Recent advancement on photocatalytic plastic upcycling

More than 8 billion tons of plastics have been generated since 1950. About 80% of these plastics have been dumped in landfills or went into natural environments, resulting in ever-worsening contamination. Among various strategies for waste plastics processing (e.g., incineration, mechanical recycling, thermochemical conversion and electrocatalytic/photocatalytic techniques), photocatalysis stands out as a cost-effective, environmentally benign and clean technique to upcycle plastic waste at ambient temperature and pressure using solar light. The mild reaction conditions for photocatalysis enable the highly selective conversion of plastic waste into targeted value-added chemicals/fuels. Here, we for the first time summarize the recent development of photocatalytic plastic upcycling based on the chemical composition of photocatalysts (e.g., metal oxides, metal sulfides, non-metals and composites). The pros and cons of various photocatalysts have been critically discussed and summarized. At last, the future challenges and opportunities in this area are presented in this review.


Introduction
Up till now, more than 8 billion tons of plastics have been synthesized, in which less than 20% of them are incinerated or recycled.Around 80% of used plastics are accumulated in the natural environment or ended up in landlls.  Rectly, various dealing strategies have been developed for treating plastic wastes, such as incineration, mechanical recycling, thermochemical conversion and electrocatalytic/photocatalytic techniques.  Amo them, large-scale incineration and landlls are the two most general routes because they are inexpensive, facile and adaptable to various feedstock.  Unfunately, these two traditional routes are susceptible to deleterious environmental impacts and negligible added values.  Lowalue products are manufactured by mechanical recycling, which down-cycles plastic wastes into low value products and is also restricted to single-component and few types of clean thermoplastics.In contrast, upcycling conversion of plastic wastes, e.g., thermocatalytic/ photocatalytic/electrocatalytic techniques, has attracted significant attention recently, since value-added chemicals/fuels or materials with extra economic value can be acquired via these appealing techniques.Among these techniques, photocatalysis stands out as a cost-effective, environmentally-benign and green strategy able to upcycle plastic wastes at ambient temperature and pressure utilizing renewable sunlight.A photocatalysis reaction conducted in mild conditions is anticipated to accurately activate target chemical bonds, while reserving the other functional groups, thus realizing high selectivity to desirable products.[34][35][36][37][38][39][40][41][42][43][44][45][46][47][48][49] Recently, various photocatalysts, such as metal oxides (e.g., Pt loaded P25 TiO 2 (ref.34 and 35) and Co doped Ga 2 O 3 Dr Jingrun Ran has received his PhD degree in Chemical Engineering from the University of Adelaide under the supervision of Professor Shi-Zhang Qiao.Currently he is appointed as a Senior Lecturer in the School of Chemical Engineering in the University of Adelaide and working in Professor Qiao's group.In 2020-2023, he has been recognised as a Clarivate Highly Cited Researcher. In 223, he was awarded the ARC Future Fellowship.His current research is focused on the atomic-scale design and synthesis of photocatalysts for producing energy fuels and value-added chemicals using renewable solar energy.Amin Talebian-Kiakalaieh received his bachelor's degree in petroleum engineering in 2007.Then he changed his major to Chemical Engineering and received his master's degree in chemical engineering from Universiti Teknologi Malaysia (UTM) in 2012.He joined the Prof. Shi-Zhang Qiao group in 2021 as a PhD candidate under the supervision of Dr Jingrun Ran at the University of Adelaide.Currently he is working on synthesizing advanced photocatalysts for renewable energy applications.
In this review, we for the rst time summarize and review recently reported photocatalysts for plastic upcycling based on their chemical compositions including metal oxides, metal sulphides, non-metals and composites.The accurate advantages/disadvantages of photocatalysts are critically analysed and discussed according to their chemical compositions.Especially, the perceptive reaction mechanisms for various photocatalysts under varied reaction conditions are also introduced and summarized in this review.

Fundamentals of photocatalytic plastic upcycling
Owing to the ultrastability of most plastics, such as polyolens and polyesters, it is very challenging to directly convert these plastics into valuable chemicals via photocatalysis at ambient temperature and pressure using sunlight only.40][41][42][43]49 Additionally, polyolens, e.g., polyethylene (PE), are usually hydrothermally treated in nitric acid aqueous solution, to be transformed into various short-chain carboxylic acids, e.g., succinic acid, glutaric acid, acetic acid, adipic acid and propanoic acid. 34,3946]48 In some of these cases, metal oxide photocatalysts with strongly oxidative photo-induced holes are applied. 36,37,46In the other case, an air/O 2 atmosphere and raised temperature/pressure are adopted to boost the oxidative cleavage of the strong C-C/C-H bonds in these untreated plastics. 44Another key reaction condition for photocatalytic plastic upcycling is the existence/absence of oxygen (O 2 ) in the reaction system.If photocatalytic plastic upcycling is conducted under aerobic conditions, abundant reactive oxygen species (ROS), such as cO 2 − , cHO 2 − and cOH radicals, will be produced by the following reactions: 36,37,44,[46][47][48] e − + O 2 / cO 2 − (1) 7][48] But O 2 in the reaction system can compete with H + /H 2 O to obtain photo-induced electrons, resulting in a much lower H 2 yield under aerobic conditions compared to that under anaerobic conditions.In contrast, if photocatalytic plastic upcycling is performed under anaerobic conditions, these ROSs are much more challenging to form in the reaction system unless the photo-induced holes with sufficient oxidation ability can oxidize OH − /H 2 O to yield cOH radicals.Without the competition from O 2 , many more photoexcited electrons can be utilized for H 2 evolution, resulting in usually a much higher H 2 evolution rate under anaerobic conditions compared to that under aerobic conditions.

Plastic upcycling with pre-treated plastics and under anaerobic conditions
As the most reported reaction condition for plastic upcycling, 34,[38][39][40][41][42][43]49 the reaction mechanism is summarized in Fig. 1a: (i) monomers/oligomers/short-chain carbon-based molecules/surface oxygenated plastics are rst generated from the pre-treatment of various plastics; (ii) under anaerobic conditions, photo-excited electrons mainly involve in H 2 evolution, thus resulting in a relatively high photocatalytic H 2 evolution rate under these reaction conditions; (iii) photoinduced holes can oxidize the formed monomers/oligomers/ short-chain carbon-based molecules to yield value-added chemicals/fuels; (iv) in some cases, if metal oxide photocatalysts with strongly oxidative holes are utilized, these monomers/oligomers/short-chain carbon-based molecules could be over-oxidized to yield CO 2 , which could be further reduced by photo-generated electrons to form CO.

Plastic upcycling with untreated plastics and under aerobic conditions
Another oen-reported reaction condition is plastic upcycling with untreated plastics and under aerobic conditions.The reaction mechanism is shown in Fig. 1b: (i) plastic powder is directly added into the reaction system; (ii) photocatalysts with strongly oxidative photo-induced holes, e.g., Nb 2 O 5 atomic layers, 36 Co doped Ga 2 O 3 (ref.37) and ZnO coupled UiO66-NH 2 , 46 are usually adopted in this reaction system, which can oxidize

Plastic upcycling with untreated plastic and under anaerobic conditions
Photocatalytic plastic upcycling under these reaction conditions is rarely reported. 45And the reaction mechanism is not very clear.So we propose a possible mechanism in Fig. 1c as follows: (i) for the reduction reaction, under anaerobic conditions, photo-excited electrons mainly involve in H 2 evolution via reducing H + /H 2 O in the reaction system.In some cases, CO 2 could be generated due to the overoxidation of untreated plastic by strongly oxidative cOH and holes.Thus, generated CO 2 can be further reduced by photo-excited electrons to produce some carbon-based fuels/chemicals; (ii) for the oxidation reaction, the photo-induced holes with sufficient oxidation ability can react with the H 2 O molecule to generate strongly oxidative cOH.Thus, cOH can react with untreated plastics to generate organic chemicals and CO 2 .In contrast, photo-excited holes with weak oxidative capacity can hardly cleave the C-C bond in untreated plastics to produce organic chemicals.And no CO 2 would be generated under these conditions.Overall, photocatalytic plastic upcycling under these reaction conditions is rarely reported and more studies under these reaction conditions should be conducted to understand the reaction mechanism.

Reaction thermodynamics and kinetics for plastic upcycling
Reaction thermodynamics and kinetics for photocatalytic upcycling of plastic wastes into fuels/valuable chemicals are complex.Nevertheless, compared to photocatalytic water splitting with a theoretical thermodynamic requirement of 1.23 eV (DG 0 = +237 kJ mol −1 ), the thermodynamic requirement for photocatalytic plastic upcycling is usually much lower.This is because the oxidation of plastics, especially for small molecules from pre-treated plastics, is much more facile than water oxidation for O 2 evolution.For instance, reforming ethylene glycol and lactic acid, from pre-treated PET and PLA, shows the Gibbs free energy changes of +9.2 and +27 kJ mol −1 , respectively.These are much lower than that of water splitting (G 0 = +237 kJ mol −1 ).Thus, the much less thermodynamics requirement for plastic oxidation compared to water splitting enables the utilization of a semiconductor photocatalyst with a smaller band gap, which exhibits a much broader light absorption range for the solar spectrum.Thus, nitride, sulphide, phosphate, arsenide, selenide and even tellurite catalysts, which show decient valence band potential for water oxidation, could be suitable for photocatalytic plastic oxidation, especially for small molecules from pre-treated plastics.Thus, abundant and undesirable plastic wastes serve as an efficient hole scavenger to increase the electron-hole separation/transfer efficiency, thus boosting the reduction reaction (H 2 evolution or CO 2 reduction).Simultaneously, value-added chemicals/fuels can be generated via plastic oxidation.Reaction kinetics in photocatalytic plastic upcycling are very complex and are affected by many factors, such as plastic type, pre-treatment route, reaction solution and reaction atmosphere.First, the types of plastics obviously affect the reaction kinetics.Polyolens (e.g., PE, PP, PS and PVC), which account for 57% of the total plastics, consist of inert C-C and C-H bonds with high dissociation energies.Thus, reaction kinetics for breaking these C-C/C-H bonds and upcycling polyolens are very sluggish.Besides, the hydrophobic nature of polyolens makes their dispersion in aqueous solution and adsorption on photocatalysts very challenging.In contrast, polyesters/polyamides (e.g., PET, PLA, PUR) with ester/amide bonds are much easier to decompose to yield the corresponding monomers, which are facilely adsorbed on photocatalysts for further forming value-added chemicals.
Second, the pre-treatment route also greatly affects the reaction kinetics.Without pre-treatment, the reaction kinetics for upcycling most plastics are rather slow, especially for inert polyolens.The hydrothermal pre-treatment of polyolens (e.g., PE) in nitric acid could cleave the C-C bond and convert most of them into various carboxylic acids (e.g., succinic acid, glutaric acid, acetic acid, adipic acid and propanoic acid).These shortchain water-soluble carboxylic acids could be more easily adsorbed onto photocatalysts and further yield value-added chemicals/fuels more efficiently.Furthermore, the hydrolysis pre-treatment of polyesters/polyamides (e.g., PET and PUR) in alkaline solution at elevated temperature could yield the corresponding monomers, which can be easily dissolved in aqueous solution and adsorbed on photocatalysts for more efficient conversion into valuable chemicals/fuels.Moreover, the plasma treatment of plastics at room temperature and atmosphere could gra oxygen-containing functionalities on the PE/PP/PVC surface, leading to increased hydrophilicity in water and more intimate interaction with photocatalysts.Thus, reaction kinetics for yielding fuels/valuable chemicals from pretreated PE/PP/PVC is improved to some extent.Nevertheless, since this plasma treatment doesn't convert PE/PP/PVC into short-chain monomers/oligomers, the enhancement of reaction kinetics for plastic upcycling is limited.
Third, different reaction solutions (e.g., pure water, alkaline aqueous solution and organic solvent) also obviously affect the reaction kinetics for plastic upcycling.As pure water is applied as the reaction solution, the reaction kinetics for upcycling most plastics in pure water are very slow.This is because that most of the plastics are hydrophobic, which oat on the surface of pure water or precipitate at the bottom rather than suspend in the water, making their adsorption or interaction with photocatalysts challenging.Besides, the undissolved plastics also hinder light absorption by photocatalysts to some extent.In contrast, as alkaline aqueous solution (e.g., NaOH and KOH) is applied as the reaction solution, some kinds of plastics, such as polyesters/polyamides, could be dissolved/hydrolysed in alkaline aqueous solution, yielding monomers easily adsorbed on photocatalysts for upcycling reactions.Thus, the corresponding reaction kinetics for these plastics are obviously improved.Nevertheless, this alkaline environment could corrode the photocatalysts and reduce their activity/selectivity/stability.As organic solvents (e.g., acetonitrile, tetrahydrofuran, cyclohexane and toluene) are utilized as the reaction solution, hydrophobic plastics (e.g., PE, PP, PS and PVC) can be dissolved in these organic solvents with stirring and elevated temperature.These will boost their adsorption on photocatalysts and further transformation into valuable chemicals/fuels.The activity/selectivity/ stability of photocatalysts will also be affected.Moreover, byproducts from these organic solvents might also be generated.
Fourth, the reaction atmosphere considerably inuences the reaction kinetics for plastic upcycling.Under aerobic conditions, various ROS species (e.g., cO 2 − , 1 O 2 , cHO 2 − , cOH and H 2 O 2 ) could be generated to greatly boost the oxidation and conversion of plastics into value-added chemicals/fuels.Nevertheless, strongly oxidative and non-selective ROS could overoxidize these plastics and generate CO 2 .In contrast, under anaerobic conditions, only photo-excited holes (and cOH) involve in the oxidation and conversion of plastics into valuable chemicals/fuels.Thus, it is much easier to regulate the oxidation ability of photo-excited holes for controlling the selectivity of plastic upcycling.Nevertheless, due to the lack of massive ROS species for oxidation of plastics, the reaction kinetics of plastic upcycling is reduced.

