A Metal–Organic Framework Platform Enabling Visible-to-Ultraviolet Upconversion-Driven Heterogeneous Photochemistry Authors: Subhrajyoti Ghosh, Sakura Nakagawa, Masanori Uji, Manpreet Singh, Naoto Matsumoto, Miku Inoue, Bhavesh Parmar, Nobuhiro Yanai Published: 2026 ABSTRACT Triplet–triplet annihilation photon upconversion (TTA-UC) provides an attractive strategy for converting low-energy photons into higher-energy photons under low excitation intensities, enabling diverse applications in photocatalysis, optoelectronics, and solar energy conversion. However, realizing efficient and durable visible-to-ultraviolet (vis-to-UV) TTA-UC in the solid materials remains a major challenge, primarily due to nonemissive annihilator aggregation, phase separation between sensitizers and annihilators, and poor photostability. Herein, we report, for the first time, vis-to-UV TTA-UC realized within a metal–organic framework (MOF) and demonstrate its utility in a heterogeneous photochemical reaction. A hybrid upconversion system is constructed by co-incorporating a p-terphenyl dicarboxylic acid annihilator (TP-(COOH) 2 ) and a tris­(2-(5-carboxypyridin-2-yl)-3,5-difluorophenyl)­iridium sensitizer (Ir­(dFppy-COOH) 3 ) into a Zr-based UiO-68 framework (UiO: University of Oslo). The resulting MOF, composed of π-conjugated organic linkers and highly stable Zr–O clusters, enables triplet exciton migration while maintaining high photo- and chemical-stability. Upon 445 nm excitation, the THF suspension of the hybrid framework exhibits UV emission with an upconversion efficiency of 0.60%, representing respectable performance for MOF-based vis-to-UV TTA-UC materials. Importantly, the rigid and porous MOF architecture effectively suppresses strong chromophore aggregation, affording excellent photostability. Leveraging these features, the MOF-based upconversion system functions as a mercury- and UV-lamp-free UV light source, enabling spatiotemporally controlled uncaging of MNI-caged l-glutamate in a heterogeneous manner. This work establishes a rational, physicochemically stable MOF-based platform for vis-to-UV TTA-UC in suspension and opens new avenues for sustainable photochemistry, photocatalysis, and light-driven functional materials. Introduction Ultraviolet (UV) light (λ < 400 nm) is a crucial and valuable source of energy, essential for numerous photochemical processes, including sterilization, , wastewater treatment, , solar hydrogen production, and various organic and inorganic reactions. − However, the solar spectrum lacks sufficient UV light (∼5%), necessitating the development of efficient methods for generating UV light. − Current methods, including mercury excimer-based lamps and xenon arc lamps, are inefficient, environmentally harmful, energy-intensive, have a short lifetime, and do not efficiently penetrate the glass reactor. , Photon upconversion (UC) offers a promising solution to the aforementioned drawbacks by converting low-energy photons to higher-energy ones. Several mechanisms of UC have been reported, including two-photon absorption (TPA), , second harmonic generation (SHG), , multistep excitation of lanthanides, , and triplet–triplet annihilation (TTA). ,− Among the reported UC mechanisms, TTA-UC is particularly attractive due to its wide spectral tunability and compatibility with weak excitation light sources. − TTA-UC also offers significant potential for driving high-energy photochemical reactions while minimizing the risk of unwanted side reactions. ,,, Typically, TTA-UC involves a triplet sensitizer (or donor) and a triplet annihilator (or acceptor). Photosensitizers initially produce triplets via the absorption of low-energy photons, then provide triplet energy to annihilators through Dexter-type triplet energy transfer (TET). When two triplets of annihilators are converged, a new singlet excited state is formed, which emits a high-energy photon (Figure a). ,− While visible (vis)-to-vis and near-infrared (NIR)-to-vis TTA-UC systems have been widely explored, there is an increasing practical demand for vis-to-UV upconversion for various applications. ,− Efficient vis-to-UV UC materials can address several key challenges: (i) lowering the energy input required to generate UV light, (ii) enabling UV-driven reactions using low-energy visible light in standard laboratory glassware (as direct UV irradiation typically requires quartz vessels), and (iii) facilitating diverse photocatalytic transformations, particularly those requiring enhanced selectivity in product formation. ,,− (a) TTA-UC energy diagram and application of MOF-based TTA-UC for photocatalysis. (b) Benchmark literature reports on MOF-based TTA-UC in different wavelength regions. Reproduced with permission from ref . Copyright 2020, Wiley-VCH. Reproduced with permission from ref . Copyright 2021, American Chemical Society. Reproduced with permission from ref . Copyright 2018, American Chemical Society. Despite extensive research over the past decade aimed at improving vis-to-UV TTA-UC through diverse chromophore pairs, − most reported chromophoric combinations operate exclusively in solution. − Our group previously demonstrated various vis-to-UV UC systems, including the most efficient to date (27.6%), based on the TIPS-Nph/Ir­(C6)2(acac) chromophore pair. ,,, Building on this, we subsequently developed a solid-state monolith and porous film capable of vis-to-UV UC. , Other groups, including Kerzig