The CIS Newsletter

Spring/Summer 1996

In This Issue

In Memoriam: A Tribute to David Packard

by Prof. John G. Linvill

David Packard, Wm. Hewlett, Frederick Terman, 1952

Given the land resource which became Stanford Industrial Park, Fred Terman's vision of a community of technical scholars was realized, and our environment coupling graduate study and industry is unmatched in the world. The correspondence in the early 50s between Terman and industry leaders (particularly Packard and Hewlett) regarding the Honors Cooperative Plan (HCP) is very illuminating. The requirements that HCP students have the same admission standards as full-time students, that they be able to complete the MS in about two calendar years, that industry pay the full cost (double tuition) were demanding conditions. Experience has verified the concept. The best graduates from across the country have been recruitable, bringing new talent to the new opportunities springing up here.

I should like to recount a personal experience with David Packard preceding the CIS formation but relevant to its later emergence. The idea occurred to me that a liaison program between our solid-state electronics faculty with their doctoral students and companies working in that then-new field would be interesting and beneficial to both sides. We proposed an annual review of Stanford work in solid-state electronics. It would provide a forum for discussion of new ideas and results, be stimulating for our PhD candidates as potential employees, and provide a modest but unrestricted income to our research program. Dr. Terman was enthusiastic, suggested the "club" be called "The Industrial Affiliates of Stanford in Solid-State Electronics" and that a committee of three be formed to identify and approach members. The committee consisted of Terman, Packard and Linvill. With such colleagues I could not fail!

ERL Educational Wing Plaque

Over the past year we have been exuberant about the scheduled new building enabled by the $77.4 million pledges of David Packard and William Hewlett. It represents a magnificent base for our work into the next century. Moreover, it is a powerful sequel to earlier contributions by Hewlett and Packard. Their first Stanford building was the Educational Wing of the Electronics Research Laboratory provided in 1952 when the annual sales of Hewlett Packard were less than $10 million. The level of courage and generosity represented by that first contribution of these remarkable friends of Stanford has been sustained to this day. Moreover, it is a singular example of what comes both to the institution and to such alumni in a lifelong connection in enterprises of mutual and public interest.

On January 30, Bill Gates spent the day at Stanford sharing his vision for the future including the dedication of the building itself. On that day, Stanford also honored many others who have promoted and supported the vision of developing a Science Quad in the near west side of campus. Bill Hewlett and David Packard, in the front row, both congratulated Bill Gates on his generosity to their alma mater. John Linvill, in our feature article, shares his personal insights concerning the key role David Packard has played for Stanford generally as well to the School of Engineering more specifically. David Packard's death on March 26, 1996 represents a loss of one of our university's greatest supporters and mentors. I would like to add a few words on how that mentorship affected my own career here at Stanford.

David Packard, in addition to his legacy in business and generous support of the community, also served his country as Deputy Secretary of Defense (1969-71) during a very contentious period of social unrest arising from tensions over the Vietnam War. His response to the challenges and criticism related to the need for new funding models to support a stronger industrial base for research was typically direct; he backed up his words with good deeds. Beginning in about 1975, David Packard established a Corporate Engineering group at HP with the specific task of promoting university-industry collaborations. In the June, 1981 issue of the HP Journal he wrote: "I am very pleased about the way these programs are going, and I hope that as time goes along we will be able to initiate similar programs in new areas of interest to our company. I also hope that the experience and success we at HP have had in university-industry cooperative research will encourage other companies throughout industry to develop, or expand, similar programs in the future." Indeed, within a few years broad-scale programs such as SRC were launched and HP was a charter member and key promoter in the founding of CIS.

As an Assistant Professor in 1975, I was the fortunate recipient of several grants from the HP program. In fact, in addition to promoting the rapid development of Computer-Aided Design (CAD) tools like the SUPREM program for process models, the equipment grants helped us develop capabilities for transistor characterization (the TECAP program, a feature article in the same 1981 HP Journal issue) which later became broadly used in the industry. What I remember best about that era however, was the way in which HP helped promote technical reviews that directly connected our students to dozens of potential mentors and ultimately users of our research results. These HP staff would come to Stanford from nearly a dozen divisions and spend quality time in these reviews. In fact, many of our best practices within CIS such as the FMA and SPIE programs, as well as the SRC TAB Reviews, are much a reflection of the visionary path which David Packard promoted two decades ago. That combination of personal dedication and corporate commitment can serve as an inspiration to us all. The long-term benefits of developing people, in fact, the hallmark of the "HP way," is the true legacy that David Packard has helped to foster.

