Catalyst Neutralisation Endpoint Effects on Trimethylolpropane Methyl Ester Filterability

The manufacturing route to trimethylolpropane methyl ester via transesterification of trimethylolpropane with fatty acid methyl esters derived from coconut, palm kernel, or synthetic C8–C10 carboxylic acids is a batch process in which the catalyst neutralisation endpoint directly governs downstream filterability. A stoichiometric excess of methyl ester, typically a molar ratio of 1:3.0 to 1:3.3, is used to shift the equilibrium toward the triester, while methanol is removed continuously under reduced pressure. The catalyst is normally sodium methoxide, introduced as a 30 % solution in methanol. After the reactor reaches the target conversion, the catalyst is neutralised with an aqueous acid, and this single unit operation fixes the salt loading, residual soap concentration, free water content, and acid number of the crude trimethylolpropane methyl ester. The neutralisation endpoint therefore controls not only the chemical quality of the ester but also its subsequent filterability. If the endpoint leaves residual alkalinity, sodium carboxylate soaps persist and stabilise water-in-oil microemulsions that pass through 10–20 µm filter cloth and later blind the medium. If the endpoint is driven too far, free acid hydrolyses ester bonds during hot drying and generates free fatty acids and partial esters that increase haze and downstream acid number. The analytical definitions of the endpoint are established by potentiometric acid number titration in accordance with ASTM D664-18e2, water content by Karl Fischer coulometry in accordance with ASTM D6304-20, residual sodium and silicon by inductively coupled plasma optical emission spectrometry in accordance with ASTM D5185-18, and filterability by the procedure described in ISO 13357-1:2002. Production-scale filtration hardware for this service includes recessed-plate filter presses, pressure leaf filters, and agitated Nutsche filters built from AISI 316L stainless steel and fitted with monofilament polyester, polypropylene, or PEEK filter cloths. The dominant field failure is not coarse plugging but progressive cloth blinding by compressible soap layers and submicron salt nuclei, which requires early batch rework rather than simple media replacement.

Neutralisation endpoint control for sodium methoxide-catalysed trimethylolpropane methyl ester synthesis

In sodium methoxide-catalysed transesterification, the catalyst is typically neutralised at 65–85 °C with 85 % phosphoric acid or 50 % citric acid solution. Acid addition is carried out over 30–60 min under moderate agitation to avoid local low-pH zones that hydrolyse the ester. The endpoint is controlled by the acid number of the crude organic phase and by the conductivity of a 10 % aqueous extract. A crude ester acid number of 0.08–0.15 mg KOH/g before water washing is commonly used; the corresponding water-extract conductivity after salt removal is product-specific but must be low enough to confirm that dissolved sodium salts have been removed. Potentiometric titration is preferred over colorimetric titration because dark crude esters obscure the colour-change endpoint of ASTM D974-22e1. The neutralisation acid determines the salt identity. Phosphoric acid produces sodium phosphate salts, which can form gelatinous precipitates and compressible filter cakes. Citric acid produces sodium citrate salts, which may crystallise as needles if the neutralisation pH is below 5.5 and the cooling rate is rapid. These two salts do not respond identically to filter aid addition. A filter leaf test conducted at constant pressure of 0.2 MPa through a 0.8 µm glass fibre filter is used to rank filterability; the time required to collect a defined filtrate volume is compared against a reference batch with known salt content and particle size distribution.

The catalyst charge and neutralisation stoichiometry must be matched to the methyl ester feed quality. Feedstocks with high free fatty acid content consume catalyst and form additional sodium soaps; the neutralisation endpoint then cannot be defined solely by the original catalyst charge. In such systems, the endpoint acid number should be confirmed after a 60–90 min hold at 75–85 °C, and the batch should be sampled from the top and bottom of the reactor because salt and soap concentrations can stratify in low-agitation vessels. Incomplete neutralisation transfers residual sodium methoxide to the filter press, where it continues to react with residual water and ester during filtration. The resulting soap gel concentrates on the filter cloth and produces a steep increase in pressure drop at a relatively low total solids loading. Correcting the batch at this stage by adding acid to the filter feed is ineffective because the soap is already distributed in the cake and the acid cannot contact it uniformly. The practical control strategy is to terminate the reactor neutralisation step using a combination of acid number, water extract pH, and turbidity after an ageing hold, before the ester is transferred to filtration.

