Monolithic Porous Polymer Architectures for Visible‐to‐UV Upconversion‐Driven Photochemical Reactions

ABSTRACT

Triplet–triplet annihilation‐based photon upconversion (TTA‐UC) expands the accessible spectral window for photochemical transformations by generating high‐energy photons from low‐energy excitation. Although TTA‐UC has been widely demonstrated in solution and in various solid‐state formats, examples of polymer materials capable of efficient visible‐to‐ultraviolet (vis‐to‐UV) upconversion under low excitation intensities remain scarce. Achieving vis‐to‐UV TTA‐UC in a porous solid is particularly challenging because it requires balancing chromophore proximity for TTA with the suppression of aggregation‐induced quenching, while simultaneously preserving substrate accessibility. Here, we introduce porous polymer monoliths that enable solid‐state TTA‐UC specifically designed for heterogeneous photochemistry. By embedding sensitizer‐annihilator dye pairs into polymer matrices with controlled co‐continuous porosity, we create monolithic architectures that display upconverted UV emission from blue light at excitation intensities below a few mW cm^{− 2}, with the operational wavelength further extended to green light excitation. The porous architecture enhances energy transfer to reaction substrates, providing a versatile platform for heterogeneous photochemical processes. Owing to their robustness and tunability, these monoliths are promising candidates as heterogeneous photoreactors, opening new opportunities for scalable and sustainable photochemistry.

Porous polymer monoliths embedded with sensitizer–annihilator dye pairs enable efficient solid‐state visible‐to‐ultraviolet triplet–triplet annihilation‐based photon upconversion under low‐intensity blue and green light. Their co‐continuous porous architectures balance chromophore interactions with substrate accessibility, providing robust and tunable heterogeneous photoreactors for sustainable photochemical transformations.

1. Introduction

Photon upconversion offers a powerful approach to expand the effective spectral range of light for photochemical conversions by enabling the in situ generation of high‐energy photons [1, 2, 3, 4, 5, 6]. Among available mechanisms, triplet–triplet annihilation‐based photon upconversion (TTA‐UC) is particularly promising because it operates efficiently at low excitation intensities, such as solar and LED irradiation [7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20]. In the TTA‐UC process, a sensitizer absorbs low‐energy photons and undergoes intersystem crossing (ISC) into the triplet state. The triplet energy is then transferred to an annihilator via Dexter‐type triplet energy transfer (TET). When two annihilator triplets come into contact with each other, they undergo triplet–triplet annihilation (TTA). This process generates a higher‐energy singlet state that emits upconverted photons (Figure 1a). Despite significant progress in TTA‐UC for photochemical reactions, most reported photoreactions have been limited to homogeneous systems, including solutions, micelles, or gels, which fall short of a truly heterogeneous UC reactor [3, 5, 21, 22, 23, 24].

FIGURE 1.

(a) Energy level diagram of the TTA‐UC process mediated by triplet exciton diffusion among annihilators. ISC: intersystem crossing, TET: triplet energy transfer, TTA: triplet–triplet annihilation. (b) Schematic illustration of porous polymer monoliths doped with high concentration chromophores, showing triplet exciton diffusion and upconversion in the porous framework. (c) Chemical structures and energy levels of the blue‐light‐absorbing sensitizers CBDAC and Ir(C6)₂(acac), the green‐light‐absorbing sensitizer Ir(btb)(ppy)Cl, and the annihilator TIPS‐Nph. The T₁ energy levels of CBDAC and TIPS‐Nph were determined from the 0–0 emission peaks of the phosphorescence spectra measured at 77 K in toluene and that of Ir(C6)₂(acac) was determined from the phosphorescence peak, while that of Ir(btb)(ppy)Cl was determined from the intersection of normalized triplet absorption and emission spectra [56, 57, 59]. The S₁ energy level of TIPS‐Nph was determined from the absorption peak [56]. All energy level values were taken from previously reported literature.

Triplet energy transfer and TTA‐UC systems in the solid state, including polymeric, inorganic, and hybrid materials have been explored [25, 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, 36, 37, 38]. Among these, immobilization of sensitizer and annihilator combinations in porous media, including porous polymers, covalent organic frameworks (COFs), and metal‐organic frameworks (MOFs), potentially provides opportunities to promote the interactions between the chromophores in the solid and substrates in solution [34, 39, 40, 41, 42, 43]. Despite these advances, most reported systems operate in the visible‐to‐visible region, which limits their applicability to UV‐driven photochemistry. Furthermore, many of these materials are typically obtained as powders, nanoparticles, or soft matrices, which may complicate handling and limit mechanical robustness and scale‐up.

For practical photochemistry, the upconverting medium should be easy to handle and readily recoverable as a heterogeneous solid photocatalytic platform. Porous monolithic polymers, a class of porous polymeric materials characterized by interconnected pores and walls are promising platforms in this regard [44]. They can be synthesized in bulk and facilitate straightforward isolation of reaction products. Owing to their nanoscale porosity resulting in a large surface area, they have long been exploited as adsorbents, chromatographic columns, and catalysts [45, 46, 47, 48, 49, 50]. Although various photofunctional monoliths including titania‐ and silver‐supported systems have been reported [51, 52, 53], light harvesting is largely confined to direct visible excitation without upconverted UV output, which narrows the scope of feasible photoreactions [54]. Previously, we reported crystalline organic monoliths, which could be exploited for Dexter energy transfer from triplet sensitizers at the pore surface [55]. However, the absence of a nanoporous architecture restricts surface‐mediated energy transfer, the crystalline scaffold is mechanically fragile, and it is difficult to integrate the triplet sensitizer into the monolith framework. Together, these limitations restrict their applicability as monolithic reactors.

