What is injection molding?

This schematic illustrates the process of injection molding. A hot, molten polymer is injected into a cold mold. A screw apparatus can be used to inject the polymer into the mold, as shown in the schematic. After the part cools and solidifies, the mold is opened and the part is ejected. No chemical reaction occurs during the molding process. (Any reaction that occurs would be a degradation reaction, which should be avoided!)

Typical products
Injection molding is widely used in a wide variety of industries: automotive, appliance, computer, communication and industrial equipment. Examples of injection molded parts in the automotive industry include sail panels, radiator end caps, door panels, lamp housings, and fuel rails. Both amorphous and crystalline thermoplastic resins are used in injection molding. Short glass fibers are commonly used as reinforcements.

Automobile parts

Automotive Radiator Sail Panel End Cap

Fuel Rails with Injectors

Injection molding is a very old process. While it is not a "low tech" process, there are a lot of "low tech" items which are made via injection molding.

Basic Process Factors in Injection Molding

The basic process factors in an injection molding process can be categorized as: 1. Material Parameters o Amorphous, Semicrystalline, Blends and Filled Systems o Pressure-Volume-Temperature (PVT) Behavior o Viscosity 2. Geometry Parameters o Wall Thickness of Part o Number of Gates o Gate Location o Gate Thickness and Area o Type of Gates: Manually or Automatically Trimmed o Constraints from Ribs, Bosses and Inserts 3. Manufacturing Parameters o Fill Time o Packing Pressure Level o Mold Temperature

Melt Temperature

The performance of an injection-molded part is dependent on the interaction of these three groups of parameters.

Material Parameters
Polymer materials used in injection molding can be classified as amorphous, semicrystalline, blends and filled systems.

Amorphous, Semicrystalline, Blends and Filled Systems

The chains in an amorphous polymer are randomly distributed. Some examples of amorphous polymers are Polycarbonate, Polyphenylene oxide and Acrylonitrile Butadiene Styrene. In a semicrystalline polymer, chains are partially ordered to form crystallites. Some examples of such polymers are Polybutylene terephthalate and Nylon. Engineering polymers are often blends of two or more polymers. Polymers are frequently blended in a twin screw extruder Fillers such as short glass fibers, coloring agents, fire retardants etc. are frequently added to the polymers to enhance their mechanical properties or other performance characteristics. Semicrystalline polymers have the highest shrinkage followed by blends and amorphous polymers. Systems with fillers have the lowest shrinkage.

Pressure-Volume-Temperature (PVT) Behavior

PVT behavior refers to the change in specific volume with temperature and pressure changes.

The specific volume is defined as volume per unit mass. The specific volume, v of a polymer changes with variations in temperature and pressure. Volumetric expansion data for polymeric materials are obtained under equilibrium. Such data represent fundamental thermodynamic properties of the material and reflect the transitions as the material moves from glassy to crystalline to melt state. In the figure, as the pressure and the temperature change from P and T to P ' and T', the volume of the same mass m changes from V to V'. P V T data can be measured using standard equipment. The specimen is heated in an enclosed cell and the change in its volume is measured when it is subjected to a range of pressures. Specimens may be either in the form of polymer pellets or they may be cut from the molded plaques. PVT behavior of materials plays a critical role in relating processing to the final part performance.

The shrinkage of a molded plastic part can be as much as twenty percent by volume when measured between the processing temperature and the ambient temperature. Semi-crystalline polymers have higher shrinkage than amorphous polymers because of the ordering and folding of chains in a semicrystalline polymer below its freezing point. This leads to a greater difference in specific volume ( v) between the melt phase and the solid phase for semi-crystalline materials. The presence of fillers like talc or short glass fibers reduces the difference in specific volume ( v) between the melt phase and the solid phase.

Viscosity
Viscosity behavior of materials is important in determining the flow length and the amount of viscous heating generated during the melt flow.

Most polymer melts exhibit shear-thinning behavior, which translates to lower, viscosity with higher shear rate. Hence the viscosity of the melt varies across the thickness of the part due to the variation in shear rate. Melt viscosity decreases with temperature but the sensitivity varies among thermoplastics. For example, the viscosity of polystyrene and polypropylene are considerably more sensitive to temperature than that of polyethylene. At pressures of several thousand p s i the viscosity increases with pressure. The presence of fillers increases melt viscosity. The viscosity is critical for determining the injection pressure with a given rate or the flow length with a given maximum pressure. The viscosity of a polymer melt is adequately described by the Cross-model. This model treats viscosity as a function of temperature (T), pressure (P), and shear rate ( ). This model handles both the Newtonian and the shear thinning flow regimes. The transition between the two regimes is characterized by t*. t* is the shear stress at which shear thinning behavior begins to manifest itself. The slope of the shearthinning region can be characterized in terms of a shear thinning index, n. This model is often adopted for simulating the filling stage of injection molding.

Geometry Parameters
Geometry parameters in part design and mold design are the wall thickness of the part, the number of gates, gate location, gate thickness and area, the type of gate and constraints from ribs, bosses and inserts. Three factors primarily govern the choice of wall thickness for a part

1. part design for stiffness 2. cooling time 3. flow length

The deflection of a part under a given load is a strong function of the wall thickness, decreasing as the wall thickness is increased. A ribbed part can meet part stiffness specifications with a lower wall thickness.

A major issue in injection molding part design is to determine nominal wall thickness. The wall thickness depends on the structural requirements, the gating scheme, process window and the material in use. Specific structural requirements determine the minimum wall thickness. The material viscosity determines the achievable flow length. Standard data charts of recommended wall thickness as a function of flow length are available as shown here. The gating scheme and the process window can be manipulated to achieve the desired flow length for a given material and the nominal part thickness.

Mold cooling accounts for more than two thirds of total cycle time. Uniform cooling improves part quality by reducing residual stresses and maintaining dimensional accuracy and stability. Cooling time is a function of mold wall temperature, melt temperature, material properties and part wall thickness.

The equation shown here gives a rough estimate of the minimum cooling time needed before part ejection. is the thermal diffusivity of the material, h is the plate thickness, Tw is the mold wall temperature, capital Tm is the melt temperature, capital Te is the ejection temperature. An example calculation has been shown here with typical values for the different variables. In the example, the minimum cooling time for the part centerline to reach the ejection temperature is calculated to be 23 seconds.

This plot shows the variation of estimated cooling time with the wall thickness of the part. Cooling time increases in a non-linear fashion with increasing part wall thickness. The cooling time for a semi-crystalline material like Polybutylene Terephthalate is always higher than that for an amorphous material like a blend of Polycarbonate and ABS.

Design Guidelines for Gates The gate provides the connection between the runner and the molded part. It must permit enough melt to enter and fill the cavity during filling and packing stages. Broad design guidelines for gate design have been summarized here: Gate Location The gate should be located at thicker sections of the part. The gate location must allow expulsion of air towards the vent. Multiple gates should be located to allow weld lines to form at appropriate positions in the part.

Gate Size Gates must be sized such that they freeze off after sufficient packing. Jetting of polymer melt should not occur. The gate thickness is usually 50% to 75% of the part thickness.

Number and Location of Gates

Multiple gates may be used to control the pressure distribution within the cavity. This will introduce weld lines. Weld lines are formed when two melt flow fronts meet. Such regions are weaker than the other regions of the part. Gate locations and the number of gates affect the severity and location of weld lines. They also affect the dimensional stability of the molded part. This example shows the formation of a weld line at the center of a rectangular part filled with two gates:

Multiple gates are often used in injection molding: 1. to avoid the limitation in flow length imposed by machine pressure and the gate area. 2. to reduce the variations in pressure distribution inside the part, which in turn reduces the nonuniform, pressure shrinkage. 3. to reduce the cycle time or to increase the production rate. 4. to mold parts with variations in part thickness. The number of weld lines formed with multiple gates and inserts is given by, N = C + G - 1 where N is the number of weld lines, C is the number of insert cores and G is the number of gates.

The effects of different gate locations on filling patterns (left) and the final states of the molded part on the (right) are shown on the two sides of the figure. The center-gated mold produces a radial filling pattern with a non-uniform pressure distribution. This leads to a warped final part. The design with two edge gates reduces the extent of pressure variation across the part. However, it introduces a weld line at the center of the part. The single edge gated design is better if the flow length of the material is adequate. Although, there is greater potential for uneven shrinkage in this case, there's no weld line in a single gated part.

Types of Gates
After selecting the location and size, we must decide on the type of gate. Gates are classified in two categories based on the method of degating Manually Trimmed: 1. 2. 3. 4. Direct of Pin Gate Tab Gate Edge Gate Fan Gate

Automatically Trimmed: 1. Pinpoint Tab Gate 2. Submarine or Tunnel Gate

Manufacturing Parameters
Other design parameters in the manufacturing of an injection molded part are fill time, packing pressure, mold temperature and melt temperature.

A gate is an inlet port to the mold cavity and it provides a connection between the runner and the mold cavity. The gate being thinner than the part usually freezes first during packing. No more material can be injected after the gate freezes off.

