A plastic IC package is a silicon die and a metal leadframe encased in filled epoxy — cheap, light, and made by the billion, but not hermetic. That single fact, that moisture can diffuse through the bulk plastic, sets almost every limit these packages have. Plastic IC packages dominate commercial and industrial electronics; ceramic is reserved for the cases where hermeticity or extreme temperature leave no choice. This guide covers what they are made of, where they stop, and how to decide.
Key takeaways
- The encapsulant is an epoxy molding compound filled 70 to 90 percent by weight with fused silica, which drops its CTE from around 76 ppm per degree C unfilled to roughly 16 to 20 ppm per degree C below the glass transition, per published EMC data.
- Plastic is non-hermetic. Its moisture permeability is several orders of magnitude higher than ceramic or metal, which is why moisture-sensitivity handling (MSL) exists at all.
- The leadframe choice is a trade: copper alloy C194 conducts heat at about 260 W per meter-kelvin but expands at about 17 ppm per degree C, while Alloy 42 (Fe-42Ni) sits near silicon’s expansion at about 4.5 ppm per degree C but carries heat at only 12 to 15 W per meter-kelvin.
- Standard commercial plastic parts top out around 85 to 125 degrees C; AEC-Q100 Grade 0 qualifies plastic to minus 40 to plus 150 degrees C, but the package, not the silicon, usually sets the ceiling.
- Non-hermetic no longer means unqualified for space. MIL-PRF-38534 Appendix D and NASA’s Class Y define qualification paths for non-hermetic parts.
What a plastic IC package is made of
A plastic IC package has four material systems: the epoxy molding compound that forms the body, the metal leadframe or laminate substrate that carries signals out, the die-attach adhesive that bonds the chip down, and the fine wire or bump interconnect that links die pads to the leadframe. Everything about cost, size, and thermal behavior follows from those four choices.
The reason the package works, and the reason it fails, is that these materials expand at different rates when heated. Silicon barely moves, at about 2.8 ppm per degree C per Electronics Cooling’s materials data. The copper leadframe and the epoxy body move six times as much. That mismatch, repeated over thermal cycles, drives the delamination and cracking modes covered later.
| Material | Role | CTE (ppm/°C) | Thermal cond. (W/m·K) | Note |
| Silicon die | Active device | ~2.8 | ~150 | The reference everything is matched against |
| Fused-silica EMC | Package body | ~16–20 (below Tg) | ~0.6–0.9 | Filler cuts CTE from ~76 unfilled |
| Copper C194 leadframe | Signal + heat path | ~16–17 | ~260 | Best conductor; worst CTE match to Si |
| Alloy 42 (Fe-42Ni) leadframe | Signal + heat path | ~4–4.7 | ~12–15 | Matches Si; poor heat, high resistivity |
| Alumina (ceramic, contrast) | Hermetic body | ~7 | ~20–30 | Non-plastic alternative |
[IMAGE 1: cutaway of a molded plastic IC showing die, die attach, leadframe, wire bonds, and epoxy body | alt: “Plastic IC packages materials cross-section showing epoxy mold compound, leadframe, die attach and wire bonds”]
The mold compound: the material that sets most of the limits
The body is an epoxy molding compound (EMC), almost always an epoxy cresol novolac resin loaded heavily with fused-silica filler. CAPLINQ’s materials data puts a typical ratio near 80 percent filler to 20 percent resin, which yields a compound CTE around 60 ppm per degree C in that mix and far lower when optimized.
Filler content is the main lever. As published silica-filler studies show, raising filler lowers CTE, water absorption, and flow, while raising thermal conductivity and flexural strength. A well-filled EMC reaches a CTE of roughly 16 to 20 ppm per degree C below its glass transition and 25 to 73 ppm per degree C above it, against about 76 and 174 ppm per degree C for unfilled epoxy, according to work in the Journal of the European Ceramic Society.
Glass transition temperature (Tg) is the number that matters most for limits. The workhorse novolac compounds run a Tg of about 150 to 175 degrees C. Above Tg the compound softens and its expansion rate jumps to the higher CTE2 figure, so a part cycled past Tg sees far larger dimensional swings and stress at the die interface. Standard EMC thermal conductivity is low, roughly 0.6 to 0.9 W per meter-kelvin, though thermally enhanced compounds with alumina or boron-nitride filler reach up to about 4 W per meter-kelvin.
