The use of pulsed lasers for microprocessing material in several manufacturing industries is presented. Microvia, ink jet printer nozzle and biomedical catheter hole drilling, thin-film scribing and micro-electro-mechanical system (MEMS) fabrication applications are reviewed.
© 2000 Optical Society of America
Although material ablation by pulsed light sources has been studied since the invention of the laser [1,2], reports in 1982 of polymers etched by uv excimer lasers [3,4] stimulated widespread investigations aimed at using the process for micromachining. In the intervening years scientific and industrial research in this field has proliferated to a staggering extent - probably spurred on by the remarkably small features that can be etched with little apparent damage to surrounding unirradiated regions of material. Manufacturing industry now uses laser micromachining in many high-tech application areas for which microfabrication is an enabling technology.
It is now known that clean ablative etching can also be achieved using pulsed laser sources at wavelengths other than the ultraviolet. Provided photons are absorbed strongly in submicron depths in timescales less than the time it takes for heat to diffuse away from the irradiated region, then pulsed lasers like copper vapor (CVL), CO2 and Nd and its harmonics can be as effective for ablative micromachining. For a particular micromachining application the choice of laser is now judged as much by criteria such as process speed, part throughput, reliability, service intervals, capital and operating costs of the overall machine tool rather than solely by the quality of the processed part.
The degree of industrial takeup of a technology is a good yardstick for assessing its utility and state of maturity. This paper discusses some industrial applications of laser micromachining and its market importance to several sectors. In this context ‘industrial application’ is taken to mean a process proven to add value to a manufactured product. Not discussed are areas which use pulsed lasers for non-ablative surface treatments like photoresist exposure as used in deep-uv photolithography, marking, annealing, hardening, smoothing and secondary ablative processes like pulsed laser deposition of thin films, surface cleaning and smoothing.
1. Hole drilling
The ability to drill ever smaller holes down to ~1µm diameter, is an underpinning technology in many industries that manufacture high-tech products. By providing solutions to critical problems in manufacturing integrated circuits, hard disks, displays, interconnects, computer peripherals and telecommunication devices, laser micromachining is a key enabling technology allowing the current revolution in information technology to continue. The requirement for material processing with micron or submicron resolution at high-speed and low-unit cost is an underpinning technology in nearly all industries that manufacture high-tech products. The combination for high-resolution, accuracy, speed and flexibility is allowing laser micromachining to gain acceptance in many industries.
As shown in Fig. 1, a practical illustration of laser micromachining can be seen when comparing the quality of holes drilled by an excimer laser and a mechanical twist drill. While the mechanically hole is close to the minimum size that can be drilled, the improved quality of this and much smaller holes drilled by lasers is obvious.
1.1 Microvia hole drilling in circuit interconnection packages
Almost as important as the rapid improvements in speed and memory of integrated circuits (IC’s) are the parallel developments in interconnection packaging made during the last 20 years. So speed, power and area (real estate) are not compromised, packages on which chips are mounted for connection to other devices have had to keep pace with the rapid advances made in IC’s. Thus there is a demand for an ever-increasing packing density of interconnections - for example mountings in current mobile phones and camcorders have around 1200 interconnections/cm2. There are now more than a dozen generic types of chip interconnection packages which include multichip-modules (MCM’s), chip-scale-packages (CSP’s) like ball-grid-arrays (BGA’s), chip-on-boards (COB’s), tape-automated-bonds (TAB’s). Generally these consist of multilayer sandwiches of conductor-insulator-conductor with electrical connection between layers made by drilling small holes (vias) through the dielectric and metal plating metal down the hole. Such blind via holes provide high-speed connections between surface-mounted components on the board and underlying power and signal planes while minimizing valuable real estate occupation. For example, due to difficulties in soldering IC’s with greater than ~200 pins, peripheral lead mounting packages like TAB’s must be made larger than the chip. By placing microvia connections in the package at the base of the chip instead of around its periphery, a BGA is no larger than 20% the size of the chip. Typically then the requirement is to drill 100µm diameter microvias on ~500µm centers. The cost for drilling these vias on such high-density packages can represent 30% of the overall cost of the board.
Drilling microvias by ablation was first investigated in the early 1980’s using pulsed Nd:YAG and CO2 lasers[5,6]. Excimer lasers led the way in applying it to volume production when the Nixdorf computer plant introduced polyimide ablative drilling of 80µm diameter vias in MCM’s - as used to connect silicon chips together in high-speed computers . Other mainframe computer manufacturers such as IBM rapidly followed suite and installed their own production lines for this application [8,9]. With fewer process steps than other methods, laser-drilling is regarded as the most versatile, robust, reliable and high-yield technology for creating microvias in thin film packages. Trillions of vias have now been drilled with excimer lasers at yields >99.99% whose mean time between failure (MTBF) has been logged at >1,000 hours.
