Nowadays there is great interest in the application of reflection high energy electron diffraction (RHEED) to monitor the growth of ultra thin films. This popularity of RHEED arose from the discovery of RHEED intensity oscillations in the early eighties. Namely, it was experimentally found that if the growth of a material follows a layer-by-layer mode then regular oscillating changes in the intensity of the specular beam occur, and the period of the oscillations corresponds to the deposition of one atomic layer of the material. These findings are of great practical importance and consequently RHEED is an important experimental technique used in nanoscale engineering. Nevertheless, the basic question of why RHEED oscillation are observed still remains open. In the past the Philips group suggested that intensity oscillations are a consequence of periodic changes in the roughness of the surface. About the same time, the University of Minnesota group claimed that during the growth the electron wave is reflected by two terraces (in an idealised case) and periodic changes of the interference conditions imply RHEED oscillation.

Since the original observations by Duncumb in 1962, a number of studies have been conducted on the effects of electron channel on characteristic x-ray emission and microanalysis. Most of the recent studies have concentrated upon using the phenomenon to perform site specific distributions of impurity elements in ordered compounds using the ALCHEMI methodology. Very few studies have attempted to accurately measure the effect as a function of orientation and compare these results to theory. In this study, two dimensional high angular resolution studies of channeling enhance x-ray emission were performed and herein the results are compared to theoretical calculations of Allen etal.

All experimental measurements presented here were conducted on a Philips EM 420T analytical electron microscope. The instrument was operated in the TEM mode, at 120 kV using an LaB6 electron source. The characteristic x-ray emission was measured using an EDAX ultra thin window Si(Li) detector having a FWHM of ∼ 145 eV at Mn Kα.

Introduction

We report here results additional to those presented earlier on single amorphous carbon nano-tips grown by carbon contamination on substrates of well-defined geometry in a scanning electron microscope (SEM). This method offers beneficial accuracy in controlling the properties of the nanoemitters such as length, aspect ratio and positioning on the given substrate.

The work function, emission field, current density, emitting area, solid angle of emission, energy distribution and short and long term stability of these emitters are all of interest. Since the literature does not provide, as far as we know, the work function of contamination-grown amorphous carbon, its value must be determined experimentally, in our case through Fowler-Nordheim (F-N) plots. This quantity and others can be derived once the electric field F at the emitting surface is known. Since the geometry of our emitters is reasonably well known from the fabrication procedure and subsequent electron microscopy, it is possible to find a satisfactory value for F by computation.

In 1970, Breedlove and Trammel predicted that, because of radiation damage, the direct imaging of individual organic molecules by electron or X-ray microscopy was not possible. This follows from an analysis, above IkeV, of the beam energy dependence of the ratio of inelastic scattering cross sections (some of which causes damage) to elastic cross sections (used to form the image). To minimize damage, we have therefore constructed a transmission electron microscope with nanometer resolution, which operates at 7 - 300 eV, below the threshold for most inelastic processes. Recently we have obtained our first images of TMV with this instrument. Isaacson first suggested that the K-shell ionization of carbon may be the most important damage process. Although less probable than valence excitation, it dumps more energy into the system, and the decay products are highly damaging. A subsequent study of electron beam damage in several π- bonded hydrocarbon crystals down to about 200 eV showed that a damage threshold does exist at about IkeV incident energy which can be explained by the carbon K-shell ionization. (Multiple scattering increases the energy threshold from 285eV).

The structure of small particles attracts great interest for their distinctiveness. Small particles of several materials undergo lattice parameter change or crystal structure change and sometimes form multiply twinned particles (MTPs) [1]. Configurational and structural fluctuations which they show under electron beam irradiation are also interesting phenomena [2]. This quasi-molten state are well explained theoretically [3], but the exact cause of it is still a matter of debate [2,4].

To clarify the behavior of nanoparticles under electron beam more precisely, we grew indium, palladium and molybdenum nanoparticles on Si (110) TEM samples in situ and investigated with an ultrahigh vacuum field emission transmission electron microscope (UHV-FE-TEM). It is a JEOL-2000VF operated at 200 kV with a base pressure of the column less thari 2.0 x 1O−8 Pa. Two kinds of evaporators are attached to the column in order to perform in situ evaporation of metals; an electron beam (EB) evaporator for high melting temperature material and a thermal heating evaporator with tungsten wire for low melting temperature ones.

