Comparison of secondary mode to topographic mode

Comparison of secondary mode (left side) to topographic mode (right side) using Agilent’s Novelx mySEM. The sample consists of a 20 nm thick carbon film on a standard copper TEM grid. The holes are 1.2 µm in diameter on a 2.5 µm pitch. The four quadrant collector plate used to collect the images is also shown. The topographic image is formed by combing the four segments as (A+D)-(B+C) or (A+B)-(C+D). Images: Agilent

The playing field for surface analyzing microscopes is hardly level as continual advances keep options interesting for users—and competitors on their toes.

The old saying goes: don’t judge a book by its cover. But for one multi-billion-dollar industry, and the thousands of researchers and technologists it represents, the “cover” is the thing that matters most. Surface characterization, where near-atomic scale lumps and bumps fascinate researchers and stump developers, doesn’t concern itself with the mysteries of bulk properties.

Surfaces are the interfaces for interaction. We depend on these properties in metals, ceramics, organic materials, and semiconductors to make many of the products we rely on work. Analytical results through the use of microscopy helps perfect these indispensible items, from automobile engines to computer chips.

It’s not surprising then, that the drive to create better microscopes has given the modern lab a dizzying array of product choices for the mission of surface characterization.

According to Jeff Rosner, chief technologist, Materials Science Solutions, Agilent Technologies, Santa Clara, Calif., what microscopy companies are offering today is a broad array of functionality for analyzing the surface of materials.

“That driving force to get data ever more specific and quantitative at very near-to-surface level as opposed to bulk-like properties—this plays out in electron microscopy,” says Rosner, referring to one of the highest-resolution methods for examining materials. He says that although we think of it as a surface, the very near-to-surface activity is more complex than a lone interface of air and matter.

“At the typical 1 keV and above (beam power in an electron microscope), we are exciting multiple electrons, and rely on sorting out the detected signals to retrieve only those signals that arise from near the surface,” he says.

But gathering those signals together is a difficult task, complicated by the fact that as the voltage increases, so does the noise. As a result, Rosner says, companies are building better detectors to improve results at lower voltages. This produces less waste, in both beam energy, and efforts to compensate for poor signal.

This improvement in electron microscopy comes just as other companies are improving their own products. Anasys Instruments, Santa Barbara, Calif., for example, recently introduced an atomic force microscope—a common surface analyzer—with an integrated infrared spectrometer to gather both physical and chemical data. On the optical side, Olympus America, Center Valley, Pa., has placed in a number of labs a relatively new laser confocal microscope that combines ease-of-use with the power of differential interference contrast imaging.

Hillocks, terraces and defects

Hillocks, terraces and defects on 6H-SiC sample (left), observed using Agilent’s Novelx mySEM. The fine structure (right) on the terraces was determined to be sub-nanometer atomic steps. Images: Maboudian Lab, COINS, Univ. of California, Berkeley

Choosing the right system isn’t easy, but it helps to take a look at the competition.

AFM learns an unusual pitch
Atomic force microscopy is a near ubiquitous instrument in materials labs. Both highly accurate and versatile, they can be used from wet spectroscopy all the way to materials analysis. Like profilometers, they are non-optical and operate by scanning a probe tip on a cantilever across a surface. They are especially useful as quantitative mappers of surface materials.

In 2008, Agilent won an R&D 100 Award for its Scanning Microwave Mode AFM, which capitalized on the potential to gather more than just deflection information from the tip of the cantilever. According to Rosner, the microscope generated a lot of excitement, both from users and from the development team at Agilent.

“Near-field optical measurement has been a holy grail because it would allow us to do optical microscopy on very small volumes on a surface,” says Rosner. Unfortunately, the extremely low throughput of photons means it is impossible to generate a lot of signal.

However, he says, Agilent found that they could exploit other radiation, namely microwaves. These are channeled in materials much better than visible photons or ultraviolet light, and vibrational spectroscopy functions there as well, from the 1 to 20 GHz frequency range. Technologists build an accessory for an AFM that would collect the microwave signal. But until they built it, they had no idea if the information would prove useful.

“From day one it was a solution looking for a problem. It solves the physics for delivering an interesting portion of the electromagnetic spectrum to the sample and collecting information,” Rosner says.