There are two opposing factors that govern non-isothermal melt flow inside a mold cavity. Melt close to a wall cools faster as the mold wall temperature is much lower than the melt temperature. This leads to a frozen layer close to the wall. However, the strain rates are higher near a mold wall leading to significant viscous heating. There are three distinct flow regions inside the mold cavity. 1. Near or at the gate 2. In the cavity - Hele Shaw Flow 3. Advancing front - Fountain Flow This figure shows schematically the different regions of flow in an end gated rectangular part: Complex entrance effects and secondary flows may be present in the vicinity of the gate. In the fully developed flow region, the flow pattern does not change with distance. This flow can be described by a two dimensional, creeping flow of an inelastic fluid in a narrow gap H --- also termed Hele-Shaw flow. In the vicinity of the flow front, there is a fountain effect by which the material from the center is carried to the outer section of the cavity.

This figure shows a plot of maximum injection pressure against the fill time for a center-gated radiator end cap made of short glass fiber reinforced Nylon 6 6. The filling is done at a constant rate of injection. The melt temperature is two hundred seventy degrees Celsius and the mold wall temperature is 75 C. To the left of the minimum in the curve, isothermal viscous flow dominates and the injection pressure drops as the fill time rises. But to the right of the minimum, heat transfer is controlling; material freezes up inside the cavity and the injection pressure rises again. In this case a fill time of 1.5 to 2 seconds is chosen.

The figure shows a typical temperature history in the cavity during the injection molding cycle. The gate freezes off first because it is thinner than the cavity. Once the part temperature is well below the polymer solidification temperature, the part is ejected.

Typical cavity pressure variations over the entire injection molding cycle are shown here. The cavity pressure rises rapidly during the filling stage, which is followed by the holding or the packing stage. Once the gate freezes off, the cavity pressure decays with time till the part is ejected. At high pressures of thousands of P S I, a polymer melt is compressible, allowing additional material to be packed in the mold cavity after mold filling is complete. This is necessary to reduce non-uniform part shrinkage, which leads to part warpage. Excessive packing results in a highly stressed part and may cause ejection problems whereas insufficient packing causes poor surface, sink marks, welds and non uniform shrinkage. The pressure distribution inside the mold cavity changes with distance from the inlet gate. This figure shows a simple part geometry with pressure variations among the points one, two and three respectively. Further away from the gate, pressure rises slowly and it decays quicker than at the points closer to

the gate. The pressure in the mold cavity Should be more uniform to minimize part warpage.

This plot shows the effect of packing on linear shrinkage: As expected, linear shrinkage decreases with increasing pressure.

Materials for injection molding

Material selection depends to a large extent on the functional constraints of the part as described earlier. Both amorphous and crystalline thermoplastic resins are used in injection molding. Short glass fibers are commonly used as reinforcements

Thermoplastic compounds commonly used in injection molding include:

Acrylonitrile butadiene styrene (ABS) * Acetal Acrylic Polycarbonate (PC) Polyester Polyethylene Fluoroplastic Polyimide Nylon Polyphenylene oxide Polypropylene (PP) ** Polystyrene (PS) Polysulphone Polyvinyl chloride (PVC)

* a high performance engineering thermoplastic ** copolymers of PP and polyethylene are expected to become increasingly common in applications in the future because the copolymer properties can be tailored. Material selection can be a complicated task. A wide variety of materials are used in common, everyday items:

Car bumper--polyurethane Bottle--high density polyethylene (HDPE) Compact disc--polystyrene Car body panel--polyester Intake manifold--nylon 6,6

The physical properties of the materials (density, thermal conductivity, melting temperature, etc.) must be considered in light of the required mechanical properties of the finished part (strength, stiffness, hardness, etc.). This schematic illustrates the performance spectrum of a variety of plastic materials.

Various additives may be added to injection molding compounds to accomplish various purposes. This table summarizes some of them. Additive Filler Plasticizer Antioxidant Colorant Flame retardant Stabilizer Reinforcement Function Examples increase bulk density calcium carbonate, talc, limestone improve processability, reduce product brittleness phthalate esters, phosphate esters prevent polymer oxidation phenols, aromatic amines provide desired part application color oil-soluble dyes, organic pigments reduce polymer flammability antimony trioxide stabilize polymer against heat or UV light carbon black, hydroxybenzophenone improve strength E-glass, S-glass, carbon, Kevlar fibers

Processing temperature is an issue in material selection for injection molding.

Polymer physical properties dictate temperature processing window (Tg < Tproc < Tdeg) The operating temperature must lie in the range between the glass transition temperature and the degradation temperature of the polymer. Example: Nylon 6,6 has its glass transition at 240 C - 265 C, but its suggested processing temperature range is 270 C - 305 C. Degradation occurs at temperatures above 350 C or 400 C. The maximum use temperature of nylon 6,6 parts is about 150 C. (Note that this is considerably below the glass transition temperature.)

Example: Material Selection Guidelines for Radiator End Cap

Material Selection Guidelines Fatigue Strength Semi-crystalline e.g. Nylon 66 is superior Impact Strength Polycarbonate is good but it has poor fatigue strength High Temperature Solvent Resistance Cost Phenolics have good high temperature performance, but they have poor impact strength Phenolics, PPS, Nylon Phenolics < Nylon < PPS

Semi-crystalline polymers like nylons are superior to amorphous polymers in fatigue strength. Some amorphous polymers like polycarbonate have good impact strength. But polycarbonate has very poor fatigue strength and it is notch sensitive. For high temperature environments, phenolics offer a good option: however the impact strength of phenolics is very low. For solvent environments, phenolics, polyphenylene sulfide and nylon are good choices. Finally, from the cost point of view, phenolics are cheaper than nylon and nylon is cheaper than polyphenylene sulfide. Nylon is chosen as the material for the radiator end cap.

Design for Manufacturing

The conventional approach to mold design for injection molding is based on allowances for post mold shrinkage of the part. Experienced designers account for possible warping by allowing "windage" in tool design. The part is subsequently molded and examined to check whether it meets specifications. This iteration involves modification of the existing mold repetitively until the part is within specifications. This procedure results in long and expensive product development time.

The approach discussed in this tutorial accounts for interactions among part design, mold design and part performance at the design stage to arrive at a mold design. The tool is cut only after detailed design to mold the part according to specifications. This reduces the product development cycle time and cost by eliminating the iterations on tool geometry

The injection molding cycle can be divided into several stages. They are mold closing, filling, packing, cooling, mold opening and ejection as shown here.

Cooling time is a major fraction of the total cycle time. The first step in the design for manufacture of an injection molded part is to establish the part specifications. This includes defining various constraints: 1. Functional Constraints 2. Assembly Constraints 3. Performance Constraints An example of such constraints is discussed for a radiator end cap. Functional requirements of the part impose certain restrictions on the part geometry. In the case of a radiator end cap, it must fit at the end of the radiator core with a flange, second, it must provide hook ups for coolant inlet and outlet, and it must have provision for coolant drain. Assembly constraints arise due to the interaction of the radiator end cap with other parts. The radiator end cap must interface with the radiator core and the coolant hose. The molded radiator end cap must be dimensionally stable in terms of shrinkage and warpage. Performance constraints for a part are established based on the mechanical performance expected from the part. These constraints play a critical role in selecting materials as well as in establishing part geometry and part thickness. The radiator end cap must withstand a cyclic thermal variation from -30 degrees to 300 F and an internal pressure load of 24 P S I G. Part design for a radiator end cap must consider the effect of internal pressure load. The material selection must take into account the effect of solvent and moisture on the mechanical properties of the material.

Cycle time in injection molding

The injection molding cycle can be divided into several stages. They are mold closing, filling, packing, cooling, mold opening and ejection as shown here. Cooling Time is a major fraction of the total cycle time.

More about "pack and hold" time More about cooling time

This graph shows a typical pressure profile for injection molding. In this example, the cycle time is 35 seconds. Filling requires just a few seconds. A slight increase in pressure occurs when the mold is closed. Filling raises the pressure; during packing the mold is held at this pressure. When gate freeze off occurs, melt can no longer be forced into the cavity and the pressure drops. When the part cools below its glass transition temperature, the mold can be opened and the part ejected. At this stage, the pressure drops to ambient level.

Relationship between fill pressure and fill time

This graph shows a typical temperature profile for injection molding of thermoplastic ABS. Two curves are shown. The solid line represents the temperature inside the part, and the dashed line represents the temperature near the gate. The glass transition temperature for ABS is 100 C to 110 C. When the gate temperature drops below this level, gate freeze-off occurs. The inside of the part cools more slowly than the gate section. When the centerline temperature reaches the ejection temperature, the part can be ejected.

More about "pack and hold"

A part is "filled" when molten polymer has flowed into all the extremities of the mold. Packing squeezes more polymer into the mold after filling is finished. The pressure of the melt holds the cavity pressure high. Packing thus takes advantage of the slight compressibility of the melt to push as much polymer into the mold as possible. This is done because cooling introduces shrinkage. Care must be taken to design the mold such that gate freeze-off does not occur too early, because no further packing can occur after gate freeze-off.

More about cooling time

Mold cooling accounts for more than two thirds of total cycle time. Uniform cooling improves part quality by reducing residual stresses and maintaining dimensional accuracy and stability.

Cooling time is a function of : mold wall temperature melt temperature material properties part wall thickness The equation shown here gives a rough estimate of the minimum cooling time needed before part ejection. is the thermal diffusivity of the material, h, is the plate thickness, capital Tw is the mold wall temperature, capital Tm is the melt temperature, capital Te is the ejection temperature. An example calculation has been shown here with typical values for the different variables. In the example, the minimum cooling time for the part centerline to reach the ejection temperature is calculated to be 23 seconds.

