Nanoelectronics December 2011/January 2012 Viewpoints
2011: The Year in Review
Several segments of importance to nanoelectronics came under strain in 2011, notably solar cells but also hybrid and plug-in electric vehicles, despite rapidly falling battery prices. The solar-cell business underwent substantial consolidation during the year, with oversupply and falling prices leading to several high-profile failures, including the demise of Solyndra LLC (Fremont, California), and this development may well deter investors from funding new approaches.
Sales of hybrid and full electric vehicles in 2011 continued to lag expectations, even though nanoelectronics approaches remain critical to improving the performance and economics of battery powering. Instead of gearing up for a substantially growing hybrid-electric-vehicle business, battery makers are looking toward growing demand for 12-volt stop-start batteries. Though this segment does not require advanced lithium-ion batteries and nanostructured electrodes, developers such as A123 Systems (Waltham, Massachusetts) are beginning to address this growing segment. In its yearly guidance in November, A123 Systems projected record revenues for the year—up 70% to 85%, to reach between $165 million and $180 million. However, the company stated that a major customer, Fiskar Automotive (Irvine, California), had unexpectedly lowered fourth-quarter orders because of its high inventory.
In a number of other segments, nanoelectronics developments were very positive, even if commercialization remained limited. Several new players ramped up research on spin-transfer-torque random-access memory (STT-RAM) as a successor to flash memory, and Crocus Technology (Grenoble, France) formed several alliances to commercialize its nanomagnetic technology, which can integrate memory and logic components monolithically. Cambrios Technologies Corporation (Sunnyvale, California) started commercializing its printed silver nanowires as a replacement transparent conductor for use in smartphones, and both Nanosys Inc. (Palo Alto, California) and Nanoco Group PLC (Manchester, England) demonstrated the possibilities for quantum dots in liquid-crystal-display (LCD) backlighting units and lighting.
Electronics and Opto: InGaAs and Nanowires
Intel's decision in May 2011 to move to a 3D FinFET gate structure for the 22 nm node bore out the conclusion by ITRS (International Technology Roadmap for Semiconductors) working groups that near-term changes to the conventional planar metal-oxide-semiconductor field-effect transistor would likely involve different geometry but rely on the same silicon platform and materials. But what will come after 3D silicon? During the year, the ITRS also decided to move the development oversight of two maturing n-type and p-type channel materials—indium gallium arsenide (InGaAs) on the n-type side and germanium for future p-type devices—to the Front End Processing work group and Process Integration, Devices, and Structures working group.
In December, at the 2011 International Electron Devices Meeting in Washington, DC, Mark Bohr discussed the transition from an era of homogeneity (silicon) to heterogeneity (mixed materials), reporting that Intel is working on new III-V channel materials, particularly InGaAs, that provide a factor of 30 improvement in electron mobility. Bohr also noted that Intel engineers prefer to work with low-voltage channel materials that give good performance rather than with higher-voltage, higher-performance materials, pointing out that InGaAs allows a working voltage as low as 0.5 volt.
The increasing interest in InGaAs as an n-type channel material for future IC making has caused a number of organizations to take a second look at monolithic optoelectronics integration—given that InGaAs is an optically active material. During the year, several organizations demonstrated new approaches to integrating this III-V material, which suffers from a lattice mismatch with silicon and a high growth temperature that is incompatible with silicon processing. Probably the most important development in the year came from the University of California, Berkeley (Berkeley, California), where researchers developed a monolithic back-end-of-line (BEOL) process to create nanoscale pillar lasers of InGaAs, with gallium arsenide shells, on a silicon substrate at a growth temperature of only 400°C. The implications of this development could be substantial, for the BEOL process does not interfere with the other steps of IC-device making. However, commercialization is still years away and will require changes in chip architectures.
Two IC applications for nanoscale self-assembly also caught my eye during the year:
- With Intel already making 3D transistor structures, and with on-chip interconnects remaining a serious issue, the IC community is looking more seriously at stacked chip architectures, which would likely first incorporate processor and memory ICs. Such an approach would depend on through-silicon-via interconnects, and Intel has described bottom-up growth of vertical nanowires, including carbon nanotubes, as a promising approach.
