Rutherford Backscattering Spectrometry (RBS) is a non-destructive ion beam analysis technique that quantitatively determines the elemental composition, thickness, and depth distribution of thin films without requiring matrix-matched reference standards. This technical article explains the physical principles behind the RBS thin film experiment, the sample types and measured parameters involved, standard testing procedures, and key industrial applications. Engineers and researchers in semiconductor processing, optical coating, and surface engineering will find practical guidance on the capabilities and limitations of RBS for thin film characterization.
Technical Overview
Rutherford Backscattering Spectrometry (RBS) is a quantitative surface and thin film analysis technique that uses a mega-electron-volt ion beam, most commonly 1.5–3.0 MeV helium ions. A collimated beam is directed onto the specimen inside a high-vacuum chamber, and a small fraction of the ions undergoes elastic Coulomb scattering (Rutherford scattering) with atomic nuclei near the surface. The energies of these backscattered ions, recorded by a semiconductor detector at a fixed scattering angle of typically 150°–180°, encode both the mass of the target atoms and their depth below the surface. Because the scattering kinematics and cross sections are known from first principles, RBS delivers absolute, standardless quantification of composition and areal density, making it one of the most trusted reference methods for thin film metrology.
The measurement relies on three interrelated physical phenomena: the kinematic factor, the scattering cross section, and the stopping power of ions traveling through matter.
Kinematic Factor and Elemental Identification
When an incident ion of mass M₁ and energy E₀ scatters elastically from a stationary target nucleus of mass M₂, the backscattered ion retains a fraction K of its initial energy. This kinematic factor depends only on the scattering angle θ and the mass ratio of the collision partners, and for backscattering near 180° it approaches the simple expression K = ((M₂ − M₁)/(M₂ + M₁))². Measuring the energy of the backscattered ions therefore identifies the mass of the scattering atoms: heavy atoms in the film generate high-energy signals, while light atoms generate lower-energy signals. Mass resolution is excellent for light and medium-mass elements but gradually degrades for heavy elements, where spectrum simulation software is often needed to separate overlapping signals.
Scattering Cross Section and Quantitative Composition
The probability of Rutherford scattering scales with the square of the product of the projectile and target atomic numbers (Z₁Z₂)² and inversely with the square of the projectile energy. Consequently, heavy elements scatter ions very efficiently and are detected with high sensitivity, whereas light elements in heavy matrices have relatively poor sensitivity. Because the cross section is calculable analytically, the measured yield of backscattered particles—combined with the integrated beam charge collected by a Faraday cup and the known detector solid angle—yields the areal density of each element in atoms per square centimeter. This is an absolute measurement that does not depend on calibration standards or element-specific sensitivity factors.
Stopping Power and Depth Profiling
Ions continuously lose energy through electronic excitations as they penetrate the solid before and after the scattering event. This energy loss converts the energy axis of the measured spectrum into a depth scale. For a film of known density, the spectral width of the film signal determines its thickness; conversely, a known thickness helps refine the composition. The near-surface depth resolution for 2 MeV helium ions is typically 5–15 nm and degrades gradually with depth as energy straggling accumulates. The practical analysis depth is approximately 1–2 μm for standard helium beams and can be extended to roughly 10 μm using high-energy protons or heavier projectiles.
Instrumentation
A typical RBS thin film experiment employs a megavolt electrostatic accelerator to generate the ion beam, a beamline with collimating apertures and a Faraday cup for charge integration, a precision goniometer for sample positioning, and a silicon surface-barrier or passivated implanted planar silicon (PIPS) detector with an energy resolution of about 10–20 keV. Beam spots of 1–3 mm and integrated charges of 5–20 microcoulombs allow a spectrum to be acquired in 10–30 minutes. Rotatable sample stages support angle-resolved geometries that enhance near-surface depth resolution.
