Smaller Keeps Getting Bigger: Marvell Lab Kicks into Gear

By Gordy Slack

It was 1962. The 1,200-square-foot Integrated Circuits Laboratory (ICL) first opened in UC Berkeley’s Cory Hall. Integrated circuits were brand new. Few people dreamed that they would turn the world’s economy on its head, launching a new chapter in the history of technology. One Berkeley engineering professor had an inkling, though: Don Pederson, the driving force behind ICL, the world’s first university-based integrated circuits lab. ICL and the Microlab, which was established in 1983, birthed numerous innovations. Silicon was first etched into micromotors, new 3-dimensional FinFET transistors led the way to scaling beyond the limits of ordinary transistors, and microfluidic self-assembly techniques enabled the integration of RFID tags with a wide range of products. Work conducted in the Berkeley Microlab also gave rise to the computer program that fueled the design of new chips around the world, Simulation Program with Integrated Circuit Emphasis, or SPICE. The Microlab has enabled $320 million dollars in research funding in the past decade alone. It has also been the training theater for more than 2000 MS/PhD/ and Postdoc students over the years. More than 1,300 undergraduates were trained here as well.  

The Marvell Nanolab at Sutardja Dai Hall.

By the late 1990s, however, it had been two decades after its last major facelift, and the Microlab needed to modernize and expand. Several fundraising efforts were launched and then retired in frustration. Finally, a decade ago, when the plan for CITRIS was proposed to the State of California, the next-generation of the microfabrication laboratory became an integral part of its proposal. CITRIS would need a world class Nanolab for its work on sensors, smart dust, bioMEMS, and integrated networks. Part of the promise of the proposal, though, was that the lab would continue to be available to researchers beyond the CITRIS community as well. State funding that established CITRIS also attracted private donors to the project; the Sutardja Dai and the Banatao families, in particular, made large enough gifts that the lab could become a world-class nanotechnology center.

Today, the latest descendant of Pederson’s lab— the two-story, 15,000-square-foot Marvell Nanofabrication Laboratory in CITRIS’s $130M headquarters—is poised to give rise to the next revolution in small. “At any given time, we have at least a dozen projects in the lab that have the potential to become the next large-impact success,” says Ming Wu, Berkeley EECS professor and Faculty Director of the Marvell Nanofabrication Laboratory.

Only time will tell what industries and how many jobs will emerge from the new lab, says Wu. But its predecessor provided critical technology support for 86 companies, including 76 start-ups that together are responsible for about 1,000 technology jobs in California.

Advances in robotics, wireless communication, smart dust, ubiquitous computing and smart materials are all growing by leaps and bounds. And, says Wu, “with the advent of low-cost, personal genome analysis, this is also the dawn of a new biotechnology age that will partly come about in the Marvell Lab.”

Equipment at the NanoLab, including this Atomic Layer Deposition, is moving in on an ongoing basis to accomodate users' needs.

Equipment began to make its way last August from the site of the Microlab in Cory Hall to the new Marvell Lab in Sutardja-Dai Hall. But do not ask Bill Flounders, the lab’s technology and operations manager, when the move will be finished.

“First,” he says, “we do not call it a ‘move,’ we call it a ‘migration.’ ‘Move’ suggests down-time, and we are not willing to accept that. Second, although we will be situated in CITRIS and out of Cory by the end of the year, work on the lab is never finished. We are always anticipating what our users’ needs will be and evolving toward them.”

The same technology used to etch tiny circuits into silicon chips now includes numerous other applications. The Marvell Lab is capable of thin-film deposition and etching, photo lithography, pattern transfer, and doping: all techniques that require both specialized equipment and an extremely clean environment. Whereas these processes were originally employed by electrical engineers, the users are now more diverse, including researchers from mechanical engineering, chemical engineering, molecular biology, chemistry, and physics.

Of particular interest to CITRIS is the lab’s development of micro-electro-mechanical systems (MEMS), which are developed there using modified semiconductor fabrication techniques. These tiny sensors are central to CITRIS’s work in monitoring the environment, health, and energy production and use. Access to the lab will allow this work to continue and to accelerate, says CITRIS Director Paul Wright.

The Marvell Lab is already a magnet for industrial research, drawing both fledgling entrepreneurs who rely on it as a innovation center and testbed for their new ideas, as well as industry leaders such as Intel and HP, who use the lab to explore new materials for more advanced transistors and other micro electronic devices. With an operating budget of approximately $3.2M per year, the lab breaks even by distributing costs according to use across its members. It currently has about 500 active members, including 100 PIs from across CITRIS and beyond.

“The lab encourages cross-pollination by bring researchers from a diverse range of research areas into a single center,” says Flounders. “Scientists have access to a much wider complement of research tools than they would have if using the resources exclusively in their own laboratory, so they are able to apply for grants they would otherwise not qualify for,” he says.

In February, for example, the National Science Foundation funded the Center for Energy Efficient Electronics Science (E3S), a new group dedicated to achieving a radical reduction in the energy consumption of information processing through a redesign of the basic logic switch. The initial grant of $25M for 5 years is extendable for a further $25M in a second 5-year term. Professor Eli Yablonovitch, the Center’s Director, calls the Marvell NanoLab a “regional resource and a key part of the work our team will pursue.” Devices in each of E3S’s main thrust areas will be fabricated in the Marvell Lab.

Professor Ali Javey pursues novel nanotechnology research, including the development of next-generation photovoltaics, in the Marvell Lab.

Another example of work conducted at the lab is that of Ali Javey, a professor in EECS at UC Berkeley, whose research focuses on the application of nanomaterials to sensors, energy harvesting, and programmable matter. Javey is developing inexpensive photo-voltaic cells that can be grown on different materials and could be orders of magnitude less expensive to produce than silicon. The Marvell Lab provides essential infrastructure for much of Javey’s research, which bridges electrical engineering, materials science, chemistry, and physics.

“The diversity of users makes the lab an exciting work environment,” says Wu. “And CITRIS is the right place on campus for the lab, partly because the cross-fertilization of ideas is so much a part of both CITRIS and Marvell Nano Lab cultures.”

By this coming December, the migration will be complete, everything should be in place, the new Marvell Lab will be fully functional. Flounders will still have to work long days and weekends to keep the lab running at its peak potential and anticipate the needs of its diverse users. However, at least he can stop running back and forth between the two buildings. Though construction will be complete by the end of the year, the Marvell Lab will not be finished until it is retired by its successor in—if past is any indicator— another twenty years or so.