How solvent choice affects flow, drying, and application in lustre glaze work

Before you ever see the glimmer of gold on a fired piece, there’s a quiet liquid workforce doing most of the real labour. Solvents don’t get the glory of a shimmering metallic finish, but without them everything falls apart, often quite literally. They’re the difference between a level, obedient film and an uneven, streaky, dust-covered mess.

Early on, I thought solvents were just thinners. Something to dissolve the ingredients and make the mixture brushable. That assumption cost me a lot of failed tests. In a lustre system, solvents are active participants. They control timing, flow, stress, and ultimately whether the chemistry survives long enough to reach the kiln intact.

WHAT SOLVENTS ACTUALLY DO

In practical terms, solvents are responsible for four things:

• dissolving otherwise awkward or stubborn chemistry

• controlling how fast the film dries

• shaping how the lustre flows and levels under the brush

• managing stress as the film shrinks during drying

Get those wrong and no amount of clever chemistry elsewhere will save you.

When you apply a lustre, you’re laying down a very thin liquid film made of dissolved solids. As the solvent evaporates, that film has to level, stop moving before it slumps, avoid cracking as it shrinks, and stop being sticky quickly enough not to become a dust magnet. Solvents decide how gracefully all of that happens.

FAST, MID AND LONG TAIL BEHAVOUR

Rather than thinking about individual solvents, it’s more useful to think in terms of evaporation roles.

Fast-flash solvents evaporate quickly. They help the lustre “set” shortly after brushing, reducing runs and dribbles, especially on vertical surfaces. Push them too far and the surface skins over before it has levelled, locking in brush marks and uneven thickness.

Mid-range solvents are the levellers. They give the film time to relax and smooth itself out. Most of the visual quality of a dried lustre film lives here. Too little and the film looks rushed. Too much and it stays mobile longer than you’d like.

Long-tail solvents evaporate slowly. They linger after everything else has left, allowing the polymer matrix to gently reorganise as it dries. Used well, they relieve stress. Overused, they turn the surface into a dust magnet.

No single solvent does all three jobs well. That’s why blends exist.

WHY BLENDS BEAT A SINGLE SOLVENT

I tried single-solvent systems early on. It’s an obvious instinct: fewer variables, simpler chemistry.

It doesn’t work.

A solvent that levels beautifully often dries too slowly. One that flashes quickly enough to control runs usually locks in defects.

Historical lustres leaned heavily on solvent blends for a reason. Turpentine, eucalyptus oil, pine oil, these weren’t chosen for romance. Potters were compensating, through experience, for exactly the same stresses we still deal with today. Modern solvents just let us do that in a much more refined way using a huge family of solvents.

SOLVENTS AT THE BENCH

Some solvents earn their keep long before anything goes near a pot.

Chloroform (CAS 67-66-3) : the gold collector.

When forming gold mercaptides, chloroform is invaluable. Gold species dissolve into it readily and, crucially, it doesn’t mix with water. That clean phase separation makes it easy to isolate the gold compound with minimal loss.

Once the gold has transferred into the chloroform layer, you can repeatedly wash it with fresh water to strip away leftover dimethyl sulphide, excess mercaptan, and other by-products. They migrate happily into the water, leaving the gold behind.

It’s one of those rare moments where the chemistry feels cooperative rather than hostile.

Methanol (CAS 67-56-1): the cleaner.

Once the gold mercaptide is dissolved in chloroform, methanol takes over. Methanol dissolves chloroform but not the gold mercaptide, so adding it forces the gold compound out of solution while carrying the chloroform and remaining impurities away.

On paper this is called liquid–liquid extraction followed by precipitation. On the bench, it’s simply the least painful way I’ve found to collect gold with minimal losses.

Solvents in the lustre itself

Once the chemistry is built, solvents take on a different role: controlling how the lustre behaves during application and drying.

Toluene (CAS 108-88-3): the flashy performer.

Toluene shows up mainly for bismuth, gold and rhodium compounds. It dissolves them quickly and cleanly, then flashes off rapidly once applied. In small amounts this is extremely useful. It helps the lustre thicken almost immediately after brushing, ensuring it stays where the brush left it.

Cyclohexanone (CAS 108-94-1): the almost-perfect answer.

Cyclohexanone sits neatly in the middle of the evaporation range. It dries slowly enough to allow the film to level properly, but quickly enough to avoid prolonged tackiness and excessive dust pickup. For a long time, it looked like the ideal backbone solvent. On paper, and in patents, it’s hard to fault. If you ask Johnson Matthey, it’s very much a go-to solvent for lustre systems.