Indeed, the development of people is the strategic role of the university and the industrial contributions to this educational process are many. The FMA program captures the essential spirit of the need both financially to support students and to build a partnership between a student Fellow, an industrial Mentor and the faculty Advisor. I am pleased to report that several of our partner companies have contributed supplemental FMA grants in order to further strengthen their connections to our students' research programs. In the face of ongoing governmental budget cuts and corporate downsizing in terms of basic research, the benefits of creating a strategically linked network of researchers is all the more critical. The interplay between company-supported fellowships and the ARPA-funded Computational Prototyping of 21st Century Device Structures project offers a number of exciting models. The following example is representative of the kind of interactions we hope will continue to grow in the future.

In support of both the CIS research program and applications of parallel computing for the Computational Prototyping program, IBM donated a 16 node SP/1 machine to CIS in 1994. In 1995, they supported two students fellowships that have directly contributed to computational prototyping program. Ken Wang is the developer of a "virtual integrated process" (VIP) simulator for creating 3D structures and spent last summer working at IBM, East Fishkill on information models to help integrate IBM's 3D tools and databases. Bruce Herndon has developed a parallel version of Stanford's PISCES program that runs not only on the IBM SP/2, achieving 550 MegaFLOPS performance on our 16 node machine, but also is portable to Intel, Sun and other PC and workstation environments due to use of MPI standards. He has also ported the PISCES-MP code to the 128 node SP/2 NASA machine at Ames Research, supported under the NAS program, achieving 9.7 GigaFLOPS. NASA sponsored a workshop on March 28-29, 1996 to promote broader connection to the semiconductor industry of the invited speakers, a meeting attended by representatives of more than half our CIS partners. Both Bruce Herndon and Vijay Naik (IBM) discussed progress in parallelization of IBM's FIELDAY code, a 3D device simulator. This parallel 3D tool will provide essential leverage in exploring deep submicron devices. In addition to it running on the Stanford SP/2 machine, NASA has expressed interest and willingness to have the 3D tool available on their much larger 128 node SP/2 as well.

In summary, there exciting new developments at Stanford, especially at the intersection of CIS and Gates buildings. The growing involvement of our industrial partners in supporting students through fellowship programs such as FMA as well as the direct partnerships in support of broader research activities such as the Computational Prototyping project mentioned above suggest a promising future for academic-industrial-government interactions. We strongly encourage your greater involvment in these synergistic developments.

"AMD is pleased to support Kathryn Wilder's graduate studies under Professor Calvin Quate, a co-inventor of the atomic force microscope, a powerful new tool for imaging nanometer scale objects and surfaces. We believe that by working with Kathryn and Professor Quate, AMD has a better window into the possibilities AFM offers microlithography and the future of IC fabrication." Bill Arnold, AMD

2) FELLOW: David Bang, Electrical Engineering Department
MENTOR: Zoran Krivokapic, AMD's Submicron Development Center
ADVISOR: Krishna Saraswat, Electrical Engineering Department
PROJECT: Three Dimensional Process Modeling

"We've been doing work on three dimensional process modeling for VLSI metallization. These include metal and dielectric deposition and etching. AMD has provided some test wafers so I can verify my models.

"At Stanford my thesis work focuses on modeling and simulating metal deposition for VLSI fabrication. AMD has processed some wafers using a test pattern I created over a summer internship. The advantage of having the wafers processed at AMD over Stanford, is that AMD has access to the most recent manufacturing equipment; so my thesis work is more relevant to industry. " David Bang, Ph.D. candidate, EE

"The main achievement of the AMD FMA program was a 3D deposition code we are using to predict feasibilities of future technologies, mainly dual damascene structures. David developed 3D version based on 2D Speedie and interfaced it to commercial gridding software. Results of the work were presented at SISDEP and IEDM conferences." Zoran Krivokapic, AMD

3) FELLOW: Patrick Canupp, Aero/Astro Department
MENTOR: Ralf Brinkmann, Siemens AG
ADVISOR: Robert MacCormack, Aero/Astro Department
PROJECT: Plasma reactor simulation

"We are developing a numerical technique for simulating the neutral gas component of etching plasmas. The fellowship has fostered close interaction between myself and the Equipment Simulation Group at Siemens." Patrick Canupp, Ph.D. candidate, Aero/Astro

