ABSTRACTS

The following is a list of the abstracts for papers which will be presented in THE SECOND INTERNATIONAL SYMPOSIUM ON ADHESION MEASUREMENT OF THIN FILMS AND COATINGS. The listing is alphabetical by presenting author. This list is updated continually to add abstracts as they become available and make appropriate corrections. This list may be conveniently searched by using the editor provided with most popular browsers (e.g. Microsoft Explorer, Netscape, ... etc.)

Repeatability in the Micro-scratch Measurement for the Adhesion of Thin Films and the Physical Meaning.

The microscratch tester can measure the film detachment very sensitively. We employed scratching styluses of various tip radii and tests were repeated on the specimen of MgF2 and Cu films of about 100 nm thick, which were prepared under almost the same condition on glass substrates. The scatter of the critical load vallues and the dependence on the stylus radius were studied. As for the magnesium fluoride a Weibull distribution was applied to the analysis of the scatter, though the shape of the scatter was not so clear despite of about 500 tests. With the increase in the stylus radius from 0.015 to 0.100 mm, both the minimum and the mean critical load increased naturally, but the critical shear stress of Benjamin-Weaver decreased. On the other hand, the interfacial energy values calculated by taking the friction force into account were found almost the same for different stylus radius. The critical load was found to correlate with the friction, which was not so strong. Fresh copper films, which had not been exposed to air for more than two days, showed a ductile fracture. The critical load of Cu films decreased a bit with the air exposure, and the fractured surface changed to be ragged as if they were crumbled. The scatter of the critical load was also observed, but the distribution was rather close to the Gaussian.

Mode II Edge Delamination of Compressed Thin Films

Ceramic coatings deposited on metal substrates generally develop significant compressive stresses when cooled from the temperature at which they are processed as a result of thermal expansion mismatch. One of the main failure modes for these coatings is edge delamination. For an ideally brittle interface, the edge delamination of a compressed thin film involves mode II interface cracking. The crack faces are in contact with normal stress acting across the faces and behind the advancing tip. Experimental measurements suggest that the mode II fracture toughness is larger than predicted by theoretical calculations. Frictional shielding of the crack tip has been shown to increase the apparent fracture toughness and it is likely that roughness effects associated with the separating faces also contribute. A model of mode II edge delamination that incorporates friction between the delaminated film and substrate as well as surface roughness effects is proposed and analyzed. This model is used to determine if frictional shielding and surface roughness effects are sufficient for explaining the large apparent mode II fracture toughness found in experiments.

Mechanistic Studies of Adhesion of Thermal Spray Materials

(Abstract not yet available)

Relative Adhesion Measurement for Thin Film Microelectronics Structures

There is a difference between technologically important adhesion or practical adhesion, and fundamental or basic adhesion. What is important in the understanding of fundamental adhesion may be of insignificant interest to technology. A manufacturer is interested in how to improve the reliability of the structure being built should an interface problem exist, rather than knowing the precise value of the fundamental adhesion. It is not possible to measure fundamental adhesion for technologically important structures due to the inability to account for all energy dissipating processes during the test. Adhesion measurements are plagued with the mode of interface loading issue: the resemblance of test interface loading to that of the actual manufactured part. What technology needs is a simple adhesion test method that is practical for product development, giving reliable information about the interface integrity. The present paper compares the value of two adhesion tests for microelectronics applications and emphasizes the importance of locus of failure analyses. A realignment of structure reliability modeling is suggested by the usage of effective fundamental adhesion instead of the standard undeterminable fundamental adhesion.

Can the Scratch Adhesion Test Ever be Quantitative?

The scratch test has been used to assess the adhesion of thin hard coatings for some time now and is a useful tool for coating development and quality assurance. However, the test is influenced by a number of intrinsic and extrinsic factors which are not adhesion related and the results of the test are usually regarded as only semiquantitative. The stress state around a moving indenter scratching a coating/substrate system is very comples and it is difficult to determine the stresses which lead to detachment. Furthermore the interfacial defect state responsible for failure is unknown. However, by a careful analysis of the observed failure modes in the scratch test (not all of which are related to adhesion) it is possible to find some which occur in regions where the stress state is relatively simple and quantification can be attempted.

