Thermo-desorption measurements during N-doped Ge-rich Ge2Sb2Te5crystallization.

Ge-rich Ge2Sb2Te5(GGST) is considered as one of the best candidates for industrial phase change memory production. GGST memory cells are generally embedded with Si or Ti nitride layers to prevent oxidation, as it leads to an undesired decrease of the GGST crystallization temperature. Furthermore, GGST films are usually doped with elements such as N, C, O, or Bi, aiming to delay GGST crystallization during the fabrication process as well as during memory cell operation. In this work, ultrahigh vacuum thermal desorption spectroscopy (TDS) was performed during isochronal annealing of a N-doped GGST film covered by a 10 nm-thick TiNxlayer. Desorption is observed before GGST crystallization, but the comparison between TDS andin situx-ray diffraction measurements shows that the main desorption peak, observed between 653 K and 703 K, occurs after GGST full crystallization. The most prominent desorbing species are Ar, N2, H2, and H. These results show that the TiNxpolycrystalline layer cannot prevent N atoms from leaving the GGST layer during annealing, suggesting a progressive change of the N-doped GGST chemical composition during thermal annealing and crystallization.

Kinetic Monte Carlo simulations of Ge-Sb-Te thin film crystallization.

Simulation of atomic redistribution in Ge-Sb-Te (GST)-based memory cells during SET/RESET cycling is needed in order to understand GST memory cell failure and to design improved non-volatile memories. However, this type of atomic scale simulations is extremely challenging. In this work, we propose to use a simplified GST system in order to catch the basics of atomic redistribution in Ge-rich GST (GrGST) films using atomistic kinetic Monte Carlo simulations. Comparison between experiments and simulations shows good agreements regarding the influence of Ge excess on GrGST crystallization, as well as concerning the GST growth kinetic in GrGST films, suggesting the crystallized GST ternary compound to be off-stoichiometric. According to the simulation of atomic redistribution in GrGST films during SET/RESET cycling, the film microstructure stabilized during cycling is significantly dependent of the GST ternary phase stoichiometry. The use of amorphous layers exhibiting the GST ternary phase stoichiometry placed at the bottom or at the top of the GrGST layer is shown to be a way of controlling the microstructure evolution of the film during cycling. The significant evolution of the local composition in the amorphous solution during cycling suggests a non-negligible variation of the crystallization temperature with operation time.

3D high resolution imaging for microelectronics: A multi-technique survey on copper pillars.

In microelectronics, recently developed 3D integration offers the possibility to stack the dice or wafers vertically instead of putting their different parts next to one another, in order to save space. As this method becomes of greater interest, the need for 3D imaging techniques becomes higher. We here report a study about different 3D characterization techniques applied to copper pillars, which are used to stack different dice together. Destructive techniques such as FIB/SEM, FIB/FIB, and PFIB/PFIB slice and view protocols have been assessed, as well as non-destructive ones, such as laboratory-based and synchrotron-based computed tomographies. A comparison of those techniques in the specific case of copper pillars is given, taking into account the constraints linked to the microelectronics industry, mainly concerning resolution and sample throughput. Laboratory-based imaging techniques are shown to be relevant in the case of punctual analyses, while synchrotron based tomographies offer highly resolved volumes for larger batches of samples.

Retrieving overlapping crystals information from TEM nano-beam electron diffraction patterns.

The diffraction patterns acquired with a transmission electron microscope (TEM) contain Bragg reflections related to all the crystals superimposed in the thin foil and crossed by the electron beam. Regarding TEM-based orientation and phase characterisation techniques, the nondissociation of these signals is usually considered as the main limitation for the indexation of diffraction patterns. A new method to identify the information related to the distinct but overlapped grains is presented. It consists in subtracting the signature of the dominant crystal before reindexing the diffraction pattern. The method is coupled to the template matching algorithm used in a standard automated crystal orientation mapping tool (ACOM-TEM). The capabilities of the approach are illustrated with the characterisation of a NiSi thin film stacked on a monocrystalline Si layer. Then, a subtracting-indexing cycle applied to a 70 nm thick thin foil containing polycrystalline tungsten electrical contacts shows the capability of the technique to recognise small nondominant grains.

TEM illumination settings study for optimum spatial resolution and indexing reliability in crystal orientation mappings.

The spatial resolution and the indexing quality obtained with an automated orientation and phase mapping tool are analyzed for different Transmission Electron Microscope (TEM) illumination settings. The electron probe size and convergence angle are studied for two TEM configuration modes referred as microprobe and nanoprobe modes. Using a 10µmC2 aperture in a FEI Tecnai F20 (S)TEM, the nanoprobe mode is used to get a small convergent electron beam while the microprobe mode provides a nearly parallel illumination at the cost of a larger probe size. The nanoprobe configuration enables to increase the spatial resolution (∼1nm vs 3nm) but also affects the fraction of mis-indexed points (15% vs 1%). Indexing errors are attributed to the increase by a factor of three of the convergence angle with respect to the microprobe mode. While intermediate optimum settings may be found and are potentially achievable on electron microscopes providing a 'free lens' control or a larger choice of C2 apertures, it is emphasized that the spatial resolution cannot be considered without reference to the indexing quality and, consequently to the convergence angle.