Here we report monolithic upconversion materials that combine a substrate‐accessible porous architecture for efficient interfacial energy transfer with scalable processability. In this fully solid monolithic material, visible‐to‐UV TTA‐UC is expected to occur via exciton diffusion among dyes that are dispersed in the polymer phase of the porous framework (Figure 1b). We incorporated the previously reported sensitizers 3,3’‐carbonylbis(7‐diethylaminocoumarin) (CBDAC) and Ir(C6)₂(acac) (C6 = coumarin 6, acac = acetylacetone) [56, 57], and the annihilator 1,4‐bis((triisopropylsilyl)ethynyl)naphthalene (TIPS‐Nph) into a monolithic polymer matrix (Figure 1c). The resulting polymer monoliths maintained their co‐continuous structures while efficiently upconverting blue (445 nm) light to the UV region (< 400 nm). Furthermore, by incorporating a recently developed triplet sensitizer Ir(btb)(ppy)Cl (btb = 1,3‐Bis(benzo[d]thiazol‐2‐yl)benzene, ppy = 2‐phenylpyridine) [58, 59], which has not yet been used in TTA‐UC, we also achieved green (532 nm)‐to‐UV TTA‐UC, thereby enabling UV‐driven photochemical reactions under green‐light irradiation. This study highlights the potential of these materials as a scalable and efficient platform for heterogeneous photocatalysis, paving the way for broader applications in sustainable photochemical transformations.

2. Results and Discussion

2.1. Synthesis of Porous Polymer Monoliths

A key strategy for fabricating porous polymer monoliths involves polymerization‐induced phase separation, where polymerization and phase separation occur simultaneously, resulting in a continuous, interconnected pore network [60]. As polymerization progresses, the initially homogeneous mixture separates into polymer‐rich and solvent‐rich phases, ultimately forming a well‐defined porous architecture. This operationally simple approach efficiently immobilizes dispersed dyes by simply blending them into the precursor solution, affording broad dye compatibility, precise control over porosity, and tunable morphology to improve interfacial energy transfer [44].

Our design goal was to create a rigid network with a composition suitable for dye loading and a maximized specific surface area to enhance the efficiency of interfacial energy transfer to substrates. In this study, we selected styrene (St) and divinylbenzene (DVB) as the monomer and the cross‐linker, respectively [49, 61, 62]. Styrene‐based polymers have excellent dye compatibility and optical transparency in the UVA region (>315 nm), where the upconverted emission occurs (Figure S4). DVB provides a well‐defined porous framework and increases the matrix rigidity [44]. Toluene and 2‐propanol were used as the porogenic solvent pair to promote preferential incorporation of chromophores into the polymer phase. We optimized the monolith composition (St: 100 µL, DVB: 100 µL, toluene: 50 µL, and 2‐propanol: 250 µL) to obtain a rigid monolith with a well‐defined porous structure (Figures 2a,b and S5–S8, Table S1–S3). Scanning electron microscopy (SEM) images confirmed an interconnected morphology with the average pore wall thickness and pore size of 0.45 ± 0.15 µm and 0.74 ± 0.16 µm, respectively (Figure 2c). The nitrogen adsorption isotherm was characteristic of a Type II profile (Figure S10), indicating predominantly macroporous (> 50 nm) characteristics [63]. Although the Brunauer–Emmett–Teller (BET) surface area of 31.6 m² g⁻¹ is relatively low, the interconnected macroporous structure provides a large interfacial contact area with the reaction medium. Such micrometer‐scale porous structures are expected to induce strong light scattering. Although this may limit photon penetration to some extent, multiple scattering within the porous framework can also increase the effective optical path length, which can be advantageous for heterogeneous photochemical applications [64].

FIGURE 2.

(a) Schematic illustration of the preparation procedure for UC monoliths. (b) Photograph of the obtained UC monoliths. SEM images of the monoliths (c) without and (d) with 15 wt% TIPS‐Nph. (e) TGA traces of polymer monoliths (gray), polymer monoliths doped with 15 wt% TIPS‐Nph (blue, solid), polymer monoliths doped with 15 wt% TIPS‐Nph after exposure to 20 mM SDS aqueous solution (blue, dashed), and TIPS‐Nph (red). (f) Normalized fluorescence spectra of different concentrations of TIPS‐Nph (1–40 wt%) in monoliths (λ _ex = 320 nm). (g) Normalized excitation spectra of TIPS‐Nph in toluene (red, solid, [TIPS‐Nph] = 10 µM), in monoliths (blue, solid, [TIPS‐Nph] = 15 wt%), and as neat powder (green, solid) (λ _em = 375 nm), and fluorescence spectra of TIPS‐Nph in toluene (red, dashed, [TIPS‐Nph] = 10 µM), in monoliths (blue, dashed, [TIPS‐Nph] = 15 wt%), and as neat powder (green, dashed) (λ _ex = 320 nm).

