Fatty acids and soap qualities: what the numbers mean

An oil matters to the soap only through the fatty acids it carries. Knowing what each acid does, and how a calculator turns acids into five quality numbers, lets you read a recipe at a glance and see where the numbers stop meaning anything.

Two properties decide most of it

Soapmaking fatty acids differ in two ways: how long the chain is, from 12 to 18 carbons for the ones that matter most, and how many double bonds it carries, from none to three. Ricinoleic acid, from castor oil, adds a third feature, a hydroxyl group halfway along.

The effect of chain length shows in how readily each sodium soap dissolves. Below a temperature called its Krafft point a soap barely dissolves; above it, the soap forms micelles and lathers. For the saturated sodium soaps the Krafft point rises steeply with chain length: about 25 °C for laurate (C12), 45 °C for myristate (C14), 60 °C for palmitate (C16) and 71 °C for stearate (C18). So in a warm wash, laurate is in solution at once, which is fast, big lather and strong degreasing, while stearate hardly dissolves at all, which is what makes a bar last, and adds a slow, fine, creamy lather in a blend.

A double bond in its usual natural form bends the chain, and bent chains cannot pack into a tight crystal. Sodium oleate therefore dissolves far more readily than sodium stearate, the saturated soap of the same length, and feels mild and slippery, and a bar rich in it stays soft until it has lost its water. Each extra double bond softens the soap further and, as a later section explains, makes it quicker to go rancid.

Real soap is always a mixture, and mixtures do not simply average. Unilever researchers measured the blend of sodium laurate and sodium oleate that soap chemists call the soap maker’s eutectic and found its Krafft point depressed by more than 15 °C at about half and half. Mixing matters for mildness too. Pure sodium laurate irritated the skin of most volunteers in a repeated open-exposure test, and the authors cautioned that ordinary soap, a mixture of less irritating chain lengths, cannot be judged by that result.

The eight acids at a glance

Ranges from the Codex standards and the other primary documents on each oil’s page; shea from Akihisa et al. (2010).

Two more acids deserve a line. Coconut oil is 5–10 % caprylic (C8) and 5–8 % capric (C10) acid, whose soaps are more soluble still, and the quality numbers count them toward nothing.

Watson puts the difference in lather plainly: fluffy, big-bubbled lather from coconut and palm kernel oil, and a denser, more stable lather from oils like olive.

How the five numbers are made

The quality numbers are straight sums of fatty-acid percentages in the blend, which is each oil’s profile weighted by its share of the oils:

• Hardness = lauric + myristic + palmitic + stearic
• Cleansing = lauric + myristic
• Conditioning = oleic + linoleic + linolenic + ricinoleic
• Bubbly lather = lauric + myristic + ricinoleic
• Creamy lather = palmitic + stearic + ricinoleic

For the 40/30/30 bar the blend is 14.8 % lauric, 6.1 % myristic, 21.2 % palmitic, 3.4 % stearic, 41.9 % oleic, 8.6 % linoleic and 0.3 % linolenic. Hardness is 14.8 + 6.1 + 21.2 + 3.4, which the calculator rounds to 45; cleansing and bubbly are both 21; conditioning is 51; creamy is 25. Our classic olive, coconut and palm bar trades 5 % of the coconut for castor oil, which lifts bubbly and creamy above cleansing.

Look at the formulas again and the five names turn out to be four sums in different combinations: the lauric and myristic pair, the palmitic and stearic pair, ricinoleic, and the unsaturated C18 acids. Hardness and conditioning are close to mirror images, because between them they count nearly every acid in most oils; for the 40/30/30 bar they add to 96, and the missing few percent are acids the convention ignores. The recommended bands say the same thing twice: hardness 29–54 and conditioning 44–69 add up to 98 at either end. Bubbly is cleansing plus ricinoleic, so without castor oil the two are identical. Balancing five qualities is really balancing three shares, the lauric pair, the palmitic and stearic pair and the unsaturated rest, with castor on the side.

The convention also has blind spots in what it counts. Caprylic and capric acid, 14 % of coconut oil, count toward nothing. So does palmitoleic acid (C16:1), a fifth of macadamia oil, and so do the 20- and 22-carbon acids of meadowfoam and jojoba. This site adds γ-linolenic acid and pomegranate’s conjugated acids to conditioning, and labels that as its own extension on the formula page.

Where the ranges come from

The usual ranges are hardness 29–54, cleansing 12–22, conditioning 44–69, bubbly 14–46 and creamy 16–48, with iodine value 41–70 and INS 136–165 alongside. They are a convention that hobby lye calculators have shared since the mid-2000s. We know of no published study behind them; they describe where popular mixed-oil recipes sat, which makes them useful as a map of familiar ground and poor as a verdict on anything outside it.