Leadframe and interconnect: the copper-versus-Alloy 42 trade-off
Most plastic parts use a stamped or etched copper-alloy leadframe because copper conducts. Alloy C194, at roughly 97.5 percent copper, carries heat at about 260 W per meter-kelvin and has low electrical resistivity near 1.7 micro-ohm-cm. Its penalty is expansion: at about 16 to 17 ppm per degree C it is badly mismatched to silicon.
The alternative is Alloy 42, an iron-42 percent nickel alloy whose CTE of about 4 to 4.7 ppm per degree C sits close to silicon and ceramic. That match buys mechanical stability, which is why it persists in some fine-pitch and hybrid parts. The cost is thermal and electrical: Alloy 42 conducts heat at only 12 to 15 W per meter-kelvin and has resistivity near 70 micro-ohm-cm, roughly forty times copper’s, per leadframe-alloy references. Pick copper when heat dominates and the die-attach can absorb the stress; pick Alloy 42 when dimensional match outweighs heat spreading.
Interconnect is usually gold, copper, or silver bonding wire. Copper wire is cheaper and stiffer but demands tighter bonding control and inert-gas coverage; gold is forgiving but forms brittle intermetallics against aluminum pads over time at elevated temperature.
The limits: temperature, moisture, and no hermetic seal
Plastic packages are non-hermetic, and that is the defining constraint. As the TJ Green hermeticity review explains, moisture permeability through bulk plastic is several orders of magnitude greater than through glass, metal, or ceramic, so water vapor reaches the die over time regardless of how clean the assembly was.
That permeability creates two failure paths. Absorbed moisture flashes to steam during reflow and can crack a part ramped above about 100 degrees C, the popcorn mode. And condensed moisture on the die can drive corrosion and leakage across conductors. It also means the standard hermeticity test, MIL-STD-883 Method 1014, does not describe a plastic part at all; it was written for cracks in glass-to-metal seals, not bulk diffusion.
| Property | Plastic (molded) | Ceramic (hermetic) |
| Hermetic seal | No | Yes |
| Moisture path | Diffuses through bulk | Negligible |
| Max temperature | ~85–150 °C (grade-dependent) | 200 °C+ |
| Thermal conductivity | Low (EMC ~0.6–0.9 W/m·K) | High (alumina/AlN) |
| Cost | Low | High |
| Size / weight | Small, light | Larger, heavier |
| Typical use | Consumer, industrial, automotive | Space, military, high-temp sensing |
Non-hermetic no longer bars a part from harsh service. The most recent MIL-PRF-38534 carries an Appendix D defining qualification and testing for non-hermetic packages, and NASA has driven a Class Y category for non-hermetic monolithic devices in space, per the same hermeticity review. Plastic-encapsulated parts now fly where only ceramic once qualified.
Compliance: temperature grades, RoHS, and halogen-free
Temperature capability is graded, not universal. For automotive-qualified plastic parts, AEC-Q100 defines the operating ranges, below. Match the grade to the ambient at the part, not the room.
| AEC-Q100 grade | Operating range | Typical use |
| Grade 0 | −40 to +150 °C | Under-hood, harsh automotive |
| Grade 1 | −40 to +125 °C | General automotive |
| Grade 2 | −40 to +105 °C | Passenger cabin |
| Grade 3 | −40 to +85 °C | Mild automotive / industrial |
| Grade 4 | 0 to +70 °C | Consumer / commercial |
Moisture handling is governed by MSL under JEDEC J-STD-020. An MSL 3 part has 168 hours of floor life after the dry-pack is opened; MSL 5a allows only 24 before the reel must be baked. Skipping that step is what turns absorbed moisture into cracked packages at reflow.
On materials compliance, RoHS restricts lead and other substances, while green usually means low-halogen. Under JEDEC JS709C, a solid-state device is low-halogen when each plastic material holds under 1000 ppm bromine from brominated flame retardants and under 1000 ppm chlorine from chlorinated sources, as TI’s Green statement and JS709C both specify. That differs from the stricter halogen-free board definition in IEC 61249-2-21, which caps total halogens at 1500 ppm with under 900 ppm each of bromine and chlorine.
