Surfi-Sculpt®..... yes, but using lasers

TWI Bulletin, January - February 2009

The power beam process Surfi-Sculpt^® was invented at TWI several years ago....using electron beam technology and has now been demonstrated using lasers

As Paul Hilton reports the Surfi-Sculpt technique enables controlled surface features to be produced on a range of substrates such as metals, polymers and ceramics. Such surface features were first demonstrated using electron beams, using electromagnetic coils first to focus the beam and then deflect this focused beam over the material surface in a rapid and controlled manner.

The material melted by the beam moves, in part, due to the surface tension generated by a temperature gradient created across the molten material surface. How the material moves, and ultimately the shape of the features produced, can be determined by precisely controlling the beam path and speed over the surface.

This article describes a laser variant of the Surfi-Sculpt process that uses optical power to melt and displace material, thereby creating the surface feature. High brightness fibre delivered laser beams, using both fibre and disclasers as the source of the laser light, have been used. Because of their high beam qualities, these lasers have the advantage that their beams can be focused to small spots of high power density, whilst still using a beam focusing lens of long focal length.

This large distance between the focusing lens and the focused spot is a necessity for the laser Surfi-Sculpt process, because the rapid manipulation of the laser beam required is achieved by orthogonally mounted and galvanometer driven beam scanning mirrors. In the work described, both disc and fibre lasers have been used at relatively modest laser powers of less than 2kW, in conjunction with two different and commercially available laser beam scanning systems, capable of handling these power levels and developed primarily with laser welding applications in the automotive industry in mind.

Objectives

As laser Surfi-Sculpt is a completely new process the objectives of the work were simply to demonstrate production of Surfi-Sculpt process features using laser beams.

Experimental approach

High brightness lasers at both IPG Photonics and Trumpf GmbH were used in this work. The initial experiments at IPG were performed with an IPG YLR-2000 laser. The laser beam spot used was estimated at 330µm in diameter at the beam waist, which was always positioned on the workpiece surface. The laser beam was manipulated using a 'Fibre Elephant' scanning head (manufactured by Arges Lasertechnik, Germany), operating, for this work, in the sweep speed range of 8-20m/min.

The scanning head was mounted on a Kuka robot arm which was stationary during the trials. The beam quality of the IPG laser was such that even after passing the beam through the scanning optics, a distance of about 500mm was created between the exit of the optics and the sample. At IPG the only material investigated was titanium, and due to time constraints, all processing was performed in air. A later series of experiments was subsequently performed at Trumpf with a TruDisk 8002 disc laser. The laser beam spot used was estimated at 340µm in diameter at the beam waist, which was again positioned on the sample surface. In this case the working distance between the output of the opticalsystem and the sample was 290mm. The laser beam was manipulated using a Trumpf PFO scanning head, operating, for this work, in the sweep speed range of 8-24m/min.

The Trumpf scanning head was also mounted on a Kuka robot arm, which was again kept still during the trials. During the work at Trumpf, to investigate the effect of an inert gas environment better or indeed a vacuum, on the formation of laser Surfi-Sculpt process features, a small vacuum chamber was used. The sample was located inside the chamber and the laser beam interacted with the sample through a quartz window at the top of the chamber.

The chamber had two outlets, one to an argon gas line and one to a rotary vacuum pump, allowing the chamber to be evacuated or filled with Argon. A final set of experiments was conducted with the samples, which now included stainless steel, mild steel, Inconel, and aluminium, as well as the original titanium, either shielded with argon gas (sitting in a small chamber) or in air. At both IPG and Trumpf, the paths described by the beams were programmed using bespoke software packages developed for the 'remote' welding applications described earlier. While these were not ideal for the repetitive and cyclic nature of the required patterns, both systems were easy to use by virtue of their advanced graphical interfaces.

