FIB Cross-sections Research Articles

Growth defects in magnetron sputtered PVD films deposited in UHV environment

The growth defects in magnetron sputtered coatings have been well studied in industrial environment. In this work, on the other hand, the emphasis is in observing the same phenomena in a lab-scale UHV environment. TiN and CrN films were deposited at a thickness of 1–2 μm. The statistical evaluation of defect density was based on stylus profilometry scans (2 mm2 area), and consequent analysis of hillock density in dependence of threshold height. Morphology of defects was observed by SEM, while on selected defects we made FIB cross-sections. In this way we were able to measure the chemical composition of individual seeds – starting points for growth defects. In contrast to expectations, there was only a relatively minor reduction of defect density by using UHV environment.

Preparation of Pt@Cu Core-Shell Particles Incorporated in Poly-(3,4-ethylenedioxythiophene) for Catalytic Applications

Hybrid materials offer the possibility to combine advantageous properties of the individual components. In the macroscopic world adobe or reinforced concrete are well known examples. The same principle can also be applied to electrodes for electro-catalytic reactions. The usually precious metal catalyst like Pt for fuel cells (or Au for sensors) is reduced in size (and thus in costs) to smaller and smaller particles which are bonded to supporting surfaces. An interesting approach is to prepare a porous matrix in which these catalytic particles can be homogeneously incorporated. Promising candidates are conducting polymers like poly-(3,4-ethylenedioxythiophene) (PEDOT) whose structure can be tuned by the used electrolyte and/or the supporting electrolyte. To homogeneously incorporate the particles they should be prepared in situ with the polymer which is possible by cyclic voltammetry or (reversed) pulse methods. As Pt2+ has the possibility to oxidize EDOT directly we deposited Cu as an intermediate with the polymer by cyclic voltammetry. In a second step a chemical exchange reaction was performed using a Pt4+containing solution. This approach can yield Pt@Cu core shell particles embedded in a polymer matrix. A positive effect of this is also that the electronic structure of the Pt layer is modified by the underlying Cu core which can lead to enhanced catalytic activities. The cyclic voltammograms of the preparation of PEDOT/Cu films show a superposition of PEDOT and Cu features. At potentials more positive than ca. 0.9 V vs. Ag/AgCl sat. KCl polymerization of EDOT occurs while at more negative potentials the capacitive behaviour of the PEDOT film dominates next to the Cu redox waves at ca. 50 mV. SEM/EDX analyses showed that thus prepared films contain 10 to 17 wt% Cu depending on the switch-off potential. If the preparation stopped before Cu oxidation Cu crystallites could be observed on the PEDOT surface and the Cu content was higher. In the case of well oxidized films no distinct Cu centres could be found (see figure 1). The chemical exchange reaction with Pt was monitored by a quartz crystal microbalance (QCM). According to QCM data the deposited mass of Pt corresponds to ca. 60 monolayers on the geometric area. This contribution will discuss the preparation of the PEDOT/Pt@Cu hybrid films as well as the electrochemical and mechanical characterization of these films. This will include CV and impedance (EIS) data as well as QCM data. Fig. 1. SEM images of a PEDOT/Cu film with the preparation stopped at potentials negative to the Cu redox potential. (left) top view of the surface and (right) FIB cross section. Figure 1

Damage initiation and evolution in silicon nitride under non-conforming lubricated hybrid rolling contact

This study focuses on the damage mechanisms in silicon nitride rolling elements used in hybrid (ceramic–metal) bearings that operate under non-conformal contact. To get an insight into the prevailing damage modes compared to the real application, a rolling contact experiment was designed to mimic the contact conditions. Hertzian contact pressures ranged from 3.0 to 5.9GPa (500–4150N). In order to approach pure rolling, the experiments were run without inducing any gross slip. Extensive surface and subsurface damage analysis was performed using conventional ceramography as well as FIB cross-sectioning. Finite element simulations were carried out to illustrate the stress state prevailing under different loading conditions. Surface damage to rollers subjected to contact pressures up to 5.1GPa (2500N) was mainly dominated by micro-spalling, which was induced due to the presence of snowflake structures. At the highest applied loads, damage appeared as a combination of macro-cracking and micro-spalling. Crack propagation was attributed to different mechanisms: (a) fatigue-induced fracture and (b) lubricant-driven crack propagation in the subsurface.