Interconnection densities on rigid and flexible printed circuit boards (PCB’s and FPC’s) are also increasing, driving the requirement for drilling ever-smaller vias in these packages . In such lower cost packages the current common practice is to mechanically drill the vias. As diameters decrease to <100µm it is generally recognized lasers will displace mechanical drills, although for these packages excimer lasers are too slow and expensive. Because ~100µm diameter tungsten-carbide drills are expensive, frequently break and rapidly wear, drilling costs skyrocket to several $ per 1,000holes. Using TEA, rf-excited slab CO2 or Q-switched Nd:YAG lasers, drilling speeds for precisely positioned vias can be as high as 200holes/sec at costs as low as 0.6¢ per 1,000holes. Trepanning the hole with a small focal spot under galvo-mirror scanner control allows hole positions and sizes to be programmed from CNC drill files containing the circuit layout - see Figure 2(a). Since copper is highly reflective at 10µm, Q-switched Nd:YAG lasers (fundamental or 3rd-harmonic) are used to drill the metal, while either CO2 (rf-excited or TEA) or 3rd-harmonic YAG lasers drill the dielectric material. When drilling blind microvias, CO2 lasers have the advantage that drilling naturally self-terminates at the copper level below without damaging it. Holes defined previously in the top copper (either by a YAG laser or chemically etched using photolithography) can be used as a conformal mask for cleanly drilling the dielectric material - see Fig. 2(b).
There is a large potential market for laser microvia drilling tools. Many companies now use pulsed CO2 lasers for drilling 80–100µm blind vias through dielectric layers on MCM’s, CSP’s, and TAB’s which then get incorporated into flat panel displays, hard-disk drives, printers, cameras, cellular phones, photocopiers, fax machines, notebook and palmtop PC’s. With laser-drilling now producing twice as many microvias than any other method, the annual market for laser-drilling tools in Japan alone is estimated to be ~600 units.
1.2 Ink jet printer nozzle drilling
Inkjet printers comprise a row of small tapered holes through which ink droplets are squirted onto paper. Adjacent to each nozzle, a tiny resistor rapidly heats and boils ink forcing it through the orifice. Increased printer quality is achieved by simultaneously reducing the nozzle diameter, decreasing the hole pitch and lengthening the head. Modern printers like HP’s Desk Jet 800C and 1600C have 300×28µm input diameter nozzles giving a resolution of 600 dots-per-inch (dpi). Earlier 300dpi printers consisted of a 100nozzle row of 50µm diameter holes made by electroforming thin nickel foil. Trying to fabricate more holes with smaller diameters reduced even further the already low 70–85% production yield. Laser-drilling of nozzle arrays allowed manufacturers to produce higher performance printer heads at greater yields. At average yields of >99%, excimer laser mask projection is now routinely used for drilling arrays of nozzles each having identical size and wall angle . Most of the ink jet printer heads sold currently are excimer laser drilled on production lines in the US and Asia. Figure 4(a) shows some excimer laser drilled nozzles in a modern printhead.
Fig. 4(b) shows nozzles with nonlinear tapers to aid the laminar flow of the droplet through the orifice. More advanced printers sometimes use piezo-actuators. Rather than being constrained to give shapes characteristic of the process, excimer laser micromachining tools with appropriate CNC programming can readily engineer custom-designed 2½D and 3D structures. Fig. 5(a) shows an example of a rifled tapered hole that spins the droplet to aid its accuracy of trajectory, while Figure 5(b) shows an array with ink reservoirs machined behind each nozzle.
1.3 Hole drilling in biomedical devices
As in microelectronics and its associated technologies, the drive for increasing miniaturization with improved device functionality is crucial to the rapid progress being made in the biomedical industry . Precision microdrilling with excimer lasers is routine when making delicate probes used for analysing arterial blood gases (ABG’s) . ABG sensors measure the partial pressures of oxygen (PaO2), carbon dioxide (PaCO2) and hydrogen-ion concentration (pH) used for monitoring the acid-base concentration essential for sustaining life. In intensive care units, ABG results are used to make decisions on patient’s ventilator conditions and the administration of different drugs. The use of fiber-optic sensors for ABG analysis provide clinical diagnostics at the patient’s bedside without the need for taking blood samples .
