scholarly journals TEM Specimen Preparation Technique For Small Semiconductor Devices

2004 ◽  
Vol 12 (2) ◽  
pp. 38-41 ◽  
Author(s):  
Mark Hudson ◽  
John Benedict ◽  
Philip Flaitz

One of the methods of preparing samples for analysis by Transmission Electron Microscopy (TEM) is the well-known procedure using a Tripod polisher and the wedge technique (1-3), developed in our laboratory. Though developed explicitly for preparing samples of integrated circuit structures built on silicon, the technique has been used in our lab for a wide variety of other materials, including metals and ceramics. In general when working with silicon samples, we have the luxury of starting with large wafer pieces or chips that are generally at least a couple of millimeters square or larger. Recently, however, we needed to work for an extended period on GaAs based lasers, where the devices requiring analysis were individual lasers extracted from individual packages. These small samples, measuring 100x100x50 microns, are too small for the manual handling involved in routine mechanical cross sectioning methods.

2000 ◽  
Vol 6 (S2) ◽  
pp. 498-499
Author(s):  
M. V. Hudson

This technique is used to prepare cross sections of semiconductor devices in small samples for analysis by Transmission Electron Microscopy (TEM). These small samples, measuring 100 X 100 X 50 microns, are too small for the manual handling involved in routine mechanical cross sectioning methods. A larger sample, for easier manual handling, is made by gluing the original small sample between a larger piece of silicon and a larger piece of dimpled quartz. The dimpled depression in the quartz is just large enough to surround the original sample. The sample is then mechanically thinned down using a Tripod polisher and the wedge technique. The quartz piece, glued to the top of the original sample, allows the progress of the polish to be monitored as the first side of the cross section is being mechanically polished. The silicon piece, glued to the bottom of the original sample, is used to gauge the final thickness of the wedge produced when polishing the second side of the cross section using the wedge technique.


Author(s):  
K. Doong ◽  
J.-M. Fu ◽  
Y.-C. Huang

Abstract The specimen preparation technique using focused ion beam (FIB) to generate cross-sectional transmission electron microscopy (XTEM) samples of chemical vapor deposition (CVD) of Tungsten-plug (W-plug) and Tungsten Silicides (WSix) was studied. Using the combination method including two axes tilting[l], gas enhanced focused ion beam milling[2] and sacrificial metal coating on both sides of electron transmission membrane[3], it was possible to prepare a sample with minimal thickness (less than 1000 A) to get high spatial resolution in TEM observation. Based on this novel thinning technique, some applications such as XTEM observation of W-plug with different aspect ratio (I - 6), and the grain structure of CVD W-plug and CVD WSix were done. Also the problems and artifacts of XTEM sample preparation of high Z-factor material such as CVD W-plug and CVD WSix were given and the ways to avoid or minimize them were suggested.


Author(s):  
Chin Kai Liu ◽  
Chi Jen. Chen ◽  
Jeh Yan.Chiou ◽  
David Su

Abstract Focused ion beam (FIB) has become a useful tool in the Integrated Circuit (IC) industry, It is playing an important role in Failure Analysis (FA), circuit repair and Transmission Electron Microscopy (TEM) specimen preparation. In particular, preparation of TEM samples using FIB has become popular within the last ten years [1]; the progress in this field is well documented. Given the usefulness of FIB, “Artifact” however is a very sensitive issue in TEM inspections. The ability to identify those artifacts in TEM analysis is an important as to understanding the significance of pictures In this paper, we will describe how to measure the damages introduced by FIB sample preparation and introduce a better way to prevent such kind of artifacts.


Author(s):  
P. J. Smith ◽  
M. V. Kulkarni ◽  
H. A. Troutman

A new preparation technique for transmission electron microscopy samples has been developed which allows an individual device within a silicon integrated circuit to be thinned.Typical devices within present silicon integrated circuits are located in the top few microns of a silicon chip, Fig. 1, and have lateral dimensions on the order of a fraction of a millimeter. Sample preparation for electron microscopy is usually accomplished by jet-etching the bottom side of the wafer, directly under the device of interest; thickness is estimated by the color of transmitted light. It is extremely difficult to ensure that the sample will be thinned in exactly the area of interest since the jet is incident on the bottom side of the sample while the device is on the top side, making exact alignment difficult; the jet is large compared to transistor dimensions; the presence of differently doped regions leads to non-uniform thinning.


Author(s):  
K. Fortunati ◽  
M. Fendorf ◽  
M. Powers ◽  
C.P. Burmester ◽  
R. Gronsky

Transmission electron microscopy, in particular high-resolution TEM, is proving to be a valuable tool in the continuing effort to characterize and understand the “high-Tc” superconducting oxides. Since specimen quality is of critical importance in high-resolution studies, care must be taken to choose the most appropriate specimen preparation technique for the material under study. The BiCaSrCuO material investigated here was in the form of small, sintered pellets with a porous microstructure which consists of small, randomly oriented, poorly connected, plate-like grains (see Figure 1). We have found that this morphology can significantly effect the production of suitable TEM specimens.The simplest and most rapid specimen preparation method employed consists of crushing a small amount of the starting material to a fine powder in an agate mortar and suspending the powder in pure ethanol or propanol. An eye dropper or syringe is then used to transfer 4-6 drops of the suspension onto a holey carbon film supported on a mesh grid, thus effectively dispersing the powder across the grid. A strong tendency for the crystal to cleave along (001) planes, due to the weak bonding between BiO layers, results in flake-like particles which exhibit a preferred [001] orientation on the grid. A high-resolution image of a specimen prepared using this method is shown in Figure 2. We have observed that some specimens produced in this manner are unstable under a 200kV beam (with LaB6 filament), with heavy damage occurring within the time that a through-focus series of micrographs can be exposed. It is also important to note that since separation along grain boundaries occurs during crushing, this method is not an appropriate choice for imaging grain boundary structures.


Sign in / Sign up

Export Citation Format

Share Document