Catalyst Characterisation. The AgI/Mg2Al1LDH, with a silver content of 9.4 wt.% was obtained as a light-yellow powder (see Fig. 1a), that exhibited the optimal photocatalytic performance. The results of the X-ray diffraction (XRD) analysis (Fig. 1b), the energy-dispersive X-ray spectroscopy (EDX) elemental mappings (Fig. 1c) and transmission electron microscopy (Fig. 1d) show that the as-prepared Mg2Al1LDH photocatalyst was loaded with AgI nanoparticles (NPs) dispersed on LDH support. The high-resolution TEM (HR-TEM) images reveal the lattices of AgI and LDH (Figs. 1e and f), which are in accord with the results of XRD analysis.
Control experiments and performance of the catalysts. The direct arylation of benzene was chosen as a model reaction for control experiments, optimising the photocatalysts and the reaction conditions. These experiments elucidated the roles played by the catalyst components and the influence of light irradiation. Catalysts with different Mg/Al molar ratios were used for catalysing the reaction under visible light irradiation. As shown in Supplementary Table 1, the AgI/Mg2Al1LDH photocatalyst exhibits the best catalytic performance without any requirement for the addition of base or any other additives. The desired product biphenyl can be obtained and identified by mass spectrometry (Supplementary Fig. 1), with I2 also being produced as a visibly evident byproduct (Fig. 1a) confirmed by UV-Vis spectroscopy (Supplementary Fig. 2). A high biphenyl yield of 93% was achieved by AgI/Mg2Al1LDH catalyst under LED light irradiation at 400 nm wavelength with an intensity of 0.1 W cm-2 (Entry 1 in Table 1). The yield of biphenyl varies moderately with Mg/Al ratio of the catalysts (Supplementary Table 1). The absence of either the irradiation or the catalyst resulted in only a trace yield of the target product (Table 1, entries 2 and 3). The absence of iodobenzene or benzene gave none or only a trace of biphenyl (Table 1, entries 4 and 5), suggesting that the product is formed through a cross-coupling process.
A significant discovery involves the successful utilization of natural sunlight for photocatalytic direct arylation. As shown in Supplementary Fig. 3, a 60% biphenyl yield was attained at relatively low reaction temperatures, obviating the requirement for supplementary energy input, and highlighting the system's impressive solar energy utilization efficiency.
The fact that both the Mg2Al1LDH support and AgI/ZrO2 individually exhibit low catalytic activities is worth noting (Table 1, entries 6 and 7). The content of AgI in AgI/ZrO2 catalyst is similar to that in AgI/Mg2Al1LDH (Supplementary Table 2, entries 1 and 2). These results indicate that there is a synergistic effect of AgI NPs and Mg2Al1LDH in the photocatalytic system. Furthermore, the addition of Na2CO3, which is the base used for preparing Mg2Al1LDH, to the system using AgI/ZrO2 catalyst did not significantly increase the biphenyl yield (Table 1, entry 8). This indicates that CO32- anions made little contribution to the catalytic performance under these conditions. The photocatalytic system demonstrated a notable tolerance towards moisture and atmospheric oxygen, as evidenced by achievement of a moderate yield of biphenyl (56.4%, Entry 9 in Table 1) under an air atmosphere.
Standard reaction conditions: iodobenzene (0.1 mmol), benzene (1 mL), catalyst (50 mg), Ar atmosphere, irradiated under a 400 nm wavelength LED light with 0.1 W cm-2 of intensity, the reaction was conducted at 60℃ for 20 h. Yields are determined by GC analysis.
We also monitored the formation of biphenyl during the direct arylation process using the AgI/Mg2Al1LDH photocatalyst. The results, summarised in Supplementary Fig. 4a, show that the yield of biphenyl reaches 80% within the first 10 h, with 90% yield reached over 20 h. Additionally, as illustrated in Supplementary Fig. 4b, the increase in biphenyl yield does not follow a linear trend with addition of more catalyst, presumably due to screening effects within the photocatalytic system20.