Filterability of the neutralised ester follows cake filtration theory. The liquid flux J through a filter medium and cake under constant pressure is described by J = ΔP / [μ (α c V/A + R_m)], where J is in m³/m²/s, ΔP is the Darcy pressure drop in Pa, μ is the dynamic viscosity in Pa·s, α is the average specific cake resistance in m/kg, c is the mass of dry solids deposited per unit filtrate volume in kg/m³, V/A is the cumulative filtrate volume per unit area in m, and R_m is the clean medium resistance in m⁻¹. In neutralised trimethylolpropane methyl ester, the solids concentration c is set by the catalyst loading and neutralisation acid, while α is determined by salt crystal habit, residual soap content, and cake compressibility. For compressible cakes, α increases with applied pressure, so operation at higher ΔP does not provide proportional throughput. This is the central reason that filterability cannot be inferred from a simple mass balance of neutralisation salts; the morphology and surface chemistry of the precipitated solids control the pressure response. A filtration test performed at constant pressure with a 0.8 µm membrane and reported as the ratio of filtered volume to total test volume under ISO 13357-1:2002 provides a practical comparison, but it does not replace a filter leaf test with the production filter cloth and filter aid. When residual soap is present, the cake may be elastic rather than brittle, and the filtration curve shows a prolonged early stage in which filtrate clarity is poor. The eventual pressure rise is then caused not by an accumulating granular cake but by a surface gel layer that is difficult to remove by backwashing.

What Limits Filtration Through Recessed-Plate Presses When Residual Sodium Methoxide Is Inadequately Neutralised?

When neutralisation is incomplete, the limiting flux through production-scale recessed-plate presses is controlled by soap accumulation and microemulsion formation rather than by the total mass of salt. Residual sodium methoxide reacts with trace moisture and ester groups to form sodium carboxylate soaps, which are surface-active and stabilise a fine dispersion of water in the ester. The first observed deviation is usually a reduction in filtrate clarity, not an increase in pressure drop, because water droplets and soap micelles below 1 µm pass through 10–20 µm monofilament cloth. As the run continues, the soap accumulates on the cloth and in the filter cake, reducing open area and producing a nonlinear pressure trend. A recessed-plate press with 12 mm chambers and a maximum differential pressure rating of 0.6 MPa can reach its terminal pressure within 30–45 min after a clear first filtrate if the soap load is high. This is often misinterpreted as a filter cloth defect, but replacing the cloth does not correct the flux decline. The appropriate correction is to return the batch to the neutralisation vessel, reheat to 70–80 °C, add the remaining acid incrementally, and hold for 60–90 min before refiltration through a diatomaceous earth precoat. Published data for this specific configuration is limited; however, the operational pattern is well recognised in batch ester processing, and the root cause is confirmed by measuring residual alkalinity and turbidity after the initial neutralisation.

The main process limit is the compressible cake resistance, not the installed filtration area. When the under-neutralised batch is filtered, the soap gel compacts under pressure. Increasing ΔP from 0.2 MPa to 0.6 MPa does not increase flow proportionally because the cake consolidates and its specific resistance rises. In plate-and-frame equipment this produces a characteristic response in which the initial flow appears acceptable, the pressure then climbs steeply, and the filtrate volume declines. Batch-to-batch variability in filterability is often traced to endpoint measurement methods that are too coarse. Glass-bulb pH measurement cannot reliably distinguish residual alkalinity equivalents in a non-aqueous buffered ester system. A potentiometric acid number titration using a mixed solvent with ASTM D664-18e2 is more reproducible. The endpoint should be checked after a short ageing period because neutralisation reactions in viscous ester media are not instantaneous. If the batch is transferred too early to the filter press, the neutralisation reaction continues in the filter cake and the resulting soap forms a skin on the cloth that cannot be removed by normal washing.

Operational boundaries are defined by water content, acid type, and materials of construction. Processing at relative humidity above 60 % without nitrogen blanketing introduces atmospheric moisture into the crude ester; this moisture dissolves a portion of the neutralisation salt and reduces the effective particle size of the solids. The neutralised ester should be dried to a water content below 200 mg/kg before final polishing filtration. Vacuum drying at 90–110 °C is typical for trimethylolpropane methyl ester streams, but the temperature must be limited if the product contains unsaturated fatty acid methyl esters to avoid thermal degradation. Addition of hydrochloric acid as a neutralising acid is generally avoided unless the entire downstream line is constructed from AISI 316L or higher molybdenum alloys because chloride ions can initiate pitting in stainless steel filter housings and plate packs. Sulfuric acid neutralisation produces sulfate salts that can require additional water washing and can leave a haze if the final ester is not adequately dried. The neutralisation acid must also be compatible with the intended application; for food-grade lubricant base stocks, citric acid and phosphoric acid are commonly used, and the final product should be checked against the residual metal limits in FDA 21 CFR 178.3570 where applicable. These boundaries are not additive afterthoughts; they constrain the neutralisation endpoint because a process designed for rapid filtration at low acid number may fail if the selected acid produces a hygroscopic salt that cannot be fully removed at the washing station.

The following analytical verification matrix is applied to neutralised trimethylolpropane methyl ester after catalyst neutralisation and before filtration release.