et al., and Murakami et al., have also proposed alternative chromophore pairs for solution-based vis-to-UV TTA-UC. ,− However, homogeneous vis-to-UV TTA-UC suffers from several intrinsic limitations, including the solubility of chromophores, their aggregation, and poor photostability under high-intensity excitation. − Moreover, diffusion-controlled TTA-UC in solution is inherently unsuitable for heterogeneous photocatalysis that allows energy-efficient isolation of products. Therefore, identifying a heterogeneous TTA-UC system that emits high-energy UV radiation with a good photostability would significantly advance sustainable photochemistry. Heterogeneous porous materials, such as metal–organic frameworks (MOFs), show potential for efficient solid-state UC. , Proper design of heterogeneous porous materials can overcome aggregation-induced quenching, significantly improving the photostability. Incorporating both sensitizer and annihilator in a single system, while maintaining a desirable distance, is feasible with MOFs. ,− These materials can be easily employed as a heterogeneous catalyst for photochemical reactions, which can be further utilized in flow chemistry-type reaction systems (Figure a). , Moreover, some MOFs show excellent stability in many solvents, allowing us to use any solvent for photocatalysis purposes. − Despite these advantages, a few MOF-based TTA-UC systems have been developed. ,− , For instance, Zhou et al. developed an aqua-stable TTA-UC method using a Zr­(IV)-based MOF for bioimaging. Morris et al. synthesized three proton-convertible MOFs using anthracene dicarboxylate derivatives and Zr-nodes. Stoddart et al. achieved a maximum UC efficiency (ηUC) of 1.95% by using a Zn-MOF (Figure b). However, MOF-derived vis-to-UV TTA-UC and their utilization for photocatalysis remain unexplored (Table S6). Recognizing the research gap, we developed a novel vis-to-UV TTA-UC MOF by introducing a Ir­(dFppy-COOH) 3 sensitizer into a Zr-UiO-68 MOF composed of a p-terphenyl dicarboxylic acid linker (TP-(COOH) 2 ). The resulting hybrid framework exhibits UC of blue light (445 nm) to UV light. This pioneering MOF-based vis-to-UV UC system achieves a maximum UC efficiency of 0.60% under 445 nm excitation, with a threshold intensity of 4.3 W cm–2 and a triplet lifetime of 1.1 ms. The rigid, porous UiO-68 framework effectively suppresses any strong chromophore aggregation and displays excellent photostability. Importantly, this MOF-based hybrid UV emitter establishes the first example of heterogeneous photochemical reactions based on TTA-UC MOF, delivering a UV light source-free, reusable approach for the spatiotemporally controlled uncaging reaction. Selective product formation using visible light and a short reaction time are the key benefits of this system. Overall, these results highlight the potential of a heterogeneous MOF-based vis-to-UV TTA-UC platform, offering new opportunities for sustainable photochemistry and photocatalysis. Results and Discussion Previous studies have established that the combination of p-terphenyl and Ir­(ppy)3 chromophores is an effective pair for upconverting blue light to UV light. Building on this foundation, we synthesized a highly physicochemically robust and photostable MOF incorporating the carboxylic acid derivatives of this UC pair (Figure a). Specifically, a Zr­(IV)-based UiO-68 framework was prepared, and Ir­(dFppy-COOH) 3 was introduced in situ during the MOF synthesis (Figure a). Ir­(dFppy-COOH) 3 was chosen as the sensitizer for these MOF-based UC studies because its carboxylate functionality enables strong coordination to the Zr6 clusters, ensuring stable incorporation within the framework. Importantly, this Ir-complex displayed absorption in the blue region and possesses a triplet energy of 2.49 eV in THF (as determined from the 0–0 emission peak). Since we could not detect any phosphorescence of UiO-68 even at liquid nitrogen temperature, we estimated the T1 energy level of the linker (1,1′,4′,1″-terphenyl)-4,4″-dicarboxylic acid (TP-(COOH) 2 ) using DFT calculations. To reduce computational cost, the calculations were performed on the isolated linker, and we assumed that the T1 energy level of TP-(COOH) 2 does not change significantly upon incorporation into the MOF, since they are not in close contact. The calculated T1 energy of TP-(COOH) 2 (2.48 eV) is close to that of the sensitizer, indicating that TET is energetically feasible (Figure S2). Moreover, the triplet energy values employed in this study should be regarded as approximate estimates. The local coordination environment within the MOF may introduce variations in these energy levels. In particular, under these near-isoenergetic conditions between the sensitizer and annihilator, even subtle energetic differences could influence the efficiency of forward triplet energy transfer, as well as the extent of reversibility and back-transfer processes. Notably, TP-(COOH) 2 also offers excellent thermal and photoluminescence stability. Its compatibility with the well-known, highly stable ‘UiO’ family also facilitates the synthesis of a robust MOF structure. Consequently, this MOF-based hybrid UC system can effectively enhance photostability, mitigate solvent-related limitations, and suppress aggregation-induced quenching, collectively enabling more efficient vis-to-UV TTA-UC. (a) Reaction scheme for the synthesis of Ir-TP-MOF using a sensitizer (Ir­(dFppy-COOH) 