To quantify the pore volume of the polymer monoliths, the dried monoliths were soaked in an aqueous solution of sodium dodecyl sulfate (SDS). The surfactant was required to facilitate water uptake into the highly hydrophobic monoliths. Thermogravimetric analysis (TGA) of the water‐soaked monoliths showed a high‐water content of 60.9 wt%, which corresponds to a pore volume of 58.7 vol% (Figure S5). This substantial uptake of reaction medium is a desirable feature for heterogeneous photochemistry.

We also evaluated the dispersion of the annihilator TIPS‐Nph in the polymer matrix by analyzing its photophysical properties. To balance the suppression of aggregation with sufficient dye content for triplet energy migration and TTA, we monitored the concentration‐dependent shift in steady‐state fluorescence. When the TIPS‐Nph concentration in the precursor mixture was increased up to the solubility limit (40 wt%, where wt% is defined throughout as relative to the total mass of St, DVB, and TIPS‐Nph), we observed a broad structureless emission band around 410 nm above 20 wt%, which is attributed to aggregate formation (Figure 2f) [65]. The 0–0 emission peak around 360 nm was not observed at higher loadings due to the inner filter effect. Consistently, the powder X‐ray diffraction (PXRD) pattern for the sample with 40 wt% TIPS‐Nph showed stronger peaks attributable to crystalline TIPS‐Nph, which became more pronounced as the loading increased (Figure S11). We therefore selected 15 wt% as the optimal concentration of TIPS‐Nph in the polymer matrix.

Due to the strong light scattering arising from the macroporous structure of the monoliths, reliable UV‐vis absorption spectra could not be obtained. Therefore, we instead used excitation and fluorescence spectroscopy to assess the dispersion of TIPS‐Nph in the polymer matrix. The excitation and fluorescence peak positions of 15 wt% TIPS‐Nph embedded in the polymer monoliths were close to those observed in solution, with almost no spectral shift, and the peaks were similarly well defined (Figure 2g). In contrast, neat TIPS‐Nph powder and monoliths with higher TIPS‐Nph loading showed broad, undefined, and red‐shifted emission spectra (Figure 2f), indicating strong aggregation of TIPS‐Nph. These spectral features indicate that strong aggregation of TIPS‐Nph is suppressed in the polymer monolith with 15 wt% loading. Furthermore, 15 wt% TIPS‐Nph in the polymer monoliths retained a high fluorescence quantum yield of 68.6%, which is comparable to that under dilute conditions (74.0%) [56]. These results indicate that the 15 wt% TIPS‐Nph remained largely molecularly dispersed within the polymer matrix. After incorporating 15 wt% TIPS‐Nph, the monoliths retained the characteristic co‐continuous porous morphology. SEM showed no noticeable collapse or densification of the pore framework, and the average pore wall thickness and pore size remained comparable to those of the dye‐free monoliths (0.46 ± 0.11 µm and 0.75 ± 0.18 µm, respectively; Figure 2d). Consistently, nitrogen adsorption analysis indicated that the porous architecture was largely preserved, giving a BET surface area of 27.6 m² g^{− 1} (Figure S10). TGA of the monoliths incorporating 15 wt% TIPS‐Nph after soaking in aqueous SDS solution gave a water content of 63.2 wt%, which corresponds to a pore volume of 61.1 vol%. These results confirm that the porosity is well retained after the introduction of TIPS‐Nph (Figure 2e).

Interestingly, a nonporous bulk poly(St‐co‐DVB) synthesized without solvent and with 15 wt% TIPS‐Nph (see the SEM image shown in Figure S9) exhibited a more pronounced, red‐shifted aggregate emission in its fluorescence spectrum (Figure S12). This result suggests that the phase separation process during monolith formation enhances dye dispersion. This could be attributed to toluene swelling of the polymer‐rich phase during phase separation [66, 67]. Toluene is a good solvent for both the growing polymer network and TIPS‐Nph and therefore keeps the matrix relatively mobile by lowering the effective glass transition temperature and delaying vitrification. This favors molecular dispersion of TIPS‐Nph and suppresses nucleation of crystalline aggregates.

We further verified that the UC dyes remained intact during radical polymerization. The actual amount of TIPS‐Nph in the monolith was determined by measuring the absorbance of the ethanol solution used for solvent exchange. An initial TIPS‐Nph loading of 15.0 wt% was employed, and the final content was calculated to be about 14.8 wt%. TGA further revealed a weight loss of approximately 15 wt% above 200°C attributed to TIPS‐Nph decomposition, indicating that nearly all the doped TIPS‐Nph was successfully incorporated into the polymer matrix (Figure 2e). Furthermore, since poly(St‐co‐DVB) does not dissolve in common solvents due to its crosslinked structure, polystyrene was used instead for the dissolution experiment. After dissolving TIPS‐Nph‐doped polystyrene monoliths (without DVB) in toluene, the UV‐vis absorption spectrum remained unchanged compared with that of a TIPS‐Nph solution (Figure S13).

