Evolving laser sources enable novel applications

Material processing represents one of the largest and most diverse application areas for all types of lasers. Applications range from high-power carbon dioxide (CO₂) laser welding and cutting of metals used in auto-body parts to diode-pumped solid-state devices that can repair computer memory chips. Some applications are well established, while others are still experimental.

Several of the newer applications are of particular interest. The increasing output power of diode lasers, for example, means these devices are now taken seriously for certain material-processing jobs. Meanwhile, in the areas of rapid prototyping and micromachining, new materials and new requirements are opening up opportunities for laser-based techniques to play a major role.

In Europe, many research organizations and commercial entities are developing such techniques; a necessarily small sampling of these developments provides some insight into the scope of their activities (see "Funding").

Diode-laser processing

During the past few years, the potential of diode lasers for material processing has been investigated by many research groups and companies in Europe. At the same time, the cost of high-power semiconductor lasers has dropped to a point that makes the technology competitive with more-established material-processing lasers such as Nd:YAG and low-power CO₂ devices. The price per watt of diode-laser output has dropped about tenfold, while the available output power has increased by the same amount. The benefits of diode lasers in an industrial setting include high reliability, small size, low weight, and low (single-phase) power consumption, as well as easy installation and, typically, no requirement for cooling water. As a result, several new semiconductor-laser products have been introduced, and the outlook for diode-laser penetration of industrial applications is bright.

Cutting, welding, drilling, and marking of nonmetals is one area that shows considerable promise. The basic requirement is that the materials absorb or can be sensitized to absorb at the diode-laser emission wavelengths. There is little to differentiate the performance of Nd:YAG lasers from diode lasers in these applications, while CO₂ lasers are often complementary. Thus, lap welding of a transparent plastic to a colored plastic is straightforward with a diode laser but nearly impossible with CO₂ lasers because they melt or vaporize only the surface layers of the material.

Conversely, thin polymer films are often impervious to treatment by diode or Nd:YAG lasers but are easily processed with CO₂ systems. Compared to nonlaser techniques—such as ultrasound or heat—diode lasers have a clear advantage in laser welding of plastics with complex geometries. They can also weld heterogeneous combinations such as polymer to metal or thermoplastic elastomers to rigid thermoplastics (see Fig. 1).

The Fraunhofer Institute for Lasertechnik (ILT; Aachen, Germany) is marketing a range of industrial diode-laser systems based on its applications research. Engineers there have developed a technique that allows the high-power output from diode-laser bars to be shaped and coupled into a fiber by two microstep mirrors. Such beam shaping addresses the very different beam-quality factors (M²) in the two orthogonal axes of the laser output. These are typically a value of 1 in the direction perpendicular to the p-n junction of the diode and 1700 in the direction parallel to the junction.

The ILT solution uses two identical microstep mirrors, which each have N highly reflecting surfaces and which together turn the beam through two 90 angles while introducing offsets between adjacent parts of the incident beam (see Fig. 2). The recombined beam has the property that the M² in one axis is increased by the multiplying factor N while the M² in the orthogonal axis is decreased by the dividing factor N. Hence, a trade-off can be made, aiming at equalizing the M² in the two directions, which allows direct coupling of the output into a fiber. Using this technique the designers achieved an overall diode-to-fiber coupling efficiency of 71%.

The Dioweld 40 from ILT is based on high-power diode-laser bars mounted on water-cooled heat sinks. Optics convert the light from the bars into a circular spot, and the system delivers 40 W at either 806 or 940 nm from an 800-µm-diameter fiber. The system is designed primarily for welding plastics, and welding speeds of up to 10 m/min have been achieved. Another product, the Diomark 10, is intended for marking. It delivers 10 W and incorporates an acousto-optic modulator that produces 30-ns pulses at repetition rates up to 50 kHz.

