



Finishing, Combining and Troubleshooting
Entry 07 / 15
12 min read
Once the glass has left the flame, the work continues: annealing so the piece lasts, finishing its surface, combining materials, and understanding what went wrong when a piece cracks, clouds or bubbles. This entry gathers that knowledge, with temperatures for borosilicate 3.3.
Annealing: making the piece last
As glass cools, the outside stiffens before the inside. When the core finally contracts, it pulls against a surface that has already set, and that tension stays locked in the glass as stress. Stressed glass can look perfect and still crack days or months later. Annealing removes it by holding the piece where the glass can relax without losing its shape, then cooling it slowly and evenly. See What Glass Is.
The temperatures that matter
| Point | Borosilicate 3.3 (approx.) | What it means |
|---|---|---|
| Working point | about 1,250 °C | Soft enough to gather, shape and join in the flame |
| Softening point | about 820 °C | The glass begins to sag under its own weight |
| Annealing point | about 560 °C (1,040 °F) | Internal stress is relieved within minutes |
| Strain point | about 510 °C (950 °F) | Below this, stress can no longer be relieved in practical time |
These are manufacturers' figures for laboratory borosilicate 3.3 glasses such as Duran and Pyrex 7740. Coloured borosilicate rods are close but not identical, so check your glass maker's recommendation. Many lampworkers set their kilns to around 565 °C (1,050 °F) for boro.
Kiln-carrying and batch annealing
The safest practice is to kiln-carry (also called garaging): keep the kiln at annealing temperature while you work, and put each piece in as soon as it leaves the flame, before it has had a chance to cool unevenly. When the session ends, the kiln runs its soak and cooling cycle on everything inside.
Batch annealing means letting finished pieces cool in a fibre blanket or vermiculite and annealing them together later. Small, solid, evenly thick boro pieces often survive this, but complex pieces, thick-to-thin joins and inclusions can crack before they reach the kiln. If you batch anneal, bring the kiln up slowly from room temperature so the cold pieces are not shocked.
An example schedule
Schedules depend on thickness, the size of the piece and your kiln, so treat this as an illustration, not a rule:
- Garage at about 565 °C while you work.
- Soak once the last piece is in: commonly 30 to 60 minutes for small jewelry-scale pieces, longer for thick ones.
- Cool slowly through the strain point, for example at a few tens of degrees per hour, to around 480 °C (900 °F). A fast cool here puts stress back in.
- Let the kiln cool naturally to room temperature with the door closed.
Don't push the soak temperature much higher to save time: closer to the softening range, fine details begin to slump.
Re-annealing, not "thermal cycling"
A correctly annealed piece does not get stronger by being heated and cooled again. Re-annealing is a repair, for pieces that cooled without a proper anneal, show stress, or have been reworked. Never put a cold, annealed piece straight back into the flame: warm it gradually in the kiln, work it, then anneal it again.
Checking stress with polarised light
Stress in glass is invisible to the eye but visible in polarised light. A polariscope is two polarising filters set at right angles over a light, so the field looks dark. Unstressed glass between them changes nothing. Stressed glass becomes doubly refracting and lets light through, so stress shows as bright or coloured bands, typically around joins, thick areas and inclusions. A well-annealed piece looks even and quiet. Two sheets of polarising film over a light make a simple version. It is not a perfect test, but it shows what your annealing is actually doing.
Finishing: cold working and sandblasting
Cold working covers every technique used to finish a piece after it has been annealed. Always anneal first: cold working a stressed piece is a reliable way to crack it.
Safety first: dust
Grinding, drilling, polishing and blasting all produce fine dust. Glass itself is amorphous, not the crystalline silica that causes silicosis, but fine glass and abrasive dust is still harmful to breathe. The core rules:
- Work wet. Run grinders, drills and rotary tools with a constant flow of water. It keeps dust out of the air and the glass cool.
- Wear a respirator rated for fine particles (FFP3 or P100) for any dry work, sandblasting included, plus safety glasses or a face shield.
- Clean wet. Wipe and mop rather than sweep, so settled dust doesn't go back into the air.
- Keep electrics safe around water: use equipment made for wet use on a protected circuit.
Sandblasting
Despite the name, glass should never be blasted with sand. Silica sand breaks down into respirable crystalline silica, which causes severe and permanent lung disease, and its use for blasting is banned or tightly restricted in many countries. Use a dedicated abrasive in an enclosed blasting cabinet with a dust collector:
| Medium | Character | Typical use |
|---|---|---|
| Aluminium oxide | Hard, sharp, reusable, the usual choice | Frosting, carving and deep etching |
| Silicon carbide | Harder and faster cutting, more expensive | Fast, deep work and hard surfaces |
| Glass bead | Round, gentle | A soft satin finish; not for carving |
- Masking: vinyl, resist tape or photoresist film protects the areas you don't want to etch, for precise, clean designs.