When plastic is the right call — and when to go ceramic
For most designs, plastic is correct by default. It is cheaper, smaller, lighter, and available in far more parts, and modern EMCs and qualification flows cover the commercial, industrial, and automotive temperature ranges without difficulty.
Reach for ceramic or a hermetic package only when a specific requirement forces it: sustained operation well above an EMC’s Tg, a moisture-driven reliability target that bulk plastic cannot meet, a UV-erasable memory needing a quartz window, or a space or military program that has not yet adopted the non-hermetic qualification routes. Even there, the MIL-PRF-38534 Appendix D and Class Y paths mean the decision is now about qualification cost and risk tolerance, not a hard material wall.
Failure modes that trace back to the plastic
Most plastic-package failures come from a short list, and all of them start with the material set.
Popcorn cracking and delamination lead it: moisture absorbed into the EMC vaporizes at reflow and separates the compound from the die or leadframe. CTE-mismatch fatigue is the slow version, where the roughly 14 ppm per degree C gap between a copper leadframe and silicon works the interfaces over thermal cycles until a wire or die-attach bond lifts. Mobile-ion contamination from a poor-grade compound can shift device thresholds, which is why ionic content is specified. And once moisture reaches the die, corrosion and electrochemical leakage across conductors degrade the part in the field. Each mode is designed against with filler tuning, MSL handling, and clean feedstock, but none is eliminated.
FAQ
What are plastic IC packages made of?
Four material systems: an epoxy molding compound body filled 70 to 90 percent by weight with fused silica, a copper-alloy or Alloy 42 leadframe that carries signals out, a die-attach adhesive bonding the chip down, and gold, copper, or silver bonding wire connecting die pads to the leads. The filled epoxy gives the package its shape, strength, and most of its limits.
Are plastic IC packages hermetic?
No. Plastic packages are non-hermetic, meaning moisture diffuses slowly through the bulk epoxy to the die. Their moisture permeability is orders of magnitude higher than ceramic or metal. This is why plastic parts carry a moisture-sensitivity level and must be baked after their floor life expires, and why the MIL-STD-883 hermeticity test does not apply to them.
What is the maximum temperature for a plastic IC package?
It depends on the compound and qualification, not a single number. Standard commercial parts run to about 85 to 125 degrees C; automotive AEC-Q100 Grade 0 qualifies plastic to plus 150 degrees C. The epoxy glass transition, typically 150 to 175 degrees C, marks where the body softens and expansion accelerates, so the package usually sets the ceiling before the silicon does.
What is the mold compound in an IC?
It is an epoxy molding compound, normally an epoxy cresol novolac resin heavily loaded with fused-silica filler. The filler, often around 80 percent by weight, lowers the compound thermal expansion and water absorption while raising its strength and heat conduction. It encapsulates the die and wires, providing mechanical protection and electrical insulation.
Plastic or ceramic IC package — which is better?
Neither universally. Plastic wins on cost, size, weight, and part availability and covers commercial through automotive ranges. Ceramic wins on hermeticity, high-temperature stability, and thermal conductivity, and is standard for space, military, and high-temperature sensing. Choose ceramic only when hermeticity or extreme temperature forces it; otherwise plastic is the default.
Are plastic IC packages RoHS and halogen-free?
Most modern parts are both, but the terms differ. RoHS restricts lead and specific substances. Low-halogen under JEDEC JS709C means each plastic material holds under 1000 ppm each of bromine and chlorine from flame retardants. The stricter halogen-free board limit in IEC 61249-2-21 caps total halogens at 1500 ppm. Check the manufacturer material declaration for the exact claim.
What to do
Default to plastic and pick the parts deliberately: choose the AEC-Q100 grade that covers the ambient at the package, confirm the MSL and bake handling before your line touches the reel, and read the material declaration for the RoHS and low-halogen claim you actually need. Weigh the leadframe question only when thermal or dimensional stress is real, copper for heat, Alloy 42 for expansion match. Specify ceramic or a hermetic package only when sustained temperature above the compound’s Tg, a hard moisture-reliability target, or a program requirement demands it, and check whether the MIL-PRF-38534 Appendix D or Class Y non-hermetic routes let a qualified plastic part serve instead.