ICMCTF2016 Session H1: Advanced Microstructural Characterization of Thin Films and Engineered Surfaces

The understanding of morphological and microstructural development during thin film growth is of particular relevance to control islands shape, nucleation and growth of nanoparticles, phase transformation, texture or surface roughness. Due to oversaturated vapor fluxes employed in physical vapor deposition (PVD) techniques, dynamics usually prevails over thermodynamics in dictating growth and microstructural evolution in PVD films. Depending on the material mobility, different growth modes occur, driven by kinetically limited surface diffusion processes and/or interfacial reactions. This has important implications in areas where nanoscale materials are involved, such as plasmonics, spintronics or next-generation of nano-devices.

Iron aluminides (Fe-Al) are attractive intermetallic compounds as lightweight and structural materials. From powder metallurgy routes, various oxide-dispersion-strengthened (ODS) versions of the FeAl intermetallic alloy have been developed - such as the FeAl40 Grade 3 alloy developped by CEA - to enhance the mechanical properties of the bulk material [1, 2]. However, the tribological performances of such materials remain poor and surface treatments are needed. In this context, the present communication presents some results on plasma assisted nitriding (PAN) for metallurgical surface modificationof the FeAl40 Grade 3 (ODS Fe-40at%Al) intermetallic alloy to improve the surface hardness and the associated wear resistance. Investigations are conducted to gain a deeper understanding on the mechanisms involved during nitriding of iron based aluminides, particularly those containing large amounts of Al (e.g. above 10 %).

Plasma nitriding was done using a 800 Hz pulsed d.c. plasma discharge containing a 95% N₂ / 5% H₂ gas mixture. The influences of the nitriding time (from 5 min to 48h) and temperature (from 400 to 645 °C) on the nitrided layer characteristics (thickness, morphology, element composition, crystallographic phase and hardness) were investigated. The nitrided layers were characterized at various scales (from 10 µm to 10 nm) with different analytical techniques: scanning electron microscopy (SEM), X-ray diffraction (XRD), electron-probe micro-analysis (EPMA), transmission electron microscopy (TEM) and atom probe tomography (APT).

Our results show that the growth kinetic follows a parabolic growth law as theoretically predicted by the Wagner’s theory where the mechanism of nitriding is mainly controlled by nitrogen diffusion in the a-Fe phase. It is also shown that, whatever the nitriding conditions used, the overall nitrided layers are divided into a thin outer sublayer of γ’-Fe₄N phase (~1 µm in thickness) and a thick inner sublayer of hex. AlN and α-Fe phases (~10 – 100 µm in thickness).The combination of TEM and APT evidences a “eutectoid-like” morphology within the inner sublayer where α-Fe lamellae alternate periodically with α-Fe/ AlN lamellae (~300 nm in length and ~10 nm in width). The origin and influence of the nitriding conditions on this lamellar morphology will be discussed.

[1] MA. Munoz-Morris et al. Acta Mater. 2002;50:2825-36.

[2] F. Moret et al. J. Phys. IV 1995;6:281-289.

Multilayer systems can offer an efficient way of controlling residual stress, improve adhesion and enhance toughness of coated systems. This work aims at the development of multilayer coating with improved adhesion, based on numerical design approach. The numerical model of titanium–titanium nitride (Ti–TiN) multilayer has been formulated with multi-physics FEM and analytical modeling, to find the optimal thickness of individual layers in a multilayer that can decrease interfacial axial and in-plane shear stress. These coatings configurations are experimentally produced to quantitatively evaluate the scratch adhesion, in-plane residual stresses, nanoindentation hardness and elastic-modulus. The multilayer in comparison with standard single-layer shows significant improvement (22%) in adhesion under decreased interfacial stress conditions, without any affect on overall coating stiffness and hardness. The multilayer coating in comparison with different configurations was also investigated. Result shows an increase in scratch adhesion of 18% and 27% for the optimal position and thickness of interlayers respectively. Qualitative comparison of in-plane residual stress shows higher stress in bi-layer and lower stress in multilayer with optimal thickness of interlayer. The approach in the study could be used to develop stress-optimized coatings for wear resistance applications.