Ideally engineers would like a material parameter (such as work of adhesion or interfacial toughness) which can be used in an appropriate model of the coating substrate system stress state to determine if detachment will occur under the loading conditions experienced in service. This data is not usually available and the development of such models must be seen as a long term goal. It must be questioned whether the scratch test can ever give quantitative information suitable for this modeling analysis. This paper will highlight the main adhesion-related failure modes from hard coatings, thermally grown oxide scales and brittle organic coatings. The use of empirical calibration studies and quantification by finite element methods will be discussed.

Scratch Rest Failure Modes and Performance of Organic Coatings for Marine Applications

Although the scratch test is widely used for the assesment of thin hard coatings, its use for the testing of organic coatings is much less widespread. This is partly because these organic layers are much thicker than the hard coatings and hence interfacial loadings are much smaller and partly because the low hardness of the coatings means that they will plastically deform considerably during the test. Organic coatings for the protection of marine structures are often based on epoxy resins or similar materials which are brittle compared to other organic resins. Furthermore, in service these materials undergo photo-oxidation which can increase their brittleness or introduce thin modified layers which have poorer properties than the as-received material. Corrosion of the underlying substrate may also lead to changes in coating adhesion. The scratch test is an ideal way of making quantitative comparisons between these materials both in the as received state and after exposure to UV or salt water. In this study a number of organic coatings and base resins used for marine protection have been assessed by the scratch test. Both through-thickness cracking and interfacial detachment have been observed when the coatings are less than about 200 microns thick with mainly plastic deformation for thicker coatings. Interfacial detachment is usually observed as a consequence of through-thickness cracking and initiates behind the moving indenter, contrary to what is usually ovserved for hard coatings. The critical loads for through-thickness cracking and interfacial detachment are reduced after exposure to salt water. The use of the scratch test to monitor performance of these coatings will be discussed.

Xiang "Sam" Dai, Mark V. Brillhart¹, Michael Roesch, and Paul S. Ho²; Hewlett-Packard Co., 1501 Page Mill Rd., MS 5L-C, Palo Alto, CA 94304

1) Cisco Systems, Inc., 170 West Tasman Drive, San Jose, CA 95134-1706

2) University of Texas, Materials Lab for Interconnect & Packaging, PRC/MER, MC R8650, Austin, Texas 78712

Adhesion Measurement and Toughening Mechanisms at Underfill Interfaces for Flip-Chip-On-Organic-Substrate Packaging

The flip-chip-on-organic-substrate (or flip-chip-on-laminate) packaging technology utilizes a particulate reinforced structural epoxy (underfill) to adhere the chip to the package or board. Although the adoption of underfill encapsulation leads to improved reliability of flip-chip solder interconnections, delamination at various interfaces becomes a major concern for assembly yield loss and package reliability. In spite of their great importance, the adhesion and fracture behaviors of UF/PCB and UF/Si interfaces have not been investigated until recently. A great deal of confusion and controversy still exists over the effects of underfill formulation and the adhesion and toughening mechanism of the interfaces. The present work focuses on investigating the effects of several key variables for UF/chip and UF/organic substrate systems on the interface adhesion strengths. These variables are underfill organosilane content, filler particle content, rubber particle content, surface morphology and chemistry of the chip and organic substrates. The approach is to quantify interfacial fracture energy using the double-cantilever-beam (DCB) specimens developed previously and study the effect of individual variables in UF/chip and UF/organic substrate systems. Discussions will be focused on the effect of underfill filler content, organosilane content, rubber particles, and surface morphology and chemistry of the chip and organic substrates. It is found that the UF/substrate interface adhesion and fracture behaviors are controlled by several distinct mechanisms operating and competing with each other, such as formation of primary bonds, crack-pinning by glass fillers, debonding of glass filler from epoxy matrix(defect formation), and cavitation and shearing induced by rubber particles. Fundamental understandings of the interfacial adhesion and toughening mechanisms can provide guidance for developing new processes and materials to enhance interfacial adhesion and reliability.

K.J. Koivusaari, Microelectronics and Material Physics Laboratories and EMPART Research Group of Infotech Oulu, Department of Electrical Engineering University of Oulu, PL 444, FIN-90571 Oulu

Adhesion Testing of Nanophase Diamond Coatings on Steel and Carbide Substrates

Films of nanophase diamond are deposited in vacuum onto almost any substrate by condensing multiply-charged carbon ions carrying keV energies. These ions obtained from the laser ablation of graphite at high intensities. The high energies of condensation produce interfacial layers between the film and substrate materials, resulting in levels of adhesion which can support the protection of substrates subjected to harsh environmental conditions. In this work, we give details of the adhesion testing performed on steel and carbide substrates coated with nanophase diamond. A commercially available scratch tester and a modified sand blaster were used to simulate high impact operating conditions. Data analysis were presented to assess the adhesion of films. The characterization studies in this work demonstrated nanophase diamond as a highly adherent coating suitable for industrial applications.

