions generally arises from the breakdown of double bonds, phenolic group, N-dimethyl group, and the amine group of the tetracycline molecule (Wang et al. 2018b). These sites are highly functional groups and electron-rich positions that can be attacked by active radicals, including O2 · −, OH·, h+, and H2O2 (Ren et al. 2019). Extensive studies have reported on the possible photocatalytic degradation pathways of tetracycline antibiotics. Herein, we summarized the five plausible routes for photocatalytic mineralization of tetracycline based on the initial damage sites of tetracycline caused by reactive oxygen species.

First possible pathway: attack on aromatic ring

The first breakdown pathway and relative intermediates are shown in Scheme 1. The aromatic ring of the tetracycline molecule is easy to undergo hydroxylation after the attack by OH·, thus forming isomers A1a and A1b with m/z = 461 (Xia et al. 2020). Product A2 (m/z = 433) is generated when the N-methyl group of the A1a fragment is damaged by h+. Lu et al. (2019b) found that the phenolic group of tetracycline also may take place an addition reaction to form A3 (m/z = 450). A4 (m/z = 248) is formed when the A3 is further broken down, producing A5 (m/z = 190) after the loss of hydroxyl and methyl groups.

Scheme 1
scheme 1

First proposed degradation route of tetracycline during heterojunction photocatalysis. Reactive oxygen species from photocatalysis are likely to attack the aromatic ring of tetracycline

It has been proposed that O2 · − and H2O2 can also attack the benzene ring of tetracycline to generate product A6 (m/z = 525). O2 · − radical and H2O2 can further oxidize N-dimethyl group at C12 and -C(O)NH2 group at C2 to produce A7 (m/z = 496) and A8 (m/z = 451), respectively. Subsequently, oxidization at the double bond of A8 can then form A9 (m/z = 367) (Xie et al. 2018). A9 can then be transformed to an identical ketone derivative A10 (m/z = 351). In the end, O2 · − and H2O2 continue decomposing A10 to carboxylic acids A11 (m/z = 255), CO2, and H2O.

For fragment A1b, -OH and -CH3 in N-dimethyl group and C8 site are removed after attacking by h+, producing A12 (m/z = 447) and A13 (m/z = 433) (Yan et al. 2021b). Then, A13 is further transformed into A14 (m/z = 288) and A15 (m/z = 211). Finally, A15 split into small fragments, including A16a (m/z = 135) and A16b (m/z = 121).

Second possible pathway: attack on amide group

The attack of the acylamino group at the C1 site is the second possible degradation site for tetracycline via heterojunction photocatalysis (Scheme 2). The attack of OH· results in the deamidation of the tetracycline structure, forming an intermediate B1 (m/z = 429) (Ma et al. 2019). By-products B2 (m/z = 386) (Shi et al. 2020a) and B3 (m/z = 352) are obtained via the cleavage of N, N-dimethyl group and subsequent dehydration of B1. The loss of the aldehyde group and breaking of the carbon–carbon diene bonds of B3 form B4 (m/z = 344) and B5 (m/z = 332) (Zhang et al. 2021). A further attack by reactive oxygen species produces B6 (m/z = 276). B7 with m/z = 246 can also be produced via the cleavage of the double bonds of B3. Finally, further dehydroxylation and hydrogenation of B7 form B8 (m/z = 218).

Scheme 2
scheme 2

Second proposed degradation route of tetracycline during heterojunction photocatalysis. Attack of amide group by reactive oxygen species results in the deamidation of tetracycline

It is inferred that acylamino group of tetracycline lost to boost B9 (m/z = 402). OH· radicals can damage the carbon ring of B9, thus forming B10 (m/z = 418) via hydroxylation (Wang et al. 2019). Two molecular hydrogens can then be removed from B10 to form B11 (m/z = 416) via dehydrogenation reaction. (Chen et al. 2020b) proposed that -NH2 group of tetracycline could be oxidized to produce B12 (m/z = 446). B12 dehydroxylation, deamidation, demethylation, and deamination form B13 (m/z = 408), B14 (m/z = 390), and B15 (m/z = 320).

