Extended abstract of a paper presented at Microscopy and Microanalysis 2013 in Indianapolis, Indiana, USA, August 4 – August 8, 2013.

Extended abstract of a paper presented at Microscopy and Microanalysis 2013 in Indianapolis, Indiana, USA, August 4 – August 8, 2013.

Extended abstract of a paper presented at Microscopy and Microanalysis 2013 in Indianapolis, Indiana, USA, August 4 – August 8, 2013.

In ductile metals, sliding contact induces plastic deformation resulting in subsurfaces, the mechanical properties of which are different from those of the bulk. This article describes a novel combination of nanomechanical test methods and analysis techniques to evaluate the mechanical behavior of the subsurfaces generated underneath a wear surface. In this methodology, nanoscratch techniques were first used to generate wear patterns as a function of load and number of cycles using a Hysitron TriboIndenter. Measurements were made on a (001) single crystal plane along two crystallographic directions, <001> and <011>. Nanoindentation was then used to measure mechanical properties in each wear pattern. The results on the (001) single crystal nickel plane showed that there was a strong increase in hardness with increasing applied load that was accompanied by a change in surface deformation. The amount of deformation underneath the wear patterns was examined from focused ion beam cross-sections of the wear patterns.

Decreasing scales effectively increase nearly all important mechanical properties of at least some “brittle” materials below 100 nm. With an emphasis on silicon nanopillars, nanowires, and nanospheres, it is shown that strength, ductility, and toughness all increase roughly with the inverse radius of the appropriate dimension. This is shown experimentally as well as on a mechanistic basis using a proposed dislocation shielding model. Theoretically, this collects a reasonable array of semiconductors and ceramics onto the same field using fundamental physical parameters. This gives proportionality between fracture toughness and the other mechanical properties. Additionally, this leads to a fundamental concept of work per unit fracture area, which predicts the critical event for brittle fracture. In semibrittle materials such as silicon, this can occur at room temperature when the scale is sufficiently small. When the local stress associated with dislocation nucleation increases to that sufficient to break bonds, an instability occurs resulting in fracture.

Atomistic simulations of nanoindentation of a 20-nm-thick Ni thin film oriented in the [111] direction were carried out to study the effects of indenter velocity and radii, interatomic potentials, and the boundary conditions used to represent the substrate. The simulation results were compared directly with experimental results of Ni thin film of the same thickness and orientation. It was found that the high indenter velocity does not affect the hardness value significantly. Different radii used for indentation also have negligible effects on the hardness value. Two different interatomic potentials were tested, giving significantly different hardness values but both within 20% of the experimental result. Different boundary conditions used to represent the substrate have a significant effect for relatively deep indentation simulations.

Dynamic indentation techniques are often used to determine mechanical properties as a function of depth by continuously measuring the stiffness of a material. The dynamics are used by superimposing an oscillation on top of the monotonic loading. Of interest was how the oscillation affects the measured mechanical properties when compared to a quasi-static indent run at the same loading conditions as a dynamic. Single crystals of nickel and NaCl as well as a polycrystalline nickel sample and amorphous fused quartz and polycarbonate have all been studied. With respect to dynamic oscillations, the result is a decrease of the load at the same displacement and thus lower measured hardness values of the ductile crystalline materials. It has also been found that the first 100 nm of displacement are the most affected by the oscillating tip, an important length scale for testing thin films, nanopillars, and nanoparticles.

Understanding the hardening response of small volumes is necessary to completely explain the mechanical properties of thin films and nanostructures. This experimental study deals with the deformation and hardening response in gold and copper films ranging in thickness from 10 to 400 nm and silicon nanoparticles with particle diameters less than 100 nm. For very thin films of both gold and copper, it was found that hardness initially decreases from about 2.5 to 1.5 GPa with increasing penetration depth. Thereafter, an increase occurs with depths beyond about 5–10% of the film thickness. It is proposed that the observed minima are produced by two competing mechanisms. It is shown that for relatively deep penetrations, a dislocation back stress argument reasonably explains the material hardening behavior unrelated to any substrate composite effect. Then, for shallow contacts, a volume-to-surface length scale argument relating to an indentation size effect is hypothesized. A simple model based on the superposition of these two mechanisms provides a reasonable fit to the experimental nanoindentation data.

Reliability of thin polymer films is strongly dependent on interfacial adhesion. This paper concerns the application of indentation combined with atomic force microscopy (AFM) for the evaluation of polymer-non-polymer adhesion. Interfacial toughness calculations have been made using AFM measurement of geometry of delaminations induced by indentation. Fracture surfaces characterized with AFM are also presented here.

