Workshops > EFUG workshop

Workshop European FIB Users Group (EFUG)

The EFUG activities are currently coordinated by Frank Altmann (Fraunhofer, Halle), Guillaume Audoit (CEA, Grenoble), Hugo Bender (IMEC, Leuven), Hélène Chauvin (Thales, Toulouse).

The European Focused Ion Beam Users Group (EFUG) aims to stimulate the interaction and collaboration between Focused Ion Beam users by the exchange of information, the organization of meetings and the organization of FIB tutorials.

The EFUG includes members of industrial users, universities, various independent institutes and manufacturers.

Program Thursday, September 28 at 9:00. 

Focused Ion Beam is one of the key preparation techniques for failure analysis at microelectronic and microsystem devices. There are multiple applications e.g. for defect screening and site specific preparation on integrated circuits and packaged devices, circuit edit, mask repair and many more. EFUG Workshop (Tuesday, September 20, 2016, 15:10-18:10) will provide a platform to exchange new information and practical experiences regarding the following Topics:

 

–       New instruments : HIM, Plasma-FIB, Laser preparation Tools

–       Efficient preparation workflows : automated TEM preparation, Plasma-FIB, Laser milling, Atom probe sample preparation

–       Circuit edit : application procedures, new gas chemistries

–       Nanoprobing

–       Process monitoring for wafer manufacturing, in-line FIB and wafer return

–       Failure analysis case studies : Si devices, wide bandgap, solar cells 

9:10             Formation of amorphous damaged layer during focused ion beam milling

A. Lotnyk, U. Ross, I. Hilmi, B. Rauschenbach, Leibniz Institute of Surface Modification, Leipzig (GE)

Focused Ga ion beam milling (FIB) is widely used for the production of TEM lamella. The interaction of ions with a specimen surface results in the formation of surface damage (e.g. amorphous layer) which is undesirable for high-resolution (S)TEM. This process can be simulated by Monte-Carlo approach implemented in the Stopping and Range of Ions in Matter (SRIM) package. However, SRIM simulations are often considered as semi-quantitative regarding the formation of amorphous layer during FIB milling. Consequently, this work aims to study amorphisation processes during ion beam milling. Experimental data are compared with SRIM simulations. It was found that the simulations slightly overestimate the thickness of the amorphous layer. The threshold for amorphisation can be characterized by the displacements per target atom (DPA) value for a given ion energy and material system. The composition of the amorphous layer was found to be of the same composition as the target material.

9:30           FIB preparation of samples for dopants imaging using SCM

F. Lorut, STMicroelectronics, Crolles (FR) 

Scanning Capacitance Microscopy is an AFM based technique that provides information about dopants distribution in a semiconductor device. As a mandatory point, sample to be observed must be kept crystalline, and thus, sample preparation is made either by polishing or by cleaving.

When dealing with FA cases where a specific device has to be analyzed, the above mentioned preparation techniques appear too complicate and/or not adapted to target a cross-section at a specified position.

We propose to detail a workflow using a FIB system that enables to pinpoint to the exact location of the fail, while being compliant with the SCM requirements. 

9:50         In situ Delayering and Nanoprobing for Streamlined Failure Analysis

A. Rummel^a, K. Schock^a, M. Kemmler^a, T. Hrncir^b, G. Goupil^b, Sharang^b, M. Sikula^b, S. Kleindiek^a
aKleindiek Nanotechnik, Reutilingen (GE), ^bTESCAN Brno s.r.o, Brno (CZ)

 With the continuous decrease in IC device dimensions, delayering devices using conventional mechanical polishing methods can be problematic. The delayering tasks have become more challenging due to the presence of physically thinner layers and ultra-low-k dielectric materials in recent, more sophisticated circuit designs. Moreover, the delayered surface has to be ready for nanoprobing without obstacles such as oxide layers, mechanical damage or intermixing between layers.

Delayering using a Xe plasma FIB accompanied by in-situ nanoprobing is a promising new method, which is able to address the above mentioned issues. During the FIB delayering process, gases are necessary to equalize FIB milling rates on different materials and to thus obtain perfectly flat surfaces. End point detection is performed by FIB secondary electron or ion signals supported by high resolution SEM imaging at low beam energies in order to prevent damaging any beam-sensitive semiconductor devices. The freshly delayered surface can be immediately investigated via nanoprobing by utilizing the Prober Shuttle integrated into the FIB-SEM chamber. AFM measurements were used to confirm the surfaces' flatness and their suitability for nanoprobing (slightly raised metal/via and transistor contacts over the insulator/dielectric layers). This approach allows nanoprobing on perfectly clean and well prepared surfaces. Using the integrated nanomanipulators, the full range of electrical characterization techniques is accessible without breaking vacuum. In this way, the Failure Analysis sequence: delayering, nanoprobing, and site-specific TEM sample preparation can be performed in one seamless process without switching platforms. 

