"It was a good learning experience," says Amick, vice president of Colin Gordon & Associates in San Bruno, Calif., who in June 2004 evaluated specialized areas of the interdisciplinary research center according to a new set of vibration and acoustic metrics. "When the design effort started in early 2000, the word nanotechnology wasn't even being applied to the project. We thought we were dealing with a cleanroom for an engineering research building."

In a span of just four years, as Amick and his colleagues embark on their 14th such facility, nanotechnology research buildings have evolved into a genre of their own. Vibration control, however, is still a moving target. Requirements for two primary nanotech activities, imaging and molecular manipulation, demand an increasingly challenging—yet still attainable—degree of precision.

With this sensitivity such a critical issue, Amick's review of spaces in Duffield Hall offers helpful details about the design and remediation techniques essential to create the appropriate physical environment for the study of minute materials and structures.

Vibration Standards

What's the best way to measure vibration in a building? One general rule of thumb posits that any perceptible movement is excessive for advanced technology research. Delicate nanotech operations involving molecular imaging and manipulation require more specific parameters.

In his evaluation of a partially occupied Duffield Hall, Amick employed the set of recordable vibration criteria (VC) advanced by the Institute of Environmental Sciences and Technology. This scale ranges from 2000 microinches per second, dubbed VC-A, down to 125 microinches per second, known as VC-E.

In general, VC-A and VC-B apply to generic laboratories, while VC-D and VC-E are the most stringent criteria used in the semiconductor market," Amick explains.

While less sensitive nanotech photolithography, imaging, and manipulation can be conducted in the VC-D and VC-E environments, cutting-edge research requires even more rigorous conditions. The National Institute of Standards and Technology (NIST) has refined the VC-E standard even further, incorporating a displacement criterion driven by optics in order to define the stability necessary for imaging at the two-angstrom level or below. The most demanding increments on this scale are NIST-A and NIST-A1.

According to Amick, the two most important questions to ask when assessing the sensitivity of a nanotech environment are what resolution of imaging will be done and whether there will be probe development. Probes, used to grab molecules for manipulation, are often a research focus in themselves. Scientists developing probes typically don't have the resources for vibration isolation, so the lab surroundings must be as quiet and stable as possible.

Peaks and Spikes

Amick's study of the three-story Duffield Hall started with the least demanding spaces from a vibration perspective, in a typical lab on the second floor designed to the VC-A standard.

"It gets very expensive very fast if you try to make multiple floors of a building meet something tighter than VC-A," Amick says.

Measurements showed the lab was on target. The primary excitation Amick recorded was an easily explainable curve produced by the movement of people walking.

The constant-bandwidth plot also displayed some broad peaks and narrow spikes, indicating additional sources of vibration. The broad peaks stemmed from factors beyond human control, ranging from external sources such as the movement of vehicles to the natural building operation—for example, as it responds to the force generated by air flowing through ductwork or water through piping.

The spikes, on the other hand, reflected equipment with rotating parts like vacuum and water pumps or fans.

"The frequency of the spike indicates what motor speed to look for in doing diagnostics," says Amick. "For instance, if a pump is running at 3600 rpm, it will show up as a sharp spike at 60 Hz."

Overall, however, the lab met the VC-A goal and was performing as intended.

Watch Out for Maturation

The results were somewhat different when Amick assessed more sensitive areas in Duffield Hall's first floor cleanroom, where 20,000 square feet are divided into three segregated zones where vibration control becomes progressively more important: process support, photolithography, and critical microsocopy. Measurements showed that the process support area performed more poorly than the VC-A labs overhead.

The process support space is located in the half of the cleanroom built on a suspended slab over the subfab in the basement. In contrast, the other half of the cleanroom sits on a slab on grade, a configuration that creates a "stiffer" floor, providing greater resistance to force and thus minimizing vibration.

However, that is not the reason Amick cites for the process support's poor showing on the vibration chart. Rather, he ascribes the problem to "maturation," a phenomenon that also crops up in semiconductor manufacturing. Essentially, maturation occurs due to declining adherence to original vibration control specifications over the life of a building.

"You have a very good facility to start with, and then over the years it gets worse," states Amick. "Maturation is almost always due to events that occur after construction, such as bringing in new equipment without paying proper attention to isolation techniques. It can also include maintenance problems, such as springs drifting out of alignment."