The interface chosen for study is that between polystyrene (PS) and a glass substrate. Spin coating has been used to prepare PS films. PS, dissolved in toluene, was spun onto the soda-lime glass and then dried at 60°C for 3h to remove residual solvent. The thickness of the PS film was approximately 660nm. To enhance the driving force for delamination, poly(methyl methacrylate) (PMMA) overlayers were applied on top of the PS films. The PMMA overlayer, of thickness approximately 1.5 μm, was first solvent cast onto a second glass slide, floated off the glass onto deionized water, and then laid on top of a first glass substrate which had been coated with PS.

Copper films of different thicknesses of 0.2, 0.5, 1 and 2 microns were electroplated on top of the adhesion-promoting barrier layers on <100> single crystal silicon wafers. Controlled Cu grain growth was achieved by annealing films in vacuum.

The Cu film microstructure was characterized using Atomic Force Microscopy and Focused Ion Beam Microscopy. Elastic modulus of 110 to 130 GPa and hardness of 1 to 1.6 GPa were measured using the continuous stiffness option (CSM) of the Nanoindenter XP. Thicker films appeared to be softer in terms of the lower modulus and hardness, exhibiting a classical Hall-Petch relationship between the yield stress and grain size. Lower elastic modulus of thicker films is due to the higher porosity and partially due to the surface roughness. Comparison between the mechanical properties of films on the substrates obtained by nanoindentation and tensile tests of the freestanding Cu films is made.

In MEMS packages and silicon devices, the adhesion of interconnects to the substrate is a critical reliability issue. A Precracked Line Scratch Test (PLST) is among one of the available tests to measure the thin line adhesion. In the test, an initial crack is introduced at the interface between the thin line and the substrate. The line is then loaded from the precracked end. The load is recorded continuously while the crack propagates before and after the line buckles. This precracked line scratch test has been applied earlier to tungsten thin lines on silicon wafers [1]. A macroscopic version of the test was also performed to evaluate the analytical model [2]. In the macroscopic tests, polycarbonate lines were bonded to steel substrates with cyanoacrylate.

In this paper, finite element analysis is performed for the Precracked Line Scratch Test before line buckling. The energy release rates and phase angles are calculated based on the corresponding load and crack length. The results are then compared to the closed-form solution. Macroscopic experimental model along with the finite element solution has provided a way to derive the interface fracture toughness as a function of the crack length based on the load and crack length history. With the analysis in place, the precracked line scratch test can be used conveniently to study the adhesion of interconnects to passivation layers, MEMS devices and packages on different scales.

Numerous mechanisms have been identified as fundamental to the adhesion of thin metallic films. The primary mechanism is the thermodynamic work of adhesion of the interface, which in its most basic description is the difference between the surface energies of the two materials and that of the interface. This quantity is often described as leveraging the contributions of other mechanisms. One of the more important mechanisms is that of plasticity occurring in a process zone in the vicinity of the delamination boundary. A quantitative model to characterize the contributions of plastic energy dissipation has been developed and used to rationalize experimental adhesion assessments. This model incorporates the functional dependence of the film thickness and constitutive properties. Orders of magnitude increases in the practical work of adhesion were both observed and predicted. Experimentally, the films used for model comparison were sputter-deposited copper ranging from 40 to 3300 nm in thickness, with and without a thin 10 nm Ti underlayer. Nanoindentation induced delamination of the Cu from SiO2/Si wafers were evaluated in the context of composite laminate theory to determine adhesion energies ranging from 0.6 to 100 J/m2 for bare Cu and from 4 to 110 J/m2 for Cu with the Ti underlayer.

Nanomechanical devices constitute an important and growing field, as they allow for new understanding of the mechanical properties at interfaces and surfaces. As an example, a newly developed nanoindentation device has been used to accomplish μN load indents into GaAs. First, it is shown that a plastic zone can be measured and is comparable to theory. Also, it is shown that the rate of indentation affects both the depth and upset zone of low load indents, implying a strain-rate sensitivity effect at room temperature. This is reinforced by observation of what appears to be a glide-based relaxation process.

