Superfat and lye discount: why recipes leave some oil unsaponified

A superfat is a deliberate shortage of lye. The calculator works out how much lye your oils could turn into soap, then has you weigh a few percent less, so a few percent of the oil never meets a hydroxide and stays in the bar as oil. This guide covers the arithmetic, the reasons behind it, where makers set it for each kind of soap, and what the calculator can tell you about how much margin you really have.

Superfat and lye discount are the same number

The two names describe one act from opposite ends. Lye discount is what you do: work out the lye that would saponify every gram of oil, then use less. Superfat is what you get: the oil left unsaponified.

Take 1000 g of olive oil. At its representative saponification value of 190 mg KOH/g, it needs 135.4 g of sodium hydroxide to turn completely into soap. At a 5 % superfat the calculator asks for 128.7 g, which is 95 % of that. The last 5 % of the oil’s demand for lye is never met, so about 50 g of oil stays in the bar.

The calculator always applies the superfat to the lye, never to the oils: all of the oils go into the pot, and the lye is the full requirement times (1 − superfat ÷ 100), corrected afterwards for the purity of your lye. The formula page writes it out with its sources.

Some older books superfat the other way round: they calculate the lye for the recipe’s oils and then add extra oil. The result is nearly the same. Saponify 1000 g of olive oil with its full 135.4 g of sodium hydroxide, add 50 g more oil, and the superfat is 4.8 %, a shade under 5 % because the extra oil is 5 % of the original oils but only 4.8 % of the new total. When Kevin Dunn and his students compared the two practices in the laboratory, they found no real difference between them. The belief that the added oil is the one left over did not survive, as the section on trace below shows.

100 g of oils at a 5 % superfat95 g: the lye you weigh is enough for this much, and it becomes soap5 g stays oil in the barWhere makers set itLye-heavy: free alkaliLiquid soapEveryday barsSalt barsLaundry-5 %0 %5 %10 %15 %20 %

Why leave lye out at all

Your oil’s saponification value is a range

Every oil’s saponification value varies from lot to lot with origin, season and refining, which is why food standards publish a range for each oil: Codex gives olive oil 184 to 196 mg KOH/g and coconut oil 248 to 265. A calculator has to weigh the lye from one figure inside each range. If your particular oil sits lower in its range than that figure, it needs less lye than you weighed, and the difference comes straight out of your superfat.

The gap between table and bottle is not hypothetical. Dunn measured the palm oil he had bought at 196 mg KOH/g, about 3 % below the 203 his supplier quoted, and both figures sit inside the published range of 190 to 209.

Our calculator carries every published range through to the result. For a bar of 80 % olive and 20 % coconut oil at a nominal 3 % superfat, the real superfat could be anywhere from −0.2 % to 6.0 %, depending on where the two oils fall in their ranges. The negative end is a lye-heavy bar.

This is what the craft’s habit of about 5 % buys you. For olive and coconut blends, a nominal superfat of about 3.2 % is the smallest that keeps even the worst case at or above zero. For the classic bar of 40 % olive, 30 % coconut and 30 % palm oil it is 3.7 %, because palm oil’s published range is wider.

Scales and lye are not perfect

A scale that reads to the gram can be a gram out on the lye, and one gram is 1.4 % of the 72.1 g of sodium hydroxide that a 500 g batch of that 40/30/30 bar needs. The smaller the batch, the bigger the share a single gram becomes.

Purity can push either way. Potassium hydroxide is sold at around 90 %: industrial flake is about 91 % KOH and 8 % water, and reagent-grade pellets are guaranteed only to 85 %. Enter the wrong purity for your KOH and the lye is several percent out, possibly in the dangerous direction. Sodium hydroxide that has absorbed moisture from the air weighs heavier than the NaOH in it, which errs the safe way, towards a softer bar with more superfat. The NaOH and KOH guide covers purity in detail.

A little free oil is milder

The last reason is feel. A few percent of unsaponified oil makes a bar feel less stripping; much more and the bar turns soft and the lather thin. Watson puts the usual range at 5 to 8 % and warns against calculators that default to zero; Faiola gives 3 to 8 % as the common range. This part is a convention of the craft. No one has measured the superfat at which a bar stops feeling harsh, and the answer would depend on the oils, the cure and whose skin it is.

Where makers set it

Conventions of the craft, as the recipes on this site are written. The superfat drawing above uses the same bands.

The calculator flags the edges. A negative superfat gets a danger warning, because it means more lye than the oils can use, which is only ever done on purpose, for laundry soap. Above 20 % it warns of a soft bar, weak lather and a higher chance of rancidity. With potassium hydroxide alone it adds a note from 3 to 4 %, the margin a liquid soap may need for the spread in saponification values, and a warning above 4 %.