Post-neutralisation salt crystal habit, ageing time, and filter aid selection

Following catalyst neutralisation, the filterability of the trimethylolpropane methyl ester is governed by the crystal habit and particle size distribution of the precipitated salt rather than by its total concentration alone. Sodium citrate salts produced by citric acid neutralisation often crystallise as needle-like particles with high aspect ratios if the endpoint is acidic and cooling is rapid. Needle-like crystals interlock to form a compressible cake, can bridge pores in 10–20 µm filter cloth, and are difficult to wash because the cake develops channels. Sodium phosphate salts produced by phosphoric acid neutralisation may be plate-like or gelatinous depending on temperature, endpoint acid number, and water content. The precipitation process is improved by controlled ageing at 55–65 °C for 30–90 min after acid addition. During ageing, small crystals and amorphous nuclei dissolve and redeposit onto larger particles, reducing the submicron population that blinds filter media. The ageing step is not a trivial holding time; it must be carried out under sufficient agitation to maintain suspension but not so intense that the fragile crystals are sheared. The optimum ageing time is batch-specific and is determined by particle size analysis in accordance with ISO 13320:2020 and by filter leaf ranking.

Filter aid addition follows the particle habit. Diatomaceous earth grades with permeabilities of 0.5–2.0 darcy are used as precoat and body feed. For gelatinous phosphate precipitates, a more permeable filter aid is required because the cake permeability is controlled mainly by the filter aid skeleton. Precoat loadings of 1–2 kg/m² and body feed doses of 0.1–0.5 wt% are typical for pressure leaf and Nutsche filtration. When the body feed is matched to the salt concentration, filtration proceeds at a lower pressure drop and the terminal pressure is reached at a higher total solids capacity. If body feed is too low, the salt forms a thin impermeable layer on the precoat and the flux collapses early. If body feed is too high, the filter aid itself contributes unnecessary solids and can increase silicon release into the filtrate. Perlite or cellulose filter aids are selected when silicon release must be minimised. The silicon concentration in the final ester is measured by ASTM D5185-18 for metalworking and electronic-grade applications where silicon is controlled. Overmixing after filter aid addition is avoided because it can shear filter aid particles and reduce permeability. The body feed should be added continuously to the filter feed tank, not dumped as a single charge, to maintain a consistent cake structure.

When phosphoric acid is used to neutralise a sodium methoxide-catalysed transesterification to an acid number below 0.10 mg KOH/g

Phosphoric acid neutralisation is used when the final trimethylolpropane methyl ester must maintain low colour and stability under subsequent vacuum stripping. The reaction between sodium methoxide and phosphoric acid produces monosodium and disodium phosphate species, depending on endpoint acid number and water content. At an endpoint acid number below 0.10 mg KOH/g, the aqueous phase after washing is slightly acidic, and the phosphate salts are predominantly in the NaH₂PO₄ form. This salt holds water of hydration, and if the ester is not dehydrated after washing, hydrated phosphate crystals can pass through a 1–5 µm polishing filter and later release water during storage. The resulting turbidity and acid number drift are then misattributed to oxidation, when the root cause is incomplete removal of a hygroscopic neutralisation salt. The recommended procedure is to add phosphoric acid as a dilute aqueous solution over 30–60 min, hold at 75–85 °C for 45–60 min, water-wash at 70–80 °C until the wash water conductivity is below 10 µS/cm, and dry under vacuum at 90–110 °C. The dry ester is then filtered through an absolute-rated 1–5 µm filter element before storage.

A specific limitation of phosphate-neutralised systems is the compressible nature of the filter cake. The phosphate precipitate forms a gel-like layer that consolidates under high differential pressure, so increasing pressure to compensate for low flux is counterproductive. In a filter leaf test, this appears as a deviation from linearity in the t/V versus V plot, indicating that the specific cake resistance is not constant. Published data for this specific configuration is limited; however, the behaviour can be correlated empirically with total phosphate content, acid number, and residual water. The ester should not be held at temperatures above 60 °C for prolonged periods in the presence of residual phosphate salts because phosphate-catalysed hydrolysis can increase acid number and release free fatty acids. If the downstream formulation includes amine-based antioxidants or rust inhibitors, the residual acid number should be below 0.05 mg KOH/g to avoid acid-amine adduct formation and subsequent haze. The combined endpoint for phosphoric acid systems is therefore defined by acid number, water content, and water-extract conductivity, not by a single pH reading. This combined endpoint is the process variable that determines whether the subsequent filtration run is routine or becomes a rework event.

In plants processing multiple methyl ester feedstocks, batch-to-batch variability in filterability is reduced when the neutralisation temperature, acid addition rate, and ageing interval are maintained within narrow ranges. The rate of acid addition influences local salt supersaturation and crystal nucleation, and a fast addition can generate submicron nuclei that are slow to filter. A slow addition under controlled agitation produces larger, more uniform crystals that release filtrate more readily. The neutralisation endpoint therefore functions as a process control variable that sets the population of filterable solids rather than as a single analytical number. Transfer from laboratory filter leaf data to production-scale filter presses requires pilot confirmation with the same filter cloth, filter aid, and pressure profile, because filterability is medium-dependent and cake compressibility is scale-sensitive. The final filtered ester is packaged under nitrogen blanket and monitored for acid number, water content, and turbidity before release.