- Directed self-assembly (DSA) already has seen some use in IC fabrication, with IBM (East Fishkill, New York) describing its use, in 2007, to create a low-dielectric "air-gap" material. In March, JSR Micro, Inc. (Sunnyvale, California), together with research partner IBM Almaden, announced the development of a sub-20-nanometer half-pitch node. The half pitch builds on the use of conventional 193 nm lithography but avoids the dual exposure step, thereby having the potential to reduce processing times and lower the cost of ownership. JSR also sees DSA as a complementary approach to other emerging lithographic technologies, such as extreme ultraviolet lithography.
In 2011, new magnetic memory stole the limelight as a series of announcements suggested that modified magnetoresistive random-access memory (MRAM) and STT-RAM are beginning to gain some traction as next-generation nonvolatile memories—a development that is in line with the latest update of the ITRS. This shift toward MRAM and variants looks likely to be at the expense of phase-change RAM (PRAM), which has suffered as research suggests NAND flash will be viable for another five years at geometries below 20 nm. PRAM, which offers only limited scaling itself, may remain stuck in niche applications: Samsung revealed during the year that system designers have been reluctant to adopt the technology in volume.
The year saw the following important announcements about memory developments:
- Samsung acquired STT-RAM intellectual-property (IP) developer Grandis Inc. (Milpitas, California) in August. Founded in 2002, Grandis has received considerable funding from the US Defense Advanced Research Projects Agency and has IP-licensing agreements with companies such as Renesas Electronics Corporation (Tokyo, Japan) and Hynix Semiconductor Inc. (Icheon, South Korea). Also in the year, Samsung announced the world's smallest perpendicular magnetic-tunnel-junction cell using 17 nm process technology.
- In July, Toshiba and Hynix announced an agreement to codevelop STT-RAM—another sign that companies are positioning themselves for commercialization. Toshiba also announced that it had used 30 nm design rules to produce a perpendicular magnetic tunnel junction with a low switching-current density of 5 x 105 amperes per square centimeter. Toshiba has seemed increasingly bullish about STT-RAM and appears to be targeting relatively low-density (512 kbit) STT-RAM as a fast, low-power memory hybrid cache for "usually-off" computing.
- The industry's main supplier of toggle MRAM—Everspin Technologies (Austin, Texas)—announced that it would launch an STT-RAM product in 2012, and another leader in the memory industry, Micron Technology (Boise, Idaho), teamed with A*Star (Singapore, Republic of Singapore) to develop STT-RAM.
Though most next-generation memory developments in 2011 focused on STT-RAM, Crocus Technology was also very active during the year in commercializing its novel thermally assisted MRAM technology. In May 2011, Crocus reached a $300 million investment agreement with Rusnano (Moscow, Russia) that will lead to an MRAM-production facility in Russia, and in October, it signed a joint technology and patent-license agreement with IBM. Crocus's thermally assisted switching MRAM technology lends itself to the creation of combined magnetic logic and memory devices for rugged, high-temperature applications. In December, it signed a technology and wafer-manufacturing agreement with China's semiconductor foundry service provider Semiconductor Manufacturing International Corp. (Shanghai, China) to apply Crocus's Magnetic-Logic-Unit technology to the automotive sector and use it in future smart cards.
ITRS working groups are also studying redox resistive RAM as a potential next-generation memory, given its good prospects for scaling, low power consumption, and speed. In this vein, news during the year that HP Inc. (Palo Alto, California) will team with Hynix to commercialize its two-terminal memristor technology was important. The companies plan initial products in 2013.
For many solar-cell companies, 2011 proved a difficult year. Customers benefited from declining prices—due to substantial industry overcapacity and a substantial fall in the cost of polysilicon—but suppliers' profitability was under strain, and a number went bankrupt, notably CIGS (copper indium gallium diselenide) solar-cell developer Solyndra.
