Dynamic Self‐Focusing Photothermal Localization Induced Mild Solvent‐Free Upcycling of Polystyrene

ABSTRACT

The pervasive accumulation of nondegradable plastic waste, particularly microplastics, represents a critical environmental crisis demanding advanced recycling strategies. Here, we introduce a fully green, photo‐driven process for polystyrene (PS) upcycling, enabled by a synergistic thermal‐focusing and space‐confined catalysis. Using earth‐abundant MoO_3‐x, it can efficiently absorb sunlight to generate localized heat, dynamically melting adjacent PS and spontaneously forming a core‐shell MoO_3‐x@PS structure. Critically, the intrinsically low thermal conductivity of PS creates a self‐forming insulating layer, which traps thermal energy at the catalyst‐polymer interface. This self‐focusing thermal localization effect sustains a microscopic high‐temperature reaction zone under ambient conditions, dramatically enhancing energy efficiency. The concentrated heat cleaves inert C─C bonds, while photogenerated charge carriers facilitate selective oxidative degradation. Consequently, we achieve 75.0% PS conversion with 70.5% combined yield of valuable products, predominantly benzoic acid crystals that spontaneously separate post‐reaction. Operating without solvents or external heating, our strategy transforms waste PS into valuable chemicals using only sunlight. This work establishes a new paradigm for solid‐state photothermal upcycling, leveraging interfacial thermal localization to enable a truly sustainable light‐to‐chemicals circular path.

A dynamic self‐focusing photothermal localization strategy through the spontaneous formation of a MoO_3‐x@PS core‐shell structure is developed. By harnessing the low‐cost catalyst, minimal energy input, and self‐separating product, this approach achieves efficient solvent‐free, mild‐condition, and cost‐effective upcycling of real waste polystyrene into crystalline benzoic acid.

1. Introduction

The ubiquity of plastic pollution, especially in the form of persistent microplastics has escalated into a global environmental and public health emergency [1, 2]. These recalcitrant fragments pervade ecosystems from ocean depths to mountain peaks, posing ingestion risks to wildlife and entering the human food chain, with potential toxicological impacts ranging from inflammatory responses to endocrine disruption [3, 4]. Addressing this crisis necessitates not only source reduction but also revolutionary advancements in waste management. Chemical recycling, which depolymerizes waste plastics into valuable monomers or platform chemicals, offers a promising route towards a circular plastics economy. However, its widespread adoption is hindered by significant technological hurdles, particularly for plastics with robust carbon–carbon (C─C) backbones like polystyrene (PS) [5, 6].

Conventional chemical recycling of PS, such as pyrolysis or catalytic cracking, typically requires harsh conditions to overcome the high bond dissociation energy of C─C bonds, resulting in substantial energy consumption and complex product mixtures that demand costly separation [7, 8, 9]. Solvent‐based processes, while sometimes milder, introduce another layer of environmental and economic burden due to solvent use, recovery, and potential emissions [10, 11]. Recently, significant advances have been achieved in photodriven PS upcycling, including intrinsic photothermal agent design [12], carbon photothermal materials [13], mixed plastic synergistic conversion [14]. However, current photothermal catalytic systems for plastic upcycling still face the challenges including the dependence on external heating device and increasing cost and complexity [15, 16], poor controllability of thermal fields due to non‐uniform temperature distribution [6, 17], and incompatibility with mild conditions, as most systems require high‐flux solar simulators or harsh operating conditions [18, 19]. Moreover, solid‐phase reactions present a solvent‐free alternative but are fundamentally limited by poor interfacial contact between solid catalysts and solid reactants. This inefficiency leads to inadequate heat and mass transfer, causing sluggish reaction kinetics, low energy utilization, and insufficient interaction between photogenerated active species and the polymer substrate, ultimately compromising activity and selectivity [20, 21]. Therefore, developing a process that operates under ambient conditions, eliminates solvents, utilizes sustainable energy input, and directs products toward high‐value chemicals remains a formidable, yet paramount, challenge.

