An electronic–photonic microprocessor chip manufactured using a conventional microelectronics foundry process is demonstrated; the chip contains 70 million transistors and 850 photonic components and directly uses light to communicate to other chips. The rapid transfer of data between chips in computer systems and data centres has become one of the bottlenecks in modern information processing. One way of increasing speeds is to use optical connections rather than electrical wires and the past decade has seen significant efforts to develop silicon-based nanophotonic approaches to integrate such links within silicon chips, but incompatibility between the manufacturing processes used in electronics and photonics has proved a hindrance. Now Chen Sun et al. describe a 'system on a chip' microprocessor that successfully integrates electronics and photonics yet is produced using standard microelectronic chip fabrication techniques. The resulting microprocessor combines 70 million transistors and 850 photonic components and can communicate optically with the outside world. This result promises a way forward for new fast, low-power computing systems architectures. Data transport across short electrical wires is limited by both bandwidth and power density, which creates a performance bottleneck for semiconductor microchips in modern computer systems—from mobile phones to large-scale data centres. These limitations can be overcome1,2,3 by using optical communications based on chip-scale electronic–photonic systems4,5,6,7 enabled by silicon-based nanophotonic devices8. However, combining electronics and photonics on the same chip has proved challenging, owing to microchip manufacturing conflicts between electronics and photonics. Consequently, current electronic–photonic chips9,10,11 are limited to niche manufacturing processes and include only a few optical devices alongside simple circuits. Here we report an electronic–photonic system on a single chip integrating over 70 million transistors and 850 photonic components that work together to provide logic, memory, and interconnect functions. This system is a realization of a microprocessor that uses on-chip photonic devices to directly communicate with other chips using light. To integrate electronics and photonics at the scale of a microprocessor chip, we adopt a ‘zero-change’ approach to the integration of photonics. Instead of developing a custom process to enable the fabrication of photonics12, which would complicate or eliminate the possibility of integration with state-of-the-art transistors at large scale and at high yield, we design optical devices using a standard microelectronics foundry process that is used for modern microprocessors13,14,15,16. This demonstration could represent the beginning of an era of chip-scale electronic–photonic systems with the potential to transform computing system architectures, enabling more powerful computers, from network infrastructure to data centres and supercomputers.

/pdf/single-chip-microprocessor-that-communicates-directly-using-3e2ii9mb9i.pdf

Single-chip microprocessor that communicates directly using light

As silicon-based electronics approach the limit of improvements to performance and capacity through dimensional scaling, attention in the semiconductor field has turned to graphene, a single layer of carbon atoms arranged in a honeycomb lattice. Its high mobility of charge carriers (electrons and holes) could lead to its use in the next generation of high-performance devices. Graphene is unlikely to replace silicon completely, however, because of the poor on/off current ratio resulting from its zero bandgap. But it could be used to improve silicon-based devices, in particular in high-speed electronics and optical modulators.

A role for graphene in silicon-based semiconductor devices

Graphene is a relatively new material with unique properties that holds promise for electronic applications. Since 2004, when the first graphene samples were intentionally fabricated, the worldwide research activities on graphene have literally exploded. Apart from physicists, also device engineers became interested in the new material and soon the prospects of graphene in electronics have been considered. For the most part, the early discussions on the potential of graphene had a prevailing positive mood, mainly based on the high carrier mobilities observed in this material. This has repeatedly led to very optimistic assessments of the potential of graphene transistors and to an underestimation of their problems. In this paper, we discuss the properties of graphene relevant for electronic applications, examine its advantages and problems, and summarize the state of the art of graphene transistors.

Graphene Transistors: Status, Prospects, and Problems

This paper presents an overview of process variation effects, including examples of mitigation strategies and test methods. Experimental and theoretical comparisons are presented for 45 nm and 65 nm RDF. SRAM matching and interconnect variation is discussed for both 65 nm and 45 nm, including examples of process and design mitigation strategies. Use of ring oscillators for detailed measurement of within-wafer and within-die variation is illustrated for 65 nm and 45 nm products.

Reducing Variation in Advanced Logic Technologies: Approaches to Process and Design for Manufacturability of Nanoscale CMOS

This paper presents photonic devices with 3 dB/cm waveguide loss fabricated in an existing commercial electronic 45 nm SOI-CMOS foundry process. By utilizing existing front-end fabrication processes the photonic devices are monolithically integrated with electronics in the same physical device layer as transistors achieving 4 ps logic stage delay, without degradation in transistor performance. We demonstrate an 8-channel optical microring-resonator filter bank and optical modulators, both controlled by integrated digital circuits. By developing a device design methodology that requires zero process infrastructure changes, a widely available platform for high-performance photonic-electronic integrated circuits is enabled.

