Soaps, detergents and agrochemicals: Making soap and recovering glycerine

2.1 Saponification: a fat plus caustic soda gives soap and glycerine

Saponification is the reaction of a triglyceride with an alkali: triglyceride + 3NaOH → 3RCOONa + glycerine. RCOONa is soap, the sodium salt of a fatty acid. Its long hydrocarbon tail dissolves in grease and its charged head dissolves in water, which is how it lifts dirt.

Two practical problems shape every soap plant. The first is that fat and caustic solution do not mix, like oil and vinegar. Kent explains that high-shear mixing, high temperature and enough pressure to stop boiling raise the rate. The reaction also speeds itself up: soap is an emulsifier, so the more soap already present, the finer the droplets and the faster the rest reacts. Continuous reactors exploit this by recirculating 13-25 times as much reacted mass as fresh feed.

The second problem is separating the soap from glycerine and excess alkali. Soapmakers do this with salt: the right brine strength splits the mass into a soap layer and a heavier salt solution, the lye, which carries glycerine and impurities. Kent gives the densities as about 0.97 g/cc for neat soap and 1.05 for spent lye. The gap is small, so separators need time and a centrifuge often follows.

The caustic dose follows from the saponification value. Kent gives 14.09 g of NaOH per 100 g of beef tallow and 18.32 g for coconut oil, whose shorter chains carry more ester links per kilogram. With 50% caustic, 100 g of tallow needs 28.2 g of solution.

Neat soap leaves the process at about 30% water. Vacuum spray dryers take it to 8-18% moisture as pellets of 70-82% total fatty matter. Finishing lines add perfume and colour, mill, extrude under vacuum, cut and stamp the bars.

• Kent (2012), Ch 36 Soap, Fatty Acids, and Synthetic Detergents · pdf pp. 1431-1440
• Shreve (1984), Ch 29 Soap and Detergents · pdf pp. 563-566
• Dryden (1997), IIIB Soaps and Detergents; Glycerin · pdf p. 311

2.2 Kettle or continuous: a trade between fat quality, labour and glycerine

The kettle is the traditional batch route. Fat is saponified with caustic in a large boiling vessel, and the soap is kept in contact with lye, which carries away glycerine and water-soluble impurities (Kent). A final "fitting" step separates a small quantity of lower-quality soap. Shreve says the whole sequence takes more than a week, and that by the 1980s the kettle survived mainly in smaller factories.

Continuous plants won on almost every count. Shreve reports that a computer-controlled continuous saponifier makes in 2 hours what batch methods make in 2 to 5 days, at more than 300 tonnes a day. Kent lists the reasons large plants switched: savings in utilities, simpler operator training, no operator bias in control, and a more consistent soap base. Dryden adds higher glycerine recovery (over 80%), less discoloured product, and less space and manpower.

There is a price. In the kettle, fitting drops 1-2% of poorer "slurry soap", which takes colour and odour bodies with it. A continuous plant has no such step, so colour components equal to roughly 5-10% of a kettle charge stay in the product, and Kent warns that fats for continuous saponification must start with better natural colour and odour. In effect the continuous route swaps labour and steam for better feedstock and instruments.

Scale varies widely. Dryden gives Indian soap plants of only 2-15 tonnes a day and notes that much Indian soap was made in villages by crude methods, with no glycerine recovered at all. At that scale a recovery plant cannot pay for itself.

• Shreve (1984), Ch 29 Soap and Detergents · pdf pp. 563-566
• Kent (2012), Ch 36 Soap, Fatty Acids, and Synthetic Detergents · pdf pp. 1431-1435, 1440
• Dryden (1997), IIIB Soaps and Detergents; Glycerin · pdf pp. 311-313, 318

2.3 Splitting the fat first gives fatty acids, sweet water and a simpler soap plant

Splitting is hydrolysis: the fat reacts with water instead of caustic, triglyceride + 3H2O → 3RCOOH + glycerine. It is an equilibrium, so the trick is to keep removing glycerine. In the continuous countercurrent splitter that Shreve and Kent describe, a stainless steel tower about 20 m high and 60 cm across runs at around 250 C and 4-6 MPa. Fat enters at the bottom and rises; hot demineralised water enters at the top and falls. The falling water collects glycerine and leaves as "sweet water" with 10-25% glycerine, while fatty acids leave the top with 97-99% of the fat split.