2.2. TTA‐UC Properties

2.2.1. Blue‐to‐UV Photon Upconversion

We evaluated the TTA‐UC properties of the monoliths using either CBDAC or Ir(C6)₂(acac) as sensitizers. Both sensitizers exhibit strong visible‐light absorption, weak UV absorption, high ISC efficiency, and triplet energy levels higher than those of TIPS‐Nph, and therefore provide excellent visible‐to‐UV TTA‐UC with TIPS‐Nph [56, 57]. Upon irradiation of the poly(St‐co‐DVB) monoliths doped with the annihilator TIPS‐Nph (15 wt%) and sensitizer CBDAC (2.0 mol% relative to TIPS‐Nph) or Ir(C6)₂(acac) (0.5 mol% relative to TIPS‐Nph) using a 445 nm continuous‐wave (cw) blue laser under an argon atmosphere, a distinct upconverted emission appeared in the UV region, peaking at 375 nm for monoliths prepared with either sensitizer (Figure 3a,b). As evidence for the sensitized TTA‐UC mechanism, no upconverted photoluminescence was observed for monoliths without the sensitizer. The sensitizer concentration was optimized to maximize the TTA‐UC quantum yield (Figure S20).

FIGURE 3.

TTA‐UC studies of monoliths with 15 wt% TIPS‐Nph and 2.0 mol% CBDAC (red) and 0.5 mol% Ir(C6)₂(acac) (blue) relative to TIPS‐Nph under Ar (λ _ex = 445 nm, 425 nm short‐pass filter). Upconversion photoluminescence (UCPL) spectra of the monoliths doped with (a) CBDAC/TIPS‐Nph and (b) Ir(C6)₂(acac)/TIPS‐Nph at different excitation intensities. (c) TTA‐UC quantum yield (theoretical maximum: 50%) of the monoliths doped with CBDAC/TIPS‐Nph (red squares) and Ir(C6)₂(acac)/TIPS‐Nph (blue circles) at different excitation intensities. (d) Excitation intensity dependence of UCPL at 375 nm on a double‐logarithmic scale for the monoliths doped with CBDAC/TIPS‐Nph (red squares) and Ir(C6)₂(acac)/TIPS‐Nph (blue circles).

The TTA‐UC quantum yield (Φ_UC; theoretical maximum = 50%, see Supporting Information for details of the definition) was determined to be 0.24 ± 0.03% for the monoliths doped with CBDAC/TIPS‐Nph and 0.64 ± 0.05% for those doped with Ir(C6)₂(acac)/TIPS‐Nph (Figure 3c). These values were obtained using a relative method [68], which employed the photoluminescence of the sensitizer as an internal standard because the strong scattering of the polymer monoliths hampered precise absorbance measurements. The triplet energy transfer efficiency (Φ_TET) from Ir(C6)₂(acac) to TIPS‐Nph was determined to be 64% from the phosphorescence quantum yield of Ir(C6)₂(acac), whereas that from CBDAC to TIPS‐Nph could not be determined because CBDAC does not exhibit phosphorescence and strong scattering hindered the transient absorption measurements.

Singlet energy transfer and/or reabsorption can be a detrimental energy loss pathway in solid state TTA‐UC [69]. To evaluate the possibility of singlet energy transfer from the excited singlet state of the annihilator to the sensitizer, the fluorescence lifetime of TIPS‐Nph was investigated in the absence and presence of sensitizers (Figure S15) [34]. The fluorescence lifetime of TIPS‐Nph in the absence of sensitizer was 7.48 ns, while it slightly decreased to 6.65 ns in the presence of CBDAC. A similar trend was observed for Ir(C6)₂(acac), yielding a lifetime of 6.75 ns. The singlet quenching efficiency was estimated using Φ_bET =  1 − (τ_FL/τ_FL,0), where τ_FL and τ_FL,0 represent the fluorescence lifetime of TIPS‐Nph in the presence and absence of sensitizer, respectively. The calculated values were 9.8% for CBDAC and 11.2% for Ir(C6)₂(acac), indicating that this process is unlikely to be the primary factor responsible for the difference in TTA‐UC quantum yields between the sensitizers.

The threshold excitation intensity (I _th) is another critical parameter of TTA‐UC systems. TTA‐UC typically transitions from a two‐photon (quadratic) regime at low excitation intensities, where triplet deactivation dominates over TTA, to a pseudo‐one‐photon (linear) regime at higher intensities, where triplet consumption by TTA becomes dominant [70, 71, 72]. The I _th value is defined as the intensity at which this transition occurs and is estimated from the intersection of the fitted lines on a double‐logarithmic plot of UC emission intensity versus excitation intensity. The UC monoliths exhibited the expected quadratic‐to‐linear transition and notably low I _th values of 1.50 ± 0.09 mW cm⁻² with CBDAC/TIPS‐Nph and 2.13 ± 0.33 mW cm⁻² with Ir(C6)₂(acac)/TIPS‐Nph (Figure 3d). Although a strictly quantitative comparison with the corresponding solution systems is not straightforward because the dye concentrations are substantially different, the I _th values obtained for the present monoliths are lower than those reported for the corresponding blue‐to‐UV TTA‐UC systems in solution [56, 57]. These values are also comparable to the solar irradiance of 1.4 mW cm⁻² at 445 ± 5 nm [73], underscoring the potential for solar energy harvesting applications.

Based on the upconversion photoluminescence (UCPL) decays on a millisecond time scale, the triplet lifetimes were 6.21 ms for CBDAC/TIPS‐Nph and 3.90 ms for Ir(C6)₂(acac)/TIPS‐Nph (Figure S16) at the respective optimized concentrations, notably longer than those in toluene [56, 57]. This elongation is likely due to restricted molecular vibrations of TIPS‐Nph within the rigid polymer matrix, contributing significantly to lowering I _th [26, 74].