The fracture behavior of the silicon/Pyrex glass anodic bond was investigated using the methods of linear elastic bimaterial fracture mechanics. Due to the high bond strength, interfacial cracks were invariably observed to kink away from the interface into the more compliant glass under approximately mode I remote tensile loading. Kink angles measured by profilometry increased from 14 to 28°as bonding temperature increased from 300 to 450 °C. A regime of stable cracking accompanied penetration of cracks into the glass, with maximum load and corresponding fracture toughness measurement occurring at a location significantly removed from the interface. Approximately mode I, near-interface, plane-strain fracture toughness values (KIC) measured by rising load testing of chevron-notched and straight-thru-cracked compact-tension specimens increased from 0.63 to 0.68 MPa-m1/2 and 0.66 to 0.75 MPa-m1/2, respectively, as bonding temperature increased from 300 to 450 °C. In addition, XPS measurements revealed a sodium depletion zone of decreasing size and depletion magnitude with increasing bonding temperature over the same range. The near-interface region of the glass also experiences compressive residual stresses which decrease linearly with distance from the interface according to linear elastic computations. These stresses increase in magnitude with increasing bonding temperature due to enhanced differential thermal contraction upon cooling to room temperature. It is proposed that the trends in toughness and in kink angle with bonding temperature can be at least partially accounted for by variation of crack-tip shielding with compressive residual stress magnitude, the effects of interfacial. crack-tip shear stresses induced by the thermal mismatch, and by an increase in Young's modulus of the near-interface glass accompanying sodium depletion.

The micro-wedge scratch technique was applied to characterize interfacial properties of copper fine lines on PMDA-ODA polyimide. A 20 pm wide, 0.1 μm radius wedge-shapeddiamond tip was used to mechanically debond Cu fine lines from polyimide substrate. Copper fine lines varies from 0.2 to 0.4 μm thick, and from 1.μm to 100 μm wide. Both cohesive and adhesive delaminations were obtained. Loaddisplacement curves were recorded and the scratch morphologies were observed by SEM. From a first order calculation, the fracture toughness of the polyimide and interfacial strength of Cu-PI were obtained.

The mechanics for plane strain geometry are presented for thin film fine lines by considering the case of microwedge indentation of a fine line subject to no residual stress. The analysis is separated into three parts depending on the absence or presence of buckling. During indentation, unbuckled or double-buckled configurations may exist, while single-buckling may occur when the indenter tip is removed. A microwedge tip was employed in the indentation testing of unstressed rf-sputtered tungsten fine lines to demonstrate some of the analysis.

The microstructure of Ta2N thin films deposited by d.c. magnetron sputtering on (1120) surface of Al2O3 is investigated using transmission electron microscopy. The effects of exposing the thin film structure to a 600°C air environment are also explored. It will be shown that under the standard deposition conditions, stresses exist in the thin film structure which leads to the formation of a textured structure in the as-deposited Ta2N. Exposure of the thin film structure to an air environment transforms the Ta2N to Ta2O5 in the orthorhombic structure. In addition, evidence for a epitaxial relationship between the Ta2O5 and Al2O3 will be presented.

Direct observations of dislocation arrangements about a range of microindentations into the [100] face of an Fe−3 wt. % Si single crystal have been accomplished. Dislocations of both large loop character initiated from the indenter and small loop character initiated as secondary reactions are found. Analysis of these allows the various contributions to the plastic strain gradient around the indentation to be assessed, with the experimental observations being reasonably consistent with continuum models.

Atomic force microscope measurements of elastic-plastic indentation into an Fe−3 wt. % Si single crystal showed that the volume displaced to the surface is nearly equal to the volume of the cavity. The surface displacement profiles and plastic zone size caused by a 69 nm penetration of a Vickers diamond tip are reasonably represented by an elastic-plastic continuum model. Invoking conservation of volume, estimates of the number of dislocations emanating from the free surface are reasonably consistent with the number of dislocations that have formed in the plastic zone to represent an average calculated plastic strain of 0.044.

A modified form of nanoscratch test, continuous micro-wedge scratch technique was applied to characterize the adhesion strength of 0.2 μm thick, 10 μm width sputtered copper fine lines on 50 μm thick solid PMDA-ODA polyimide films. A 20 μm width wedge-shaped diamond tip was used to mechanically debond Cu fine lines to polyimide substrate. Using a microindenter under control depth testing mode, critical loads for delaminations were obtained. The scratch morphology and the debonded areas were observed by scanning electron microscopy. The scratch test results by conical-shaped tip on plain films are also presented here and are compared to the results by wedge-shaped tip in fine metal lines. Practical work of adhesion of 4.1 J/m2 was calculated by FEM based on bimaterial fracture mechanism