A lens blank can come off the saw looking perfect and still be scrap. The most expensive defect in precision optics is subsurface damage optical lens makers cannot see — a layer of micro-cracks sitting below the cut surface, invisible to the eye, to bright-light inspection and to any profilometer. It survives rough grinding. It usually announces itself after polishing, when the part already carries most of its cost.
Knowing where that damage comes from during cutting is what lets you decide how much stock to leave for it. This article covers five key causes of subsurface damage optical lens production has to design around, how the damage layer is actually measured, how deep it runs in the common infrared and optical materials, and how the grinding sequence removes it. For the wider process context, our overview of the optical lens manufacturing process covers the full route from blank to finished optic.
What Subsurface Damage Optical Lens Defects Actually Are
Subsurface damage (SSD) is the zone of micro-cracks and plastically deformed material extending below a machined surface. Every abrasive process creates it — sawing, generating, lapping — and it is not the same thing as the two numbers most buyers ask about:
- Surface roughness (Ra, Sq) describes the texture you can see and measure optically. It says nothing about what is underneath.
- Figure error (PV, irregularity) describes the shape of the surface against its nominal form. Also a surface property.
- Subsurface damage is depth into the bulk. A part can polish to an excellent Ra and still carry a damaged layer that reduces strength and scatters light.
That distinction is the whole reason subsurface damage optical lens defects cost so much: every inspection tool on a normal optical shop floor measures the surface, and the defect is not on the surface.
1. Mechanical Stress from Cutting Forces
During cutting, forces exerted by the wire or tool create stress beneath the surface. In brittle optical materials these stresses generate radial and lateral micro-cracks that propagate past the nominal cut plane.
Feed rate is the variable most people get wrong. Pushing the feed shortens the cut and lengthens the damage layer, and the trade is not linear — the last 20% of cycle time you save can cost far more than that in extra grinding. On our SG-series machines the Y and Z axes resolve to 0.01 mm with ±0.01 mm repeatability and run from 0 to 1000 mm/min, which sounds generous until you work on something like silicon carbide, where a realistic production feed is around 0.2 mm/min. Brittle material does not care what the machine is capable of.
More on how cut quality is judged in practice: optical cutting quality.
2. Thermal Effects and Heat Accumulation
High-speed cutting generates localized heat, especially in materials like BK7, fused silica and germanium. Thermal gradients create tensile stress zones that leave irreversible subsurface deformation and cracking behind.
This is where a lot of published advice goes wrong. Wire tension is not a single number you set once — it follows wire diameter, and the working values are lower than most people assume:
| Endless loop wire OD | Core wire | Grit | Working tension |
|---|---|---|---|
| 0.30 mm | 0.21 mm | D54 | 60 N |
| 0.35 mm | 0.21–0.25 mm | D76 | 80 N |
| 0.45 mm | 0.28–0.30 mm | D76 | 100 N |
| 0.50–0.55 mm | 0.33 mm | D76 / D91 | 130 N |
| 0.60–0.65 mm | 0.36–0.39 mm | D91 | 140 N |
| 0.80–1.00 mm | 0.51–0.60 mm | D107 | 170 N |
The machines hold tension automatically, and in practice we run 10–20 N either side of the table value depending on material and cut depth. Wire linear speed sits in a similar band: 52 m/s maximum on the SG40 germanium wafer slicing machine, 58 m/s on the SG20, 61 m/s on the SGI 40. Anyone quoting 80 m/s for optical work is quoting a large-block stone saw, not a machine that slices lens blanks.
Speed and tension only matter if coolant is actually reaching the cut. A correctly tensioned wire running dry will do more thermal damage in one pass than a badly tensioned wire that is properly flooded.
3. Tool Wear and Abrasive Degradation
Worn wire and worn abrasive generate uneven forces and unstable cutting conditions. The result is localized indentation and more cracking beneath the surface.
The trap here is that diamond wire does not fail loudly. It degrades. Cut time creeps up by 10%, then 20%, the operator compensates by increasing feed, and the damage layer grows quietly for weeks before anyone connects it to a yield drop at polishing. As a working rule, endless loop wire gives roughly 1–3 m² of cut area depending on material hardness, and guide wheels last two to three months in normal production. Track cut area, not calendar days — consumable drift is the most common root cause of subsurface damage optical lens problems that appear with no change to the program.
