How to Fix Splay and Silver Streaks: Moisture, Shear and Mold-Side Sources

A polycarbonate lens cover went silver on shot 40 of a production run. By the end of the first shift, 22% of parts showed a fan of white streaks from the gate, and every one failed gloss inspection. The dryer had logged 90 °C instead of 120 °C, and the moisture analyzer read 0.07% on hopper pellets. One drying correction returned rejects to zero.

Splay is that fan of silver streaks or white haze. It appears when gas flashes out of the melt at the flow front, and the bubbles stretch into streaks as the front moves. The gas can be steam from wet resin, degradation gas from over-shear, or trapped air. Each source needs a different fix, and fixing the wrong one wastes a full week of trial shots.

Read this from the mold side. You are the tooling engineer deciding gate geometry, runner layout, venting, and gate land finish before steel is cut. Drying and barrel settings enter as inputs to those decisions, not as the main event. Values marked typical come from processing guides and resin datasheets, so confirm each one against your grade.

The 9-defect troubleshooting overview covers splay in a paragraph. This guide goes deep on moisture, shear, and mold-side sources, with the numbers and the diagnosis order that end the defect.

The Snapshot

• Splay is a fan of silver streaks or white haze from gas flashing out of the melt at the flow front. Reject share typically runs 10 to 30% of a shift.
• Moisture is the dominant cause. Hygroscopic resins absorb 0.1 to 0.7% moisture in ambient air, and PC must dry at 120 °C for 3 to 4 hours to below 0.02%.
• Each 0.1% of moisture releases roughly 2.5 ml of steam per gram of melt at a 300 °C barrel. That vapor has nowhere to hide in the cavity.
• Over-shear follows moisture. Gates below 50% of wall thickness push shear rates past the 40,000 to 60,000 s-1 limit typical for PC.
• Check in this order: moisture first, screw speed second, gate and runner third, purge last. A 5,000 to 15,000 USD dryer protects against a 10 to 30% reject day.

The Silver Fan Near the Gate

Look at the failure before you touch a dial. Thin silver lines fan out from the gate area like frost on glass. On transparent resins the streaks look metallic and directional. On dark gloss parts the same defect shows as white haze across 20 to 40% of the surface. A textured surface hides it, which is why splay surfaces first on automotive trim and consumer electronics.

Position is the first clue. Gas that flashes at the gate produces streaks that radiate from the gate. Gas flashing all over the part produces a uniform haze, and that signature points to wet resin. A fan at the end of fill or at weld lines points to trapped air, which is a venting story, not a drying story.

Molders misread splay in two directions. Some call every streak jetting and slow the injection speed until the part short-fills. Others blame every haze on moisture and re-dry resin that was already dry. Meanwhile the real cause sits in a cold slug well that was never machined. Both mistakes cost a full day of trials.

Jetting leaves worm tracks from a free jet that snaked across the cavity. That is a different mechanism with a different cure, covered in our jetting guide. Glass-flow marks are a third neighbor: glossy bands with no streaks at all. Splay, jetting, and burn marks each get a row in the table below. Misreading one for another is why splay survives a week of failed fixes.

What Splay Actually Is: Gas in the Melt

Splay starts with dissolved gas. Melt under 50 to 150 MPa of injection pressure can hold gas in solution. The flow front runs near atmospheric pressure, so gas comes out of solution there and nucleates into bubbles. Fountain flow stretches those bubbles into long streaks, and the skin freezes them within 0.1 to 1 second of fill.

The gas has three sources, and each behaves differently. Water vapor dominates on hygroscopic resins, flashing everywhere because moisture sits in every pellet. Degradation gas, mostly CO2 and monomer fragments, forms where melt overheats or over-shears. Trapped air from poor venting shows up at last-filled locations.

The volume math explains why moisture is so destructive. One gram of PC at 0.1% moisture carries 0.001 g of water. That water becomes roughly 2.5 ml of steam at a 300 °C barrel temperature. A 100 g part at 0.1% moisture tries to push about 250 ml of steam through a cavity with nowhere to put it.

Not every surface streak is gas. A rough gate land tears the melt skin and leaves silver lines that contain no bubbles. Flow marks from hesitation leave ripples with no gas at all. The fastest discriminator is a short shot at 60 to 70% fill. Gas splay shows bubbles frozen at the front. Torn skin shows a smooth but damaged surface.

The table below is the field reference. Use the signature and location columns to pick the root-cause family before you change anything.

