Water Assisted Injection Moulding: a Study on the Influence of Melt and Process Parameters on the Residual Wall Thickness and the Occurrence of Part Defects for Pa6.

KATHOLIEKE HOGESCHOOL LIMBURG Departement Industriele Wetenschappen en Technologie Water Assisted Injection Moulding: A Study on the Influence of Melt and Process Parameters on the Residual Wall Thickness and the Occurrence of Part Defects for PA6. Gielen Bjorn, Heynickx Lien, Van den Brink Dominique, Vandersteegen Jochen Water Assisted Injection Moulding: A Study on the Influence of Melt and Process Parameters on the Residual Wall Thickness and the Occurrence of Part Defects for PA6. Gielen Bjorn, Heynickx Lien, Van den Brink Dominique, Vandersteegen Jochen Received 7 March 2011 Abstract Although water assisted injection moulding is a relatively new and advanced technique to produce hollow parts, several problems which reduce the product quality can occur. Namely, water inclusions in the wall, severe fingering and double wall defects reduce the product part stiffness. This study investigated the effect of the process parameters water volume flow rate, water injection delay time and melt temperature on the part defects to obtain a better understanding of the formation of these part defects. The effect on the residual wall thickness and the weight of the products was also investigated. Three materials were used: two polyamides and one polypropylene as a reference. The experiments were carried out on an injection moulding machine equipped with a water-injection unit to produce hollow handles. The parts were weighed before and after drying, their wall thickness was measured and their percentage of defects was calculated in order to determine the influence of the processing parameters. IR images and pressure profiles ware also used to investigate possible detection methods for part defects. It was found that all the investigated parameters have a significant influence on the residual wall thickness and part defects. Their ‘ideal’ settings depend on the material and the desired properties since small wall thicknesses and few part defects do not always correspond. IR images can give an indication for the presence of defects, but the internal structure of the surface can not be examined using this technique. Key words: water assisted injection moulding, polyamide, processing parameters, residual wall thickness, fingering, double wall 1. 1. Introduction The development of the water assisted injection moulding (WAIM) technique has led to a breakthrough in the manufacturing of hollow or partly hollow plastic parts, which is not possible with the conventional injection moulding process [1] [2]. These hollow parts are characterised by the lighter weight of the products, relatively lower resin costs per part and less shrinkage and warpage compared to conventional moulded products [3]. Several techniques are available for the production of hollow parts. Two of the most important techniques for hollow parts are gas assisted injection moulding (GAIM) and water assisted injection moulding (WAIM). The principle of water assisted injection moulding is basically similar to gas assisted injection moulding, but the use of water has some interesting advantages over gas assisted injection moulding such as a shorter cycle time and an average smaller residual wall thickness of the products. This is due to the fact that the thermal conductivity and the heat capacity of water are respectively 40 and 4 times greater than those of gas [1], providing a better cooling of the part and therefore a faster production. Water assisted injection moulding is better to mould larger parts with a smaller residual wall thickness. Another advantage of water assisted injection moulding is the absence of the foaming phenomenon in the internal surface which can occur by using gas assisted injection moulding , since water does not dissolve or diffuse in the polymer melt during injection moulding. Disadvantages of water assisted injection moulding are part defects such as fingering and double wall arising in the internal surface. Also, the process is more complex so experience with conventional injection moulding is no guarantee for understanding water assisted injection moulding. The control f the process is much more critical, the design of the mould is more complicated and requires more knowledge of the dynamic interaction between the polymer and the filling water [4]. In addition, new water related processing parameters (water pressure, water injection delay time, water temperature, †¦ ) are involved and require proper adjustment [5]. The technology can be used to mould a large variety of polymers, includi ng glass fibre reinforced composites. Liu [6] reported the successful moulding of PP, PE, ABS and glass fibre filled polyamide 6 (PA6) composites with water assisted injection moulding. According to our investigation, few researchers have described the moulding of pure polyamide 6 with the water assisted injection moulding technique. By understanding the water assisted injection moulding process for this polymer, one is able to optimize the production and improve the product quality of PA6 parts. The water assisted injection moulding process can be applied in four variations [1] [6], which differ by the amount of injected polymer and introduction and evacuation of the water. In the short shot moulding process the cavity is first partially filled with polymer and in the next sequence, highly pressurised water is injected into the polymer core. The pressure is maintained during the cooling sequence to minimize the shrinkage and to further cool down the part. After cooling, the water is evacuated and the part is ejected. This technique is useful for thick parts, but the important limitation is the switchover mark on the surface [7]. This limitation can be eliminated by using the full shot moulding process in which the cavity is first completely filled with polymer. Next, the injection of water pushes the melt into a cavity positioned outside the cavity of the product. This overspill cavity arises by pulling back a core to enlarge the mould cavity for the overspill polymer which has to be removed from the product afterwards. Full shot moulding makes it possible to mould parts with a more uniform residual wall thickness through the end of the part in comparison to short shot moulded parts. Michaeli et al. [6] stated that the melt-push-back process is also capable to produce parts with a more uniform residual wall thickness distribution. In this process the water is injected in the reversed direction of the polymer injection and pushes the excess of melt back into the plasticizing unit. Finally, the core-pulling process uses a series of steel cores that are pulled back during water injection to increase the cavity volume and the product volume. This method produces parts with more uniform internal profiles in comparison to the