Characterization of Thermoplastics Additive Manufacturing by ...
CHARACTERIZATION OF NEW THERMOPLASTICS FOR ADDITIVE MANUFACTURING BY SELECTIVE LASER SINTERING D. Rietzel, F. Kuehnlein, D. Drummer Institute of Polymer Technology, University of Erlangen-Nuremberg Am Weichselgarten 9, 91058 Erlangen / Germany
Abstract Additive manufacturing technologies like selective laser sintering offer the potential to create complex, individualized parts in series. Predominantly polyamide12 can be used for direct part generation. This leads to restrictions for many industrial and medical applications. Thus research on further polymers plays a major role for applying additive manufacturing to serial production of individual products. Currently, great efforts are made to process new technical thermoplastics by selective laser sintering. The suitability and processing behaviour by means of melting, (isothermal) crystallization, resulting morphology and part-properties of thermoplastics like polyoxymethylene, polyethylene and polypropylene is presented and compared with commercially available powders in this paper.
Figure 1: Complex SLS part generated with POM powder
Introduction
Motivation
Formerly, techniques of additive manufacturing used to serve merely for prototype construction for special products, Figure 1. This has clearly changed. They are now used for a vast number of new fields of application. Especially for technical parts, products made by additive manufacturing have grown in importance, now being much more than mere objects for demonstration [1]. In particular, powder-based plastics processing techniques, such as selective laser sintering (SLS) and selective mask sintering (SMS) generate good mechanical component properties [2]. In contrast to this benefit, the range of materials suitable for processing by this technique is very much restricted to polyamide 12 (e.g. PA2200, EOS GmbH), polyamide 11 (e.g. Primepart DC, EOS GmbH), polystyrene (e.g. Primecast PS), and, in very few cases, thermoplastic elastomers [3]. Currently, there is a growing number and variety of applications which require components that can withstand high mechanical load, have a high density and a variety of other characteristics. A material such as PA12 is therefore no longer sufficient to meet the high demands posed to components suited for serial production. This is why there are new semi-crystalline thermoplastics, e.g. polypropylene (PP) or polyetheretherketone (PEEK) on the verge of entering the market.
The investigations presented in this paper are mainly concerned with reducing the existing restrictions in terms of materials, eventually aimed at widening the range of applications for the laser sintering of plastics. Regarding the options of usage for other types of thermoplastic, technologies such as SLS may be successful too, if material properties are to be different from those of PA12. This processing technique demands that the material fulfils a series of requirements. It is therefore essential that the engineer knows and understands the significant mechanisms of interaction as well as subsequent material characterization, in order to adequately process the material. Especially, melting and crystallization behaviors must be known because they are of outstanding importance for the laser sintering process. Much of the knowledge available on the process at present is based on the processing of PA12 as a semi-crystalline thermoplastic. Thermoanalytical methods of analysis for various semi-crystalline thermoplastics are suitable to show which of the major material properties have an effect on the processability by SLS. In this study, PA12, POM, PE-HD, PP and PEEK were submitted to testing with regard of their processing behaviors, and components were tested as to their properties.
ANTEC 2010 / 2247
SLS Process and Demands on Thermoplastics During manufacturing, the plastics powder is applied, layer by layer. The powder that is located at the interface with the component is fused selectively by a CO2 laser, and thus connected insolvably to the melt layer below. The surrounding powder that was not molten supports the generated melt. Only if all layers of the component have been produced is the powder bed with its components inside cooled down.
Requirements to the Plastic Powder The properties of laser sintered parts, e.g. density, surface topography, accuracy of detail and dimensional stability are determined by the process parameters interacting with the material. In investigations it was found, for instance, that the geometry of particles is a decisive factor that substantially determines component coarseness [4,5]. Melt behavior (e.g. viscosity and surface tension) is just as significant as powder density and flowability of laser sintering powders. The size distribution and geometry of particles are of major importance for the sintered parts’ porosities [6,7]. High powder density thus leads to higher densities, dimensional accuracy and strength in the sintered parts, yet it may deteriorate flowabilities. Commercially available laser sintering powders with good flowabilities consist of spherulite particles with a narrow size distribution of d = 60 µm, and with a low share of fine particles of d = 10 µm. [8,9] Considering the long building times, the plastic material must particularly resist thermal degradation, since it is kept at a temperature close to the crystalline melting point for several hours. At the end of the production process, the thermally aged powder that had not molten is separated from the laser sintered component and will be recycled. In an ideal case, the powder should not be agglomerated, so that the part can be separated from the remaining powder merely by making use of the force of gravity. Due to ageing during building, the powder’s material properties usually change, which is why the remaining material must be mixed with roughly 30 wt-% of virgin powder, in most cases.
