Effects of Extrusion Blow Molding Internal Cooling Technology on HDPE Container Performance

from site : idspackaging.com (original site)

Introduction

Blow molding machines are used for the production of plastic parts in many industries including packaging and automotive. Packaging, however, is the single largest user of blow molded thermoplastic containers in the US with approximately 70 percent of market (Rosato et. al., 2004).

The rigid and semi-rigid plastic packaging industry is an economic power that comprises 21 percent of the $115 billion packaging industry (Ernst and Young, 2002). Plastic containers alone totaled $11.4 billion dollars in 2001. Plastic bottle demand grows by more than four percent each year and is expected to reach 11 billion pounds by 2006 as shown in Figure 1.

Figure 1. Current and Projected US Plastic Container Demand (Freedonia Group, 2002)

The increasing demand for plastic bottles has forced bottle suppliers to install equipment and develop procedures that can produce the increased production required to meet customer demand. The steps taken to increase output have included:

• installing more equipment
• retrofitting existing equipment
• replacing existing machinery with new state-of-the-art equipment.

According to a survey by the Packaging Machinery Manufacturer’s Institute (PMMI), 55.2 percent of respondents reported replacing existing machinery with new models rated for higher output, while 39 percent reported upgrading their existing equipment with state-of-the-art retrofit kits (PMMI, 2003). The first quarter of 2003 saw $21.8 million in blow molding machinery sales (SPI, 2003). Total sales for blow molding machinery in all industries in 2003 are estimated at $505 million compared to around $350 million for 1999 (Rosato et. al., 2004).

Extrusion Blow Molding

The extrusion blow molding industry has been growing since its first successful commercially produced item, the “Stopette” deodorant squeeze bottle by Plax Corporation, in 1945. After this success, nearly every major company who made rubber machines and injection machines began to develop blow molders. Material limitations of the time made only small container blow molding possible (Belcher, 1999). High Density Polyethylene, made available in 1956, led to the success of low cost, reliable extrusion blow molded manufacturing in North America and allowed for the production of larger containers (Lee, 1990). The process of extrusion blow molding was developed from modifications made to early injection molders as well as glass container machines.

Extrusion blow molding machines use the following process (Figure 2):

• Plastic pellets of a specific material are fed into the extruder to be melted
• Melted plastic is extruded into a tube called a parison and introduced into the mold chamber
• An aluminum mold with a cavity in the shape of the desired bottle is clamped around the parison clamping the bottom shut; the parison is then cut off at the top
• A blow pin drops into the neck opening of the bottle
• Plastic is mechanically forced into the “finish” of the bottle by the blow pin tip forming the threads and mouth of the container
• The blow pin blows air into the tube, inflating the plastic tube like a balloon
• The tube is inflated into the walls of the mold cavity
• The mold cavity is water-cooled; when plastic comes in contact with the mold, the plastic cools and becomes hard
• The mold opens, the bottle flash is removed, and the cooled and formed bottle is dropped onto a conveyor to be carried to a packing or filling operation

Extrusion blow molding machines are capable of producing multiple containers simultaneously by using multiple-cavity molds. Many models, including the one owned by the Michigan State University School of Packaging, operate on a continuous rather than stepwise operating process. During the steps when a bottle is being blown and cooled, the next parison is already being formed. As soon as one bottle is cooled and dropped onto the conveyor, the mold is shuttled over and to pick up the next parison for molding, greatly reducing cycle times. Other machines use rotary technology (Figure 3) where a continuously formed parison is wrapped around a wheel and several operations are performed on the parison as it advances around the wheel.

Figure 3. Rotary Extrusion Blow Molder

Figure 2. Conventional Extrusion Blow Mold Process

The wheel contains multiple molds and bottles are formed in each one. The wheel contains the air blowing mechanism, which punctures the parison to feed air inside. Rotary machinery is very expensive, due to the cost of the multiple molds and blow pins around the machine, but can acheive very high production rates. Rotary machines must produce millions of bottles to remain economical (Lee, 1990).

Manufacturers that do not produce extremely high volumes of bottles, but still need to add capacity, are looking for new technologies that will improve the throughput of traditional blow molding equipment. One of the best areas to reduce cycle times is in the cooling and blowing stages. Total molding time for a single 22oz bottle on the current machine is 9.51 seconds; 6 seconds of this is the time required to inflate the parison into the mold and cool it, forming the bottle. In addition, 0.5 seconds is used to exhaust the air from the bottle before the cooled bottle is dropped out of the machine. These two processes make up more than two-thirds of the total cycle time. Reducing the time required for each of these stages would drastically lower cycle times and increase bottle output.

The blowing stage is required to inflate the bottle and to cool the plastic. The pressurized air is injected into the inside of the container to hold the plastic against the mold cavity. The mold cavity is water-cooled and heat is transferred out of the plastic by conduction to cool it to a temperature where it can maintain its shape and dimensions. The sooner the bottle is brought down to a stable temperature, the sooner it can be removed from the mold and another bottle can be formed.

