Multi-kW Laser Cladding using Cylindrical Collimators and Square-formed Fibers

Transcription

1 Multi-kW Laser Cladding using Cylindrical Collimators and Square-formed Fibers SPIE Photonics West conference in San Francisco, January Submitted version - Mats Blomqvist, Stuart Campbell, Jyrki Latokartano, Jari Tuominen OPTOSKAND TAMPERE UNIVERSITY OF TECHNOLOGY

2 ABSTRACT In industrial laser cladding applications various new possibilities have opened up by introduction of laser sources with powers over 10 kw. Higher laser power allows higher deposition rates, which enables new applications for example in heavy engineering. However, to fully utilize the high power, beam area in focus needs to be increased significantly compared to for example welding. For high brightness lasers, this often requires complicated processing optics as the beam is usually Gaussian when defocused. In most surface treatment applications process would benefit from homogenous intensity distribution instead of a Gaussian one. In this paper we present ideas for cladding applications using a 12 kw disc laser coupled into a square-formed fiber with a 1000x1000 µm-core. The output of the fiber is collimated by a newly developed collimator based on cylindrical lenses with an 1:3.3 aspect ratio of focal lengths. The asymmetrically collimated beam is then condensed to a homogeneous rectangular spot on the work-piece using an f=500 mm focusing unit. With this setup we reach a spot size of 7.4x2.2 mm = 16.3 mm 2, implying laser power densities up to 740 W/mm 2. The asymmetric collimator is based on efficiently water-cooled cylindrical lenses with different focal lengths. Having interchangeable fiber connector interfaces and Optoskand s standard exit interface, the collimator can easily be implemented in optical heads. We present results on the optics performance including power transmission, image quality and focal shifts at power levels up to 12 kw. Results of preliminary cladding tests using the asymmetrical optics and off-axis tandem wire feeding will also be presented orally. Deposition rate and efficiency using high power levels will be investigated. Analyses of cladding bead geometry and microstructure will be performed. Keywords: high power lasers, cylindrical collimator, focal shift, laser cladding, processing optics mats.blomqvist@optoskand.se; phone ; fax ; 1

3 1 INTRODUCTION The technique of industrial laser cladding is becoming increasingly important in applications dealing with wear and corrosion of components. By depositing corrosion resistant super alloys, such as inconels (austenitic nickel-chromiumbased super alloys) and stellites (cobalt-chromium hardfacing alloys) on more common metal construction materials, such as steel, an inexpensive high-performing component can be manufactured. With the introduction of laser sources with output power over 10 kw new possibilities have arisen. Higher laser power allows higher deposition rates, either depositing at higher speed or larger areas. This enables new applications for example in heavy engineering. Similarly in laser hardening, where for example hardenable steel is heat treated just under its melting point to create a selfquenched martensitic layer with great hardness, larger areas with consequently higher throughputs can be addressed. When high power disc or fiber lasers are used for surface treatments the beam needs to be expanded considerably. There are various methods achieve this: 1. Defocus the beam with ordinary welding optics. 2. Use homogenizing optics or mirrors. 3. Use oscillating scanners. Option 1 is usually not suitable because defocused beam resembles Gaussian distribution. Options 2 and 3 can produce preferred homogeneous intensity distribution but may be complicated and costly to realize. While top-hat intensity distribution is beneficial, cladding applications may also benefit from rectangular spot. Wire tip positioning, cladding bead height/width ratio and dilution effects should benefit from rectangular beam shape. In hardening applications rectangular intensity distribution gives clear benefits, when trying to minimize the annealed zone between parallel hardening tracks. For all surface treatment applications the long focal length is usually also beneficial. Longer working distance helps minimizing the effects of contamination of process optics. Material feeding is usually also less complicated with longer focal distances. There are various ways of achieving a rectangular spot in focus. Using standard circular-core fibers the process optics usually includes a homogenizing optical element with micro-lens arrays. Advantages with such components are the flexibility to change the shape in focus, but the systems are complex and thus costly. For hardening applications one can also use scanning optics that moves the round focus back and forth very rapidly. Alternatively the use of rectangular shaped fiber cores will also create a rectangular focus using standard collimating and focusing optics. The limitation is the aspect ratio which will be fixed for a specific rectangular shape fiber design. We propose a simple, flexible and rugged design. Using a square-formed fiber in combination with an asymmetric collimator based on two efficiently water-cooled cylindrical lenses with different focal lengths, the size of the rectangular spot can be designed using focusing optics with different focal lengths and the aspect ratio can be varied by changing the collimator instead of the complete optical fiber. The limited number of optical elements will vouch for stable, low-loss transmission with low focal shifts. In this paper we present results from applications using a 12 kw disc laser coupled into a square-formed fiber with a 1000x1000 µm-core. The output of the fiber is collimated by a newly developed collimator based on cylindrical lenses with a 1:3.3 focal length aspect ratio. The asymmetrically collimated beam is then condensed to the work piece using an f=500 mm focusing unit. With this setup we reach a spot size of 7.4x2.2 mm = 16.3 mm 2, implying laser power densities up to 740 W/mm 2. 2 CYLINDRICAL COLLIMATOR DESIGN In this section the water-cooled collimator design based on cylindrical lenses will be described. Adjustments and optical limits of the design will also be addressed. 2.1 Optomechanical design To transform the output beam from a square-formed fiber to a rectangular focus, the basic idea is to use two orthogonally mounted cylindrical lenses with different focal lengths in one collimator. Mounting the lenses at focal length distance from the fiber end plane a collimated beam of asymmetric profile will exit the optics. Combining this collimator with a standard spherical focusing lens a rectangular focus will be attained, where the aspect ratio of the 2

