Laser milling for micro tooling D T Pham, S S Dimov, P V Petkov and T Dobrev Manufacturing Engineering Centre,Cardiff University, UK Abstract Laser milling provides a new method of producing components for the micro tool manufacturing industry. This paper considers the technical capabilities of laser milling when applied to the machining of components for micro tooling. Characteristics important for laser milling as a micro tooling manufacturing process are reviewed. These characteristics are surface finish, aspect ratio, dimensional accuracy and minimum feature size. Laser systems use a broad spectrum of pulse durations, ranging from microseconds to femtoseconds. A study is presented of the impact of pulse duration on the surface quality as represented by the surface roughness and extent of the heat affected zone (HAZ) and recast layer. Keywords: Laser micro machining, Micro machining, Laser pulse duration 1. Introduction In the area of micro tool manufacture, laser milling has to compete with various other machining methods. Often structure quality needs to be balanced against efficiency. A proper laser milling strategy is a prerequisite for high accuracy and good quality of the machined micro tools. Characteristics affecting micro tool manufacturing are discussed and their impact on the quality illustrated.
2. Surface finish Surface finish is an important aspect of tool manufacturing, particularly with the current trend towards miniaturisation of every consumer product imaginable. In the manufacture of microtools, the importance of surface roughness is many times higher than for conventional tools, because the roughness can frequently be comparable to the feature sizes. It is expected that surface roughness is affected by the following three laser milling variables [1]: Laser lamp current, I Laser pulse frequency, f Scanning speed, V Laser power is related to the electric current through the laser flashlamp and signifies the amount of energy that is contained in a single laser pulse. The laser frequency gives the repetition rate of laser pulses, and the scanning speed is the speed with which the laser spot travels on the target surface. The relationship between the frequency and the scanning speed defines the distance between the centres of two neighbouring laser craters. An experimental study was carried out to identify the relationship between surface finish and the laser parameters I, f and V. Table 1 shows the values of the parameters used. In all experiments, the layer thickness was kept at 2 µm. # I, % f, khz V, mm/s 1. 90 45 400 2. 86 30 300 3. 82 20 200 4. 76 13 150 5. 71.5 9 100 Table 1 Laser process parameters For the surface finish test a straightforward geometry was employed: a square with a pre-defined size. The chosen size will depend on the requirements of the roughness measuring equipment. In order to avoid interference from the texturing of the test surface, the test piece was prepared in advance with its test surface polished to under R a = 0.2 µm. The measurements of the surface finish were to be made using contact surface roughness measurement equipment (SJ- 201 Roughness Tester). As the test workpiece was polished, the depth of the feature was chosen to be 50 µm, giving a total of 25 layers to be machined. In normal working settings, the best surface finish is achieved by machining each separate layer at a random hatching angle. In order to obtain a true representation of the attainable surface finish, throughout the experiment the scanning direction of the laser beam was restricted to only the x- and y-axes.