The clearest proof is a soap older than any calculator. A bar of pure olive oil scores hardness 17, cleansing 0, conditioning 82, bubbly 0, creamy 17, iodine 85 and INS 106: outside six of the seven ranges. Yet olive-oil soaps such as those of Castile, Marseille and Aleppo have been made around the Mediterranean for centuries, and a well-cured castile is a firm, long-lasting, mild bar. The numbers go wrong for plain reasons. Hardness counts only saturated acids and ignores water and time: sodium oleate is a solid soap, and a castile that is soft and slippery at six weeks is hard after a long cure. Cleansing 0 says only that there is no lauric or myristic acid; olive soap cleans perfectly well. Bubbly 0 describes a lather that is slow to build and low, which is how a castile lathers.

If a bar works for you, your own figures are a better guide than the table. With a Maker account the calculator can save a target captured from a recipe that worked, and judge new recipes against that instead of the usual ranges.

Saturated, unsaturated, and the acid that goes rancid

Double bonds are where oxygen attacks a fat, and the most vulnerable spot is a carbon sitting between two double bonds, whose hydrogens come away easily. Oleic acid has no such carbon, linoleic has one and linolenic two. In a classic measurement, each extra double bond at least doubled the rate of oxidation (Holman and Elmer, 1947), and later work found that oxidisability rises in step with the number of those in-between carbons (Cosgrove et al., 1987). Oleic-rich soap keeps well; soap high in linoleic and linolenic acid is the kind that develops rancid smells and orange spots.

The convention counts all three toward conditioning, so a recipe can raise its conditioning number with the very acid most likely to shorten its shelf life. Hemp seed and linseed oil score high on conditioning and are among the least stable oils on a soapmaker’s shelf. The quality panel will not warn you; the iodine value and the linoleic and linolenic totals will.

Ricinoleic acid is the odd one out. Castor oil is 85–92 % ricinoleic acid, whose soap is so soluble that Failor treats castor as partly a solvent in transparent soap. In a bar it props up the lather of the other soaps, and it is the only acid counted toward three qualities at once, so 5 % castor oil lifts conditioning, bubbly and creamy by about four points each. Too much gives a soft, tacky bar; most recipes keep castor between 5 and 10 %.

What the numbers cannot see

The five qualities read the fatty-acid profile of the oils and nothing else. They cannot see:

• the superfat, which changes how a bar feels far more than a point or two of conditioning;
• the water and the cure, which decide how hard a bar is on the day you use it;
• the unsaponifiable matter, such as the few percent of sterols and triterpene alcohols in shea butter;
• additives, from sugar that boosts lather to clay and salt;
• the acids outside the convention, and the difference between a stable oleic chain and a fragile linolenic one.

Treat the numbers as a fast way to compare recipes that are otherwise alike, which is what they were built for.

Designing by acids: a worked example

Suppose you want an everyday bar without palm oil: moderate cleansing, the hardness of a palm bar, some lather from castor, and few polyunsaturated acids for a long shelf life. Work it acid by acid.

1. Cleansing from a lauric oil. Coconut oil is 68 % lauric plus myristic, so 25 % coconut gives a cleansing of 17, in the middle of the usual range.
2. Hardness from palmitic and stearic. Shea butter is 45 % palmitic plus stearic, so 20 % shea adds 9 points; 40 % olive oil adds 6.6, the coconut 3.0 and the rest about 1. With the 17 from cleansing, hardness lands at 37.
3. Lather from castor. 5 % castor oil brings 4.4 % ricinoleic acid, which lifts bubbly to 21 and creamy to 24.
4. The rest oleic-rich. Olive oil and high-oleic sunflower oil fill the remaining 50 %, which keeps linoleic at 7.8 % and linolenic near zero, and holds the iodine value at 59.

That is this site’s palm-free classic: 40 % olive, 25 % coconut, 20 % shea, 10 % high-oleic sunflower and 5 % castor oil. The calculator gives hardness 37, cleansing 17, conditioning 60, bubbly 21, creamy 24, iodine 59 and INS 145, and 138.5 g of sodium hydroxide for 1000 g of oils at a 5 % superfat. It also shows what the acids cannot: across the published saponification ranges of those oils, the 5 % superfat could really be anywhere from 2.1 % to 8.0 %. Our first-batch palm-free bar is the simpler cousin, 60 % olive, 25 % coconut, 10 % shea and 5 % castor, at hardness 35 and cleansing 17.

The recipe design guide takes the same approach through three full recipes, and compare oils puts up to six profiles side by side when you are choosing between candidates.