Scratch Test for Film-Substrate Adhesion: Principles, Standards, Methods, and Applications

The scratch test is one of the most widely used quantitative methods for evaluating the adhesion strength between a thin film or coating and its substrate. By drawing a diamond stylus across the coated surface under a progressively increasing normal load, the test determines the critical load at which coating failure begins, delivering objective data for quality control, process optimization, and material qualification. This article explains the technical principles, applicable standards, sample requirements, testing procedures, and typical applications of scratch-based film-substrate adhesion testing.

Technical Overview

The scratch test is a well-established quantitative mechanical method for assessing the adhesion of thin films and hard coatings deposited on rigid substrates. A hard stylus, most commonly a diamond Rockwell C cone with a tip radius of 100 to 200 micrometers, is drawn across the coated surface at constant speed while the applied normal force increases linearly. As the load rises, the stress field beneath and ahead of the stylus intensifies, producing a combination of elastic-plastic deformation, compressive shear stress at the interface, and tensile stress behind the indenter. When the interfacial stress exceeds the practical adhesion strength of the film-substrate system, characteristic failure events such as cracking, buckling, spallation, and delamination occur.

The normal load at which the first well-defined failure event appears is defined as the critical load, which serves as the practical adhesion strength indicator for the coating system. The scratch method is favored in both industry and research because it is fast, reproducible, applicable to very hard coatings that cannot be evaluated by tape or pull-off methods, and directly referenced by international standards such as ASTM C1624, ISO 20502, and EN 1071-3.

Testing Samples

Hard Protective Coatings on Tools

Wear-resistant PVD and CVD coatings such as TiN, TiAlN, AlCrN, CrN, TiC, and diamond-like carbon (DLC) deposited on cemented carbide, high-speed steel, and tool steel substrates.

Semiconductor and Microelectronic Films

Metal interconnect layers, diffusion barriers, dielectric layers, and passivation films on silicon wafers or other rigid substrates, where interfacial delamination directly affects device yield and long-term reliability.

Optical and Functional Coatings

Anti-reflective multilayers, mirror coatings, transparent conductive oxides such as ITO, and glass coatings used in optics, display panels, and photovoltaic modules.

Decorative and Consumer Product Coatings

PVD finishes on watch components, sanitary hardware, eyewear, mobile device housings, and automotive trim parts where appearance retention and wear durability are essential.

Biomedical and Specialized Engineering Coatings

Coatings on orthopedic and dental implants, surgical instruments, and corrosion- or chemical-resistant layers used in demanding engineering environments.

Sample Preparation Requirements

• Flat, rigid specimens with a test area of at least 10 mm by 10 mm and sufficient substrate thickness, generally 1 mm or more, to prevent bending during scratching.
• Low surface roughness, ideally with Ra below approximately 0.1 micrometer, because high roughness masks acoustic emission signals and scatters critical load values.
• Coating thickness typically between about 0.1 micrometer and 30 micrometers for conventional micro-scratch testing; nanometer-scale films are better suited to nano-scratch instrumentation.
• Clean, dry, defect-free surfaces free of grease, dust, and loosely adherent particles.
• At least three to five valid scratch tracks spaced far enough apart, typically 1 mm or several times the scratch width, to avoid mutual interaction.

Testing Items

• Critical load Lc1, corresponding to the onset of cohesive failure within the coating, such as conformal or Hertzian cracks.
• Critical load Lc2, corresponding to the onset of adhesive failure at the film-substrate interface, such as delamination, buckling, spallation, or chipping.
• Critical load Lc3, corresponding to complete penetration of the coating with exposure of the substrate.
• Tangential friction force and friction coefficient as a function of the applied normal load.
• Acoustic emission event count and signal intensity versus load.
• Penetration depth and residual scratch profile where depth-sensing capability is available.
• Post-test failure mode identification and classification by optical microscopy, profilometry, or scanning electron microscopy.

Testing Methods

Progressive Load Scratch Test

A single scratch track, typically 3 to 10 mm long, is produced while the normal load ramps linearly from a low initial value up to a final value selected to guarantee substrate exposure. Final loads of several tens to more than one hundred newtons are common for hard tool coatings, while milli-newton to newton ranges are used for very thin functional films. The progressive mode locates failure thresholds efficiently in a single run and is the default approach of ASTM C1624 and ISO 20502.

