Multitopic acid modulation unlocks a new dimensionality in defect-engineered MOFs Authors: Ming-Shuo Sun, Yu Gai, Guo-Ying Han, Tong Liu, Pascal Van Der Voort, Xiao Feng Published: 2026 ABSTRACT Defect engineering is a powerful strategy to tailor the physicochemical properties of metal–organic frameworks (MOFs). Among various approaches, modulation synthesis – typically employing monotopic modulators – has emerged as the most widely used method to introduce structural defects. However, conventional modulation strategies aiming for high defect densities typically rely on using large quantities of monocarboxylic acids and/or strong acidity. Herein, we introduce a third dimension to modulate the defects of MOFs by employing multitopic carboxylic acids as modulators, wherein the topicity of the modulator exceeds that of the pristine linker. Using UiO-66 as a model system, multitopic modulators with varying topic numbers and carboxylate positions significantly influence defect incorporation. In particular, ortho-substituted acids and higher-topic modulators efficiently generate higher defect densities at much lower dosages without relying on harsh acidic conditions. Catalytic epoxide ring-opening studies demonstrate that defect engineering enhances activity by creating accessible active sites. This establishes multitopic modulation as a tunable strategy for controlling MOF structure–function relationships. Introduction Metal–organic frameworks (MOFs) are a class of porous crystalline materials constructed from metal nodes and organic linkers, known for their high surface area, tunable structures, and diverse functionalities.1–4 In recent years, defect engineering – a strategy that intentionally introduces structural imperfections such as missing linkers or clusters – has emerged as a powerful tool to tailor the properties of MOFs.5–8 These engineered defects can enhance chemical reactivity, improve mass transport, and create new active sites, thereby broadening the scope of MOF applications in areas such as catalysis, gas separation, and sensing.9–14 Among the various strategies for defect engineering in MOFs—including de novo synthesis and post-synthesis—coordination modulation synthesis has emerged as the most widely adopted approach because defects can be introduced directly during crystallization while largely preserving the framework topology and synthetic simplicity.15,16 Introduced by Kitagawa and co-workers in 2009, the concept of “coordination modulation” involves the addition of monocarboxylic acids as modulators during MOF synthesis.17 These modulators compete with linkers for coordination at the metal nodes, thereby slowing the metal-linker assembly process and improving crystallinity.18–21 This approach has been initially employed to control crystal size and morphology.22–25 Subsequent studies revealed that when modulators remain coordinated to the metal nodes without being displaced, they can occupy coordination sites and deliberately generate defects within the framework structure.26–32 Common modulators include monocarboxylic acids such as formic acid (FA),33 acetic acid (AA),34 difluoroacetic acid (DFA), trifluoroacetic acid (TFA),35–37 benzoic acid (BA),38,39 citric acid (CA)40 and proline.41 Shearer and colleagues systematically investigated two key factors that influence defect formation – the concentration and acidity (pKa) of the modulator – demonstrating that higher amounts and stronger acidity typically lead to increased defect incorporation.35,38 In defect-engineered MOFs, modulators are often structural analogues or partial fragments of the linkers, designed to compete with metal nodes for coordination. Their topicity is usually lower than that of the linkers of MOFs. For example, UiO-66—one of the most extensively investigated platforms for defect engineering—comprises Zr6O4(OH)4 secondary building units coordinated by twelve 1,4-benzenedicarboxylate (BDC) linkers,42 whereas commonly employed modulators for this system are monotopic carboxylic acids such as formic acid and acetic acid. In MUV-10, which uses a tritopic linker, modulators such as ditopic isophthalic acid or monotopic benzoic acid have been employed.43,44 However, conventional modulation strategies often depend on elevated modulator amounts to achieve noticeable defect levels, as exemplified by the use of excess formic acid or stronger competitors like trifluoroacetic acid (TFA) to generate defects in UiO-66.16,45 Such conditions raise concerns related to environmental impact and synthetic cost. More fundamentally, conventional modulation remains almost exclusively dependent on tuning modulator quantity and acidity. Although increasing the amount of monotopic acid promotes defect formation, it also increases proton concentration, suppresses linker deprotonation, and slows framework assembly. Consequently, defect generation is intrinsically coupled with crystallization kinetics, leading to a trade-off between defect