Imine‐Orientation and Confinement‐Engineered Programming of Polyiodide Speciation in Nitrogen‐Rich COFs for Tailored Iodine Capture

ABSTRACT

Achieving precise control over the aggregation state of volatile iodine (I₂) within porous adsorbents is critical for developing high‐performance materials that go beyond mere capacity metrics. Herein, we report a rational design strategy for covalent organic frameworks (COFs) in which polyiodide speciation is programmed through meticulous manipulation of nitrogen site environments and spatial confinement. Two nitrogen‐enriched COFs, Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2, are constructed with deliberate imine orientation and interlayer slippage. Despite nearly identical topologies and comparably high iodine uptake (5.0 vs. 4.7 g·g⁻¹), they exhibit distinctly different confined iodine chemistry. Comprehensive spectroscopic analyses reveal that Py‐Trz‐COF‐1 stabilizes a higher proportion of I₃ ⁻ species, whereas Py‐Trz‐COF‐2 favors I₅ ⁻ formation. Density functional theory calculations attribute this divergence to site‐specific electronic modulation: in Py‐Trz‐COF‐1, localized electron density at the terminal imine nitrogen enhances charge transfer and stabilizes I₃ ⁻, while in Py‐Trz‐COF‐2, enhanced π‐delocalization around the pyridine‐triazine cavity, coupled with larger confinement space, promotes the evolution toward I₅ ⁻. This work demonstrates that polyiodide distribution in COFs can be deliberately engineered through structural precision at the molecular level, offering a new design paradigm for tailoring iodine chemistry in porous materials.

Imine orientation and confinement effects are deliberately engineered in nitrogen‐rich COFs to program polyiodide speciation during iodine uptake. Two isoreticular frameworks with subtle structural variation exhibit distinct I₃ ⁻ and I₅ ⁻ distributions, demonstrating that iodine aggregation behavior can be precisely regulated through molecular‐level design of the local framework environment.

1. Introduction

The safe management of radioactive iodine isotopes (¹²⁹I and ¹³¹I) remains a pressing challenge in nuclear fuel cycle closure, owing to their high volatility, environmental persistence, and radiotoxicity [1, 2, 3]. Effective capture and confinement of these species requires adsorbents that not only exhibit high uptake capacity but also provide a controlled chemical environment to stabilize iodine in a desired aggregation form, which directly influences long‐term storage stability and material regeneration [4, 5]. While a variety of porous materials, including activated carbons, zeolites, and metal–organic frameworks (MOFs), have been explored for iodine capture [6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17], they often lack the synthetic tunability to systematically tailor host–guest interactions at the molecular level.

Covalent organic frameworks (COFs) have emerged as an ideal platform for such precision engineering. Their crystalline, modular structures enable atomic‐level control over pore geometry, surface functionality, and local electronic environments [18, 19, 20, 21, 22, 23, 24, 25, 26, 27, 28, 29]. In the context of iodine adsorption, numerous COFs have demonstrated impressive capacities, primarily attributed to the incorporation of π‐conjugated systems and nitrogen‐rich building blocks such as pyridine and imine units [3, 5, 30, 31, 32, 33, 34, 35, 36]. Compared with traditional iodine adsorbents such as activated carbons and polymeric materials, COFs offer well‐defined and tunable porous structures with periodically arranged functional sites. This structural precision enables a clearer understanding of the relationship between framework structure and iodine adsorption behavior [37, 38, 39]. However, the prevailing design focus has largely been on maximizing uptake through increasing nitrogen content or porosity, treating iodine adsorption as a capacity‐driven process. A critical, yet underexplored, dimension is the deliberate programming of iodine's chemical state within the confined pores—specifically, the ability to direct whether iodine aggregates as I₃ ⁻ or larger polyiodides like I₅ ⁻. This speciation is governed by subtle differences in the host's electronic structure and spatial confinement, factors that remain poorly understood and rarely designed for [36].

