Due to dependence of milling rates on copper grain orientation, Cu remains a circuit edit challenge. First it was Cu over SiO2 and now Cu over lower-k dielectrics. The choice of etch-assisting chemistry is important as it affects the degree of dielectric over-etch when cutting Cu. An ideal precursor should uniformly etch Cu and protect underlying dielectric. Cu does not have non-corroding, simple volatile compounds, which survive ion bombardment (such as AlI3 AlBr3 and AlCl3 for aluminium). Milling uniformity, therefore, must be achieved through atomic mixing of copper atoms with precursor components so as to suppress channeling in Cu grains and yield non-conductive sputtered products. Oxygen containing compounds would seem to be the best precursors for this role. However, O may readily destroy organic dielectrics since it volatilises carbon (CO2 and CO), essential to organic dielectrics or carbon doped oxides (CDO). We will discuss precursors that help Cu etching yet reduce etching rates of known organic dielectrics.
New Integrated Circuit designs frequently use two top metal interconnect layers as power and ground bases, what makes accessing the signal nodes and circuitry modification through the front side very difficult. Accessing the nodes through silicon requires precision dose control for not only milling the access via, but also for the via filling process. We present an overview of experimental sample absorbed current monitoring set-up and examples of the "stage current" plots, enabling precision end-pointing of via milling and in-via deposition processes in backside circuit edit application.
The use of focused ion beam (FIB) system for failure analysis is greatly increasing. FIB system is employed for both fault localisation and sample preparation. In particular, FIB voltage contrast plays an important role during full failure analysis. Here we will present a series of examples, which are done in Philips Semiconductors Nijmegen
Voltage contrast like imaging of n-wells in a grounded p-substrate is presented as a new method for assisting FIB microsurgery from the backside of the die. The contrast can be produced on the bare silicon as a transient signal that can be used for end-pointing of the backside trenching process in order to separate between bulk and active silicon volume. It can be made permanent by a spontaneous XeF2 etch followed by the deposition of a SiOx layer. Applications include direct CAD – FIB alignment and visualisation of the wells for analysis purposes. Basic effects of the active devices in Silicon that are the origin of the contrast are presented. The contrast variation on a SiOX film has been investigated in detail as a function of the parameters beam energy, beam current, scanning speed and magnification. Further projects on new processes are introduced.
As the voltage contrast is routinely used for fault localisation in test structures, some basic experiments have been done to determine the critical resistivity for voltage contrast imaging of a failing structure. Furthermore, the resistivity of FIB deposited metal lines has been measured to evaluate how these metal lines can be used to ground certain test structures. The usage of Kleindiek micromanipulators for SEM/FIB in situ micro probing will be described.
The FEI Dualbeam FIB at Delft University of Technology is equipped with an Electron Backscattering Diffraction (EBSD) system for local crystallographic analysis. The combination of both techniques is an interesting field to investigate. To do so, lift-out TEM specimens were prepared with the FIB and deposited on the carbon film of a copper TEM grid. EBSD was performed on these very thin specimens made of different materials (Cu3Au, Cu, Cr). This method might produce high quality diffraction patterns with improved signal-to-background ratios and allow orientation mapping on cross section with a better spatial resolution than other techniques. Advantages and disadvantages of the technique will be discussed.
In recent years, lift-out techniques for TEM sample preparation have received considerable attention as an alternative to FIB classical trench milling. In particular, removing the slice of interest and attaching it to a support in-situ seems to be a very promising solution, as it allows one to further thin the specimen after TEM analysis and guarantees a wide tilting range as no thick sidewalls are present. In this work, we employed a slightly modified version of the in-situ lift-out method to prepare and analyse TEM specimens from different materials. These include thick (> 10 µm) SiGe layers, sub-quarter-micron HMOS devices and Cu pins. In addition, Convergent Beam Electron Diffraction (CBED) was used for a preliminary study of the effects of this type of sample preparation on the strain distribution within the material, as compared to standard FIB trench milling. Problems encountered as well as possible solutions are discussed here, together with some suggestions for future developments of the technique.