In the UK, surgical diode-laser pioneer Diomed Ltd. (Cambridge) last year announced its intention to expand into the material-processing arena. With a 60-W unit already on the market, the company plans to launch a 120-W system emitting at wavelengths centered around 810 nm that is currently being developed with UK government support. This system will be targeted at applications on the edge of feasibility at low power, such as seam brazing and ceramic-to-metal brazing. Diomed uses free-space optics to simultaneously image several discrete diode lasers—rather than diode-laser bars—into the delivery fiber. Anamorphic magnification of the laser emissions enables the outputs of many diode lasers to be combined into the aperture of a single focusing lens that images the individual emitters into the linear and angular aperture of the fiber. In principle, many kilowatts of power can be delivered this way.

Diomed also has been investigating soldering of both leaded components and surface-mounted devices (SMDs). This is an application requiring CW or long-pulse powers in the region of 10 W and is one in which diode lasers perform well. The advantages of compactness, low power consumption, and air cooling give diode lasers an edge over Nd:YAG devices in this application. Brazing is an emerging application area, partly because of the need to eliminate soldering flux, which is an environmental contaminant. As with cutting and welding of metals, the power currently available is marginal for wide application coverage, but there are a number of accessible niches.

Rapid prototyping

While early research on laser-based rapid prototyping focused on production of nonstructural parts, the emphasis now is on producing engineering parts with full mechanical strength to be used as a working system component or as tooling. This, of course, means working with metals rather than resins or polymers. The aim is to achieve one-step production in small volumes of usable parts. Research in Europe is moving in several directions.

The group at ILT has undertaken a detailed investigation of the material properties of laser-generated parts. The decision to use this method of producing functional prototypes depends on how similar the parts are to those that would be produced in a long-run production process. Stainless steel (316L) was the material tested. The results showed that the material properties of laser-generated parts are very similar to those made from sheet metal, although the generated parts are harder and less flexible than the sheet-metal parts. The yield strength and Young`s modulus are of the same order of magnitude for both types of part, and, interestingly, there is no difference for a force applied parallel to or perpendicular to the orientation of the layers.

The tensile strength of the laser-generated parts is about 15% lower and the elongation at break shows a significant difference—10% for the laser generated part and 38% for the sheet metal part. Overall, these tests indicate that in many applications representative functional prototypes can be made by laser generating.

Prototype parts also can be made by laminating cut sheets of material. The Bremen Institute of Applied Technology (BIAS; Bremen, Germany) has developed a process called laser-assisted sliced prototyping. Three-dimensional computer-aided-design (CAD) data are used to produce a computer model of the component cut into thin slices, in the same way as would be done for stereolithography. For complex shapes the individual layers need to be thinner. A computer-aided-manufacturing (CAM) program takes the model and generates paths for laser cutting of the sheet steel that will form the component. The slices are then spot welded to build the prototype.

In the UK, a project involving three universities and seven industrial partners is intended to define a new laminated manufacturing process for the large tooling needed primarily in the aerospace industry. The LAST-FORM (Large Scale Tooling For Rapid Manufacturing) project is investigating several options, including laser sintering, laser welding and cladding, and laser bonding and stacking. The project was initiated by British Aerospace Military Aircraft Division (Preston, England), which has a long-term strategic interest in low-cost, short-lead-time tooling. The other industrial partners are Rolls Royce (Derby), Rover (Warwick), British Aerospace Airbus (Filton), Shorts (Belfast), Delcam (Birmingham), and Quantum Laser Engineering (Coventry). The academic partners are the universities of Leeds, Liverpool, and Warwick.

The tooling requirements set by the aerospace industry are tough to meet. Tooling can be up to 2 × 1 × 1 m in size, and the intent is to produce workable tooling without secondary processes. There are additional needs for high-temperature use of some tools—operations at room temperature include metal pressing, at medium temperature (up to 450C), injection molding, and at high temperature (up to 950C), superplastic forming diffusion bonding of titanium. The project goal is to be able to manufacture usable tools in all areas, which will save at least 30% of the time and cost when compared to existing practices.

Each university is evaluating a different technique, and all will start developing small tools, building up to larger tools as the project progresses. At the University of Leeds the investigators are building on their expertise in the area of powder metallurgy and laser sintering of powders. Liverpool University will examine laser welding and cladding with multiple-axis beam manipulation, while Warwick is developing laminated tools based on two-dimensional laser cutting followed by stacking the sections and bonding them into dimensionally accurate three-dimensional (3-D) laminated tools. Tools are likely to be metal, but British Aerospace has not ruled out the use of other materials, such as ceramics, for some applications.