Herein, we shift the design paradigm from capacity optimization to chemical‐state programming. We report two isoreticular, nitrogen‐enriched COFs—Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2—that are strategically designed to investigate how imine orientation and induced interlayer slippage dictate polyiodide speciation (Figure 1). By integrating concentrated pyridine and triazine domains, we enhance local charge density and create tailored confinement spaces. The key structural distinction lies in the orientation of the imine linkage (─C═NH─Ph─NH═C─vs. ─NH═C─Ph─C═NH─), a subtle yet powerful design variable that spatially and electronically decouples the nitrogen sites. We hypothesize that this difference will bias the local electron density and the effective confinement volume, thereby steering the preferential stabilization of I₃ ⁻ versus I₅ ⁻. Through a combination of high‐performance iodine capture experiments, in‐depth spectroscopy (Raman, XPS), and density functional theory (DFT) simulations, we demonstrate that these COFs, despite nearly identical topologies and capacities, program fundamentally different iodine aggregation pathways. This work establishes imine orientation and confinement engineering as powerful molecular‐level tools for programming iodine chemistry within COFs, moving beyond adsorption capacity toward the rational design of adsorbents with tailored iodine speciation for enhanced stability and handling.

FIGURE 1.

(a) Schematic illustration of the molecular design strategy for iodine‐adsorbing COFs, highlighting enhanced local charge density and engineered confinement effects. (b) Synthetic routes to Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2.

2. Results and discussion

Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 were synthesized via a solvothermal condensation reaction (Figure 1b and Figure S1, Supporting Information) and were characterized using powder X‐ray diffraction (PXRD), Fourier‐transform infrared (FTIR) spectroscopy, and solid‐state ¹³C cross‐polarization magic‐angle‐spinning (CP‐MAS) NMR spectroscopy (Figure 2 and Figure S2, Supporting Information). To address the high molecular weight and limited solubility of the monomers, the synthetic protocol was deliberately optimized. In addition to acid activation of the aldehyde monomer, the stoichiometric ratio of PDA or TPA was substantially increased to bias the equilibrium toward imine formation and improve framework crystallinity. Structural analysis of Py‐Trz‐COF‐1 by PXRD indicates a high degree of crystallinity, as evidenced by three sharp peaks at 2θ = 2.8°, 4.7°, and 5.5°, which are indexed to the (010), (02$\overline{1}$), and (02$\overline{2}$) reflections, respectively. A similar diffraction pattern is observed for Py‐Trz‐COF‐2, with three peaks at 2θ = 2.8°, 4.7°, and 5.4°, corresponding to the (010), (01$\overline{2}$), and (002) reflections, suggesting topologically analogous framework architectures. Pawley refinement of the PXRD data yields a space group of P1 for Py‐Trz‐COF‐1, with unit cell parameters of a = 7.02 Å, b = 37.60 Å, c = 37.60 Å, α = 120.0°, β = 90.0°, and γ = 90.0° (R_p = 2.42%, R_wp = 3.45%). For Py‐Trz‐COF‐2, the refinement also indicates a P1 space group, with unit cell parameters of a = 6.92 Å, b = 38.22 Å, c = 38.22 Å, α = 120.0°, β = 90.0°, and γ = 90.0° (R_p = 6.93%, R_wp = 9.30%) (Table S3, S4, Supporting Information). The close agreement between the experimental and simulated PXRD profiles not only validates the reliability of the structural refinements but also confirms that Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 are topologically identical, providing an ideal platform to isolate the effects of imine orientation and interlayer packing. In the FTIR spectra, the characteristic stretching vibration of the –C = N– bond appears at 1623 cm⁻¹ for Py‐Trz‐COF‐1 and 1620 cm⁻¹ for Py‐Trz‐COF‐2 [40, 41]. Meanwhile, the aldehyde and amine signals from the corresponding monomers are significantly reduced, indicating efficient consumption of the precursors and successful imine condensation. Furthermore, solid‐state ¹³C CP‐MAS NMR spectroscopy shows a broad resonance in the range of 150–160 ppm, consistent with the formation of extended imine‐linked frameworks, which further confirms the successful synthesis of the COFs⁴⁰.

FIGURE 2.

PXRD profiles of (a) Py‐Trz‐COF‐1 and (b) Py‐Trz‐COF‐2. FT‐IR spectra of (c) Py‐Trz‐COF‐1 and (d) Py‐Trz‐COF‐2. (e) Solid‐state ¹³C CP/MAS NMR spectrum of Py‐Trz‐COF‐1. (f) N₂ sorption isotherms of Py‐Trz‐COF‐1.