This means that one layer will be machined with the laser scanning the x-axis and the next layer with the laser scanning along the y-axis. For each test sample, a total of 4 measurements of the surface roughness were made, split in two groups at 90 angle to each other. This way, the influence of the scanning direction of the laser beam was estimated. All 5 tests were machined on the same workpiece, and a total of 25 surface roughness measurements were made. Table 2 gives the results of the roughness measurements in both the x- and y- directions. Notice that test 3 has 3 measurements produced in the x-axis. This was required because of the considerable difference in the results of the two previous measurements of the same test. Test R a measurements in x and y, µm #x1 #x2 #x3* #y1 #y2 Avg X Avg Y R a avg 1 1.442 1.482 1.563 1.594 1.462 1.579 1.520 2 1.094 1.079 1.155 1.160 1.087 1.158 1.122 3 0.977 1.150 1.057 1.176 1.084 1.061 1.130 1.089 4 1.071 1.064 1.132 1.104 1.068 1.118 1.093 5 1.042 0.962 1.032 1.066 1.002 1.049 1.026 *third measurement of test 3 sample Table 2 Roughness measurement results Total Avg 1.136 1.207 1.170 The lowest average result is obtained from test 5 and is just above the 1 µm surface roughness. The lowest single measurement also comes from test 5 and is R a = 0.962 µm. The highest average surface roughness is measured on test 1, and is over a half a micrometer more. Figure 1 shows the results from the surface roughness measurements against the experiment parameters. It is obvious that the combination of the lower parameter values produces the best result, although if one disregards the lowest and the highest results, the surface roughness seems rather constant around the 1.1 µm value. Nevertheless, the results suggest that lower lamp currents, speeds, and frequencies are more appropriate when the surface finish is an issue. This will of course mean longer machining times. Good surface finish is very important as an attribute, if laser milling is to be established as a micro manufacturing process. 3. Aspect ratio The aspect ratio is defined as the ratio of the height to the width or the gap size of a concave structure. When manufacturing high aspect ratio free-standing structures, e.g. pillars or walls, the requirements on the manufacturing process
are different from those associated with a concave geometry. Since almost all manufacturing processes utilise some kind of tool, the aspect ratio in the case of a concave geometry causes significant factors to be examined, such as tool diameter and length. Unlike conventional milling, in laser milling the achievement of high aspect ratio is hampered by difficulties in producing vertical walls. Achieving a plane wall that is parallel to the laser z-axis is an issue for laser milling. This is due to the high dependence of the material removal process on the size of the laser spot, which enlarges when the beam is located close to a plane surface parallel to the laser axis [2]. As a result laser milled walls have a draft angle. The draft angle is typically in the order of 5 to 12 degrees, and is directly proportional to the aspect ratio of the microstructure. 100 90 450 400 % or khz 80 70 60 50 40 30 20 10 0 0 1.000 1.100 1.200 1.300 1.400 1.500 1.600 Ra avg, microns Figure 1 Variations of the three parameters against the surface roughness I,% f, khz V, mm/s 350 300 250 200 150 100 50 mm/s The successful manufacturing of vertical walls has been achieved through the use of an angular offset of the laser beam when machining the border cuts along the wall. There are two parameters in the laser milling process that are expected to influence the aspect ratio: Minimum approach angle Maximum approach angle The two approach angles [1] refer to the minimum and maximum angles that limit the location of the laser beam during vertical wall manufacturing, and they lie in the vertical plane normal to the given wall (see Figure 2). During vertical wall machining, the laser axis can be positioned anywhere within the range
specified by the two approach angles. In case of incorrect settings, damage on the surrounding faces will be present as can be seen in Figure 3. Minimum approach angle Maximum approach angle Laser beam Figure 2 Laser beam position Figure 3 Result of incorrect setting Tests conducted in this work show that the laser milling process equipped with the vertical wall machining option is still limited in its ability to produce grooves of high aspect ratio. The achieved result is in the low aspect ratio range of the replication process employed for producing similar sized features. 4. Accuracy A nominal accuracy experiment was designed to identify the deviation of the lateral dimensions of laser milled features from the nominal along the x- and y- axes. Due to the unique thermal character of the laser/material interaction, and the location of the laser spot relative to the contour of the feature, a distinction between internal and external dimensions was considered in this experiment. The experiment was carried out on two different metals to investigate the influence of the material properties. The two materials selected for the experiment are H13 tool steel and copper. The selected materials have previously shown a distinctive difference in the response to laser irradiation, and thus require a wide range of laser milling parameters in the material removal process. The major difference in the material properties of the two metals is in the thermal conductivity. Copper exhibits a thermal conductivity that is 2.5 times higher than that of tool steel. Due to the thermal nature of the material removal process during laser milling, the parameter settings for the machining of the two materials will be significantly different. The nominal accuracy test measures the deviation of the dimensions of the machined features from the nominal along both the x- and y-axes. Hence the workpiece should allow for the extraction of various measurements in these directions. Also, the test geometry should have internal and external dimensions, so as to provide the possibility for analysis of the influence of the effective laser spot relative to the contour of the geometry. Figure 4 explains the meaning of the dimension types. It shows that for the external dimensions, the hatching is
