Within integrated propylene oxide derivative plants, 1,2-propanediamine is obtained by continuous reductive amination of 1,2-propanediol over a transition-metal hydrogenation catalyst in the presence of excess ammonia and hydrogen. The primary reaction network proceeds through the intermediate 1-amino-2-propanol, generated by substitution of the primary hydroxyl group; subsequent amination of the secondary hydroxyl yields the vicinal diamine. Byproduct suppression in this system is governed not by a single catalyst parameter but by the interaction of metal dispersion, support acidity, ammonia partial pressure, hydrogen partial pressure, and residence time at elevated temperatures. Industrial fixed-bed reactors processing 1,2-propanediol typically operate within a temperature envelope of 160–200°C and hydrogen pressures of 8–16 MPa. Under these conditions, the most persistent byproducts include 2-methylpiperazine, 2,5-dimethylpiperazine, higher polyamines, and, on catalysts with high hydrogenolysis activity, isopropylamine and propylamine. The selection of a catalyst therefore begins with matching the metal’s hydrogenation function to the desired C–N bond-forming steps while suppressing C–N and C–C hydrogenolysis. Because the reaction is exothermic and produces water as a co-product, alumina-containing supports require stabilization against hydrothermal phase transformation, especially at the reactor inlet where ammonia and water concentrations are greatest. A catalyst that is active at lower temperature can reduce the equilibrium contribution of cyclic amine formation but may exhibit insufficient activity for the second amination step from 1-amino-2-propanol to 1,2-propanediamine, leaving an intermediate that must be recycled. Conversely, a catalyst with high activity at 200°C may convert 1,2-propanediol completely but produce an unacceptable level of cyclic amines. These competing demands create a narrow processing window in which catalyst composition, promoter loading, and support architecture must be co-optimized.

Thermal Degradation Pathways in Ni-Al₂O₃-Catalysed Amination

Supported Ni-Al₂O₃ fixed beds exhibit a sharp acceleration in selectivity loss to 2-methylpiperazine and 2,5-dimethylpiperazine at temperatures above 185°C, because intramolecular condensation of 1-amino-2-propanol and surface-bound imine intermediates is favoured over the desired second amination. The exotherm for the two-step amination is sufficient to create axial hot spots in adiabatic beds when the liquid feed distributor produces flow maldistribution; in a commercial trickle-bed reactor with a bed diameter of 1.2 m and packed height of 7.5 m, axial temperature deviations of ±5°C around the 175°C set point have been associated with a measurable increase in total cyclic amines. Supported nickel catalysts containing 20–30 wt% Ni on alumina are particularly sensitive to hydrothermal attack because water produced in the reaction can hydrate γ-Al₂O₃ to boehmite at extended exposure above 180°C, causing loss of crush strength and an increase in pressure drop. The use of silica-stabilized alumina or a SiO₂ support reduces this degradation pathway but may lower the metal dispersion to 10–15%, requiring higher nickel loading or higher temperature. Promoters such as lanthanum or cerium at 1–3 wt% adjust support acidity and stabilize nickel crystallite size, but if promoter loading exceeds 5 wt%, the surface becomes too basic for efficient imine hydrogenation and the overall rate declines. The catalyst reduction protocol also influences byproduct profiles: incomplete reduction of nickel oxide leaves NiO domains that catalyse alcohol dehydration to propylene oxide intermediates and oligomeric residues, while excessive reduction temperatures above 500°C sinter the nickel crystallites and increase hydrogenolysis to isopropylamine. Therefore, the thermally stable operating envelope for Ni-Al₂O₃ catalyst is frequently limited to 160–180°C for amination, with a preferred hot-spot control window of no more than ±5°C across the bed.

Where ammonia-to-diol molar ratios fall below 8:1 in a fixed-bed operation, the equilibrium concentration of the monoamine intermediate remains high, and the availability of surface-bound primary amine species for intramolecular cyclization increases. The consequence is a nonlinear rise in 2-methylpiperazine concentration that cannot be corrected solely by increasing hydrogen pressure or decreasing space velocity. A high ammonia excess of 12:1 to 20:1 is the most reliable method for suppressing secondary amine formation because ammonia competes with the amine intermediate for surface sites and shifts the amination equilibrium toward 1,2-propanediamine. However, ammonia-rich operation introduces a downstream separation burden: excess ammonia and water must be flashed at 2.0–3.5 MPa and recycled through an ammonia recovery column, and the energy penalty for recompression becomes significant above 15:1. The process conflict is therefore between selectivity and energy cost. In a commercial facility with a 25,000 t/year 1,2-propanediamine capacity, increasing the ammonia-to-diol ratio from 8:1 to 15:1 may lower cyclic byproducts by 40–60%, but the additional ammonia recycle can raise steam consumption by 1.1–1.4 GJ per tonne of product. Catalysts based on cobalt or Ni-Co alloys tolerate lower ammonia ratios than pure nickel because cobalt shows a lower tendency toward C–N hydrogenolysis and a higher intrinsic selectivity for primary amines, allowing operation at 10:1 with similar byproduct levels to nickel at 15:1. The selection of a catalyst with high primary amine selectivity therefore relaxes the severity of the ammonia recycle loop and reduces the energy penalty, provided that the catalyst also maintains sufficient hydrogenation activity for the second amination step. At ratios above 20:1, further selectivity gains are marginal, and the risk of ammonium carbamate precipitation in cooler downstream sections increases if CO₂ is present in the ammonia feed. This threshold defines the upper practical limit for ammonia-to-diol molar ratio in continuous fixed-bed synthesis.