Metallic nanoparticles finely dispersed onto high surface-area supports play an important role in heterogeneous catalysis. The performance of a supported metal catalyst can be directly related to the size and spatial distribution of the metallic nanoparticles. With the recent development of highresolution SEM instruments, it is now possible to observe nanoparticles in a field emission SEM. At low voltages, surface details of catalyst supports as well as metallic nanoparticles can be observed. The particle contrast in low voltage SEM images, however, is still not well understood. We have previously shown that the contrast of metallic particles can be enhanced if a small positive potential is applied to the sample. It is suggested that backscattered electrons (BE) significantly contribute to the visibility of metallic nanoparticles in high-resolution SE images. In this paper, we report further study on the origin of particle contrast in high-resolution SE images.

Figure 1 shows a set of SE images of the same area of a carbon supported Pt catalyst.

The development of the very large and ultra large scale integrated technology technology is based on the shrinkage of the lateral dimensions of semiconductors devices to a sub-micron level. Under these conditions the precise control of the total amount of dopant atoms dissolved in the silicon substrate, and of their depth and lateral distribution, become crucial requirements. One of the most promising device characterization technique is that based on transmission electron microscopy (TEM) analysis. Indeed, by means of TEM observations the device morphology can be imaged with a spatial resolution of the order of 0.1 — 0.3 nm; and, after an appropriate sample preparation method, based on selective chemical etch of doped silicon regions, the 2D dopant concentration profiles can be delineated with a resolution of the order of 5nm.

The chemical solution used for this aim consists of a mixture of both nitric and hydrofluoric acid.

The epitaxial growth of CdTe on Si is currently being investigated as a means of generating an alternative substrate for the subsequent growth and processing of HgCdTe based infrared detecting devices. The most favorable orientation for the growth of high-quality HgCdTe is CdTe (111)B, which has been demonstrated to be grown on miscut Si (001) substrates with very high quality . However, twinning islands and multiple domains are present at the nucleation stage, which has a detrimental effect on the device properties.

An alternative approach, which has the potential to overcome the multiple domain problems in the nucleation stage, is to grow CdTe (111) B on Si (111) substrates. However, previous reports have shown that the direct growth of CdTe onto Si (111) results in an A-polarity for the films . In an attempt to overcome this limitation, passivation of the substrate surface has been investigated.

Narrow band gap semiconductors such as ReSi2-x (Eg ∽0.12 eV) are potential materials for infrared detectors [1]. Further, ReSi2-x is in thermodynamic equilibrium and has a very small lattice mismatch with Si offering the potential of developing ReSi2-x /Si heterojunction devices where the detector element and the signal processing circuitry can be integrated on one Si chip. In a previous study, strong crystallographic alignment between ReSi2-x film and (001)Si substrate was observed by ion channeling [1]. In this study, a transmission electron microscopy (TEM) study has been performed on ReSi2-x epitaxial films on (001) Si. Comparisons are made to our previous microscopy study [2] on the defect structures in bulk single crystals of ReSi2-x.

ReSi2-x films were prepared by reactive deposition epitaxy (RDE) technique by evaporating Re onto (001) Si wafer at 650 °C. A cap layer of Cr was evaporated at room temperature. A bright field (BF) TEM image of the interface in cross-section is shown in Fig. 1 with the corresponding selected area diffraction pattern (SADP) shown in Fig. 2.

The rate of heteroepitaxial strain relaxation via misfit dislocation introduction is strongly limited by the kinetics of dislocation nucleation, propagation and interaction. Here we describe real-time observations of the interaction between moving threading dislocations and pre-existing interfacial misfit dislocations using in-situ transmission electron microscopy. This was accomplished both during in-situ observations of the heteroepitaxial growth process using a modified ultrahigh vacuum TEM (UHV-TEM) equipped with chemical vapor deposition capabilities, as well as during in-situ anneals of metastable structures grown by molecular beam epitaxy.