Progress in Measuring Residual Stresses in Coatings Through Bilayer Curvature

Residual stresses are known to play a significant role in the delamination and failure mode of adhesive layers and coatings . Curvature measurements of bilayer systems have been employed for several centuries to obtain information related to relative shrinkage and expansion of the layers. Timoshenko and Corcoran have laid the groundwork for beam and plate bending models, respectively, associated with this phenomena. Recent work has sought to determine optimal geometries which improve the sensitivity of the techniques for measuring the coefficients of thermal expansion for the coating. Common, commercially available DMA (dynamic thermal analysis) and TMA (thermal mechanical analysis) equipment provides an especially attractive approach for quantifying the displacements under carefully controlled temperature exposure. The purpose of this paper is to present recent results showing how this approach can be used to quantify residual deformation states within coatings, and how these states may change with aging and other processing variables. Guidelines for appropriate specimen configurations are also given. This simple technique may prove useful in modifying processing steps to minimize residual stresses or increase dimensional stability of coated or bonded systems.

Measurement of Adhesion of Hard Coatings on Ductile Substrates

(Abstract not yet available)

Methods for Measuring the Adhesion of Films and Substrates on Ductile Substrates

(Abstract not yet available)

1) Department of Physics, University of Aveiro, 3810 Aveiro, Portugal

2) Department of Mechanical Engineering, University of Aveiro, 3810 Aveiro, Portugal

* E-mail: fan@fis.ua.pt

Raman Spectroscopy of Residual Stresses in Diamond Films

Diamond coatings are attractive for potential applications due to their extreme properties, such as a large band gap, wide range of optical transparency, the highest hardness and Young's modulus as well as thermal conductivity of any known material. Most of the applications require that the films adhere adequately to the substrate. It has been generally recognized that the presence of residual stress is an important aspect of the coating reliability affecting the film/substrate adhesion. So, a quantitative evaluation and understanding of the film stresses are necessary for optimizing process conditions and designing reliable diamond coatings. Residual stresses in chemical-vapor-deposited (CVD) diamond films are composed of two sources, namely, intrinsic stress due to lattice mismatch with substrate and defects in the film, and thermal stress originated from a thermal-expansion mismatch between the film and substrate on cooling. In this presentation, we report the determination of residual stresses in CVD diamond films by micro-Raman spectroscopy combined with a bi- metal theory and a plate bending theory. Two important results are obtained from this work, i.e., a) the residual stresses are not uniformly distributed in the film. b) The diamond Raman peak broadening can be explained assuming the bending of the film caused by the thermal mismatch.

Measurement of Interfacial Fracture Energy in Microelectronic Multifilm Applications

(Abstract not yet available)

Improvement and Test of Diamond Film Adhesion on Steel Substrate by Cu/tic Sputtering Layer

Diamond thin films were deposited by hot filament CVD method on Cu/TiC sputtering interlayer on steel substrates. Characterization of X-ray, SEM and Raman spectroscopy was carried out for film surface morphology and structure. A vertical tensile test was performed to determine diamond film adhesion strength. Since the poor adhesion strength of diamond film with most materials,it is many affords to improve diamond film adhesion strenth. However, it is few quantitative results of diamond adhesion strength. In order to ensure the fracture during tensile test occurs and only occurs at the interface of diamond film and the substrate, a thin metal sputtering thin layer was covered on fresh diamond film and subsequently a surface treatment was carried out to enhance the coherence between diamond film surface and the tensile bar. Therefore, quantitative results of film adhesion strength could be obtained. The results indicated that the sputtering intermediate layer plays an important role not only on promoting diamond nucleation but also on diamond film adhesion strength on steel substrates.