← swipe to scroll →

Table note: Signatures and fix order are typical industry practice for unfilled engineering resins. A part can carry two defects at once, so confirm with a short shot first.

Moisture: The Dominant Cause and the Drying Numbers

Moisture causes most splay cases, and the reason is chemistry. Hygroscopic resins such as PA, PC, PET, PBT, ABS, TPU, PPSU, PSU, PEI, and PEEK pull water out of humid shop air. Ambient absorption lands in the 0.1 to 0.7% range depending on resin, humidity, and storage history. None of them mold acceptably at those levels.

Each family has a published drying spec. PC dries at 120 °C for 3 to 4 hours to below 0.02%. ABS runs at 80 to 90 °C for 2 to 4 hours to below roughly 0.1%, with 0.05% typical for gloss parts. PA66 takes 80 °C for 4 hours to below 0.2%. PBT needs 120 to 130 °C to below 0.02%, and TPU runs 80 to 100 °C for 2 to 4 hours.

The 0.02% rule is the one to memorize for condensation polymers. PC, PET, PBT, and the PSU family hydrolyze above that threshold, which means the polymer chains actually break. The damage is not cosmetic only. A part molded from wet PC loses impact strength even when the streaks are invisible on texture.

Believing the dryer is the classic error. A dryer at 90 °C instead of 120 °C, a clogged filter, a dew point above -40 °C, or a hopper that gives only 90 minutes of residence. All four produce wet resin behind a healthy-looking display. The fix is verification, not trust. Pull pellets from the hopper throat and run a moisture analyzer. Below 0.02% passes for the condensation family, and the analyzer beats any dryer display.

Moisture splay has a recognizable fingerprint. The haze is uniform and covers the whole surface, often with tiny surface bubbles on the first shots after a shutdown. Parts fine at start-up and bad by mid-shift mean pellets are picking up moisture in an open hopper. Parts bad from shot one after a weekend mean the resin sat wet in the machine hopper.

Shear and Degradation: Gate and Screw Sources

When the resin is dry and splay persists, look at shear. Melt forced through an undersized gate sees enormous shear rate. Above a resin-specific ceiling, shear tears polymer chains into degradation gas. The gate geometry that decides this was drawn before steel was cut, so it is a mold decision with a process consequence.

The gate depth rule is the practical trigger. Edge gates at 50 to 80% of wall thickness are the typical working range. Below about 50% of wall thickness, shear splay becomes likely on engineering resins. The damage concentrates at the gate because velocity peaks there. A pinpoint gate on a fast fill can exceed 100,000 s-1, while the practical ceiling for PC sits near 40,000 to 60,000 s-1.

Melt temperature stacks on top of shear. Barrel temperatures near the top of the window add thermal degradation, and the two insults multiply because sheared melt is more heat-sensitive. PC above roughly 320 °C releases degradation gas even when bone-dry. The barrel profile, screw speed, and back pressure set that temperature, so the process can either respect the gate or punish it.

Screw speed and back pressure are the second shear source. A screw running at maximum RPM works the melt in the barrel. High back pressure forces melt back over the screw flights repeatedly. Both generate frictional heat that barrel heaters cannot remove. If the gate and runner were sized for the flow length, the screw should never need to work that hard.

The fix ladder is short. Reduce screw speed by 20 to 40% and watch 10 to 20 shots. Drop back pressure toward the minimum that holds shot consistency, typically 0.3 to 1.0 MPa. If streaks thin but do not die, the gate owns the defect. Enlarge gate cross-section by 20 to 50% and shorten the land toward 0.5 to 1.5 mm.

Mold-Side Sources: Cold Slugs, Vents and Gate Land

Three mold-side features create splay that no drying fixes. The first is the cold slug. Melt at the sprue tip and at runner ends is colder than the flowing melt, often by 20 to 30 °C. It touches cool steel between shots. That cold plug injects straight into the cavity on the next shot. It does not flow or wet the surface, and it reads as a silver streak at the gate.

The cure is a cold slug well at the end of the sprue and at every runner end. Industry practice sizes the well at 0.5 to 1.0 times the runner diameter deep, with an undercut so the slug pulls out. On a sprue, the well is the first 5 to 10 mm past the sprue bushing exit. A mold without cold slug wells shows gate splay on every shot, and no process setting removes it.

The second mold-side source is poor venting. Gas in the cavity must escape through parting-line vents, and vent depth is resin-specific. The vent must pass gas without flashing. Typical land vents run 0.01 to 0.04 mm deep. Low-viscosity semicrystalline grades such as PP and PA66 sit at the shallow end, near 0.01 to 0.02 mm. PC and ABS run 0.03 to 0.04 mm. Land length stays near 1 to 3 mm.

