Characterisation of carbon fibre-reinforced polyamide manufactured by selective laser sintering

Characterisation of carbon fibre-reinforced polyamide manufactured by selective laser sintering

Additive Manufacturing xxx (2015) xxx–xxx Contents lists available at ScienceDirect Additive Manufacturing journal homepage:

3MB Sizes 0 Downloads 7 Views

Additive Manufacturing xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Additive Manufacturing journal homepage:

Full length article

Characterisation of carbon fibre-reinforced polyamide manufactured by selective laser sintering Anton Jansson ∗ , Lars Pejryd Örebro University, Department of Mechanical Engineering, Fakultetsgatan, 1, 701 82 Örebro, Sweden

a r t i c l e

i n f o

Article history: Received 4 June 2015 Received in revised form 26 August 2015 Accepted 14 December 2015 Available online 21 December 2015 Keywords: Computed tomography Carbon fibre–reinforced polymer Selective laser sintering Additive manufacturing

a b s t r a c t Polymers and reinforced plastics are employed in various load-bearing applications, from household objects to aerospace products. These materials are light, strong, and relatively cheap but can be difficult to form into complex geometries. However, the development of additive manufacturing processes has made it easier to manufacture reinforced plastics in complex shapes. The aim of this work was to study the internal features and mechanical properties of carbon fibre-reinforced polyamide (CF/PA12) fabricated with the additive manufacturing technique of selective laser sintering. The test specimens were studied using computed tomography to analyse the internal geometry, and the material proved to be highly porous. Moreover, the test specimens revealed an internal layered structure, which was found to have a great effect on the tensile properties of the material. The results highlight that there is room for further optimisation of the manufacturing parameters for CF/PA12, because the layered structure makes it challenging to design end user parts with acceptable mechanical properties. © 2015 Elsevier B.V. All rights reserved.

1. Introduction In recent years, a new branch of manufacturing techniques, called additive manufacturing (AM), has evolved from rapid prototyping, which has been around since the 1980’s [1]. In AM, small portions of material are added layer by layer to create the end product instead of removing material from a larger bulk. The AM technique has significant advantages over classical manufacturing methods, and designers can create parts with any geometry without being restricted by the limitations of milling, lathing, and moulding. There are several materials available for AM, such as a variety of plastics with and without reinforcement fillers and a selection of metallic alloys. Selective laser sintering (SLS) is an AM method that is based on the powder-bed technique. In SLS, powder is raked in thin layers over a build table that moves downwards in steps, and lasers melt a portion of the added layer according to the geometry of the part to be built. This creates a thin slice of the part and fuses this section to that beneath it. Today, the technique has advanced to a point where the produced parts no longer serve only as prototypes but

Abbreviations: CF/PA12, carbon fibre–reinforced polyamide; SLS, selective laser sintering; CFRP, carbon fibre–reinforced polymer; CT, computed tomography. ∗ Corresponding author. E-mail addresses: [email protected] (A. Jansson), [email protected] (L. Pejryd).

possess material properties that are suitable for end user products [2]. One of the new SLS materials being developed for end user parts is carbon fibre-reinforced polyamide (CF/PA12). The material itself is well known in the industry and has been used for many years for fabricating complex and light parts by injection moulding [3]. The material in its raw form is a powder consisting of polyamide spherical particles with diameters in the range of 50 ␮m mixed with carbon fibres of diameter 10 ␮m and length 100–200 ␮m. The fibres normally undergo chemical treatments before they are blended with the polyamide to achieve greater adhesion between the fibres and the plastic matrix. A higher percentage of fibres increases the strength of the material. Yan et al. found that a fibre percentage of approximately 40% was suitable for the SLS process and that higher percentages caused problems when applying new coats of powder layers [4]. The mechanical properties of the material are generally strengthened by the fibres, making it stiffer, stronger, and lighter. However, the material properties are influenced by the fibre orientation. For example, in injection-moulded parts, the fibres themselves align along the melt flow direction, increasing the material strength in that direction but decreasing the strength in other directions [5]. The inhomogeneity in the material strength can restrict the part designs suitable for injection moulding, because the options for the melt flow directions are limited. 2214-8604/© 2015 Elsevier B.V. All rights reserved.