4. Brittle Material Properties
Optical materials are low fracture toughness substrates. Minor perturbations during cutting induce cracks below the surface, and infrared materials like germanium and zinc selenide are the most sensitive of the group to both mechanical and thermal stress. Material selection therefore changes the allowance before it changes anything else — see our summary of optical lens materials and the published physical data in Crystran’s germanium material sheet.
5. Why It Goes Unnoticed on the Shop Floor
Visual inspection and conventional profilometers cannot detect defects below the surface, so damage that will scrap a batch passes incoming inspection without comment. Without in-process monitoring the cracks form, propagate through the grinding stages, and turn into polishing time and material loss at the most expensive point in the route. By the time a subsurface damage optical lens problem is visible, the cheapest corrective action — changing the cut — is several stages behind you.
How Is the Damage Layer Actually Measured?
This is the question that separates suppliers who have measured subsurface damage optical lens depth from suppliers who have read about it. Every method that returns a real depth number is destructive.
| Method | How it works | What you get | Destructive? | Shop-floor practicality |
|---|---|---|---|---|
| Etch and inspect | Chemical etch opens the cracks until they are optically visible | Exposed crack depth | Yes | High — the usual choice |
| Taper / angle polish | Polish at a shallow angle so depth is magnified across the face | Depth profile | Yes | Medium |
| Cross-section + microscopy | Section, mount, polish, image | Single-point ground truth | Yes | Low — slow, one point per sample |
| MRF spot / dimpling | Remove material in controlled layers, inspect each layer | Depth distribution | Yes | Low — needs dedicated equipment |
| White-light interferometry, confocal, AFM | Optical surface metrology | Surface only — no SSD depth | No | High, but answers a different question |
There is no non-destructive way to measure subsurface damage in a finished optic. Every method that returns a real depth figure destroys the part it measured. If a supplier quotes an SSD depth and calls the measurement non-destructive, ask what was actually measured — it was the surface.
What about OCT, ultrasonic C-scan and X-ray CT? They appear in a lot of SSD articles, and they are genuinely useful techniques — in thick composites, in weldments, in packaged assemblies. On a 5 mm optical blank they are the wrong instrument: their spatial resolution is one to two orders of magnitude coarser than the damage layer you are trying to find, so a clean scan proves nothing. We have not found a production case where they replaced an etch test. If someone has, we would like to see the data.
How Deep Does It Go in Germanium, ZnSe and Sapphire?
We do not publish absolute depth figures, for a reason given in the next section. What is reliable is the ordering — how the common materials rank against each other under the same cut:
| Material | Fracture behaviour | Relative damage depth | What actually bites you |
|---|---|---|---|
| Optical glass (BK7 and similar) | Isotropic, most forgiving | Lowest | Predictable; the easy baseline |
| Fused silica | Isotropic, harder | Low to moderate | Wheel loading, thermal drift |
| Germanium | Cleaves on preferred planes | Moderate to high | Cleavage steps rather than random cracking |
| ZnSe / ZnS | Soft, low toughness, polycrystalline | High | Grain pull-out and edge roll-off, not just cracks |
| Sapphire | Hardest, strongly anisotropic | Orientation dependent | Same cut, different result on C, A and R planes |
The allowance you need follows that ordering, and kerf compounds it: the thinnest endless loop wire we supply is 0.35 mm and the resulting kerf is about 0.4 mm, so material consumed per slice is target thickness plus roughly 0.4 mm before any damage allowance. The full accounting is in our note on kerf loss in optical cutting.
Removing It: the Grit Ladder
Each stage must remove more material than the damage depth left by the stage before it. Break that rule anywhere in the chain and the damage is not removed — it is carried forward, thinned down, and revealed at final polish, when the part is worth the most.
In practice that means rough grinding has to eat the saw marks and the cracked layer underneath them, which is deeper than the visible marks. Fine grinding then removes the rough-grind damage, and polishing only has to deal with what fine grinding left. Polishing is the wrong tool for damage removal — the removal rate is too low, and by the time you have polished deep enough to clear a rough-grind crack you have destroyed the figure.
This is also why the finishing target and the cutting parameters are one decision, not two. Our discussion of achieving a fine germanium lens surface finish starts from the same premise. Controlling subsurface damage optical lens defects is a cutting and grinding problem that happens to be discovered at the polishing bench.
