Additive layer manufacture of tensile test specimens in stainless steel 316L by laser consolidation.

Sewell, Neil Thomas ; Bassoli, Elena ; Gatto, Andrea 等

1. INTRODUCTION

As Additive Layer Manufacturing (ALM) systems mature, more is expected of the functional parts produced using the technology. For ALM to be accepted as an industrial manufacturing technique, parts need to be produced reliably to high tolerances and with good mechanical properties. Currently, one of the major limitations for ALM is the difficulty in predicting part characteristics to ensure their robustness and repeatability.

Laser Consolidation (LC) is a powder-fed based technology which produces net-shape parts without the need for additional finishing. Powder-fed ALM offers promising advantages: metallurgical soundness, high strength and ductility of parts; ease of change of the powder supplied to enable multi-material or graded parts (Domack & Baughman, 2005); operation using standard CNC know-how enabling acceptance in most industries. As in every ALM system, the building process in LC is inherently anisotropic. In particular, the non-coaxial nature of the LC head assembly and its path are likely to cause direction dependant features varying not only along and perpendicular to the build direction, but also within each layer.

Previous work has suggested that parts produced using LC exhibit excellent mechanical properties (Xue & Islam, 2000) but often leave out important information regarding the experimental procedure and processes used (Toyserkani & Khajepour, 2006). More information is available for a similar powder-fed system known as Laser Direct Deposition (LDD), using a continuous CO2 laser coaxial with the powder nozzle where fully dense specimens were made in a nickel based superalloy resulting in high strength and remarkable ductility (Zhang et al., 2007). Studies on Titanium alloys (Gao et al., 2007) proved that parts can be successfully built by LDD with higher mechanical properties than obtained by casting and equal to those obtained by wrought annealed parts. These authors also investigated and proved part anisotropy.

With LC, the direction of build, either axially aligned with the head or orthogonal to it, will affect the characteristics of the part due to the non-coaxial head configuration. This paper examines the use of LC to produce some simple dog-bone samples in stainless steel 316L using two different orientations, parallel and orthogonal to the build head. The results illustrate that although the mechanical and microstructural properties of the parts are excellent, there are some differences due to part orientation. Explanations for the differences are investigated.

2. EXPERIMENTAL PROCEDURE

2.1 Part Manufacture

Parts for testing were manufactured using the Accufusion Laser Consolidation System, owned by Airbus and based at the University of Exeter. This system is currently involved in a UK Government Technology Strategy Board sponsored project, named DAMASCUS, to investigate the ALM of parts for aerospace and automotive. The LC system head assembly consists of a non-coaxial powder feeder nozzle and a pulsed Nd:YAG laser. Powder is fed into the laser beam at its focal point and consolidates as the laser source is removed. Building starts on a substrate that is moved beneath the head assembly to produce a consolidated bead. The head unit is raised incrementally to build up the part focusing the beam on the previously built section. In this experiment, beads of material were deposited one on top of another at 0.1mm increments to create rectangular parts just over 20mm high (200 layers) by 180mm long. Bead width depends on the amount of powder that consolidates within the laser spot: in this experiment parts were 1mm thick.

Effectively LC is a head assembly mounted on the Z-axis of a CNC system with the X and Y axis moving in the horizontal plane beneath. Additional rotational axes are available. A simple CNC program was designed to deposit the 180mm long beads of material on the substrate, a 30mm thick steel plate. For the specimens with the axis parallel to X direction the head scan during build was parallel to the direction of the laser beam and the powder fed (Figure 1). Reciprocating axial passes were used, i.e. the first part being built from -X to +X (X+ve), the second from +X to -X (X-ve) and repeated. After manufacturing eight test samples along the X axis (4 for each vector), parts were made along the Y axis in a similar manner. For Y specimens the head unit moved orthogonally to the head assembly plane. Parts were then removed from the substrate along the first few beads as allowed for by the 2mm overbuild added. The parts then had waists added by spark erosion obtaining geometry consistent with standard BS EN ISO 527-2 (reduced section width 10mm, length 75mm).

[FIGURE 1 OMITTED]

2.2 Processing Parameters

Parts were produced using previously derived settings for processing stainless steel 316L by LC. 10ms pulses at 30Hz with 5J per pulse were used with a 5g per minute 316L powder feed rate. An initial build speed of 300mm per minute was used, deliberately lower that usual, to allow for the heat absorbance of the substrate affecting build height for the first 10 layers (Toyserkani & Khajepour, 2006). After this a rate of 375mm per minute was used, derived from a desire to overlap laser spots by a minimum of 50%. Given that the laser spot diameter is 0.5mm, the laser needs to fire once every 0.25mm. At 30Hz with a 50% overlap, substrate movement would be 450mm per minute, however, preliminary experiments showed that inconsistent build-up occurred at rates higher than 375mm per minute in the axis perpendicular to the laser head assembly.

