Acoustic emission monitoring in the lapping process.

Dobrescu, Tiberiu ; Dorin, Alexandru ; Jiga, Gabriel 等

1. INTRODUCTION

Lapping is a loose abrasive process used in manufacturing many precision components. One major application of the lapping process is the machining of the air bearing surface of the read-write head for magnetic disk data storage devices. A root-mean-square surface roughness of 5 nm or better is commonly required for the air bearing surface. Small abrasive sizes (typically, 0.3 [micro]m) are used without re-supply to ensure slow material removal rate (MRR) and good surface roughness characteristics. Because of a lack of a detailed understanding of the process mechanisms, fine-tuning or process development for a new product has always been an empirical process with success dependent upon the skill of the machine operator and engineer (Inasaki et al., 1993).

The investigations performed in this study are focused on two parts of the lapping problem. First the mechanisms involved in the lapping process were studied, with considerations of the effect of machining time. Acoustic emission (AE) sensors are utilized here as a sensitive feedback technique for process modeling and investigation. Acoustic emission sensors detect the surface elastic stress waves generated by energy released during plastic deformation and brittle fracture in the material. Since the elastic waves can be transmitted through a variety of media, acoustic emission sensing offers great flexibility in sensor location and sensitivity, especially for cases with small material removal rate. The second part of this study is to investigate the correlation between acoustic emission signals and the material removal rate mechanisms. This correlation can be used, in the future, as a basis for in-process monitoring for lapping process control.

2. EXPERIMENTAL SETUP

Figure 1 shows the experimental setup used for lapping on hand polishing table. A removable lap plate (76.2 mm diameter) was used to facilitate observations of the plate surface. To control the surface roughness of the lap plate, it was first mirror finished on a precision milling machine before each experiment. [Al.sub.2][O.sub.3] abrasives were used in the experiment with an initial mean grain diameter of 3 [micro]m, and were mixed with water in a 1:6 weight ratio. The workpiece was a 5 mm by 5 mm piece of soda lime glass attached to a copper work holder. The total load on the workpiece was 3.45 N applied by dead weight.

[FIGURE 1 OMITTED]

The lap plate rotated at 180 rpm, and the non-rotating workpiece was held stationary by a fixture allowing only vertical movement. A wide band acoustic emission sensor was mounted directly on the back side of the work holder.

The material removal rate was determined by periodically measuring the workpiece thickness using a dial indicator.

The signals detected by piezoelectric sensor are amplified and filtered so that they have the amplitude required by the input stage of the digital oscilloscope. Once the acquisition is completed, data are stored on the QUADRA 950 hard disk or on magneto optical disks. Acquisition procedures realized through LabVIEW (by National Instruments) allow to control the LeCroy 9400 oscilloscope.

3. MECHANISMS OF LAPPING PROCESSES

Lapping is a free abrasive machining process in which the abrasives are allowed to rotate between the workpiece and the lap plate--three body abrasion. When an abrasive particle penetrates the plate surface, the abrasive can become embedded in the plate material and form an abrasive/workpiece interaction similar to that in the fixed abrasive or two body abrasion process. The deeper the penetration the more likely the abrasive will be fixed in the plate. Figure 2 illustrates the possible interaction between workpiece, lap plate and abrasive (Dobrescu, 1996).

[FIGURE 2 OMITTED]

Lapping is a very complicated and random process resulting from the variations of abrasive grains by its size and shapes and from the numerous variables which have an effect on the process quality.

Lapping is a finish method used to obtain good surface quality. Important variables affecting lapping efficiency are abrasive grain size, lapping pressure, lapping speed, quantity of lapping compound supplied and viscosity of the compound. Comparison of the effects of variables on the overall process efficiency is not yet clear owing to the complexity and randomness of the process. The former does not produce any chips by cutting but causes the workpiece to deform plastically which may result in strain-hardening and, finally, the microfracture of the workpiece. On the other hand, the latter produces microchips by the cutting operation. In addition, some small abrasive grains are driven into the workpiece by the relatively large grains. The quantitative ratio between rolling grains and sliding grains affects the lapping efficiency and it is very difficult to predict that ratio by analytical approach because it is affected by numerous environmental and process variables which are irregularly varied. It is recommended to use the experimental approach rather than only the theoretical approach to analyze the lapping process.

A "critical size ratio" for the characteristic particle size, the longest particle diagonal, to film thickness, which divides the two body and three body material removal mechanisms. Through experiments with various abrasives and work materials, the critical size ratio was found to be two.

The ratio of the hardness of the workpiece and the plate is also a significant factor. Two body abrasion will occur when the abrasive indentation on the soft plate is deep in comparison to that on the workpiece. Increase of two body abrasion has been reported to cause high shaft wear rates with a very soft bearing linear (Chang, 1995).