Early tests backed that up. Films laid down well, levelled beautifully, and behaved exactly as you’d hope during application.

The problems only appeared later, once the system became more complete. As soon as bismuth octoate entered the mix, cyclohexanone started causing trouble. Almost immediately the bismuth would aggregate, thicken, and eventually drop out of solution. Sometimes this took seconds, sometimes hours, but it was consistent and repeatable. Fresh batches that looked perfectly fine would quietly turn cloudy or clumpy far sooner than they should have.

What made this especially frustrating was that nothing looked wrong at first. Application was fine. But after hours or days, that tell-tale haze would appear in the lustre solution, or worse, a gel-like sludge would form at the bottom of the vial. A quiet sign that the bismuth had decided to evacuate the mix without leaving a forwarding address.

The lesson wasn’t that cyclohexanone is a bad solvent. It was that solvents don’t operate in isolation. A solvent can behave impeccably on its own and still destabilise the wider chemistry once polymers, metals, and fluxes are all present.

From that point on, solvents stopped being something I chose on paper specifications alone, i had to ensure they were compatable with each of the other components too.

Isophorone (CAS 78-59-1): the Backbone That Actually Behaved.

Isophorone entered the picture as a direct response to the cyclohexanone headaches.

On paper, it made sense. It sits in the same broad solvent family, similar polarity, similar solvency for polymers and metal organics, but with a significantly slower evaporation rate. My thinking at the time was simple and, in hindsight, only partially right: if cyclohexanone was too abrupt and was upsetting the bismuth, then something slower and more forgiving might keep the system calm for longer.

And, crucially, it did play nicer with bismuth octoate.

Back then, before I’d solved the long-term stability of the bismuth itself, a lot of my effort went into trying to keep it happy through solvent choice alone. Isophorone caused far fewer immediate problems. The solution stayed clear for longer. The bismuth didn’t clump or gel as aggressively. Compared to cyclohexanone, it felt like progress.

Its slow evaporation rate also seemed promising for another reason. PMIB films were cracking during drying, and I hoped that a slower solvent would allow the film more time to relax and shed stress before locking up. In practice, it helped a little, but it wasn’t the real fix. Rosin turned out to be the correct answer there, not solvent.

In other words, isophorone helped with some symptoms, but not the underlying causes.

With hindsight, that makes sense. Once the bismuth itself was properly stabilised, and once the PMIB film behaviour was modified with rosin, isophorone stopped being essential. It remains a solid, well-behaved backbone solvent, but it’s no longer doing the heavy lifting it once was.

At this point, with a more robust system overall, I could probably revisit other solvent options that were previously off the table. Isophorone earned its place by getting me through a fragile phase of development, not because it’s the only solvent that can work.

That distinction matters.

The Saviour: Acetic Acid (CAS 64-19-7).

Acetic acid deserves a proper mention here.

You might not instinctively think of it as a solvent, at least not in the classical “neat, well-behaved lab solvent” sense, but it absolutely is one. And during what I now think of as the glorious failures of the bismuth years, it turned out to be indispensable.

In many ways, if it hadn’t been for acetic acid and a fairly unhealthy level of stubbornness on my part, this entire journey would have ground to a halt.

Acetic acid sits in a strange middle ground. It dissolves things that refuse to cooperate elsewhere, tolerates heat, mixes with water when needed, and crucially, it doesn’t immediately provoke bismuth into throwing itself out of solution. That last point mattered more than I realised at the time.

It wasn’t just acting as a passive medium either. In the bismuth work, acetic acid quietly did several jobs at once: dissolving starting materials, buffering conditions into a survivable range, and allowing water to be driven off slowly without collapsing the chemistry. It gave me just enough control to keep pushing forward when almost every other route failed.

It also announced its presence loudly and repeatedly to my nose, but at that stage I was willing to accept almost anything that didn’t end with another beaker full of dead white sludge.

Acetic acid isn’t elegant. It doesn’t stay politely in the background. But when things were falling apart everywhere else, it kept the door open long enough for the rest of the system to finally come together.

IN SUMMARY

Solvents may not sparkle, but they’re indispensable:

• chloroform isolates gold cleanly with minimal loss

• methanol removes chloroform and purifies the gold compound

• toluene dissolves stubborn metals and helps the film set quickly

• isophorone allows smooth, even film formation but must be balanced

• acetic acid converts, solvates, and occasionally commits chemical sabotage

Over the course of this project I accumulated an absurd collection of solvents. From the outside it looks excessive. From the bench, it became obvious why. For almost every reaction or behaviour, there is a solvent that makes life easier and another that quietly ruins your day.

Learning which was which wasn’t optional.