This plot shows the variation of estimated cooling time with the wall thickness of the part. Cooling time increases in a non-linear fashion with increasing part wall thickness. The cooling time for a semi-crystalline material like Polybutylene Terephthalate is always higher than that for an amorphous material like a blend of Polycarbonate and ABS.

This table gives thermal data for various unfilled polymers. Material k [W/m K] Cp [J/kg K] [kg/m3] Tm [C] Tw [C] Te [C] ABS 0.264 1314 1040 238 57 82 PC 0.190 1298 1200 302 82 118 PC/ABS 0.246 1252 1120 260 79 93 PBT 0.264 1741 1310 243 41 116 NYLON 0.250 4400 961 260 85 170

For fiber reinforced compounds, k increases compared to the unfilled polymer. This decreases the cooling time.

Relationship between fill pressure and fill time

This graph shows the relationship between the fill pressure and the time required for fill calculated for a bumper endcap. The minimum which occurs in the graph is typical of most fill pressure/fill time relationships. The minimum separates two regions of fill behavior. In the first region, the fill behavior can be predicted from isothermal analysis. Viscosity controls the fill pressure. In the second regime, heat transfer controls fill pressure. A significant amount of cooling takes place during fill in this region. Because of this, the flow cross section is being constricted during fill. The resin viscosity also goes up during fill. These two changes act to increase the required fill pressure. The "ideal" regime is marked on the graph. However, it is not always possible to operate in this region because of possible problems with getting the desired morphology. For example, high crystallinity is desired, very fast fill can cause low crystallinity and porosity on the polymer surface.

This graph compares fill pressure/fill time curves for unreinforced and glass fiber reinforced nylons. The fillers cause an additional effect here related to thermal diffusivity and thermal conductivity. The top curve is for glass fiber reinforced nylon. It has roughly the same viscosity characteristics at 270 C as does

unreinforced nylon at 290 C. The thermal conductivity of the glass fiber reinforced nylon makes the curve a little steeper. Except for that, the shapes are about the same.

Flow in injection molding

Flow in injection molding is non-isothermal.

There is a frozen layer near the mold wall which develops with time. Therefore, the cross section for flow changes with time. Viscous heating and mold wall cooling occur.

There are three distinct regions in the flow:

Near or at the gate--three dimensional flow which is typically not modeled by standard software packages In the cavity: Hele-Shaw flow (two dimensional flow in narrow gaps which is standard in software flow modeling packages) Advancing front: fountain flow

All three types of flow affect fiber orientation in different ways. More about:

The frozen layer Flow with shear

The frozen layer

The flow issues in injection molding are complicated because the thermoplastic polymer is cooling and solidifying as it flows. During mold filling, the flow profile is not a typical parabola because the polymer close to the surface contacts the cold mold and solidifies into a shell before the core does. This schematic illustrates:

Another issue in the flow and cooling characteristics is the effect of surface cooling on morphology. The layer of polymer close to the wall contacting the cold mold cools rapidly. (It solidifies within a few seconds). This rapidly cooled skin will have amorphous characteristics, while the core, which cools more slowly, might have a crystalline or semicrystalline center. The anisotropic cooling characteristics can present a problem when uniform properties are desired. Here is an computer modeling example of the "frozen skin" effect.

Example of frozen skin in IM

This picture shows a fairly simple injection molded part which has uniform wall thickness. The picture shows the flow front advancement over the five second fill time. The blue region was first to fill; the red region was last to fill.

This picture shows the predicted frozen layer thickness profile for the part. The red region is predicted to have a 70% frozen layer thickness. Conversely, the blue region is predicted to have only a 10 to 20% frozen layer thickness. Clearly, a great deal of variation in frozen layer thickness is possible in injection molded parts.

Flow with shear

Fluid viscosity changes when the fluid is sheared. The shear viscosity is modeled by the following relation:

, the viscosity in the absence of shear, is a function of temperature and pressure, as shown explicitly in the numerator of the above expression. The strain rate ( ) dependence of viscosity is indicated in the denominator. This graph shows a typical viscosity-strain rate dependence for shear thinning materials:
0

This graph shows shear stress/shear rate data for polyethylene at various pressures. Viscosity is significantly affected by pressure. The graph shows pressures ranging from 1 to 1000 atm. A tenfold change in shear stress at the same strain rate occurs between these two pressures, meaning that viscosity also increases tenfold.

Since typical injection molding pressures are about 10,000 psi (680 atm), the effect of pressure on viscosity is significant in injection molding applications.

Performance of Molded Parts

The performance of a molded part is gauged in terms of several criteria.

Part stiffness refers to the deflection under a load.This depends on the part geometry and the modulus of the material along the appropriate direction. The strength refers to the stress at which the part breaks or fails in static mechanical tests. The impact strength refers to the maximum energy that can be absorbed by the part before it fails. The fatigue strength refers to the failure stress under repeated loading. Dimensional stability refers to a comparison of the molded part shape after de-molding and over time, with the mold shape; this was discussed in the previous module.

A typical stress versus strain curve is shown here for a glass fiber filled thermoplastic material tested in tension along the mold flow direction. The modulus is given by the slope of the stress strain curve at small strains, i.e. in the linear elastic region. The tensile strength of the material is the stress at which it breaks. The modulus as well as the strength drop at elevated temperatures. The elongation to failure or failure strain is usually a good indicator of the toughness of the material.

For example, at room temperature, the modulus of a glass fiber reinforced polybutylene terephthalate along the flow direction is 8 billion Pascals or 8 giga Pascals. The tensile strength of the same material along the flow direction is 112 million Pascals or 112 mega Pascals. The strain to failure for this material is 0.034 or 3.4 %. The tensile modulus of steel is much higher -- 207 giga Pascals; but the ratio of modulus to density is comparable.

Impact
The impact strength of a material refers to its capacity to absorb energy in high speed collisions with other objects. The units are energy per unit area (Joules per meter squared). This is quite sensitive to the impact velocity, and to the masses of the objects. Typical Strain Rates in Impact Event The rates of strain experienced by two different automotive components in collisions at two different speeds are compared in the following table. An automotive bumper colliding at 5 miles per hour experiences strain rates of 10 to 100 % per second. A dashboard colliding at 15 miles per hour experiences strain rates of 100 to 1000 % per second. Impact testing must be carried out at strain rates relevant to the application. Impact Event 5mph automotive bumper impact 15mph automotive instrument panel (dashboard) impact Strain Rate ( % /s) 10 - 100 100 - 1000

Impact and Part Performance

Materials like polycarbonate that are ductile at room temperature have a higher impact strength than materials like polystyrene that are brittle at room temperature. Long chain polymers are more flexible and have a higher impact strength than shorter chains. The impact strength also depends on the geometry of the part through the moment of inertia. Features like ribs that stiffen the part can also lower the impact strength of the part, as shown for the two materials: Polycarbonate Geometry Flat Disk Disk with rib Geometry Flat Disk Disk with Rib Failure Mode Ductile Ductile Polyetherimide Failure Mode Ductile Brittle Failure Load (lbs) 2100 1800 Failure Load (lbs) 3400 400

The geometry of the two different types of test specimens used to obtain the data is shown in the following figure. The ribbed part is stiffer but will have lower impact strength.

Fatigue

What is Fatigue? Fatigue is Repeated Loading

Cyclic or repeated loading can cause failure at lower stresses than static loading. This aspect is central to fatigue performance. The highest frequencies with low amplitudes are characteristic of noise and vibration studies while the lower frequencies with moderate amplitudes represent classical fatigue testing. Molded test specimens or samples cut out of the molded part are tested in a chosen mode -- tensile or flexural -- for thousands or millions of cycles. The yield stress for a given number of cycles is termed the fatigue strength. The fatigue life of a part is the number of cycles to failure at a given fluctuating load.

High Cycle/ High Rate/ Low Amplitude o Noise and Vibration Low Cycle/ High Rate/ High Amplitude o Impact Fatigue High Cycle/ Low Rate/ Moderate Amplitude o Classical Fatigue

S-N Curve

S: Stress amplitude N: Number of cycles to break S is plotted against log N

S-N data can be used reliably for design only if the test conditions for generating S-N data match the service conditions for the component. The most critical choice for tests is between load controlled or displacement controlled cyclic loading. Other test variables are temperature, mean stress, amplitude of fluctuation and frequency. Elevated temperatures hasten failure. Higher frequencies shorten the test time. Hence the test frequency should be as high as possible without having significant heating due to viscous dissipation in the polymeric material. Such heating is less of a concern with fiber filled plastics than with neat plastics.

The breaking stress is lowered as the number of cycles is increased. Semi-crystalline polymers such as PBT have a much more gradually varying S-N curve than amorphous polymers like polycarbonate, as seen on the figure. This means that semi-crystalline polymers perform better with respect to fatigue.

Case Study: Radiator End Cap

We now turn to a design case study of a radiator end cap molded from 33 wt% glass fiber reinforced nylon 66. The effect of non-uniform fiber orientation and anisotropy in mechanical properties on the stresses in service and the fatigue performance will be illustrated with this example. The failure modes and fatigue strength of the part will depend on the varying fiber orientation and loading along the part. The moduli and other thermo-elastic properties are anisotropic and will depend on direction. The radiator end cap is clamped with a flange at the top of the radiator core. It has four different ports for inlet, outlet and draining. The part is 471.5 millimeters long and 57 millimeters wide. The wall thickness is 2.54 mm except for the flange where it is 3. 8 mm. Sixteen ribs are spaced equally along the part, each of them being 5.08 mm high and 1.90 mm thick. This long part is filled through a central circular gate of 6.35 mm diameter at the top. The flow pattern is radial near the gate and longitudinal after several ribs, on either side of the gate.