Ralf Brinkmann/Patrick Canupp

"Both our experiences with the FMA program have demonstrated to us that a collaboration between CIS and Siemens Corporate Research is possible, even though there are several thousand miles (and nine hours) in between. Especially helpful was the willingness of Peter (Smeys) to come to Siemens quite often; also Patrick's stay a couple of months ago was fruitful." Ralf Brinkmann, Siemens AG

4) FELLOW: David Shen, Electrical Engineering Department
MENTOR: Gitty Nasserbakht, Texas Instruments
ADVISOR: Bruce Wooley, Electrical Engineering Department
PROJECT: Microchip radio

"Texas Instruments was responsible for the fabrication of my radio receiver prototype, presented in the February Solid State Circuits Conference. I have recently met with several members of the Integrated Systems Laboratory at Texas Instruments during the ISSCC conference."

"While working on my chip design, I created design rule checking, layout entry, and layout extraction files in order to use the CAD tools at Stanford to design the chip. The data from the Stanford tools was then translated into a format that TI uses internally." David Shen, Ph.D. candidate, EE

"The semiconductor R&D division of TI has long been involved in collaborative research with Stanford. One of the best forms of such collaboration has been through the FMA program. Geographical limitations have not kept us from staying in close contact.

"I have met with David on several occasions during my on-campus visits, as well as meetings at ISSCC conferences in San Francisco. David also visited TI on one occasion to do a final review of his design. David's research resulted in an IC chip fabricated in TI's BiCMOS technology and shipped to Stanford for testing.

"As a result of the success of our current FMA program with David, we plan to continue our support of Prof. Wooley's research activities by extending another FMA fellowship to the student who will be continuing David's research. The student will be spending the summer quarter at TI -- no doubt the start of another very successful FMA relationship, and I look forward to reporting another success story in the near future." Gitty Nasserbakht, TI

5) FELLOW: John Heinlein, Electrical Engineering Department
MENTOR: Dave Archer, Intel Corporation
ADVISOR: Anoop Gupta, Computer Science Department
PROJECT: Leveraging a programmable protocol processor for accelerated multiprocessor communication and synchronization

John Heinlein

"My Intel mentor and I have been working together for several years now, beginning in 1994 when Intel and Stanford were collaborating on the design of a next generation supercomputer. My role was to design and implement a protocol for high bandwidth message passing communication within a hardware cache coherent shared memory system. More recently I have finished the implementation of this protocol for the Stanford: FLASH (FLexible Architecture for SHared memory) multiprocessor system. I recently traveled to Portland, Oregon, to participate in the Intel Graduate Foundation Fellowship Forum, in which recipients of Intel fellowships from around the country came to describe their research." John Heinlein, Ph.D. candidate, EE

"The FMA program in general, and my mentoring relationship with John Heinlein in particular, show great promise for mutual benefit of industry and academia. Interaction between Stanford students and Intel technical staff has increased the understanding and technical excellence of both. The mentoring relationship offers unique benefits to student and mentor alike. The sole reservation I've noted in this endeavor to date is that there's never enough time available for mentor involvement to obtain the optimal benefit for student and mentor." Dave Archer, Intel

6) FELLOW: Justin Leung, Electrical Engineering Department
MENTOR: Barbara Vasquez, Motorola
ADVISOR: Simon Wong, Electrical Engineering Department
PROJECT: Development of an active substrate membrane probe card

Justin Leung

"My relationship with Motorola began when I had the opportunity to spend a summer working in Dr. Vasquez's group in Phoenix, characterizing known-good-die burn-in technology. Afterwards, we kept in touch periodically and entered into a more formal mentoring relationship this past year as part of the FMA fellowship. One of our interactions included an invitation by Dr. Vasquez to present a talk describing our research at the Motorola Science Advisory Board Associates meeting held in San Antonio, TX last October." Justin Leung, Ph.D. candidate, EE

"Recently, Justin has been the recipient of support through Motorola's university research program sponsored by the Motorola Scientific Advisory Board Associates (SABA), of which I am a member. In his research under the direction of Professor Simon Wong, Justin has been responsible for significant improvements on a novel membrane probe card technology. At our request, Justin presented his work at a Motorola technical review last October. Of all the university research sponsored by the SABA, Justin was awarded the 'Best Student Paper of the Year' award. I had the pleasure of delivering the plaque to Justin during a small celebration of his award at Stanford last December." Barbara Vasquez, Motorola

Stanford's ability to produce master's and doctorate graduates who are among the best in the world depends, in part, on its ability to attract the top bachelor's graduates into the University. Stanford's world-wide reputation ensures that the vast majority of the best undergraduates seeking to go to graduate school in such areas as electrical engineering, computer science and related fields, will apply to, and be accepted by, the University. Yet, acceptance does not always mean enrollment. Most of our highest ranked applicants also apply to, and are accepted by, other top U.S. schools such as Carnegie Mellon University, the Massachusetts Institute of Technology, the University of California at Berkeley, and the University of Illinois. While there is much competition among these schools, it is always Stanford's intention to enroll more than its "fair share" of the best students.