Photocatalysts for plastic upcycling
6][47][48][49] We will introduce the research in these elds according to the above four categories.

Metal oxide based photocatalysts for plastic upcycling
Various metal oxides, such as TiO 2 , 34,35  ) can rst oxidize polyolens to form hydroxyl, carbonyl and carboxyl functionalities.Then, the strong C-C/C-H bonds of these polyolens will be polarized and signicantly weakened, resulting in the much easier cleaving of these C-C/C-H bonds and upcycling of these polyolens into value-added chemicals/fuels.
First, we will introduce two studies using the most extensively studied metal oxide photocatalyst, TiO 2 , for photocatalytic plastic upcycling under anaerobic conditions with pre-treated polyolens. 34,35Both these studies adopt commercial P25 TiO 2 loaded with the Pt cocatalyst.The Reisner 34 group rst used 6% nitric acid and a hydrothermal reaction at 180 °C for 4 hours to convert ∼40% polyethylene (PE) into a variety of liquid chemicals including succinic acid (44%), glutaric acid (22%), acetic acid (21%), adipic (9%) and propanoic acid (4%).Succinic acid and glutaric acid are identied as the major products from PE conversion (Fig. 2a).Then, they synthesized 1 wt% Pt nanoparticle (NP) loaded P25 TiO 2 using a chemical reduction route.Aer a 96 hour reaction using the PE decomposition solutions, the as-synthesized 1 wt% Pt NP loaded P25 TiO 2 exhibits photocatalytic performance for the evolution of ethylene (0.017 mmol g −1 ), ethane (0.25 mmol g −1 ), propylene (0.007 mmol g −1 ), propane (0.14 mmol g −1 ), H 2 (6.3 mmol g −1 ) and CO 2 (5.9 mmol g −1 ).Thus, ethane and propane are detected as the major alkane products from photoreforming of PE decomposition solution (Fig. 2a).They found that only a small amount of ethylene is generated in this reaction, attributed to the efficient transfer of adsorbed hydrogen to the intermediate radical on 1 wt% Pt NP loaded P25 TiO 2 .Some of the generated CO 2 and H 2 arise from the decarboxylation reaction.Besides, due to the strong oxidation abilities of photogenerated holes in TiO 2 , mineralization of the acquired chemicals (e.g., succinic and glutaric acid) aer pretreatment can occur, also leading to CO 2 and H 2 evolution. 13C-labelled succinic acid using 1 wt% Pt NP loaded TiO 2 via 1 H-nuclear magnetic resonance ( 1 H NMR) spectroscopy conrms that the evolved ethane is generated from the succinic acid.They found that without Pt as the co-catalyst, larger amounts of ethylene and smaller amounts of ethane are observed.Ethane became the major product again as MoS 2 was adopted as the co-catalyst.They have also synthesized Pt loaded cyanamide-regulated carbon nitride powder ( NCN CN x -jPt).Aer a 72 hour photocatalytic reaction, the major alkane products from P25jPt and NCN CN x jPt are ethane at 56.3 and 7.2 mmol g −1 h −1 (Fig. 2b and c).Then, they have set up a ow photocatalytic reactor system (Fig. 2d), in which a continuing generation of ethane and propane was achieved for both P25jPt and NCN CN x jPt (Fig. 2e and f), together with the constant production of ethylene and propylene for NCN CN x jPt (Fig. 2f).In another study, a plasma pre-treatment strategy was reported to treat polyolens to partially cleave the C-C/C-H bonds and generate oxygenated functional groups on the backbones of PE, PP or PVC. 35FTIR and XPS spectra together conrm the formation of hydroxyl, carboxyl and carbonyl functionalities on the surface of PE.Contact angle measurement further reveals the gradually reduced contact angle of PE with water as plasma treatment time increases, suggesting the increased hydrophilicity of treated PE.This will lead to better dispersion of treated PE in aqueous solution and more intimate contact between the photocatalyst and treated PE.Molecular dynamics computations reveal a stronger interaction between the TiO 2 surface and plasma treated PE compared to that between the TiO 2 surface and untreated PE.
Pt NPs with sizes of 5-15 nm are loaded on P25 TiO 2 via photo-deposition to synthesize a Pt-TiO 2 photocatalyst.Photocatalytic experiments show that H 2 evolution rst rises and then decreases with increasing plasma treatment time on PE using Pt-TiO 2 .This is because the rst activity increase arises from the generation of abundant -OH functionalities on the surface of PE.The subsequent activity decrease is due to the formation of massive carbonyl and carboxyl functionalities on treated PE.Reforming of PE also generates CH 4 , C 2 H 4 and C 2 H 6 .But no liquid products were found by 1 H NMR for short-/ long-term photocatalysis tests.This plasma treatment effect also increased the H 2 evolution activity on treated PP or PVC compared to untreated PP or PVC, suggesting the universality of this route.When comparing the above two studies using the same photocatalyst (Pt loaded P25 TiO 2 ) under similar reaction conditions (Table 1), we can nd that the H 2 /C 2 H 4 /C 2 H 6 /CO 2 evolution rates of PE upcycling using the hydrothermal pretreatment are much higher than those of PE upcycling using the plasma pre-treatment.These are easy to understand Fig. 2 (a) Hydrothermal pre-treatment of PE to form dicarboxylic acid (i) followed by conversion into gaseous hydrocarbon using photocatalysis to yield alkanes (ii) or electrolysis to yield alkenes (iii).(b) Photocatalytic reforming of succinic acid in 0.1 M HNO 3 using (b) P25jPt or (c) NCN CN x jPt.Reaction conditions: AM 1.5G (100 mW cm −2 ), 25 °C, 2 ml of 10 mg ml −1 succinic acid in 0.1 M HNO 3 (pH set to be 4), and 2 mg ml −1 photocatalyst.(d) Image of the photocatalytic flow setup.The pre-treated PE solution (not displayed in this image) in the reservoir is continuously pumped with a peristaltic pump into the photoreactor with an irradiation area of 25 cm 2 .Then, the solution is pumped into the reservoir again; the generated gaseous products are sampled and studied by gas chromatography.Photocatalytic product generation using a flow setup with (e) P25jPt and (f) NCNCN x jPt.Reproduced with permission from copyright 2021, American Chemical Society. 34 (ref.37), respectively, for directly photocatalytic upcycling of polyolens and PET in an air atmosphere without pre-treatment.In one study, 36 the authors have designed (258.9 mmol g −1 h −1 ), 24 h reaction a general strategy of converting different plastic wastes (PE, PP and PVC) into CO 2 followed by photo-reduction to form acetic acid as a C 2 fuel under simulated natural environmental conditions (Fig. 3a).First, they designed and synthesized Nb 2 O 5 atomic layers using the as-synthesized niobic acid atomic layers as the precursor followed by annealing in air.The earthabundant and robust Nb 2 O 5 is chosen due to its suitable conduction band (CB) and valence band (VB) positions (+2.5 V vs. SHE for the CB and −0.9 V vs. SHE for the VB at pH = 7).Thus, Nb 2 O 5 can generate highly oxidative cOH radicals (+2.32 V vs. SHE at pH = 7) to degrade plastics and photo-generated electrons to reduce CO 2 (−0.6 V vs. SHE at pH = 7).Nb 2 O 5 atomic layers can degrade PE, PP and PVC with identical numbers of carbon in 40, 60 and 90 hours, respectively.The generated CO 2 amounts increase gradually and reach the highest values in the corresponding time (Fig. 3b).They found that the overall numbers of moles of carbon in the generated CO 2 gas and CO 2 dissolved in solution are almost equivalent to that in pure PE, PP or PVC.These results conrm that plastics are completely degraded to form CO 2 gas.Furthermore, the generated CH 3 COOH amounts are also gradually increased (Fig. 3c) and averaged CH 3 COOH generation rates on PE, PP and PVC are ∼47.4,40.6 and 39.5 mg g −1 h −1 , respectively (Fig. 3d).

Chemical
To obtain insightful understanding of the reaction mechanism on photocatalytic conversion of plastics into CH  37 has synthesized Co doped Ga 2 O 3 nanosheets (Co-Ga 2 O 3 ) for photocatalytic conversion of pulverized powder from PE bags, PP boxes or PET bottles into syngas (CO and H 2 ) along with CO 2 in the presence of water under ambient conditions.First, they reveal the CB and VB edge positions for both Co-Ga 2 O 3 and Ga 2 O 3 , conrming that both of their photo-induced electrons and holes can drive some pivotal reactions, e.g., H 2 O oxidation or CO 2 /O 2 /H 2 O reduction.Then PE bags, PP boxes and PET plastic bottles were crushed into powders using a pulveriser.Aerwards, Co-Ga 2 O 3 or Ga 2 O 3 was utilized to convert these powders in pure water with simulated solar light (AM 1.5G, 100   1.

