Optically Clear Silicone Molds (LSR): Achieving Glass-Like Transparency for Lenses

Glass-like LSR lenses fail from haze, particles, and fogging, not from “the silicone was not clear enough on the TDS.” Optical-grade platinum LSR is a low-yellowing, high-purity addition rubber. It still copies every scratch in the tool, every dust speck in the shot, and every cyclic siloxane that later condenses on a reflector.

This article is LSR injection of lenses, light guides, TIR optics, and LED overmolds. It is not poured craft RTV that looks water-clear in a cup. It is not a translucent industrial gasket. Phenyl-containing optical grades can raise refractive index; dimethyl optical LSR sits near the usual PDMS index. Pick the grade for the optical design, then build a cell that can keep it clean.

Optical LSR is a volatile and cleanliness spec first

Optical grades are formulated for high transmittance and low color. They also tend to carry more low-molecular-weight cyclics (D3-D6 and cousins) than a stripped industrial LSR. In a hot lamp or a sealed luminaire those species leave the part and fog a nearby reflector or lens. That is why low-volatility / low-fogging optical LSR plus a post-cure is the default production path, not an optional bake.

What “optical” has to mean on the RFQ:

• Named optical LSR, not “transparent 40A”
• Transmittance and haze methods (for example ASTM D1003 language) on the production thickness
• Yellowness or a colorimetric limit after heat/UV aging if the lamp is hot or outdoor
• Volatile / fogging requirement if the part sits in a sealed optic
• Platinum only. Peroxide HCR yellowing is the wrong chemistry for a lens. Tin-cure RTV is a mold rubber.

A general-purpose translucent LSR can look “clear enough” on a 2 mm plaque and still scatter in a 10 mm TIR or yellow after a 150 °C soak. Do not prototype optics in that drum and expect the production grade to match the CAD index.

The tool surface is the optical surface

LSR does not get a secondary polish like a glass blank. The cavity finish is the lens finish.

SPI A-1 / A-2 (diamond buff) is the usual conversation for high-clarity faces. VDI spark on an optical face is a designed diffuser, not a defect, and it must be intentional. Parting lines, ejector pins, and gate vestiges on a clear aperture are refractive errors. Put them on a flange, a gate tab you clip, or a non-optical wall.

Draft still helps. Optical LSR can strip with little draft, but rubbing a polished steel face dulls the tool and prints haze into later shots. A small draft on non-critical sides protects the polish. Steel selection and corrosion control matter: a rust bloom in a cooling circuit becomes pits in the optic.

Vacuum or high-quality venting belongs in the tool design. A trapped bubble in a TIR is a reject, not a cosmetic nick. Thick sections fill well in LSR compared with thermoplastic, but they still trap air if the gate and vents are guessed.

Dust is a yield problem, not a housekeeping slogan

One particle in a 3 mm lens is a dark spec at the customer. Optical LSR cells run like medical cells even when the part is a lamp: covered hoppers, filtered air, no cardboard at the press, gloves that do not shed, tools wiped as if they were surgical.

Mixers and color lines are contamination machines. Natural optical LSR plus a shop-tinted “just a little grey” is how you buy streaks. If the optic is tinted, use a compatible optical pigment at the supplier’s ratio, introduced in a closed system.

Moisture and inhibition still apply. An undercured platinum shot is sticky and hazy. Sulfur, amines, and tin-contaminated handling gear kill the face the same way they kill a mold rubber. Keep optical LSR on dedicated hoses.

First-article inspection for optics is not a caliper. It is a light box, a haze check, and a particle map. CMM the flange. Photograph the aperture.

Post-cure is how you keep it clear in service

Mold-cure makes a handleable lens. It does not finish the network or drive off cyclics. Circulating-air post-cure (grade-dependent, often in a high-temperature window used for LSR generally) is the usual lever against later fogging, odor, and some yellowing. Vacuum bake is a further lever when the part sits next to sensors or in a sealed optic.

Post-cure also moves dimensions. Optical LSR can shrink more than a filled industrial grade when the bake continues the network and loses volatiles. Cut the tool for the post-cured size, or you will fight focal length after the oven. Thickness changes the bake: a 1 mm film is not a 12 mm TIR. Follow the grade, then verify on the real section.

Skip the bake and you may still ship a pretty part. The reflector fogs in the field. That is an optical failure with a chemistry cause.

Geometry that refuses to look like glass

Uniform walls help, but LSR optics exist because freeform and thick-to-thin transitions are possible. The fights are:

• Sink and shrink in thick bosses attached to a thin lens: the optic pulls. Decouple mechanical bosses from the optical body.
• Weld lines in the aperture: two-gate fills can put a refractive seam on axis. One gate plus overflow or a well-placed weld on a non-optical rim is cleaner.
• Soft grades on large lenses: a low Shore A optic sags under its own weight and the prescription moves. Harder optical LSR exists for a reason.
• Overmold onto LEDs or PCBs: the package must survive the tool temperature and the platinum cure. Residues on the die yellow the interface.

Refractive index must match the optical file. Dimethyl optical LSR is not a drop-in for a phenyl high-index grade. Do not “clear up” a lighting design by swapping glass to silicone without re-tracing.

Measure glass-like. Do not eyeball it.

Shop-floor “looks water-clear” is not a spec. Write:

“Optical platinum LSR [grade], natural. Tool optical faces SPI [A-1/A-2] or equivalent measured polish. Production in a controlled-particle cell. Post-cure [T / t / airflow or vacuum] on the production thickness. Acceptance: transmittance and haze per [ASTM D1003 or named method] at [thickness]; yellowness after [heat/UV] as agreed; no particles in the clear aperture above [map]. Volatile/fogging requirement [named] if sealed. Dimensions apply after post-cure. Tin-cure and peroxide grades excluded.”

That block buys the grade, the polish, the bake, and the measurement. Anything less is a translucent part with a lens drawing attached.

FAQ

Can I pour optically clear platinum RTV instead of injecting optical LSR?

For a one-off prototype, sometimes. Production lenses that need repeatable prescription, low fogging, and a polished tool path are LSR injection jobs. Poured RTV picks up bubbles, cup dust, and a different volatile profile.

Why does an optical LSR part haze after a few weeks in a hot lamp?

Low-molecular-weight siloxanes left the rubber and condensed on a cooler face, or the grade was not a low-volatility optical LSR, or post-cure was skipped. Heat also yellows the wrong chemistry. Specify fogging and a bake, not only as-molded clarity.

Is SPI A-2 polish enough for a TIR optic?

Often yes for lighting; imaging optics may need a tighter face and a different conversation. The rule is that the steel finish copies into the silicone. Spark texture and parting lines in the aperture are optical features, wanted or not.

Does post-cure change the lens size enough to matter?

Yes. Continued crosslinking and volatile loss shrink the part. A tool cut to as-molded CAD will throw the focal geometry after a real bake. Qualify dimensions on post-cured parts.

Will a general translucent LSR pass as optical if I polish the tool?

Unlikely. Optical grades control color, scattering, and often volatiles. A gasket-grade translucent rubber can still haze, yellow, or fog a reflector. Polish cannot fix the compound.