The JKR Technique of Measuring Adhesion

The Johnson-Kendall-Roberts (JKR) technique provides a practical means to measure the work of adhesion of surfaces. The basic theory behind this technique is reviewed in this work using a fracture mechanics approach. Basic experimental techniques are also reviewed. Some limitations of this theory, together with possible corrections, are discussed. Extension of this technique to dissipative systems are also discussed. Such systems will exhibit hystersis during the loading and unloading phase of a JKR test. Hystersis can occur due to surface rate dependent effects or bulk inelastic deformation (e.g.viscoelasticity). Recent theoretical and experimental results for these two distinct sources of hystersis are reviewed. Emphasis is placed on the distinction between the energy approach and the stress intensity factor approach. Some suggestions are given for the interpretation of experimental data of viscoelastic systems.

Residual Stress Evolution During the Tensile Failure of submicron Multilayers

Micromechanical tensile tests on submicron multilayers, deposited on a pure Ti substrate were performed in a scanning electron microscope. During the tests, the evolution of damage with increasing tensile strain was followed. The thin layers cracked until a critical strain was reached. Depending on the composition of the layers, some multilayers showed debonding, before or after reaching crack saturation. It was observed that debonding was caused by the deflection of cracks at the interfaces. The stresses in the films were determined by X-ray diffraction. Also, energy dispersive spectroscopy was used to determine which interfaces were exposed after cracking. By combining the measured stress-strain data, the observations during testing in the scanning electron microscope, the X-ray diffraction and energy dispersive spectroscopy data, we were able to identify the weak interfaces in the multilayers. Analytical models and calculations are discussed and applied, in order to determine critical parameters of the multilayer on substrate systems; for example the fracture energies of the films and of the weak interfaces.

Jet Business Unit, Corvallis, Oregon, USA

Blade Adhesion Test Applied to Polyimide Films on Silicon

The blade adhesion test characterizes adhesion strength by measuring the load required to remove a film from a semi-rigid substrate with a sharp blade. The blade adhesion test was used to quantify the adhesion strength of polyimide films spun on Si with thicknesses varying from 8 um to 30 um. Adhesion strength was altered by applying Ta- or SiC-coatings to the silicon substrate and by subjecting assembled specimens to elevated temperature aqueous environments. Results indicate that blade tangential force generally increases with increasing polyimide thickness and decreases as a result of elevated temperature exposure. Results quantify the sensitivity of the measured response to interface strength. A finite element analysis of interface separation by the blade was developed to assess the interface strength from the coating removal force data and from geometry of the film as it is removed by the blade. The effect of the polyimide residual stress, the blade to substrate friction, and blade to polymer friction on the measured adhesion strength was considered. The influence of the input parameters including, blade angle of incidence and blade normal force was also modeled.

Application of the blade adhesion test to practical adhesive strength measurements is discussed. The generalization of the blade adhesion test to other materials systems and other dimensional scales is considered.

Energetic Approach for the Evaluation of Adhesion of Sputter Deposited TiC Films Measured by Scratch Test

TiC thin films were deposited onto glass substrates in a double target (Ti and C alternate magnetron sputtering apparatus. Adhesion of the thin film was measured by a scratch method using a diamond tip. The adhesion was found to depend on the composition of the thin film. Internal stress, hardness and wear of the film were also measured in relation to the adhesion. The mechanism of the scratching is considered by the energetic approach. The scratching is essentially a process of pressing the film an of giving the film the momentum of a diamond tip. This process is a combination of a compressive and shear stress application by a diamond tip to the normal and parallel direction of the film surface. The sear stress application and the momentum transfer are related to the roughness of the film surface and the shape of the diamond tip. The evaluation of these effects is not simple and here, we evaluate the energy accumulation due to the compressive stress. The value of the energy is estimated by a simple model. The relation between the accumulated energy and the film adhesion is discussed. The efect of the spontaneously generated internal stress in the film on the adhesion, which is equivalent to the elastic strain energy in the film, is taken into account.

Quantification of Coating Adhesion Using Laser Induced Decohesion Spectroscopy

We present a new technique, laser induced decohesion spectroscopy (LIDS), which is capable of measuring the practical work of adhesion G between a transparent polymer coating and an opaque coating or substrate. In LIDS, a laser pulse directed onto the sample creates a blister at the transparent/opaque interface. The blister's internal pressure depends on the laser pulse energy, and at a critical pressure the sample fractures, creating an annular debond similar to that obtained in the standard blister test. By measuring physical variables such as the curvature of the blister, and its radius and thickness, it is possible to deduce G. Here we measure G between an automotive clearcoat and four opaque basecoats of various pigmentations (black, white, red, metallic green) as a function of clearcoat thickness. We find that G depends on pigmentation due to the various pigment volume concentrations (PVC's) and specific pigment-binder interactions. Also, G depends on the clearcoat thickness when the thickness is comparable to the size of the plastic zone, Rp. We will also present data on other interfaces of interest to automotive coatings, such as basecoat-primer, primer-electrocoat, and clearcoat-TPO (thermoplastic olefin).