Broad substrate scope. The practicality of the AgI/Mg2Al1LDH photocatalyst in the direct arylation of arenes depends on the substrate and operation life in addition to the sustainability and environmental impact of all the products. The aryl iodides, contain electron-donating or electron-withdrawing groups, exhibited excellent reactivity towards benzene under irradiation of 400 nm light (Table 2). The corresponding products were obtained in good yield from 87–99%, while the reactions conducted in the dark showed negligible yields of target products. Moreover, the reactions between benzene and iodides containing heterocycles such as thiophene and pyridine afford the desired arylated heteroarenes with good yield (94% and 91%, 6–7), demonstrating a broad substrate scope for this photocatalytic system. It is noteworthy that iodobenzene possessing an electron-deficient nitrile substituent group exhibits a higher reactivity (99% yield) under the same conditions compared to that with an electron-donating methoxy group (91% yield), suggesting the involvement of a SET process in the reaction9.
Importantly, direct arylation of inactive arenes beyond benzene with iodobenzene occurred using the photocatalytic system. As shown in Table 2, good yields are obtained from arenes with substituent groups containing heteroatoms (74%-98%, 9–12), while moderate yields from arenes with hydrocarbon substituent group (27%-63%, 8, 13–16). Notably, no products related to benzylic activation were detected. This achievement is particularly challenging due to the significantly stronger C(sp2)-H bonds compared to the C(sp3)-H bond (by about 100 kJ mol− 1)21. The reaction involving toluene yielded three products (2-, 3-, and 4-methyl biphenyls) with a ratio of 62:24:14 (8), indicating potential involvement of radical intermediates in the pathway for the photocatalytic reaction22,23. When comparing the yields of biaryl products from the direct arylation of arenes with iodobenzene, we observed a trend of the reactant arenes: benzene > arenes with substituent groups containing heteroatoms > toluene > arenes with multiple or larger hydrocarbon substituent. The large difference in the yields reflects the heavy dependence of the arylation on the reaction between the aryl radicals and the arene. Significantly higher yields are attained from the arenes with heteroatom substituent groups compared to those with alkyl substituent groups.
Reaction conditions: aryl iodide (0.1 mmol), arene (1 mL), AgI/Mg2Al1LDH catalyst (50 mg), Ar atmosphere, irradiated under a 400 nm wavelength LED light. Yields were determined by GC analysis. aReaction was conducted at 60℃ for 20 h, and the light intensity was 0.1 W cm-2. bReaction was conducted at 80℃ for 20 h, and the light intensity was 0.1 W cm-2. cReaction was conducted at 60℃ for 48 h, and the light intensity was 0.1 W cm-2. dReaction was conducted at 85℃ for 48 h, and the light intensity was 0.2 W cm-2. To enhance the yield of the products 1–7, the catalyst was filtered out at half of the reaction time during photocatalysis, followed by the addition of 50 mg of fresh catalyst to the liquid mixture. eReaction was conducted at 60℃ for 20 h, and the light intensity was 0.2 W cm-2. fReaction was conducted at 60℃ for 20 h, and the light intensity was 0.1 W cm-2. Iodobenzene was used in producing the products 8–16.
The surface OH groups of LDH are consumed during the photocatalytic reaction to generated •OH radicals. Supplementary Fig. 5 illustrates the impact of this consumption on the biphenyl yield when using a used AgI/Mg2Al1LDH photocatalyst under 400 nm irradiation, showing a decrease from 80–48%. However, a simple regeneration method (as detailed in the Method section) involving the addition of Na2CO3 aqueous solution to a sealed flask, followed by heating to 85°C with stirring for 1 hour, can effectively restore the product yield. X-ray photoelectron spectroscopy (XPS) and Zeta Potential analysis presented in Supplementary Fig. 6 indicate that the surface hydroxyl groups of LDH can be recovered by regeneration. The catalyst can be recycled multiple times.
