Glycerine Rivers: Trying To Understand Them

My previous post, Glycerine Rivers: Secret Revealed, about water concentration as a contributing factor to the formation of glycerine rivers, got lots of feedback all of which I’m very thankful for. Many said that they had never made the connection between water concentration and the rivers before, but that it made sense now that they looked at their own soaping experience from this perspective. Someone felt that they had had the opposite experience with water concentration and rivers. Two people came forward questioning the method of gradually adding water to one half of the batch mixing it in by hand, suggesting that it was in fact insufficient mixing rather than the water content that caused the formation of streaks in the high-water soap. This was a valid concern as it addressed the only variable other than water content in the experiment. Since believing is good but knowing is better, and since it was easy enough to find out, I decided to replicate my initial experiment correcting the method slightly.

I used the same formula as in the first experiment: 20% coconut oil, 40% olive oil, 40% palm oil with a 5% superfat and 1 tsp of titanium dioxide as colourant for my batch of 1000g oils plus some lemongrass essential oil for fragrance.

This time I made two separate lye solutions: I divided the total amount of NaOH in two and made a 1/1.4 lye/water solution with one half and a 1/2.4 lye/water solution with the other half. I then let both lye solutions cool down to room temp.

Like I did in the earlier experiment I added the colourant and fragrance to the oils and stickblended carefully. The previous time I had divided the batch in two at trace. This time I divided the oil mixture between two identical jugs, weighing them to make sure that I had equal halves. I then added the lye solutions each to its own half of the oil mixture and stickblended each batter to light/medium trace. Technically I now had two different soaps made from the same ‘masterbatched’ oil, colourant and fragrance mixture. Like I did the previous time I used a lengthwise divider and poured both soaps side by side in the same lined, wooden mould. I then placed the mould in the oven, pre-heated to 60C, left it there for 4h and then turned off the oven leaving the soap in the cooling oven just like I had the first time. The first time I took out the mould about 1.5h after turning off the oven. This time I left the mould to cool in the oven overnight.

The outcome of the second experiment is virtually identical to the outcome of the first experiment, the low-water soap showing no sign of rivers while the high-water soap is full of them throughout the log. This time I also got an interesting vertical yellow line throughout the log exactly in the spot where my corrugated plastic divider had been. I’ll get back to this yellow line later on.

In light of this second experiment I’m pretty confident that I can rule out the theory that it was insufficient mixing rather than the water concentration that caused the streaks in the high-water soap. In the second experiment the only variable was water concentration and the pictures pretty much speak for themselves: with the oil formula, colourant, fragrance, temperature and method being identical, the high-water soap showed streaks whereas the low-water soap did not. While uneven incorporation of colourant may cause some form of streaking in soap that did not happen in my experiment. The conclusion is then that in an ambient temp of 60C, water concentration was definitely a significant factor contributing to the formation of glycerine rivers in the soap. Beyond all reasonable doubt.

Having said that, I need to point out, once again, that by keeping all other variables constant I only tested for water concentration. This means that there may be a row of other factors such as ambient heat, sugar content, oils used, fragrance etc. that influence the formation of glycerine rivers at any given water concentration.

So, I popped my titanium dioxide soap in the oven for four hours and ended up with glycerine rivers. Big deal – that’s hardly groundbreaking news. Much more interesting is how they formed and why I didn’t get them in the low-water soap.

In a comment to my previous blog post, DeeAnna Weed made an interesting observation:

“Water content is not really what causes mottling and streaking, although I can see why one would think it is. What water content DOES do is affect whether the soap is likely to gel or not at relatively low temperatures. High water soaps go to full gel at much lower temps than low-water soaps and are thus more likely to show mottling and streaking than soaps that don’t gel.

Mottling and streaking is more likely if the soap reaches a full gel state and is allowed to very slowly cool, so the different soaps (stearic soap, oleic soap, palmitic soap, etc.) can crystallize at different times.

Opaque pigments are more likely to concentrate in the soaps that crystallize last. Also some of the soaps themselves are more opaque and some are more translucent, which in itself can cause mottling and streaking even without adding pigments.”