In the realm of nanoelectronics, solar cells had a brighter year than the industry backdrop would suggest. Nanosolar Inc. (San Jose, California) began to ramp up production of its printed CIGS solar modules and announced in November that two production sites belonging to EdF Energies Nouvelles (Paris, France) would each install 3 megawatts of modules. In October, Nanosolar also announced that the National Renewable Energy Laboratory (NREL; Golden, Colorado) had certified an aperture efficiency of 17.1% for a CIGS solar cell printed onto a flexible foil substrate. This figure compares with Nanosolar's 2007 report of a small-area efficiency of 14% for a printed solar cell on glass, highlighting the rapid progress the company is making.
News of Nanosolar's progress, as well as the positive development of increased investment in CIGS technology as a whole, was tempered at the year's end by an announcement by First Solar Inc. (Tempe, Arizona) that it would abandon its CIGS research program to concentrate on lowering the cost of its cadmium telluride (CdTe) process. First Solar remains by far the largest provider of non-silicon-based solar cells, with some 45% of the thin-film-solar-cell market.
Reflecting the increasing pressure to improve module efficiencies, Dupont Inc. (Wilmington, Delaware) acquired Innovalight Inc. (Sunnyvale, California) in July 2011 to gain access to that company's silicon-nanoparticle-ink technology. Innovalight's technology, which adds a screen-printed absorbing layer, can increase solar-cell efficiencies by some 1%, and Innovalight has several deals with Chinese cell manufacturers, including JinkoSolar, Yingli Green Energy, JA Solar, and Solarfun.
Another approach to improving cell and module efficiencies is to integrate antireflection coatings on the front surface. Many current approaches use thin-film coatings on glass, but Mitsubishi Rayon Co. Ltd. (Tokyo, Japan) and SolOptics, a division of Genie Lens Technologies LLC (Englewood, Colorado), are each developing nanostructured coatings for this application, either on plastic or glass substrates. Use of plastic substrates allows installers to coat modules in the field with a roller and hence addresses the installed base of solar panels. In January, Mitsubishi Rayon announced a moth-eye biomimetic coating consisting of an acrylic resin with nanopillars 200 nanometers tall and 50 nanometers wide at the top. The company stated that the coating could increase annual efficiency by 5% to 6% depending on latitude. In September, SolOptics announced that NREL had measured efficiency improvements of 4% using its Fusion coating, with the biggest improvement coming at wide angles of incidence.
In 2010, three of the major companies developing organic solar cells—Konarka, Heliatek, and Solarmer Energy—all announced cell efficiencies above 8%. Konarka and Solarmer use wet printing processes, and Heliatek uses vacuum deposition and a tandem cell. In December 2011, Heliatek raised the bar again, this time by a full 1.5%, achieving 9.8% efficiency for a 1.1 cm2 tandem cell, a rate that translates into large-area-module efficiency of some 7.5%, according to the company. Heliatek is targeting mobile, off-grid applications, but with price declines in the mainstream photovoltaics market, all third-generation cell developers continued to struggle to bring out commercial products in 2011.
Quantum Dots for Displays and Lighting
Developments in quantum dots (QDs) were positive in 2011, with most companies focusing on improving LCDs backlit by light-emitting diodes (LEDs). Nanosys's (Palo Alto, California) QuantumRail technology, which the company licensed to LG Innotek (Seoul, South Korea) and Samsung in 2010, has morphed into a technology it now calls a quantum-dot enhancement film, which has a form factor and terminology that are very much in line with the LCD industry's use of enhancement films between the backlight unit and the LCD panel itself.
Nanosys competitor Nanoco Group PLC (Manchester, England) continues to manufacture its quantum dots for display applications, although its approach is to deliver raw materials rather than a more forward-integrated enhancement film. Nanoco is one of the few companies to have developed cadmium-free quantum dots, and in August, it announced a joint development agreement with a major unnamed company to create LED lighting with improved performance. Another company targeting lighting and displays with its quantum-dot light-enhancement technology is QD Vision (Lexington, Massachusetts). In May, QD Vision received new funding of $22 million to invest in a new production facility for its quantum dots, in anticipation of making product announcements in 2012.
