Catalytic Hydration of Cyclohexene to Cyclohexanol for Nylon-6 Production: A Review

Abstract

Nylon-6 (PA6) is a widely used engineering plastic, with cyclohexanol as the key intermediate of the benzene–cyclohexene–caprolactam chain. Cyclohexene hydration is considered a green alternative to cyclohexane oxidation owing to its high atom economy and selectivity. This review compares direct and indirect hydration within a four-level framework (catalyst, process, system and industrial chain). Direct hydration is constrained by thermodynamic equilibrium and oil–water interfacial mass transfer, with an industrial single-pass conversion of about 10–13% under excess-water conditions. Reported intensification routes include synergistic acidity–hydrophobicity–diffusion design of HZSM-5, cosolvent screening, and RD → HPRD → RDWC integration. Indirect hydration reconstructs the reaction network via esterification–hydrogenation (or hydrolysis/transesterification), can reach conversions above 99%, but adds process complexity, catalyst-stability demands and hydrogen consumption. Benchmarking the schemes on a per-ton basis, route preference is discussed within a three-dimensional framework (plant scale, hydrogen price and carbon constraints). The data required for a fully quantitative ranking are identified. The framework may aid process selection under given hydrogen and carbon-market conditions.

1. Introduction

Nylon-6 (PA6), produced by ring-opening polymerization of caprolactam, is one of the most widely used engineering plastics and synthetic fibers [1,2]. As illustrated in Figure 1, industrial caprolactam production is dominated by three benzene-based routes: cyclohexane oxidation, cyclohexene hydration (direct and indirect) and phenol hydrogenation, all converging on cyclohexanone, which is converted to caprolactam via ammoximation and Beckmann rearrangement against a worldwide cyclohexanone-oxime capacity exceeding 5 Mt/a (industrial capacity) [3,4]. Cyclohexanol is the key intermediate of the cyclohexene route, and its production efficiency largely determines that route’s competitiveness [2,3].

Figure 1. Industrial benzene-based synthetic network to nylon-6.

Table 1 summarizes the four industrial routes. Cyclohexane oxidation is mature but suffers from low KA-oil selectivity, large waste output and explosion hazards and still accounts for about 90% (estimated) of cyclohexanone production [2,3,4,5,6]. Phenol hydrogenation offers high conversion and selectivity (Pd-based catalysts approaching 100%) but is constrained by feedstock availability and precious-metal costs [2,5]. Direct and indirect cyclohexene hydration circumvent the oxidation step with high carbon atom utilization and are regarded as the most promising green alternatives [6]. China, the world’s largest caprolactam producer, faces a need under its “dual-carbon” strategy for low-energy, high-atom-economy cyclohexanol processes; with volatile hydrogen prices reshaping the chain, the practical question is which route is preferred under a given plant scale, hydrogen price and carbon constraint. The hydrogen price largely determines the operating cost of hydrogen-consuming routes such as esterification–hydrogenation; carbon constraints turn the emission gap between grey and green hydrogen into a cost; and plant scale decides whether the high recycle energy of direct hydration or the high fixed investment of indirect hydration dominates. No route is intrinsically superior, and the answer shifts with these three conditions.

Table 1. Comparison of major industrial routes for cyclohexanol production.

Process Route	Conversion	Selectivity	Atom Economy	Major Drawback
Cyclohexane oxidation	≤5%	~80–90%	~80–85%	Explosion risk, large waste output
Phenol hydrogenation	>99%	~99%	~95.5%	High feedstock precious-metal catalyst cost
Direct hydration	~10–13%	>99%	>99%	Mass-transfer limitation, low equilibrium conversion
Esterification–hydrogenation	>99%	99.4%	99.4%	Long process flow hydrogen-source dependence
Esterification–hydrolysis	≈56%	≈94%	~100%	Water–carboxylic acid separation energy

Direct hydration: ~10–13% is for industrial equilibrium single-pass conversion; higher catalyst/cosolvent values are laboratory apparent conversions and are not directly comparable.

Direct hydration is the one-step addition of water to cyclohexene over a solid acid catalyst. Its single-pass conversion is limited by two intrinsic, intertwined constraints: thermodynamic equilibrium (about 10–13% conversion near 393 K, with a measured thermodynamic equilibrium constant Ka of 0.125–0.135 at 393.15 K, activity basis [7]) and oil–water interfacial mass transfer (cyclohexene dissolves in water to only about 200–500 ppm, the 500 ppm value being an Aspen Plus simulation [6]). The two constraints interact: raising the temperature accelerates mass transfer and kinetics but shifts the equilibrium further left, while increasing the water-to-oil ratio favors the equilibrium but dilutes the oil-phase reactant at the catalyst surface; every intensification lever is therefore a trade-off within this “dual-constraint plane” rather than a simple activity increase. Indirect hydration instead reconstructs the reaction network by introducing a carboxylic acid medium: the equilibrium-limited hydration is split into esterification and hydrogenation (or hydrolysis/transesterification) steps, bypassing the equilibrium bottleneck as a whole and making conversions above 99% achievable, at the price of hydrogen consumption, process complexity and copper-catalyst stability.

Most existing reviews treat catalyst modification in isolation [8] or stop at route-specific process summaries, and none benchmarks direct and indirect hydration against each other under a unified framework. The present work therefore proposes a four-level assessment framework (catalyst → process → system → industrial chain), performs cross-route benchmarking under unified functional units and system boundaries, and constructs a three-dimensional decision framework incorporating plant scale, hydrogen price and carbon constraints to provide systematic decision support for process selection.

This review was compiled through a structured search of Web of Science, Scopus and patent databases (Espacenet/Derwent), covering 1990–2026 with keyword combinations such as “cyclohexene hydration”, “cyclohexanol”, “HZSM-5”, “reactive distillation”, “cyclohexyl acetate hydrogenation/hydrolysis” and “transesterification”. Records were screened by title/abstract and full text; priority was given to peer-reviewed experimental studies, validated simulations, pilot or industrial demonstrations, and patents disclosing operating data. Because the two route families are conventionally reported from different feedstock boundaries, all quantitative data throughout the manuscript are labeled by a unified evidence-type code (a = optimized simulation, b = pilot-scale, c = measured/experimental, d = engineering estimate, e = theoretical), and cross-boundary comparisons are kept semi-quantitative.

2. Direct Hydration: Catalysis and Process Intensification

The four-level framework begins with the direct-hydration family at its catalyst and process levels. As illustrated in Figure 2, direct cyclohexene hydration proceeds through protonation of the C=C bond to a cyclohexyl carbenium ion, which reacts with water to give cyclohexanol [2]; protonation is the rate-determining step [8]. The route has a short flowsheet, high selectivity, no hydrogen consumption and 100% theoretical atom economy. The only industrially demonstrated process is the Asahi slurry route, which suspends HZSM-5 in the water phase and operates near the equilibrium ceiling (about 10–13% for single-pass conversion), with the slurry configuration entailing catalyst attrition, solid–liquid separation and regeneration [2,4,9,10,11,12]. The industrial window of 373–403 K is a kinetic–selectivity compromise: higher temperature accelerates kinetics but favors oligomer by-products [2]. Recent modeling shows that the catalyst-layer efficiency factor is governed by two-phase fluid properties and reaction conditions rather than packing geometry, so catalyst “activity” and “accessibility” must be treated separately [13].

Figure 2. Mechanism of direct cyclohexene hydration. Step 1 (slow, RDS): protonation of the C=C bond forms the secondary cyclohexyl carbocation; Step 2 (fast): water nucleophilic attack forms the oxonium ion; Step 3 (fast): deprotonation yields cyclohexanol and regenerates H⁺.