incorporation and structural evolution. These limitations raise an important question: can a fundamentally different molecular parameter be introduced into modulation chemistry to regulate defect formation? Unlike conventional monotopic modulation, which regulates defect formation indirectly through acidity and concentration, topicity modulation directly controls the coordination competition between modulators and framework linkers by varying the intrinsic coordination capacity of the modulator molecule.46–49 Instead of further varying the quantity or acidity of monotopic modulators, we propose manipulating the intrinsic coordination capacity of the modulator itself. Specifically, we introduce multitopic modulators whose topicity exceeds that of the parent linker, establishing a new modulation dimension that is fundamentally different from that of conventional monotopic modulation (Fig. 1). This strategy introduces a new dimension of control in modulator design beyond conventional parameters such as quantity and acidity. Using UiO-66 as a prototype, small amounts of multitopic acids generate a high density of defects, eliminating the need for large modulator excesses or highly corrosive acids reported previously.16,50–52 Comparison of the new dimension of modulation in this work with two conventional dimensions. Defect introduction is further influenced by both the spatial orientation and the topicity of the carboxylate groups: ortho-substituted acids promote defect formation more efficiently than their meta- or para-substituted counterparts, and modulators with higher topicity tend to induce defects more readily. Catalytic tests with the styrene-oxide ring-opening reaction reveal an activity trend in which moderate defect concentrations enhance performance by exposing open metal sites, whereas excessive defects diminish activity because of framework collapse and restricted site accessibility. These findings establish topicity as a fundamental molecular design parameter for modulation synthesis, providing a general strategy for rational defect engineering beyond conventional monotopic modulation. Results and discussion Characterization To explore the topicity of modulators (refers to the number and spatial arrangement of coordinating groups) as a third dimension in defect engineering, the proposed topicity modulation strategy was explored using a series of multitopic acids as modulators. These included phthalic acid (2O), isophthalic acid (2M), 1,2,3-benzenetricarboxylic acid (3O), 1,3,5-benzenetricarboxylic acid (3M), and 1,2,4,5-benzenetetracarboxylic acid (4OM) (Table S4). The resulting MOFs were named UiO-66-nX-Y, where n denotes the topicity of the modulator and X indicates the relative position of the carboxylic group: O stands for ortho and M stands for meta and Y represents the ratio of the modulator to the linker. This selected modulator is used to investigate systematically how both the topicity and spatial arrangement of modulator carboxylate groups influence defect introduction. Benzoic acid, a monotopic and studied modulator, was used as a reference. By varying the amount of each modulator, the defect amount was evaluated to compare the effect of multitopic modulators in defect generation. The crystallinity of the MOF samples was evaluated by powder X-ray diffraction (PXRD), while defect density was quantified via thermogravimetric analysis (TGA) (Fig. S1–S4).35,53,54 PXRD confirmed that UiO-66 could be synthesized in the absence of any modulator. When isophthalic acid (2M) was employed as a multitopic modulator, the addition of only 0.2 equivalents significantly increased the defect density, reaching 1.0 defects per Zr6 cluster. The modulation behavior displayed strong sensitivity to the concentration of the modulator; increasing the dosage to 0.35 equivalents raised the defect density to 2.2 defects per cluster—comparable to values achieved using a large excess of conventional monocarboxylic acid modulators.16,35,38,52,55,56 PXRD patterns revealed a gradual loss of crystallinity with increasing amounts of isophthalic acid (Fig. 2a). At 0.35 equivalents, the characteristic diffraction peaks of UiO-66 were barely retained, and further increases in modulator concentration led to the disappearance of these peaks, indicating a breakdown of long-range order. This trend aligns with previous findings that excessive defect incorporation often compromises MOF crystallinity.9,57,58 Notably, despite further increases in modulator concentration beyond 0.35 equivalents, the defect density plateaued at approximately 2.1 defects per cluster (Fig. 2b). This suggests that the framework of UiO-66 has a structural limit in terms of defect tolerance, beyond which additional modulators cannot be incorporated into the framework. (a) PXRD and (b) TGA curves of UiO-66 modulated with isophthalic acid (2M). (c) FT-IR spectra and (d) Raman spectra of UiO-66-2M-X series samples. (e) Nitrogen adsorption isotherms obtained on