Solar energy represents a clean and sustainable resource. If one can efficiently convert the incident sunlight into localized high temperature, it would open a pathway toward truly solvent‐free and mild‐condition catalysis. Achieving such thermal localization, however, necessitates an effective insulating material. Polystyrene (PS), with its intrinsically low thermal conductivity presents an ideal candidate to confine heat precisely at the reaction interface. By coupling this property with a photothermal semiconductor, such as molybdenum oxide (MoO_3‐x), capable of generating reactive radicals, we can establish a self‐insulating catalytic system. Herein, we report a photodriven, entirely green process for the upcycling of real PS waste (Figure 1), where the polymer itself serves as both feedstock and thermal insulator to enable highly efficient, localized photothermal conversion. Upon light irradiation, the catalyst rapidly generates heat, melting the adjacent PS. We discover that the molten PS, possessing very low thermal conductivity, dynamically forms an insulating layer coating the catalyst particles. This layer reduces heat dissipation significantly, creating a self‐focusing thermal localization effect that sustains a high‐temperature micro‐reaction zone at the catalyst‐polymer interface. It allows for the efficient cleavage the C─C bonds of PS under ambient air pressure and light irradiation, without any solvent or external heating. The process selectively converts waste PS into high‐value aromatic compounds, with benzoic acid as the primary product crystallizing spontaneously for easy recovery. This work demonstrates a closed‐loop light‐to‐chemicals path, transforming a problematic waste stream into valuable resources using only sunlight, and presents a novel design principle for efficient solid‐state photothermal reactions.

FIGURE 1.

Schematic illustration of the self‐focusing thermal localization effect and the photothermal upcycling of PS to benzoic acid in this work.

2. Results and Discussion

2.1. Characterization of MoO₃, MoO_3‐x, and MoO_3‐x@PS

The fabrication of MoO_3‐x is schematically depicted in Figure S1. As temperature increases, the precursors of Ammonium molybdate tetrahydrate (AHM) initially undergoes dehydration, followed by successive ammonium ion decomposition. After proceeding an intermediate phase of hexagonal [(H₂O/NH₃)](MoO₃)₃ (247°C), orthorhombic α‐MoO_3‐x was obtained [22]. During this process, the rapid evolution of NH₃ and H₂O induces partial lattice oxygen loss, while the restructuring of the layered architecture during the transformation of the intermediate to α‐MoO_3‐x collectively contributes to the formation of oxygen defects [23]. Scanning electron microscopy (SEM) (Figures 2a and S2) and transmission electron microscopy (TEM) (Figure 2b) images present the as‐prepared catalyst consists of sub‐micrometer in size with a distribution ranging from 100 to 800 nm. The selected area electron diffraction (SAED) pattern reveals its single crystalline feature (Figure 2c), which is consistent with the fast Fourier transform (FFT) pattern (Figure S3). The inverse fast Fourier transform (IFFT) images (Figures 2d‐f, S4, and S5) give of the lattice fringes of MoO₃ with a lattice spacing of 0.38, 0.26 and 0.33 nm, corresponding to the distance between the (110), (111), and (021) crystal facets in α‐MoO₃, respectively [24]. In addition, the uniformly distributed Mo and O elements in EDX mapping (Figures 2g,h, and S6) demonstrate the successful preparation of molybdenum oxide. Moreover, pronounced local lattice fringe discontinuities and distortions can be observed in the TEM images of MoO_3‐x (Figure 2i–j, as highlighted by the yellow circles), which is attributed to atomic displacements induced by oxygen vacancies [25].

FIGURE 2.

(a) SEM image, (b) TEM image, and (c) SAED image of MoO_3‐x. (d) HRTEM image of MoO_3‐x. (e) IFFT pattern of green box area in Figure 3d. (f) Line profile of IFFT image in Figure 2e. Elemental mappings of (g) O and (h) Mo of MoO_3‐x. (i) Enlarged view of blue box in Figure 2d. (j) IFFT pattern of yellow box area in Figure 2i. (k) XRD pattern of MoO₃ and MoO_3‐x. (l) UV–vis spectra and (m) photothermal response curves of PS, MoO₃, and MoO_3‐x.