2.1. HZSM-5 Catalyst Design and Performance Evolution

At the catalyst level, HZSM-5 is the most widely used solid acid catalyst for direct hydration [8,14]. Its ten-membered-ring (10-MR) channels suppress side reactions and keep cyclohexanol selectivity above 99% [12]: ZSM-5 reached 99.2% selectivity at 7.5% conversion, whereas mordenite/ZSM-12 dropped to about 74% and 12-ring beta zeolite to only 35.4%, both through dicyclohexyl ether formation, with the hydrophobicity threshold requiring SiO₂/Al₂O₃ ≥ 20 [12]. Conventional HZSM-5 has two key drawbacks, micropore diffusion resistance and surface hydrophilicity [8], and research has focused on three levers: acidity regulation, hierarchical-pore construction and surface hydrophobization.

(1) Acidity regulation. Conversion correlates with framework strong Brønsted acid sites, while Lewis acid sites initiate side reactions [2,8]. HNO₃ treatment raised conversion from 9.0% to 12.2% at 99.7% selectivity (measured) [14]; Ce³⁺ incorporation created Si–OH–Ce strong Brønsted sites (strong Brønsted/Lewis ratio 9.2 vs. 2.3), giving 12.7% conversion and 99.9% selectivity [15]; and a SAR-tunable series showed that the strong Brønsted acid density peaks at SAR = 50, where the yield is highest [16]. The strong-acid-site logic is further supported across the modification strategies: Ga incorporation raised the Brønsted density from 0.4 to 0.5 mmol/g with conversion rising from 8.6% to 10.1%, and mild Na₂CO₃ alkali treatment raised the strong acid density from 330 to 350 µmol/g with conversion increasing from 10% to 13% [15].

(2) Hierarchical-pore construction. Alkaline treatment creates mesoporosity by desilication: across the Na₂CO₃ concentration series, the external surface area and mesopore volume reached 150 m²/g and 0.19 cm³/g, and the 0.5 M sample improved conversion/selectivity from 10.0%/98.1% to 13.0%/99.3% (measured). The window is narrow, since ≥1 M treatment destroys the MFI framework and lowers zeolite recovery from 86% to 68% [13]. NaOH, evaluated as an alternative base, was optimal at 0.2 M but slightly less effective than Na₂CO₃ (12.7%/99.2%) [17]. Acid treatment forms mesopores by dealumination (BET 320 → 368 m²/g) [14], and reducing the b-axis from 700 to 30 nm shortens diffusion paths (external area 28 → 80 m²/g) [18].

(3) Surface hydrophobization. Reducing surface polarity enriches cyclohexene at the pore mouths and promotes cyclohexanol desorption [19,20,21]. OTS grafting raised the water contact angle from 0° to 122.6–146.1° and the conversion from 4.15% to 20.87% at 403 K (measured) [19,20]. Combining acid treatment with OTS grafting (HCl-OTS-HZSM-5) reached 24.07% conversion with the hydration barrier lowered from 73.78 to 46.67 kJ/mol (measured) [21]. Stirring/static comparisons show that the Pickering emulsion eliminates the interfacial limitation: stirring raised unmodified HZSM-5 from 5.03% to 9.81% but barely affected OTS-HZSM-5 (20.87% static vs. 21.13% stirred) [19]. Acid-only hydrophobization is milder (contact angle 17° → 23.65°) and only partially relieves the limitation, since stirring still markedly improves the acid-treated catalyst [21]. The cost is acid-site coverage (total acidity 441 → 382 µmol/g), and the organic layer is prone to hydrothermal detachment [8,19]. Covalently grafted sulfonic acid (SFZ-15) doubled the Brønsted density (294 → 551 µmol/g), giving >12% conversion and >99% selectivity (measured) over five cycles, with the unoxidized thiol intermediate reaching only 11.4% conversion and confirming the sulfonic acid groups as the active sites [22].

(4) Crystal morphology and stability. A 30 nm b-axis nanosheet (NS-HZ5-30) raised initial conversion from 2.9% to 7.5% and reached 14.0% with 100% selectivity (TOF 3.1×, measured), with only 2.5% carbon after 20 cycles versus 4.88% for the bulk catalyst [18]. Template-free microparticles sustained 9.6% conversion over 1200 h with 96.8% selectivity (measured) [23]. Deactivation has two sources: reversible coking and irreversible dealumination, which together influence the regeneration strategy [23].

(5) Shaping. For reactive-distillation internals, powder catalysts must be shaped, but alumina binders dilute active sites, and zeolite thin films install only limited catalyst [4]. Template-free, seed-induced synthesis gives performance comparable to commercial ZSM-5 (10.7% vs. 10.2% yield) at lower cost [16].

Direct-hydration catalyst research has shifted from single-site acidity enhancement toward synergistic design of acidity, pore structure and wettability. Because activity gains were measured under different conditions, direct numerical comparison is unreliable; the qualitative pattern in Table 2 supports the working hypothesis that hydrophobization (20.87–24.07%) and cosolvent-assisted mass transfer (up to 47.5%) outperform acidity regulation alone (12.2–12.7%), with mass transfer and interfacial contact dominating apparent activity in the practical window. Testing this requires a unified kinetic basis, which the literature lacks.

Table 2. Comparison of catalytic performance of modified HZSM-5 catalysts for cyclohexene hydration.

Catalyst (Modification)	Modification Feature	Reaction Conditions	Conv./Sel./Yield (%)	Ref.
HZSM-5 (HNO3-treated)	Dealuminated, mesoporous	393 K, 4 h	12.2/99.7/12.2	[14]
SFZ-15 (sulfonic-acid-grafted)	-SO3H-grafted, enhanced Brønsted acidity	393 K, 2 h	14.3/99.8/14.3	[22]
PEG-HZSM-5 (PEG-modified)	PEG-templated mesopores, exposed acid sites	403 K, 4 h	15.93/99.35/15.8	[24]
Ce-HZSM-5 (Ce3+-incorporated)	Strong Brønsted sites (Si–OH–Ce)	398 K, 3 h	12.7/99.9/12.7	[15]
HZSM-5 nanosheet (b-axis 30 nm)	Shortened diffusion path, high acid-site accessibility	398 K, 3 h	14.0/100.0/14.0	[18]

Data were obtained under different conditions and are not normalized; direct cross-catalyst comparison is not valid.

2.2. Cosolvent-Induced Phase-Behavior Regulation and Mass-Transfer Intensification

The catalyst-level levers in Section 2.1 raise intrinsic activity but leave the oil–water mass-transfer bottleneck untouched; cosolvents act directly on this bottleneck by regulating phase behavior [25]. Mechanistically, the hydrophilic solid acid catalyst is surrounded by the aqueous phase, so hydration proceeds there and the aqueous-phase cyclohexene concentration directly governs conversion; cosolvents reconstruct the phase behavior of the cyclohexene–water system, lowering the interfacial tension and enhancing local mass transfer [26]. Zhang et al. [25] built a dual-indicator (mutual-solubility, partition-coefficient) screening framework: from 2440 solvents, candidates that were unstable, reactive or unsuitable for distillation recovery were successively removed, leaving 306 for COSMO-RS prediction and finally 10 for experimental validation; m-cresol gave the highest cyclohexanol yield (47.5%, measured), with phenolic solvents showing the highest yields in the reported data (39.6–47.5%, vs. 36.3–36.7% for alcohols and 23.4–26% for ketones; 13.7% without solvent). Molecular dynamics traced the gain to the amphiphilicity of m-cresol: it thickens the interfacial layer (0.46 → 1.13 nm) and raises the cyclohexanol partition coefficient (4.88 → 23) [25]. Wang et al. [27] extended the molecular screening to process design: among about 40 solvents tested, only o-methylphenol, benzyl alcohol and acetophenone raised the yield without side reactions, with the effective solvents being partly miscible with water (extraction of cyclohexanol from the aqueous phase matters more than dissolving cyclohexene); acetophenone raised the yield from 12.3% to 27.6% (measured) but cut whole-process TAC and emissions by only about 8% and 17% (simulated), the cosolvent-recovery separation energy consuming most of the yield dividend, an economic ceiling for the cosolvent route. Mechanistically, the gain is not purely solubility: ethylene glycol lowers the apparent activation energy from 77.69 to 67.23 kJ/mol (measured) [7], and isophorone is the cosolvent of choice in reactive distillation (yield 12.0% → 24.3% in the patented process) [26].

