Search

The influence of the type of alumina used as catalyst support on the evolution, activity, and lifetime of the catalyst during water-assisted CVD growth (or ‘supergrowth') of single-walled carbon nanotube (SWNT) carpets has been studied. The catalyst consisted of a thin Fe film supported on alumina films deposited by different methods: atomic layer deposition (ALD), e-beam, and magnetron sputtering. In order to fully understand the influence of the type of alumina on SWNT carpet growth, crystalline alumina (c-cut sapphire) and annealed alumina deposited by e-beam were also used as catalyst supports. The activity and lifetime of Fe catalyst during SWNT carpet growth showed a strong dependence on the type of alumina used as support. Fe supported on sputtered alumina (sputtered/Fe) showed the highest catalytic activity and lifetime, which was closely followed by e-beam/Fe while Fe supported on sapphire (sapphire/Fe) showed the least catalytic activity and lifetime. AFM, XPS depth profile, variable angle spectroscopic ellipsometry (VASE) studies revealed that the catalyst evolution and the porosity of the different alumina supports correlate with the lifetime and activity of the catalysts.

Dynamic friction, wear volumes and wear morphology have been studied for sliding wear in polysilicon in ambient air at μN normal loads using on-chip micron-scale test specimens. With increasing number of wear cycles, the friction coefficients show two distinct types of behavior: (i) an increase by a factor of two and a half to a steady-state regime after peaking at three times the initial value of about 0.10 ± 0.04, with no failure after millions of cycles; (ii) an increase by a factor larger than three followed by failure after ∼105 cycles. Additionally, the average nano-scale wear coefficient sharply increased in the first ∼105 cycles up to about 10−4 and then decayed by an order of magnitude over the course of several million cycles. For both modes of behavior, abrasive wear is the governing mechanism, the difference being attributed to variations in the local surface morphology (and wear debris) between the sliding surfaces. The oxidation of worn polysilicon surfaces only affects the friction coefficient after periods of inactivity (>30 min).

Nanoindentations were created in the GaAs(100) surface to act as strain centers for the controlled nucleation and patterning of InAs quantum dots. A systematic approach is taken to understand the indent deformation processes that may lead to precision patterning capabilities for a broad range of nanostructures. Indents were generated using loads below 450 μN with a sharp cube corner indenter. Site-specific cross-sectional thinning of nanoindents (down to 100 nm in size) has been achieved using the in-situ ‘lift-out’ technique. This allowed for observation of subsurface deformation by transmission electron microscopy (TEM). The crystal was observed to deform solely by dislocation activity with no evidence of amorphization, twinning, fracture, or phase transformation. It is shown that the single phase deformation of GaAs can be well characterized and controlled, which should allow for exploitation of the dislocation strain field to bias nucleation of self-assembled quantum dots.

The presence of impurities in aluminum alloys is of great interest with respect to microstructural properties, specifically, the effect of solute on texture and anisotropy. This paper presents new evidence of the pronounced effect of solute drag based on in-situ annealing and Electron Backscatter Diffraction experiments of Zr-rich Al alloys subject to prior strain. A compensation effect was found for grain boundary mobility maxima for specific boundary types. Trends in activation energy as a function of boundary type support the observations of a compensation effect with respect to temperature. Evidence for irregular motion of boundaries from in-situ observations is discussed in reference to new theoretical results that suggest that boundaries migrating in the presence of solutes should move sporadically provided that the length scale at which observations are made is small enough. A study of both boundary motion and solute segregation to specific boundary types using Scanning Transmission Electron Microscopy and in-situ TEM is presented.

The deformation behavior of Al and Al-Mg thin films has been studied with the unique experimental approach of in-situ nanoindentation in a transmission electron microscope. This paper concentrates on the role of solute Mg additions in the transfer of plasticity across grain boundaries. The investigated Al alloys were deposited onto a Si substrate as thin films with a thickness of 200–300 nm and Mg concentrations of 0, 1.1, 1.8, 2.6 and 5.0 wt% Mg. The results show that in the Al-Mg alloys, the solutes effectively pin high-angle grain boundaries, while in pure Al considerable grain boundary motion is observed at room temperature. The mobility of low-angle grain boundaries is however not affected by the presence of Mg. In addition, Mg was observed to affect dislocation dynamics in the matrix.

