by Michael A. Fury, Techcet Group
May 27, 2009 – The Spring 2009 meeting of The Electrochemical Society kicked off on Memorial Day, May 25 at the San Francisco Hilton with an official attendance 1629, about half of normal — yet another testament to the depth of the current global economic downturn. The number of last-minute presentation cancellations seemed to this observer to be greater than at the MRS meeting last month — since ECS is a smaller organization than MRS, the perceived impact is huge. By mid-afternoon, many symposia were running ahead of schedule due to cancellations, and rooms that were set up for 300 people drew 30, though there were some notable exceptions.
CMP Symposium
The one-day CMP symposium opened with a review from yours truly of significant milestones in the first quarter-century of CMP. (Old people do these sorts of things from experience; youngsters just look them up on the Internet.) For the first time in the history of CMP consumables market statistics, the Techcet Group is forecasting a decrease in CMP slurry and pad revenues in 2009. In previous downturns, CMP revenues slowed in growth rate but continued to climb nonetheless as CMP expanded throughout semiconductor manufacturing. The revenue decline along with other process material sectors is a clear indicator of market saturation and maturity.
C. Hsu of Feng Chia University in Taiwan presented some Cu CMP work using nano-polystyrene particles as the abrasive in a citric acid/H2O2 slurry. Acceptable removal rates were achieved with fewer Cu scratches on blanket wafers. Follow-up work needs to include patterned wafers to characterize dishing and erosion effects.
Rob Rhoades of Entrepix presented work in collaboration with Medtronic on CMP for direct wafer bonding of hermetically sealed structures for implantable medical devices. Engineers accustomed to working with demanding semiconductor specifications still find reasons to cringe when adding FDA requirements to their repertoire. Excellent bonding was obtained with several different silicon and glass substrates and combinations thereof.
F. Liu at Applied Materials presented a framework for CMP of phase-change memory materials, demonstrated on PVD Ge2Sb2Te5. This material is softer and more fragile than Cu, and is prone to delamination if subjected to an inappropriate chemical environment. A commercial slurry and PCMP cleaning solution were used to demonstrate a viable process window, though the suppliers of these materials were not disclosed.
Y. Chen of Applied Materials revisited the process alternatives for PCMP cleaning of Cu features in a hydrophobic low-k dielectric environment. The different propensities for the Cu and the low-k surfaces toward particle adhesion require that an integrated approach be used in defining the details of the clean-rinse-dry sequence. A combination of brush cleaning and Marangoni IPA drying, with careful attention to maintaining wafer wetness during the transfer step between the two, was found to be optimum.
E. Shalyt of ECI Technology demonstrated the use of near-infrared monitoring for both slurry and PCMP solutions in real time, taking advantage of the NIR transparency of the industry-standard Teflon distribution plumbing.
R. Nagarajan of IIT Madras revisited the use of ultrasonics for PCMP cleaning of low-k dielectric surfaces. Frequencies <80kHz are still avoided for fear of ultrasonic cavitation device damage, but the acoustic streaming associated with megasonics >200kHz lack the mechanical aggressiveness needed for removing the smallest slurry particles that may have been negligible at 65nm, but not at 32nm. The Taguchi DOE work identified a process window in the 100-200kHz range that provided a mild cavitation boost to the megasonics that was effective in meeting the target specifications.
G. Banerjee of Air Products reviewed the PCMP challenges for nodes beyond 45nm. The materials being evaluated for Cu barriers at 45nm (Ta/TaN), 32nm (Ru, Ag) and 22nm (Ru, Mn, Co) each present unique challenges for tuning the PCMP chemistry so that there is no significant galvanic current at the interface of the barrier with the Cu. While thermodynamics provides the traditional guideline, in practice the ability to retard the corrosion kinetics enables formulators to open some process windows that are thermodynamically ill-advised.
CNTs, fuel cells, batteries
The non-CMP world was also well-represented on this opening day. The University of Notre Dame sent a pair of papers on combinatorial materials discovery methods for methanol fuel cell catalysts. Y. Zhang presented work on a quaternary Pt-Co-V-Cr system that screened concurrently for methanol oxidation activity, oxygen reduction activity, and corrosion resistance; the researchers observed best overall performance in compositions that tended not to form crystalline intermetallics. M. Jeon presented similar work on a ternary Pt-Ni-Cr system.
M. Steiner at IBM Watson Research presented a body of work on the optical properties of carbon nanotubes for integration into nano-photonic devices. The huge variability in individual CNTs’ properties can be reduced in practice by placing the emitting CNT in a micro-cavity between two reflecting mirrors, which results in spectral narrowing and greater consistency between devices — opening the door to resolving some volume manufacturing considerations.
M. Fuhrer at the U. of Maryland fabricated high-frequency electronic devices with a single CNT element across the Pt/Au and Cr/Au electrodes. Several theoretical predictions were elegantly confirmed, but the fabrication method needs a bit more work. Placing a catalyst site near the electrodes, growing the CNTs, and hoping one and only one of them will lay properly across the electrodes is better suited for Las Vegas than for Silicon Valley.
M. Bockrath at CalTech, in collaboration with startup Tangidyne Corp., showed a nanomechanical resonator fabricated from a single CNT suspended across a gap. Such a device is capable of detecting the presence of a single atom landing on the CNT, provided that the atom does not land on a stationary vibration node.
An invited presentation by E. Takeuchi of U Buffalo entitled “Broad-Based Challenges in Battery Implementation” drew an overflow crowd of several hundred people — more were listening in the hallway than were present in many of the other sessions! This overflow condition persisted through three additional talks (all invited) in a session about “Grand Challenges & Opportunities in Energy Conversion and Storage — Energy Storage in Batteries.”
Michael A. Fury, Ph.D, is senior technology analyst at Techcet Group, LLC, P.O. Box 29, Del Mar, CA 92014; email [email protected].