A Pulsed Laser Technique to Evaluate the Adhesion Strength of Oxide Films on Metallic Substrates

A pulsed laser technique (Nd:YAG laser 1.06µm, FWHM 14.5ns), based on conventional laser acoustics testing, has been developed to define the spallresistance of oxide films on a metallic target. Direct irradiation of the coated side of the sample generates an acoustic source at the innerfilm-metal interface.A qualitative model of the laser induced epicentral displacements has beenproposed to evaluate the critical laser irradiance corresponding toprogressive damages at the inner film-metal interface, i.e. mainly the interface delamination. Experimental measurements of the displacements generated by the laser irradiation (few nanometers at 50 Khz) were performed with a heterodyne interferometric probe both at the opposite irradiated face and also directly at the epicenter of the irradiated face. More recently, a quantitative numerical model of the thermoelastic source at the inner interface has been defined to calculate the adhesion strength of the

coatings. This model allows us to calculate the contribution of both the thermal expansion and the acoustic wave generation to the stress generated at the inner film-metal interface leading to the decohesion of the film (buckling). The behavior of sputtered and plasma sprayed Al2O3 films of various thickness (1 to 200 microns) deposited on stainless steels substrates has been tested.

Characterization of Thin Film Adhesion with the Nano-Scratch Tester (NST)

As the thickness of functional coatings continually decreases to satisfy the structural and protective needs of modern day thin film applications, quantitative instrumentation has become a necessity for adequate evaluation of material properties, particularly scratch resistance and adhesion at the film-substrate interface. The Nano-Scratch Tester (NST) is a new instrument overcoming the limitations of both the classical stylus scratch test (normal force range) and the scanning force microscope (SFM) technique (short sliding distances), allowing scratch lengths of up to 10 mm. Tangential force and penetration depth are simultaneously measured during the scratching process, in a multipass contact fatigue mode. For high resolution inspection of the deformed or damaged area, a scanning force microscope is integrated into the system. Experimental results are presented for a range of applications, particularly curved sample geometries and the use of scratch mapping. The results indicate very good reproducibility and confirm the application of this nes instrument for the accurate characterization of elasticity, hardness, adhesion and mechanical integrity in coated systems where the film thickness is less than 1 micron.

Adhesion of Thin Hard Films to Polymers - a Working Model

The use of polymers is ubiquitous in the current times. There are many forms of applications of which one is the use of coated polymers. The deposition of coatings on polymers can be done in many ways. This publication describes the working model for the adhesion of Titanium Nitride thin hard films on the thermoplastic polymers poly(butylene terephthalate), poly(amide) and poly(carbonate). Also the change of the coating morphology with process parameters according to the model of Thornton is demonstrated.

The deposition method investigated herein is Magnetron Sputtering Ion Plating (MSIP). The physical mechanism for attaining adhesion of the thin films on the polymers is described. The deposition process is subdivided into several steps, which have been derived from depth profiling with Ar ion beam etching and X-ray photoelectron spectroscopy, Secondary Ion Mass Spectroscopy (SIMS) and Optical Emission Spectroscopy (OES).

The model shows how the matter generated in the sputter deposition process interacts with the polymer substrate to form an intimately intermixed region at the surface of the polymer. This region exhibits a remarkably high content of oxygen, which is shown to result directly from the deposition mechanism.

The structure of the coatings shows a dependency on the kind of polymer used as substrate.

Adherence Of Organic Coatings On Metals: The Role Of The Interphase And Its Residual Stresses.