- Textures: adjust pressure, distance, grit size and time for anything from a subtle frost to a deep carve.
- Contrast: sandblast some areas and leave others polished, or use it to bring out details in sculpted work.
Grinding, polishing and engraving
- Diamond tools: diamond wheels, discs and bits for precise grinding, removing excess material and refining edges.
- Polishing: work through progressively finer abrasives, removing the scratches of each grit before moving to the next, then finish with a polishing compound such as cerium oxide on a felt or cork wheel to restore clarity and shine.
- Engraving: for fine details or a signature, use small diamond burrs in a rotary tool, kept wet; experiment with different tips.
- Fire polishing: a quick pass through the flame restores shine but softens detail, and the piece must be annealed again.
Graphite for smoother finishes
Finishing also begins at the torch. Graphite's smooth surface and high heat tolerance make it ideal for marvering hot glass, and custom-shaped graphite tools can create specific textures or refine shapes with great precision, which means less cold work later. A slightly cooled graphite or tungsten punty also releases more cleanly, leaving fewer marks during punty transfers.
Combining glass, metals and other materials
The real artistry in borosilicate often lies in combining techniques and materials. The rule behind all of it is compatibility: everything fused together has to expand and contract together.
Borosilicate and soft glass
Borosilicate and soft glass cannot be fused together. Borosilicate 3.3 has a coefficient of expansion (COE) of about 33 on the scale lampworkers use, while soft glasses are typically around 90 to 104: roughly three times as much. As a joined piece cools, the soft glass tries to shrink far more than the boro, and the tension cracks it, often right away and sometimes later. They also work at very different temperatures: in a flame hot enough for boro, soft glass boils, burns and loses its colour. Scientific glassblowers join them only through a graded seal, a series of intermediate glasses that step the expansion gradually. At the bench: keep rods separate, label them, and never mix scrap.
Compatibility also matters within borosilicate. Coloured boro is formulated to sit close to clear, but some colours expand a little more or less than others. Test a new pairing in a small piece, and check it in the polariscope, before using it in a finished design.
Metals in borosilicate
It is tempting to think a metal works as long as it is compatible with the glass's COE. In reality almost no metal matches borosilicate, which expands less than nearly any metal. What decides whether a metal survives is how thin it is, whether it melts or oxidises at torch temperatures, and how far its expansion differs from the glass.
| Metal | Expansion vs boro (about 3.3) | In practice |
|---|---|---|
| Fine silver | About 18–19, melts at 962 °C | Mostly used for fuming: vaporised in the flame, it gives yellows and ambers on clear glass and blues and greens over dark colours. Foil and leaf tend to dissolve or strike into colour rather than stay metallic. Thicker wire is likely to crack the glass. |
| Gold (fine or high-karat) | About 14, melts at 1,064 °C | Fumed or used as leaf for pinks and reds; heavy fuming can become metallic. |
| Tungsten | About 4.3–4.5, melts at about 3,400 °C | The closest match. Fine tungsten wire is sealed into hard glass in scientific work, ideally through a sealing glass. In the studio it is mainly a tool metal (picks, punties). |
| Platinum | About 9, melts at 1,768 °C | Stable in the flame and doesn't oxidise, but its expansion matches soft glass, not boro. Only very fine wire or foil is worth testing. |
| Copper | About 17, melts at 1,085 °C | Usually problematic in boro: wire and mesh crack the glass on cooling, and copper oxidises heavily in the flame. Very thin foil can sometimes survive because it is soft enough to give, but test first. |
Two safety points. Fuming produces metal vapour, so only fume under strong local extraction that pulls fumes away from your face. And never encase plated, coated or unknown metals: plating can flake, and coatings and alloys can release toxic fumes.
Organic and found materials
- Organic inclusions: leaves, fibres and other organic matter char and give off gas when encased in hot glass, leaving dark marks and bubbles that can weaken the piece. What can be encased is the mineral ash left once organic matter is fully burned, or a material that has been prepared and tested.
- Found materials and residue: in my own practice I combine glass with residue and found materials, which give it a different kind of permanence.
- Multiple layers: combine encasing, fuming and patterning in a single piece; each layer adds a dimension.
Encapsulating movement
- Imply motion: sculpt flowing forms such as spirals, waves or flames by stretching, twisting and rotating the hot glass.
- Layer for depth: build up clear and coloured layers to create the illusion of depth and movement.