former mentioned techniques. The investigated polymer, polyamide 6, is a semi-crystalline material that is classified as an engineering plastic. Beneficial properties of polyamide 6 over a wide range of temperatures and humidity are its high strength, high stiffness and toughness, good wear and abrasion resistance, low coefficient of friction and good chemical and electrical resistance [8], [9], [10], [11]. Further modification can be obtained by adding additives like fillers and plasticizers. Fillers achieve a better strength and stiffness, plasticizers cause higher toughness [12], [13]. Modification of the molecular weight has also a distinguishable influence on the polymer properties. The main drawback of polyamide 6 is the high moisture absorption affecting the rheological properties and deteriorating the processability. High moisture absorption causes part defects and results in a bad internal surface quality reducing the stiffness of the end product [14], [15], [16]. The absorbed water in the polymer has two important roles. Firstly, the water molecules act as plasticizers reducing the hydrogen bond interactions between the polyamide chains. Secondly, due to the hygroscopic nature of polyamide 6, the absorbed water influences the condensation reaction equilibrium depending on the initial moisture content [17], [18]. If the moisture content is less than prescribed thermodynamically, the reaction proceeds towards further condensation and an increase in molecular weight, increasing the viscosity exponentially. If the moisture content is higher than determined by the equilibrium, a hydrolysis reaction occurs. This causes a decrease in viscosity as well as degradation of the polymer. To properly process polyamide 6, the pellets are preferably dried to small moisture levels. Possible part defects that may occur in water assisted injection moulding are fingering and double wall. Liu and Lin [19], [20] investigated the problem of fingering in water assisted injection moulded composites. The fingering phenomenon comprises the non-uniformly penetration of water bubbles or steam into the polymer wall outside the designed water channels causing the formation of finger-shaped branches and small cavities in the polymer wall, which lead to a bad internal surface quality and a reduced part stiffness of the end product. Fingering depends on the dimension of the part: thick parts usually exhibit less fingering pattern than thin parts. Three factors might affect the formation of fingering. The first one is shrinkage of the polymer during cooling. During post-filling, the polymer undergoes volumetric shrinkage allowing water to penetrate into the parts. The more the polymer shrinks, the more water will penetrate into the part and causes fingering. However, this statement is in contradiction with the statement that thick parts exhibit less fingering, so research is necessary to investigate which conclusion is relevant for PA6. Shrinkage also depends on the crystallinity of the polymer. PA6 is a crystalline polymer and thus expected to show more shrinkage than an amorphous polymer. Also, water injection moulded parts show more fingering than gas injection moulded parts since water has a higher cooling capacity than gas and has a higher cooling rate of the polymer. This leads to a greater non-uniform temperature distribution in the polymer material which worsens the uniformity of water penetration inside the parts, resulting in fingering [20]. This is equal to the statement that more shrinkage leads to more fingering. The second factor is the viscosity of the polymer melt. During water injection, water follows the path with the least resistance. While the water enters into the mould cavity, it cools the polymer melt and increases the viscosity. It then becomes more difficult for the water to penetrate into the core of the parts. The third factor is the flow resistance in the channel and in the polymer. The higher the flow resistance in the channel and the lower the flow resistance in the polymer, the more water fingering will be induced. The researchers [19] concluded that water fingering worsens when a combination of a higher water pressure, a smaller melt short shot size and a shorter water injection delay time is used. Other parameters which need to be paid attention to are the temperatures of the melt, the mould and the water. Increasing these parameters decreases the cooling rate and the viscosity and results in more fingering because it is easier for water to penetrate in the part. Another possible part defect in water assisted injection moulding is double wall [22]. Double wall is the appearance of a second polymer wall inside the hollow product which deforms the wall significantly. During the water hold time water inclusions develop in the wall. The double wall is caused by the pressure drop after the water hold time. The water in the inclusions becomes steam and expands the internal material surface causing a second wall in the polymer. Two important parameters are responsible for the forming of double wall. The first parameter is the water hold time which should be minimized in order to eliminate double wall. Increasing the water hold time compensates the shrinkage in the wall but enlarge the water inclusions in the wall worsening the double wall. The second parameter is the volume flow rate. Increasing the volume flow rate causes more double wall because more water penetrates in the polymer forming water inclusions which expand into double walls after the water hold time. There is not much known about the phenomenon of double wall. A lot of investigators do not even make a difference between fingering and double wall and characterize it as the same part defect. Hollow parts are mainly characterized by their residual wall thickness that should be minimized in order to obtain a good product quality with minimal defects. A decline in wall thickness is mainly obtained by choosing water assisted injection moulding instead of gas assisted injection moulding [4]. However, Michaeli et al. [22] concluded that in some cases gas assisted injection moulding yields parts with a smaller residual wall thickness. Several process settings can be applied for water assisted injection moulding, each having an influence on the residual wall thickness. Huang and Deng [23] concluded that besides the short shot size there are two other main parameters affecting the residual wall thickness in polypropylene samples. The first parameter is the melt temperature. Increasing the melt temperature leads to reduction of the solidified layer thickness and thus to a decrease in wall thickness. The second parameter is the water injection delay time. When increasing the water injection delay time, the residual wall thickness exhibits an increment because the solidified layer of the melt becomes thicker. Liu and Wu [24] researched the effect of the injection pin on the residual wall