How to Design Overmolding: Bonding, Shrinkage Match and Tooling

Run 41 of a 300,000-piece connector program stopped at the leak bench when 7 of 32 TPU gaskets failed a 0.1 MPa air test. The worst part showed 0.4 mm of peel-back at the PA66 bond line. The tool drawings explained why: no primer, no lock, and a 4 mm soft wall standing over a 1.5 mm rib.

Overmolding presses a soft elastomer layer onto a rigid substrate inside one molded assembly. Typical jobs are a TPE grip over a PP handle, a TPU seal around a PA66 housing, and a strain-relief boot over a cable. TPE and TPU are thermoplastics, so both layers can be injection molded in sequence.

These parts fail two public ways. Delamination peels the soft layer at edges, holes or gate stubs. Leaks open a channel between the layers where water or air travels. Both failure modes trace to decisions made before the steel was cut.

This guide walks product engineers and mold buyers through substrate prep, bond mechanics, shrinkage, geometry, tooling and cost. Values are typical industry ranges from supplier literature and shop practice. Verify every number against your grade datasheet and toolmaker before you quote.

The Snapshot

• Bond peel strength runs 2 to 10 N/mm for chemically bonded pairs such as SEBS on PP, and 0.5 to 2 N/mm where mechanical locks carry the load alone.
• Keep the overmold melt 15 to 30 °C below the substrate melting point so the second shot never remelts the first.
• Shrinkage mismatch between the two materials spans 0.3 to 2.5 percent, and typical soft walls run 0.5 to 3 mm.
• Where chemistry cannot bond, add undercuts or pinholes with 0.3 to 0.8 mm of engagement.
• The overmold stage adds 10 to 40 seconds of cycle and typically 30 to 70 percent to tool cost.

Why Overmolds Delaminate and Leak

Peel failures carry real cost. On a 16-cavity grip tool, a 3 percent delamination rate turns a 30 second cycle into rework, sorting and returns. Molders report 5 to 15 percent first-trial scrap on TPE overmolds. The dominant causes sit on the drawing: a wrong material pair, a dirty substrate or a cold mold.

Delamination starts where peel stress concentrates. Thin soft edges, gate vestiges, through-holes and sharp corners lift first. The crack then runs along the interface at 2 to 10 N/mm for bonded pairs and 0.5 to 2 N/mm for mechanical grip alone. Once air or water enters, the interface becomes a leak path.

Seal programs fail on the bench before they reach the field. Leak tests at 0.05 to 0.3 MPa air are standard on connector gaskets. A 0.1 mm void at the bond line can fail them. Hot water attacks TPU bonds by hydrolysis over time, so wet-duty material choice matters as much as geometry.

Peel testing is the standard release gate. Mold a 2 to 3 mm soft plaque over a substrate plaque, cut 25 mm strips, and pull at 90 degrees. Compatible pairs typically hold 2 to 10 N/mm, and anything below 1 N/mm will not survive handling. Ask the molder for the test method in the RFQ so results compare across suppliers.

Substrate Preparation: Clean, Dry, Treated

The substrate surface is the bond. Mold release carried over from the first shot is the most common contaminant. A few ppm of silicone at the interface blocks wetting across a whole cavity. Degrease molded parts, or purge the first-shot tool of sprays before overmolding.

Cleaning has a hierarchy. Molded parts should travel from shot one to the overmold press without mold release, fingerprints or dust. Dry ice blasting and atmospheric plasma remove residue without solvent, and both work inline at line speeds of 1 to 5 m per minute. Barrier gloves and covered bins stop re-contamination between stations.

Nylon is moisture-critical in both shots. PA6 and PA66 absorb 2.5 to 3 percent moisture at saturation. Dry pellets at 80 to 90 °C for 2 to 4 hours to below 0.2 percent before molding. Molded nylon parts re-absorb quickly, so overmold within 4 to 8 hours or re-dry them for 2 to 4 hours at 70 to 90 °C.

Polypropylene will not bond out of the box. Untreated PP sits at 29 to 31 mN/m surface energy, below the wetting threshold of most TPE melts. Flame, plasma or corona treatment lifts it to 38 to 44 mN/m. Shops screen with dyne pens or a water contact angle below 60 to 70 degrees.

Treatment decays within hours to days, so overmold in the same shift. A chlorinated polyolefin primer is the alternative for PP when timing slips. Metal substrates get no treatment shortcut: degrease them, and consider a silane primer for TPU on aluminum.

Bonding: Chemical, Mechanical and Mechanical-Lock

Three mechanisms carry an overmold joint: chemical bond, micro-mechanical grip and macro locks. Chemical bonding needs compatible polarity. Styrenic TPEs such as SEBS pair with PP, PE and ABS, while polar TPUs suit nylon, PC and PBT better. Each pair behaves differently in peel, and our TPE material guide summarizes the families.

