C. Y. Kong, P. A. Carroll, P. Brown, R. J. Scudamore
TWI Technology Centre (Yorkshire) Ltd, Wallis Way, Catcliffe, Rotherham, S60 5TZ, UK
Paper presented at 14th International Conference on Joining of Materials, Sunday 29 April 2007 - Wednesday 2 May 2007, Helsingør, Denmark.
This study investigated the effect of average powder particle size on deposit quality, surface finish and powder deposition efficiency, using the process of direct metal laser deposition (DMLD). Five samples of Inconel 625 powder with a range of particle sizes, chosen because they are commonly used in the DMLD process, were considered. In this work, a coaxial powder deposition nozzle capable of producing a small powder focus was used in conjunction with aTrumpf DMD505 laser deposition machine. The DMD505 comprises a 2kW CO2 laser and has a single cantilever 5-axis Cartesian gantry with CNC, a powder feed system, a powder delivery nozzle, and a patented feedback and control system for producing layers of constant height. The experimental work showed that, for the Inconel 625 used, the average particle size has a direct and indirect effect on the deposit height and efficiency. Smaller particle sizes can be focused more easily resulting in better efficiencies and larger deposits. If the particle size is too small however, problems with powder delivery can occur causing a reduction in deposit height and efficiency. Powder particle size distribution varied significantly within the powders used hence, because no systematic approach to powder distribution was adopted, conclusions on the effect of powder size distribution could not be drawn. It is clear that, as part of optimising a laser deposition application, powder size is one of the key variables. The results from this study will enable a more informed powder size selection decision to be made for applications ranging from repair to cladding.
The direct metal laser deposition (DMLD) technique has shown great potential for the fabrication of solid free form objects and repair of high-value machine tools and turbine components. [1-2] The process has many advantages including: [3-4]
- Low heat input is required when compared to other powder-feed processes, such as thermal spraying. This minimises thermal distortion and results in a small heat-effected zone.
- A wide variety of commercially available powders, developed for other forms of processing, can be used.
- Less than 5% dilution into the substrate can be achieved, therefore it is possible to deposit onto a very thin base material.
- Deposits are metallurgically bonded to the substrate with minimum defects, compared to other additive processes.
The complexity of the process itself however, has hindered wider industrial application. Many parameters govern the process, and consequently affect the quality of the deposit. Therefore, a thorough understanding of the deposition process is imperative. The fundamental variables affecting the process are: 
- Laser power and the focused spot size of the beam.
- Nozzle velocity with respect to the deposit, also known as feed rate or processing speed.
- Nozzle alignment with respect to the machine tool centre point (TCP).
- Mass flow-rate of powder delivered to the work-piece.
- The interaction of the powder, laser beam and substrate.
- The powder particle size.
An initial investigation into the influence of average powder particle size on the deposition quality, surface roughness and powder deposition efficiency is reported here.
2.0 Equipment and materials
Deposition was carried out using a Trumpf DMD505 laser deposition system. This comprises a 2kW CO 2 laser and has a single cantilever 5-axis Cartesian gantry system with a processing envelope of 2m wide, 1.1m deep and 0.75m height. It has a CNC control system, a powder feed system, a powder delivery nozzle, and a patented feedback and control system for producing layers of constant height. Figure 1 shows the system installed at TWI.
Fig.1. The Trumpf DMD505 Laser DMLD system installed at TWI
The system uses an integrated process control system to maximise the quality of the deposit. The two components of this system are active height control, and a facility to link laser power and tool feed rate. For this project, only the latter component, known as 'Tip Control', was employed. This permitted the reduction of laser power during the speed ramp-up and ramp-down, associated with gantry movement during starting and stopping.
For this experiment, a high precision coaxial nozzle was used. The nozzle is predominantly made from a copper alloy. Within a coaxial nozzle, the focusing laser beam passes through the central axis of the head and powder is fedco-axially through an annulus, around the outside of the cone of the focusing beam. Figure 2 shows the capability of this nozzle to produce a fine focus in the powder stream, with little extraneous powder outside the cone of the focus. During deposition the chosen laser beam waist diameter was positioned at the position of the powder focus and this point was also coincident with the TCP of the deposition machine.