thickness distribution. In their research they concluded that a pin consisting of a sintered porous surface can mould large parts with a more uniform residual wall thickness distribution. Application of this high flow rate pin could significantly improve the product quality in terms of smaller wall thickness. The non-uniformity of the residual wall thickness distribution usually occurring in curved product sections, was found to be less in water assisted injection moulded parts than in gas assisted injection moulded parts. Fluids try to follow the path with the least resistance; this is the shortest path. However, water has a higher mass inertion than gas and therefore gives a better concentricity of the residual wall thickness over a cross-section. The uniformity of the residual wall thickness could be improved by adopting different mould temperatures [25]. In order to mould PA6 parts with high product quality using WAIM, one has to apply the optimal settings to reduce the part defects (fingering and double wall) and decline the residual wall thickness. The water injection delay time influences fingering and the residual wall thickness. A shorter water injection delay time causes a smaller residual wall thickness but results in more fingering [19] [23]. The water volume flow rate is an important parameter on fingering, double wall and the residual wall thickness. A higher water volume flow rate gives a smaller residual wall thickness but more fingering and double wall [21]. The melt temperature has an influence on fingering and the residual wall thickness. Increasing the melt temperature leads to a smaller residual wall thickness but to more fingering [19] [20] [23]. Further experiments in this research focuses on the influence of the water injection delay time, the water volume flow rate and the melt temperature on the residual wall thickness and the occurrence of part defects for PA6, using a design of experiments. 2. Experimental procedure 2. 1 Materials To investigate the residual wall thickness and part defects like water inclusions, fingering and double wall of water assisted injection moulded products, two polyamides (PA F223-D and PA F130-E1 from DSM) and a polypropylene (PP 400-GA05 from Ineos) were used. The melting and crystallization temperature of the materials were measured with differential scanning calorimetry (TA Instruments 2920 CE). The viscosity was measured with a capillary rheometer (CEAST Smart Rheo 2000 twin bore). The complex viscosity, storage and loss modulus were measured with a parallel plates rheometer (AERES strain controlled rheometer). The measurements show that the zero-shear viscosity of the polypropylene is higher than those of the polyamides. PA F223-D has a higher zero-shear viscosity than PA F130-E1. This indicates that the molecular weight is higher for polypropylene than for PA F223-D and that PA F130-E1 has the lowest molecular weight. The decrease in viscosity at higher shear rates occurs at lower frequencies for polypropylene in comparison with the polyamides. This indicates a higher molecular weight distribution for polypropylene than polyamide. PA F223-D has a higher molecular weight distribution than PA F130-E1. . 2 Moulded parts The part that is moulded for this experiment is displayed in figure 1. It contains four curved sections varying in radius of curvature. The polymer and water enter the cavity at the bottom side. Water enters when a movable injector is pushed forward and a core at the end of the part is pulled allowing the polymer to be removed out of the inner core of the part. Figure 1: moul ded part containing four curved sections 2. 3 Equipment The injection moulding was executed with an Engel 80-ton injection moulding machine ES 330H/80V/80HL-2F. The dimensions of the machine are 4,80 m x 2,22m x 2m. The single screw diameter is 50 mm and the plastification unit can operate with a maximum injection rate of 152 cm? /s. A volume flow rate controlled water injection unit brings the water into the mould. It can operate with a maximum water volume flow rate of 30 l/min, delivered by a maximum pressure of 200 bar. During the tests a maximum pressure of 160 bar was applied. Three pressure sensors type Priamus are localized in the mould. An IR-camera type FLIR is used to observe the temperature distribution over the produced handles after moulding. The IR-images were used to compare with the internal section to discover the cause of part defects, because water in the product gives lower temperatures on the IR-image. The dried hollow polymer handles are sawed in lengthwise direction with a band saw. The band saw is type Metabo bas 260 swift. After sawing the hollow polymer handles, the internal section were scanned with a scanner type Cannon FG17500. A metric software 8. 01 plus was applied to analyze the surface quality of the polymer. With the aid of the software, the surface area of the wall and the part defects can be determined. . 4 Method DOE procedure The experiment investigated four processing parameters that were selected after a preliminary literature study, assuming these will have a significant influence on the residual wall thickness and part defects. These parameters are the water volume flow rate, water injection delay time and melt temperature. Using the design of experiments approach, each parameter was measured on a low and high setting; making this a 23 experiment. In addition three centre points were measured bringing the total number of experiments to 11 for each material. Response functions were constructed for residual wall thickness, weight and defects. The significance of each parameter was checked by comparing its effect to the deviation of the centre points, neglecting the parameter if its effect is smaller than the centre point deviation. Experimental procedure After steady state was reached, five shots were produced for each setting of water volume flow rate, water injection delay time, melt temperature. For each last shot of an experiment, a print screen of the pressure evolution and an IR-image was taken. During the process the following parameters were measured: * real injection time; * absolute shot volume; * real cushion; * flow number. After moulding, the parts were dried with compressed air to remove water of the inner- and outer surface of the product. Then the injection side of the product was removed and both halves were weighed. The sawed injection product was controlled on the presence of defects giving a first indication of the quality of the product. Afterwards the handle was dried in a furnace for 44 hours on 90  °C. After drying, the handles were weighed again obtaining the weight of the water inclusions in the product wall. Three representative handles were sawed in lengthwise direction and the breams were removed. The handles were scanned to determine the residual wall thickness and the percentages of defects like water inclusions, fingering