Third possible pathway: attack on N-dimethyl group

The attack on the N-dimethyl group is the third possible degradation route of tetracycline during the photocatalysis process (Scheme 3). It is suggested that the decomposition of tetracycline after successive losing of methyl on the N-dimethyl group (Mahamallik et al. 2015) and C8 site induces the formation of D1 (m/z = 416) and D2 (m/z = 402) (Shi et al. 2020c). The attack of h+ promotes the process of N-demethylation. Separation of -C(O)NH2 group from cyclic structure and hydroxylation addition reaction on C2 transforms D2 into D3 (m/z = 377). Intermediates D4 (m/z = 274) and D5 (m/z = 186) associated with the destruction of carbon–carbon ring of D3. D5 can be further fragmented into D6 (m/z = 160) via losing -CH = CH2 group.

Scheme 3
scheme 3

Third proposed degradation route of tetracycline during heterojunction photocatalysis. The degradation of tetracycline may start with the loss of N-dimethyl group structure

Pervious research found that area of high energy in D1 molecule can be further damaged to form D7 (m/z = 400) and D8 (m/z = 353) (Ren et al. 2019). D8 then can be decomposed into D9a (m/z = 233) and D9b (m/z = 183) due to the synergistic effect of OH·, O2 · −, and h+. Further attack of O2 · − and h+ will result in the formation of A11. Jiang et al. (2017) suggested that D2 can be transformed into D10 (m/z = 333). Also, the removal of carboxyl groups and addition reaction on C7 generate D11 (m/z = 244).

Lai et al. (2021) found that N-dimethyl group in tetracycline can be directly oxidized into carbonyl group and generate E1 (m/z = 415). Then, E1 can undergo ring opening reaction to form E2 (m/z = 332), which is further converted into E3 (m/z = 318). Intermediate E4 (m/z = 430) generates when -CH3 is separated from the N-dimethyl group. The removal of -NH-CH3 and -CONH2 group of E4 forms E5a (m/z = 358), which splits into E6 (m/z = 276), CO2, and H2O.

Xie et al. (2018) indicated that E4 can be transferred into D1 or E5b (m/z = 395) after the attack of h+. Ring opening due to the fracture of carbon–carbon single bone forms E9 (m/z = 359). E7 (m/z = 359) is generated from the detachment of the hydroxyl functional group on C8 site and -CONH2 group of D1 (Yang et al. 2020). Further attack of free radicals transforms E7 into E8 (m/z = 248). Eventually, all intermediates are decomposed into inorganic substances.

It has been suggested that C-N bond of tetracycline is easily fractured, and substitution of the hydroxyl group may occur in position of C12, forming E10 (m/z = 389). Several compounds, such as E11 (m/z = 387), E12 (m/z = 371), and E13 (m/z = 355) can be formed after carbonylation, demethylation, and dihydroxylation of the E10 derivate. The removal of benzene ring from E13 directly forms E14 (m/z = 337), which further mineralized into small molecule fragments, including E15 (m/z = 297), E16 (m/z = 223), and E17 (m/z = 107).

Fourth possible pathway: attack on double bonds

The unsaturated double bonds could be the major oxidization sites in tetracycline due to the high electrophilicity of hydroxyl radicals (Scheme 4). In the fourth degradation route of tetracycline, two different intermediates (F1a and F1b) with m/z = 461 are generated after the substitution reaction on the conjugate bond between C5 and C6 or C2 and C13 (Ren et al. 2019). Further, OH· damage on different positions of F1a and F1b forms isomers F2, F3, and F19 (m/z = 477) via hydroxylation or oxidization reactions (Yang et al. 2018). The breaking of the C–C bonds in the ring structure of F3 and F1a generates F4 (m/z = 277), F5 (m/z = 209), and F6 (m/z = 149).