Metal sulphide based photocatalysts for plastic upcycling
9][40] Thus, they usually exhibit outstanding H 2 evolution rates under visible light irradiation under anaerobic conditions when applied on photocatalytic plastic upcycling.On the other hand, these metal sulphides possess moderate oxidation abilities owing to the S 3p derived VB.Thus, their photo-induced holes can oxidize the pre-treated plastics without over-oxidation to form CO 2 under anaerobic conditions.However, owing to their moderate oxidation abilities, it is challenging for metal sulphide based photocatalysts to directly upcycle the plastics without pre-treatment.9][40] Several strategies, such as surface engineering, 38 loading cocatalysts 39 and band structure engineering, 40 have been adopted to increase the photocatalytic efficiencies.The Reisner 38 group has presented a strategy to photoreform plastic wastes to yield H 2 and value-added organics using photocatalysts in water under sunlight (Fig. 4a).They have designed and synthesized a CdS/CdO x quantum dot (QD) photocatalyst.They found that when CdS QDs were added into aqueous NaOH, a thin cadmium oxide/hydroxide (CdO x ) is generated to impede photo-corrosion.Ligand-free QDs are found to work with most substrates due to their exposed surfaces.In comparison, oleic acid-capped QDs can only work with PET, probably owing to the hydrophobic effect beneting the substrate-QD interaction.First, CdS/CdO x was adopted for photocatalytic reforming of various polymers including polylactic acid (PLA), PET, polyurethane (PUR), polyvinylpyrrolidone (PVP), polyethylene glycol (PEG), LDPE, PVC, poly(methyl methacrylate) (PMMA), polystyrene (PS) and polycarbonate (PC), respectively.Among them, only PLA, PET and PUR are found to achieve higher photocatalytic H 2 evolution performances while much less photocatalytic H 2 evolution is observed on the other substrates.So these three polymers are selected for photocatalytic reforming.CdS/CdO x QDs were used for photocatalytic reforming of PLA, PET and PUR in NaOH aqueous solution in a N 2 atmosphere.With optimised reaction conditions, CdS/CdO x QDs exhibit the photocatalytic H 2 evolution activities of 64.3 ± 14.7, 3.42 ± 0.87 and 0.85 ± 0.28 mmol g −1 h −1 , respectively.Isotope-labelling experiments reveal that the H 2 produced arises from water, not the substrate.As a comparison, 5% Pt loaded TiO 2 only exhibits the H 2 evolution rates of 0.011 ± 0.004 and 0.074 ± 0.029 mmol g −1 h −1 under identical reaction conditions.And without expensive Pt as the co-catalyst, bare TiO 2 exhibits no H 2 evolution.Besides, ZnSe QDs as a Cd-free catalyst show no H 2 production under the same reaction conditions.These highlight the advantages of strong visible-light absorption, no use of co-catalysts and fast oxidation of complicated substrates for CdS/CdO x QDs.To further increase the rate, a pre-treatment route was developed to hydrolyse PET, PUR and PLA in 10 M NaOH aqueous solution for 24 hours at 40 °C to release the monomers.Aer pre-treatment and removing the undissolved polymer by centrifugation, the supernatant was used to reduce the absorbance and scattering, thus leading to more photons absorbed by CdS/CdO x QDs and higher rates.Pre-treated PET and PUR solutions were found to obviously increase the activities of CdS/CdO x QDs, compared to Raw PET and PUR, respectively (Fig. 4b).Compared with Raw PLA, pre-treated PLA exhibits almost no inuence on the efficiency of CdS/CdO x QDs (Fig. 4b).This is because PLA is easily dissolved in NaOH aqueous solution.Another advantage of CdS/CdO x QDs is that they can function even in highly alkaline solution.Then, 1 H-NMR spectroscopy is adopted to study the reaction solutions and organic oxidation products.It was found that PLA is rst hydrolysed to form sodium lactate followed by oxidation to generate a pyruvate-based compound (Fig. 4c).As for PET, it is rst hydrolysed to generate terephthalate, ethylene glycol and isophthalate, followed by the formation of formate, glycolate, ethanol, acetate and lactate (Fig. 4d).Photo-reforming terephthalic acid doesn't generate H 2 , indicating that only the aliphatic component of PET yields the oxidation products.And terephthalic precipitates as a disodium salt to be easily recovered as a valuable chemical.As shown in Fig. 4e, PUR is hydrolysed to form an aliphatic component (propylene glycol) and aromatic component (2,6-diaminotoluene).While propylene glycol is oxidized to form formate, acetate, pyruvate and lactate (Fig. 4e), 2,6-diaminotoleune remains intact.The overall conversion of all polymers is lower than 40%, since CdS/CdO x QDs cannot completely mineralise these polymers into CO 2 .And no CO 3 2− or CO 2 is detected.These polymers are just partially oxidized to form various chemicals.CdS/CdO x QDs are also shown to achieve photo-reforming of a PET bottle and pre-treated PET bottle to generate H 2 and value-added chemicals (Fig. 4f).
In another study, a cocatalyst/photocatalyst MoS 2 /CdS system was synthesized for photocatalytic upcycling of pretreated polyester/polyolen. 39 The Qiu group 39 designed this system, MoS 2 -tipped CdS nanorod (MoS 2 /CdS), to reform the pre-treated PE, PLA or PET to generate various value-added chemicals and H 2 , as shown in Fig. 5a.The as-synthesized MoS 2 /CdS composite exhibits the accumulation of photogenerated electrons and holes on the MoS 2 tip and sidewalls of CdS, respectively (Fig. 5a).The TEM image (Fig. 5b and c) and HRTEM image (Fig. 5d) directly reveal the intimate coupling of MoS 2 on the tip of CdS NSs.The strong electronic coupling between MoS 2 and CdS is conrmed by the surface-sensitive high-resolution XPS technique, showing that electrons are transferred from CdS to MoS 2 .Results of steady-state PL, photocurrent response and electrochemical impedance spectroscopy (EIS) all conrm the more efficient charge separation/ migration in MoS 2 /CdS.Furthermore, selective deposition reactions indicate that MnO x nanosheets and Pt nanoparticles are selectively loaded on the sidewalls of CdS and the MoS 2 tip, respectively.These results also indicate that photo-generated electrons and holes are accumulated on the MoS 2 tip and sidewalls of CdS, respectively.Linear sweep voltammetry (LSV) curves further indicate the signicantly higher HER activity of MoS 2 /CdS compared to MoS 2 alone.Then, the electron spin resonance (ESR) technique is further utilized to detect the active species on MoS 2 /CdS, which conrms the existence of photo-     5g). 13C NMR spectroscopy and the self-built route together conrm the production of CO  5h).The 1 H NMR test shows that the concentration of terephthalate is not changed but the concentration of ethylene glycol is altered, suggesting that these formed carboxylate chemicals arise from ethylene glycol but not from terephthalate.
Holes were conrmed to be the major active species for ethylene glycol oxidation via control experiments.MoS 2 /CdS was also utilized for photo-reforming of the pre-treated PE by nitric acid in a hydrothermal reaction (Fig. 5i).HPLC tests conrm the existence of formic acid, succinic acid, glutaric acid, acetic acid, propionic acid and adipic acid, with the major products of succinic acid and glutaric acid (Fig. 5i).Then, MoS 2 /CdS was utilized to photo-reform the pre-treated PE, generating a H 2 evolution rate of 1.13 ± 0.06 mmol g −1 h −1 (Fig. 5j) Chemical Science Review adopted to analyse the degradation products of photocatalytic PET conversion.Ethylene glycol, terephthalic acid and glycolate were detected in the pre-treated solutions before reaction.Finally, the pre-treated PET was oxidized to yield formate, methanol, acetate and ethanol.
The above three studies have demonstrated the immense potential of metal sulphide based photocatalysts, especially Cdbased catalysts, to upcycle these pre-treated plastics into valuable chemicals/fuels using simulated solar light.But the toxicity of Cd-based catalysts, together with the insufficient stability and oxidation capacity of metal sulphide based photocatalysts, seriously restricts their realistic applications in industrial scale solar plastic upcycling.All the performances and reaction conditions of the studies in this section are summarized in Table 2.

Non-metal based photocatalysts for plastic upcycling
2][43][44] Owing to the moderate oxidation ability of C x N y , it is very challenging for them to directly upcycle plastics into value-added chemicals/fuels at room temperature and under anaerobic conditions.the Reisner group have developed a cocatalyst/photocatalyst (Ni 2 P loaded CN x ) system to photoreform various pre-treated plastics to acquire value-added chemicals without generating CO 2 or even CO 3 2− . 41They have reported the photo-reforming of plastic wastes to produce valueadded organics and H 2 using a Ni 2 P loaded cyanamidefunctionalized carbon nitride ( H2N CN x jNi 2 P) photocatalyst (Fig. 6a).The TEM image of H2N CN x jNi 2 P (Fig. 6b) shows the loading of Ni 2 P NPs on the surface of H2N    1.6 mg ml −1 Ni 2 P/CN x , 2 ml 1 M KOH, 5 mg ml −1 pre-treated microfibers, 25 mg ml −1 pre-treated PET bottle or 25 mg ml −1 pre-treated PET bottle with 5 mg ml −1 soybean oil and simulated sunlight (AM 1.5G, 100 mW cm −2 ).(i) Image of the upscaled photocatalytic reactor.(j) Upscaled photocatalytic reforming of polyester microfibers.Reaction conditions: 1.6 mg ml −1 Ni 2 P/CN x , 120 ml 1 M KOH, 5 mg ml −1 pre-treated microfibers and simulated sunlight (AM 1.5G, 100 mW cm −2 ).Reproduced with permission from copyright 2019, American Chemical Society. 41hemical Science Review noted that while 99.1% Ni content of Ni 2 P still remains on CN x , XPS results show that at least the surface of Ni 2 P is converted to Ni(OH) 2 aer reaction in the KOH aqueous solution.Then, they further upscaled the photocatalytic reactor to 120 ml (Fig. 6i).Using this upscaled reactor, a photocatalytic H 2 evolution of 53.5 mmol g sub −1 was achieved via photo-reforming of polyester microbers in ve days (Fig. 6j).Furthermore, the Reisner group 42 has assembled a photocatalyst panel loaded with CN x -jNi 2 P to photo-reform various wastes including plastics.They utilized a simple low temperature and drop casting route to synthesize scalable photocatalyst panels with carbon nitride/ nickel phosphide (CN x /Ni 2 P).Ni 2 P nanoparticles were dispersed on CN x as the co-catalyst.First, they optimised the CN x /Ni 2 P panels for the highest H 2 evolution, light harvesting and recyclability on a scale of 1 cm 2 .Aerwards, they adopted a 25 cm 2 panel to photo-reform PET, a-cellulose and municipal solid waste (MSW), respectively.They found that the illumination conguration (front illumination or back illumination) plays the key role in the H 2 evolution activities.To demonstrate the feasibility of "real world" application of this system, it is operated in seawater and 20% sunlight irradiation (20 mW cm −2 ), and still shows ∼50% H 2 evolution activity compared to that in pure water and 1 sunlight irradiation (100 mW cm −2 ).The Yang group 43 has fabricated a graphitic carbon nitride (CN)/carbon nanotubes (CNTs)/NiMo nanoparticle (CN/CNTs/ NM) photocatalyst for photo-reforming of pre-treated PET.The intimate combination between CN and in situ generated CNTs was conrmed by the FTIR, Raman and TGA techniques.CN/CNTs/NM shows the highest activity for H 2 evolution from photo-reforming of PET.Carbon nitride (CN), carbon nitride/ carbon nanotubes (CN/CNTs) and CN/CNTs/NM were utilized to photo-reform pre-treated PET or PLA.The 1 H NMR spectrum conrms the existence of ethylene glycol (EG), terephthalate (TPA) and other small molecules aer the pre-treatment of PET.The 1 H NMR technique also conrms the formation of glyoxal and glycolate aer photo-reforming of pre-treated PET.The highest H 2 evolution was observed on CN/CNTs/NM for photoreforming of PET, about 14 times larger than that of CN.Additionally, CN/CNTs/NM exhibits higher H 2 evolution activity for photo-reforming of PLA compared to that for photoreforming of PET.CN/CNTs/NM also exhibits good robustness for H 2 evolution from photo-reforming of pre-treated PET.CN/ CNTs/NM was also adopted to photo-reform a pre-treated PET bottle, which shows slightly smaller activity compared to that of pure PET, due to the existence of many different additives in the PET bottle.Single-particle PL tests further conrm that the PL of CN/CNTs/NM is obviously quenched compared to that of CN/ CNTs or CN.The PL of CN/CNTs is also quenched compared to CN.These results indicate that the efficient electron transfer from CN to CNTs and further to NM obviously decreases the charge recombination.These are in accordance with the electrochemical impedance spectroscopy (EIS) and photocurrent density measurements.A single-particle PL study further conrms that with the addition of EG, the PL intensity of CN/ CNTs/NM is signicantly decreased because the photogenerated holes are captured by EG.Based on the above results, they propose the mechanism as follows: aer the photo-excitation, the photo-generated electrons transfer from CN to CNTs and further to NM, where photo-generated electrons reduce protons to produce H 2 gas.Photo-induced holes oxidize the EG to form glyoxal and carboxylate.
Although it is very challenging to photo-reform untreated plastics at room temperature, in aqueous solution and under anaerobic conditions using C x N y , Cao et al. 44 have successfully utilized C 3 N 4 to photocatalytically convert PS into aromatic oxygenates, such as benzoic acid, acetophenone and benzaldehyde in acetonitrile at 80-150 °C, with light irradiation and in air.They have adopted various well-known, environmentally benign and simple-to-synthesize photocatalysts including TiO 2 , ZnO, ZnS and C 3 N 4 to upcycle PS into aromatic oxygenates at 80 °C, with light illumination and in air.TiO 2 , ZnO, ZnS and C 3 N 4 show a PS conversion of 13%, 21%, 12% and 55%, 64% and 60%, respectively.TiO 2 shows the lowest selectivity of 15%, since the main products for TiO 2 are CO 2 and CO.This indicates that TiO 2 is not a suitable photocatalyst for upcycling PS.Besides, despite that both ZnO and ZnS exhibit good selectivity, their PS conversion is not good.Thus, C 3 N 4 was adopted.A range of metals, such as 0.1% Au, 0.5% Au, 0.5% Pt, 0.5% Fe and 0.5% Cu, were loaded on C 3 N 4, respectively.Although these metal loadings could increase the conversion to some extent, the selectivity was reduced, compared to C 3 N 4 without metal loading.These might be boosted by the overoxidation of intermediates/products.Then, they have investigated the photocatalytic oxidation of PS using g-C 3 N 4 .The time-evolution experimental results (Fig. 7a) exhibit an obvious induction period in the rst 3 hours, followed by the fast accumulation of various inorganic/organic products (CO x , benzoic acid, acetophenone and benzaldehyde) in 24 hours.Based on the spectroscopic result, gel permeation chromatography (GPC) and liquid product analysis, it was found that in the induction period of reaction, reactive oxygen species partially oxidize the PS.The reaction mechanism is disclosed in Fig. 7b  As shown in Fig. 7c, they can utilize this strategy to acquire pure chemicals, such as 240 mg benzoic acid, via using column chromatography separation.As presented in Fig. 7d, via appropriately regulating the ratio of substrate/catalyst (5 : 2) and reaction time (8 hours), they can acquire a stable yield rate of various organics (10 mg g −1 h −1 ) with a selectivity of 76% in 18 cycles for photocatalytic oxidative conversion of PET plastic pellets (500 mg).Furthermore, they utilized a ow reaction system to optimize the activity and selectivity towards specic products via tuning the weight hourly space velocity (WHSV).They found that better selectivity for benzaldehyde (51%) and acetophenone (31%) is acquired with an optimized WHSV of 0.9 h −1 (Fig. 7f).
The above four studies show the great potential of using C x N y based photocatalysts to upcycle pre-treated polyester/polyolen at room temperature and under anaerobic conditions or even untreated polyolen at raised temperature and in air/O 2 .Owing to the signicant advantages of earth-abundance, costeffectiveness, excellent absorption of light and suitable redox abilities, further investigation on C x N y based photocatalysts is anticipated.All the performances and reaction conditions of the studies in this section are summarized in Table 3.

