Selective production of methylindan and tetralin with xylose or hemicellulose

The study addresses the challenge of producing valuable bicyclic aromatics, specifically indan and tetralin, from renewable biomass. While substantial effort has focused on producing monocyclic aromatics (benzene, toluene, xylene) from lignocellulosic platforms, selective synthesis of bicyclic aromatics remains limited. Indan and tetralin are important as jet fuel additives and intermediates for high thermal-stability fuels (e.g., JP-900) and various chemicals. Hemicellulose, a major biomass component readily extractable from residues, offers an abundant and renewable source of sugars (e.g., xylose). The authors propose a two-step strategy: (1) convert xylose or hemicellulose directly to cyclopentanone (CPO) using a non-noble metal catalyst; (2) transform CPO to methylindan and tetralin over acidic zeolites. The work aims to develop selective, scalable routes using inexpensive catalysts and to elucidate mechanisms enabling high yields of bicyclic aromatics.

Prior research has achieved notable progress in converting lignocellulosic platform molecules to monocyclic aromatics and transportation fuels over zeolites and metal catalysts (e.g., ZSM-5 catalyzed furan conversion; catalytic fast pyrolysis routes to BTX). However, selective synthesis of bicyclic aromatics from biomass is scarce. Indan and tetralin are conventionally produced from fossil resources at low yields, underscoring the need for renewable routes. Recent work demonstrated direct CPO production from xylose/hemicellulose using Ru/C, but reliance on precious metals is undesirable. Literature also establishes: (i) chloride salts can act as Lewis acids promoting isomerization (xylose→xylulose) and dehydration to furfural; (ii) Cu-based catalysts can hydrogenate/hydrogenolyze furfural to CPO; (iii) H-ZSM-5’s 10-ring channel structure favors aromatics formation; and (iv) self-aldol condensation of cycloketones can yield dimer intermediates (e.g., [1,1'-bi(cyclopentylidene)]-2-one) that relate to bicyclic aromatic formation under acidic conditions.

Feedstocks and extraction: Hemicellulose was extracted from poplar wood and corncob using methyltetrahydrofuran (MTHF) and 0.1 M oxalic acid in a 50 mL batch reactor at 428 K for 2.5 h (400 rpm), followed by phase separation, solvent recycling, and HPLC analysis.

Catalysts: Zeolites (H-ZSM-5 with various SiO2/Al2O3 ratios, H-Y, H-B, H-USY, SAPO-34) were calcined at 823 K for 4 h before use. Nano H-ZSM-5 (SiO2/Al2O3=160) averaged 100 nm size. Metal catalysts (Cu/SBA-15, Ni/SBA-15, Co/SBA-15; 30 wt% metals) were prepared via ammonia-evaporation deposition on SBA-15, calcined at 673 K, then reduced in 20% H2/Ar at 673 K for 4 h. Bimetallics (Cu-La/SBA-15, Cu-Ce/SBA-15, Cu-Y/SBA-15) were made by incipient wetness impregnation (3 wt% RE), calcined/reduced as above. A NaCl-treated Cu-La/SBA-15 was prepared by contacting in 3 wt% NaCl aqueous solution at 433 K for 4 h. A commercial Pd/C was used for dehydrogenation of cyclopentanol-containing streams.

Step 1 reaction (hydrogenolysis to CPO): Conducted in a 50 mL stainless-steel batch reactor in an organic–aqueous biphasic system. Typical conditions: substrate (0.3 g xylose or hemicellulose equivalents), 10 mL 3–5 wt% NaCl aqueous solution, 10 mL toluene (or other organic solvent), 30 mg catalyst (Cu/SBA-15 or Cu-La/SBA-15), 3 MPa H2 at room temperature (after five H2 purges), heated to 433 K (40 min ramp), 4 h at 500 rpm, then quenched. Optimization considered salt type/concentration, solvent, temperature, and initial xylose concentration. Organic phase products were analyzed by GC; aqueous by HPLC. Carbon yields calculated as carbon in product divided by carbon in substrate fed. Catalyst reusability was tested over multiple cycles; regeneration by calcination in air at 773 K for 4 h followed by reduction in 20% H2/Ar at 773 K for 4 h. Pd/C dehydrogenation: toluene-phase product was treated at 463 K for 4 h to convert cyclopentanol to CPO.