Initially, both eclipsed AA and staggered AB stacking types were constructed for Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 (Figures 2a,b,3a, and Figure S4, S5, Supporting Information). Comparison with the experimental PXRD patterns reveals that the AB stacking type is inconsistent with the data, while the AA stacking type provides only a rough match—in particular, the relative intensities of the reflections in the 2θ range of 4–6° deviate noticeably from the experimental profiles. Considering the highly concentrated distribution of nitrogen atoms within the pyridine and 1,2,4‐triazine units, which may modulate interlayer electrostatic interactions, we further simulated a nearly eclipsed AA stacking configuration [42, 43, 44]. Comparison between the simulated and experimental PXRD patterns indicates that both COFs are better described by this nearly eclipsed arrangement. This preference is reasonably attributed to electrostatic repulsion between adjacent nitrogen‐rich regions, which disfavors perfectly eclipsed stacking and induces a slight lateral offset between neighboring layers. Specifically, for Py‐Trz‐COF‐1, the intensity of the (02$\overline{1}$) reflection is weaker than that of the (02$\overline{2}$) reflection in the eclipsed AA stacking type, whereas the nearly eclipsed AA stacking type shows a reversal of this intensity relationship, with a stronger (02$\overline{1}$) reflection relative to (02$\overline{2}$), which more closely reproduces the experimental pattern. Py‐Trz‐COF‐2 exhibits a similar trend, further corroborating the assignment of a nearly eclipsed AA stacking mode for both frameworks. Consistent with this structural model, the two COFs display distinct textural characteristics. The observed difference in BET specific surface areas is tentatively attributed to variations in monomer solubility, imine formation reversibility, and framework growth kinetics during solvothermal polymerization, which can influence crystallization quality, defect density, interlayer stacking order, and ultimately the accessible porosity of the resulting COFs. It should be noted that the BET specific surface area obtained from N₂ sorption reflects the experimentally accessible porosity of the activated bulk material, rather than the intrinsic geometric surface area of an idealized framework. Therefore, even though Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 are topologically analogous, substantial differences in measured BET specific surface area may still arise from differences in pore accessibility in real samples. To verify the reproducibility of this textural difference, N₂ adsorption–desorption measurements were performed using three independently prepared batches of each COF. The average BET specific surface areas were determined to be 266 ± 93 m²·g⁻¹ for Py‐Trz‐COF‐1 and 20 ± 8 m²·g⁻¹ for Py‐Trz‐COF‐2. These results confirm that Py‐Trz‐COF‐2 consistently exhibits much lower N₂‐accessible porosity than Py‐Trz‐COF‐1 across different fabrication batches. Given their isoreticular framework models, it is reasonable to assume that the intrinsic geometric porosities of the idealized structures should not differ drastically. Thus, the large experimental discrepancy in BET specific surface area is more conservatively interpreted as mainly reflecting differences in experimentally accessible porosity, rather than serving as direct evidence of a fundamental difference in framework topology. (Figure 2f and Figure S3, S4, Supporting Information). Both COFs exhibit predominantly type I N₂ adsorption–desorption isotherms, indicating their microporous nature. Notably, open hysteresis loops are observed, which are commonly associated with complex pore structures and non‐ideal adsorption–desorption equilibration in porous materials [45]. In the present case, this behavior is consistent with restricted pore accessibility in the COF frameworks, and is more pronounced in Py‐Trz‐COF‐2, in agreement with its lower N₂‐accessible BET specific surface area and more constrained pore environment. Pore size distribution analysis reveals that both COFs possess dominant pore diameters centered at approximately 1.9 nm. These values are significantly smaller than the theoretical pore sizes predicted for a perfectly eclipsed AA stacking model, but are in good agreement with those expected for the nearly eclipsed AA configuration (Figure 3a), thereby providing independent textural support for the proposed stacking arrangement. Furthermore, scanning electron microscopy (SEM) was employed to examine the morphology of the two COFs (Figure S7, S8, Supporting Information). Py‐Trz‐COF‐1 predominantly exhibits a coral‐like morphology, whereas Py‐Trz‐COF‐2 displays a sheet‐like structure. Such morphological differences may further contribute to the disparity in specific surface areas, with the more open and branched architecture of Py‐Trz‐COF‐1 favoring greater exposure of internal pores. Transmission electron microscopy (TEM) was performed; however, clear lattice fringes could not be reliably resolved under the applied imaging conditions. The observation of lattice fringes in COFs is known to be sensitive to electron‐beam damage, low electron‐scattering contrast, sample thickness, and the degree of layer exfoliation or delamination [46, 47, 48]. Therefore, the structural assignment in this work is mainly supported by PXRD analysis, Pawley refinement, and structural simulation (Figure S9, Supporting Information). TGA results (Figure S14, Supporting Information) show that both COFs are thermally stable up to ∼400°C, which is well above the operating temperature range (room temperature to 75°C) used for iodine adsorption. The COFs show limited stability under strong acidic and basic conditions but remain stable in common organic solvents, which is consistent with the solvent‐based desorption process used in this work (Figure S16, S17, Supporting Information).