and double wall. Table 1: Processing variables and settings used for the experiments | A| B| C| Process parameters| Water volume flow rate (l/min)| Water injection delay time (s)| Melt temperature ( °C)| -1| 10| 1,5| 240| 0| 20| 5| 260| 1| 30| 8,5| 280| A| B| C| Experiment| Water volume flow rate (l/min)| Water injection delay time (s)| Melt temperature ( °C)| 1| 20| 5| 260| 2| 10| 1,5| 240| 3| 30| 8,5| 240| 4| 30| 1,5| 240| 5| 10| 8,5| 240| 6| 20| 5| 260| 7| 30| 8,5| 280| 8| 10| 1,5| 280| 9| 10| 8,5| 280| 10| 30| 1,5| 280| 11| 20| 5| 260| Table 2:Experiment schedule for parameter settings The part defects are calculated as follows: The parameters whic h have an influence on the product quality are the water volume flow rate, the water injection delay time and the melt temperature. The tested parameter settings are presented in table 1. A code using -1 for low, 0 for center point and 1 for high setting is applied. The combination of parameter settings (Table 2) are determined using design of experiments. Using the design of experiments method makes is possible to obtain a better understanding of the formation of the defects and the influence of the processing parameters. Eventually a moulding window will e determined in which the defects are minimized. 2 Results and discussion 3. 1 Effect of water volume flow rate Figure 2 shows the effect of the water volume flow rate on the residual wall thickness and the product defects. Following paragraphs discuss these diagrams. Influence on the residual wall thickness and product weight Considering the residual wall thickness, polypropylene has a smaller residual wall thickness than the residual wall thickness of polyamide for all experiments. Polypropylene has a higher molecular mass, causing a wider water flow front. The higher molecular weight distribution (MWD) of PP causes a block velocity profile of the water resulting in a higher velocity on the polymer- water interface. The higher velocity causes a higher shearing and consequently a lower viscosity, resulting in a higher removal. This results both in a smaller residual wall thickness and a lower part weight. The density of polypropylene is lower compared to compared to the density of polyamide, which is also an explanation for the lower weight of product. Regarding the product weight and the residual wall thickness, the two polyamides do not differ much and correspond to each other. This is probably due to the small difference in molecular mass and molecular weight distribution. Figure 2: Effects of water volume flow rate on water inclusions, residual wall thickness, fingering and double wall The residual wall thickness of polyamide seems not to be significantly affected by the water volume flow rate. Polypropylene however, tends to increase in residual wall thickness by application of a lower water volume flow rate. These results correspond to the suggestion that a higher water volume flow rate gives a smaller residual wall thickness, as was found in preliminary literature study. In general a material with a high MWD, like PP, will be more influenced by the flow rate than other materials like polyamides. Influence on water inclusions In general the water inclusions tend to decline when the water flow rate increases. A higher flow rate provides more shearing, giving a lower viscosity. The water can easily remove the internal polymer and does not penetrate into the walls of the polymer. PP completely corresponds to this theory, but the polyamides show an irregularity during high flow rates. this can be explained by the strong variations in injection time and a chance in the water flow profile to a profile called recirculation flow. It pushes the water into the walls causing an increase in water inclusions. The strong variations that occur with PP are probably caused by its high MWD. This property also played an important role regarding the RWT. Influence on fingering Fingering is probably caused by differential shrinkage, making a material with a high degree of crystallization like polypropylene vulnerable to this defect. The defect diminishes at high flow rates where the wall thickness is smaller, making it more difficult for polypropylene to shrink and for fingering to occur. PA F130-E1 exhibits a low variation in fingering in comparison to the other materials. This material is not in its area of shear tinning when working with low flow rates, resulting in a constant viscosity and residual wall thickness. Because of this low variation in RWT and corresponding shrinkage with changing flow rate the effect of the flow rate will be very small. In general materials with a high MWD and MM like PP show more fingering than materials with low MWD and MM like PA F130E1. PA F223D will take a intermediate position taking into account that its centre point is not completely representative. This trend for fingering is visible for all parameters. Influence on double wall The influence of the flow rate shows an overall optimum profile. Low and high flow rates cause more double wall than an intermediate setting. A first theory relates this effect to the pressure inside the part. Low flow rates exert low pressure on the polymer melt and the polymer will not form a rigid layer, making it possible for the water to penetrate into the polymer walls. High flow rates cause high pressures inside the part forcing the water into the walls. The second theory is based on the tendency of water to follow the ath of least resistance and can only explain the effect with low flow rates. The low rate will cause almost no shear stress inside the polymer so it will not reach the shear tinning area. The polymer will not easily be pushed back by the water and the water will penetrate the walls where the resistance is much lower. The effect of the water volume flow rate on double wall is the same as for water inclusions, which is explained by the rel ation between water inclusions and double wall. Water inclusions can grow during the water hold time, collapsing into double walls in the product after pressure removal. The higher double wall defects for PAF130-E1 in comparison to PA F223-D are probably caused by the difference in molecular mass. PA F130-E1 has a lower molecular mass and thus a less viscous melt resulting in a more easily penetrable material. Another noticeable result is that of the investigated materials, PA F223-D shows high part defects but relatively the smallest double wall defects. PP, the material with the highest MM shows more double wall at high flow rates for which no explanation can be found. 3. 