Results and discussion

In the first series of experiments with the fibre laser, after confirming that a linear Surfi-Sculpt feature was possible, an attempt was made to construct a more complex feature from a series of radially symmetric linear swipes ofthe laser beam, based on the pattern shown in Fig.1, with the hope of producing a conical feature on the titanium plate. The sequence consisted of 60 repeats of eight radially sequential single swipes. These results were obtained in air and did not use the vacuum normally required for EB work. The resulting feature and the processing conditions are given Fig.2. In this case the height of the protrusion was about 5mm, and its manufacturing time was about five seconds using a laser power of 1kW. Although this feature was heavily oxidised, it was noticeably smoother than the initial linear swipes and showed less spatter.

Fig.1. Scan pattern used to produce a star shaped feature. The arrows in the diagram indicate the direction of the movement of the laser beam, sequentially from arm to arm

Fig.2. SEM images of a conical laser Surfi-Sculpt feature created in air. The parameters were: laser power, 1kW, scanning speed 16m/min, number of swipes per arm, 60, time delay between swipes, 0.5ms. The total time required to create this feature was approximately 5s

During the second set of processing trials, a major objective was to study the effects of atmospheric condition in the processing region using the small hermetic chamber described earlier. The same programmed laser beam path as that shown in Figure 1, was used to create features in a 100% argon atmosphere and in vacuum, as well as in air, all on the same titanium plate. The process parameters were kept the same as those shown in the caption to Figure2, but only ten swipes per arm of the star shape were used. The laser beam was switched off between swipes, the only delay encountered being that produced by the scanner moving to its new start position, which was difficult to estimate for the equipment used. The results of this experiment showed a far greater degree of dependence on the processing atmosphere than had been expected.

In air the feature size was the largest but the feature did not have a smooth surface and showed evidence of spatter and significant oxidation with some heat marking of the substrate surface. In vacuum, the feature appeared muchsmoother and with no evidence of oxidation but the feature size was much reduced from that made in air.

In a 100% argon atmosphere, there was practically no heat marking at all, no oxidation, no visible spatter, and the feature produced very closely represented the actual programmed path of the laser beam, in that the 'gap' between the linear swipes in Figure 1, was quite obvious, which was not the case for either the result in air or in vacuum. These effects are probably associated with the thermal conductivity of the processing environment used and how this has affected the viscosity, movement and solidification of the material during the process.

Following on from this work a second set of trials was conducted at Trumpf, this time with the objective of optimising further the process on titanium with inert gas shielding, to look at other materials and to compare this work with results produced in air. To do this, firstly, star shaped features were produced, using parameters very similar to those established earlier, on both commercially pure titanium and Ti-6Al-4V, in a gaseous argon environment, (asfor all the experiments hereafter reported, unless otherwise stated), with both a 'zero' separation between the arms, as well as with the small gap between the arms, as depicted in Figure 1. No obvious differences could be seen in the resulting features due to material composition. Figure 3 shows an SEM image of the former feature, showing the high degree of uniformity of the surface, which was unoxidised and showed little spatter. In developing this process greater attention was paid to thermal management of the laser beam energy incident on a particular feature. This process involved the introduction of slight delays between beam swipes, both in terms of the swipe pattern per arm and the beam progression from arm to arm. This approach was seen to maintain build up of the feature, by providing the necessary time for beneficial solidification of the molten material.

Fig.3. Star shaped feature produced using a short delay between swipes to enhance the management of heat input into the sample (compare with results shown in Figure 2)

a) SEM image of the whole feature

b) Macrosection through the peak

The feature shown in Figure 3, for example, was produced by programming a delay between the swipes of half a second. This, of course, significantly increases the time to make a feature but when producing an array of features, this 'delay' can be used effectively to process an adjacent feature. Indeed, the same technique is used in the production of electron beam induced Surfi-Sculpt features. Figure 4a shows an SEM image of the feature produced using a scan pattern with a distance of 0.5mm programmed between each pair of arms. Once again, very uniform features were produced. Of particular interest is the shape of the top of the central cavity produced in this feature, which is clearly octagonal in shape, consistent with the eight arm pattern. Figure 4b shows this more clearly; the length of each section of the octagon formed is about 200µm in length, which can be compared to the size of the focused laser spot used, which was over 300µm in diameter.

