(72c) Continuous Transfer Hydrogenolysis of Lignin Model Compounds in Hot-Compressed 2-Propanol with Ni Catalyst Bed

Lignin, a phenolic polymer of lignocellulose, is gaining attention as a sustainable feedstock to supply renewable aromatic chemicals owing to its structure and abundance [1]. However, most lignin is currently burned as a low-valued fuel and unutilized as a chemical feedstock due to its complex, heterogeneous, and rigid structure [2]. Hydrogenolysis reaction facilitates scission of the strong C-O linkages in lignin, thus promoting its depolymerization to produce phenolic monomers while suppressing unfavored sub-reactions such as repolymerization, frequently observed in other reactions (pyrolysis, acid/base catalysis, etc.) [3]. However, hydrogenolysis generally requires compressed H₂ (usually more than 1 MPa) and noble metal catalysts [4]. Therefore, the reaction has been usually conducted with small batch reactors, with few reports on H₂ flow reaction systems [5]. For the future implementation of the process offering the capacity to convert enormous amounts of lignin into the desired light aromatics, a study on the safer and inexpensive flow hydrogenolysis system is indispensable.

Hot-compressed alcohol, especially 2-propanol (IPA), has been known as an effective hydrogen donor, enabling the replacement of the compressed H₂, referred as transfer hydrogenolysis. The authors reported batch transfer hydrogenolysis of diphenyl ether (DPE), a simple model compound in lignin, employing IPA and non-noble metal catalysts [6]. Kinetic analysis of the hydrogen transfer route (IPA to DPE) in this study indicated the reaction proceeds through the direct hydrogen transfer between IPA and DPE in the liquid phase. Although the simultaneous H₂ formation from IPA (dehydrogenation) was also observed, the analysis result implies the possibility of continuous reaction in simple IPA liquid flow. Despite such research progress, the study on flow transfer hydrogenolysis has not been reported to date. With this regard, continuous transfer hydrogenolysis of DPE was performed. Reaction behavior depending on time on stream was discussed from the viewpoint of the reaction durability and mechanism reported in our previous report [6]. Also, a comparison of the reaction performance between the flow and batch reactions was made.

Methodology

Flow reactor system configuration is shown in Fig. 1. The system is composed of two HPLC pumps, three-way valves for a pulse tracer injection, a pre-heating coil, a reactor main body (3/8” o.d. SUS tube), a cooling coil immersed in an ice water bath, a back pressure regulator, a pressure transducer, and a gas/liquid sampling setup. Two pumps were used for feeding IPA (pump 1) and DPE solution (0.50 M in IPA, pump 2). Three-way valves were used for the pulse tracer experiment to calculate the contact time (time required to pass the fixed bed). A thermocouple well (TCW), made of 1/8’’ o.d. SUS tube, was arranged to measure the temperature inside the reactor. The reactor (4.2 mL) was packed with 2.85 g of Ni/Al₂O₃-SiO₂ (65 wt% Ni content, powder) supplied by Sigma-Aldrich. The apparent void volume of the reactor main body after catalyst packing was determined to be 2.6 mL by the tracer experiments.

At the reaction, IPA was flowed under the objective conditions (feeding rate: 0.50-1.0 mL/min (values at room temperature), reactor temperature: 160°C-180°C, back pressure: 6.0 MPa). The average catalyst contact time was in the range of 2.2 min to 4.4 min. When the reaction system conditions became stable, 20% of pump 1 flow (IPA) was substituted with pump 2 (0.50 M DPE in IPA) to flow the 0.10 M DPE solution in IPA without a change in the total flow rate. This time was recorded as 0 min on time on stream (TOS). A liquid sample was collected with small vials (5 mL) at the desired TOS. The obtained samples were analyzed using GC-FID and HPLC. Conversion of DPE and yield of the DPE derivatives were evaluated based on the initial DPE concentration. Yield of acetone, a resulting product of IPA’s hydrogen donation to DPE or H₂ formation, was calculated based on the feed IPA concentration.

Results and discussion

Fig. 2 presents the TOS profile of the flow reaction experiment at the following conditions: 4.4 min contact time, 160°C. The concentration of DPE and its derivatives increased with TOS and reached a plateau. At TOS = 90 min, conversion of DPE reached 87% with 48% benzene, 20% phenol, 35% cyclohexane, 54% cyclohexanol, and 1.1% cyclohexanone yields. Benzene and phenol are the primary products of the C-O dissociation in DPE. Cyclohexane, cyclohexanol, and cyclohexanone should derive from the over-hydrogenation of the aromatic products. Approximately 90% of DPE and its derivatives were quantified as those compounds, indicating that the reaction proceeded selectively over the other unintended reactions. A significant decrease or change in the product composition was not detected in the range of 3 h after reaching the plateau, suggesting the stability of the developed flow reactor system with Ni catalyst fixed bed.