performed on the outer side of the contour and thus the laser spot is also located outside. Conversely, positioning the laser spot on the inside of the closed contour forms the internal dimensions. The reason behind this distinction is that the measured dimensions will depend on the laser/material interaction [3] and any difference between the diameter of the actual laser spot and the assumed diameter will result in a uniform error around the contour. Therefore, it is expected that the measured external dimensions will be larger than the nominal. Hence, the opposite is anticipated for the internal dimensions. Two experimental parts were machined (see Figure 5) one in each of the test materials, with laser milling parameters previously identified as giving the best results and with reasonable differences in the laser scanning speeds for each material. On each part, two identical features were machined. The machined parts were to be cleaned in an ultrasound bath, and then the measurements to be carried out on a commercially available vision system. Border cuts Cavity Nominal dimensions E Internal Laser spot Figure 4 Dimension types Figure 5 Test part Comparison between the results of the two test materials will provide an insight into the influence that the material properties have on the accuracy of the laser milling process. There were two test features machined on each of the test pieces. After ultrasound bath cleaning of the samples, measurements in both x- and y- directions were taken and analysed. Figure 6 presents a comparison of the deviation from the nominal dimensions for all test features. In this figure, the first six dimensions show the results for the x-axis, whereas the next six are in the y-axis. The dimensions for each axis were split into the two dimension types (see Figure 4), the first three being internal and the next three, external. In addition, the dimensions are arranged in ascending order from the nominal. From the results, it is obvious that copper provides better accuracy than H13 tool steel. Both copper tests have lower average deviation. Another consideration would be that the feature edges of the H13 test were not clearly defined, which led to the accumulation of errors during test measurement. Under the magnification of the measuring equipment, the H13 test piece displayed a considerable recast layer, larger than for copper, which further complicated the process of measuring. The difference in the accuracy obtained with both materials is largely
attributed to the different physical properties, and in particular the thermal properties. 0.06 0.04 0.02 Internal Internal Deviation, mm 0-0.02-0.04 External External H13_1 H13_2 Cu_1 Cu_2-0.06-0.08-0.1 d12 d13 d11 d1 d5 d6 d10 d8 d9 d2 d3 d4 Dimension designation Figure 6 Deviations from the nominal dimensions 0.09 H13_1 H13_2 Cu_1 Cu_2 0.08 0.07 Absolute error, mm 0.06 0.05 0.04 0.03 0.02 0.01 0 0.6 0.7 1.2 1.4 2.1 2.6 2.7 3.3 3.4 3.8 4.3 5 Nominal value, mm Figure 7 Nominal dimension versus absolute error As can be seen in Figure 7, the deviation does not exhibit any dependence on the size of the nominal dimensions. The distribution for both materials is consistent along the whole range. This result confirms the expectation that the error is independent of the dimension value and that it is related to a more stable factor, such as the size of the crater left from a single laser shot. Furthermore, this is proof of the fact that the dimensional accuracy of the laser milling process is dependent on the settings of the process.
5. Minimum feature size The minimum feature size is a very important attribute for the tool manufacturing process especially in the production of micro tools. As with the aspect ratio, the tool operated by the process predefines the minimum feature size. Experiments were conducted to determine the minimum feature dimensions that can be achieved with the laser milling process [4]. General feature types were used, namely grooves and ribs (or walls). When machining a free-standing feature, such as a wall, without any neighbouring features that could act as obstacles for the tool, a different aspect ratio is generally anticipated. In this situation, other restrictions apply, such as tool/workpiece contact, tool-induced vibrations and the workpiece material properties. Therefore, laser milling being a non-contact process has a distinct advantage. In addition the vertical walls option was included in the experiment, so as to avoid the formation of a draft angle on the walls of the structure. 40 µm groov 40 µm wall Figure 9 Top view Figure 10 Close up Figure 8 CAD model The only variable expected to have an effect on the minimum feature size is the diameter of the focused laser beam spot. ( 45 and 80 µm). Since the minimum was being sought the test was carried out with the smallest possible aperture. A CAD model of the structure can be seen in Figure 8. After cleaning of the workpiece, some measurements were taken on the vision system, but the smallest grooves proved too difficult for any conclusive result. Therefore, the resulting structure was observed and estimates were produced under scanning electron microscope (SEM). Figure 9 shows an overall top view of the fabricated structure. From the resulting shape, it is obvious that the microstructure was close to exceeding the capabilities of the process. The 40 µm groove was not machined to the final depth of 120 µm. The achieved depth was difficult to measure because the bottom surface of the groove was not flat. On the other hand, the 40 µm thick walls are well formed (see Figure 10), and with their depth of 120 µm the obtained aspect ratio for a wall was equal to 3.