The impact of light intensity and wavelength on the catalytic performance of AgI/Mg 2 Al 1 LDH photocatalyst. Figures 2a and 2b demonstrate the sensitivity of the AgI/Mg2Al1LDH catalyst's performance to intensity and wavelength of the irradiation, in the context of direct benzene arylation. Higher irradiation intensity leads to increased biphenyl yield. Notably, even under the conditions of extremely low light intensity, as low as 0.01 W cm-2, the catalyst achieves a substantial 40% biphenyl yield, indicating its efficiency. It exhibits an apparent quantum yield (AQY) of 3.3% at this low light intensity, highlighting its exceptional utilization of visible light for catalysis.
Figure 2b presents an action spectrum, which provides valuable insight into the impact of irradiation wavelength on the direct arylation using the AgI/Mg2Al1LDH catalyst. The determination of AQY involves normalising the number of target molecules generated to the incident photons within the photocatalytic system24. Notably, irradiation by LED light centred at 350 nm and 400 nm results in substantially elevated AQY values, while longer wavelengths (550 and 660 nm) yield negligible AQYs. This clear dependence of AQYs on the irradiation wavelengths aligns with the light absorption characteristics exhibited by the AgI/Mg2Al1LDH catalyst. This behaviour contrasts with prior research, specifically direct photolysis, where the optimal yield was achieved using deep-UV light, but a significant decrease was observed at a longer wavelength.25 It is evident that the primary driver of the reaction is the light absorbed by AgI nanoparticles, given that the Mg2Al1LDH support has minimal absorption at wavelengths longer than 350 nm (Fig. 2b). Consequently, the AgI nanoparticles effectively function as efficient light absorbers for the catalyst system.
Photo-generated hydroxyl radicals to facilitate direct arylation. The remarkable observation that a relatively modest amount of absorbed light energy leads to a substantial biphenyl yield implies the existence of low activation energy barriers within the photocatalytic reaction pathway. An in-depth kinetic study on the reaction between iodobenzene and benzene over AgI/Mg2Al1LDH under light irradiation reveals a first-order kinetics relationship with respect to the substrate, as depicted in Supplementary Fig. 7. Employing the Arrhenius equation (raw data shown in Fig. 3a) we estimate an apparent activation energy of 46.3 kJ mol-1 for the photocatalytic reaction, which is significantly lower than 61.7 kJ mol-1 reported for a thermally driven system 6 (as shown in Fig. 3b).
Comparison of the relative amounts of surface OH groups in the AgI/Mg2Al1LDH catalyst before and after the photocatalytic reaction, as determined from the XPS spectra of O 1s, reveals a noticeable reduction in surface hydroxyl groups during the course of the photocatalytic reaction. Within the XPS spectra (Figs. 4a and 4b), the O 1s peaks at 532.7, 531.4, and 530.2 eV correspond to various oxygen species in the catalyst, including metal-hydroxyl (M-OH), metal oxide (M-O), and carbonate (CO32-) of LDH, respectively26–28. It is noteworthy that the concentration of M-OH on the catalyst after the photocatalytic reaction is significantly lower than that on the pristine catalyst. This decline can be attributed to the transformation of surface hydroxyl groups of LDH into •OH radicals, which are consumed during the reaction, leading to the formation of sites with unpaired electron confined within oxygen vacancies. As depicted in Fig. 4c the electron paramagnetic resonance (EPR) analysis of the AgI/Mg2Al1LDH catalyst before and after photocatalysis exhibits a prominent EPR signal characterized by a g factor of 2.004 in the used catalyst, unequivocally indicating the existence of such sites29. Consequently, it can be inferred that surface OH groups are consumed in the process, generating •OH radicals. The activation energy barriers in the reaction mediated by •OH radicals are notably low. The recycling experiments regenerated the surface OH groups.
