Conjugated covalent organic frameworks (COFs) [RN1.1]offer tunable band structures and π-electron delocalization for photocatalytic hydrogen production [
1-
2]. However, efficient light harvesting does not necessarily ensure productive charge transfer, as the linkage chemistry strongly influences backbone rigidity, electronic delocalization, and the competition between charge recombination and catalysis [
3]. In particular, flexible imine and vinyl linkages can undergo out-of-plane motions that disrupt orbital overlap and promote vibrational energy dissipation [
4-
5]. Previous studies have highlighted the importance of framework functionality and linkage locking in photocatalysis [
6-
7]. Therefore, rigidifying COF linkages to enhance conjugation, suppress energy loss, and prolong charge-carrier lifetimes is critical for improving photocatalytic hydrogen evolution.
In a recent work published in
Nature Synthesis, Zhao and co-workers reported a one-pot route to a coumarin-linked conjugated COF and compared it with imine- and vinyl-linked counterparts [
8]. Base-catalysed Knoevenagel condensation followed by intramolecular cyclization forms the coumarin units (
Fig. 1a), eliminating the concentrated hydrochloric acid treatment required in an earlier route [
9]. Diffraction and spectroscopic measurements support linkage formation and a crystalline framework consistent with AA stacking. The advance therefore lies in a simplified synthetic route and the investigation of how this linkage influences charge dynamics and photocatalytic interfaces.
The rationale for selecting coumarin becomes clearer when the three linkages are compared computationally (
Fig. 1b). Imine and vinyl connections permit out-of-plane pedal motion that can disrupt orbital overlap. Incorporating a cyclic coumarin unit raises the calculated rotational energy barrier to 60 kJ mol
-1, compared with 32 and 39 kJ mol
-1 for the imine and vinyl models, respectively. The corresponding orbital delocalization indices decrease from 7.05 to 4.71 and 4.38, with the lower value indicating greater delocalization (
Fig. 1c). These results support a backbone that better maintains planar conjugation and limits torsional energy dissipation. Crucially, the proposed benefit is not simply a larger conjugated structure: the linkage also constrains the molecular motions that can undermine its electronic function.
The optical and photoelectrochemical measurements are consistent with this structural interpretation. The vinyl-linked material has a photoluminescence peak intensity 533 times that of the coumarin-linked COF (
Fig. 2a). Weak emission alone does not establish productive charge separation; the stronger light-induced electron paramagnetic resonance (EPR) signal, higher photocurrent and lower impedance provide additional support for improved carrier generation and transport. The absorption edge shifts to 732 nm for the coumarin-linked COF, compared with 533 nm for the imine analogue, and its reported optical bandgap is 1.69 eV (
Fig. 2b). Calculated frontier orbitals further suggest an internal donor-acceptor arrangement, with the lowest unoccupied molecular orbital concentrated mainly on the coumarin units. Together, these observations connect linkage rigidity with altered excited-state behavior rather than relying on photoluminescence quenching as the sole mechanistic indicator. Femtosecond transient absorption spectroscopy reveals that the coumarin-linked COF exhibits markedly prolonged excited-state lifetimes (
${\mathit{\tau }}_{1}=45.42\end{equation}$ ps and
${\mathit{\tau }}_{2}=1031.39\mathit{ }\end{equation}$ps), substantially longer than those of the imine- and vinyl-linked COFs (
Fig. 2c). This behavior indicates that the enhanced electronic delocalization effectively stabilizes the charge-separated state, providing sufficient time for electron migration to the catalytic sites. In the presence of chloroplatinic acid as a cocatalyst precursor, the photogenerated electrons are further transferred to Pt nanoparticles within approximately 407 ps, where they participate in proton reduction (
Fig. 2d). This time should be distinguished from the fitted decay components and from charge dynamics under sustained catalytic conditions.
The improved charge dynamics are accompanied by high sacrificial photocatalytic hydrogen evolution activity. The reported protocol uses 1.5 mg of COF in 10 mL of 0.14 M aqueous ascorbic acid, with Pt introduced by in situ photodeposition from chloroplatinic acid (Pt content of 2 wt%). Under optimized 440 nm irradiation, the coumarin-linked COF reaches 531.2 mmol g
-1 h
-1 over the s
table 12-20 h interval, compared with 34.6 mmol g
-1 h
-1 during the initial 0-4 h. Under a 300 W Xe lamp with a >420 nm cutoff, the reported rate is 166 mmol g
-1 h
-1, versus 0.016 and 3.5 mmol g
-1 h
-1 for the imine- and vinyl-linked materials. The cross-study comparison in
Fig. 2e provides context only: catalyst amount, Pt content, illumination and sacrificial reagent vary among entries (
Table 1), precluding a direct ranking of intrinsic activity. The apparent quantum yield (AQY) reaches 37.95% at 405 nm (
Fig. 2f). These wavelength and time-window distinctions are essential for interpreting the reported activity. The results demonstrate efficient hydrogen evolution in a sacrificial system; they do not establish an equivalent solar-to-hydrogen efficiency for overall water splitting. Post-reaction diffraction and infrared measurements support retention of the framework structure over the examined catalytic operation.
The coumarin linkage also affects the catalyst-water and catalyst-metal interfaces. The original study attributes the observed dispersion of photodeposited Pt nanoparticles to interactions with coumarin oxygen atoms. Broadening and red-shifting of the carbonyl infrared band upon hydration support carbonyl-water hydrogen bonding, while a water contact angle of 34° indicates favourable wetting. These observations suggest that electronic and interfacial effects may jointly influence activity, consistent with earlier work on framework order, metal-binding sites and hydrophilic channels [
10].
The value of this study is the combination of a one-pot synthetic route with evidence linking linkage chemistry to excited-state dynamics and interfacial electron extraction. However, the three COFs are not controls that isolate rigidity alone. They differ in morphology, wetting and cocatalyst interactions, and their Brunauer-Emmett-Teller surface areas are 399, 41 and 143 m2 g-1 for the imine-, vinyl- and coumarin-linked materials, respectively. The highest activity therefore does not simply follow the largest surface area, but the independent contributions of rigidity, donor-acceptor character and interfacial properties remain unresolved. Comparisons with matched morphology and independently quantified Pt loading and dispersion, together with charge-dynamics measurements in the catalytic medium, would help distinguish these contributions.
Several questions remain for extending this approach. Despite absorption extending to 732 nm, the reported AQYs fall to 0.026% at 490 nm and 0.019% at 550 nm, showing that broad absorption does not ensure efficient use of lower-energy photons. Resolving this mismatch could guide donor–acceptor tuning while retaining structural rigidity. Longer operation, earth-abundant cocatalysts and coupling hydrogen evolution to a sustainable oxidation reaction also require investigation. Coumarin linkages thus provide a useful platform for jointly tuning backbone dynamics and interfaces; testing this strategy across other frameworks will establish its broader applicability.
PDF