2.3 Tests

Ultimate Tensile Strength (UTS) tests were carried out on at least 5 parts in each orientation at a speed of 2mm per minute with strain measured using an extensometer acting on a gage length of 50mm. After testing, rupture surfaces and un-melted powder were observed using the scanning electron microscope (SEM) to investigate failure mechanisms and joining phenomena between the particles.

3. RESULTS AND DISCUSSION

Table 1 shows the main results of the UTS tests. The mean and Standard Deviation (SD) were calculated for the positive and negative vectors of X and Y directions separately and together. The separate build vectors are inconclusive as there were too few parts for a meaningful SD. Yet, no remarkable differences can be seen and the stress-strain graphs are very similar. The X direction results show high consistency in their behaviour in both directions, whereas the Y direction data is less consistent. UTS of parts built in the Y direction is on average 40MPa lower than that of the X specimens and the maximum strain is almost half. All LC parts show high strength when compared with sheet stainless steel 316L values. Elongation at break is only slightly smaller for X parts. Typical rupture surfaces of X and Y direction parts are shown in Figure 2.

[FIGURE 2 OMITTED]

X specimens (Fig. 2a) exhibit uniform ductile failure morphology based on micro voids forming under strain; porosity previous to rupture is almost absent and joining between the layers is not visible. For Y parts, instead, many large voids are evident throughout the surface (Fig.2b), up to 100 [micro]m wide. The smooth inner surface suggests they were formed during laser consolidation. In Y direction the powder is sprayed across and not along the section buing built, which can likely be the cause for less material consolidated in the building area.

4. CONCLUSION

In conclusion, this study confirmed that LC parts show high mechanical properties, equal to or better than parts produced traditionally. In particular, ductility is remarkably high compared with other additive processes. Anisotropy within the building plane was analysed, finding that the specimens created in the X direction exhibit higher strength and much higher ductility when compared with parts created in the Y direction. The parts in the X direction demonstrate consistent results throughout the test samples, where as the parts in the Y direction displayed greater variation between the results. SEM observation of the rupture surfaces revealed numerous voids in the structure of the parts in the Y direction; it was proposed that these void could be a dominant factor in tensile properties. Further study will be carried out to clarify this speculation and to investigate the factors contributing to void formation. Polished sections will be observed to calculate the percentage porosity in X and Y directions. No definitive result has been obtained as to positive and negative scan vectors, that will be further investigated. As a direct consequence of this research, a specific CNC programming employing the rotation axis of the worktable could ensure maximum mechanical response in the load direction.

5. REFERENCES

Domack, M.S. & Baughman, J.M. (2005). Development of nickel-titanium graded composition components. Rapid Prototyping J., Vol. 11, No. 1, 41-51, ISSN: 1355-2546.

Gao, S.Y.; Zhang, Y.Z.; Shi, L.K.; Du, B.L.; Xi, M.Z. & Ji, H.Z. (2007). Research on Laser Direct Deposition Process of Ti-6Al-4V Alloy. Acta Metall. Sinica (English Letters), Vol. 20, No. 3, (June 2007) 171-180, ISSN: 1006-7191

Toyserkani, E & Khajepour, A. (2006). A mechatronics approach to laser powder deposition process. Mechatronics, Vol. 16, No. 10, (Dec. 2006) 631-641, ISSN: 0957-4158

Xue, L. & Islam, M.U. (2000). Free-form laser consolidation for producing metallurgically sound and functional components. J. of Laser Applications, Vol.12, Num.4, (Aug 2000) 160-165, ISSN: 1042-346X

Zhang, K.; Liu, W. & Shang, X. (2007). Research on the processing experiments of laser metal deposition shaping. Optics & Laser Technology, Vol. 39, 549-557, ISSN: 0030-3992.

Tab. 1. Tensile tests results.

Sheet AISI 316L UTS = 485MPa b

 No. UTS b
 parts (MPa) (%)

 mean SD mean SD

X +ve 4 589 5.6 31.0 3.4
X -ve 3 582 7.9 34.7 1.0
X all 7 586 8.1 32.8 2.8
Y +ve 3 537 34.8 15.6 8.1
Y -ve 2 555 0.0 18.6 0.2
Y all 5 544 28.4 16.6 6.4