This review focuses on engineering initiatives in manufacturing submillimeter parts with micrometer-scale features. Conventional rapid prototyping is not very dimensionally accurate, but newer processes are closing the gap. New micromilling machines can produce an array of tiny features with accuracies measured in micrometers. The chapter highlights various efforts by different companies in the field of micromachining and other micromanufacturing processes. American Society of Mechanical Engineers member Dick DeVor and his team of researchers received a grant from the National Science Foundation to demonstrate a machine that could mill parts with micrometer-scale features. A three-year grant followed, enabling the researchers to improve the technology and probe the radical differences between micromachining and conventional milling. In addition, a microfactory has been developed by the Swiss Center for Electronics and Microtechnology, which links several small delta robots that can assemble three parts per second with 5-micrometer accuracy.


By 2000, it was obvious that mechanical parts were shrinking ASME fellow Dick DeVor and his team of researchers at the University of Illinois at Urban-Champaign saw it in cell phones, automotive sensors, hearing aids, and hundred of other product. The march to miniaturization led to a big idea. As DeVor put it,"Why do you need a machine tool as big your kitchen to make a part as small as your fingernail?" DeVor was not the only one asking that question. The trends were equally visible from Europe and Asia. In consumer electronic, small sensors, tiny electrical connectors, and diminutive hard drives filled an increasingly large role in cameras, GPS system, and music player. Smaller parts made medical devices like hearing aids,pacemaker,and orthopedic fixtures smaller and less noticeable. They also made possible new diagnostics and innovative devices to remove cardiovascular clots.


The military wanted to shrink the equipment soldiers carried, and develop cameras and engines for tiny tactical drone aircraft. Makers of laboratory equipment sought to compress pumps, valves, and mixing devices onto chip size laboratories to run hundreds or even thousands of experiments on a single sample droplet. Semiconductor manufacturers were interested in building deep X-ray lithography masks with finer features, and in better ways to handle the minuscule parts used in printed circuits.

"These components are around you everywhere today, but we don't think about it," said Andy Phillip, one of DeVor's former graduate students. "Hearing aids need small knobs that rotate to adjust the volume, and small electrical connectors to transmit the signal. Cell phones have hundreds of individual components. Arthroscopic surgical devices, pacemakers, and orthopedic implants for bones in the hand have lots of macroscale components that need microscale fasteners and other devices."

Back in 2000, U.S. researchers interested in submillimeter-scale parts worked primarily with semiconductor processes to make micro electromechanical systems, or MEMS, chip-size devices with simple moving parts. European and Asian researchers, on the other hand, were investigating a wider range of processes, such as machining and extrusion, often as an enabling technology for a new product or technology. Highly focused government initiatives fund ed many of their efforts.

Some of the work was compelling. Japan's National Institute for Advanced Industrial Science and Technology, for example, built a diamond-tooled lathe with a base scarcely larger than a quarter, as well as a briefcase-size micromanufacturing line where two robots shuttled parts among three milling machines. Japan's government subsidized the development of ROBOnano at automation giant Fanuc Ltd. The $1 million unit consisted of a five-axis mill, grinder, and shaper whose lathe achieved nanometer-scale linear resolutions.

By 2007, Germany's Hanover Fair was displaying several micromachining products. They ranged from extruders that produced parts with micrometer-scale features to robotic assembly lines that fit on a tabletop. One, from the Swiss Center for Electronics and Microtechnology, was a finalist for the fair's $150,000 Hermes Prize for new technology. It featured miniature high-speed robots designed to work with millimeter-scale parts. Its developers claimed that the small robots worked faster and more precisely than conventional robots when assembling watches, micromotors, sensors, circuits, and optical systems.

DeVor entered the race in 2000 with a grant from the National Science Foundation to demonstrate a machine that could mill parts with micrometer-scale features. A three-year grant followed, enabling the researchers to improve the technology and probe the radical differences between micromachining and conventional milling. A

grant from the U.S. Army funded integrated microfactories that combined machine tools with materials-handling robots, instrumentation, and metrology. Then, several ofnDeVor's graduate students did what mechanical engineers often do when they see an opportunity and understand the technology. They founded a company.

{Why build a machine as big as your kitchen to make a fingernail-size part?}

Taking Research to Market

The new firm, Microlution Inc., headquartered in Chicago, is headed by Phillip, who became president, and Andrew Honegger, vice president. Another of DeVor's former students, Onik Bhattacharyya, heads sales and marketing. Microlution rolled out its first product, the 310-S, in 2007. With its 4- foot- square footprint and 2.S- cubic inch workspace, it looks more like an office printer than a machine tool.

Yet the 31 O-S is a true 3-axis milling machine with the same CNC controls and codes as larger tools. It has 2-micrometer positioning accuracy, micrometer repeatability (without re-registration), and much greater spindle speeds than those of conventional high-speed mills. It even has an optional automated tool and pallet changer.

Moreover, the 310-S was designed specifically for the milling of very sn1.all parts. Its specialized focus and small size translate into a $90,000 price tag, significantly lower than larger competitors that can manage a wider range of sizes. In its first year, Microlution sold six machines, Bhattacharyya said.