To support slice-based laser prototyping, the KAEMaRT group at the universities of Parma and Milano in Italy is developing new slicing software. The goal of the research is to investigate advantages and drawbacks of producing stereolithography data from direct slicing of a 3-D solid object. The group has developed a direct slicing module using the ACIS 3-D Toolkit for Windows NT platform. Sliced models are created by direct slicing of 3-D solid objects, using the geometrical and topological information present in the solid model to control the slicing procedure and ensure the sliced model correctness.

The system imports a solid object using a standard format, allows the setting of slicing parameters, performs direct slicing on the solid object, and then provides a set of consistency checks to evaluate the sliced model and identify critical regions that can introduce errors during the fabrication process. Once the model has been found to be correct, output data for stereolithography machines can be automatically produced. The system supports HPGL (Hewlett-Packard Graphics Language) format and is expected to extend its capabilities with CLI (Common Layer Interface) and STL format. This project is run in collaboration with the University of Tokyo in Japan.

Laser forming. Researchers at Dundee University (Scotland) have had success developing a prototyping method using laser forming in conjunction with laser cutting. This has advantages over stereo- lithography and other layer-by-layer processes, because it can produce thin section components from sheet metal and make a prototype that represents the full engineering properties of the final component.

Laser forming uses a focused or partially focused beam that is tracked over the surface of the laser-cut blank. The Dundee group uses a 1-kW CO₂ laser. The degree of focus, tracking speed, laser pulse duty cycle, and laser power are adjusted to bring about softening or melting of the metal below the beam to a depth usually between one-third and one-half the full thickness of the sheet. As the bottom surface remains largely unaffected, thermal distortion occurs when the top surface metal along the laser track expands, contracts, and changes phase from solid to liquid and back again. Cooling is either by natural means or forced more rapidly with a gas or liquid jet following the laser beam. The overall effect is a bending up toward the laser of the sheet along the track. On a single pass the deflection induced by this is typically 2. Repeating the process over the same path a number of times enables a sharp bend to be achieved. Alternatively, by offsetting each track slightly, radii can be changed—the closer the track pitch, the tighter the radius.

According to the Dundee researchers, the folds and curves produced using laser forming are as rigid as those produced by physical folding and bending, although some surface marking is inevitable. Tests done on 1- and 2-mm stainless steel show that, with the optimum parameters, 90 folds can be achieved in both with fewer than 40 double passes. The duty cycle is higher and the feed rate is lower for the thicker metal, so this can become a slow procedure but it still takes only minutes instead of hours.

Laser caving. Another laser-based technique for rapid manufacture of metal parts is being optimized at the Bayerisches Laserzentrum (BLZ; Erlangen, Germany). Laser caving uses high-power CO₂ lasers to ablate material from a solid metal block. Controlled laser-beam caving enables the direct conversion from a CAD data set to, for example, a forging die or injection mold. The method has the advantage that the part formed has the full material properties of the metal block.

Material removal is achieved in one of two ways in the BLZ work. Melt ablation, in which the laser melts the metal that is then removed with a gas jet, is a fast process that can remove more than 1000 mm³/min—it leaves a poor-quality surface, however. The alternative method is reactive ablation, in which the metal is melted in an oxygen atmosphere. The metal oxide so created detaches in chips from the bulk metal when it cools; reactive ablation is also known as chip removal. The material-removal rate is much lower, but the surface quality and working accuracy are better. BLZ has optimized the process using a combination of melt ablation and chip removal.

At BLZ, a gantry machine, Lasercav LC 500, is used for the 3-D processing. The machine, which is equipped with a 750-W CO₂ laser, is manufactured by the project`s industrial partner, LCTec (Pfronten, Bavaria). An optical depth-measuring system at the workpiece drives a feedback loop to the processing laser to provide machining accuracy of caving depth on the order of a few microns. Software development is also necessary and currently relies on "slicing" the workpiece`s CAD model and generating a numerical-control data set to feed the machine control. The laser beam is delivered to the workpiece by adaptive optics for maximum flexibility.