3 ) and annihilator (TP-(COOH) 2 ). (b) PXRD profile of the different MOFs. (c) BET surface area and porosity distribution plots (inset) of TP-MOF and Ir-TP-MOF 100 . (d) PXRD profile of the recovered Ir-TP-MOF 100 after dispersing it in different solvents. (e) Normalized absorbance/excitation (dash) and photoluminescence (solid) spectra of Ir­(dFppy-COOH) 3 (blue) (λex = 374 nm), and Ir-TP-MOF 100 (red) (λex = 274 nm), in deoxygenated THF. To construct the hybrid MOF-based TTA-UC system, we first synthesized the Ir­(dFppy-COOH) 3 sensitizer. The detailed synthetic procedure and characterization data for this complex are provided in the Supporting Information (Schemes S1–S4 and Figures S3–S10). MOFs without and with various ratios of Ir­(dFppy-COOH) 3 relative to the TP-(COOH) 2 linker were then prepared via an in situ solvothermal method using ZrCl4 and trifluoroacetic acid (TFA) modulator, in DMF (Figure a). The resulting materials were named as TP-MOF (sensitizer-free), Ir-TP-MOF 50 (Ir­(dFppy-COOH) 3 :TP-(COOH) 2 feed ratio = 1:50), Ir-TP-MOF 100 (1:100), and Ir-TP-MOF 200 (1:200). The actual sensitizer-to-linker ratios in the final materials were determined by HF digestion followed by UV–vis analysis in DMSO, and they are reasonably close to the feed ratios (Table S1). The reaction between ZrCl4 and the TP-(COOH) 2 linker consistently yielded the UiO-68 topology. The PXRD patterns of all synthesized MOFs closely matched the simulated UiO-68 pattern, confirming their phase purity (Figures b and S11). The modest value of R wp and R p obtained from the Pawley refinement of the slow-scan PXRD data for Ir-TP-MOF 100 further validated the structural agreement (Figure S12). The UiO-68 structure contains [Zr6O4(OH)4]12+ secondary building units, where each Zr­(IV) center is coordinated by four μ3-O and four μ3-OH groups. Six TP-(COOH) 2 linkers bridge these SBUs to form the extended framework. It features two types of pores: larger octahedral pores and smaller tetrahedral pores, interconnected by tetragonal windows (Figure S13). The tetrahedral pores arise from assemblies of four Zr6 clusters, while the octahedral pores are built from six clusters. Ir-TP-MOF 50 , Ir-TP-MOF 100 , and Ir-TP-MOF 200 exhibited PXRD patterns nearly identical to TP-MOF, indicating that incorporation of small amounts of Ir­(dFppy-COOH) 3 does not significantly disturb the UiO-68 structure (Figure b). Although the exact coordination environment of Ir­(dFppy-COOH) 3 within the MOF cannot be definitively established in the absence of single-crystal structural evidence, insights from previous literature suggest that it most likely coordinates to the highly oxophilic Zr6-oxo cluster of the MOF. , Importantly, Ir-TP-MOF 100 retained its crystallinity after solvent removal and thermal activation, demonstrating its excellent structural robustness (Figure S11). The successful formation of coordination bonds between the carboxylate groups of TP-(COOH) 2 linker and Zr­(IV) was ensured by FT-IR analysis. Ir-TP-MOF 100 displayed two distinct absorption bands at 1652 and 1584 cm–1, which originated from the Zr-coordinated asymmetric and symmetric stretching of the carboxylate groups of the linker molecules. In the linker, these peaks were located at 1676 and 1602 cm–1, respectively. The shift of these peaks ensures the formation of coordination bonds between Zr­(IV) and the carboxylate groups of the linker (Figure S14). The homogeneous distribution of micrometer-sized crystalline particles in the FE-SEM images of TP-MOF and Ir-TP-MOF 100 validated the phase purity and crystalline nature of the synthesized MOFs (Figures S15–S16). The measured N2 surface areas for activated TP-MOF and Ir-TP-MOF 100 are 1392 and 1146 m2 g–1, respectively (Figure c). This measurement confirms the highly porous nature of these MOFs. The decrease in surface area of Ir-TP-MOF 100 is due to the incorporation of Ir­(dFppy-COOH) 3 in the modified framework. The pore-size distribution plot displayed that most of the micropores of TP-MOF and Ir-TP-MOF 100 are centered at 1.09 nm (inset of Figure c). Thermogravimetric analysis of both TP-MOF and Ir-TP-MOF 100 displayed thermal stability up to 490 °C under N2 atmosphere. In the as-synthesized MOFs, the guest solvent molecules were observed to be removed up to 220 °C (Figure S17). The presence of guest solvents once again supported the porosity of the MOFs. In contrast, the activated sample exhibits a single-step weight loss due to the thermal decomposition of the MOF structures at 490 °C (Figure S17). The TGA curves corresponding to the TP-(COOH) 2 linker and free Ir­(dFppy-COOH) 3 were completely different from the MOFs, which evidenced the absence of unreacted linker in the MOF powders (Figure S18). Moreover, the absence of peaks corresponding to the linker in the PXRD profile of the MOFs further validates this observation (Figure S19). The chemical stability of Ir-TP-MOF 100 was evaluated by dispersing the material in various solvents (THF, DMF, MeOH, and CHCl3) for 2 h at room temperature. PXRD analysis of the recovered samples demonstrated no noticeable changes compared to the pristine MOF, confirming that the framework maintains its structural integrity across a wide range of solvent environments (Figure d). This stability broadens the material’s potential applicability, particularly for catalytic reactions in diverse organic media. After successful synthesis and