Fig.2. The coaxial powder nozzle delivering powder to a focus of approximately 0.5mm diameter
The powder used in this study was Micro-Melt ® Inconel 625 Nickel alloy. Inconel 625 was selected because this alloy widely used in laser deposition for turbine blade repairs, demonstrating high strength and toughness up to 1093°C.
Five samples of powder, with a range of particle sizes commonly used in the DMLD process, were chosen for this study and were labelled as A to E respectively, see Table 1. A is the largest powder through to E being the smallest. The mesh size equivalent to the particle size in micrometers is also presented. Prior to deposition the powder samples were fully characterised for chemistry and particle size distribution. Chemical analysis was carried out using the LECO analytical technique and powder size distribution was determined using a Malvern Matersizer Microplus laser diffractometer.
Table 1 Powder particle size in the Inconel 625 samples
|Element||Sample A||Sample B||Sample C||Sample D||Sample E|
|Particle size range
A toolpath designed to produce a rectangular block was used in this experiment to make test samples. The dimensions of the test blocks were 20 mm long, 10 mm wide and 3 mm tall. To introduce uniform heat input throughout the deposit and ensure deposit quality, the toolpath was varied so that each layer was deposited at right angles to the previous layer. As previous work had shown no benefit in using an oxygen-free chamber, all experiments described in this paper were conducted with localised inert gas shielding only. 
The deposits were subjected to optical microscopy and surface profile analysis using a 3D stylus profile measuring technique (contact measurement). Contact measurement was used because of the surface roughness of the DMLD deposits. A powder deposition efficiency measurement was carried out to determine the percentage of powder that forms the deposit. Powder deposition efficiency was calculated by comparing the powder mass flow rate over the time taken for the deposition process and the volume of the actual deposit.
3.0 Results and discussion
3.1 Process parameters
To allow a comparison between powder samples, a generic set of deposition procedures was used, which were based on the metallographic results that produced the minimum amount of defects in all five samples. Preliminary experiment scarried out for each powder samples are not presented here. Listed below are the process parameters chosen to make the deposits:
- Laser power = 1000W
- Beam diameter at powder focus position = 1.0mm
- Nozzle velocity = 600mm/min
- Powder feed rate = 1.5g/min
- Z increment = 0.16mm
- Track overlap distance = 33% of track width
3.2 Powder morphology
The particle size distribution within each of the samples A to E is presented in Figure 3 and Table 2. The results show that the powder samples have very different size distributions and possess wider size range compared to the size range specified. The 50% value was used as the 'average powder particle size' for this study.
Figures 4 (a) and (b) show SEM images for powder samples A and D, at the same magnification, to compare the observed particle sizes. Figures 5 (a) and (b) show cross-sections of these same powders. The sections show minimal internal porosity in these Inconel 625 powder samples.
Fig.3. The powder size distribution of the samples, analysed using the Malvern Matersizer Microplus laser diffractometer technique
Table 2 Cumulative particle size distribution at 10%, 50% and 90% of the given powder samples. The d(0.5) or 50% value was used as the 'average powder particle size' for this study.
|Element||Sample A||Sample B||Sample C||Sample D||Sample E|
Fig.4. SEM images of
a) sample A and
Fig.5. Cross sections of powder from a) sample A and
b) sample D, showing no significant internal pores
3.3 Powder deposition efficiency
Figure 6 shows the powder deposition efficiency for each powder sample. The highest powder efficiency of 54% was found using sample C.
Analysis of the powder streams exiting the nozzle, showed that when the coarsest powder was used, the powder focus was greater than 1.5mm in width. When the finest powder was used, a very small powder focus of approximately 0.5mmwas obtained. Though a very small powder focus was achieved in the powder with the smallest particle size range, sample E, in this case a low deposition efficiency was found. During observation of the powders exiting the nozzle, it was evident that the powder from sample E appeared to pulse. This could be due to the finest particles in Sample E, which have diameters less than 20 µm, coagulating within the nozzle, due to their large surface area, resulting in the blocking of the nozzle. Figure 7 shows an image of the particle sizes found in sample E. The powder is delivered using a carrier gas, hence when sufficient carrier gas pressure built-up, the powder was forced out of the nozzle. This blocking and unblocking resulted in the pulsing of the powder.
Two obvious effects are influenced the efficiency; Large particle size powders are produce poor powder focus, hence deposition efficiency is reduced because the powder spot size is significantly larger than the laser spot size. If too small a powder size is used, problems in delivering the powder consistently to the substrate can be encountered, which again causes a reduction in powder efficiency. As is indicated by the results in Figure 6, the powder efficiency affects the layer height, because if the process is more efficient, more material is deposited per pass.