The typical X‐ray diffraction (XRD) patterns of MoO₃ and its defect one (MoO_3‐x) (Figure 2k) exhibit similar diffraction peaks, confirming their identical structure (JCPDS No.05‐0508) [26]. X‐ray photoelectron spectroscopy (XPS) survey spectra (Figure S7) demonstrated that the prepared MoO_3‐x contained O and Mo elements, and the two distinct peaks at binding energies of 530.7 and 531.7 eV can be attributed to lattice oxygen atoms (O_L) and vacancy of oxygen atoms (O_V), respectively (Figure S8) [27]. Significantly, the EPR spectra (Figure S9) show a signal with g value of 2.003, confirming the presence of O_V in molybdenum oxide nanosheets [28, 29]. The ultraviolet‐visible diffuse spectra (UV–vis DRS) shows the PS exhibited strong absorption exclusively in the wavelength range below 300 nm, whereas MoO₃ and MoO_3‐x possesses an absorption onset extending to 400 nm, with a slightly absorption throughout the 400–800 nm range (Figure 2l, S10). Moreover, the enhanced absorption of MoO_3‐x in near‐infrared range (800‐900 nm) was also appeared while MoO₃ exhibited negligible optical absorption in this range, which is attributed to the relatively localized electrons induced by surface defect of oxygen [30, 31]. The fluorescence intensity follows a trend of initially increasing and subsequently decreasing with the rise in the proportion of PS, originating from the generation and recombination efficiency of photogenerated charges (Figure S11) [32]. These phenomena suggested that the substrate ratio can modulate the light harvest and charge transfer process, and only an optimal ratio can ensure the efficient progression of the catalytic reaction. Notably, the pure PS exhibited a negligible temperature elevation (from 27°C to 40°C) during the initial 5min irradiation period (Figure 2m), which is attributed to its inherently weak optical absorption. This observation confirms the minimal contribution of pristine PS to heat generation in photothermal catalytic systems. While the MoO₃ undergoes a steady temperature increase, reaching 132°C by the completion of the 5 min irradiation under the same experimental conditions. In contrast, the oxygen vacancy‐rich MoO_3‐x exhibited significantly enhanced photothermal conversion efficiency, evidenced by a rapid temperature increase from ambient conditions (27°C) to 150°C within the first 2 min of irradiation and continued to rise thereafter, ultimately reaching a maximum of 184°C (52°C higher than MoO₃). These results provide compelling evidence that oxygen vacancies substantially enhance the photothermal conversion efficiency, thereby laying the foundation for the subsequent formation of the dynamic core‐shell structure.