What Limits Liquid Hourly Space Velocity in Fixed-Bed Raney Nickel Operations?

A reduction in liquid hourly space velocity from 1.0 h⁻¹ to 0.4 h⁻¹ does not linearly improve 1,2-propanediol conversion in fixed-bed Raney nickel pilot reactors because the second amination step becomes equilibrium-limited, while byproduct formation continues to progress through irreversible condensation pathways. Raney nickel catalysts, whether used as slurry in batch autoclaves or as shaped fixed-bed formulations, exhibit high initial activity but also high hydrogenolysis activity when the liquid hourly space velocity is low and hydrogen partial pressure is high, leading to isopropylamine formation. The fixed-bed Raney nickel operation at LHSV values below 0.3 h⁻¹ is therefore not recommended for continuous processes despite high conversion, because total cyclic amines and light hydrogenolysis products rise simultaneously. At LHSV values above 1.5 h⁻¹, the conversion of 1-amino-2-propanol to 1,2-propanediamine becomes incomplete, and the recycle stream becomes enriched in the intermediate, which then enters the reactor and forms additional cyclic amines. A further constraint is pressure drop: fixed-bed Raney nickel particles with a mean particle diameter of 2–4 mm and bed depth of 3–6 m develop pressure drops of 0.15–0.35 MPa at liquid hourly space velocities of 0.8–1.2 h⁻¹, depending on ammonia-to-diol ratio and hydrogen flow. If the particle size is reduced below 1 mm to improve mass transfer, the pressure drop increases beyond acceptable limits for standard ammonia recycle compressors, and catalyst attrition generates fines that contaminate the product and increase filtration load. Consequently, the optimal LHSV window for fixed-bed Raney nickel is typically 0.5–1.0 h⁻¹, with an upper threshold determined by intermediate breakthrough and a lower threshold determined by secondary amine formation. Process control systems using online gas chromatography are required to maintain the LHSV within this window because manual sampling intervals of 4–6 h fail to capture transient excursions that alter byproduct profiles.

Simultaneously, chloride-promoted Cu-ZnO-Al₂O₃ catalysts have been evaluated for direct amination of 1,2-propanediol because copper suppresses C–C hydrogenolysis and therefore avoids propylamine and isopropylamine formation. However, the hydrothermal stability of Cu-ZnO-Al₂O₃ is limited when the water partial pressure exceeds 2.0 MPa at reaction temperatures above 190°C, and residual chloride from the preparation can promote corrosion of stainless steel downstream equipment if not washed to less than 50 ppm. Copper catalysts also show lower hydrogenation activity for the second amination step, requiring temperatures 10–20°C higher than nickel or cobalt catalysts to achieve comparable 1,2-propanediol conversion. This higher temperature partially negates the advantage of lower hydrogenolysis. Published data for the specific Cu-ZnO-Al₂O₃ configuration in continuous 1,2-propanediol amination is limited, but fixed-bed pilot studies indicate that total cyclic byproducts are higher than for Co/SiO₂ at equivalent conversion because the higher operating temperature favours intramolecular condensation. A cupric oxide–zinc oxide–alumina catalyst may be suitable only when the feedstock contains impurities that poison nickel or cobalt catalysts, such as sulfur species that would otherwise require an upstream guard bed. Even then, the water sensitivity of the alumina component demands a pre-drying step for the feedstock and strict control of the water content in the recycled ammonia stream to below 0.2 wt%. Without this control, the crush strength of Cu-ZnO-Al₂O₃ extrudates declines over 500–1000 h of operation, and the resulting fines accumulate in the lower sections of the reactor, leading to channeling and localized hot spots.