/pdf/open-foundry-platform-for-high-performance-electronic-471vdswkg5.pdf

Open foundry platform for high-performance electronic-photonic integration

We report record RF performance in 45-nm silicon-on- insulator (SOI) CMOS technology. RF performance scaling with channel length and layout optimization is demonstrated. Peak fT's of 485 GHz and 345 GHz are measured in floating- body NFET and PFET with nearby wiring parasitics (i.e., gate- to-contact capacitance) included after de-embedding, thus representing FET performance in a real design. The measured fT's are the highest ever reported in a CMOS technology. Body- contacted FETs are also analyzed that have layout optimized for high-frequency analog applications. Employing a notched body contact layout, we reduce parasitic capacitance and gate leakage current significantly, thus improving RF performance with low power. For longer than minimum channel length and a body-contacted NFET with notched layout, we measure a peak fT of 245 GHz with no degradation in critical analog figures of merit, such as self-gain.

Record RF performance of 45-nm SOI CMOS Technology

The scaling of semiconductor technologies from 90- to 45-nm nodes highlights the need for accurate and predictive compact models that address the regime where small-scale physical effects become dominant. These demanding requirements on compact models extend beyond the core model to a suite of design tools that include extraction tools and statistical methods to account for unpredictable variation (e.g., random dopant fluctuations and polysilicon linewidth variation) and predictable variation (e.g., transistor response differences that are layout dependent). Layout-dependent or local environment differences are driven by factors such as lithography and novel performance-enhancing process techniques such as dual-stress nitride liner films. Sources of variation such as rapid thermal annealing temperature, low-frequency noise, and modeling of back-end-of-line elements need to be considered. The modeling of intradie and interdie variations, updated for small geometries, should be properly positioned in the design flow. This paper presents the challenges and results of compact modeling at the 65-nm node and beyond

Modeling of Variation in Submicrometer CMOS ULSI Technologies

Extremely thin SOI (ETSOI) MOSFET is an attractive candidate for 22nm technology and beyond due to its excellent short channel control, low leakage current, and immunity to random dopant fluctuation [1–5]. Short channel effects are mainly controlled by channel thickness, so there is no need for aggressive scaling of the gate dielectric. Thus the gate leakage is reduced beyond what is achievable in high-k bulk technologies. Low-power operation is further enhanced by negligible GIDL current due to the undoped channel. In addition, ETSOI devices have inherently no junction leakage by the virtue of thin silicon channel. Higher gate voltage overdrive is achieved for a given supply voltage compared to bulk technologies due to smaller subthreshold slope. This enables low-V DD logic operation. Moreover, low-V DD SRAM functionality is supported by small V T - mismatch in undoped channel [5]. In conventional CMOS technologies, complete device redesign is needed if V T changes are required. In ETSOI, however, threshold voltage is tuned through gate workfunction modulation without change in the channel doping. Thus V T tuning is to a large extent decoupled from device scaling.

Fully depleted extremely thin SOI for mainstream 20nm low-power technology and beyond

We report record RF FET performance in 65 and 90-nm silicon-on-insulator (SOI) CMOS technologies featuring measured gate lengths from 27 to 43 nm and analyze factors contributing to that performance The effect of layout and geometry optimization as well as channel length scaling is investigated to improve RF performance, namely fT, and fMAX A peak fT of 330 GHz is measured in a fully-wired 65-nm NFET A complete de-embedding method to accurately determine RF characteristics of the intrinsic 90-nm SOI NFET results in a peak fT of 290 GHz and an fMAX of 450 GHz

Record RF performance of sub-46 nm L/sub gate/ NFETs in microprocessor SOI CMOS technologies

Field Effect Transistors on SOI offer inherent capacitance and process advantages. The flow of heat generated at the drain junction may be impeded by dielectric isolation but an assessment must also account for conduction of heat through the gate stack and through the device contacts, and its impact on device characteristics should be captured by the scalable model to enable accurate circuit design. A quantitative comparison to 45nm planar SOI shows that while the scaled FinFET on dielectric devices show higher normalized thermal resistance, as expected from device scaling, the characteristic time constant for self heating is still well below the operating frequency of typical logic circuits, hence resulting in negligible self heating effect. For cases where the self heating becomes a factor, e.g., in high-speed I/O circuits, the same design methods can be applied for both planar and FinFET devices on dielectric isolation.

Sungjae Lee

Papers

Record RF performance of 45-nm SOI CMOS Technology

Modeling of Variation in Submicrometer CMOS ULSI Technologies

Fully depleted extremely thin SOI for mainstream 20nm low-power technology and beyond

Record RF performance of sub-46 nm L/sub gate/ NFETs in microprocessor SOI CMOS technologies

Experimental analysis and modeling of self heating effect in dielectric isolated planar and fin devices