laser cutting, Laser Microjet, laser sources, water jet, moderate peak power, laser beam, laser source, Frank Wagner, pulse length, nm wavelength, peak powers, peak power, pulse lasers, pulse duration
The Laser Microjet® Technology 10 Years of Development (M401) Frank Wagner, Ochйlio Sibailly, Nandor Vбgу, Rafal Romanowicz, Bernold Richerzhagen Synova SA, Ecublens, CH ABSTRACT In 1993, scientists at the Institute of Applied Optics of the Swiss Federal Institute of Technology Lausanne, Switzerland, demonstrated the feasibility of a hybrid laser system using a water jet as light guide. The laser beam was focused into a nozzle while passing through a pressurized water chamber. Once coupled to the water jet, the laser light is guided inside like in any optical waveguide. The light guiding jet hits the sample, which is machined by the laser pulses contained within. The stable water jet thus works as liquid waveguide delivering the laser light to the work piece. The jet also efficiently expels the melt from the cut, and cools the cut, thus avoiding thermal damage to the sample. 10 years after the demonstration of the principle, engineers at Synova present a summary of the development paths that this technology has undertaken. From the very first trials till today's efficient industrial machines, the Laser Microjet®, as its inventors called it, underwent many refinements, concerning mainly the coupling device, stability of the water jet, and the laser sources employed. The evolution of these three points is reported briefly, and illustrated by some application examples for the different laser types. Research continues aimed at better understanding of the cutting phenomenon, further miniaturization of the water jet, and a wider choice of laser sources. 1. INTRODUCTION As water is the most common transparent object in nature, the idea of using water as an optical element is as old as the optical sciences themselves. Already in 1842 Colladon described in "Comptes Rendus" water jets, which guide light 1. Thereafter, the invention of the laser, and in particular the development of laser micro machining, enabled the step from the purely esthetical illuminated fountains to a versatile cutting tool based on the same principle. The light guiding water jet in its modern form, the water jet-guided laser, was realized in the framework of a project for laser dentistry in 1993. The original idea at the time was that pain in laser dentistry is caused by heat conduction to the nerve of the tooth, and that a collinear water jet will cool the tooth 2. Later on it became clear that the water jet addresses not only the problem of the heat-affected zone (HAZ) but also other problems of modern laser cutting, like focussing problems, burr formation, material deposition, and too high forces on the sample. In consequence, the Laser Microjet® is today an established cutting tool especially in the domain of cutting, grooving and drilling of heat-sensitive or mechanically sensitive materials. The success of the Laser Microjet is based on a continuous research effort
in order to improve the overall performance of the system. After giving some more details on the operating principle and its realization, we will discuss in the following section the key points of the combining device, stability of the water jet, and the choice of the laser source. The development of these points was most important for the evolution from the first steps in 1993 to today's reliable industrial machines. 2. THE OPERATING PRINCIPLE A sketch of the water-jet guided laser cutting system is shown in Figure 1. We use pure de-ionized and filtered water at 5 to 50 MPa for the water-jet. The nozzles are made out of sapphire or diamond in order to generate a long stable portion of the water jet. The laser beam, coming from the fiber delivery of the laser, is collimated, passes a beam expander and is then focused through a quartz window into the nozzle. The coupling unit is very similar to a usual fiber-coupling unit, except that the intensity distribution of the light in the nozzle is flat-top and not Gaussian, due to the mode mixing in the fiber delivery of the laser and the imaging properties of the setup. Once in the water jet, the light is reflected at the air-water interface due to the refractive index
step (Figure 2). The samples are fixed onto a CNC controlled translation stage, and the samples are moved under the water-jet guided laser beam in one direction during the cutting process, and the optical head moves in the perpendicular direction. The z-variation of the stage is only necessary in order to adapt to the different working distances of differently sized nozzles at different water pressures, and is not used during the cutting procedure.