Thermal Boundary Conditions Apart from material application, temperature control during building is of major significance for the part’s profile of properties. For part density, a closed melt film is crucial, making it necessary for the powder material to have low melt viscosity [9]. Due to their wide range of softening and the resulting possibility of achieving low component densities, as well as dimensional stability,
2248 / ANTEC 2010
amorphous thermoplastics are employed e.g. as lost cores in precision casting [3]. Thus for direct part manufacturing, only the semi-crystalline polymers are relevant at present and PA12 (EOS PA2200) is the only commercially available SLS powder. Semi-crystalline plastics, however, are heated to the point above glass transition temperature, close to the crystalline melting point, with the laser merely fusing the crystalline shares. Due to the high chain mobility, decrease in viscosity is much steeper after exceeding this narrow range of crystalline melting. [10,11] In an ideal case, the building process would lead to the model state of quasi-isotherm laser sintering, where melt and solid powder exist side by side. The building process is carried out in a sort of two-phase mixing state. The laser merely introduces the energy that is necessary for the material to exceed the point of phase transition. Temperature increase in the surrounding powder bed should be as small as possible, here. This leads to another requirement of this process, in terms of material properties: The plastic material’s crystallization temperature should be clearly below the crystalline melting point [9]. Dynamic Differential Scanning Calorimetry (DSC), among others, can describe phase transitions, and thus the differences in temperatures between crystalline melting and crystallization. On this basis, the possible range of processing temperatures during SLS production may be defined. In case this range is exceeded, powders will get out of control and melt, whereas, if temperature is below this range, the plastic melt produced up to then will start to crystallize, causing shrinkage or curling. Following this model, powder and part are slowly cooled down only when the building process is completed. A low temperature gradient is applied, which provides for components with little residual stress and high dimensional stability.
Materials and Methods Commercially available laser sintering powders made from polyamide 12 (PA2200) and polyetherketone (PEEK HP3) were tested as to their processing properties and compared to PE-HD and PP sintering powders. Grinding tests were also conducted on PA12, PP and POM pellets, in order to look into the potential included in this method of preparation. Additionally, tensile bars made of PP laser sintering powder by 3D-Systems were used as a reference for the parts made from the polymer type by the LKT. In Table 1 an overview of the material properties for the investigated properties is shown [12]. The POM powder was mixed with 0.2 wt.-% of Aerosil®, in order to step up flowability. In IR measurements conducted on PE-HD before, the material was found to have a high transmission coefficient when submitted to the wavelength of the used CO2 laser (λ = 10.6 µm). By adding 0.4 wt-% of Carbon Black® the penetration depth of
the laser could be limited to approximately 100µm. As a result, the parts were black, had a high degree of absorption and better flowability. Table 1: Properties of the investigated polymers [12]
PA12
E-modulus [N/mm2] 1400
σYield [N/mm2] 50
εBreak [%] ~ 200
Tmelt [°C] 170 - 180
PP
1400-1800
25 - 40
> 50
160 - 165
PE-HD
600 - 1400
18 - 30
> 50
125 - 135
POM
2600-3200
60 - 75
20- >50 (H) 15-40 (C)
175 (H) 164-172 (C)
PEEK
3700
100
> 50
335
the determined beginning of crystallization. The exothermal heatflows resulting during crystallization were recorded. The PA 12 material was submitted to previous drying (below N2) at 120 °C for 15 minutes in DSC, since otherwise the values are very much scattered.
SLS Parameter Studies Having determined the admitted building temperatures in previous thermoanalytical tests, reference specimens could be produced from POM, PE-HD, PP and PA 12. For this purpose, tensile bars were made following different irradiation strategies. The parameters were varied according to the following principle, until the layers were bonded completely:
EL = Generating spherulite particles directly from polymerization is not possible with all types of plastics. As an alternative, powders can be made from pellets. Cryoscopic grinding is a well-proven method to make powder particles with a size less than 100µm. The pellets are cooled down to Tmill=-50°C in a cooling section and fed into a counterrotating pinned disc mill (impact crusher principle). After milling, the powder is classified and screened down to the desired particle diameter. The fine grain with diameters below 80 µm was sieved and investigated.
