Duffield Hall is the first of three new facilities at the core of Cornell's Life Sciences Initiative. Dedicated in October 2004, the $60-million, 155,000-gsf building supports the University's Advanced Science and Technology Initiative (ASTI), housing teaching and research programs in materials science, nanotechnology, micromechanics, electronics, and optoelectronics. A nanofabrication facility provides space for production of semiconductor devices, optoelectronic devices, micro-electromechanical systems, and polymers, and for nanotech biomaterials and molecular construction research.

The building design includes dedicated imaging rooms for scanning electron microscopy (SEM), transmission electron microscopy (TEM) and scanning transmission electron microscopy (STEM) on the first floor, while the second and third floors house mainly laboratory and office space for faculty and graduate students. The main electrical room is located in the basement.

Measuring and Mitigating EMI

Current-generation SEMs, TEMs, and STEMs typically require EMI levels of 1 milligauss (mG) or less in order to perform as designed; many of the next-generation technologies on the horizon will require significantly lower levels (0.1-0.3 mG). Since levels for even current-generation nanotech equipment are significantly below any that has been associated with increased health risks, performance requirements for high-tech instrumentation are often the leading driver in determining acceptable EMI levels in areas housing those tools.

Patterns of electromagnetic interference (EMI) in a given space can be extremely complex, especially in areas where multiple sources are contributing significantly to the EMI "landscape." For this reason, predicting real world EMI levels based only on pre-construction ambient site measurements, project drawings, and equipment specifications—although an important part of planning any nanotech facility project—is almost never enough.

Once construction is completed and equipment is in place, empirical measurements can be taken. This is where the good news begins: EMI is proportionally linear with respect to load. This means the calculations necessary to resolve any remaining interference issues are straightforward.

"In other words, if you double the load in a building, all you have to do is double any measurement you've taken to get the projected interference level at that location at the new load," says Vitale.

Vitale's assessment of Duffield Hall focused on measuring ambient direct current (DC) emissions (0-10 Hz) and alternating current (AC) extremely low frequency (ELF) emissions (60 Hz). This included measuring emissions in the switchgear room—substations, transformers, panels, conduits, etc.—as well as from the transformer/substation bus to the ceiling and into the first floor area above, and checking for ground current problems on all available grounding conductors. He also measured levels in upper floor labs and offices, including emissions from electrical rooms and utility corridors, in the ground floor cleanrooms and imaging areas, and around the external perimeter of the building. Measurements were taken at 17 percent load.

(Radiofrequency interference from broadcast signals, cell phones, and microwave radiation can also cause both increased health risks and operating problems for high-tech instrumentation. The original plan for Vitale's assessment of Duffield Hall included measuring RF levels, but this task was ultimately cancelled for logistical reasons.)

Based on his measurements, Vitale made mitigation recommendations where the data indicated that projected EMI levels under normal operating circumstances would exceed appropriate thresholds.

Main Electrical Room and Surrounds

The building's main electrical room houses a 2500 kVA substation and 13.8 kV primary feeders, which Vitale describes as typical. Levels inside the room are predictably fairly high, with a maximum of almost 150 mG next to the switchgear itself. In the hallways just outside the room, levels are generally less than 10 mG, with a peak of nearly 70 mG (again, at the point closest to the switchgear) and smaller spikes in a few places due to equipment panels.

"The point here is that levels in the hallway outside a switchgear room are not that high—not acceptable for research, but generally not a health concern either," says Vitale.

The situation is more serious above the transformers, switchgear, and busses, where levels can be extremely high—including into the floors above. Directly on top of the main bus, Vitale measured an EMI level of 811 mG. On top of the first floor slab above, readings were about 13 mG. This means that under a normal load of 60 percent, they would be over 40 mG. Two meters higher—i.e., about head-height for a tall person standing—the levels at normal load would be close to 20 mG.

"That's a bit high for long-term exposure. It's nothing very serious, but I would have shielded the floor. In most buildings you would put a flat aluminum plate shield in the concrete slab, and then you will never have an EMI problem above the switchgear," says Vitale.

To minimize health risks from EMI, Vitale recommends a maximum long-term exposure level for human beings of 5 mG in offices and 10 mG in labs, based on many years of experience and work on thousands of buildings. Both figures are significantly below common industry standards.

"You have to do the right thing about health. I have seen cancer cases— clusters of cancer cases—in people exposed to just 100 milligauss over several years," he says.