The instrument is still in the early stages of development (technologists are working on a broader frequency range), but the potential is there for major applications in dopant distribution in semiconductors, coatings analysis, fuel cell membranes, and more.

Other vendors have expanded the abilities of their AFMs. In February 2010, Anasys introduced its nanoIR platform, a powerful new measurement tool that reveals the chemical composition of samples at the nanoscale.

The system uses the tip of an AFM probe to measure infrared (IR) absorption. The sample is illuminated by a tunable IR source. When IR radiation is absorbed by a region of the sample, the region heats up. The heat generates a rapid thermal expansion pulse that can be detected by the AFM cantilever tip.

This technique beats the “far field” optical diffraction limit, according to Craig Prater, Anasys chief technology officer, because the absorbed radiation is measured by the tip in the extreme near field. The result is higher resolution than even Fourier Transform IR spectroscopy can attain.

But for all the capabilities of AFM, users will require visual information. And their options typically boil down to a choice: electron, or optical. Both types have seen striking improvements even as the fundamental principles behind each technique remain fundamentally unchanged.

Materials under the gun
The core principle behind scanning electron microscopy (SEM) is secondary electron emission. Electron beam scans an object causing it to emit secondary electrons, which form a pattern that produces a 3-D image. Other common signals created by beam emissions include back-scattered electrons, x-rays, light, specimen current, and transmitted electrons (in transmission electron microscopy).

Electron microscopy is often an ideal for surface characterization for a number of reasons. First, the signals result from interactions between the electron beam and atoms at or near the surface of the sample. At beam strengths of around 1 keV and up, this produces resolutions from about 5 nm down to below 1 nm. In addition, large depth-of-field resulting from the narrow electron beam yields rich 3-D information.

 Optical light microscopes can’t match the power of an electron gun for resolution, but the price for the performance is, of course, the limitations that electron microscopy imposes: higher cost, more stringent environmental, use of lab space, and a higher learning curve for operation. There is also the question of sample preparation, which is a mixed bag. Metals are typically ideal for electron microscopy because they are electrically conductive (interacts well with the beam) and electrically grounded (prevents electrostatic buildup). But other materials, particularly anything organic, must be specially prepared. Also, sample size is generally quite limited.

Earlier this year, Agilent—an experienced manufacturer of AFMs—purchased Novelx, a scanning electron microscopy manufacturing. Novelx’s flagship product, mySEM, was a 2009 R&D 100 Awards winner that effectively deals with some of the drawbacks of SEMs mentioned previously. It is a compact field emission SEM that offers capabilities previously only available in more expensive and much larger conventional field emission SEMs. According to Jim Rynne, vice president of marketing at Novelx, Lafayette, Calif., the mySEM provides sub-10 nm resolution and several low-voltage imaging techniques that reveal surface information.

Surface features at the nanoscale features can be enhanced using a variety of techniques, says Rynne. Lowering the primary beam voltage to < 1 keV decreases the penetration depth of the beam (typically 30-40 nm) and accentuates surface features that may otherwise be masked at higher voltages. Operating the electron detector in a differential or “topographic” mode, where the common background signal is subtracted from the image, also improves surface contrast. With crystalline materials, using the relatively recent technique of electron channeling contrast imaging (ECCI), vertical resolution can be further improved resulting in the ability to image individual dislocations, atomic steps, and other defects on or near the surface.

For example, says Rynne, a mySEM operating at low-voltage (0.5keV–1.2keV) in topographic mode can resolve sub-nm steps on the surface of a 6H-SiC polished wafer. Verification of step height is then accomplished with AFM. The ability to collect such high resolution information is the job of the mySEM’s quad-segmented microchannel plate detector, located directly below the objective lens. This is a picky instrument, only accepting secondary electrons depending on the bias of the front plate.

Helios NanoLab 650

The Helios NanoLab 650 (left) integrates FEI’s extreme high-resolution scanning electron microscope technology with a new, high-performance focused ion beam for nanoprototyping in academic and industrial research labs. The 450S provides imaging and milling capability for applications in semiconductor labs.

