RBS thin film stress testing combines Rutherford Backscattering Spectrometry with residual stress analysis to deliver accurate thickness, density, and composition data for semiconductor, MEMS, and optical coating applications. This integrated approach enables engineers to quantify thin film stress, optimize deposition processes, and improve device reliability. Learn how RBS ion beam analysis supports advanced thin film stress evaluation in this comprehensive technical guide.
Technical Overview
Rutherford Backscattering Spectrometry (RBS) is a quantitative ion beam analysis technique that uses a collimated beam of MeV-energy helium ions to probe the elemental composition, thickness, and areal density of thin films. When incident ions collide elastically with atoms in the sample, a fraction is backscattered toward a solid-state detector. The measured energy of each backscattered ion carries two key pieces of information: the kinematic factor reveals the mass of the target atom, while the cumulative energy loss of the ion traveling through the film provides depth information. The result is a depth-resolved elemental profile that is inherently quantitative, because the scattering cross sections follow well-established Rutherford physics and do not require matrix-matched standards.
In thin film stress evaluation, RBS plays a critical supporting role. Residual stress is most commonly calculated from substrate curvature using the Stoney equation, which requires accurate knowledge of film thickness and, for stress gradient studies, film density. RBS directly measures areal density in atoms per square centimeter with a typical accuracy of one to three percent. When combined with an independent thickness measurement, this yields the absolute mass density of the film, a parameter that strongly influences intrinsic stress in sputtered, evaporated, and plasma-enhanced deposited layers. Furthermore, RBS in channeling geometry can assess lattice strain, crystalline quality, and defect levels in epitaxial films, all of which are directly linked to stress states in compound semiconductor and epitaxial structures.
Testing Samples
Semiconductor Thin Films
Single-layer and multilayer films deposited on silicon, silicon-on-insulator, gallium arsenide, and other compound semiconductor wafers are routinely analyzed. Typical examples include gate oxides, high-k dielectrics, metal gates, silicides, and epitaxial layers such as SiGe and GaN. RBS quantifies stoichiometry, thickness uniformity, and interface composition with excellent depth resolution.
Metal and Alloy Films
Metallization stacks containing aluminum, copper, titanium, titanium nitride, tantalum, tantalum nitride, tungsten, and nickel are analyzed to determine thickness, composition, and interfacial mixing. These parameters are essential inputs for electromigration studies and residual stress modeling in interconnect structures.
Dielectric and Optical Coatings
Silicon dioxide, silicon nitride, hafnium oxide, aluminum oxide, titanium oxide, and multilayer optical stacks can be characterized for atomic ratio, density, and hydrogen-free compositional analysis. RBS is particularly valuable for verifying the nitrogen-to-oxygen ratio in oxynitride films used in optical and passivation applications.
Multilayer Stacks and Barrier Films
Complex multilayer structures such as diffusion barriers, reflective Bragg stacks, and low-dimensional superlattices are analyzed layer by layer. The technique resolves individual layers with a near-surface depth resolution of approximately five to ten nanometers and a total accessible depth of roughly one to two micrometers for helium beams.
Testing Items
- Elemental composition and atomic concentration profiles for elements from boron to uranium
- Film thickness and areal density measured in atoms per square centimeter
- Mass density determination through combined RBS and profilometry or X-ray reflectometry data
- Stoichiometry verification for compounds such as SiNx, SiOxNy, HfO2, TiN, and TaN
- Interface analysis including intermixing, interdiffusion, and reaction layer thickness
- Heavy element impurity detection at levels down to approximately 0.1 atomic percent in light matrices
- Crystalline quality assessment through channeling minimum yield measurements
- Lattice strain and epitaxial relationship evaluation using angular scan channeling analysis
- Residual stress correlation by combining RBS-derived thickness and density with wafer curvature or X-ray diffraction stress data
Testing Methods
Standard RBS Measurement
A collimated beam of helium ions, typically He+ or He2+ accelerated to 2.0 to 3.05 MeV, is directed onto the sample under high vacuum. Backscattered ions are collected by a silicon surface-barrier detector positioned at a scattering angle of approximately 160 to 170 degrees. The recorded energy spectrum is converted into a depth-concentration profile, with heavy elements appearing at higher backscattered energies and providing the best sensitivity. The measurement is fast, requiring roughly ten to thirty minutes per point, and is essentially non-destructive for most inorganic films.