It took seven long years to get there. Fortunately, Microlution could build on research at De Vor's Laboratory for MicroManufacturing Research and Education at the University of Illinois. "One of the great things about the research at the university is that we showed how well it works," Phillip said.

Microlution needs every bit of credibility it can muster because it must compete with a dizzying array of new micromanufacturing technologies. On the micromachining side, many traditional machine-tool makers want a piece of the lucrative medical device market. Here a company must contend with Fanu c and such top CNC powerhouses as Japan's Makino Milling Machine Co. and California's Haas Automation Inc.

Microlution also has smaller, specialized competitors, such as Moore Nanotechnology Systems LLC. Moore specializes in ultrahigh-precision CNC machines designed to manufacture optics. The company's Nanotech 250UPL uses a groove compensated air bearing work spindle to achieve motion accuracy of less than 25 nanometers.

In 2004, DeVor and several top academic experts completed a study of European and Asian advances in micromanufucturing for the World Technology Evaluation Center Inc. It found an enormous range of technologies capable of turning out parts with micrometer-scale features.

Some, like micromachining, are subtractive, slicing away at a blank workpiece until only the finished part is left. One method is electrical discharge machining, which uses controlled electrical arcs to erode a workpiece. Since tool and workpiece do not touch, there is no risk of crushing or cracking fragile tools or parts.

Japan, Korea, and Taiwan all have major electrical discharge programs. Panasonic Factory Automation has a micro-EDM system capable of drilling 50-micrometerdiameter holes in stainless steel with 1 micrometer positional accuracy. Such small holes are used to attach very small gears, springs, molds, precision optics, and video camera magnetic heads. At National Taiwan University, mechanical engineers have used micro-EDM to carve an intricate pagoda from a wire only 1.25 mm thick. Meanwhile, in Germany, Laser Zentrum Hannover e.V has used femtosecond pulses of laser light to cut grooves 200 nanometers wide into a substrate.

Another approach is to form net shape parts. Many of these processes-microinjection molding, hot embossing, microforging, microstamping, hydrostatic extrusion-resemble those used to manufacture automobiles. Researchers at Taiwan's Metal Industries Research & Development Center have used many of these techniques to make bearing parts, deep-drawn profiles, and titanium-based prosthetics.

Among the additive processes , rapid prototyping has attracted significant attention. The process starts with special polymers or ultrafine metal or ceramic powders. It then uses a light source, usually a bright ultraviolet light or laser, to chemically bond or physically fuse the powders with one another to form a layer of solid material. Working from a CAD model, prototypers can build complex parts (often with movable features) layer by layer.


Conventional rapid pro to typing is not very dimensionally accurate, but newer processes are closing the gap. Microstereolithography, a rapid prototyping process originally developed 15 years ago at Japan's Nagoya University, now generates layers only 200 nanometers thick. The developers say it takes about five minutes to fabricate a gear with a is-micrometer outer diameter. The Swiss Federal Institute of Technology in Lausanne has used its own microstereolithography system to sculpt a finely detailed 2-millimeter horse and an even smaller turbine fan. Not to be outdone, Laser Zentrum Hannover has chiseled miniature statues and spiders from a light-sensitive polymer.

Material Variation

Against so much competition, Microlution hopes to leverage De Vor's reputation as a top researcher in the field. "We're not coming to market with a new idea, but with something that has been very thoroughly developed and scientifically proven," Phillip claimed.

Potential customers need to hear this because, while microprocessing techniques look like their macroscale cousins, the interaction between process and material is often very different. For example, take injection molding, which shoots molten plastic into a mold where it solidifies (a process usually helped along by cooling the mold). Try that with millimeter-scale parts, though, and the molten plastic solidifies before it ever fills the mold. The reason? The cold mold instantly absorbs the heat from the minuscule amount of plastic entering it.

Machining metals raises many issues as well. Forces that scarcely merit attention on conventional machines, like slight vibrations or infinitesimally small deflections of the spindle, can rip a tiny piece apart. Thermal problems abound, especially in small parts with limited ability to absorb heat. Even equipment sensitive enough to measure distance in micrometers may not have the resolution to measure those distances on 20 or 40 parts on a workpiece the size of a quarter.

DeVor's team is an excellent case study of some of the issues that occur when scaling down to smaller parts. Take fixturing, for example.

Securing a part in place is often taken for granted on a macro scale. DeVor recalled working with hearing aid parts that were only slightly larger than a flake of dandruff. To make them, he secured a relatively large metal blank onto a sacrificial layer with a water-soluble glue, then milled 50 to 100 parts in a single operation. " Afterwards," he said, "we'd dissolve the glue and pick out the parts with tweezers."

Milling dozens of parts at a time works great in the lab, where graduate students can spend a day or two swapping and recalibrating tools and metrology instruments between operations. That wouldn't cut it in industry.

In stead , Microlution developed a kinematic mount, a flat adapter plate with round balls in its four corners. The balls fit into four V-shaped slots in the machine and are held in place with strong magnets. "You can take off the spindle head, slap on a laser metrology head, measure, and then put the spindle back on with submicron repeatability," Bhattacharyya said. "The same with the workpiece, and with our automated workpiece changer, you can run a new workpiece without re-registering the machine." Then, of course, the machine had to actually cut micrometerscale features. That turns out to be very, very different from ordinary machining.