Fig. 6(a) shows an example a ABG catheter for monitoring blood in prematurely borne babies. The hole at the side of the PVC bilumen sleeving tube through which blood is drawn is machined using a KrF excimer laser. In this case the clean cutting capability of the laser provides the necessary rigidity that prevents kinking and blockage of the tube when inserted into the artery.
More important components of this catheter are the PaO2 and PaCO2 sensors. These consist of a spiral of up to five ~50×15µm rectangular holes machined in a 100µm diameter acrylic (PMMA) optical fiber with an ArF laser. The holes are filled with a reagents whose optical transmission depend on the PaO2 and PaCO2 levels of the surrounding blood. Using a fully-automated workstation shown in Figure 6(b) that has computer-controlled reel-to-reel fiber-feeding and laser-firing, all five holes shown in Figure 7(a) are drilled in the fiber. By spatially-multiplexing a single excimer beam into five smaller ones, holes are drilled simultaneously through the fiber.
Preferential excimer laser etching of plastics compared to metals is applied to the stripping of insulation from fine diameter wires prior to soldering connections. The process relies on the threshold for excimer laser ablation of the polymer being much lower than for damaging the copper or silver core. As shown in Figure 7(b), excimer lasers are used to cleanly strip away the polyurethane insulation sleeving of wires which form the pH resistivity sensor in the ABG catheter above. Such pulsed laser wirestripping is also in widespread use for preparing connection wires to computer hard-disk reader heads.
2. Laser scribing of thin films
Although the cost for producing power from the best large-area thin-film silicon (TFS) solar cells is still ~25¢/kWhr which is more than three times that for fossil and nuclear fuel power stations, recent technological improvements in cell design have stimulated a rapid growth for domestic and commercial use on buildings as a local source of electricity. Compared to crystalline devices, TFS panels use far less active material and because interconnections between cells are intrinsic to their fabrication are cheaper to manufacture. First introduced commercially in the early 1990’s, TFS solar cells comprise a 5-layer thin-film sandwich on a float-glass substrate with each layer only a fraction of a micron thick. Light passes through the glass and the first film of a transparent conductive oxide (TCO) material like indium tin oxide (ITO).
As illustrated in Figure 8, electron-hole pairs and a photovoltaic voltage are generated between p-i-n Si-diode junction layers. Individual cells are segregated and interconnected by scribing narrow isolation tracks in each film and collecting the photocurrent at the end of the panel. The fabrication steps are: (i) chemical vapor deposition of TCO; (ii) cell segregation by laser-scribing; (iii) plasma deposition of p-i-n doped amorphous Si films, each layer laser-scribed; (iv) deposition of conductive film of aluminum or TCO material; (v) laser-scribing of top conductor; (vi) laminate protective plastic or glass covering on top.
The cell width is generally varied to give the required voltage while its length is changed to produce the requisite current. For incorporating into individual products completed panels are cut into smaller sizes. To maximize efficiency, isolation tracks need to be kept as narrow as possible conducive with maintaining high electrical resistivity between the collection strips. The ability to achieve high lateral spatial resolution with precision depth-control without inducing damage to underlying layers are the reasons pulsed excimer and Q-switched YAG laser ablation is used for scribing cells . Tracks are typically ~25–50µm wide displaced by ~30–50µm in each film giving inter-strip impedances >1MW. Glass panels to ~0.5m size are processed with the laser operating ‘on the fly’ taking typically ~1min/layer to scribe. The efficiency for generating electricity from the best TFS cells is now ~7%, so in full sunlight ~70W is produced per square meter of cell. Currently the world manufacturing capacity for TFS cells is ~40MW.
There is intense ongoing research on various types of flat panel display devices (FPD’s) which will supersede conventional cathode-ray tubes (CRT’s) in most high-definition and large area applications. These include active matrix liquid crystal displays (AMLCD’s), polymer light-emitting diodes including polymer devices (LED’s and pLED’s) and plasma displays panels (PDP’s). This display market is now worth around $30B/year. As the demand for larger panels grows more conventional lithography-etch methods of production become problematic and expensive. Lasers are increasingly being used to process FPD’s in scribing operations that segment and define interconnect electrode circuitry in a similar manner to that used for scribing TFS solar panels.
Fig. 9(a) shows a thin film of ITO on glass scribed with 25mm wide tracks by the Q-switched 3rd-harmonic Nd laser machine in Fig 9(b). In addition excimer lasers are now commonly used for low temperature annealing of the silicon layer in thin-film transistors (TFT’s) used to switch and hold the transmissive state of pixels on AMLCD’s.
