New analytical approaches
January 5, 2018
Innovative device architectures containing new materials require new solid state analyzes.
In order to follow Moore’s law, the semiconductor industry is pursuing the development of new device architectures using innovative materials. This in turn requires new possibilities for solid state analysis, whether for material characterization or online metrology.
In addition to basic R & D, these capabilities are established at critical points in the semiconductor device manufacturing chain, for example to measure thickness and thin film composition, doping profiles of regions source / drain of the transistor. This approach is used to reduce “time to data”, which basically means that we can not wait until the end of the production line to know if a device is working as expected or not.
Each step of the process is expensive and a fully functional device can take months.
Recent advances in instrumentation and computing power have opened the door to many new and exciting analytical possibilities.
An interesting example concerns the development of coherent sources. Until now, coherent photon sources have been used to study the atomic and electronic structure of materials, but only in large synchrotron radiation devices. However, recent developments have introduced table coherence photon sources that may soon be generating demand in the semiconductor laboratory / environment.
Importantly, the increased processing power that is now within our reach also allows engineers to make the most of these sources and other sources through imaging techniques such as psychography.
Ultrasound allows the processing of complex patterns resulting from a coherent interaction of electrons or photons with a sample in recognizable images at a wavelength resolution close to that of the source without the need for lenses (lenses tend to introduce aberrations). Possible fields of application range from nondestructive imaging of surfaces and background structures to the investigation of chemical reactions in the lower-femtosecond range.
Detector developments also benefit from many of the analytical techniques currently used. A good example is Transmission Electron Microscopy (TEM) and Scanning Electron Microscopy (STEM), which can visualize heavy and light elements at atomic resolution.
By combining increased computing power, imaging approaches such as tomography, holography, psychography, differential phase contrast imaging, and more. can be developed. All this allows TEM / STEM to be much more detailed not only on atoms, for example in 2D materials such as MoS2, but also offers the ability to map electric fields and magnetic domains with unprecedented resolution.
The semiconductor industry is growing very fast. Since the beginning of the 21st century, the industry has produced many disruptive technologies; Technologies needed in an increasingly fragmented application space, not just the central processing unit (CPU).
Other applications ranging from the Internet of Things to autonomous vehicles, to portable human-machine interfaces, are underway, each meeting the requirements and analytical requirements.
“In the long run, the industry will have to think about new methods of characterization / measurement, with one possible scenario, namely MEMS-based analyzers”
– Paul van der Heide
In this changing landscape of semiconductors, we as an industry face the major challenge of developing the right support infrastructure, with the right analytical tools, with the right people in the right place. And while it is true that we can literally analyze everything, the question that must always be asked is “What is important?”
Analyzes are expensive and many of the larger manufacturing companies do not always have time to fully explore all potential areas of interest, and there are many. As a result, the sector is finding greater collaboration, particularly with research institutes such as IMEC.
If we look in ten or fifteen years, we can expect a different landscape. Although I am sure that existing techniques such as TEM / STEM are still widely used – probably more than what we are currently doing (we already see that TEM / STEM will be extended to the workshop).
We will also see developments that move the boundaries of the possible.
This could range from the increased use of hybrid metrology (combining the results of several analytical techniques and different process steps) to the development of new innovative approaches.
This is illustrated by the example of secondary ion mass spectrometry (SIMS) – see picture on the left. With SIMS, a high-energy ion beam is directed to the solid sample of interest, causing the atoms to emerge in the near-surface region. A small percentage of them are ionized and pass through a mass spectrometer that separates the ions according to their mass / charge ratio. In SIMS dynamic mode, a depth profile of the sample composition can be derived.
Today, with this technique, we can not focus the incoming energy ion beam on a limited volume, i. H. At a point close to the size of a transistor. At imec, however, we were able to present innovative concepts that resulted in what is known as SIMS 1.5D and SIMS autofocus (SF-SIMS). These approaches rely on the detection of constituents in repeating array structures that provide averaged and statistically significant information. In this way, the spatial resolution limit of the SIMS has been overcome.
There are also exciting developments at imec in other analytical fields such as atomic probe tomography (APT), photoelectron spectroscopy (PES), Raman spectroscopy, Rutherford backscatter (RBS) and scanning probe microscopy ( SPM).
An important step was the development of FFT-SSRM (Fast Fourier Transform SSRM) technology, which measures carrier distributions in FinFETs with unprecedented sensitivity. This feature has also been translated into a commercial product installed on several imec partner sites.
Above: SIMS uses an energy ion beam directed at a solid sample to obtain a profile of its composition
New analytical approaches?
Like the rest of the electronics industry, the biggest challenges in material characterization and online metrology over the next ten to fifteen years are likely to be lower costs.
This will force us to think about new approaches and methods. Today, we use highly specialized techniques developed on mutually exclusive and expensive platforms. But why not use micro-electro-mechanical systems (MEMS) capable of simultaneously performing parallel analysis and perhaps even in situ?
It is quite possible to envision scenarios in which an army of such units could scan an entire platelet in the fraction of time it needs now or, alternatively, include such units in platelet test structures.
In addition, they could be reusable or disposable, and if that were done, it would change the game.
This is a big challenge, but if it were possible, it would have a significant impact on the semiconductor industry.