The flow and crossflow moduli for the 33 wt.% glass fiber reinforced Nylon 66 were taken from tensile data obtained on plaques molded from this material. The plots show that the moduli drop by several fold as the temperature increases from 25 to 75 degrees Celsius. The anisotropy ratio increases only slightly by comparison. Moisture will reduce the moduli further to 2/3 of the dry values.

This still picture shows a side view of the final state of fiber orientation in the shell.

Stress Analysis Design

The maximum stress location is predicted to be far from the gate where the fiber orientation is across the rib. This area must be further reinforced to avoid failure. Anisotropic and non-uniform material properties must be accounted for.

This photograph shows a radiator end cap that has developed fatigue cracks across the ribs at one end. The ribs in this region are positioned perpendicular to the melt flow direction in the base of the part.

Experimental tests confirm that the crossflow material properties dominate the fatigue performance of the part when it is molded with a pin gate:

See the "Developing Fiber Orientation During Fill of a Radiator Cap" video. (This movie shows the developing fiber orientation in the shell during mold-fill of a radiator end cap. Near the pin gate, the fibers orient downward along the rib direction. A few ribs away from the gate, the flow direction is turned at an angle of 45? to the ribs. Further downstream, the shell fiber orientation is along the length of the part and at right angles to the ribs. )
The maximum stress location is predicted to be far from the gate where the main flow and the fiber orientation in the base is across the rib. Experimental observations bear out this result of the calculation. This area must be further reinforced to avoid failure. Most injection molded parts with chopped fiber reinforcement fail at locations where the cross flow properties are dominant. So, anisotropic and non-uniform material properties must be taken into account in a realistic analysis of part performance.

How flow develops microstructure

Most of the flow in injection molding can be described by the Hele-Shaw model, which describes twodimensional flow in narrow gaps. Most predictions and models for melt flow advancement are based on Hele-Shaw flow.

This diagram shows flow at the fill front. The thickness direction is enlarged. The melt advances through the cavity by what is called fountain flow. A no-slip condition holds along the wall. Instead, the melt advances along the wall by spilling over from center of the molding. The big impact of fountain flow in injection molding is the kind of fiber orientation it induces. At the middle of the flow front, the melt stretches parallel to the front, causing fibers to be aligned parallel to the flow front. (i.e., they are aligned across the flow, perpendicular to the wall.) The fibers along the wall, however, are aligned parallel to the wall. There are two basic guidelines for flow induced alignment:

Shear flows align fibers in the direction of flow. Stretching aligns fibers in the direction of stretch.

Converging and diverging flow provide examples of these guidelines. In converging flows, the direction of strength corresponds with the direction of flow. Thus, the fibers end up parallel to the flow. For diverging flow, the stretch is biaxial stretch, but more fibers are aligned perpendicular to flow. For flow in a center-gated disk (radial flow), the melt is stretched as it flows radially so the fibers end up aligned as shown in the diagram. The flow pattern during mold fill in injection molding, governs the fiber rotation. Shear flows align fibers in the direction of the flow while stretching flows align fibers in the direction of stretch.

For example, in diverging flows, the fluid is stretched bi-axially in the plane perpendicular to the flow direction and the fibers tend to be aligned in the cross flow direction. But in converging flows, the fluid is stretched along the flow direction and the fibers tend to be aligned in the flow direction.

The shell-core structure

Crystallinity Fiber Loading Fiber Aspect Ratio Fiber Orientation Distribution Skin/Core Ratio Weld Lines

The local characteristics of microstructure in an injection molded part are the degree of crystallinity of the polymer, the fiber aspect ratio and the fiber orientation distribution. These may vary from location to location in the molded part. In particular, the fiber orientation varies across the thickness of the part, forming a skin-core structure. The fibers are aligned out of the plane in the core and aligned more or less with the flow near the mold wall. The two outer layers -- called the skin -- typically amount to 75 % of the thickness for injection molded parts. Weld lines are formed by the meeting of two flow fronts; the part is usually weaker near the weld lines than away from them. The fountain flow and resulting fiber orientation characteristics lead to the shell-core structure in IM parts. Fibers in the core are aligned transverse to the flow, perpendicular to the fibers in the shell. The shell fibers are aligned with the flow. This diagram shows the shell-core microstructure. The core is exaggerated in this diagram. A typical core in IM parts is quite small compared to the shell, usually occupying only 30% of the part volume. The above diagram is a simplification; in reality there are more layers which can be distinguished in an IM part. For instance, melt near the wall freezes very quickly. The resulting frozen skin has a somewhat different fiber orientation than does the shell.

This plot shows the shell-core structure in injection molded parts. The orientation parameter a11 is graphed as a function of distance from the center of a part. The closer a11 is to 1, the greater the flow-direction alignment. From the graph, it is apparent that the distinction between "shell" and "core" is somewhat

arbitrary. However, it is clear that fibers in the center of the part are strongly oriented transverse to the flow and fibers at the edges of the part are strongly oriented parallel to the flow.

By cutting samples from an injection molded plaque and testing their tensile moduli, the effect of the shell-core microstructure can be seen. This graph shows flow and crossflow tensile moduli as a function of fiber content. There is a significant property difference between the flow direction and the cross flow direction. The flow direction is stronger because most of the fibers (70%) are aligned with the flow. The modulus and strength of fiber reinforced polymers are anisotropic. They are higher along the flow direction because of fiber orientation closer to the flow direction. The degree of anisotropy may be measured by the ratio of the modulus along the flow direction to that in the cross flow direction. This ratio increases with fiber loading as seen in the data presented for glass fiber reinforced polystyrene.

Fiber orientation in IM
Basic characteristics of fiber orientation in injection molding:

Three dimensional orientation is possible Virtually no aggregation of fibers occurs Principal orientation direction is along the flow

Short Fibers are used to reinforce thermoplastics in applications which:

Involve high volume production Require moderate stiffness Require moderate loads

Thermoplastics are usually reinforced with chopped glass fibers for applications requiring high volume production. The shorter fibers can provide moderate stiffness and the molded parts can be used under moderate loads. Short fibers also improve the fatigue life of a part. The mode of failure is then controlled by fiber orientation.

Effect of Crystallinity on Mechanical Properties of Un-oriented PPS Film

The degree of crystallinity has an important bearing on mechanical properties. Higher crystallinity leads to higher modulus. For example, the tensile modulus of polyphenylene sulfide is increased by about 40% as the degree of crystallinity is increased from 0 to 35%. At the same time, the elongation to break goes down by a factor of more than five fold indicating that crystallinity reduces the polymer ductility and toughness. The triangular diagram is convenient for depicting trends in performance with composition. As we move away from the vertex marked filler, the cost rises. The strength rises as we move toward the vertex marked fiber. The degree of crystallinity may be inferred from density measurements. It increases upon annealing or with slower cooling. The rate of cooling in the molding cycle is determined by a variety of factors including local part thickness, the presence of ribs and the mold wall temperature. Thicker sections cool more slowly and develop more crystallinity. Property Percent crystallinity* Density, gm/cc Tensile modulus MPa Elongation at break, % Quick-Quenched, Unannealed 0 1.3094 1926 20.0 Quick-Quenched, Annealed at 200 C 30.2 1.3458 2574 4.8 Slow-Cooled, Unannealed 34.8 1.3514 2709 3.4

* From density measurements: ra = 1.314 g / cc, rc = 1.43 g / cc

Anisotropy
The degree of anisotropy is increased by increasing:

Fiber Aspect Ratio L/D Fiber Orientation Fiber Loading

Longer fibers and a greater extent of fiber orientation also, like greater fiber loading, lead to greater anisotropy in properties. The degree of anisotropy is greatest for perfectly aligned long fibers. But with perfectly aligned fibers, the material is isotropic in a plane, and termed transversely isotropic. This feature simplifies design and analysis to a great extent. Longer fibers are also more easily oriented in shear flows. However, the mean fiber aspect ratio (fiber length to diameter ratio) is decreased by the feed screw and by the high shear rates experienced by the polymer melt in the mold cavity.

Effect of Fiber Length on Properties

Experimentally measured values of tensile modulus and tensile strength for two different Polybutylene terephthalate molding compounds are presented. Both the modulus and the strength are significantly higher for the molding compound with longer fibers on the average. Glass Fiber Reinforced Polybutylene terephthalate Average Fiber Tensile Modulus Tensile Strength Length (GPa) (MPa) 154 m 8.40 0.22 114.12 0.87 221 m 9.60 0.15 136.00 0.71

There is a distribution of fiber lengths in any molding compound. This distribution shifts to smaller lengths in regrind -- polymer obtained from recycled products by grinding. Typical distributions of fiber length for three different compounds are shown here. The average fiber aspect ratio -the ratio of length to diameter -- is about 27 for the nylon molding compound and only half as much for the polybutylene terephthalate and polycarbonate molding compounds.

It is particularly important to understand trends in tensile modulus because it is a good indicator of trends in other properties such as fatigue strength. The Halpin-Tsai rule is a simple and effective model for relating the tensile modulus of aligned discontinuous fiber composites to various parameters. Plots based on this rule are presented here. These plots show that greater fiber loading and a greater ratio of fiber modulus to matrix modulus magnify the effect of fiber aspect ratio on the tensile modulus for the aligned, chopped fiber composite.

