CeCl₃-Promoted Simultaneous Photocatalytic Cleavage and Amination of C_α–C_β Bond in Lignin Model Compounds and Native Lignin

Introduction

The increasing global demand for energy and fuels has urged us to explore a sustainable alternative to the depleted fossil resources. The production of valuable aromatic compounds from the depolymerization of lignin, the second most abundant non-edible biomass resource other than cellulose and the most abundant aromatic resources on earth,¹ have attracted intense attention in recent years. However, only less than 2% of lignin is utilized to deliver commercial products² due to its composition of a variety of distinct and chemically different bonding motifs,^1–4 with each demanding different reaction conditions for selective depolymerization. Various significant advancements have been made with C–O bond cleavage in both lignin model compounds and native lignin.^5–14 In sharp contrast, it remains a major challenge to achieve the selective cleavage of the C–C bond due to their intrinsic inertness and ubiquity.¹⁵ Some progress has been achieved in the selective cleavage of C–C bond in lignin model compounds with the catalysis of transition metal–based complexes (M = Cu, V, Fe, Ru, Ir, Rh)^16–25 or the highly efficient, photoredox catalyst systems.^17,25,26 More recently, Knowles and co-workers²⁷ demonstrated that the photoredox catalysis could enable the β-scission of both tertiary alcohols and cyclic aliphatic alcohols, as well as the lignin model compounds at room temperature (RT). Our group also disclosed that iridium (Ir) complex could achieve the redox-neutral, photocatalytic cleavage of C–C bond in a wide range of lignin models, as well as the native lignin at RT.²⁸ However, the use of the noble metal–based catalyst would certainly, increase the production cost. Therefore, it is more desirable to develop a low-cost photocatalyst system, which would possess good stability and high efficiency for the practical application of lignin depolymerization. On the other hand, to maximize the lignin valorization, it is of crucial importance to produce high value-added platform chemicals from lignin depolymerization.

Cerium is the most abundant rare earth element, and cerium (III) compounds exhibit widespread application in solid luminescent materials and organic synthesis.^29–31 As an inexpensive, commercially available Lewis acid, cerium(III) trichloride (CeCl₃) also demonstrates a vital role in photoredox catalysis.^32–36 Most recently, Zuo et al³² successfully employed cerium salts to realize the photocatalytically selective C–H amination of methane through the catalysis of alkoxy radicals generated from simple alcohols.³⁴ They discovered that CeCl₃ could also be applied to the photocatalytic ring opening of cycloalkanols. These excellent examples indicated that the presence of a hydroxyl (OH) group is critical for such reactions. It is well-known that the benzylic alpha hydroxyl groups (α–OH) are abundant in lignin as β-O-4 structures featuring a secondary benzylic hydroxyl group at the C_α position and a primary hydroxyl group at C_γ position, accounting for ∼ 50% of all lignin linkages.³⁷ Therefore, we envisioned that CeCl₃ could promote the cleavage of C_α–C_β bonds in lignin; additionally, because cerium is readily abundant in nature, it would make CeCl₃ an ideal catalyst for lignin depolymerization, as it would cut the product cost substantially. Herein, we demonstrate a CeCl₃-promoted, one-pot, visible light–driven photocatalytic strategy to not only achieve the selective cleavage of C_α–C_β bonds in a wide-range of lignin model compounds such as β-O-4 and β-1 models but also furnish aldehyde (up to 97%) and N-containing products (up to 95%) through the amination of C_α–C_β bond (Scheme 1) at RT, without any pretreatments and breakdown of C–O bonds or loss of γ-CH₂OH group. More importantly, this method is efficient for natural lignin depolymerization and affords the corresponding functionalized products. To our best knowledge, our current study is the first example of simultaneous cleavage and amination of C–C bonds in both lignin model compounds and native lignin, and thus, it provides an economic-viable way for the synthesis of useful N-containing complexes.