When epoxy resins are applied onto metallic substrates and cured, internal stresses are developed within the organic layer. These stresses reduce the adhesive strength and occasionally induce cracking, buckling of coatings. Mechanical properties (residual stresses, Young's modulus and practical adhesion) of organic layers (DGEBA DER 332 epoxy resin and IPD hardener) with different thicknesses were measured. The coatings were deposited on aluminum (5754) and titanium (Ti6Al4V) alloys after different surface treatments or on gold coated metallic sheets. By using one and two dimensional approaches based on the plate and curve beam theories, either bi-layer (assuming a perfect polymer/metal interface with a null thickness) or tri-layer models (with an real polymer/metal interphase) were considered to evaluate residual stresses and Young's modulus of the entire coating, the polymer/metal interphase and the remaining part of the coating having the bulk properties. Physical and chemical properties of thin, thick coatings, interphase and bulk properties were determined by using differential thermal analysis (DSC) and FTIR spectroscopy. Young's modulus, curvature radii of coated samples and practical adhesion were determined by 3 point flexure tests. The locus of the failure initiation was determined by scanning electron microscopy (SEM), electron microprobe (EMPA) and FTIR spectroscopy. Results show the limitation domain of the bi-layer model when the coating/substrate interphase is formed. For the same coating formulation and the same curing conditions, residual stresses, interphase Young modulus, practical adhesion, conversion and Tg of coatings depend on the nature of the substrate, the surface treatment and on the coating thickness. A gradient of residual stresses and modulus were observed according to the different surface treatments. When adhesive failure was observed, the adherence increases when internal stresses within the polymer/metal interphase decrease.

Scratch Adhesion Testing of Thin Hard Tribological Coatings - Challenges and New Opportunities

Nowadays coatings are deposited in numerous ways for a variety of applications. The coatings possess different properties to fulfil the requested requirements, but the most important property of the coating is the adhesion to the substrate. The tribological coatings are typically exposed to high surface pressures and velocities and therefore require sufficiently good adhesion in order to be functional in the application. The practical adhesion, which is the function of fundamental adhesion and method-dependent factors, signifies "the force of the work required to detach the coating from the substrate". The practical adhesion can be measured with several different techniques. One of the techniques that has received widespread scientific and industrial acceptance, besides all the criticism, is the scratch adhesion testing. During the early days rather simple instruments with only an acoustic monitoring of the scratch incident were applied, but lately many technical improvements and developments have been carried out and nowadays computer controlled systems are available. Several factors affect the scratch testing, including the indenter and its condition, bulk and surface properties of the coating and the properties of the substrate. The major effects and problems related to scratch adhesion testing of hard tribological coatings will be reviewed in the presentation. Also the latest work carried out to improve the accuracy and reliability of the scratch test method will be discussed.

On the Evaluation of Substrate-coating Adhesion by Indentation Experiments

Adhesion between a substrate and a coating is discussed as a technological performance criterion which is usually classified and evaluated with respect to specific loading conditions. Common principles for the evaluation of adhesion are summarized and the particular meaning of indentation experiments are outlined. Depth sensing indentation measurements are described in more detail. A simple approach for modeling the indentation processes as a superposition of elastic and plastic deformation portions is presented. Consequences for the response behavior of substrate-coating systems are derived. The discussion of the thus expected effects is combined with a summary of experimental results from investigations of different substrate-coating systems such as hard coatings on tool steels or coatings for microsystem applications. Relations between specific coating conditions and adhesion related effects in the indentation response are discussed. The meaning of internal coating stresses is particularly underlined. The present state of the evaluation of substrate-coating adhesion by depth sensing indentation measurements is critically analyzed. Consequences for the instrumentation of the measurements and for further investigations are derived.

(1) White Oak Semiconductor, 6000 Technology Blvd, Sandston, VA 23150, USA

(2) Fraunhofer USA, Bradley University, 1501 W. Bradley Avenue, Peoria, IL

61625, USA

(3) Fraunhofer Institute for Materials and Beam Technology, Winterbergstr.

16, 01277 Dresden, Germany

A Comparative Study of Adhesion Test Methods for Hard Coatings Including Aspects of Non-destructive Testing