The •OH radicals are regarded essential for oxidation in photocatalysis30, being the main reactant for various fundamental processes such as SEO, double bond addition, and hydrogen abstraction31. Considering both SEO and hydrogen abstraction steps are involved in direct C-H arylation9, which •OH radicals can participate in and facilitate, LDH can serve as a reservoir for hydroxyl groups, supplying them internally to produce •OH radicals that drive the reaction without the need for any peroxide addition. This is supported by fact that Mg2Al1LDH has the maximum crystallite size, and the AgI/Mg2Al1LDH catalyst exhibits the highest biphenyl yield (Supplementary Table 3 and Supplementary Fig. 8a). The size of crystallite is directly proportional to the quantity of surface hydroxyl groups on LDH32,33. In contrast, other factors, such as their AgI NPs contents (represented by the silver content in Supplementary Table 2) and specific surface areas (Supplementary Fig. 9), have a lesser impact on the performance of AgI/MgxAlyLDH photocatalysts.
The presence of •OH radicals is substantiated by multiple experimental observations. Terephthalic acid (H2BDC) was employed as a fluorescent probe to detect •OH radicals within our photocatalytic system, following a well-documented procedure34. In the fluorescent experiment, a characteristic fluorescence signal at 425 nm is prominently observed (see Fig. 4d), while the absence of any the following components: AgI NPs, Mg2Al1LDH support, H2BDC and irradiation, yields no discernible fluorescence peaks. The result provides direct evidence that hydroxyl can be converted to •OH radicals in the photocatalytic system. Besides, the concentration of •OH radicals varies in the order of AgI/LDH > AgI/ZrO2 > LDH, which is consistent with their catalytic activities shown in Table 1. The results not only provide compelling evidence for the generation of •OH radicals from LDH but also imply that the illuminated AgI NPs play a pivotal role in driving this process.
The excellent performance of AgI/Mg2Al1LDH catalyst under dry conditions, as demonstrated in Table 1, entry 1 (in the absence of moisture), further underscores the origins of •OH radicals from surface OH groups. Furthermore, the photocatalytic reaction was conducted in the presence of isopropanol (IPA) as a •OH scavenger35. As depicted in Fig. 4a, the yield of biphenyl decreased from 80–57% upon addition of 1 mmol of IPA into the reaction (2nd column). These findings point to the indispensable role played by •OH radicals in the photocatalysis.
In the literature, it has been reported that UV irradiation on a composite of TiO2 and MgAl-LDH, as well as MgAl-LDH with partial Ni2+ substitution for Mg2+, and zinc-tin (ZnSn)-LDH can induce the generation of •OH radicals, which can effectively eliminate environmental pollutants36–38. In this study, we have discovered that the visible light irradiation can induce the conversion of surface hydroxyl groups within MgAl-LDH into reactive •OH radicals, facilitating direct C-H arylation. Using visible light instead of UV irradiation enhances solar energy utilization, reduces the likelihood of generating undesired byproduct, and minimises potential safety risks. Therefore, this finding has practical implications for future sustainable chemical synthesis.
The contribution of photo-generated charge carriers in the photocatalysis. As shown in Fig. 5a, the use of KI as an effective quencher for surface-bound free radicals and a hole trapper39 (3rd column) caused a sharp decline in yield from 80–40%, while adding Mn(Ac)3 as an electron scavenger40 leads to a moderate decrease in the yield (4th column). This indicates that photogenerated holes and the resultant surface-bound free radicals that form are critical for achieving high biphenyl yield.