UiO-66-2M samples at 77 K. (f) TGA of UiO-66 samples modulated with low concentrations of monocarboxylic acids. In comparison, benzoic acid – used as a monotopic modulator control – had a negligible effect on defect density under the same low-equivalent conditions. Notably, achieving a similar level of defect incorporation with benzoic acid has been previously reported to require over 50 equivalents (Fig. S5).38 Fourier-transform infrared spectroscopy (FT-IR) of UiO-66-2M-X shows a pronounced red shift of the µ3-O stretching vibration from 680 cm−1 (pristine UiO-66) to lower wavenumbers, indicating deformation of the Zr6O4(OH)4 cluster arising from missing-linker defects (asymmetric vibrations of Zr–O–Zr bonds) (Fig. 2c and S6–S9).59 Raman spectroscopy measurements were conducted on the 2M sample series, 2M modulator and BDC linker. Comparative Raman analysis revealed a distinct peak at 1003 cm−1 in 2M-modulated samples and 2M modulator, absent in pristine UiO-66 and the BDC linker, with intensity scaling proportionally to 2M concentration. This peak is assigned to vibrational modes of the benzene ring in the incorporated 2M modulator that are symmetry-forbidden in the BDC ligand (Fig. 2d). Nitrogen adsorption–desorption analysis reveals progressive structural disruption, as the BET surface area of UiO-66 decreases with increasing incorporation of the 2M modulator (Fig. 2e and S10–S13). This trend parallels the loss of long-range order observed by PXRD and is consistent with previously reported correlations between missing-linker defects and reduced surface area.9,57,60 Meanwhile, our pore size analysis indicates that all samples except for UiO-66-3M-0.1 show no detectable pore features attributable to missing clusters (Fig. S14–S18). Scanning electron microscopy further reveals a substantial morphological change—the particle size decreases from about 1 µm in pristine UiO-66 to roughly 100 nm in UiO-66-2M-X—demonstrating that multitopic modulation does not enlarge crystals, in contrast to monotopic acid modulation (Fig. 3). A comparable trend is observed for other multitopic carboxylic acid modulators (Fig. S19–S22). Large excesses of monocarboxylic acids protonate the pristine linkers, slow their deprotonation, and consequently decelerate framework assembly, enabling the growth of larger, more ordered crystals. In the multitopic case, however, only a small amount of modulator is required to generate a high density of missing-linker defects. Thermodynamically, a multitopic modulator binds to Zr6 nodes with an affinity comparable to that of the ditopic linker, so its incorporation competes directly with framework formation without substantially lowering the overall lattice energy. By contrast, coordination of a monotopic acid to the node stabilizes only a cluster-level intermediate and formation of the extended framework remains energetically favoured, so additional time is needed for the system to reorganise into a lower-energy crystalline structure. This distinction reconciles the apparent contradiction of the traditional strategy—slow growth yet limited defect density—with the topicity modulation strategy, where a minimal additive rapidly yields crystals rich in defects. SEM images of (a) UiO-66, (b) UiO-66-2M-0.10, (c) UiO-66-2M-0.20 and (d) UiO-66-2M-0.35. Other multitopic modulators (2O, 3O, 3M and 4OM) displayed the same behavior as 2M: increasing the modulator loading raised the defect density but progressively disrupted long-range order, ultimately driving partial amorphization. When crystallinity was preserved, as indicated by retention of the principal UiO-66 diffraction, the defect concentration plateaued at roughly 1.6–1.7 defects per Zr6 cluster. Even without regard to crystallinity, an upper limit remained: 2O, 3O, and 4OM produced a maximum of about 2.0 defects per cluster (1.9–2.3), whereas 3M reached only 1.8 defects per cluster (Tables S2, S3 and Fig. S23–S45). At the same topicity number, ortho-substituted carboxylic acids (2O) introduced more defects than their meta-substituted counterparts (2M). Although phthalic acid exhibits slightly higher acidity than isophthalic acid, as reflected by its lower pKa, acidity alone does not fully explain the observed difference. To isolate this effect, we employed low concentrations of acidic modulators—trifluoroacetic acid (TFA, pKa = 0.23) and formic acid (FA, pKa = 3.75)—alongside benzoic acid (BA, pKa = 4.20) as a sterically unhindered control. Despite the significant pKa difference between TFA and the other modulators (FA and BA), all three acids show negligible defect generation at low concentrations (Fig. 2f). Given the low concentrations used (0.1–0.3 equivalents), the spatial arrangement of carboxylate groups—specifically the ortho position—appears to enhance coordination competition, leading to more effective defect formation. The enhanced defect-inducing ability of ortho-substituted modulators may be attributed to the increased steric hindrance