Jeung-hyun Jeong

Film-thickness considerations in microcantilever-beam test in measuring mechanical properties of metal thin film

Since metal thin films have broad applications in micromechanical devices, their mechanical properties are very important in designing and controlling such devices. In this study, the Young’s modulus and yield strength were estimated by using a microcantilever beam-bending technique. Al and Au cantilever beams of various thicknesses and lengths were fabricated by silicon bulk micromachining and were loaded in bending by a nanoindenter; displacement was simultaneously obtained at high resolution. The load–displacement data showed that the yield strengths of the two films increase with decreasing film-thickness. However, the Al and Au films showed different mechanical behaviors: the yield strength of the Au film followed the grain-size dependency of Hall–Petch type, but that of the Al film was not fully described by the grain-size effect alone. This difference arises mainly from the different surface conditions of each film. The effect of the surface on a film’s mechanical properties is discussed and a microstructure effect was suggested that a modified misfit dislocation theory used to predict the yield strength. 2003 Elsevier Science B.V. All rights reserved.

Evaluation of elastic properties and temperature effects in Si thin films using an electrostatic microresonator

Laterally driven microresonators were used to estimate the temperature-dependent elastic modulus of single-crystalline Si for microelectromechanical systems (MEMS). The resonators were fabricated through surface micromachining from silicon-on-glass wafers. They were moved laterally by alternating electrostatic force at a series of frequencies, and then a resonance frequency was determined, under temperature cycling in the range of 25/spl deg/C to 600/spl deg/C, by detecting the maximum displacement. The elastic modulus was obtained in the temperature range by Rayleighs energy method from the detected resonance frequency. At this time, the temperature dependency of elastic modulus was affected by surface oxidation as well as its intrinsic variation: a temperature cycle permanently reduces the resonance frequency. The effect of Si oxidation was analyzed for thermal cycling by applying a simple composite model to the measured frequency data; here the oxide thickness was estimated from the difference in the resonance frequency before and after the temperature cycle, and was confirmed by field-emission scanning electron microscopy. Finally, the temperature coefficient of the elastic modulus of Si in the direction was determined as -64/spl times/10/sup -6/[/spl deg/C/sup -1/]. This value was quite comparable to those reported in previous literatures, and much more so if the specimen temperature is calibrated more exactly.

Mechanical analysis for crack-free release of chemical-vapor-deposited diamond wafers

Chemical-vapor-deposited (CVD) diamond thick films for electronic applications must be released without cracks from the substrate as freestanding wafers. In this study, the mechanism of cracking in the CVD films was investigated experimentally and theoretically. Experimental observations showed that cracks initiated at the edge of the diamond wafer and then propagated towards the center. Finite-element analysis (FEA) reveals that, during cooling, compressive thermal stresses concentrate at the thick films edge and additional tensile stress acts circumferentially. This was verified by the experimental analysis of diamond films deposited on Si, Mo and W substrates. Observations on low interfacial adhesion and crack-free film on the W substrate indicated that, in addition to the high thermal stress, low interfacial adhesion plays an important role in cracking. Thus, film cracking depends on the fracture strength of the film and its relative magnitude with respect to interfacial adhesion. Methods of crack suppression were suggested on the basis of this cracking mechanism: increase of film thickness and minimization of the substrates CTE and interfacial adhesion. The analysis was confirmed by successful suppression of cracking by application of a low-adhesion interlayer prior to deposition of diamond film.

A new indentation cracking method for evaluating interfacial adhesion energy of hard films

Abstract The linear dependency of interfacial crack length on indentation load in a spherical brale C indentation (the so-called indentation cracking test), d P /d c , has been generally used as the criterion of adhesion for a hard film on a soft substrate. However, this criterion is a relative yardstick that is valid only in materials of same kind. For a more reliable evaluation of adhesion, the energy release rate concept was introduced to relate the dependence of interfacial crack on indentation load to the release of stored energy in the film; here the stress distribution around the indentation is required. The stress state in the film was deduced using the condition of strain continuity at the interface from the stress state in the substrate and then the critical energy release rate G c was obtained as an analytical function of d P /d c and material properties. This method was applied to the adhesion measurement of diamond-like carbon film deposited on AISI D2 steel. The effects of deposition condition and residual stress on adhesion were discussed on the basis of the G c results.

Intrinsic stress in chemical vapor deposited diamond films: An analytical model for the plastic deformation of the Si substrate

The intrinsic stress in diamond film deposited on a Si substrate is difficult to measure because high-temperature deposition induces plastic deformation in the Si and so renders useless an elastic solution. In this study, an analytical model is proposed to estimate intrinsic stress using a substrate-curvature technique and considering the plastic deformation of substrate. The stress distribution of the as-deposited film is affected not only by the intrinsic stress of the film but also by the bending and plastic deformation of the substrate. In this model, the distribution is formulated, based on elastic/plastic plate-bending theory, in terms of substrate curvatures, intrinsic stress in the film, and yield stress of the substrate. The intrinsic stress of the film together with the yield stress of the substrate can be obtained from experimentally measured substrate curvatures by solving two equilibrium equations and a moment-relaxation equation describing the film removal. Diamond films were deposited by mi...