• Download figure
• Download PowerPoint

Methods

Materials and general procedures

All chemical syntheses and air- and moisture-sensitive materials manipulations were carried out in inflamed Schlenk-type glassware on a dual-manifold Schlenk line or a high-vacuum line or an argon-filled glovebox. Phenol; 2-methoxyphenol; diisopropylamine; 2-bromoacetophenone; 2-bromo-4′-methoxyacetophenone; 2-bromo-3′-methoxyacetophenone; sodium borohydride; ethyl (2-methoxyphenoxy)acetate; benzylmagnesium chloride; 2-bromo-4'-hydroxyacetophenone; formaldehyde; inorganic salts; solvents; and the standard substances benzaldehyde, 4-methoxybenzaldehyde, 3,4-dimethoxybenzaldehyde, 3,4,5-trimethoxybenzaldehyde were purchased from Adamas-beta^®, Shanghai, China. n-BuLi (1.6 M solution in hexanes), guaiacol, CeCl₃, nBu₄NCl, nBu₄NI, nBu₄PCl, nBu₄NOAc, and di-tert-butyl azodicarboxylate (DBAD) were purchased from J&K (Beijing, China). All chemicals obtained were used as received unless otherwise specified. Namely, tetrahydrofuran (THF) was dried over sodium/potassium alloy and distilled under a nitrogen atmosphere before use for the synthesis of the lignin model compounds. Acetonitrile (CH₃CN) was dried over CaH₂ distilled under a nitrogen atmosphere prior to use for light-induced catalytic reactions.

Each reaction was monitored by thin-layer chromatography (TLC), visualizing with ultraviolet (UV) light. Column chromatography purifications were performed using silica gel. The reaction mixtures were analyzed using Alliance High-Performance Liquid Chromatography (HPLC) system (Waters, Milford, MA), equipped with autosampler, C18 column (length: 75 mm, internal diameter: 4.6 mm, temperature: 35 °C), and UV/vis detector (λ = 220 nm). A mixture of methanol/water (CH₃OH∶H₂O; 40∶60) was used as the mobile phase, with a flow rate of 1.0 mL/min. Nuclear magnetic resonance (NMR) spectra were recorded on an Avance II 500 (500 MHz, ¹H; 126 MHz, ¹³C) instrument (Bruker, Shanghai, China) at RT. Chemical shifts for ¹H and ¹³C spectra were referenced to internal solvent resonances. Mass spectra recordings were obtained using the Bruker MicroTOF Q II. Gas chromatography (GC) was analyzed using the Shimadzu GC-2014 (Beijing, China), and a flame ionization detector with the following parameters: carrier gas: N₂ (2 mL/min); program temperature: 60 °C (1 min), 5 °C/min to 210 °C (7 min), and 5 °C/min to 220 °C (5 min); injector temperature: 250 °C; detector temperature: 250 °C; split ratio: 30; and injection volume: 5 μL. Yields were determined by GC analysis using naphthalene as an internal standard.

Light-induced catalytic reactions

A 10 mL Schlenk tube was charged with a lignin model compound (0.2 mmol, 1 eq.), CeCl₃ (2 mol %), tetrabutylammonium chloride (nBu₄NCl; 5 mol %), DBAD (1.1 eq), and 1.0 mL anhydrous acetonitrile (CH₃CN) at N₂ atmosphere. The reaction was irradiated with a 30 W blue LED lamp (λ = 460 nm) at RT for 12 h. The resulting mixture was dissolved to constant volume in a 50 mL volumetric flask and measured by HPLC.

Extraction and Depolymerization of Lignin

Extraction of lignin

We charged a round bottom flask with 10.0 g pine sawdust, 50 mL 1,4-dioxane, and 1.7 mL HCl (37 wt %), and heated to reflux at 85 °C in an oil bath for 3 h. After cooling to RT, the mixture was added with 3.36 g sodium bicarbonate (NaHCO₃), stirred for another 30 min, after which it was filtered and washed with 10 mL of dioxane. Then the solution was concentrated at 40 °C under reduced pressure. The resulting dark-brown oil was diluted with 30 mL ethyl acetate (EtOAc) and added dropwise to 500 mL of hexane to precipitate the lignin. After filtration, the collected lignin was washed with hexane (50 mL), followed by diethyl ether (50 mL) for 5 min each while sonicating. The recovered lignin was dried overnight at RT in a desiccator to afford 1.01 g pine lignin.

Depolymerization of lignin

Under atmosphere N₂ conditions, 50 mg of pine lignin, 4.9 mg of CeCl₃, 4.9 mg nBu₄NCl, and 50 mg DBAD were dissolved in 1.0 mL dry CH₃CN, in a 10 mL Schlenk bottle. The reactor was illuminated under a Kessil H150B LED Grow Light (Richmond, CA, USA) at RT for 12 h. Then, still under nitrogen condition, the second batch of 4.9 mg of CeCl₃ and 4.9 mg nBu₄NCl was added, with subsequent illumination for another 12 h.