Composite photocatalysts for plastic upcycling
Currently, the reported composite photocatalysts for plastic upcycling are based on two photon absorbers, both of which can absorb photons and generate photo-induced electrons/holes for upcycling plastics.The effectiveness of composite photocatalysts is determined by the junction type and internal interaction between the components: (i) the most common type II heterojunction can increase light absorption and boost charge separation, but the redox abilities are compromised; (ii) Z-scheme heterojunctions can enhance light harvesting, accelerate electron-hole dissociation and reserve strong redox capabilities, simultaneously; (iii) strong and intimate interaction between different components, which are connected by chemical bonds rather than physical bonds, can signicantly improve the charge separation efficiency and stability of overall composite photocatalysts, thus leading to greatly increased performances.7][48][49] An inorganic composite photocatalyst reported is a 0.5 wt% Pt loaded CdO x /CdS/SiC photocatalyst (Pt-CdO x /CdS/SiC). 45The as-synthesized Pt-CdO x /CdS/SiC was adopted to photo-reform various organic wastes (e.g., PE, PS, IR, cellulose, lignin, albumin and keratin) in 10 M NaOH aqueous solution at 70 °C.The photocatalytic H 2 evolution activities of 25.0, 19.4 and 36.7 mmol g −1 h −1 were observed in the presence of PE, PS and IR, respectively.It was found the raised temperature and increased basicity of the reaction system could signicantly increase the photocatalytic H 2 evolution of Pt-CdO x /CdS/SiC in the presence of a-cellulose, albumin or PE.The photocatalytic stabilities for reforming cellulose, albumin and PE were tested, respectively.Pt-CdO x /CdS/SiC exhibits good stability for reforming cellulose and albumin aer re-addition of the substrate.In comparison, Pt-CdO x /CdS/SiC shows poor stability for reforming of PE even aer re-addition of the substrate, probably because PE gels cover the photocatalysts at raised temperature.The increased photocatalytic performance of CdO x /CdS/SiC is attributed to the formation of a type II heterojunction between CdS and SiC resulting in enhanced charge separation/transfer.
Reported inorganic/organic composite photocatalysts are categorized into inorganic/MOF based composite photocatalysts 46,47 and inorganic/C 3 N 4 based composite photocatalysts, 48,49 respectively.The Zhang group have reported two research studies on inorganic/MOF based composite photocatalysts. 46,47The Zhang group 47 have synthesized a zinc oxide (ZnO)/UiO66-NH 2 composite via a partial calcination route, with ultra-small-sized ZnO nanoparticles (NPs) conned into the Review Chemical Science framework of UiO66-NH 2 .The as-synthesized ZnO/UiO66-NH 2 composite photocatalyst is utilized for photocatalytic valorisation of PLA and PVC.First, they utilized a post-synthesis route to acquire Zn-UiO66-NH 2 via coordinating Zn 2+ with the -NH 2 group in Zn-UiO66-NH 2 (Fig. 8a).Then, ZnO/UiO66-NH 2 with a porous structure was acquired via annealing Zn-UiO66-NH 2 in air at 350 °C (Fig. 8a).The SEM image of ZnO/UiO66-NH 2 exhibits a uniform rhombic octahedral morphology exposed with a smooth surface (Fig. 8b), indicating that the raw structure of UiO66-NH 2 is reserved aer the synthesis.The TEM image of ZnO/UiO66-NH 2 shows a uniform particle size of about 200 nm (Fig. 8c).The STEM-HAADF image and corresponding elemental mapping images of ZnO/UiO66-NH 2 (Fig. 8d) show the homogeneous distribution of O, Zn and Zr elements on ZnO/UiO66-NH 2 .As for PLA valorisation, the ZnO/UiO66-NH 2 composite exhibits a higher acetic acid yield of 14.4%, compared to those of ZnO (3.3%) and UiO66-NH 2 (4.7%), respectively (Fig. 8e).They found that the acetic acid generation in the preliminary stage is low followed by gradual enhancement with increasing time (Fig. 8f).This is attributed to the small exposed surface area and low hydrophilicity of big-sized PLA particles in the beginning.Nevertheless, as reaction time increases, these big-sized PLA particles are gradually transformed into small-sized PLA particles with increased hydrophilicity, leading to a larger exposed surface area and higher activity.The ZnO/UiO66-NH 2 composite shows an excellent selectivity of 91.6% for acetic acid via valorisation of PLA (Fig. 8g).As for the control experiments, the ZnO/UiO66-NH 2 composite shows a much higher acetic acid yield (14.4%) than ZnO (3.3%) and physically mixed ZnO/UiO66-NH 2 (2.0%), respectively.These results indicate the great importance of intimate interaction between ZnO and UiO66-NH 2 .Fig. 8h exhibits the total organic carbon (TOC) concentrations of ZnO/ UiO66-NH 2 (2.1 g l −1 ), ZnO (0.6 g l −1 ) and UiO66-NH 2 (0.9 g l −1 ).The obviously larger TOC concentration for ZnO/UiO66-NH 2 suggests its capability of continuously transforming PLA into soluble organic chemicals, e.g., acetic acid, boosted by the synergistic effect between UiO66-NH 2 and ZnO.Control experiments (Fig. 8i) exhibit much lower acetic acid generation for physically mixed ZnO/UiO66-NH 2 (2.0%) and negligible acetic acid generation without light, a catalyst or PLA, suggesting the key role of strong interaction between ZnO and UiO66-NH 2 .
Further study shows that no acetic acid was detected as a N 2 atmosphere is adopted for photocatalytic PLA conversion, indicating the key role of O 2 in photocatalytic plastic conversion.Additionally, ZnO/UiO66-NH 2 also exhibits excellent stability for photocatalytic PLA valorisation with no obvious change observed on the compositions, structures and morphologies.Besides, ZnO/UiO66-NH 2 also exhibits the generation of H 2 in photocatalytic PLA valorisation, with the TON (26.36) and TOF (0.75 h −1 ) observed for H 2 evolution (Fig. 8j).The ZnO/UiO66-NH 2 composite was also utilized for