- Optical effects: lenses, clear glass and trapped bubbles make the movement shift as you look from different angles.
Troubleshooting
Most problems come back to three things: heat, cleanliness and compatibility.
| Problem | Likely causes | What to try |
|---|---|---|
| Crack while working | Thermal shock, uneven heating | Introduce glass slowly; keep the whole piece warm |
| Crack days later | Poor annealing, mismatched materials | Kiln-carry; polariscope; test combinations |
| Grey, dark or metallic patches | Reducing flame, inner cone | Neutral or oxidising flame; burn off soot |
| Muddy colour | Overheating, flame chemistry, reactions | Less heat; check flame and pairing |
| Fizzing colour | Boiling from overheating | Work it cooler, further out |
| Bubbles | Trapped air, dirty glass | Wrap from one end; lance while hot |
| Cloudy, crusty surface | Devitrification | Clean glass; less reworking |
Cracking
When a piece cracks tells you why. During work it is usually thermal shock: glass introduced too fast, or a thin section cooling while you heat another. After finishing it usually means stress: no anneal, too short a soak, too fast a cool, or mismatched materials. Heat evenly, always anneal in a kiln, and use the polariscope to see where stress gathers.
Scorching and sooting
Scorch marks appear when the glass sits in a reducing flame, one with too much fuel and too little oxygen, or when it is held inside the inner cone. Soot settles on the surface, and metal-bearing colours, especially silver and copper ones, are reduced and turn grey, brown or metallic. Adjust the flame towards neutral or slightly oxidising (see The Flame and Reading Heat) and work just beyond the tip of the inner cone, where the flame is hottest and cleanest. Surface soot usually burns off in an oxidising flame; soot trapped under encasing stays.
Muddy colour
If colours look muddy or dull, the glass may be overheated, the flame chemistry may be wrong, or the colours may be reacting with each other. Experiment with temperature, flame and colour combinations, and see Colour in Glass for why metals behave as they do.
Boiling colours
Many coloured rods tolerate less heat than clear. Pushed too hot, they boil: the surface fizzes, tiny bubbles form and the colour can grey. Preheat colour gently and melt it further out in the flame.
Bubbles
Bubbles come from air trapped while handling the glass: when wrapping a gather, encasing, folding, or laying frit, which holds air between its grains. Dirty or damp glass adds more. Work slowly, wrap from one end so air is pushed out, and keep your glass clean. An unwanted bubble can often be heated, lanced with a pointed tool and smoothed over.
Devitrification
A cloudy, frosted or crusty surface means the glass has started to crystallise. It happens when glass spends too long in the range where crystals grow, usually through overworking, and dirt (fingerprints, dust, marker residue) gives crystals somewhere to start. Some colours, often opaques, are more prone. Clean rods with alcohol and avoid unnecessary reheating. See What Glass Is for why glass crystallises at all.
Thermal gradients
Uneven heat causes many of these problems, but used deliberately it is a tool. For sharp detail without deforming the whole piece, keep the base slightly cooler while concentrating intense heat on the area you are working. Return the whole piece to the flame regularly so the cooler part doesn't crack.
All of this knowledge only becomes useful at the bench, through repetition. The last entry, Practice: Tips, Masters and the Journey, is about how that practice grows.
Test yourself10 questions
Choose one answer for each question, then check your results at the end.
-
01What does annealing remove?
Stressed glass can look perfect and still crack days or months later.
-
02What happens below the strain point (about 510 °C)?
So pieces are held near the annealing point, then cooled slowly through the strain point.
-
03What does a polariscope show?
Two crossed polarising filters over a light: well-annealed glass looks even and quiet.
-
04Why grind and drill glass wet?
Fine glass and abrasive dust is harmful to breathe. Clean wet too: wipe and mop rather than sweep.
-
05What should glass never be blasted with?
Silica sand produces respirable crystalline silica, which causes severe lung disease.
-
06Can borosilicate and soft glass be fused together?
Scientific glassblowers join them only through a graded seal of intermediate glasses.
-
07Which metal comes closest to the expansion of borosilicate?
Tungsten (about 4.3) is the closest, but in the studio it is mostly a tool metal.
-
08What happens to leaves or fibres encased in hot glass?
Only fully burned mineral ash, or a prepared and tested material, can be encased.
-
09What usually causes grey, dark or metallic patches?
Adjust towards a neutral or slightly oxidising flame; surface soot usually burns off.
-
10Is heating and cooling a finished piece again ("thermal cycling") good practice?
A correctly annealed piece does not get stronger from extra cycles.
0 / 10 answered


