Although the triplet lifetime of the annihilator is, in principle, an intrinsic property of the annihilator, the fitted values differ between the two monoliths. To understand the origin of this difference, the performance of the two sensitizer‐based monoliths was compared in terms of the relevant parameters (Table S5).

As expected, the CBDAC/TIPS‐Nph monoliths, which possess a longer triplet lifetime, exhibited a lower I _th than the Ir(C6)₂(acac) monoliths. However, while ${\tau }_{\mathrm{T}}^2$ differs by approximately 2.5‐fold, the difference in the observed I _th is about 1.4‐fold. Because the TTA rate constant is expected to be essentially identical for both sensitizer systems, the difference in I _th can be discussed in terms of absorption coefficient, τ_T, and TET efficiency (see SI for details). Although precise absorbance determination is difficult due to strong scattering, the apparent absorbance at 445 nm obtained in an integrating sphere was slightly higher, 0.073 for CBDAC and 0.064 for Ir(C6)₂(acac) in the monoliths. Therefore, the optical absorption coefficient (α) does not explain the relatively low I _th observed for Ir(C6)₂(acac)/TIPS‐Nph monoliths. Instead, considering the shorter triplet lifetime of TIPS‐Nph in the Ir(C6)₂(acac) system, this result suggests that the TET efficiency is likely higher than that of the CBDAC system. This higher TET efficiency compensates for the faster deactivation, thereby maintaining a sufficient steady‐state triplet population to achieve the relatively low I _th. That is also consistent with the higher TTA‐UC quantum yield. Assuming a higher TET efficiency from Ir(C6)₂(acac) to TIPS‐Nph, Ir may introduce nonradiative deactivation of nearby annihilator triplets through external spin‐orbit coupling by the heavy central atom or additional triplet pathways arising from local proximity and/or intermolecular interactions, leading to the shorter apparent τ_T [75, 76].

Consistent with this interpretation, analysis of the time‐resolved UC emission gave β values that were closer to unity for Ir(C6)₂(acac) (0.997) than for CBDAC (0.964) under the same excitation intensity (Figure S16). Following the definition, β is a dimensionless parameter describing the fraction of annihilator triplets that initially decay via TTA and therefore reflects the TTA efficiency [77]. The larger β value for Ir(C6)₂(acac)/TIPS‐Nph suggests a higher effective triplet population, indicating that the initial triplet decay is more strongly governed by TTA. This trend is also consistent with the higher TTA‐UC quantum yield observed for the Ir(C6)₂(acac)/TIPS‐Nph system and provides a plausible explanation for the relatively low threshold excitation intensity.

We should note that UC monoliths have limited photostability. Under continuous irradiation at 120.1 mW cm⁻² for 1 h, the UCPL intensity gradually decreased to 40% of its initial value (Figure S17). For reliable use of these materials in photochemical reactions, further enhancement of their long‐term photostability will be essential. The gradual decrease in UC emission intensity is likely associated with irreversible photochemical degradation of the embedded dyes under continuous irradiation. In the solid monolith, the sensitizer and annihilator are immobilized in the polymer phase and are repeatedly exposed to excitation, whereas photodegraded dyes cannot be replaced by diffusion of intact molecules as in solution. In addition, higher excited states generated during visible‐to‐UV TTA‐UC may promote irreversible reactions involving the embedded dyes and their local environment [78, 79]. The triplet lifetime of TIPS‐Nph in the monolith reaches the millisecond regime; thus, long‐lived TIPS‐Nph triplet states may undergo irreversible photochemical reactions with nearby species, including locally aggregated dyes or the surrounding polymer matrix, leading to the formation of non‐emissive degraded species. Therefore, minimizing local dye aggregation and optimizing the combination of polymer matrix and embedded dyes could improve the photostability of the UC monoliths.

Furthermore, the TTA‐UC performance of the monolith doped with Ir(C6)₂(acac)/TIPS‐Nph, which showed the highest UC quantum yield among the examined monoliths, was evaluated under aqueous conditions in the presence of SDS. Notably, the UC quantum yield remained essentially unchanged (Φ_UC = 0.83%), while the threshold excitation intensity increased to I _th = 17.8 mW cm^{− 2}, even after immersion in degassed water for 6 h. These results suggest that the porous poly(St‐co‐DVB) monolith provides a stable hydrophobic microenvironment for the sensitizer/annihilator pair (Figures S18 and S19). Such water compatibility represents a significant advantage over conventional solution‐based TTA‐UC systems, for which the range of UC dyes compatible with aqueous media remains very limited and UC quantum yields are generally low (Table S4) [2, 80, 81, 82, 83]. This highlights the potential applicability of these UC monoliths in aqueous photochemical systems.

We also compared the TTA‐UC performance of porous monoliths with nonporous bulk polymers (Figure S21). As described above, the phase separation process during monolith preparation enhances dye dispersion and suppresses aggregation. In contrast, significant aggregation was observed in the bulk polymer at equivalent dye concentrations, which shortened the triplet lifetime (Figure S22). As a result, the bulk polymers showed substantially higher I _th, exceeding 85.3 mW cm⁻² for CBDAC/TIPS‐Nph and 227.3 mW cm⁻² for Ir(C6)₂(acac)/TIPS‐Nph. Furthermore, obvious photodegradation was observed before the linear regime was reached for both sensitizers, implying that the true I _th values exceed these apparent estimates (Figure S23). These results indicate that improved molecular dispersion of the dyes within the polymer matrix plays a crucial role in reducing aggregation, extending triplet lifetimes, and enabling operation at reduced excitation intensities.