Please cite this article in press as: A. Jansson, L. Pejryd, Characterisation of carbon fibre-reinforced polyamide manufactured by selective laser sintering, Addit Manuf (2015),


A. Jansson, L. Pejryd / Additive Manufacturing xxx (2015) xxx–xxx

Fig. 1. The manufacturing orientation of the tensile bars in the build chamber. The powder spreading rake travels in the x-direction.

SLS could prove to be a powerful alternative manufacturing method in which the fibre alignment could be controlled more effectively. CF/PA12 parts built by SLS exhibit different material properties in different directions, and there are claims that this is linked to the fibre orientation in the material. It is thought that the fibres align themselves along the direction in which the powder is spread in the build chamber, i.e. the x-direction. However, the SLS process is a complex method with a large number of parameters that affect the outcome of the produced material, such as powder composition, laser beam settings, temperature of the powder bed, and powder layer thickness. Moreover, there have been reports of porosity in other carbon fibre–reinforced polymers (CFRPs) manufactured by SLS, such as in the PA12 material studied by Van Hooreweder et al [6]. The group compared the mechanical properties of SLS parts and injection-moulded parts that used the same material and found that the SLS parts were weaker. They also detected porosity in the SLS parts by examining their cross sections and postulated that the porosity was homogenous throughout the samples. Other studies on SLS parts have also indicated that porosity is caused by this process [7,8]. To study complex materials such as CFRPs, there is a need to investigate their internal features in a non-destructive manner. It has been recognised that computed tomography (CT) is particularly suited for analysing complex parts built by AM methods [9]. The use of industrial CT systems has escalated in recent years, and as the complexity of parts increases, new reliable non-destructive methods of investigation have become necessary. CT has proved to be an effective tool for studying composite materials, as has been shown by Meneghetti et al. [10]. among others. The aim of this work was to characterise carbon fibre-reinforced polyamide parts manufactured by SLS, to broaden our understanding of CF/PA12 and its properties. 2. Materials and methods The SLS system used in this study was from one of the major equipment manufacturers in the area (EOS P396), using the recommended parameters set by the equipment manufacturer. The CF/PA12 material analysed was also one of their products and is marketed as having ‘outstanding mechanical properties charac-

Fig. 2. Cross sections from a y direction tensile bar (a) cross section normal to the build direction. (b) cross section from the centre of the tensile bar, showing a layered structure that continues throughout the specimen. (c) cross section along the fusing area of two build layers.

terised by extreme stiffness and strength’. The material is known to exhibit variations in properties depending on the build direction in the SLS build chamber. Tensile bars complying with ISO-527 [11] were built in six directions in the build chamber: x, y, xy, x 45◦ , y 45◦ , and xy 45◦ . The bars labelled 45◦ had a 45◦ tilt angle with respect to the z-axis of the build chamber, as illustrated in Fig. 1, whereas the x, y, and xy tensile bars were built in the plane of the build table. Three tensile bars was built in each build direction. The thickness of the bars was 4 mm, and that of a newly spread powder layer was 0.15 mm. Therefore, each of the non-tilted and tilted bars was built up by approximately 27 and 800 powder layers, respectively. The tensile bars were built using the cross-directional laser sweep technique. The laser starts each layer by melting the contour of the part slice and then alternatively sweeps along the x- and ydirections to fill the slice. Thus, if the first layer was melted along the x-direction, the next layer will be melted along the y-direction. The tensile bars were examined in the state ‘as delivered’, which meant that they had been exposed to a light abrasive treatment using glass beads. A 20 × 40 × 5 mm piece of standard PA12 was also fabricated, using SLS, as a reference object for porosity studies.

Please cite this article in press as: A. Jansson, L. Pejryd, Characterisation of carbon fibre-reinforced polyamide manufactured by selective laser sintering, Addit Manuf (2015),

A. Jansson, L. Pejryd / Additive Manufacturing xxx (2015) xxx–xxx


Table 1 Tensile properties of test specimens with different build directions.

