A furnace sight glass is one of the hardest optical jobs in industry. The window looks at molten metal through a wall of radiant heat, takes spatter from the pour, and is cycled from a cold start to full temperature every shift, which is a harder environment than most laboratory optics ever see. Most of the failures reported on foundry lines are not optical in the usual sense: the glass has not been scratched, it has crazed, devitrified, cracked or fogged from the outside. This note walks through the failure modes that actually occur and what to change in the specification or the use.
Optical Quartz and Crystal Glass in Foundry and Molten Metal Inspection
On a foundry line the optical head is either a fixed sight glass on the furnace wall or a portable device carried to the pour lip, and in both cases the element in front of the beam is a quartz or crystal glass window chosen because it transmits in the near-infrared as well as the visible and because it holds its structure at temperature. Fused silica in particular is selected for its low expansion coefficient and its ability to take thermal shock that would craze ordinary glass. The window is usually thicker than an optical window would otherwise be, because it is being used as a pressure and heat barrier as much as an optical element. What makes this application distinctive is that the dominant degradation is cumulative and invisible until it is severe: the surface that faces the melt changes over time, and the degree of change is what decides how long the sight glass lasts before the operator can no longer read the metal level or the meniscus.
Inspection and measurement
Checking a sight glass in this environment is not a bench measurement. Optical transmission and surface quality are conventionally assessed using the general optical drawing and test conventions, but the values that matter on a furnace are better judged on the installed part at working temperature than in a laboratory at room temperature. The practical approach is to record the point at which operators stop trusting the view, to treat that as the actual service end rather than waiting for a visible crack, because a crazed glass that still transmits is already giving a degraded image. Where investment in process visibility is justified, the useful measurement is the change in the transmitted signal over the service interval at the wavelengths the head works at, which turns replacement scheduling from an opinion into a number. It is also worth checking the window at cold start, because the thermal shock that cracks a window occurs in the first minutes of a shift, not during steady operation.
Coating and deposition considerations
Coatings on a furnace sight glass are usually applied for the opposite reason to a normal optical coating: they are meant to keep the hot face clean or to reject part of the radiation, not to improve the image. A thin film that reflects across the near-infrared can lower the heat load reaching the optics behind the window and extend the life of the camera, and a hydrophobic or release-type top layer can make spatter and slag less likely to bond. Both need to be treated cautiously. Any layer on the hot face is itself exposed to the full thermal cycling, and a coating with a expansion mismatch to the substrate will craze or delaminate in a way that bare glass would not. A coating intended to repel molten metal is a consumable decision, and it is worth confirming the durability figures from the coating supplier against cycling rather than a single hot test, because the number of cycles between the cold state and full temperature is what the part actually experiences.
Design rules that reduce cost
The first money saver is thickness. A sight glass is often over-specified in thickness on the grounds that it is a pressure boundary, and the extra thickness adds material, weight and the thermal mass that makes the cold-start shock worse. The second is to separate the functions: a replaceable, cheap protective sleeve in front of a more expensive finished window means the expensive element is not the one taking the thermal cycling. The third is to use a stepped or refracting geometry where the line of sight allows it, so that part of the beam path is through air rather than through glass, which reduces both the heat load and the amount of glass exposed. The fourth is to fix a service interval and a recorded end-of-life criterion rather than replacing on complaint, since an emergency replacement during a pour is far more expensive than a planned swap. Standard catalogue discs are almost always cheaper than cut-to-size custom shapes for this duty.
Common failure modes and how they show up
Five modes account for most foundry sight glass problems. Thermal shock cracking appears in the first minutes of a shift as a radial crack, usually at the edge where the glass is constrained by its holder, and the fix is a low-expansion material with a relieved edge and a controlled warm-up rather than a different thickness. Devitrification shows up as a milky or cloudy surface on the hot face after extended service at high temperature, and it is irreversible, so the response is to shorten the service interval or to move the window away from the hottest zone rather than to polish it. Coating delamination on the hot face appears as flaking or as a sudden change in the image, and it means the coating was not qualified for cycling. Edge chipping from spatter or from the holder torque appears as a local fracture, and it is usually a mechanical fitting problem. Finally, deposition and oxidation on the outer face gradually darkens the view, which is the most benign failure and the one most often mistaken for devitrification when the glass has in fact only got dirty.
Requirements specific to Foundry and Molten Metal Inspection
A foundry sight glass should be specified as a heat and pressure boundary first and an optical element second. Choose a low-expansion quartz or crystal material, keep the thickness to what the pressure duty actually requires, and relieve the edge so the holder does not constrain the glass during a cold start. Treat any hot-face coating as a consumable with a stated cycle count, and set a recorded service interval based on the point at which operators stop trusting the view rather than on visible damage. Separating a cheap replaceable sleeve from the finished window is usually the largest single saving.
- Low-expansion quartz or crystal, thickness matched to pressure duty only
- Edge relieved so the holder does not restrain the glass on cold start
- Hot-face coatings treated as consumables with a stated cycle count
- Replaceable cheap sleeve in front of the finished element
- Service interval recorded against loss of image trust, not visible damage
Framework references: surface quality, dimensional tolerances and test methods for optical elements are conventionally stated in ISO 10110, and the current issue should be consulted rather than assumed. Statements about temperature limits and cycling behaviour reflect general material properties of fused silica and crystal glasses and are qualitative; confirm the specific grades with the material supplier before fixing a service interval.
Selection data at a glance
| Parameter | Typical value | Notes |
|---|---|---|
| Failure mode | How it appears | Usual fix |
| Thermal shock crack | Radial crack at first minutes | Low-expansion material, relieved edge, warm-up |
| Devitrification | Milky hot face over time | Shorten interval, move off hottest zone |
| Coating delamination | Flaking, sudden image change | Qualify coating for cycle count |
| Edge chipping | Local fracture from spatter | Review holder torque and fit |
| Outer face deposition | Gradual darkening | Scheduled clean or sleeve |
| Thickness over-spec | Higher cost and worse shock | Match to pressure duty only |
Frequently asked questions
Why does the sight glass crack on the first minutes of a shift rather than at the hot end of the day?
Because the thermal shock is largest at the cold start, when the surface of the glass heats far faster than its interior. A window that survives a steady pour can still fail in the first minutes if its edge is constrained by the holder, so the fix is edge relief, a warm-up period and a low-expansion material rather than more thickness.
Can devitrified quartz be polished back into service?
Generally no. Devitrification is a structural change in the surface of the material with a permanent milky appearance, so the practical responses are a shorter service interval or relocating the window away from the hottest zone, not refinishing.
Is a coating on the hot face a good idea?
It can be, for radiation rejection or release, but only if the supplier has qualified it for repeated thermal cycling. A coating chosen for a single hot test tends to flake on the first cold start, which turns an optical improvement into a debris problem inside the housing.
Need a quotation for optical quartz and crystal glass?
Send your drawing, target specification or coating requirement and an optical engineer will reply within 24 hours with a price and a lead time.
Request a Quote Email an engineer