Scheme 4
scheme 4

Fourth proposed degradation route of tetracycline during heterojunction photocatalysis. Double bonds generally have a high affinity with reactive oxygen species and can be the main attack sites

Dealkylation of N-dimethyl group of F1a after the attack by h+ generates compound F7 (m/z = 432). Cyclic structure of F7 immediately breaks to form F8 (m/z = 363) or F13 (m/z = 376). F13 is further oxidized to form F14 (m/z = 279) and F15 (m/z = 121), which is completely decomposed into CO2, H2O, and other smaller molecular inorganic products (Chen et al. 2020a). F20 (m/z = 493) is formed after hydroxyl substitution on the aromatic ring of F19. Both F19 and F20 can form F7 (Yang et al. 2020). The attack of h+ and reactive oxygen species also can degrade F7 into F21 (m/z = 387). Subsequent destruction of carbonyl group forms F22 (m/z = 214). One of the methyl groups of -N(CH3)2 in F19 and F1a can be oxidized by OH· to generate F23 (m/z = 491) and F24 (m/z = 475), respectively.

Xia et al. (2020) also indicated that the double bond between C2 and C13 of tetracycline could be attacked by h+ radical, inducing the cleavage of unsaturated bond and forming an amino-free fragment F9 (m/z = 389). A series of processes including loss of -N(CH3)2-CHO group, dihydroxylation, and demethylation of F9 yield products F10 (m/z = 346), F11 (m/z = 304), and F12 (m/z = 258). Yu et al. (2020) suggested that the degradation pathway of tetracycline → F16 (m/z = 459) → F17 (m/z = 485) → F18 (m/z = 559) occurs via the substitution reaction on conjugated bonds.

Fifth possible pathway: attack on hydroxyl of C8 site

The fifth photocatalytic degradation pathway (Scheme 5) begins with the loss of hydroxyl at the C8 site of tetracycline to form G1 (m/z = 426). G1 dealkylation or dehydration then produces E5b and G2 (m/z = 383). Further splitting of the ring structure converts of G2 into E9 or G2 (m/z = 301) (Lai et al. 2021; Xie et al. 2018). An intermediate G4 (m/z = 337) is obtained after a series of continuous reactions including dehydration, demethylation, and deamination of G2 (Ma et al. 2019). Ring opening and multi-dehydroxylation reactions then yield G5 (m/z = 318) and G6 (m/z = 282).

Scheme 5
scheme 5

Fifth proposed degradation route of tetracycline during heterojunction photocatalysis. Attack of reactive oxygen species results in the loss of hydroxyl in C8 site

It has been revealed that tetracycline can also simultaneously lose hydroxyl and methyl groups at C8 to form G7 (m/z = 413), which can then undergo dehydrogenation to produce G8 (m/z = 411). G8 has two main photocatalytic mineralization pathways: (1) G8 → G9 (m/z = 318) → G10 (m/z = 270) or G11 (m/z = 280) → G12 (m/z = 148) → G13 (m/z = 104) → CO2, H2O, and NH3; (2) G8 → G14 (m/z = 389) → G15 (m/z = 359) → G16 (m/z = 337) → E15 → E16 → E17 → CO2, H2O, and NH3.

In summary, hydroxylation, dealkylation, dehydration, oxidation, deamination, and deamidation are the fundamental mechanistic steps for the decomposition of the molecular structure of tetracycline by various reactive oxygen species.

Performance of tetracycline degradation by heterojunction systems

Previous studies have been recorded on the development of various heterojunction photocatalysts for tetracycline photodegradation under light illumination (Guo et al. 2019a; Yan et al. 2019; Zheng et al. 2021). Therefore, it is necessary to assess their practical performance to validate their efficiency in removing tetracycline. The existing literature on the performance of heterojunction photocatalysts only focused on removal efficiency and degradation rate under identical conditions. Photocatalysis reaction is highly dependent on varying parameters (e.g., pH, temperature, initial substrate content, photocatalysts mass, the types of catalysts, and especially light wavelength and intensity), and performance of photocatalysts cannot simply be determined via degradation efficiency and rate alone (Anwer et al. 2019; He et al. 2021b). Also, degradation efficiency cannot directly reflect the actual performance or practical application value of different photocatalytic materials. In this regard, more objective and credible numeric metrics that can include as many parameters as possible are pressingly needed to be introduced to quantitatively evaluate the performance of photocatalytic systems (Vikrant et al. 2019).

