A friend shared a video on Facebook of a process for glazing pottery called "mocha diffusion." A light-coloured glaze is applied to a pot and then, while it's wet, a black liquid is dripped into it. It doesn't just diffuse but forms a dendritic pattern, like this:

enter image description here (image source.)

The video claims that this is due to acid-base chemistry. A quick look around the web reveals that a mild acid is added to liquid containing the dark pigment, but I haven't found an explanation of why this would lead to these kinds of patterns.

The patterns look like viscous fingering, but that doesn't explain why an acidic solution is needed. I realise there may be some chemistry involved here, but my question is what causes the fluid dynamical instability that forms these patterns. Is it a chemical reaction (and if so how does it couple to the fluid flow), or is it just viscous fingering (in which case why would the acid be needed), or could it be some other physical process?

  • 1
    $\begingroup$ This could form the basis for the openning of the next Dr. Who! $\endgroup$ Commented Jun 22, 2016 at 6:04
  • $\begingroup$ have a look at fractals en.wikipedia.org/wiki/Fractal $\endgroup$
    – anna v
    Commented Jun 22, 2016 at 6:05
  • 1
    $\begingroup$ This link could be useful: physics.utoronto.ca/~smorris/edl/mochaware/mochaware.html There is also some kind of an explanation, even if it is really qualitative. Looks like it is a combination of hydrodynamic instability (Marangoni effect) and chemical reaction. $\endgroup$
    – valerio
    Commented Jun 22, 2016 at 6:14

2 Answers 2


I should start with the disclaimer that I don't know the answer to this, however I have seen very similar patterns in flocculating systems and I would guess that the same principles are involved.

The patterns are produced by adding a drop of pigment to a layer of slip. Both of these are colloidal suspensions. Slip is a suspension of aluminosilicate particles in water. This is normally at a slightly alkaline pH and the aluminosilicate particles carry a net negative charge. This negative charge makes the particle repel each other so the colloidal suspension is stable.

I can't find a definitive description of the pigment, but coloured pigments are usually particle suspensions of transition metal oxides and carbonates. In an acidic suspension these particles carry a net positive charge. As with the slip, the positively charged particles repel each other and this keeps the suspension stable.

If you mix the pigment and the slip several effects combine to destabilise the colloids and cause flocculation of the particles. Firstly there is the rather obvious one that the negative aluminosilicate particles and the positive transition metal oxide/carbonate particles will attract each other and bind together to produce a flocculate. Secondly the mixing of the solvents lowers the pH of the aluminosilicate slip dispersion and raises the pH of the pigment dispersion. In both cases this reduces the surface charge density and this also encourages flocculation.

Flocculated interface

So at the interface between the pigment and slip you will get a layer of flocculated dispersion. This is not in itself unstable, but if the osmotic pressures of the two dispersions are different water will flow through the floccuate layer and cause a pressure difference. This pressure difference will rupture the flocculate layer, and at the rupture point a bubble of one dispersion will flow into the other. This process then restarts and a further rupture and bubble growth occurs. The end result is a dendrite of one dispersion growing into the other.

Dendritic growth

Producing a dendrite will be critically dependent on lots of factors. We need the flocculate to form as a reasonably solid layer, and we need a big enough osmotic pressure difference to cause the interface to rupture. The viscosities of the dispersions will need to be low enough for one to flow into the other, but high enough that they don't immediately mix into a muddy mess. Predicting all this from first principles is well nigh impossible - all we can say is that slight changes in the dispersion properties are likely to produce considerable differences in the pattern formed.

Finally, there is an obvious example of a process like this in the crystal gardens that many of us will have grown in elementary school science lessons.

  • $\begingroup$ Cool, this sounds pretty plausible to me! $\endgroup$
    – N. Virgo
    Commented Jun 22, 2016 at 7:19

I am a young student of ceramics class. I have just started exploring the Mocha diffusion process. I have viewed a lot of videos of the mocha pottery being created and lot of articles about it. They all explain the process to be a chemical reaction of the acidic mocha solution on the alkaline slip. However I note that all kind of agents have been used, both acidic and alkaline. Alcohol, whiskey, acetone, mouth wash, tobacco juice all are alkaline. Soy sauce, lemon juice and other acids have been used and are acidic. So, the mocha diffusion is not likely to be due to acid-alkali reaction. The process of dispersion of the drop of solution placed on the slip appears to be a function of the surface tension of the agents. The slip is basically providing a slick wet surface with water as the substrate with a surface tension of about 72. All the agents that have been used for the mocha dispersion are low surface tension. Alcohol is 22, acetone (nail polish remover) 24, mouth wash beer wine are around 32-38. Tobacco juice around 38-43. The lower the surface tension, faster the speed of dispersion. So I think the mocha dispersion phenomenon is purely a function of surface tension and is not dependent on the acidity of the mocha solution. In the coming few days I am about to test this hypothesis.

  • $\begingroup$ I look forward to hearing about your results in testing this! If you can show that it happens when the pH of the dye solution is the same as that of the glaze, then we'll know it doesn't require an acid-base reaction. If the patterns are caused by viscous fingering then lowering the surface tension would make sense, since surface tension will tend to work against the viscosity effects. So I think your idea is quite plausible. $\endgroup$
    – N. Virgo
    Commented Feb 19, 2017 at 2:20

Not the answer you're looking for? Browse other questions tagged or ask your own question.