Backscattered electron SEM (BSE‐SEM) characterization shows the MoO_3‐x and PS particles are only in simple physical contact after the physically mixing of MoO_3‐x and PS powders (Figure 3a), while it change to a tightly coating after 8 min of light irradiation (Figure 3b,c) [33]. The interface between the MoO_3‐x (outlined by the red dashed curve) and the PS (within the yellow dashed curve) is no longer distinct, and the brightness of the MoO_3‐x regions is markedly reduced. These morphological and compositional changes indicate that the melted PS has compactly coated the MoO_3‐x nanoparticles, resulting in the in situ construction of a core‐shell structure. Additionally, the proposed morphological evolution is also corroborated by TEM (Figures S12 and 3d‐g), that the PS component melts and coats the surface of MoO_3‐x upon irradiation, laying the structural foundation for the induction of the dynamic self‐focusing thermal localization effect. The characteristic peaks at 996, 820 and 666 cm⁻¹ can be ascribed to A_g, v _as M═O stretch, A_g, v _s M═O stretch and B_2g, B_3g, v _as O–M–O stretch of MoO_3‐x, which were eventually diminished with increasing temperature (Figure 3h) [34]. Since Raman is surface‐sensitive, this decreasing signal indicates the surface of MoO_3‐x is covered during irradiation, which could reflect the core‐shell structure formation process (Figure 3i). Differential scanning calorimetry (DSC) curve of the employed PS, indicating that the melting temperature of PS is approximately 93.6°C (Figure 3j). Notably, for both MoO_3‐x and MoO_3‐x@PS composite, they maintain similar tempera rising rates (K _T0 = 247.6, K represents the slope of the tangent line to the temperature curve) when the temperature is below the melting point of PS (Figures 2j‐l and S13). However, once the temperature exceeds the PS melting point, PS melts and in situ forms a compact insulating layer on the MoO_3‐x surface, leading to a pronounced divergence in heating rates between MoO_3‐x and MoO_3‐x@PS. The heating rate of MoO_3‐x decreases significantly to 82.2. This decline is attributed to the increasing temperature difference between MoO_3‐x and the ambient environment as its temperature rises, resulting in substantial heat losses through radiation, convection, and conduction [35]. In contrast, for the MoO_3‐x@PS composite, the in situ formed PS insulation layer effectively confines heat dissipation, allowing it to retain a considerably higher heating rate (K _T = 165.1), resulting an additional temperature increase of 50°C compared to its individual components, reaching a peak temperature of 236°C (on MoO_3‐x@PS‐3, MoO_3‐x coating with 250 mg PS) under identical irradiation conditions. This result clearly corroborates the successful establishment of the dynamic self‐focusing thermal localization effect. Particularly, the optimized MoO_3‐x@PS composite system achieved the observed photothermal phenomena can be attributed to a sequential process initiated by the temperature elevation within the catalytic system. And an appropriate thickness of the PS insulation layer is paramount for achieving the optimal dynamic self‐focusing thermal localization effect, which is determined by the PS mass (250 mg) within the composite (Figure S14). This parameter is crucial for maximizing the overall photothermal conversion efficiency.

FIGURE 3.

(a) SEM image of physical mixed MoO_3‐x and PS with BSE mode. (b and c) SEM images with BSE mode. Elemental mappings of (d) C, Mo, O, (e) C, f) Mo, and (g) O of MoO_3‐x@PS‐3 (physical mixed MoO_3‐x and PS irradiated for 8 min). (h) Temperature‐dependent in situ Raman spectroscopy and (i) schematic illustration of the core‐shell structure formation process of MoO_3‐x@PS‐3. (j) DSC curve of PS. (k) photothermal response curves and (l) photothermal IR images of different samples.