collimating lens fiber delivery laser
optical head on y-z-stage beam expander focusing lens water pump coupling unit sample on x-stage
Laser Water entry
Diamond nozzle Sample to be cut
Figure 1: Schematic of the water-jet guided laser setup
Figure 2: Detailed sketch of the coupling unit
Since over five years this tool is applied to (micro) machining problems of various domains. During this time the water jet-guided laser proved its numerous advantages in industrial applications compared to classical laser cutting. The following list qualitatively describes some advantages:
1. The guidance of the light has two advantages: It allows processing of different material thickness' without refocusing; because there is no focal point, light intensity is constant over the whole length of the water jet. The parallelism of the light and the fact that the light guidance avoids irradiation of sensitive surfaces near the cutting track, generates parallel cutting kerfs and avoids rounded top edges. This is especially helpful when cutting hard materials that are several millimetres thickness, as for example machine tool inserts made out of cubic boron nitride (CBN).
2. The expulsion of the melt by a water jet, rather than with cutting gas of several bars pressure, is much more efficient, because the density and in consequence the momentum of the water jet are much higher than density and momentum of the cutting gas stream. This is important for avoiding burrs on the sample backside when cutting metal foils, as for example stencils. A second important consequence is that practically no material can be ejected against the water jet and redeposit on the topside of the sample, thus strongly reducing the debris on the sample's surface.
3. The fact that the water jet does not expand between the nozzle and the work piece ensures a nearly force-free cut compared to classical laser cutting. The main reason for the force on the sample in classical laser cutting being the deviation of a large part of the assist gas stream by the sample surface, these forces are avoided as the water jet passes nearly entirely through the cutting kerf.
4. The last advantage mentioned here, and in many cases the most important one, is the cooling effect of the water-jet. As the sample is cooled right after the end of the laser pulse at exactly the site that was heated before, heat conduction into the workpiece is efficiently decreased and in consequence the thermal load of the sample is strongly reduced. The "cold cut" is not only generated by the water-jet cooling of the cut surface, it is also supported by the fact that we have a nearly flat-top intensity distribution (see point 1, 3), and is aided by the efficient expulsion of the hot melt (see point 3).
In short these advantages ensure a new market for this special technique of laser cutting. The present contribution however, will not concentrate on the applications of the Laser Microjet® but on the development of key points that today guaranty highly reliable industrial cutting machines based on this fascinating principle.
3. COMBINING DEVICES As in the beginning the water jet-guided laser was planned as a tool for dentistry, the first device that coupled the laser to the water jet was a prototype of a dentistry tool. This prototype was not yet equipped with an alignment camera that allows visualizing of the relative positions and sizes of laser focus and nozzle entry. In this first stage relatively large ruby or sapphire nozzles were used to generate the water jet and pressures of only 20 bars or less were applied using a simple pump. The laser employed was however already a pulsed flash lamp pumped Nd:YAG laser as is the case for numerous applications today. Over the years, the combination of light and water jet was then implemented in several devices: The starting point of the development was a prototype for industrial applications that used an open optical setup on an optical table. Most important, this prototype setup introduced the usage of the alignment camera and additional illumination for the alignment procedure, but also it allowed for variable demagnification factors ranging from 2 to 8. For the first industrial machine already an identical setup was implemented in a sealed optics head comprising exactly the components used in the industrial prototype. This first optics head was still rather bulky (Figure 3, left side) and was subsequently improved and miniaturized. Figure 3, right side, shows the optics head as it is used today for 1064nm and 532nm systems. The new optical head is smaller lighter and more rugged than the former versions. Figure 3: Left side: first sealed optical head. Right side: today's optical head. Both images are approximately at the same scale. 4. WATER JET STABILITY One of the key issues when realizing a water jet-guided laser for material process
ing is the generation of a highly stable water jet. A cylindrical liquid surface is not stable with respect to external and internal perturbations. One of the destabilizing factors is the surface tension of the liquid, because drops have higher volume content at the same surface as a cylinder. Besides surface tension, the velocity profile of the jet and the surrounding atmosphere mainly define the stability problem 4,5. Due to the inherent instability of liquid jets, it is only a question of time when surface wave
s, starting out of noise, are amplified to such an extent that droplets detach from the jet. The distance of the end of the continuous liquid column from the nozzle outlet is called the break-up length lb. The temporal growth rate
of the surface waves depends on the water jet's speed 5, thus leading to an optimum in the "stability curve", i.e. the break-up length versus the jet's speed (Fig 4).