At 0 % there is no margin left at all. Put our all-coconut laundry bar into the calculator and its worst case across the published ranges is −3.4 %. That is harmless in a washing machine and a good reason not to wash your face with it. At the other end, the pink salt bar at 18 % runs from 15.3 % to 20.6 % across the same ranges: once the superfat is this high, the uncertainty stops mattering.

High superfat and orange spots

Orange spots are rancidity, the oxidation of unsaturated fat, and polyunsaturated oils, rich in linoleic and linolenic acid, go first. It is tempting to blame the free oil alone, but the superfat is a smaller lever than it looks. When Dunn aged olive oil soaps in oxygen at 60 °C (140 °F), a bar with a 10 % lye discount did end up more orange than one with none, but the one with none went noticeably orange in about the same time. His conclusion was that lowering the superfat is not a cure for orange spots.

What changed the outcome was the oil itself, since coconut oil soap never became noticeably orange, and additives: rosemary extract and the chelator EDTA each held the orange off past the end of his 300-hour test. Traces of iron and copper, heat, light and damp storage speed rancidity up. None of that makes a very high superfat wise: the usual ranges top out at 8 %, and Faiola links superfats above 10 % to a shorter shelf life. The orange spots guide and the antioxidants and chelators guide take it further.

Oil added at trace is not the oil left over

A popular technique is to hold back a luxury oil and stir it in at trace, on the theory that the lye is already spoken for and the last oil in will be the superfat. Dunn tested it. His students made pairs of soaps with identical ingredients, one with every oil mixed before the lye went in and one with an oil held back until trace, held both at 71 °C (160 °F) for four hours to finish the reaction, then extracted the unsaponified oil and analysed it by NMR spectroscopy.

From Dunn’s study for the Handcrafted Soapmakers Guild (2007). The order of mixing made little or no difference; how fast each oil reacts decided what was left.

Dunn’s explanation is that olive and grapeseed oil react more slowly than coconut and palm, so they were over-represented in whatever oil survived, whether they went in first or last, while castor reacts faster, so less of it remained even when it was the last oil in. Whatever lye is still free when the extra oil arrives treats it like any other oil.

The practical upshot is to mix all your oils before the lye. Stirring a special oil in at trace buys nothing, and Dunn points out that it risks mixing it unevenly, so that some bars carry more free oil than the recipe intended and others less.

Hot process: after the cook, you do choose

Hot process is different because the cook runs the reaction to completion before the soap goes into the mould. There is no free lye left, so oil stirred in afterwards stays as oil. Classic Bells cautions that it keeps that form for a while but is not guaranteed to last, and that saturated and monounsaturated oils keep best.

In the calculator, choose Hot process and a tick box appears beside each oil for adding it after the cook. The lye is then worked out on the other oils, at whatever discount you set for the cook (makers often use 0 to 1 %), and the calculator tells you the superfat the held-back oil gives.

That superfat is counted in lye, not in grams. Cook 950 g of olive, coconut and palm oil (40 %, 25 % and 30 % of the recipe) at a 1 % discount, stir in 50 g of shea butter afterwards, and the calculator asks for 141.1 g of sodium hydroxide and reports a superfat of about 5.3 %. Hold back 50 g of avocado oil instead and it reports 5.4 %: avocado oil’s saponification value is a little higher than shea butter’s, so the same weight stands for a little more lye that was never supplied.

What the calculator shows you

The Superfat box is the lye discount, 5 % unless you change it. Two parts of the results tell you what that number really amounts to.

Effective superfat, among the result tiles, counts fat you might not think of as oil: fat carried in by milk or cream, and oils stirred in after a hot-process cook. Add 300 g of heavy cream to a batch of 1000 g of oils at 5 % and the effective superfat is 14.3 %, because cream is 36 % fat and none of that fat was in the lye calculation.

How well is this pinned down? takes the published range of every oil and shows what your superfat really is across them: the superfat you asked for, the worst case if the oils saponify low, the best case if they saponify high, and the lye the ranges allow. When the worst case goes below zero it says so plainly and names the lowest nominal superfat that avoids it. The worst case takes every oil at the bottom of its range at once, which real lots rarely do: it marks the furthest the published data could put you, and the real spread is narrower. If your supplier gives you a certificate of analysis with the measured saponification value of your lot, enter it under Measured values from a certificate of analysis and that oil’s uncertainty disappears.

Superfat is not the only thing that decides how mild a bar is, and it does little to the pH: the free oil sits beside the soap without neutralising anything. How to check a finished bar for free lye is in the soap pH guide, and why saponification values vary in the first place is in the SAP values guide.