The effect of colorants on the properties of rotomolded polyethylene parts.

INTRODUCTION

The basic principles of rotational molding are simple. The main steps of the process are:

1) polymer (powder, pellet, or plastisol) is added to a hollow mold and the mold is closed,

2) the mold is biaxially rotated and heated to cover its inner surface with the polymer,

3) the rotating mold is cooled by water spray or air to solidify the polymer,

4) the mold is opened and the part removed.

Rotational molding existed long before plastics were developed. In 1865, Lovegrove patented the rotational casting of shot and artillery shells (1); later, hollow chocolate articles were made by rotomolding (2). Industrial scale rotomolding of plastics started with poly(vinylchloride) (PVC) plastisols in the 1940s followed by polyethylene (PE) in the 1950s (3). The latter still dominates the market.

Despite the long history of rotational molding, relatively few scientific papers have been published on the subject. A monograph on rotational molding containing contributions from many different authors was edited by Bruins in 1971 (3). Throne (4-6) and, more recently, Crawford and his coworkers (7-11) studied the heat transfer, melting, and solidification process in rotomolding. Studies carried out in our laboratories have concentrated on heat transfer (12), warpage (1317), and regrind/recycle (18). It was found that warpage and global shrinkage are caused by residual quench stresses, can be correlated with levels of densification and rates of cooling during the solidification process, and can be minimized by mold pressurization (16),

All the above studies were carried out with unpigmented polyethylene, although industrially rotomolded polyethylene parts often contain colorants. Colored rotomolded products include toys, playground equipment and recreational, household, and marine items. Even large rotomolded tanks come in various colors.

Pigments are usually simply dry blended with polyethylene powder. For better pigment dispersion, often a small amount of surfactant is added. There are different techniques of dry blending. We may distinguish between (i) low intensity mixing and (ii) high intensity mixing. Drum tumblers and ribbon blenders are typical low intensity mixers, operating at low speed and relatively long mixing time (in excess of 10 min). Henschel mixers are a typical example of high intensity mixers operating at high speed (several hundred rpm) with short mixing time (a few minutes).

Practical experience shows that even 0.1-0.5% pigment can cause severe problems regarding warpage and shrinkage and deteriorate the mechanical properties of rotomolded parts. These problems can be diminished by using an extruder-compounded colored polyethylene powder, but of course, compounding on extruder and subsequent grinding are more expensive than dry blending. Compounding also decreases the flexibility in production programming of the rotomolder. Even small amounts of parts with different colors can be made economically with dry blending, but small amounts of color compounded rotomolding grade polyethylene are not readily available.

The authors are not aware of any detailed scientific study published on the effect of colorants on the mechanical properties, warpage, and shrinkage of rotomolded polyethylene parts. The aim of this study was to gather information and gain a better understanding of the effect of colorants on part properties. We have to emphasize that the colored polyethylene samples used in the present study were not commercial materials, but were manufactured exclusively for the present study with various pigments, mixing techniques, and with or without the addition of surfactant.

EXPERIMENTAL

Materials

A rotomolding grade linear low-density polyethylene (LLDPE) resin was used, with a density of 0.935 g/[cm.sup.3] and a melt index of 5.5. The pigments included three inorganic particulates with predominantly less than 1 [[micro]meter] particle size (titanium dioxide, iron oxide red, and cadmium oxide yellow), carbon black (medium jet mean particle size 27 [[micro]meter], surface area about 95 [m.sup.3]/g), and phthalocyanine blue (particle size less than 12 [[micro]meter], surface area about 60 [+ or -] 30 [m.sup.2]/g), an organic pigment. The titanium dioxide used was a surface-coated grade. In [TABULAR DATA FOR TABLE 1 OMITTED] some samples, 770 ppm of zinc stearate was used as a surfactant and, in one unpigmented LLDPE sample, an extra 500 ppm of anti-oxidant, Ciba-Geigy Irganox 1076 (octadecyl 3,5-bis (1,1-dimethylethyl)-4-hydroxy-benzene propanoate), was added. The composition of samples as well as the mixing conditions are listed in Table 1.

Dry Blending and Compounding

A series of samples was prepared by high intensity mixing of the LLDPE powder, pigment, and zinc stearate surfactant (samples 5-7). For comparison, a portion of the unpigmented LLDPE was also high intensity mixed without any additive, another portion was high intensity mixed with zinc stearate, and a further sample was high intensity mixed with zinc stearate and antioxidant (samples 2-4). Mixing time was 3 min in a Henschel-type mixer.