Focused Ion Beams (FIB) are widely used in semiconductor industry for local deposition of metal layers. Such a deposition may be used for device modifications or as a protective layer for subsequent milling operations prior to FIB sections or TEM sample preparations. An organometallic compound (e.g. PtC9H16 for Platinum deposition) is introduced into the chamber. The deposition process relies on ion assisted CVD, i.e. the molecules that are adsorbed on the surface are dissociated by the energetic ions. The main drawback of the process is that ion-sample interactions result in damage (milling, amorphisation) of the top sample surface, which may be the layer of analytical interest. As an alternative, the e-beam assisted platinum deposition has been investigated in a dual beam system in which both an ion column and an electron column are integrated in the same chamber. E-beam energy, beam current density and scanning modes influence metal deposition rate. Low accelerating voltage and high current densities increase the deposition rate while fast scanning modes improve deposition uniformity. Experiments, process characterisation and applications that have been studied in the Physical Characterisation group at Crolles will be discussed.
The examination of the metal-oxide interface of materials, oxidised in different environments will bring information about the mechanisms involved in the oxidation. Preparation of transverse TEM samples however in such cases is a challenge as the interface is a single plane (even though not a monolayer). Standard techniques used till recently include ion beam milling, ultra-microtomy and to a certain extent tripod polishing. In the cases where the oxide has a large compressive stress the sample preparation is even more delicate as the thin regions are vulnerable to shattering. The possibility of using FIB to prepare such samples has made the preparation and examination of such samples extremely accessible. In this presentation the use of FIB will be demonstrated in such cases and the different advantages of the technique will be shown in the preparation of metal-oxide interfaces. The precautions to be taken during the sample preparation and the probable artefacts produced will be discussed. Examples of TEM analyses will be given to illustrate the advantage of the technique.
The FIB is a valuable tool to prepare TEM lamella because it allows to select the position of the lamella with high precision. This advantage often outweighs other disadvantages of the FIB lamella preparation method as, for example, the higher level of surface damage. Here, a new kind of preparation artefact is presented and analysed. In the TEM image, it appears as dark, spherical spot with diameters up to a few hundred nanometers. In the SEM, spherical particles were detected on the surface of the TEM lamella. EDX microanalysis revealed a high concentration of Ga. The probable root cause is re-deposition of elemental Ga onto the surface of the lamella. This may occur during the preparation procedure in the FIB, depending on the processing conditions. This Ga probably stems from the Ga ion beam of the FIB tool.
We have developed the method to fabricate nano scale 3-dimensional structures using FIB deposition and etching. We have suggested that nano scale 3-dimensional structures should be used as the micro mold. We made the fundamental experiments for micro molds fabricated by FIB. We want to report some results.
While the FIB is frequently applied to a variety of technical materials, it is much less commonly, but also successfully, employed in the TEM sample preparation and micro-machining of materials of biological origin. The advantages that this method provides, particularly with respect to biological samples, are that (i) many materials can be observed in their natural state and arrangement, that (ii) sections exposing the three-dimensional structure of a material or structure can be cut, that (iii) deformation and damage due to mechanical cutting of the specimen are avoided, and that (iv) all this is possible at high resolution and without elaborate sample preparation such as embedding and serial sectioning. Using nutshells, eggshells, wood and insect cuticle as examples, the strengths and weaknesses of the FIB as applied to biological materials are illustrated.
The ability to fabricate structures with nanometer precision is of fundamental importance to any exploitation of nanotechnology. The focused ion beam nanolithopgrahy has been explored to realize novel applications in nanoelectronics, photonic communication devices, bionanotechnology, nanomaterials and stencilling of complex structures and materials. I will go over a wide range of scoping from lower level modification to complete device fabrication. This method is effective and flexible in creating novel nanostructures both with properties otherwise unattainable and with greater complexity. I will discuss its limitations and potentials in the future development of nanotechnology.
