Experimental results are presented for stress evolution, in vacuum and electrolyte, for the first monolayer of Cu on Au(111). In electrolyte the monolayer is pseudomorphic and the stress-thickness change is -0.60 N/m, while conventional epitaxy theory predicts a value of +7.76 N/m. In vacuum, the monolayer is incoherent with the underlying gold. Using a combination of first-principles based calculations and molecular dynamic simulations we analyzed these results and demonstrate that in electrolyte, overlayer coherency is maintained owing to anion adsorption.
The stress evolution during electrodeposition of NiMn from a sulfamate-based bath was investigated as a function of Mn concentration and current density. The NiMn stress evolution with film thickness exhibited an initial high transitional stress region followed by a region of steady-state stress with a magnitude that depended on deposition rate, similar to the previously reported stress evolution in electrodeposited Ni [S. J. Hearne and J. A. Floro, J. Appl. Phys. 97, 014901-1 (2005)]. The incorporation of increasing amounts of Mn resulted in a linear increase in the steady-state stress at constant current density. However, no significant changes in the texture or grain size were observed, which indicates that an atomistic process is driving the changes in steady-state stress. Additionally, microstrain measured by ex situ x-ray diffraction increased with increasing Mn content, which was likely the result of localized lattice distortions associated with substitutional incorporation of Mn and/or increased twin density.
Transmission electron microscopy and x-ray diffraction were used to assess the microstructure and strain of Al{sub x}Ga{sub 1?x}N(x = 0.61-0.64) layers grown on AlN. The compressively-strained AlGaN is partially relaxed by inclined threading dislocations, similar to observations on Si-doped AlGaN by P. Cantu, F. Wu, P. Waltereit, S. Keller, A. E. Romanov, U. K. Mishra, S. P. DenBaars, and J. S. Speck [Appl. Phys. Lett. 83, 674 (2003) ]; however, in our material, the dislocations bend before the introduction of any Si. The bending may be initiated by the greater lattice mismatch or the lower dislocation density of our material, but the presence of Si is not necessarily required. The relaxation by inclined dislocations is quantitatively accounted for with the model of A. E. Romanov and J. S. Speck [Appl. Phys. Lett. 83, 2569 (2003)], and we demonstrate the predicted linear dependence of relaxation on layer thickness. Notably, such relaxation was not found in tensile strained AlGaN grown on GaN [J. A. Floro, D. M. Follstaedt, P. Provencio, S. J. Hearne, and S. R. Lee, J. Appl. Phys. 96, 7087 (2004)], even though the same mechanism appears applicable.
Using in situ wafer-curvature measurements of thin-film stress, we determine the critical thickness for strain relaxation in Al{sub x}Ga{sub 1-x}N/GaN heterostructures with 0.14 {le} x {le} 1. The surface morphology of selected films is examined by atomic force microscopy. Comparison of these measurements with critical-thickness models for brittle fracture and dislocation glide suggests that the onset of strain relaxation occurs by surface fracture for all compositions. Misfit-dislocations follow initial fracture, with slip-system selection occurring under the influence of composition-dependent changes in surface morphology.
The ability of future integrated metal-semiconductor micro-systems such as RF MEMS to perform highly complex functions will depend on developing freestanding metal structures that offer improved conductivity and reflectivity over polysilicon structures. For example, metal-based RF MEMS technology could replace the bulky RF system presently used in communications, navigation, and avionics systems. However, stress gradients that induce warpage of active components have prevented the implementation of this technology. Figure 1, is an interference micrograph image of a series of cantilever beams fabricated from electrodeposited Ni. The curvature in the beams was the result of stress gradients intrinsic to the electrodeposition process. To study the sources of the stress in electrodeposition of Ni we have incorporated a wafer curvature based stress sensor, the multibeam optical stress sensor, into an electrodeposition cell. We have determined that there are two regions of stress induced by electrodepositing Ni from a sulfamate-based bath (Fig 2). The stress evolution during the first region, 0-1000{angstrom}, was determined to be dependent only on the substrate material (Au vs. Cu), whereas the stress evolution during the second region, >1000{angstrom}, was highly dependent on the deposition conditions. In this region, the stress varied from +0.5 GPa to -0.5GPa, depending solely on the deposition rate. We examined four likely sources for the compressive intrinsic stress, i.e. reduction in tensile stress, and determined that only the adatom diffusion into grain boundaries model of Sheldon, et al. could account for the observed compressive stress. In the presentation, we shall discuss the compressive stress generation mechanisms considered and the ramifications of these results on fabrication of electrodeposited Ni for MEMS applications.