10:10            Combination of precise laser and FIB milling for TEM based IC failure analysis

F. Altmann, M. Simon-Najasek, S. Hübner, M. Lejoyeux
Fraunhofer Institute for Microstructure of Materials and Systems IMWS, Halle (GE) 

Defects in the crystal structure of the active silicon substrate of integrated circuits (IC) could cause leakages between p-n junctions or source-drain areas of MOSFET transistors, thus resulting in reduced electrical performance or malfunction. Such defects can be localized within the IC structure by photoemission microscopy (PEM) or optical beam induced resistance change (OBIRCH). Unfortunately, the physical root cause analysis with subsequent FIB/SEM cross sectioning and TEM analysis often fails, because the PEM or OBIRCH spot is not precise enough for lamella positioning. Furthermore, SEM observation during FIB milling is not sufficient enough to hit the nm sized defects. Alternatively, a plane view TEM investigation of the active silicon can be performed, with a high success rate, in order to investigate the defect by TEM and characterize the leakage path formation. In this work, we present a new and effective preparation routine to achieve plane view TEM samples from a located IC area including precise laser-milling by MicroPrep XL-ChunkTM technique and subsequent Plasma- and Ga-FIB polishing. In a first step, the localized defective IC structure is separated by laser milling in a 2000μm by 400μm area in less than 10min. Then, the XL chunk is positioned at a perforated Cu ring and fixed with the surrounding IC surface by epoxy glue. Further laser milling is applied to thin the silicon substrate until 50μm residual thickness. After ultrasonic cleaning, the lamella is pre-thinned by fast Plasma-FIB preparation. Finally, the lamella is locally thinned down to electron transparency at the active silicon surface by Ga-FIB under SEM observation to guarantee a uniform plan view sample. The following TEM analysis of the leaky MOSFET transistor reveals a pin like defect in the channel area between source and drain. To get further information about the defect depth and elemental composition of the defect, a second TEM lamella containing the defect structure is cut out of the plan view lamella. The lamella is transferred into a clip holder by in-situ lift out technique and finally polished by Ga-FIB. Our sample is then investigated again by TEM and STEM/HAADF imaging. It could be seen, that the upper 25nm of the channel area was amorphized. This was caused by the necessary FIB removal of the upper IC structure for plane view TEM investigation. The defect could be imaged by STEM-HAADF and was located directly in the silicon surface of the affected transistor channel. Additional EDX mapping revealed that the defect consisted mainly of iron. The root cause of the local iron contamination should be process related e.g. contamination from a process tool for wafer cleaning or gate oxidation.

10:30 COFFEE BREAK

10:50             Pico-Second Laser and Broad Argon Beam Tools for Characterization of Advanced Packages and Devices

Vincent Richard^a and M. Hassel Shearer^b
aGATAN FRANCE - ROPER SCIENTIFIC, Evry (FR), ^bGATAN Inc. Pleasanton (USA)

 “More than Moore” is emerging as one of the solutions to continue the means for increasing speed and functionality of semiconductor devices since geometry scaling is nearly impossible going forward. Although, this methodology relaxes the pressure to find new materials and extremely expensive front end production tools used by process engineering it may increase the difficulty confronting Failure Analysis Engineers.

 A key tool set for Failure Analysis Engineers has been and will continue to be electron and ion microscopes. Electron microscopy imaging and analytical techniques have been pushed to 0.1nm resolution over planar areas as large as 10x 10 microns with device geometries shrinking to the 5 to 10 nm design rules. More than Moore maintains the same planar resolution but with advanced packages and stacked chips, cross sectional depths increase to perhaps as large as a few millimeters. Until now, mechanical polishing or FIB have been the tools of choice for preparing cross sections for Failure Analysis.  Mechanical preparation may be limited due to the increased fragility of the silicon devices thinned to less than 50 microns within these advanced packages. While Ga FIB’s and newer Plasma FIB’s may not have sufficient milling speed necessary to expose regions within a large thick package or cross sectional surfaces of stacked chips. 

An alternative or compliment to FIB, may be a precision pulsed laser tool designed specifically for these applications. These tools have “milling rates” orders of magnitude faster than FIB’s. This higher speed can be utilized to cross section complete packages.