Two successful applications undertaken include the manufacture of forging tools and of surface-texturing tools. The pattern to be generated on a surface is taken from a photograph by converting the gray levels into ablation depths. Hence, a leather pattern can be ablated onto a tool that can texture the surface of plastics with the same pattern. The process is still being developed to commercial viability by BLZ, and further optimization in terms of process flexibility and cost will open up new application areas.

Micromachining

Whereas there is a need to develop metal prototypes for macroscopic objects, there is also a need to develop prototypes in other materials on a microscopic scale for components such as sensors, biomedical products, and other micromechanical devices. This is an area in which other types of lasers take over—the shorter-wavelength systems that provide higher resolution. Laser micromachining can produce usable devices directly or can make masters for higher-volume replication in a way that is flexible and rapid.

UV micromachining. The use of pulsed ultraviolet (UV) lasers to produce 3-D features on polymer or other substrates by the process of laser ablation is well known. The pressures to produce smaller and smaller structures come from many areas—the semiconductor industry being an obvious source. Today`s most-advanced storage devices, such as 256-Mbit DRAMs, require structures as small as 0.25 µm. Within ten years, 16-Gbit DRAMs will be in use, requiring features down to 0.085 µm. In addition, miniaturization of sensors and processes, for example in the field of biosensors and bioprocessors, calls for higher- and higher-resolution micromachining. This drives demand for UV laser micromachining because only short-wavelength lasers can meet the resolution requirement at the leading edge.

German laser manufacturer Lambda Physik (Göttingen) has been developing excimer-laser applications for many years. Growing application areas include deep-UV lithography and thin-film-transistor (TFT) annealing (see "Annealing of liquid-crystal displays"). As these applications are taken up in industry, the company has seen significant growth—25%-30% per year for the past three years. Accurate hole drilling is required in the electronics industry, for example, for vias and for ink-jet printer nozzles—but at sizes around 10 µm these do not test the lasers` ultimate performance.

A resolution of 0.18 µm can be achieved with argon fluoride (ArF) excimer lasers emitting at 193 nm. Exitech (Long Hanborough, England) offers a commercial system based on a Lambda Physik 193-nm laser, and this is currently being used for development of very-deep-UV photoresists and other materials necessary for 193-nm lithography. The next step is trying to harness the shorter wavelength of the F2 laser (157 nm). A commercial system suitable for industrial use is under development but is still some way from viability.

Not all groups, however, are using excimer lasers. Laser Zentrum (Hannover, Germany), has assessed frequency-quadrupled Nd:YAG lasers emitting at 266 nm as UV machining sources alongside its excimer lasers. The Nd:YAG laser assessed was built at the center, and with optimized optics a focused spot size smaller than 0.5 µm was measured. This laser proves especially effective at mask repair—it is used in single-shot mode to remove opaque defects and at high repetition rate to evaporate chromium by laser-based chemical-vapor deposition to fill clear defects. The researchers are also investigating frequency-doubled argon-ion lasers emitting at 248 nm for marking masks, wafers, and other planar substrates.

At the Microelectronics Centre in the Technical University of Denmark (Lyngby, Denmark), an argon-ion laser emitting at 488 nm is effective for a range of operations. The laser output is CW, so the method of operation and the range of processes available are different from pulsed lasers. Micromachining of silicon is achieved with a tightly focused spot, sized below 0.5 µm, to machine under a chlorine atmosphere. The molten material reacts with the chlorine, forming silicon chlorides that can be removed as gases. The sample can be moved at speeds of up to 100 mm/s during the writing process while the beam is modulated on the fly.