Step 2 reaction (CPO or CHO to bicyclic aromatics): Performed in a 316L tubular fixed-bed reactor. Catalysts were diluted with 40–70 mesh quartz; reactor packed with 20–40 mesh quartz and sealed with quartz wool. Catalyst was pretreated in N2 at reaction temperature for 0.5 h. Liquid feed (CPO or cyclohexanone, CHO) was co-fed with N2 using an HPLC pump. Typical conditions for CPO: 623–723 K, 0.1 MPa N2, weight hourly space velocity (WHSV) 0.45–1.2 g g−1 h−1, initial N2/CPO molar ratio 36:1; products collected after 2 h-on-stream for liquid analysis (Agilent GC with 1,4-dioxane internal standard), and gas analyzed online. For CHO: 613–663 K, 0.1 MPa N2, WHSV ~0.6–1.2 g g−1 h−1, initial N2/CHO=36:1. Conversions defined by feed minus unreacted; carbon yields referenced to converted carbon.

Catalyst characterization: TG-DTA-MS of used bimetallic catalysts (air, 303–873 K, 10 K min−1) for coke assessment; STEM-EDX for elemental mapping; XPS; N2 physisorption (ASAP 2010) for surface area/pore analysis; NH3-chemisorption and NH3-TPD (Autochem 2910) to quantify total acidity and acid strength distribution; FT-IR with pyridine probes to assess Brønsted/Lewis acidity at 423/623 K desorption.

Mechanistic probes: Identification and conversion of intermediates [1,1'-bi(cyclopentylidene)]-2-one (from CPO) and [1,1'-bi(cyclohexylidene)]-2-one (from CHO) over H-ZSM-5 at varied temperatures/WHSV; comparison of feeds (CPO, dimer, 1:1 mixture) to probe pathways; low-temperature runs (e.g., 623 K) to capture C10 oxygenates (HHNO, MTHIO) as intermediates. Influence of NaCl on catalytic behavior assessed via pretreated catalysts and biphasic media studies.

• Non-noble Cu/SBA-15 outperformed Co/Ni analogs for direct xylose hydrogenolysis to cyclopentanone (CPO) in a toluene/NaCl biphasic system; optimized CPO carbon yield ~60% at 433 K, 3 MPa H2, 4 h.
• NaCl (3–5 wt%) and an organic phase (toluene) promoted xylose isomerization/dehydration (xylose→xylulose→furfural) and suppressed over-hydrogenation to xylitol/THFA, thereby increasing CPO formation. Mechanistically, chloride acted as a Lewis acid (promoting isomerization), partially covered Cu particles and withdrew electron density (restraining furan-ring hydrogenation), and reduced catalyst activity toward xylose→xylitol hydrogenation.
• Cu-La/SBA-15 exhibited markedly improved reusability relative to Cu/SBA-15, attributed to reduced Cu leaching and sintering; activity loss was mainly due to coke (polymerization of furfuryl alcohol intermediates) and was reversible by calcination/reduction.
• Dehydrogenation of cyclopentanol by-product over Pd/C (463 K, 4 h) increased CPO carbon yield from 58.5% to 65.8%.
• Applicability to other feeds: Using xylulose, arabinose, and hemicellulose solutions extracted from poplar wood or corncob afforded ~55–60% CPO carbon yields under similar conditions (Table 1). Examples (conversion >99%): xylulose (CPO 60.4%), arabinose (CPO 55.1%), hemicellulose_poplar (CPO 57.5%), hemicellulose_corncob (CPO 54.9%).
• Over acidic zeolites, H-ZSM-5 gave the highest activity/selectivity to C10 bicyclic aromatics (methylindan and tetralin). Tuning SiO2/Al2O3 showed an optimum at H-ZSM-5 (160); after optimizing temperature and WHSV, total carbon yield of methylindan+tetralin reached 65%.
• The dimer [1,1'-bi(cyclopentylidene)]-2-one was identified as an intermediate from CPO self-aldol condensation; feeding this dimer increased C10 aromatics yields under the same conditions and did not require prior trimerization. At lower temperatures (e.g., 623 K), C10 oxygenates (HHNO, MTHIO) were observed as rearrangement intermediates en route to methylindan/tetralin.
• Catalyst stability: Conventional H-ZSM-5 (160) deactivated with time-on-stream due to coking but was fully regenerable by air calcination at 823 K. Nano H-ZSM-5 (160) showed stable performance with no evident activity loss over 24 h for both CPO and CHO feeds.
• Replacing CPO with cyclohexanone (CHO) switched the main products to C12 bicyclic aromatics, dominated by dimethyltetralin; [1,1'-bi(cyclohexylidene)]-2-one was confirmed as an intermediate. About 20% total carbon yield of C6–C9 aromatics was co-produced, which can be hydrogenated to jet-fuel additives (e.g., cyclohexane/alkylated cyclohexanes). Nano H-ZSM-5 (160) offered slightly higher CHO conversion/selectivity due to improved mass transfer.
• Process assessment: Sankey analyses for xylose/hemicellulose routes showed an overall jet-fuel-range hydrocarbon carbon yield of 53.6%, second only to a prior report (73.6%) for hemicellulose to chain alkanes; inclusion of a dehydration step increased overall jet-fuel-range aromatics yield to 60.3%.

