Sunday, February 9, 2014

Historical Lye Making, Part 2

On Thursday, we posted about historical methods to drip lye and some of the chemistry associated with it.  Today, we wanted to talk about measuring the strength of the lye to check if it's concentrated enough for making soap.  In the olden days, that was often done with some sort of density test, either by floating an egg or potato, or matching the density of the lye solution with a saturated table salt solution (sodium chloride, NaCl).  An egg has a density of 1.03-1.1 g/mL, and a saturated NaCl solution is a little more precise at right around 1.2 g/mL, but in our soap calculations, we normally use a solution that is 1.3 g/mL, according to density-sodium hydroxide concentration correlations.  One initial conclusion from that is that old-time soap makers probably used less concentrated lye solutions.  Similarly, older soap recipes often call for cooking the lye water and fat together (i.e., hot process soaps), which probably boils off a lot of the extra water.

For modern homesteaders, who might have a scale and measuring cup handy, it would be much more precise just to measure the density of the lye solution directly.  (A graduated cylinder would make this calculation--and other density calculations you might want to do--more precise, but a measuring cup, used judiciously, should be good enough.

Another technique is to use a pH indicator and either dilute a small (representative) portion of the lye water or titrate it with an acid (such as vinegar) to find the strength of it.  The pH indicator would also be useful during the soapmaking process to check the progress of the saponification reaction.

Let's take a look at each of those techniques in more detail.

To make soap, we normally use a ratio of something like 2.89 oz NaOH to 7.87 oz water, which works out to about 26.9 wt% NaOH.  According to the above calculator, that should give us a solution density of 1.29 g/mL.  (The analogous numbers for KOH lye would be 4.06 oz KOH to the same amount of water, giving 34.0 wt% KOH, and a density of 1.33 g/mL.)  If we take an egg density of 1.1 g/mL, we can calculate the amount of water that should be displaced by the egg if we know the volume of it.

A typical large egg has a mass of 57 g, corresponding to a volume of 51.8 mL.  Buoyancy dictates that the egg should displace 57 g of the lye solution, which will correspond to a volume less than 51.8 mL if the lye solution is more dense than the egg (which it should be if the egg is floating). As an approximation, we can find an equation for an egg and make a graph to see how much of the egg should be above the water for a quarter-sized interface.

The egg equation came from here, but we normalized it to match the dimensions of an actual egg.  We assumed that an egg was sufficiently symmetrical to use a 2-D projection and calculate areas instead of using a 3-D model and calculating volumes.  In reality, the egg will sit with the skinny end slightly lower in the water since the air pocket is toward the flatter end.  In any case, leaving an area the size of a quarter above the surface would require a lye density of 1.13 g/mL, which is considerably less dense than our standard recipe, which has a density closer to 1.3 g/mL.  If the egg were a little less dense (toward the 1.03 g/mL end), it would sit higher.  As a point of reference, a potato has a density near 1.09 g/mL, in the same range as an egg.

This is a real egg in our standard lye solution (using NaOH).  The solution is yellow because we were testing pH indicators with it (described below).  The real egg looks not too far off of the graphical one, but there are more precise ways to test the lye's strength. 

For example, using data found here (and their related NaOH calculator), we can make a correlation, measure the density of the lye directly, and use the correlation to calculate the concentration.  It would help to have a digital scale and a graduated cylinder, but you can probably get at least as close as the egg/potato method with an old spring-loaded scale and a measuring cup.  We dripped a small batch of lye recently and were doing tests with it, but accidentally spilled it in the kitchen sink before we could test this method.  For lye dripped from ashes, use the KOH equation.  Note that if our solution density is 1.3 g/mL, our lye concentration (as KOH) is about 34 wt%, or 5.2 molar.  This density method will be our favorite lye strength test going forward.

Another way to test the strength is with a pH indicator.  One natural pH indicator is cabbage juice, which contains anthocyanidin pigments.  (As an aside, we noticed similar color changes in elderberry juice and wondered why; elderberries have a similar set of pigments.)  These pigments change structure as the pH of a solution changes, with each structure having a different color.  See here for more info.