“Typical electron microsopes have an Eberhard-Thornly detector, a large bulky detector that extracts secondary electrons and gives you a traditional shadowed view,” says Rosner, who has spoken with numerous users of the mySEM system. The new detector, he says, gives four different viewing angles.

The further step of operating the detector in differential mode, says, Rynne, produces an effect similar to positioning the optical in a light microscope at angle incident to the light source.

The detection of steps, hillocks, voids, and other surface features is an important task for semiconductor related applications, but typically these require dedicated microscopes. The small size and general-purpose usability of the mySEM—essentially a desktop SEM—belies the strong analytical results available from the system, says Rynne.

“It is quite remarkable for such a compact system that can be made conveniently available to researchers,” he says.

FEI Company, Hillsboro, Ore., recently upgraded its offerings with the introduction of Helios NanoLab x50 DualBeam Series in April 2010. The company says this most powerful and versatile DualBeam system available on the market today. It integrates FEI’s extreme high-resolution scanning electron microscope (XHR SEM) with a new, high-performance focused ion beam (FIB), to deliver a better level of imaging and milling capability for leading-edge applications in semiconductor and materials science research and development.

The company commercialized its first the DualBeam microscope more than 10 years ago, and the XHR SEM platform has been successful in the marketplace, particularly in Magellan form. The Helios NanoLab x50 differs from previous Helios microscopes primarily by the inclusion of a new high-performance Tomahawk focused ion beam, originally introduced in the specialized V400ACE microscope and now equipped with fast-switching technology. The new FIB provides SEM and FIB live monitoring of milling operations, a smaller FIB spot for more precise milling control, and higher beam currents for faster material removal on large structures, such as through-silicon vias (TSVs).

“At some level, this instrument is an evolution. We were going for more nodes, more critical areas in analysis,” says Richard Young, technology manager at FEI.

This applies particular to transmission electron microscopy, which communicates rich chemical and crystal structure information from reading electrons passing through a sample. But this sample preparation is difficult. With the Helios, Young reports, overall throughput of advanced TEM lamella preparation has been improved by 40%.

“If you can hold off some of those samples from going to TEM, there is a huge benefit,” says Young. The instrument is also geared for labs that support fabrication of tough 3-D packaging for chips, such as through-silicon vias. This industry was among the first to adopt the DualBeam, but FEI reports that more than 30 Helios units have sold to pure materials labs.

As a result, FEI has split its DualBeam offerings in two, gearing the 450(S) for semiconductor labs, and the 650 for academic and industrial research centers involved in nanoprototyping. The sub-nanometer resolution of the Helios 650 at extremely low beam energies provides surface-specific imaging that, until now, was unavailable in a dual beam instrument.

“This is the first DualBeam able to resolve below 1 nm at 1keV,” says Young. This distinction is important, he says, because low beam energies produce less noise that must be filtered.

“The focused ion beam is very good with low landing energies,” says Todd Templeton, products marketing manager for FEI’s Electronics Division. The Tomahawk “makes use of a new solid-state directional backscatter detector. It is a very sensitive detector with regard to bias. And it’s a good detector for getting rid of charges and good signal-to-noise ratio.”

Light steps up to the plate
In comparison to electron microscopy, light microscopy doesn’t stand a chance then? Not so. More than a year ago, Olympus launched a new microscope, the LEXT OLS4000. A confocal microscope, the new system was designed specifically to match the company’s optical expertise with the quality and accuracy requirements of the semiconductor industry.

While the intended market for the new LEXT microscope didn’t materialize as Olympus had hoped, a pleasant surprise to the company, says David Rideout, product manager for micro-imaging at Olympus, was how many other microscopists were eager to put the LEXT’s performance to the test. This was particularly true, he says, of materials analysis.

“We’ve been speaking regularly with Chris Brown, a professor at Worcester Polytechnic Institute, Worcester, Mass. He says he’s been getting some fantastic results using this instrument in his lab for lots of different materials analysis,” says Rideout.

Brown’s research relies on innovative techniques for examining in extreme detail the geometry of the surfaces of various materials to judge wear, and what the surface roughness reveals about the material’s behavior. Current projects at his lab include the measurement of paper, granite, and grinding wheels.