Most cosolvents remain at the laboratory stage: weakly acidic solvents may attack HZSM-5 acid sites, cosolvent-based RD schemes can fail with real feeds containing cyclohexane, which must be purged from the organic decanter phase with large purge flows and severe cosolvent loss [11], cosolvent dosage has an optimum (excess cosolvent occupies active sites and distorts the equilibrium [7]), and cosolvent addition and surface hydrophobization may be partially mutually exclusive. These couplings constitute a clear research gap.

2.3. Reaction–Separation Coupling and Process-Intensification Routes

Catalyst and cosolvent levers improve the reaction rate, but the back-mixing and delayed product removal of conventional stirred-tank reactors still leave single-pass conversion equilibrium-limited. Reaction–separation coupling instead removes cyclohexanol in situ and shifts the equilibrium forward; with sufficient reboiler duty and catalyst holdup, each of the three schemes below can drive the hydration to near-complete conversion (99.9%, simulated specified for RD [11]; 99.86% optimized for RDWC [28]). These schemes belong to the process-intensification family [29,30]; catalytic RD is an industrially established technology [31,32] whose design is supported by established guidelines for tray hydraulics, catalyst holdup and feasibility screening [33,34] and by equilibrium- and rate-based column models [35].

(1) Reactive distillation (RD). The conventional RD configuration for cyclohexene hydration is shown in Figure 3. Chen et al. [11] designed an excess-water flowsheet with a water/cyclohexene molar ratio of 2.0 and a top decanter and stripper, specifying 99.9% conversion: the reboiler duty is 46.76% lower than the Asahi benchmark (simulated) [36] and 32.86% lower than the formic acid two-column flowsheet of Katariya et al. [37], the two baselines being non-comparable; Hu et al. [26] and Sun et al. [38] report no energy or TAC baseline. The high conversion is not free: for stoichiometric feed the simulated conversion drops to 99.4% (simulated), the reflux ratio and total reboiler duty rise to 91.1 and 28.42 MW, and the required 5 atm operation violates the Amberlyst-15 temperature ceiling of 393 K; the excess-water design needs only 6.55 and about 3.07 MW, which is why it makes the 99.9% specification both feasible and economical [11]. Sun et al. [38] experimentally validated a catalytic-distillation process over A-36 resin with isophorone as cosolvent, obtaining a simulated conversion of 40.63% under optimized conditions with 12 reaction trays, a reflux ratio of 3 and a distillate-to-feed ratio of 0.25; this value and Chen’s 99.9% specification come from different models and boundary conditions and are not directly comparable. A side-reactor configuration SRC cuts TAC by 11.41% (simulated) relative to the traditional process when catalyst effectiveness is low: in the evaluation of Ye et al., the traditional, idealized-RD, effectiveness-corrected-RD and SRC designs give annual costs of USD 4.12M, USD 3.90M, USD 6.60M and USD 3.65M, respectively [39]. Hu et al. [26] proposed a three-column process with 99.3% conversion and ≥99.2% purity with plant-wide control schemes.

Figure 3. Conventional reactive distillation (RD) configuration for cyclohexene hydration.

(2) Heat-pump-assisted reactive distillation (HPRD). The organic-phase-refluxed RD with side stripper (ORDS) configuration is shown in Figure 4. Liu et al. [9] proposed organic-phase-refluxed RD with a side stripper ORDS, which cuts energy by 51% relative to organic-phase-refluxed conventional RD, OCRD; adding mechanical vapor recompression gives HPRD with 58% energy saving, 24% TAC reduction and 62% CO₂ reduction relative to ORDS, with TAC −58% relative to OCRD, at a specific energy of only 718.78 kJ/kg versus 1388 kJ/kg (simulated) for Chen, 2607 for Asahi and 3827 for Zheng. Across the four RD configurations and HPRD, specific energy decreases stepwise from 3927 kJ/kg for ACRD through 3564 for OCRD, 2037 for ARDS and 1727 for ORDS to 719 for HPRD, with CO₂ emissions falling from 0.26 to 0.04 kg/kg cyclohexanol; most of the ORDS saving comes from the RD integration itself, the heat pump being a secondary intensification [9]. Heat-pump assistance is feasible only when Q/W > 10 by the Plesu criterion: the Q/W ratios are 3.71 for ACRD and OCRD, 12.1 for ARDS and 25.5 for ORDS, which is why only ORDS qualifies [9]. The temperature-difference requirement is echoed in other systems: vapor-recompression integration is worthwhile only for close-boiling systems [40,41], and a double-heat-pump variant in an ethyl-acetate hybrid separation eliminated external heating steam entirely, cutting cooling-water demand by 86% and total CO₂ emissions by about 59% under mixed/renewable electricity [42]. As a cross-system reference not transferable to the cyclohexene route, internally heat-integrated RD, R-HIDiC, achieves energy, TAC and CO₂ reductions of 29.37%, 32.56% and 52.79% versus conventional RD [43].

Figure 4. Organic-phase-refluxed reactive distillation with side stripper (ORDS) configuration.

(3) Reactive dividing-wall column (RDWC). RDWC integrates reaction and all separation functions in one column shell. Zheng et al. [28] optimized a cyclohexene-hydration RDWC by response-surface methodology with NSGA-II: 99.132 wt% cyclohexanol purity, 99.86% conversion, and energy-cost and TAC reductions of 7.76% and 5.74% relative to the pre-optimized design, with energy dominating the operating cost (reboiler duty and TAC correlate at 0.9965); the optimization used response-surface surrogates validated against rigorous Aspen Plus simulations (deviation below 1% over 100 Pareto-optimal solutions) with the final trade-off selected by CRITIC weighting, and the process is isophorone-assisted, linking it to the cosolvent-design discussion in Section 2.2. RDWC gains in other esterification systems (e.g., isopropyl acetate −21.31% TAC [44], ethyl acetate [45,46], heat-pump/dividing-wall combinations [40,41,47]) are system-specific and not transferable to the cyclohexene route.

As shown in Table 3, the three schemes follow a staged selection rule: RD for retrofitting and control simplicity [26,39], HPRD when deep energy and CO₂ reduction dominate [9], and RDWC for grass-roots large plants [28]; in all three, catalyst holdup and two-phase hydraulics, rather than tray count alone, should be the primary design variables [43].

Table 3. Comparison of three process-intensification schemes.

Scheme	Single-Pass Conversion	Equipment Integration	Equipment Investment	Control Difficulty	Energy-Saving Potential	Application Scenarios	Ref.
RD	~99% (simulated)	Reaction + distillation coupling	Medium	Low	Moderate	Retrofitting existing columns	[11,39]
HPRD	~99% (simulated)	Reaction + distillation + MVR	Medium	Medium	Moderate–high	Deep low-carbon upgrading	[9]
RDWC	~99% (simulated)	Complete separation within the tower	Medium–high	High	Significant	New large-scale plants	[28]

3. Indirect Hydration: Catalytic Pathways and Process Intensification

Because the equilibrium conversion of direct hydration cannot be fundamentally raised by engineering means, indirect hydration instead reconstructs the reaction network itself. The classic acetic-acid-mediated route comprises two steps (Figure 5):

Figure 5. Indirect hydration reaction scheme.

