SIMS2015 Session FS1+MS1-MoA: FIB-SIMS (2:00-3:40 pm) + MeV SIMS (4:00-5:40 pm)

In order to increase density, performance and functionality, semiconductor circuits may be integrated in 3-D (chips stacked vertically). In order to connect them, through silicon vias (TSVs) are used. These TSVs are usually made from copper and require efficient barrier layers to prevent diffusion of copper into the surrounding silicon. Full wafer samples with different barriers (varying thickness and deposition conditions for example) can be routinely characterized by SIMS to characterize their efficiency, but doing so in actual devices is challenging. This is however of great interest as the barrier quality may change significantly between planar and 3-D geometries. Furthermore the barrier thickness may vary as a function of via depth. Also the filling of the vias can only be addressed for 3-D sample geometries.

Conventional dual beam ToF-SIMS depth profiling is not adapted because of the deep profiles required (TSVs may be of the order of 80 µm deep), surface topography, as well as the shift of the analysed zone with depth. This is also the least favorable geometry for observing any lateral diffusion from the TSVs.

A protocol was developed to allow depth profiling of vias using precise wafer cleaving followed by an ex-situ plasma-FIB cut that leaves the sidewalls of the TSVs almost at the surface (covered by a thin SiO₂ layer). Dual-beam depth profiling is then used on the cleaved and FIB polished side of the sample which is placed flat in the sample holder. The analyzed field of view can then easily contain 3 or 4 complete vias and the diffusion profiles of copper analyzed at different via depths and between vias.

In order to obtain 3-D information on the filling of the copper via an in-situ Ga FIB was used to to cut the TSVs. At each successive FIB cut a ToF-SIMS image is acquired thus allowing a 3-D volume to be acquired using a slice and view approach. This approach enables voids in the TSVs to be observed and defects in the barrier layer to be identified.

By combining the two approaches both copper diffusion and voids in the TSV can be characterized.

Acknowledgements: This work was supported by the French "Recherches Technologiques de Base" Program and was performed on the Nano Characterization Platform (PFNC) of the CEA Grenoble.

In energy applications of functional materials very often thin layers are formed at the interface between two materials during operation of the device or thin interlayers are specifically deposited. The detailed characterisation of these functional layers is often very important for the elucidation of the functionality. If thicker cover layers cover these layers of interest it is not possible to perform classical SIMS depth profiling. Depending on the materials and layers properties, the preparation of cross sections can be challenging or impossible due to smearing effects or breaking of the layers. Therefore the production of a FIB crater followed by crater wall imaging can be a helpful tool for the investigation of these intermediate layers. Here we report on two examples in the context of energy related research.

Thanks to its versatility, sodium cobaltate is still investigated in manifold scientific studies. Its large Seebeck coefficient combined with good temperature stability and low toxicity makes it particularly interesting as thermoelectric material. The present study deals with inhomogeneities of the chemical composition when a temperature gradient is applied parallel to its natural layered structure – a common situation in thermoelectric systems. Highly epitaxial thin films on sapphire (001) and lanthanum aluminate (111) are prepared via pulsed laser deposition. Due to pronounced phase instabilities against atmospheric conditions, an additional aluminum oxide protective layer was inevitable for subsequent experiments. The mass spectrometric characterization of the about 1.6 µm thick intermediate sodium cobaltate layer was done by crater wall analysis. In order to expand the surface of the cross section the FIB beam was adjusted in a very flat angle of incident to artificially increase the observable depth scale approximately by a factor of ten.

The multidimensional chemical characterization of complex sample systems is of major interest in many research fields. TOF-SIMS is a powerful chemical imaging technique which provides elemental as well as comprehensive molecular information. The combination of high resolution primary ion beams with conventional high dose sputter beams also allows 3D chemical characterization of inorganic and organic samples. By choosing the appropriate sputter species this conventional dual beam approach offers the intrinsic possibility to chemically modify the analyzed surface. This allows for optimized measurement conditions e. g. higher secondary ion yields, quantitative MCs analysis, organic depth profiling.

Unfortunately, this so-called dual beam approach has its limitations for the analysis of extremely rough samples and samples with voids and material that exhibits strong local variations in density or sputter yield. Not only is the initial surface topography unknown but it is also modified and in many cases even roughened by the sputtering process. To overcome these limitations we have developed a FIB/SIMS system [1] which makes 3D tomography SIMS analysis of rough or porous samples possible. However, this new approach has its limitations as well. To learn more about the advantages and the limitations of the different approaches we analyzed different sample systems with both methods.

In this contribution we compare the results of the FIB/SIMS analysis with results of conventional dual beam depth profiling. The investigated sample systems include oxidized metal surfaces and lithium ion battery electrodes. We observed that the conventional cross section views are partially distorted and that the FIB/SIMS images reveal the original geometry and composition of the analyzed volume. Since the FIB/SIMS information can be used to correct the conventional SIMS data we consider the combination of both techniques the most comprehensive approach.