2 Effect of the water injection delay time Figure 3 shows the effect of the water injection delay time on the residual wall thickness and the product defects. Following paragraphs discuss these diagrams. Influence on the residual wall thickness and product weight A shorter water injection delay time causes a decline in the residual wall thickness and weight of product for polyamide and polypropylene, as was predicted by literature. When the water injection delay time is short, the polymer is still hot and has a low viscosity, giving the water the opportunity to hollow out the polymer more. A higher water injection delay time causes a higher residual wall thickness and weight of product because the melt viscosity increases. The residual wall thickness is higher for polyamide than for polypropylene, for the same reasons as discussed under the previous paragraph. Influence on water inclusions An increase in delay time is accompanied by an increase in viscosity that makes it more difficult for the water to penetrate into the walls. However, if the viscosity is too high, the water can also be forced into the walls, which explains the slight increase in water inclusions for PP 400-GA05 and PA F130-E1. Figure 3: Effects of the water injection delay time on water inclusions, residual wall thickness, fingering and double wall The strong decrease in water inclusions that occurs with polypropylene is caused by the high heat capacity of this material making it cool down faster than the polyamides and increasing its viscosity faster. PA F130-E1 shows less defects in comparison with PA F223-D, probably caused by its strong temperature related viscosity. During the delay time the viscosity strongly increases, making it more difficult for the water to penetrate. Influence on fingering Increasing the delay time increases the residual wall thickness and should result in a higher chance of fingering. In general this trend can be observed from the diagrams. More fingering occurs within PP 400 GA05 caused by its higher degree of crystallization. Furthermore, this polymer is processed at much higher temperatures than its melting point, leading to further shrinkage and corresponding fingering. The presence of defects during processing with long delay times is caused by a greater residual wall thickness, which leads to more differential shrinkage. Processing with short delay times increases the presence even further, but no plausible statement can explain this. Therefore further research will be necessary. The results for the centre point of PA F223-D are probably not representative, since it is based on a single experiment of three shots. Influence on double wall The water injection delay time shows an overall optimum profile, which can be related to the shift of the rigidity in time, caused by the temperature drop over time. Production with a short delay time will push water inside the low viscosity polymer which makes it easy for the water to penetrate. Increasing the delay time also increases the viscosity and thus the rigidity of the polymer making it harder to penetrate the walls. However, if the viscosity is too high, the water will also be forced into the walls. PA F130-E1 shows more double wall because its viscosity is highly temperature related. Longer delay times will lead to larger temperature drops and a corresponding increase in viscosity. 3. 3 Effect of the melt temperature Figure 4 shows the effect of the melt temperature on the residual wall thickness and the product defects. Following paragraphs discuss these diagrams. Influence on the residual wall thickness and weight of product Considering the residual wall thickness, polypropylene has a lower residual wall thickness and weight of product compared with polyamide for all experiments. This might be due to the higher molecular weight and the higher molecular weight distribution of polypropylene. The higher molecular weight distribution shows a square blocked velocity profile causing a higher velocity on the wall and a higher shearing. This results in a lower viscosity and therefore more removal of the polymer decreasing the residual wall thickness and weight of product. PA F130-E1 and PA F223-D are quite similar and do not differ much in residual wall thickness and weight of product. Increasing the melt temperature leads to a smaller residual wall thickness as was suggested by literature, because a higher melt temperature makes the polymer less viscous so more polymer can be removed by the water. The residual wall thickness of PP decreases from 240 °C to 260 °C but increases to 280 °C. The increase of the melt temperature is difficult to explain and is possible a unreliable point. In fact, also for polypropylene a decreasing tendency is remarkable. Influence on water inclusions There is a trend towards higher water inclusions when a higher melt temperature is applied. This parameter thus shows the opposite effect of the delay time. PA F130-E1 differs from the other materials, probably due to the strong variations in injection time during production. Furthermore this polymer shows a lot of internal heating (viscous heating) when Figure 4: Effects of the melt temperature time on water inclusions, residual wall thickness, fingering and double wall high pressure is exerted, leading to a drop in viscosity. This effect usually appears in combination with low melt temperatures where the polymer has a lot of pressure to bear. So low temperatures will lead to viscous heating and therefore to high intern temperatures. The water can then easily enter the walls of the polymer due to the lower vicosity. Influence on fingering When processing the material at low temperatures, the material will have to shrink less and thus less fingering will occur. This trend is clearly visible for the polyamides, taken into account that the centre point of PA F223-D is not representative. In fact, the effects should show an opposite effect from the delay time, but this is not clearly noticeable. Further research will be necessary to explain these irregularities. Influence on double wall A higher melt temperature causes an overall increase in double wall defects for all three materials. This effect is related to the viscosity of the polymer which varies with temperature. High temperatures will lead to a decrease in viscosity making the polymer more penetrable for water. Another theory is based on the water flow profile inside the polymer. Higher temperatures usually correspond to smaller residual wall thicknesses changing the flow profile into a recirculation flow where water will be pulled into the walls. 3. 