4 two sides of the rectangle depends on the relation between the focal lengths of the two cylindrical lenses. Below, the lens with shorter focal length will be denoted f 1 and the lens with the longer focal length will be denoted f 2. To handle power levels exceeding 10 kw with potential strong back-reflections from the work-piece, efficient water cooling has to be applied. By absorbing scattered light and controlling the temperature of the lens holders thermal lensing effects are minimized. Since there is a tolerance of ±100 mrad on the rotation of a square-formed fiber core relative to the fiber connector mount it is necessary to rotationally align the fiber to the collimating lenses. This adjustment is performed by rotating the fiber connector mount as the specific fiber is connected. By using the guide laser the adjustment can be safely made before turning on the high-power laser. Figure 1. Section view of a water-cooled collimator with cylindrical lenses giving an aspect ratio of 1:3.3. In Fig. 1. a section view of the newly developed collimator is shown. It has an aspect ratio of 1:3.3 using an 30x30 mm f 1 =68 mm cylindrical lens as first collimating lens and an f 2 =226 mm cylindrical lens (octagonal shape based on a 50x50 mm lens) as second collimating lens. Both lenses are of the plano-convex singlet type made of fused silica with antireflective coating. The lenses are mounted in holders where water cooling is applied to a stainless steel housing surrounding the holders. The exit interface follows Optoskand s standard M58x1 thread interface for pre-aligned optics. On the input side the fiber connector can be rotated to align the fiber with respect to the optics. The fiber connector can be replaced between QB, QD (LLK-D) and Q5 (LLK-B) interfaces. 2.2 Aspect ratios and design limits As mention above, the aspect ratio is determined by the relation between the focal lengths of the two cylindrical lenses. When designing the optics there are a few parameters to keep in mind. Having a maximum lens diameter set by the mechanical design, normally the second collimating lens (f 2 ) limits the acceptance numerical aperture (NA acc ) the collimator can handle. On the contrary the focal length of the first collimating lens (f 1 ) cannot be too short because of the necessary free space of the fiber connector in front of the fiber end plane. Also having too large curvature of f 1 might deteriorate the imaging properties of the collimator. The combination of these two design limitations will put a limit to the possible aspect ratios and also to the magnification and working distance of the complete optics with focusing unit and process adapters. In this paper two collimators have been evaluated; a 1:3.3 collimator (f 1 =68/f 2 =226) and a 1:2.0 collimator (f 1 =113/f 2 =226). The collimators have the QB or QD (LLK-D) fiber connector interface. Based on the clear aperture of the f 2 =226 mm lens, NA acc =0.10. This might be limiting for a range of lasers especially when using square-formed fibers with large cores (e.g. 1000x1000 µm) that results in a higher output NA. Thus in the latest design the second cylindrical lens has been changed to f 2 =170 mm giving an NA acc =0.13. With a new, shorter fiber connector mount and f 1 -lenses of focal lengths 57 mm, 68 mm, 85 mm, and 113 mm a set of aspect ratios (1:3.0, 1:2.5, 1:2.0, and 1:1.5) are available. 3