An injection molded fiber reinforced plastic part is like a sandwich structure of two outer layers also termed the shell and one mid-plane layer termed the core. There is usually a distribution of fiber orientations in any of these layers. The shell is stronger in the flow direction because the fibers are oriented closer to this direction. The fibers in the core tend to orient along the cross flow direction, or be randomly oriented in the layer.

Fiber orientation parameter, fp

When the fibers are all in parallel planes - as is the case with thin layers, the distribution of fiber orientations can be represented by a single parameter fp. With a fixed reference direction, such as the bulk flow direction, this parameter ranges from negative one to one. This parameter is 0 when all orientations are equally likely, and 1 when all the fibers are perfectly aligned along a reference direction. When the fibers are all aligned at right angles to the reference direction, fp is negative one. The images presented here correspond to various values of this parameter between 0 and 1, i.e. between random orientation and perfect alignment. The range from 0 to 1 is sufficient to describe the variation in properties with fiber orientation.

fp = 0.0 random

fp = 0.3 slightly aligned

fp = 0.6 moderately aligned

fp = 0.9 highly aligned

The effect of fiber orientation in a layer on its thermo-mechanical properties can be seen by carrying out calculations with the help of a computer program from the University of Delaware. The plots here show the predicted modulus along the flow direction and the predicted modulus transverse to the flow direction, for a layer of glass fiber reinforced nylon 66, as a function of fp . The fiber loading is fixed at 33 wt% and the fiber aspect ratio is fixed at 25. The longitudinal modulus is seen to increase with increasing orientation while the transverse modulus decreases with increasing orientation along the flow direction. The anisotropy ratio for the plane increases from 1 to over 3 as fp varies from 0 to 1. The plane of isotropy coincides with the plane of fibers when they are randomly oriented (fp =0). But the plane of isotropy becomes perpendicular to the plane of the fibers when they are perfectly aligned (fp=1). A fiber orientation experiment

An experiment
Here are the specifications for a fiber orientation experiment:

3" by 11" plaques were injection molded from polybutylene teraphthalate (PBT), polycarbonate (PC), and nylon 6,6 (polyamide, or PA) o PBT: GE Plastics Valox 420 (30% glass filled) L/D = 12.4 Tfill = 1.89 or 5.56 seconds (time was varied) Temperature = 280 C

PC: GE Plastics Lexan 3413 (30 % glass filled) L/D = 13.5 Tfill = 1.74 seconds Temperature = 343 C

PA: DuPont Zytel 70G33L (33% glass filled) L/D = 24.4 Tfill = 1.57 seconds Temperature = 306 C Side view and end view samples were taken from three locations. One column of photos, divided into ten slices in the z direction. o End view: ~60 fibers per slice o Side view: ~20 fibers per slice
o

This diagram illustrates the mold:

The rectangular mold has an end gate roughly half the thickness of the part. Samples were taken from three locations at different distances from the gate. The orientation distribution was characterized at each location. This graph illustrates the viscosity behavior of nylon. It is strongly shear thinning at all shear rates.

This graph shows the viscosity behavior of PBT. It is shear thinning only at high shear rates.

How do these differences in molding compound rheology and differences in fiber aspect ratio affect fiber orientation? Resins which are more shear thinning exhibit a sharper velocity gradient next to the wall. Most of the flow will have a smaller gradient. The shear thinning properties affect the shear strain rates seen in flow. Results:

Comparison between shell and core Summary of trends in fiber orientation

Comparison between the shell and the core

How the shell-core structure is affected by...

Distance from the gate Material type (fiber length) Fill time

The shell-core structure and distance from the gate

This graph gives the fiber orientation parameter a11 at several distances from the gate for a rectangular plaque molded from nylon. The shell-core structure is evident at all three locations. a11 is close to 1 in the core, indicating the core fibers are strongly oriented with the flow.

The shell-core structure and fiber length

This graph is a summary plot of the shell orientation for polycarbonate, nylon, and PBT plaques for a two second fill time. All materials exhibit roughly the same orientation in the shell.

This graph shows the core orientations for the three materials. Notice that the PC and PBT core orientations are close to random, while the nylon plaque shows a more transverse orientation in the core.

Fill time in the shell and core

This graph shows the effect of fill time in the shell for a PBT plaque. Lengthening the fill time has essentially no effect on fiber orientation in the shell.

This graph shows the effect of fill time in the shell for a PBT plaque. Lengthening the fill time leads to a greater degree for orientation in the core.

Summary of trends in fiber orientation

Two factors are varied in these experiments: viscosity behavior of the molding compound and the fiber aspect ratio. Since both factors can affect fiber orientation, the results are difficult to interpret. The reason the nylon material appears to be more orientable is believed to be the effect of the longer fibers used in the nylon molding compound. Summary of trends in fiber orientation for PA, PC, and PBT:

All plaques show thick shell with narrow core. Alignment in shell PA > PC PBT Random in-plane orientation in core (PBT, PC); more transverse orientation in PA The differences between materials reflect the fact that the L/D ratio for these fibers is 24 for PA and 13 for PBT and PC. Higher overall alignment with longer fill times (PBT)

Residual Stress
Shear stresses develop within the melt during the filling and packing stages. These flow stresses cannot relax completely when the part is cooled rapidly below the glass transition temperature, leaving residual stresses. Non-uniform cooling with gradients of temperature and crystallinity also lead to residual stresses in the molded part.

Through-Thickness Effect
Pressure and temperature are time-dependent and non-uniform through the part:

Different layers of plastic solidify under different pressure and temperature (different PVT paths) Different amounts of shrinkage in different layers develop residual stress

Residual Stress Development: Mechanism

Consider a hot pool of the polymer melt at a temperature above Tg, in a cool mold maintained at a temperature below the polymer Tg. As the polymer in contact with the mold cools below Tg, it transitions from a viscous fluid incapable of sustaining a stress, to an elastic solid with a load carrying capacity. As this polymer region near the mold wall cools through the glass transition temperature, its thermal shrinkage is unconstrained by the material surrounding it, since the surrounding material is still above the Tg. When the next volume element of the polymer cools below Tg, its thermal shrinkage is now constrained by the polymer material towards the mold wall. Hence, the outer layer sees an incremental compressive stress, while the inner layer of the polymer sees an incremental tensile stress. As this cooling process continues, a parabolic residual stress profile develops, with compressive stresses on the outside towards the mold wall and tensile stresses on the inside.

Surface layer cools below Tg and can shrink unconstrained.

The thermal contraction of the next layer to cool below Tg is constrained by the solid layer above it.

The final residual-stress profile through-the-thickness is parabolic in shape, with compressive stresses on the outside and tensile stresses on the inside.

Predicted Trends in Residual Stress

Before investigating the residual stress development in composites, we shall first look at residual stress development in neat amorphous and semi-crystalline polymers. The properties of polycarbonate are used as representative of an amorphous polymer, while Polyether-etherketone, or PEEK, is used as a representative of a semicrystalline polymer. The calculations show that as the cooling process continues, a parabolic residual stress profile develops, with compressive stresses on the outside towards the mold wall and tensile stresses on the inside.

Neat Polymers Short-Fiber Composites

Residual Stress - Neat Polymers

The specimen modeled is an infinite plate of thickness 2b, and at an initial temperature Tm. This plate is cooled symmetrically from both sides to a quench temperature Tf, with the heat transfer at the surface being characterized by a heat transfer coefficient, h.

Let us then consider the residual stress development in amorphous polycarbonate resin. This figure shows the effect of the quench temperature on the residual stress profile through the thickness of a 3mm thick slab of polycarbonate, which is cooled from a melt temperature of 275C. The solid line represents the residual stress profile for a quench temperature of 22C, while the dashed line represents the residual stress profile for a quench temperature of 120C. As is evident, the quench temperature has a significant effect on the residual stress profile. The lower the quench temperature, the more severe are the temperature gradients, resulting in higher residual stresses. The characteristic residual stress profile with compressive stresses towards the surface and tensile stresses towards the center, is self-equilibrating, that is, the area under the tensile portion of the profile is equal to the area under the compressive portion. The coefficient of thermal expansion, , has a profound effect on the residual stresses, as it is the thermal strains which result in the subsequent build up of thermal stresses. The higher the coefficient of thermal expansion, the larger are the thermal strains. Therefore, the residual stresses increase with increasing . Secondly, PEEK exhibits a bilinear coefficient of thermal expansion, with its value above Tg being twice as large as its value below Tg. If the modulus increase due to the presence of the crystallinity is substantial above Tg, then the stress buildup occurs in a region where the coefficient of thermal expansion is also higher. All of these factors lead to an increase in the residual stress levels.

