Device failure analysis by scanning electron microscopy
Scanning electron microscopy for semiconductor device failure analysis
SEARCH · Search NASA
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Scanning electron microscopy for semiconductor device failure analysis
Scanning tunneling microscopy (STM) methods are used to directly control the barrier height of a metal tunnel tip-semiconductor tunnel junction. Barrier behavior is measured by tunnel current-voltage spectroscopy and compared to theory. A unique surface preparation method is used to prepare a low surface state density Si surface. Control of band bending with this method enables STM investigation of semiconductor subsurface properties.
Scanning probe microscopy (SPM) is an important tool for performing measurements at the nanoscale in imaging bacteria or proteins in biology, as well as in the electronics industry. An essential element of SPM is a sharp, stable tip that possesses a small radius of curvature to enhance spatial resolution. Existing techniques for forming such tips are not ideal. High-aspect-ratio, monolithically integrated, as-grown carbon nanofibers (CNFs) have been formed that show promise for SPM applications by overcoming the limitations present in wet chemical and separate substrate etching processes.
Scanning laser acoustic microscopy was used to nondestructively evaluate solar cell interconnect bonds made by resistance welding. Both copper-silver and silver-silver welds were analyzed. The bonds were produced either by a conventional parallel-gap welding technique using rectangular electrodes or new annular gap design with a circular electrode cross section. With the scanning laser acoustic microscope, it was possible to produce a real time television image which reveales the weld configuration as it relates to electrode geometry. The effect of electrode misalinement with the surface of the cell was also determined. A preliminary metallographic analysis was performed on selected welds to establish the relationship between actual size and shape of the weld area and the information available from acoustic micrographs.
A three-phase microcircuit failure analysis procedure is presented. In the first phase, the device is nondestructively tested and the data obtained is correlated with the circumstances surrounding the failure event. The device is then subjected to a limited nondestructive analysis dominated by optical and scanning electron microscopy (SEM). Knowing that microcircuits usually fail from manufacturing defects or misuse, the analyst looks for one of the probable mechanisms. The SEM has two special operating modes that provide unique data for nondestructive analysis: voltage contrast and conductive current. These are used with other SEM modes and surface analysis techniques to analyze complicated microcircuits. The last phase includes the destructive procedures. The end product of these analyses is a corrective action that will yield a highly reliable system.
The Stardust spacecraft returned the first undoubted samples of cometary dust, with many grains embedded in the silica aerogel collector . Although many tracks contain one or more large terminal particles of a wide range of mineral compositions , there is also abundant material along the track walls. To help interpret the full particle size, structure and mass, both experimental simulation of impact by shots and numerical modeling of the impact process have been attempted. However, all approaches require accurate and precise measurement of impact track size parameters such as length, width and volume of specific portions. To make such measurements is not easy, especially if extensive aerogel fracturing and discoloration has occurred. In this paper we describe the application and limitations of laser confocal imagery for determination of aerogel track parameters, and for the location of particle remains.
Scanning electron mirror microscope advantages over electron mirror and scanning electron microscopes for examining integrated circuits
Scanning tunneling optical resonance microscopy (STORM) is a method, now undergoing development, for measuring optoelectronic properties of materials and devices on the nanoscale by means of a combination of (1) traditional scanning tunneling microscopy (STM) with (2) tunable laser spectroscopy. In STORM, an STM tip probing a semiconductor is illuminated with modulated light at a wavelength in the visible-to-near-infrared range and the resulting photoenhancement of the tunneling current is measured as a function of the illuminating wavelength. The photoenhancement of tunneling current occurs when the laser photon energy is sufficient to excite charge carriers into the conduction band of the semiconductor. Figure 1 schematically depicts a proposed STORM apparatus. The light for illuminating the semiconductor specimen at the STM would be generated by a ring laser that would be tunable across the wavelength range of interest. The laser beam would be chopped by an achromatic liquid-crystal modulator. A polarization-maintaining optical fiber would couple the light to the tip/sample junction of a commercial STM. An STM can be operated in one of two modes: constant height or constant current. A STORM apparatus would be operated in the constant-current mode, in which the height of the tip relative to the specimen would be varied in order to keep the tunneling current constant. In this mode, a feedback control circuit adjusts the voltage applied to a piezoelectric actuator in the STM that adjusts the height of the STM tip to keep the tunneling current constant. The exponential relationship between the tunneling current and tip-to-sample distance makes it relatively easy to implement this mode of operation. The choice of method by which the photoenhanced portion of the tunneling current would be measured depends on choice of the frequency at which the input illumination would be modulated (chopped). If the frequency of modulation were low enough (typically < 10 Hz) that the feedback circuit could respond, then the voltage applied to the piezoelectric tip-height actuator could be measured by use of a lock-in amplifier locked to the modulation (chopping) signal. However, at a high modulation frequency (typically in the kilohertz range or higher), the feedback circuit would be unable to respond. In this case, the photoenhanced portion of the tunneling current could be measured directly. For this purpose, the tunneling current would be passed through a precise resistor and the voltage drop would be measured by use of the lock-in amplifier.