Build direction

Tensile strength (MPa) Average Standard deviation

Strain at breaking point (%) Average Standard deviation

Young’s modulus (GPa) Average Standard deviation

x y xy x 45◦ y 45◦ xy 45◦

66.7 54.0 56.7 31.3 31.9 30.9

3.0 5.1 4.6 2.7 2.5 2.4

6.3 3.6 4.1 2.4 2.1 2.1

0.3 1.1 0.7 1.3 0.6 1.2

0.6 0.1 0.3 0.9 0.3 0.9

0.2 0.3 0.1 0.1 0 0.1

The CT system used for inspection of the specimens was a Nikon XT H 225 with a micro-focus X-ray source. The target material was molybdenum, the acceleration voltage was 62 kV, and the filament current was 153 ␮A. CT scans of 18 tensile bars were performed at the centres of the bars, capturing approximately 25–30 mm of the total 80 mm gauge length. The scanned data was then analysed using commercially available software. The fibre orientations of the tensile bars were studied by optical examination of the samples. Pieces of the tensile bars were mounted in epoxy and grinded/polished down to the centre of the bar. The orientation of the fibres was determined by studying several cross sections taken in three different directions, one along the build layers, one perpendicular to the build direction, and one along the side of the building plane.

3. Results The results from the tensile tests, shown in Table 1, were mostly consistent with the material properties provided by the supplier. Some discrepancies were observed for the x-direction specimens, for which the tensile strength was found to be approximately 7% lower than the value specified by the manufacturer. The data points was collected from three tensile bars in each build direction. In comparison to standard PA12 from the same manufacturer, the CF/PA12 exhibits a 28% higher tensile strength for the xdirection and a 12% increase for the y-direction. The increase in strength does come with a cost in form of a more brittle material; the strain at break for the x-direction is reduced by 80% compared to PA12. The Young’s modulus for the x-direction and y-direction is increased by 73% and 53% respectively compared to PA12. The CT data revealed a highly porous structure inside the tensile bars, and the porosity was found to be concentrated and homogenous in the interlayer planes. The material was also found to have a high degree of poor fusion at the part edges where the layers from the build had failed to fuse together. The porosity and poor fusion of the layers in a bar built in the y-direction can be seen in Fig. 2. In addition, warpage was detected in all of the tensile bars. The bars that were built in the plane of the build chamber were only warped normal to the build direction, whereas the tilted bars were warped along the build direction as well. Because the porosity was concentrated at the interlayer planes, the cross-sectional areas of such regions were significantly lower. The tilted tensile bars showed the smallest cross-sectional areas, so the highest stresses in those bars were along the build layers rather than across the cross section normal to the build direction. Consequently, the tilted bars all fractured along the building layers. A cross section of a bar built in the xy 45◦ direction and its fractured surface are shown in Fig. 3. The porosity seen in Fig. 2 was quantitatively measured and compared with the total volume of the test specimens. The total volume was defined to be along the contours of the poorly fused part edges. The results are listed in Table 2.

Fig. 3. (a) Cross section of a xy 45◦ tensile bar. (b) The fracture after tensile testing. The tensile bar fractured at the bonding area between layers.

Table 2 Internal porosity of test specimens.

Build direction

Internal porosity (%) Average

Standard deviation

x y xy x 45◦ y 45◦ xy 45◦

11.8 12.3 15.9 13.2 11.4 12.3

0.2 0.2 0.7 1.3 1.3 1.7

The results in Table 2 can be seen as the bulk porosity of the material. There were variations in the average bulk porosity depending on the build direction, and within each direction, there were differences of as much as 4%. Moreover, the specimens that were built in the xy direction in the plane of the build chamber were found to have the highest bulk porosity. The reference PA12 piece displayed a porosity value of 8.2%. The PA12 porosity level is in agreement with results from other groups, such as Kruth et al. [12]. The sizes of the pores in CF/PA12, range from 0.0001 mm3 up to 0.05 mm3 and the larger porosities often persist through several build layers within the bulk of the material. A few examples of pores can be seen in Fig. 4. The figure also includes a pore from the reference PA12 piece.

