RBS Semiconductor Thin Film Testing: Principles, Capabilities, and Applications

RBS semiconductor thin film testing (Rutherford Backscattering Spectrometry) is a powerful, non-destructive ion beam analysis technique used to determine the elemental composition, layer thickness, and depth distribution of thin films in semiconductor devices. As chip architectures shrink and film stacks become increasingly complex, RBS provides the absolute, standards-free quantification that process engineers and materials scientists need for reliable quality control and failure analysis. This article explains how RBS works, what samples and properties it can measure, its analytical methods, and its most important applications in the semiconductor industry.

Technical Overview

Rutherford Backscattering Spectrometry (RBS) is a quantitative surface and thin film analysis technique in which a collimated beam of light ions—most commonly helium ions (He+ or He2+) accelerated to energies of 1 to 3 MeV—is directed onto a sample. A small fraction of these ions undergo elastic Coulomb collisions (Rutherford scattering) with the nuclei of atoms in the specimen and are backscattered toward a detector, typically positioned at a scattering angle of 160 to 170 degrees. The measured energy spectrum of the backscattered ions encodes two fundamental pieces of information about the sample.

First, the kinematic factor, which describes the fraction of energy retained by an ion after colliding with a target nucleus, depends only on the masses of the ion and the target atom. Heavier target elements therefore produce backscattered ions of higher energy, allowing elemental identification. Second, as the incident ions travel into and the scattered ions travel out of the material, they lose energy continuously through electronic stopping. This energy loss converts the depth at which a scattering event occurs into an energy shift in the spectrum, giving RBS its intrinsic depth profiling capability.

Because the scattering cross sections for these collisions are well known from fundamental physics, RBS is one of the few analytical techniques that delivers absolute quantitative results—in areal density units of atoms per square centimeter—without requiring reference standards. Typical detection limits range from approximately 0.1 to 1 atomic percent for elements heavier than the substrate (for example, heavy metals on silicon), while depth resolution at the surface is on the order of 5 to 20 nanometers depending on geometry and detector resolution. The probing depth for standard 2 MeV helium beams extends roughly 1 to 2 micrometers into the sample.

For semiconductor manufacturing, these characteristics make RBS a reference method for verifying film stoichiometry, thickness, and interface quality, and it is frequently used to calibrate or validate other techniques such as ellipsometry, X-ray reflectometry, and SIMS.

Testing Samples

RBS is well suited to a wide variety of semiconductor materials and layered structures. Samples must be reasonably flat, smooth, and stable under vacuum and ion bombardment. Typical specimen sizes range from small pieces of a few millimeters square up to wafer sections or full wafers, depending on the chamber design of the instrument.

Gate Dielectrics and High-k Stacks

High-k dielectric films such as hafnium oxide (HfO2), zirconium oxide (ZrO2), hafnium silicate, and aluminum oxide (Al2O3), along with interfacial SiO2 layers, metal gates, and capping layers in advanced CMOS gate stacks, are routinely characterized for composition and thickness.

Metal Layers, Barriers, and Interconnects

Thin films of copper, aluminum, tungsten, titanium, tantalum, platinum, and their nitrides (TiN, TaN) used in metallization schemes and diffusion barriers can be measured for thickness, purity, and interfacial reaction.

Silicides and Contact Structures

Nickel silicide (NiSi), cobalt silicide (CoSi2), titanium silicide (TiSi2), and platinum silicide layers formed by rapid thermal processing are analyzed to determine stoichiometry and the degree of silicon consumption.

Compound and Wide-Bandgap Semiconductors

Epitaxial films and heterostructures based on GaAs, InP, GaN, AlGaN/GaN, SiGe, and SiC are tested for elemental ratios, layer thicknesses, and compositional grading.

Photovoltaic, Optical, and Functional Films

Transparent conductive oxides such as ITO, thin-film photovoltaic absorbers including CIGS and CdTe, optical coatings, magnetic multilayers, and hard protective coatings are also common RBS samples.

Testing Items

RBS semiconductor thin film testing covers a comprehensive set of measurable properties:

• Areal density and film thickness: determination of atoms per square centimeter, convertible to physical thickness when film density is known.
• Elemental composition and stoichiometry: absolute quantification of atomic ratios in alloys, oxides, nitrides, and compound films.
• Depth profiles: concentration-versus-depth distributions of elements within the top micrometer of the sample.
• Multilayer analysis: individual layer thicknesses and compositions in complex stacks, superlattices, and gate structures.
• Interface characterization: detection of intermixing, interdiffusion, and reaction layers between films and substrates.
• Heavy element contamination: identification and quantification of trace high-Z impurities in low-Z matrices.
• Ion implantation dose verification: measurement of total implanted fluence and integrated dose for process calibration.
• Crystalline quality assessment: evaluation of lattice disorder, defect density, impurity lattice location, and epitaxial layer quality using Rutherford backscattering channeling (RBS/C).
• Film density determination: combined thickness measurement by microscopy or profilometry plus RBS areal density yields film density.