Energy level alignment of the catalyst components and reactant was investigated to understand the transport of photo-induced charge carriers. The LUMO energy level of iodobenzene was measured using cyclic voltammetry with Ag/AgCl as the reference electrode and Fc/Fc+ couple used as the internal standard (Fig. 5b). The ELUMO value of iodobenzene is calculated to be -4.32 eV relative to vacuum or -0.12 eV relative to the standard hydrogen electrode (SHE)41, as depicted in Fig. 5d. Valence band X-ray photoelectron spectroscopy (VB-XPS) method was used to determine the potentials at the valence band maximum (VBM) edge of the Mg2Al1LDH and AgI, which were found to be 2.10 eV and 2.30 eV, respectively (Fig. 5c). The EVB,SHE of Mg2Al1LDH and AgI are 1.96 and 2.16 eV, respectively (Fig. 5d). Band gap energies (Eg) of 5.37 eV for Mg2Al1LDH and 2.85 eV for AgI were obtained from their Tauc plots, (Supplementary Fig. 10). Thus, the corresponding conduction band maximum (CBM) potentials for Mg2Al1LDH and AgI are − 3.41 and − 0.69 eV, respectively (Fig. 5d). This data revealed that AgI NPs efficiently absorb visible light and generate electron-hole pairs. The electrons then transfer to the LUMO of the iodobenzene reactant, activating it, while the holes are captured by the surface hydroxyl groups of Mg2Al1LDH, generating •OH radicals that are crucial for the direct arylation of benzene (see discussion in Fig. 4). This is consistent with the result in Fig. 5a, where the hole scavengers substantially inhibit the biphenyl yield.
The influence of LDH support on the efficiency of •OH radical generation was also explored by comparing the AgI/Zn2Al1LDH catalyst with AgI/Mg2Al1LDH catalyst. The VBM difference between AgI to Zn2Al1LDH is smaller than that between AgI and Mg2Al1LDH42, as discussed in Supplementary Fig. 11. This leads to a slower transfer rate of photo-induced holes in AgI/Zn2Al1LDH, which, in turn, affects the biphenyl yield of this catalyst (Fig. 5a).
Mechanism of the photocatalytic system. Combining the discussions presented in Figs. 4 and 5 above, it can be concluded that the photoinduced charge carriers, •OH radicals, and oxygen vacancies play crucial roles in reducing the energy requirement for the direct C-H arylation.
To confirm the conversion of iodobenzene to benzene radical, induced by the photo-excited electrons from AgI, we used time-dependent ATR-FTIR spectroscopy during visible light irradiation. Specifically, we conducted experiments using the AgI/Mg2Al1LDH catalyst and 4-chloroiodobenzene reactant as a representative sample, because the characteristic peaks of the catalyst and iodobenzene overlap. As displayed in Fig. 6a, the intensity of the C-I bond vibration at approximately 804 cm− 1 consistently diminishes over time, while the signals originating from M-O bond at 551 cm− 1 remain unchanged. An intriguing observation is a slight red-shift at around 651 cm− 1, which corresponds to M-OH bond in the catalyst and can be attributed to the lattice distortion resulting from increased defects, such as oxygen vacancies, within the LDH support44. It is important to note that our comparative analysis revealed no significant spectral changes when testing each individual sample (as demonstrated in Supplementary Fig. 12).
Of particular note is the absence of any homo-coupling side products arising from the corresponding aryl halides. The remarkable selectivity towards the desired cross-coupling product can be attributed to the low concentration of aryl iodide in the system (0.1 M, molar ratio of arene to aryl iodide is about 100). To confirm this, we deliberately increased the concentration of aryl iodide in the reaction system, as evidenced in Supplementary Fig. 13, resulting in the detection of homo-coupling products.
The labelling experiment is a commonly employed technique to determine whether C-H bond cleavage is the rate-determining step9,12,45. We determined a kinetic isotope effect (KIE) value of 1.36 through 1H NMR analysis (see Supplementary Fig. 14), which slightly exceeds previously reported values9,12. This result implies that C-H activation is not the rate-determining step in the direct C-H arylation and lends support the involvement of •OH radicals in the C-H activation. Nevertheless, it is important to note that the rate-limiting step can vary when different arenes are used as coupling partners. For instance, under identical reaction conditions, arenes with electron-withdrawing groups (9–12, Table 2) coupled more efficiently with iodobenzene than those containing electron-donating groups (6, and 13–16, Table 2), highlighting the importance of C-H bond acidity in the arylation process10,12.