Furthermore, the simulation results also revealed the self‐focusing thermal localization effect induced by the core‐shell structure (Figure 4a). Typically, the MoO_3‐x structure was simplified as a spherical core (orange region), while the PS film (blue region) was approximated as a homogeneous continuous layer. Heat conduction through the film was simulated using a porous media methodology to maintain physical consistency while ensuring computational tractability. The MoO_3‐x core serves as the primary heat source, exhibiting an overall elevated temperature that gradually decreases along the radial direction (Figure 4a). The thickness of the PS film simultaneously influences both light absorption (i.e., photothermal generation) and thermal localization effects. The light absorption behavior can be readily interpreted with the PS film thickness increases, thereby diminishing photothermal generation. The temperature distribution demonstrates that with increasing PS film thickness, the temperature decay rate within the PS layer slows (as indicated by the decrease in the k‐value from −0.02137 to −0.01176), signifying a strengthened thermal confinement effect (Figure 4b). An optimal PS film thickness is essential to balance heat generation and thermal retention, thereby achieving the most pronounced self‐focusing thermal localization effect, which is consistent with experimental measurements (Figure S14). Additionally, there is a significant shift in the XPS spectrum of Mo 3d (Figure 4c), indicating the tightly coating of PS on the surface of MoO_3‐x nanoparticles. In detail, the peaks located at 236.1 and 233.1 eV can be ascribed to Mo⁶⁺ 3d_3/2 and Mo⁶⁺ 3d_5/2, and peaks centered at 234.5 and 231.6 eV originated from Mo⁵⁺ 3d_3/2 and Mo⁵⁺ 3d_5/2 [36]. This spectral feature indicates the coexistence of Mo^{6 +} and Mo^{5 +} species within the material, owing to the existence of oxygen vacancies [37]. The Mo 3d peaks of MoO_3‐x@PS composite shows a distinct shift towards lower binding energy, indicating that the Mo species has gained electrons [38]. Given that PS is an electron‐donating entity, this observation suggested that the core‐shell structure formed is not merely a result of simple physical coating, but rather involves chemical interactions accompanied by charge transfer. And such electronic interaction is highly favorable for the proceeding of catalytic reactions. The charge transfer dynamics at the interface of MoO_3‐x@PS were investigated through photocurrent measurements (Figure S15). Due to the inherent electrical insulating nature of PS, the MoO_3‐x@PS composites exhibit a lower overall photocurrent intensity compared to pristine MoO_3‐x. Crucially, an I/I ₀ ratio below 1 for MoO_3‐x@PS‐x (x = 1, 2, 3, 4 and 5 corresponding to PS of 150, 200, 250, 300 and 350 mg) indicates localized charge accumulation at the electrode interface, rather than on the MoO_3‐x electrode itself [39]. The contrasting I/I ₀> 1 for pure MoO_3‐x confirms that photogenerated charges partially transfer from the catalyst into the adjacent PS component, which likely serves as a charge reservoir, potentially beneficial for both light harvesting and the spatial separation of photogenerated charge carriers. To further investigate the charge transfer behavior, Kelvin probe force microscopy (KPFM) was conducted. The dark‐state surface potential of the MoO_3‐x is 189.6 mV (Figure 4d,h), which is much higher than that of MoO_3‐x@PS composite (Figure 4f,i, in the range of 95.6–100.7 mV). This can be ascribed to the fact that the surface of the MoO_3‐x@PS composite is predominantly constituted by PS. However, under irradiation, the magnitude of the potential increased to 31.7‐37.6 mV for MoO_3‐x@PS composite (Figure 4i), which is higher than that of pure MoO_3‐x (Figure 4h, 25.7 mV). This result indicates that the MoO_3‐x@PS complex exhibits a significantly enhanced surface charge accumulation than pure MoO_3‐x under light irradiation [40]. Furthermore, since PS itself does not produce any charge under light exposure, it implies that the charges was photogenerated on MoO_3‐x and then migrate to PS upon irradiation. This charge transfer process facilitates the separation of photogenerated charges, thereby ensuring the efficient catalytic reaction.

FIGURE 4.

(a) Simulation of temperature distribution of self‐focusing thermal localization induced by different‐thickness PS layers via COMSOL. (b) Simulated radial temperature profile from the sphere center outward via COMSOL. (c) High‐resolution XPS spectra of Mo 3d of MoO_3‐x and MoO_3‐x@PS. KPFM image of MoO_3‐x (d) in darkness and (e) under light irradiation and MoO_3‐x@PS (f) in darkness and (g) under light irradiation. (h) Surface potential of MoO_3‐x obtained from the sample areas (Figures 4d,e, white line). (i) Surface potential of MoO_3‐x@PS obtained from the sample areas (Figures 4f,g, white line).