Fig. 3 shows the corresponding acetone yield. With the absence of DPE and its derivatives, acetone formation (0.34 M, 2.6% yield on IPA basis) was observed at TOS = 0 min. This acetone formation can be attributed to the IPA dehydrogenation, which does not contribute to the targeted DPE hydrogenolysis. With the increase in DPE derivatives concentration in the product, acetone yield also became higher, reaching 0.69 M at a plateau. The increment can be attributed to the IPA hydrogen donation to DPE (targeted transfer hydrogenolysis reaction). It must be noted here that the formation of acetone observed at TOS = 0 needs to be suppressed because this acetone indicates the unnecessary IPA consumption to form gaseous H₂, which may cause safety issues on the process. In this case, the acetone yield (2.6%) was significantly reduced compared to that of the batch reaction (approximately 12% [6]). It probably owes to the high back pressure condition shifting the IPA dehydrogenation reaction equilibrium to the left direction. In contrast, it implies the difficulty of the process operation under ambient pressure.

Phenol was easily hydrogenated to cyclohexanol and had been the minor product in the hydrogenolysis reaction [7]. Since lignin-derived monomers usually possess phenolic structures, inhibition of such over-hydrogenation is important. Fig. 4 shows the conversion vs phenol selectivity plot using the batch reaction results [6] and the flow reaction results (this study). Note that batch reactions were conducted under the temperature of 218°C-253°C, 7.5 min-16 h reaction time, while the flow conditions were set to 160°C-180°C and 2.2 min-4.4 min catalyst contact time. The plots showed a clear right upper shift at the flow reactions, meaning the higher phenol selectivity could be accomplished with the same DPE conversion. When the pseudo-first-order consecutive reaction model assuming DPE to phenol (k₁) and phenol to cyclohexanol (k₂) was applied to the plot, the k₁/k₂ values were 0.029 for the batch and 0.45 for the flow experiments, indicating that the flow reaction was more than ten times selective for C-O dissociation.

In this study, the flow reaction was conducted with the packed bed reactor having higher catalyst mass per reactor volume compared to that of the batch reactions (the catalyst was dispersed in the IPA solution). Such high packing density of the catalyst might have realized high DPE conversion at lower temperature, where the undesired over-hydrogenation was suppressed. Additionally, different diffusion behaviors might have changed the apparent reaction rate. These aspects are under consideration and further discussion is planned to be presented with additional experimental results.

Conclusion

This study conducted DPE transfer hydrogenolysis with IPA in the Ni catalyst fixed bed reactor system. As a result, continuous C-O dissociation of DPE to mono-aromatics or cyclic products was demonstrated with more than 90% balance and without any loss in catalytic activities during 3 h TOS. TOS profile of acetone concentration made it possible to obtain insight into the role of IPA as a hydrogen-donating solvent. In the same manner as the batch reaction, IPA was used both for hydrogenolysis and H₂ production (sub-reaction). The sub-reaction could be suppressed compared to the batch reaction, which will be advantageous in future practical applications. Additionally, phenol selectivity was discussed as the representative of lignin-derived monomers. The flow reactor offered the favored conversion-selectivity profile compared to the batch reaction, demonstrating the feasible reaction performance for lignin transfer hydrogenolysis toward the selective production of aromatics. According to these results, further considerations on phenol selectivity, catalyst reactivation tests, and the usage of other model compounds are in progress. Overall, this study gives important insight into transfer hydrogenolysis (C-O dissociation) with the flow reactor system and contributes to the implementation of safer and inexpensive lignin utilization technologies.

References

[1] F. Brienza et al., RSC Sustainability, 2, 37 (2024).

[2] Y. Wu et al., Energy Fuels, 37, 5206 (2023).

[3] Z. Xue et al., Fuel Process. Technol., 248, 107835 (2023).

[4] M. M. Kumar et al., Molec. Catal., 549, 113474 (2023).

[5] W. Lan et al., Green Chem., 23, 320 (2021).

[6] T. Dowaki et al., Ind. Eng. Chem. Res., 63, 2991 (2024).

[7] H. Wu et al., ACS Sustainable Chem. Eng., 6(3), 2872 (2018).