No more streaks! cure polycarbonate molding maladies.

The key to molding good-looking PC parts is to dry the resin properly and use the right screw, check valve and end-cap/nozzle assembly.

Black and brown streaks and splay are among the most common ills seen in polycarbonate parts (see PT, Nov. '95, p. 15). There's no mystery about where these cosmetic flaws come from. Streaks are the visible signs of heat degradation caused by a combination of time and temperature. Another degradation effect, splay, is bubbles of gas that form at the flow front. Typical causes of both problems include moisture from improperly dried resins, exposure of the melt to an iron-rich surface, or machine-design factors such as the screw, check valve, end-cap, or temperature control. Severity of streaking or splay can increase significantly when these factors combine to create a corrosive environment in which moisture-degraded PC reacts with exposed iron surfaces.

The critical importance of these factors - especially those related to equipment design - was demonstrated by a laboratory study performed by Dow Plastics on a molding project of a manufacturer of HVAC equipment. During the evaluation of materials to be used in an appliance cover, the molder observed streaking in parts molded of a beige ignition-resistant (IR) polycarbonate.

CHECK OUT YOUR HARDWARE

Improper injection screw design can mechanically degrade polycarbonate and form volatiles. The result can be the same streaked appearance caused by processing improperly dried PC. Older end-cap designs often have stagnant flow areas and typically do not seal well against the higher pressures required to mold polycarbonate and some other engineering thermoplastics. Over time, material that stagnates in the flow path or leaks at a parting line degrades and then shows up as black or brown streaks or specks in the molded part. This problem is aggravated when metal surfaces in the polymer flow path are high in iron or when resin is insufficiently dried.

Temperature-control systems on many older presses are generally not optimized for polycarbonate processing. Having too few control zones is a common problem. Other contributing factors include improperly sized or instrumented temperature-control zones, or inefficient heating from slow-response heater bands. (In general, mica bands have faster response than ceramic types.)

Another common flaw in temperature controls is using the last thermocouple on the barrel, rather than a dedicated zone, to control the end-cap/adapter area. This set-up means that the forward face (nozzle side) of the end cap is heated primarily by the melt (through the steel) or by the nozzle zone and can be 100 [degrees] F cooler than the setpoint of the last zone on the barrel. As a result, streaking can appear near the gate area of the part. Also, cold, high-viscosity plastic in the cushion area of the end cap can cause material stagnation and degradation, which can appear as streaks, especially near the end of fill.

Many of these common equipment faults were evident in the molding study for the HVAC part. Samples were first molded on a 700-ton injection machine, on loan to Dow's lab, which was not optimized for molding IR polycarbonate. The press was fitted with a 70-oz., 3.5-in. diam. screw of general-purpose polyolefin design with 20:1 L/D, 2.3:1 compression ratio, and a flight profile of 10/4/6 turns in the feed/transition/metering zones. This is among the least desirable designs for IR PC. What's more, large areas of the chrome plating had worn off at the screw roots, exposing the steel.

The screw was fitted with a three-piece, polyolefin-type non-return valve with deep discharge flutes. This is again one of the least appropriate designs for processing IR PC. It was constructed of a D2 steel and showed enough wear on the forward and rear seats to allow leakage during injection.

The end cap was a multi-piece design constructed of 4140 nitrided steel. It had several discontinuous areas in the flow path where material could stagnate, and it did not seal efficiently when processing high-temperature polymers. The barrel had only three temperature-control zones, with the end cap included in the last barrel zone. The forward surface of the end cap was 75 [degrees] F cooler than the setpoint, creating the potential for material stagnation in the cushion.

To test the effects of equipment design, moisture, and temperature, resin samples were molded into sprue-gated plaques measuring 1 x 2 ft x 0.125 in. thick (representative of the housing size). A sample of beige IR polycarbonate with 20 MFI taken from the most recent trial by the molder was used as a standard for the test evaluations. A natural (uncolored) sample of the same resin was included to investigate the effect of colorants on the formation or intensification of streaking. The effect of a different color and resin viscosity was observed by evaluating a sample of gray IR PC with 15 MFI.