Low intensity mixed dry blends were prepared by drum tumbling LLDPE with the additives (samples 8-15). One sample was compounded on a 1.5-inch single-screw extruder, and the melt was extruded through a 20 mesh screen. The melt temperature in the die area was [approximately]430 [degrees] F (220 [degrees] C). The compound was pelletized and subsequently ground to 35 mesh on a Wedco grinder (sample 16).

Molding Conditions

Parts were molded in a McNeil Akron Rotocast 800 machine. This machine consists of three tire-driven rotating arms, an oven, and a cooling chamber. In the load station, powdered polymer is added to the mold. The closed mold containing PE powder is then transferred into the oven where it undergoes biaxial rotation, which spreads the powder over the internal surface of the mold. The heat supplied by the oven melts the thermoplastic powder, which then fuses together into a polymeric shell attached to the internal surface of the mold. The oven temperature was set to 608 [degrees] F (320 [degrees] C), the same temperature that was used in the warpage studies reported earlier (12-15). Oven time was varied between 9 and 12 min. The nominal rate of rotation of the minor arm was set to 18 rpm (effective rate 6 rpm), and the major arm rotating rate was 12 rpm, giving a ratio of major/minor = 2. The parts were cooled by 3 min water mist followed by 4 min fan blowing. The offset arm of the machine was used in all experiments.

A rectangular steel sheet mold was used, approximately 73 by 52 by 21 cm (29 by 20 by 8 inch); 4000 g polyethylene were applied, resulting in [approximately]3.6 mm (0.142 inch) average wall thickness of the rotomolded part.

Freecote 33 (Hysol Aerospace and Industrial Prod. Div., NJ) was applied as mold release in two thin layers using an aerosol can. After 6-8 parts, the mold release was applied again.

Molding in the McNeil Akron machine was first performed in a series of experiments in air. Another series was made with nitrogen gas in the mold. The overpressure of [N.sub.2] was 0.14 bar (1.14 bar absolute pressure); i.e., 2 psig (16.7 psia).

Testing

Low temperature impact tests, often referred to as dart impact tests, were performed on 12.5 by 12.5 cm (5 by 5 inch) test specimens cut from the flat bottom part of the boxes. The test was carried out according to the Association of Rotational Molders (ARM) standard at -40 [degrees] C (-40 [degrees] F). In all cases, more than 20 test specimens were used to obtain at least 10 "events," i.e., 10 failures or nonfailures, whichever may be less. On one series of samples, the dart impact test was carried out also at room temperature with 10-20 test specimens.

Izod impact tests were carried out at room temperature on notched test specimens, according to ASTM D256, using a 0.907-kg (2-lb) pendulum. Tensile testing was performed at room temperature on a Monsanto Tensile Tester T-10 with a 50-mm (2-inch) gauge length at a 50 mm/min (2 inch/min) strain rate on ASTM D638 type I test specimens. The morphology of the LLDPE powder and the rotomolded parts was investigated by optical microscopy in a transmission mode on a Leitz Laborlux 12 POL type microscope.

Some rotomolded parts were cut into about 5-mm (0.2-inch) pellets with a laboratory scale Brabender granulator. The melt index of the pellets was measured at 190 [degrees] C according to ASTM D1238, with 2.16-kg load.

The global shrinkage of the parts was measured both along the length and width of the box. Parts were kept at room temperature in an upright position for at least 24 h before measuring shrinkage or warpage. To measure shrinkage, two small cuts were made into the mold wall at the parting line at exactly 300-mm (11.81-inch) distance. These cuts showed up in the PE parts as well-defined sharp signs, and the distance between the signs was measured. Warpage was measured on the "flat" bottom part of the boxes in a 50 by 50 mm (2 by 2 inch) mesh as the vertical distance of a mesh-node from the ideal plane. The mold itself had a measurable warpage -10.1 mm (-0.398 inch) at the deepest point.

RESULTS

Optimization of Molding Conditions

Before we started the actual study on the effect of colorants, two series of boxes were molded of the unpigmented base polymer on the McNeil-Akron machine by varying oven time at 608 [degrees] F (320 [degrees] C). One series was performed with [N.sub.2] as the inert gas and another series without inert gas. The low temperature dart impact strength was measured, and the data are summarized in Fig. 1. The ARM impact strength of the parts rotomolded in air increases up to 9-10 min oven time, then a rapid decrease is observed. The data show that with just 1 min overcure, i.e., the increase of oven time from 10 to 11 min, the low temperature impact strength decreases from 64 ft[center dot]lbs to 20 ft[center dot]lbs, a decrease of nearly 70%. On the other hand, if molding is carried out in [N.sub.2], the impact strength remains around 67 ft[center dot]lbs even at 14 min oven time.