Laser LIGA. A process developed in Europe over the past few years that allows large-scale production of micromachined parts is laser LIGA. Developed originally by a team at the Institut for Mikrotechnik (IMM) in Mainz, Germany, laser LIGA uses replication techniques that were developed for the LIGA process, which is a combination of lithography, microelectroplating, and micromolding processes for low-cost replication. Whereas the x-ray lithography used traditionally in LIGA can produce very fine features, it does not have the flexibility to machine nonuniform structures and has difficulty coping with nonvertical walls. Laser LIGA provides a fast and flexible complementary technique.

Applications potential is increasing for polymer microstructures. In the field of micro-optics, for example, many parts are needed in volume, including mechanical precision alignment and coupling structures as well as optical elements such as grating structures, zone plates, and waveguides. Microfluidics also needs parts such as filters and meshes for particle separation and flow channels for mixing and reaction of components. Whereas laser-based micromachining can produce these structures, the cost for volume production is prohibitive, hence the need for a lower cost-replication route.

Laser LIGA starts with a laser ablation process. A polymer (PMMA or novolak, for example) layer is prepared on a substrate and micromachined with a UV laser; the initial work at IMM used an Exitech micromachining station fitted with a Lambda Physik laser emitting at 193 nm. The "microlandscape" of features of different sizes and depths produced is electroplated with nickel and then backfilled to make templates—inverse structures that can be used for the injection-molding step. The polymer is separated from the nickel template produced by shock-freezing with liquid nitrogen.

Microinjection molding from these templates needs special techniques because the features are so fine. The mold cavity must be evacuated because any slits made in the mold insert to allow air to escape would be of the same size as the features and would clog immediately under the filling conditions used. To fill very fine features, or those with a high aspect ratio, the polymer needs to be heated close to its melting temperature to reduce the viscosity. After filling, the system is allowed to cool so that the polymer is sufficiently rigid to allow it to be removed from the mold. Using this technique, the IMM group has replicated features as fine as 20 µm.

Laser Zentrum (Hannover) has also investigated laser LIGA and an alternative method of laser machining ceramic molds directly, giving a simpler process. Further research is needed, however, to assess which materials will work best and whether they will wear well enough to be economically viable as commercial molds. The group has found that some ceramics provide acceptable alternatives to the conventional metal molds. Whereas nearly all types of technical ceramic can be machined with high accuracy, the best results so far have been obtained with nonoxide ceramics such as silicon carbide and silicon nitride.

In a joint project with the Fraunhofer-Institut fur Werkstoffphysik and Schichttechnologie (Dresden, Germany), the group is also examining use of reactive gases to assist the ceramic workpiece processing.

Micromachining with CVLs. Copper-vapor lasers (CVLs), although emitting at wavelengths longer than UV systems, offer some advantages in micromachining applications. Holes as small as 10 µm in diameter can be drilled at high speeds; the CVL parameters are useful here—primarily high repetition rate and high pulse energy in a short pulse. As with excimer-laser processing, there is no microcracking or glassification, and there is excellent control over taper and dimensional tolerances. Development of CVL-based processes has been spearheaded by Oxford Lasers (Oxford, England). Commercially successful applications include drilling microscopic holes for ink-jet printers, flow and dosage regulators, and, most recently, petroleum and diesel fuel-injector nozzles.

Working with an automotive partner, Oxford Lasers has just completed a two-year study into diesel-fuel-injection nozzles. Researchers have been able to produce nozzles with diameters ranging from 70 to 200 µm. The process speed and reproducibility and the hole roundness and edge quality are well within the requirements for a production process. In addition, the laser can drill with a controlled taper. Drilling of very small holes is required for the next generation of diesel nozzles, necessary to meet the latest emission standards.

Oxford Lasers has also just completed an applications study on a part destined for use on the Large Hadron Collider at CERN. The ceramic part is a self-aligned fiber-ribbon termination for a modulator array. Because of the small dimensions of the semiconductor component and limitations of optical access to this device, the fiber must be aligned within an angle of 1 to prevent significant deviation of the light. Holes were drilled in electronics-grade alumina to provide precise lateral and angular constraint for single mode fiber ribbons.