Constant Load Scratch Test

Several parallel scratches are made at fixed normal loads bracketing the suspected critical load. Because every point along each track experiences identical loading, this mode verifies thresholds with improved statistical confidence and is useful for very smooth or homogeneous samples and for confirming progressive-load results.

Instrumentation and Signal Acquisition

Modern scratch testers integrate a closed-loop normal load actuator, a tangential force sensor for friction measurement, an acoustic emission transducer positioned near the stylus to capture the energy released by brittle cracking and interfacial detachment, optional depth sensing before, during, and after scratching, and an optical microscope or profilometer for post-test imaging of the scratch track.

Critical Load Determination

The critical loads are determined by combining several complementary criteria: abrupt spikes in the acoustic emission curve, sudden slope changes or discontinuities in the friction force record, abrupt steps in penetration depth, and direct microscopic evidence of specific failure modes along the track. The reported adhesion value is normally the mean critical load of the valid scratches together with its standard deviation, accompanied by images documenting the failure morphology.

Reference Standards

• ASTM C1624, covering adhesion strength and mechanical failure properties of ceramic coatings by quantitative scratch testing.
• ISO 20502, covering the determination of adhesion of ceramic coatings by scratch testing within fine ceramics.
• EN 1071-3, covering the determination of adhesion and other mechanical failure modes by a scratch test for advanced technical ceramic coatings.

Applications

• Coating process development, including optimization of deposition temperature, bias voltage, plasma etching pretreatment, and interlayer design in PVD and CVD production lines.
• Cutting tool qualification, including incoming inspection of coated inserts, drills, and end mills, and correlation of adhesion data with tool life.
• Automotive and aerospace engineering, for wear- and corrosion-resistant coatings on engine components, transmission parts, and structural hardware.
• Consumer electronics and decorative hardware, for scratch durability assessment of PVD finishes on device frames, watches, and sanitary fittings.
• Semiconductor manufacturing and packaging, for adhesion screening of metallization, barrier layers, and encapsulation interfaces.
• Optics and display industries, for durability evaluation of anti-reflective, hydrophobic, and transparent conductive coatings.
• Medical devices, for reliability verification of biocompatible coatings on implants and surgical instruments.
• Research and failure analysis, for comparative benchmarking of novel coating systems and root-cause investigation of delamination field failures.

FAQ

What is the basic principle of the scratch adhesion test?

A diamond stylus is drawn across the coated surface under an increasing normal load. The resulting stresses intensify until the coating fails cohesively or adhesively, and the normal load at which the first well-defined failure event appears is recorded as the critical load, a practical measure of adhesion strength.

How do Lc1, Lc2, and Lc3 differ?

Lc1 marks the first cohesive cracks inside the coating, Lc2 marks the first adhesive delamination or spallation at the interface, and Lc3 corresponds to complete coating removal with substrate exposure. For adhesion assessment, Lc2 is normally the reported value.

Why is acoustic emission monitoring important?

Brittle cracking and interfacial detachment release elastic waves that the acoustic emission sensor detects in real time, allowing failure loads to be pinpointed precisely on the load ramp even when optical contrast between the coating and the failure feature is weak.

How many scratches are required per sample?

International standards generally require at least three to five valid, non-interacting scratches per sample, and the mean critical load with its scatter is reported, because local microstructure and surface roughness cause natural variability in the results.

What are the limitations of the scratch method?

The critical load is a system property that depends not only on interfacial adhesion but also on coating hardness and thickness, substrate hardness, residual stress, friction coefficient, and stylus geometry. Results are therefore best used comparatively under fixed test conditions rather than interpreted as absolute bond energies, and compliant substrates or very soft polymeric coatings may be better evaluated by pull-off or tape methods.

Can this test be outsourced to external laboratories?

Yes. Accredited third-party laboratories routinely perform scratch adhesion testing to ASTM C1624, ISO 20502, and EN 1071-3 for coating suppliers, tool manufacturers, electronics makers, and research organizations, issuing reports that include critical load values, friction data, acoustic emission curves, and failure-mode micrographs.