Thermoanalytical Investigations To investigate the melting and crystallization behavior of the thermoplastic material employed, the SLS process was simulated by DSC measurements. According to DIN 53765, 10 or 20 K/min is the standard heating or cooling rate, respectively, for thermoplastics. Laser sintering is however a slow process of part generation, which makes standard measurement unsuited to sufficiently describe the real process. The molten contour is kept to high temperature over a long period of time. The temperature is close to the crystalline melting point, and new preheated layers are applied on the previously applied ones. For these processes, models are available [9], which were added the time-related process of crystallization in the tests performed. For this purpose, several heating and cooling rates (10/5/1 K/min) were set and the resulting heatflows measured. For ideal quasi-isotherm laser sintering, scientists started from the assumption that the melt does not crystallize for a long period of time at the point just below the crystalline melting point. With crystallization related to time and temperature, further DSC tests were carried out alongside the laser sintering process. The specimens were heated to T = Tm+20 K at 10 K/min, cooled down to measurement temperature at a cooling rate of 40 K/min and kept in the isothermal state at different temperatures above
PL vS ⋅ h S
EL = energy density PL = laser power hS = hatch distance vs = scanning speed
Component Testing Layer bonds and resulting morphologies of the tensile bars’ were determined by examining images of microtome cuts taken by a transmitted light microscope. Tensile tests were conducted on the laser sintered parts according to DIN EN ISO 527-1,-2, and the fracture surfaces examined under a scanning electron microscope (SEM).
Results and Discussion Powder Preparation In the grinding tests, only POM revealed large quantities of powders with particle sizes below 80 µm, Figure 2. Thus an economic powder production can be realized if this material enters the market. This process generates irregular particle geometries and a broad range of different gradings with a high share of fine sizes. In Table 2 you can see the various semi-crystalline thermoplastic pellets after grinding in comparison to commercially available SLS powder. Those powders were then sieved below 80µm to fit the SLS process. For PP cryogenically milled powder could not be investigated and compared to precipitated PP powder in further steps.
ANTEC 2010 / 2249
Figure 2: SEM Image of PA 2200 powder (left) and cryoscopically grinded POM powder (right) Table 2: Distribution of particle sizes in cryoscopically grinded thermoplastics Size [µm] < 250 < 125
PA12 [%] 88.5 49.5
< 63
13.5
PP [%] 28.5 4.5 -
POM [%] 97.5 88.0 53.5
Thermoanalytical Investigations The findings presented in Figure 3 show that commercially available thermoplastics start to crystallize even before the standard crystallization temperature is reached. The range of processing parameters determined in the standard process is thus unsuited to serve as the sole processing criterion. Besides the determination of the building temperature via the difference between crystallization and melting peak also isothermal curves were measured for different temperatures between the phase changes. Thus the time-stability of the two-phase area could be analyzed.
2250 / ANTEC 2010
Figure 3: Heating, cooling and isothermal soaking of commercially available SLS powders by means of DSC In the shown DSCs (Figure 3 and Figure 4) the time until the crystallization peak was reached is illustrated. Above the highest shown temperatures no defined crystallization peak could be observed. This was the minimum temperature used for the generation of samples in the SLS process. In general PA 12 (PA 2200), after pre-drying, is the material that features the latest crystallization point among all examined thermoplastics, which means that above Tiso = 172 °C one may assume there is a two-phase mixing area over the entire building process.
layer bonding, while keeping energy density as low as possible. Figure 5 presents embedded microtome cuts taken from tensile test bars in a parameter study looking into POM at different energy densities, and it shows PA 2200 as a reference. In the top images, the laser fails to penetrate the applied powder layer completely. At the rear of the individual layers, there are powder particles visible that have not fused. In case of incomplete layer bonding, as presented in Figure 5 (bottom), a homogenous component with a high degree of crystallinity (KPOM = 85 %) is obtained.
Figure 5: Transmitted light images with polarized light taken of microtome cuts from laser sintered tensile bars t. l.: POM (EL = 35.0 mJ), t. r.: POM (EL = 40.0 mJ) b. l.: POM (EL = 60.0 mJ), b. r.: PA 2200 (EL = 60.0 mJ)
Figure 4: Heating, cooling and isothermal soaking of new investigated powders by means of DSC Accordingly, to set the building temperature right, it is not sufficient to know the range of possible temperatures determined by DSC. For instance, if a low building temperature is set, crystallization may start during layer generation, which typically reveals in the component contour bending upwards. Especially, if new thermoplastics like PE (Figure 4) with a narrow range of possible temperatures between melting and crystallization are processed, this can be an unfavorable aspect, which is why high building temperatures are usually recommended here. SLS Parameter Studies The Laser sintering tests were mainly focused on determining the building parameters that permit for complete
Figure 6: Transmitted light images in polarized light taken of microtome cuts from laser-sintered tensile bars left: PE-HD (EL = 80.0 mJ) right: PP (rectangular built, EL = unknown) The transmitted light images in Figure 6 show the crosssection of a tensile bar (left) from PE-HD, that was sintered in horizontal position, as well a vertically-built tensile bar from PP (right), made by 3D-Systems.