The work demonstrates a selective, renewable route to valuable bicyclic aromatics from hemicellulosic sugars by combining a tailored non-noble metal hydrogenolysis step with shape-selective acid zeolite catalysis. The Cu-La/SBA-15 catalyst, aided by a toluene/NaCl biphasic medium, enables high-yield production of CPO directly from xylose/hemicellulose by channeling the pathway through xylulose/furfural while suppressing over-hydrogenation. The La promoter stabilizes Cu dispersion against leaching and sintering, enhancing reusability. The second step over H-ZSM-5 exploits the zeolite’s pore architecture and acidity to convert CPO via a self-aldol dimerization to [1,1'-bi(cyclopentylidene)]-2-one, followed by rearrangement and aromatization to methylindan and tetralin. Identification and successful conversion of dimer and downstream oxygenated intermediates (HHNO, MTHIO) confirm the proposed cascade mechanism. Optimizing zeolite acidity (SiO2/Al2O3=160) and addressing coking with nano-sized H-ZSM-5 deliver up to 65% carbon yield of C10 bicyclic aromatics with good stability. Extending the approach to CHO shifts selectivity to C12 bicyclic aromatics (dimethyltetralin), broadening applicability to lignin-derived cycloketones. Collectively, the findings address the initial goal of selectively producing high-value bicyclic aromatics from abundant, inexpensive biomass components and elucidate mechanistic underpinnings that guide catalyst and process design.

A two-step biomass-to-aromatics process was developed: (1) direct hydrogenolysis of xylose or extracted hemicellulose to cyclopentanone (CPO) over non-noble Cu-La/SBA-15 in a toluene/NaCl biphasic system with ~60% carbon yield; (2) conversion of CPO to methylindan and tetralin over H-ZSM-5 via cascade self-aldol condensation/rearrangement/aromatization, achieving up to 65% total carbon yield of C10 bicyclic aromatics with optimized H-ZSM-5 (SiO2/Al2O3=160). Mechanistic studies identified [1,1'-bi(cyclopentylidene)]-2-one and C10 oxygenates (HHNO, MTHIO) as key intermediates. Nano H-ZSM-5 mitigated coking and maintained stable performance over 24 h. Substituting CHO produced C12 bicyclic aromatics (dimethyltetralin) via analogous pathways. The process offers a promising route to renewable bicyclic aromatics and thermal-stable jet fuel components from lignocellulose. Future work should target further increasing CPO yield/selectivity in step 1, enhancing aromatics selectivity/yield in step 2, and scaling with solvent and salt recycling to improve overall sustainability and economics.

• Catalyst deactivation due to coke formation was observed for H-ZSM-5 during CPO conversion; while regenerable by air calcination, coking remains a limitation for long-term continuous operation (partly alleviated by nano H-ZSM-5).
• Bimetallic Cu-based catalysts still showed slight deactivation over cycles in xylose hydrogenolysis; regeneration is required.
• The step-1 process relies on an organic–aqueous biphasic system with added NaCl; although NaCl solution is reusable, the need for salt and solvent handling may add process complexity.
• Overall carbon yields, while high (e.g., ~60% CPO; up to 65% C10 aromatics), leave room for improvement to approach theoretical maxima; authors note the need for further optimization of both steps.
• Product distributions include co-produced lighter aromatics (C6–C9) and dehydrogenation products (methylindene, naphthalene), indicating some selectivity limitations.