Methyl Ester Yield from Triglyceride Transesterification with Catalyst Predissolution in Anhydrous Methanol

Production of fatty acid methyl ester by alkali-catalyzed transesterification of refined triglycerides is commonly controlled by measuring ester content according to EN 14103:2011 and total glycerol by ASTM D6584-20a after phase separation, methanol recovery, and washing. In catalyst predissolution practice, sodium methoxide is delivered as a 25 wt% to 30 wt% solution in anhydrous methanol and diluted further with fresh methanol containing ≤0.10 wt% water before charging into a closed reactor. The stoichiometric minimum is 3 mol methanol per 1 mol triglyceride, but industrial operation at a 6:1 molar ratio shifts the equilibrium toward methyl ester and compensates for methanol lost to vapor during atmospheric reflux. A refined rapeseed oil charge with acid value ≤0.50 mg KOH/g, phosphorus ≤10 ppm, and moisture ≤0.05 wt% typically yields 97.5–99.0% methyl ester content after 60 min at 60–65°C when sodium methoxide is fully predissolved and the reactor is adequately baffled. The central advantage of predissolution is not increased thermodynamic equilibrium but a shorter induction period and reduced localized saponification: solid sodium methoxide particles, if discharged directly into a warm oil/methanol dispersion, can create local alkoxide concentrations that hydrolyze to sodium hydroxide and consume free fatty acids or ester bonds before catalytic transesterification reaches the bulk phase. In continuous plants, the same predissolution principle is implemented by injecting a controlled 2–5 wt% sodium methoxide stream into dry methanol ahead of the first reactor nozzle, with static mixing and interlocked flow ratio controllers to prevent high-alkalinity slugs. This configuration is used for rapeseed, soybean, and palm olein feedstocks where free fatty acid content is below 0.5 wt%, and it reduces batch-to-batch yield variance compared with dry solid catalyst addition because the catalyst concentration at the interface is consistent within the first 2–5 min of reaction.

What Limits Methyl Ester Yield When Water Enters the Catalyst Predissolution Stage?

When water ingress exceeds 0.10 wt% in methanol, catalyst speciation changes because sodium methoxide reacts with water to form sodium hydroxide and methanol. The hydroxide ion is a competent transesterification catalyst but is also the active species for saponification of triglycerides and methyl ester. Each mole of sodium hydroxide that neutralizes free fatty acid produces water, generating an autocatalytic hydrolysis path that increases mono-, di-, and triglyceride-bound glycerol. In a 10 m³ batch reactor processing refined soybean oil at 6:1 methanol-to-oil molar ratio and 0.5 wt% NaOCH₃ on oil, increasing methanol water content from 0.05 wt% to 0.30 wt% is associated with a yield reduction of 2–5 percentage points and a rise in total glycerol from 0.18 wt% to 0.35–0.50 wt% when downstream washing and glycerol separation are identical. The mechanism is observable as a tighter emulsion at the methyl ester/glycerol interface: sodium soaps formed from free fatty acid or ester saponification reduce interfacial tension and stabilize glycerol droplets below 20 µm, slowing gravity settling and requiring high-speed disc-stack centrifuges at 2500–4500 rpm to achieve acceptable separation. Water also reduces the selectivity of the reaction because hydrolysis followed by saponification consumes two moles of alkali per mole of fatty acid released, leaving less methoxide for the desired transesterification. For this reason, methanol storage tanks are fitted with nitrogen blankets, desiccant breathers, and molecular sieve drying units to hold water below 0.10 wt%, and incoming oil is dried under vacuum at 105–110°C and 80–100 mbar until moisture is ≤0.05 wt%.

In batch reactors larger than 5 m³, methyl ester yield is often mass-transfer limited during the first 10–20 min because methanol and triglyceride form a two-phase liquid system. The reaction can proceed in the polar methanol phase where methoxide anions attack glyceride esters at the interface, but the initial interfacial area is controlled by impeller shear and power dissipation. A single six-blade Rushton turbine with diameter 0.33 T operating at 120–150 rpm in a baffled reactor may dissipate 0.7–1.2 W/kg and generate methanol droplets with Sauter mean diameter 80–150 µm. As conversion increases, methyl ester acts as a co-solvent, the mixture becomes more homogeneous, and the kinetic regime shifts from mass-transfer-limited to reaction-rate-limited. Predissolved sodium methoxide removes the solid dissolution boundary layer but does not eliminate the need for droplet dispersion; therefore, reactors are often equipped with a second 45° pitched-blade turbine above the Rushton impeller to improve axial circulation. Batch records from 20 m³ stirred vessels show that a drop in tip speed below 3.0 m/s during catalyst injection can produce a yield loss of 1.5–3.0% at otherwise identical temperature and molar ratio. Static mixers at the methanol/catalyst feed line, with 6–12 helical elements and liquid velocity 1.5–2.5 m/s, are used to dilute the 25 wt% sodium methoxide solution with dry methanol immediately upstream of the reactor nozzle to prevent unmixed high-alkalinity slugs from entering the oil phase.