The pigment structures of the cabbage anthocyanidins look something like this, with the different colors as shown.  Part of the reason the change from red to purple happens over such a wide pH range is the colorless intermediate.  Similarly, the yellow compound starts to form at pH > 8, but doesn't become the dominant form of the molecule until much higher pH (the presence of both yellow and blue make the solution green, kind of like a Ziploc bag).  The "R" groups are glucosides (i.e. substituted glucose molecules).  Sources for this figure came from here, here, and here.  If you took note of the concentrations above (i.e., that our standard soap recipe calls for 5.2 molar lye) and you are familiar with the pH scale, you might realize that there's a bit of a problem here.  That is, our lye should be at pH 14.7, but our indicator will be yellow at every pH > 11.

Fortunately, we can dilute a small, representative portion of the lye to bring it into the pH range where the indicator is effective.  On the far left in this picture is an undiluted lye solution we dripped from some wood ashes a few weeks back; it's yellow, which means the pH is at least 11.  Since pH is measured on a log scale, diluting by a factor of 10 (conveniently 1 teaspoon solution plus three tablespoons water) should decrease the solution pH by one unit (assuming the water is actually neutral).  On the first dilution, the solution is already green!  That means the undiluted solution was not much over pH 11.  The further dilutions (using one teaspoon of the first dilution plus three tablespoons water, etc.) are consistent with that conclusion, looking similar to pH 9 and pH 7-8  solutions above.  The upshot of this technique is basically (heh) that if the lye is concentrated enough for soapmaking, it should take at least four 1:10 dilution steps to show a color other than yellow.  Alternatively, that means that we should concentrate our lye solution by a factor of 1000 before using it to make soap.  Unfortunately, we only made around a quart to begin with, so we'll only be left with a few drops at the end--not enough to do much with (even dissolve a feather, which was another test of lye strength we were going to try).

Another approach would be to titrate the lye with an acid, and figure out how much acid we needed to observe a color change.  Maybe that will be the subject of a future post.

Guess we'll just have to dry it down with waste heat from the oven (after baking bread or something) and store it in a jar until we can make some more!

Also, in case you're interested, here's how we made the pH indicator solution.  We chopped about a third of a cabbage to give around four cups chopped cabbage.

Then we poured about two cups boiling water onto the cabbage and let it steep for about two hours.

Then we strained out the cabbage (and made coleslaw!), leaving this dark purple-colored liquid.  We add about a teaspoon of this 'cabbage tea' to a cup of liquid to test the pH. It's a little-known fact that a hot jar of this liquid was the inspiration for both the band name 'Deep Purple' and their hit single 'Smoke on the Water.'  (Don't bother looking that up.)

Have you dripped lye from wood ashes?  What did you use it for?  How did you test the strength?  Let us know in the comments section below!


  1. I have dripped lye from wood ashes! Actually, I soaked the ashes in boiling water and then used the water to soak dirty linens in. I'm interested in historical laundry methods and soaking linen in lye was a ubiquitous first step (sometimes followed by washing with soap, but soap was expensive in the Renaissance). Thanks so much for your post--I'm looking to figure out how concentrated my lye solutions should be for optimal cleansing properties. Do you have any thoughts on that matter?

    1. Thanks for stopping by!

      I think the more concentrated the solution, the better it would be for cleaning, but at some point it would start to also dissolve dyes on the cloth. I'd have to read more about the chemistry of Renaissance dyes, but if I had to guess, I'd shoot for a pH between 10 and 11 as a starting point. You can probably get a more precise answer by finding the molecular structure of a dye of interest and then finding it's pKa; staying at least one pH unit below the pKa should keep it pretty insoluble.

      Although, if you're washing the linens by hand, the durability of the skin on your hands might become the limiting factor. It's hard to put a solid number on it, but I'd start to get nervous about leaving my hands in a solution with a pH higher than about 9 for extended periods of time. This paper is somewhat consistent with that range.

      Hope that helps!

  2. This pair of posts is amazingly useful. I have here a 16th century recipe for making what I can only call octopus-lutefisk (from, believe it or not, Transylvania). And if I’m going to make it, of COURSE I have to make my own lye solution...

  3. wow... real folk scientists... thanks : )

  4. hey soapies... just in case it makes u froth, i'm making soap with my homeschooled kids but needed to reverse engineer the amount of koh in our solution as im going to steal a little for making a traditional japanese metalsmiths patina called rokusho... yours was the only site that gave all necessary parameters. thanks again :)