The scaling approach that has been the technological mainstay of the semiconductor industry for the last 30 years is beginning to face limits to continued miniaturization with commensurate performance enhancement. Prof. Saraswat has been working on new and innovative materials, structures, and process technology of silicon, germanium and III-V devices and interconnects for VLSI and nanoelectronics. Areas of his current interest are: new device structures to continue scaling MOS transistors, 3-dimentional ICs with multiple layers of heterogeneous devices, metal and optical interconnections and high efficiency and low cost solar cells.
CURRENT MAJOR PROJECTS:
Materials, Structures, Devices Physics and Technology for Nanoelectronics
Si CMOS technology has dominated the microelectronics industry, with continued scaling. However, future Si CMOS scaling is reaching practical and fundamental limits. To go beyond these limits novel materials and structures are being aggressively studied. A channel material with high µ and therefore high injection velocity can increase drive current and reduce delay. Currently, strained-Si is the dominant technology for high performance MOSFETs and increasing the strain provides a viable solution to scaling. However, performance enhancement due to strain is beginning to saturate with scaling of nanoscale MOSFETs. Looking into future scaling it becomes important to look at higher mobility materials, like Ge and III-V together with innovative device structures which may perform better than even very highly strained Si. Heterogeneous integration of the high mobility Ge and III-V materials on Si with novel device structures can take us to sub-10 nm regime. However there are many daunting questions. Ge PMOS shows much promise but the NMOS has many problems to be solved. Similarly there are many examples of high performance III-V NMOS, but PMOS is still problematic. Is Ge PMOS and III-V NMOS integration on Si feasible or a headache for the manufacturing folks? Can we have all Ge or all III-V CMOS or it is just a fantasy? Prof. SaraswatÕs research is trying to answer these questions.
Interconnects for Nanoelectronics
Continuous scaling of VLSI circuits can pose significant problems for interconnects, especially for those responsible for long distance communication on a high performance chip. Our modeling predicts that the situation is worse than anticipated in the ITRS, which assumes that the resistivity of copper will not change appreciably with scaling in the future. We show that resistance of interconnect wires in light of scaling induced increase in electron surface scattering, fractional cross section area occupied by the high resistivity barrier and realistic interconnect operation temperature will lead to a significant rise in the effective resistivity of Cu. As a result both power and delay of these interconnects is likely to rise significantly in the future. In the light of various metal interconnect limitations, alternate solutions need to be pursued. His group focuses on two such solutions, optical interconnects and three-dimensional (3-D) ICs with multiple active Si layers.
High Efficiency attribute of solar cells is a powerful enabler to reduce the $/Wp metric by cutting across the entire PV supply chain. In addition, a higher efficiency cell, especially if it is higher Voc, can result in a much larger energy yield (KWHr/KWp) per given efficiency or peak power. These factors go a long way in enabling the goals set by U.S. Department of Energy of 1. delivering modules at under $0.40/W, 2. reducing the BOS cost, 3. ultimately achieving an LCOE of 5 cents/KWhr. To meet these requirements we are developing a novel junctionless solar cell using metal/insulator/semiconductor (MIS) structures as carrier selective contacts on Si, GaAs and other absorbers.
AFFILIATED Ph.D. STUDENTS:
Research: Germanium Based Laser and Modulator for Photonic Interconnects.