distance from nozzle outlet (mm)
Figure 4: Stability curves for different nozzle diameters d ranging from 50 to 200 microns, according to the simple model proposed by Sterling and Sleicher 6. Smaller nozzles exhibit this maximum at higher pressures and the maximum obtainable break-up length is smaller than for larger nozzle diameters. Guided light is scattered out of the jet as soon as the amplitude of the surface waves exceeds a certain threshold value
7. A perfectly stable water jet is needed for water jet-guided laser machining. Within the working distance the jet needs to be stable enough to avoid droplet formation when inserting a work piece perpendicularly to the jet's axis, and also needs to be insensitive to droplet formation during work piece insertion. The maximum working distance, lw, defined by the above criterion roughly follows the stability curve. However, for a given nozzle, the maximum in lw occurs at lower pressures than the maximum in lb, and we observe naturally lw < lb for all pressures (Figure 5). 160 140 120 100 80 60 40 20 0 0 50 100 150 200 250 300 jet speed (m/s) Figure 5: Break-up length, lb (crosses), and maximum working distance, lw (circles), for the same nozzle. as a function of the average jet speed. Nozzle diameter was 75 microns. An upper limit for the working distance is given by the temporal fluctuations of the break-up length. For systems with strong noise the minimum break-up distance, lb,min , and the maximum working distance, lw, approach each other (Figure 6, lower jet), for systems with little noise there is an important difference between these two values (Figure 6, upper jet). Figure 6 shows photographs of two water jets, one emerging from a "low noise" nozzle (upper part) and one emerging from a "high noise" nozzle (lower part), superposed with the photograph of a ruler (mm scale). Flow direction is from left to right. As long as the coupled laser light is guided in the jet, the latter is not visible because
only little ambient light was used during the photograph. The guided light is scattered out of the jet at the instantaneous break-up point, lb(t), that fluctuates rapidly leading to the impression of a zone of scattered light ranging from lb,min to lb,max . Figure 6: Images of water jets emerging from a "low noise" nozzle (upper part) and a "high noise" nozzle (lower part). Flow direction is from left to right. For both nozzles lb,min , lb,max and lw are indicated. Note that the distance between lb,min and lw is much higher in the case of the low-noise nozzle. "Noise" in the above paragraph may be anything related to velocity profile perturbations of the water jet. In particular, the constancy of the water pressure and the symmetry of the flow towards the nozzle are important for obtaining stable jets. Concerning the water pumps, Synova started with simple rotational pumps. The pressure fluctuations were decreased using pumps with several pistons and later on nitrogen loaded damping elements. A big step forward towards better pressure constancy was the usage of a pump with two hydraulically driven and separately regulated cylinders. This principle is still in use today, even though the electronics have evolved remarkably and today's pumps are completely microprocessor controlled and communicate with the machine tool by the industrial Profibus® protocol. Another source of "noise" is the nozzle itself, which has to be polished, sharp-edged and without chipped surfaces (Figure 7, right side). These features lead to a separation of the jet from the nozzle edge, which in turn results in a non-parabolic very stable velocity profile. Due to these requirements we changed from sapphire nozzles to diamond nozzles where sharper edges can be achieved. In addition diamond offers higher tensile fracture strength and higher heat conductivity compared to sapphire, which is also advantageous for resisting the thermal shocks induced by the laser pulses. Figure 7: Scanning Electron micrographs
of two diamond nozzle entries. Left side: 20-micron nozzle, bad edge quality, rough walls, rough surface and downstream taper are observed. Right side: 50-micron nozzle, sharp edge, no downstream taper and a smooth surface are observed. In spite of these extreme stability requirements, the nozzle diameters that can be used in Microjet cutting are now close to the kerf width values obtained during classical laser cutting. The smallest nozzle today has a diameter of only 40 microns and further miniaturization is possible.
