Application of iterative path revision technique for laser cutting with controlled fracture

Abstract

Laser cutting using the controlled fracture technique has great potential to be employed for the ceramic substrate machining. The heat produced on the surface of a ceramic substrate by the laser separates the substrate controllably along the moving path of the laser beam. Because the extension of the breaking frontier is lager than the movement of the laser spot, the actual fracture trajectory deviates from the desired trajectory when cutting a curve or cutting an asymmetrical straight line. To eliminate this deviation, the iterative learning control method is introduced to obtain the optimal laser beam movement path. The fracture contour image is grabbed by a CCD camera after laser cutting completion. A new image processing system is proposed to detect the deviation between the desired cutting path and the actual fracture trajectory. The laser-movement path for the next trial can then be determined according to the iterative path revision algorithm. The actual fracture trajectory converging to the desired cutting path is assured after a few path revisions. The experimental materials used in these experiments are alumina ceramics and the laser source is CO₂ laser. The proposed system can achieve a machining precision of about 0.1 mm.

Introduction

The laser cutting technique with controlled fracture proposed by Lumley [1] is very important in machining brittle materials. The applied laser energy produces a mechanical stress that causes the substrate to separate along the moving path of the laser beam. This material separation is similar to a crack extension with controllable fracture growth. Lumley [1] successfully used this technique to cut brittle materials such as alumina ceramic substrate and glass using a CO₂ laser. The required laser power is less than that for conventional laser evaporative cutting, or laser scribing, and the cutting speed is much higher.

Grove et al. [2] proposed a related controlled fracture method for cutting glass that used a higher cutting speed. In recent years, Kondratenko [3] and Unger and Wittenbecher [4] further investigated using a low power laser to separate glass with an additional coolant, water, which produced tensile stress along the cutting path. This improvement in the controlled fracture method has been recognized as providing good prospects for future laser cutting applications.

Tsai and Liou [5] gave an account of the fracture mechanism explaining why the material separation is controllable. They mentioned that the cutting process could be divided into three stages. The first is the initiation stage, the fracture is initiated due to the tensile stress at the specimen edge. The second is the stable growth stage where the stress near the laser spot is highly compressive. After laser passes, the plastic compressive stress is relaxed, and then induces a residual tensile stress, that makes the fracture grow from upper to lower substrate surfaces. The final stage is the unstable fracture. The stress near the crack tip is entirely tensile stress along the thickness direction making the crack extend unstably.

In the experimental results of Tsai and Liou [5], the crack tip leaves behind the laser spot and the distance between the crack tip to the laser spot is variable during the cutting process. Because the propagation of the crack tip at the cutting frontier is lager than the movement of the laser spot, the fracture trajectory deviates from the desired trajectory for cutting a curve or cutting an asymmetrical straight line. To ensure accurate splitting, the stress field must be symmetrical with the crack. For most cutting conditions the cutting path is not symmetrical and the crack extension trajectory deviates from the desired cutting path.

To diminish the deviation between the actual fracture trajectory and the desired cutting path, Tsai and Liou [6] proposed an on-line crack detection system to revise the moving path of the laser beam. During the cutting process of controlled fracture, the crack tip of the separation frontier is detected on-line using an image processing system. The distance measurement between the crack tip and the laser beam can be accomplished from the analysis of crack image processing. From this distance, the motion path compensation of the laser beam can be determined.

The method proposed by Tsai and Liou [6] must be performed within the proper environment. The bright burning due to the laser heat would influence the crack detection. Because there is a delay in on-line crack detection and the cutting speed is highly restricted, this method is not suitable for mass production. In this paper, an alternative method for eliminating the deviation between the actual fracture trajectory and the given desired cutting path is proposed. The iterative path revision technique is introduced to off-line revision of the applied laser path. The idea of path revision is based on the iterative learning control theory [7] that has been successfully applied in the trajectory control of robots.

In this paper, an image processing technique using clip operation for the contour measurement is proposed to evaluate the cutting deviation. After the first cutting, the fracture contour is measured by the image processing system. The laser path for the next operation is then revised according to the iterative learning control method. The final revised laser path can be obtained as the actual fracture trajectory converges to the desired cutting path. The final revised laser path is the optimal path that can be used in mass production.