Fig.6. Graph showing the deposition efficiency and the layer height achieved for the powders used. A has the largest average powder size, through to E having the smallest
Fig.7. SEM image of the sample E powder. Fine particles of less than 20 µm diameter are present
Figure 8 shows the macro section from a deposited sample using powder C. The total height of the 18-layer sample was approximately 4.4 mm, from which it is possible to calculate an average layer height of 0.24 mm. The average layer height of deposits produced using the other powder samples was lower. For example, the layer height in deposited sample A was 0.08 mm, while in sample D it was 0.19 mm. The deposited layer height is showed in Figure 6.
Fig.8. Cross section of a deposited sample produced using powder sample C
Figures 9 (a) to (e) show cross-sections of deposited samples produced using powder samples A to E, respectively. Uniform deposition tracks can be observed in all cases after etching. No significant solidification defects were found in the deposited samples. A slightly higher level of porosity was observed in samples A and B, which could have originated from inherent porosity within the powder. Small isolated porosity was the only obvious issue with all the deposits, although no significant pores were found in the deposit produced using powder E.
Fig.9. Cross-sections of deposited samples showing low levels of porosity and uniform deposition tracks
a) Sample A
Fig.9 e) Sample E
3.5 Surface topography
Figures 10 (a)-(e) show three-dimensional surface maps of the top surface of the deposits made. The Sa values shown are the 3D surface roughness equivalent to the Ra values, which would be found from a 2D surface. The deposited sample C possessed the lowest Sa value, this could be because the process parameters used were close to optimum.
Deposits A and B produced high Sa values, it is likely that this was due to the large powder particle sizes involved. The powder delivery issues involved with deposits D and E showed up in the analysis below. The Sa value for D was high, approaching that of deposits A and B, because of small interval inconsistencies in the powder delivery. The sample E powder delivery was significantly more inconsistent, hence waviness is evident, when compared to other samples, which can be seen from the red area at the top corner of the surface map in Figure 10(e). Further work is required to determine the effect of processing parameters on surface topography because of the number of key variables involved.
Fig.10. 3D surface roughness maps of the deposited samples produced using autocorrelation analysis
(courtesy University of Huddersfield)
a) Sample A, Sa = 38.17 µm
b) Sample B, Sa = 39.06 µm
c) Sample C, Sa = 24.65 µm
d) Sample D, Sa = 37.79 µm
e) Sample E, Sa = 28.50 µm
The direct metal laser deposition process was used to study the effect of powder particle size on the deposit quality, surface roughness and powder deposition efficiency, using a single set of processing parameters for each powder investigated. Five powder samples of Inconel 625 nickel alloy were compared. The conclusions arising from this study are:
- Average particle size has a direct and indirect effect on the deposit height and efficiency. Smaller particle sizes can be focused more easily resulting in better efficiencies and larger deposits. If the particle size is too small however, problems with powder delivery can occur causing a reduction in deposit height and efficiency.
- In this study, coarse powder produced high surface roughness values, because of the large powder particle sizes involved. Poor Sa values also occurred with the smallest powders where delivery problems occurred.
- No significant solidification defects were found in all five deposited samples. Small isolated porosity was the only obvious issue with all the deposits, although a slightly higher level of porosity was observed in samples containing larger particle sizes (samples A and B).
- G. Bi, A. Gasser, K. Wissenbach, A. Drenker and R. Poprawe: 'Characterization of the process control for the direct laser metallic powder deposition'. Surface and Coatings Technology, v 201, n 6, 2006, p 2676-2683.
- C.P. Paul, P. Ganesh, S.K. Mishra, P. Bhargava, J. Negi and A.K. Nath: 'Investigating laser rapid manufacturing for Inconel-625 components'. Optics & Laser Technology, v 39, n 4, 2007, p 800-805.
- G. Bi, A. Gasser, K. Wissenbach, A. Drenker, R. Poprawe: 'Investigation on the direct laser metallic powder deposition process via temperature measurement'. Applied Surface Science, v 253, 2006, p 1411-1416.
- P.A. Carroll: 'Development of Waspaloy ® deposition procedures'. TWI Report No. 16059, 2005.