A Cutting Difference

In the macro world, sharp cutting tools shear thin slices of material off the workpiece. This works because the radius of the tool's edge is very small compared with the thickness of the material sheared off in anyone pass. Think of it as cutting thin slices of a Thanksgiving turkey with a newly sharpened carving knife.

As features grow smaller, though, the relationship between tool edge and the amount of sheared material breaks down. Microscale tools are not any sharper than their larger cousins, but they need to cut far thinner slices.

This gets harder and harder as the radius of the tool edge approaches the thickness of the material being removed.

Instead of trying to carve ordinary slices from the turkey, imagine trying to cut slices that are 10 or 15 times thinner. Instead of slicing through the turkey cleanly, the knife tends to push the meat out of the way. This is called plowing. On milling machines, plowing annihilates accuracy, sabotages surface finish, and dulls cutting tools.

It takes speed to avoid microscale plowing. Typical milling machines operate at 5,000 to 10,000 rpm, and high-speed machines reach 40,000 rpm. Microlution's 310-S runs its spindles at 50,000 to 200,000 rpm. Its spindle skims along the surface, using a combination of low forces and high speeds to shave off a few micrometers of material per pass. High speeds also make it possible to finish a part in a reasonable amount of time. It 's relatively simple to keep spindle speeds high when moving in a straight line. But what happens when it comes time to mill square corner pockets? Conventional machines come in fast, decelerate as they reach the corner, stop, and then speed up as they move in the other direction.

If a micro mill acted like that, it would plow the metal going into and out of the corner. To prevent this, the spindles on De Vor's systems accelerate and decelerate at 5 gs , two-and-a- half to five times greater than conventional mills.

This calls for a fundamentally different type of design, Phillip asserted. " If you take a traditional machine, shrink it down, and accelerate it to 5 gees, you'd have poor dynamic performance," he said. "Its small bearings are not as stiff as larger bearings, so compliance of the machine would be a problem."

To overcome that problem, Microlution redesigned the machine's motion stages. "The idea," Phillip said, "is to position the center of effort of the motor going through the center of mass of the stage holding the workpiece so there is as little deflection as possible."

Microlution, like other emerging makers of micromanufacturing equipment, has scored some early successes. For biotechnology firm iCyt Mission Technology Inc. in Champaign, Ill., it machined a 0.070 mm-diameter nozzle orifice with 0.001 mm repeatability for use in sorting cells.

It also prototyped small hearing aid parts for Knowles Electronics Inc. The company had made prototypes by microinjection molding, but the small molds cost up to $30,000 and took eight weeks or longer to build. Microlution said that machining the parts slashed turnaround to three weeks while cutting costs.

Putting It Together

As impressive as micromachining and other micromanufacturing processes are, the technology to handle small parts demonstrated by the Swiss Center for Electronics and Microtechnology at last year's Hanover Fair is equally breathtaking.

The company's microfactory is based on its PocketDelta, which it describes as the world's fas test robot. It is capable of up to 3 cycles per second with 5-micrometer accuracy. The robot itself resembles larger delta robots, whose effector, the business end of the robot, hangs from three arms. Each arm consists of two lightweight carbon fiber tubes that give the robot excellent stability at high speeds.

Each 120 x 120 x 200 mm module includes the PocketDelta robot, component transfer system, feeder mechanism, motion drives, control electronics, and computing.

The robot operates in an 80 x 80 x 30 mm workspace, using mechanical and vacuum grippers to grab and position parts weighing up to 20 grams.

Lining up several modules in a row would create an assembly line capable of building watches, gluing electronic parts into place, building optical systems, or manipulating cells.

The Swiss Center is especially interested in applying the technology to the watch industry. Ordinarily, manipulating tiny watch gears and cogs is a laborious undertaking that goes offshore to nations with cheap labor. Robots might keep some of that work in Switzerland and other high-wage nations. Granted, robots will replace workers on assembly lines. Yet they also open new jobs to people who design, build, program, and repair robots.

In fact, it is hard to say exactly where this new generation of shrinking mechanical parts will lead. In the past, chip-size microelectromechanical systems supplied many of the most impressive gains from small mechanical parts.

Yet MEMS are rather simple devices. Commercial MEMS usually consist of just one or two parts that move in response to physical forces , such as acceleration , weight, or current. Yet engineers have used them to do everything from stabilize cars that have begun to skid and guide GPS systems when they lose their satellite beacon to filter cell phone signals and project images on television screens.

Imagine, then, what engineers could do with true, chip-size mechanical devices that can do everything that's done by their larger counterparts. Imagine packing the function of a true mechanical system into a cell phone. Or building a traffic surveillance aircraft the size of a pencil. Imagine creating medical devices that robotically clean out a diseased artery or attack a cancerous growth.

Just imagine anything you want. Then imagine it smaller.

{Imagine what engineers could do with true, chip-size. mechanical devices. }