FIGURE 3.

(a) Structure models of the vertical view of Py‐Trz‐COF‐1. Color code for space‐filling diagrams: H, white; C, gray; and N, blue. (b) Time‐dependent I₂ vapor uptake of Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 at 75°C under ambient pressure. (c) I₂ breakthrough profiles of Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 at 25°C. Breakthrough time is defined at C/C₀ ≈ 1. (d) High‐resolution XPS spectra of the N 1s region for Py‐Trz‐COF‐1 before and after iodine adsorption. (e) High‐resolution XPS spectra of the I 3d region for I₂@Py‐Trz‐COF‐1 and I₂@Py‐Trz‐COF‐2. (f) Raman spectra of Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 before and after iodine adsorption.

To evaluate how the identified structural and textural distinctions translate into iodine capture behavior, vapor‐phase iodine adsorption experiments were conducted using non‐radioactive ¹²⁷I as a chemical surrogate for radioactive isotopes ¹²⁹I and ¹³¹I, given their essentially identical chemical properties. Each COF sample was exposed to iodine vapor at 75°C under ambient pressure, conditions chosen to approximate typical nuclear fuel reprocessing environments. The adsorption process was monitored by recording the weight gain of the samples at regular time intervals [2]. As shown in Figure 3b, both Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 exhibit rapid initial iodine uptake, with their mass increasing steeply during the first hour of exposure. Adsorption reaches saturation after approximately 40 h, as indicated by the plateau in weight change. Concomitantly, Py‐Trz‐COF‐1 undergoes a gradual color transition from yellow to dark brown during iodine loading, visually confirming the successful incorporation of iodine vapor. Quantitatively, Py‐Trz‐COF‐1 achieved an iodine uptake capacity of 5.0 g·g⁻¹, while Py‐Trz‐COF‐2 reached a slightly lower value of 4.7 g·g⁻¹ under identical conditions. Notably, despite the large difference in BET specific surface area, the two COFs exhibit comparable iodine uptake capacities. This suggests that iodine adsorption in this system is not governed solely by accessible surface area or external surface deposition. Instead, the adsorption process is dominated by host–guest interactions within the framework, as further supported by the uniform iodine distribution observed in EDS mapping and the formation of polyiodide species revealed by Raman and XPS analyses. Despite their nearly identical topologies and comparable uptake capacities, Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 exhibit distinctly different iodine adsorption kinetics, which can be directly correlated with the variation in imine orientation and the resulting differences in local electronic structure. In Py‐Trz‐COF‐1, the imine nitrogen is located at the terminus of the linkage and is spatially decoupled from the electron‐deficient triazine and pyridine units. This configuration preserves a relatively localized lone‐pair electron density at the imine nitrogen, enabling stronger iodine–framework interactions, likely involving charge transfer. As a result, the adsorption process is governed by site‐specific interactions, giving rise to pseudo‐second‐order kinetics (R² = 0.996, k₂ = 0.0353 g·g⁻¹·h⁻¹) (Table S1, Supporting Information). In contrast, in Py‐Trz‐COF‐2, the imine nitrogen is positioned in closer proximity to the heterocyclic cores, which enhances π‐electron delocalization and reduces the local electron density at the nitrogen sites. Consequently, iodine adsorption is less dominated by specific binding sites and more influenced by mass transport and diffusion processes, leading to pseudo‐first‐order kinetic behavior (R² = 0.997, k₁ = 0.0824 h⁻¹). Although Py‐Trz‐COF‐1 exhibits a comparatively slower apparent uptake rate, its stronger iodine–framework interaction enables more sustained iodine retention, as reflected by the significantly prolonged breakthrough time. The stronger N→I charge‐transfer interactions stabilize iodine species within the framework, delaying breakthrough despite the slower adsorption process. In contrast, Py‐Trz‐COF‐2, governed by pseudo‐first‐order kinetics, shows faster initial uptake but weaker iodine retention, leading to earlier breakthrough. Accordingly, the breakthrough curves show that Py‐Trz‐COF‐1 retains iodine vapor for up to 23 h, whereas Py‐Trz‐COF‐2 reaches breakthrough after about 13 h (Figure 3c). Here, the breakthrough time is defined as the point at which the outlet iodine concentration approaches the inlet concentration (C/C₀ ≈ 1), corresponding to near‐complete breakthrough. Furthermore, the coral‐like morphology and higher accessible surface area of Py‐Trz‐COF‐1 likely contribute synergistically to its superior iodine capture performance, compared to Py‐Trz‐COF‐2, which possesses a more compact, sheet‐like structure. Collectively, these results highlight the cooperative interplay between local electronic structure and textural characteristics in governing both iodine uptake kinetics and retention behavior, extending iodine capture design beyond capacity metrics alone.

