Position-dependent carboxyl functionalization in covalent organic frameworks for selective photocatalytic CO2 reduction

Imported from: /opt/downloadPDF/chemwiki_pubstore/02f38cc35ab9b55b67f835955a216ceb.pdf/fchem-14-1892359.pdf

Abstract Summary

This article examines photocatalytic carbon dioxide reduction using two carboxyl-position isomeric covalent organic frameworks, TpBdda and TpBdad, as heterogeneous organic photocatalysts. Both materials were tested for light-driven conversion of CO₂ in water under Xe-lamp irradiation. The central finding is that changing only the spatial arrangement of carboxyl groups within otherwise similar COF frameworks strongly alters product selectivity. TpBdda mainly produced CO with nearly complete CO selectivity, whereas TpBdad favored deeper reduction to CH₄ with about 90% CH₄ selectivity. Spectroscopic and photoelectrochemical data support the interpretation that the two frameworks differ in how they stabilize and transform surface-bound reduction intermediates.

Advances and Special Progress

A key advance is the use of positional isomerism of a functional group, rather than changing the overall framework composition, to tune photocatalytic CO₂ reduction selectivity. The study shows that subtle rearrangement of carboxyl groups in a COF can redirect the reaction from a shallow two-electron pathway giving CO to a deeper multielectron pathway giving CH₄.

The work also provides mechanistic insight beyond simple activity comparison. In situ diffuse reflectance infrared Fourier transform spectroscopy was used to compare surface intermediates formed on the two COFs during illumination under CO₂. The data support different intermediate evolution patterns: TpBdda favors *CO-related species and CO release, whereas TpBdad more strongly stabilizes hydrogenated intermediates, especially methoxy-related species associated with CH₄ formation.

Another meaningful aspect is that the system operates in deionized water without any added molecular photosensitizer or sacrificial electron donor being reported. The article presents these COFs as light-absorbing photocatalysts themselves, and relates their performance to framework structure, pore environment, optical absorption, and charge-transport behavior.

Additional Remarks

The chemical significance of this study lies in selective control over two different CO₂ reduction outcomes: CO, a two-electron product, and CH₄, a much more demanding deep-reduction product. Demonstrating a switch between these outcomes in closely related organic frameworks is important because product selectivity is a major challenge in CO₂ photoreduction chemistry.

The reported activities are modest in absolute terms, but the article emphasizes selectivity control rather than record productivity. The experiments were conducted for relatively short irradiation times, and the article presents cycling stability tests qualitatively in figure form, but does not report numerical long-term durability metrics in the text.

Control experiments without catalyst or without CO₂ gave no detected CO or CH₄, supporting that the gaseous carbon-containing products arose from photocatalytic CO₂ conversion under the reported conditions. The article does not report H₂ evolution, formate production, apparent quantum yields, or isotope-labeling evidence. Therefore, the mechanistic picture should be understood as based mainly on product analysis, optical and photoelectrochemical characterization, and in situ infrared detection of surface intermediates.

Content of the Published Article in Detail

The article studies two covalent organic frameworks prepared from 2,4,6-triformylphloroglucinol and two different carboxyl-position isomeric diamine monomers. The resulting photocatalysts, TpBdda and TpBdad, have similar overall compositions but different spatial arrangements of carboxyl groups on the biphenyl units. The purpose of this design is to isolate the effect of functional-group position on CO₂ reduction behavior.

For photocatalytic testing, 10 mg of catalyst was dispersed in 10 mL of deionized water by ultrasonication. Before irradiation, the sealed quartz reactor was purged with Ar for 5 min to remove residual air and oxygen. Then 5 mL of CO₂ was injected into the reactor. Irradiation was performed with a 300 W Xe lamp without an optical filter for 2 h. Gas samples were taken from the reactor headspace every 0.5 h and analyzed by gas chromatography. CO and CH₄ were identified by comparison with standard gases and quantified using calibration curves.

The article does not describe a separate molecular photosensitizer, sacrificial electron donor, or added proton donor. Instead, the COFs themselves are treated as the photoactive materials. UV-vis diffuse reflectance spectra show that both TpBdda and TpBdad absorb in the UV and visible regions, indicating that both can harvest light. TpBdad shows broader visible-light absorption and a red-shifted absorption edge compared with TpBdda.

Mott-Schottky and Tauc analyses were used to examine semiconductor properties. Both materials showed positive Mott-Schottky slopes, which the article interprets as n-type semiconductor behavior. The flat-band potentials were estimated as about −0.29 V for TpBdda and −0.107 V versus NHE for TpBdad. Optical band gaps were reported as about 2.433 eV for TpBdda and 1.971 eV for TpBdad. Transient photocurrent measurements showed that both materials generate photocurrent under intermittent irradiation, supporting formation and transport of photoinduced charge carriers. The article interprets the difference in photocurrent intensity as evidence that carboxyl-group positional isomerism affects charge separation and migration.

Product analysis showed clear selectivity differences. TpBdda mainly formed CO at about 5.2 μmol g⁻¹ h⁻¹, with CH₄ barely detected. TpBdad mainly formed CH₄ at about 1.8 μmol g⁻¹ h⁻¹, with only a small amount of CO, about 0.2 μmol g⁻¹ h⁻¹. The article states that TpBdda showed nearly complete selectivity to CO, while TpBdad showed about 90% selectivity to CH₄. Control experiments without catalyst or without CO₂ produced neither CO nor CH₄.