Anhydrous Methanol Storage, Molecular Sieve Drying, and Catalyst Dissolution Exotherm Control

Sodium methoxide predissolution releases heat when solid sodium methoxide dissolves or when concentrated 25–30 wt% methanolic solution is diluted with dry methanol. The heat of dissolution is neutralized in jacketed plug-flow dilution elbows or in a small recirculation loop with a plate heat exchanger sized for 5–10°C temperature rise at 0.5–1.0 wt% NaOCH₃ on oil dosage. The diluted catalyst stream is charged at 30–40°C to avoid flashing methanol in the reactor, which is maintained at 60–65°C under total reflux. Anhydrous methanol is dried with type 3A molecular sieve beads in a parallel-regeneration dryer skid; the dried methanol specification is ≤0.10 wt% water, ≤10 ppm sodium, ≤5 ppm chloride, and ≤0.1 vol% non-methanol organics. The catalyst dissolution tank is equipped with a relief vent sized to API 2000 for methanol vapor and a rupture disk set at 3.5 barg, with a nitrogen sweep to maintain oxygen below 8 vol% to stay outside methanol flammability limits. Solid sodium methoxide addition to an open vessel is avoided because the material is hygroscopic and reacts with atmospheric moisture; sealed dissolvers with load cells and lock hoppers are standard in continuous plants.

Refined soybean oil with an acid value below 0.50 mg KOH/g and phosphorus below 10 ppm is suitable for direct sodium methoxide-catalyzed transesterification, but higher free fatty acid levels alter catalyst consumption and yield. Each 1.0 wt% free fatty acid in oil corresponds to approximately 2.0 mg KOH/g acid value and consumes sodium hydroxide or methoxide to form soap and water. A 1.0 mg KOH/g increase in acid value can consume roughly 0.1 wt% sodium methoxide relative to oil and reduce measured methyl ester yield by 0.5–1.5% unless additional catalyst is dosed. Feedstock with acid value above 2.0 mg KOH/g is typically pretreated with acid esterification or steam stripping before alkali transesterification; otherwise, the soap concentration can exceed 1.0 wt% and cause centrifuge bowl plugging. In a 15 m³ batch with decanter-type separator, soybean oil acid value rising from 0.2 mg KOH/g to 1.5 mg KOH/g produced an emulsion layer that occupied 12–18 vol% of the settling vessel and lowered separated methyl ester yield from 98.0% to 94.5% under identical process settings. Phospholipids and trace metals from poor degumming further deactivate catalyst and stabilize emulsions; phosphorus should be below 10 ppm and calcium/magnesium combined below 5 ppm for sodium methoxide systems. Published data for the exact interaction of phosphorus with predissolved sodium methoxide in high-FFA feedstocks is limited, but industrial practice applies dilute phosphoric acid pretreatment followed by bleaching clay to restore feedstock quality.

When Potassium Hydroxide Predissolution Is Compared With Sodium Methylate in Anhydrous Systems

With 0.5 wt% sodium methoxide on oil, the active catalyst is present as sodium and methoxide ions, while potassium hydroxide predissolved in methanol forms potassium methoxide and water. The liberated water can increase soap formation, so KOH-catalyzed systems frequently use a slightly higher methanol-to-oil ratio, often 6.5:1 to 7:1, to compensate for water dilution and maintain phase separation. For the same molar alkali loading, potassium methoxide is kinetically faster than sodium methoxide in the initial triglyceride-to-diglyceride step, but it also generates soaps more readily and produces a denser glycerol phase that can retain methanol and catalyst. Comparative batch data from 1 m³ pilot reactors using palm olein show methyl ester yields of 96.5–98.0% for potassium hydroxide predissolved in anhydrous methanol versus 97.5–99.0% for sodium methylate under 60°C, 60 min, and 6:1 molar ratio. The sodium methylate system is preferred when the feedstock is refined and dry, while potassium hydroxide can be selected when the acid value is between 0.5–1.5 mg KOH/g because the potassium soaps are softer and less likely to form rigid films on heat exchangers. Above 1.5 mg KOH/g, neither alkali catalyst should be used without acid pretreatment because yield loss and separation failures become statistically frequent.

Kinetic Parameters and Reversible Transesterification Equilibrium Under Sodium Methoxide Catalysis