In Figure 1, we present our UHV-TEM observations of the epilayer thicknesses and compositions at which dislocation interactions result in blocking of the propagating threading segment, and compare these with the theoretical predictions of Freund and those of Schwarz and Tersoff. Our results show that dislocation blocking can significantly affect the overall rate of strain relaxation in these structures. This is important, as blocked threading segments introduce undesired band gap states into electronic devices and can act as easy diffusion paths for impurities and dopants.

Nowadays, High Resolution Electron Microscopes are capable to resolve structures on a scale below 100 pm. They can be equipped for Electron Holography in order to detect electric / magnetic fields and for chemical analyses (Electron Energy Loss Spectroscopy & Energy Dispersed X-ray's) that can be performed with a lateral resolution of 0.5 to 1 nm [1]. With the aid of computer sciences it became also possible to quantify local strain. We utilize Philips CM200 and CM300 field emission instruments with attached image filters, the JEOL Atomic Resolution Microscope and specialized software [2] to perform these tasks. On the other hand, a recent highlight in materials sciences is the development a GaN technology that is driven by a fast trial and error approach and aims to revolutionize lighting [3]. It was unavoidable that basic materials properties of the nano-structured thin films are barely understood because of the rapid progress [4].

Cathodoluminescence (CL) microanalysis (spectroscopy and microscopy) enables both pre-existing and irradiation induced defects in the bulk and surface defect structure of wide band gap materials (i.e. semiconductors and insulators) to be monitored and characterized with high spatial resolution and sensitivity. The local micro-volume of specimen may be selected for investigation by varying the electron beam parameters. CL micro analytical techniques allow the in situ monitoring of electron irradiation induced defects and the investigation of irradiation induced electromigration of mobile charged defect species. Irradiation can result in the formation of defects and /or the transformation of existing defect precursors. CL emissions from a material are usually associated with native and impurity defects of the host lattice, however in special cases CL microanalysis can provide direct or indirect evidence for the presence of interstitial molecular species in a material. Atomic displacements from the normal bonding (i.e. defect free) sites induced by an electron beam can result from either knock-on, or radiolytic processes, depending on the incident electron beam energy.

The performance of III-V nitride-based microelectronic and optoelectronic devices relates directly to the micro structure of these materials. Selective lateral overgrowth has been exploited to produce GaN heteroepitaxial films with low defect density [1]. Si is a promising substrate due to its low cost, large size, and the potential for the intergration of GaN-based optoelectronic devices with Si-based electronics. It is also possible to produce high-quality GaN material for devices using lateral overgrowth on a Si substrate [2]. At present, only limited information on the defect structure in GaN heteroepitaxial films grown by selective lateral growth is available, especially those grown on Si substrate. Recent work [3] on GaN pyramids grown on (111) Si substrates by this method has shown that in the center, or core, of the GaN pyramid (at and above the window area) dislocations thread through the pyramid nearly perpendicular to the substrate surface and the dislocation density is quite high.

The possible dislocations and slip systems in the wurtzite structure are the same as in hcp structure [1]. The Burgers vectors of these dislocations are . The dislocations can lie on either the (0001) basal plane or prism planes. The dislocations lie on pyramidal planes. TEM studies have revealed that there are predominately three types of dislocations in a wurtzite GaN epilayer which has not been grown by selective overgrowth (e. g. [2, 3]). The majority of the dislocations are threading dislocations with Burgers vector which appear randomly in the epilayer, they result from the growth errors during the growth process. The other two types of dislocation are halflpops with a [0001] or Burgers vector. The [0001] dislocation half-loop lies on the prism plane and the dislocation half-loop lies on the (0001) basal plane which usually appears near the epilayer/substrate interface.

This report provides analysis data on black deposits observed on gold-plated test socket pads. The socket pads are an integral part of a burn-in test where the test simulates an accelerated aging process and is routinely performed on new products. The test consists of placing a device, as shown in Figure 1, with lead/tin plated copper leads in a test socket then placing that assembly into a “burn-in” oven set to a specific elevated temperature, 150° C in this case, for a specific time. Typically, a bias voltage is applied through the pads on the socket during the test. Historically, during extended burn-in tests of Quad Flat Pack (QFP) devices, problems arose when several devices unexpectedly exhibited open socket-to-device contacts. The “opens” problem repeated itself when new QFP devices were installed in the sockets . Upon visual inspection of the sockets, black deposits were observed in the contact regions, as shown in Figure 2.