There are several ways to reduce bottle cooling time. Traditionally, mold-makers have built cooling channels in the mold to reach all parts of the bottle, but focusing on the thickest parts of the container, which require the most cooling. Some companies have taken cooling a step farther, opting for internal cooling to speed up cool-down times, thereby reducing blow and exhaust cycle times.

Internal cooling is the process of cooling the blow-molded part from the inside out. A standard blow pin inflates the container with room temperature, dry air. The air is forced into the container, where it remains during the entire blow-cycle and is allowed to escape during the exhaust cycle. While the air sits in the inflating container, it remains stagnant and does little to contribute to cooling since the hot plastic heats it. Water spray, liquid CO 2, nitrogen, circulated air, and supercooled air are all methods of internal cooling. Water spray methods work by spraying an air and water mix into the part. The water helps to cool the part when it is evaporated into steam. The continuous air stream cycles the steam out of the part through a pressure release valve. This method requires that the water mix be introduced after the inside walls of the container are solidified to prevent the water from affecting the internal surface finish. Finally, dry air must be blown in to remove any additional water left behind. Air/water sprays can reduce cycle times by as much as one-third but any residual water left behind can be undesirable in applications involving moisture sensitive product or where contamination is an issue (Lee, 1990). Liquid CO 2 and nitrogen systems work in similar ways. Internal cooling through cold circulated air is the basis of the Fasti Blow Mold Booster system.

Fasti System

The Fasti system is to internally cools the container. The most important feature of this technique is the continuous removal of hot air while the system simultaneously introduces cold air into the part. This process of cycling the air is achieved by using a recirculating blow pin.

Recirculating blow pin

The blow pin must be engineered for each specific machine and application. The pin consists of a central exhaust pipe, which pulls hot air out of the bottle and a fitting which channels cold air into the bottle around the outside of the exhaust pipe (See Figure 9). In addition, small channels direct air around the circumference of the blow pin externally at the cooling sleeve to cool the moil or top flashing.

Figure 4. Recirculating Blow Pin

Blowing of the bottle becomes a three-stage process with the Fasti system described below and shown in Figure 5.

1. The pre-blow phase uses chilled air through both blow pin channels to inflate the parison inside the mold as quickly as possible, while the blow pin is entering the bottle finish. The high-pressure pre-blow process forms the container, forcing out ambient air between the mold and parison through the mold vents. The bottle is inflated and comes in contact with the water-cooled cavity surfaces, which promotes the cooling of the bottle.
2. The blowing phase is used to do the actual cooling of the bottle with cold air. Upon inflation of the bottle in the pre-blow stage, only a small amount of back-pressure is required to maintain contact with the mold cavity walls. During this stage, air flow through the center channel of the blow pin is reversed, allowing hot air to escape while cold air is blown in through the outer channel. Cold air entering the container forces the cold air out the center channel. This allows the cold air to circulate instead of trapping the hot air inside the container as with a conventional process.
3. Finally, the venting stage where the air pressure is balanced between the container and the outside. At this stage, the container will have cooled sufficiently to maintain its dimensionality.

Figure 5. Fasti Blow Stages

A new recirculating blow pin costs from six hundred to seven hundred dollars, comparable to the cost of a traditional pin, but can vary due to size and complexity.

Cooling system The air cooling system itself is quite simple. Air is forced into the system and cooled by cold water in a heat exchanger. A refrigeration system or cold water can be used. The cooled air is then directed through the blow pin for bottle blowing.

Costs 
Total cost of the unit is listed in Table 1 along with an estimate of equipment payback time based on a three-shift production schedule at 80% efficiency.

Installation Installation of the equipment can require significant modifications to the machine, such as the mounting brackets for the cooling unit and the valves that were installed for the project reported in this paper. The objective was to locate the cooling unit at an elevation higher than the blow pin and with as short a distance between the blow pin, cooling unit, and valves as possible. To accomplish this, the cooling unit was mounted on top of the machine surround with a substantial bracket able to support its 90 pound weight. The mounting tray consisted included a three-point support with the main weight of the unit being carried by the machine surround and a third mount point inside the cabinet on a mounting plate. The mounting tray included a lip to prevent the unit from vibrating off. The valve units were mounted to the sheet metal sides of the machine surround using sandwich plates to stabilize them. This spread out the load further.

Efficiency 
Fasti USA claims to increase production by reducing air blow cycles by as much as 35%. The entire forming process for one bottle is 9.51 seconds, with 6 seconds of that being used for blowing of the ottle. If a bottle machine produces one bottle every 9.51 seconds, that machine will produce 379 bottles in a 1-hour period. Reducing forming time to 7.42 seconds will allow the machine to produce 485 bottles in 1 hour, a 28% production increase. Production improvements like this could have significant impacts on any industry that uses blow-molding processes.