characterization of the sensitizer-incorporated MOFs, we evaluated their photophysical properties. All measurements were carried out in deoxygenated, stabilizer-free THF after preparation of samples inside an Ar-filled glovebox (O2 < 0.1 ppm). The UV–vis absorption spectrum of a 100 μM THF solution of Ir­(dFppy-COOH) 3 exhibited a distinct absorption band in the visible region with a molar extinction coefficient (ε) of 2415 M–1 cm–1 at 445 nm (Figure e). Upon excitation at 374 nm, Ir­(dFppy-COOH) 3 showed emission peaks at 497 and 535 nm (Figure e). From the photoluminescence spectrum, the triplet energy (T1) of the complex was determined to be 2.49 eV, calculated from the 0–0 emission peak at 497 nm (Figure e). The measured phosphorescence quantum yield (QY) and phosphorescence lifetime of Ir­(dFppy-COOH) 3 were 90.1% and 1.22 μs, respectively (Figure S20 and Table S2). TP-(COOH) 2 was selected as the annihilator due to its characteristic UV emission peaks at 350 and 368 nm with a unity fluorescent quantum yield (Figure S21, Table S2). Upon incorporation into TP-MOF, the emission maxima of TP-(COOH) 2 underwent a noticeable red shift to 380 and 401 nm, and the fluorescence lifetime was significantly prolonged to 19.4 ns compared with that of the free linker (0.85 ns) (Figures S22–S24). This bathochromic shift is attributed to excitonic interactions between the linkers within the framework. , Importantly, the high fluorescence quantum yield of 66.1% was still maintained in the MOF (Table S2). Having validated the individual performance of both sensitizer and annihilator, we integrated them within a single Zr-based framework to minimize intermolecular distances, thereby facilitating efficient TET and enabling multidimensional pathways for rapid triplet migration. The strong Zr­(IV)–carboxylate connectivity also helps to reduce chromophore aggregation, preserving structural fidelity in the solvent phase. Notably, in situ incorporation of Ir­(dFppy-COOH) 3 during MOF synthesis produced an assembly of sensitizer and annihilator exhibiting a UV emission with an emission maximum at 380 nm, which is blue-shifted compared with TP-MOF (Figure S22). This blue shift may be due to some changes in the coordination environment of Zr clusters within the MOF induced by Ir­(dFppy-COOH) 3 . The change in coordination environment was also supported by the slight shift of excitation spectra of the MOF after Ir­(dFppy-COOH) 3 incorporation (Figure S22). The appearance of the characteristic phosphorescence peak of the sensitizer at 540 nm (red-shifted by 5 nm upon coordination) further confirms its successful incorporation into the Zr-oxo cluster (Figures e and S22). Meanwhile, as the Ir­(dFppy-COOH) 3 loading in MOF increases, the phosphorescence emission peak at 540 nm becomes more pronounced. Notably, no red shift was observed upon increasing the sensitizer concentration, indicating the absence of strong aggregation of either the Ir-complex or the TP-based annihilator within the MOF channels (Figures S22 and S25). The phosphorescence peak of Ir­(dFppy-COOH) 3 showed a large red shift to 545 nm in the bulk solid state compared with that in solution (Figure S25). However, these observations do not conclusively define the distribution of the sensitizer or the local aggregation or heterogeneity within the MOF. Moreover, the incorporation of this complex in TP-MOF reduced the phosphorescence QY of the Ir­(dFppy-COOH) 3 due to TET to the neighboring TP-(COOH) 2 ligand. The measured phosphorescence QY of Ir­(dFppy-COOH) 3 encapsulated in Ir-TP-MOF 50 , Ir-TP-MOF 100 , and Ir-TP-MOF 200 were 49.9, 46.4, and 52.2%, respectively (Table S3). Additionally, the photoluminescence lifetime of the free Ir­(dFppy-COOH) 3 was found to be decreased upon incorporation into the MOF matrices (Figure S26). Together, the reduced phosphorescence QY and shortened lifetime provide clear evidence of TET from the Ir-based sensitizer to the TP annihilator inside the framework. Based on the phosphorescence QY values, the TET efficiencies for Ir-TP-MOF 50 , Ir-TP-MOF 100 , and Ir-TP-MOF 200 were estimated to be 44.7%, 48.5%, and 42.1%, respectively (Table S3). These values confirm that increasing the sensitizer loading within the well-defined MOF environment led to consistent TET efficiency while preventing detrimental aggregation. However, the observed TET efficiency for all the UC-MOFs is relatively low. To investigate the underlying cause, we performed phosphorescence lifetime measurements of all UC-MOFs under upconversion irradiation conditions using the Ir sensitizer. The measurements reveal delayed phosphorescence from the Ir-sensitizer, likely due to back-energy transfer from the TP linker’s triplet state to the Ir complex’s triplet state (Figure S26). Notably, the contribution of this delayed component increases with increasing Ir-sensitizer concentration (Figure S26). These findings support the hypothesis of back energy transfer from the triplet TP-linker to the triplet Ir-complex. After completing the basic optical measurements of the THF suspension (3 mg of MOF in 3 mL of THF) for MOFs with different sensitizer ratios, we proceeded to measure TTA-UC. Since the Ir­(dFppy-COOH) 3 sensitizer shows sufficient absorption at 445 nm, all samples were excited using a 445 nm diode laser. Remarkably, upon 445 nm excitation, all MOF suspensions exhibited pronounced upconverted emission in the UV region (Figure a). The