Testing Samples
RBS places relatively few constraints on specimens, but sample form and preparation directly influence data quality. Specimens must be vacuum compatible, reasonably flat, and large enough to cover the beam footprint (a few millimeters). Insulating samples may require charge compensation with an electron flood gun to prevent beam-induced charging that would distort the spectrum. Handling should follow recognized practices such as ASTM E1829 and ISO 18117 to avoid adventitious surface contamination.
Semiconductor Thin Films
Blanket and patterned layers on silicon wafers represent the largest sample group, including gate oxides, high-k dielectrics such as HfO₂ and ZrO₂, metal silicides, diffusion barriers, and epitaxial or ion-implanted layers.
Metallic and Conductive Films
Deposited metal layers and stacks—for example Al, Ti, Ni, Cu, W, Pt, Au, and Ta/TaN barriers—together with alloy films and thin metallizations used in microelectronics and packaging.
Dielectric and Optical Coatings
Transparent oxide and nitride films such as SiO₂, Si₃N₄, TiO₂, Ta₂O₅, and Al₂O₃, as well as transparent conductive oxides such as indium tin oxide used in displays and photovoltaics.
Multilayer and Surface-Modified Structures
Alternating nanometer-scale multilayers (X-ray mirrors, magnetic superlattices, hard coatings), nitrided, carburized, oxidized, or ion-implanted surface layers, and any smooth film-on-substrate system within the technique's depth range.
Testing Items
A standard RBS thin film experiment delivers the following quantitative results:
- Elemental composition: atomic concentrations or stoichiometric ratios of all elements above their detection limits, typically 0.1–1 at% for elements heavier than the substrate.
- Areal density: absolute atoms/cm² values for each layer, obtained directly from scattering yields.
- Film thickness: reported in nanometers, derived from areal density and the assumed or fitted atomic density.
- Depth profiles: element concentration versus depth, revealing gradients, interdiffusion, and compositional non-uniformity.
- Interface analysis: degree of intermixing or reaction-layer formation between adjacent films.
- Heavy-element contamination: detection limits in the range of 10¹²–10¹⁴ atoms/cm² for high-Z impurities on light substrates.
- Crystalline quality assessment (channeling mode): lattice damage profiles, implantation damage, and impurity lattice sites in single crystals.
- Light-element quantification: hydrogen, carbon, nitrogen, and oxygen content measured with complementary ERDA or NRA when required.
Testing Methods
Standard RBS Measurement Procedure
The conventional workflow of an RBS thin film experiment proceeds as follows:
- Mount the specimen on the goniometer and evacuate the chamber to high vacuum, typically better than 10⁻⁶ mbar.
- Direct a 2 MeV helium beam onto the sample and record the integrated charge, typically 5–20 microcoulombs.
- Acquire the backscattered energy spectrum at a fixed detector angle, commonly 165°–170°.
- Construct a layer model (thickness, composition) and iteratively simulate the spectrum with dedicated RBS simulation software until the calculated and measured spectra match.
- Repeat the measurement at tilted incidence geometries when improved near-surface depth resolution is required.
- Report composition, areal density, thickness, and depth profiles with stated uncertainties.
RBS Channeling Mode
For single-crystal samples, the beam is aligned with a major crystallographic axis using the goniometer, which dramatically suppresses scattering from regular lattice atoms. This reduces substrate background, enhances sensitivity to interstitial impurities, and allows the channeling minimum yield to quantify lattice disorder. Channeling RBS is widely used to characterize ion implantation damage and annealing behavior, epitaxial layer quality, and defect depth distributions.
Complementary Ion Beam Techniques
Elastic Recoil Detection Analysis (ERDA) detects atoms knocked forward out of the sample and quantifies hydrogen down to roughly 0.1 at%. Nuclear Reaction Analysis (NRA) uses low-energy nuclear reactions to measure light elements such as nitrogen, oxygen, carbon, fluorine, and boron with high sensitivity and excellent depth resolution. Particle-Induced X-ray Emission (PIXE) adds trace-element capability. These methods share the same accelerator infrastructure as RBS and are frequently combined in a single experimental session.