5 3 EXPERIMENTAL The experiments are divided into two parts. In the first part the collimators were evaluated with respect to power handling capability and power transmission. Astigmatism, image quality and focal shifts at power levels up to 12 kw were measured. In the second part hardening and cladding tests were performed. 3.1 Experimental Setup Two different experimental configurations were used. For the astigmatism measurements and initial evaluations a fiber laser (IPG YLR-4000-SS) with 4 kw maximum output power at approximately λ = 1070 nm was used as beam source. The output fiber was connected via a fiber-to-fiber optic switch and a process fiber to the measurement object. Using a 600x600 µm core process fiber and an f=120 mm focusing lens in combination with the two collimators a small enough spot sizes with reasonably short Rayleigh length were achieved to measure the beam parameters accurately with a Primes instrument. Figure 2. Process head mounted vertically to evaluate the beam quality using a Primes instrument. The same configuration was used for the process tests. For most of the measurements and for the process tests a 12 kw cw disc laser (Trumpf TruDisc 16002) was used. The laser also emits light at λ = 1070 nm. Two different process fibers were used a 400 µm core Trumpf LLK-D fiber cable and a 1000x1000 µm core Optoskand QD fiber cable. The collimators under test were mounted on a robotic arm and an Optoskand f=500 mm focusing unit was applied, see Fig. 2. The beam parameters were measured with a Primes FocusMonitor. A typical caustic result at a fixed power level is shown in Fig. 3. The caustics are presented using the 2 nd moment s algorithm. The z-position and the Rayleigh length are used in the calculations of the normalized System Focal Shift Factor (SFSF). 1,2 For a square-formed fiber it is also possible to estimate how astigmatic this beam is, that is how much the z-position of the focal planes differs between the two orthogonal axes of the rectangular shape image. 4

6 To monitor component heating during process a Flir ThermaCAM E45 was used. Also ΔT and flow rates of the cooling water of the collimator were measured. Using the Optoskand QD fiber cable with integrated sensors it is also possible to monitor back-reflected light and ΔT of the cooling water of the fiber connector. Figure 3. The beam parameters for the f113/f226 mm collimating lens and f=500 mm focusing lens combination at 12 kw using a 400-µm-core fiber. The x-axis corresponds to the narrower width as a result of the f=226 mm collimating lens. The z-position and the Rayleigh length of the x-axis are used in the calculations of the SFSF. To evaluate the performance of the developed optics a simple cladding setup was built. As the beam is rectangular it was possible to test so called tandem wire setup. Two 1.2 mm Inconel 625 wires can be either leading or trailing the beam. Wires were fed with two synchronized fire feeding units controlled by a Fronius Time Digital In cladding, the beam is travelling in the direction of the shorter side of the beam as illustrated in Fig. 4. Figure 4. Tandem wire feeding in trailing setup with illustration of the focused beam (image from below). As the beam is rectangular in shape and has an even intensity distribution, it was assumed that the positioning of the wire tips against beam position should not be as critical as with Gaussian beams. Using two wires with some distance between them should in theory also generate good cladding bead shapes when considering height/width ratio. Lower bead height is beneficial as the cladding material is usually expensive. Thicker cladding means also more machining in the final product which increases both the costs and the amount of wasted material. The employed long focal length of the focusing optics (f=500 mm) is beneficial in leading wire setup. This feeding position is usually preferred by most authors 3 but may be difficult to realize as the wire feeding angle against the substrate should be rather small. The setup illustrated above should enable small enough wire angles for smooth process. Used setup is rather simple and does not have too many adjustable parameters. For optimal cladding results some adjustments may need to be done. However the main focus in these tests was the process performance of the developed optical system and hence these limitations of the wire feeding system were not considered too restrictive. 5