Elongation at break shows a maximum both for parts molded in air and [N.sub.2] [ILLUSTRATION FOR FIGURE 2 OMITTED], but the maximum is much narrower in air than in [N.sup.2], indicating a narrow optimal range of oven time in air. The maximum is around 10 min oven time, in agreement with the optimum in low temperature impact strength.

Global Shrinkage and Warpage

Both shrinkage and warpage showed relatively poor reproducibility with the exception of the series of experiments where nitrogen pressurization was applied. In the light of earlier studies on warpage (12-14), this result is not surprising. Warpage is determined by a complex combination of different factors such as cooling rate, crystallization kinetics of the PE, wall thickness, and shape of the part, as well as adhesion of PE to the mold. While the former factors can be kept fairly constant, adhesion of PE to the mold is difficult to control; it changes from part to part. Even the composition of material molded earlier may affect adhesion since some surfactant will stick to the mold surface and change its adhesion properties.

As a general trend, the compounds containing zinc stearate as surfactant exhibit considerably larger warpage than those without surfactant. An example is shown in Fig. 3 where the warpage along the center line is plotted in the longitudinal direction of the box. (Warpage is regarded positive if the bottom of the box is warped upward, i.e., in the inside direction.) The maximal warpage data are given in Table 2. A statistical evaluation of warpage data (Table 3) shows that the average warpage of surfactant-containing samples is about double of those containing no surfactant, but both series of data are scattered in a wide range. Nitrogen pressurization decreases warpage to a few millimeters (0.1-0.2 inch), and also the standard derivation decreases by a factor of 5, indicating highly improved uniformity. No significant difference was found in the warpage of surfactant containing low and high intensity mixed samples.

Global shrinkage gave the same trends as warpage: large variations from 1.61 to 3.22% were observed if the mold was not pressurized (Tables 2 and 3), the average shrinkage of parts made in the absence of surfactant was 1.87 [+ or -] 0.17%, the presence of surfactant increased shrinkage by a factor of 1.31 to 2.45 [+ or -] 0.39%. Very uniform and low shrinkage was found with [N.sub.2] pressurization (1.70 [+ or -] 0.05%) for samples without surfactant. The presence of surfactant increased the shrinkage only slightly (by a factor of 1.07) to 1.82 [+ or -] 0.14%. The correlation of warpage and shrinkage is shown in Fig. 4.

It has to be noted that experimental conditions were chosen to obtain high cooling rates and easily measurable warpage and shrinkage data, so a generalization of the above numerical results would be dangerous. It has been shown in previous work (12-15) that shrinkage varies with the molding and cooling conditions; thus, it is not possible to establish a conversion factor between shrinkage of unpigmented LLDPE and surfactant plus pigment-containing blends.

Mechanical Properties

The low temperature ARM impact test results of colored rotomolded boxes showed large variations and, with some pigments, a drastic decrease was observed. The data are summarized in Table 4. All colored samples except the compounded white (sample 16), the black (sample 14), and low intensity mixed red without surfactant (sample 13) gave low impact resistance at 9 min oven time. The wall of some boxes contained small bubbles on the inside of the box, indicating undercure. An increase of cure time to 9.5 and 10 min didn't bring a noticeable increase in impact strength with the exception of the yellow low intensity mixed sample without surfactant (sample 11). With the surfactant-containing red samples (7 and 12), signs of oxidative degradation were found at 10 min oven time and the dart impact strength decreased significantly.