Please cite this article in press as: A. Jansson, L. Pejryd, Characterisation of carbon fibre-reinforced polyamide manufactured by selective laser sintering, Addit Manuf (2015),


A. Jansson, L. Pejryd / Additive Manufacturing xxx (2015) xxx–xxx

Size distribution of pores

log(Number of pores/mm3)

1000 CF/PA12



10 1 0.1 0.01 0.001 0.000


0.019 0.029 0.039 Pore volume (mm3)



Fig. 5. Size distribution of pores in CF/PA12 and PA12. From the diagram it can be deduced that the CF/PA12 material has a larger amount of small pores per mm3 than PA12 does. CF/PA12 also has a larger amount of big pores. PA12 displays a more concentrated size distribution than CF/PA12.

Porosity of tensile bars 35%

Level of porosity

30% 25%

Porosity excluding poorly fused edges Porosity including poorly fused edges

20% 15% 10% 5% 0% x


xy x 45° Build direction

y 45°

xy 45°

Fig. 6. Porosity level of tensile bars, including/excluding the poorly fused edges.

Fig. 4. (a) A large size porosity found in CF/PA12. The porosity stretch through several building layers and has a volume of 0.0071 mm3 . (b) A medium size porosity found in CF/PA12, the volume of the pore is 0.0012 mm3 . (c) A small size porosity found in CF/PA12. These are the most common pores with sizes such as 0.0005 mm3 . (d) A large size pore from PA12. Unlike the large pores found in CF/PA12 this pore is compact in its shape.

As shown in Fig. 4 the larger porosities in CF/PA12 stretches through several build layers in the material while the large porosities in PA12 are not as far reaching. The size distribution of the CF/PA12 porosities in comparison with the porosities found in the reference PA12 piece is shown in Fig. 5. The distribution is collected from approximately 4000 mm3 of CF/PA12 and PA12 respectively. Because all the CF/PA12 specimens showed poor fusion at the edges of the parts, the edges of the specimens could not be clearly detected. If the poorly fused edges of the layers were used as the edges of the specimens (as they would be if measured by tactile/optical methods), then the resulting porosity is dramatically changed. The results from such measurements and a comparison between including/excluding the delaminated layers are shown in Fig. 6. The difference in determining the defect volumes is explained further in Fig. 7. The edge of the material was determined as a line that connects the ends of the poorly fused layers. Furthermore, the high level of porosity and the poor fusion at the part edges observed from the CT data were verified by optical examination of the fractured surfaces of the tensile bars, as shown in Fig. 8. The edge phenomena stands in contrast with the reference PA12 piece where no such effects were found. CT slices from the PA12

Please cite this article in press as: A. Jansson, L. Pejryd, Characterisation of carbon fibre-reinforced polyamide manufactured by selective laser sintering, Addit Manuf (2015),

A. Jansson, L. Pejryd / Additive Manufacturing xxx (2015) xxx–xxx


Fig. 9. Cross sections from the PA12 piece. (a) A cross section in the build plane of the piece as well as a side view. (b) Cross section taken across the build direction of the piece.

Fig. 7. (a) Cross section of a tensile bar, porosity and poor fusion can be seen. (b) The cyan borders shows the internal defects of the material. (c) Here, the cyan border includes both the internal defects and the poorly fused edges.

piece, Fig. 9, show a noticeably more homogeneous distribution both in size and location than the pores in the CF/PA12 material. A total of 3000 fibres were studied to assess their orientations, and the measurement procedure is illustrated in Fig. 10. The preferred direction of the fibre orientation was found to be in the x-direction of the build chamber, and no fibres were seen oriented perpendicular to the build plane, i.e. in the z-direction. A comparison between the orientations in the x and y tensile bars can be seen in Fig. 11. 4. Discussion

Fig. 8. (a) Fractured surface from a x direction tensile bar. The level of porosity and poor fusion is in accordance with the CT-data. (b) Magnification of (a).