Ability of Chromia-former alloys to form a thermal oxide scale allows reducing oxidation kinetics of the materials. However, durability of the metal/ceramic system depends on the mechanical integrity of this scale, and also of the scale damage such as buckling which could appear during oxidation and then cooling. These scale damages are closely correlated to the magnitude and sign of the residual stresses present in the oxide scale as well as to its microstructure i.e. grain size. In this work, an accurate determination of residual stresses in both Ni-30Cr and Fe-47Cr alloys at macroscopic scale in the adherent oxide and at local scale through damaged areas allows a comparison with models describing thin film delamination. A multi-scale approach is then proposed:

- At macroscopic scale, residual stress magnitudes are determined thanks to conventional XRD using sin²ψ method and Raman spectroscopy. The influence of metallurgical parameters on scale microstructural states and damage is studied.

- At microscopic scale, residual stress mapping through different types of delamination are performed thanks to Raman micro spectroscopy and Synchrotron micro diffraction. In this last case, the ability of switching the x-ray beam wavelength from poly to mono chromatic while keeping the beam at the same place on the sample surface allows studying alternatively the substrate and the oxide scale respectively.

Stress release modes by diffusional creep or delamination are studied and quantified. Morphological information and associated residual stresses allows confrontation to buckling mechanic and calculation of the interface toughness.

[1] M. Guerain, P. Goudeau, J. L. Grosseau-Poussard, Journal of Applied Physics 110, 093516 (2011)

[2] M. Guerain, P. Goudeau, B. Panicaud, J. L. Grosseau-Poussard, Journal of Applied Physics 113, 063502 (2013)

[3] M. Guerain, P. Goudeau, J.L. Grosseau-Poussard, Scripta Materialia 109, 15-18 (2015)

Analysis and control of residual stresses in advanced engineering materials are important issues for reliability assessment at small scales, e.g. for micro-electromechanical systems (MEMS) and nano-crystalline and amorphous bulk and thin film materials. This presentation gives an overview of the recent advances in the field of sub-micron scale residual stress assessment by the use of focused ion beam (FIB)-controlled material removal techniques.

The two step method consists of incremental FIB ring-core milling combined with high-resolution in-situ SEM-FEG imaging of the relaxing surface and a full field strain analysis by digital image correlation (DIC). The through-thickness profile of the residual stress can be also obtained by comparison of the experimentally measured surface strain with finite element modelling using Schajer’s integral method.

In this presentation, we will review the most recent advances in the field of FIB-DIC methods for residual stress assessment at the micro and nano scales, with focus on recent efforts for development of automated procedures for local residual stress analysis of (i) thin films, (ii) microelectronics devices and (iii) polycrystalline and amorphous bulk materials.

Practical applications of the method on several systems will be described and discussed. In particular, the issues of residual stress assessment on very thin films and micro-devices, stress depth profiling, stress measurement on amorphous materials and the effects of ion induced damage and elastic anisotropy on the relaxation strains will be reviewed.

One of the important questions is the limitation of the technique due to Ga implantation. In the present work we attempt to identify the zones impacted by FIB milling and the extent of Ga penetration and to determine the resolution limit of the method due to ion affected zones. Transmission Electron Microscopy (TEM) and Atom Probe Tomography (APT) are employed on a variety of materials to visualize and evaluate the Ga-induced zones. TEM is used to assess the nature and spatial distribution of the defects and to quantify the distribution of elements using Energy Dispersive Spectroscopy (EDS). A change of the structure in a Si-specimen, where the slightly thicker part of a TEM lamella was still crystalline, while the thinner part transformed to a fully amorphous state, could be observed.