emission maximum at 380 nm closely aligns with the fluorescence spectrum of the corresponding Ir-incorporated MOFs in THF. These MOF-assembled UC pairs therefore produce an anti-Stokes shift of 0.48 eV, demonstrating their clear vis-to-UV UC capability. Moreover, no similar emission was observed for the TP-MOF, underscoring the critical role of the Ir-sensitizer in enabling the UC process within the framework (Figure S27). (a) The TTA-UC spectra of the suspension of Ir-TP-MOF 100 in THF with increasing excitation intensity (3 mg of MOF in 3 mL of deoxygenated THF, λex = 445 nm, the UC emission below 425 nm is blocked by the 425 nm short-pass filter). (b) TTA-UC efficiencies of MOFs at various excitation intensities. (c) Excitation intensity dependence of the TTA-UC intensity on the MOFs. (d) TTA-UC emission decays of MOFs under pulsed excitation at 445 nm, measured at 380 nm. Fitting was performed using equation S4. Previous reports suggested that the UC performance depends strongly on the sensitizer–annihilator ratio. , The overall TTA-UC efficiency of a UC pair is governed by the product of (i) the singlet production probability by TTA (f) (ii) the intersystem crossing efficiency of the sensitizer (Φ ISC), (iii) the triplet energy-transfer efficiency from sensitizer to annihilator (Φ TET), (iv) the probability of TTA between annihilator triplets (Φ TTA), and (v) the fluorescence quantum yield of the upconverted singlet state (Φ F). Therefore, to quantify the TTA-UC efficiency across the measured systems, we consider the following standard equation: , ηuc=fΦISCΦTETΦTTAΦF Because TTA-UC is intrinsically a bimolecular mechanism requiring two low-energy photons to produce one higher-energy photon, the theoretical maximum UC quantum yield (Φ UC) is limited to 50%. For ease of comparison across systems, this value is typically normalized to 100% by defining the TTA-UC efficiency (ηUC) as twice the experimentally determined Φ UC. The measured ηUC for Ir-TP-MOF 50 , Ir-TP-MOF 100 , and Ir-TP-MOF 200 were 0.31%, 0.60%, and 0.09%, respectively (Figure b). Overall, the results demonstrate that the Ir-TP-MOF 100 exhibits the highest UV-light generation efficiency among the three different sensitizer concentrations. This composition provides the optimal sensitizer–annihilator balance, and deviations in either direction diminish the UC performance. The prolonged annihilator triplet lifetime τT = 1.1 ms in Ir-TP-MOF 100 (Figure d) compared with that of p-terphenyl in solution (τT = 80 μs) indicates a reduction in nonradiative relaxation and a more sustained triplet population on the annihilator, thereby promoting efficient exciton diffusion and TTA. Thus, the 100:1 ratio represents an optimal regime in which sensitizers remain sufficiently spaced to avoid self-quenching while maintaining a high enough density to produce triplet excitons at a rate supporting efficient annihilation. In contrast, Ir-TP-MOF 50 , which contains the highest sensitizer loading, exhibits a lower ηUC (0.31%) despite its relatively long triplet lifetime (1.2 ms) (Figure d). As the Ir­(dFppy-COOH) 3 strongly absorbs in the upconversion emission region, the increased reabsorption at elevated Ir­(dFppy-COOH) 3 concentration is primarily responsible for this behavior. Such behavior is consistent with the previous report. Furthermore, the measured shorter fluorescence lifetime of the MOF at the higher sensitizer concentration indicates enhanced FRET-based back energy transfer from the S1 state of the annihilator to the S1 state of the sensitizer (Figure S28). Interestingly, a significantly shorter triplet lifetime of 0.45 ms was observed for Ir-TP-MOF 200 with the lower sensitizer loading. This cannot be explained by the triplet back energy transfer, rather probably due to the existence of more triplet defect sites at the lower sensitizer loading. This short triplet lifetime reduces TTA efficiency and, as shown below, increases the threshold excitation intensity, preventing TTA saturation and resulting in the lowest ηUC of 0.09%. Power-dependent UC measurements of the THF suspensions of the MOFs reveal a clear quadratic-to-linear transition characteristic of TTA-UC behavior, with slopes of approximately 2 and 1 in the double-logarithmic plots. Although the threshold intensities (I th) of the other combinations are comparable (4.2 and 4.3 W cm–2 for Ir-TP-MOF 50 and Ir-TP-MOF 100 , respectively), the Ir-TP-MOF 200 displays a markedly higher I th (>4.6 W cm–2) (Figure c). The I th value for Ir-TP-MOF 200 could not be determined accurately because its slope did not saturate in the slope 1 region. The threshold excitation intensity can be expressed as follows: , Ith=1αΦISCΦTETγTT(τT)2 where α is the absorption coefficient and γTT is the TTA rate constant. The lower sensitizer concentration (smaller α) and short τT (0.45 ms) substantially raises I th for Ir-TP-MOF 200 (Figure d). The overall finding clearly indicates that a low sensitizer concentration generates an insufficient triplet exciton density to reach the saturated TTA regime under moderate excitation. Although Ir-TP-MOF 50 contains a higher sensitizer density, which increases the absorption coefficient (α), its I th remains comparable to that of Ir-TP-MOF 100 . The observed trend suggests that increasing sensitizer concentration may promote triplet back energy transfer, consistent with our phosphorescence lifetime measurements (λex = 405 nm, λdetection = 500 nm) attributed to the Ir-sensitizer. Collectively, these findings show that the Ir-TP-MOF 100 provides the most favorable interplay among light absorption, triplet generation and energy transfer, triplet diffusion, and annihilation dynamics within the MOF. This optimized combination produces the highest UC efficiency in the series while maintaining a threshold intensity similar to that of Ir-TP-MOF 50 . Reproducibility is a critical criterion for any UC material. Accordingly, we evaluated the reproducibility of UC performance across all UC-MOFs by conducting independent measurements under identical conditions. The results, summarized in Table S4, demonstrate consistent performance across repeated trials, thereby confirming the good reproducibility of the prepared MOFs. To know the O2-sensitivity of the TTA-UC assignment of the Ir-TP-MOF 100 , we conducted the UC experiment using O2-saturated THF under open air. However, under these experimental conditions, we have not observed any UC emission spectra, even after significantly increasing the laser power (Figure S29a). We also attempted to collect phosphorescence spectra of the Ir-complex under identical conditions, but did not observe any emission (Figure S29b). We further tried to measure the triplet lifetime of the Ir-TP-MOF 100 , and the observed decay profile was almost completely overlapped with the IRF signal (Figure S29c). All of the above control experiments indicated that the Ir-TP-MOF 100 underwent complete triplet quenching under O2-saturated conditions, further supporting the TTA-UC assignment. Despite the sufficient TET efficiency and long triplet lifetimes of the developed UC systems, the measured I th values remain relatively high. This behavior may arise from slow triplet energy migration, which is likely governed by the relative orientation and spatial arrangement of annihilators within the MOF. In addition, the close alignment of the T1 energy levels of the sensitizer and annihilator increases the probability of back energy transfer, which can adversely affect the UC performance of the MOFs. Therefore, future optimization of the MOF architecture, i.e., improving structural order, employing MOFs with closer annihilator packing, and judicious selection of chromophores, could facilitate faster triplet energy migration and help further reducing I th. Enhancing the photostability of the UC system is a key objective in developing the MOF-based sensitizer–annihilator assembly. To assess this, the Ir-TP-MOF 100 suspension was continuously irradiated with a high laser power of 49.4 W cm–2 for 30 min, during which consistent UCPL emission was observed (Figure ). Notably, the UC emission intensity was almost maintained during this continuous excitation. Furthermore, the photoluminescence spectra of the MOF suspension before and after the UC measurements were nearly identical (inset of Figure ). These results confirmed the excellent photostability of the MOF-based UC system. Furthermore, the PXRD pattern of the recovered MOF after the UC experiment retained its characteristic reflections, with a slight decrease in peak intensity, demonstrating that the framework largely retained structural integrity under irradiation conditions (Figure S30). To verify the possibility of chromophore leaching into the solution, the supernatant obtained after the UC experiment of Ir-TP-MOF 100 was collected and subjected to UC measurements at 445 nm excitation. No upconverted emission was detected, indicating that the UC-active components remain immobilized within the MOF throughout the experiment (Figure S31). The potential leaching of chromophores from the MOF was further evaluated by measuring the absorbance of the supernatant obtained after irradiating the MOF suspension at a laser power corresponding to the UC-saturated region (slope = 1). No characteristic absorbance peaks attributable to the chromophores were observed, indicating that significant leaching does not occur under the applied experimental conditions (Figure S32). Time-dependent upconversion photoluminescence (UCPL) intensity of Ir-TP-MOF 100 at 380 nm under 445 nm excitation with the laser power of 49.4 W cm–2. Inset: photoluminescence spectra of Ir-TP-MOF 100 in THF before and after the UC experiment. The significant UV light-generation capability, broad solvent compatibility, and outstanding photostability of this MOF-based UC system motivated us to explore its applicability in photocatalysis. The generation of mild, in situ UV light via TTA-UC, substantially reduces the likelihood of side-product formation compared to direct UV irradiation. In addition to reusability, the heterogeneous nature of the MOF-based UC platform enables facile separation of the desired product, unlike homogeneous UC systems. Notably, most previously reported UC-light-driven photocatalytic reactions have been conducted exclusively in homogeneous media, which poses a significant limitation for product purification. − The present heterogeneous MOF-based UC system, therefore, offers a compelling solution to these challenges. To demonstrate the photocatalytic utility of this MOF-based UC platform, we selected the photochemical uncaging of MNI-caged l-glutamate (Figure a). This chemical transformation conventionally requires direct UV irradiation and suffers from limited practical applicability due to phototoxicity and poor light penetration. In contrast, we demonstrate that