Once again looking at a chart of vibration frequency measurements, Amick traced the process support area's sub-standard rating to the proliferation of vacuum pumps, which had simply been set on the floor.

"The design for the space was based on the premise that all mechanical equipment would be isolated," he says. "As soon as the people at Cornell were aware of our readings, they knew what the problems were and how to fix them."

Specific remediation measures depend on the vibration frequency emitted by the offending machinery. In the range of 60 Hz or higher, placing the equipment on rubber pads is usually sufficient. Larger gear generating frequencies in the range of 29 to 30 Hz requires the installation of springs.

"The vacuum pumps can be fixed by simply lifting them and sliding a rubber pad underneath," Amick notes. "It's just a matter of finding them all. It's a different story for bigger items that might involve heavy lifting and piping realignment. It can get very expensive to do this as a retrofit, but if you address the need from the beginning, it's a non-issue."

Excellent Imaging Environment

The two other zones of the cleanroom turned in an excellent performance in Amick's studies. Measurements of the area dedicated to photolithography show that it meets the standards of VC-D, and some parts of the space attain a rating of NIST-A. The imaging suite fares even better, performing much better than the criteria for NIST-A.

"This is one of the world's best spaces," Amick says. "Not too many spaces I've evaluated are in the same league. Everything has to come together right to make that happen, and it did here."

Contributing to the high-level performance are factors such as the location of the suite as far from the street and from the building"s mechanical systems as possible. (The latter distance also has a positive effect in minimizing electro-magnetic interference, or EMI).

Breaking Up the Slab

The design of the cleanroom foundation, which challenges conventional notions about the wisdom of a single slab, also plays a significant role in the space's success. The photolithography zone sits on a six-inch thick concrete slab on grade, while the critical microscopy zone is subdivided into three sections that are further separated from their surroundings by being sited over concrete islands three feet thick. In addition, a separation joint around each island isolates it from the effects of peripheral movement such as walking.

Amick explains that this approach represents a departure from traditional thinking, which held that the use of islands to break up a uniform slab would reduce its stiffening and make the vibration environment worse. However, new information has emerged to document the improved vibration control of the island configuration--as long as the islands are a different thickness from the slab itself.

"Several years ago our data discouraged the use of islands," Amick relates. "But some literature in the microscopy world advanced the idea that a two- or three-foot-thick island offered more vibration protection, so if something dropped on the slab 10 feet away, it wouldn't have much impact. Now we have enough measurements to confirm that the island is a better choice."

Hush!

Acoustic metrics are based on the human-driven scale of NC, noise criteria, with the quietest environments reflected by the lowest numbers. Amick's Duffield Hall readings populated a range from NC-70 down to NC-25. At present, the best recording studio has earned a rating of NC-17.

Although taken at a later stage of occupancy than the industry norm (the building start-up, where mechanical equipment is running but research tools are not), Amick's measurements indicate the spaces come in close to their original targets. A cleanroom area without tools registered just a few points above the goal of NC-55; with a fume hood in operation that figure jumped to NC-65. The electron-beam suite measured NC-60, an increase of five to seven points over the anticipated rating for start-up space.

Sections of the cleanroom that house process equipment rated the highest, from NC-65 to NC-70, due to equipment noise as well as the massive amounts of air cycling through the space. While that high noise level does not affect the work conducted in the area, it does make it difficult for occupants to converse.

"If you are doing something that is communication critical, you need to do it somewhere else," comments Amick, noting that "it is impossible to achieve a quiet cleanroom in the range of class 100 or better. It gets noisier as the class number gets smaller."

As with vibration, noise control is most critical in the microscopy suite, where even a whisper can distort an image. The two microscope rooms are performing as intended, reaching an average between NC-25 and NC-30.

Cornell hung acoustically absorbent material wrapped in nylon fabric, AlphaFlex™ Banners, on all four walls, in front of the door, and below the ceiling to help noise decay more quickly.

"It won't inherently make the room quieter because the primary noise source is the incoming air," says Amick, "but if there is a perturbation, such as an operator clearing her throat, it will not last as long."

By Nicole Zaro Stahl