Component Testing
ANTEC 2010 / 2251
The test specimens produced were tested as to their mechanical properties in a tensile test and compared to each other, Figure 7. Accordingly, with horizontally built tensile test bars, values of E-modulus as well as tensile stress at break are comparable to the expected material values from literature [12]. The tested tensile bars by 3D-Systems had been built vertically, and cannot be compared to the horizontally built bars. For all specimens the values of elongation at break achievable up to now are clearly below the expected values [12]. Apart from the structure resulting from long sintering times, and from the high degree of crystallinity, the poor elongation at break is due to the coarse surface and the residual porosity of the parts. Fractures start mainly in the notches on the coarse surface. POM shows a relatively homogeneous and plane surface, making it less prone to fail from notches than other measured components. #
Figure 7: mechanical properties of laser sintered tensile bars Figure 8 shows the fracture mirrors for a sintered vertically-built 3D-Systems PP bar as well as for a sintered POM tensile bar made by the LKT. The PP part shows a large ductile fracture onset starting at the lower boundary layer (see figure, left). In the POM specimen, the fracture starts in a large pore and soon ends in a brittle residual fracture.
Figure 8: SEM image of fracture onsets in laser-sintered tensile bars from PP (left) and POM (right), respectively
Summary and Outlook
2252 / ANTEC 2010
It could be shown that is possible to generate components of thermoplastic materials other than PA 12 by laser sintering under consideration of optimized processing parameters. This broader spectrum of usable materials opens up a wide range of applications for the process. Cryogenically milling is a good technology to generate fine powders for SLS but it is not possible for every thermoplastic. Knowledge of the process has also been improved for this technique. Considering process behavior, PA 2200 certainly is an extremely robust material. However, thanks to improvements in machine engineering, the potential of other laser sintering powders for commercialization has been stepped up too. Materials such as PP and PEEK are currently entering the market. Apart from the investigations presented in this study, concerned with materials’ fusing and crystallization behaviors, other techniques of thermoanalytical measurement are growing in importance. They are required to qualify these new materials comprehensively. These techniques comprise aspects of absorption behavior of the powder fill, film formation based on melt viscosity, but also factors of scaling because of the specific volumes changing at phase transition. Especially the existence of the two-phase SLS model has to be studied to know more about the beginning of melting and crystallization. Moreover, it was possible to generate, from the mentioned powders, components with typical basis material properties – with regard to stiffness and strength, whereas fracture behaviors were very brittle with all the materials under investigation. Finally there is high potential for further improvements of process-specific measurements techniques and the processing of new polymers.
References 1.
Rietzel, D., Kühnlein, F., Feulner, R., Hülder, G., von Wilmowsky, C., Fruth, C., Nkenke, E., Schmachtenberg, E., Proceedings SPE – European Conference on Medical Polymers 2008, Belfast, UK, (2008) 2. Kühnlein, F., Rietzel, D., Wendel, B., Feulner, Hülder, G., Schmachtenberg, E. Plastverarbeiter, 08 (2008) 3. Pfister, A., Dissertation, Albert-Ludwigs-Universität Freiburg i.Br. (2005) 4. Gebhardt, A.: Rapid Prototyping – Werkzeuge für die schnelle Produktentstehung. Carl Hanser Verlag, München, 2000 5. Podszun, W., Harrison, D., Alscher, G., Patent DE 19820725.5 (1999) 6. McGeary, R., Journal of the American Ceramic Society, 44, pp. 513-522, (1961) 7. Seul, T., Dissertation, RWTH Aachen (2004) 8. Nelson, C., Dissertation, University of Texas at Austin, USA (1993) 9. Alscher, G., Dissertation, GH-Essen (2000) 10. Nöken, S., Dissertation, RWTH Aachen (1997)
11. Eyerer, P., Kunststoffe, 83 (1993) 12. Ehrenstein, G.W.: Polymeric Material. Carl Hanser Verlag, München, 2000
Key Words Selective Laser Sintering, Additive Manufacturing, Isothermal Crystallization, POM, PP, PA2200, PE
ANTEC 2010 / 2253
Abstract Additive manufacturing technologies like selective laser sintering offer the potential to create complex, individualized parts in series. Predominantly polyamide12 can be used for direct part generation. This leads to restrictions for many industrial and medical applications. Thus research on further polymers plays a major role for applying additive manufacturing to serial production of individual products. Currently, great efforts are made to process new technical thermoplastics by selective laser sintering. The suitability and processing behaviour by means of melting, (isothermal) crystallization, resulting morphology and part-properties of thermoplastics like polyoxymethylene, polyethylene and polypropylene is presented and compared with commercially available powders in this paper.