To further evaluate the practical applicability and operational robustness of these materials, iodine retention, desorption, and recyclability were investigated [2, 49]. After iodine loading, both Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 retain over 70% of the adsorbed iodine after 10 days of ambient storage, demonstrating good retention stability and suitability for long‐term containment of volatile iodine species (Figure 4a). Desorption experiments show that iodine can be partially released within 10 min by rinsing the iodine‐loaded COFs with ethanol, with ∼50% of the adsorbed iodine removed in the first desorption step (Figure 4b). Time‐dependent UV–vis spectra of the desorption solution (Figure S18, S19, Supporting Information) show characteristic I₃ ⁻ absorption bands (∼290 and ∼350–370 nm) that increase with time, indicating the release of iodine as soluble polyiodide species. The desorption plateau primarily reflects the total amount of releasable iodine in the two COFs. Py‐Trz‐COF‐1 exhibits a higher desorption plateau, indicating a larger fraction of iodine can be released, whereas Py‐Trz‐COF‐2 shows a lower plateau, corresponding to a more limited releasable iodine fraction. This difference can be attributed to the distinct polyiodide speciation in the two COFs. In Py‐Trz‐COF‐2, the higher fraction of I₅ ⁻ species, owing to their larger size and stronger confinement within the framework, leads to a significant portion of iodine being more tightly retained and less accessible for desorption. In contrast, Py‐Trz‐COF‐1, which favors smaller polyiodide species (I₃ ⁻), exhibits a larger releasable iodine fraction. These results indicate that iodine release is governed not only by adsorption behavior but also by polyiodide speciation and the associated confinement effects within the framework. A more detailed discussion of polyiodide speciation is provided in a subsequent section. Both Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 exhibit competitive iodine uptake capacities compared with representative COF‐based adsorbents reported in the literature (Figure 4c) [35, 50, 51, 52, 53, 54, 55, 56, 57, 58, 59, 60, 61, 62, 63]. To further elucidate the structure–iodine adsorption relationship, a comparison with representative COF‐based iodine adsorbents is summarized in Table S2 (Supporting Information). Most reported imine‐linked COFs adopt conventional eclipsed AA stacking structures and achieve iodine capture primarily through nitrogen‐rich frameworks or N–I interactions. In contrast, Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 share identical topologies but exhibit a nearly eclipsed stacking configuration, leading to distinct local electronic environments and confinement effects. Notably, despite their comparable iodine uptake capacities, the two COFs display different polyiodide speciation behaviors. This distinction is not commonly observed in previously reported systems, where adsorption performance is typically discussed in terms of capacity rather than iodine speciation. These results highlight that subtle structural variations, such as imine orientation and interlayer slippage, can play a decisive role in regulating iodine chemistry within COF frameworks. Recyclability tests were performed over five consecutive iodine adsorption–desorption cycles. Despite the loss of long‐range crystallinity during repeated cycles, both materials retain over 50% of their initial iodine adsorption capacity after five cycles. This indicates that the adsorption functionality is largely preserved, even though structural ordering is diminished (Figure S16 and S22, S23, Supporting Information).

FIGURE 4.

(a) I₂ retention performance of the two COFs upon exposure to air at 25°C under ambient pressure. (b) I₂ desorption behavior of iodine‐loaded Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 in ethanol solution. (c) Comparison of iodine adsorption capacities among different adsorbents. (d) Adsorption‐density calculations of Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 using the Sorption module in Materials Studio.