Step 1 (esterification): C₆H₁₀ + CH₃COOH ⇌ C₆H₁₁OOCCH₃ (cyclohexyl acetate);

Step 2 (conversion): path a (hydrolysis) C₆H₁₁OOCCH₃ + H₂O ⇌ C₆H₁₁OH + CH₃COOH; path b (hydrogenation) C₆H₁₁OOCCH₃ + 2H₂ → C₆H₁₁OH + CH₃CH₂OH.

This reconstruction transfers the equilibrium bottleneck: esterification is a homogeneous liquid-phase reaction with an equilibrium conversion above 68% (versus only about 10–13% for direct hydration) [6]. The hydrogenation step is strongly driven and near-quantitative, consuming 2 mol H₂ per mole of product and coproducing ethanol; the hydrolysis step consumes no hydrogen but is equilibrium-limited and requires large acetic acid recycle and dehydration energy. The integrated esterification–hydrogenation process reaches a measured benzene-based atom economy of 99.4% [6]. Hydrogen from natural-gas reforming raises whole-process carbon intensity, whereas green-electrolysis hydrogen can lower it below that of the hydrolysis path; this is why hydrogen price and source are treated as a decision dimension.

3.1. Esterification Catalytic Systems and Medium Selection

In the reconstructed network, the first bottleneck is the esterification step itself: the catalytic difficulty is that the acidity must activate the cyclohexene double bond while suppressing dimerization and isomerization, and the catalyst must remain stable in the water- and acid-containing liquid phase; because the carbenium-ion intermediate can also react with cyclohexene to form dimers, stronger Brønsted acidity raises activity while aggravating this side reaction, making selectivity a trade-off in acid-strength design [29]. Guang et al. [48] established a pseudo-homogeneous kinetic model for cyclohexene–acetic acid esterification over Brønsted–Lewis bifunctional heteropolyacid-based ionic-liquid solid acids (BLA-HPA-ILs); the optimal ½Cu catalyst gave 65.92% conversion and 98.61% selectivity (measured) with stable activity over six reuses; the pseudo-homogeneous model spans 333.15–363.15 K with forward and reverse apparent activation energies of 56.21 and 92.34 kJ/mol (measured), and in the doping spectrum (Cu, Co, Ni, Zn, Fe, Al) 1/2Cu gave the best conversion (65.92%) while excess copper raised the Lewis acidity to the point of hindering acetic acid attack and the other metal-doped analogues (Fe, Co, Ni, Zn, Al) reached 51–56% conversion, while excessive copper doping reduced the conversion to 2.3–45%, with a reaction enthalpy of −43.7 kJ/mol [48].

In medium selection, the acetic acid system is the most industrially mature: less corrosive than formic acid, with a suitable boiling-point difference and full compatibility with the caprolactam chain. The formic acid route is more reactive but has intrinsic drawbacks: formic acid decomposes above about 333 K, forcing vacuum operation (below about 0.6 bar), and its azeotropic behavior demands dedicated NRTL parameters: a binary heterogeneous minimum-boiling cyclohexene–formic acid azeotrope and a ternary reactive azeotrope give rise to liquid–liquid phase splitting, requiring NRTL parameters from dedicated equilibrium experiments [4]. Imam et al. [36] showed from first principles that formic acid reacts faster below about 353 K and gives a reactive-distillation-feasible volatility ordering, whereas acetic acid makes high-purity separation from the reaction zone difficult; their idealized reactor–separator analysis indicates that a temperature-shifted equilibrium could in principle cut the energy requirement to about one-quarter of the Asahi benchmark (simulated), while the actual reactor–separator concepts (RS and two-stage R1S1 + R2S2) show that the hydrolysis step remains constrained by phase splitting and formic acid recovery [36]. Catalyst stability further favors the acetic route: formic acid collapses sulfonic-acid-functionalized UiO-66 within three reuses, whereas cyclohexanecarboxylic acid leaves it intact over five cycles [49,50].

As summarized in Table 4, esterification catalysts are dominated by solid acids: cation-exchange resins are active but thermally limited (about 393 K) [4,6,11]; zeolites offer tunable acidity but need suitable pore geometry for the bulky ester [2]; MOF-based solid acids allow molecular-level tuning of acid-site density: defect engineering (acetic acid modulator) raised the total acid content (Brønsted + Lewis, Lewis-dominated) of sulfonic-functionalized UiO-66 to 244.5 µmol/g at 1345 m²/g (measured) and restored activity, whereas the non-defective UiO-66–SO₃H sample (367 m²/g) was completely inactive; the defect-engineered catalyst gave 50.5% conversion and 86.6% selectivity (measured) remaining at 45.4% conversion and 87.0% selectivity after five cycles [49], and Jiang et al. [50] further tuned the acid density up to 249.5 µmol/g (83.4% conversion, 99.8% selectivity in formic acid esterification); heteropolyacid-based solid acids combine strong acidity with good reusability [48]. Sun et al. [51] integrated esterification and hydrolysis on one hierarchical polystyrene-based solid acid bed, reaching 90.2% cyclohexanol yield (80.3% after ten recycles) with a per-active-site TOF of 447.3 h⁻¹ (measured), versus 51.5% (TOF 228.7 h⁻¹) for Amberlyst-15 and 31.4% (TOF 164.4 h⁻¹) for HZSM-5.

Table 4. Comparison of catalytic performance of representative systems for cyclohexene esterification.

Catalyst	Acid	Conditions	Conv./Sel./Yield (%)	Ref.
BLA-HPA-ILs	Acetic acid	333.15–363.15 K	65.92/98.61/65.00	[48]
UiO-66-SO3H-def	Cyclohexanecarboxylic acid	423 K, 8 h, 3 wt%	50.5/86.6/43.7	[49]
HP CLPS-SO3H	Acetic acid	353 K, 6 h, 10 wt%	Yield 90.2	[51]

Data were obtained under different acid-to-olefin ratios, temperatures and loadings and are not normalized.

A caveat applies to the esterification comparison in Table 4: the conversions were obtained at different acid-to-olefin ratios, temperatures and catalyst loadings, so the apparent ranking reflects reaction conditions at least as much as intrinsic activity. A fair comparison would require a common kinetic basis such as turnover frequency at matched conversion, which is currently missing, an important standardization gap for the field.

3.2. Conversion Pathways of Cyclohexyl Esters: Hydrolysis and Hydrogenation

Once esterification is established, the route choice splits at the ester intermediate. Hydrolysis generates cyclohexanol and the carboxylic acid by nucleophilic substitution: its advantages are high material utilization, no hydrogen consumption and 100% theoretical atom economy, but it is equilibrium-limited, requires excess water and incurs large water–acid separation energy; this trade-off of separation energy for zero hydrogen consumption favors high hydrogen prices and disfavors high steam/electricity prices. Independent analyses characterize the hydrolysis step as complex and energy-intensive, and Freund and coworkers deemed the route economically unviable without an appropriate catalyst [6]. Liu et al. [52] developed dual-SO₃H heteropolyacid-based solid acids reaching 90.56% conversion with 94.86% selectivity (versus H₂SO₄ 95.77%/60.63%, measured), reusable five times; the bare heteropolyacids are water-soluble and difficult to separate; the temperature window is bounded by β-elimination pyrolysis of the ester to cyclohexene and acetic acid, so hydrolysis runs under mild conditions. Kumar et al. [4] validated the two-step reactive-distillation concept experimentally with formic acid as a reactive entrainer in a 2.35 m mini-plant esterification column packed with SULZER-DX and KATAPAK-S (Amberlyst-15): esterification reached 98.3% conversion with 75.5 mol% ester (mini-plant, measured) in the bottom product, and the hydrolysis column gave 45–51% ester conversion at about 50% cyclohexanol purity in an unoptimized multi-purpose column, with optimized coupled-column simulations projecting near-complete conversion and >99 mol% purity (simulated). Ionic liquids can serve as phase-transfer catalysts reducing the water-to-ester ratio and suppressing acid volatilization, but their recovery cost and impurity accumulation during recycling remain bottlenecks to scale-up.