Figure 1: Complex SLS part generated with POM powder
Introduction
Motivation
Formerly, techniques of additive manufacturing used to serve merely for prototype construction for special products, Figure 1. This has clearly changed. They are now used for a vast number of new fields of application. Especially for technical parts, products made by additive manufacturing have grown in importance, now being much more than mere objects for demonstration [1]. In particular, powder-based plastics processing techniques, such as selective laser sintering (SLS) and selective mask sintering (SMS) generate good mechanical component properties [2]. In contrast to this benefit, the range of materials suitable for processing by this technique is very much restricted to polyamide 12 (e.g. PA2200, EOS GmbH), polyamide 11 (e.g. Primepart DC, EOS GmbH), polystyrene (e.g. Primecast PS), and, in very few cases, thermoplastic elastomers [3]. Currently, there is a growing number and variety of applications which require components that can withstand high mechanical load, have a high density and a variety of other characteristics. A material such as PA12 is therefore no longer sufficient to meet the high demands posed to components suited for serial production. This is why there are new semi-crystalline thermoplastics, e.g. polypropylene (PP) or polyetheretherketone (PEEK) on the verge of entering the market.
The investigations presented in this paper are mainly concerned with reducing the existing restrictions in terms of materials, eventually aimed at widening the range of applications for the laser sintering of plastics. Regarding the options of usage for other types of thermoplastic, technologies such as SLS may be successful too, if material properties are to be different from those of PA12. This processing technique demands that the material fulfils a series of requirements. It is therefore essential that the engineer knows and understands the significant mechanisms of interaction as well as subsequent material characterization, in order to adequately process the material. Especially, melting and crystallization behaviors must be known because they are of outstanding importance for the laser sintering process. Much of the knowledge available on the process at present is based on the processing of PA12 as a semi-crystalline thermoplastic. Thermoanalytical methods of analysis for various semi-crystalline thermoplastics are suitable to show which of the major material properties have an effect on the processability by SLS. In this study, PA12, POM, PE-HD, PP and PEEK were submitted to testing with regard of their processing behaviors, and components were tested as to their properties.
ANTEC 2010 / 2247
SLS Process and Demands on Thermoplastics During manufacturing, the plastics powder is applied, layer by layer. The powder that is located at the interface with the component is fused selectively by a CO2 laser, and thus connected insolvably to the melt layer below. The surrounding powder that was not molten supports the generated melt. Only if all layers of the component have been produced is the powder bed with its components inside cooled down.
Requirements to the Plastic Powder The properties of laser sintered parts, e.g. density, surface topography, accuracy of detail and dimensional stability are determined by the process parameters interacting with the material. In investigations it was found, for instance, that the geometry of particles is a decisive factor that substantially determines component coarseness [4,5]. Melt behavior (e.g. viscosity and surface tension) is just as significant as powder density and flowability of laser sintering powders. The size distribution and geometry of particles are of major importance for the sintered parts’ porosities [6,7]. High powder density thus leads to higher densities, dimensional accuracy and strength in the sintered parts, yet it may deteriorate flowabilities. Commercially available laser sintering powders with good flowabilities consist of spherulite particles with a narrow size distribution of d = 60 µm, and with a low share of fine particles of d = 10 µm. [8,9] Considering the long building times, the plastic material must particularly resist thermal degradation, since it is kept at a temperature close to the crystalline melting point for several hours. At the end of the production process, the thermally aged powder that had not molten is separated from the laser sintered component and will be recycled. In an ideal case, the powder should not be agglomerated, so that the part can be separated from the remaining powder merely by making use of the force of gravity. Due to ageing during building, the powder’s material properties usually change, which is why the remaining material must be mixed with roughly 30 wt-% of virgin powder, in most cases.