A European collaborative project on CVL-based micromachining has been underway for just more than one year. COMPALA (cost-effective series and mass production of high-precision metallic microparts and optical structures in mold by laser submicron machining) is funded by the European Union under the Brite-Euram scheme. Partners in this program are Philips (Eindhoven), Robert Bosch, TNO (Delft), Oxford University, and Oxford Lasers. The CVL used for this project is a master-oscillator-power-amplifier configuration producing 40-W average power and 130-kW peak power in a beam that is only two times diffraction limited.

Power stabilization of the output beam is important for reproducible results, and a stability of better than ۪.25% has been achieved with a stabilization unit. Spectroscopic investigations of the laser-produced plasma show the presence of strong spectral features that may be used in process control further into the project in areas such as focus position, depth control, and multilayer milling. Fiberoptic beam delivery may also be possible. Although results are not yet released the partners claim to be excited about the potential of this program after a successful first year.

Funding

Funding for material-processing projects comes from several sources. The projects closest to market development are funded mostly by industry. The European Eureka program funds some work, and industry contributes financial resources, too. Eureka is funded by the individual governments; it is not funded by the European Union (EU), and close-to-market projects are allowed. Programs such as the Industrial and Materials Technologies (IMT) research program are aimed at generic research, which is defined as precompetitive and so not close to market, so it is funded by the EU. Such projects usually involve groups from several countries cooperating in a major application area.

IMT is the second largest of the 15 specific programs under the EU 4th Framework Programme. It is also known as Brite-Euram 3. The total budget for the program between 1994 and 1999 is 1700 billion European Currency Units (ECUs). The technical scope of the program includes production technologies, materials and technologies for product innovation, aeronautics technologies, and technologies for surface transport means. All of these are pertinent to laser-based projects, and considerable laser material-processing work is funded under this program.

Some of the work described in this report is also funded under another EU program (ESPRIT (European Strategic Programme for Research and Development in Information Technology). And other work is funded under national programs sponsored by individual governments.

Annealing of liquid-crystal displays

Sopra (Bois-Colombes, France) started work on excimer-laser annealing of amorphous silicon about four years ago (see Laser Focus World, May 1996, p. 101). Since this pioneering work, other laser manufacturers are developing products for this application.

Laser manufacturer Lambda Physik (Göttingen, Germany) has developed the Microlas system for thin-film-transistor (TFT) annealing. The demands placed on active-matrix liquid-crystal displays (AMLCDs) are increasingly severe. They are such that the advantages of polycrystalline silicon over amorphous silicon are needed—the key features are higher carrier mobility, faster charging times, and CMOS process compatibility. Whereas various techniques have been investigated for producing polycrystalline silicon, including furnaces, lamps, and argon-ion lasers, excimer-laser annealing is generally considered superior to all other techniques. In this process a glass substrate is coated with amorphous silicon, which melts under the UV laser beam and recrystallizes into the polycrystalline form.

The Microlas 308-nm laser-based optical system has been developed in close cooperation with industrial partners such as the Japan Steel Works (JSW). At the JSW, the system has been integrated into a complete annealing automat with loading and unloading cassettes, fully automated steppers, handling robots, and central system controllers. The first systems have been installed in high-volume production lines across the world.

Lasers and art

Excimer lasers emitting at 248 nm are preferred for this application; the combination of short pulse length and strong absorption limits the effect of the laser to a small thickness with little thermal damage outside this area. The energy density used is between 10 and 200 mJ/cm². The material removal per pulse is a layer between 0.5 and 1 µm thick, and, as both the paint and varnish layers are of the order of tens of microns thick, the process is fully controllable.

In addition to cleaning the paint, the laser can also clean the support material of the picture. This has been successfully demonstrated on canvas, wood, and paper; the black fungi that commonly affect old paper can be removed directly with UV laser irradiation, for example. Overpainting can also be removed, revealing the older or unmodified work behind the newer version.

On-line process control is achieved with optical diagnostic systems. Image processing, for example, is used with a module for broadband reflectography that detects the increasing brightness of the reflected light as the contaminated layers are removed. More-analytical techniques are being evaluated and include wavelength-tunable reflectography, which reveals layer by layer information at different depths for each wavelength. Laser-induced fluorescence also can analyze the paint pigments.