While checking two lighting transformers, Vitale discovered a problem with a grounded current: current levels of nine or ten amps near ground busses, where levels should be below one amp. This indicates that one of the neutral lines of the lighting circuit is making contact with the ground, so that the current is returning via an alternate path.

Fortunately, University planners labeled every conduit, which according to Vitale means they should be able to resolve the issue in one or two days.

"Label every conduit in the building. It can mean the difference between minutes and weeks in the time required to find and fix a problem," says Vitale.

According to Vitale, grounded currents are actually quite common. He estimates an error rate of three percent on most buildings, generally due to National Electrical Code (NEC) violations.

"Ground currents are also a significant source of interference, and are difficult to shield. About eight years ago I discovered the secret to shielding ground currents: a five-sided magnetic shielding box. A single flat plate over a conduit of the ground current has no effect at all, but if you cover five sides, it attenuates it," says Vitale.

Offices and Labs

By placing most lab and office spaces on upper floors, away from the building power source, Duffield's planners reduced occupants' EMI exposure. This helps minimize the requirement for additional mitigation measures.

Vitale's measurements of EMI levels on the second and third floors showed that they are extremely low, with offices and labs throughout registering less than 0.1 mG, and less than 0.04 mG (the gaussmeter's minimal resolution) in most environments.

Hallways are generally quite low as well, but a small magnetic field of 25-30 mG shows up in one location. According to Vitale, planners didn't realize that they would have panels and transformers adjacent to the wall in one electrical room. Under full load, levels will be 100-150 mG, which is high enough to be a problem. A similar problem shows up near a transformer adjacent to one lab. In both cases, Vitale recommends shielding on the adjacent wall to mitigate the problem.

Nanotech Labs and Nanoproduction Area

Most of the critical research and nanoproduction areas in Duffield Hall are located on the first floor. Vitale's measurements of AC ELF interference showed extremely low levels inside the semiconductor cleanroom—less than 1 mG in most areas, and many below 0.1 mG—and in the adjacent biosciences area. Due to some local peaks, Vitale recommends a three-sided wall-plus-floor shield for the biosciences lab, and an L-shaped shield for the semiconductor room.

In the TEM room, levels ranged from 0-3 mG, but the higher levels are restricted to local areas. Near the TEMs themselves levels were 0.5 mG or lower.

From the standpoint of minimizing EMI levels, the most impressive part of the facility is the SEM/STEM area. AC ELF interference throughout the area, which consists of two adjacent rooms, is essentially non-existent.

"I have never measured a zero SEM/STEM area. To achieve that, they used PVC conduits in the wall," he says.

Vitale also measured DC fields in the SEM/STEM area and found them to be generally acceptable, except for some unexplained spikes.

"This is one of those complicated rooms. We discovered that whenever they turned on the magnetic lens of the STEM, it had an effect on the DC fields throughout the area. When you put tools too close to one another, one of the rooms has to be shielded. The magnetic fields from one tool will affect the others," he says.

Mitigation: Avoidance, Active Cancellation, and Passive Shielding

In many cases, EMI can be avoided simply by moving the source of the interference away from the area where it is causing a problem. In Duffield Hall, for instance, planners put the switchgear room in the front of the building and the sensitive equipment at the opposite end. While such a strategy can be a good starting point, most facilities will require further mitigation measures. The choices are between active mitigation systems and passive shielding.

"Active cancellation technology can be effective in addressing uniform fields such as transmission lines, distribution lines, or conduits carrying a current. It does not work in places where you have a complicated or mixed set of currents coming from different directions. For that kind of situation, the only solution is really a passive approach," says Vitale.

According to Vitale, magnetic shields work three dimensionally: wherever the field is, it works on that particular area. But near-field magnetic shielding is very different from far-field, where requirements can be mathematically calculated.

"Lately, it has been very difficult to sell shields in the scientific world. In some ways, shielding is almost like a black art, and understanding it comes from years of experience. But, it's not that complicated when you know what you are doing, and it's not expensive. A lot of companies will spend $3 million on a tool, but they won't shield it," says Vitale.

For companies that do invest in EMI shielding, Vitale has a warning:

"When you buy a shield, get a performance guarantee. If you don't, you are going to be very sorry," he says.

By John Treat

We welcome your Questions and Comments

Copyright 2008 Tradeline Inc.
All Rights Reserved
ISSN: 1096-4894