2.2. Photothermal Catalytic PS Upcycling

The homemade catalytic reaction device for upcycling of PS was shown in Figure 5a. The reaction substrate was placed within the reaction cell, and the reaction was initiated via bottom‐up irradiation. Due to the photothermal effect, the air at the bottom of the device was heated first and rises, thus establishing a circulating airflow within the device. Meanwhile, the degradation products of PS were also carried upward by this airflow. When the product attached on the cooler walls, it will nucleation, condensation and crystallization to a bigger crystal, which facilitated the spontaneous separation of the products from the unreacted substrate. Liquid‐state ¹H and ¹³C nuclear magnetic resonance spectroscopy (NMR) analysis shows all the resonance signals exclusively corresponding to benzoic acid, indicating the high purity of product (Figure 5b). GC‐MS analysis further indicates in addition to the primary product benzoic acid (yielding 51.3%), another two byproducts of benzaldehyde (yielding 14.8%) and acetophenone (yielding 4.4%) of MoO_3‐x@PS‐3 can be also detected with the light‐to‐chemical conversion efficiency (LCCE) of ∼0.509% (Tables S32–S36) [41], alongside trace amounts of other products (<0.5%). A comparative analysis of the PS degradation rate and product yields between MoO₃@PS (MoO₃ coating with 250 mg PS) and MoO_3‐x@PS‐3 are illustrated in Figures S16 and S17, respectively. Under identical experimental conditions, MoO_3‐x@PS‐3 exhibited a PS conversion rate of 75.0%, which is much higher than that of MoO₃@PS (39.0%). Additionally, the benzoic acid yield of MoO_3‐x@PS‐3 reached 51.3%, representing a 1.9‐fold enhancement compared to MoO₃@PS (27.6%).

FIGURE 5.

(a) Photograph of the photothermal catalytic reaction device and benzoic acid crystal product captured during the reaction process. (b) ¹H and ¹³C NMR spectrum of the individually collected crystalline product. (c) GC‐MS spectrum of all products dissolved in acetonitrile. (d) PS upcycling rates over MoO_3‐x@PS‐x. (e) Product yields from photothermal catalytic reactions with different MoO_3‐x‐to‐PS ratios. (f) The GPC of degraded PS over MoO_3‐x@PS‐x and raw PS. (g) Benzoic acid yields of photothermal catalytic upcycling for real PS plastics. (h) Comparison with some reported literature on the recycling of plastic.