Although bulk silicon is ostensibly immune to cyclic fatigue and environmentally-assisted cracking, the thin film form of the material exhibits significantly different behavior. Such silicon thin films are used in small-scale structural applications, including microelectromechanical systems (MEMS), and display ‘metal-like’ stress-life (S/N) fatigue behavior in room temperature air environments. Fatigue lives in excess of 1011 cycles have been observed at high frequency (∼40 kHz), fully-reversed stress amplitudes as low as half the fracture strength using surface micromachined, resonant-loaded, fatigue characterization structures. Recent experiments have clarified the origin of the susceptibility of thin film silicon to fatigue failure. Stress-life fatigue, transmission electron microscopy, infrared microscopy, and numerical models have been used to establish that the mechanism of the apparent fatigue failure of thin-film silicon involves sequential oxidation and environmentally-assisted crack growth solely within the nanometerscale silica layer on the surface of the silicon, via a process that we term ‘reaction-layer fatigue’. Only thin films are susceptible to such a failure mechanism because the critical crack size for catastrophic failure of the entire silicon structure can be exceeded by a crack solely within the surface oxide layer. The growth of the oxide layer and the environmentally-assisted initiation of cracks under cyclic loading conditions are discussed in detail. Furthermore, the importance of interfacial fracture mechanics solutions and the synergism of the oxidation and cracking processes are described. Finally, the successful mitigation of reaction-layer fatigue with monolayer coatings is shown.

Although bulk silicon is not known to be susceptible to cyclic fatigue, micron-scale structures made from mono and polycrystalline silicon films are vulnerable to degradation by fatigue in ambient air environments. Such silicon thin films are used in small-scale structural applications, including microelectromechanical systems (MEMS), and display “metal-like” stress-life (S/N) fatigue behavior in room temperature air environments. Previously, the authors have observed fatigue lives in excess of 1011 cycles at high frequency (∼40 kHz), fully-reversed stress amplitudes as low as half the fracture strength using a surface micromachined, resonant-loaded, fatigue characterization structures. Stress-life fatigue, transmission electron microscopy, infrared microscopy, and numerical models were used to establish that the mechanism of the fatigue failure of thin-film silicon involves the sequential oxidation and environmentally-assisted crack growth solely within the native silica layer, a process that we term “reaction-layer fatigue”. Only thin films are susceptible to such a failure mechanism because the critical crack size for catastrophic failure of the entire silicon structure can be exceeded by a crack solely within the native oxide layer. The importance of the interfacial geometry on the mechanics of the reaction-layer fatigue mechanism is described.

In this research, we have focused on the morphological evolution of a model metal film / silicon substrate system. When aluminum (Al) is physical vapor deposited on (100) oriented single crystal silicon (Si) at 280 °C it grows heteroepitaxially. Crystallographically, the resulting films are a Potts model system with degeneracy two; their topology simulates a polycrystalline grain structure, with the two types of grains enveloping each other in a convoluted, maze-like arrangement. The morphological evolution of the films during annealing was determined via a combination of transmission electron microscopy (TEM) and atomic force microscopy (AFM). Films with thickness' of 100 to 500 nm were annealed in-situ in the TEM from 250 °C to 550 °C, and characterized using a (220) two-beam condition from one of the grain orientations. This yields predominantly white on black images, which are amenable to quantitative image processing techniques. As anticipated, the grain size increased linearly during the first annealing cycle. Also so-called sub-grain boundaries were observed. These low angle boundaries originate during film deposition when arrays of dislocations are introduced in the areas where coalescing islands of the same orientation meet. After cooling and reheating the sub-grain boundaries disappear. This happens because upon cooling, misfit dislocations were introduced into the films to relieve the thermal misfit strain, resulting in a dense array of dislocations at the film / substrate interface. These thermal misfit dislocations interact strongly with pre-existing sub-grain dislocations; this, in effect, heals the microstructure.

Recent research has established that for silicon structural films used in microelectromechanical systems (MEMS), the susceptibility to premature failure under cyclic fatigue loading originates from a degradation process that is confined to the surface oxide. In ambient air environments, a sequential, stress-assisted oxidation and stress-corrosion cracking process can occur within the native oxide on polycrystalline silicon (referred to as reaction-layer fatigue); for the structural films of micron-scale dimensions, such incipient cracking in the oxide can lead to catastrophic failure of the entire silicon component. Since the degradation process is intimately linked to the thin reaction layer on the silicon, modification of this surface and the access of the environment to it can dramatically alter the fatigue resistance of the material. The purpose of this paper is to evaluate the efficacy of modifying the fatigue behavior of polycrystalline silicon with alkene-based monolayers. Specifically, 2-μm thick polysilicon fatigue structures were coated with a monolayer film based on 1-octadecene and cyclically tested to failure in laboratory air. By applying the coating, the formation of the native oxide was prevented. Compared to the fatigue behavior of untreated polysilicon, the lives of the coated samples ranged from 105 to >1010 cycles at stress amplitudes greater than ∼90% of the ultimate strength of the film. The dramatic improvement in fatigue resistance was attributed to the monolayer inhibiting the formation of the native oxide and stress corrosion of the surface. It is concluded that the surprising susceptibility of thin structural silicon films to premature fatigue failure can be inhibited by such monolayer coatings.

