Haze Is a Number Before It Is a Complaint
A buyer holds a nipple against a window and calls it cloudy. A laboratory reports 8.4 percent haze on a 2 mm plaque. Both describe transmission loss and only one of them belongs in a purchase specification. ASTM D1003 defines haze as the share of transmitted light that deviates from the incident beam by more than 2.5 degrees, and that definition fixes the geometry of the measurement.
A finished nipple cannot be measured directly with much confidence. Curvature, wall thickness variation and the radius at the tip bend light before it reaches the detector. Anysil measures a 2 mm compression molded plaque cast from the same lot as the production run and reports that value beside the part number of the high-transparency nipple grade. When a buyer's incoming inspection reads a molded nipple on a bench instrument, the two figures will differ, and the difference is geometry rather than material.
Source One: Silica Loading and Catalyst Ratio
Transparent LSR depends on fumed silica with a controlled surface area and a platinum to SiH ratio held inside a narrow band. Raising silica loading improves tear strength and costs transparency. The trade is settled at the formulation stage, and a molder receiving compound outside the band cannot recover the loss at the press.
Two observations from the production floor. An under-cured part looks clearest immediately after demolding and turns hazier over the next 48 hours as the network continues to build, so a first-article check taken at the press passes and the following morning's audit fails. Anysil holds transparent parts 24 hours before the transparency check for that reason. Separately, a platinum ratio drifting low leaves unreacted SiH that fogs the surface during post-cure. The drift sits inside the supplier's own certificate of analysis and appears as a lot-to-lot haze band of 1.5 to 2 percentage points, not as a single bad lot.
Source Two: Mold Surface and Venting
Optical inserts are polished to SPI A-2 or better, and polish is half the story. Vent depth matters more than most buyers expect. A vent at 0.004 mm passes air and blocks material. A vent at 0.008 mm passes both, and the hairline it leaves near the gate scatters light. Anysil holds optical inserts at 0.003 to 0.005 mm and gauges vent depth at every insert service, because polishing removes steel and a re-polished insert ends up with vents deeper than the drawing allows.
Mold deposit is the second mechanism. Every few thousand shots a film builds on the cavity from volatiles in the compound, and the film transfers a matte texture to the part. On a 64-cavity nipple tool running a 16 second cycle, deposit becomes visible in transmitted light somewhere between 15,000 and 20,000 shots. That interval is an observation rather than a specification, and it shifts with the compound lot. Cleaning on a fixed calendar produces the slow haze drift that nobody traces back to the tool.
Source Three: Post-Cure and Volatile Bloom
A post-cure drives off low molecular weight siloxane and residual volatiles. It also oxidizes the surface when the oven air is dirty or the temperature overshoots. A 200 °C post-cure with a 5 °C overshoot on a 1.5 mm section can leave a faint surface haze that polishing will not remove, because the change is chemical rather than physical.
The oven matters as much as the profile. Air moving through a galvanized duct picks up zinc and oil mist, and both land on the part. Anysil runs transparent work in a filtered forced-air oven with a stainless chamber, loads a single layer on stainless trays, and keeps polyethylene bags out of the chamber.
Fault Matrix for a Haze Complaint
Most haze complaints resolve within one shift once the pattern is read correctly. The spread of the defect across cavities carries more information than its severity.
|
Symptom |
Likely cause |
How to confirm |
Corrective action |
|
Even haze across all cavities |
Silica loading or catalyst ratio outside the window |
Plaque haze on the retained sample from the same lot |
Re-blend with retained lot, verify ratio at incoming inspection |
|
Haze concentrated near the gate |
Vent depth over spec after re-polish |
Vent depth gauge at the gate insert |
Re-cut vents to 0.003 to 0.005 mm |
|
Haze appears 24 to 48 hours after molding |
Under-cure, network still building |
Read haze at 2 hours and again at 48 hours |
Add 1 to 2 seconds to the cycle or raise mold temperature by 3 to 5 °C |
|
Matte patch on one side of a cavity |
Mold deposit |
Inspect the cavity under oblique light |
Clean the insert and shorten the cleaning interval |
|
Faint bloom after post-cure |
Oven contamination or temperature overshoot |
Compare oven inlet air with part surface, read the oven log |
Filter the chamber air, tighten the controller band |
|
Haze in one cavity only |
Cooling imbalance on that cavity |
Thermal image of the cavity block |
Rebalance cooling, check for a blocked channel |
The single-cavity row is the one that saves a tool. When haze appears in one position and nowhere else, the compound is almost never the cause, and a compound change at that point will cost a week and return nothing.
What the Cell Controls on Every Shift
The tool runs a 64-cavity cold runner with needle valve gates at 0.8 mm, mold temperature held at 175 °C within 2 °C, and cavity pressure monitored on four positions. Cycle time is 16 seconds for a 1.8 mm wall teat, and the valve gate opening sequence is staggered across three groups so the runner does not see a single pressure spike.
First-article inspection runs every 2 hours and covers part weight, critical dimensions and a plaque haze reading. Weight is tracked on a control chart with a 2 percent band, because weight is the earliest signal of a fill or cure shift and it moves before haze does. When weight drifts toward the upper band, the cell adjusts before a single cloudy part reaches the inspection bench.
Visual inspection sits at a backlit station with 800 lux and a black matte background, and the operator rotates each part through 90 degrees. Haze that survives that rotation is bulk haze. Haze that appears and disappears with rotation is surface scatter from a scratch or a deposit, and the two are routed to different corrective actions.
Agreeing an Acceptance Limit With the Buyer
A haze clause is only useful when it names the specimen, the instrument and the light source. Anysil writes the clause for baby care LSR with three parts. A 2 mm plaque from the production lot, measured per ASTM D1003 with illuminant D65, and a limit agreed per grade, typically 6 percent for a premium transparent teat and 10 percent for a standard one.
The clause should also state the ageing condition. A reading taken 2 hours after molding and a reading taken after 24 hours are not the same number, and a supplier measuring early will always look better. Fixing the hold time before measurement removes the argument before it starts.
|
A production detail worth knowing when comparing quotations. A 64-cavity tool with balanced cooling produces a haze spread of about 0.8 percentage points between the best and worst cavity, and that spread is normal. A tool with one starved cooling circuit shows a 3 point spread, and the difference is visible on the plaque set before any instrument is switched on. Asking a supplier for plaques from first, middle and last cavity exposes the tool faster than any audit question. |
Next Step
Ask for a transparency sample set from the current high-transparency nipple grade, supplied as three 2 mm plaques with the haze value and lot number printed on each. Buyers comparing two compounds should request plaques rather than finished nipples, because only the plaque gives a number the two suppliers can be held to.