Financial aid is a necessary, but not sufficient condition in such recruitment efforts. We also need to show these students what is so special about Stanford/CIS and why it will be to their benefit to attend Stanford.

When CIS began its recruitment efforts a decade ago we had to explain to prospective students what the Center was about and why it presented them with a unique opportunity. In the last few years the situation has reversed itself, and CIS has become a magnet that actually draws some applicants to Stanford. The Center's outstanding facilities, its Fellow-Mentor-Advisor (FMA and SPIE programs, and the potential for significant industrial interactions, all contribute to the attraction, as of course do its outstanding affiliated faculty and graduate students.

CIS recruitment activity, focusing on top applicants interested in CIS related research areas, is part of a larger Electrical Engineering Department recruitment effort under the direction of Professor Fabian Pease (Admissions Chairman), Lori Fabrizio (Director of Admissions), and Patricia Goodrich (Admissions Administrator).

Personal telephone calls and extensive e-mail exchanges among applicants and electrical engineering students, staff and faculty is a key part of the recruitment effort. So too, are weekly mailings to applicants of The Stanford Daily and Stanford Report newspapers. (We want them to feel they are part of our family even before they enroll).

Of particular significance is the electrical engineering department's EE Recruitment Day, held in March, and to which all applicants are invited. This event draws over 100 applicants, features presentations by faculty, tours of electrical engineering facilities, informal discussions with graduate students, and a panel of graduate students who speak frankly about their Stanford experiences. Our key effort is to make students feel that we really want them to come to Stanford. We realize we have a lot to offer them, and know they have much to offer Stanford. We therefore seek every opportunity to convey this message to them.

So, how did we do? Preliminary results for this year's recruitment efforts (for the incoming 1996-97 class) indicate that approximately 40% of the top 70 applicants have decided to enroll at Stanford. This figure, which could still increase slightly when all decisions are made, is particularly impressive considering that the vast majority of these 70 applicants also applied to the other top schools mentioned above. Final figures for women and minority students are not yet in, but preliminary suggest an increase over last year.

Of course, having the top students decide to enroll at Stanford/CIS is only the beginning. In our next issue I will highlight some of the things CARES is doing to enrich their experiences while at CIS.

Philips, looking to pull in some of the leading ideas with no direct applications as yet, values university relationships, and takes full advantage of the collaboration between industry and academia, an environment which allows these ideas to develop. According to Bar-Gadda, the Stanford atmosphere, especially the work being done at Stanford on RISC (Reduced Instrution Set Computing) and the MIPS chip, influenced both the compiler and the software architecture of TriMedia.

Bar-Gadda stated: "I'd liken it to a biological process, with cross-pollination between industry and the university, particularly in a world of support for product development people. The path is sometimes direct and conscious, sometimes indirect and unconscious, but the Stanford University environment provides you with an atmosphere of creativity in which new ideas jostle people's brains and lead to a chain reaction. Of course how it happens is pretty mysterious, and you can't point to any moment, any single event, in particular, as it is a stochastic, and not a deterministic process, but the academic context, the chemistry that occurs being near a major university, is what makes it possible.

"If one is only in, say a corporate, environment, one's view of the outside world is filtered by a certain viewpoint. A university connection lets those ideas the space and freedom to roam, to explore. And while it is very difficult to trace exactly how the ideas developed over the past 10 years, in that Philips did not benefit directly from their CIS visiting fellows at any moment in particular, those people learned things by being at CIS, and they communicated their thoughts, in both direct and indirect ways, to people at Philips."

According to Bar-Gadda, "Looking for high-compute performance beyond RISC, we got into multimedia -- a technology looking for an application." Knowing that VLIW architecture was capable of executing many operations within each clock cycle, it was used to develop accelerated video processing. "We brought in Julien Labrousse, a Philips designer with experience in interactive compact disc (CDI) chips, and the group ended up creating a multimedia processor, originally termed LIFE, first publicly announced at the 1989 Philips Concern Research Exhibition in Eindhoven, and now available as Trimedia."