To completely exclude the possibility of oxidation as the cause of the low ARM impact strength, the dart impact test was carried out also at room temperature and the melt index of the polymer after molding was measured. (It is known that oxidation of the internal surface drastically decreases the low temperature impact resistance (10), but does not affect the room temperature impact test much.) It was found that the room temperature dart impact test data correlate well with the dart impact strength measured at -40 [degrees] F (Table 4, [ILLUSTRATION FOR FIGURE 5 OMITTED]), although the range of data was narrower in the room temperature test. The melt index at 9 min oven time shows only minor variations and is close to that measured on the virgin PE powder (Table 5). At 10 min oven time, a significant decrease is observed but the change of the melt index does not correlate with the impact strength; thus, oxidation can be excluded as the source of low impact strength found with several pigmented samples. Even molding under [N.sub.2] and extension of cure time to 12 min didn't bring a noticeable [TABULAR DATA FOR TABLE 2 OMITTED] [TABULAR DATA FOR TABLE 3 OMITTED] improvement. The low temperature dart impact resistance was similar to optimal values obtained in air with 9, 9.5, or 10 min oven time. This finding shows that the low impact strength obtained in air is not due to undercure or overcure but rather an "inherent" property of some colorant-PE blends.

The results of the Izod impact tests are given in Table 6. While the unpigmented PE does not break in this test, most dry-blended samples had an impact strength of 0.1-0.15 J/notched mm (2-3 ft.lbs/notched inch) some even less than 0.1 J/notched mm (2 ft.lbs/notched in). The correlation with the ARM impact test is not too good, probably because the mechanisms of failure in the two tests are quite different. The ARM impact test is very sensitive to crack initiation, while the notch cut into the Izod impact test specimen before testing serves to initiate the crack.

The tensile strength at yield and the elongation at break data are given in Table 6. The tensile strength at yield shows only little variations, but elongation at break correlates with the dart impact results. The highest elongation is obtained with those samples that gave the best ARM impact resistance (unpigmented, red and yellow low intensity mixed without surfactant, black, and compounded white - samples 1, 11, 13, 14, and 16).

[TABULAR DATA FOR TABLE 4 OMITTED]

Optical Microscopy Study of Colored Rotomolded PE

Transmission optical photomicrographs were obtained of the cross section of all colored LLDPE boxes rotomolded with 9 min oven time. The photomicrographs show that dry-blended pigments are concentrated along the fusion lines of the individual particles. An example is shown in Fig. 6. The microscopy did not reveal such differences in the color distribution that would explain the big differences in the mechanical properties of the parts made with different dry-blended colors.

In some photomicrographs of parts cured for 9 min, bubbles were seen. At longer cure times, most of the bubbles disappear.

The photomicrographs of the parts made from the white compounded sample show, as expected, no pigment concentration along the fusion lines. The individual particles can still be distinguished indicating that the pigment distribution is not quite uniform at the microscopic level [ILLUSTRATION FOR FIGURE 7 OMITTED].

Experiments to Improve Mechanical Properties

Here we will summarize the results of a limited number of experiments aimed to improve the mechanical properties of parts made with dry blended colors. LLDPE with phthalocyanine blue was chosen for these experiments since it gave the lowest impact strength.

In one experiment, 3000 g of phthalocyanine blue and surfactant-containing sample 15 was added into the mold and 1000 g of unpigmented LLDPE powder was placed in a 10-[[micro]meter] thin PE bag and fixed in the center of the mold through the venting holes. The pigmented PE melted first, then, when the PE bag reached its melting point, the unpigmented powder fell into the mold. (The same result could have been achieved by using a dump box.) A two-layer structure formed. The cross section is shown in Fig. 8. The microscopic picture shows that there is virtually no pigment in the innermost 10-20% layer of the part. The dart impact strength improved from MFE = 3 J (2 ft.lbs) to MFE = 41 J (30 ft.lbs). (The molding conditions were: oven time 10 min, cooling 3 min mist and 4 min fan.) In a control experiment, the 3000 g blue pigmented dry blend (sample 15) and 1000 g unpigmented LLDPE powder were added into the mold at the same time and mixed in the mold for 2 min. The dart impact strength was found to be only 1.9 J (1.4 ft.lbs).

The results suggest that colored parts could be manufactured with improved mechanical properties if [TABULAR DATA FOR TABLE 5 OMITTED] [TABULAR DATA FOR TABLE 6 OMITTED] the dry-blended pigment-containing PE forms the outer layer of the part and unpigmented PE forms the inside layer.

DISCUSSION

Pigments in rotomolding are traditionally dry blended. In rotational molding there is, however, virtually no shearing or stretching flow during the process. Thus dry-blended pigments do not distribute evenly, but occupy the fusion lines between the particles. This may lead to a serious weakening of the interface and a point of crack initiation. Even small amounts of finely powdered pigments can cover a large percentage of the fusion area.