during coordination (Fig. 4). (a) Correlation between modulator concentration and defect density in synthesized defective UiO-66. (b) Influence of carboxylic acid arrangement (number and relative position) in modulators on their defect-generation capability. This spatial constraint could interfere with the binding of linkers at adjacent sites on the Zr6 cluster, thereby facilitating defect formation. Furthermore, for modulators with the same spatial arrangement (e.g., all ortho- or all meta-substituted), a higher topicity number consistently resulted in higher defect densities. While increased topicity number is often associated with stronger acidity (lower pKa), this alone does not account for the observed effects at low modulator concentrations. Modulators with higher topicity likely promote defect generation by offering multiple coordination sites, which increases the probability of competitive binding with the metal nodes. These findings highlight that both the number and positional arrangement of coordination sites in the modulator significantly influence defect incorporation efficiency. In particular, multitopic modulators with ortho-positioned carboxylate groups and a higher topicity number exhibit superior capability for defect generation. Compared to the conventional strategy, the topicity modulation approach results in high defect densities with much lower modulator dosages and avoids the use of large amounts of acid or the use of strong, corrosive acids. The two strategies differ fundamentally in their synthetic dynamics and outcomes. In traditional modulation, large excesses of monocarboxylic acids generate high proton concentrations that suppress linker deprotonation, retard framework assembly, and afford ample time for crystallographic error correction. This results in conventional coordination modulation that generally exhibits a trade-off between defect generation and structural order. Under many reported synthetic conditions, increasing the amount of monocarboxylic acid beyond a certain level no longer results in proportional increases in defect concentration while progressively affecting framework properties.61,62 The topicity strategy circumvents this constraint. Multitopic modulators bind to metal nodes, compete with the pristine linkers, and promote defect formation. This behavior is consistent with the substantially lower modulator dosage required in topicity modulation, which likely minimizes the proton-buffering effect commonly associated with large excesses of monocarboxylic acids. Because the reaction mixture lacks the proton-buffering capacity characteristic of conventional modulation conditions, long-range order diminishes rapidly with the increase in defect density. Performance optimization using the topicity strategy requires a deliberate balance: enough defects to generate abundant active sites, yet sufficient structural order and surface area to maintain accessibility and overall functionality. We anticipate that the concept of topicity modulation can be readily extended to other Zr-based MOFs because they share similar coordination environments and modulation mechanisms with UiO-66. More broadly, this strategy also provides a promising framework for defect engineering in non-Zr MOFs. However, its practical implementation will depend on the coordination chemistry and crystallization behavior of each metal system. For example, defect modulation in Ti-based MOFs remains particularly challenging because the instability of Ti-oxo clusters during synthesis often leads to unidentified crystalline phases or amorphous products rather than well-defined defective frameworks.8 Similar metal-specific limitations may also exist in other MOF families. Future studies should therefore focus on developing modulation conditions tailored to the coordination characteristics of individual MOF systems, thereby expanding the applicability of topicity modulation across a wider range of porous materials. Catalysis Epoxide ring-opening reactions are fundamental transformations in organic synthesis, providing access to a broad range of functionalized alcohols and key intermediates for pharmaceuticals, fine chemicals, and polymer materials.63,64 Previous studies have established a correlation between defect density and catalytic activity in defect-engineered Zr-based MOFs.65,66 In this work, the same reaction was employed under standardized conditions to investigate the relationship between defect concentration and catalytic performance within the sample of this work. Styrene oxide was used as the epoxide substrate, and isopropanol served both as the nucleophile and the reaction solvent (Fig. 5a). (a) Reaction scheme for the ring-opening of styrene oxide (reaction conditions: 55 °C, 1.5 mol % catalyst, stirring, 24 h). Catalytic performance in the ring-opening reaction of styrene oxide using defective UiO-66 modulated using different modulators: (b) 2O and (c) 2M. (d) Comparison of defects and conversion