TriMedia, the first fully programmable DSP/CPU, is capable of concurrently processing audio, video, graphics and communications data all on one chip, in what Philips terms/calls "concurrency." Instead of adding DSP features to a CPU, which would simply add media capability, but not result in multimedia, the TriMedia chip adds CPU features to a DSP, with the result that the TriMedia chip is fully programmable. In the course of their interactions at Stanford over the past 10 years, Bar-Gadda and others at Philips saw that the technology being developed in connection with Stanford had promise and a potential market presence. Their contact with CIS assisted in the growth and development of that idea, to say the least, and so gets some of the credit.

John DeGenova (TI), graduate student Simon Fang, and Bob Helms with Ellipsometry System

The vision for the Center is to create a unique and focused program in which faculty and students from various disciplines and institutions come together with industry and government agencies, and work toward these objectives:

• Conduct fundamental research studies which will provide the basis for making rational changes in semiconductor processes for environmental improvement.
• Develop a methodology for incorporating the environmental safety and health factors in developing new processes, new tools and new protocols for semiconductor manufacturing. The focus is on process modification, factory integration, chemical use minimization and recycling.
• Apply this methodology to specific processes which are of significant environmental concern.
• Develop new educational programs and modify existing ones to teach environmental factors as an integrated part of the engineering courses and projects. The purpose is to train a generation of engineers and scientists who will help industry achieve its long-term environmental goals.
• Provide a pre-competitive forum for bringing experts from industry, research institutions and regulatory agencies together to promote a pro-active and preventative approach to environmental concerns on semiconductor manufacturing.

The Center management team consists of the Center director (Farhang Shadman, professor of chemical and environmental engineering, University of Arizona), and two associate directors (C. Robert Helms, professor of electrical engineering, Stanford University and Rafael Reif, professor of electrical engineering, MIT). The Stanford team includes Bob Helms as overall Center co-director, Chris Chidsey of the chemistry department, and Greg Kovacs and Jim McVittie in electrical engineering.

An essential issue for the semiconductor industry is the development and use of environmentally sound manufacturing processes. One focus is the large amount of highly purified water required to rinse and clean a silicon wafer. The semiconductor industry is interested in reducing water use through innovative technology development. There is also a need to minimize the amount of energy required to manufacture microchips.

The new Center will work in partnership with firms from the semiconductor industry. These firms will interact with faculty and students in research, and provide opportunities for student interns both to learn the manufacturing process firsthand and also to contribute solutions for industrial problems while they are still in school. ERC for EBSM graduates will be part of a new generation of engineers capable of integrating environmentally conscious concepts up-front in the design of manufacturing processes.

For more information, contact:
Prof. Bob Helms
helms@ee.stanford.edu

In addition to Rissman, HP Labs speakers included Richard Marconi, Manager of technical communications, who gave an Introduction to HP Labs, Art Muto, Project Manager of the High Speed Electronics Department, Wes Higaki, of the Measurement Systems Department, Ed Middlesworth of the Intergrated Circuits Business Division's (ICBD) Technology Development Department, Rolf Jaeger, Manager Device Technology Department on Micromachining Research at HP Labs, and Stephen Laderman, Materials Characterization Project Manager.

The presentations highlighted just a few of HP's exciting technologies. Among the recent contributions of HP Laboratories presented were an air-cooled package for the latest Precision microprocessor capable of dissipating 80 Watts, ASICS for interfacing communication systems, new materials and processes for optoelectronics, and a thermally actuated micromachined valve for a gas chromatography system.

A typical day at the HP Labs

In addition to the presentations,Russ Parker and Paul Rissman led four groups of approximately ten students each on thirty-minute tours of both the Measurement Research Center and the Computer Research Center.

The Measurement Research Center has technical programs in photonics, high-speed electronics, measurement systems, medical and analytical instrumentation and systems, epitaxial materials, micromechanics, device processing, precision instruments and manufacturing technology.

Programs in the Computer Research Center focus on computer architecture, printing technology, mass storage, database management, object technology, multimedia systems and software engineering.

The visit, tours, presentations and meals were appreciated by all in attendance, as evidenced by the plethora of questions. The food was great, the interchange of questions and answers encouraging, and the interrelationship between the basic research (invention) occurring at universities including Stanford and the applied research (innovation) occurring at HP Labs was much in evidence.

Editor:
Harrianne Mills
650/725-3626

Send comments, suggestions to: coordinator@cis.stanford.edu

Updated 9/19/96