Ruthenium–Carbon Catalyst Deactivation Mechanisms and Regeneration Constraints

On carbon-supported ruthenium catalysts, lower temperature operation at 130–160°C reduces cyclic amine formation by lowering the equilibrium concentration of imine intermediates, but oxygenated intermediates and steam can gasify the carbon support at localized hot spots above 180°C. Ruthenium is also highly sensitive to sulfur poisoning, and feed sulfur concentrations as low as 0.5–1.0 ppm can cause irreversible loss of activity in the first 200 h of operation. For this reason, a sulfur guard bed containing ZnO or a commercial Cu-Zn adsorbent is installed upstream when ruthenium-based catalysts are used. Regeneration of deactivated Ru/C catalyst requires oxidative burn-off of carbonaceous deposits at 300–350°C in a dilute air stream, but this regeneration window is narrow: below 300°C the removal of oligomeric deposits is incomplete, and above 350°C the carbon support itself undergoes gasification, causing loss of metal and surface area. After oxidative regeneration, the catalyst must be reduced at 250–300°C under hydrogen, with a slow heating ramp of 1–2°C/min to prevent sintering of ruthenium crystallites. The selectivity of Ru/C is also sensitive to the presence of residual alkali promoters; potassium added at 1–2 wt% suppresses support acidity but can leach into the aqueous ammonia feed and contaminate the product. In continuous trickle-bed operation with a bed length of 2–4 m, ruthenium catalysts achieve complete 1,2-propanediol conversion at lower temperatures than nickel, but the byproduct profile shifts toward 2,5-dimethylpiperazine if the feed is rich in 1-amino-2-propanol. Published data for long-term Ru/C operation in 1,2-propanediol amination is limited to pilot-scale campaigns of less than 1000 h, and industrial adoption remains constrained by the cost of ruthenium and the sensitivity of the carbon support to water.

Residual 2-methylpiperazine levels in high-purity polymer intermediate applications are controlled to ≤0.3 wt% by a two-column vacuum distillation train, with water content determined by ASTM E203-16, APHA color by ASTM D1209-05, and density by ASTM D4052-18; total piperazine derivatives and polyamines are monitored by gas chromatography using a flame ionization detector calibrated under ISO/IEC 17025:2017, clause 7.6.1.

To reduce strong acid site density on alumina and silica-alumina supports, magnesium, potassium, and cerium promoters at loadings of 1–5 wt% are employed, thereby suppressing dehydration of 1-amino-2-propanol to propylene oxide intermediates and subsequent oligomerization. Temperature-programmed desorption of ammonia on a Ni-Co/SiO₂ catalyst has shown that residual strong acid sites above 1.0 mmol/g correlate with an increase in total cyclic amines; neutralizing those sites with potassium carbonate can lower 2-methylpiperazine concentration by 25–35% in continuous pilot runs. However, excessive promoter loading above 5 wt% leads to pore mouth blocking and a loss of accessible metal surface area, reducing the observed first-order rate constant by 15–20%. The promoter must also be resistant to leaching in the aqueous ammonia feed; potassium and sodium are prone to leaching at levels of 10–50 ppm over 1000 h, while cerium and lanthanum remain structurally bound. This distinction is critical in fixed-bed catalyst replacement cycles because leached alkali metals can accelerate corrosion in stainless steel distillation columns and produce metal carboxylates that foul reboilers. A supported Ni-Co catalyst with ceria promotion at 2 wt% on silica-stabilized alumina demonstrates a stable selectivity profile over 1200 h in a 1.5 m trickle-bed pilot unit, whereas a corresponding potassium-promoted analog loses 30% of its ceria-normalized selectivity advantage after 800 h due to potassium migration to the reactor walls. The mechanical crush strength of the catalyst extrudates, measured according to ASTM D4179 or an equivalent radial crush method, must remain above 1.5 MPa after hydrothermal exposure to prevent the formation of fines that cause channeling and localized hot spots.

When Catalyst Regeneration Skips Low-Steam Hydrogen Reduction

If a fixed-bed Ni-Co catalyst is regenerated by oxidative burn-off followed by direct hydrogen reduction without a low-steam purge, the initially high exotherm from nickel oxide reduction can create internal temperatures above 500°C even when the bulk bed temperature is controlled at 350°C. This localized overheating sinters the Ni-Co alloy and increases the proportion of cobalt-rich phases, which subsequently shifts selectivity toward unsaturated intermediates and raises 2-methylpiperazine formation. The correct regeneration protocol begins with a nitrogen purge to remove ammonia and hydrogen, followed by controlled oxidation with 0.5–1.0 vol% oxygen in nitrogen at 300–350°C until the carbon monoxide and carbon dioxide concentration in the off-gas stabilizes. A low-steam purge of 1–3 vol% steam in nitrogen is then applied at 250–300°C to passivate surface carbides and remove residual chlorine or sulfur compounds before hydrogen reduction. Reduction is performed with a hydrogen-in-nitrogen stream starting at 0.5 vol% hydrogen and increasing stepwise to 100 vol% over 8–12 h, maintaining a bed outlet temperature no more than 15°C above the inlet temperature. Skipping the low-steam purge is particularly detrimental for catalysts containing alkali or alkaline-earth promoters because the residual surface species react with water during reduction, forming hydroxides that block active sites. After regeneration, the catalyst typically recovers 90–95% of its initial activity, but the selectivity to 1,2-propanediamine may remain 1–2 percentage points lower than fresh catalyst, a loss attributed to irreversible restructuring of the Ni-Co bimetallic surface. In continuous commercial operation, regeneration cycles are therefore planned at intervals of 800–1500 h, with catalyst unloading avoided until the activity loss exceeds 20% or the pressure drop exceeds 0.35 MPa at design LHSV.