between the 2M-series and 2O-series. (e) Comparison of defects and conversion for the 2M-series and 3M-series. For UiO-66-2M, catalytic activity increases initially with higher modulator dosage but subsequently decreases, despite a continuous rise in defect density (Fig. 5c). The decline in activity can be attributed to reduced crystallinity, which limits the accessibility of active sites, as supported by PXRD and nitrogen adsorption results. Similar trends were observed for the other four modulators (Fig. 5b and S46–S49). Thus, achieving optimal catalytic performance requires balancing the introduction of defects to generate active centers with maintaining sufficient crystallinity to ensure their accessibility. Both the number and the spatial arrangement of carboxylate groups in the modulator influence not only defect generation but also catalytic activity. In terms of carboxylate positioning, meta-substituted modulators (e.g., 2M) outperform ortho-substituted counterparts (e.g., 2O) in catalytic efficiency. For instance, 2M achieves maximum activity at a relatively low defect density (X ≈ 0.2), whereas 2O requires higher defect densities (X ≈ 0.11) but still exhibits lower efficiency, likely due to a more pronounced loss of long-range order (Fig. 5d). A similar distinction is observed between UiO-66-2X and UiO-66-3X series, where meta-substitution results in more favorable catalytic outcomes. With respect to topicity, modulators containing more carboxyl groups (e.g., 3M and 4OM) display a stronger ability to generate defects, and their catalytic efficiency peaks at relatively low defect concentrations (X ≈ 0.1 for 3M and X ≈ 0.15 for 4OM). However, the effect of topicity is modulated by substitution position: the O-series attain peak efficiency only at higher defect densities (1.7–1.8 defects per cluster), while the M-series achieve optimal performance at lower defect levels (<1.5 defects per cluster) (Fig. 5e). Notably, although the M-series modulators produce fewer defects than the O-series, they exhibit higher catalytic efficiencies. Overall, higher topicity and meta-substitution facilitate defect incorporation yet accelerate the loss of long-range order, resulting in reduced catalytic performance gains. When comparing the best-performing samples from each modulator, catalytic activity shows a downward trend relative to the defect number, highlighting that crystallinity plays a decisive role in catalytic performance.62 This observation indirectly confirms that topicity modulation indeed bypasses the intrinsic trade-off in conventional modulation, in which competition between modulator coordination and linker deprotonation balances defect generation and framework perfection. To further probe the accessibility of defect-associated active sites, alcohol nucleophiles with different steric demands were employed (Fig. S51). The enhanced and size-dependent activity observed for defect-engineered UiO-66 suggests that defect formation not only increases the number of accessible Zr sites but also modifies substrate accessibility within the pore environment. Conclusions Building on the established parameters of modulator quantity and acidity (pKa), this study introduces a new dimension of control—modulator topicity—and establishes a novel modulation strategy, termed the topicity modulation policy, for defect engineering in MOFs. This approach employs multitopic modulators with higher coordination capacities than the framework linkers to enable more efficient and tunable defect introduction. Compared to the conventional modulation approach, the topicity modulation strategy enables the incorporation of a higher density of defects under mild conditions, without the need for large excesses of modulators or corrosive acids. Systematic investigation of the modulator topic number and carboxylate position reveals their critical roles in tuning defect concentration, with ortho-positioned and higher-topic modulators showing superior defect-inducing capability. Catalytic evaluations using the epoxide ring-opening reaction confirm that while defects enhance activity by generating accessible acid sites, excessive defects compromise framework crystallinity and surface area, ultimately reducing performance. This work underscores the importance of balancing defect density and structural integrity in defect-engineered MOFs and provides a generalizable strategy for advancing MOF-based functional materials. Author contributions Ming-shuo Sun: data curation, formal analysis, methodology, writing – original draft, writing – review & editing; Yu Gai: methodology, writing – review & editing; Guo-ying Han: writing – review & editing; Tong Liu: methodology, writing – review & editing; Pascal Van Der Voort: writing – review & editing; Xiao Feng: conceptualization, supervision, funding acquisition, resources, project administration, writing – review & editing. Conflicts of interest There are no conflicts to declare. Supplementary Material