Size and plasticity effects in zirconia micropillars compression

The micropillar compression technique has shown the potential for activating the brittle-to-ductile transition in ceramic monocrystals when testing reduced volumes. In this work, the role of size is studied by comparing the mechanical response of polycrystalline tetragonal zirconia micropillars and macroscopic specimens under compression. In micropillars, the absence of the natural defect population typical of bulk zirconia increases considerably the strength, allowing the activation of plastic deformation mechanisms and their study, showing in this way that the brittle-to-ductile transition is not limited to ceramic monocrystals only. The main mechanism of plastic deformation is transformation-induced plasticity, which is shown to be size dependent. The deformation behavior is studied in detail by loading-unloading tests at constant and increasing peak stresses, while the microstructure evolution is revealed by FIB cross-sections, TEM and STEM observations performed on lamellas extracted from pillars retrieved before failure. Finally, a failure mechanism is proposed, based on the damage induced by phase transformation.

Digital laser printing of metal/metal-oxide nano-composites with tunable electrical properties

We study the electrical properties of aluminum structures printed by the laser forward transfer of molten, femtoliter droplets in air. The resulting printed material is an aluminum/aluminum-oxide nano-composite. By controlling the printing conditions, and thereby the droplet volume, its jetting velocity and duration, it is possible to tune the electrical resistivity to a large extent. The material resistivity depends on the degree of oxidation which takes place during jetting and on the formation of electrical contact points as molten droplets impact the substrate. Evidence for these processes is provided by FIB cross sections of printed structures.

Analysis of 16 QFN Device I/O Pads for Solderability Failures

In-process failures were experienced during printed circuit board (PCB) SMT assembly of a 16 Quad Flat No Leads (QFN) device. The failures appeared to be solderability related with QFN unit I/O pads not soldering robustly and sometimes leading to QFN detachment following board mounting. When assembly did take place on affected QFN units, the resulting solder joint was observed to be weak. This paper reports on very systematic analyses of the QFN device I/O pads using optical inspections, AES surface, AES depth profiling, SEM/EDX, SIMS, FIB and TEM cross-sectional measurements to determine the root cause of the failure and the failure mechanism. The detached QFN units, suspect and good unsoldered units, passing and failing units obtained from customers were examined. The industry standard surface mount solderability testing was performed on good and suspect parts, and all were observed to pass as evidenced by >95% coverage of the I/O pads. Optical inspections and a wide variety of physical analysis of the pads on fresh parts showed no anomalies with only the expected Au over Pd over Ni found. AES analysis was performed including depth profiling to look for any issues in the NiPdAu over base Cu plating layers that could be contributing the solderability failures. The AES depth profiling indicated AuPd film on the Ni under layer for the I/O pads as expected. No unexpected elements or oxide layers were observed at any layer. Then, one failing and one passing units were compared by doing FIB cross-section, FIB planar section and TEM cross-section analysis. The cross-sectional analysis showed rough Ni surface for the failing units, while the Ni surface was relatively smooth for the passing unit. Further, finer Cu grains and Ni grains were observed on the passing units. Additionally, the lead frame fabrication process mapping showed rough Cu, Ni “texturing” and use of low electro chemical polishing (ECP) current on the bad units compared to that of the good units. All affected bad units were confirmed coming from a second source Cu supplier with the rough Cu. The weak and irregular NiSn IMC formation on the bad units caused IMC separation and possible spalling during board solder reflow primarily due to the rough base Cu and irregular grain sizes and resulting lower ECP lead frame plating current. A possible final factor was marginally low Pd thickness. In conclusion, the 16 QFN device solderability failure root cause summary and the lessons learned from a wide variety of analysis techniques will be discussed.

Coupling In Situ TEM and Ex Situ Analysis to Understand Heterogeneous Sodiation of Antimony.

We employed an in situ electrochemical cell in the transmission electron microscope (TEM) together with ex situ time-of-flight, secondary-ion mass spectrometry (TOF-SIMS) depth profiling, and FIB-helium ion scanning microscope (HIM) imaging to detail the structural and compositional changes associated with Na/Na(+) charging/discharging of 50 and 100 nm thin films of Sb. TOF-SIMS on a partially sodiated 100 nm Sb film gives a Na signal that progressively decreases toward the current collector, indicating that sodiation does not proceed uniformly. This heterogeneity will lead to local volumetric expansion gradients that would in turn serve as a major source of intrinsic stress in the microstructure. In situ TEM shows time-dependent buckling and localized separation of the sodiated films from their TiN-Ge nanowire support, which is a mechanism of stress-relaxation. Localized horizontal fracture does not occur directly at the interface, but rather at a short distance away within the bulk of the Sb. HIM images of FIB cross sections taken from sodiated half-cells, electrically disconnected, and aged at room temperature, demonstrate nonuniform film swelling and the onset of analogous through-bulk separation. TOF-SIMS highlights time-dependent segregation of Na within the structure, both to the film-current collector interface and to the film surface where a solid electrolyte interphase (SEI) exists, agreeing with the electrochemical impedance results that show time-dependent increase of the films' charge transfer resistance. We propose that Na segregation serves as a secondary source of stress relief, which occurs over somewhat longer time scales.