2.2.2. Green‐to‐UV Photon Upconversion

To utilize longer wavelengths for photochemical transformations, we synthesized and employed a green‐light‐absorbing sensitizer, Ir(btb)(ppy)Cl [58, 59], in combination with the annihilator TIPS‐Nph. Ir(btb)(ppy)Cl shows a relatively strong singlet‐to‐triplet (S‐T) absorption band at 575 nm with a molar extinction coefficient of 6.7×10² M⁻¹ cm⁻¹. It has been reported that the use of an S‐T absorption sensitizer is a powerful strategy to enlarge the anti‐Stokes shift by circumventing the energy loss associated with ISC [84, 85, 86]. The triplet energy level of Ir(btb)(ppy)Cl (2.16 eV) is high enough to sensitize TIPS‐Nph (2.12 eV) (Figures S24 and S25). Compared with Ir(C6)₂(acac) and CBDAC, Ir(btb)(ppy)Cl absorbs further into the green, thereby affording a larger anti‐Stokes shift and broadening the accessible excitation window for TTA‐UC. In deaerated THF under 532 nm cw laser irradiation, the Ir(btb)(ppy)Cl/TIPS‐Nph pair showed UV upconversion emission (Figure S26). The efficiency of this green‐to‐UV TTA‐UC was estimated to be Φ_UC = 7.1 ± 0.3%. The anti‐Stokes shift was 0.98 eV, defined as the energy gap between the excitation photon energy (532 nm) and the maximum UC emission (375 nm) [87]. Previous work reported record anti‐Stokes shifts of 0.98 eV with Φ_UC = 3.2% and 0.91 eV with Φ_UC = 3.8% for visible‐to‐UV upconversion [21]. Our result demonstrates a comparably large anti‐Stokes shift of 0.98 eV with a higher Φ_UC of 7.1%. Furthermore, no noticeable decrease in the upconverted emission intensity was observed even after 1 h continuous irradiation at an excitation light power density as high as 11.7 W cm⁻², demonstrating the high photostability of the system (Figure S27).

When embedded in the monoliths, the TTA‐UC properties of Ir(btb)(ppy)Cl/TIPS‐Nph were partially retained (Figure S28). Upon 532 nm green cw laser irradiation under Ar, upconverted UV emission was clearly detectable, although with a reduced UC quantum yield (Φ_UC = 1.5 × 10⁻²%). Relative to solution, the triplet lifetime was prolonged in the monolith to 1.73 ms, corresponding to an 11‐fold elongation relative to deaerated THF containing 100 µM Ir(btb)(ppy)Cl and 10 mM TIPS‐Nph (τ_T = 155 µs). This prolongation was accompanied by a lower I _th = 918.2 mW cm⁻², compared with I _th = 4.9 W cm⁻² in deaerated THF. However, as in the case of the blue‐to‐UV dye pairs, the green‐to‐UV UC monolith exhibited a gradual decrease in UC emission intensity under continuous irradiation (Figure S29), suggesting that the solid polymer environment can affect the photostability of embedded dye pairs. Despite this current limitation, these results still suggest that this system could serve as a promising solid‐state platform for UV‐driven photochemical reactions under green light excitation.

2.3. Photochemical Reactions with UC Monoliths

As a proof of concept for heterogeneous UV‐driven photochemical reactions, we examined photochemical reactions using the UC monoliths, taking advantage of their ease of handling and recoverability. The monoliths can be cut into a desired size, directly immersed in the reaction solution, and removed after irradiation using tweezers.

CMNB‐caged fluorescein (fluorescein bis(5‐carboxymethoxy‐2‐nitrobenzyl) ether) was selected as a model substrate to demonstrate that upconverted UV photons can trigger bond cleavage in aqueous media (Figure 4a) [6, 22]. CMNB‐caged fluorescein displays an absorption band in the UV region (<400 nm), providing good spectral overlap with the UC emission maximum of our monoliths, while the released fluorescein affords a bright readout in the green region (Figure S32) [88]. Under 525 nm LED irradiation, solutions containing CMNB‐caged fluorescein and monoliths doped with Ir(btb)(ppy)Cl/TIPS‐Nph showed a 5.6‐fold increase in fluorescein emission at its maximum. In contrast, no significant increase was observed in the absence of either TIPS‐Nph or the monoliths (Figure 4b), confirming that photouncaging is driven by the UC process. Under shorter‐wavelength irradiation at 467 nm, using monoliths doped with Ir(C6)₂(acac)/TIPS‐Nph, a 32‐fold larger fluorescence enhancement was observed than that in the absence of monoliths (Figure S33). This enhancement may be associated with the higher TTA‐UC quantum yield and lower threshold excitation intensity of the sensitizer system under the experimental conditions, although a direct quantitative comparison between different excitation wavelengths is not straightforward.

FIGURE 4.

