The rapid, quantitative and non-destructive determination of trace level dopants and elucidation of nanostructures is in high demand in nanotechnology and semiconductor research. Laboratory-based micro-XRF provides unique insight into these structures and facilitates flexible, accurate, quantitative analysis that provide several key advantages over existing techniques.
This demand is driven by large advances in electronics and materials science, which is resulting in increasingly complex systems, such as modern 3D finFET transistors. These transistors have a non-planar structure and tend to use single-digit nanometer high-K dielectric insulators instead of SiO2 gate oxides.
These new 3D systems, with novel compositions, dimensions and geometries, in addition to their increasingly smaller size (e.g. 50 µm pads for transistors), make analysis difficult using conventional analysis methods.1,2
Current Approaches: SIMS and TEM
Secondary Ion Mass Spectrometry (SIMS) has been the conventional choice for microelectronics analysis. SIMS uses a focused ion beam to sputter the surface of a specimen; then, the secondary ions ejected from the specimen surface are analyzed with the aim of determining the specimen’s composition.
However, SIMS encounters problems when working with new devices and materials that can cause variations in sputtering rate which, in turn, results in inaccurate data. These sputtering variations can result from the specimen having impurities (in the case of high-k gate hafnium dielectrics)3 or from the specimen having a non-planar structure.1
In addition, SIMS experiments can take around 30 minutes to run, which can produce a bottleneck in the research process.
Transmission electron microscopy (TEM) is a popular alternative used to avoid these issues. TEM measures the way in which electrons are transmitted through a specimen. Due to this, regions of interest from the specimen must be selectively sliced to form an ultrathin lamella of <100 nm.
The need for extensive sample preparation makes TEM a low-throughput method with a high level of labor required. In addition, preparing the specimen can accidentally remove interesting features from the region to be imaged.
Figure 1. Semiconductor MOSFET designs: flat dielectric layers such as in traditional 2D designs (shown in a) are now moving to complex 3D structures in 3D FinFETs (b) and proposed vertical nanowire designs (c), resulting in new analytical challenges. A Moore and L Shi, “Emerging challenges and materials for thermal management of electronics.” Materials Today 2014
Figure 2. Current approaches to measure thin films are SIMS or TEM sectioning, both which are low-throughput and destructive. Shown above is a TEM image of a 16-nm finFET. D James, “Moore’s Law Continues into the 1x-nm Era.” 21st Itnl Conference on Ion Implanation Technology 2016.
A Novel Approach
Using novel developments in X-ray sources and optics Sigray has developed the AttoMap microXRF, an XRF system with sensitivities below a femtogram. The relative concentration of different components can be accurately determined with no need for standards, and the absolute concentration of high-k dielectric materials can be measured in 2 minutes at a repeatability of 1%.
In addition, XRF is a non-destructive method meaning the specimen remains intact following measurement, allowing the system to be used as an upstream technique and enabling multiple investigations on the same specimen. The AttoMap microXRF’s high spatial resolution of 10 µm means it is perfect for pre-SIMS and TEM analysis, where it can be used to identify regions of interest for further analysis.
The AttoMap was used to analyze specimen of thin films on a silicon substrate, prepared by a third party, as a means of validating its capabilities and determining its lower limit of detection (LLD). The AttoMap’s multi-target X-ray source allowed several targets to be chosen for simultaneous analysis, providing an optimized XRF signal. This technology is unique to the AttoMap and allows lower LLDs to be achieved, as demonstrated in Table 1 for Co.
Table 1. Lower Limits of Detection with 3-sigma Confidence at 400s: LLDs of well below sub-angstrom can be obtained with Sigray’s AttoMap non-destructively. Moreover, as can be seen from the Co thin film rows, the x-ray source target selection can have a remarkable impact on LLD: for example, Cu has a ~10X better LLD than a Mo target. For this reason, AttoMap uses a patented multi-target x-ray source that enables easy instrument optimization for each new investigation.
||Lower Limit of Detection with 99.7% Confidence
||Moly (k-a: 17.4 keV)
||Copper (k-a:8 keV)
||Moly (k-a: 17.4 keV)
In addition, quantitative linearity was achieved (for components such as Co, Ni and HfO) with thin films of different thicknesses (5, 10, 20 Å).
Figure 3. Co Thin Film Spectra: Spectra of 5 Angstrom thick Co (brown peak) peak fitted from background using Sigray’s software.
Figure 4. Co Thickness Linearity: Linearity of 5, 10, and 20 Angstrom Co films showing a r2 linear regression of 0.9999. Counts/s shown are at a “flat” geometry; a 20X increase in counts/s can be achieved at higher angles.
Sigray AttoMap gives researchers an ultra-sensitive and non-destructive means of determining dopant concentrations and thin film thickness. Its novel x-ray optics and high brightness x-ray source allow sensitive experiments to be carried out quickly and, using its multi-target feature, low limits of detection can be achieved for most elements of interest.
The high spatial resolution of the AttoMap means it can be used to quickly carry out upstream analyses, where regions of interest can be identified in a matter of seconds-minutes. Following identification of these areas, downstream methods, such as SIMS and TEM, can be carried out more effectively.
It is possible to use the system for single layer analysis (as explored above) but also for the analysis of multiple elements and multi-layer systems, with the microXRF capable of simultaneous detection of different elements.
References and Further Reading
- AA Budrevich and W Vandervost. “Chapter 5: SIMS Analysis on the Transistor Scale: Probing Composition and Dopants in Nonplanar, Confined 3D Volumes,” Metrology and Diagnostic Techniques for Nanoelectronics. Eds: Z Ma and DG Seiler (2017) Pan Stanford Publishing Pte. Ltd.
- J Bennett, et al. “SIMS depth profiling of advanced gate dielectric materials,” Applied Surface Science 203 (2003).
- T Hasegawa, S Akahori. “High reliable quantification analysis of impurities in high-k gate dielectrics by SIMS,” Special Issue on the Depth Profiling of Ultra Thin Films 28:11 (2007): 638-641.
This information has been sourced, reviewed and adapted from materials provided by Sigray, Inc.
For more information on this source, please visit Sigray, Inc.