The third mold-side source is the gate land. A land that is too long keeps melt channeled past the point where it should spread. A rough land tears the melt skin, and both leave silver lines that mimic splay. Edge gate lands belong in the 0.5 to 1.5 mm range. Cosmetic gate zones are polished to an SPI A-2 mirror at roughly 0.05 to 0.1 µm Ra.

Our venting guide covers depth, location, and land rules for every resin family. All three mold-side sources share one property: they are invisible in flow simulation and obvious on the first trial shot. That is why the mold trial review checks cold slug wells, vent depth, and gate land finish before anyone touches injection speed.

Contamination: Regrind and Masterbatch

Contaminated feed makes clean molds look guilty. Regrind is the first suspect. Sprues and runners ground on a dull blade produce dust and fines, and dusty regrind carries surface area that soaks up moisture fast. Most datasheets cap regrind at 10 to 30% of the feed, and the cap drops toward zero for gloss and optical parts. Regrind left open in a bin behaves like wet virgin resin regardless of the dryer.

Lubricated regrind is a nastier variant. Parts sprayed with external mold release carry the lubricant into the regrind. The lubricant vaporizes at barrel temperature into white streaks that no dryer touches. The fingerprint is splay on high-regrind parts with a clean virgin pellet check.

Colorant masterbatch is the third vector. Masterbatch carries its own moisture, and its carrier resin may differ from the base resin. A 2 to 4% load of damp carrier can push the blend past 0.02% even when the base resin is bone-dry. Dry the blend, not just the base resin, and keep opened masterbatch bags sealed.

The isolation test costs one hour. Run 50 to 100 shots on 100% virgin resin from a sealed bag, dried to spec. If splay disappears, the contamination is in the feed stream. If it survives virgin material, the mold or the process owns the defect.

Diagnosis Workflow: What to Check in Order

Work the splay checklist in this order, from the cheapest check to the most expensive change. Skipping the order is how a mold gets modified for a moisture problem, or a dryer gets blamed for a gate undersized on the drawing.

Step 1: verify moisture first. Check the dryer log for the last 8 hours. Temperature must sit within spec, dew point below -40 °C, and hopper residence at or above the drying time. Then pull pellets from the hopper throat and run a moisture analyzer. Below 0.02% passes for PC, PET, PBT, and the PSU family. PA66 passes below 0.2%. Wet resin explains 70 to 80% of splay cases before mold work is justified.

Step 2: cut the shear inputs. Reduce screw speed by 20 to 40% and trim back pressure. If the melt runs hot, pull the barrel profile down 10 to 20 °C toward the datasheet midpoint. Run 10 to 20 shots and compare. Shear splay thins as the screw slows. Moisture splay ignores screw speed, which makes this a clean discriminator.

Step 3: audit the gate and runner. Measure gate depth against wall thickness. Below 50% of wall, enlarge the gate by 20 to 50% in cross-section and shorten the land toward 0.5 to 1.5 mm. Confirm a cold slug well exists at the sprue and runner ends at 0.5 to 1.0 times diameter deep. Polish a rough EDM gate land to the SPI A-2 range. These steel changes happen once and hold for the life of the tool.

Step 4: purge the barrel. Degraded resin from a shutdown or a wrong-temperature excursion contaminates every shot. Purge with a cleaning compound for 10 to 20 minutes near the top of the processing window, then run virgin material. The purge also clears residue from a previous, incompatible resin.

Step 5: escalate to a vented barrel. For PA and PC on wet-prone, long-cycle jobs, a vented barrel with a two-stage screw is the last resort. The vent opens at the barrel, and a vacuum pump pulls volatiles out before the melt reaches the nozzle. Typical two-stage screws run an L/D near 24 to 32. A vented barrel cannot fix cold slugs or mold venting, so use it only after steps 1 through 4 are proven clean.

Verification follows a fixed window. Run the first 50 shots and inspect every one under shop lighting. Then check gloss on the cosmetic surface with a 60° gloss meter against the approved sample, with acceptance typically within 3 to 5 gloss units. A fix that survives 50 shots is a fix. A defect that returns on shot 60 was masked, not cured. On new tools this verification belongs in the mold trial and T1 sampling phase, where the process window is documented before production starts.