Elemental analysis (EA) confirms that the two COFs possess very similar nitrogen contents, consistent with their identical framework compositions. The slight variation observed in semiquantitative EDS results can be attributed to measurement limitations. Despite the comparable nitrogen contents, the two COFs exhibit different polyiodide speciation behaviors, indicating that iodine adsorption is governed not by the total nitrogen content, but by the local nitrogen environment and its electronic and spatial characteristics (Figure S10, S11, Table S5, Supporting Information). X‐ray photoelectron spectroscopy (XPS) further substantiates this finding. The survey spectrum of each iodine‐loaded sample displays two new I 3d signals centered at ≈ 630 eV and 620 eV, confirming the successful incorporation of iodine species. High‐resolution N 1s spectra (Figure 3d) provide direct evidence for charge‐transfer interactions between nitrogen sites and iodine. Prior to adsorption, two components at 399.0/398.2 eV (Py‐Trz‐COF‐1) and 399.0/398.3 eV (Py‐Trz‐COF‐2) correspond to imine and pyridine/triazine nitrogen, respectively. After iodine uptake, both peaks shift to higher binding energies (399.6/399.0 eV) and a new shoulder emerges at ≈ 400.8 eV, consistent with the formation of N→I charge‐transfer complexes. High‐resolution I 3d spectra (Figure 3e) show the expected spin–orbit doublets: for Py‐Trz‐COF‐1, I 3d_3/2 peaks at 631.3 and 629.8 eV and I 3d_5/2 peaks at 619.8 and 618.3 eV; for Py‐Trz‐COF‐2, the corresponding values are 631.7/630.0 eV and 620.2/618.4 eV. These energies are characteristic of the polyiodide species I₃ ⁻ and I₅ ^−64‐67. Taken together, the uniform iodine distribution, the emergence of polyiodide I 3d signatures, and the systematic N 1s chemical shifts demonstrate that nitrogen sites in Py‐Trz‐COFs actively participate in iodine chemisorption. Raman spectroscopy provides complementary vibrational evidence for polyiodide formation within the frameworks (Figure 3f) [68, 69, 70, 71]. Prior to iodine exposure, the Raman spectra of both Py‐Trz‐COFs are featureless in the 80–200 cm⁻¹ region. After adsorption, two new bands appear for each sample: In Py‐Trz‐COF‐1, two characteristic peaks were observed at 111 and 167 cm⁻¹, corresponding to I₃ ⁻ and I₅ ⁻, respectively. Similarly, in Py‐Trz‐COF‐2, the peaks at 109 and 165 cm^{− 1} were assigned to I₃ ⁻ and I₅ ⁻, respectively. Notably, both Raman and XPS analyses consistently indicate that the I₅ ⁻ signal is more pronounced in Py‐Trz‐COF‐2, whereas Py‐Trz‐COF‐1 exhibits a relatively higher contribution from I₃ ⁻. It should be noted that the relative distributions of I₃ ⁻ and I₅ ⁻ are inferred from spectroscopic features and represent qualitative or semi‐quantitative trends rather than exact compositional ratios. Nevertheless, these consistent observations clearly demonstrate that, although both polyiodide species coexist within the COF frameworks, their relative distributions are systematically governed by imine orientation and the associated local confinement environments.

To further support this conclusion, UV–vis diffuse reflectance spectroscopy was performed to probe the optical response upon iodine adsorption (Figure S24, S25, Supporting Information). After iodine uptake, Py‐Trz‐COF‐1 shows a moderate increase in absorption in the 400–500 nm region, which can be attributed to the formation of smaller polyiodide species such as I₃ ⁻. In contrast, Py‐Trz‐COF‐2 exhibits a broader and more pronounced absorption enhancement extending to longer wavelengths, indicative of the presence of larger polyiodide species (e.g., I₅ ⁻). These optical features are fully consistent with the Raman and XPS results, providing complementary evidence for the distinct polyiodide speciation in the two COFs.

To rationalize the experimentally observed divergence in polyiodide speciation, it is reasonable to hypothesize that subtle variations in the local electronic environment, arising from imine orientation and stacking‐induced confinement, govern the evolution of iodine species within the two frameworks. In Py‐Trz‐COF‐1, the imine nitrogen atoms are positioned farther from the π‐deficient triazine and pyridine centers, which preserves a more localized electron density at the imine site and thereby favors the stabilization of smaller polyiodide species such as I₃ ⁻. In contrast, in Py‐Trz‐COF‐2, the closer spatial proximity of the imine site to the heteroaromatic cores enhances orbital overlap and π‐electron delocalization, potentially facilitating the further association of iodine into larger aggregates such as I₅ ⁻. Moreover, the enlarged confinement space generated by interlayer slippage in the Py‐Trz‐COF‐2 framework may further bias the local microenvironment toward I₅ ⁻ formation.