Chemical Science Review
photocatalytic PVC valorisation, with an acetic acid yield of 9.2% (Fig. 8k) as well as TON of 33.13 and TOF of 0.95 h −1 for H 2 evolution (Fig. 8l).Compared to ZnO and UiO66-NH 2 , the increased photocatalytic activity of the ZnO/UiO66-NH 2 composite arises from the broad light absorption, rapid charge dissociation/migration and highly exposed active sites.Then, they use FTIR spectroscopy to test the intermediate products in PLA valorisation by the ZnO/UiO66-NH 2 composite.As time increases, the rising intensities of two peaks at 1760 and 3400 cm −1 , ascribed to the C]O and -OH of the carboxylic acid, respectively, are observed (Fig. 8m).These results also conrm the capability of the ZnO/UiO66-NH 2 composite to transform PLA into carboxylic acid containing substances in photocatalysis.Furthermore, they also conrm that the developed ZnO/UiO66-NH 2 composite could convert LDPE and PET through a photocatalysis reaction.They also demonstrate that the ZnO/UiO66-NH 2 composite can valorise a commercial PLA bag and PLA straw via photocatalysis.Photoelectrochemical (PEC) current density measurement (Fig. 8n) shows the largest PEC current density for ZnO/UiO66-NH 2 , again conrming the key role of intimate interaction between ZnO and UiO66-NH 2 .
Based on the above results, the excellent PLA conversion of the ZnO/UiO66-NH 2 composite is attributed to the following reasons: (i) partial annealing route reserves the highly porous structure of ZnO/UiO66-NH 2 , thus supplying numerous active centres; (ii) the combination of porous UiO66-NH 2 and ZnO leads to efficient interfacial charge separation/migration; (iii) the combination of ZnO with UiO66-NH 2 optimises the electronic structure.ESR results and quenching experiments together reveal that the radicals of cO 2 − and cOH together play a key role in photocatalytic PLA conversion (Fig. 8o).Besides, Zscheme charge separation/transfer is increased for the ZnO/ UiO66-NH 2 composite (Fig. 8o).Finally, a possible reaction pathway is proposed for photocatalytic PLA conversion, and active radicals preferentially cleave the C-O bond of the PLA chain in photocatalysis, leading to the gradual cracking of the PLA carbon chain.Then, PLA is transformed into PLA plastic fragments or oligomers by these active radicals.At last, these PLA plastic fragments or oligomers are transformed into acetic acid by the ZnO/UiO66-NH 2 composite.For the conversion of PVC, they propose that the active radicals rst cleave the C-Cl bond, followed by further oxidation to oxygen-containing organic intermediates released into the reaction solution.
Finally, active radicals further oxidize these organic intermediates to generate acetic acid.In another study, the Zhang group 47  The small sizes of Ag 2 O NPs incorporated in the pores of Fe-MOF ensure the exposure of abundant active sites, leading to enhanced conversion of MPs.Total organic carbon (TOC) analysis results disclose a remarkable increase in TOC for Ag 2 O/ Fe-MOF and Fe-MOF, since the photocatalyst starts to transform the MPs into water-dissolvable long-chain fragments.This also results in the obvious reduction of the weight on PEG in the beginning of photocatalysis.Further analysis indicates the formation of small amounts of acetic acid in 0-5 hours of photocatalytic reaction, suggesting that in the beginning of photocatalysis radicals convert PEG MPs into long-chain fragments.The generation of ethanol and formic acid is also disclosed.The increased TOC concentration is much higher than the formation of formic acid, ethanol and acetic acid in total, indicating that some of the PEG MPs were transformed into soluble MPs and other intermediates.They also tested the photocatalytic reforming of PE and PET MPs for all the samples.Ag 2 O/Fe-MOF exhibits much higher PE and PET MP weight loss, compared with bare Ag 2 O and Fe-MOF, respectively.Ag 2 O/Fe-MOF also exhibits the increased photocatalytic H 2 evolution activities of 1.7 and 1.9 mmol g −1 h −1 for PE and PET MPs, respectively.The above results not only corroborate that the incorporation of Ag 2 O into porous Fe-MOF can be adopted for photocatalytic reforming of PEG/PE/PET MPs, but also conrm that the active centres arising from structure defects can boost the plastic upcycling.The other two studies are based on inorganic/polymerized C 3 N 4 based composite photocatalysts. 48,49One work 48 reports a heterostructure composed of V-substituted phosphomolybdic acid clusters coupled with g-C 3 N 4 nanosheets (VPOM/CNNS).The FTIR spectra of VPOM/CNNS exhibit distinctive vibration modes of CNNS and Keggin units of VPOM, conrming the successful combination of CNNS with VPOM.Besides, VPOM/ CNNS also shows a similar surface area to CNNS (103.51 m 2 g −1 ), suggesting that CNNS reserving its ultrathin nanosheet morphology aer combining with VPOM clusters.TEM results show that the VPOM/CNNS composite reserves the distinct twodimensional layered structure.Aberration-corrected high-angle annular dark eld scanning transmission electron microscopy (AC-HAADF-STEM) combined with elemental mapping analysis conrms the uniform distribution of VPOM clusters on the surface of CNNS.The XPS results conrm the electron transfer from CNNS to VPOM in the VPOM/CNNS composite.The newly formed peaks in the XPS O 1 s spectra indicate the generation of C-O-Mo or C-O-V bonds, again revealing the coupling of VPOM with CNNS.VPOM/CNNS exhibits increased light harvesting in the range of 460-600 nm, in contrast with bare CNNS, again revealing the existence of VPOM in VPOM/CNNS.Via combining the XPS VB and UPS results, they found that VPOM and CNNS construct a type II hetero-junction with a built-in electric eld pointing from CNNS to VPOM.Furthermore, in situ XPS results conrm the accumulation of photo-induced electrons and holes in CNNS and VPOM, respectively, with light illumination.These results conrm the Z-scheme charge transfer in the VPOM/ CNNS composite.Femtosecond transient absorption spectroscopy (fs-TAS) was adopted to study the photo-induced electron/ hole kinetics in VPOM/CNNS.Via applying AgNO 3 as the electron scavenger, they found that the peak at ∼686 nm is ascribed to the CNNSc − absorption and the signal at 550 nm is principally attributed to the photo-induced holes of CNNS.VPOM/ CNNS exhibits increased CNNSc − absorption at ∼686 nm, compared to bare CNNS, suggesting the more effective separation/transfer of photo-generated electrons/holes.Furthermore, they found an additional decay component (s 3 = 19.42ps) for the hole species of VPOM/CNNS, which is ascribed to the Z-scheme charge transfer pathway in the heterostructure interface.VPOM/CNNS composites all exhibit increased decay lifetimes of electron species, compared to CNNS alone, suggesting the more efficient dissociation/ migration of photo-generated electrons in VPOM/CNNS.ESR experiments further show the obvious enhanced signals of DMPO-cO 2 − and DMPO-cOH signals compared with CNNS or VPOM alone, again conrming the increased charge kinetics in the Z-scheme junction.Then, the as-synthesized photocatalysts were adopted for photocatalytic reforming of a range of plastics.First, they were utilized to photo-reform PE as it is extensively applied and not biodegradable.The optimised VPOM-CNNS composite shows an outstanding photocatalytic HCOOH generation rate (24.66 mmol h −1 g −1 ), about 262 times larger than that of CNNS alone.The optimised VPOM-CNNS composite also exhibits better photocatalytic activity than the mechanically mixed VPOM and CNNS.These further corroborate that the intimate interaction between VPOM and CNNS could obviously increase the charge separation/transfer efficiency.Additionally, a 100 hour stability test was also conducted on the VPOM/CNNS composite.Excellent stability of HCOOH generation was observed on the VPOM/CNNS composite for photocatalytic reforming of PE.Aer a 100 hour reaction, no apparent alteration can be found in the composition/structure of the VPOM/CNNS composite.Besides, VPOM/CNNS was also used for photocatalytic reforming of PP, PVC, PEG and PAM.HCOOH is identied as the upcycled product.The HCOOH H + and cO 2 − were identied as the principal reactive species towards photocatalytic reforming of plastics. 1H NMR spectroscopy was further adopted to study the photocatalytic reforming reaction.Aer 36 hour visible-light illumination, apart from HCOOH as the principal product, large amounts of long-chain alcohols and a trace amount of formaldehyde were identied in the liquid phase.These are common intermediates in the electron transfer-oxygen transfer oxidation reaction of vanadium compounds.This is further conrmed by the observation of an eight-line signal of V IV aer light illumination in an argon atmosphere, suggesting that some of the V V species in VPOM are reduced to form V IV in the photocatalytic reforming of PE reaction.IR spectra reveal the generation of new carbonyl groups in the range of 1710-1760 cm −1 in photocatalytic reforming of PE plastic bags.And the generation of peroxides was also conrmed to arise from the reaction between alkyl radicals and reactive oxygen species.DFT computations were also conducted to acquire the insightful understanding.The computation results indicate that the C and N elements in CNNS and O elements in VPOM serve as the principal reactive sites towards photocatalytic reforming of plastics.On the basis of all the results, they propose a photocatalytic mechanism: with visible-light illumination, both VPOM and CNNS will be excited to generate abundant photo-induced electrons and holes.Via the ligand to metal charge transfer (LMCT), the electrons from the O atom in the HOMO of VPOM is excited to an antibonding orbital of the LUMO in the transition metal centres (V or Mo).For CNNS, with light illumination, electrons are excited from the HOMO or N 2p states to the LUMO or hybridized C 2p and N 2p states.Then, the photo-induced electrons and holes will migrate and dissociate following the Z-scheme scheme.Aerwards, the photo-induced holes remaining in the HOMO (O 2p states) of VPOM will from oxocentred radicals.Finally, the highly active and photo-excited VPOM clusters would boost the oxidative cleavage of the C-C bond, leading to the formation of formaldehyde and a carboncentred alkyl radical.Then, reactive oxygen species will oxidize the alkyl radical to generate alkyl peroxide groups, which is transformed into long-chain alcohols.At the same time, formaldehyde will be oxidized to generate formic acid, by the generated cO 2

−
. Furthermore, Gong et al. 49 have developed a metal-free photocatalyst composed of carbonized polymer dot-graphitic carbon nitride (CPDs-CN).The as-synthesized carbonized polymer dots (CPDs) possess a big conjugated graphitic sp 2 carbon combined with sp 3 carbons, as conrmed by 1 H and 13 C NMR spectroscopy.FTIR and XPS spectroscopy techniques together conrm the existence of carboxylic, hydroxyl and amino functional groups in CPDs.The TEM images show the CPDs-CN composite comprising CPDs with sizes of 1.9-2.4nm loaded on the surface of CN sheets.The XPS results conrm the combination of CPDs with CN sheets via forming amide bonds.The coupling of CPDs with CN sheets also leads to the change of colour from light yellow for CN to dark brown for CPDs-CN, thus increasing the light absorption in the whole visible-light range (400-800 nm).The assynthesized CPDs-CN was utilized for photocatalytic The above ve studies underscore the appealing prospects of composite photocatalysts for plastic upcycling, which exhibit both an extended absorption range of light and efficient charge separation/transfer. [45][46][47][48][49] Especially, MOF based composite photocatalysts show excellent performances for upcycling untreated polyesters/polyolens in air and in organic solvent (e.g., acetonitrile). 46,47Besides, poly-oxalate based composite photocatalysts also exhibit outstanding activities/selectivity for upcycling untreated polyesters/polyolens in an O 2 atmosphere and in organic solvent (e.g., acetonitrile). 48All the performances and reaction conditions in this section are summarized in Table 4.

Conclusion and outlook
The above studies introduce the current achievements in various photocatalysts for plastic upcycling.These photocatalysts were categorized into four different types: (i) metal oxide based photocatalysts; (ii) metal sulphide based photocatalysts; (iii) non-metal based photocatalysts and (iv) composite photocatalysts.Usually, metal oxide based photocatalysts (e.g., TiO 2 ) possess very positive valence band potentials and strongly oxidative photo-excited holes, which can generate highly oxidative cOH radicals to directly oxidize the inert/robust plastics (e.g., PE, PP and PVC) and cleave the strong C-C/C-O/C-H bonds to form value-added short-chain chemicals/fuels under aerobic/anaerobic conditions and at room temperature.In some work, metal oxide based photocatalysts, e.g.Nb 2 O 5 , can even fully oxidize the inert polyolens (e.g., PE, PP and PVC) to yield CO 2 in air and at room temperature, which can be further reduced to form a C 2 chemical (acetic acid).But the evolution efficiencies of these chemicals/fuels for metal oxide based photocatalysts are restricted owing to the wide bandgap width and weak absorption of solar light.In contrast, metal sulphide based photocatalysts (e.g., CdS/CdO x and Cd 0.5 Zn 0.5 S) with suitable bandgap widths exhibit strong absorption of light and favourable redox abilities.Photoinduced holes of metal sulphide based photocatalysts possess moderate oxidation abilities to upcycle the pre-treated polyester/polyolen solutions and yield value-added chemicals without their overoxidation to form CO 2 .On the other hand, photo-excited electrons of metal sulphide based photocatalysts with sufficient reduction capacity could efficiently yield H 2 fuel under anaerobic conditions.Even though metal sulphide based photocatalysts exhibit the above advantages, the insufficient stability and notorious toxicity, especially for Cd-based photocatalysts, signicantly restrict their industrial scale applications for solar plastic upcycling.Furthermore, non-metal based photocatalysts, e.g., C x N y based catalysts, have shown many attractive edges, such as cheapness, high abundance, suitable bandgap width, adequate redox capacities, strong robustness against photo-/chemical corrosion and regulable structure/ composition/functionalities.Actually, C x N y based catalysts coupled with a cocatalyst, such as Ni 2 P, have displayed some activities/selectivity/stability for yielding valuable chemicals/ fuels via upcycling the monomers (e.g., ethylene glycol) released by pre-treating the polyesters (PET and PLA) at room temperature and under anaerobic conditions.Interestingly, C x N y based photocatalysts, e.g., C 3 N 4 , even exhibit some activity/ selectivity/robustness for photocatalytic upcycling of untreated PS into value-added chemicals at raised temperature, in organic solvent (acetonitrile) and in air.But the efficiencies of nonmetal C x N y based photocatalysts are still much lower compared to those of metal-based photocatalysts, making their industrial utilization impossible at the moment.Compared with the above single-photon-absorber based photocatalysts, composite photocatalysts comprising two or even more photoabsorbers are also summarized in this review.They are categorized into inorganic based composite photocatalysts and inorganic/organic based composite photocatalysts.Only one inorganic composite photocatalyst is discussed in this review.The Pt cocatalyst loaded CdO x /CdS/SiC (Pt-CdO x /CdS/SiC) photocatalyst exhibits low H 2 evolution activities from upcycling the untreated PE at raised temperature, in concentrated alkaline solution and under anaerobic conditions.These are mainly attributed to the strong C-C/C-H bonds of inert PE and lack of strongly oxidative ROS, e.g., cOH radicals, owing to the unfavourable valence band edge potentials of CdS or SiC together with anaerobic reaction conditions.For inorganic/organic composite photocatalysts, two MOF based composites, Ag 2 O/ Fe-MOF and ZnO/UiO66-NH 2 , are introduced in this review.at room temperature and in an O 2 atmosphere.This activity arises from the strong oxidation ability on the photo-induced holes from VPOM and the efficient Z-scheme charge transfer in VPOM/CNNS.Currently, although some advancements have been realized in this eld, tremendous challenges are required to be overcome in the future, which is anticipated to bring about numerous opportunities in this eld.For instance, the realistic application of photocatalytic plastic upcycling is severely restricted by the insufficient activity, selectivity and stability together with the low cost-effectiveness of photocatalysts.First, the inefficient use of the whole solar spectrum seriously restricts the maximum efficiency for solar-to-chemical conversion using the photocatalysis technique.Second, the high recombination rate of photo-excited electrons and holes in photocatalysts further limits the efficiency of photocatalysts seriously.This mainly arises from the different time scales for the generation/recombination of electron-hole pairs (picosecond to nanosecond) and surface redox reactions (millisecond to second).Third, the easy aggregation of photocatalyst powder suspended in aqueous solution signicantly impedes the performance of powder-form photocatalysts for large-scale applications.Besides, recycling use of these photocatalysts via regeneration/re-activation is also challenging, since the separation of these powder-form photocatalysts from the suspension reaction system is difficult.Lastly, the long-term use of photocatalysts will lead to the gradual reduction of performance due to the reduced crystallinity and covered active sites by the product.Thus, herein, we summarize these challenges and possible opportunities in the following three sections:

