Focused Ion Beam (FIB) Microscope - Nanomechanical Characterization in the Focused Ion Beam Microscope by Omniprobe

Topics Covered

Background

Omniprobe, Inc. is an industry leader in the manufacturing of accessories enabling "nano-lab" capabilities for electron and ion beam microscopes. Products include innovative nano-manipulation products for electrical/mechanical testing and sample preparation as well as gas injector systems. The original flagship product is the AutoProbe™ 100.1 for TEM sample prep via the "in-situ lift-out" method.

Incorporated in 1999, Omniprobe has leveraged their vision of "time-saving tools designed by expert users" into a successful enterprise that has become the number one volume seller of in-situ chamber-mounted manipulation tools for TEM sample prep. The products service all sectors of the science community, from nanotechnology to biology to semiconductor.

Omniprobe's development team includes several employees that formerly held Senior Member of Technical Staff or higher positions at Texas Instruments in Dallas, as well as diverse education backgrounds ranging from Microbiology to Materials Science, Physics and Electrical Engineering.

Introduction

The availability of the Focused Ion Beam (FIB) microscope with its excellent imaging resolution, depth of focus and ion milling capability have made it an appealing platform for materials characterization at the sub-micron, or “nano” level. The FIB has already become popular for the preparation of samples for TEM from bulk samples. Nano-mechanical characterization in the FIB is another extension of the FIB capabilities into the realm of nano-technology.

Advantage of Materials Characterization in the FIB

A very important advantage of materials characterization in the FIB is the ability to image and even isolate small structures with the ion beam. In this way, the local mechanical properties of thin films or beams can be directly measured. Extrapolating mechanical properties of small structures from bulk property data is not always appropriate. Also, “edge effect” properties of thin films become more important as the surface to volume ratio increases. The ability to measure the mechanical response of tiny structures can be critical for rapid technology development. For example, in the integrated circuit industry, new materials challenges have been posed by the conversion to copper traces, low dielectric constant insulating layers and Pb-free solders. Rapid integration of these new materials requires accurate finite element modeling of material behavior, which depends on accurate thin film mechanical data of unfamiliar materials.

Methods and Analysis

A sensitive strain gauge has been embedded in the probe shaft of an automated in-situ nanomanipulator for the FIB or Scanning Electron Microscope (SEM). The strain gauge used here is a tiny resistive gauge connected to a sensitive Wheatstone bridge amplifier. The strain gauge is positioned to maximize sensitivity to deflection of the probe shaft. The most straightforward application of this system is automated surface contact detection that is appropriate for a fine tip and independent of electrical continuity to the FIB ground. Beyond this purpose, this system can be used for such mechanical tests as bending beam, nano-hardness and fracture limit tests. This system can be used effectively in a qualitative mode which is quite sufficient for many experiments. It can also be used to obtain quantitative data with an NIST-traceable calibrated load cell. In addition, nano-mechanical testing can be combined with simultaneous electrical testing and video recording of the image.

Figure 1. The result of a particle fracture test

Figure 2. Load vs. overtravel graph for a particle fracture test.

Figure 3. The visualization of a scrub mark made with the probe tip point.

An example of the application of nano-mechanical testing to the FIB environment is the fracture testing of tiny particles. This simple test provides important information about the particle’s mechanical properties. Figure 1 shows the result of such a fracture experiment. The ultimate fracture strength of tiny particles can be assessed in a direct method by loading the particles to failure. Figure 2 is the corresponding load vs. overtravel curve for such a measurement.

Figure 4. The contact resistance versus overtravel (µm) diagram for a single scrubbing event as depicted in Fig. 3.

Figure 5. FIB image of a bond pad scrubbing test involving multiple touchdowns. One event ran off the edge of the pad.

Figure 6. Probe load measurements recorded for 10 repeated touchdown events. The 7th touchdown having lower force correlated to scrubbing off the edge of the pad. The first touchdown was held a long period of time prior to retracting.

Overtravel is defined as the displacement of the probe shaft towards the center of the particle after the probe tip first makes contact with the surface of the particle. Because of the size and brittleness of the particle, these data include some compliance of the fine probe tip as well as the elastic response of the particle. The ultimate fracture strength of the particle (roughly 6mg load at fracture) is determined by calibrating the measurement shaft with an NIST-traceable load cell. With an automated testing routine, hundreds of particles can be analyzed within two hours, and the data compiled can be used for comparison of particles from a different production process.

In another example, the process of in-line bond pad probing is simulated with the nanomanipulator to determine the conditions responsible for subsurface cracking of the Cu/low-k dielectric stack beneath the bond pad (Fig. 3).

This is a condition that can lead to an effect called “cratering” in which the bond (wire bond of flip chip bump) breaks away from the die in the package and removes a portion of the dielectric stack. In-line probing typically involves “scrubbing” the pad with the probe tip to break through any surface oxide and give a reproducible contact resistance. In these experiments, the pointed probe tip typically used for testing tiny objects is replaced with a standard probe tip with a flat bottom, which is used for in-line testing. An interesting graph that shows the physical behavior of a probe tip while performing the scrubbing movement is shown in Fig. 4.

Figure 4 shows the effect of surface debris on contact resistance during the scrubbing action. Figure 5 shows the appearance of a bond pad after repeated touchdown and scrubbing events. The probe tip loading information is shown in Fig. 6.

Surface damage and loading information can be combined with subsurface damage revealed by subsequent in-situ liftout and TEM inspection of the probed areas. This in-situ lift-out can be performed with the same nanomanipulator that is used for the probing simulations. In addition, real time video of the touchdown and scrubbing events can be captured simultaneously with the electrical and mechanical data.

Study of MEMS Structures with Nanomechanical Testing

The study of MEMS structures with nanomechanical testing offers a great opportunity for understanding such processes as wear, stiction and elastic behavior of a dynamic compound system. Figure 7 shows the testing of the load required to deflect a Digital Light Processing™ (DLP™ a Texas Instruments Technology) mirror device. Figure 8 shows a typical load vs. displacement curve for this device. This is an excellent example of the combination of the high resolution imaging capabilities of the SEM or FIB with nanomechanical testing performed by sensitive strain gauge attached to a nanomanipulator. Exact placement and simultaneous visual monitoring of the experiment increase efficiency and add a great value to the data. Unlike in previous examples discussed in this study, in this case the loads required for deflection are in the micro-gram range.

Figure 7. Mirror deflection experiment using the Texas Instruments DLP™ chip.

Calibration in this region often requires extrapolation of calibration data, or special calibrated load cells. In the case shown in Figs. 7 and 8, we extrapolated the calibration data from an NIST-traceable load cell that operates in the milligram range. This certainly provides a repeatable qualitative standard with which mirrors of different lots can be accurately compared. More accurate quantitative calibration would require a dedicated calibration source.

Conclusions and Acknowledgments

The examples highlighted here demonstrate the power and flexibility of nanomechanical testing in the FIB or SEM with a probe shaft that includes a built-in strain gauge. Loads that range from grams to micrograms are achievable. Calibration is limited only by the availability of calibrated load cells in the smallest load ranges. Deflections in the range of a few nanometers range can be accurately applied. Simultaneous electri- cal, mechanical and visual data can be combined to provide a revealing study of physical behavior of complex and dynamic nanostructures.

Figure 8. Load in mg vs. deflection in mm for a mirror deflection using the Texas Instruments DLP™ chip.

Source:Omniprobe

For more information on this source, please visit Omniprobe