Costs
Blow Mold Booster Unit Cost	$11,710.00
Blow Valve Blocks	$1,490.00
Blow Pin Design Cost	$1,100.00
Blow Pin Machining Cost	$700.00
Installed Investment	$15,000.00
CostsValues in US$
	Conventional	Fasti-BMB
Operating Efficiency	80%	80%
Base Machine-Hour Cost	$80.00	$80.00
Number of Operators per Machine	1	1
Cost per Man-Hour	$20.00	$20.00
Productive Hours per Day	19.2	19.2
Product Weight, grams	25	25
Cost of Resin per Kilo	$1.25	$1.25
Cost of Resin per Part	$0.03	$0.03
Energy Consumption, KW/h	(Included in Base Cost)	7
Energy Cost per KW/h	$0.10	$0.10
Additional Energy Cost per Hour	$0.00	$0.70
Air Consumption, Nm3/h	(Included in Base Cost)	20
Cost of Compressed Air/Nm3/h	$0.02	$0.02
Additional Air Cost/h	$0.00	$0.39
Total Additional Costs	$0.00	$1.09
Manufacturing Cost per Hour	$100.00	$101.09
Increase in Production, %	0%	28%
Number of Parts Produced per Hour	379	485
Cost of Resin Consumed per Hour	$11.37	$14.55
Total Cost per Part	$0.29	$0.24
Savings per Part	N/A	$0.06
Savings per Day	N/A	$516.06
PAYBACK PERIOD, DAYS		29.1

Table 1. Fasti Cost Analysis and Payback Period Estimate (FastiUSA, 2004)

Figure 6. Machine Surround Before                         Figure 7. Machine Surround After

Objective 
The question that remains is:

What effects do the cooler temperatures and reduced cycle times
have on the physical properties of the finished bottle?

Crystallinity 
Forming temperatures have a direct effect on the crystallinity and density of HDPE containers. Polymer crystallinity and density have a direct correlation with such properties as clarity, permeability, column crush strength, and impact strength (Hernandez, Selke, and Culter, 2000). The objective the research was to quantify the effect(s) that lower processing temperatures have on bottle performance in these areas. Table 2 shows the effect of decreased crystallinity on various bottle properties

Crystallinity	decreases
Density	decreases
Permeability	increases
Opacity	decreases
Blocking	increases
Tensile Strength	decreases
Compression Strength	decreases
Clarity	increases
Tear Resistance	increases
Impact Strength	increases
Toughness	increases
Ductility	increases
Ultimate Elongation	increases
Heat Sealing Temperature	decreases
Heat Sealing Range	increases

Table 2. Effect of Decreased Crystallinity in Polymers
(Hernandez et al, 2000)


As is evident in the table, a reduction in crystallinity can have significant effects on bottle performance. Most notable for those involved in packaging are increased impact strength, reduced compression strength, and increased permeability.

“During processing, the major difference between amorphous and crystalline polymers is that amorphous polymers gradually lose their molecular mobility as the material cools, whereas crystallizing polymers (like HDPE) change suddenly from mobile liquids to crystalline solids at a sharply-defined melting/freezing point” (Rosato et. al., 2004). For this reason, crystallizing polymers are more difficult to blow mold because of their narrow workable temperature range. The rate of crystallization can be controlled by the cooling process and ultimate crystallinity may be reduced by quenching (Rosato et. al., 2004).

“Crystallization is useful in blow molding because (1) it freezes [the container] in the stretch orientation and thus gives the oriented structure permanence; and (2) it improves many end-use properties of particular importance in food packaging, including rigidity, dimensional stability on reheating, and impermeability. However, crystallization also tends to degrade some useful properties, such as ultimate elongation, impact strength, transparency, and environmental stress crack resistance” (Rosato et. al., 2004).

If we assume thatis proportional to the % crystallinity of the test ΔH_f specimen and if we know the ΔH_f of the test specimen in pure crystalline form (100% crystallinity), we can compute the % crystallinity as follows:

Where:

ΔH_f = heat of fusion of semi-crystalline polymer, J/g
ΔH_f* = heat of fusion of 100% crystalline sample, J/g. For PE, this value is 286.2 J/g (Selke and Xiong, 2003).

It was expected that the Fasti unit’s internal cooling technology would be likely to affect crystallinity of the finished container due to the quicker cooling and therefore shorter period of time during which the bottle is at its crystallization temperature.