Focused ion beam (FIB) instruments are useful for high spatial resolution milling, deposition, and imaging capabilities. As a result, FIB specimen preparation techniques have been widely accepted within the semiconductor community as a means to rapidly prepare high quality, site-specific specimens for transmission electron microscopy (TEM) [1]. In spite of the excellent results that have been observed for both high resolution (HREM) and standard TEM specimen preparation applications, a degree of structural modification is inherent to FIB milled surfaces [2,3]. The magnitude of the damage region that results from Ga+ ion bombardment is dependent on the operating parameters of the FIB (e.g., beam current, beam voltage, milling time, and the use of reactive gas assisted etching).

Lattice defects occur as a consequence of FIB milling because the incident ions transfer energy to the atoms of the target material. Momentum transferred from the incident ions to the target atoms can result in the creation of point defects (e.g., vacancies, self interstitials, and interstitial and substitutional ion implantation), the generation of phonons, and plasmon excitation in the case of metal targets.

Development of future semiconductor processes and improving existing ones requires a strong capability to identify and eliminate the sources of particle defects. The current industry approach for accomplishing this task is to analyze the particles on the wafer surface. The most commonly used technique for this purpose is SEM with EDX (SEM/EDX). This approach provides information about defect chemistry and morphology, which in turn provides a direct path to the reevaluation of processes and hardware parts needed to determine the root cause of a problem. However, the SEM/EDX has three fundamental limitations: (a) it is a large volume technique and therefore, the sensitivity can be insufficient for sub micron particles and thin patches; (b) there is no chemical bonding information due to poor energy resolution; (c) charging problem for dielectrics. In some cases there is a need to use other techniques to support or substitute the EDX results.

Devices for character recognition as well as cellular phones require computational elements that work at higher speeds with lower current requirements. Separation by IMplanted Oxygen (SIMOX) is one type of Silicon On Insulator (SOI) technology that shows great promise in meeting the future demands for faster and more efficient applications. The ultra-thin SIMOX substrates produced by low energy/low dose implant methods make possible the construction of large scale integrated circuits with fully depleted CMOS devices. One of the great challenges in the production of SIMOX technology is achieving high quality Si and Si02 layers. High energy implantation of 0+ ions causes damage to the Si crystal and therefore requires a high temperature annealing step to repair it. Annealing of SIMOX takes place in a mixed atmosphere of argon and oxygen. Having oxygen in the ambient creates a superficial Si02 layer. This reduces the thickness of the SOI layer but also protects the surface from pitting.

To serve the needs of future submicron electronic applications such as cellular communications and character recognition, the computational elements have to be drastically reduced in size as well as power consumption. A promising approach is based on implantation of oxygen ions into a silicon substrate and subsequent high temperature annealing. The process when properly done, results in a very thin silicon layer separated from the bulk silicon substrate by a thin layer of silicon oxide. In these SIMOX elements the top silicon layer is seriously degraded by defects introduced by the ion implantation process, consequently affecting the performance of devices built on them. A careful selection of processing parameters such as implant energy and dose is essential in order to minimize the defect densities. It is believed that Si interstitials are the main reason for the formation of most defects in SIMOX [1]. Consequently, the reduction of the implant oxygen dose minimizes the Si interstitials, and finally lowers the defect densities in both the top Si and buried oxide layers.

Wafer bonding allows the production of Compliant Universal substrates that are made by bonding a thin (< 10 nm) layer twisted ∼45 degrees to the underlying substrate. Subsequent growth on this twisted layer results in defect free films even when the growth material has a significant lattice mismatch with the substrate. Defects on the bonding interface are a common observation when bonding GaAs to many substrates, but the exact nature of these defects has not been clear. We have studied this bonding layer in GaAs-GaAs twist bonded structures by Scanning Transmission Electron Microscopy and Electron Energy Loss Spectroscopy and established that the defects are voids with a portion being partially filled with gallium. Two general sizes of voids are seen. The larger voids are approximately 45 nm in diameter and 22 nm in the wafer normal direction and are distributed in an approximately linear relationship.