Search

The existence of electromigration in thin films has been acknowledged since the early sixties. Electromigration is described as the main transport for atoms in a conductor under a current stress. Initial interest had been of a theoretical nature as electromigration had little impact on circuit reliability. With the maturing of Very Large Scale Integrated Circuit (VLSI) technology, current densities are exceeding 106 Amps/cm2 while linestripes are reaching into the submicron range. In this environment, electromigration can cause unwanted open or short circuits in thin films. This has serious implications on the reliability of any integrated circuit. By 1990, millions of transistors may be fabricated on a chip with feature sizes smaller than the wavelength of visible light.

There is a great deal of interest in the study of the morphology of cured epoxy resin. Generally the surfaces of the cured samples or fracture surfaces, unetched or suitably etched, are examined on the transmission electron microscope using the two-stage replica method. The replicating fluid commonly used is 25% polyacrylic acid (MW ∼ 250,000) solution in water. It dries to a brittle replica which is removed by mechanical prying with the help of tweezers; extreme caution is required to ensure that it is not damaged. It has been shown that for fractographic studies of epoxy resin, replication based on PAA can give rise to artifacts and hence it is necessary to back it uo with other evidence, e.g. from the SEM. The standard replicating fluid, containing 10% cellulose acetate in acetone, gives a softer and easily removable replica but the acetone can affect the surface of epoxy resin. To circumvent the latter problem, a modified replication method was developed and is described here.

Standard sample preparation techniques of various metals, composites, ceramics and polymer specimens can be very time consuming (such as dimpling and ion milling etc.), yet still introroduce artifacts during specimen preparation. With many of these materials numerous hours and sometimes even days can be spent preparing samples for TEM studies. In the past, ultramicrotomy has been used primarily for biological applications, but over the years has been tried on various polymers, metal and ceramic powders, oxide layers, etc. The main concern has been deformation introduced to the specimen while cutting. This study will demonstrate successful ultramicrotomy over a variety of non-biological specimens.Materials chosen for this experiment consist of Rapidly Solidified (RS) Al-V powders, P-120 Carbon fibers, Silicon Carbide (MMC) fibers, Kevlar and an IC semiconductor chip. Routine SEM micrographs were taken of all material as-received. All the specimens were embedded using a low viscoscity resin, L. R. White (hard grade). They were microtomed with a diamond knife at a 55-60° cutting angle using an AO/Reichert Ultracut Ultramicrotome with a slow cutting speed of 0.4 mm/sec which obtained the best results. All sections were picked up on Hex 700TB Ni grids.

Electromigration is a well known phenomenon in the aluminum films of an integrated circuit and occurs at elevated temperatures under the influence of current densities exceeding 10 amps/cm2. The failure modes are characteristically cracks, voids or hillocks in the metal lines, causing open-circuits. The majority of research in the electromigration of aluminum conductors examines the formation of these voids and hillocks in the linestripes. This is unfortunate since experiments have shown that short-circuits caused by whisker formation between adjacent stripes or between multi-level structures can be equally damaging to device performance. Most previous work has involved examining whiskers grown by annealing, however, very little work has been carried out on whiskers produced by electromigration. In this investigation aluminum whiskers formed by electromigration were studied with a transmission electron microsope to determine their crystallographic orientation.

Fiber reinforced titanium alloys hold immense potential for applications in the aerospace industry. However, chemical reaction between the fibers and the titanium alloys at fabrication temperatures leads to the formation of brittle reaction products which limits their development. In the present study, coated SiC fibers have been used to evaluate the effects of surface coating on the reaction zone in the SiC/IMI829 system.IMI829 (Ti-5.5A1-3.5Sn-3.0Zr-0.3Mo-1Nb-0.3Si), a near alpha alloy, in the form of PREP powder (-35 mesh), was used a茸 the matrix. CVD grown AVCO SCS-6 SiC fibers were used as discontinuous reinforcements. These fibers of 142μm diameter contained an overlayer with high Si/C ratio on top of an amorphous carbon layer, the thickness of the coating being ∽ 1μm. SCS-6 fibers, broken into ∽ 2mm lengths, were mixed with IMI829 powder (representing < 0.1vol%) and the mixture was consolidated by HIP'ing at 871°C/0. 28GPa/4h.