Wall thickness 
In addition to impacting crystallinity, the Fasti system may also affect the wall thickness of the finished containers; reducing the amount of time that the plastic has to flow out into the mold may affect material distribution. Distribution changes may be solved using the parison programming to adjust the profile and maintain uniform wall thickness throughout the bottle and between forming methods

Dimensionality 
Bottle dimensions may be affected by cold air blowing. It is likely that dimensional stability will be affected by internal cooling. The quicker cooling and therefore “freezing” of the container shape will result in less warpage. Dimensional stability of the container is critical to maintain tolerances for the filling operations in terms of volume as well as dimensions in the finish area to allow the closure to work well with the container.

Hypothesis An experiment was designed to test the hypothesis:

H_o: The use of internal cooling does not significantly impact the physical properties of extrusion blow molded HDPE bottles.

Experimental Design and Test Methods

Materials and Setup Setup of the stagnant blow settings was accomplished with the help of Bekum America. The settings are set so that the container is formed in the shortest amount of time possible that would still allow complete formation of the container. The cycle time of this setup as shown in Table 3 is 9.51 seconds. The parison program was designed to create approximately equal wall thickness throughout the wall and heel of the container. These set points are saved in the machine as “16OZ BOTTLE THESIS” along with the parison program.

Action	Time (sec)
Extend Knife Delay	0.63
Retract Knife Cut Delay	0.63
Mold Close Delay Time	0.00
Carriage Down Delay	0.25
Blowing Delay	0.12
Blowing Time	6.00
Exhaust Time	0.50
Deflash Delay	0.00
Blow pin 1st Step Delay Time	0.14
Blow pin 2nd Step Delay Time	0.10
Container Blowoff Delay	0.00
Container Blowoff Time	0.00
Carriage Up Delay	0.30
Fasti Delay	~~~
Knife Pulse Cut	0.16
Carriage Up First Cycle Delay	2.00
Mold Crack Time	0.25
Mold Crack Hold	0.40
Controlled Support Air Delay	2.00
Controlled Support Air Time	2.00
Machine Cycle Timer	9.51
Blow Pressure (psi)	65.00
Back Pressure (psi)	~~~
Extruder Temperature Set points (deg F)	350.00

Table 3. Bekum Blow Molder Stagnant Blow Set Points


Fasti Cold Air Setup Setup of the Fasti cold air blow settings was accomplished using trial and error. The settings are set so that the container is formed in the shortest amount of time possible that would still allow complete formation of the container. The cycle time of this setup as shown in Table 4 is 7.42 seconds. The parison program was designed to create approximately equal wall thickness throughout the wall and heel of the container. Another experiment, shown in Appendix C, attempted to determine the connection between various timer and pressure settings with container volume or shrinkage. The findings from this study were taken into consideration when setting up the machine. These set points are saved in the machine as “16OZ BOTTLE FASTI THESIS” along with the parison program.

Action	Time (sec)
Extend Knife Delay	0.63
Retract Knife Cut Delay	0.63
Mold Close Delay Time	0.00
Carriage Down Delay	0.25
Blowing Delay	0.12
Blowing Time	4.50
Exhaust Time	0.35
Deflash Delay	0.00
Blow pin 1st Step Delay Time	0.14
Blow pin 2nd Step Delay Time	0.10
Container Blowoff Delay	0.00
Container Blowoff Time	0.00
Carriage Up Delay	0.30
Fasti Delay	0.85
Knife Pulse Cut	0.16
Carriage Up First Cycle Delay	2.00
Mold Crack Time	0.25
Mold Crack Hold	0.40
Controlled Support Air Delay	2.00
Controlled Support Air Time	2.00
Machine Cycle Timer	7.40
Blow Pressure (psi)	85.00
Back Pressure (psi)	18.00
Extruder Temperature Set points (deg F)	350.00

Table 4. Bekum Blow Molder Cold Air Set Points


Controls (Constants) The following pieces of equipment were used in the experiments and setup of the machine. These items remained constant through all tests.

Plastic Type

• Union Carbide UNIPOL Polyolefin DMDA-6220 NT7 UNIVAL

Mold

• Manufactured by MC Molds, Williamston, MI

Blow Mold Machine

• Model H-111S Bekum America, Williamston, MI

Fasti Blow Mold Booster II (BMB II)

• FastiUSA, Elgin, IL

Closure

• Rexam Closures and Containers Evansville, IN
• Cap style:      28 DECO CC2 SPECIAL
• Color:             Any
• Material:        PLS 10
• Liner:             827
• Description:   W01 base/GA4 lid crabsclaw o/s
• • Orifice:        0.155

Experimental Methods

Sampling The sampling procedure was designed with the objective to obtain consistent samples in a regulated manner to facilitate labeling and tracking of containers.

Containers were manufactured by the two different manufacturing methods: Fasti Cold Air blow and Conventional blow. Short run times were necessary due to machinery limitations, namely, air supply was inconsistent and only allowed production of 30 containers before pressures fell below the specified settings.

