Role of flux compounds like bismuth in lowering firing temperature and improving adhesion

THE PROBLEM WITH BISMUTH

(Why one percent of the mix took ninety-nine percent of the time)

Once I’d accepted that lustres live or die by solubility, bismuth quickly became the next unavoidable step.

In a lustre firing you’re nowhere near the temperatures needed to properly melt or even soften most glaze surfaces. If you simply paint gold onto a fired glaze and heat it gently, it will sit there politely and then wipe straight off again. No bonding. No permanence.

That’s the role of the flux.

By introducing a low-melting flux into the lustre layer, you locally soften the glaze surface just enough for the metal to bond during the firing. Not a full melt. Just a brief, controlled softening at the interface. Without that, the lustre has nothing to grab onto.

For low-temperature lustres, the practical flux options are limited. Lead works beautifully, but for obvious reasons it’s a non-starter. Boron had a short trial, but as a glass former in its own right it brought too many side effects.

That left bismuth.

On paper, bismuth is ideal. Low melting behaviour. Long history in lustres. Widely referenced in both historical texts and modern patents. In practice, getting it into a usable form was where the real suffering began.

WHY BISMUTH IS AWKWARD

The problem with bismuth isn’t what it does in the kiln. It’s getting it there in the first place.

Most readily available bismuth compounds are completely unsuited to lustres. Oxides, subnitrates, carbonates, hydroxides. All perfectly respectable ceramic materials but all hopelessly insoluble

What a lustre needs is a bismuth compound that:

• dissolves cleanly in organic solvents

• forms a uniform, brushable film

• mixes at a molecular level with the rest of the system

• survives storage without turning into sludge

I went through the usual dead ends. Subnitrates that dissolved in acids but refused to behave afterwards. Nitrates that looked promising and then turn into useless fluff. Metal–resin soaps that felt like they should work but stubbornly didn’t.

Each one taught me something. Mostly what i couldnt use.

THE FORM THAT ACTUALLY WORKS

Modern lustre patents and SDS sheets consistently point to one family of compounds: metal carboxylates, and for bismuth in particular, bismuth 2-ethylhexanoate (CAS 67874-82-2). Usually called bismuth octoate.

This form of bismuth is oil-soluble, compatible with modern lustre mixes, and capable of forming stable, even films. There’s a reason it shows up again and again in commercial formulations.

Buying it in sensible quantities, however, turned out to be a fantasy. Industrial suppliers were happy to sell it by the drum. Nobody wanted to know me when I asked for a few grams.

So my only route was to make it.

Making Bismuth Octoate (The Practical Version).

On paper, there are lots of ways to make bismuth octoate. Textbooks & patents make it look straightforward, almost routine.

In practice, that optimism doesn’t survive contact with the real world. I spent months trying nearly every route I could find and can genuinely say nothing seemed to come close.

What follows isn’t the most elegant reaction, and it certainly isn’t one i found from a textbook or guide. It’s the only route I found that actually worked and could be repeated.

This method starts with bismuth subnitrate. Other starting materials can be used, but they added extra steps and extra failure modes and usually a much dirtier end product. Subnitrate, i found was simply the least bad place to begin.

During this phase of the work, one word kept creeping into my notes, my thoughts, and my nightmares:

HYDROLYSIS

This is where bismuth does what bismuth loves to do.

It grabs water, polymerises itself, drops out of solution, and turns back into a useless solid. Once that happens, its effectively dead. No amount of stirring, heating, or positive thinking brings it back into play.

Hydrolysis isn’t a subtle failure mode. If the pH drifts, hydrolysis. Too much water in the mix, hydrolysis. Rapid temperature changes, hydrolysis. Blink at it the wrong way and you’re filtering another insoluble disappointment out of the flask.

To put some scale on this: this stage of the project consumed roughly 500 g of bismuth subnitrate, 250 g of bismuth trioxide, and 100 g of bismuth carbonate. Most individual experiments used about 3g at a time. On paper, many of those routes should have worked. On the bench, almost none of them did.

In theory, there are more elegant approaches. A trained chemist would likely reach for controlled metathesis, carefully managing pH, stoichiometry, and water activity to exchange ligands cleanly. I tried that route. Repeatedly. In a small bench setup it proved brutally unforgiving. The system was fragile, and bismuth punished every minor deviation with immediate hydrolysis.

What eventually became clear is that bismuth doesn’t reward elegance.

So instead of elegance, I chose persistence. I stopped trying to persuade the bismuth and just started to bully it.

That’s where glacial acetic acid earned its place. It sits in a Goldilocks pH zone for bismuth, acidic enough to keep it soluble, but not so aggressive that it creates new problems. More importantly, it allows water to be removed gradually by evaporation. As the reaction proceeds and the system dries out, hydrolysis becomes harder and harder for the bismuth to fall back into.

Through repetition rather than theory, 80 °C emerged as the sweet spot. Hot enough to maintain momentum and drive off water. Gentle enough to avoid destabilising the system. Not refined chemistry. Just chemistry that works.