Correction of absorption-edge artifacts in polychromatic X-ray tomography in a scanning electron microscope for 3D microelectronics.

X-ray tomography is widely used in materials science. However, X-ray scanners are often based on polychromatic radiation that creates artifacts such as dark streaks. We show this artifact is not always due to beam hardening. It may appear when scanning samples with high-Z elements inside a low-Z matrix because of the high-Z element absorption edge: X-rays whose energy is above this edge are strongly absorbed, violating the exponential decay assumption for reconstruction algorithms and generating dark streaks. A method is proposed to limit the absorption edge effect and is applied on a microelectronic case to suppress dark streaks between interconnections.

Reduction of the scanning time by total variation minimization reconstruction for X-ray tomography in a SEM.

Total variation minimization is applied to the particular case of X-ray tomography in a scanning electron microscope. To prove the efficiency of this reconstruction method, noise-free and noisy data based on the Shepp & Logan phantom have been simulated. These simulations confirm that Total variation minimization-reconstruction algorithm better manages data containing low number of projections with respect to simultaneous iterative reconstruction technique or filtered backprojection, even in the presence of noise. The algorithm has been applied to real data sets, with a low angular sampling and a high level of noise. Two samples containing micro-interconnections have been analyzed and 3D reconstructions show that Total variation minimization-based algorithm performs well even with 60 projections in order to properly recover a 500 nm diameter void inside a copper interconnection.

Deep sub micrometer imaging of defects in copper pillars by X-ray tomography in a SEM.

The potential of X-ray nanotomography hosted in a SEM in presented in this paper. In order to improve the detail detectability of this system, which is directly related to the X-ray source size, thin metal layers have been studied and installed in the equipment. A 3D resolution pattern has been created in order to determine the smallest detectable features by this setup. This sample is a 25 µm diameter copper pillar in which size-controlled holes have been milled using a plasma-focused ion beam. This pattern has then been scanned and the resulting 3D reconstruction demonstrates that the instrument is able to detect 500 nm diameter voids in a copper interconnection, as used in 3D integration.

Three-dimensional imaging of copper pillars using x-ray tomography within a scanning electron microscope: a simulation study based on synchrotron data.

While microelectronic devices are frequently characterized with surface-sensitive techniques having nanometer resolution, interconnections used in 3D integration require 3D imaging with high penetration depth and deep sub-micrometer spatial resolution. X-ray tomography is well adapted to this situation. In this context, the purpose of this study is to assess a versatile and turn-key tomographic system allowing for 3D x-ray nanotomography of copper pillars. The tomography tool uses the thin electron beam of a scanning electron microscope (SEM) to provoke x-ray emission from specific metallic targets. Then, radiographs are recorded while the sample rotates in a conventional cone beam tomography scheme that ends up with 3D reconstructions of the pillar. Starting from copper pillars data, collected at the European Synchrotron Radiation Facility, we build a 3D numerical model of a copper pillar, paying particular attention to intermetallics. This model is then used to simulate physical radiographs of the pillar using the geometry of the SEM-hosted x-ray tomography system. Eventually, data are reconstructed and it is shown that the system makes it possible the quantification of 3D intermetallics volume in copper pillars. The paper also includes a prospective discussion about resolution issues.

Chemical 3D tomography of 28nm high K metal gate transistor: STEM XEDS experimental method and results.

A new STEM XEDS tomography technique is proposed thanks to the implementation of multi EDX SDD detectors in analytical TEMs. The technique flow is presented and the first results obtained on a 28nm FDSOI transistor are detailed. The latter are compared with 2D XEDS analysis to demonstrate the interest of the slice extraction in all directions from a large analyzed volume without any 3D overlap effect issues.

Three-dimensional semiconductor device investigation using focused ion beam and scanning electron microscopy imaging (FIB/SEM tomography).

Three-dimensional focused ion beam/scanning electron microscopy (FIB/SEM tomography) is currently an important technique to characterize in 3D a complex semiconductor device or a specific failure. However, the industrial context demands low turnaround time making the technique less useful. To make it more attractive, the following study focuses on a specific methodology going from sample preparation to the final volume reconstruction to reduce the global time analysis while keeping reliable results. The FIB/SEM parameters available will be first analyzed to acquire a relevant dataset in a reasonable time (few hours). Then, a new alignment strategy based on specific alignment marks [using tetraethoxylisane (TEOS) and Pt deposition] is proposed to improve the volume reconstruction speed. These points combined represent a considerable improvement regarding the reliability of the results and the time consumption (gain of factor 3). This method is then applied to various case studies illustrating the benefits of the FIB/SEM tomography technique such as the precise identification of the origin of 3D defects, or the capability to perform a virtual top-down deprocessing on soft material not possible by any mechanical solution.