This document summarizes research of reactively deposited metal hydride thin films and their properties. Reactive deposition processes are of interest, because desired stoichiometric phases are created in a one-step process. In general, this allows for better control of film stress compared with two-step processes that react hydrogen with pre-deposited metal films. Films grown by reactive methods potentially have improved mechanical integrity, performance and aging characteristics. The two reactive deposition techniques described in this report are reactive sputter deposition and reactive deposition involving electron-beam evaporation. Erbium hydride thin films are the main focus of this work. ErH{sub x} films are grown by ion beam sputtering erbium in the presence of hydrogen. Substrates include a Al{sub 2}O{sub 3} {l_brace}0001{r_brace}, a Al{sub 2}O{sub 3} {l_brace}1120{r_brace}, Si{l_brace}001{r_brace} having a native oxide, and polycrystalline molybdenum substrates. Scandium dideuteride films are also studied. ScD{sub x} is grown by evaporating scandium in the presence of molecular deuterium. Substrates used for scandium deuteride growth include single crystal sapphire and molybdenum-alumina cermet. Ultra-high vacuum methods are employed in all experiments to ensure the growth of high purity films, because both erbium and scandium have a strong affinity for oxygen. Film microstructure, phase, composition and stress are evaluated using a number of thin film and surface analytical techniques. In particular, we present evidence for a new erbium hydride phase, cubic erbium trihydride. This phase develops in films having a large in-plane compressive stress independent of substrate material. Erbium hydride thin films form with a strong <111> out-of-plane texture on all substrate materials. A moderate in-plane texture is also found; this crystallographic alignment forms as a result of the substrate/target geometry and not epitaxy. Multi-beam optical sensors (MOSS) are used for in-situ analysis of erbium hydride and scandium hydride film stress. These instruments probe the evolution of film stress during all stages of deposition and cooldown. Erbium hydride thin film stress is investigated for different growth conditions including temperature and sputter gas, and properties such as thermal expansion coefficient are measured. The in-situ stress measurement technique is further developed to make it suitable for manufacturing systems. New features added to this technique include the ability to monitor multiple substrates during a single deposition and a rapidly switched, tiltable mirror that accounts for small differences in sample alignment on a platen.
Strain-induced self-assembly during semiconductor heteroepitaxy offers a promising approach to produce quantum nanostructures for nanologic and optoelectronics applications. Our current research direction aims to move beyond self-assembly of the basic quantum dot towards the fabrication of more complex, potentially functional structures such as quantum dot molecules and quantum wires. This report summarizes the steps taken to improve the growth quality of our GeSi molecular beam epitaxy process, and then highlights the outcomes of this effort.
Erbium hydride thin films are grown onto polished, a-axis {alpha} Al{sub 2}O{sub 3} (sapphire) substrates by reactive ion beam sputtering and analyzed to determine composition, phase and microstructure. Erbium is sputtered while maintaining a H{sub 2} partial pressure of 1.4 x 10{sup {minus}4} Torr. Growth is conducted at several substrate temperatures between 30 and 500 C. Rutherford backscattering spectrometry (RBS) and elastic recoil detection analyses after deposition show that the H/Er areal density ratio is approximately 3:1 for growth temperatures of 30, 150 and 275 C, while for growth above {approximately}430 C, the ratio of hydrogen to metal is closer to 2:1. However, x-ray diffraction shows that all films have a cubic metal sublattice structure corresponding to that of ErH{sub 2}. RBS and Auger electron that sputtered erbium hydride thin films are relatively free of impurities.
Strained-layer semiconductor films offer tremendous potential with regards to optoelectronic applications for high speed communications, mobile communications, sensing, and novel logic devices. It is an unfortunate reality that many of the possible film/substrate combinations that could be exploited technologically are off limits because of large differences in lattice parameters, chemical compatibilities, or thermal expansion rates. These mechanical, chemical, and thermal incompatibilities manifest themselves primarily in terms of lattice defects such as dislocations and antiphase boundaries, and in some cases through enhanced surface roughness. An additional limitation, from a production point of view, is money. Device manufacturers as a rule want the cheapest substrate possible. Freeing the heteroepitaxial world of the bonds of (near) lattice matching would vastly expand the types of working devices that could be grown. As a result, a great deal of effort has been expended finding schemes to integrate dissimilar film/substrate materials while preserving the perfection of the film layer. One such scheme receiving significant attention lately is the so-called compliant substrate approach.