A Note on the Numbers You Will See Quoted Elsewhere
Absolute SSD depth figures circulate widely in equipment marketing, usually with no stated measurement method, sample size or material grade. We have chosen not to repeat them. Damage depth depends on the specific ingot, the wire condition, the feed and the fixture — a figure measured on someone else’s germanium tells you very little about yours.
What we do instead: send us your material and we will cut it. The trial is free across the cutting and grinding range, your material comes back along with the process sheet we used, and the damage assessment is made on your parts rather than on a brochure. That is the only honest way to put a number on subsurface damage optical lens work for a specific job.
Effects on Optical Performance
- Light scattering from the damaged layer reduces modulation transfer function (MTF).
- Mechanical strength drops, raising failure risk in thin windows and domes.
- Excessive polishing is needed to clear hidden defects, which raises cost and threatens figure.
- Long-term reliability suffers under thermal cycling and mechanical load, as buried cracks propagate.
Mitigation Strategies
To minimize subsurface damage optical lens defects in production:
- Set feed, wire speed and tension as a set, not individually. Use the tension table above as the starting point for the wire diameter in use.
- Track wire consumption by cut area, not by shift. Replace before cut time drifts, not after.
- Stage the removal. Coarse cut, rough grind, fine grind — each stage sized against the damage left by the previous one.
- Flood the cut. Coolant delivery failures cause more thermal damage than aggressive parameters do.
- Verify destructively on a sample. One etched sample per material and parameter set is worth more than any amount of surface metrology.
Surface imperfection tolerances and how they are called out on the drawing are defined in ISO 10110-7, the right reference to cite in a purchase specification — note that it governs surface imperfections, so subsurface requirements still have to be stated separately in the contract.
Where This Matters Most
The industries where subsurface damage optical lens control decides the yield are the ones working hard, brittle and expensive materials:
- Infrared imaging and thermal optics in germanium, ZnSe and ZnS
- Defence and aerospace windows and domes
- Sapphire windows, substrates and watch crystals
- Quartz and fused silica optics for high-power and UV use
- Optical components for semiconductor inspection and metrology
These are optical fabrication applications — not ophthalmic Rx laboratory work and not architectural glass. Different materials, different tolerances, different machines.
Practical Next Steps
If you are chasing a subsurface damage optical lens yield problem that shows up at polishing, work backwards in this order: etch one sample from the batch to confirm the damage depth, check wire condition and cut area since the last change, then check whether rough grinding is removing enough stock to clear it. In our experience the answer is in the first two steps more often than the third.
If you are specifying a new line, decide the cutting allowance and the grinding sequence together — our walkthrough of the germanium blank to lens process shows how the stages interlock, and the infrared optics manufacturing equipment overview covers the machines for each stage.
FAQ
What is subsurface damage in optical lens cutting?
Micro-cracks and plastic deformation beneath the visible surface, caused by cutting forces, thermal stress and the low fracture toughness of optical materials. It is a depth into the bulk, not a surface texture, which is why subsurface damage optical lens defects pass normal inspection.
Can subsurface damage be detected visually?
No. It requires a destructive method — chemical etching followed by microscopy is the usual choice, with taper polishing or cross-sectioning where a depth profile is needed. Optical surface metrology such as white-light interferometry measures the surface only.
Can polishing alone remove subsurface damage?
No. Polishing removal rates are far too low to clear a saw- or rough-grind-induced damage layer, and attempting it destroys the figure. Damage has to be removed by grinding; polishing only handles what grinding leaves.
How much stock should I leave for it?
Enough that each stage removes more than the previous stage’s damage depth, on top of kerf and any bow from the cut. Since the depth depends on your material and your parameters, the workable approach is a trial cut on your own material rather than a table value.
How does subsurface damage affect lens performance?
It increases scatter and reduces MTF, lowers mechanical strength, raises polishing cost, and shortens service life under thermal or mechanical cycling.
How can manufacturers minimize it?
Match feed, wire speed and tension to the material; replace wire on cut area; stage the removal so no damage is carried forward; keep the cut flooded; and verify destructively on samples. Controlling subsurface damage optical lens defects is a process-discipline problem more than an equipment problem.