6. Laser Source The laser source employed to produce the microtooling inserts (features) has a direct influence on the surface quality. In recent years a wide range of laser sources have become commercially available. Laser pulse durations may vary from microseconds to a few femtoseconds. Experiments were conducted to assess the impact of the laser type (pulse duration) on surface quality. Four different laser types were employed having femto, pico, nano and microsecond pulse durations respectively to produce simple square shapes (1 mm 1 mm). Figure 11 presents the maps of the resulting surfaces (taken by white light interferometry). (a) femtosecond pulse duration (b) picosecond pulse duration (c) nanosecond pulse duration Figure 11 Scanned surfaces representation (d) microsecond pulse duration The best result appeared to be from the picosecond laser - roughness achieved Ra 0.29 µm and the worst result, as expected, from the microsecond laser roughness achieved Ra 2.18 µm (see Figure 12). Further investigation into the heat affected zones also shows significant differences. For the surface produced by the microsecond laser, the recast layer was about 30 µm with the heat affected zone about 4 µm. The surface produced by the picosecond laser was virtually free of heat affected zones (undetected using the same equipment). Using ultrashort pulse lasers improves dramatically the surface quality and opens up the possibility for direct manufacturing of microtooling inserts (features) [5]. Surface roughness is no longer such a major obstacle. Figure 12 Line profiles for pico- and micro-second lasers
7. Conclusions This paper has reviewed characteristics that are important to laser milling as a micro tool process: surface finish, aspect ratio, accuracy, minimum feature size and laser pulse duration. The main conclusions are: Laser milling is capable of producing adequate surface finish for micro tools. Features with an aspect ratio of 2.5 are achievable with the laser milling process. Accuracy is directly influenced by laser-material interaction. Material properties have to be considered in order to improve process accuracy. Thin walls down to 40 µm can successfully be produced while the smallest groove machined is 120 µm. Shorter pulse durations will offer highly improved surface finish and accuracy when employed in the laser milling process. ACKNOWLEDGEMENTS The authors would like to thank the European Commission, the Welsh Assembly Government and the UK Engineering and Physical Sciences Research Council for funding this research under the ERDF Programmes Micro Tooling Centre and Supporting Innovative Product Engineering and Responsive Manufacture and the EPSRC Programme The Cardiff Innovative Manufacturing Research Centre. Also, this work was carried out within the framework of the EC Networks of Excellence Innovative Production Machines and Systems (I*PROMS) and Multi-Material Micro Manufacture: Technologies and Applications (4M). REFERENCES 1.Lasertech GmbH, Operating manual. Gildemeister Lasertec GmbH, Tirolerstrasse 85, D 87459 Pfronten, Germany, 1999 2.Pham, D. T., Dimov, S. S., Petkov, P. V., Petkov, S. P. Laser Milling. Proc Instn Mech Engrs, Vol 216, Number 5, Part B: J Engineering Manufacture, pp. 657 669, 2002 3.Jandeleit, J., Horn, A., Weichenhain, R., Kreutz, E. W., Poprawe, R., Fundamental investigations of micromachining by nano- and picosecond laser radiation. J. Applied Surface Science 127-129, pp. 885 891, 1998 4.Madou, M J, Fundamentals of Microfabrication The Science of Miniaturisation. 2 nd Ed. CRC Press, Boca Raton, Florida, 2001 5.Heyl, P., Olschewski, T., Wijnaendts, R. W., Manufacturing of 3D structures for micro-tools using laser ablation. Microelectronic Engineering 57 58, pp. 775-780, 2001