The transesterification network proceeds through three consecutive reversible steps: triglyceride to diglyceride, diglyceride to monoglyceride, and monoglyceride to glycerol, releasing one methyl ester per step. Pseudo-first-order treatment of refined rapeseed oil in excess methanol gives apparent rate constants for the three steps that differ by roughly an order of magnitude, with the monoglyceride-to-glycerol step often rate-limiting under high viscosity and poor methanol dispersion. At 60°C and 6:1 methanol-to-oil molar ratio, the forward reaction approaches equilibrium within 15–30 min when predissolved sodium methoxide is fully distributed; however, industrial reaction time is extended to 60–120 min to reduce the bound glycerol sum below 0.25 wt%. The equilibrium constant for methanolysis is not a single parameter because the three stepwise reactions have different constants and because the mixture becomes non-ideal as methyl ester and glycerol form separate phases. Excess methanol above the stoichiometric 3:1 ratio drives conversion, but ratios above 9:1 add recovery duty without measurable yield gain and may increase the solubility of glycerol in ester, raising total glycerol after washing. Temperature is limited by methanol boiling point at atmospheric pressure; moving to 100–120°C requires 5–10 barg pressure reactors and reduces cycle time but increases catalyst degradation and soap-forming side reactions. Published kinetic data for this specific configuration are often aggregated in terms of an overall apparent activation energy near 33–50 kJ/mol, but the effective activation energy under mass-transfer-limited conditions can be lower because diffusion coefficients improve with temperature.

Because bound glycerol and residual glycerides co-elute in some simple chromatographic methods, analytical error in methyl ester yield is frequently dominated by unreacted monoglycerides and diglycerides that remain in the ester layer after water washing. The ester content reported by EN 14103:2011 is based on gas chromatographic separation with methyl heptadecanoate internal standard on a polar poly(ethylene glycol) capillary column 30 m long, 0.25 mm internal diameter, and 0.25 µm film thickness. Bound glycerol is measured by ASTM D6584-20a after silylation with N-methyl-N-trimethylsilyltrifluoroacetamide and detection by flame ionization; the sum of monoglyceride, diglyceride, and triglyceride is converted to total glycerol. If reactor samples are taken before complete settling, dispersed glycerol droplets in the ester phase can artificially raise total glycerol readings by 0.05–0.15 wt%. Therefore, sampling valves are located downstream of a decanter or centrifuge, and samples are centrifuged at 3000–4000 rpm for 10 min before injection. Calibration curves are verified at 0.01 wt%, 0.10 wt%, and 1.00 wt% levels for each glyceride species to maintain linearity across the range needed for an accurate mass balance.

On a 25 m³ production line equipped with two disc-stack centrifuges in series, the most frequent yield-limiting failure is not low gross conversion but poor glycerol separation caused by soap accumulation. Sodium soap levels above 0.3 wt% in the ester phase correlate with a transfer of mono- and diglycerides into the lighter phase and an increase in total glycerol above the 0.25 wt% limit even when the reactor conversion is 97%. The first centrifuge is typically set to 2500–3000 rpm with a light-phase backpressure of 1.5–2.0 barg; the second centrifuge polishes with a smaller bowl and 3500 rpm. When feedstock moisture rises to 0.15 wt%, the soap content can exceed 0.5 wt%, causing bowl sludge deposits that require cleaning every 4–6 h instead of every 24 h. Operators use a quick soap titration method calibrated against FTIR soap absorbance to decide when acidulation washing with 0.1–0.2 wt% citric acid solution is required before the second centrifuge. These field observations explain why methyl ester yield is not solely a reactor metric; separation losses in the glycerol/ester interface can account for 1–3% yield loss and must be included in the plant mass balance.

Because Catalyst Predissolution Shortens Induction Time, Feedstock Metering and Static Mixer Design Determine Attainable Yield

In continuous plants where methanol and catalyst solution are combined in a static mixer upstream of the first reactor, the predissolved sodium methoxide stream must be diluted to 2–5 wt% NaOCH₃ before contacting oil to prevent saponification at the nozzle. A typical static mixer is a DN50 pipe with 12 helical elements, operating at 1.0–1.5 m/s liquid velocity and Reynolds number above 5000; under these conditions the coefficient of variation of alkali concentration at the reactor inlet is below 5% within 0.5 s. The first continuous stirred tank reactor has a residence time of 30–45 min, and the second reactor in series provides an additional 30–45 min to reduce mono- and diglycerides to specification. If the feed pump loses suction or the static mixer bypass opens, localized high-alkalinity regions form, producing soaps and a yield drop of 3–6% that cannot be recovered by downstream catalysis because the active catalyst has been consumed. Therefore, flow ratio controllers, static mixer differential pressure transmitters, and conductivity probes in the methoxide dilution line are interlocked to stop oil feed when the catalyst-to-methanol ratio exceeds 10% of setpoint.

The final methyl ester yield after washing and vacuum drying is determined by the entire sequence of reactor conversion, phase separation, water washing, and distillative methanol removal. Water washing with 5–10 wt% deionized water at 60–65°C is used to remove glycerol, soaps, and residual catalyst; the wet ester is then dried at 90–100°C and 200 mbar to a water content below 500 mg/kg as specified in EN 14214:2012. A dry final ester with 98.5% methyl ester content and total glycerol 0.20 wt% represents stable operation; excursions are usually traced to water carryover in methanol or incomplete washing rather than insufficient catalyst activity. Published data for the specific combination of sodium methoxide predissolution and continuous countercurrent washing at production scale is limited, but the process is routinely validated by sampling the final methyl ester for EN 14103:2011, EN 14105:2011, and EN ISO 12937:2000 across 10 consecutive batches.