TTA-UC enables efficient uncaging under visible light excitation. The strong UV absorption of MNI-caged l-glutamate, analogous to the UC-emission region, further motivated its selection as a model substrate. , Experimental details of the photocatalytic process are provided in the Supporting Information. Reaction progress was monitored by 1H NMR and UV–vis spectroscopy. (a) Reaction scheme for the uncaging of MNI-caged l-glutamate. (b) 1H NMR and (c) UV–vis absorption spectra of MNI-caged l-glutamate before and after irradiation by 445 nm laser in the presence of various reaction components (gray color in 1H NMR represents the starting material (-OMe) peak, and green color represents the corresponding peak in the product). (d) The possible reaction mechanism of MNI-caged l-glutamate uncaging reaction (ET: energy transfer) (inset: color change of the reaction mixture after reaction). The uncaging of MNI-caged l-glutamate was conducted in deoxygenated MeOH by using our best-performing UC-active MOF (Ir-TP-MOF 100 ). The measured ηUC, I th, and τT of Ir-TP-MOF 100 in MeOH were 0.06%, 6.9 W cm–2, and 0.23 ms, respectively (Figures S33–S35). The deterioration in TTA-UC performance in MeOH relative to THF is probably due to solvent-induced nonradiative deactivation of excited states. As a polar protic solvent, MeOH introduces high-energy O–H vibrational modes that may couple with the excited triplet states of the sensitizer/annihilator, thereby accelerating nonradiative decay and shortening the triplet lifetime. In contrast, THF, being a polar aprotic solvent, lacks such vibrational quenching pathways and better preserves long-lived triplet states. Additionally, hydrogen-bonding interactions between MeOH and the oxo- and hydroxo-bridges of the Zr-UiO-68 MOF framework or chromophoric units may perturb the excited-state electronic structure, further increasing internal conversion rates and reducing emission efficiency. The shortened triplet lifetime in MeOH consequently limits triplet diffusion and reduces the probability of TTA, leading to a significant decrease in ηUC. These intrinsic and persistent effects led to reduced UC performance of Ir-TP-MOF 100 in MeOH. Although the UC performance of Ir-TP-MOF 100 in MeOH was inferior to that of its THF suspension, the photochemical reaction was performed in MeOH owing to the high solubility of the substrate. Importantly, under 445 nm laser irradiation, Ir-TP-MOF 100 generates in situ UV emission in MeOH, which selectively cleaves the MNI protecting group, affording a significant color change and an isolated yield of 6.9% of uncaged MNI (Figures b-c and S36–S38, Table S5). The obtained 1H NMR peaks of uncaged MNI closely matched those of the corresponding peaks obtained after irradiation of the starting material with a 370 nm LED (Figure S38). Time-dependent studies reveal a gradual increase in product absorption, reaching a maximum after 90 min; accordingly, the reaction time was fixed at 90 min for subsequent experiments (Figures S36–S37). Control experiments clearly demonstrate that irradiation of the starting material alone, in the presence of a TP-MOF lacking upconversion emission under identical 445 nm excitation, produces no measurable yield of uncaged MNI. Furthermore, performing the reaction in the presence of only Ir­(dFppy-COOH) 3 results in a negligible yield (0.4%), thereby effectively ruling out both direct photolysis and photoredox pathways (Figures b-c, S38, and Table S5). It should be noted that since no uncaged product was detected in the NMR and UV–vis spectra for TP-MOF alone, no photochemical conversion occurred in TP-MOF. Instead, partial adsorption of the starting material occurred due to coordination of its free carboxylic group to the oxophilic Zr6 clusters of the MOF. Distinguishing adsorption from photochemical reaction based on the NMR peaks of the starting material is challenging, making it difficult to determine the exact conversion. However, considering that nearly half of the starting material is adsorbed by TP-MOF, it is reasonable that the 6.9% isolation yield obtained by irradiating Ir-TP-MOF 100 with visible light is lower than the 22.6% isolation yield under direct UV irradiation. The reusability test of recovered Ir-TP-MOF 100 after washing displayed a yield of 2% (Figure S38). The MOF-based material can be reused, providing good proof of concept and demonstrating the first use of heterogeneous UC-MOF for organic photochemical reactions. To verify the heterogeneous nature of the reaction, we performed a filtrate reactivity test. In this experiment, the uncaging reaction was carried out in the presence of Ir-TP-MOF 100 for 30 min. The solid MOF particles were then removed from the reaction mixture inside the glovebox, and the resulting supernatant was further irradiated for an additional 60 min. Notably, no appreciable increase in reaction yield was observed; the final yield (∼4.8%) remained comparable to that obtained after the initial 30 min of irradiation (Figure S39). This control experiment supports the heterogeneous nature of the reaction and rules out significant leaching of the chromophores under the reaction conditions. This experiment also suggested that the reduced uncaging yield observed upon reuse of the MOF is unlikely to arise from chromophore leaching from the framework. The framework stability during the reaction was also verified by PXRD, FT-IR, and TGA analyses of the