Hydrogenation breaks the ester C–O bond to give cyclohexanol and the corresponding alcohol; unlike the equilibrium-limited hydrolysis route, it is strongly exothermic with a theoretical conversion approaching 100%, allowing high conversion and selectivity under mild conditions with a suitable catalyst. Copper-based catalysts are the most widely used in industry [53], their activity originating from the synergy of Cu⁰ (H₂ dissociation) and Cu⁺ (ester-carbonyl polarization). The ester hydrogenation network also includes transesterification with ethanol to ethyl acetate, pyrolysis to acetic acid and cyclohexene, and cyclohexanol dehydration/dehydrogenation to cyclohexanone and cyclohexane as by-product paths that set the selectivity ceiling [53].

Across these strategies, catalyst design has converged on the same activity model and a small set of practical levers. Because sintering or reduction of Cu⁺ destroys the Cu⁰/Cu⁺ synergy, the catalysts differ mainly in how the Cu particles are fixed, how Cu⁺ and oxygen vacancies are stabilized, and which support is chosen to suppress side reactions.

Sintering is the dominant deactivation mode of Cu catalysts: the Hüttig and Tamman temperatures of Cu are only 407 and 678 K [54], and confinement is a common countermeasure. The LDH-derived Cu/MgO/Al₂O₃ holds Cu in its layered structure and keeps 99.59% conversion and 98.94% selectivity for 500 h (measured), with the space–time yield falling only from 0.91 to 0.82 g g⁻¹ h⁻¹, against 0.92 to 0.50 for commercial Cu/ZnO/Al₂O₃ [54]. The carbon framework confines Cu-CeO_x in the same way: after five recycles the Cu⁺ fraction moved from 45.4% to 45.0% and the oxygen-vacancy content from 72.1% to 69.7%, with no Cu²⁺ detected [55]. The same confinement strategy removes the need for hydrogen pre-reduction, since nascent CO from carbonization reduces the oxide in situ [55]. The MOF-derived Cu/CeO₂ fixes Cu on its carrier skeleton, and the Cu loading fell only from 38.9 to 37.1 wt% over eight recycles [56]. Noble-metal systems use the same idea: in Au–Ag/SBA-15, Ag prevents Au aggregation, keeping the catalyst stable for 200 h, whereas monometallic Au loses 25% of its activity within 100 h [57].

Cu⁺ and oxygen vacancies are regulated through oxide supports or lattice doping, and the preparation method matters as much as the composition. The sol–gel Cu₃Zr₇ reached 97.4% conversion and stayed stable for 72 h (measured) at the low H₂/ester ratio of 20, whereas its coprecipitated and impregnated counterparts managed only 9.4% and 8.7% despite the coprecipitated sample having smaller Cu particles (31 vs. 41 nm). Activity is thus governed by surface Cu⁺/oxygen-vacancy sites and the acid–base character of ZrO₂, not by particle size [58]. The ternary Cu₁Al₁Mn₁ spinel takes a different balance: a high Cu⁰ fraction (Cu⁺/(Cu⁰ + Cu⁺) of only 21.6%), the highest high-valent-Mn ratio of 4.2 and the most abundant oxygen vacancies (41.8% of total oxygen). With this combination it outperformed both binary spinels and ran for more than 500 h without Cu aggregation, the vacancies adsorbing and activating the carbonyl while Cu⁰ dissociates H₂ [53]. The CeO₂ systems give the clearest comparison: the Cu⁰–CeO₂ electronic interaction raised the Cu⁺ fraction and oxygen-vacancy content to the highest of the series, and the Ce-free control fell to 64.1% conversion and 51.5% selectivity [55]; MOF-derived CeO₂ lowered the apparent activation energy to 52.8 kJ/mol from 63.3 and 70.5 kJ/mol for cube- and rod-shaped CeO₂ [56]. The LDH-derived catalyst also shows that the Cu⁺ fraction must stay within a defined window: conversion followed a volcano with reduction temperature, 56.8%, 78.9% and 56.4% at 453, 533 and 593 K, around an optimal Cu⁰/(Cu⁰ + Cu⁺) of about 0.70 [54].

Support identity strongly influenced selectivity: Cu/ZrO₂ converted 84.3% but delivered only 15.4% selectivity, while Cu/Al₂O₃ and Cu/SiO₂ reached 73.1% and 43.2% [56]. Interface studies explain why at the molecular level. At the Cu–O–Zr³⁺–O_v interfacial site (O_v, oxygen vacancy), H₂ dissociates on interfacial Cu and spills over onto ZrO₂; the TOF correlates linearly with the surface Zr³⁺ fraction rather than with Cu⁰/Cu⁺, the activation energy drops from 118 to 72 kJ/mol, and the catalyst loses less than 3% activity over 120 h while unmodified Cu deactivates within 80 h [59]. At the Cu–O–SiO₂ interface, H₂ is activated heterolytically into Cu–Hδ⁻ and SiO–Hδ⁺; NaOH poisoning removes this pathway, the kinetic isotope effect falls back to 1.6, and DFT indicates that heterolytic splitting is thermodynamically favored [60]. The same balance between intermediate adsorption and H₂ dissociation holds for subnanometric Cu clusters, whose nuclearity can be tuned from single atoms through Cu₃/Cu₄ to nanoparticles: they maintained a 95% intermediate yield and stayed stable for 100–150 h at a batch scale of at least 1 kg [61].

The pilot results are consistent with these design rules. Lanthanum promotion of Cu/ZnO/SiO₂ strengthened surface basicity, suppressed acid-catalyzed side reactions, and raised the cyclohexanol selectivity from 98.3% to 99.7% without compromising conversion [6]. From laboratory screening to industrial scale-up, the logic of catalyst modification is the same: stabilize the active sites and suppress side reactions rather than simply push up activity. The hydrogenation pathway shows higher reported conversions and selectivities and has pilot-scale validation; the hydrolysis pathway keeps its structural advantages in atom economy and zero hydrogen consumption, but its process energy consumption remains under-reported, and a complete techno-economic comparison under the same boundary is a clear research gap.

3.3. Whole-Process Integration and Pilot Validation

Catalyst-level performance is already near-optimal; the remaining gains lie in whole-process integration. Industrialization of the hydrogenation pathway relies on direct esterification–hydrogenation coupling, in which the esterification-column bottom stream feeds the hydrogenation reactor directly. The Sinopec pilot (8000 t/a, >1000 h) achieved esterification and average hydrogenation conversions above 99%, cyclohexanol and ethanol selectivities of 99.4% and 99.5%, and a benzene-based atom economy of 99.4% (pilot-scale), with the reaction heat recovered in a waste-heat boiler for cascade steam use; the demonstration catalyst was La-promoted Cu/ZnO/SiO₂, giving 99.5% conversion and 99.7% selectivity at 483 K, 5.0 MPa (pilot-scale) and a H₂/ester molar ratio of 30 [1,6]. Later pilot studies point to two further conclusions. Structured catalytic packings are technically viable: in esterification RD over HY@SiC foam packing, raising the acid-to-olefin ratio increased the cyclohexene conversion while excess cyclohexene had the opposite effect, and a double-recycle design approached complete conversion, although the zeolite catalysis itself remained a research gap [62]. Process configuration and intermediate purity matter as much as catalyst performance: the transesterification route was validated on seepage catalytic packing in a reactive-column plus pressure-swing-column design, closing the material loop by hydrolyzing the methyl-acetate by-product back to acetic acid and methanol at lower cost than the baseline [63]; and intermediate purity is not always beneficial, since the 99.90 wt% scheme gave the lowest TAC while its methyl-acetate hydrolysis unit dominated the whole-process gas emissions [64].