The manufacturing process was started up. The first five containers retrieved from the machine were discarded. From then on, each container made was removed in order and placed inverted (finish down) in a divided, numbered sample tray. The containers were inverted to give them time to cool and to prevent the bottom pinch-off from becoming fused to the container. 30 Samples were made per run. The bottles were laid out in the sample trays as shown in Figure 9.

Figure 8. Sample Tray Layout

Five minutes after the cycle was complete, the pinch-offs were removed. The containers were then labeled by tray location and placed right-side-up in a new sample tray. The sample tray was labeled with the date, run number, and manufacturing method (Fasti or Conventional).

Conditioning The container samples were conditioned after processing to ensure that the results were not influenced by environmental conditions .

Conditioning was performed in accordance with ASTM D-618-00 Standard Practice for Conditioning Plastics for Testing. The labeled sample trays were stored for a minimum of 40 hours in the room where the bottles were formed before being tested. Average temperature, checked by thermometer, of this room was 24.2 degrees Celsius.

Dimensions The container samples were measured to determine the overall dimensions of a container and determine variance among samples. Equipmenmt used included: 1) Scherr-Tumico Industries Model 20-3500 Optical Comparator (0.0001”), 2) Mitutoyo Model CD-6”BS Digital Caliper (0.0005”), and 3) Mettler PM2000 Scale (0.01g) Magna-mike (0.0001”)

Container samples were measured according to ASTM D 2911-94 Standard Specification for Dimensions and Tolerances for Plastic Bottles, as follows:

Finish : Finish dimensions were measured using the optical comparator.
Typical tolerances for a finish of this size have a range of around 0.020 inches.
The T dimension indicates the diameter of the finish at the tips of the threads.
The E dimension indicates the diameter of the finish at the base of the threads.
The H dimension indicates the distance from the top of the bottle to the transfer bead and the I dimension indicates the inside diameter of the finish area. A diagram of the location of the various dimensions is shown in Figure 10.

Figure 9. Finish Dimensions

Volume: To determine container volume, the containers were weighed when empty, and then filled with water conditioned according to ASTM C2911-94 and weighed again. Container volume was calculated as follows:

B_v(mL)=(B_f−B_e)/997.0

Body Dimensions: Width was taken to be the average of the measurements at the parting line and then rotated 90 degrees. The width was measured using calipers 3” from the bottom of the container.

Range of Dimensions			Width Tolerance
	inches		inches
0	up to but not including	1	0.03
1	up to but not including	2	0.05
2	up to but not including	4	0.06
4	up to but not including	6	0.08
6	up to but not including	8	0.09
8	up to but not including	10	0.11

Table 5. Bottle Body Dimension Tolerances (ASTM D 2911-94)


Body Wall thickness: Wall thickness was measured using the Magna-mike with measurements taken at 0.25” increments up the container wall, as shown in Figure 11, as well as measurements in the heel. These measurements were taken at the parting line by the bottom detent at 12:00 as shown in Figure 12 and then repeated every 90 degrees around the container for a total of 64 measurements.

Figure 10. Magna-Mike Measurement Locations

Figure 11. Container Rotation Callouts

Compression Testing The objective of the column crush test was to collect information about the crushing properties of blown thermoplastic containers. Column crush properties include the crushing yield load, deflection at crushing yield load, crushing load at failure, and apparent crushing stiffness. The compression tests were performed with a Lansmont Corporation Squeezer Compression Tester (0.1 lbs, 0.001 in) ansd a Mettler AE 160 Scale (0.0001g).

Crush testing was performed in accordance with ASTM D 2659-95 Standard Test Method for Column Crush Properties of Blown Thermoplastic Containers (ASTM D 2659, 1995). Twenty samples from each manufacturing method were tested as shown in Figure 13 to determine crushing yield load, deflection at crushing yield load, crushing load at failure, and apparent crushing stiffness. A modified closure was applied to the container. The closure had a vent hole which allowed air to escape during testing as shown in Figure 14. The crown of the closure prevented the hole from sealing and causing pressure to build in the container which could affect compression strength.

Figure 12. Compression Testing Setup

Figure 13. Compression Testing Vent Hole

The data from the compression tester was exported to an Excel file for analysis. The crushing yield load, deflection at crushing yield load, and apparent crushing stiffness were extrapolated from the data as follows:

Crushing Yield Load was determined as the point on the crush load/deflection curve at which an increase in deflection occurs without an increase in crush load and was expressed in lbs. to three significant figures (Figure 15).

Deflection at Crushing Yield Load was the reduction in height (x-axis of Figure 15) of the sample at the crushing yield load expressed in inches to three significant figures.

Apparent Crushing Stiffness was calculated by selecting a point on the straight line segment of the crush load/deflection curve as shown in Figure 15 and dividing force at this point by the corresponding deflection expressed in pounds per inch to three significant figures.