To systematically examine these hypotheses, we first performed adsorption‐density calculations using the Sorption module in Materials Studio, followed by density functional theory (DFT) simulations to quantitatively evaluate site‐dependent binding energies and charge‐transfer characteristics of iodine at different adsorption locations. As illustrated in Figure 5a, three representative spatial positions within the adsorption environment were defined as P1, P2, and P3. Specifically, P1 corresponds to the cavity region formed by the pyridine and triazine units, P2 is located at the space behind the triazine moiety, and P3 represents the site associated with the imine nitrogen generated through the Schiff‐base linkage. Together, these positions capture the primary interaction regions accessible to incoming iodine molecules within the COF channels. In the present calculations, only molecular iodine (I₂) was considered as the adsorbate, and DFT simulations were employed to evaluate its binding energies at the three adsorption‐relevant positions (P1, P2, and P3). Although polyiodide species (e.g., I₃ ⁻ and I₅ ⁻) are formed after adsorption under confined conditions, molecular iodine (I₂) serves as the initial interacting species with the framework. Therefore, the DFT calculations based on I₂ adsorption are designed to probe the initial binding and charge‐transfer behavior, which governs the subsequent evolution of iodine into different polyiodide species under confinement. The adsorption‐density maps (Figures 4d,e) provide a direct visualization of preferred iodine accommodation sites. Notably, Py‐Trz‐COF‐2 shows a markedly higher adsorption density at the P1 region, indicating that the pyridine–triazine cavity functions as the dominant adsorption site. This preliminary result suggests that I₂ is preferentially stabilized within the P1 cavity of Py‐Trz‐COF‐2. In contrast, the adsorption‐density distribution of Py‐Trz‐COF‐1 does not reveal a clear preference among the three regions, necessitating further energetic analysis to resolve its dominant adsorption site. Figure 5a summarizes the optimized configurations of I₂ interacting with the three defined positions in Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2, together with their relative energies calculated at the B3LYP/6‐31g(d,p) level [72]. For Py‐Trz‐COF‐1, the configurations at P1, P2, and P3 exhibit relative energies of 1.31, 4.52, and 0.69 kcal mol^{− 1}, respectively, indicating that the imine‐associated P3 site is thermodynamically most favorable. In Py‐Trz‐COF‐2, the corresponding energies are 0, 6.56, and 4.15 kcal mol^{− 1}, unambiguously identifying the pyridine–triazine cavity (P1) as the preferred adsorption site. Although the lowest‐energy configurations of the two COFs occur at different positions, their energies are comparable (0.69 kcal mol^{− 1} for Py‐Trz‐COF‐1‐P3 vs. 0 kcal mol^{− 1} for Py‐Trz‐COF‐2‐P1), suggesting similar intrinsic affinities toward molecular iodine. However, the distinct spatial stabilization environments—imine‐centered in Py‐Trz‐COF‐1 versus cavity‐centered in Py‐Trz‐COF‐2—are expected to steer the subsequent evolution of confined iodine toward different polyiodide species, consistent with the experimental observations discussed above. It is also noteworthy that, during geometry optimization of the Py‐Trz‐COF‐1‐P1 configuration, the I₂ molecule spontaneously migrates toward the P3 region regardless of its initial placement, indicating that P1 does not represent a stable adsorption site in Py‐Trz‐COF‐1; accordingly, this configuration is excluded from further analysis.

FIGURE 5.

(a) Three configurations of I₂ interacting with the N‐sites in Py‐Trz‐COF‐1/2 with the relative energies (in kcal/mol) at the B3LYP/6‐31g(d,p) level of theory. (b) Hirshfeld charge population of Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 (negative values denote the loss of electrons) (c) ESP maps of Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 ligands obtained at the B3LYP/6‐31g(d,p) level of theory.