Three intensification paths are relevant in whole-process integration [1,6]: cascade utilization of reaction heat, closed-loop management of carboxylic acid recycling, and cross-unit catalyst sharing (bifunctional catalysts with both esterification and hydrolysis activity); esterification–transesterification variants, by contrast, suffer from ether and olefin side reactions and are equilibrium-limited [6]. The competitiveness of the indirect route depends more on system-level factors than on remaining catalytic gaps: acetic acid synergy, hydrogen price and intermediate purity. The research focus shifts accordingly from catalyst discovery toward process integration and techno-economic optimization [63]. Beyond the 8000 t/a pilot, Sinopec has slated first commercialization at 200 kt·a⁻¹ [1].

4. Unified Comparison and Industrial-Chain Decision-Making

Section 2 and Section 3 reviewed the catalyst- and process-level intensification levers of the two route families; the remaining question is which route is preferred at the system and industrial-chain levels. Table 5 addresses this by benchmarking the four representative routes on a unified functional unit (1 t cyclohexanol); because the two families are conventionally reported from different feedstock boundaries, the comparison is kept semi-quantitative, with the evidence-type labels defined in the footnote to Table 5.

Table 5. Cross-route comparison of main technical routes for cyclohexanol production.

Indicator	Direct Hydration	Indirect Hydrolysis	Indirect Hydrogenation	Indirect Transesterification RD
Conversion	99.86% a (RDWC)	98.3% c (esterification); 45–51% c (hydrolysis)	>99% b	33.1% b
Selectivity/purity	>99% c; 99.132 wt% a	~50% c	99.4% b	>99.5% b; 99.90 wt% a
Atom economy	>99% d	~100% e	99.4% c (benzene-based)	~100% e
Energy/TAC	718.78 kJ/kg; TAC −24% a	~2500–3500 kJ/kg d; n.r. (expected higher than direct hydration)	n.r.	422,117 USD/a (1000 t/a; 99.90 wt% scheme)
Technology maturity/pilot scale	Industrial (Asahi); intensification: lab	Lab	Pilot 8000 t/a, >1000 h b	Pilot

^a = optimized simulation; ^b = pilot-scale; ^c = measured (experimental); ^d = estimated; ^e = theoretical. n.r. = not reported. Benzene-based atom economy is defined as the mass fraction of benzene carbon atoms retained in the final cyclohexanol, calculated on the Trost atom-economy basis. Values reflect reaction-, process- and chain-level metrics; percentages are relative to their own baselines and are not additive, and simulated and experimental/pilot values are not directly comparable. The indirect-hydrolysis energy range is an engineering estimate, for trend reference only.

4.1. Cross-Route Benchmarking and Industrial-Chain Carbon Context

These comparisons rest on established methodological tools: atom economy [65], the E-factor [66], greenness metrics [67,68] and formal LCA under defined functional units and system boundaries [69]. Three conclusions follow from the benchmark (Figure 6), which is semi-quantitative because the route families are reported from different feedstock boundaries and some cells are not reported.

Figure 6. Semi-quantitative cross-route comparison of cyclohexanol production routes. (a) Specific energy consumption of direct-hydration reaction–separation schemes; (b) benzene-based atom economy of the four industrial routes.

First, HPRD shows the lowest specific energy among direct-hydration schemes, 718.78 kJ/kg, 48–81% below Chen, Asahi and Zheng (simulated). These specific energies are equivalent-heat-load values defined as

SEC = (W_e/η + Qth)/mout, η = 0.384

(1)

where W_e is the compressor power, η the thermal efficiency, Q_th the flash/stripper heat duty and m_out the product mass flow; the values thus include compressor power converted at a common thermal efficiency, from Aspen Plus simulations [9]. The 1388/2607/3827 kJ/kg comparison values are transcribed from [9] and may use different energy bases.

Second, indirect esterification–hydrogenation reaches a benzene-based atom economy of 99.4% [6] (i.e., the mass fraction of benzene carbon atoms retained in the final cyclohexanol, calculated on the Trost atom-economy basis) versus 83.7% for cyclohexane oxidation, 95.5% for phenol hydrogenation and 99.3% for cyclohexene hydration, with per-pass yields of 3.7% (cyclohexane oxidation), 5.1% (cyclohexene hydration) and 14.9% (phenol hydrogenation) [6]; the transesterification RD flowsheet shows a total annualized cost of about 422,117 USD/a at 1000 t/a (≈422 USD/t for the 99.90 wt% intermediate scheme, simulated) [64], and a process-level comparison on the same 1 kt/a basis without separation places transesterification with a TACP of 74.46 USD/t below hydrolysis at 78.99 USD/t and far below ester hydrogenation at 313.60 USD/t, which is dominated by the high-pressure reactor [44]. Third, as compiled in Table 6, the reported TAC and CO₂ savings are consistently positive but baseline-dependent and mutually non-additive, so they should be read as directional evidence rather than absolute benchmarks. Under the current data conditions, only condition-dependent preferential judgments can be offered, and the “not reported” cells define the measurement agenda for the field.

Table 6. Benchmark of direct-hydration process-intensification schemes on a semi-quantitative basis.

Configuration	Energy/TAC Improvement	Baseline
Simulated RD	Reboiler duty −46.76%/−32.86%	Asahi/Katariya
Simulated SRC	TAC −11.41%	Traditional reaction-separation process
HPRD	Energy −58%, CO2 −62%, TAC −24%/−58%	ORDS/OCRD
RDWC	Energy cost −7.76%, TAC −5.74%	Pre-optimized RDWC
Cross-system R-HIDiC	Energy −29.37%, TAC −32.56%, CO2 −52.79%	Conventional RD

All improvement percentages are relative to the baseline process cited in each original study; baselines differ across entries and are not directly comparable.

Beyond the route-level benchmark, the industrial chain adds its own carbon and feedstock constraints: Hong et al. [70] showed that steam, electricity, benzene and process chemicals dominate the environmental impact of a Chinese 200 kt/a caprolactam plant (GWP 6.9–7.5 × 10³ kg CO₂-eq/t, measured/LCA), with energy-source switching (GWP 6.4 × 10³ kg CO₂-eq/t) and NO_x/ammonia abatement (respiratory-inorganics, terrestrial-ecotoxicity and GWP cuts of 15.8%, 11.1% and 14.3%) as quantified levers. Downstream recycling matters as much: nylon-6 waste alcoholysis cuts GHG emissions by 63% (with 65% and 75% reductions in fossil-resource and water use) and converts real waste streams to ε-caprolactam at 67–76% yields (measured), at a recycled minimum selling price 30% below the five-year virgin-market average [71]. Grey hydrogen shifts emissions upstream, whereas renewable-power electrolysis can bring the hydrogenation route close to the carbon level of the hydrolysis route under the same boundary [69,72]; a cradle-to-gate LCA of the esterification–hydrogenation route as a function of hydrogen source remains missing and is the key data gap for the decision framework. Industrial practice has already demonstrated chain-level atom-economy optimization: in the Eversun process, rectification–dehydrogenation of the by-product cyclohexane back to benzene raises the benzene efficiency from 78% to 98% [3].