Silicon photonics is considered to be a key enabling technology for future CMOS systems since it has the potential to alleviate the bandwidth-power-density bottleneck of electrical interconnects. In this work we are developing semiconductor beam structures for light emission and modulation applications. For example, we consider a strained crossed nanobeam structure for a low-threshold germanium laser. It includes a germanium (Ge) photonic crystal nanobeam resonator and a silicon nitride (SiN) stressor nanobeam in the perpendicular direction. While the photonic crystal nanobeam enables light confinement in a subwavelength volume with minimal optical loss, the SiN nanobeam induces high tensile strain in the small region where the optical mode is tightly confined. As maintaining a small optical loss and a high tensile strain reduces the required pumping for achieving net optical gain beyond cavity losses, this technique can be used to develop an extremely low-threshold Ge laser source. We are also developing a compact Ge electro-absorption modulator (EAM) based on Franz-Keldysh effect according to which an applied electric field causes band-tilting thereby increasing the absorption coefficient in the weakly absorbing regime. The Ge EAM devices are integrated in SOI Si platform. These devices have improved performance including higher modulation speed, higher efficiency and lower capacitance, owing to the strong confinement of optical and electrical field enabled by the submicron Ge/Si waveguide platform.
Raisul Islam and Koosha Nassiri Nazif
Research: Junction-less multi-material high efficiency low cost solar cell.
To meet the requirements of high efficiency and low cost we are developing a novel junctionless solar cell technology using metal/insulator/semiconductor (MIS) structures as carrier selective contacts on silicon, germanium, III-Vs and other absorbers. In this novel approach diffused p-n junction are replaced by MIS heterojunctions using oxides such as ZnO or TiO2 and NiO, which have band lineups such that it enables a selective contact for either electron/hole while reflecting the other carrier away from the contact. This structure, which eliminates the need of p-n junctions, is a low cost solution to a high efficiency solar cell. Using this approach we conceived the idea of a 3 terminal multijunction solar cell using MIS heterojunctions. This structure, which eliminates the need of p-n junctions, is a low cost solution to a high efficiency solar cell where two subcells can work independently. The design is not limited by lattice matching or current matching. Thus each sub cell can be optimized independently for maximum efficiency.
Research: Integration of III-V FETs on Silicon
As scaling of Silicon CMOS becomes increasingly difficult, use of alternate materials with high carrier mobilities is being explored extensively. Despite burgeoning interest in III-Vs for realizing high performance transistors at low power, III-V MOSFETs have been plagued by (a) poor surface passivation resulting in high Dit; (b) poor PMOS performance, which makes realizing CMOS in a single material system difficult; (c) low density of states (DOS) for electrons; (d) poor dopant activation in the source-drain (S/D) and (e) high contact resistance. We are developing novel solutions to overcome these problems. We are working on the 6.1-6.2 lattice constant system with InGaSb as the channel material because of its advantages in terms of band engineering and high mobility/offsets for both electrons and holes. The goal is to achieve high electron/hole mobility in the same channel material for n- and p-MOS devices through the optimization of surface passivation, stoichiometry, heterostructure design and novel contact/interface engineering. Hetero-integration on a silicon substrate is one of the key challenges in enabling III-V CMOS. III-Vs also inherently suffer from degraded short-channel effects and are more vulnerable to variability issues. Alternative device architectures such as SOI gate-all-around finFETs are needed to improve the device performance. Current work focuses on using the rapid-melt-growth (RMG) technique to fabricate III-V FinFETs with multiple channel materials on a silicon substrate.
Ju Hyung Nam
Research: Selective Heteroepitaxial Growth of Ge on Si for Electronic and Optical Devices.
Recently Ge has emerged as a viable candidate to augment Si for continued scaling of CMOS devices and providing high performance on- and off-chip opticle interconnects. However, Ge substrate is not easy to handle and is not easily available. Therefore to utilize above mentioned advantages of Ge it is imperative to develop technology for heterogeneous integration of Ge on Si-based platform. It is crucial to be able to grow high quality Ge layers selectively. Ge has large (4.2%) lattice mismatch with Si, which causes strain to Ge layers grown directly on Si which results in high density of dislocations and rough surface. In our past work we have demonstrated that growth of Ge on Si by Multiple Hydrogen Annealing for Heteroepitaxy (MHAH) technique yields very low dislocation density and surface roughness. In this work we are focusing on selective growth of Ge on Si through a SiO2 windows combined with H2 anneal and extend it to grow Ge laterally on SiO2 for GOI applications. This will allow fabrication of high performance MOSFETs and optical devices in Ge integrated on Si with future promise of continued CMOS scaling.