The resulting parabolic residual stress profile in the semicrystalline PEEK is shown by the solid line in this figure. In order to determine the contribution of the crystallization process to the residual stress development, the residual stress profile was also determined after artificially suppressing the crystallization to obtain an amorphous material. The resulting residual stress profile for this case is shown by the dashed line. It can be seen that the crystallization process increases both the compressive stresses at the surface as well as the tensile stresses at the center. We shall now explore the mechanisms for this crystallinity-induced increase in residual stresses. Another issue revolves around the contribution of the crystallization shrinkage to the residual stress development. This figure shows the resin modulus and fraction crystallinity as a function of temperature as the polymer is slowly cooled from the melt temperature. Crystallization occurs well above Tg, before the resin modulus starts to build up. Therefore, the shrinkage associated with crystallization occurs in a region where the resin modulus is still very low. Also, the crystallization shrinkage strains are very low compared to the thermal shrinkage strains. Hence, the role of crystallization shrinkage in the residual stress development in neat PEEK is insignificant. In summary, the thermal history plays a dominant role in the development of residual stresses, and the residual stress levels increase with an increase in the thermal gradients through the thickness of the material. Therefore, any processing factors which increase the thermal gradients will result in an increase in the residual stresses. The residual stresses were found to increase with decreasing quench temperatures, but are relatively insensitive to the melt temperature. The residual stress levels decrease with decreasing coefficient of thermal expansion, elastic modulus and thickness, and with increasing thermal conductivity.

Residual Stress - Short-Fiber Composites

We now turn our attention to residual stress formation in short fiber reinforced composites. As before, we shall first investigate residual stress formation in glass-reinforced amorphous polycarbonate composites, followed by glass-reinforced semicrystalline PEEK composites.

It is well established that injection molded composites typically exhibit the skin-core microstructure shown here schematically. In flat, injection molded sections, the fibers in the skin region tend to align along the flow direction, while the fiber orientation in the core region depends on the mold geometry. We shall assume for the purpose of this analysis, that the fibers in the skin region are aligned parallel to the flow, while the fibers in the core region are aligned perpendicular to the flow. The material anisotropy introduced by the fiber orientation effects will give rise to different residual stress distributions in the flow direction and in a direction transverse to the flow. Finally, we look at the effect of crystallization on the residual stress development in a skin-core microstructure. The solid line shows the residual stress profile in the flow direction under slow cooling with uniform crystallinity through the thickness. In the flow direction, the skin layer has a higher modulus and a lower coefficient than the core. The contraction of the core is therefore constrained by the skin layer, resulting in compressive residual stresses on the surface and tensile stresses in the interior. As before, to determine the contribution of the crystallization process to the residual stress development, the residual stress profile was also determined after artificially suppressing the crystallization to obtain an amorphous matrix material. The resulting residual stress profile for this case is shown by the dashed line. It can be seen that the crystallization process increases both the compressive stresses at the surface as well as the tensile stresses at the center. In a direction transverse to the flow, the residual stress profile is reversed, with tensile stresses on the surface and compressive stresses in the interior. This occurs because, in the transverse direction, the skin layer has a lower modulus and a higher coefficient of thermal expansion than the core. The contraction of the skin layer is therefore constrained by the core, resulting in tensile residual stresses in the skin and compressive stresses in the core. The crystallization has the same effect of increasing the residual stresses.

Residual Stress and Warpage

Thus far, we have restricted our attention to cases where the thermal history, the material properties and the resulting residual stresses are symmetrical about the mid-plane of the specimen. But asymmetry in the microstructure and thermal history can result in an asymmetric residual stress profile. Such an unbalanced residual stress profile leads to out-of-plane deformation or warpage, of the composite. As an example, let us consider a 4mm thick composite section with the skin-core microstructure. As before, the fibers in the skin region are aligned parallel to the flow or in the x-direction, while the fibers in the core region are aligned perpendicular to the flow or in the y-direction. The skin thickness at the bottom is fixed at 1mm, while the thickness of the top skin is varied between 0 and 1mm.

A skin thickness of 1mm represents the symmetric case with no out-of-plane curvature or an infinitely large radius of curvature. As the top skin thickness is reduced, the amount of asymmetry in the geometry increases, resulting in large warpage in both the x- and y-directions. This increase in warpage is characterized by a smaller radius of the out-of-plane curvature. Note that the direction of the out-of-plane curvature is opposite in the x-and y-directions, due to the anisotropy of the skin and core layers.

Residual Stress Trends in Composites

Thus, we have seen that the presence of a reinforcing glass phase, the crystallization process and the skincore microstructure typical of injection molded composites, all increases the residual stress state in the material. Further, an important observation was even on slow cooling, significant residual stresses can persist in materials with the skin-core microstructure. Finally, asymmetry in the microstructure and thermal history can result in an asymmetric residual stress profile, which in turn leads to out-of-plane deformation or warpage, of the composite.

Glass reinforcement phase increases the level of residual stresses. Crystallization increases residual stress. Skin/core microstructure results in higher levels of residual stress. Residual stresses in skin/core microstructures persist even in the absence of thermal gradients (isothermal cooling). Asymmetry in the microstructure and thermal history can result in warpage.

Gate locations

This diagram is a simple illustration of how to gate and how not to gate. If a part with sections of varying thickness is gated in the thin part, the thin part freezes off too soon to allow the thick part to be sufficiently packed. Severe shrinkage occurs in the thick part and warpage results. Gating in the thick part allows both parts to be adequately packed before gate freeze-off, allowing for more uniform shrinkage.

Here is a more detailed example of the effect of gate location.

This diagram is an idealization of an automotive sail panel. The part is 100 cm long and exhibits a curvature. Two different gate locations were used in doing shrinkage calculations for this part. The original gate location is shown on the diagram.

This diagram shows the new gate location.

This diagram illustrates the shrinkage and warpage predicted for the sail panel filled with neat polymer from the old gate location. The shrinkage has been exaggerated by a factor of ten for visual purposes.

The shrinkage and warpage predicted with the new gate location are shown here. Nonuniform shrinkage occurs with this gating arrangement, causing the part to be warped. A curvature is introduced into the part shape.

The melt advancement profiles explain how the gating arrangement affects the shrinkage characteristics of the sail panel. This picture shows the melt advancement for the original gate location. Red represents the last region to fill. The thin part is the last to fill so maximum shrinkage occurs there, causing the shortening of this part seen above.

Here is an analogous diagram showing the melt advancement for the new gate location. Both ends fill last, so both ends experience the greatest shrinkage. Since the central region of the part shrinks less, warpage occurs as shown above due to nonuniform shrinkage.

Multiple gates
Reasons for multiple gated filling:

If there is a flow length limit for given machine pressure and reasonable gate area To reduce non-uniformity of pressure and shrinkage If thickness variations and inserts are required, locate gates so they feed into thicker regions To reduce cycle time

The disadvantage to multiple gated filling is that using more than one gate causes weld lines in the injection molded part. The number of weld lines can be calculated simply. Given:

C, the number of insert cores G, the number of gates

Then the number of weld lines N = C + G -1 As mentioned above, use of multiple closely spaced gates can minimize pressure gradients. If a significant pressure distribution causes a problem in injection molding, use of multiple gates is warranted. An interesting side-note is the fact that the pressure distribution in packing is different from the pressure distribution at the end of fill.

Weld lines

Definition and causes of weld lines Strength issues with weld lines Minimizing the effect of weld lines

Causes of weld lines

Weld lines refer to weaker regions formed by the impingement of two separate flow fronts. These may originate from multiple gates as shown here. The fibers near an advancing melt front are parallel to the front or perpendicular to the flow direction. This abrupt change in fiber orientation weakens this region. Weld lines may be formed also by splitting and rejoining of flow fronts around inserts. The orientation of the weld line is affected by the position of the insert relative to the gate, as shown on the diagram here. A weld line strength factor may be defined as the ratio of the composite strength in the presence of a weld line to the composite strength without weld lines . Causes of weld lines in IM:

Multigated molds leading to head-on impingement of two separate flow fronts.

Splitting and rejoining of flow fronts due to presence of inserts.

This diagram shows the flow splitting up before the insert and rejoining after it. At sufficiently high flow rates, instead of flowing uniformly, the polymer may shoot out of the gate until it impinges against the mold boundary and is compressed against the opposite wall. Backfilling fills in the rest of the mold. This phenomenon creates many weld lines.

Strength issues with weld lines

Weld lines tend to weaken an injection molded part. The weld line factor describes the weakening caused by the presence of a weld line. Weld line factor = strength with weld line / strength without weld line.

This graph shows how weld line factor is affected by fiber concentration. Increasing the fiber concentration decreases the weld line factor. The reason weld line factor drops with the addition of fibers is found in fiber orientation. Fibers at the flow front are oriented parallel to the front. The collision of two fronts to form a weld line results in fiber orientation parallel to the weld line. Thus, the strength across the weld line drops dramatically with respect to regions of the part without weld lines.

Minimizing the effect of weld lines

Guidelines to minimize the effect of weld lines: 1. Locate weld lines closer to a gate to make them strong. It is important to have enough pressure to get good packing. 2. Provide venting at the weld line. 3. Increase part thickness at the weld line 4. Increase melt temperature more flow. 5. Increase injection pressure and speed.