Figure 14. Compression Data Example

Differential Scanning Calorimetry The objective of the studies using Differential Scanning Calorimetry was to determine the melting point and percent crystallinity of HDPE samples.

The instruments used were: 1) TA Instruments DSC Q 100 Differential Scanning Calorimeter and 2) Mettler AE160 scale (0.0001g). The procedure used is outlined below.

• Handle all samples with tweezers
• Cut a 9-10 milligram sample from the container
• Weigh the sample in the bottom aluminum pan
• Record sample weight
• Apply the top pan and crimp sample closed as shown in Figure 16
• Place sample in DSC centered on thermocouple
• Using the settings shown in Table 6, run the experiment
• Integrate the curve shown by the analysis program shown in Figure 17 with the curve starting at 60°C and ending at 150°C
• Calculate percent crystallinity from the result using 286.2 J/g as the baseline for 100% crystalline HDPE (Selke and Xiong, 2003)

Figure 15. DSC Sample Pan Crimper

1	Ramp 20.000°C/mi to 180.00°C
2	Mark end of Cycle 0
3	Isothermal for 2.00 min
4	Ramp 20.000°C/mi to 40.00°C
5	Mark end of Cycle 1
6	Isothermal for 2.00 min
7	Ramp 20.000°C/mi to 180.00°C
8	Mark end of Cycle 2

Table 6. DSC Heat/Cool/Heat Setup Method

DATA AND RESULTS

Cycle time improvements 
The improvement of cycle time was significant. Use of the Fasti system accomplishing a 22% decrease from 9.51 seconds per cycle to 7.42 seconds per cycle. Containers manufactured using the two methods were complete and correctly formed and similar in appearance. Dimensionality

Comparison of dimensions of containers made with the Fasti and conventional manufacturing methods reveal several differences as shown in Table 7. P-values shown in bold show statistical significance. The most striking difference between the containers is an overall shrinkage of the Fasti containers. While shrinkage after the forming process is common, this effect appeared to be magnified by the Fasti system. While the Fasti containers shrank more than the conventional bottles, the shrinkage was consistent across containers as displayed by the low standard deviation found among samples. Shrinkage does not entirely account for the reduced volume of the Fasti containers. This is explained in part by the increased bottle weight. More resin in the container walls makes the volume inside the container smaller.

Warpage after the forming process is also a common occurrence. Warpage of the body area of the Fasti containers was significantly lessened with the Fasti process as can seen in the “Body Diameter Difference” category of Table 7. This category is a calculation of the differences in diameter of the container across the parting line versus turned 90 degrees from the parting line. The decreased warpage is best attributed to the uniformity of the wall thickness at the measuring point. The ontainer diameter was measured 3 inches from the bottom of the container at point 8 of the wall thickness measurements. Review of the wall thickness found in Appendix A proves that decrease of warpage is not related to varying wall thickness around the container. The likely reason for reduced warpage of the container is the internal cooling. Container thickness at the diameter measuring point is relatively thin and would receive the most cooling. This proves that the Fasti unit will actually “freeze” the container in place with more thorough cooling.

Table 7. Dimensional Test Result Comparison – Conventional vs. Fasti

The Fasti containers had a more consistent wall thickness throughout the container while conventional containers had thinner walls near the top as shown by the wall thickness data in Appendix A. The most statistically significant points (those with p-values less than 0.001) are shown in bold in Table 7. These values coincide with the thinnest spots on the conventional containers. This can be explained by the extrusion speed. The Fasti containers used a higher extrusion rate in order to prepare the parison faster. The slower extrusion rate of the conventional process allowed the parison to stretch under its own weight and therefore thin out near the top. This issue could be easily resolved by altering the parison profile or possibly by lowering the melt temperature in the conventional process.

Table 8. t-Score p-Values for Wall Thickness

Compression Strength Compression strength tests performed on the finished containers revealed very little difference between processes. In both cases, failure was seen in the heel area resulting in buckling of the bottle walls. Table 9 shows a comparison of the compression test results for conventional versus Fasti containers. The p-values in the table are less than 0.01 in each category showing that the data is statistically significant. The statistical significance however, does not speak of the practical significance of the effect the Fasti system has on the container. The higher crushing yield load of the Fasti container is likely attributed to the slightly higher average thickness of the container at the crush failure point displayed in the data in Appendix A.

Table 8 shows significant data difference in the apparent crushing stiffness of the containers to theorize that the Fasti containers may have a lower percent crystallinity and therefore lower stiffness as expressed in Table 9. This theory was tested in Section 3.4. All compression strength data is displayed in Appendix B.

Table 9. Compression Test Results – Conventional vs. Fasti

Crystallinity

Crystallinity of the container is determined by processing conditions. The average crystallinity of the virgin HDPE resin pellets, as displayed in Table 10, is 76.32%.