This was the pattern repeated throughout this project. Practical experiments came first. Understanding followed. And somewhere in that loop, the solution finally appeared.

STEP-BY-STEP PROCESS

1. Dissolve the bismuth

Add 5g of bismuth subnitrate (CAS 1304-85-4) to a 100ml borosilicate beaker.
Add 40mL of glacial acetic acid (CAS 64-19-7).

Heat the mixture to 110 °C and hold it there for around one hour with a watchglass to slow evaporation, stirring continuously. The solution will gradually turn clear or faintly yellow.

Once the solution is completely clear, reduce the temperature to 80 °C.

2. Introduce the fatty acid

Over the course of one hour, slowly add 10–12 mL of 2-ethylhexanoic acid (CAS 149-57-5). Add it gradually. Rushing this step drops the temperature and causes the bismuth to fall out of solution. keep the beaker half covered with a watch glass whilst adding

3. Bully the bismuth into submission

Maintain the mixture at 80 °C and leave it uncovered and running overnight.

During this long, slow cook, the acetic acid steadily evaporates. Any water produced during the reaction leaves with it. As the volume shrinks and the environment dries out, the bismuth has fewer and fewer options left. Eventually, reacting with the 2-ethylhexanoic acid becomes the only way forward.

This slow evaporation does two critical things:

• It removes water, allowing the reaction to proceed rather than stall or reverse

• It forces completion by steadily concentrating the solution removing the acetic acid and giving busmuth no choice but to combine with the 2-ethylhexanoic acid

Think of it less as fine chemistry and more as leaning on the system until it gives up.

4. Recognising completion

The reaction is complete when you’re left with a thick, pale yellow, viscous oil and there is no detectable acetic acid smell. If you’re unsure, add a small additional amount of fresh 2-ethylhexanoic acid and continue heating for a few more hours.

Your nose is a surprisingly reliable tool here. Vinegar notes mean you’re not done.

5. Yield check

Allow the mixture to cool slightly, then weigh it.

The easiest way to do this is to tare the scales with an empty 100 mL borosilicate beaker (the same type used for the reaction), then place the octoate beaker on the scales.

From 5g of bismuth subnitrate, the theoretical yield of bismuth octoate is around 11g, in practise i was never far away.

6. Dilution and filtration

Once the mass is known, dilute the product with toluene to create a 15 % w/w bismuth octoate stock solution. So for 11g you add 62g of toluene

Then filter the solution through a 0.47µm syringe filter or filter paper.

(A glass syringe is essential here. Toluene and plastic syringes do not get on)

The result is a clear, pale yellow solution suitable for direct use in lustre

A NOTE ON STABILITY

At this stage, the bismuth octoate will already function perfectly well as a lustre flux. No further processing is required for it to work, and if you make it as described, it will fire correctly and do the job it’s meant to do.

What it won’t do particularly well, without further intervention, is sit on a shelf.

Over time, unstabilised bismuth octoate will haze, clump, and eventually drop out of solution. The chemistry in the kiln still behaves, but dosing becomes less reliable and consistency starts to drift. Anyone who has found a suspicious sludge at the bottom of an old lustre bottle has already met this problem.

After many months of focused work, I did eventually solve this stability issue. So I now have a way to dramatically improve the shelf life of bismuth octoate, and by extension the long-term reliability of the finished lustre itself.

However, for now, I’m choosing not to include that step here. It improves storage and repeatability, not the underlying firing chemistry, and this guide will still get you to a fully functional, working lustre without it. The stabilisation details may make their way into the guide later, once the pain of that particular rabbit hole has faded.

HOW MUCH BISMUTH IS ENOUGH

One of the more important realisations came late in testing: the amount of bismuth added directly controls the temperature the lustre needs to reach in the kiln.

Bismuth lowers the effective softening point of the glaze surface. More bismuth means bonding occurs at a lower peak temperature. Less bismuth pushes that bonding window higher.

All of my testing was carried out on cone 6 glazed ware, with lustre firings aimed at around 840 °C. That temperature became the fixed point. From there, the bismuth level was adjusted until adhesion was reliable without pushing the firing hotter than necessary.

There was a second constraint as well.

BISMUTH AFFECTS COLOUR

At higher additions, bismuth tends to change the tone of the gold, often introducing a subtle purplish or violet cast. This effect is well documented in lustres and becomes more pronounced as the bismuth content rises. It isn’t a flaw, and in some contexts it’s actively desirable.

For my work, it wasn’t.

I was chasing the cleanest, palest, truest gold I could achieve. That meant using the lowest bismuth addition that still gave reliable bonding at my chosen firing temperature. Any more than that and the colour began to drift.

WHY THIS PAGE EXISTS

Bismuth looks like a footnote on paper. In reality, it was the single most time-consuming part of the entire system.

Not because it’s complex in theory, but because its moody and unreliable on the bench.

Once it was finally understood and brought under control, everything else was tackled with confidence