For a systematic study of the Ga amount FIB has been used with different settings to implant Ga to a Si wafer and to a Si PSM (pre-sharpened mount), used for atom probe tomography. Raman spectroscopy has been performed on the Si wafer showing that no peak shift occurs and hence that there is no stress due to the implanted Ga. At the same time the peak intensity decreases with prolonged exposure time. Upon annealing the Si signal reappears, while Ga is visible as droplets on the surface. The local Ga concentration variations are analyzed as a function of ion implantation parameters. APT is used to assess the spatial distribution of Ga, after its implantation into the PSM via FIB and to visualize and quantify local concentration gradients and segregations that might be of interest.

[A] www.istress.eu

[C] Korsunsky et al., 2009, Mater. Lett., 63:1961–1963

[D] Larson et al., 1999, Ultramicroscopy, 79:287-293

[E] Thompson et al., 2007, Ultramicroscopy, 107:131–139

Two different electrodeposited copper films, ranging from 2.5 to 20µm thickness, were studied with regard to their thermo-mechanical fatigue behaviour via fast thermal cycling. Film A shows a constant grain size with varying film thickness, which is related to high incorporation of additives, whereas Film B is composed of a pure copper, leading to a film thickness dependency with respect to grain size. To study the film evolution during cycling, an advanced characterization method was implemented, combing area-specific topographical information using atomic force microscopy and microstructural information using electron backscatter diffraction. Assessing the information from the two techniques, obtained at the same area, enables to study the evolution of global parameters such as roughness, texture, average grain size and grain boundary characteristics. Furthermore, it allows the classification of the on-going fatigue mechanisms in the two film systems by examining the evolution of certain microstructural features and the mechanical properties by local nanoindentation. In addition, in-situ film stress measurements and film cross-sections were used for a more comprehensive interpretation. It was observed that starting with a similar initial microstructure, as it is the case for 5 µm film thickness, Film A shows a more stable microstructure, whereas Film B shows significant grain coarsening and greater film surface roughness, indicating that remaining additives play an important role in stabilizing grain boundaries during thermo-mechanical fatigue.

Wurtzite-type (WZ) semiconductors, including ZnO, are composed of cation and anion layers alternating along the <001> direction. This leads to the anisotropic properties of the basal surfaces of the WZ-type materials and their interfaces. [1-3] Recently, it was shown that it is possible to orient the growth of ZnO films with a chosen polarity by RF-sputtering deposition. [4] This presents a pending interest for applications related to ZnO interfaces with metals. In particular, for ZnO interfaces with Ti used either for Ohmic contacts, [5,6] or in thin film stacks in glazing industry, where the ultrathin Ti layer serves as an adhesion promoter in conducting glass. [7] Such interfaces are very sensitive to annealing. During a thermal treatment ZnO decomposes to oxidize Ti, and a ZnTiO_x compound can be formed at the interface as a result of species diffusion. [8,9]

We report on an extended study of the ZnO films RF-sputtered on silicon wafers. At an elevated temperature the growth of ZnO thin films begins with the substrate oxidation. Zn vapour pressure is high, and most of the Zn atoms re-evaporate from the wafer surface. After a while, first ZnO grains nucleate. At this point the deposition conditions determine the orientation of the film which cannot be inversed after the nucleation stage. We demonstrate that the Zn orientation can be obtained only in a narrow range of the deposition conditions, as shown in Fig. 1. The grain size and defects in the ZnO films studied by a number of techniques vary as function of the deposition temperature and substrate bias. These properties of ZnO strongly affect its reactivity with Ti.

The Hard X-Ray Photoemission Spectroscopy (HAXPES) allows studying buried interfaces in a non-destructive way. This technique provides the information about the chemical environment of the sample with a nanometre in-depth resolution. We studied by HAXPES the reactivity of the obtained ZnO thin films covered by a 4 nm thick Ti layer as deposited and after annealing at temperatures up to 550 °C. We claim that a higher deposition temperature leads to more stable ZnO films. Thus, oxygen diffusion onset temperature is increased, and Ti stays under-oxidized even after annealing at 550 °C.