3. Future trends in laser micromachining
‘Micro-electro-mechanical systems’ (MEM’s) or ‘Microsystems technologies’ (MST) bring together mechanical, electrical and optical technologies to create an integrated device that employs miniaturization to achieve high-complexity in a small volume . This generally involves fabricating mm-µm size structures with µm-nm tolerances. Existing products include devices such as computer hard-disk drive heads, inkjet printer heads, heart pacemakers, hearing aids, pressure and chemical sensors, infrared imagers, accelerometers, gyroscopes, magnetoresistive sensors and microspectrometers. Recent market analysis reports predict the market for MEM’s products will continue to grow at a rate of 18%/annum reaching a value of $34B/year by 2002. Emerging products like drug delivery systems, magneto-optical heads, optical switches, lab on a chip, magneto-optical heads and micromotors will add an additional $4B to this market. The success of microengineering comes from miniaturization and its consequences: high-sensitivity, short-measurement times, low-energy consumption, good-stability, high-reliability, self-calibration and testing. Microsensors detecting local parameters like pressure, flow, force, acceleration, temperature, humidity, chemical content etc, have in the last decade been engineered into the engine and performance management systems of cars and aircraft. They also provide the key to electromechanical microcomponents such as ink jet printer nozzles, gas chromatographs, gyroscopes, galvanometers, microactuators, micromotors, micro-optics etc. Devices like implantable drug delivery systems containing sensors, valves and control system with power source capable of operating for many years are being developed. There is no doubt microengineering will be a key underpinning technology of the 21st century.
Examples of the types of surfaces that can be structured by excimer laser ablation are shown in Fig. 10. Blazed grating and pyramid-like structures can be readily fabricated on surfaces by mask-dragging techniques .
Such methods can be used for making micro-optical surfaces like those shown in Figure 11. The ‘black’ anti-reflective property of the ‘moth’s-eye’ type of surface machined on the CsI crystal substrate shown in Fig. 11(a) is being used to prevent ghost images in very large infrared optical telescopes. The microlens array shown in Fig. 11(b) is used for shaping beams from laser diodes. Each lenslet in this array has a focal length of 1mm.
Currently most MST devices are manufactured using photolithography to define the surface shape of silicon or quartz material that is selectively removed below in a subsequent chemical or plasma etching process. Material removal is almost always used to achieve the topography of the MST device being fabricated. Microparts having a 2½D topology are fabricated by undercutting material from the planar surface using anisotropic etching along crystallographic planes. Upward building is achieved by depositing additional layers on top of processed ones below and repeating the lithography-etch cycle. Adaption of silicon lithography and etch batch-processing as developed by the semiconductor industry is currently the dominant MEMS fabrication method. However being restricted to just one material (silicon) surface and bulk etched in only three directions - along (110), (100) and (111) crystallographic-planes, other more flexible micromachining methods including pulsed uv-laser ablation are being evaluated for MEMS. The advantages of laser micromachining are: (i) few processing steps, (iii) highly-flexible CNC programming of shapes for engineering prototyping, (iii) capable of serial and batch-mode production processing, (iv) no major investment required in large clean-room facilities and many expensive process tools, (v) applicable to a wide range of polymers, ceramics, glasses, crystals, insulators, conductors, piezomaterials, biomaterials, non-planar substrates, thin and thick films, (vi) compatible with lithographic processes and photomask manufacture.
Advances in personal healthcare and environmental monitoring are driving the development of various diagnostic chip-based devices for performing analytic functions. Applications include diagnostics of food and water supplies, drug delivery systems, personal drug administration, DNA analysis, pregnancy testing and blood monitoring. Although many sensor device technologies are well developed, the inherent bio-incompatibility of silicon is driving the move towards developing microengineering techniques suitable for use with other materials. Since most biocompatible materials do not lend themselves to lithography-etch processing laser machining of biodevices is becoming increasingly important. Fig. 12 shows the layout of a 55×40mm biochip that uses travelling-wave dielectrophoresis to sort and sense cells . It comprises 2½D laminations of channels, chambers and electrode conveyor tracks. An excimer laser is used to both ablate a conductive layer of gold to leave 10mm wide electrode tracks as well as drill microvia holes in the polymer layer .
Using only a low-voltage AC power supply, dielectrophoresis is used to control cell transport in the chip. Microfluidic channels and ramps like those shown in Fig. 13 are used to transport the sample from inlet ports to analysis sites.
Constant and varying depth laser-micromachined channels are also used for locating and securing fiber optics in telecommunication devices as shown in Fig. 14.
As illustrated in Fig. 15, controlled 3D-structuring of materials by excimer laser etching  can produce the basic building blocks of bridges, diaphragms, pits, holes, ramps, cantilevers, etc needed for microengineering devices like gyroscopes, galvanometers, gas chromatographs, microactuators, micromotors, etc.