Wetting starts the chemical bond. A melt wets a surface when the substrate energy exceeds the polymer energy by a useful margin. The practical targets are 38 to 44 mN/m on the substrate and a melt that reaches the bond zone hot. Mold temperature does the wetting work: bond tools typically run 40 to 80 °C on both halves.

Micro-mechanical grip comes from surface roughness. Blast or texture the bond zone to about Ra 0.8 to 3.2 µm to add area. Avoid texture so deep it traps air, because interface voids turn into leaks.

Macro locks carry the joint when chemistry is impossible. Pinholes and undercuts with 0.3 to 0.8 mm of engagement stop the soft layer from peeling. Through-holes should run at least 1.5 to 2.0 times the soft wall diameter so the melt locks through. Metal inserts, POM and PTFE need this treatment every time.

← swipe to scroll →

Table: typical molded-sample bond ranges from supplier literature, geometry, wall and process shift the results, so verify against your grade datasheet.

Bond strength peaks on the freshest surface. Overmold within hours of shot one when you can, because PP treatment and nylon dryness both decay with shelf time. A tie layer of compatible TPE, 0.3 to 0.5 mm thick, bridges pairs that will not wet each other. The price is one extra shot.

Adhesion promoters fill the gap between treatment and locks. Chlorinated polyolefin primers for PP dry to a 5 to 15 µm film. Silane primers improve TPU bonds to metals and glass. Check VOC rules and supplier cure windows before committing.

Shrinkage Mismatch and Warpage Control

Every polymer shrinks as it cools, and overmolding joins two materials that shrink at different rates. Styrenic TPEs shrink 0.5 to 2.0 percent and TPUs 0.3 to 1.2 percent. Rigid substrates sit lower. ABS runs 0.4 to 0.7 percent, PC 0.5 to 0.7, PBT 0.3 to 0.8 and PA66 0.7 to 1.8. The practical mismatch window runs 0.3 to 2.5 percent, and it decides warpage, stress and seal grip.

Mismatch can help or hurt. A TPE that shrinks more than its substrate contracts onto the core part, which holds seal lips in compression. That same contraction pulls the rigid substrate into tension, and thin PC or PMMA ribs can crack under it.

The arithmetic is easy to see. Across 100 mm of bonded length, a 1 percent shrinkage difference wants to move the interface about 1 mm. The bond resists, so the stack either warps or stores stress.

A two-material plate curls like a thermostat strip. Long flat bonds curl most, so designers break them with grooves or gaps every 20 to 50 mm. End the soft layer in a radius, never a square edge, so peel cannot start.

Temperature controls the result. Keep both mold halves in the same band so the two faces cool symmetrically. A cavity-to-core spread beyond 10 to 20 °C tilts the part. Most crystalline substrates melt between 200 and 260 °C. Keep the TPE melt 15 to 30 °C below that ceiling so shot two never remelts shot one.

Glass-filled substrates shrink less, which widens the mismatch. A 30 percent glass PA66 part shrinks near 0.3 to 0.6 percent against a TPE at 1 percent. Plan locks and accept higher interface stress. Datasheet shrinkage comes from plaques, so confirm it on your own mold trials.

Metal substrates change the game. TPE shrinks onto a steel insert and grips it, which is how strain relief sleeves hold. The same grip stores stress at the interface, so knurls or locks of 0.3 to 0.8 mm keep the sleeve located through thermal cycling.

Cooling layout follows the bond. Run water channels 8 to 12 mm below the bond surface on both halves. Channel diameters of 3 to 5 mm and a 10 to 20 °C inlet spread keep the two faces even.

Wall Thickness, Flow Length and Geometry Rules

Soft walls occupy a narrow design band. Overmold grades span 30 Shore A to 70 Shore D, and typical walls run 0.5 to 3 mm. Soft-touch grips use 1.0 to 2.5 mm, while sealing beads run up to 3 mm for crush resistance.

Thin walls fill fast and cool fast. Below 0.5 mm, TPE short-shots at ribs, gates and far corners. Thick walls sink and slow the cycle. Cooling time scales with the square of wall thickness, so a 3 mm section cools about four times longer than a 1.5 mm section.

Flow length decides how far the soft melt travels before freezing. Plan 100 to 150 times the wall for Shore A grades above 50 Shore A. Below 50 Shore A, viscosity climbs and the practical limit drops toward 100:1. A fan gate or multiple gates rescues long grips.

Radius every internal corner 0.3 to 0.5 mm so peel cannot start on a knife edge. Keep soft edges 0.8 to 1.0 mm minimum so they do not tear in handling. Draft soft walls 1 to 3 degrees, because a feather-thin TPE edge is a delamination waiting to happen.