Given that charge‐transfer complexes are formed during the chemisorption of I₂ onto Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2, the iodine adsorption capacities of these COFs were further analyzed using Hirshfeld charge calculations [73], as shown in Figure 5b. The Hirshfeld charge distribution quantitatively reflects the extent of electron transfer from the framework to the adsorbed iodine molecule. For Py‐Trz‐COF‐1, Py‐Trz‐COF‐1‐P3 exhibits the largest charge transfer (−0.739), while for Py‐Trz‐COF‐2, Py‐Trz‐COF‐2‐P1 shows the most negative Hirshfeld charge value (−0.245). These results indicate that the preferred adsorption sites in both frameworks donate the greatest amount of electron density to I₂, thereby exhibiting the strongest initial iodine–framework interactions, in excellent agreement with the relative energy analysis discussed above. In addition to the Hirshfeld charge distribution, electrostatic potential (ESP) mapping further elucidates the distinct iodine–framework interaction environments in the two COFs. Figure 5c illustrates the ESP maps for the five optimized configurations. As a reliable descriptor of ionic interaction propensity, the ESP map shows that regions in red represent positive ESP values, whereas blue regions correspond to negative values. Although the positions of minimum ESP values are broadly similar across configurations, their magnitudes and spatial distributions differ, with the exception of Py‐Trz‐COF‐2‐P1, for which the minimum ESP is notably enhanced. Moreover, the two iodine atoms within each I₂ molecule experience distinct ESP values, reflecting site‐dependent electrostatic asymmetry imposed by the framework.

This overall trend is consistent with the experimentally observed speciation preference, where Py‐Trz‐COF‐1 enriches I₃ ⁻ while Py‐Trz‐COF‐2 promotes I₅ ⁻ formation. In Py‐Trz‐COF‐1, the terminal imine nitrogen provides a more localized and electron‐rich microenvironment, enabling stronger electron donation to the initially adsorbed I₂, as evidenced by the larger Hirshfeld charge transfer. Such localized charge transfer favors the stabilization of smaller polyiodides, in line with the relatively stronger I₃ ⁻ signatures in Raman spectra.

By contrast, in Py‐Trz‐COF‐2 the imine site is electronically coupled to the N‐rich heteroaromatic backbone, leading to enhanced π‐delocalization and diminished local electron density at the binding site, consistent with the reduced Hirshfeld charge transfer and the ESP characteristics at the preferred P1 cavity. Together with the enlarged pyridine–triazine confinement space, this delocalized electronic landscape biases the confined iodine evolution toward larger aggregates such as I₅ ⁻, matching the stronger I₅ ⁻ contributions observed in both Raman and XPS. Collectively, these results establish a coherent structure–electronics–confinement relationship: imine orientation and stacking‐governed confinement jointly program the dominant polyiodide speciation in nitrogen‐rich COFs, thereby rationalizing the distinct adsorption/retention behaviors of Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2.

3. Conclusion

In this work, we demonstrate a chemical‐state programming strategy for iodine capture by constructing two nitrogen‐enriched covalent organic frameworks, Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2, in which pyridinic and triazine nitrogen atoms are spatially concentrated within a rigid backbone. This architectural motif induces characteristic interlayer slippage and well‐defined confinement microenvironments, enabling iodine stabilization through coupled electronic and spatial effects rather than simple physical adsorption. Although Py‐Trz‐COF‐1 and Py‐Trz‐COF‐2 share nearly identical topological frameworks and comparable iodine uptake capacities, a subtle inversion in imine linkage orientation gives rise to markedly different local electronic environments, which in turn program the speciation of confined polyiodides. Py‐Trz‐COF‐1 features a terminal, electron‐rich imine nitrogen that supports strong charge transfer to iodine and preferentially stabilizes smaller polyiodide species such as I₃ ⁻. In contrast, Py‐Trz‐COF‐2 exhibits enhanced π‐electron delocalization through stronger electronic coupling between the imine site and the electron‐deficient heterocyclic backbone, together with a more accommodating pyridine–triazine cavity, collectively favoring the formation and stabilization of larger polyiodides such as I₅ ⁻. These programmed differences in iodine aggregation behavior are consistently supported by complementary spectroscopic and theoretical analyses, including Raman and XPS measurements, adsorption‐density mapping, binding‐energy calculations, Hirshfeld charge analysis, and electrostatic potential evaluation. Collectively, this work establishes that polyiodide speciation within COFs can be deliberately engineered through precise control of nitrogen positioning and local electronic structure, even in frameworks with otherwise similar architectures. By decoupling iodine uptake capacity from iodine aggregation state, this study highlights confinement‐aware molecular design as a powerful paradigm for programming iodine chemistry in porous frameworks, and provides general design guidance for the development of advanced adsorbents aimed at the safe, stable, and regenerable immobilization of volatile radioactive iodine.

Conflicts of Interest

The authors declare no conflict of interest.

Supporting information

Supporting File: smll74193‐sup‐0001‐SuppMat.docx.

Data Availability Statement

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