Photon scanning tunneling microscopy (PSTM) is used to characterize Si3N4/Si02 optical channel waveguides being used for integrated optical-micromechanical sensors. PSTM utilizes an optical fiber tapered to a fine point which is piezoelectrically positioned to measure the decay of the evanescent field intensity associated with the waveguide propagating mode. Evanescent field decays are recorded for both ridge channel waveguides and planar waveguide regions. Values for the local effective refractive index are calculated from the data for both polarizations and compared to model calculations.
Electron beam welding studies were conducted in the Skylab M551 metals melting experiment, on three different materials; namely 2219-T87 aluminum alloy, 304L stainless steel, and commercially pure tantalum (0.5 wt % columbium). Welds were made in both one gravity and zero gravity (Skylab) environments. Segments from each of the welds were investigated by microhardness, optical microscopy, scanning microscopy, and electron probe techniques. In the 2219-T87 aluminum alloy samples, macroscopic banding and the presence of an eutectic phase in the grain boundaries of the heat affected zone were observed. The stainless steel samples exhibited a sharp weld interface and macroscopic bands. The primary microstructural features found in the tantalum were the presence of either columnar grains (ground base) or equiaxed grains (Skylab). The factors contributing to these effects are discussed and the role of reduced gravity in welding is considered.
While scanning Auger micrographs are used extensively for illustrating the stoichiometry of complex surfaces and for indicating areas of interest for fine point Auger spectroscopy, there are many problems in the quantification and analysis of Auger images. These problems include multiple contrast mechanisms and the lack of meaningful relationships with other Auger data. Collection of multielemental Auger images allows some new approaches to image analysis and presentation. Information about the distribution and quantity of elemental combinations at a surface are retrievable, and particular combinations of elements can be imaged, such as alloy phases. Results from the precipitate hardened alloy Al-2124 illustrate multispectral Auger imaging.
The first set of hydrogen reduction experiments to use actual lunar material was recently completed. The sample, 70035, is a coarse-grained vesicular basalt containing 18.46 wt. percent FeO and 12.97 wt. percent TiO2. The mineralogy includes pyroxene, ilmenite, plagioclase, and minor olivine. The sample was crushed to a grain size of less than 500 microns. The crushed basalt was reduced with hydrogen in seven tests at temperatures of 900-1050 C and pressures of 1-10 atm for 30-60 minutes. A capacitance probe, measuring the dew point of the gas stream, was used to follow reaction progress. Experiments were also conducted using a terrestrial basalt similar to some lunar mare samples. Minnesota Lunar Simulant (MLS-1) contains 13.29 wt. percent FeO, 2.96 wt. percent Fe2O3, and 6.56 wt. percent TiO2. The major minerals include plagioclase, pyroxene, olivine, ilmenite, and magnetite. The rock was ground and seived, and experiments were run on the less than 74- and 500-1168-micron fractions. Experiments were also conducted on less than 74-micron powders of olivine, pyroxene, synthetic ilmenite, and TiO2. The terrestrial rock and mineral samples were reduced with flowing hydrogen at 1100 C in a microbalance furnace, with reaction progress monitored by weight loss. Experiments were run at atmospheric pressure for durations of 3-4 hr. Solid samples from both sets of experiments were analyzed by Mossbauer spectroscopy, petrographic microscopy, scanning electron microscopy, tunneling electron microscopy, and x-ray diffraction. Apollo 17 soil 78221 was examined for evidence of natural reduction in the lunar environment. This sample was chosen based on its high maturity level (I sub s/FeO = 93.0). The FeO content is 11.68 wt. percent and the TiO2 content is 3.84 wt. percent. A polished thin section of the 90-150 micron size fraction was analyzed by petrographic microscopy and scanning electron microscopy.
