Cleanrooms: Sweating the Small Stuff

Using cleanroom grease, vacuum, careful cable design, and appropriate coatings, system designers can achieve motion control in cleanroom conditions.

Success, the saying goes, comes from focusing on the small stuff. In cleanrooms, they focus is on the really small stuff -- particles, to be exact. U.S. Federal cleanroom standards range from Class 1 to Class 100,000, based on the number of particles per cubic foot of air. ISO cleanroom standard 14644 defines Classes 1 through 9, Classes 3 through 8 of which correspond to the U.S. standards.

In a cleanroom, any source of particles is a problem. The problem is that just about anything that moves is a source of particles, and since motion control is all about movement…well, you get the picture.

Of course, not all cleanrooms are created equal. A Class 10,000 cleanroom is far more forgiving than a Class 1 cleanroom. “That’s the first thing you need to know when you’re designing these systems -- what class is needed and where?” notes Paul Smith, director of business development, SSi Robotics Life Science, PaR Systems Inc. (Shoreview, MN). “Where is the product? Is it processed throughout the cleanroom or do you just need to keep a certain area within your machine clean within a certain rating? Knowing that beforehand will guide your design.”

Keeping It Clean
The most basic approach is to keep the particles away from the work area. That means locating the motion system, where possible, below the region that needs to be clean and using laminar flow to whisk away the particles generated. “The first rule of motion in a cleanroom is get it below your work zone,” says Jim Monnich, Parker Hannifin’s engineering manager at the Daedal location (Irwin, PA). “That’s not always possible but that’s what you try to do because once you have a laminar flow there, you can have something that’s relatively dirty and it won’t affect your process.”

Smith recently worked on a system for a biomed lab that required one region to be very clean for conducting cell culture. “We directed HEPA-filtered air toward the reagents and located the robot downstream,” Smith says. “If the robot generates particles or the chain and cables generate particles, they’ll be blown away from the critical area.”

Indeed, those cables and chains he mentioned can be major sources of particulates -- and headaches, depending on the class of cleanroom. Cables are available sheathed with Teflon-based material, and some vendors field cleanroom-specific cabling products. “Many times if it’s linear motion you don’t want to use round cable, you want to use flat cable so there’s no twisting going on,” says Monnich. “You make specialty carriers so that there isn’t rubbing. There’s a lot of care taken in the selection of the different materials so that you’re not generating particulates.”

At Asyst Technologies Inc. (Fremont, CA), automation supplier for the semiconductor and related industries, designers limited the amount of cabling in their wafer engine system by using a centralized controller but distributed servo driver boards, says vice president and CTO, Anthony Bonora. “There are typically many connectors or cables that are required between the encoder, the motor, driver amplifier, limit sensors, and controller. The wafer engine uses servo drivers located directly adjacent to the motors, so we minimize cabling to the CPU.”

This wafer sorting/loading system transfers silicon wafers from Class 1 containers to process and measurement tools. The system combines up to four loadports and a clean air system with the Asyst Wafer Engine. It can incorporate more than 20 axes and supports on-board real-time monitoring and diagnostic capabilities (Image courtesy of Asyst Technologies Inc).

Monnich advises avoiding hinged carriers with linkages. Extruded, formed carriers made of specialty cleanroom materials provide a cleaner option. Even better, some manufacturers have developed structured cabling that contains embedded strength members and dispenses with the need for cable carriers entirely. The tradeoff, of course, is cost.

The mention of hinges brings up another issue -- that of coatings. Painted surfaces, for example, tend to particulate, making hard anodized aluminum a more appealing choice for cleanrooms. For more demanding medical applications, 316 stainless steel with its high nickel and chrome content offers good performance. Passivating the stainless steel -- etching out the iron -- leaves a surface that doesn’t oxidize. “It’s a very rust- and corrosion-resistant surface treatment,” says Smith. “If you want to go very clean, aseptic, and sterile, or you have certain corrosive chemicals in the environment, then 316 is often used.” Certain plastics provide good options, as well.