4 Comparison results of response functions and graphical tendencies Table 3 gives an overview of the best parameter settings to reduce each part defect and the total defects for each material, comparing the results of a response function (before /) and the graphics in figure 1/2/3 (after /). The response functions were constructed using the design of experiments method. Regarding the defects, only one response function was made which ncludes both fingering and double wall. Ideal settings according to response functions Using the response functions several conclusions can be made for the ideal settings of the parameters. A high water injection delay time creates less water inclusions in both polyamides. In addition, the polyamides show the same optimal settings of all parameters for the residual wall thickness. A low water volume flow rate is beneficial for the reduction of the residual wall thickness, fingering and double wall. Other settings conflict to give minimal part defects in polyamide. In general, PA F223-D has minimal part defects for a low water volume flow rate, a high water injection delay time, and a low melt temperature. PA F130-E1 requires a low water volume flow rate as well, but a low water injection delay time and a high melt temperature. The optimal settings for the minimal percentage water inclusions, fingering and double wall are highly contradictory for polyamide and polypropylene. PP 400-GA05 requires a high water volume flow rate for a smaller residual wall thickness but the other parameters have the same optimal settings for the residual wall thickness as polyamide 6. The water volume flow rate and the water injection delay time do not influence fingering and double wall significantly, so no conclusions can be made for these settings to obtain minimal fingering and double wall. If also the water inclusions and the residual wall thickness are taken into account, PP400-GA05 shows minimal part defects for a high water volume flow rate, a low water injection delay time and a low melt temperature. There can be concluded that for all three materials the best parameter settings for residual wall thickness, fingering and double wall are not the same. Since other effects may probably have an influence on the materials, further investigation will be necessary. Ideal settings according to previous diagrams The following best parameter setting are based on the previous graphics. Regarding the residual wall thickness, the best parameters for PP and PA F130D are an intermediate water flow rate, a low water injection delay time and a high melt temperature. For PA 223D, a low water flow rate is required and the other parameters are the same. For fingering, a high water volume rate, an intermediate water injection delay time and melt temperature give the best results for PP. PA F223D gives the best results for a low water volume flow rate and water injection delay time and an intermediate melt temperature. PA F130E1 in contrary, the best parameters setting are a high water volume flow rate and a low water injection delay time and melt temperature. The parameters setting are different for the three materials and show a lot of irregularities. Considering double wall, the best parameter settings are a low water flow rate, water injection delay time and melt temperature. For both polyamides the best parameter setting are equal, an intermediate water flow rate, delay time and melt temperature. For the water inclusions, the best results were observed for a high water volume rate and an intermediate water delay time and melt temperature. PA F223D shows less water inclusions when an intermediate flow rate, a low delay time and in intermediate melt temperature is used. PA F130E1 in contrary gives the best results for a intermediate flow rate and delay time and a low melt temperature. Ideal settings for PA F223-D| | Optimal setting| Response factor| Target figure| Water volume flow rate| Water injection delay time| Melt temperature| Water inclusions| Low| High/High| High/High| Low/Low| Residual wall thickness| Low| Low/High| Low/Low| High/High| Fingering and double wall| Low| Low/High| High/Low| Low/Low| Total| | Low/High| High/Low| Low/Low| Table 3: Overview of ideal settings for the materials for each part defect Ideal settings for PA F130-E1| | Optimal setting| Response factor| Target figure| Water volume flow rate| Water injection delay time| Melt temperature| Water inclusions| Low| Low/High| High/High| High/Low| Residual wall thickness| Low| Low/High| Low/Low| High/High| Fingering and double wall| Low| Low/High| Low/Low| Low/Low| Total| | Low/High| Low/Low| High/Low| Ideal settings for PP 400-GA05| | Optimal setting| Response factor| Target figure| Water volume flow rate| Water injection delay time| Melt temperature| Water inclusions| Low| High/High| Low/High| High/Low| Residual wall thickness| Low| High/High| Low/Low| Low/High| Fingering and double wall| Low| -/High| -/Low| Low/Low| Total| | High/High| Low/Low| Low/Low| Once again, here can be concluded that the best parameter settings for residual wall thickness, fingering and double wall are not the same. Since other effects may probably have an influence on the materials, further investigation is necessary. . 5 Detection of part defects 3. 5. 1 IR-images IR images were tested if they can be used to obtain an indication of the quality of the product. This could be an important tool during production as it can deliver information about the internal structure without damaging the part. It was found that a uniform heat distribution in the part usually corresponds to few defects in the walls. The polymer equally shrinks and cools, causing no stress in the walls which could lead to defects. Figure 5 shows the section and an infrared image of a hollow part from PP 400-GA05 without defects. The corresponding IR picture shows no significant Figure 6: Scan and IR image of PP 400-GA05 product with defects irregularities in the heat distribution. The areas where the part has a higher or lower Figure 5: Scan and IR image of PP 400-GA05 product without defects temperature than the bulk are caused by respectively larger and smaller residual wall thicknesses. Small defects like fingering can occur where colder areas are completely surrounded by larger hot material making it isolated cold regions. If the cold area is gradually spread throughout the part and is therefore no isolated region, fingering is often not present. The isolated colder regions are most likely filled with water which provides more efficient cooling than the rest of the part, resulting in uneven shrinkage and corresponding fingering cavities. Large defects like double wall are often present where a relative large cold area spreads throughout the complete section and is surrounded by hot material. These defects are also likely to occur in regions with higher temperature than the rest of the part. However, the latter can also correspond with a larger residual wall thickness and is thus not conclusive for the presence of double wall defects. Figure 6 shows a part with double wall defects and the corresponding infrared image where the isolated hot (white) and cold areas (yellow) are indicated. The texture of the internal surface can not be judged with an infrared image. PA F130-E1 tends to show a rougher surface in regions with a much lower temperature, but not all parts share this conclusion. Within the other materials a rough surface does not corresponds with lower temperature areas thus the infrared does not delivers clear information about this property. 3. 5. 