A review of research investigations of several ultrasonic evaluation techniques applicable to structural ceramics for advanced heat engines is presented. This review highlights recent work conducted under the sponsorship of and at the Lewis Research Center. Results obtained with scanning acoustic microscopy, scanning laser acoustic microscopy, photo acoustic microscopy, and scanning electron acoustic microscopy are compared. In addition to these flaw imaging techniques, microstructure characterization by analytical ultrasonics is described. The techniques were evaluated by application to research samples of monolithic silicon nitride and silicon carbide in the form of discs and bars containing naturally occurring and deliberately-introduced flaws and microstructural anomalies. Strengths and limitations of the techniques are discussed.
A review is presentod of research investigations of several acoustic microscopy techniques for application to structural ceramics for advanced heat engines. Results obtained with scanning acoustic microscopy (SAM), scanning laser acoustic microscopy (SLAM), scanning electron acoustic microscopy (SEAM), and photoacoustic microscopy (PAM) are compared. The techniques were evaluated on research samples of green and sintered monolithic silicon nitrides and silicon carbides in the form of modulus-of-rupture bars containing deliberately introduced flaws. Strengths and limitations of the techniques are described with emphasis on statistics of detectability of flaws that constitute potential fracture origins.
Near-field scanning optical microscopy (NSOM) also called scanning near-field optical microscopy (SNOM) is now well accepted as a powerful tool for sub-wavelength (nanoscale in the optical region) spatial resolution microscopy and a large number of related tasks. The importance lies in the fact of strategic advantages of standard microscopy but with significantly enhanced resolution. Since many modern optical diagnostic techniques have found useful applications in space, it is logical to consider the future role of NSOM in such situations. For example, protein crystal growth study under microgravity conditions is a valid candidate. If applied successfully, processes at molecular level can be studied during the growth. NSOM has already been demonstrated to be useful for the study of such crystals here on earth. The basic principle of NSOM can be illustrated. The illumination-collection mode is shown although several other possible approaches exist. In this, the sample is illuminated and the light from the sample is collected through the same tiny aperture opening. A tapered optical fiber is scanned near the sample surface. The tip is coated generally with a metal with a sub-wavelength aperture opening. The tip-sample distance is maintained constant while scanning. Thus, the optical signal available for collection is generally a function of the optical properties of the sample surface. Since the aperture is sub-wavelength in diameter and the tip is held very close (again in the sub-wavelength domain) to the surface, the lateral resolution in the sub-wavelength domain is obtained. Thus, the typical wavelength- order resolution of ordinary microscopy can be significantly enhanced while maintaining the strategic advantages (no need of sample in vacuum chamber, electron beams, etc). Commercial NSOM systems play a key role in the success and widespread acceptance of the tool. These commercial systems work fairly well in laboratory conditions on earth. However, they may not be deployed on a space mission as such with comparable outcome. In this communication we discuss several potential problems and some possible solutions.
A new tip for scanning tunneling microscopy and a tip repair procedure that allows one to reproducibly obtain atomic images of highly oriented pyrolytic graphite with previously inoperable tips are reported. The tips are shown to be relatively oxide-free and highly resistant to oxidation. The tips are fabricated with graphite by two distinct methods.
Scanning electron microscopy for reliability studies in semiconductor devices
A unique plasma gun employing coaxial carbon electrodes was used in an attempt to deposit thin films of amorphous diamond-like carbon. A number of different structural, compositional, and electrical characterization techniques were used to characterize these films. These included scanning electron microscopy, scanning transmission electron microscopy, X ray diffraction and absorption, spectrographic analysis, energy dispersive spectroscopy, and selected area electron diffraction. Optical absorption and electrical resistivity measurements were also performed. The films were determined to be primarily amorphous, with poor adhesion to fused silica substrates. Many inclusions of particulates were found to be present as well. Analysis of these particulates revealed the presence of trace impurities, such as Fe and Cu, which were also found in the graphite electrode material. The electrodes were the source of these impurities. No evidence of diamond-like crystallite structure was found in any of the film samples. Details of the apparatus, experimental procedure, and film characteristics are presented.