7 4 RESULTS In this section experimental results will be presented. For the hardening and cladding test only preliminary hardening results will be presented the cladding results will be available for the oral presentation. 4.1 Power handling and beam quality The initial evaluation of the new collimators was performed using a 600x600 µm core fiber and an f=120 mm focusing lens. After rotationally aligning the fiber connector mount with respect to the collimating lenses a uniform flat-top rectangular shape image with sharp edges was achieved. By evaluating the 2 nd moment s values of the caustic using the built in algorithm in the Primes software, the beam parameters could be determined for the two orthogonal axes, see Fig. 4. The difference in z-position of the x- and y-axes was determined to 0.4 mm, which is significantly below the Rayleigh length. Because of the difference in focal length of the collimating lenses the Rayleigh length has to be calculated for the two orthogonal axes instead of taking the average value shown in Fig. 5. For the test shown in the figure (1:2.0 collimator), these values are Z Rx = 5.4 mm and Z Ry = 1.9 mm. Up to 3.5 kw only very small focal shifts were measured (SFSF < -0.1). Figure 5. The beam parameters for the f113/f226 mm collimating lens and f=120 mm focusing lens combination at 3.5 kw using a 600x600 µm core fiber. The y-axis corresponds to the narrower width as a result of the f=226 mm collimating lens. The z-positions of the x- and y-axes are used in the astigmatism evaluation of the optics. The first part of the high-power evaluation using the 12 kw disc laser a round 400 µm core fiber was used in combination with an f=500 mm focusing unit. With this fiber the image will not be as well defined as for a squareformed fiber. The resulting ellipsoid will distort as one moves outside the focus as shown in the 3D representation of the Primes measurement in Fig. 6 (1:2.0 collimator). The power was increased in 2-kW steps up to 12 kw. The laser beam was terminated in a beam dump without possibility to measure transmitted power. Instead the optics was monitored using a thermal camera and ΔT logging of the cooling water connected to the collimator. At 12 kw, ΔT was stable at approximately 1.0 K, which with a water flow of 0.6 l/min implies power absorption in the collimator P ΔTmcv = = kw W, (1) Δt 60 abs 40 where the specific heat capacity of water is c v = 4.19 kj/kg K. 6

8 Figure 6. 3D representation of the Primes measurement at 12 kw for the f113/f226 mm collimating lens and f=500 mm focusing lens combination using a 400-µm-core fiber. The collimator was at room temperature except for the fiber connector mount that heated up to approximately 60 C. This heating is mainly attributed to absorption inside the fiber connector mount due to scattered light. In the next generation of collimators (Sec. 2.2) the connector mount is more integrated into the collimator housing where it is better cooled. Also the air-cooled focusing unit heated up to about 50 C. This heating is attributed to scattered light absorption as the collimated beam is very close to the clear aperture of the lens. Based on the Primes caustic measurements thermal shift and calculated normalized SFSF could be determined for the axis corresponding to the f=226 mm collimating lens, which has the smallest magnification and thus the shortest Rayleigh length. The result is shown in Fig.7., where the 2 kw measurement was the lowest power that gave repeatable results. In principle one should normalize the measurement at zero power, but from the results using the fiber laser it is expected that the thermal shifts up to 2 kw corresponds to an SFSF < The negative thermal shift implies a shorter focal length following the definition of focal shifts. 1 At 12 kw the SFSF is -0.5, well below SFSF of 1 which indicates a change in the spot size by approximately 40 %. As rule of thumb this is the maximum change acceptable for an application. One would expect the change in system focal shift factor to be linear, which is not the case in Fig.7. This is explained by the long time constant of the focal shifts. The power should be kept constant at each power level for at least 10 minutes, which was not the case for the lower power levels but for the 12 kw measurements. 7