Under the soft layer, substrate ribs and bosses follow normal rules with margin. A boss that supports a 0.3 mm soft pad needs wall thickness near its base, because the soft layer does not stiffen it. Soft-side tolerances run ±0.1 to ±0.3 mm, while the bond line position is set by steel at ±0.05 mm.

Knit lines are weak spots in any soft layer. Where two flow fronts meet, the bond to the substrate is often thinner, so place the meeting point away from seal lips and visible grips. Fill simulation on the DFM shows the weld line before steel is cut.

Tooling Approaches: Pick-and-Place vs Core-Back vs Shuttle

Three tool concepts dominate insert overmolding, and each changes cycle and cost. Core-back keeps the part in one tool. The first shot fills the full cavity, a core retracts to open a second space, and shot two fills it.

Core-back alignment is automatic because the same tool holds both shots. The cycle grows by 10 to 25 seconds, and the tool adder typically lands at 20 to 40 percent.

Pick-and-place runs two tools, often on two machines. A robot lifts each molded substrate into the overmold cavity, which suits large parts, metal inserts and mixed materials. Expect 15 to 40 seconds of added cycle for the transfer. Retention and loading rules are covered in our guide to insert mold design.

A shuttle mold slides the bottom half between two top halves. One side molds the first shot while the other side runs shot two. It adds 10 to 30 seconds per part and fits mid-volume jobs where a robot is overkill.

Cavity count and locating decide which concept wins. Substrates locate on three points or a nest with ±0.05 mm repeatability typical. A 16-cavity overmold tool with pick-and-place needs a robot cycle under 10 seconds per set to keep the 15 to 40 second adder honest.

Two-Shot Rotary: One Comparison

For completeness, a rotary two-shot tool molds both layers in one clamped tool on an index plate, with two injection units and no robot. Handling time drops near zero, but the stack is taller and the tool adder runs 40 to 80 percent, which pays only at high volume. The deep design rules live in our guide to two-shot mold design. The concepts above are the insert-style routes for the parts this article covers.

← swipe to scroll →

Table: typical industry planning ranges, cavity count, part size and automation level move each number.

Cool both halves whatever the concept. Bond quality depends on interface temperature, so each side runs its own controller at 40 to 80 °C typical. Keep the two within 10 to 20 °C of each other, because an uneven tool warps the stack and peels one side of the bond.

Tool trials confirm the choice. Run both approaches in DFM before quoting, because cavity count and part size move the winner more than habit.

Gating, Venting and Ejection for Overmold Tools

Gate the soft layer at its thickest section and feed it in one continuous front. Submarine gates at 0.8 to 1.2 mm bore keep the vestige small on cosmetic grips. Fan or film gates spread melt across long seals. Keep the land short at 0.5 to 1.0 mm so soft material does not hang and blush.

Soft materials flash before rigid ones do. Vent depths of 0.01 to 0.02 mm suit TPE and TPU, against 0.02 to 0.05 mm for commodity rigid resins. Grind vents shallow and put them at last-to-fill. Trapped air at the bond interface becomes voids that leak. Depth and location rules are in our guide to mold venting.

Vent the interface, not just the cavity. Bond zones trap air where the melt meets the substrate. Place 0.01 to 0.02 mm vents along the bond line at 25 to 50 mm spacing. A vacuum assist of 20 to 40 kPa helps thin seals fill before the interface seals off.

TPE grips steel like a suction cup on ejection. Use stripper plates or rings and large-area pins instead of thin ejectors. Add air poppets for deep soft sections. Hold the mold closed until the part cools below 40 to 50 °C so it does not deform.

Polish the soft-side cavity to SPI A-3 or B-3 finish. Coarse texture tears soft TPE on the way out. Flexible undercuts strip without slides when the lip is under about 0.5 mm, but confirm it in tool trials. Core pins that form lock holes need full support, because a 0.5 mm pin buckles under TPE pressure.

Runners and cold wells matter for soft materials. Use a cold runner with a 6 to 8 mm diameter for TPE and a 10 to 15 mm cold slug well at the sprue. Soft sprues stretch and snap at different points each cycle, so automate runner removal above about 50,000 parts a year.

Applications That Justify Overmolding

Soft-touch handles justify overmolding by feel alone. Power tools, shavers and hand tools carry a 40 to 70 Shore A TPE layer of 1.5 to 2.5 mm over PP or ABS. The layer adds grip and damping, and a peel test is the release gate because failure here is cosmetic and warranty-driven.

Seals make overmolding functional. A gasket molded onto a connector housing deletes the separate O-ring and its assembly step. The bond line becomes part of the sealing system, and IP67 testing pushes 1 m of water for 30 minutes. Design gasket lips for 15 to 25 percent compression. More squeeze over-compresses TPE and causes compression set.