2 Pressure and injection rate profiles The pressure profiles and the injection rate during the injection of the melt can be used to declare some of the defects shown in the different polymer samples. The injection rate was set on 45 mm/sec and the maximum feasible pressure of the WAIM equipment was 160 bar. For the polyamides, this maximum pressure was exceeded in some cases during filling of the mould. This caused a temporary reduction of the injection rate and possibly some of the part defects of polyamide. Polypropylene does not show this trend because it has a lower flow number than PA. Figure 7 shows an example of this phenomenon in PA F223-D. In general it can be concluded that when the temperature increases the measured pressure and the flow number decrease. This reduction in pressure can be explained by the fact that the viscosity of the material will be lower at a higher temperature, so the polymer can flow easier and less pressure has to be applied. However it is not guaranteed that an irregularity in the injection profile of the polymer results in problems during water injection. This experiment was to limited to make clear conclusions about this effect. Figure 7: Pressure and injection rate profile for PA F223-D (green=pressure, blue=injection rate) 4. Conclusions 4. 1 General This report has studied the influence of melt and processing parameters on the residual wall thickness and the occurrence of part defects of PA6 parts produced with water assisted injection moulding. The study examined two polyamides (PA F223D, PA F130E1) and polypropylene 400 GA05 that was used as a reference material. According to a design of experiments, the water volume flow rate, the water injection delay time and the melt temperature are the investigated processing parameters. The water volume flow rate settings are 10, 20 and 30 l/min, the injection delay time settings are 1,5s, 5s and 8,5s and the melt temperature settings are 240 °C, 260 °C and 280 °C. It was found that the residual wall thickness showed a large distribution for all three materials, even for shots from a single experiment and therefore an extra analysis was made on the weight of the products. The product weight can give an indication for the wall thickness as a low weight usually corresponds to small thicknesses. Based on this theory it was found that the water volume flow rate has little influence on the weight and wall thickness. A larger delay time increases the viscosity of the polymer and leads to higher weight and a larger wall thickness. The melt temperature has the opposite effect, decreasing the viscosity at higher temperatures and therefore decreasing the wall thickness and weight. PP has a lower weight comparing with the two polyamides. This can be explained that PP has a higher molecular weight and molecular weight distribution. A higher molecular weight gives a parabolic velocity profile and a higher molecular weight distribution shows a square blocked velocity profile causing a higher velocity located on the wall. The higher velocity causes a higher shearing giving a lower viscosity. This results in more removal of the polymer and the residual wall thickness decreases. In fact, a lower residual wall thickness exhibits a lower weight of product. The water inclusions for the three materials tend to decline when the water flow rate increases although PP shows some irregularities. The delay time has an overall optimum profile. Intermediate settings exhibit less inclusions in comparison to other settings. The melt temperature has the opposite effect of the delay time so more water inclusions occur when higher temperatures are applied. Other defects like fingering and double wall are also influenced by the investigated processing parameters. The water flow rate shows an optimum profile for double wall, so an intermediate setting will yield the best results. For fingering the three materials all act in a different way to the flow rate. An optimum profile also occurs for double wall when the delay time is investigated. In general, fingering increases when the delay time increases, but the materials show a lot of irregularities. An increase in melt temperature causes an overall increase of fingering and double wall. Comparing PA6 and PP, PP shows an overall smaller residual wall thickness than PA6, but on average the same percentage of part defects. This means that a reduction in residual wall thickness does not influence the product quality for part defects positively. When comparing the different polyamides 6, the investigated low viscosity polyamide 6 PA F223-D show less water inclusions and fingering but more double wall than the medium viscosity polyamide 6 PA F130-D. The best choice for a polyamide depends on the relative importance of the negative influence of each part defect on the product quality. This is discussed in detail in the next paragraph. To get an indication of the quality of the product IR images can be used. It was found that a uniform heat distribution corresponds to few defects and irregularities point to the presence of defects like fingering or double wall. The internal texture of the surface can not be examined with IR. This study concludes that the investigated parameters have a significant influence on the presence of part defects and on the residual wall thickness. However, there are still irregularities in the results of the fingering phenomenon that can not be explained yet with the current knowledge of the water assisted injection moulding process. Therefore further analysis will be necessary to examine this part defect. Further research can be done by expanding the design of experiments and by producing more shots for each experiment. Producing more than three shots of each experiment should give more corresponding results, because the standard deviation of three shots is usually high. 4. 2 Selection of the material for WAIM Three materials were used in this experiment namely PA F223D, PA F130E1 and PP as reference. The residual wall thickness, weight of product, water inclusions and part defects like fingering and double wall must be compared with each other to choose the best material. This by varying the selected parameters like water volume rate, water injection delay time and melt temperature. PP has in general the lowest weight of product and residual wall thickness for the three investigated parameters: water volume rate, water injection delay time and melt temperature. As a result of a smaller residual wall thickness, PP has the lowest weight of product. Considering the water inclusions a decreasing tendency is obtained when a higher water volume flow rate and water injection delay time is used. There is a trend toward higher water inclusions when higher melt temperature are applied. Regarding to fingering and double wall, an opposite effect is noticeable. Increasing the water volume rate and water injection delay time gives a decreasing fingering tendency but more double wall effects. For