(a) Photouncaging of CMNB‐caged fluorescein by upconverted UV light generated in the monoliths. (b) Fluorescence spectra of released fluorescein after 6 h irradiation with a 525 nm LED: before irradiation (light gray), after 525 nm irradiation without monoliths (light blue), after 525 nm irradiation in the presence of Ir(btb)(ppy)Cl‐only monoliths (blue), and after 525 nm irradiation in the presence of Ir(btb)(ppy)Cl/TIPS‐Nph monoliths (dark blue) (λ _ex = 490 nm). The solutions containing CMNB‐caged fluorescein (13 µM) and SDS (20 mM) in phosphate buffer (10 mM, pH 7.4) were degassed by five freeze‐pump‐thaw cycles prior to irradiation. After irradiation, the monoliths were removed and all spectra were recorded using the same instrument sensitivity settings. The fluorescence intensity after irradiation with Ir(btb)(ppy)Cl/TIPS‐Nph monoliths (dark blue) was 5.6‐fold higher than that obtained without monoliths (light blue) at the emission maximum (λ _em = 524.4 nm). The three control spectra (before irradiation, after 525 nm irradiation without monoliths, and after 525 nm irradiation with Ir(btb)(ppy)Cl‐only monoliths) show nearly identical fluorescence profiles and thus appear overlapped.

We further showcased the utility of the UC monoliths in a heterogeneous photochemical reaction by performing the visible‐to‐UV UC‐assisted debromination of 2‐bromopyrazine (2‐BP) using 1,3,6,8‐pyrenetetrasulfonic acid tetrasodium salt (PTS) as a singlet energy acceptor (Figure 5a) [89]. The absorption spectrum of PTS overlaps well with the upconverted emission from TIPS‐Nph but shows negligible absorption in the visible region (Figure 5b). Thus, the high S₁ state energy of TIPS‐Nph can be harnessed to generate the excited state of PTS via Förster resonance energy transfer (FRET). Upon 525 nm LED irradiation for 6 h, 2‐BP was converted to the debrominated product in 6.1% yield in the presence of monoliths containing both Ir(btb)(ppy)Cl and TIPS‐Nph, as confirmed by shifts in the ¹H NMR spectra (Figures 5d, S34, S35, and Table S6). A control experiment using sensitizer‐only monoliths (without the annihilator) gave a much lower yield (0.6%), approximately one‐tenth of that obtained with both dyes, and omission of PTS afforded no detectable product.

FIGURE 5.

(a) Schematic illustration of the debromination of 2‐BP using a TTA‐UC system (Ir(btb)(ppy)Cl as the sensitizer and TIPS‐Nph as the annihilator) with PTS as an interfacial FRET acceptor. Reaction conditions: 10 mM 2‐BP, 12 mM PTS, 10 mM sodium L‐ascorbate (NaHAsc), 100 mM dodecyltrimethylammonium chloride (DTAC) in D₂O after five freeze‐pump‐thaw cycles. (b) UV‐vis absorption and normalized emission spectra. UV‐vis absorption spectra of 250 µM Ir(btb)(ppy)Cl in THF (orange) and 12 mM PTS in 100 mM DTAC solution (blue), UCPL spectrum of the monoliths doped with Ir(btb)(ppy)Cl/TIPS‐Nph (purple line) under 532 nm cw laser irradiation after five freeze‐pump‐thaw cycles with a 510 nm short‐pass filter, and 525 nm LED lamp spectrum (green). The spectral overlap between the UC emission and PTS absorption is highlighted with a black border. (c) Fluorescence lifetime of monoliths doped with 15 wt% TIPS‐Nph in 100 mM DTAC solution (black) and in 12 mM PTS and 100 mM DTAC solution (blue) (λ _ex = 320 nm, λ _det = 360 nm). The decay corresponds to a FRET efficiency of 57.4%. (d) Debromination product yield determined by the integration of the ¹H NMR (500 MHz D₂O).

A control experiment in homogeneous aqueous solution with the UC pair used in the monoliths that showed the 6.1% yield is not feasible. This is due to limited solubility in aqueous conditions and potential mechanistic shifts, such as direct triplet sensitization that would yield the PTS radical anion in a one‐photon reaction in homogeneous solution in the presence of ascorbate [89]. However, benchmarking against the only homogeneous TTA‐UC system that is known for initiating dehalogenation reactions in water—the [Ru(bpy)₃]²⁺‐anthracene‐9‐propionate pair—demonstrated that our heterogeneous monoliths provide significantly higher debromination yields (6.1% vs. 1.8%, see Section 4.5 of the Supporting Information for details) [81]. Beyond efficiency, which is always a problem for homogeneous TTA‐UC systems in water (see also Table S4) [2, 80, 81, 82, 83], the heterogeneous approach offers superior selectivity and simplified purification, as the immobilized components prevent complex product mixtures and eliminate the need for separating annihilator decomposition products.

We also confirmed that dye leaching from the monolith was negligible under the present reaction conditions. In the ¹H NMR spectra of the reaction solution after photocatalysis, no obvious peaks derived from the embedded sensitizer or annihilator were detected (Figures S34, S35). Furthermore, a dye‐leaching test using a highly TIPS‐Nph‐loaded monolith showed no detectable dye‐derived absorption in the supernatant after immersion in the reaction solvent for 6 h (Figure S14). Although the reactivity needs to be optimized in further studies, these considerations represent an important proof‐of‐concept regarding the usage of solid visible‐to‐UV upconversion materials in heterogeneous photocatalytic applications. With further control over pore structure, dye localization near the pore surface, and photon penetration, the monolithic platform could offer a versatile route toward recoverable solid upconversion photoreactors.