Thermal Boundary Conditions Apart from material application, temperature control during building is of major significance for the part’s profile of properties. For part density, a closed melt film is crucial, making it necessary for the powder material to have low melt viscosity [9]. Due to their wide range of softening and the resulting possibility of achieving low component densities, as well as dimensional stability,
2248 / ANTEC 2010
amorphous thermoplastics are employed e.g. as lost cores in precision casting [3]. Thus for direct part manufacturing, only the semi-crystalline polymers are relevant at present and PA12 (EOS PA2200) is the only commercially available SLS powder. Semi-crystalline plastics, however, are heated to the point above glass transition temperature, close to the crystalline melting point, with the laser merely fusing the crystalline shares. Due to the high chain mobility, decrease in viscosity is much steeper after exceeding this narrow range of crystalline melting. [10,11] In an ideal case, the building process would lead to the model state of quasi-isotherm laser sintering, where melt and solid powder exist side by side. The building process is carried out in a sort of two-phase mixing state. The laser merely introduces the energy that is necessary for the material to exceed the point of phase transition. Temperature increase in the surrounding powder bed should be as small as possible, here. This leads to another requirement of this process, in terms of material properties: The plastic material’s crystallization temperature should be clearly below the crystalline melting point [9]. Dynamic Differential Scanning Calorimetry (DSC), among others, can describe phase transitions, and thus the differences in temperatures between crystalline melting and crystallization. On this basis, the possible range of processing temperatures during SLS production may be defined. In case this range is exceeded, powders will get out of control and melt, whereas, if temperature is below this range, the plastic melt produced up to then will start to crystallize, causing shrinkage or curling. Following this model, powder and part are slowly cooled down only when the building process is completed. A low temperature gradient is applied, which provides for components with little residual stress and high dimensional stability.
Materials and Methods Commercially available laser sintering powders made from polyamide 12 (PA2200) and polyetherketone (PEEK HP3) were tested as to their processing properties and compared to PE-HD and PP sintering powders. Grinding tests were also conducted on PA12, PP and POM pellets, in order to look into the potential included in this method of preparation. Additionally, tensile bars made of PP laser sintering powder by 3D-Systems were used as a reference for the parts made from the polymer type by the LKT. In Table 1 an overview of the material properties for the investigated properties is shown [12]. The POM powder was mixed with 0.2 wt.-% of Aerosil®, in order to step up flowability. In IR measurements conducted on PE-HD before, the material was found to have a high transmission coefficient when submitted to the wavelength of the used CO2 laser (λ = 10.6 µm). By adding 0.4 wt-% of Carbon Black® the penetration depth of
the laser could be limited to approximately 100µm. As a result, the parts were black, had a high degree of absorption and better flowability. Table 1: Properties of the investigated polymers [12]
PA12
E-modulus [N/mm2] 1400
σYield [N/mm2] 50
εBreak [%] ~ 200
Tmelt [°C] 170 - 180
PP
1400-1800
25 - 40
> 50
160 - 165
PE-HD
600 - 1400
18 - 30
> 50
125 - 135
POM
2600-3200
60 - 75
20- >50 (H) 15-40 (C)
175 (H) 164-172 (C)
PEEK
3700
100
> 50
335
the determined beginning of crystallization. The exothermal heatflows resulting during crystallization were recorded. The PA 12 material was submitted to previous drying (below N2) at 120 °C for 15 minutes in DSC, since otherwise the values are very much scattered.
SLS Parameter Studies Having determined the admitted building temperatures in previous thermoanalytical tests, reference specimens could be produced from POM, PE-HD, PP and PA 12. For this purpose, tensile bars were made following different irradiation strategies. The parameters were varied according to the following principle, until the layers were bonded completely:
EL = Generating spherulite particles directly from polymerization is not possible with all types of plastics. As an alternative, powders can be made from pellets. Cryoscopic grinding is a well-proven method to make powder particles with a size less than 100µm. The pellets are cooled down to Tmill=-50°C in a cooling section and fed into a counterrotating pinned disc mill (impact crusher principle). After milling, the powder is classified and screened down to the desired particle diameter. The fine grain with diameters below 80 µm was sieved and investigated.