2-μm thick structural films of polycrystalline silicon are shown to display “metal-like” stress-life fatigue behavior in room air, with failures occurring after > 1011 cycles at stresses as low as half the fracture strength. Using in situ measurements of the specimen compliance and transmission electron microscopy to characterize such damage, the mechanism of thin-film silicon fatigue is deduced to be sequential oxidation and moisture-assisted cracking in the native SiO2 layer. This mechanism can also occur in bulk silicon but it is only relevant in thin films where the critical crack size for catastrophic failure can be exceeded within the oxide layer. The fatigue susceptibility of thin-film silicon is shown to be suppressed by alkene-based self-assembled monolayer coatings that prevent the formation of the native oxide.

Heteroepitaxial films of aluminum bicrystals grown on silicon provide a model system in which to study plasticity in polycrystalline metal thin films. For the bicrystal films, dislocations are confined to move on two different slip plane orientations because of the orientation of the crystals on the substrate. In-situ transmission electron microscopy (TEM) observations during thermal cycling have shown two threshold temperatures for dislocation motion on cooling. A simple model uses the resolved shear stress on the possible slip planes to explain the TEM observations. Mechanisms responsible for the dislocation behavior are studied in-situ during thermal cycling between room temperature and 450°C with x-ray diffraction. The strains are determined using a sin2(Ψ) analysis at each temperature. Direct comparisons are made between the TEM observations, the model and x-ray diffraction results.

Laser lift-off and bonding has been demonstrated as a viable route for the integration of III-nitride opto-electronics with mainstream device technology. A critical remaining question is the structural and chemical quality of the layers following lift-off. In this paper, we present detailed structural and chemical characterization of both the epitaxial layer and the substrate using standard transmission electron microscopy techniques. Conventional diffraction contrast and high resolution electron microscopy indicate that the structural alteration of the material is limited to approximately the first 50 nm. Energy dispersive electron spectroscopy line profiles show that intermixing is also confined to similar thicknesses. These results indicate that laser lift-off of even thin layers is likely to result in materials suitable for device integration. Additionally, because the damage to the sapphire substrate is minimal, it should be possible to polish and re-use these substrates for subsequent heteroepitaxial growths, resulting in significant economic benefits.

A novel pattern transfer technique, based upon direct contact between nanofabricated printheads and amorphous films/substrates is being developed. Transfer of features from printhead to substrate is via an induced crystallization mechanism, with a goal of controlling crystal nucleation at dimensions comparable to the printhead features. To realize such a printing mechanism, appropriate mediums/materials that can be crystallized through contact with the printhead elements must be identified. From a range of candidate materials we focus on Ge and Indium Tin Oxide (ITO) films grown by molecular beam epitaxy or sputtering. The crystallization may be induced directly by heating (thermally), pressure (mechanically), or electrical conduction through the contact points. Crystallization of many amorphous materials is known to be self sustaining (so called “explosive crystallization”), raising the need to identify mechanisms for abruptly stopping crystallization if feature dimensions are to be constrained. A set of amorphous or nanocrystalline samples have been studied during crystallization by a heated printhead fabricated with a Focused Ion Beam (FIB). A simplified continuum model suggests the feasibility of achieving lateral confinement of the induced crystallization. In order to characterize fully the transitional and final substrate states, real-time observation of nanoscale crystallization in the transmission electron microscope (TEM) is in progress. A special TEM goniometer is being developed which incorporates a piezoelectrically driven tip and a heating capability to enable observation of tip-substrate contact at elevated temperatures that should give invaluable insight into crystallization mechanisms.

