Book Description

Metal silicides have been used in silicon technology as contacts to achieve high device performance and desired device functions. The growth and applications of silicide materials have recently attracted increasing interest for nanoscale device applications. Nanoscale silicide materials have been demonstrated with various synthetic approaches. Solid state reaction wherein high quality silicides form through diffusion of metal atoms into silicon nano-templates and the subsequent phase transformation caught significant attention for the fabrication of nanoscale Si devices. Very interestingly, studies on the diffusion and phase transformation processes at nanoscale have indicated possible deviations from the bulk and the thin film system. Here we studied growth kinetics, electronic properties and device applications of nanoscale silicides formed through solid state reaction. We have grown single crystal PtSi nanowires and PtSi/Si/PtSi nanowire heterostructures through solid state reaction. TEM studies show that the heterostructures have atomically sharp interfaces free of defects. Electrical measurement of PtSi nanowires shows a low resistivity of ∼28.6 μΩ*cm and a high breakdown current density beyond 108 A/cm2. Furthermore, using single-crystal PtSi/Si/PtSi nanowire heterostructures with atomically clean interfaces, we have fabricated p-channel enhancement mode transistors with the best reported performance for intrinsic silicon nanowires to date. In our results, silicide can provide a clean and no Fermi level pinning interface and then silicide can form Ohmic-contact behavior by replacing the source/drain metal with PtSi. It has been proven by our experiment by contacting PtSi with intrinsic Si nanowires (no extrinsic doping) to achieve high performance p-channel device. By utilizing the same approach, single crystal MnSi nanowires and MnSi/Si/MnSi nanowire heterojunction with atomically sharp interfaces can also been grown. Electrical transport studies on MnSi nanowire shows an abrupt resistance reduction due to the spin ordering at ~29.7 K.A negative magnetoresistance (MR) ~1.8% under 5 Tesla at 1.6 K is achieved, demonstrating the ferromagnetic behavior of MnSi. Furthermore, using the MnSi/p-Si/MnSi heterostructure, we have studied the charge injection at various temperatures via the Schottky barrier, and the spin scattering was observed through magnetotransport studies of MnSi/p-Si/MnSi heterojunction. Our results represent the first report of magnetic contact fabrication through the formation of single crystal heterojunction nanowires and the first demonstration of spin injection and detection in such Si nanowire devices. The magnetic silicides approach thus opens a new pathway to create ferromagnetic/semiconductor junction with clean and sharp interface, and maysignificantly impact the future of spintronics. Beyond those applications, silicide phase control at nanoscale is investigated. Three nickel phases, Ni31Si12, Ni2Si and NiSi2 are observed in one step annealing at 550 oC. NiSi2 grows initially through the Si NW and then the area close to nickel pad transforms into the nickel-rich phase, Ni31Si12. With prolonged annealing over 5 minutes, the Ni2Si starts to show up in between Ni31Si12 and NiSi2. The growth sequence is different from the thin film system where Ni2Si usually appears as the initial phase in the beginning as the annealing temperature is higher than 400 oC. Interfacial energy differences and surface free energy are believed to play an important role here at the nanoscale, which lead to the formation of normally unfavorable silicide phases in Si NWs. In addition, Si/SiOx core/shell NW structure is used to explore the phase transformation of silicides in the structure-confined nano environment. Nickel silicides in the structure-confined core/shell Si NW shares the similar phase formation sequences as those appeared in the bared SiNWs, while the growth rate is significantly retarded. This may be attributed to the high compressive stress built-in in the core/shell NW structure that retards the diffusion of the nickel atom as well as limits the volume expansion of the metal-rich phases. As a result, the high stress at this finite scale hinders the continuous growth of Ni31Si12 into the core/shell NWs and totally eliminates the formation of Ni2Si in core/shell NWs with thick oxide shells (~ 50 nm). Through these studies, we have demonstrated first time the phase formation sequences of nickel silicides in Si and Si/SiOx NW structures, which is of great importance for reliable contact engineering for Si NW devices. Furthermore, we have provided a clear picture of the hindered nickel silicide growth in confined nanoscale environment and showed the deviated behavior of silicides growth under stress. The information rendered here will be useful for Si NW device applications as well as for the silicon device engineering at nanoscale in general. To further investigate the oxide shell effect, Mn5Si3 and Fe5Ge3 NW were grown within various oxide thickness to explore the nucleation and growth in the nanowire structure. A oxide shell exerted a compressive stress on the silicide or germanide materials will make those materials with single-crystal properties. Interestingly, single-crystal growth of contact materials can be also implemented for germanide materials. The iron-rich germanide, Fe5Ge3, was successfully grown with single-crystal properties. It shows ferromagnetic properties with a Curie temperature above the room temperature verified by magnetic force microscope (MFM). Two different epitaxial relations found at germanide/germanium interface due to the different sizes of the germanium NW templates. These two different crystal structures exhibited magnetic anisotropy in magnetic force microscope (MFM) measurement, showing differently preferred domain orientations. In-plane and out-of-plane magnetization in the Fe5Ge3 NWs are observed in our experiment. The crystal orientation or engineering stress may have influence on the magnetic domain structure. This ferromagnetic contact material may open the way for spintronics to grow the magnetic materials on the semiconducting materials and control the direction of magnetization in the future. Those silicide studies indicated silicide metal-heterojunction field effect transistor has excellent device performance. In addition, Si channel region can be shrunk to less than 10 nm and also keep semiconducting properties without high leakage current. This approach has the potential for future nanoelectronics. However, silicide phase transformation shows a deviated behavior from the studies in bulk system. It may be associated with stress effect or nucleation behavior at nanosclae, leading the different formation phase or sequence. For those interesting phenomena, it has attracted more and more attention and may gain more insight studies in the near future.