This Small Business Innovation Research (SBIR) Phase I will investigate breakthrough concepts to overcome the limitations imposed by the fundamental physical properties of silicon that prevent it from emitting and sensing light in the infrared range of wavelengths used in telecommunications. These limitations are the most important barrier to using light, instead of electrical signals, to transmit information within a CMOS chip, and between CMOS chips. Using optical, rather than electrical, interconnects will increase performance, decrease power dissipation (heat generation), and reduce manufacturing costs. The research objectives are the identification and demonstration of a new silicon-based material, a superlattice incorporating Silicon, Germanium and Carbon atoms, whose optoelectronic properties are comparable to those of III-V semiconductors from which the LASERs and Photo-Detectors currently used in optical communications, are made. The project will begin with theoretical modeling and simulation of several superlattice compositions, in order to identify the one with the most promising properties, which will then be fabricated and characterized, both as a stand-alone film and by incorporation into a basic photo-diode. It is anticipated that a new class of Si-Ge-C superlattice materials will enable high-efficiency silicon-based devices for light-emission and light-sensing in this range of wavelengths.
The broader impact/commercial potential of this project will be in the area of Silicon Photonics, which is a core technology with applications to several fields. The most important field is CMOS manufacturing, where silicon photonics can help Moore's Law maintain its trajectory, overcoming the barrier posed by the limitations of electrical interconnects by replacing them with optical interconnects, within a chip and from chip-to-chip. Optical interconnects will increase performance, improve reliability, and lower power dissipation, while reducing manufacturing costs of leading-edge CMOS technology. Other high-volume applications include fiber optics communications for Fiber-To-The-Home (FTTH) by enabling more compact equipment, capable of more functionality at lower cost, and the replacement of legacy electrical connections for Ethernet, HDMI, DisplayPort, and USB, with lower cost optical cables. By extending the functionality of CMOS to handle light efficiently, both emission and absorption across a wide range of wavelengths, new applications will open up for imaging technology and true silicon photonics.
This project had the goal of studying new materials, Si-Ge-C superlattices, which will be used to develop highly-efficient light-absorbing devices, such as silicon-based LASERs. Superlattices are artificial crystals that are fabricated through the sequential deposition of atomic layers, in which the composition of the atomic planes as well as the number of atomic planes for each composition, are engineering parameters that can be varied to modify the optoelectronic properties of the superlattices. The optoelectronic properties of superlattices can be radically different with respect to those of the constituent elements of the superlattic; Si, Ge, C, etc. Superlattices with different combinations of atomic planes with the same composition can also have very different optoelectronic properties. This project investigated superlattices that can be integrated with current and future CMOS technologies. This project comprised both theoretical work (computer simulations) and experimental work (film growth). For the computer simulations, pre-existing codes were adapted and fine-tuned to produce accurate results for known materials, and new codes were developed to investigate a variety of superlattice cells, with multiple compositions for the atomic planes, multiple periodicities for each composition, and straining to multiple crystalline orientations of the silicon substrate. A key goal of the simulations was to determine which superlattices showed the most promising optoelectronic properties so that subsequent development efforts can be focused on growing those exact compositions. A large number of superlattices with optoelectronic properties that make them good candidates for the development of silicon-based LASERs, Photo-Diodes (PDs), and other optoelectronic devices, were identified. These will enable the development of photonic circuits with a high density of integration and very large number of active elements (LASERs, PDs, etc.). The superlattices studied in this project covered a wide range of wavelengths in the infrared (IR), which open up a number of different applications. A few superlattices have band-gaps near the 1.55 micron wavelength, which makes them suitable for applications such as optical transceivers (emitters and receivers), which can be used for optical interconnects within a chip and from chip-to-chip. Optical interconnects have been recognized to be a crucial solution to the continuation of "Moore’s Law". During this project, the computer codes for ab-initio simulation of the optoelectronic properties of the superlattices were continuously improved for better accuracy and higher speed. The functionality of the ab-initio codes was also extended to deliver "k.p models" (pronounced k dot p) for the band structure of semiconductors, including superlattices, which can be used for TCAD simulation of devices incorporating the superlattices. TCAD simulations are used to estimate basic electrical and optical behavior of devices incorporating superlattices, which is necessary to design circuitry that will interact with these devices. TCAD device simulations have shown that one of the superlattices, with a band-gap suitable for light emission and absorption near the 1.55 micron range, has a quantum efficiency only 20 times smaller than that of GaAs. Although GaAs is one of the main materials used today for photonic applications, the fabrication of GaAs devices is generally more expensive because it does not benefit from the economies of scale of silicon manufacturing, and tends not to be used for large integrated circuits. In addition to the development of the codes and modeling of various superlattices, preliminary experimental efforts were made to grow superlattice films according to various specifications. Two methods were used. One method was solid-source molecular beam epitaxy (MBE), which had some limited success, but did not result in carbon levels at the desired concentrations. Films from this method were used to make diodes which were characterized. These diodes did not display any unusual optoelectronic properties, due in part to the low concentration of Carbon in the films. The primary parameters which can be varied with this type of MBE include the temperature of the solid source, type of solid source, background pressure, and growth rate. The other method used to grow superlattices was chemical vapor deposition (CVD). This method is widely used in the semiconductor industry to grow silicon based crystal layers. Also in this case, the films produced did not incorporate enough carbon. Some of the parameters that can be varied, include temperature, vapor pressure, chemical precursors, and gas flow rate. The superlattices investigated in this project have many future applications beyond optical communications. These include: 1) Photo-Diodes in CMOS image sensors, enabling higher efficiency absorption of the longer wavelengths of the visible (i.e., Red), as well as CMOS-based solutions for image sensing in the Short-Wavelength (SWIR), Mid-Wavelength (MWIR), and Long-Wavelength Infra-red (LWIR); 2) Silicon-based single-crystalline Multi-Junction Photo-Voltaic Cells covering a much wider range of wavelengths than silicon with higher efficiency; 3) Base regions of Heterojunction Bipolar Transistors (HBTs) in BiCMOS, enabling better circuit performance for wireless communications; and 4) transistor regions of advanced CMOS devices operating at lower voltages to reduce power consumption.