5. LASER SOURCES AND IMPORTANT APPLICATIONS The employment of water as the jet medium enables a large range of laser types to be used, because of its large transmission window. When defining a threshold of tolerable absorption of 0.3 cm-1, the transmission window ranges from 1120 nm in the infrared until 190 nm in the ultra-violet wavelength spectrum (Figure 8).
Abs. koeff. (1/cm)
Figure 8: Water absorption spectrum from UV to IR wavelength 8. Historically, we first used long pulse Nd:YAG lasers at 1064 nm wavelength. At this wavelength nearly any pulse duration and average laser power
can be purchased as industrial grade laser sources. For machining of thinner materials Q-switched lasers are also employed since 1999 resulting in superior cut quality, and higher cutting speeds. Using Q-switched Nd:YAG lasers for industrial cutting applications is somewhat unusual when considering the situation from the point of view
of classical laser cutting. In classical laser cutting, the higher peak power of Qswitched lasers that generates vapour in the cutting kerf is a disadvantage compared to long pulse lasers, which only melt the material. Due to vapour generation in classical Q-switched laser cutting melt droplets are ejected against the cutting gas stream onto the front side of the work piece, where they deposit. The high momentum water jet, having a similar density as the melt, eliminates this problem. In this case ejected droplets cannot reach the surface against the water flow and the advantages of the Q-switched lasers, i.e. high pulse repetition rate and short pulse lengths become beneficial for cut quality. Nevertheless rather long Q-switched pulses (> 200 ns) with not too high peak powers (< 4 kW) are employed in water-jet guided laser cutting. Naturally, the allowed peak power value for good cut quality depends on the nozzle diameter, the pulse duration, and the material to be cut. A good example for the benefit of moderate peak power is the important domain of silicon cutting (Figure 9).
Figure 9: Cut quality dependence on laser peak power. Cutting of 100-micron thick silicon wafers at 1064nm and a pulse duration of 80 ? s. The allowed peak power for good quality is pulse length dependent. The nozzle diameter for all cuts was 50 microns. More recently Synova also uses frequency doubled and frequency tripled Nd:YAG lasers at 532 nm and 355 nm wavelengths respectively. As the jet diameter determines the cut width, and the peak powers of these lasers are limited, the possibility to focus these shorter wavelengths to smaller spots is not exploited when using them in the
Laser Microjet systems. The shorter wavelength has nevertheless the advantage of stronger absorption in some important materials, as for example copper and GaAs. By this means, the range of possible applications of the water jet-guided laser technology was further enlarged (Figure 10). Figure 10: Samples processed with the 532nm laser. Left: 100-micron thick copper foil; cut width is 100 microns. Right: Dicing of thin GaAs wafer. Besides the moderate peak power, which is optimum in Laser Microjet® cutting, the requirement to use a constant irradiated surface on the sample defines the specifications of the laser to be used. The typical laser system for water jet-guided laser cutting has average powers of 50 W and more (not yet in the UV), peak powers of less than 4 kW (depending on pulse length), and moderate beam quality (due to the use of multi-mode lasers). 6. MACHINE TOOL IMPLEMENTATION AND DESIGN In this paragraph, we shortly illustrate the development of the design of the machine tool itself with some photographs. Common points in all models are the linear motor driven translation stages, the NC-type machine control, and the modular concept of the machines. The modular concept enables the use of different types of laser on one machine by simply changing the fibre delivery, and it allows space saving in the production hall (clean room) by placing the laser, pump and machine tool in different locations. Figure 11 shows the first industrial machine including water pump, laser and control electronics. Figure 12 shows the central machine tool of the newer models combining the axis, the control electronics and the user interface
. The LDS200-A also integrates a cleaning station and automatic cassette-to-cassette handling of the wafers. Figure 11: First industrial machine build by Synova SA.