Review
4.1 Regulation of the characteristics of photocatalysts for increased activity/selectivity/stability (1) Currently, no studies report photocatalysts with atomic-scale active sites in this eld.So catalysts with atomic-scale reactive sites, such as single atoms, bi-single atoms and clusters, can be screened and developed for photocatalytic upcycling of various types of plastics (e.g., PE, PP, PVC, PS and PET) into targeted value-added chemicals/fuels with excellent activity/selectivity/ stability.
(2) Only limited engineering methods, e.g., loading cocatalysts, element doping, morphology controlling and constructing Z-scheme/type II junctions, have been applied in this eld.Thus, those advanced engineering routes of photocatalysts, e.g., phase engineering, defect engineering, facet engineering and band structure engineering, can also be utilized for photocatalytic plastic upcycling.
(3) The cocatalyst plays a signicant role in enhancing the activity/selectivity/stability of the photocatalyst.But currently only a few studies report the loading of cocatalysts (e.g., Pt NPs, Ni 2 P NPs, MoS 2 and NiMo) for photocatalytic plastic upcycling.And no insightful studies on the functional mechanism of these cocatalysts have been performed and reported.So more studies can be focused on engineering the composition/ structure of the cocatalyst and its interaction with photocatalysts for tuning their activity/selectivity/stability for specic upcycling reactions.
(4) Currently, all the metal sulphide based photocatalysts reported in this eld are based on Cd-based photocatalysts and suffer from notorious toxicity in realistic applications.Thus, Cdfree metal sulphide based photocatalysts can be screened and developed for photocatalytic plastic upcycling.
(5) Certain photocatalysts, e.g., metal sulphides/selenides/ phosphides, usually suffer from inferior photo-/thermal-/ chemical-stability, compared to those of metal oxides.Their stability can be enhanced by the following strategies: (a) combining with other photocatalysts/co-catalysts (e.g., metal oxides and metals) to boost electron-hole separation/transfer with reductive/oxidative electron/hole transfer to other photocatalysts/co-catalysts for avoiding self-reduction/-oxidation; (b) coating with a metal oxide layer to avoid chemical corrosion from the acidic/alkaline reaction environment.
(6) Cheap and robust C x N y catalysts can be studied more owing to their unique advantages of earth-abundance, strong absorption of light and suitable oxidation abilities.

Advanced characterization and theoretical computations for revealing the structure-activity relationship and insightful reaction mechanisms
(1) A variety of in situ characterization techniques, such as in situ XPS, in situ ESR, in situ FTIR, in situ Raman, in situ AFM-KPFM, in situ XAS, in situ TAS, in situ SPV and in situ TSPL, can be utilized to reveal the structure-activity relationship and reaction mechanism, especially the in situ time-resolved characterization to study the femtosecond-scale kinetics of electron/hole separation/transfer/trapping/recombination in catalysts.
(2) It still remain unknown how the ROS is involved in photocatalytic plastic upcycling reactions.Various in situ characterization techniques, especially in situ ESR and in situ FTIR, can contribute to the study of ROS involved reactions, in which inert and untreated polyester/polyolens can be efficiently upcycled into valuable short-chain chemicals/fuels.
(3) Online GC-TCD/FID and HPLC systems can be established and utilized to track and monitor the intermediates and

Chemical Science Review
products in photocatalytic plastic upcycling for revealing the insightful reaction mechanism under realistic conditions.(4) Based on experimental results, theoretical computations, especially operando computation approaches, can be utilized to gain further insights into the structure-performance correlation in photocatalysts for plastic upcycling.They can also be applied to study the reaction mechanism via revealing the reaction thermodynamics/kinetics in plastic upcycling.

Advanced technologies for the realistic application in this area
(1) Currently, most of the photocatalytic reactions were conducted at 25 °C.The infrared region of the solar spectrum should be utilized to raise the reaction temperature for enhancing the reaction rates.Inexpensive and scalable reactors should be developed, which can efficiently utilize infrared light to raise the temperature of the reaction system and reserve the heat inside the reactor to keep the reaction system at a desirable temperature without external heating.Thus, the rates of photocatalysis reactions can be enhanced greatly.
(2) Seawater can be utilized to upcycle these plastic wastes to avoid the use of limited fresh water resources.
(3) Flow reactors can be utilized to avoid the overoxidation of chemicals to form tremendous CO 2 generated in photocatalytic upcycling.
(4) In photocatalytic plastic upcycling, abundant CO 2 might be generated due to the overoxidation of plastics, especially when an air atmosphere is applied.Thus, CO 2 concentration should be monitored in photocatalytic plastic upcycling.And the efficient capture of the yielded CO 2 and its further conversion into value-added carbon-based chemicals using identical photocatalysts should be studied.
(5) Earth-abundant and cheap cocatalysts can be developed to signicantly increase the rate, selectivity and stability of photocatalysts for large-scale and cost-effective plastic upcycling using sunlight.Especially, efficient, low-cost and scalable loading techniques should be explored to atomically disperse these highly effective cocatalysts onto photocatalysts.(6) Studies should be more focused on one-step photocatalytic upcycling of plastics without any pre-treatment.
(7) Currently, most reactions are conducted in aqueous solution.However, it is very hard for plastics to be suspended well in aqueous solution.More research should be conducted in some organic phase solvent (e.g., acetonitrile and dichloromethane) to better suspend and/or dissolve plastics and ensure better interaction between the catalyst and reactant/ intermediate together with more rapid mass transfer.
(8) Currently, all the photocatalytic plastic upcycling is conducted based on one reactor system, which cannot meet the requirements for realistic applications.Reaction systems containing multiple reactors with photocatalysts possessing different functions can be designed and constructed.For examples, one reactor containing metal oxide photocatalysts can be used to cleave the C-C/C-O/C-N/C-H bonds of plastics and yield monomers/oligomers/small molecules.Furthermore, these yielded monomers/oligomers/small molecules can be further transferred to another reactor containing metal sulphides/C x N y , which possess mild oxidation abilities to transform these chemicals to acquire value-added chemicals without over-oxidizing them to yield CO 2 .
(9) A solar simulator (AM 1.5G, 100 mW cm −2 ) is utilized in most reactions for photocatalytic plastic upcycling.For realistic applications in the future, solar concentrators can be applied to increase the photon intensity to achieve signicantly increased efficiency.
(10) More efficient and cost-effective pre-treatment strategies can be developed and adopted to be combined with the photocatalysis technique for catalytic upcycling of plastics into value-added chemicals/fuels via environmentally benign and cost-effective routes.(11) In realistic applications, it is very challenging to separate plastics and many of them are mixed with each other.Thus, more studies on photocatalytic upcycling of mixed plastics should be conducted to accelerate the development of realistic plastic upcycling techniques.

Fig. 1
Fig.1Schematic images for (a) photocatalytic plastic upcycling using pre-treated plastics and under anaerobic conditions, (b) photocatalytic plastic upcycling using untreated plastics and under aerobic conditions and (c) photocatalytic plastic upcycling using untreated plastics and under anaerobic conditions.

Fig. 3
Fig. 3 (a) Schematic figure showing the conversion of various plastic wastes into C 2 fuels via a designed two-step reaction pathway under simulated natural environmental conditions.(b) Generation of CO 2 in photocatalytic oxidation of pure PE, PP and PVC using Nb 2 O 5 atomic layers.In this reaction, the molar ratio of carbon in each plastic and Nb 2 O 5 atomic layers is about 50 : 1. (c) The production amounts of CH 3 COOH and (d) generation rates of CH 3 COOH and CO in photocatalytic conversion of pure PE, PP and PVC, together with the photocatalytic reduction of pure CO 2 in water.Schematic illustration for (e) the band edge potentials of Nb 2 O 5 atomic layers as well as the potentials for CO 2 , H 2 O, H 2 O 2 and O 2 redox couples at pH = 7. (f) The increased two-step C-C bond cleavage and coupling mechanism for conversion of PE into CH 3 COOH under simulated natural environmental conditions.Reproduced with permission from copyright 2020, Wiley-VCH.36

Fig. 4
Fig. 4 (a) Schematic illustration for photo-reforming various plastic wastes using CdS/CdO x QDs in alkaline aqueous solution.(b) Photocatalytic reforming of various plastics by CdS/CdO x QDs.Reaction conditions: 1 nmol CdS/QDs, plastic powders (50 mg ml −1 PLA, 25 mg ml −1 PET, PET bottle or PUR), without pre-treatment or pre-treated in 2 ml 10 M NaOH aqueous solution. 1H-NMR spectra for (c) PLA, (d) PET and (e) PUR prior to (pre-PR) and after (post-PR) 24 hour light illumination using 1 nmol CdS/CdO x QDs in 2 ml 10 M NaOD in D 2 O. (f) Photocatalytic reforming of the PET bottle to generate H 2 using CdS/CdO x QDs.Reaction conditions: 1 nmol CdS/CdO x QDs, ground PET bottle (25 mg ml −1 ) directly used or pre-treated in 2 ml 10 M NaOH aqueous solution.(f) Inset shows the image of the PET bottle and H 2 bubbles on the surface of plastics.Reproduced with permission from copyright 2018, Royal Society of Chemistry.38

38
Fig. 4 (a) Schematic illustration for photo-reforming various plastic wastes using CdS/CdO x QDs in alkaline aqueous solution.(b) Photocatalytic reforming of various plastics by CdS/CdO x QDs.Reaction conditions: 1 nmol CdS/QDs, plastic powders (50 mg ml −1 PLA, 25 mg ml −1 PET, PET bottle or PUR), without pre-treatment or pre-treated in 2 ml 10 M NaOH aqueous solution. 1H-NMR spectra for (c) PLA, (d) PET and (e) PUR prior to (pre-PR) and after (post-PR) 24 hour light illumination using 1 nmol CdS/CdO x QDs in 2 ml 10 M NaOD in D 2 O. (f) Photocatalytic reforming of the PET bottle to generate H 2 using CdS/CdO x QDs.Reaction conditions: 1 nmol CdS/CdO x QDs, ground PET bottle (25 mg ml −1 ) directly used or pre-treated in 2 ml 10 M NaOH aqueous solution.(f) Inset shows the image of the PET bottle and H 2 bubbles on the surface of plastics.Reproduced with permission from copyright 2018, Royal Society of Chemistry.38

Fig. 5
Fig. 5 (a) Schematic illustration for photocatalytic reforming of pre-treated plastics using the MoS 2 /CdS composite.(b and c) TEM images and (d) HRTEM image of the MoS 2 /CdS composite.(e) Photocatalytic reforming of pre-treated PLA for H 2 evolution using 21.8 wt% MoS 2 loaded CdS in various KOH concentrations.(f) Effect of MoS 2 loading on the H 2 evolution from photocatalytic reforming of pre-treated PLA in 10 M KOH using MoS 2 /CdS composites with different MoS 2 contents.(g) The concentrations of lactate and formate in a 5 hour photo-reforming reaction.(h) The concentrations of acetate and formate in a 5 hour photo-reforming reaction.(i) Schematic image showing the conversion of PE into carboxylic acid followed by a photo-reforming reaction on MoS 2 /CdS.(j) H 2 generation from 25 hour photocatalytic reforming of pre-treated PE. (k) Generation activities of alkane from photocatalytic reforming of pre-treated PE in a 5 hour reaction.(l)In situ ESR spectra in the presence of various substrates via using the spin-trapping agent.Reproduced with permission from copyright 2022, American Chemical Society.39

Fig. 6
Fig. 6 (a) Schematic illustration of plastic photo-reforming using a Ni 2 P/CN x photocatalyst.(b) TEM image of Ni 2 P/CN x .(b) Inset shows the lattice distance of Ni 2 P NPs.(c) High resolution XPS spectra of C 1s for CN x and Ni 2 P/CN x .(d) High resolution XPS spectra of Ni 2p for Ni 2 P and Ni 2 P/CN x .(e) Effect of Ni 2 P loading on the H 2 evolution activities of Ni 2 P/CN x for 20 hour PET photo-reforming.(f) Effect of KOH concentration on the H 2 evolution activities of Ni 2 P/CN x for 20 hour PET photo-reforming.(g) Long-term photocatalytic reforming of PET and PLA.Reaction conditions: 2 wt% Ni 2 P/CN x (1.6 mg ml −1 ), pre-treated PET (25 mg ml −1 ), 2 ml 1 M KOH aqueous solution, simulated sunlight (AM 1.5G, 100 mW cm −2 ) and 25 °C.(h) Long term photocatalytic reforming of polyester microfibers, a PET bottle and a PET bottle coated with soybean oil.Reaction conditions:1.6mg ml −1 Ni 2 P/CN x , 2 ml 1 M KOH, 5 mg ml −1 pre-treated microfibers, 25 mg ml −1 pre-treated PET bottle or 25 mg ml −1 pre-treated PET bottle with 5 mg ml −1 soybean oil and simulated sunlight (AM 1.5G, 100 mW cm −2 ).(i) Image of the upscaled photocatalytic reactor.(j) Upscaled photocatalytic reforming of polyester microfibers.Reaction conditions: 1.6 mg ml −1 Ni 2 P/CN x , 120 ml 1 M KOH, 5 mg ml −1 pre-treated microfibers and simulated sunlight (AM 1.5G, 100 mW cm −2 ).Reproduced with permission from copyright 2019, American Chemical Society.41 as follows: (1) oxidative functionalization of PS with OH groups at C a or C phenyl sites as well as OH or C]O groups at C b sites occurs under both thermo and light illumination conditions using C 3 N 4 ; (2) with light illumination, C 3 N 4 generates photo-excited electrons/holes to form cO 2 − and possible carbon radical intermediates; (3) reactive oxygen radicals or oxidative photo-generated holes easily attack PS-O to generate the C-C-Oc intermediate, resulting in the breakage of the C-C bond and scission of the polymer backbone via the b scission procedure.Overoxidation of carbon containing reactants/intermediates/products could occur in any of the above steps, thus generating undesired CO 2 .