Table 10. Percent Crystallinity of Virgin HDPE Resin

	J/g	%Cry
Pellet 1	220.8	77.15%
Pellet 2	213.8	74.70%
Pellet 3	220.7	77.11%
Avg	218.4	76.32%

Initial crystallinity tests involved removing a sample from three points on the bottle: in the heel, in the body 3 inches above the bottom, and in the shoulder. These samples were taken without regard to their orientation on the container as far as container rotation. This may explain the large variance percent crystallinity between container samples shown in Table 11. For example, the shoulder sample from conventional bottle 28 from sample tray as shown in Figure 9 was taken at the 12:00 position while conventional bottle 29 was taken from 5:00. Wall thickness data in Figures 18, 19, and 20 in Appendix A show significant wall thickness differences from point to point around the container at the shoulder (thickness measurement 16).

Table 11. Percent Crystallinity of Conventional vs. Fasti Containers

	Body	Shoulder	Heel
Conventional
Bottle 28	218.7	219.1	231.6
Bottle 29	219.2	234.4	226.5
Bottle 30	220	228.8	223.5
Avg	219.3	227.4333	227.2
%-Crystallinity	76.62%	79.47%	9.39%
Fasti
Bottle 28	230.8	233.2	220.3
Bottle 29	234.8	227.9	209.8
Bottle 30	230.9	228.9	215.5
Avg	232.1667	230	215.2
%-Crystallinity	81.12%	80.36%	75.19%

More thorough and controlled testing was done on a single sample container for each conventional and Fasti container at the heel. A sample was taken at each of the four points of rotation as shown in Figure 12 after the container was measured for thickness. The thickness measurement at the sample area corresponds to measurement point 2 of the data shown in Figure 20 and Figure 23 in Appendix A. Table 12 shows the thickness at each measuring point and its corresponding percent crystallinity. This data shows that there is no significant correlation between wall thickness and percent crystallinity in the heel area.

Table 12. Percent Crystallinity at Points in Heel

	12:00	3:00	6:00	9:00
Conventional
Thickness (in.)	0.0261	0.0175	0.0256	0.0192
J/g	220.5	222.8	226.8	234.5
%-Crystallinity	77.04%	77.85%	79.25%	81.94%
Fasti
Thickness (in.)	0.0223	0.0180	0.0289	0.0177
J/g	225.8	225.7	231	236.7
%-Crystallinity	78.90%	78.86%	80.71%	82.70%

Of further interest is the difference in crystallinity between conventional and Fasti containers. The data in Table 12 shows no significant difference in percent crystallinity between manufacturing methods.

CONCLUSIONS AND RECOMMENDATIONS

The addition of Fasti internal cooling technology to the Bekum blow molder significantly increased production rates of the machine. Production increases as high as 22% were seen between conventional and Fasti molding methods despite limitations with our equipment including insufficient air-supply. An increased air supply would likely drastically lower the blow air temperature released from the Fasti unit. Decreased blow temperatures would result in more thorough cooling and more stable container properties.

Dimensional analysis of containers proves that reduced warpage is a positive effect of internal cooling. The proof is found in the more consistent container diameter of Fasti containers, which is not associated with wall thickness. The thorough cooling and therefore setting of the body walls with low blow air temperatures reduced warpage after being released from the mold. Further proof of this could come from continued research upon implementation of a more reliable air supply. Fasti container formation resulted in a higher overall container weight which caused a reduction in volume. Without container and closure drawings it is difficult to determine if the container finish dimensions fell within desired tolerances but a low standard deviation among samples of the same manufacturing method displayed consistency and low standard deviation.

Compression tested containers showed very little difference in performance that could be directly attributed to changes in crystallinity, container dimensions, or wall thickness. Compression strength could be improved for certain applications by increasing wall thickness in the heel area.

Study of the containers by Differential Scanning Calorimetry at various container locations reveals that the percent crystallinity of the container is independent of container wall thickness. Furthermore, the percent crystallinity did not appear to be dramatically affected by blow temperature.

It should be noted that the tests were run on containers produced through short run times. Increasing the reliability of the air supply and therefore lengthening run cycles would produce greater sample sizes and allow more thorough testing.

In summary, the Fasti internal cooling technology greatly increased output without making considerable changes to container performance. Any shortcoming found with the system can easily be programmed out with a combination of changes in processing temperatures, air flow, and parison programming. The positive effects of the Fasti system including reduced cycle times and reduced warpage would likely be magnified with the use of more air volume resulting in lower temperatures.

RECOMMENDATIONS FOR FUTURE RESEARCH

Redesign of Mold 
The mold that was supplied with the machine uses a finish which is very outdated. The threads used are a custom thread design available only from Rexam Closures. Various attempts to acquire sample caps, finish drawings, production tolerances, turned up little information. A very small number of caps were acquired for studies but the hinged lid is not acceptable for many performance tests. It would be desirable to redesign the finish area of the mold. This section is removable and could be replaced with a more standard thread for which closures are more readily available.