The use of micro- and nanoparticles to create inks for printing low-cost electronic materials and components saw several new developments in 2011. Foremost among the promising announcements was the news that Nissha Printing Co., Ltd. (Kyoto, Japan), had incorporated Cambrios's ClearOhm transparent conductive material, consisting of printed silver nanowires, into the touch sensor and module of an Android smartphone. This news suggests that printed electronics solutions may soon challenge sputtered indium tin oxide (ITO) in applications that require a transparent conductor. Another challenger to ITO—graphene—saw development in 2011, with Vorbeck Materials (Jessup, Maryland) announcing that MeadWestvaco Corporation (Richmond, Virginia) had incorporated its Vor-ink conductive ink into an antitheft packaging solution due for commercialization in 2012.
To take advantage of the potential to print low-cost components using roll-to-roll processing, curing must be rapid but also gentle enough not to damage plastic film substrates. In 2011, several companies made progress in rapid curing. Dupont Microcircuit Materials (Wilmington, Delaware) announced that it would join a research program at the Holst Centre (Eindhoven, Netherlands) to optimize printed metallic structures on flexible substrates, including low-energy sintering. And in April, Showa Denko K.K. (Tokyo, Japan) and NovaCentrix Inc. (Austin, Texas) entered into a development, manufacturing, and sales agreement for NovaCentrix's PulseForge photonic sintering platform and its silver and nanoparticle-based conductive inks.
- Batteries. Development of new electrode materials for batteries continued in 2011. Though most development has focused on new noncobalt cathode materials containing nickel, manganese, iron, or phosphate, the year saw the development of new anode nanomaterials that can absorb (lithiate) more lithium ions within the atomic structure without suffering the deleterious effects of lattice expansion. Among the organizations active in this emerging area are Nexeon Ltd. (Abingdon, England), which spun out of Imperial College (London, England); Nanosys; Argonne National Laboratories (Albuquerque, New Mexico); and Hitachi Maxell Ltd. (Osaka, Japan).
- Ultracapacitors. Researchers at Angstron Materials (Dayton, Ohio) demonstrated a graphene-based ultracapacitor with an energy density of close to 90 watt-hours per kilogram of electrode material, some five times higher than that of conventional activated-carbon-based electrochemical double-layer capacitors. And in October, FDK Corp. (Tokyo, Japan) and Asahi Kasei Corporation announced a joint venture to commercialize lithium-ion capacitors consisting of one lithium battery electrode and one ultracapacitor electrode.
- Quantum computing. In 2011, D-Wave Systems (Burnaby, British Columbia) published experimental results that showed, for the first time, that its quantum annealing processor operates via quantum effects, even if not specifically via quantum entanglement. Quantum computing began to penetrate the commercial market during the year as D-Wave announced that Lockheed Martin (Bethesda, Maryland) would purchase its first quantum computer.
Look for These Developments in 2012
- Further consolidation in the solar-cell industry is likely in 2012 if continued oversupply pushes down sale prices. With the climate for introducing wholly new solar-cell technologies becoming very difficult, we can expect more efforts to leverage existing solar technologies, such as new absorbing inks that contain silicon or CdTe. Further development of antireflective coatings or off-angle absorbing nanostructures is also likely.
- The development of graphene as a material for touchscreens, or as a transparent conductor, has increased dramatically in the past couple of years, and prototypes (as opposed to demonstrators) could start appearing in 2012. A company such as Samsung might see the development of a graphene touchscreen as an innovation that would create a buzz in the market, as it continues to challenge market leader Apple Inc. (Cupertino, California) in the smartphone and tablet markets.
- With the decline in plasma TVs, LCDs are now unchallenged in the large-area-screen market, and competitive differentiation centers on aspects such as 3D, color rendering, and brightness. The first commercial quantum-dot-enhanced backlight unit is likely to appear in 2012, and other companies will seek to enter the market, particularly those offering forward integration of QDs into brightness-enhancing films.
- MRAM development has been something of a roller coaster, first with the development of toggle MRAM, which failed to live up to initial expectations but has since seen commercialization in selected niche applications. More recently, it has seen application with STT-RAM, which is gaining some critical mass. In 2012, we may see limited commercialization of low-density devices, but clients should also monitor the use of MRAM devices in combined logic and memory components.