Figure 12: Newer models of Microjet® machines; shown without laser or pump. 7. SUMMARY AND OUTLOOK From the first steps in dentistry, the realization of water jet-guided laser cutting was constantly improved over the last 10 years. The optical head, coupling the laser into the water jet, was miniaturized in order to obtain more compact machines and better dynamics of the axis. It today comprises an alignment camera and it is aligned with a simple vision system. The water jet was rendered more stable and its diameter was decreased. In spite of the extreme stability requirements of water jet-guided laser cutting we can use today jet diameters down to 40 microns and further miniaturization will be possible. The laser sources employed were diversified, comprising today different pulse durations at 1064nm and 532nm. Together with the laser manufacturers pulsed sources exhibiting high average power (> 50W) and moderate peak power (<4kW) were developed. These new sources led to high quality machining results in many application domains. Clearly the potential of this young laser material processing technique is still huge. Future development projects
include further miniaturization of the water jet, further diversification of laser sources, and machines with a wider working area (800 x 800 mm2). These developments will enable the Laser Microjet® to conquer wider domains of applications. REFERENCES 1. Hecht J.. City of Light, New York:Oxford University
Press, 1999. 2. Richerzhagen B., "Entwicklung und Konstruktion eines Systems zur Ьbertragung von Laserenergie fьr die Laserzahnbehandlung," doctoral thesis, EPFL, 1993. 3. Couty P., Wagner F. and Hoffmann P.. Mode noise in a multimode waterjet waveguide. J. Opt. Eng. to be published, 2003. 4. McCarthy M.J. and Molloy N.A.. Review of Stability of Liquid Jets and the Influence of Nozzle Design. Chem. Eng. J. Vol. 7, 1974; pp. 1-20. 5. Lin S.P. and Reitz R.D.. Drop and spray formation from a liquid jet. Ann. Rev. Fluid Mech. Vol. 30, 1998; pp. 85-105. 6. Sterling A.M. and Sleicher C.A.. The instability of capillary jets. J. Fluid Mech. Vol. 68, 1975; pp. 477-495. 7. Vбgу N., Spiegel Б., Couty P., Wagner F.R. and Richerzhagen B.. New technique for high-speed microjet breakup analysis. Exp. Fluid. in press, 2003. 8. Nikogosyan D.N.. Properties of optical and Laser-related Materials, Chichester:John Wiley & Sons
MEET THE AUTHORS Frank Wagner worked during his studies in the field of ultra-fast laser development (M.Sc., University of Gцttingen, Germany) and thereafter in the field of UV-laser ablation (PhD, Swiss Federal Institute of Technology Lausanne, Switzerland). Mr. Wagner was hired as head of the Process R&D Group at Synova SA in 2000. Main interest is the research and development
of the water-jet guided laser technique. Ochйlio Sibailly worked during her studies in the domain of Material Science
s in particular the numerical simulation
of shaping and forming processes. In 2002 she received her MSc degree in micro engineering from the Swiss Federal Institute of Technology Lausanne, Switzerland, and is now employed at Synova SA. Nбndor Vбgу received his MS degree in engineering physics from the Budapest University of Technology and Economics and started PhD school in 1999. With a Swiss federal scholarship he continued his PhD research at the Swiss Federal Institute of Technology in Lausanne in a joint Research Program
with Synova SA from January 2000. His research interest is in microjet stability. Rafal Romanowicz Following his Ph.D. and M.Sc. in Micro-engineering at EPFL, he acted for 3 years as Managing Director of Sensile Technologies, a company specialised in industrial pressure sensors and fuel delivery logistics systems. He joined Synova in 2003 as Marketing Manager Bernold Richerzhagen received his MSc in mechanics from the Technical University of Aachen, Germany, and his PhD in micro engineering from the Swiss Federal Institute of Technology Lausanne, Switzerland. He is the inventor of the award winning water jet-guided laser technology. Today he is the CEO of SYNOVA SA, Lausanne, an incorporated company manufacturing high precision laser machines, which he has founded in 1997.
F Wagner, O Sibailly, N Vágó