Cable strain reliefs and boots use TPU for tear and abrasion resistance at 70 Shore A to 55 Shore D. A boot overmolded onto the jacket survives repeated flex and pull tests of 30 to 100 N on consumer electronics. Automotive harnesses push the same joints through salt spray and 105 °C soak.

Overmolding wins when it deletes a secondary process. A bonded gasket removes an O-ring groove, a loose seal and an assembly step. A soft boot over a metal shell removes a glued part. Ask whether the soft layer seals, cushions or decorates, because each duty sets a different wall and hardness target.

Automotive and electronics programs pay for overmolding because it removes assembly steps and seals better than loose gaskets. As a planning rule, volumes below 50,000 to 100,000 parts per year rarely justify a dedicated overmold tool. Field sealing failures can change that math.

Cost and Cycle Impact

Budget cycle time as a step change. The overmold stage adds 10 to 40 seconds per part. Core-back adds 10 to 25 seconds inside one tool, pick-and-place adds 15 to 40 seconds for transfer, and shuttle lands between. Rotary two-shot adds near zero handling time and buys that speed with more tooling.

Tooling cost follows the same ladder. An overmold tool typically costs 30 to 70 percent more than a single-shot tool of equal cavity count. Pick-and-place adds automation, and rotary two-shot stacks run 40 to 80 percent higher.

Run the unit economics before the tool order. At an $80 per hour machine rate, 20 extra seconds adds about $0.44 per part. At 1 million parts a year, that adder reaches $440,000. Material and scrap push it higher, which is why the cycle adder belongs in the RFQ spreadsheet, not the tool room.

Material cost widens the per-part gap. TPE sells at roughly 1.5 to 3 times polypropylene per kilogram, and TPU costs more again. On a thin soft layer the resin adder stays small, but the soft material share of part cost climbs fast. Early trials absorb 5 to 15 percent scrap on TPE runs, so mold test plaques and peel coupons before production.

Put bond tests in the RFQ. Specify a 90 degree peel method with a 2 to 4 N/mm minimum for soft-touch parts, or a leak test pressure for seals. Ask for temperature control on both mold halves. A DFM pass that flags chemistry and shrinkage before steel is cut is cheaper than a re-cut. Our mold design DFM service writes those checks into the tooling quote.

FAQ: Overmolding Design

Q1. Why does my TPE layer peel off a PP handle?

PP needs a surface energy lift before a SEBS melt can wet it. Untreated PP at 29 to 31 mN/m sits below the 38 to 44 mN/m target. Flame, plasma or corona treatment, or a chlorinated polyolefin primer, closes the gap. Screen with a water contact angle below 60 to 70 degrees and overmold the same shift.

Q2. Can TPE or TPU bond directly to metal?

Not by chemistry alone. Thermoplastic melts do not wet metal oxide surfaces strongly. Add mechanical locks with 0.3 to 0.8 mm engagement, knurls, or through-holes at 1.5 to 2.0 times the soft wall. Locked joints peel at 0.5 to 2 N/mm, enough for strain reliefs but not for pressure seals.

Q3. What wall thickness should the soft layer use?

Plan 0.5 to 3 mm, with 1.0 to 2.5 mm for soft-touch and up to 3 mm for sealing beads. Below 0.5 mm the layer short-shots at ribs and gates. Above 3 mm, cooling time grows with the square of thickness and the layer sinks.

Q4. How much slower is an overmold cycle?

Expect 10 to 40 extra seconds per part. Core-back adds 10 to 25 seconds inside one tool, and pick-and-place adds 15 to 40 for the transfer. A 30 second base cycle becomes a 45 to 70 second cycle before optimization.

Q5. When do I need temperature control on both mold halves?

For any bonded overmold. Interface temperature drives wetting, so run both halves at 40 to 80 °C and hold them within 10 to 20 °C of each other. One cold half gives one-sided bond strength, a warped part, and peel on the cold side only.

The Bottom Line

An overmold survives when the numbers line up. You need chemistry or locks, surface energy above 38 to 44 mN/m, and a melt 15 to 30 °C below the substrate melting point. Shrinkage mismatch must be understood at 0.3 to 2.5 percent before the tool is cut.

Then accept the 10 to 40 seconds of added cycle and the 30 to 70 percent tool adder. The geometry and cooling that protect the bond line are not optional.

DieStrike builds production molds under IATF 16949, with cavity steel hardened to HRC 62 and standard machining at ±0.005 mm. Its connector overmold programs for TE Connectivity, Amphenol and Luxshare start with a DFM that fixes bonding strategy and shrinkage before steel is ordered. Production molds ship in 2 to 5 weeks. Start with chemistry, then geometry, then tooling, and build the lock in from day one if the pair will not bond. A peel test on a molded coupon beats a field failure every time.