Surface degradation of cemented carbides in scratching contact with granite and diamond–The roles of microstructure and composition

Cemented carbides are composite materials comprising metal carbide grains in a more ductile metallic binder. This microstructure results in a combination of high hardness and toughness, making them commonly used as rock crushing parts in rock drilling, where they usually show wear on a very fine scale. The hardness and toughness can ultimately be tuned for the application by adjusting the carbide grain size, binder fraction and composition.In the present investigation, the initial micro-scale deformation and wear of polished cemented carbide surfaces is studied by micro scratching with diamond and granite styli, and also by instrumented micro and nanoindentation. The deformation and wear is evaluated on the sub-micrometer scale using high resolution FEG-SEM and FIB cross sectioning. The selected microstructures include besides four cemented carbide grades that are commonly used in rock drilling also binderless and Ni containing grades. This wider range of cemented carbides is used to gain fundamental insights into the relations between microstructure and micro-scale deformation and wear. The results are discussed with respect to their significance for wear of cemented carbides in rock drilling operations.

Fault isolation in a case study of failure analysis on Metal–Insulator–Metal capacitor structures

Metal–Insulator–Metal (MIM) capacitors are extensively used in many microelectronic products such as BICMOS for automotive radar applications, but failure analysis process is very laborious due to peculiar structure. In this paper a possible fault isolation process flow was shown in a failure analysis case study. In particular due to the fact that it is a customer return device, many checks were done in fault isolation steps before physical analyses. Finally a FIB cross section was performed on OBIRCH hot spot and the results confirmed the validity of the fault isolation process.

Erosion Mechanism of MoS2-Based Films Exposed to Atomic Oxygen Environments.

The erosion mechanism of magnetron sputtered MoS2 films exposed to the atomic oxygen environment was studied and compared with the Ti-doped MoS2 and MoS2/Ti multilayer films. The compositional and structural changes were investigated as a function of incident fluence by Rutherford back scattering (RBS) and focused ion beam combining with scanning electron microscopy (FIB&SEM). The RBS results indicate that the sulfur atoms are eroded by the incident atomic oxygen atoms and the removed sulfur amount increases but the erosion rate decreases with increasing of incident fluence. For pure MoS2 films the erosion process turns to saturate at the end of investigated fluence of 4.8×10(21) O cm(-2), and for Ti-doped and MoS2/Ti multilayer films the saturation of sulfur erosion is much earlier around incident fluence of 5.2×10(19) and 2.6×10(19) O cm(-2), respectively. FIB cross-section results reveal that pores structures present in the as-deposited MoS2 films provide a reaction highway, which allows the incident atomic oxygen to be able to reach and react with the sulfur at bottom. Introducing titanium doping or MoS2/Ti multilayer structures definitely reduce the density of pores and defects in the initial films, consequently, erosion process is suppressed or blocked, and the instinct lubricant properties of MoS2 phases can be well-retained in vacuum sliding conditions.

Imaging Structural Damage of Bacteria on Antibacterial Copper – a Scanning Electron Microscopy Study Using FIB Cross-Sectioning

Abstract The combination Scanning Electron Microscopy/Focused Ion Beam sectioning represents an established method for the three-dimensional characterization of material microstructures. This work examines whether this method of analysis is also applicable in case of a high resolution structure elucidation of bacterial cells in interaction with material surfaces. Here, the focus is on imaging the mechanism of bacteria cell damage induced by antibacterial copper materials.

Examination of the Aging Effects of Silicone Molds During Vacuum Casting Processes via Scanning Electron Microscopy

We present an in-depth investigation of contaminated silicone molds by use of scanning electron microscopy, that provides additional insight into the aging process of the molds. The molds lifetime in vacuum casting processes of polyurethane is limited due to contamination with the diisocyanate compound of the PUR resin. SEM investigations, combined with FIB cross-sections and EDX measurements show that three different types of particles are distributed in the contaminated silicone matrix. The particles differ in shape and size, with a range from 50 nm to 50 µm. This explains the loss in transparency and the hardened haptics of contaminated molds.