The mechanical, thermal, chemical, and tribological properties of diamond make it an ideal material for the fabrication of MEMS components. However, conventional CVD diamond deposition methods result in either a coarse-grained pure diamond structure that prevents high- resolution patterning, or in a fine-grained diamond film with a significant amount of intergranular non-diamond carbon. At Argonne National Laboratory, we are able to produce phase-pure ultrananocrystalline diamond (UNCD) films for the fabrication of MEMS components. UNCD is grown by microwave plasma CVD using C60-Ar or CH4-Ar plasmas, resulting in films that have 3-5 nm grain size, are 10-20 times smoother than conventionally grown diamond films, and can have mechanical properties similar to that of single crystal diamond. We used lithographic patterning, lift-off, and etching, in conjunction with the capability for growing UNCD on SiO2 to fabricate 2-D and 3-D UNCD-MEMS structures. We have performed initial characterization of mechanical properties by using nanoindentation and in-situ TEM indentor techniques. The values of Hardness (∼88 GPa) and Young's modulus (∼ 864 GPa) measured are very close to those of single crystal diamond (100 GPa and 1000 GPa respectively). The results show that UNCD is a promising material for future high performance MEMS devices.

In-situ TEM annealing experiments on the Pd (20 nm) / a-Ge (150 nm) / Pd (50 nm) GaAs ohmic contact system have permitted real time determination of the evolution of contact microstructure. As-deposited cross-sectional samples of equal thickness were prepared using a focused ion beam (FIB) method and then subjected to in-situ annealing at temperatures between 130-400 °C. Excluding Pd-GaAs interactions, four sequential solid state reactions were observed during annealing of the Pd:Ge thin films. First, interdiffusion of the Pd and Ge layers occurred, followed by formation of the hexagonal Pd2Ge phase. This hexagonal phase then transformed into orthorhombic PdGe, followed by solid state epitaxial growth of Ge at the contact / GaAs interface. The kinetics of the solid state reactions, which occur during ohmic contact formation, were determined by measuring the grain growth rates associated with each phase from the videotape observations. These data agreed with a previous study that measured the activation energies through a differential scanning calorimetry (DSC) method. We established that the Ge transport to the GaAs interface was dependent upon the grain size of the PdGe phase. The nucleation and growth of this phase was demonstrated to have a significant effect on the solid phase epitaxial growth of Ge on GaAs. These findings allowed us to engineer an improved two step annealing procedure that would control the shape and size of the PdGe grains. Based on these results, we have established the suitability of combining FIB sample preparation with in-situ cross-sectional transmission electron microscopy (TEM) annealing for studying thin film solid-state reactions.

We present real time observations of the interaction of dislocations in heteroepitaxial strained layers using a specially modified ultrahigh vacuum transmission electron microscope equipped with in-situ deposition capabilities. These observations have led to delineation of the regime of epilayer thickness and composition where dislocation interactions result in blocking of the propagating threading segment. It is found that both the blocking probability as well as the magnitude of the dislocation interaction force are strongly dependent on the Burgers vectors of the dislocations involved, with the greatest effects observed when the Burgers vectors of the two dislocations are parallel with respect to each other. Frame-by-frame analysis of the motion of the dislocation threading segment during interaction is used to extract the magnitude of the interaction stresses as a function of both the level of heteroepitaxial strain and the dislocation geometry. Finally, by continuing growth following observations of blocking during annealing, we find that blocked dislocations are likely to remain in that configuration until substantial additional heteroepitaxial stresses are incorporated into the layer. These results have direct relevance to the successful integration of strained layer heterostructures into electronic device applications. This is because blocked threading segments result in the introduction of undesired band gap states, enhance impurity diffusion, modify surface morphology and act to limit the dislocation density reductions achievable in graded buffer structures.

In this paper we demonstrate, using both SIMS and transistor electrical characteristics, that substitutional carbon fractions of 0.5% in heavily doped Si0.8Ge0.2 base heterojunction bipolar transistors (HBTs) reduce both thermal diffusion and transient enhanced diffusion (TED) of boron. Furthermore we show that carbon suppresses TED of boron in carbon-free regions that surround the carbon layers.

We report the effects of growth conditions on the strain and crystalline quality of lowtemperature (LT) grown GaP films by gas-source molecular beam epitaxy. At temperatures below 160 °C, poly-crystalline GaP films are always obtained, regardless of the PH3 low rate used, while at temperatures above 160 °C, the material quality is affected by the PH3 flow rate. Contrary to compressively strained LT GaAs, high-resolution X-ray rocking curve measurement indicates a tensile strain of the LT GaP films, which is considered to be due to PGa antisite defects. The strain is found to be affected by the PH3 flow rate, the growth temperature, and post-growth annealing. Contrary to LT GaAs, no P precipitates are observed in cross-sectional transmission electron microscopy.