The temporal dynamics of PS degradation efficiency and benzoyl product formation over a 24‐hour irradiation period under varying weight ratios of MoO_3‐x to PS are presented in Figure 5d,e. Notably, both parameters exhibit a non‐monotonic dependence on PS concentration. When the concentration of PS increases, the degradation rate initially ascends to a peak value before subsequently declining, while the total yield of benzoyl products follows a parallel ascending‐descending trend, demonstrating that an optimal thickness of the insulation layer is crucial for ensuring the efficient progression of the reaction. The gel permeation chromatography (GPC) revealed that the number‐average molecular weight (M _n) of the irradiated PS substrates was significantly reduced to a minimum of 1506, which is two orders of magnitude lower than the initial value of 359830 (Figures 5f and S18 and Table S1). This substantial reduction unequivocally demonstrates the exceptional depolymerization efficiency of the applied photocatalytic system toward PS substrates. Furthermore, the PS depolymerization degree exhibited a trend of initial increase followed by a decrease with increasing PS feed amount, consistent with the PS conversion rate and product yields. The reason may be that when the proportion of PS is too low (PS: MoO_3‐x ≤ 3: 10 in weight, Figure S19a), the resulted thin thermal insulation layer on the MoO_3‐x surface leads to heat dissipation, thereby resulting in low reaction efficiency. Conversely, an excessively high PS proportion (PS: MoO_3‐x ≥ 7: 10 in weight, Figure S19b) results in an overly thick insulation layer. This enhances optical reflection, thereby weakening light absorption and reducing the generation of photogenerated charges and photothermal energy, but also hinders the outflow of reaction products, ultimately suppressing the reaction process and causing a precipitous decline in yield. Only when the thickness of the PS insulation layer is optimal as PS: MoO_3‐x ≈ 1: 2 in weight, the system can effectively balance efficient optical absorption, effective heat preservation, and smooth product diffusion (Figure S19c). Furthermore, we extend the catalytic decomposition into real‐life PS‐based materials under daily‐use scenarios, which exhibited comparable levels of commercial PS powders under the same testing conditions (Figure 5g), indicating that additives and impurities in these waste PS materials have negligible effect on the reaction efficiency. Finally, the structure of the catalyst remained unaltered before and after 24 h reaction, confirming its superior structural stability (Figures S20 and S21). Furthermore, the MoO_3‐x@PS system exhibits excellent stability over five consecutive cycles (120 h total), maintaining benzoic acid yields above 40% with minimal decline from the initial 48% (Figure S22). The consistent kinetic profiles across cycles, including similar induction periods and reaction rates, indicate preserved accessibility of catalytic active sites. This slight decline is attributed to carbon deposition on the catalyst surface (Figure S23), which can be readily removed through calcination regeneration step, largely restoring the catalytic activity (Figures S22,S24, and S25). We also examined the single‐batch long‐term stability, the system maintained adequate photothermal temperature and sustained catalytic activity throughout prolonged reaction time (Figures S26,S27). Compared to the previously reported literature on plastic recycling, this work presents a comparison superiority in terms of cost control, severity of reaction conditions, product value, and selectivity (Figure 5h, Table S2) [10, 15, 16, 17, 19, 42]. To provide a rigorous and transparent assessment, we established a unified techno‐economic calculation framework applicable across PS recycling pathways including mechanical recycling, solvent dissolution‐reprecipitation, catalytic pyrolysis, solvent‐chemical catalytic recycling and solvent‐photocatalytic recycling (Tables S3‐S31). As shown in the Table S31, this work achieves a return of investment (ROI) of 1,327%, a payback period (Payback) of 27.5 days, and a gross profit (π_net ) of $2,048 USD/t PS, 1–2 orders of magnitude superior to all other processes, highlighting the advantages of this work. Interestingly, there were no products can be detected under dark conditions with thermal activation, indicating the impossibility of a purely thermos‐catalytic pathway (Figure S28). Similarly, when the system was conducted under light irradiation with a 20°C water‐cooled bath, there is still no products can be detected, which demonstrated that the process does not proceed via a single photocatalytic mechanism. Afterwards, we subsequently investigated the effects of PS particle size, light intensity, reaction atmosphere and reactor design. PS size had minimal impact on reaction efficiency, except that excessively fine particles induced over‐oxidation (Figures S29,S30). Light, as the predominant energy input, exerted a pronounced influence on the catalytic process (Figure S31). As expected for an oxidation reaction, the introduction of O₂ enhanced conversion while air represent an acceptable alternative from the perspectives of economy and operational safety (Figure S32). In addition, optimized reactor design can further enhance the reaction efficiency (Figure S33). Furthermore, a scaled‐up reaction system with 10‐fold PS feed capacity (2.5 g) was constructed, incorporating an air pump, four xenon lamps, a modified quartz tube reactor and a spherical condenser (Figure S34). Consequently, 1.026 g of benzoic acid was efficiently collected as pure crystals in the condenser, corresponding to 85.2% of the small‐bath yield (250 mg PS).