4.2. Three-Dimensional Decision Framework: Scale, Hydrogen Price and Carbon Constraints

Both the route-level benchmark and the chain-level carbon context leave route choice context-dependent; the framework below makes that context explicit. Within each route, the framework also applies at the process-structure level: the same chemistry can be arranged either as a conventional reaction-followed-by-distillation sequence, R-F-D, or as reactive distillation, RD. Direct hydration offers the most direct evidence: the industrial Asahi benchmark is an R-F-D-type flowsheet with a single-pass conversion near the equilibrium ceiling of about 10–13% [10], whereas RD raises the simulated conversion to 40.63% (model experimentally validated [38]) or to a design-specified 99.9% [11] and cuts the reboiler duty by 46.76% relative to Asahi [11]; Ye et al. [39] caution, however, that RD is not always superior, since low catalyst effectiveness inside the column favors a side-reactor configuration, and in the indirect routes RD has been validated at mini-plant [4] and pilot [44] scale. Under strong carbon constraints, the RD family, HPRD and RDWC, inherits this structural advantage [9,28]. Based on Table 5, this review proposes a three-dimensional decision framework with plant scale, hydrogen price and carbon constraints as its axes, as a heuristic rather than a predictive model. Dimension 1, plant scale: direct hydration has low fixed investment but high operating cost from large recycles; indirect hydration is the reverse, with a scale crossover around 10⁴–10⁵ t/a (estimated), illustrative and not calibrated. The upstream benzene-to-cyclohexene segment (selective hydrogenation of benzene) is not benchmarked here; the cost and purity of the cyclohexene feed propagate into both route families and remain an explicit data gap for chain-level techno-economic assessment. Dimension 2, hydrogen price and source: this dimension affects only the hydrogenation route; the hydrogen term is $m_{{\mathrm{H}}_2}H$, about 40 kg H₂ per tonne of product (stoichiometric), and the carbon intensity of the hydrogen source determines the emission benefit: grey hydrogen shifts emissions upstream, whereas green hydrogen can make the hydrogenation route the lowest-carbon option. Quantitatively, the life-cycle global-warming potentials of hydrogen production span roughly an order of magnitude, from about 11.3–12.3 kg CO₂-eq/kg H₂ for fossil and nuclear-thermal routes to about 1.0–2.4 kg CO₂-eq/kg H₂ (measured/LCA) for wind and solar electrolysis [72]. Dimension 3, carbon constraint: tight carbon constraints internalize emissions as $e_{\mathrm{r}\mathrm{o}\mathrm{u}\mathrm{t}\mathrm{e}}C$ and favor heavily heat-integrated configurations (HPRD, RDWC).

Formally, the unit-cost gap between the indirect and direct routes is:

$$\Delta C_{\mathrm{u}\mathrm{n}\mathrm{i}\mathrm{t}}(S,H,C)={\Delta }_0\left(S\right)+m_{{\mathrm{H}}_2}H+(e_{\mathrm{i}\mathrm{n}\mathrm{d}\mathrm{i}\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{t}}-e_{\mathrm{d}\mathrm{i}\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{t}})C$$

(2)

where $m_{{\mathrm{H}}_2}\approx$ 40 kg H₂ per tonne of product is the stoichiometric hydrogen demand of the indirect-hydrogenation route, the carbon term internalizes route-specific emission intensities at carbon price C, and ${\Delta }_0\left(S\right)$ is the baseline unit-cost gap before the hydrogen and carbon terms. For the grey-hydrogen case considered here, $e_{\mathrm{i}\mathrm{n}\mathrm{d}\mathrm{i}\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{t}}-e_{\mathrm{d}\mathrm{i}\mathrm{r}\mathrm{e}\mathrm{c}\mathrm{t}}\approx 0.45$ t CO₂/t product (estimated), which mainly reflects the assumed grey-hydrogen contribution of the indirect route rather than a complete route-level life-cycle emission difference (from $e_{\mathrm{S}\mathrm{M}\mathrm{R}}\approx$ 9–13 kg CO₂/kg H₂ [72]), and ${\Delta }_0\left(S\right)$ is an illustrative estimate (−90 USD/t at 100 kt/a, band −60 to −100 USD/t); Δ₀(S) assumes a 100 kt/a plant and a fixed-capital difference between the two route families, but its CapEx amortization basis has not been independently derived and the value must not be treated as a quantitative result. The quantitative thresholds are site-specific and deliberately not fixed; the framework specifies which data are needed to locate them in a given context.

Combining the three dimensions gives qualitative route preferences under four illustrative scenarios. First, at small scale with a high hydrogen price and weak carbon constraint, direct hydration with cosolvent enhancement appears the most suitable. Second, at large scale with a low hydrogen price from green hydrogen and a strong carbon constraint, indirect esterification–hydrogenation appears most advantageous. Third, at large scale with a high hydrogen price and moderate carbon constraint, transesterification–reactive distillation deserves priority evaluation, with no hydrogen consumption and pilot validation [63]. Fourth, under a strong carbon constraint with surplus steam/electricity, HPRD superimposed on direct hydration is worth considering [9]. The electricity source is the core carbon lever: with renewable power, HPRD CO₂ drops from 0.04 to 0.03 kg/kg cyclohexanol [9]; cross-system evidence suggests that the electricity source largely determines the value of heat-pump integration, with thermodynamic efficiency rising from 10.01% to 16.95% under hydropower and CO₂ savings widening from 23.56% to 25.80% in a dividing-wall column case [47].

To illustrate how the framework is applied with local data, Table 7 defines a benchmark case: 1–3 USD/kg H₂ brackets the IEA (2024) gap between low-emissions and unabated fossil hydrogen projected for 2030, while today’s unabated natural-gas hydrogen costs 0.8–5.7 USD/kg (renewable hydrogen currently 1.5–6× more expensive [73]; IRENA projects green hydrogen ≤5 USD/kg by 2030 and <2 USD/kg by 2050 in most regions [74]; solar/wind alkaline-electrolysis green hydrogen is projected at 2.4–2.9 USD/kg by 2030 [75]); EU ETS spot prices fluctuated within about 51–90 EUR/t (≈55–100 USD/t, market data) over 2023–2026, with annual averages of 83, 65 and 74 EUR/t (market data) [76]. Figure 7 presents the crossover as a preference map; sensitivity-based route assessments of this type have been used in other energy-intensive chemical systems, where hydrogen and carbon prices are controlling exogenous factors [75,77]. A concrete example is the carbon-price-scheduled flexible methanol process of Pan et al. [77], which switches between renewable and fossil modes at a 144.2 USD/t carbon-price threshold and outperforms the purely renewable route by USD 10.65–11.21 M under Monte-Carlo-perturbed carbon prices, showing how route flexibility hedges price risk, a lesson directly transferable to the choice between the hydrogen-consuming and hydrolysis options of the cyclohexanol chain.

Table 7. Benchmark-case parameters underlying the decision framework and Figure 7.

Parameter	Benchmark Value	Basis/Note
Plant scale	100 kt/a cyclohexanol	Estimate
Hydrogen price range	1–3 USD/kg H2	[73]
Current unabated natural-gas hydrogen	0.8–5.7 USD/kg H2	[73]
Current renewable-hydrogen premium	1.5–6×	[73]
Green-hydrogen outlook	≤5 USD/kg by 2030; <2 USD/kg by 2050 in most regions	[74]
Carbon price range	55–100 USD/t CO2	[76]
Hydrogen demand m(H2)	About 40 kg H2/t cyclohexanol	Stoichiometry
Emission gap e(indirect) − e(direct)	About 0.45 t CO2/t product (grey hydrogen)	Grey-H2 contribution assumed; not a full route-level LCA difference [72]
Baseline cost gap Δ0(S)	−90 USD/t at 100 kt/a (band −60 to −100)	Estimate
Hydrogen source basis	Grey (SMR) hydrogen	Assumption behind Figure 7

Figure 7. Illustrative preference map between direct hydration and indirect esterification–hydrogenation as a function of hydrogen price and carbon price (basis: 100 kt/a; grey/SMR hydrogen). The map is qualitative and site-dependent (a heuristic, not a calibrated model); the plant scale is fixed at 100 kt/a and the scale crossover at 10⁴–10⁵ t/a (Section 4.2) is illustrative. The ~0.45 t CO₂/t emission difference (Table 7) principally reflects the assumed grey-hydrogen contribution, not a complete route-level life-cycle emission difference; the favorable regions are therefore illustrative only under the assumed parameters.