Figure 16 shows examples of uv-laser micromachined 3D-structures in polymers which when used with the LIGA process of electroforming (from the German acronym: lithographie galvanoformung abformung), can be replicated in metal - a process now known as laser-LIGA. Once a master has been made by excimer laser micromachining such methods allow high volumes of replica parts to be manufactured at low unit costs.
Already recognized by government-supported initiatives in Japan and the European Union, laser micromachining will be a key manufacturing tool in emerging nanotechnologies. The economic advantages of mass production at low unit cost is of the highest importance and will open up many new industrial application areas.
It is a pleasure to thank J Fieret, D Milne, N Rizvi, P Rumsby, and D Thomas of Exitech Ltd who contributed much of the original experimental material contained in this paper.
References and links
1. N. Bloembergen“Laser-material interactions, fundamentals and applications,” in Laser Ablation: Mechanisms and Applications II , AIP Conf. Proc. J. C. Miller and D. B. Geohegan eds., 288 (1994).
2. J. C. Miller, ‘History, scope and the future of laser ablation’, in Laser Ablation, Principles and Applications, J. C. Miller ed., Springer-Verlag (1994).
3. Y. Kawamura, K. Toyoda, and S. Namba, “Effective deep ultraviolet photoetching of polymethyl methacrylate by an excimer laser,” Appl. Phys. Lett. 40, 374 (1982). [CrossRef]
4. R. Srinivasan and V. Mayne-Banton, “Self-developing photoetching of poly(ethylene terephthalate) films by far ultraviolet excimer laser radiation,” Appl. Phys. Lett. 41, 576 (1982). [CrossRef]
5. S. D. Allen, M. Bass, and M. L. Teisniger, “Comparison of pulsed Nd:YAG and pulsed CO2 lasers for hole drilling in printed circuit board materials,” CLEO Conference Summary (1982).
6. M. N. Watson, “Laser drilling of printed circuit boards,” Circuit World , 11, 13 (1984). [CrossRef]
7. F. Bachman, “Excimer lasers in a fabrication line for a highly integrated printed circuit board,” Chemtronics 4, 149 (1989).
8. J. R. Lankard and G. E. Wolbold, “Laser ablation of polyimide in a manufacturing facility,” Appl. Phys. A54, 355 (1992).
9. R. S. Patel, T. F. Redmond, C. Tessler, D. Tudryn, and D. Pulaski, “Via production benefits from excimer laser tools,” Laser Focus World (Jan 1996).
10. H. Holden, “Microvia PCB’s:the next generation of substrates & packages,” Future Circuits International 1, 71 (1997).
11. C. Rowan, “Excimer lasers drill precise holes with higher yields,” Laser Focus World (Aug 1995).
12. A. South, “Miniaturization of Medical Products: The Development Challenge,” Medical Device Technology 9, 30 (1998). [PubMed]
13. M. C. Gower“Excimer lasers for surgery and biomedical fabrication,” in Nanotechnology in Medicine and the Biosciences, R R H Coombs and D W Robinson eds, Gordon & Breach (1996).
14. R. S. Gifford and D J Bartnik. “Using optical sensors to measure arterial blood gases,” Opt. & Photonics News 9, 27 (Mar 1998). [CrossRef]
15. S. Kiyama, T. Matsuoka, Y. Hirano, S. Nakano, M. Osumi, and Y. Kuwano, “Laser patterning of integrated-type a-Si solar cell submodules,” JSPE , 11, 2069, (1990).
16. A. B. Frazier, R. O. Warrington, and C. Friedrich, “The Miniaturization Technologies: Past, Present and Future,” IEEE Trans. on Industrial Electronics , 42, No5, 423 (1995). [CrossRef]
17. E. C. Harvey and P. T. Rumsby, ‘Fabrication techniques and their application to produce novel micromachined structures and devices using excimer laser mask projection’ in Micromachining and microfabrication process technology III, Proc. SPIE3223, 26 (1997).
18. R. Pethig, J. P. H. Burt, A. Parton, N. H. Rizvi, M. S. Talary, and J. A. Tame, ‘Development of biofactory on a chip technology using excimer laser micromachining’ J. Micromech. Microeng. 8, 57 (1999). [CrossRef]
19. W. Bacher, W. Menz, and J. Mohr, “The LIGA Technique and its potential for Microsystems - A Survey”, IEEE Trans. on Industrial Electronics , 42, No5, 431 (1995). [CrossRef]
20. P. McKeown, ‘Nanotechnology’ in Emerging Technology Series: New and Advanced Materials, UN Industrial Development Organization 1 (1997).