Results and Discussion

Optimization studies

In our initial studies, we observed the generation of aldehyde and hydrazinium products through this photocatalytic depolymerization and functionalization of simple β-O-4 lignin model A with DBAD under N₂ atmosphere. A 95% yield of benzaldehyde ( aa) and 94% yield of hydrazinium ( ba) were obtained for the reaction performed with 2 mol % of CeCl₃, 5 mol % of tetrabutylammonium chloride (nBu₄NCl), whereas ∼ 90% product yields could still be maintained with the reduced amount of CeCl₃ (0.5 mol %) and nBu₄NCl (1.5 mol %) (entries 1 and 2, Table 1). Without nBu₄NCl, the product yields drastically decreased, and most of the substrates remained unreacted (entry 3, Table 1). Replacing nBu₄NCl with nBu₄NI or nBu₄NOAc quenched the reaction completely, whereas nBu₄PCl afforded comparable product yield (entries 4–6, Table 1), probably resulting from the drastic enhancement of blue-light absorption by the chloride anion, thus giving rise to a more efficient photoexcitation.³² Moreover, CeCl₃, DBAD, and light were all indispensable for the reaction (entries 7–9, Table 1). Without blue LED irradiation, no product was formed, even when heated at 80 °C for 12 h, confirming the photocatalytic nature of the reaction (entry 10, Table 1). Further, the reaction was not moisture or air sensitive, since the addition of one equivalent of water still furnished aa in 88% yield and ba in 89% yield (entry 11, Table 1). In sharp contrast, the reaction was shut down completely by Ir-based photocatalyst in the presence of oxygen²⁸; however, performing the reaction in the air afforded aa in 83% yield and ba in 82% (entry 12, Table 1).

Table 1 | Studies of the Depolymerization of Lignin Model A.^a

Entry	Variation From the “Standard” Conditions (Entry 1)	Yield (%)b
		aa	ba
1	CeCl3 (2 mol %), nBu4NCl (5 mol %)	95	94
2	CeCl3 (0.5 mol %), nBu4NCl (1.5 mol %)	89	90
3	CeCl3 (2 mol %)	43	40
4	nBu4NI, instead of nBu4NCl	0	0
5	nBu4NOAc, instead of nBu4NCl	0	0
6	nBu4PCl, instead of nBu4NCl	91	90
7	No CeCl3	0	0
8	No DBAD	0	0
9	No light	0	0
10	No light, 80 °C	0	0
11c	Water (1.0 eq.)	88	89
12c	Air atmosphere	83	82

^aStandard condition: 0.2 mmol substrate, 2 mol % CeCl₃, 5 mol % nBu₄NCl, 1.1 eq. DBAD, 1.0 mL dry CH₃CN, N₂ atmosphere, RT, 12 h, 30 W blue LED. ^bThe yields were measured by HPLC. ^c50 W LED, 24 h.