DeeAnna continues:

“When the old soap makers of the 1700s and 1800s deliberately wanted to make mottled soap, they worried about many of the same issues — in particular they wanted to pour the soap in the mold as hot as was practical with as little stirring as possible (they wanted large mottles) and then cool the soap in the mold as slowly and quietly as possible.

Mottled soap in those days was desirable because only pure soap could be mottled. The mottling proved the soap didn’t have any fillers that made the soap cheaper and reduced its cleaning power.”

(Anybody interested can read more about mottled soap here: https://web.archive.org/web/20130123080239/http://www.1911encyclopedia.org/Soap and here: https://archive.org/stream/modernsoapscandl00lambrich#page/n5/mode/2up.)

This ties in with what Kevin Dunn tells us about the saponification behaviour of low-water vs. high-water soap in his book Scientific Soapmaking The Chemistry of the Cold Process. Low-water soaps saponify faster than high-water soaps creating more heat in a shorter period of time and reaching a steeper and higher initial heat peak than corresponding high-water soaps. Low-water soaps go through gel phase (Kevin Dunn refers to soap going through full gel phase as ‘neat soap’: “a lamellar, lyotropic liquid-crystalline phase of soap and water“) at a higher temperature than high-water soaps and they go through gel phase more rapidly than high-water soaps.

What does all this tell me about the soap in my experiment? By using a high water concentration I increased the likelihood of the soap reaching gel phase and the high-water soap may well have reached a full gel even without the high ambient heat, ie the heat generated by the saponification process itself would probably have been enough to make it go through gel phase had I just insulated the mould by wrapping it in a towel or so. It’s also possible that the high-water soap would have gone through full gel at a lower temperature than the 60C that I now forced on it in my oven. By keeping the ambient temperature high for a prolonged period of time I may also have forced the high-water soap to stay in its gel phase longer than it would have had I left it to its own devices insulated with a towel. A prolonged gel phase in combination with slow cooling would create ideal conditions for gradual crystallization and streaking to occur.

If a full gel phase and the slow cooling that makes gradual crystallization possible are prerequisite for glycerine rivers or streaks to form, how should I then interpret the fact that my low-water soap doesn’t show any signs of streaking? There seems to be two possible scenarios for why the low-water soap shows no streaks:

A. The low-water soap didn’t go through gel phase, ie it didn’t crystallize. That’s a pretty radical thought given that the soap was saponifying, ie generating heat by itself in addition to spending 4h in an ambient temp of 60C. Also, it spent that time rubbing shoulders with its high-water friend who evidently did go through full gel, yet the low-water soap shows no signs of what is known as ‘partial gel’. In partial gel some parts of the soap go through gel phase while others don’t due to uneven heat distribution. Visually, this tends to manifest itself as ‘crop circles’; darker areas where the soap has crystallized and lighter areas where it hasn’t. But, I never saw the low-water soap gelling; I put my experiment in the oven and went grocery shopping – I must remind myself not to multi-task next time I’m conducting important scientific experiments . When I took a first peek, the soap was well into its third hour in the oven and the high-water soap was gelling; dark in colour and soft to the touch, but the low-water soap looked like it does now and was firm to the touch. Since we know that low-water soaps go through gel phase at a higher temperature than high-water soaps it’s not entirely impossible that my ambient 60C wasn’t enough to force the low-water soap to gel so I’m not discarding this alternative.

B. The low-water soap did crystallize, but aided by the warm ambient temperature it moved in and out of its gel phase at such speed that the slow cooling conditions required for gradual crystallization and streaking didn’t occur. The quick initial rise and fall in the temperature of low-water soaps described by Kevin Dunn would support this. If the low-water soap did go through gel phase it will have done so at a higher temperature than the high-water soap and probably at a temperature well higher than the 60C in my oven. By the time I looked at it in its third hour in the oven, it would have completed its gel phase, finished crystallization and cooled down to 60C again.