Ion Channeling Analysis
For crystalline samples, the beam is aligned with a major crystallographic axis such as the silicon 100 or 110 direction. Regular lattice atoms shadow one another, dramatically reducing the backscattering yield from the bulk. The normalized minimum yield quantifies crystalline perfection, while angular scans across the channeling dip reveal strain-induced lattice tilts and mismatches in epitaxial heterostructures. This makes channeling RBS a powerful complement to conventional stress measurement methods.
Data Simulation and Stress Correlation
Acquired spectra are fitted using simulation packages such as SIMNRA or RUMP to extract layer thickness, composition, and interface sharpness. The areal density output is then combined with independently measured geometric thickness to calculate absolute film density. Together with substrate curvature data processed through the Stoney equation, or with X-ray diffraction measurements using the sin-squared-psi method, these RBS results enable complete and traceable residual stress determination for the film under investigation.
Applications
- Advanced gate stack qualification for high-k metal gate CMOS process development
- Residual stress management in MEMS structural layers such as polysilicon, silicon nitride, and electroplated metals
- Silicide formation studies including thickness and stoichiometry of TiSi2, CoSi2, and NiSi contacts
- Diffusion barrier integrity verification for copper interconnect metallization
- Optical coating density and interface control for laser mirrors, filters, and antireflection stacks
- Epitaxial layer characterization and strain analysis for SiGe, GaN, and HEMT device structures
- Thin film photovoltaic compositional optimization for CIGS, CdTe, and tandem cell architectures
- Failure analysis of delamination, blistering, and stress-driven voiding in multilayer packaging
- Research and development support for universities, national laboratories, and semiconductor foundries
FAQ
Can RBS measure thin film stress directly?
RBS does not measure mechanical stress as a primary quantity. Instead, it provides the thickness, areal density, composition, and lattice strain information required for accurate stress calculation. When paired with wafer curvature or X-ray diffraction stress measurements, RBS significantly improves the reliability of the resulting stress values.
What is the detection limit and depth capability of RBS?
For heavy elements in a light matrix, detection limits reach approximately 0.1 atomic percent. Light elements in heavy matrices are less sensitive and may require helium or proton elastic scattering variants. The analyzable depth is roughly one to two micrometers with standard helium beams, with a near-surface depth resolution of five to ten nanometers.
What sample sizes and forms are accepted?
Typical coupons range from small pieces of five by five millimeters up to full wafers of two hundred millimeters or larger, depending on the chamber configuration. The surface should be reasonably smooth and flat for optimal depth resolution, and moderately insulating samples can usually be measured with charge integration corrections.
Is RBS a destructive testing technique?
In most cases RBS is considered non-destructive. The MeV helium ion dose is low enough to avoid measurable damage or heating in metals, semiconductors, and dielectrics. Organic or radiation-sensitive materials may exhibit minor beam damage and are evaluated case by case.
Why is RBS considered a reference-quality metrology method?
Because the scattering cross sections are governed by well-known Coulomb interaction physics, RBS quantification does not rely on calibration standards. This gives RBS a typical absolute accuracy of one to three percent for areal density, making it a benchmark technique for calibrating and validating other thin film measurement tools.
How long does a typical RBS analysis take?
A standard spectrum acquisition requires ten to thirty minutes, while full channeling analysis with angular scans may take one to two hours. Turnaround time including simulation, interpretation, and reporting is typically two to five working days depending on sample complexity.