The SLS manufacturing process is highly complex, and the cause of the porosity is likely a combination of many different factors. The porosity was found to have a great effect on the material properties, more so than the fibre orientation in some build directions. It could be argued that a high degree of porosity decreases the material weight and might therefore be beneficial. However, the pores were filled with powder, and any weight-reducing effect was therefore diminished. From pressureless sintering of CFRP, it is well known that, because the fibres do not shrink, porosity may be a problem. The results of the present study agree with this assessment. When comparing the porosity of CF/PA12 to PA12 the higher levels of porosity found in CF/PA12 can have many causes. The packing of raw CF/PA12 powder in the build chamber is likely more problematic than pack a more homogenous powder such as PA12. The relatively long fibres in the CF/PA12 powder can cause voids in the packed powder that during the sintering process can cause pores. The thermal properties of the CF/PA12 powder is more complicated than it is in pure powders because of the combination of materials with different thermal properties. Small local variations in the composition of the powder can have effect on the size of the melt pool and thus giving porosity problems. A way to increase the density of the part would be to increase the beam energy, causing the laser to melt a larger volume of the powder during each pass. This strategy, however, would influence the part accuracy in a negative direction. A balance between part accuracy and part density must be achieved. It may be that for

Please cite this article in press as: A. Jansson, L. Pejryd, Characterisation of carbon fibre-reinforced polyamide manufactured by selective laser sintering, Addit Manuf (2015),


A. Jansson, L. Pejryd / Additive Manufacturing xxx (2015) xxx–xxx

Fig. 10. (a) Original optical image of fibres in a x tensile bar (b) Colour coded fibre directions (c) Orientation chart.(For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.).

some applications parameters producing higher density but lower accuracy are desirable, while the opposite might be true for other applications. Extensive parameter studies may therefore be needed to aid in the discovery of optimal parameters for different applications of the material. The bulk porosity was approximately the same for all the build directions except for the xy direction in the build Table √ plane. This could be because more laser sweeps (by a factor of 2) were used per layer to build those specimens than the others in the same plane. The reason that this porosity increase is not observed in the tilted xy specimens might be their smaller layer areas, which greatly reduce the amount of laser sweeps used for each layer.

Because the samples were abrasively treated by the manufacturer before delivery, nothing conclusive could be said about the lack of fusion depth at the part edges other than that it was substantial. As the tensile data corroborated well with the material data provided by the equipment manufacturer, we assume that the material is supposed to look this way and that it was not due to a faulty set of tensile bars. In the tensile strength data, there were three clear groups; x, y and xy, and the tilted bars. In terms of porosity, the x- and ydirections were very similar, so it can be inferred that the difference in material properties is mostly due to the fibre orientation. Moreover, the preference for fibre orientation in the x-direction of the build chamber could be because the rake that spreads the powder moves in that direction. The fibre lengths are typically of the same order or greater than the new powder layer that the rake spreads. Therefore, a fibre that is originally placed in an orientation perpendicular to the build plane is very likely to get hit by the rake as it spreads the powder, as illustrated in Fig. 12. This repeated mechanical action between the rake and the fibres will align most of the fibres in the x-direction of the build chamber, and only the fibres that were originally oriented in the build plane can maintain their orientation. Therefore, even if the orientation composition of the original powder is random, it will strongly tend towards the x-direction after being spread in the chamber. This theory also implies that very few or no fibres will have an orientation perpendicular to the build plane, which is in agreement with the results. All the tilted bars showed similar material properties. This may be because they have approximately the same cross-sectional areas, independent of their build directions. The fibres are oriented solely in the build plane and therefore offer no strength in between the layers. In fact, the fibres could have a negative impact on the interlayer strength of the material because the adhesion strength between the fibres and the polyamide is much less than the bond within the polyamide. The procedure for measuring the volume fraction of porosity is an important factor to be considered. As can be seen from the results, when the poorly fused edges is included in the measurement, there is not much impact on the volume fraction of the tensile bars built in the plane. However, the effect on the tilted bars is significant. Using the edges of the poorly fused layers as the part edges also highlights the large amount of poor fusion that is present in the tilted bars. Moreover, the higher level of porosity found in the tilted bars could also be because they consist of approximately 30 times more build layers than the bars built in the build chamber plane. Because all the layers show poor fusion, this effect would significantly increase the porosity. The large directional variations in the material properties would make designing parts for the SLS process a difficult task. We conclude that much of the versatility of the manufacturing process may be lost because of the limitations in the material properties.