All PC samples were evaluated undried (pellet moisture greater than 0.05%) and after being dried to less than 0.02% moisture. To test the impact of inefficient drying, improperly dried samples with moisture content above 0.05% were simulated by allowing dried resins to absorb moisture for at least 24 hr.

Processing thermal stability of each material was evaluated by raising the nozzle-zone temperature above the highest recommended processing temperature, then adjusting it to the lowest recommended temperature for each sample. Results were compared with molding at intermediate nozzle temperatures.

All properly dried PC that was processed on the original general-purpose equipment showed random occurrences of black or brown streaks and specks. The streaks were more obvious and more common on samples exposed to greater heat, whether through a "heat soak" or a higher nozzle temperature. But even at moderate temperatures, all samples showed streaks. Streaking was localized to the sprue area and formed a distinctive five-pointed star pattern. This pattern was more evident during material changes and on samples processed with a cool nozzle. On gray and beige parts, streaking was discernible only during a color change or with extended heat soaks and was not as pronounced as on the natural parts.

Higher temperatures caused more consistent gate-area splay in colored resins (not just during color changes). This suggests that the processing thermal stability of the resins was affected by the process of coloring the product or by the colorant itself. For both the gray and beige samples, splay occurred when the thermocouple reading was above 565 F.

MOISTURE MAKES IT WORSE

The change in resin condition from properly dried to improperly dried had a significant effect on the black streaks and splay for all color samples. Streaks on natural samples were most noticeable because of the translucency of the plaque. Brown streaks emanating from the star pattern tended to elongate and become darker. Black streaks first appeared when processing the undried natural resin and occurred randomly, even at moderate nozzle temperatures. Heat soaks caused black and brown streaks to women and sometimes cover the part.

Improperly dried beige PC showed slight black streaking over the last third of the part to fill. Changes in nozzle temperature did not alter the frequency or severity of these streaks. The star pattern appeared on the beige samples only during color and material changes.

It was difficult to find evidence of streaking on gray parts produced with the lower nozzle temperature. However, both moderate and high nozzle temperatures caused black streaks to form in the last half of the part. Heat soak resulted in excessive splay and black streaks even near the gate.

While splay was noticed on all parts, it was less severe when a cooler nozzle was used and more severe with a hotter nozzle. This supports anecdotal reports that the presence of moisture, combined with iron exposure, seems to decrease thermal stability of the polymer. Streaks and splay could be greatly reduced by lowering the melt temperature.

DOING IT RIGHT

For the second part of the study, the molding machine was upgraded from its original condition to determine how improvements would affect the occurrence of streaks and splay. The original screw was replaced with a highly distributive mixing section that is known to melt PC efficiently. The check valve was replaced with a four-piece design that has an outstanding track record in difficult PC applications. The end cap and nozzle were replaced with an improved design featuring a constant-taper flow path, which eliminates the potential for material stagnation. The one-piece end cap also eliminates melt leakage into the parting line between the pieces, thus avoiding degradation. The new end cap has chromed internal surfaces to improve polymer flow and virtually eliminates exposure to iron-rich metals.

The upgraded equipment significantly improved part appearance under all molding conditions. None of the properly dried resin samples showed signs of splay. At moderate nozzle temperatures, streak-free parts were obtained consistently.

The equipment upgrades also yielded a noticeable improvement in the appearance of parts molded from improperly dried resins. Streaking was almost completely eliminated.

Samples molded on the upgraded equipment showed no evidence of the star-patterned streaking observed when processing with the original equipment. The streaks on the original samples matched the pattern of five extremely deep discharge flutes on the original non-return valve. The bottom of those flutes (not present on the new NRV) held a large amount of degraded plastic that was being transferred to the parts. This discolored material is likely the result of a reaction between the tool steel and the polymer as it stagnates in the flutes.

Brent Salamon is a senior development engineer with the Applied Fabrication Technology Group of Dow Plastics in Midland, Mich.

Since the completion of this study, Mike Martin has joined Dupont Dow Elastomers in Freeport, Texas.