Structural characterization was used to relate framework properties to catalysis. Powder X-ray diffraction showed ordered COF formation in both cases, but with different main low-angle diffraction peaks, indicating different lattice spacing and packing. FTIR confirmed similar backbone structures with some differences in local chemical environment. XPS showed C, N, and O in both frameworks and suggested differences in local electronic environments, especially in carbon and oxygen chemical states. Solid-state ¹³C NMR further supported distinct carbon environments related to the isomeric structures. Nitrogen sorption measurements showed different pore sizes and surface areas: TpBdda had a characteristic pore size around 2.3 nm and a specific surface area of 28.472 m² g⁻¹, whereas TpBdad had a dominant pore size around 3.9 nm and a specific surface area of 41.8646 m² g⁻¹.

The mechanistic evidence comes mainly from in situ DRIFTS during CO₂ adsorption and photocatalytic reduction. After CO₂ introduction, both COFs showed adsorption-related signals, indicating CO₂ uptake and preliminary activation. Upon light irradiation, new bands appeared in the 2000–800 cm⁻¹ range. The article associates bands near 1612, 1543, 1425, and 1040 cm⁻¹ with framework-related vibrations, monodentate carbonate species, bicarbonate species, and methoxy-related intermediates, respectively. A band around 1650 cm⁻¹ is assigned to adsorbed water.

For TpBdda, bands near 1690 and 1540 cm⁻¹ are assigned to *COOH and *CO-related intermediates. The methoxy-related band near 1040 cm⁻¹ is relatively weak. Based on these observations, the article proposes that TpBdda mainly follows a shallow pathway in which CO₂ is converted to *COOH, then to *CO, followed by rapid CO desorption. This interpretation is used to explain the high CO selectivity.

For TpBdad, DRIFTS signals in the 1700–1500 cm⁻¹ region and near 1040 cm⁻¹ become stronger and continue to intensify with irradiation time. The article interprets the stronger methoxy-related signal as evidence that TpBdad more effectively stabilizes hydrogenated intermediates involved in deep reduction. On this basis, it proposes that the local pore environment, surface chemistry, and electronic structure of TpBdad promote continued proton-coupled electron transfer and further hydrogenation, ultimately favoring CH₄ formation.

The article therefore presents the selectivity difference as a combined consequence of several experimentally observed factors: distinct framework packing, different pore environments, different surface chemical states, different optical absorption and band structures, different charge-transport behavior, and different intermediate evolution under reaction conditions. However, the mechanistic sequence beyond the observed infrared bands is proposed rather than directly proven step-by-step.

Catalyst

The catalysts are TpBdda and TpBdad, two carboxyl-position isomeric covalent organic frameworks. They are heterogeneous, metal-free, organic framework photocatalysts. Both were prepared from 2,4,6-triformylphloroglucinol and a biphenyl diamine containing carboxyl groups, but the carboxyl groups occupy different positions in the two diamine monomers.

Chemically, the two materials have similar framework compositions but distinct structural and surface properties. The article reports differences in crystallographic peak positions, pore size distribution, specific surface area, XPS peak shapes, solid-state ¹³C NMR patterns, visible-light absorption, flat-band potential, and photocurrent response. These differences are linked to different catalytic roles in CO₂ reduction: TpBdda is CO-selective, while TpBdad is CH₄-selective.

The article attributes the selectivity difference to how carboxyl-group arrangement changes the local pore environment and the stabilization of intermediates. TpBdda favors formation and desorption of *CO-related intermediates. TpBdad more strongly stabilizes hydrogenated intermediates, especially methoxy-related species, which the article connects to deeper reduction toward CH₄.

Photosensitizer

No separate photosensitizer is reported. The article treats the covalent organic frameworks themselves as the light-absorbing photocatalytic materials. UV-vis diffuse reflectance data show that both frameworks absorb in the UV and visible regions, and this light-harvesting ability is presented as the basis for photocatalysis.

Because no distinct molecular photosensitizer is included, there is no reported photosensitizer identity, concentration, or separate quenching pathway. The article also does not report reductive quenching or oxidative quenching of a photosensitizer by a sacrificial donor or catalyst. Instead, it discusses photoinduced charge generation and transfer within the COF frameworks, supported by transient photocurrent and semiconductor analyses.

Investigation

	cat	cat conc [µM]	PS	PS conc [mM]	e-D	e-D conc [M]	.	.	solvent A	additives	.	.	.	.	λexc [nm]	.	TON CO	.	.	TON CH4	TON H2	.	TON HCOOH	.	.
1.	Without catalyst	not reported	not reported	not reported	not reported	not reported			deionized water	not reported					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "WXelamp" is not declared as a valid unit of measurement for this property.
2.	Without CO2	not reported	not reported	not reported	not reported	not reported			deionized water	not reported					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "Molecule:100493" is not in the list (nitrogen, argon, CO2, CO, O2, H2, none) of allowed values for the "Feedstock gas" property. "WXelamp" is not declared as a valid unit of measurement for this property.
3.	TpBdda	not reported	not reported	not reported	not reported	not reported			deionized water	not reported					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "WXelamp" is not declared as a valid unit of measurement for this property.
4.	TpBdad	not reported	not reported	not reported	not reported	not reported			deionized water	not reported					not reported		not reported			not reported	not reported		not reported			"not reported" is not a number. "WXelamp" is not declared as a valid unit of measurement for this property.

Investigation-Name: inv0ExportRefresh