Prevention on the Production Floor

Prevention is a set of habits with numbers attached. Run desiccant dryers continuously at -40 °C dew point or better. Size the hopper so pellet residence matches drying time. PC needs 3 to 4 hours, ABS and TPU need 2 to 4 hours, and PA66 needs 4 hours.

Seal the material path. Opened bags, open gaylords, and uncovered regrind bins let hygroscopic resin climb past 0.02% in hours at typical shop humidity of 50 to 70% RH. Dried PC exposed to open air reabsorbs measurably within 15 to 30 minutes. The rule is simple: dried resin never sees open air between the dryer and the machine throat.

Schedule the mold-side checks. Vent channels need cleaning in preventive maintenance at typical intervals of 50,000 to 100,000 shots. Gate lands get re-checked when a mold returns from repair, because hand polishing rarely matches the original land length. Cold slug wells are verified at every T0 trial and after any runner modification.

Track the data that catches drift early. A dryer log with temperature, dew point, and material code per shift turns a 22% reject day into a 15-minute root-cause search. A first-shot inspection that flags splay at shot 1 instead of shot 4000 stops a bad shift before it starts. The plants that fix splay fastest measured moisture before they blamed the mold.

Resin Drying Quick-Reference Table

This table is the field reference for the moisture half of splay. Temperatures, times, and targets are typical industry practice from resin datasheets. Drying time assumes a desiccant dryer at -40 °C dew point with adequate airflow. The targets are the pass lines used for moisture analyzer checks.

← swipe to scroll →

Table note: Values are typical industry practice. Confirm temperature, time, and target moisture against the specific grade datasheet before setting the dryer.

Read the table with the analyzer rule in mind. For PC, PBT, PET, and the PSU family, 0.02% is the ceiling, and a reading above it justifies stopping the press. PA66 passes at 0.2% because nylon tolerates more water before hydrolysis shows. Non-hygroscopic PP and PE rarely splay from moisture. Their splay points to shear degradation or venting. The fix order flips to the mold side first.

FAQ: Splay, Silver Streaks and Gas Traps

Q1. What is the difference between splay, jetting, and burn marks?

Splay is gas flashing out of the melt at the flow front, leaving a fan of silver streaks or white haze. Jetting is a free jet snaking across the cavity in worm tracks, with no gas involved. Burn marks are charred deposits where trapped air passed roughly 300 to 400 °C. The differentiation table in this guide gives the field signature of each.

Q2. How do I tell moisture splay from shear splay?

Use location and uniformity. Moisture splay is a uniform haze across the whole surface, because water sits in every pellet. Shear splay concentrates at the gate in a directional fan. Then confirm with a moisture analyzer on hopper pellets. Above 0.02% for PC and the condensation family means drying is the fix, and below that the gate owns the problem.

Q3. How long can dried PC sit before it reabsorbs moisture?

Not long at typical shop humidity. Dried PC exposed to open air at 50 to 70% RH can climb back above the 0.02% target within 15 to 30 minutes. Keep dried pellets sealed between the dryer and the hopper. The analyzer reading at the hopper throat is the only number that matters.

Q4. Can process settings alone fix splay?

Only for the shear family. Cutting screw speed by 20 to 40% and lowering back pressure can thin out shear splay. Moisture splay needs dry resin. Cold slug, venting, and gate land splay need steel changes. Add a well at 0.5 to 1.0 times diameter deep, vents at 0.01 to 0.04 mm, and a land at 0.5 to 1.5 mm. Process settings that mask the defect return on the next material lot.

Q5. Is a new dryer cheaper than the rejects it prevents?

Almost always. A desiccant dryer with -40 °C dew point capability runs in the 5,000 to 15,000 USD range. A splay day on a busy machine rejects 10 to 30% of output. One bad shift on one machine often exceeds the monthly cost of the dryer. Drying is the cheapest insurance in the molding plant.

The Bottom Line

Splay is gas in the melt, and moisture causes most of it. Dry hygroscopic resin below 0.02% for the condensation family. Verify with an analyzer. Cut screw speed by 20 to 40%. Then fix the gate, cold slug wells, and vents. Cheap checks first, steel changes last, and the defect stays gone.

DieStrike builds injection molds under IATF 16949 across 120+ machines. Molds ship in 2 to 5 weeks with geometry at ±0.005 mm standard and ±0.002 mm precision. Hardened cavity steel runs at HRC 62. Standard parts ship in 3 to 7 days. Connector OEMs including TE Connectivity, Amphenol, and Luxshare run production on the same discipline.