Although the use and exact definition of the term "ductile" as applied to the machining of non-metallic materials has been recently discussed, the ductile regime is traditionally referred to as material removal by plastic deformation. Chip formation indicating ductile machining has been observed by several researchers for a wide range of brittle materials (Moriwaki et al., 1992).

Brittle regime machining is associated with crack generation-based material removal. Because of the distribution of abrasive sizes, material removal may include both ductile and brittle machining (Trumpold et al., 1994).

Lapping mechanisms can be divided into four states. For material removal in ductile regime machining, the volumetric removal rate is associated with the size of the plastic deformation zones. In the case of two body ductile abrasion, the material removal mechanism is similar to that for metals. In a physical lapping process, three body ductile mode machining contributes very little to the total removal rate. For brittle machining, the volumetric removal rate is related to the lateral crack area, assuming the lateral cracks propagate to the surface. When three body abrasion occurs, the interaction between the abrasive grains and the workpiece is similar to that in an indentation test. For the two body abrasion mode in brittle regime machining, the removal rate is related to the groove length and, due to fracture, the depth and length of the lateral cracks created (Dobrescu, 1998).

For a given abrasive size distribution, the total material removal rate can be estimated by the summation of material removal rate from each of the process states (Chang, 1995). The distribution of abrasive size and the total number of grains in the contact area is dependent on time due to abrasive wear during the process. The lapped workpiece surfaces from the tests conducted here were examined under an optical microscope. At the beginning, the first 30 seconds of lapping, the surfaces were composed mainly of rough pits from indentations, indicating mainly three-body brittle machining. After 120 seconds of lapping, the glass surface showed more groove/sliding marks and smooth areas, suggesting a transition from three-body brittle to two-body ductile machining. This transition was also obvious when comparing the surfaces after 30 and 300 seconds of lapping.

As illustrated in Figure 2, several of the acoustic emission sources in lapping are not related to material removal. To distinguish the significance of these sources, acoustic emission signals were also recorded for non-standard lapping operations. The dry Al2O3 glass signal was generated from rubbing the glass against the same amount of fresh dry abrasive powder used in the slurry during lapping. The dry copper-glass signal was generated by direct plate-work contact with neither abrasive nor water, while the wet copper-glass signal was with water between plate and work. The relative low wet copper-glass signal level suggested low noise amplitude from the fluid. After lapping for 600 seconds, a sudden drop of 300 mV was observed in the acoustic emission signal level, followed by the occurrence of audible vibration noise. Examination of the lap plate at this stage indicated the abrasive slurry had dried.

The decrease in material removal rate with time can be explained by the transition from brittle to ductile machining, since the crack area in brittle machining is as much as ten times the plastic deformation area in ductile machining.

An image processing program was used to analyze the work surface after lapping. With this it was possible to distinguish the fracture marks from the smooth surface by noting the difference in the gray scale image.

4. CONCLUSION

Two criteria were found to be important in describing the material removal mechanisms in lapping, i.e. ductile versus brittle machining and two body versus three body abrasion. Using these two criteria, the mechanisms can be classified into four states. Due to the abrasive size distribution, multiple mechanisms usually coexist during lapping operation. The time effect on material removal rate can be explained by the change of abrasive size distribution. A transition from mostly three body brittle to mainly two body ductile machining was observed in the first 120 seconds of lapping with 3 [micro]m [Al.sub.2][O.sub.3] abrasive. An acoustic emission (AE) sensor setup was used to monitor the material removal rate (MRR) and observe the reduction in removal due to changes in abrasive size with lapping time.

5. REFERENCES

Chang, Y. P. (1995). Monitoring and Characterization of Grinding and Lapping Processes, Ph.D. Dissertation, Department of Mechanical Engineering, University of California, Berkeley, U.S.A

Dobrescu, T. (1996). Surface Grinding on a Rotary Table of Silicon Wafers, Research Reports, LAPT, University of Naples "Federico II", pp. 35-38, Italy

Dobrescu, T. (1998). Cercetari privind optimizarea masinilor de superfinisat materiale fragile, PhD Theses, University "Politehnica" of Bucharest, Romania

Inasaki, I.; Tonshoff, H. & Howes, T. D. (1993). Abrasive Machining in the Future, CIRP Annals, no. 42, pp. 723-732

Moriwaki, T.; Shamoto, E. & Inoue, K. (1992). Ultraprecision Ductile Cutting of Glass by Applying Ultrasonic Vibration, CIRP Annals, no. 41, pp. 141-144

Trumpold, H.; Hattori, M.; Tsutsumi, C. & Melzer, C. (1994). Grinding Mode Identification by Means of Surface Characterization, CIRP Annals, no. 43, pp. 479-481