Performance evaluation methods

Researchers have suggested using quantum efficiency as an elaborated index for assessing the performances of different photocatalytic systems for comparison purposes (Anwer and Park 2018; Younis et al. 2020). Recently, quantum efficiency has generally become one of the preferred performance metrics for removing various organic pollutants from water environments using photocatalytic materials (Rajput et al. 2021; Vellingiri et al. 2020). Quantum efficiency is defined as the per unit amount of absorbed photons used to effectively decompose the molecular number of pollutants (Vikrant et al. 2020). The quantum efficiency of photocatalysts during photocatalytic degradation is calculated as shown in Eq. (1) (Raza et al. 2020).

Nevertheless, quantum efficiency alone is not sufficiently standard for parallel comparing the performance of various heterojunction photocatalysts since it cannot take into account all operational variables during photocatalysis processes (He et al. 2021b). In recent years, Anwer et al. (2019) has proposed space time yield and figure of merit are proposed as new concepts for assessing photocatalytic degradation efficiency. Space time yield and figures of merit are effective performance metrics that can control the effects of catalyst dosage, pollutant content, irradiation time, and applied power. The values of space time yield and figures of merit of photocatalytic materials are obtained using Eqs. 2 and 3, respectively. The metrics have been used to determine the photocatalytic performances of the degradation of organic dye (Abbas et al. 2021; Anwer et al. 2019; Raza et al. 2021), CO2 reduction (Cortes et al. 2019), microcystin-LR removal (He et al. 2020b), gaseous benzene photomineralization (Vikrant et al. 2019), purification of volatile organic compounds (Talaiekhozani et al. 2021), and the inactivation of harmful algae (He et al. 2021b) were evaluated between various reported photocatalytic materials.

$${\text{Quantum}}\,{\text{efficiency = }}\frac{{{\text{Degradation}}\,{\text{rate}}\left( {{\text{molecules}}\,{\text{per}}\,{\text{second}}} \right)}}{{{\text{Photon}}\,{\text{flux}}\left( {{\text{photons}}\,{\text{per}}\,{\text{second}}} \right)}}$$
(1)
$${\text{Space}}\,{\text{time}}\,{\text{yield = }}\frac{{{\text{Quantum}}\,{\text{efficiency}}\,{\text{(molecules}}\,{\text{per}}\,{\text{photon)}}}}{{{\text{Photocatalyst}}\,{\text{mass}}\,{\text{(g)}}}}$$
(2)
$${\text{Figures of merit}} = \frac{{{\text{Product obtained}}\left( {\text{L}} \right)}}{{{\text{Photocatalyst mass}}\,\left( {\text{g}} \right) \times {\text{Irradiation time}}\left( {\text{h}} \right) \times {\text{Energy consumption}}\,\left( {\text{W}} \times {\text{h/mol}} \right)}}$$
(3)

Herein, recent relevant reports regarding the photocatalytic degradation of tetracycline in aqueous systems with heterojunction photocatalysts have been summarized (Table 2). Moreover, we introduced quantum efficiency, space time yield, and figures of merit to gauge the photocatalytic activity of these heterojunction composites. Their photocatalytic performance was also evaluated based on calculated quantum efficiency, space time yield, and figures of merit. To perform more intuitive assessment, the score of photocatalyst with the highest figures of merit value was adjusted to 100 as a reference, and the score of other photocatalysts was graded from 0 to 100 via conversion calculation (He et al. 2020b).