References

1. N. Ohashi, et al., Jpn. JAP, 46, L1042 (2007)
2. H. Tampo, et al., APL, 87, 141904 (2005)
3. E. Barthel, et al., Thin Solid Films, 473, 272 (2005)
4. J. R. Williams, et al., APL, 103, 042107 (2013)
5. K. Ip, et al., JVST B, 21, 2378 (2003)
6. U. Ozgur, et al., JAP, 98, 041301 (2005)
7. S. Grachev, et al., Thin Solid Films, 518, 6052 (2010)
8. J. Chen, et al., JES, 153, G462 (2006)
9. R. Knut, et al., JAP, 115, 043714 (2014)

There are practical limitations to depth profiling for probing the sample chemistry beyond the surface region which include preferential sputtering and accumulated sputter beam damage. Both effects result in a distortion or complete loss of the true 3D chemical distribution of multi-layer thin films.

An alternative approach to achieve 3D chemical imaging of complex matrix chemistries is to utilize in situ FIB milling and sectioning in conjunction with TOF-SIMS chemical imaging, or 3D FIB-TOF tomography [1]. This approach has already proven successful for inorganic thin film samples such as batteries and solid oxide fuel cells. In this presentation we will present new data where we have extended this method to organic multi-layer thin films.

The present study investigated samples from two classes of materials: one metal-organic mixed matrix composition and one mixed organic phase comprised of two polymer moieties. Since there was no preferential sputtering, an immediate result of the FIB-TOF imaging was the accurate determination of the depth scale. We have collected characteristic molecular information from each sample for the purposes of 2D and 3D imaging. Cluster ion beam polishing (e.g. Ar_2,500⁺) was necessary to remove the FIB beam-induced damage, and the new instrument configuration allows cluster ion polishing to be accomplished with ease. We will highlight the instrumentation development and the methodologies which make FIB-TOF of organic thin films possible.

1. A. Wucher, G.L. Fisher and C.M. Mahoney, Three-Dimensional Imaging with Cluster Ion Beams (p. 207-246) in Cluster Secondary Ion Mass Spectrometry: Principles and Applications, C.M. Mahoney (Ed.), Wiley & Sons, N.J. (2013).

Continued device miniaturization in microelectronics calls for a fundamental understanding of diffusion processes and damage mechanisms in the Cu metallization/TiN barrier layer system. The starting point of the present study is a combined experimental and theoretical examination of lattice diffusion in ideal single-crystal TiN/Cu stacks grown on MgO(001) by unbalanced DC magnetron sputter deposition. After a 12 h annealing treatment at 1000 °C, a uniform Cu diffusion layer of 7-12 nm is observed by scanning transmission electron microscopy and atom probe tomography (APT). Density-functional theory calculations predict a stoichiometry-dependent atomic diffusion mechanism of Cu in bulk TiN, with Cu diffusing on the N sublattice for the experimental N/Ti ratio of 0.92.

These findings are extended to a comparison of grain boundary diffusion of Cu in dense polycrystalline TiN sputter-deposited on Si at 700 °C and underdense polycrystalline TiN grown on Si without external substrate heating. While the Cu diffusion path along dense TiN grain boundaries can be restricted to approximately 30 nm after a 1 h annealing treatment at 900 °C as visualized by 3D APT reconstructions, it already exceeds 500 nm after annealing at 700 °C in the underdense low-temperature TiN barrier. In this case, the formation of the Cu₃Si phase, which characteristically grows along the close-packed <101> directions in Si, is identified as the main damage mechanism leading to complete barrier failure.

To meet the low-temperature processing needs of semiconductor industry and at the same time exploit the improved performance of dense polycrystalline barrier layers, deposition of Ti_1-xTa_xN barriers on Si is demonstrated by a reactive hybrid DC/high-power impulse magnetron sputtering process, where barrier densification is achieved by pulsed irradiation of the growth surface with only a few at.% of energetic Ta ions without external substrate heating. These barrier layers delay the onset of Cu grain boundary diffusion to temperatures above 800 °C (1 h annealing time) and are therefore capable of competing with TiN barriers deposited at 700 °C.