In this paper, a workflow consisting of a pico-second laser tool (microPREP TM) followed by a Broad Argon Beam tool is proposed as a solution for extremely large area preparation with surfaces suitable for SEM and FIB based analysis directly after polishing in the broad argon beam tool.  The advantage of the broad argon beam tool vs. a FIB is that areas mm’s by mm’s in size can be observed

Results will be presented discussing the advantages of this workflow in terms of speed, size and quality of the surface for electron microscopy analysis. Anticipated enhancements that will come as the technique is developed will be discussed. 

11:10             FIB automation for TEM lamella and alignments

V. Brogden, Thermo Fisher, Eindhoven (NL)

As our demand for immediate media and data available at our finger tips increases, so does the production of IC’s that utilize shrinking devices, novel architectures and new materials. In order to meet the challenge of characterizing these new production processes and examining new kinds of failures, semiconductor engineers now face the need for an unprecedented amount of TEM analysis.

Creating a FIB-prepared TEM sample requires a highly-skilled operator to perform hundreds of small tasks sequentially, being careful to not make any mistakes or skip any steps. As chip designers continue their innovation, failure analysis, with its insatiable need for TEM samples, becomes the bottle neck of the yield ramp. Automation has been shown to increase throughput, reduce process mistakes, decrease operator injuries and most importantly, allow managers to utilize their valuable employees in more interesting tasks, diminishing operator fatigue and burn-out. This presentation will explore the applications of FIB automation in order to produce high-volume, high-quality TEM samples with identical lamellae at precise locations. I will also explore ways FIB automation can be used to keep a dual-beam system in excellent condition and ready to lift out the next sample.

11:30             Leveraging F.A. capabilities thanks to 3D Slice and View technique

K. Rousseau, V. Mourier, T. Monniez, C. Hodeau, O. Glorieux, Serma Technologies, Grenoble (FR)

By considering smaller and smaller dimensions of devices being developed in the semiconductor industry and requiring more and more new arrangement of materials for getting the desired functions, running a failure analysis can be mazed without proper tools to make progress. Only a clean sample preparation can pinpoint the failure mechanism of an unexpected behaviour and will drive to an accurate explanation of the situation. FIB-SEM tomography, also known as 3D Slice and View technique (3DSV), was recently introduced in order to tackle down discrepancies raised during F.A. investigations.

In this contribution, the great interests of 3DSV technique will be presented through many complex investigations in repetitive chips such as for example image sensors or memory arrays. The difficulties in data acquisition and processing will be also discussed. 

11:50             COLDFIB:The new FIB source from laser cooled atoms

M.Reveillard^a, M. Viteau^a, A. Houel^a, A. Delobbe^a, D.Comparat^b
aOrsay Physics, Tescan Orsay Holding, Fuveau (FR), ^bLaboratoire Aimé Cotton, Université Paris-Sud, ENS Cachan, CNRS, Université Paris-Saclay, Orsay (FR)

Focused Ion Beam (FIB) column combine with a Scanning Electron Microscope (SEM) provide full control of nanofabrication or nanolithography processes. Despite the very high technological level of the available machines, research of new ion sources allowing even higher resolution and a wider choice of atomic or molecular ions for new and demanding application is very active.

Our new system, COLDFIB, wants to take up this challenge of the nanomanufacturing by the coupling of two high technologies: the laser cooling of atoms, and manipulation of charged particles.

Very innovative, this industrial solution, based on a source of ions obtained from atoms laser cooled and ionized, will allow realizing ions beam in the unequalled performances, to reach engraving’s sizes of some nanometers. This new technology offers a resolution, for example at 5KeV, 10 times better than the LMIS one, and reaches the nanometer at 30keV.

We’ll present in this talk the last results obtained. In addition to the experimental part and performances will also show some first applications.

 12:10            Correlative workflows with x-rays, electron- and ion beams

P. Gnauck, C. Jäger, F. Perez, I. Schulmeyer, Carl Zeiss Microscopy, Oberkochen (GE)

 Many challenges in material science and life sciences require the combination of complementary techniques in order to understand complex systems. E.g.in battery research a multiscale and multimodal investigation of the sample is required. While a wide field light microscope provides 2D overview images of the anode separator cathode interface a confocal microscope is used to characterize the roughness of the separator foil. Correlating this data with electron microscopy provides high resolution surface information and chemical information. Investigation of the samples by means of x-ray microscopy provides 3D and 4D data that is used to localize interesting features inside the sample. Correlating this data with Crossbeam systems provides high resolution 3D information exactly at the area of interest. Finally the focused ion beam can be combined with SIMS in order to provide e.g. high resolution Li distribution maps of a battery sample.