Performance comparison of various heterojunction photocatalysts for tetracycline degradation

As for photocatalytic decontamination of tetracycline, the best performance was displayed by a ternary composite Ag3PO4/AgBr/g-C3N4 with dual Z-scheme heterojunction, which showed quantum efficiency of 4.56 × 10–6 molecules/photon, space time yield of 9.13 × 10–5 molecules/(g·photon), and figures of merit of 7.68 × 10–11 mol·L/(g·J·h) (Yu et al. 2020). Ag3PO4/AgBr/g-C3N4 had a small arc radius, indicating low charge migration resistance and recombination of charge pairs. Ag element behaved as the electron transfer mediator during photodegradation to promote the recombination of electrons and holes of Ag3PO4/AgBr and AgBr/g-C3N4 (double Z-scheme systems). Therefore, better redox ability of the whole system could be obtained, improving the photocatalysis performance (Fig. 14). Ag3PO4/AgBr/g-C3N4 could represent one of the most promising heterojunctioned photocatalysts for the efficient purification of the antibiotic-contaminated sewage. Ag3PO4/AgBr also showed excellent photocatalysis activity for tetracycline degradation with a score of 87.5, quantum efficiency of 3.99 × 10–6 molecules/photon, space time yield of 7.99 × 10–5 molecules/(g·photon), and figures of merit of 6.72 × 10–11 mol·L/(g·J·h). Ag3PO4/AgBr was able to mineralize 70% of tetracycline under visible light (initial tetracycline content of 40 mg/L).

Fig. 14
figure 14

a X-ray diffraction patterns of g-C3N4, AgBr, Ag3PO4, Ag3PO4/AgBr, and Ag3PO4/AgBr/g-C3N4. b Photocatalytic degradation mechanism of tetracycline by Ag3PO4/AgBr/g-C3N4: metallic Ag acted as the electron transfer mediator to build the dual Z-scheme heterojunction. c The effect of AgBr content on photocatalytic degradation of tetracycline using Ag3PO4/AgBr under the visible light irradiation. d The effect of g-C3N4 content on photocatalytic degradation of tetracycline using Ag3PO4/AgBr/g-C3N4 under visible light irradiation (Reprinted with permission of Elsevier from Yu et al. 2020). e, electron; h+, hole; O2 · −, superoxide radical; ·OH hydroxyl radical. TC, tetracycline

The 2D/2D/2D configuration ternary heterojunction synthesized using g-C3N4, 15wt.% CoAl-layered double hydroxide, and 1wt.% reduced graphene oxide (CoAl-LDH/g-C3N4/RGO) (Jo and Tonda 2019) also exhibited a high value of figure of merit, correlative quantum efficiency, and space time yield of 3.30 × 10–11 mol·L/(g·J·h)), 2.35 × 10–6 molecules/photon, and 4.71 × 10–5 molecules/(g·photon), respectively. Almost 99.0% of tetracycline with initial concentration of 20 mg/L was decomposed from antibiotic-contaminated sewage by 0.25 g/L CoAl-LDH/g-C3N4/RGO photocatalyst within 60 min of visible light illumination. Excellent photodegradation performance of CoAl-LDH/g-C3N4/RGO was attributed to two factors: (i) the notable 2D/2D/2D arrangement structure increased the synergy effect of the g-C3N4, layered double hydroxide, and reduced graphene oxide, thus accelerating the interfacial charge migration direct and separation of photogenerated electron–hole pairs; (ii) Enhanced light response in the visible range of RGO and intimate interactions between the constituents. In summary, the outstanding photocatalytic performance of heterojunction catalysts against tetracycline is mainly attributed to the formation of a favorable heterojunction structure, which improving interfacial charge transfer and hindering the recombination of charge carriers.