Scope of lignin models

It is recognized that there is a close correlation between the catalyst reactivity and the lignin model substrate structure. With the optimized condition in hand, we continued to investigate the efficacy of the CeCl₃ catalyst system on the depolymerization of lignin model compounds with varying positions of the methoxyl groups of the aromatic rings. Varying the methoxy group position on either benzene ring (left, red) or the phenoxy group (right, blue) did not affect the catalytic performance. The corresponding β-scission/amination product, aldehydes (up to 96% yield) and hydraziniums (up to 95% yield) were produced in high to excellent yields for the lignin model B to H in the absence of γ–OH group without the production of acid or phenol byproducts (entries 1–7, Table 2). By using substrate B′ containing a hydroxyl group at C₄-position on the benzene ring (left, red), we investigated the effectiveness of the hydroxyl group on the catalytic performance of the CeCl₃ catalyst system and achieved a corresponding β-scission/amination products ab′ and ba in 70% and 79% yield, respectively (entry 8, Table 2). We also examined the reaction with more complex β-O-4 lignin models possessing an aliphatic hydroxyl (γ–OH) group. This CeCl₃-based photoredox catalytic system demonstrated precisely an excellent selectivity and high activity toward cleavage and amination of C_α–C_β bond in a wide array of β-O-4 lignin models ( I– O) containing both α–OH and γ–CH₂OH groups (entries 9–15, Table 2). Notably, only the C_α–C_β bondcould be cleaved selectively during the reaction whereas the C_β–C_γ or C_β–O bond remained intact, thus furnishing the corresponding aldehydes ( aa– ad, up to 97% yield) and hydraziniums ( bc– bd, up to 94% yield) in good to excellent yields. Such chemoselectivity might be attributable to the stability of the possible cleavage products. Since β-1 is another prevalent linkage found in lignin,^1–4,38 we also examined the effectiveness of this CeCl₃-based photocatalytic system on the depolymerization of β-1 lignin model P. Approximately 94% of P was converted into the corresponding cleavage and amination of C_α–C_β bond product aa and be in 86% and 88% yield, respectively (Scheme 2), which was different from our previous studies in which the noble-metal Ir-based photocatalytic system demonstrated complete ineffectiveness for such β-1 lignin model compound without –OMe group on the left benzene ring due to the incapability of the proton-coupled electron transfer (PCET) reaction.²⁸

• Download figure
• Download PowerPoint

Table 2 | Selectively Photocatalytic Cleavage and Amination of C_α–C_β Bond in β-O-4 Lignin Model Compounds.^a

Entry	Substrate	Conv.(%)b	Yield (%)b		Entry	Substrate	Conv.(%)b	Yield (%)b
1		95	ab, 91	ba, 91	9		92	aa, 87	bc, 86
2		>99	ac, 96	ba, 94	10		>99	ab, 94 (92)c	bc, 94 (81)c
3		>99	ac, 94	ba, 95	11		>99	ac, 93	bc, 93
4		96	aa, 93	bb, 92	12		>99	ad, 96	bc, 94
5		98	ab, 94	bb, 92	13		96	ab, 92	bd, 93
6		99	ac, 96	bb, 94	14		>99	ac, 95	bd, 94
7		>99	ad, 95	bb, 95	15		>99	ad, 97	bd, 94
8		82	ab′, 70	ba, 79

^aStandard condition: 0.2 mmol substrate, 2 mol % CeCl₃, 5 mol % nBu₄NCl, 1.1 eq. DBAD, 1.0 mL CH₃CN, N₂ atmosphere, RT, 12 h, 30 W blue LED. ^bThe yields were measured by HPLC. ^cIsolated yield.

Collectively, these results suggest that this photocatalytic strategy is a simple and convenient method that could not only achieve one-step, selective cleavage of C_α–C_β bond in lignin models at RT, but also furnish useful amination products. In particular, such N-containing complexes could be transformed into valuable chemicals. For example, product bc could be cyclized to generate 1,3,4-oxadiazine in 60% yield and 1,2-diazetidine in 24% yield. The 1,3,4-oxadiazine, in turn, could be utilized in the formulation of insecticide or pest control chemicals, whereas the 1,2-diazetidine is a novel N-heterocyclic compound with unique structure, versatile activities, and safety for nontarget organisms; thus, displays a wide range of potential applications in organic synthesis and medicinal chemistry (Scheme 3), including its uses in central nervous system (CNS) disorders and absorption, distribution, metabolism, and excretion (ADME) profiling.^39–43