The crude acids are then distilled under high vacuum, about 10 mbar with live steam, because they decompose if boiled at atmospheric pressure. Kent's typical still cut is 1-5% light ends, 85-91% refined product and 7-11% residue, or pitch. Fractional distillation separates chain lengths, giving for example 99.5% pure C12 from coconut. It cannot separate stearic from oleic acid, which boil too close together. Their different melting points are used instead, by pressing, solvent crystallisation or hydrophilisation.

Neutralising the distilled acids is close to trivial: RCOOH + NaOH → RCOONa + H₂O, essentially instantaneous in a high-shear mixer at 85-98 C. Some plants neutralise 70-85% of the acid with soda ash and finish with caustic; the carbon dioxide released makes the paste very viscous, so this only pays where soda ash is cheaper or purer.

The same splitter feeds the wider market for Fatty acids, about 5.7 billion lb (2.6 million tonnes) worldwide in 2000 according to Wittcoff. Fatty acids or their methyl esters are also hydrogenated at 200-300 C and 200-300 bar over copper chromite to make Fatty alcohols, the raw material for many detergent surfactants.

• Shreve (1984), Ch 29 Soap and Detergents · pdf pp. 553-557, 564-565
• Kent (2012), Ch 36 Soap, Fatty Acids, and Synthetic Detergents · pdf pp. 1440-1450, 1454
• Wittcoff (2004), Ch 13 Fats and Oils · pdf pp. 442, 451

2.4 Glycerine is worth recovering, but its supply follows soap and biodiesel

Every route from fat releases glycerine, about one molecule for every three fatty acids. Recovering it means removing water and salt. Sweet water from a splitter is the easy case. Shreve describes the Procter & Gamble scheme: sweet water at about 12% glycerine passes through a triple-effect evaporator, where vapour from one stage boils the next, and reaches 75-80% glycerine, usually with no heat beyond what the hot sweet water already carries. The crude is settled hot for 48 hours to drop fatty impurities, giving about 78% glycerine. It is then distilled under vacuum (8 kPa, about 200 C) with a little caustic, and condensed in three stages; the first gives 99% chemically pure glycerine. Carbon bleaching and filtration or ion exchange finish it.

Spent lye from saponification is harder because it is loaded with salt, which comes out in the evaporators and must be removed constantly.

Glycerine has many modest uses rather than one big one. Dryden's pattern puts alkyd resins and plastics at 35%, tobacco humidification 15%, cellophane 12%, explosives 10% and food and pharmaceuticals 10%.

The supply side is the weak point. Natural glycerine is made because soap and fatty acids are made, not because anyone ordered more glycerine. As detergents replaced soap, supply lagged demand, and from about 1948 Shell's route from propylene via allyl chloride and epichlorohydrin filled the gap; Shreve puts synthetic glycerine at about 40% of the market, and Dryden's table shows the world synthetic share falling to 30% by 1988. Wittcoff shows the opposite risk: Procter & Gamble's fat substitute Olestra, made from methyl esters, could have caused a 50% world oversupply of glycerine had it succeeded, and biodiesel from methanolysis raises the same disposal problem.

• Shreve (1984), Ch 29 Soap and Detergents · pdf pp. 567-570
• Dryden (1997), IIIB Soaps and Detergents; Glycerin · pdf pp. 316-318
• Wittcoff (2004), Ch 13 Fats and Oils · pdf pp. 454, 456-457

See it in the table

Check your understanding

Why does a continuous saponification reactor recirculate 13-25 times as much reacted mass as fresh feed?

Kent counts seven steps for neutral fat saponification but three for fatty acid neutralisation. What explains most of the difference?

Why did synthetic glycerine from propylene become important from about 1948?