On a comparative note, several heterojunction composites such as g-C3N4/Bi4O5I2, g-C3N4/BiOI, g-C3N4/Bi5O7I, In2S3/Bi2O2CO3, and g-C3N4/Ag/P3HT showed poor photodegradation ability against tetracycline under visible light (Fan et al. 2020; Huang et al. 2019; Liu et al. 2019b), and their value of figures of merit was 3.52 × 10–15, 6.64 × 10–15, 2.30 × 10–14, 3.65 × 10–14, and 4.16 × 10–14 mol·L/(g·J·h)), respectively. The unsatisfactory photocatalytic performance of these composites could be explained by some reasons as followings: (i) these photocatalysts had deficiency in at least one of parameters in Eq. (1) or/ and Eq. (3) (Anwer et al. 2019); (ii) the formation of type I heterojunction in g-C3N4/Bi4O5I2 and g-C3N4/BiOI accelerated the recombination of charge carriers and also decreased oxidation and reduction potential of the whole composite; (iii) relatively lower specific surface areas of g-C3N4/Bi4O5I2, g-C3N4/BiOI, g-C3N4/Bi5O7I, In2S3/Bi2O2CO3, and g-C3N4/Ag/P3HT compared with other photocatalytic materials in Table 2. This indicates that photochemical and photophysical properties, especially the visible light response of these photocatalytic systems should be further improved for to enhance their practical application in the future.

Besides, quantum efficiency values of excellent photocatalytic systems were suggested to be higher than 1.00 × 10–7 molecules/photon (Anwer et al. 2019). Quantum efficiency values of heterojunctioned systems are listed in Table 2 generally between 4.01 × 10–9 and 1.25 × 10–5 molecules/photon. Note that most systems had a corresponding quantum efficiency value higher than 10–7, which due to the fact that heterojunction systems generally pose better photogenerated charge-pair separation and light-harvesting properties.

A 3D architecture organic heterojunction (PANI/PDI) developed by combining polyaniline and perylene diimide exhibited the highest quantum efficiency and space time yield (4.99 × 10–4 molecules/(g·photon)) under visible light (Fig. 15) (Dai et al. 2020). Therefore, the 3D PANI/PDI per mass unit could effectively utilize a unit photon to decompose more tetracycline molecules than other photocatalysts. PANI/PDI showed excellent visible light harvest because PDI and PANI have strong absorption ability at about 713 nm and the whole UV–Vis light region, respectively (Jiang et al. 2016). Such outstanding photocatalytic ability of the 3D PANI/PDI system was due to: (i) the formation of energy-matched heterojunction and larger delocalized π–electron conjugated system caused by π–π bonds between PANI and PDI, which restrained the recombination of e-h+ pairs; (ii) structural superiority of 3D framework providing more reactive sites and charge transport channels; (iii) coupling with PANI polymer backbone enhances the strength of the PDI hydrogel, thereby increasing the catalytic stability of the whole composite. PANI/PDI might be one of the promising candidates as visible-light-driven photocatalytic system for environmental applications. Therefore, future studies should explore such photocatalytic materials with high visible light response.

Fig. 15
figure 15

a Morphologic structure of PANI/PDI. b Photocatalysis mechanism of PANI/PDI heterojunction photocatalysts under visible light irradiation: direct Z-scheme heterojunction mechanism. c and d Scanning electron microscopy images of 3D PANI/PDI: PDI were uniformly dispersed in the 3D network structure of PANI (Reprinted with permission of Elsevier from Dai et al. 2020). PANI, Polyaniline; PDI, Perylene diimide. 3D, three-dimension

A novel type II heterojunction composite Bi2W2O9/g-C3N4 also showed high quantum efficiency of 6.45 × 10–6 molecules/photon, space time yield of 3.23 × 10–5 molecules/(g·photon), and removal efficiency of 95% (3.8 and 2.6 times higher than that of the pristine Bi2W2O9 and g-C3N4 catalysts) after 90 min of simulated sunlight irradiation (Fig. 16) (Obregón et al. 2020). The enhanced photocatalytic performance of Bi2W2O9/g-C3N4 mainly attributed to the establishment of type II heterojunction caused by the difference of conduction band and valence band of both Bi2W2O9 and g-C3N4 semiconductors, which accelerated the effective separation of photoexcited charge carriers. Also, Bi2W2O9/g-C3N4 (4.3 m2/g) had a larger surface area than that of pure Bi2W2O9 (1.5 m2/g) and a lower band gap value (2.70 eV) than that of pure Bi2W2O9 (2.91 eV), enhancing the photocatalytic degradation of tetracycline. It is widely suggested that photocatalysis performance of photocatalysts is influenced by their textural, morphological, and crystalline properties such as surface area, crystalline size, porosity, pore size, crystallinity.