• Download figure
• Download PowerPoint

Mechanistic studies

In order to probe more insights into the CeCl₃ catalytic reaction, several UV–vis experiments were conducted ( Supporting Information Figure S1). It turned out that lignin model A did not display any absorbance peak above 300 nm, whereas the combination of CeCl₃ with lignin model A exhibited only a weak absorption in the region of 300–400 nm. The addition of nBu₄NCl to the above mixture induced an intense absorption at 330 nm, probably due to the interaction of the chloride anion with the 5d orbitals of cerium.^44,45 A previous report suggested that the Ce^III complex could be activated into Ce^IV species in situ [E_1/2(Ce^III/Ce^IV) = 0.41 V vs SCE in CH₃CN] by a photoinduced single-electron oxidation with DBAD* (E* = 1.66 V vs SCE in CH₃CN).³³ Therefore, we synthesized (nBu₄N)₂Ce^IVCl₆ and found its mixture with lignin model A exhibited an absorption peak at λ = 370 nm. We initiated the reaction utilizing CeCl₃ in the presence of nBu₄NCl and DBAD and observed the reaction for an induction time within 30 min. Meanwhile, we found no apparent conversion of lignin model A within this period (Figure 1a), an observation which indicated that CeCl₃ functionalized as a precatalyst.^33,34 Furthermore, UV–vis absorption spectra of a mixture of CeCl₃/lignin/nBu₄NCl/DBAD in CH₃CN showed that the absorbance at approximately 360–370 nm increased in the first 10 min and then decreased gradually with the reaction time, suggesting the transformation of Ce^III into catalytic quantities of Ce^IV species, followed by the reconversion of Ce^IV species back to Ce^III during the reaction ( Supporting Information Figure S2). Remarkably, without irradiation, the mixture of (nBu₄N)₂Ce^IVCl₆ and lignin model A under basic condition, displayed an absorption peak at λ = 370 nm, whereas after 30 min of LED irradiation, an intense absorption peak was apparent at λ = 330 nm, attributable to Ce^III species ( Supporting Information Figure S4), thereby, indicating that the Ce^IV species would have been converted back to Ce^III species after the photoinduced ligand-to-metal charge-transfer (LMCT).^32,33,45 Accordingly, we employed (nBu₄N)₂Ce^IVCl₆ to examine the mechanism of the reaction. We observed 92% of lignin model A was converted to the corresponding cleavage and amination of C_α–C_β bond product aa and ba in 89% and 85% yields, respectively in the presence of DBAD (Scheme 4a), which confirmed that the Ce^IV species is the real catalyst for the reaction. Moreover, the addition of the free-radical trapping agent, (TEMPO), quenched the reaction entirely, which revealed the free-radical nature of the reaction (Scheme 4b). As expected, the presence of the α–OH group was critical for the β-scission; this CeCl₃-based catalyst system was ineffective completely for the reaction when ACH₃ having α-methoxy group instead of α–OH group was used as the substrate (Scheme 4c).

• Download figure
• Download PowerPoint

Furthermore, the fact that the reaction was unfeasible in the dark suggested this photocatalytic cleavage and amination of C_α–C_β bond in the lignin model substrate could be activated/deactivated by light, thus leading to the controlled regulation of the photocatalytic depolymerization process. To investigate the nature of this responsiveness, we performed the reaction by switching on/off of the external light stimuli. When the experiment was exposed to a 460 nm light source, an 18% conversion was achieved in 2 h (Figure 1b). After removal of the light source, no further conversion was notable during the time course of 1 h; however, reexposure to 460 nm light, led to further reaction progress. This cycle could be repeated several times to attain a quantitative conversion as high as ∼ 99%, indicating the exhibition of precise control of the photocatalytic C_α–C_β bond scission/amination process.

• Download figure
• Download PowerPoint

Based on the above observations, we proposed a plausible mechanism for this photocatalytic β-scission/amination of C_α–C_β bond process (Figure 2), as follows: (1) In the presence of nBu₄NCl, CeCl₃ is coordinated to the lignin model A to form a Ce^IIICl_n/lignin complex through the deprotonation of the hydroxyl group, which subsequently went through the single-electron oxidation with DBAD* to generate a Ce^IVCl_n/lignin species along with the release of DBAD ^−·. (2) The photoinduced LMCT homolysis of Ce^IVCl_n/lignin species afforded the key alkoxy radical intermediate along with the release of Ce^IIICl₃ for the next catalytic cycle. (3) Such alkoxy radical intermediate could substantially weaken the adjacent C–C bond,⁴⁶ thus enabling the cleavage of C_α–C_β bond to generate the aldehyde product aa and alkyl radical. (4) Subsequently, the highly reactive alkyl radical species could be coupled with DBAD to form the nitrogen-centered radical intermediate,^47,48 which went through the single-electron-transfer/proton transfer (SET/PT) process to furnish hydrazinium product ba. In another competitive pathway, alkyl radical could also react with DBAD ^−· to produce ba.