Directly measuring the adhesion of surface coatings is still a challenge. Surface engineers however, desperately need technical quantities to evaluate the adhesion of their coatings, since this property essentially determines the applicability of their products in industrial practice. Several adhesion testing methods came into use for this reason such as the scratch test, bending test, Rockwell test, cavitation test, and impact test. These techniques gained acceptance, as they are easy to use and have a rather simple technical construction. The contrary results often reported by such testing gave reason to perform a comparison of these techniques by utilizing them to test the adhesion of TiNhard coatings on steel (film thickness 1.2 ...2.4 µm). The films were deposited by ion plating. Adjusting the duration of an in-situ cleaning process (argon ion sputtering) modified the interface conditions for the subsequent film deposition. Measurement results of the various standard imprint tests were investigated with regard to their correlation with the pre-sputtering time. Several test parameters of these methods were used to evaluate the adhesion: the friction work, acoustic emission activity and critical load of the scratch test, the critical strain and the defect density of the four-point bending test, the proportional damage area of cavitation test, the critical number of loading cycles of the impact test, and Young's modulus of the film as measured with the laser-acoustic method. For the majority of these parameters, a significant correlation with the pre-sputtering time was found, but while some of them indicated the expected improvement of the adhesion with increasing pre-sputtering, other parameters suggested an opposite effect. Summarizing these surprising results revealed that the evaluation of the film quality differs for local and global adhesion behavior. An interesting aspect of the non-destructive adhesion testing was the utilization of laser induced surface acoustic waves. The laser-acoustic technique is gaining acceptance as a quick and robust technique for determining the Young's modulus of thin films with thickness down to nano-meters. The results demonstrate the promising possibility of laser acoustic method for fast, non-destructive and inexpensive evaluation of the adhesion behavior.

1150 Hazawa-cho, Kanagawa-ku, Yokohama, 221-0863 Japan

Adhesion and Abrasion of Sputter-deposited Ceramic Thin Films on Glass

Sputtered metal nitride and metal oxide thin films are widely used in coated glass products, e.g., heat reflecting glass, low-E glass. These films show strong adhesion to glass and high hardness, making it difficult to measure the adhesion strength using conventional methods such as the scratch test. We have measured the adhesion strength of these thin films using the direct pull-off test developed for highly adherent thin films and investigated the abrasion behavior of the films.

The adhesion strength of metal nitride thin films, MNx(M: Cr, Ta, Ti, Zr), to glass was found to increase in the order: CrNx<TiNx<ZrNx<TaNx, in proportion to the strength of metal-oxide(M-O) bond. The adhesion strength at the interface of metal nitride upper layer/metal oxide under layer, MNx(M: Cr, Ti)/MOy(M: Al, Sn, Ti, Zr, Ta, Zn), was found to increase as the strength of the M-O bond in the metal oxide became weak. These results implied that the chemical bonds formed at the interface played an important role in the adhesion of the films deposited by the plasma process.

By the Taber abrasion test, metal nitride thin films, MNx(M: Cr, Ta, Ti), were abraded in a peeling mode. The peel rate increased exponentially with the decrease of the adhesion strength and the hardness and with the increase of the coefficient of friction.

*Present address

Electrotechnical Laboratory, 1-1-4 Umezono, Tsukuba, Ibaraki, 305-3586 Japan

phone: 81-298-54-5424, fax: 81-298-54-3363, email: susumu@etl.go.jp

The Dilution Effect of a Titanium Interlayer on the Adhesion of Ceramic Coatings

The present work is developed in order to study the effect of a titanium-based interlayer on the scratch test adhesion of ceramic coatings. Tungsten nitride ceramic coatings were deposited on to high-speed steel substrates after the deposition of a 0.8 m interlayer. The idea is to improve the adhesion by the introduction of an adequate interlayer between the substrate and the coating. This is a common procedure in the field of the adhesion of thin films and coatings.

It is the aim of this study to compare the introduction of pure titanium and titanium alloys interlayes with different grades of dilution such that the titanium activity changed. This choice was made due to the high titanium chemical reactivity and because it may decrease the stresses that might exist between the ceramic coating and the metallic substrate. Based on this, titanium based interlayers, with increasing amounts of aluminium as the element of dilution, were introduced prior the deposition of the ceramic coating.

The results show that for both, without interlayer and with a pure titanium one, the critical adhesive loads are significantly lower than if a titanium-aluminium alloy is used as interlayer. In the later case, the amount of aluminium does not seem to substantially affect the adhesion.

Low Cycle Friction Fatigue Testing under Sub-critical Loads

Micromechanical testing of coated materials is often retained for laboratory assessment of mechanical strength of coating and interface. Quasi-static, "single shot" experiments like indentation and standard scratch testing have a long standing record for their limited efficiency with respect to their potential of realistic evaluation of coating mechanical properties. Indeed . while useful to eliminate bad coatings, the contact loads required to produce macroscopic failure are too high (roughly 1GPa) to correspond to any realistic situation. Then the coating/substrate composite deformation is often controlled by substrate hardness.

In the present paper low cycle friction fatigue operation is adopted to produce interface-specific information yielding quantitative data on adhesion.

