- This article is filed under:
- » View All
Nanotech Vision: Semiconductor Inspection with a Twist
by Winn Hardin, Contributing Editor - AIA Posted 10/12/2006
The U.S. National Nanotechnology Initiative defines nanotechnology as: ‘‘…the understanding and control of matter at dimensions of roughly 1 to 100 nanometers, where unique phenomena enable novel applications. Encompassing nanoscale science, engineering and technology, nanotechnology involves imaging, measuring, modeling, and manipulating matter at this length scale.’‘
Although the word 'nanotech' has come to the attention of the mass media during the past five years, various technical disciplines have been working with matter on the nanometer scale for decades. Today, the smallest features (also called critical dimensions) of a microprocessor reach down to 60 nm, sometimes 35 nm. This clearly puts semiconductors within the nanotechnology umbrella. Virtually all chemistry involves molecules that fit within this range, and nano-sized fibers and particulates are often used to enhance the material properties of everything from coatings to adhesives, as well as macro materials, such as airplane wings and bicycles.
The point here is that nanotech, while seemingly the stuff of science fiction, is here today, and has been for some time. And just as the U.S. National Nanotechnology Initiative explains, a big part of nanotechnology is the ‘‘…imaging, measuring, modeling…’‘ of the very small. And what does that better than machine vision?
Challenges for Nano-Vision
The wavelength of visible light ranges from approximately 400 nm to 700 nm. Typically, a system cannot 'see' something smaller than the light waves that illuminate it, however, engineers have developed special techniques that allow systems to do just that. Near-field imaging is one, where the objective is placed closer to the surface under inspection than a single wavelength of illuminating light. Confocal microscopy is another method, as illustrated by products such as Siemens (New York) SISCAN 3D Optical Inspection Sensor and Hyphenated Systems (Burlingame, California) HS200 OP (optical profiler).
All confocal microscopes use an aperture to exclude light that originates above or below the object plane, giving them excellent depth resolution, however, the aperture also excludes information from all points in the object plane except the imaged point, thus requiring a scanning approach to image acquisition. Advanced confocal microscopy simultaneously acquires data from multiple apertures in a spinning disk to dramatically reduce image acquisition time. ACM can produce a detailed model with sub-micrometer resolution in all three dimensions in a few seconds. To create the 3D image, the microscope collects 2D images or 'slices' of an object, and then uses computers and software to recreate a 3D image from hundreds of individual slices – similar to a computed tomography system.
According to Terence Lundy, vice president and managing director of Hyphenated Systems, the HS200 can detect contrast differences with steps as small as 5 nm in the Z or vertical axis, which gives a usable measurement capability down to about 20 nm – again in the vertical. Resolution is increased in the X and Y axis.
Using data on the microscope position and settings, Hyphenated Systems uses image processing algorithms to extract 3D surface data from the confocal microscope image. Alignment of individual images can be a challenge for some confocal microscopy systems, but Lundy explains, ‘‘Slice alignment is not the same issue for the HS200 advanced confocal microscope as it is for other scanning confocal microscopes. The HS200 offers real-time 2D confocal slices, registered to the CCD, and does not have the registration difficulties that other confocal technologies have.’‘
Confocal imaging also has the capability to image benefit optically transparent materials. ‘‘We have successfully measured through hundreds of microns of SU8 photoresist or glass, for instance,’‘ Lundy says. ‘‘Naturally, the optical quality of the surfaces and the transparent material will affect measurements made through them. Surface detection is achieved using broad band illumination and detection around 550nm, so the material must show low absorption at this wavelength. We have been able to see multi-walled nanotubes in some circumstances but, in general, our technology is used to address the slightly larger (perhaps >50nm) scales (MEMS, Bio-MEMS/Microfluidics, etc.).’‘
Delving Deep with Electrons
To inspect and characterize objects with nanometer or sub-nanometer resolution, as is the case with much of the carbon single and multiwalled nanotube under commercial development; companies are looking at the once-complex world of electron microscopy. FEI Company (Hillsboro, Oregon) is making these systems more user-friendly to support the growing nanotech industry.
According to Stacey Stone, Technologist at FEI's NanoElectronics Fab Division, FEI has a large portfolio of tools related to focused ion beam and electron beam imaging technologies. In addition to electron microscopes to image and inspect nanotech features and devices, FEI can also incorporate ion beams capable of milling the material as it's inspected to create nanostructures.
‘‘Some of our systems use more then one type of beam in the same chamber, referred to dualbeam systems,’‘ Stone explains. ‘‘Titan is our state-of-the-art transmission electron microscopy (TEM) tool capable of sub-Angstrom resolution. Our automation relies heavily on machine vision techniques such as pattern recognition and sub-pixel edge detection. These machine vision techniques are used to position the focused ion beam and the scanning electron beams.’‘
FEI's electron and ion beam systems generally are used for design debug on finished computer chips and are capable of editing the circuits on the chip. ‘‘[Also] we have customers that use our systems for prototyping MEMS devices. Our large wafer dualbeam systems are used for process control, process monitoring and defect root cause analysis in the semiconductor fab and data storage production environments. The large wafer dualbeam systems are used to automatically mill sections just a few microns across in the samples to reveal subsurface features that can then be SEM (scanning electron microscopy) imaged and analyzed for its dimensions or deformations,’‘ says Stone.
FEI’s systems are used to visualize or create images of features in the tens of nanometers. TEM parallel electron energy spectroscopy (PEELS) technology is used to characterize elemental components of materials and structures, which is important to nanotechnology because of the technologies dependence on basic material science and material properties. PEELS is becoming especially important in analyzing transition regions in materials.
Looking forward, Stone notes that, ‘‘Our customers are driving for more automated and easy to use systems for TEM preparation and imaging. FEI is leading the way with our use of pattern recognition and edge finding algorithms used in conjunction with our leading edge FIB [focused ion beam], SEM [scanning electron microscopy], and TEM [transmission electron microscopy] technology to automate tasks that have historically been done manually. Also, customers have an increased need for higher resolution and material contrast on the 45 nm and 32 nm semiconductor process nodes. This need is being driven by limitations in conventional SEM imaging techniques for contrast and resolution at the nanoscale. Further, the material characterization in the semiconductor industry is starting to be driven by the size of transition regions where concentration changes of specific material components change leading up to an interface.’‘
As you can see by the examples discussed above, at the nanometer scale, the physical properties of materials change, driven by quantum effects rather than bulk properties. Resolving these materials, the transitions between materials, and the unique shapes these materials take – such as the single molecular carbon nanotube – requires the most powerful imaging techniques. As the Nanotech world continues to shrink, industry will continue to push imaging techniques to their limits. And as we've seen in the machine vision industry time and again, directly behind the imager is the computer and the vision algorithms that turn observations into objective data.
There are currently no comments for this article.
Leave a Comment:
All fields are required, but only your name and comment will be visible (email addresses are kept confidential). Comments are moderated and will not appear immediately. Please no link dropping, no keywords or domains as names; do not spam, and please do not advertise.