Evaluation of the adhesion strength in DLC film-coated systems using the film-cracking technique

A theoretical analysis of the adhesion dependence of film-cracking behavior resulting from the uniaxial tensile loading of a hard film on a ductile substrate is presented. An interface shear stress that develops due to the deformation mismatch of the film and substrate induces a normal tensile stress in the film; these stresses are analyzed as a function of external strain and crack spacing, using the shear lag theory. When the film tensile stress exceeds the film fracture strength σc, film cracking occurs at an early stage of uniaxial loading; but as the loading increases, the interface shear stress increases above the critical value τc and causes the interface failure. After that, no more film cracking occurs, since stress transfer into the film is impossible due to interface damage. Thus, the interface adhesion can be estimated in terms of the shear strength from the external strain esat and crack spacing λsat measured at the time that film cracking stops. The test results for the diamond-like carbon (...

A method of determining the elastic properties of diamond-like carbon films

The elastic properties of diamond-like carbon (DLC ) films were measured by a simple method using DLC bridges which are free from the mechanical constraints of the substrate. The DLC films were deposited on a Si wafer by radio frequency (RF ) glow discharge at a deposition pressure of 1.33 Pa. Because of the high residual compressive stress of the film, the bridge exhibited a sinusoidal displacement on removing the substrate constraint. By measuring the amplitude with a known bridge length, we could determine the strain of the film which occurred by stress relaxation. Combined with independent stress measurement using the laser reflection method, this method allows the calculation of the biaxial elastic modulus, E/(1’n), where E is the elastic modulus and n is Poisson’s ratio of the DLC film. The biaxial elastic modulus increased from 10 to 150 GPa with increasing negative bias voltage from 100 to 550 V. By comparing the biaxial elastic modulus with the plane‐strain modulus, E/(1’n2), measured by nano-indentation, we could further determine the elastic modulus and Poisson’s ratio, independently. The elastic modulus, E, ranged from 16 to 133 GPa in this range of the negative bias voltage. However, large errors were incorporated in the calculation of Poisson’s ratio due to the pile up of errors in the measurements of the elastic properties and the residual compressive stress.

Analytical model for intrinsic residual stress effects and out-of-plane deflections in free-standing thick films

An analytical model for the influence of residual stress on the out-of-plane deflection in a free-standing thick diamond film (the bowing phenomenon) is presented. The variation in residual stress with film thickness is usually believed to cause the bowing. In this study, the stress variation is assumed to be produced by a gradual increase in substrate deformation resulting from layer-by-layer deposition of the film. The model uses the infinitesimal plate-bending theory to describe the layer-by-layer film growth more exactly, considering the two deformation modes of contraction or expansion and bending. To verify the suggested model, thick diamond films were fabricated on Si, Mo, and W substrates of varying thicknesses by microwave plasma assisted chemical vapor deposition. The model’s predictions on bowing, based on the intrinsic stress value measured by the curvature method, were in good agreement with the bowing curvature of the as-released films measured by a profilometer. This confirms that the bowin...

Effect of Oxygen Addition on Residual Stress Formation of Cubic Boron Nitride Thin Films

In this study we investigated the oxygen effect on the nucleation and its residual stress during unbalanced magnetron sputtering. Up to 0.5% in oxygen flow rate, cubic phase (c-BN) was dominated with extremely small fraction of Hexagonal phase (h-BN) of increasing trend with oxygen concentration, whereas hexagonal phase is dominated beyond 0.75% flow rate. Interestingly, the residual stress in cubic-phase-dominated films was substantially reduced with small amount of oxygen () down to a low value comparable to the h-BN case. This may be because oxygen atoms break B-N bonds and make B-O bonds more favorably, increasing bonds preference, as revealed by FTIR and NEXAFS. It was confirmed by experimental facts that the threshold bias voltage for nucleation and growth of cubic phase were increased from -55 V to -70 V and from -50 V to -60 V respectively. The reduction of residual stress in O-added c-BN films is seemingly resulting from the microstructure of the films. The oxygen tends to increase slightly the amount of h-BN phase in the grain boundary of c-BN and the soft h-BN phase of 3D network including surrounding nano grains of cubic phase may relax the residual stress of cubic phase.

An Analytical Model for Intrinsic Residual Stress Effect on Out-of-Plane Deflection in Chemical-Vapor-Deposited Free-Standing Thick Film

The effect of residual stress on the out-of-plane deflection in a free-standing thick diamond films was investigated theoretically and experimentally. The deflection is believed to be caused by the variation in residual stress with film thickness. Key idea of this study is that the stress variation may be produced by gradually increasing substrate deformation resulting from the layer-by-layer deposition of the film. The layer-by-layer deposition was modeled by using infinitesimal plate-bending theory, considering the two deformation modes of contraction or expansion and bending. To verify the suggested model, several hundred micron thick diamond films were fabricated on Si, Mo and W substrates of varying thicknesses by microwave plasma assisted chemical vapor deposition. The models predictions on bowing, based on intrinsic stress value measured by the curvature method, were in good agreement with the experimentally measured curvature of the as-released films. Finally, it is concluded that the bowing of CVD thick films depends on the intrinsic stress variation of the film associated with gradual increase in substrate deformation.