Testing Methods

Standard RBS analysis is performed in a high-vacuum ion beam line. A tandem or Van de Graaff accelerator generates a monoenergetic helium beam, typically 2.0 MeV He2+, which is collimated and raster-scanned over a spot of roughly 1 to 2 millimeters on the sample. Backscattered ions are collected by a silicon surface barrier or PIPS (passivated implanted planar silicon) detector, and the resulting energy spectrum is recorded with multichannel pulse-height analysis.

The analytical workflow generally follows these steps:

• Sample mounting and alignment: the specimen is mounted on a goniometer stage, aligned in the scattering chamber, and pumped to high vacuum.
• Spectrum acquisition: backscattering data are collected at a fixed geometry, commonly a 165 to 170 degree scattering angle, with a beam dose measured by a Faraday cup for absolute normalization.
• Spectrum simulation and fitting: the measured spectrum is modeled with dedicated software such as SIMNRA or RUMP, which applies the kinematic factor, Rutherford scattering cross sections, and stopping powers to extract composition and depth profiles.
• Channeling measurements (optional): for single-crystal samples, the goniometer is oriented to align a major crystallographic axis with the beam, suppressing yields from the bulk lattice and highlighting displaced atoms or defects.
• Reporting: results are reported as areal densities, thicknesses, atomic concentrations, and depth profiles with associated uncertainties.

In practice, RBS is often combined with complementary ion beam techniques to extend its capabilities. Elastic Recoil Detection Analysis (ERDA or ERD) detects hydrogen and other light recoiled atoms that standard RBS cannot easily quantify. Nuclear Reaction Analysis (NRA) improves sensitivity for light elements such as nitrogen, oxygen, boron, and fluorine. Particle Induced X-ray Emission (PIXE) adds trace-level sensitivity across the periodic table. These combined methods provide a complete elemental picture of complex semiconductor film stacks.

Applications

RBS semiconductor thin film testing supports critical activities across the semiconductor value chain:

• Advanced gate stack development: qualifying high-k dielectric deposition processes, verifying Hf:O ratios, and monitoring interfacial layer growth in high-k metal gate (HKMG) technology.
• Metallization and barrier engineering: confirming barrier layer thickness and integrity, evaluating seed layer coverage, and studying copper or aluminum interaction with underlying films.
• Silicidation process control: optimizing rapid thermal anneal conditions by measuring silicide phase stoichiometry and thickness uniformity.
• Ion implantation metrology: calibrating implanter dose and validating simulated implantation profiles, including channeling tails in crystalline silicon.
• Compound semiconductor growth: verifying stoichiometry and interface abruptness in MOCVD and MBE grown III-V and III-nitride heterostructures for RF, optoelectronic, and power devices.
• Wide-bandgap power electronics: characterizing SiC and GaN epitaxial layers, dopant profiles, and dielectric passivation films.
• Photovoltaics and displays: analyzing absorber layers, transparent conductive oxides, and anti-reflection coatings for composition and thickness control.
• Reference metrology: serving as a NIST-traceable, standards-free reference technique for calibrating thickness and composition tools used in high-volume fabrication.
• Research and failure analysis: supporting academic materials research, new materials screening, and root-cause investigation of process excursions and device failures.

FAQ

Is RBS a destructive testing technique?

No. Under standard conditions, RBS is essentially non-destructive. The helium ion dose is low enough that it does not measurably sputter or alter most semiconductor materials, and samples can generally be returned for further testing or reuse.

What is the detection limit of RBS?

For elements heavier than the substrate matrix, detection limits are typically around 0.1 to 1 atomic percent, and can reach the parts-per-million range for very heavy elements such as gold or uranium on silicon. Sensitivity for elements lighter than the matrix is much poorer, which is why ERDA or NRA is used for hydrogen, nitrogen, and oxygen quantification when high sensitivity is required.

What depth can RBS analyze and at what resolution?

A 2 MeV helium beam probes roughly the top 1 to 2 micrometers of a sample. Near-surface depth resolution is approximately 5 to 20 nanometers and degrades gradually with depth due to energy straggling. Grazing incidence or grazing exit geometries can improve resolution for ultrathin films at the cost of probing depth.

How does RBS compare with SIMS, XPS, EDS, and XRR?

Unlike SIMS, RBS is quantitatively accurate without standards but has poorer trace sensitivity. Compared with XPS, RBS probes deeper and gives absolute concentrations rather than relative surface compositions. EDS in an SEM has much lower depth control and quantification accuracy for thin films. Versus X-ray reflectometry, RBS provides elemental identification in addition to thickness and density information. In modern workflows, these techniques are complementary, with RBS often serving as the quantitative benchmark.

Can RBS measure hydrogen in films?

Standard RBS cannot quantify hydrogen because helium backscattering from hydrogen is kinematically forbidden in backscattering geometry. Hydrogen is instead measured using Elastic Recoil Detection Analysis (ERDA) performed on the same type of ion beam platform.

What sample preparation is required?

Preparation is minimal. Samples should be clean, flat, and vacuum compatible. Contaminants such as organic residues or thick native oxides may be noted in the spectrum but usually do not prevent analysis. Full wafers, wafer fragments, and mounted cross-section pieces can all be accommodated depending on the instrument chamber.