Boundary lubrication: Influence of the size and structure of inorganic fullerene-like MoS2 nanoparticles on friction and wear reduction

The use of Inorganic Fullerene-like (IF) nanoparticles in lubricants has proved extremely effective to reduce friction and wear under severe boundary lubrication conditions. It has furthermore been suggested that the synthesis of smaller nanoparticles containing many structural defects would benefit friction and wear reduction, as they would penetrate and exfoliate more easily in the contact, leading to the quick formation of homogeneous tribofilms. In this study, four different types of IF-MoS2 were synthesized so as to be able to differentiate the influence of both the size and the morphology of the nanoparticles on their tribological behavior. Pure-sliding, reciprocating tribological testing of these four types of nanoparticles revealed their excellent friction-reducing properties in severe boundary lubrication, with splash lubrication taking place for a high number of cycles. High wear reduction was also obtained and confirmed using optical profilometry. Although the nanoparticle structure was found to have an influence on their effectiveness in time, all the nanoparticles tested – regardless of size or crystallinity – were found to achieve the same performances as long as proper oil recirculation took place, ensuring a continuous feeding of the contact in nanoparticles. The formation of MoS2 tribofilms on the wear surfaces was confirmed using XPS analyses and observed on FIB cross sections, and their nature was discussed in the light of the associated tribological results. As the size and morphology of the IF-MoS2 did not affect their performance in the range studied, their friction reducing properties were compared to those of bulk h-MoS2 tested in the same conditions. The benefits of using spherical nanoparticles such as IF-MoS2 was clearly shown.

Low temperature FIB cross section: Application to indium micro bumps

This paper presents the interest of low temperature FIB cross section on indium micro bump. Experimental setup and results which demonstrate the interest of cooling the sample are detailed. We will explain the artefacts observed during FIB milling at room temperature. The Ga ions interact with indium to create locally an eutectic alloy, with melting point below room temperature. Inside the vacuum chamber, this eutectic alloys sublimates quickly and voids appear in the cross section. Cooling the sample with cryogenic stage enables to perform “clean” cross section without these artefacts.

(Invited) Nanocrystalline Diamond for Near Junction Heat Spreading in GaN Power HEMTs

As a wide-bandgap semiconductor, gallium nitride (GaN) is an attractive material for next-generation power devices, but capabilities have been limited by self-heating effects. Attempts to mitigate thermal impairment have been limited due to the difficulties involved with placing high thermal conductivity materials close to heat sources in the device. Heat spreading schemes have involved growth of AlGaN/GaN on single crystal or CVD diamond, or capping of fully-processed HEMTs using nanocrystalline diamond (NCD). All approaches have suffered from reduced HEMT performance or limited substrate size. Recently, we have successfully demonstrated a “gate after diamond” approach, which enables scalable, large-area diamond integration and improves the thermal budget of the process by depositing NCD before the thermally sensitive Schottky metal gate, shown in the cross-section schematic and FIB image in Figure 1 [1,2]. The diamond-capped device appeared to demonstrate improved DC I-V characteristics in most key performance areas, notably improved on-resistance, saturation current, and transconductance, and reduced off-state current and gate leakage, as summarized in Table I. In addition, diamond-capped devices demonstrated improved forward blocking characteristics, manifested by the reduced off-state leakage current and a nearly 200V improvement in breakdown voltage in forward blocking mode. RF devices have produced comparable f T, f MAX, and output power relative to a reference device. Furthermore, channel temperature measurements by Raman thermography, shown in Figure 2, indicate a 20% reduction in channel temperature. Thermal mapping has also indicated a change in the temperature profile through the source-drain region by pulling the hot spot away from the gate edge and broadening the temperature distribution. These results produce a good agreement when compared to a finite-element model. A significant drawback to this approach is that the passivation layer, however thin, acts as a thermal insulator between the hottest part of the device and the heat spreading layer. This process has recently been improved to enable the deposition of diamond directly on the GaN surface, which produced an enhanced heat spreading effect as well as reduced current collapse. In this process, the most critical step is the reliable formation of the gate opening without damaging the now unprotected GaN surface in the diamond etch, which involves a high power plasma etch in O2-based chemistry. To mitigate this effect and produce a reliable gate opening, we have developed a sacrificial gate after diamond process, where the gate “recess” is formed as a SiNXpillar before diamond growth, followed by either a realignment of the gate recess or selective diamond growth outside of the gate region. This approach is expected to improve the yield and scalability of the diamond directly on GaN process. This presentation will review progress in simulation, process integration, channel temperature mapping, and electrical characterization of diamond-capped devices.1. M.J. Tadjer, T.J. Anderson, K.D. Hobart, T.I. Feygelson, J.D. Caldwell, C.R. Eddy, Jr. F.J. Kub, J.E. Butler, B.B. Pate, J. Melngailis. IEEE Electron Dev. Lett. 33, 23-25 (2012)2. T.J. Anderson, A.D. Koehler, M.J. Tadjer, K.D. Hobart, T.I. Feygelson, J.K. Hite, B.B. Pate, C.R. Eddy, Jr, F.J. Kub. 2013 CS Mantech Technical DigestFigure 1. Schematic and FIB cross-section of AlGaN/GaN HEMT with NCD heat spreading filmTable I. Table of DC parameters for reference and diamond-capped HEMTsFigure 2. Raman thermography estimation of channel temperature with and without NCD heat spreading films