So, I didn’t monitor the internal temp of the low-water soap and I didn’t see it gel, but now I was dying to know if it had gelled and crystallized. Normally, crystallized soap is deeper in colour and has a more translucent look compared to un-gelled soap. But this soap is full of well-dispersed titanium dioxide and that doesn’t help me at all in trying to determine its internal structure. Titanium dioxide can make a translucent soap completely opaque and is used specifically to make soap lighter giving it a ‘creamy’ or ‘chalky’ look depending on how you like to describe it.

Looking at the soap against the afternoon sun doesn’t give much of a clue; the light seems to penetrate through the edges pretty much the same way on the high-water and low-water sides. The titanium dioxide makes it difficult to see how opaque or translucent it really is.

Letting the sunlight reflect from the cut surface at an angle also doesn’t provide any ready answers: The high-water side shows the slight indentations typical of streaking and a fine grain in between. The low-water soap shows lots of air bubbles from this angle. The mysterious yellow line down the middle shows the same fine grain as can be seen elsewhere on the high-water soap.

So, looking at the outside and cut surfaces of this soap didn’t give any definite clues as to whether the low-water soap had gone through gel phase or not, but I remembered an interesting ‘accident’ I had a while ago. While trying to emboss a round soap made in a PVC pipe I put a little too much pressure on the soap and it shattered and split in two. I could have kicked myself for being so clumsy, but seeing the inside of the crystallized soap was quite interesting.

When you cut soap, you cut through the crystalline structure and you’re not likely to get a good view of this internal structure. When you put pressure on the soap and it splits, it’s likely to split along the grain and if you’re lucky you might get to see a split as neat as this. Here you can clearly see the sunlight reflected in the grain of the fibrous crystalline structure giving it a silken sheen.

The lyricist in Kevin Dunn is revealed as he describes the crystalline structure of soap like this: “It is as if a nation of earthworms had settled down for a long winter’s nap“. How serendipitous that my carnation soap in the image above has such an accurate earthworm colour!

I obviously needed to see what my low-water soap really looked like inside and the way to do that was to try to get it to split. I didn’t want to crush it but rather shatter it. One might device some better way of doing this but I was impatient so I just gave it a decisive whack with my rubber mallet. That caused a split and I was able to pull the bar apart in four pieces.

The two pieces in front are low-water soap. The split wasn’t quite as neat as for my carnation soap above but it does reveal some interesting texture.

I don’t have a microscope but a good camera helps when you need to magnify stuff. Along the top part of the picture you can see how the soap has broken off in ‘flakes’ against the grain suggesting a lamellar structure. If you look closely at the ridges on the left you can see fibrous texture.

Here a bit of editing in the image on the right (lowering the gamma bright) helps emphasize the reflection of the sunlight in the texture of the soap. The reflection is strongest right under the spot where the mallet hit the soap, but if you look closely you’ll see reflection all over the surface, even in the ‘valleys’ between the parallel ridges at the top corner of the split surface. It seems unlikely that you’d get that kind of texture and reflection in an un-gelled, un-crystallized soap.

In this light I’m leaning towards the second scenario of why the low-water soap doesn’t show streaking: Aided by the heat in the oven the low-water soap went in and out of a full gel phase too rapidly for the slow, streak inducing cooling conditions to occur. Since the low-water soap didn’t show continued gelling at 60C I’m going to draw the conclusion that the soap moved out of its gel phase at a temperature higher than 60C. Ie the heat in the oven did not prolong the gel phase of this soap. In fact, the heat in the oven may – or may not – have made the duration of the gel phase shorter by increasing the rate of saponification.

Had the constant temperature in the oven been higher, high enough to coincide with the temp at which the low-water soap went through full gel, it may have prolonged the gel phase and the cooling, and the low-water soap might have shown streaking as a result.