Brown is also a member of the ASME B-46 committee on surface textures and helps develop national and international standards for surface metrology. He is particularly interested in the determination of uncertainty, and noise control and management in surface measurements. His work with the LEXT may help shape these standards, in part because the system was designed specifically to handle the management of uncertainty. According to Rideout, the LEXT is the first laser microscope to provide both accuracy and repeatability guarantees:

  • Z Repeatability 50x: sn-1=0.012 µm
  • XY Repeatability 100x: 3sn-1=0.02 µm
  • Z Accuracy 0.2+L/100 µm or Less (L=Measuring Length µm)
  • XY Accuracy Measurement Value ±2%

OLS4000 is currently benefiting his research and may eventually hone these standards.

The LEXT uses a number of existing optical systems already developed at Olympus, but it is in the combination of several advanced technologies that the microscope finds its advantage.

“The key is the 405-nm laser, the slope to 85 degrees, and the traceability,” says Rideout. “With this instrument you get the versatility and everything a regular microscope gives you, but the LEXT also provides accuracy and QA/QC. Those are really the sweet spots for this microscope.”


The LEXT OLS4000 from Olympus America employs a newly-developed dual confocal system that can capture a clear image from a specimen consisting of materials with extremely varied reflectivity levels. The 405-nm laser diode is combined with specially designed optics to provide optimized image quality and minimized aberrations, particularly when used as a surface roughness measuring tool.

The LEXT represents Olympus’ relatively recent foray into a dual beam system. A conventional confocal microscope works with the behavior of interfered light to produce highly detailed imagery. The intereference effect is supplied by a pinhole; in a dual beam system such as the LEXT, there are two pinholes. The system is able to get Z data off of the smaller pinhole, resulting in images that reveal high detail at extreme angles.

Another advantage is the field of view. With profilometers, a single drag across a surface can yield many millimeters worth of solid data. With microscopes, whether electron or optical, the field of view is much narrower. However, image stitching capability of up to 500 images on the LEXT allows for up to 100-mm resolved on a single line using the new roughness-specific mode.

This system is calibrated in the same way contact surface roughness gauges are calibrated, and has the necessary roughness parameters and filters required per ISO and JQA. This allows users with contact surface roughness gauges to obtain output results from the system consistent with their existing instruments, with the advantage of greater speed and non-contact measurement.

Another important component is the laser diode, which Rideout points out is largely responsible for providing Olympus with what he says the world-leading spatial resolution. The system resolves patterns clearly to 0.12-µm line-width.

The 405-nm laser is combined with specially designed optics to provide optimized image quality and minimized aberrations. The resolving power is enhanced by confocal optics featuring an optimized circular pinhole and a high speed XY scanner developed from Olympus MEMS technology. The dual beam architecture enhances Z-resolution: 0.8 nm scale resolution and 1.0 nm display resolution.

The instrument also features an auto-thresholding function for laser intensity. The system will automatically examine the sample and determine the optimum laser and white light intensity for the sample. This provides optimal imaging and the elimination of operator error.

These characteristics led Olympus to believe it would see most of its customers from the semiconductor sector. But the appeal has been much broader as a whole.

“Despite the downturn, we had a very good year for this instrument,” says Rideout, who reports the microscope has been used to study aggregate pavement for roads, MEMS devices, coatings and finishing, corrosion, silicon VIAs, polymers, microbatteries, and solar inspection.

“The long-term goal is to have it more automated, to make it even easier to use. We plan to improve it so that you can really just push a button and walk away,” says Rideout.

As with so many other microscopy products, software technology, Rideout says, is also becoming increasingly important to the value of the microscope, particularly as it is able to allow users to better utilize the capabilities of the product.

For example, the microscope has the ability to image samples in 3-D and “true color” by combining the laser image with the full color brightfield image using the systems software capabilities. In addition, differential interference contrast imaging (DIC) can also be combined with the confocal image to provide identification of minute level differences on the surface of a specimen.

These features are all useful, but Rideout stresses an important factor in the success of the instrument: it is approachable. The interface and software are there to help the user get the most of out of the machine, and learn more from surface analysis.

Published in R & D magazine: Vol. 52, No. 3, June, 2010, pp. 24-27.