9 Figure 7. Thermal shift and calculated normalized SFSF for the f113/f226 mm collimating lens and f=500 mm focusing lens combination using a 400-µm-core fiber. The values are calculated for the axis corresponding to the f=226 mm collimating lens, which has the smallest magnification and also the shortest Rayleigh length. Next the process fiber was changed to a square-formed 1000x1000 µm core fiber. Compared to the 400 µm core fiber, a minor increase in temperature was noted on the focusing unit. This is explained by a small increase in output angle using the square-formed fiber making the beam diameter very close to the clear aperture of the focusing unit. The larger core size gives a larger image in focus and also a much larger Rayleigh length. A 3D representation of the Primes measurement at 12 kw for the f68/f226 mm collimator (1:3.3) and f=500 mm focusing lens is shown in Fig. 8. and the focal plane is shown in Fig. 9. The spot size is approximately 7.4x2.2 mm 2 = 16.3 mm 2, indicating laser power density around 740 W/mm 2. The power density profile is relatively uniform over at least 20 mm in the z-direction vouching for stable process performance. 8

10 Figure 8. 3D representation of the Primes measurement at 12 kw for the f68/f226 mm collimating lens and f=500 mm focusing lens combination using a 1000x1000-µm-core fiber. 4.2 Hardening and cladding tests Up until now only initial hardening tests have been performed using the collimated beam without focusing optics. Results from hardening and cladding tests using the setup describe in Sec will be presented orally. Figure 9. Focal plane 12 kw for the f68/f226 mm collimating lens and f=500 mm focusing lens combination using a 1000x1000-µm-core fiber. The beam size is approximately 7.4x2.2 mm 2. 9

11 5 CONCLUSIONS In conclusion, a simple, flexible and rugged design to create a rectangular shape beam focus for multi-kw laser cladding is proposed and demonstrated. Using a square-formed fiber in combination with an asymmetric collimator based on two efficiently water-cooled cylindrical lenses with different focal lengths, the size of the rectangular spot can be designed using focusing optics with different focal lengths and the aspect ratio can be varied by changing the collimator. The limited number of optical elements vouches for stable, low-loss transmission with low focal shifts. Using a disc laser at 12 kw together with a 1:3.3 aspect ratio collimator, a steady and homogeneous spot of 7.4x2.2 mm = 16.3 mm 2 was achieved implying with laser power density of 740 W/mm 2. At 12 kw the SFSF is -0.5 indicating only small focal shifts. Acknowledgements: This work has been performed within the LIFT project that has received funding from the European Community's Seventh Framework Program FP7-NMP under grant agreement n CP-IP REFERENCES [1] Blomster, O., Pålsson, M., Roos, S-O., Blomqvist, M., Abt, F., Dausinger, F., Deininger, C., and Huonker, M., Optics performance at high-power levels, Proc. SPIE 6871, 68712B-68712B-10 (2008). [2] Blomqvist, M., Blomster, O., Pålsson, M., Campbell, S., Becker, F., Rath, W., All-in-quartz optics for low focal shifts, Proc. SPIE 7912, (2011). [3] Nurminen, J., Hot-Wire Laser Cladding, Doctoral Thesis - Tampere University of Technology, publication 765, ISBN (2008). 10

12 Article authors: Mats Blomqvist*, Stuart Campbell Optoskand AB Aminogatan 30, SE Mölndal, Sweden * mats.blomqvist@optoskand.se; phone ; Jyrki Latokartano a, Jari Tuominen b Tampere University of Technology Tampere University of Technology, Korkeakoulunkatu 6, FI Tampere, Finland a Department of Production Engineering b Department of Materials Science Authors: Mats Blomqvist, Stuart Campbell, Jyrki Latokartano, Jari Tuominen Multi-kW Laser Cladding using Cylindrical Collimators and Square-formed Fibers, Proceedings of SPIE (2012) Copyright 2012 Society of Photo-Optical Instrumentation Engineers. One print or electronic copy may be made for personal use only. Systematic reproduction and distribution, duplication of any material in this paper for a fee or for commercial purposes, or modification of the content of the paper are prohibited.