Case Study: Fuel Rail (opens in new window)

Types of gates
After selecting the location and size, we must decide on the type of gate. Gates are classified in two categories based on the method of degating Manually Trimmed: 1. 2. 3. 4. Direct of Pin Gate Tab Gate Edge Gate Fan Gate

Automatically Trimmed: 1. Pinpoint Tab Gate 2. Submarine or Tunnel Gate Manually Trimmed Gates Automatically Trimmed Gates

Direct or pin gates are manually trimmed from the part. They Pinpoint tab gates and submarine gates are automatical are commonly used in single cavity molds for thick walled trimmed from the cavity after molding. large parts. This gate has minimum pressure drop but it leaves a mark on the part surface after the gate is trimmed. A pinpoint tab gate is used either in three-plate molds o The gate diameter is usually chosen to be twice the part in hot runner molds for rapid gate freeze-off and easy thickness. degating. The runner is automatically trimmed from the part with a stripper plate as the mold opens. A tab gate is typically employed for thin and flat parts to reduce the shear stress in the cavity. The high shear stress A submarine or tunnel gate is widely used in two plate generated around the gate is confined to the auxiliary tab, molds or in multi cavity molds. A tapered tunnel runs which is easier to trim off after molding. The tab thickness is from the edge of the gate to the cavity just below the typically 75% of the part thickness and such gates are parting line. This gating scheme permits automatic extensively used for molding polycarbonate, SAN and ABS degating of the part from the runner system during the types of materials. part ejection stage. Edge gates are commonly used for multi-cavity two plate

molds. An edge gate is located on the parting line of the mold and it is trimmed off after molding. Edge gates are suitable for medium and thick section parts. A fan gate is a wide edge gate with variable thickness. It is effective for flat thin moldings and for molding large heavy parts.

Definitions
The part shrinkage can be defined in terms of either the part length or the part volume. As shown here, L sub part is the length of the part after molding and L sub mold is the length of the mold. Linear shrinkage, SL is defined as a ratio of (Lmold - Lpart) / L mold. Volumetric shrinkage, Sv is defined as (Vmold - Vpart)/ Vmold. Linear shrinkage, SL is approximately one third of the volumetric shrinkage SV for isotropic parts. True shrinkage is evaluated with the dimensions of the hot mold.

The shrinkage of a molded part is not a material property; rather it is a system property that depends on material properties, part geometry and manufacturing parameters. Material properties include PVT data, thermal properties, viscosity etc. Part geometry parameters include part thickness mold constraints like ribs and bosses. Manufacturing parameters include temperature, pressure, fill time etc. Part shrinkage is a system property that depends on the combination of all of the above parameters.

Warpage causes the post mold shape of plastic parts to differ from the mold cavity shape. Warpage
results from nonuniform shrinkage. This schematic illustrates.

Here is a mathematical definition of warpage: w = hpart Click here for an example of warpage. Sink marks are created when a lot of shrinkage occurs in certain sections. They appear as depressions in the surface of the molded part. Long term dimensional stability issues include relaxation, creep, and aging phenomena.

Engineering vs. True Shrinkage

It is important to make a distinction between "engineering" and "true" shrinkage. This distinction arises because the mold expands as it is heated.

Engineering (standard) shrinkage:

True shrinkage:

This example shows the effect of the shrinkage basis chosen: Cold part: 99.0 mm Cold mold: 100.0 mm Hot mold (Hot = cold + S=10.0 mm/m True S = 11.1 mm/m

steel

T) = 100.11 mm

Shrinkage is not a material property! It depends on...

Material properties --- PVT data, Thermal properties, etc. Part geometry --- Thickness, mold constraints, etc. Manufacturing --- Temperature, pressure, flow, etc.

Shrinkage is a system property! A good predictor of shrinkage must consider all three factors.

An example of warpage
This diagram depicts a simple example of warpage due to nonuniform shrinkage. A disk is molded with the gate in the center. The pressure is high near the gate and drops off radially. Because the pressure is highest in the center, as the mold is packed, more polymer is packed near the gate than near the edges of the mold. Since less shrinkage occurs with a higher degree of packing, higher shrinkage occurs at the edge of the disk than at the center. This nonuniform shrinkage causes the final part shape to be domed. Nonuniform shrinkage is common because pressure gradients are unavoidable, especially when filling with a single gate.

Shrinkage with fibers

Shrinkage across the fiber direction is considerably greater than shrinkage along the fiber direction. The fiber constrains along its axis. Thus, fibers cause anisotropic shrinkage.

For a fiber (rod), SV = Sfiber + 2SX fiber SX fiber/Sfiber = 3-6 Where: Sfiber = length shrinkage SV = volumetric shrinkage SX fiber = shrinkage across fiber direction

Shrinkage Analysis
Analysis of dimensional stability may be carried out at three different levels. The simplest analysis, Level 1, does not account for the pressure history or the temperature history. So it is practically useless. Level 2 analysis is quite useful for simpler part shapes. It explicitly accounts for the pressure history and the temperature history. For complex shapes, computer aided finite element simulation is recommended. Level 3.

Level 1 Analysis
In level one analysis the material supplier provides the linear shrinkage number SL.

The relationship among Lmold, Lpart and SL is shown here. This method of analysis gives us a very quick estimate of linear shrinkage. Such shrinkage numbers are not based on the actual processing conditions. This method often requires trial and error in cutting the tool, as the shrinkage estimates are very crude. It also does not provide any prediction for the part warpage.

Method Available Data Quick Requires trial-and-error mold designing No prediction of warpage

Benefits

Issues

Level 2 Analysis
In the level two analysis, shrinkage estimates are obtained from P V T data. The assumptions of this method are uniform temperature and uniform pressure of material. Five states of the material marked A through E during the molding cycle are shown here:

Few studies have attempted to find good predictors for shrinkage in real fiber filled molding compounds. However, studies have identified prediction methods for shrinkage in unfilled parts. PVT diagrams can be used to predict shrinkage.

What's a PVT diagram? Factors affecting PVT diagrams Using PVT diagrams to predict shrinkage

What's a PVT diagram?

PVT data ("pressure-volume-temperature") describe how specific volume of a polymer melt changes with pressure and temperature.

This graph gives V as a function of T for two polymers. For the amorphous polymer, the slope of the curve is greater above the glass transition temperature (Tg). Greater shrinkage occurs above Tg than below Tg. For the semicrystalline PET, a much bigger jump in specific volume occurs at the melting point. This enormous drop in specific volume shows that much greater shrinkage occurs for semicrystalline polymers than for amorphous polymer. There is also a gentle change in the slope of specific volume curve below the melting point. Specific volume is also affected by pressure. Here is a PVT diagram for polycarbonate. Different curves T-V are shown for different pressures. The specific volume curves shifts down with increasing pressure. The top curve is at ambient pressure, the middle curve is 80 MPa, and the bottom curve is 100 MPa. Also, notice that the glass transition temperature increases slightly with pressure. Here is a PVT diagram for filled materials. The specific volume is lower for filled materials than for unfilled materials. (The filler increases the density). As with unfilled materials, increased pressure shifts the T-V curves downward. Filling the material also shifts the curves down and decreases the slopes. This flattening of the T-V curve is advantageous because it decreases the total shrinkage for a part. However, when fiber is used as a filler, the result is somewhat disadvantageous because anisotropy is increased. Here is a PVT diagram for a semicrystalline material. Shrinkage is typically larger for semicrystalline material than for amorphous materials.

Factors affecting PVT diagrams

The cooling rate for injection molded parts can reach thousands of degrees per second. At this very fast rate, the extent of crystallinity is minimal. At lower cooling rates, polymer sections can attain a higher degree of crystallinity. this diagram shows how specific volume-temperature data is affected by cooling rate. In summary, fast cooling leads to less shrinkage; slow cooling leads to more crystallinity and high shrinkage. Differences in cooling rates can also cause sink marks. Variations in cooling rate can be caused by:

some areas of the part being thicker than other parts variation in mold temperature due to poor cooling design local frictional heating

Using PVT diagrams to estimate shrinkage

An idealized pressure history for the mold cavity is shown on the figure with four stages: filling, packing, cooling and part ejection. The five states represent: the un-pressurized melt at the injection A temperature the pressurized melt when the melt front B reaches the end of the cavity the fully packed melt at the time the gate C freezes off D the solidified polymer E the polymer part when it is ejected from the mold

The corresponding path with five states is marked on the PVT diagram. Point C is the end of packing at gate freeze off and point E is the part ejection. The volumetric shrinkage is calculated based on the volumes at these points:

On the PVT diagram, the material moves from the melt state (point A) to the final solidified state (point E). The volumetric shrinkage SV is defined as the ratio of the difference between the mold volume and the part volume to the mold volume. From the PVT diagram, the volumetric shrinkage SV is found as ( vc - vE) / vc:

The pressure history inside the part varies with distance from the gate. This figure shows a rectangular flat plaque that is injection molded by a fan gate. The points 1 2, and 3 are different locations inside the part. The pressure histories at these points are different as shown. This difference in pressure history will cause different shrinkage at these locations:

The volumetric shrinkage is calculated for the three points from the PVT diagram for polycarbonate. The volumetric shrinkages at these points are 0.014, 0.023 and 0.04 respectively. The linear shrinkage may be estimated as 0.005, 0.008 and 0.013:

To summarize the PVT analysis method: 1. Required at each point: o Thermal history o Pressure history o Freeze-off time 2. Follow path through PVT diagram to estimate shrinkage Result: Local shrinkage --> warpage Limitation:

1. PVT path is not determined as function of time. Cooling rate has an important effect on shrinkage, especially for crystalline materials. The PVT method neglects this effect. 2. Through-thickness effects not included. 3. Residual stresses not computed.

Level 3 Analysis

This illustrates the Computer Aided simulation scheme. The material, the processing conditions, the mold geometry and convergence criteria are input to conduct flow and heat transfer analyses. The flow and heat transfer analyses provide the melt temperature and stress fields as a function of time and location. These are used to perform in-mold thermomechanical analysis to predict in-mold stresses and temperatures.