The stronger electrophilic properties of •OH radicals play a crucial role in facilitating C-H bond activation and benefiting the arylation reaction compared to species generated in other catalytic systems46,47. Furthermore, our experiment involving an equimolar mixture of benzene and d6-benzene with iodobenzene resulted in a significantly lower yield of deuterated product (78%), further substantiating the above analysis. From a green chemistry perspective, radical-mediated C-H bond activation prove to be superior to transition-metal-based reactions48.
Building upon the preceding analysis, we propose a tentative reaction mechanism for direct C-H arylation on the AgI/Mg2Al1LDH catalyst under irradiation, as depicted in Fig. 6b. Initially, photo-excited electron from AgI NPs inject into the LUMO of aryl iodide, leading to dehalogenation and the formation of an aryl radical and iodide ion (Steps I and II). Simultaneously, the positively charged holes migrate to the LDH support, where they react with surface hydroxyl groups, generating •OH radicals while vacating surface sites to create oxygen vacancies when acquiring electrons from I− ions (Steps II and III). This effectively prevents electron-hole recombination on the AgI NPs. Subsequently, the aryl radical reacts with a neighbouring arene molecule to form a biaryl radical, favoured due to its free-radical nature. Given the high molar ratio of benzene to aryl iodide (over 100) and the increased stability of a larger conjugated radical, it is reasonable to anticipate the prevalence of biaryl radical product in the reaction system compared to the aryl radical. The process of generating •OH radicals by hole transfer, as validated in Figs. 4d and 6a, is counterbalanced by a SET process (C-I bond cleavage). Although it has been reported that •OH radicals can directly react with benzene, this reaction is reversible, and the •OH-benzene adduct is unstable, decomposes into •OH and benzene at temperatures exceeding 330 K49. Furthermore, our catalytic system was deoxygenated prior to the reaction, contrary to the conventional requirement of an oxygen atmosphere for the reaction between •OH radicals and benzene50. Consequently, •OH radicals abstract a hydrogen from the biaryl radical, yielding water as a byproduct, as confirmed by the water measurement test (Supplementary Table 4). Significantly, this C-H bond cleavage involved in hydrogen abstraction is a radical reaction characterized by a low activation energy barrier.
Concurrently, iodide ions are oxidised at sites where •OH radicals are generated, yielding iodine and the oxygen vacancies with two electrons. It is worth noting that while molecular iodine is a recognized catalyst for many organic reactions,51 adding a small amount of I2 (0.1 equiv.) can completely inhibit the photocatalytic reaction. So I2’s involvement in the reaction mechanism is unlikely. Moreover, the surface structure of the AgI/Mg2Al1LDH catalyst can be restored by a regeneration process (Steps III and I).
In summary, the synergistic combination of AgI NPs and MgxAlyLDH with an appropriate x/y ratio proves to be highly effective in catalysing direct C-H arylation under mild conditions. This process stands out for its elimination of the need for additives and bases, and its potential to be driven solely by solar irradiation. The excitation of electrons and holes induced by the light absorption by AgI NPs leads to the conversion of aryl halides into aryl radicals and the generation of •OH radicals on the LDH surface, which abstract hydrogen atoms from C-H bonds. Crucially, the energy alignment within the photocatalysis system allows the promotion of aryl halides conversion into aryl radicals by the excited electrons, while the holes interact with the surface hydroxyl groups on the LDH, yielding •OH radicals capable of abstracting hydrogen atoms from C-H bonds (involving the single electron oxidation of biaryl radicals). These concurrent processes are characterized by both low activation energy barriers due to their radical nature and effectively hinder electron-hole recombination within AgI nanoparticles. Consequently, the photocatalytic direct C-H arylation can proceed efficiently under visible light or solar irradiation without the need for a base and other additives. The main byproducts, water and iodine, can be conveniently separated or recovered. This synthesis approach impeccably fulfils the requirement for a sustainable future18, encompassing efficiency, safety, solar energy utilisation, and environmental friendliness. It opens up new avenues for catalysing C-H arylation via an eco-friendly process and the development of highly efficient catalytic systems.