In the absence of PTS, no significant debrominated product was observed either with monoliths doped only with Ir(btb)(ppy)Cl or with Ir(btb)(ppy)Cl/TIPS‐Nph. These observations indicate that direct electron transfer from ³Ir(btb)(ppy)Cl*, ³TIPS‐Nph*, or ¹TIPS‐Nph* to 2‐BP is not the dominant pathway for the photocatalytic debromination of 2‐BP.

In principle, the triplet energy of Ir(btb)(ppy)Cl (2.16 eV) exceeds that of PTS (1.85 eV) [89], so direct triplet energy sensitization of PTS by Ir(btb)(ppy)Cl in the monoliths could be thermodynamically allowed. To elucidate the mechanism, we performed time‐resolved photoluminescence measurements of the monoliths. The triplet lifetime of Ir(btb)(ppy)Cl in the monolith was essentially unchanged upon addition of PTS (Figure S36), suggesting that direct triplet sensitization of PTS is inefficient. The triplet lifetime of TIPS‐Nph in the monoliths doped with Ir(btb)(ppy)Cl/TIPS‐Nph slightly shortened after soaking in a PTS solution, implying that the inefficient sensitization of PTS (18.0%) may occur via the longer‐lived ³TIPS‐Nph* species.

In contrast, the fluorescence lifetime of TIPS‐Nph in the monolith (without Ir(btb)(ppy)Cl) clearly decreased in the presence of PTS (Figure 5c). The decay profiles were well described by a biexponential model both in the absence and presence of PTS, reflecting the heterogeneous local environments in the polymer matrix, such as interfacial vs bulk regions, as well as a minor fraction of aggregated species suggested by the steady‐state spectra. Upon addition of PTS, both lifetime components became shorter. The present measurements cannot quantitatively distinguish between nonradiative Förster resonance energy transfer and radiative reabsorption of the upconverted UV emission by PTS. Nevertheless, the observed lifetime shortening supports the presence of a short‐range nonradiative energy‐transfer pathway [90]. Based on the change in the intensity‐averaged lifetime, calculated using the weighted contributions of each decay component, the apparent energy transfer efficiency was estimated to be 57.4% using the relationship Φ_ET =  1 − (τ_FL/τ_FL,0), where τ_FL and τ_FL,0 represent the fluorescence lifetimes of TIPS‐Nph in the presence and absence of PTS, respectively. At the same time, photon‐mediated transfer may also contribute to the overall energy transfer process in this scattering porous system.

The relatively large pore size of the monoliths (0.75 µm ± 0.18 µm) is sufficient to allow facile diffusion of small molecules such as PTS into the internal porous network. The molecular dimensions of PTS are much smaller than the macropore size. In addition, the monoliths exhibit a high water content (63.2 wt%) after soaking, indicating that the pores are readily filled with aqueous solution. Therefore, the pore size itself does not limit the donor (TIPS‐Nph) and acceptor (PTS) proximity required for FRET. Although we did not quantify the interfacial distribution of PTS, control experiments using a bulk (nonporous) polymer showed that the fluorescence lifetime of TIPS‐Nph was less shortened by PTS than in the porous monoliths (Figure S37). Therefore, the porous architecture increases the probability that acceptor molecules approach within the Förster radius of TIPS‐Nph, thereby enhancing the overall transfer efficiency.

Although the current reaction yield remains modest, further improvements in photostability, photon penetration, and TTA‐UC quantum yield may be achieved by controlling the monolith thickness, optimizing the pore size and pore‐wall thickness, and localizing the sensitizer–annihilator pairs near optically and substrate‐accessible pore surfaces. Overall, these considerations provide a useful design principle for employing solid‐state upconversion materials as an effective energy‐transfer platform in heterogeneous photocatalytic applications.

3. Conclusion

In summary, we have demonstrated visible‐to‐UV TTA‐UC in porous polymer monoliths and their applications in photochemical reactions. By embedding sensitizer–annihilator pairs within a rigid poly(St‐co‐DVB) framework, the monoliths exhibit upconverted UV emission and can be exploited for photochemical reactions via interfacial energy transfer. The system thus bridges the gap between homogeneous TTA‐UC systems and practical heterogeneous materials, offering a solid‐state platform that can convert visible light into usable UV energy. Although the TTA‐UC quantum yield, photostability, and reaction efficiency still require further improvement, this design concept provides a robust foundation for the development of durable and scalable upconverting materials. With such advances, these monoliths could be integrated into flow photochemical systems and solar‐driven reactors. Overall, the present study represents an important step toward sustainable photochemistry by enabling efficient utilization of visible light to drive energy‐demanding chemical transformations.

Author Contributions

Nobuhiro Yanai: conceptualization, supervision, writing – review and editing, writing – original draft, funding acquisition, methodology. Masanori Uji: validation, investigation, writing – review and editing, methodology, funding acquisition. Naoto Matsumoto: validation, investigation, writing – review and editing, methodology, funding acquisition. Sakura Nakagawa: investigation, writing – original draft, methodology, writing – review and editing. Maria–Sophie Bertrams: writing – review and editing, investigation, methodology. Christoph Kerzig: supervision, funding acquisition, writing – review and editing, methodology.

Conflicts of Interest

The authors declare no conflicts of interest.

Supporting information

Supporting File: anie73911‐sup‐0001‐SuppMat.pdf.

Data Availability Statement

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