Immaculate Motion
Lower-end cleanrooms may use vector motors or other types of synchronous motors and conveyor systems. In tighter cleanroom classes, though, servo motors provide the answer, says Monnich. “The majority of new applications are all servo motors of some kind. They’re trying to limit the motion to just what it needs to be and the best way to do that is a servo system or stepper system.”

“In the control of the high acceleration of the wafer engine, we want to have a system capable of very stable and accurate positioning with a very fast settling time,” says Bonora. “We achieve that partially by going to very-high-torque brushless motors in conjunction with high-resolution encoders, and keeping the inertia of the system as low as possible.”

Brushed motors can be a challenge, he cautions. “A brush-type motor can easily generate a million particles a minute. It’s almost like cigarette smoke, these very fine graphite and metallic particles. Unless the motor is carefully sealed, particles come out of the motor and will contaminate the zone around it.”

Rotary servo motors are the easiest to use in cleanrooms, says Monnich. “Since the rotary motor by its very design is a self-contained device, it is pretty easy to make it cleanroom-compatible.” The biggest challenge is the front bearing. The rotary motion can break loose particles and release them into the air.

High-viscosity, non-particulating cleanroom grease provides one method of capturing those particles. In addition, manufacturers typically pull a light vacuum in the motor so that any particles jarred loose will be drawn back into the motor rather than released into the room. The outflow from the motor is either exhausted outside the cleanroom or ported into the cleanroom vacuum line.

Linear Thinking
In the case of linear motion, things become a bit more complicated. The standard approach is to convert the motion of a rotary servo motor into linear motion using an actuator, such as a ball screw. Most of these devices are metal on metal, which means they produce particles. To minimize particulation, we return to the usual methods: containment by packing the actuator with cleanroom grease and removal by pulling a vacuum in the housing surrounding the actuator. It’s not as simple as it sounds, though. “A linear device is much harder to seal,” says Monnich. “That’s a real engineering effort from a materials standpoint, how to design the seals to try to minimize how many particulates will come out of the mechanism.”

Linear actuators also present real limits to speed. Even at high viscosity, the cleanroom grease can be whipped off when the speed gets too high. And that’s not the only constraint. “The amount of particulates that come off the ball screw is an exponential function of your speed,” Monnich notes. “Just because you double your speed that doesn’t mean you’re doubling how many particles you produce. That could mean you’re quadrupling them, and there are speeds where particle generation really starts to get high.”

Linear Motors
Linear motors provide an alternative. Magnetic linear motors have no contact between the moving element and the stationary portion, which eliminates particle generation. “The adoption of linear motors in cleanrooms is very rapid,” says Monnich. “They were the early adopters.” Of course, there is always a tradeoff. Linear motors require linear encoders for feedback, which ups the price tag. In addition, the motors themselves are more expensive because they incorporate a magnet bar consisting of exotic materials.

“If you have a short travel, a linear motor may not be much more expensive than a ball screw solution,” says Steve Huard, engineering manager of servos for the Parker Hannifin Automation Group. “But if you have a very, very long travel, you end up having a lot of magnet material, in which case your linear motor becomes very expensive.”

Another drawback of linear motors it that they require more cabling. You basically trade a mechanical device that generates contaminants for a clean device with cables -- which also generate contaminants. The increased throughput offered by linear motors still makes them an appealing cleanroom solution, though, so long as they provide sufficient muscle.

Looking to the future, Bonora notes that the semiconductor industry is under pressure to reduce the cost of automated materials handling, which means reduced cost per servo axis. “Ideally, we’d like to be able to see a multifold reduction in the next few years,” he says. “Part of it has to come from an integrated look at cabling, sensors, motor drivers, controllers, the motors themselves…a whole organic group of subsystems that constitute controlling motion.”

Manufacturers have already economized by relocating facilities to low cost regions like Singapore and China, he says. Now, he suggests, it’s time to move beyond. “I think many opportunities have already been realized through manufacturing supply chain changes. I think there is an equally significant and important way to achieve benefits through a comprehensive design approach.”

Vendors, the gauntlet has been thrown.