Giga-Cycle Property of a New Age-Hardened Aluminium Alloy Containing Excess Solute Magnesium

The giga-cycle property of a newly developed Al alloy, which contains 0.5wt.% excess Mg solute compared to a standard age-hardened 6061 alloy (6061-T6), was investigated by using smooth specimens subjected to ultrasonic fatigue. The fatigue strength of the new alloy was higher than that of a normal 6061 alloy particularly at relatively low stress amplitude level. Several analyses (surface crack observation, fractography, FIB cross-sectioning, etc.) were also conducted to reveal the micro-mechanism of the observed strength properties. The following results were obtained: i) No fatigue limit was confirmed for both 6061 and new alloy. ii) Total life (Nf) of 6061 and new alloys was determined by a single fatigue crack initiated from a surface PSB crack. iii) Crack initiation resistance defined by N25 (number of cycles to reach ρ = 25 mm-2, where ρ is the PSB crack number density) for new alloy was higher than that of 6061. iv) The higher fatigue strength of new alloy was explained by the effect of excess Mg solute which increased the resistance against the formation of PSB cracks.

An Evaluation of FIB Cross-Sectioning Using a Cooling Stage for Metals and Alloys with Low Melting Point

Functions of an observation and an analysis in electron microscope, such as scanning electron microscope (SEM) or transmission electron microscope (TEM) are indispensable to evaluate advanced materials. Therefore a specimen preparation technique, that is a front end of the electron microscopy, has become highly important, thus a choice of it affects a result of the evaluation. The authors was combined a cooling stage in FIB and applied it for evaluation of metals with low melting point. The electron microscopic evaluation of Lead solder, Indium, Tin and Bismuth, metals with low melting point, has been always discussed if the results represent the actual physics. Metals with low melting point are heat sensitive materials, so the comparison of cross-sectioning with room and low temperature, it can be said that low temperature cross-sectioning has less effect and keeps the actual physics of the sample. In this paper, some knowledge from comparisons of cross-sectioning with room and low temperature for metals with low melting point are reported.

Mechanical and tribological behavior of nanostructured copper–alumina cermets obtained by Pulsed Electric Current Sintering

Nanostructured Cu–Al2O3 powders obtained by the reduction of CuO with Al in a high energy ball mill were successfully consolidated by Pulsed Electric Current Sintering (PECS). The effect of the composition and microstructure of these PECS Cu–Al2O3 cermets on their strength was investigated. The friction and wear behavior of these cermets were studied under reciprocating sliding against corundum at 23°C and 50% RH, and compared to the behavior of coarse grained PECS sintered pure copper. The effect of grain size on the coefficient of friction was masked by the formation of a surface tribolayer. The wear depth recorded on Cu–Al2O3 is lesser than half the one on coarse grained copper. Surface and subsurface deformation studied through FIB cross-sections showed that delamination and oxidative wear were active on Cu and Cu–Al2O3 cermets respectively under the current sliding test conditions. PECS Cu–Al2O3 cermets showed a good thermal stability even at 600°C.

Test Structures for Characterization of Through-Silicon Vias

3-D chip stacking using through-silicon vias (TSVs) requires accurate characterization of the TSV, the thinned silicon, and the stacked dies. This paper proposes a set of test structures specifically designed to address the electrical characterization of TSV in terms of resistance, capacitance, leakage, yield, and their impact on the 2-D interconnects of the stacked dies. Examples of the use of these structures are presented, and the observed electrical behaviors are explained with the support of FIB cross-section images.

Emerging Techniques for 3-D Integrated System-in-Package Failure Diagnostics

Abstract Failure analysis is becoming increasingly difficult with the emergence of 3D integrated packages due to their complex layouts, diverse materials, shrinking dimensions, and tight fits. This article demonstrates several FA techniques, including high-frequency scanning acoustic microscopy, lock-in thermography, and FIB cross-sectioning in combination with plasma ion etching or laser ablation. Detailed case studies show how the various methods can be used to analyze bonding integrity between different materials, chip-to-chip interface structures, buried interconnect defects, and through-silicon vias at either the device or package level.