Fig. 16
figure 16

a Photocatalytic removal of tetracycline using Bi2W2O9, g-C3N4 and Bi2W2O9/g-C3N4 under sunlight illumination. b Charge-carrier separation process for Bi2W2O9/g-C3N4: different positions of the valence and conduction bands of Bi2W2O9 and g-C3N4 built a type II heterojunction system. (Reprinted with permission of Elsevier from Obregón et al. 2020). e, electron; h+, hole; O2 · −, superoxide radical; ·OH hydroxyl radical; Eg, energy gap

ZnO/GO/Ag3PO4 ternary heterojunction system, one of ZnO-based representative heterojunction, was constructed through a facile ultrasonic-assisted precipitation method and exhibited superior performance for tetracycline removal (quantum efficiency = 3.17 × 10–6 molecules/photon, and space time yield = 6.34 × 10–5 molecules/(g·photon)) (Zhu et al. 2020). ZnO/GO/Ag3PO4 system had higher tetracycline removal efficiency (96.32%) than ZnO (51.33%), Ag3PO4 (48.35%), GO/Ag3PO4 (80.61%), ZnO/GO (79.60%), and ZnO/Ag3PO4 (70%) after 75 min of visible irradiation. Enhanced adsorption ability, accelerated electron transfer, increased number of active species, and effective separation of photoproduced charge carriers enhanced photocatalytic activity of ZnO/GO/Ag3PO4.

The heterojunction systems in Table 2 with poor photocatalytic performance and quantum efficiency less than 1.00 × 10–7 molecules/photon showed a relatively low degradation efficiency < 80% for tetracycline antibiotic, except for Ag-C3N4/SnS2 (≈ 90%). For instance, a noble-metal-free p–n heterojunction CuBi2O4/MoS2 displayed a low quantum efficiency of 2.26 × 10–8 molecules/photon and removal ratio of 76.0% after 120 min of visible light illumination (Guo et al. 2019b). Bi4NbO8Cl/g-C3N4 nanosheets with {001} exposing facets prepared via a molten-salt growth method degraded 78.0% of tetracycline after 180 min of photocatalysis with a quantum efficiency of 7.72 × 10–8 molecules/photon (Xu et al. 2020b). Among three performance metrics, quantum efficiency is an essential quantitative measurement tool for performance of photocatalysts. The photocatalytic system with low quantum efficiency cannot exhibit high value of space time yield and figures of merit, indicating poor practical application potential.

Conclusion

Over the past few decades, tetracycline antibiotic is one of the most common and persistent environmental pollutants, threatening ecological and human health worldwide. Recent research has demonstrated that heterojunction photocatalysis is an ideal and promising method for improving the separation/migration of electron–hole pair and remediating tetracycline-contaminated aquatic environments. However, the particular application of heterojunction semiconductor photocatalysts still suffers from several significant limitations. Therefore, future research should address the following bottlenecks.

Firstly, performance evaluation is an objective option to screen excellent photocatalytic materials with outstanding photo-activity. A few studies have about systematically assessed and compared the photocatalytic performance of different catalysts during photocatalytic degradation. Besides, only quantum efficiency, space time yield, and figures of merit have been used as evaluation metrics, ignoring thermodynamical indices, such as temperature that can greatly influence the rate of photocatalytic degradation. Therefore, more inferred numeric metrics, performance evaluation methods, formulae and mathematical models should be exhibited in the future for an advanced photocatalytic performance assessment. Secondly, the toxic effects of the intermediates formed during photocatalytic degradation of tetracycline antibiotic are unknown. Therefore, toxicology assessment of tetracycline degradation intermediates is necessary. Future studies should also focus on accelerating the decomposition of tetracycline and other drugs into CO2, H2O, and other small molecular inorganics to decrease the production of toxic intermediates. Additionally, the plausible mechanism involving the transfer of photogenerated electrons and holes in heterojunction systems composed of more than two semiconductor photocatalysts remains unknown and requires relevant reflection and studies.