Fig. 7
Fig. 7 (a) Generation of various products for photocatalytic oxidation of PS.Reaction conditions: 10 mg PS (M w = ∼50 kDa), 50 mg g-C 3 N 4 , 30 ml acetonitrile, 300 W xenon light, 150 °C and under 10 bar O 2 .Columns with various colours denote different generated products.The standard deviation of conversion in three parallel experiments were denoted by the error bars.(b) Proposed reaction pathways for photocatalytic oxidation of PS.(c) Schematic illustration for the conversion experiments of 500 mg PS pellets.(d) Photocatalytic oxidation of 500 mg PS pellets in 20 reaction cycles.Reaction conditions: 500 mg PS pellets, 200 mg g-C 3 N 4 , 40 ml acetonitrile, 150 °C, 10 bar O 2 and 300 W xenon light illumination for 8 hours in each cycle.After every cycle, the solution is released and pure solvent is added.(e) Catalytic oxidation of PS after various pre-treatments.Reaction conditions: 20 mg PS, 50 mg g-C 3 N 4 , 30 ml acetonitrile, 150 °C and 300 W xenon light illumination for 8 hours.PS-O: thermal treatment at 150 °C in acetonitrile with O 2 .PS-1: thermal treatment at 220 °C in air.PS-2: thermal treatment at 300 °C in air.PS-3: pyrolysis at 350 °C in N 2 .(f) Performances of catalytic oxidation of PS at different weight hourly space velocities (WHSVs).Reaction conditions: 30 ml acetonitrile, 100 mg g-C 3 N 4 , 120 °C, 300 W xenon light irradiation, and a high pressure syringe pump used to pump PS solution (about 0.3 mg ml −1 in acetonitrile) into the reactor at different rates.10 ml reaction solution was drained out manually as the PS solution pumped in amounted to the same volume.The standard deviation of conversion in 3 parallel experiments was denoted by the error bars.Reproduced with permission from ref. 44 Springer Nature.

have fabricated Ag 2 O
NPs embedded in an Fe-based MOF via in situ conversion of unstable Ag sites in a Fe-Ag bimetallic MOF.First, they synthesized the Fe-based MOF followed by a postsynthesis technique to acquire the Fe-Ag bimetallic MOF.Then, the as-synthesized Fe-Ag bimetallic MOF was subjected to light irradiation to form the Ag 2 O NP enclosed Fe-based MOF, denoted as Ag 2 O/Fe-MOF.The MOF structure impedes the growth of Ag 2 O NPs and renders better dispersion of Ag 2 O NPs. XRD results conrm that the addition of Ag sites and Ag 2 O generates defects in the structure of the Fe based MOF.XPS results reveal strong electronic coupling between added Ag sites and the Fe-based MOF in the Fe-Ag bimetallic MOF.XPS results also disclose the defects generated in the structure of the Fe-Ag bimetallic MOF and Ag 2 O/Fe-MOF.The XPS results also indicate the oxygen vacancies formed in Ag 2 O/Fe-MOF, due to the cleaving of Fe-O bonds following light illumination.Besides, the etching XPS test of Ag 2 O/Fe-MOF shows the increasing peak intensity as the etching time increases, further conrming that Ag 2 O NPs are incorporated into the pores of the Fe based MOF.The incorporation of Ag 2 O NPs in the pores of Fe-MOF is also conrmed by the N 2 sorption analyses, which reveal the obviously reduced surface area of Ag 2 O/Fe-MOF (110 m 2 g −1 ) compared with that of the Fe-Ag bimetallic MOF (268 m 2 g −1 of oxygenated plastics can be further oxidatively cleaved to yield short-chain carbon-based molecules or even over-oxidized to form CO 2 ; (vi) photo-induced electrons can reduce the formed CO 2 to yield CO or even carbon-based molecules, e.g., acetic acid; 36 (vii) photo-induced electrons can also reduce H + /H 2 O to yield H 2 ; (viii) in some cases, non-metal photocatalysts with moderate oxidation ability, e.g., g-C 3 N 4 , and organic solvent, 44 e.g., acetonitrile, instead of aqueous solution, are applied in the reaction system with raised temperature (e.g., 70 °C) rather than room temperature (25 °C).
polarizing and greatly weakening the inert C-C/C-O/C-H/C-N bonds in plastics; (v) these C-C/C-O/C-H/C-N bonds Nb 2 O 5 (ref.36) and Ga 2 O 3 , 37 have been applied for photocatalytic plastic upcycling.

36
1hemical ScienceReview mW cm −2 ) at ambient temperature and pressure via a photocatalysis reaction.1HNMR spectroscopy reveals no detectable liquid product and H 2 , CO and CO 2 were found by gas chromatography (GC).Especially, Co-Ga 2 O 3 exhibits the photocatalytic evolution activities of H 2 (647.8mmol g −1 h −1 ) and CO (158.3 mmol g −1 h −1 ) from converting PE powders, about 160 and 190% times larger than those of Ga 2 O 3 .Aer 48 hour irradiation, the weight loss of PE bags can reach 81%.Co-Ga 2 O 3 also exhibits excellent stabilities for photocatalytic conversion of these plastic powders.Control experiments show that the generated H 2 arises from H 2 O rather than plastics; both O 2 and H 2 O participated in the oxidation of PE to CO 2 , which is further reduced to form CO. The in situ electron spin resonance (ESR) technique conrms the existence of cOH and cO 2 − radicals with Co-Ga 2 O 3 or Ga 2 O 3 in photocatalytic reactions.Combining with isotope-labelling experiments, they deduce that both the cOH radical and O 2 are involved in the photocatalytic oxidation of PE to CO 2 .In situ FTIR spectra further conrm that Co-Ga 2 O 3 can reduce CO 2 to CO via a photocatalytic reaction.Based on the revealed results, they propose the photocatalytic mechanism as below: (i) with light illumination, the photo-induced electrons/ holes in Co-Ga 2 O 3 or Ga 2 O 3 can react with water to form H 2 and O 2 , respectively; (ii) cOH radicals and O 2 are involved in converting plastics into CO 2 , whilst O 2 was also reduced to form cO 2 − radicals, H 2 O 2 and H 2 O; (iii) the formed CO 2 is further reduced to generate CO and O 2 is also formed via oxidation of H 2 O. Furthermore, the reasons for the increased activities of Co-Ga 2 O 3 compared to those of Ga 2 O 3 are summarized as follows: (i) the enhanced light absorption of Co-Ga 2 O 3 from d-d internal transition; (ii) the increased density of states (DOS) at the VB edge for Co-Ga 2 O 3 ; (iii) the obviously reduced charge recombination for Co-Ga 2 O 3 ; (iv) the stronger adsorption towards CO 2 for Co-Ga 2 O 3 ; (v) the lower energy barrier for CO 2 reduction or H 2 evolution on Co-Ga 2 O 3 .Both of the above two studies have adopted a one-step photocatalytic upcycling route to directly convert polyolens into value-added chemicals in an air atmosphere.With the existence of O 2 in air, various ROSs can be generated to participate in the reactions, which can help convert these plastics into oxygenated value-added chemicals/ fuels and CO 2 .The generated CO 2 can be further reduced by photo-induced electrons to form value-added chemicals, e.g., CO and acetic acid.All the performances and reaction conditions of the studies in this section are summarized in Table Nevertheless, a photoluminescence test using terephthalic acid (TPA) as the cOH scavenger only shows a negligible characteristic PL peak arising from cOH, suggesting the minimal role of cOH in photoreforming.In contrast, photo-generated holes are deemed as the major active species in photo-reforming.The MoS 2 /CdS photocatalyst was utilized for photo-reforming of pre-treated PLA, PET and PE.The largest H 2 evolution of 6.68 ± 0.10 mmol g −1 h −1 was realized by MoS 2 /CdS from photoreforming of pre-treated PLA in 10 M NaOH aqueous solution (Fig.5e).Fig.5fshows the inuence of various MoS 2 loading amounts on the tip of CdS, with 21.8 wt% MoS 2 loading reaching the highest H 2 evolution activity.To highlight the unique benet of the MoS 2 /CdS structure, MoS 2 nanosheets selectively loaded on the sidewalls of CdS nanorods were synthesized as the control sample, denoted as CdS@MoS 2 .In contrast, CdS@MoS 2 with a similar loading of 21.1 wt% only exhibits a much lower photocatalytic H 2 evolution rate of 2.21 ± 0.25 mmol g −1 h −1 , compared with that of MoS 2 /CdS under identical reaction conditions (Fig.5f).MoS 2 /CdS exhibits 200 hour long-term stability for photocatalytic H 2 evolution in pretreated PLA solution, with 72% of the activity in the 8th cycle compared with that in the 1st cycle.Furthermore, isotope labelling experiments conrm that the source of generated H 39© 2024 The Author(s).Publishedby the Royal Society of Chemistry Chem.Sci., 2024, 15, 1611-1637 | 1621 Review Chemical Science generated holes and cOH with light illumination.2 is water splitting, but not the pre-treated PLA substrate.Both 1 H NMR spectroscopy and high-performance liquid chromatography (HPLC) together indicate the generation of formate (5.37 ± 0.67 mmol l −1 ) aer 5 hour light illumination (Fig.
1hey have also tested the photocatalytic reforming of PET on MoS 2 /CdS, displaying stable H 2 evolution in a 25 hour test, with a rate of up to 3.90 ± 0.07 mmol g −1 h −1 .The 200 hour test further indicates the structural and composition stability of MoS 2 /CdS.Isotope-labelling experiments suggest that the generated H 2 arises from water splitting, but not from PET constituent monomers.Furthermore, photocatalytic reforming of a pre-treated PET bottle shows a H 2 evolution rate of 3.53 ± 0.07 mmol g −1 h −1 , further conrming the potential realistic application.Formate, acetate and glycolate can also be detected by1H NMR spectroscopy in photoreforming of pre-treated PET.The gradual increase of formate and acetate amounts can be observed in a 5 hour reaction (Fig.
3  2−arising from lactate oxidation by MoS 2 / CdS.Additionally, DMPO and Na 2 S/Na 2 SO 3 are added as scavengers of cOH radicals and Na 2 S/Na 2 SO 3 , respectively.No obvious reduction of formate was observed aer adding DMPO.In contrast, apparent reduction of formate was found aer adding Na 2 S/Na 2 SO 3 , indicating the principal role of holes in lactate oxidation.This is in accordance with the ESR and PL results.Then, it is proposed that lactate is oxidized by holes to form acetaldehyde followed by acetate, methanol and formate.But the other products are not observed by 1 H NMR spectroscopy and HPLC except formate.
40Notably, MoS 2 /CdS still exhibits robust H 2 evolution aer a 200 hour test, with good PET compositional/structural stability for the reacted MoS 2 /CdS.Isotope-labelling experiments also reveal that most of the generated H 2 arises from water splitting.Besides, MoS 2 / CdS also shows a CH 4 generation rate of up to 196.2 ± 1.76 mmol g −1 h −1 (Fig.5k) and a CO 2 generation rate of 2.75 ± 0.05 mmol g −1 h −1 .Control experiments reveal that CH 4 originates from the Kolbe photo-oxidation decarboxylation of carboxylic acid, not from CO 2 reduction.Due to the existence of abundant carboxylic acids in substrates, other gaseous alkanes, such as ethane, propane and n-pentane, are also generated with rates of 1.86 ± 0.04, 0.78 ± 0.20 and 7.6 ± 0.60 mmol g −1 h −1 (Fig.5k), respectively, via decarboxylation with a hydrogen transfer mechanism.An in situ ESR test conrms the production of carbon-centred radical species, deemed as the pivotal intermediates in the Kolbe decarboxylation reaction (Fig.5l).These results conrm that MoS 2 /CdS can induce the decarboxylation reaction.Control experiments further indicate that holes play a key role in photocatalytic decarboxylation.Finally, they propose a mechanism: (i) photo-induced holes lead to the oxidation of acetic acid to generate methane via decarboxylation; (ii) photocatalytic decarboxylation of succinic acid and glutaric acid leads to generation of ethane and propane, respectively.This successful photocatalytic decarboxylation reaction is for the rst time reported on CdS based photocatalysts.Similarly, Li et al.40combined the strategies of band structure engineering and loading cocatalysts to synthesize 4.3 wt% MoS 2 coupled Cd 0.5 Zn 0.5 S (M 4.3 /C 0.5 Z 0.5 S).The strong electronic coupling between MoS 2 and Cd 0.5 Zn 0.5 S was conrmed by the obvious peak shis in Raman and XPS spectra.Transient surface photovoltage (TR-SPV) results show the 250% higher SPV value of M 4.3 /C 0.5 Z 0.5 S compared to that of Cd 0.5 Zn 0.5 S, as well as the elongated SPV signal of M 4.3 /C 0.5 Z 0.5 S.These results indicate the much more efficient charge separation and longer lifetimes of photo-induced charge carriers in M 4.3 /C 0.5 Z 0.5 S.These are further corroborated by M 4.3 /C 0.5 Z 0.5 S via the electrochemical impedance spectra (EIS) and photoelectrochemical current (PEC) densities.The polarization curves of M 4.3 /C 0.5 Z 0.5 S exhibit increased HER activity compared to that of Cd 0.5 Zn 0.5 S, indicating that the loading of MoS 2 can increase the activities of M 4.3 /C 0.5 Z 0.5 S. Blank experiments show that without light, a photocatalyst, or NaOH, H 2 evolution or degradation of PET cannot occur.M 4.3 /C 0.5 Z 0.5 S shows the largest H 1 generation activity (15.9 mmol h −1 g −1 ) compared to 4.3 wt% MoS 2 coupled Cd x Zn 1−x S (X = 0, 0.2, 0.4, 0.8 and 1).An outstanding photocatalytic H 2 generation rate was also realized by M 4.3 /C 0.5 Z 0.5 S using PET bottles.Good H 2 evolution robustness was also observed in 5 hour irradiation in PET or PETbottle-based aqueous solution.1H-NMRspectroscopy was CN x .A lattice distance of 0.221 nm was observed in the Fig. 6b inset, attributed to the (111) facet of hexagonal N 2 P. No obvious change is detected between the high resolution XPS spectra of C 1s for H2N CN x and H2N CN x jNi 2 P, suggesting that loading the Ni 2 P co-catalyst doesn't impose obvious inuence on the surface features of H2N CN x .The high-resolution XPS spectra of Ni 2p 3/2 for