Data Analysis and Quality Assurance
Spectrum interpretation relies on simulation codes based on the single-scattering approximation, together with evaluated ion stopping power databases. Typical relative uncertainties are 1–5% for composition and 2–5% for areal density, dominated by stopping power accuracy; thickness uncertainty additionally reflects the assumed film density. Terminology follows ISO 18115 for surface chemical analysis, and accredited third-party laboratories commonly operate under ISO/IEC 17025 quality management systems, providing independent verification for industrial customers.
Applications
RBS thin film experiments serve a broad range of research and industrial needs:
- High-k gate stack metrology in advanced CMOS: composition and thickness of HfO₂, ZrO₂, and silicate dielectrics.
- Silicide formation studies: reaction monitoring and endpoint control for NiSi, CoSi₂, and TiSi₂ processes.
- Ion implantation dose verification: absolute implanted fluence checks that validate dose calibration of implanters.
- Barrier layer qualification: thickness and stoichiometry of Ta/TaN and TiN diffusion barriers.
- Optical coating quality control: stoichiometry of oxide and nitride films that directly governs refractive index and durability.
- Transparent conductive oxides: tin content and thickness of ITO correlated with electrical resistivity.
- Magnetic multilayers: layer thickness, interfacial mixing, and abruptness in spin-valve and giant magnetoresistance stacks.
- Surface engineering: depth profiles of nitrided, carburized, and hardened layers in tool steels and engineering components.
- Corrosion and oxidation research: growth kinetics and stoichiometry of thin native and thermal oxide films.
- Metal finishing: thickness and compositional conformity of electroplated and vapor-deposited coatings.
FAQ
What elements can RBS detect?
RBS can, in principle, detect every element from beryllium to uranium. Sensitivity is highest for heavy elements in light matrices, where detection limits reach 0.1 at% or better. Very light elements, particularly hydrogen, cannot be measured by conventional RBS and require complementary ERDA, while nitrogen, oxygen, and carbon are often more accurately quantified by NRA.
Is the RBS experiment destructive?
For most inorganic films, RBS is effectively non-destructive: helium energies used are below nuclear reaction thresholds, so no sample activation occurs, and beam-induced damage is negligible. Some organic and beam-sensitive materials may undergo slight modification, which should be assessed before analysis.
Why does RBS not need reference standards?
The Rutherford scattering cross section is known analytically from fundamental physics, and the detector geometry and integrated beam charge are measured directly. The elemental yields therefore convert into absolute areal densities without calibration curves, which is why RBS is often used to certify reference materials for other techniques.
What depth resolution and analysis depth can be achieved?
Near-surface depth resolution is typically 5–15 nm for 2 MeV helium and can be improved to a few nanometers with grazing incidence or exit geometries. The standard analysis depth is about 1–2 μm for helium beams, extendable to roughly 10 μm with high-energy protons.
How does RBS compare with SIMS, XPS, and AES?
RBS offers absolute, standardless quantification and moderate depth resolution but limited trace sensitivity. SIMS provides ppm to ppb detection limits yet depends on standards and suffers matrix effects. XPS and AES deliver chemical state information and superior lateral or surface sensitivity but probe only the topmost nanometers. In practice these techniques are complementary, and RBS frequently serves as the quantitative benchmark.
What sample sizes and preparation are required?
Typical specimens range from pieces of roughly 5 × 5 mm up to full wafers. Samples should be flat, clean, and vacuum compatible; insulating films may require electron flood charge compensation. No special thinning or cross-sectioning is needed, which preserves the sample for further testing.
How accurate are RBS thickness and composition results?
Composition is typically accurate to within 1–5% relative, and areal density to within 2–5%, limited mainly by stopping power data. Thickness accuracy additionally depends on the assumed film density; when density is unknown, RBS results are most rigorously expressed as areal density in atoms per square centimeter.
