Neither of these hurts anything. A bar with bubbles and a bar with rivers wash exactly as well as a perfect one, and some soapmakers cultivate rivers on purpose. They get filed together as "cosmetic flaws", which is fair, and treated as one subject, which is not: one is a technique problem you control completely, and the other you mostly decided before the lye went in.
Which one have you got?
- Air bubbles: round voids through the bar, from pinholes to peas, distributed all through it. Visible on a cut face.
- Glycerin rivers: translucent crackled veining, like a dried riverbed or crazed glaze, usually in the middle of the bar and most obvious in pale colored soap.
- Not those? Soda ash is a white powdery film on the outside that rubs off; partial gel is a dark translucent circle with a pale ring around it; hard white specks inside are stearic spots from false trace; orange or brown freckles months later are rancidity.
Air bubbles: three seconds at the blender
Much of the air in a soap bar goes in before you switch the blender on. Lower a stick blender into batter and the bell traps a pocket of air underneath it; pull the trigger and that pocket is chopped into a few thousand bubbles and distributed evenly through your loaf. Everything after that, pouring and tapping included, is damage limitation.
Submerge the head fully, then tilt it to one side and tap it gently on the bottom of the pot until you see the trapped air escape and stop. That is "burping", and it takes about three seconds. Then bring it upright, keep the head under the surface for the whole time it runs, and pulse rather than holding the trigger. A blender that breaks the surface while running whips air in as fast as it emulsifies, which is also how a lot of people arrive at a thicker trace than they wanted.
- Don't over-blend. Short bursts with the motor off between them. More blending is more air, as well as less working time.
- Give the head enough depth. You can't burp a blender in batter shallower than its guard is wide — a small batch in a wide bowl is the usual culprit. Batter depth by batch and container.
- Pour low and slow, close to the mold surface, so the batter isn't falling through air on the way in.
- Bang the mold on the counter several times after pouring. It works: bubbles rise if you give them a route.
- Pour thinner. Thick trace traps bubbles because they can't rise through it, so if bubbles are your problem, pour at light trace. It's a trade: thicker trace is one of the defenses against soda ash, so pick for the problem you have.
- Spray the top with alcohol to pop surface bubbles, which also happens to prevent soda ash.
Glycerin rivers need two things at once
Rivers are a heat effect. Saponification produces glycerin throughout the batch, and when a loaf gets hot enough the glycerin and water separate into channels rather than staying evenly dispersed; as the soap cools and hardens around them, those channels set as translucent veins.
So they need heat and they need water. Heat is mostly your mold, because whether a batch gets hot is decided by how fast it sheds heat; water is your lye concentration. Put the two together and the pattern soapmakers commonly notice falls out:
| Setup | Water in the batter | Surface per 100 mL | Rivers? |
|---|---|---|---|
| 25% solution, 5 lb slab covered and insulated | 27.4% | 46.8 cm² | Both conditions at their maximum. Expect them. |
| 33% solution, 5 lb slab covered | 20.3% | 46.8 cm² | Gels hard regardless; the water is the variable you have left. |
| 33% solution, 2.5 lb loaf covered | 20.3% | 61.9 cm² | The common case, and the one that goes either way. |
| 33% solution, 2.5 lb loaf uncovered, cool room | 20.3% | 61.9 cm² | Usually never gets hot enough. Expect soda ash instead. |
| 40% solution, 2.5 lb loaf covered | 15.9% | 61.9 cm² | Gels, but there is little free water to congeal. |
| 33% solution, 12 cavities covered | 20.3% | 138.1 cm² | Sheds heat far too fast to gel. Rivers are off the table. |
Water as a share of the whole batter for the classic bar at 1,000 g of oils and 5% superfat; surface per 100 mL of batter from the mold's filled dimensions, as in the gel phase guide. A lower surface figure sheds heat more slowly and so gels more readily. The last column is judgment from those two figures, not a measurement.
Read the top and bottom rows together. A 25% solution in a slab is 27.4% water in a mold that loses heat through only 46.8 cm² per 100 mL — the maximum of both conditions. The same recipe in 12 cavities is spread over 2.2 times the surface of a standard loaf and never gets hot enough for the question to arise. And at 40% the batter carries only 15.9% water, so there is little to congeal even when it does gel.
So discounting your water and soaping cooler both work, but they aren't the only levers. You can also just use a smaller or more open mold, which costs you nothing in cure time, and the water discounting guide lays out what the other route costs.
Titanium dioxide gets the blame. As Soap Queen's rivers article puts it, rivers are more visible with pigments, especially TD, which is denser than micas: translucent channels show against a dense white the way a watermark shows against paper. Whether rivers form without a colorant is less settled. Auntie Clara's experiments, as Wholesale Supplies Plus reports them, split one high-water batch into colored and uncolored layers and found rivers only in the colored layer. The one way TD does contribute is worth keeping in mind, though: dispersing it in water rather than in oil adds water to the batch, which is the second condition. Soap Queen also recommends dispersing it in oil. If you like TD and dislike rivers, disperse it in a light oil and the contribution goes away while the visibility, unfortunately, does not.
Are they worth fixing?
Honestly, often not. Neither affects lather, hardness, mildness or shelf life; both are invisible in a dark or heavily colored bar; and rivers in particular have a following: makers deliberately soap hot with full water to get the crackled effect, which photographs beautifully.
The case for caring is consistency rather than quality. If you sell, bars that look different from batch to batch are a problem even when they wash identically, and the two faults are unequally fixable:
- Bubbles are free to fix. Burp the blender, pulse, pour low, bang the mold. Nothing about the recipe changes, so there is no reason not to.
- Rivers cost something to fix. Discounting water shortens your working time; preventing gel changes the color and the unmold schedule. Decide whether the look matters more than the trade.
And if you do nothing: it is still soap, and the person washing with it will never know.