Fig.4a) Shows an SEM image of the same features as shown in Figure 3, but with a distance of 0.5mm programmed between each pair of arms

Fig.4b) Shows a magnified view of the octagonal shaped central hole

Figure 5 shows a series of features made using an alternative pattern involving 16 arms instead of the eight used previously, in Ti-6Al-4V. As can be seen, increasing the number of arms in the pattern increased the smoothness of the resulting feature and increasing laser power increased it's height.

Fig.5a) Shows a series of features made in Ti-6A1-4V with the following parameters: laser power between 750W and 2kW, 16 arms to the pattern, arm length 3mm, 20 beam swipes per arm, (in sets of 16), 16m/min swipe speed and a 0.25s dwell between swipes. There was also no gap between the programmed arms of the feature. The largest feature is 3mm in height.

Fig.5b) Shows close up views of the sample from Figure 5a, extreme right, made at 750W

A 'standard' set of parameters, derived for the feature with eight arms and no gap between the arms, was then used on a variety of other materials, both under the argon gas shield and in air. The results can be seen in Figure6. In all cases other than the 7000 series aluminium, a similar feature to that produced in titanium was obtained. In air, clearly a lot of oxidation was involved, and this tended to increase the height of the feature formed. For aluminium, at a laser power of 1kW, it was felt that this might not be sufficient to couple into the material and the pattern was repeated at a power of 2kW. At this power the process worked and the projection heights recorded were1.3mm in argon but 4.9mm in air, indicating increased coupling on an oxidised sample.

Fig.6. Shows the effects of using the standard set of process parameters for the 8 armed star on,

a) 7000 series aluminium

b) C-Mn steel

c) Inconel 718 and

d) 304 stainless steel. The oxidised features (right hand side) were made in air

Finally, array patterns of both eight and 16 armed features were produced in an argon atmosphere, for demonstration purposes, in Ti-6Al-4V. The eight armed feature was also used to produce an array of features in the 7000 series aluminium alloy, again in an argon atmosphere. These can be seen in Figure 7.

Fig.7. Shows photographs of arrays of features produced in

a) titanium, and

b) aluminium

Main conclusions and recommendations

The work conducted in this project, using high brightness laser sources and commercially available laser beam scanning systems and associated software, has allowed demonstration of the laser Surfi-Sculpt process. In particular, Surfi-Sculpt process features have been produced on a range of metals including titanium, stainless steel, aluminium and Inconel. While these features resembled those produced using electron beams, there are apparent differences when the processes are compared, particularly with respect to the effective beam scan speeds used. Scan speeds used with the laser process were significantly slower than those used with the EB process.

As for the electron beam process, management of the heat input to a particular feature, when producing both single and arrays of features, was found to be very important it would benefit from modifications to existing tool path generation software in order to make the laser process more commercially viable. Notwithstanding the scanning speed restrictions in the laser systems used to date, the ability to carry out the laser Surfi-Sculpt process variant with local gas shielding on any size of component is significant and may open up new application opportunities.

The fibre laser, with its compact size, availability over a very wide range of laser powers and very high beam quality, would appear to be an excellent candidate for laser Surfi-Sculpt applications. It is expected that scanning systems, using similar geometries to those used in this work, but producing focused spot sizes down to 0.13mm in diameter, will soon be available, using single mode fibre lasers. The results presented here were generated from a limited amount of experimental work using the two laser sources employed. Notwithstanding this, the potential of the laser for applications of the Surfi-Sculpt process has been effectively demonstrated on several metallic alloys and all the necessary equipment to perform the process is commercially available.