In a first order, quasi-empirical approach a common contact pressure of a series of specimens is chosen by adopting the highest load in scratch testing that will not produce any visible surface damage (cracking, chipping and spalling) in standard constant load scratch testing.

The number of cycles to adhesive failure then becomes a quantitative fatigue failure criterion of the interface. "Post mortem" analysis then indeed reveals interface delamination.

In a second approach, optimum contact conditions (indenter tip radius and contact load leading to a von Mises stress maximum centred at the interface) are computed numerically owing to a recently developed boundary element code.

The pertinence of this approach is illustrated on reference specimens (W on Silicon) with well-defined interface quality prepared by ion implantation. Some comparisons are made with results from bulge-blister testing.

1) Materials Science, Uppsala University, SE-751 21 Uppsala, Sweden

2) SAQ Kontroll AB, P.O. Box 49306, SE-100 29 Stockholm, Sweden

Delamination of Thin Hard Coatings Induced by Combined Residual Stress and Topography

Most of todays thin ceramic coatings generated by PVD or CVD techniques are given a residual stress state. This state has a number of origins and will be of either tensile or compressive nature. Irrespective of which, on topographic substrates which deviate from being perfectly smooth, flat and infinitely large, they will induce tensile stresses across the coating substrate interface. These"lift off" stresses, and the risk of delamination they represent, have been investigated using finite element calculations. Surface features such as edges and grooves in the substrate as well as coating edges, materialised e.g. as cracks or pores in the coating, have been studied.

The general experience that highly stressed PVD-coatings have to be sufficiently thin not to spontaneously detach from the substrate has been both experimentally and numerically verified. If the coating thickness is increased, interfacial lift off stresses comparable in magnitude with the residual stress, will develop and may cause coating detachment.

However, it is also shown that the interfacial stress becomes independent of the coating thicknesses once the thickness exceedes about three times the amplitude of the interface roughness. Also, the interfacial stress scales approximately linearly with the maximum inclination of the surface profile.

Various highly stressed ceramic coatings deposited on surface irregularities such as substrate corners and grinding grooves are used to demonstrate the hazardous combination of stress and topography experimentally.

Elastic-plastic Peeling of Thin Films

The peel test configuration is often used to study the failure of thin and flexible laminates bonded to substrates and a lot of investigations have been reported. In these studies, the peeling arm was usually modeled as an inextensible elastic-plastic beam in which the deformation was assumed to be dominated by bending. The decohesion energy (fracture toughness) was deduced from energy balance considerations and taken to be the only parameter governing the fracture process. However, recent studies by the authors have demonstrated that the failure process is controlled by the fracture toughness as well as the crack-tip stresses. Fail to take into account these stresses may lead to unrealistic conclusions. In addition, for thin film peeling, the deformation in the film may not be bending dominant and any analysis based on beam bending theory is not appropriate.

In this study, the embedded-process-zone (EPZ) model is to be used in analyzing the peeling of thin films from a substrate. In this approach, the decohesion process along the interface is characterized by two parameters representing the toughness and the stresses supported by the interface. The crack growth parameters can be obtained by comparing the fracture behavior observed in an experimental test to numerical simulations. These crack-growth parameters are then taken to characterize the decohesion between the thin film and the substrate, and can be subsequently used to predict failure of different peel configuration. It will be demonstrated that this approach allows quantitative predictions for the applied loads and deformations associated with the peeling of thin films.

Influence of the Interface Layers and Microstresses in the Coatings on the Adhesion of Ion-plated Tin Coatings

The interface layers first were prepared on a mild steel(A3 steel) substrate before the deposition of TiN coatings, which were Ni layers coated by electroplating and brush plating respectively, Ni-P layer deposited by chemical plating, nitration layer and boriding layer prepared by ion-nitriding and powder pack cementation process respectively. Then, the inluence of the interface layers on the adhesion of ion-plated TiN coatings have been studied by means of scratch tests. Moreover, the adhesion performance of ion-plated TiN and Ti(Y)N coatings to the substrate have been evaluated by means of constant-rate tensile tests and a scratch tests,and effect of microstresses in the coatings on adherence of TiN and Ti(Y)N coatings have been investigated.The features of the constant-rate tensile test and the scratch test are also discussed in this study. The results show that enhancement of adhesion performance for the ion-plated TiN coatings can be achieved by introducing suitable interface layers between TiN coatings and the substrate.