recovered MOF sample after the catalysis reaction (Figure S40). All the aforementioned measurements supported the MOF’s stability after the reaction. Therefore, the leaching of chromophores or break of framework might not be the actual reason for low yield during the recycle test; it may be attributed to the emergence of additional nonradiative decay pathways within the UC-MOF, potentially resulting from the coordination of – COOH groups from the starting materials and uncaged products during the first catalytic cycle. Figure d shows the possible mechanism of glutamate uncaging predominantly via upconverted UV energy generated by Ir-TP-MOF 100 . The decrease in UC emission and substantial spectral overlap between the UC emission of Ir-TP-MOF 100 and the UV–vis absorption of MNI-caged l-glutamate suggests a Förster resonance energy transfer (FRET)-type mechanism from the S1 of the emitter to the low-lying S1 state of MNI-caged l-glutamate (Figures S41–S42). The coordination of the carboxylate group of MNI-caged l-glutamate to the Zr6 clusters pulls the starting material closer to the UC-MOF, thereby making the FRET energy transfer more favorable. Consistent with this hypothesis, significant quenching of the fluorescence lifetime of Ir-TP-MOF 100 from 11.05 to 2.69 ns is observed upon addition of MNI-caged l-glutamate (Figure S43). Although the observed reduction in fluorescence lifetime is consistent with a FRET-based mechanism and excludes the static quenching mechanism for the uncaging reaction, the possibility of UV photon emission from the MOF, followed by reabsorption by the substrate and nonproductive adsorption, cannot be entirely excluded. To investigate the reabsorption mechanism, we performed the uncaging reaction using a specific experimental setup, as reported by Wenger et al. in a previous study (Figure S45-a). However, no measurable product yield or characteristic color change of the MNI-caged l-glutamate solution was observed after 90 min of irradiation (Figure S45-b). This result indicates that reabsorption of UV photons is unlikely to be the operative mechanism for the uncaging reaction. To verify the existence of a nonproductive adsorption mechanism, we performed uncaging reactions at higher concentrations of MNI-caged l-glutamate. Notably, a significant decrease in product yield at a fixed reaction time was observed with increasing substrate concentration (Figure S46, Table S5). This trend suggests that nonproductive adsorption of the starting material, as well as absorption of the excitation light by the photoproducts, leading to an enhanced inner filter effect, may contribute to the reduced yield of the uncaging reaction. Among these factors, the latter is likely predominant, as the photoproducts exhibit substantial absorption at the excitation wavelength (Figure c). Collectively, these results demonstrate that upconverted photons can perform chemically productive work in a heterogeneous MOF platform, providing a UV- and mercury-free, nonphotoredox strategy for the spatiotemporally controlled cleavage of neurotransmitters, enabling broader application of MOF-based UC materials in UV-activated photochemistry. Although the present MOF-based vis-to-UV system represents the first example of UC-driven, UV-light-promoted heterogeneous photocatalysis, it still has certain limitations. Moving forward, our goal is to develop improved MOF systems that fully exploit their high surface area and well-defined pore structures for size-selective product formation. To enhance the performance of UC-MOF-based photocatalysts, several strategies may be considered: (i) selecting substrates that do not possess strongly coordinating functional groups (e.g., free carboxylic acids), (ii) engineering the MOF surface or pore environment to minimize strong substrate binding, for example by employing metal nodes such as Zn­(II) or Ni­(II), and (iii) optimizing pore size and connectivity to facilitate efficient substrate diffusion and improve accessibility to active sites. Conclusions This work demonstrates the development of a vis-to-UV TTA-UC system by integrating Ir­(dFppy-COOH) 3 sensitizer and TP-(COOH) 2 annihilator within a Zr-based MOF. The in situ incorporation strategy enabled precise control over sensitizer loading while preserving the crystallinity, porosity, outstanding physicochemical, and photostability of the framework. The rigid architecture minimized chromophore aggregation, facilitated considerable triplet energy transfer, and supported multidimensional exciton migration throughout the framework. As a result, the hybrid material exhibited efficient UC of 445 nm light to UV emission, achieving a maximum UC efficiency of 0.60% with a moderate threshold excitation intensity and a long triplet lifetime of 1.1 ms. This work demonstrates the first example of a MOF-based vis-to-UV UC system and corresponding photocatalysis, setting a new direction for designing heterogeneous UC materials. The ability to generate UV light from low-energy visible sources using a stable, solution-processable MOF platform opens new opportunities for sustainable photochemistry, enabling the uncaging reaction. More broadly, this study establishes a versatile design strategy for next-generation UC materials that addresses long-standing limitations associated with molecular diffusion, strong chromophore aggregation, and photostability in diffusion-based systems. Supplementary Material