a higher melt temperature, an increase of fingering and double wall is observed. PA F130E1 and PA F223D are quiet similar to each other in weight of product and residual wall thickness. For the weight of product PA F130E1 and PA F223D correspond to each other and for the residual wall thickness they do not differ much for all three investigated parameters. Regarding the water inclusions, both polyamides decline when water the flow rate increases. The highest flow rate provides more shearing and therefore a lower viscosity resulting in more water inclusions. Increasing the water injection delay time leads to lower water inclusions because the viscosity increases which makes it more difficult for water to penetrate into the walls. When higher melt temperatures are applied, results in a trend towards higher water inclusions because of the lower viscosity. In general, PA F223D shows more water inclusions comparing with PA F130E1 for the water volume flow rate, water injection delay time and the melt temperature. Considering fingering and double wall, different results were observed for the two polyamides. In general it is remarkable that PA F223D shows more fingering then PA F130E1 while PA F130E1 gives more double wall then PA F223D and this for the three investigated parameters. For the water volume flow rate, the polyamides act in different ways towards fingering. Although, for double wall the flow rate shows an optimum for the polyamides so an intermediate setting gives the best results. When the water injection delay time increases there is also an increasing tendency of fingering observed but the materials show a lot of irregularities. Increasing the melt temperature gives an increase of fingering and double wall as well. In this investigation can be concluded that both polyamides have their advantages and disadvantages. A selection of the best polyamide depends on which effect is undesirable. The residual wall thickness are the same for the two polyamides. PA F223D shows more water inclusions and fingering for all investigated parameters. However, PA F223D shows less double wall and therefore the best results. PA F130E1 in contrary gives more double wall effects but less fingering and water inclusions. In fact, PA F223D is preferred when double wall must be reduced and PA F130E1 is preferred when fingering and water inclusions must be reduced. Further investigation is necessary on another type of polyamide which improves the product quality using WAIM. References [1] Knights, M. (2002). Water Injection Molding Makes Hollow Parts Faster, Lighter. Plastics Technology, pp. 42-47, 62-63. 2] Liu, S. -J. , & Lin, M. -J. , & Wu, Y. -C. (2007). An experimental study of the water-assisted injection molding of glass fiber filled poly- butylene-terephthalate (PBT) composites. Composites Science and Technology, pp. 1415- 1424. [3] Liu, S. -J. , & Hsieh, M. -H. (2007). Residual Wall Thickness Distribution at the Transition and Curve Sections of Wat er-assisted Injection Molded Tubes. International Polymer Processing, pp. 82-89. [4] Liu, S. -J. , & Chen, Y. -S. (2003). Water- Assisted Injection Molding of Thermoplastic Materials: Effects of Processing Parameters. Polymer Engineering and Science, pp. 806- 1817. [5] Liu, S. -J. , & Wu, Y. -C. (2007). Dynamic visualization of cavity-filling process in fluid- assisted injection molding-gas versus water. Polymer Testing, pp. 232-242. [6] Liu, S. -J. (2009). Water Assisted Injection Molding: A Review. International Polymer Processing, pp. 315-325. [7] Liu, S. -J. , & Chen, Y. -S. (2004). The manufacturing of thermoplastic composite parts by water-assisted injection-molding technology. Composites Part A: Applied Science and Manufacturing, pp. 171-180. [8] Chavarria, F. , & Paul, D. R. (2004). Comparison of nanocomposites based on nylon 6 and nylon 66. Polymer 45, pp. 8501-8515. [9] Li, D. , Liu, Q. , Yu, L. , Li, X. , & Zhang, Z. (2009). Correlation between interfacial interaction and mechanical properties of PA-6 doped with surface-capped nano-silica. Applied Surface Science 255 pp. 7871-7877. [10] Sannen, S. (2010). Polyamides. Intern presentation of Catholic University Leuven. [11] Su, K. , Lin, J. , & Lin, C. (2007). Influence of Reprocessing on the Mechanical Properties and Structure of Polyamide 6. Journal of Materials Processing Technology, pp. 532-538 [12] Fornes, T. , Yoon, P. , Keskkula, H. , & Paul, D. (2001). Nylon 6 nanocomposites: the effect of atrix molecular weight. Polymer, pp. 9923- 940. [13] Uribe-Arocha, P. , Mehler, C. , Puskas, J. E. & Alstadt, V. (2003). Effect of Sample Thickness on the Mechanical Properties of Injection Molded Polyamide-6 and Polyamide- 6 Clay Nanocomposites. Polymer 44, pp. 2441-2446. [14] Steeman, P. & Nijenhuis, A. (2010). The Effect of Random Branching on the Balance between Flow and M echanical Properties of Polyamide-6. Polymer 51, pp. 2700-2707. [15] Incarnato, L. , Scarfato, P. , Scatteia, L. & Acierno, D. (2004). Rheological Behavior of New Melt Compounded Copolyamide Nanocomposites. Polymer 45, pp. 3487-3496. 16] Meng, Y. Z. & Tjong, S. C. (1996). Rheology and morphology of compatibilized polyamide 6 blends containing liquid crystalline copolyesters. Polymer 39 No. 1, pp. 99-107. [17] Acierna, S. , & Van Puyvelde, P. (2005). Rheological Behavior of Polyamide 11 with Varying initial Moisture Content. Journal of Applied Polymer Science, pp. 666-670. [18] Chow, W. , Bakar, A. , & Ishak, Z. (2005). Water Absorption and Hygrothermal Aging Study on Organomontmirillonite Reinforced Polyamide6/polypropylene Nanocomposites. Journal of Applied Polymer Science, pp. 780- 790. [19] Liu, S. -J. , & Lin, S. -P. (2005,april). Study of Fingering’ in Water-assisted Injection Molded Composites. Polymer Rheology and Processing lab. [20] Lin, K. -Y. , & Liu, S. -J. (2009) . The Influence of Processing Parameters on Fingering Formation in Fluid Assisted Injection Molded Disks. Polymer Engineering and Science, pp. 2257-2263. [21] Walter Michaeli, E. H. (2010). ‘Perspective Kunststofftechnik berichte 2010’ IKV- Kolloquium Aachen 2010 pp. 19-21 (Analyse der Fehlerentstehungs-mechanismen). [22] Michaeli, W. , & Lettowsky, C. , & Groenlund, O. (2005). Understanding Residual Wall Thickness. Kunststoffeplast Europe 95, pp. 87-91. [23] Huang, H-X. , & Zhou, R-H. 2009). Preliminary investigation on morphology in water-assisted injection molded polymer blends. Polymer testing, pp. 235-244 [24] Liu, S. -J. , & Wu, Y. -C. (2006). A Novel High Flow Rate Pin for Water-assisted Injection Molding of Plastic Parts with a More Uniform Residual Wall Thickness Distribution. International Polymer Processing, pp. 436-439. [25] Lin, K. -Y. , Chang, F. -A. & Liu, S. -J. (2009). Using Differential Mold Temperatures to improve the Residual Wall Thickness Uniformit y Around Curved Sections of Fluid Assisted Injection Molded Tubes. International Communications in Heat and Mass Transfer 36, pp. 491-497.