5. Conclusions It was found that the carbon fibre-reinforced material was highly porous as a result of the SLS manufacturing process. The porosity was found to be concentrated in between the layers produced during the manufacturing process, and the layers of porosity significantly weakened the material in the direction normal to the layered structure (i.e. in the z-direction of the build chamber). Furthermore, the material exhibited different strength properties depending on the direction in which it was sintered in the build chamber. The different material properties in the build plane are mainly due to the fibre orientation, whereas the properties in the

Please cite this article in press as: A. Jansson, L. Pejryd, Characterisation of carbon fibre-reinforced polyamide manufactured by selective laser sintering, Addit Manuf (2015),

A. Jansson, L. Pejryd / Additive Manufacturing xxx (2015) xxx–xxx


Fiber orientation in tensile bars 25% x - tensile bar

y - tensile bar

Percentage of fibres





0% 90°



22,5° 0° -22,5° Fiber orientation




Fig. 11. The orientation of carbon fibres in the x and y tensile bars. The orientation is along the tensile bars build direction, 90◦ is parallel to the build direction of the specimen while 0◦ is perpendicular.

Fig. 12. Illustration of the rake spreading a powder layer in the build chamber. Fibres that are not oriented in the plane of the build chamber will become aligned by the interaction with the moving rake.

z-direction of the build chamber depend on the porosity along the build layers rather than the fibre orientation. References [1] J.-P. Kruth, G. Levy, F. Klocke, T.H.C. Childs, Consolidation phenomena in laser and powder-bed based layered manufacturing, CIRP Ann.-Manuf. Technol. 56 (2007) 730–759, [2] R.D. Goodridge, C.J. Tuck, R.J.M. Hague, Laser sintering of polyamides and other polymers, Prog. Mater. Sci. 57 (2012) 229–267, 1016/j.pmatsci.2011.04.001. [3] O. Wirjadi, M. Godehardt, K. Schladitz, B. Wagner, A. Rack, M. Gurka, S. Nissle, A. Noll, Characterization of multilayer structures in fiber reinforced polymer employing synchrotron and laboratory X-ray CT, Int. J. Mater. Res. 105 (2014) 645–654, [4] C. Yan, L. Hao, L. Xu, Y. Shi, Preparation, characterisation and processing of carbon fibre/polyamide-12 composites for selective laser sintering, Compos. Sci. Technol. 71 (2011) 1834–1841, 2011.08.013. [5] A. Bernasconi, F. Cosmi, D. Dreossi, Local anisotropy analysis of injection-moulded fibre reinforced polymer composites, Compos. Sci. Technol. 68 (2008) 2574–2581,

[6] B. Van Hooreweder, D. Moens, R. Boonen, J.-P. Kruth, P. Sas, On the difference in material structure and fatigue properties of nylon specimens produced by injection molding and selective laser sintering, Polym. Test. 32 (2013) 972–981, [7] B. Van Hooreweder, J.-P. Kruth, High cycle fatigue properties of selective laser sintered parts in polyamide 12, CIRP Ann.-Manuf. Technol. 63 (2014) 241–244, [8] K. Shahzad, J. Deckers, Z. Zhang, J.-P. Kruth, J. Vleugels, Additive manufacturing of zirconia parts by indirect selective laser sintering, J. Eur. Ceram. Soc. 34 (2014) 81–89, [9] L. De Chiffre, S. Carmignato, J.-P. Kruth, R. Schmitt, A. Weckenmann, Industrial applications of computed tomography, CIRP Ann.-Manuf. Technol. 63 (2014) 655–677, [10] G. Meneghetti, M. Ricotta, G. Lucchetta, S. Carmignato, An hysteresis energy-based synthesis of fully reversed axial fatigue behaviour of different polypropylene composites, Compos. Part B: Eng. 65 (2014) 17–25, http://dx. [11] ISO 527:2012: Plastics-Determination of tensile properties, International Organization for Standardization (2012). [12] J.-P. Kruth, G. Levy, R. Schnidel, T. Craeghs, E. Yasa, Consolidation of polymer powders by selective laser sintering, PMI (2008) 8286626/PMI08 Kruth Levy Keynote.

Please cite this article in press as: A. Jansson, L. Pejryd, Characterisation of carbon fibre-reinforced polyamide manufactured by selective laser sintering, Addit Manuf (2015),