Finally, structural analysis is performed on the ejected part with in-mold stresses, and temperatures to predict the final part shrinkage, part warpage and residual stresses. Shrinkage predictions from the level three analysis are accurate to within 10 percent provided the part is constrained in the mold. Slip inside the mold is not accounted for in this analysis. The computer-aided simulations are able to predict trends in warpage; the predicted magnitudes are not as accurate. Such analysis is excellent for initial part design.

The benefits of the level three analysis are rapid analysis for functional design, detailed analysis for tool geometry, coupling geometry, material and manufacturing. It also provides an estimate of global and local shrinkage and warpage. The level three analysis requires flow, thermal and mechanical material data. Software costs and computational time for detailed design analysis must be available.

Next we discuss two case studies that illustrate the application of computer aided level three analysis to mold design. The first case study explores the effect of material type and filler on the shrinkage and the warpage of an automotive sail panel. The sail panels are typically used on the side of a car. The second case study deals with an automotive transmission cover. Here the number and the location of gates are varied to relocate the weld lines and to minimize the part warpage.

Level 3 Analysis, Case Study: Sail Panel

This shows an idealized geometry of the sail panel. It is a top view. It is filled by a single edge gate located at the left-hand corner. The predominant flow length is 990 millimeters. The part has both an inplane curvature and an out of plane curvature. The part thickness varies between 3.3 millimeters and 4 millimeters.

Another view of the sail panel depicts the out of plane curvature.

The predicted shrinkage and warpage for this geometry with unfilled and filled Polybutylene Terephthalate. will be compared here. The first material is unfilled Polybutylene Terephthalate. It has a minimum processing temperature of 223 C, a maximum processing temperature of 250 C and an ejection temperature of 140 C. The selected fill time is 3 seconds and the rest of the cycle lasts 20 seconds. The second material is 40-weight percent short glass fiber filled Polybutylene Terephthalate. It has a minimum processing temperature of 243 C, a maximum processing temperature of 276 C, and an ejection temperature of 140 C. The selected inlet melt temperature is 260 C, the selected fill time is 5 seconds and the rest of the cycle lasts 15 seconds.

Materials and Process Conditions

Polymer Polybutylene Terephthalate Unfilled Processing temp., C Min:223 Max:225 Eject:140 Inlet melt temp., C Fill time, seconds 240 3

Post-fill tim seconds 20

Polybutylene Terephthalate Filled 40% by weight short glass fibers

Min:243 Max:276 Eject:140

260

15

A comparison of the molded part outlined in red, and the mold cavity outlined in blue, is shown for the sail panel with unfilled Polybutylene Terephthalate, in one plane. Polybutylene Terephthalate is a semi-crystalline material and hence the shrinkage is higher than for amorphous polymers.

A comparison of the molded part outline and the cavity outline is shown for the sail panel, with short glass fiber filled Polybutylene Terephthalate as the material, in that same plane. The presence of the glass fibers reduces the total shrinkage compared to the unfilled Polybutylene Terephthalate. The plotting scale factor is 1. The blue line represents the original mold cavity and the red line indicates the post-molded part size. Fiber filled polymers have lower shrinkage than the neat polymer along the flow directions; but in other directions, they shrink more. The next two view graphs show a different view of the sail panel with the original shape and the post molded shape.

This view compares the molded part and the mold cavity for the crystalline unfilled Polybutylene Terephthalate.

A comparison of the molded part and the mold cavity is shown for the short glass fiber filled Polybutylene Terephthalate. This view shows that the part warpage in the out-of-plane direction is much greater than that shown with the unfilled Polybutylene Terephthalate. This difference in warpage is due to fiber orientation.

This view shows the predominant fiber orientation developed in single gated filling of the sail panel if a short glass fiber reinforced material is used. The glass fiber orientation varies significantly across the part. This introduces anisotropic, nonuniform shrinkage across the part, which is the reason for the higher warpage with the filled Polybutylene Terephthalate. In summary, it is harder to control warpage in complex shaped parts with fiber filled materials.

Level 3 Analysis, Case Study: Transmission Cover

The second case study deals with computer aided simulation of an automotive transmission cover. This cover sits atop the gearbox. Hence, the flatness of the bottom base area is a functional requirement for the part. The shaft for changing gears comes out at the top through the shaft hole whose diameter is 35.4 millimeters.

Due to the movement of the shaft during the change of gears, the top portions of the cover experience higher load. Hence, another functional requirement is to minimize the number of weld lines in the top region. The dimensions of the part are displayed here. The thickness of the part varies from 8.9 millimeters in the flange area to 5.08 millimeters elsewhere in the part. The rib area is 3.81 millimeters. Edge gates have been used in the simulations. Three different cases have been considered. In the first case, a single edge gate is located at point 1. In the second case two edge gates located at points 3 and 4 have been considered. In the third case, one edge gate at point 2 and one edge gate at point 4 are considered. Thirty-three weight percent short glass fiber reinforced nylon 6 6 was used as the material for all three cases.

This shows the top view of the transmission cover depicting the weld lines for single gated filling. Only one weld line is formed at the critical shaft region. This is caused by the mold insert required for the shaft hole. Other weld lines are formed at non-critical areas.

This illustrates weld lines in case two where two gates are located at the opposite ends of the base flange area. The number of weld lines formed at the critical shaft area increases to four with multi-gated filling. However, changing the location of one of the gates could move the weld lines to non-critical areas. This illustrates the weld lines in case three where one of the gates is located in the top shaft area and the other one is located at the base flange area. The number of weld lines formed at the critical shaft area is reduced to one similar to the single gated case. Hence, by appropriate location of gates, one can move the weld lines to meet the functional requirement of the part. Let us look at the other functional requirement of the part now, which is the flatness of the base flange area. In all these figures, the blue line represents the original feature line and the red line represents the postmolded warped part.

For the single gate as shown here, we find a substantial amount of warpage for the base flange area, which may not be acceptable.

This shows part warpage for the double gated transmission cover when the two gates are located at the opposite ends of the base flange area. As seen from the red and the blue feature lines, part warpage is reduced drastically as compared to the single gate case. The reasons for this are two fold. First, the two gates reduce the nonuniformity of pressure distribution and second the two gates are located at the thicker section of the part compared to the single gated case, which is located at the thinner section. This gives the thicker region more time to cool compared to a single gate. In the third case, one gate is located at the top shaft section and the second gate is located at the bottom flange section. Part warpage for this gating configuration is similar to the case when the two gates were located at the opposite ends of the bottom flange area. However, as we saw earlier, the number of weld lines in the critical shaft area is also minimized in this case. Hence, from the functional requirement point of view, the third

gating configuration of two gates with one located at the shaft area and the other located at the bottom flange area provides the best result. This illustrates that using level 3 analysis one can investigate the effect of changes in mold design and material quickly without actually cutting the tool.

Factors affecting shrinkage

Here are the results of some calculations based on the PVT diagram for unfilled polycarbonate. The effect of several variables on in-plane and through-thickness shrinkage are given in the following series of graphs. This graph shows the effect of packing pressure on shrinkage. Increased packing pressure causes a decrease in shrinkage. The effect is nearly the same for both in-plane and through-thickness shrinkage. This graph shows the effect wall thickness. Thicker parts experience greater shrinkage. This effect is more pronounced in the through-thickness shrinkage than the in-plane shrinkage. Gate thickness also has an effect on shrinkage. Usually, the gate thickness is about half the wall thickness. As the gate size is increased, the through-thickness shrinkage decreases significantly, but the in-plane shrinkage is basically unaffected. Melt temperature does not have much of an effect on shrinkage. Mold wall temperature affects shrinkage because it determines the cooling rate. Increased mold wall temperature leads to decreased cooling rates and increased shrinkage. Although PVT diagrams can be used to make short, simple calculations to determine the effect of various factors on shrinkage, the dominant factors affecting shrinkage are:

In-mold constraint conditions Fiber orientation (see Microstructure section on fiber orientation)

In mold constraint conditions

Here is a picture of a suggested rib design to reduce shrinkage. The rib should be thinner than the wall to avoid sinks. A thick rib leads to a slow-cooling section with greater shrinkage, leading to voids and sinks.

Here is an sketch of a highly constrained plaque. There is a tall rib all along the edge of the plaque.

Shrinkage in this constrained plaque was measured experimentally for a variety of materials. The results were compared to an unconstrained flat plaque and a less constrained plaque having short ribs across the surface. This graph shows the crossflow shrinkage for unfilled nylon. The level of shrinkage is much greater for all three plaques, but the relationships are the same as for polycarbonate.

This graph shows crossflow shrinkage for glass filled nylon. The shrinkage decreases some with the addition of filler but is still significant. The crossflow shrinkage does not decrease as much as the flow direction shrinkage because the fibers are orientated along the flow direction.

Goals for the future in injection molding

An important future goal for injection molding is to make more structural parts than have been made in the past. The current, empirical approach to design of IM parts is "make it and break it." Major efforts by General Motors and General Electric are being expended to put science and engineering in IM design rather than relying on trial and error. The end goal is to be able to define the desired performance and select materials and arrive at a part design that will perform. In arriving at an appropriate design, it is important to remember that the injection molded parts may not always look like their metal counterparts! In summary, here are some guidelines for design for manufacturing

Design for minimum number of parts (integration of parts) In mass production, minimize parts variation. Dimensional stability and reproducibility are prime goals. Design parts for multiuse Design parts for ease of manufacturing Avoid separate fasteners Avoid flexible components (e.g., rubber hoses)