Experimental studies of dry-blended pigments in polyethylene indicate that while there is generally a reduction in impact strength of molded parts, the magnitude of that reduction varies widely depending upon resin, pigment, and processing conditions - and probably individual techniques of operators.

To understand the change of mechanical properties of rotomolded parts made with dry-blended pigments, we have to analyze the properties of the fusion surface between the individual particles. The following main factors must be taken into account: (i) coverage of the polymer particles by pigments, (ii) distribution of pigment along the fusion area, (iii) adhesion between the pigment and the polymer, and (iv) influence of the additives on the crystallization kinetics of the polymer.

Certain observations seem clear:

(i) The coverage of the fusion area by pigment should be kept as low as possible. This can be achieved by using the lowest acceptable pigment levels or coarser pigment particles.

(ii) Poor pigment distribution can lead to weak spots, which become the sites of crack initiation.

(iii) Adhesion of polymer and pigment can be influenced by surfactants. Surfactants have a very great importance for rotomolded pigmented polyethylene. The addition of zinc stearate had positive effects on color brightness and powder properties. It virtually eliminated airborne dust. It also decreased electrostatic charge. There may also be effects on mechanical properties. (Cadmium yellow and iron oxide red gave much better impact strength without surfactant.)

(iv) Pigments can also influence the crystallization of polymers. Phthalo type compounds are known as nucleating agents. The increased warpage and shrinkage well known for phthalocyanine blue, may be an example of this effect.

As shown in Tables 4 and 6, the extrusion-compounded sample has very good mechanical properties compared to the dry-blended products. This supports the hypotheses that shear mixing rather than dry blending leads to superior parts.

CONCLUSIONS

1. In rotomolded parts made with dry-blended colors, pigments concentrate in the fusion area between the polymer particles.

2. The presence of zinc stearate surfactant increases warpage and shrinkage, but improves color brightness and powder handling properties.

3. Mold pressurization virtually eliminates warpage and results in lower and more uniform shrinkage.

4. With new pigment-surfactant combinations, molding conditions should be adjusted and properties of the parts investigated.

5. Even small amounts of dry-blended pigments and pigment-surfactant combinations reduce mechanical properties (especially low temperature impact resistance). The extent of reduction varies considerably with pigment, processing technique, and other variables. Molding under [N.sub.2] does not eliminate this problem.

6. Two-layer parts with colored outer and unpigmented inner layer may be beneficial in producing parts with superior mechanical properties.

7. Compounded pigments are recommended for demanding applications.

ACKNOWLEDGMENTS

The research described in this paper was supported by the Association of Rotational Molders.

REFERENCES

1. T. J. Lovegrove, U.S. Patent 48,022 (1865).

2. J. J. Jensen, U.S. Patent 1,812,242 (1931).

3. P. F. Bruins, ed., Basic Principles of Rotational Molding, Gordon and Breach, New York (1971).

4. M. A. Rao and J. L. Throne, Polym. Eng. Sci., 12, 237 (1972).

5. J. L. Throne, Polym Eng. Sci., 12, 335 (1972).

6. J. L. Throne, Polym. Eng. Sci., 16, 257 (1976).

7. R.J. Crawford and J. A. Scott, Plast. Rubber Proc. Appl., 5, 239 (1985).

8. R. J. Crawford and P. J. Nugent, Plast. Rubber Proc. Appl., 7, 85 (1987).

9. R. J. Crawford and P. J. Nugent, Plast. Rubber Proc. Appl., 11, 107 (1989).

10. R. J. Crawford, P. Nugent, and W. Xin, Intern. Polym. Process., 6, 56 (1991).

11. P. Nugent, PhD, dissertation, Queen's University of Belfast (1990).

12. K. Iwakura, Y. Ohta, C. H. Chen, and J. L. White, Intern. Polym. Process, 4, 163 (1989).

13. K. Iwakura, Y. Ohta, C. H. Chen, and J. L. White, Modern Plastics, July 1990, p. 56.

14. Y. Ohta, C. H. Chen, and J. L. White, Kunststoffe, 79, 1349 (1989).

15. C. H. Chen, J. L. White, and Y. Ohta, Polym. Eng. Sci., 30, 1523 (1990).

16. C. H. Chen, J. L. White, and Y. Ohta, Intern. Polym. Process, 6, 212 (1991).

17. 17. S. Bawiskar and J. L. White, Polym. Eng. Sci. 34, 815 (1994).

18. S. S. Song, T. Nagy, and J. L. White, Intern. Polym. Process, 7, 274 (1992).