4.3. Technology Maturity and Scale-Up Risk

Route preference alone does not guarantee feasibility. In terms of technology maturity: indirect esterification–hydrogenation has reached pilot-scale demonstration (Sinopec, 8000 t/a, >1000 h [6]) and underpins the first 200 kt·a⁻¹ industrial unit of the green caprolactam technology (cyclohexanone oxime section) [1]; transesterification–reactive distillation has completed pilot validation [63]; the catalytic-distillation process of Kumar et al. is at the mini-plant stage [4]; direct-hydration intensification mostly remains at the laboratory and simulation stage [38]. The main scale-up risks are three bottlenecks. First, catalytic-packing engineering: powder catalysts must be shaped into structured packings, and such work accounts for a very small fraction of the literature; the risk is quantified by Ye et al. [39], where accounting for the catalyst effectiveness factor (0.35) raised the RD annual cost from USD 3.90 M to USD 6.60 M (simulated), turning RD from apparently superior to clearly infeasible. Second, long-term stability: durability data for hydrophobic modification layers are almost absent (OTS detaches at elevated temperature [8], and such layers preclude high-temperature regeneration), Cu⁰/Cu⁺ drift and sintering have been identified [54,55,56], Amberlyst-15 lost 18–20% of its initial activity (measured) over a single campaign [4], and public lifetime data remain on the order of hours to hundreds of hours, far from the thousands of hours required industrially; stability concerns extend downstream, where the TS-1 zeolite in ammoximation suffers framework silicon dissolution in the alkaline medium, mitigated industrially by reducing solvent polarity and replenishing amorphous silica [2]. Third, insufficient transparency of simulation assumptions: cyclohexene conversions set in different RD simulations span 40.63–99.9% without stated experimental bases; even design-specified conversions should pass a physical-feasibility check (Chen et al. tied their 99.9% specification to the Amberlyst-15 temperature ceiling that caps the RD pressure at 3 atm in the excess-water design [11]), and Kumar et al. [4] set a positive example by validating their column model against experimental concentration and temperature profiles. Reporting the experimental basis should become a basic norm for process-simulation studies.

5. Conclusions and Outlook

The comparative analysis leads to six conclusions. First, the two routes are constrained differently and cannot be judged by a single indicator: direct hydration is limited by two constraints—thermodynamic equilibrium (a single-pass equilibrium conversion of only about 10–13% under the industrial excess-water slurry configuration, no in situ product removal) and oil–water interfacial mass transfer. Its intensification levers—catalyst synergy, cosolvent design and RD energy integration—can act synergistically or mutually exclusively; indirect hydration bypasses the equilibrium bottleneck by reconstructing the reaction network and can reach conversions and selectivities above 99%, at the cost of process complexity, catalyst-stability demands and hydrogen consumption. Second, hydrophobicity regulation and diffusion-path shortening are the most effective levers in direct hydration, but the reported gains were measured under different conditions and are not directly comparable; the ranking of surface wettability > pore architecture > acid-site density serves as a working hypothesis that awaits unified kinetic verification [18]. Third, the cosolvent route has an economic ceiling: the yield dividend is largely consumed by the separation energy of cosolvent recovery, leaving only limited whole-process TAC and emission savings [27]. Fourth, the benefit of process intensification is largely system-dependent; simulated gains from other systems cannot be transferred directly and must be checked against catalyst effectiveness and operability. Fifth, indirect-route competitiveness depends mainly on system-level factors, namely acetic acid synergy, hydrogen price and intermediate purity, rather than on remaining catalytic gaps; the esterification–hydrogenation route has pilot validation and high atom economy [6], and a higher intermediate purity is not always beneficial [64]. Sixth, the literature suffers systematic comparability deficits, and the conclusions above are bounded by the same deficits: the quantitative thresholds are site-specific estimates rather than calibrated values, and the qualitative rankings should be read as working hypotheses. The key point of the framework is that route choice is condition-dependent: at small scale with expensive hydrogen, direct hydration with cosolvent enhancement appears the most suitable; at large scale with cheap green hydrogen, indirect esterification–hydrogenation appears most advantageous; and under strong carbon constraints with surplus renewable power, HPRD-based direct hydration or transesterification RD are worth considering. These conclusions position the review as a decision-support framework rather than a verdict on which route is intrinsically superior: the answer depends on scale, hydrogen and carbon context and on the data still needed to make the comparison quantitative.

These conclusions lead to the following recommendations. Catalyst research should treat long-term stability as the primary issue. Most stability data come from short laboratory tests under ideal feeds, lasting hours to hundreds of hours, and are not yet convincing enough to support engineering scale-up; the research focus should shift to stability under realistic feed compositions, toward the thousands of hours required industrially, while addressing the two dominant deactivation issues, hydrophobic-layer durability and Cu⁰/Cu⁺ drift, rather than merely reporting short-term performance. Process research should require all relative reductions to state their reference baselines and preferably adopt a common baseline, such as a conventional distillation process; simulation studies should also report the experimental basis of their conversion assumptions. Systems evaluation should complete a TEA and LCA comparison of the four routes under the same capacity and accounting boundary, with the hydrogen-source carbon intensity as an explicit variable. Engineering transfer should prioritize structured catalytic packings, a critical link in realizing simulated RD-type benefits.

Future work can proceed along the following directions (Figure 8). Catalyst design should shift from single-site acidity regulation toward synergistic design of HZSM-5 acidity, pore channels and hydrophobicity, with long-term stability as an explicit design target, accelerated by high-throughput experiments and machine learning. Process intensification and scale-up should deepen the coupling of RD internals, catalyst shaping and cosolvent recovery, with catalyst holdup and two-phase hydraulics as the primary design variables, to advance from laboratory to pilot. At the industrial-chain level, green-hydrogen coupling and bio-based cyclohexene will change the carbon footprint. Hydrogen dominates the unit production cost of CO₂-based fuels, accounting for 70–80%, and green hydrogen from solar/wind alkaline electrolysis is projected to fall to 2.4–2.9 USD/kg by 2030 [75], so the hydrogen-price dimension of the decision framework will shift in favor of the indirect route as green hydrogen becomes competitive. Catalyst innovation advances along the chain, from commercial Pt-based cyclohexane dehydrogenation to Au-Pd-Pt/TS-1 systems bridging in situ H₂O₂ synthesis and ammoximation [3]. Closed-loop recycling of nylon-6 via acid-catalyzed alcoholysis back to caprolactam [71] will reshape feedstock and carbon constraints on the demand side, bio-based phenol from lignin is becoming a renewable source for the cyclohexanone route [5], and life-cycle assessment is the standard for defining process greenness. Dual-functional acid–amphiphilic solid catalysts (e.g., surface-modified HZSM-5/SBA-15) that remove the oil–water phase barrier without additives, reactive-distillation configurations such as RDWC with mechanical vapor recompression (RDWC-MVR) to lower thermal energy demand below 1000 kJ/kg, and cradle-to-gate life-cycle assessments that explicitly model grid-decarbonization trajectories and compare fossil versus green hydrogen inputs are priority directions.

Figure 8. Outlook for cyclohexene hydration to cyclohexanol along the nylon-6 chain (Catalyst design, process intensification and life-cycle assessment directions; see Section 5).