“Molecular composites” are a new class of structural polymers which are high-strength, high-modulus, thermally stable, and environmentally resistant. A rigid-rod, extended chain polymer component is used to reinforce a matrix of a ductile polymer with the intent of achieving a composite on the molecular level. The critical factor in processing a “molecular composites” is that the rigid-rod reinforcing component be well dispersed and not phase separate from the matrix component at any stage of processing. For the greatest versatility, a “molecular composites” system should be amenable to fabrication with traditional thermoplastic processing techniques. We previously reported on the morphology of “molecular composites” formed by coagulation spinning from a solution of rigid-rod/stiff-coil polymer blend and from a solution of a rigid-rod/stiff-coil triblock copolymer. Although these polymer systems formed “molecular composites”, they did not have a glass transition temperature below the degradation temperature and could not be consolidated by thermal processing techniques. In this paper we are reporting on the morphology of rigid-rod and flexible-coil thermoplastic blends which are processable by precipitation and thermal consolidation.

The fabrication of the aluminum bridge test vehicle for use in the crystallographic studies of electromigration involves several photolithographic processes, some common, while others quite unique. It is most important to start with a clean wafer of known orientation. The wafers used are 7 mil thick boron doped silicon. The diameter of the wafer is 1.5 inches with a resistivity of 10-20 ohm-cm. The crystallographic orientation is (111).Initial attempts were made to both drill and laser holes in the silicon wafers then back fill with photoresist or mounting wax. A diamond tipped dentist burr was used to successfully drill holes in the wafer. This proved unacceptable in that the perimeter of the hole was cracked and chipped. Additionally, the minimum size hole realizable was > 300 μm. The drilled holes could not be arrayed on the wafer to any extent because the wafer would not stand up to the stress of multiple drilling.

The Mg-20Nd melt spun ribbonsare being examined in depth because of their potentiodynamic polarization response as compared to that of 7075-T73 aluminum. The Mg-20Nd ribbon exhibited pseudopassivation behavior at a current density approximately one order of magnitude less than the aluminum alloy. In previous work, this was found to be due to the presence of a tessellatal precipitated network which formed in the ribbons during melt spinning. EDS results on TEM foils gave similar compositional values on both matrix and precipitates. These results were not trusted due to the small size and large volume fraction of the precipitates.The Mg-20Nd ribbon was swabbed with a dilute solution of sulfuric acid to loosen the particles in the matrix. A heavy oxide film was also formed on the surface at this time. Plastic replication solution was then applied to this surface and stripped away when dry. The particle side of the plastic film was then carbon coated. This film was placed on 3mm grids and the plastic dissolved away leaving the carbon film and the oxide precipitate mixture.

Titanium alloys produced by the powder metallurgy (PM) techniques are of interest because of the potential cost savings and property enhancement above that possible by conventional ingot metallurgy approaches. In the present program the Ti-5Al-2.5Sn alloy, an all alpha system, is being studied as a model PM titanium alloy. Alloying additions of Si and Ge, which normally result in coarse precipitate regions in cast and wrought material because of their low solubility, have been made. The PM products circumvent the ingot problem because segregation is avoided and precipitates are retained in solution by either the rapid quench rate (pre-alloyed powder) or homogeneous mixing (elemental blend approach). The resulting product displays structures and behavior which differ from conventional Ti-5Al-2.5Sn product. To date Ti-5Al-2.5Sn PM alloys with 0.5 Si and with 1.0Ge w/o have been prepared by an elemental blend/cold compact/sinter process, followed by Hot Isotatic Pressing (HIP'ing). Subsequently specimens were aged at 1000°F (540°C) for 3, 66, and 500 hours in order to precipitate silicides and germanides. Thin foils were then prepared and characterized using TEM and STEM techniques.

A patent pending test holder was designed to support research in electromigration using a JEOL 100 CX STEM. The holder provides electrical connections for an aluminum stripe test vehicle which allows the electromigration process to be conducted within the microscope.The holder is specifically designed for use in conjunction with a JEOL-type electron microscope side entry goniometer. Four leads provide the ability to power three different test stripes on one test vehicle. This allows three separate electromigration runs/observations in-situ without removing the test vehicle from the STEM, thus minimizing test vehicle time outside the vacuum environment.