• Download figure
• Download PowerPoint

Practicability of the CeCl₃-based photocatalytic system

A gram-scale reaction was performed by using lignin model J as the substrate, furnishing the selective C_α–C_β bond cleavage and amination product ab and bc in 92% and 81% isolated yield, respectively ( Supporting Information Scheme S1). Moreover, the practicability of this new method could be verified by the catalyst reuse experiment using a 1∶1.1 molar ratio of lignin model A and DBAD under blue-light irradiation at RT (check Supporting Information for detailed procedure). To our great satisfaction, this CeCl₃-based photocatalytic system had a long-life of catalytic performance and maintained a constant catalytic activity. Even after repeating the reaction cycle nine times, ∼88% conversion of lignin model A could still be achieved with an 80% yield of aa and a 79% yield of ba (Figure 3). Both the reuse and the switch experiments (vide supra) demonstrate the stability of the CeCl₃-based photocatalytic system, which is crucial for the practical application.

• Download figure
• Download PowerPoint

Degradation of trimeric lignin model and real lignin

The identification of such a highly efficient and selective photocatalytic system for C_α–C_β bond cleavage and amination at RT prompted us to apply this strategy to the depolymerization of a more complex lignin model structure. A trimeric model compound ( Trimer) was examined (Scheme 5), and the results showed that 85% of Trimer was converted to the corresponding cleavage and amination product, ab, bf, and ba in 81%, 78%, and 78% yield, respectively.

• Download figure
• Download PowerPoint

The success achieved in the depolymerization of lignin model compounds and the trimer inspired us to examine further the applicability of this CeCl₃-based photocatalytic system on the depolymerization of real or natural product lignin. Through the extraction of pine sawdust with HCl/dioxane, we prepared the softwood pine lignin. After photocatalytic depolymerization, two-dimensional (2D) heteronuclear single-quantum coherence (HSQC) spectra of the lignin obtained indicated that the major peaks attributable to β-O-4 units (A) and phenylcoumaran (β-5) units (B) in the pine lignin had almost disappeared completely, and new cross peaks became apparent, which were consistent with the expected cleavage products (Figure 4). More specifically, selective cleavage of C_α–C_β bond led to the disappearance of A_α (δ_C 73, δ_H 5.0 ppm) and A_β (δ_C 84, δ_H 4.2 ppm) in β-O-4, as well as the formation of corresponding aldehyde group V_α (δ_C 192, δ_H 9.9 ppm) attributed to vanillin product (V) and H_β (δ_C 94, δ_H 5.9 ppm) attributed hydrazine product (H). Furthermore, the aromatic regions peaks obtained from NMR spectra of the depolymerized lignin (Figure 4c and d) indicated that the majority of the following G-type lignin (G) found in pine lignin [G2 (δ_C 110, δ_H 6.9 ppm), G5 (δ_C 115, δ_H 6.7 ppm), and G6 (δ_C 119, δ_H 6.8 ppm)] were converted to aromatic ring [V2 (δ_C 108, δ_H 7.4 ppm), V5 (δ_C 115, δ_H 7.3 ppm), and V6 (δ_C 125, δ_H 7.4 ppm)] attributable to unit V. The GC analysis of products confirmed that ∼11.94 wt% (note: The N-containing part of the product came from DBAD rather than lignin.) of small aromatic compounds (hydrazine converted from phenols) could be obtained after the depolymerization of pine-dioxane lignin by this CeCl₃-based photocatalytic strategy (Scheme 6 and Supporting Information Scheme S2–S4, Figure S5). Altogether, these results demonstrated that this CeCl₃-based photocatalytic strategy is an effective method for achieving the selective cleavage and amination of C_α–C_β bond in natural lignin depolymerization.

• Download figure
• Download PowerPoint

Conclusion

We developed a one-step, CeCl₃-based photocatalytic strategy to not only achieve highly selective C_α–C_β bond cleavage but also produce functionalized aromatics, such as aldehyde (up to 97%) and N-containing products (up to 95%) in high to excellent yield from both β-1 and a wide range of β-O-4 lignin model compounds at RT. The long-life catalytic performance of this CeCl₃ catalytic system was verified by catalyst reuse experiment, as well as on/off switching experiments. More importantly, this method could be utilized for the photocatalytic depolymerization of both trimeric lignin model compound and natural lignin product, thus providing a convenient and economic-viable way to access highly value-added platform chemicals through lignin depolymerization.

• Download figure
• Download PowerPoint

Supporting Information

Supporting Information is available.

Conflict of Interests

The authors declare no competing interests.