Table 2
Metal sulphide based photocatalysts for photo-reforming of plastics AE 1.76 mmol g −1 h −1 ), C 2 H 6 (1.86 AE 0.04 mmol g −1 h −1 ), C 3 H 8 (0.78 AE 0.2 mmol g −1 h −1 ), npentane (7.60 AE 0.6 mmol g −1 h −1 ) CO 2 (2.75 AE 0.05 mmol g −1 h −1 ), 5 h reaction MoS 2 -Cd 0.5 Zn 0.CN x jNi 2 P obviously shi to the lower binding energy direction, compared to those for Ni 2 P, suggesting a strong cocatalyst-support interaction with electrons transferring from H2N CN x to Ni 2 P. First, they pre-treated PET and PLA in KOH © 2024 The Author(s).Published by the Royal Society of Chemistry Chem.Sci., 2024, 15, 1611-1637 | 1623ReviewChemical ScienceH2N terephthalate in PET are released in pre-treatment.Then, all the reaction conditions, such as the loading of the Ni 2 P co-catalyst (Fig.6e) and concentration of KOH (Fig.6f), were optimized to achieve the highest photocatalytic H 2 evolution.50hoursolar light illumination on H2N CN x jNi 2 P leads to photocatalytic H 2 evolution of 82.5 ± 7.3 and 178 ± 12 mmol g −1 using pre-treated x jNi 2 P can photo-reform the pre-treated PLA into formate and acetate.CN x jNi 2 P also exhibits the feasibility of photo-reforming real-world waste of polyester microber, a PET bottle and an oil-contaminated PET bottle into H 2 and a series of chemicals (Fig.6h).Post-catalysis characterization indicates the good stability of CN x in CN x jNi 2 P.However, it should be

Table 3
Non-metal based photocatalysts for photo-reforming of plastics x (52%), conversion (95 AE 5%), 24 h reaction 44 (2022) © 2024 The Author(s).Published by the Royal Society of Chemistry Chem.Sci., 2024, 15, 1611-1637 | 1627 ).Nevertheless, this surface area of Ag 2 O/Fe-MOF (110 m 2 g −1 ) is still much higher than that of bare Ag 2 O (17 m 2 g −1 ), highlighting the advantage of the structure for Ag 2 O/Fe-MOF.As the porous Ag 2 O/Fe-MOF can supply a much larger surface area with increased active centres, compared with the Fe based MOF, Ag 2 O/Fe-MOF exhibits an obviously widened light absorption range, which will boost the photocatalytic performance.Furthermore, MS plots further indicate that the Fe based MOF and Ag 2 O are p-type and n-type semiconductors, respectively.These results reveal the formation of a p-n junction in Ag 2 O/Fe-MOF.Then, transient photocurrent (TPC) density measurements were conducted on bare Ag 2 O, the Fe based MOF and Ag 2 O/Fe-MOF.The results show that the TPC density values are as follows: Ag 2 O/Fe-MOF > Fe-MOF > Ag 2 O.The highest TPC density value of Ag 2 O/Fe-MOF is attributed to the formation of a p-n junction in Ag 2 O/Fe-MOF, leading to increased dissociation/transportation of electrons/holes.Besides, the existence of defects in the structures of Ag 2 O/Fe-MOF leads to more open frameworks and active centres, thus inducing enhanced photo-generated electrons/holes.Additionally, the results of EIS, steady-state PL spectroscopy and bode-phase plots also conrm the highest efficiency of charge separation/ transfer for Ag 2 O/Fe-MOF among these samples, in accordance with the TPC density measurement results.Then, the assynthesized Ag 2 O/Fe-MOF is adopted for photocatalytic H 2 evolution coupled with upcycling of PEG.With the introduction of 0.2 wt% Ag 2 O in Fe-MOF, Ag 2 O/Fe-MOF (0.2 wt%) exhibits the highest photocatalytic PET MP weight loss (27.5 mg in 3 hours) and H 2 evolution (6.2 mmol g −1 in 2.5 hours).In contrast, with lower (0.05 wt%) or higher (1 wt%) Ag 2 O NPs introduced, both Ag 2 O/Fe-MOF (0.05 wt%) and Ag 2 O/Fe-MOF (1 wt%) exhibit inferior photocatalytic PET MP weight loss and H 2 evolution.This is because the introduction of Ag 2 O NPs could boost the catalytic activity of Ag 2 O/Fe-MOF, whilst excessively introduced Ag 2 O NPs induce some destruction in the structure of Fe-MOF, thus reducing the number of active sites.In contrast, bare Ag 2 O and the Fe MOF exhibit inferior PEG conversion efficiency owing to their unfavourable band gap and insufficient light absorption.Additionally, Ag 2 O shows no H 2 evolution since the CB of Ag 2 O (0.12 V vs. NHE) is lower than the H 2 evolution potential (0 V vs. NHE).The physically mixed sample of Ag 2 O and Fe-MOF (Ag 2 O@Fe-MOF) exhibit a low PEG conversion (6.1 mg) and H 2 evolution (2.3 mmol g −1 ).These results reveal that the developed photochemistry route induces intimate interaction between Ag 2 O and Fe-MOF for efficient charge transport.

Table 4
Composite photocatalysts for photo-reforming of plastics

Table 4 (
13ntd.)ChemicalScienceReview reforming of PET and PLA.Photocatalytic reforming of the pretreated PET solution leads to generation of abundant EGderived chemicals, such as glycolaldehyde, glycolic acid, formic acid, ethanol, acetaldehyde and acetic acid aer 8 day photocatalytic reforming using CPDs-CN.1Hand13CNMR spectroscopy reveal that: (i) PET plastic conversion is increases monotonically with increasing time; (ii) high selectivity is achieved for glycolic acid and acetic acid; (iii) little selectivity change is observed for the intermediates, such as glycolaldehyde, formic acid, ethanol and acetaldehyde.The reaction pathway is revealed: photo-induced holes rst oxidize EG to form glycolaldehyde, followed by further oxidation to generate glycolic acid and formic acid.Besides, EG could also be dehydroxylated to generate ethanol, followed by further oxidation to acetaldehyde and acetic acid.The conventional anatase TiO 2 exhibits inferior photocatalytic activities for generating the above chemicals, compared to CPDs-CN.The photocatalytic H 2 evolution activities coupled with PET/PLA hydrolysis were also determined.Without Pt as the co-catalyst, CPDs-CN exhibits a photocatalytic H 2 evolution activity of 298 ± 58 mmol g −1 h −1 using pre-treated PET as the substrate.In comparison, with Pt loaded as the co-catalyst, CPDs-CN exhibits a photocatalytic H 2 evolution activity of 1034 ± 134 and 1326 ± 181 mmol g −1 h −1 via using pre-treated PET and PLA as the substrate, respectively.As a contrast, anatase TiO 2 loaded with Pt only exhibits a photocatalytic H 2 evolution activity of 55 ± 4 mmol g −1 h −1 using pretreated PET as the substrate.The experimental results show that CPDs can obviously increase the light harvesting and boost the dissociation/transportation of photo-excited electrons/holes, thus leading to the increased activities of CPDs/CN.Also, no cOH was detected via the uorescence experiment, suggesting that photo-excited holes play a key role in oxidation of substrates.
Chemical ScienceAg 2 O/Fe-MOF exhibits good H 2 evolution activities for upcycling the PEG/PET/PE MPs.Especially, Ag 2 O/Fe-MOF shows some activity in upcycling PEG MPs into acetic acid.Interestingly, ZnO/UiO66-NH 2 shows good efficiencies for upcycling untreated PLA or PVC into acetic acid with outstanding selectivity accompanied by some H 2 evolution, in water, at room temperature and in air.The good efficiencies of both Ag 2 O/Fe-MOF and ZnO/UiO66-NH 2 are mainly attributed to the rational synthesis strategy to yield Ag 2 O or ZnO NPs encapsulated in the pores of a MOF structure, which ensures strong interaction between Ag 2 O or ZnO NPs and the MOF structure for rapid charge separation/transfer, as well as excellent scattering of ultra-little Ag 2 O or ZnO NPs exposed with a high surface area.Another interesting study about inorganic/organic composite photocatalysts reports the synthesis of poly-oxalate combined polymeric C 3 N 4 , i.e., V-substituted phosphomolybdic acid clusters/g-C 3 N 4 nanosheets (VPOM/CNNS), for upcycling untreated PE/PEG/PP/PVC/PAA in water or organic solvent (acetonitrile) to generate a value-added chemical (formic acid)