Research: Metal/Insulator/Semiconductor Contacts for Photovoltics and Nanoelectronics.
As device scaling continues, parasitic source resistance largely dominated by contact resistance, is beginning to limit the device performance. Specific contact resistivity of a metal-semiconductor (M/S) contact is dependent on the Schottky barrier height, and the electrically active dopant density N at that interface. In M/S contacts the metal Fermi level is pinned at the charge neutrality level, ECNL, resulting in fixed electron and hole Schottky barrier heights. To obtain low contact resistivity it is essential to reduce Schottky barrier height. Historically the method to reduce contact resistivity is by increasing N to > 1E20/cm3 thereby thinning the barrier, thus allowing more tunneling current. This method works well for n type Si and p-Ge which can be doped heavily. However, it is not very practical for n-Ge, p-Si, many III-Vs and 2D materials because of inability to dope them heavily. We are investigating Fermi level depinning by inserting a thin dielectric between the metal and semiconductor and reduce barrier height. The metal electron wavefunction now decays in the insulator resulting in fewer MIGS thus depinning the Fermi level, which now pins on the insulator. Hence metal workfunction can now be used to tune the effective barrier height. For n-type ohmic contacts, metals with a low metal workfunction and insulator with ~0 conduction band offset (ZnO, TiO2, SnO2) with semiconductor should result in a near zero barrier height to electrons. For p-type ohmic contacts, metals with a high work function and insulator with ~0 valence band offset (NiO, CuAlO2) with semiconductor would be desirable. This method has been used to obtain low barrier height and thus low contact resistivity in Si, Ge and III-V semiconductors and should prove to be very useful for 2D materials.
Research: Low Resistance ohmic contacts to Ge and III-V Devices
For the nanoscale transistors in Ge, III-V and carbon based MOSFETs and optical devices low resistance contacts must be developed. The specific contact resistivity of many of these materials is too high primarily due to low doping density and Fermi level pinning. MIS contacts solve this problem using interfacial layers that depin the Fermi level leading to engineering of the Schottky barrier height and thus achieve low contact resistance. Several mechanisms, such as metal induced gap states (MIGS) and dipole formation have been proposed to explain the effect of Fermi level unpinning. Despite the large body of experimental data, there is no consensus on the underlying physics and mechanisms responsible for Fermi level unpinning. In this project the tunneling transport is being treated quantum mechanically using free electron approach using transfer matrix formalism for solution of the Schrodinger equation self-consistent with the Poisson solution for near equilibrium transport through tunnel barriers. The effect of doping density, different barrier materials, interface and fixed charges, interface dipoles and metal work functions is included. This will help to predict optimized tunnel barrier materials and device structures that can be used to minimize contact resistance.
Research: Nanoscale Germanium MOSFETs
Si CMOS technology has dominated the microelectronics industry, with continued scaling. However, future Si CMOS scaling is reaching practical and fundamental limits. To go beyond these limits novel higher mobility materials like Ge are being aggressively studied to continue scaling of MOSFETs. Development of MOS technology in these materials has been largely possible due to advances in the high-k/metal gate stack technology. High-k dielectrics with low interface state density (Dit) have been demonstrated to passivate the channel with atomic layer deposition (ALD) as the preferred technique. Therefore, Ge MOSFET is quickly becoming the choice for future CMOS. However, currently, very little is understood with respect to reliability of these new CMOS device components (high-k gate dielectrics and advanced channel materials). In this study, we propose develop technology to fabricate nanoscale Ge MOSFETs with metal gate/ high-k gate dielectric and characterize the radiation effects of this new class transistors for the first time and establish a fundamental understanding of how to improve their reliability and radiation tolerance through high mobility channel/high-k gate stack material synthesis and design enhancements.