Thermoanalytical Investigations To investigate the melting and crystallization behavior of the thermoplastic material employed, the SLS process was simulated by DSC measurements. According to DIN 53765, 10 or 20 K/min is the standard heating or cooling rate, respectively, for thermoplastics. Laser sintering is however a slow process of part generation, which makes standard measurement unsuited to sufficiently describe the real process. The molten contour is kept to high temperature over a long period of time. The temperature is close to the crystalline melting point, and new preheated layers are applied on the previously applied ones. For these processes, models are available [9], which were added the time-related process of crystallization in the tests performed. For this purpose, several heating and cooling rates (10/5/1 K/min) were set and the resulting heatflows measured. For ideal quasi-isotherm laser sintering, scientists started from the assumption that the melt does not crystallize for a long period of time at the point just below the crystalline melting point. With crystallization related to time and temperature, further DSC tests were carried out alongside the laser sintering process. The specimens were heated to T = Tm+20 K at 10 K/min, cooled down to measurement temperature at a cooling rate of 40 K/min and kept in the isothermal state at different temperatures above
PL vS ⋅ h S
EL = energy density PL = laser power hS = hatch distance vs = scanning speed
Component Testing Layer bonds and resulting morphologies of the tensile bars’ were determined by examining images of microtome cuts taken by a transmitted light microscope. Tensile tests were conducted on the laser sintered parts according to DIN EN ISO 527-1,-2, and the fracture surfaces examined under a scanning electron microscope (SEM).
Results and Discussion Powder Preparation In the grinding tests, only POM revealed large quantities of powders with particle sizes below 80 µm, Figure 2. Thus an economic powder production can be realized if this material enters the market. This process generates irregular particle geometries and a broad range of different gradings with a high share of fine sizes. In Table 2 you can see the various semi-crystalline thermoplastic pellets after grinding in comparison to commercially available SLS powder. Those powders were then sieved below 80µm to fit the SLS process. For PP cryogenically milled powder could not be investigated and compared to precipitated PP powder in further steps.
ANTEC 2010 / 2249
Figure 2: SEM Image of PA 2200 powder (left) and cryoscopically grinded POM powder (right) Table 2: Distribution of particle sizes in cryoscopically grinded thermoplastics Size [µm] < 250 < 125
PA12 [%] 88.5 49.5
< 63
13.5
PP [%] 28.5 4.5 -
POM [%] 97.5 88.0 53.5
Thermoanalytical Investigations The findings presented in Figure 3 show that commercially available thermoplastics start to crystallize even before the standard crystallization temperature is reached. The range of processing parameters determined in the standard process is thus unsuited to serve as the sole processing criterion. Besides the determination of the building temperature via the difference between crystallization and melting peak also isothermal curves were measured for different temperatures between the phase changes. Thus the time-stability of the two-phase area could be analyzed.
2250 / ANTEC 2010
Figure 3: Heating, cooling and isothermal soaking of commercially available SLS powders by means of DSC In the shown DSCs (Figure 3 and Figure 4) the time until the crystallization peak was reached is illustrated. Above the highest shown temperatures no defined crystallization peak could be observed. This was the minimum temperature used for the generation of samples in the SLS process. In general PA 12 (PA 2200), after pre-drying, is the material that features the latest crystallization point among all examined thermoplastics, which means that above Tiso = 172 °C one may assume there is a two-phase mixing area over the entire building process.
layer bonding, while keeping energy density as low as possible. Figure 5 presents embedded microtome cuts taken from tensile test bars in a parameter study looking into POM at different energy densities, and it shows PA 2200 as a reference. In the top images, the laser fails to penetrate the applied powder layer completely. At the rear of the individual layers, there are powder particles visible that have not fused. In case of incomplete layer bonding, as presented in Figure 5 (bottom), a homogenous component with a high degree of crystallinity (KPOM = 85 %) is obtained.
Figure 5: Transmitted light images with polarized light taken of microtome cuts from laser sintered tensile bars t. l.: POM (EL = 35.0 mJ), t. r.: POM (EL = 40.0 mJ) b. l.: POM (EL = 60.0 mJ), b. r.: PA 2200 (EL = 60.0 mJ)
Figure 4: Heating, cooling and isothermal soaking of new investigated powders by means of DSC Accordingly, to set the building temperature right, it is not sufficient to know the range of possible temperatures determined by DSC. For instance, if a low building temperature is set, crystallization may start during layer generation, which typically reveals in the component contour bending upwards. Especially, if new thermoplastics like PE (Figure 4) with a narrow range of possible temperatures between melting and crystallization are processed, this can be an unfavorable aspect, which is why high building temperatures are usually recommended here. SLS Parameter Studies The Laser sintering tests were mainly focused on determining the building parameters that permit for complete
Figure 6: Transmitted light images in polarized light taken of microtome cuts from laser-sintered tensile bars left: PE-HD (EL = 80.0 mJ) right: PP (rectangular built, EL = unknown) The transmitted light images in Figure 6 show the crosssection of a tensile bar (left) from PE-HD, that was sintered in horizontal position, as well a vertically-built tensile bar from PP (right), made by 3D-Systems.