In addition, the use of the Fasti system would produce better quality bottles with less shrinkage in the finish area if the water channels allowed better cooling. Redesigning the mold for the new finish would allow the opportunity to redesign these cooling channels.

Regrind 
The School of Packaging has a granulator which could be used for regrinding containers. Studies could then be conducted regarding the changes in processing temperatures and conditions with the use of the regrind material. Furthermore, studies could be done using various mixes of regrind and virgin material and their effects on container performance.

Different Containers
The use of the Fasti machine on a 16 fluid ounce container from this mold does not compare to the benefits realized from a larger container with thicker walls. According to Fasti, a common application for the Fasti unit is for the production of blow molded gas tanks. These containers have very thick coextruded walls requiring long blow times to cool the tank and set the plastic. Factors like shrinkage may be magnified in the larger part. Investigating the effect of container size and volume on cooling time and container shrinkage is recommended.

Environmental Stress Cracking 
Another important test which could be run on these containers is an environmental stress cracking test. Comparisons could be made between the two manufacturing methods following the test procedures outlined in ASTM D2561-95, Environmental Stress-Crack Resistance of Blow-Molded Polyethylene Containers.

Impact Testing 
Impact testing of HDPE containers is difficult through standard testing methods including ASTM D 2463-95 Drop Impact Resistance of Blow-Molded Thermoplastic Containers. The problem lies in the incredibly high inherent strength of the materials. Preliminary testing failed to produce any impact failure at all. Possible solutions to this problem could include freezing of the containers before testing to increase brittleness.

Torque Testing 
The different manufacturing methods could potentially produce different results in closure torque testing. Application and removal torque could be tested for each of the containers.

Optical Microscopy 
Optical microscopy is another method for analyzing the physical composition of a container including crystallinity by observing crystalline regions.

Permeability Testing 
Permeability could be conducted on different containers produced by the machine.

Resin 
Various resins could be tested in this machine with the current setup. The current extruder screw is capable of running polyethylene including HDPE and LDPE as well as polyethylene blends. Further information could be obtained from Bekum America regarding the compatibility of this screw for other materials as well as possibly acquiring a new screw capable of a wider array of material compatibility.

REFERENCES

ASTM Standard D2463-95. (1995). Standard Test Method for Drop Impact Resistance of Blow-Molded Thermoplastic Containers. Annual Book of ASTM Standards. vol 8.02. West Conshohocken, PA: ASTM International.

ASTM Standard D2659-95. (1995). Standard Test Method for Column Crush Properties of Blown Thermoplastic Containers,” Annual Book of ASTM Standards. vol 8.02. West Conshohocken, PA: ASTM International.

ASTM Standard D2911-94. (1994). Standard Specification for Dimensions and Tolerances for Plastic Bottles. Annual Book of ASTM Standards. vol 8.02. West Conshohocken, PA: ASTM International.

ASTM Standard D3417-99. (1999). Standard Test Method for Enthalpies of Fusion and Crystallization of Polymers by Differential Scanning Calorimetry (DSC). Annual Book of ASTM Standards. vol 8.02. West Conshohocken, PA: ASTM International.

Belcher, Samuel L. (1999). Practical Extrusion Blow Molding. NY: Marcel Dekker, Inc.

Ernst and Young. (2002). 2002 Global Packaging Report – The Top 100 Players.

Fasti USA Website. (2004). Internal Cooling for the Blow Molding Industry. <http://www.fastiusa.com/product3.php >

Freedonia Group. (2002). U.S. Plastic Container Demand. Plastics News.

Hernandez, Selke, and Culter. (2000). Plastics Packaging: Properties, Processing, Applications, and Regulations. Cincinnati, OH: Hanser Gardner Publications.

Lee, Norman C. (1990). Plastic Blow Molding Handbook. NY: Van Nostrand Reinhold.

Packaging Manufacturing Machinery Institute. (2003). 2003 U.S. Productivity & Profitability Trends Indicator Study Executive Summary. PMMI Publishing.

Rosato, Rosato, and DiMattia. (2004). Blow Molding Handbook. Cincinnati, OH: Hanser Gardner Publications.

Selke, S. and Xiong, L. (2003) PKG 829 Lab 3 Determination of Percent Crystallinity of Polyethylene by the Density Gradient and Differential Scanning Calorimetry.

Sichina, W.J. (2000). DSC as Problem Solving Tool: Measurement of Percent Crystallinity of Thermoplastics. <www.perkinelmer.com >

Sichina, W.J. (2000). Application of DSC to Injection Molding. < www.perkinelmer.com >

Society of Plastics Industry. (2003). U.S. blow molding equipment sales (2002-1Q 2003). Plastics News.