Now back to the mysterious yellow line running the length of the log right where I pulled out the plastic divider. How can that be explained? A few days after the soap was made the line is not as bright as it was on cutting, but it’s still clearly visible. DOS (dreaded orange spots), a visual indication of rancid oils in the soap, spring to most soapmakers’ minds when dealing with unaccounted for yellow colour. But DOS rarely shows up immediately and the spots seldom have well-defined edges like my yellow line does. Considering that the entire log was made with the same well-blended oils it’s unlikely that the centre line alone would show signs of rancidity. Looking at the surface of the log, however, does give some interesting clues:

 

 

The high-water side of the log is uneven in colour. Apart from the crackle effect which can be seen it shows yellow ‘stains’ or blotches. The low-water side of the log looks pristine without yellow ‘staining’. But the most interesting clue is the top of the soap:

From the inverted stamp pattern on top you can tell that the marks were made at about medium trace on both the low-water and high-water soaps. Before making these marks I smoothed out the entire top surface with a scraper. Now, if you look at the bar in the centre you can see that the surface of the low-water soap is smooth between the ridges. The high-water side, on the other hand, is ‘knobbly’ or ‘wrinkly’ like orange peel – or a severe case of cellulite.  So, yellow blotches on the side and a knobbly, uneven top. Off the top of my head I’d say that looks like overheating issues. Having seen these clues I’m thinking that the yellow line in the centre of the soap is also a sign of overheating. While the low-water soap was going through gel phase it would have reached a temperature way above the gelling temp of the high-water soap. Seeing that the high-water soap overheated on the side that was closest to the 60C ambient temp it’s possible that it was scorched too on the side that was right next to the much hotter, gelling low-water soap.

So, with the same formula, fragrance, colourant, mixing temp, mould and ambient temp the high-water soap overheated but the low-water soap didn’t? Looks like that just happened here.

This, of course, is speculation. Once again I have to point out that I only monitored the ambient temp, not the internal temperature of each soap. Also, I’m not knowledgeable enough to attempt any kind of explanation of what happens on a molecular level when soap overheats. But, if it does hold true that high-water soap can overheat at a temp where the identical low-water soap shows no signs of overheating, the practical implications can be much more far-reaching than just forcing or avoiding glycerine rivers. I, for one, will be a lot more cautious when oven processing high-water soaps in the future, keeping lower temperature for a shorter period of time.

Milk and beer soaps are typically made with a full liquid amount “because the liquid is what holds the goodness”. Because of this they’ll gel at a relatively low temp and often show streaking even without added colourant. The liquids also contain sugar which increases the heat generated during saponification, and these soaps are notorious for easy overheating. I regularly oven process milk soaps, but I make them with powdered milk reconstituted with a very small amount of water added to the oils and a steep water discount in the lye solution, To date I’ve never had a milk soap made like this overheat in the oven – or produce streaks. Is it because of the water discount? It could be.

I’ve read lots of accounts of tops ‘deflating’ on soaps that were oven processed, yet I’ve never had that happen myself. On the one hand I don’t raise the ambient temp above 60C because that is what the proofing setting on my oven keeps. But on the other hand I also always do a fairly steep water discount for soap logs. For soap in individual moulds I tend to go with a higher water concentration and thinking back now, that is the only context where I’ve experienced overheating issues (‘pockmarks’ forming on the outside of soaps moulded in silicone moulds) when oven processing at 60C. To remedy a situation where a soap seems to ‘melt’ from the heat during oven processing, it seems intuitive to lower the heat. While that may do the trick it’s also possible that lowering the water concentration works equally well.

Many fragrances are known to ‘heat up’ and speed up trace, ie accelerate saponification. Again, it seems intuitive to add water because that slows down saponification. But the resulting high water concentration will also bring down the temperature at which the soap gels increasing the risk of partial gel when trying to avoid gel – and increasing the risk of streaking if pigments are used and gelling is encouraged.

When it comes to avoiding gel soap often shows partial gel even though it was refrigerated or kept in the freezer to prevent temps from rising too much. While keeping temperatures low is key, lowering the water concentration may also help because it keeps the soap from gelling at low temp.

Throughout this post I’ve talked about water concentration in conjunction with heat. To sum it up both heat and water concentration are important factors in the formation of glycerine rivers or streaks. What my experiments show quite clearly is how constant external heat can do different things to saponifying soap depending on water concentration. This is definitely true for the formation of streaks, but it may also be true for other issues that we tend to associate with overheating.

All this has been very exciting and I’ve learnt a lot in the process, but now I’m off to chop up a whole lot of titanium dioxide and lemongrass soap for embeds..