The further in situ electron paramagnetic resonance (EPR) spectroscopy shows there is no obvious EPR signals appeared corresponding to superoxide radicals (·O₂ ⁻) or hydroxyl radicals (·OH) under dark or light conditions (Figure 6a‐c). However, a distinct singlet oxygen (¹O₂) signal was detected under dark conditions, which can be attributed to the oxygen vacancies on the MoO_3‐x surface facilitating the adsorption and activation of atmospheric molecular oxygen, a process that alters the electronic spin state of O₂ molecules and consequently induces ¹O₂ generation in the absence of light [43, 44, 45]. Notably, there are more ¹O₂ species generated under light‐exposed conditions. In situ FTIR shows three distinct absorption regions emerged and intensified progressively under. Irradiation (Figure 6d,e). The broad band center at 3000 cm⁻¹ (attributable to the O–H stretching vibration) and the sharp peak at 1700 cm⁻¹ (assigned to the C═O stretching of the carboxylic group) provided direct spectroscopic evidence for the accumulation of benzoic acid as the terminal product [10, 46]. Meanwhile, a sharp peak at 1735 cm⁻¹, ascribed to the ketone C═O stretching vibration of benzoylformic acid, confirmed its role as a key reaction intermediate. Collectively, these results outline the reaction pathway for PS‐to‐benzoic acid upcycling, as illustrated in Figure 6f, the ¹O₂ species initiate the abstraction of C–H bonds from the PS backbone, thereby generating benzylic radical as reaction intermediates, inducing PS undergoes conversion to benzoic acid via HAT and b‐scission pathways. These intermediates subsequently undergoes reaction with molecular oxygen to afford a peroxy radical, as documented in prior studies [42]. In addition to this pathway, ¹O₂ can directly insert into the a‐benzylic C─H bond of the polymer chain, further promoting oxidative degradation through alternative mechanistic channels [47]. The subsequent b‐scission of the peroxy radical facilitates chain cleavage, leading to the formation of two distinct chain end groups, including the phenyl ketone moiety and the phenyl alcohol fragment (Figure 6f). These newly generated species then recursively participate in the aforementioned reaction sequence, ultimately converging to the primary oxidation product, benzoic acid, as supported by experimental evidence (Figure 5c) [48]. In summary, the upcycling of waste PS can be attributed to a photothermal‐driven photocatalytic process, which involves the photothermal activation of PS molecules, followed by the stepwise cleavage of PS into high‐value organic small molecules mediated by photogenerated radicals [49].

FIGURE 6.

(a) EPR spectrum of superoxide radical (·O₂ ⁻). (b) EPR spectrum of singlet oxygen (¹O₂). (c) EPR spectrum of hydroxyl radical (·OH). (d and e) Irradiation time‐dependent in situ FTIR of MoO_3‐x@PS‐3. (f) Diagram of the reaction pathway and mechanism for PS degradation.

3. Conclusion

In this work, we have successfully established a dynamic self‐focusing thermal localization strategy through the spontaneous formation of a MoO_3‐x@PS core‐shell structure. This heat‐aggregating configuration enables efficient conversion of real‐world waste polystyrene into benzoic acid under solvent‐free and normal pressure under light irradiation with high selectivity. By harnessing the intrinsic low thermal conductivity of PS as a self‐insulating layer, the system sustains localized high temperatures at the catalyst/polymer interface using only light irradiation. A 24 h benzoic acid yield of 51.3% was achieved, with the product spontaneously crystallizing from the unreacted substrate, which simplified the further purification. This work provides a new route for solid‐state photothermal upcycling of plastic waste, offering transformative guidance for the design of green catalytic technologies tailored to real‐world waste streams.

Author Contributions

Shuang Tang: Validation, data curation, Writing – original draft, investigation. Chuntian Qiu: Data curation, investigation. Bin Zhang: Investigation. Yang‐Sen Xu: Writing – review and editing, conceptualization, methodology, funding acquisition, formal analysis, supervision. Xiang Ling: Writing – review and editing, conceptualization, methodology, funding acquisition, supervision. Nailiang Yang: Writing – review and editing, resources, formal analysis, conceptualization, funding acquisition.

Funding

National Natural Science Foundation of China (22579116, 52572118, 92163209), Shenzhen International Collaborative Project (GJHZ20220913143013025), Special Innovation Program For General Higher Education Institutions of Guangdong Province (2025KTSCX110), Scientific Research Projects of University of Guangdong Provincial Education Office (2024KCXTD064), Shenzhen Peacock Plan (Grant No. 20180703896C), Shenzhen Key Laboratory of 2D Metamaterials for Information Technology (ZDSYS201707271014468).

Conflicts of Interest

The authors declare no conflict of interest.

Supporting information

Supporting File: adma74031‐sup‐0001‐SuppMat.docx.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.