Component Testing
ANTEC 2010 / 2251
The test specimens produced were tested as to their mechanical properties in a tensile test and compared to each other, Figure 7. Accordingly, with horizontally built tensile test bars, values of E-modulus as well as tensile stress at break are comparable to the expected material values from literature [12]. The tested tensile bars by 3D-Systems had been built vertically, and cannot be compared to the horizontally built bars. For all specimens the values of elongation at break achievable up to now are clearly below the expected values [12]. Apart from the structure resulting from long sintering times, and from the high degree of crystallinity, the poor elongation at break is due to the coarse surface and the residual porosity of the parts. Fractures start mainly in the notches on the coarse surface. POM shows a relatively homogeneous and plane surface, making it less prone to fail from notches than other measured components. #
Figure 7: mechanical properties of laser sintered tensile bars Figure 8 shows the fracture mirrors for a sintered vertically-built 3D-Systems PP bar as well as for a sintered POM tensile bar made by the LKT. The PP part shows a large ductile fracture onset starting at the lower boundary layer (see figure, left). In the POM specimen, the fracture starts in a large pore and soon ends in a brittle residual fracture.
Figure 8: SEM image of fracture onsets in laser-sintered tensile bars from PP (left) and POM (right), respectively
Summary and Outlook
2252 / ANTEC 2010
It could be shown that is possible to generate components of thermoplastic materials other than PA 12 by laser sintering under consideration of optimized processing parameters. This broader spectrum of usable materials opens up a wide range of applications for the process. Cryogenically milling is a good technology to generate fine powders for SLS but it is not possible for every thermoplastic. Knowledge of the process has also been improved for this technique. Considering process behavior, PA 2200 certainly is an extremely robust material. However, thanks to improvements in machine engineering, the potential of other laser sintering powders for commercialization has been stepped up too. Materials such as PP and PEEK are currently entering the market. Apart from the investigations presented in this study, concerned with materials’ fusing and crystallization behaviors, other techniques of thermoanalytical measurement are growing in importance. They are required to qualify these new materials comprehensively. These techniques comprise aspects of absorption behavior of the powder fill, film formation based on melt viscosity, but also factors of scaling because of the specific volumes changing at phase transition. Especially the existence of the two-phase SLS model has to be studied to know more about the beginning of melting and crystallization. Moreover, it was possible to generate, from the mentioned powders, components with typical basis material properties – with regard to stiffness and strength, whereas fracture behaviors were very brittle with all the materials under investigation. Finally there is high potential for further improvements of process-specific measurements techniques and the processing of new polymers.
References 1.
Rietzel, D., Kühnlein, F., Feulner, R., Hülder, G., von Wilmowsky, C., Fruth, C., Nkenke, E., Schmachtenberg, E., Proceedings SPE – European Conference on Medical Polymers 2008, Belfast, UK, (2008) 2. Kühnlein, F., Rietzel, D., Wendel, B., Feulner, Hülder, G., Schmachtenberg, E. Plastverarbeiter, 08 (2008) 3. Pfister, A., Dissertation, Albert-Ludwigs-Universität Freiburg i.Br. (2005) 4. Gebhardt, A.: Rapid Prototyping – Werkzeuge für die schnelle Produktentstehung. Carl Hanser Verlag, München, 2000 5. Podszun, W., Harrison, D., Alscher, G., Patent DE 19820725.5 (1999) 6. McGeary, R., Journal of the American Ceramic Society, 44, pp. 513-522, (1961) 7. Seul, T., Dissertation, RWTH Aachen (2004) 8. Nelson, C., Dissertation, University of Texas at Austin, USA (1993) 9. Alscher, G., Dissertation, GH-Essen (2000) 10. Nöken, S., Dissertation, RWTH Aachen (1997)
11. Eyerer, P., Kunststoffe, 83 (1993) 12. Ehrenstein, G.W.: Polymeric Material. Carl Hanser Verlag, München, 2000
Key Words Selective Laser Sintering, Additive Manufacturing, Isothermal Crystallization, POM, PP, PA2200, PE
ANTEC 2010 / 2253