[1] F. Kollmer, W. Paul, M. Krehl, E. Niehuis, SIMS XVIII proceedings paper, Surf. Interface Anal., 2012

Li-ion batteries technology is an attractive energy storage system for many applications such as portable devices. To follow growing energy demand of new devices, their energy density needs to be improved and silicon-based anode is a serious option since it offers a specific capacity almost ten times higher than carbonaceous materials. Silicon anodes suffer from a drastic capacity fading making it unusable after few cycles. A detailed understanding of the lithiation mechanism of silicon could contribute to the design of more robust architectures.

In this work, composite silicon electrodes have been electrochemically cycled versus Li. After a controlled transfer from a glove box to a TOF-SIMS by using an air-tight vessel, they have been characterized by using an in situ Ga FIB implemented on a TOF.SIMS 5 instrument. Lithiation stage has been extensively studied, in particular during the first cycle.

During the 1^st cycle, a core-shell structure is observed, evolving with cycling. This has already been described, however, for the first time to our knowledge, thanks to the in situ FIB, it has been possible to observe the core-shell structure in Si particles throughout the entire electrode, whatever their in-depth position. Si particles close to the collector present the same structure with similar core/shell volume ratio as particles close to the surface. This demonstrates the efficiency of electrode porosity, allowing the Li to access the deeper parts of the electrode. Cross analysis by FIB-TOF-SIMS/AES has also been achieved and brings even more useful information.

After a successful study of lithiation of similar material by using AES [1], FIB-TOF-SIMS allows a better understanding of lithiation mechanisms by revealing the structure of the micrometer-sized particles, and also by providing information on Li distribution throughout the entire electrode. Such approach can be adapted to different kind of materials, and could also be considered for a multiscale analysis of Li in cycled electrodes.

[1] E. De Vito et al., J. Mater. Chem. A, 2013, 1, 4956-4965

For solid oxide fuel cell (SOFC), and especially for interconnect materials, high demands are put on electronic conductivity, thermal stability at elevated service temperatures, as well as tensile strength. After oxidation a complex mixture of different oxides forms develop on the metal surface, and often several specific elemental concentrations are low and difficult to analyze. ToF-SIMS provides the possibility to determine the lateral distribution of oxide and various elements. In addition, depth distribution using sputter depth profiling is able to simultaneously detect all elements present in a material even if the impurities in the interconnect alloy from the high-temperature-grown oxide scale are previously unknown. A unique possibility with SIMS is to the possibility to study oxidation kinetics by analyzing the distribution of stable isotopes (i.e.¹⁸O and deuterium) to gain further insight into oxide ion motion in the oxide and its relevance for longevity and performance of the material. ToF-SIMS also provides the ability to render obtained information in 3D allowing an unprecedented way of understanding oxide formation on SOFC materials.

In this study the ferritic stainless steel Crofer 22 APU, specially developed as interconnect material in SOFC, was oxidized first in an environment containing Ar and ¹⁸O₂ for one week at 850 ºc and subsequently for three further weeks in laboratory air (containing ¹⁶O₂). Two materials were oxidized simultaneous, uncoated Crofer 22 APU as well as Crofer 22 APU coated with 10 nm Ce. The oxides were analyzed using IONTOF V ToF-SIMS instruments to characterize the oxide surface at high lateral resolution followed by depth profiling using Cs⁺ and O⁺ to analyze the oxide growth and to create 3D renderings of the oxide scale. The "Collimated Burst Alignment” (CBA) mode as published by Holzlechner et. Al. was used for improved accuracy of the oxygen isotope measurements. One of the used ToF-SIMS was equipped with a Focused ion beam (FIB) which allow for in-situ preparation of trenches in the oxide materials. The walls of the trenches were analyzed with the Bi-LMIG to validate the 3D rendering results and to display the oxygen diffusion through the oxides as 2D images.

Large cluster beams were initially proposed as a tool for organic surface analysis, when Mahoney et al. used massive supersonic water/glycerol clusters to produce secondary ion mass spectra of peptides [1]. Over the years, this approach generated a range of new methods such as DESI, DICE or EDI-MS, powerful for organic molecule ionization. In SIMS, large Ar clusters, first developed for surface smoothing, have induced a breakthrough for molecular depth profiling of organic thin films [ 2,3 ] and they are also being used for molecular analysis with minimal fragmentation.

Targeting a better understanding of desorption induced by large gas clusters, this contribution follows our previous theoretical studies of Ar cluster bombardment of model polymers, using a coarse-grained representation of the molecular sample (polyethylene chains in an amorphous arrangement) [4]. Here, we report on a detailed comparison of the effects of the Ar cluster incidence angle (0° vs. 45°) and the cluster nature (CH₄ vs. Ar) on the soft sputtering of polymeric samples. In addition we compare the results of Ar cluster-induced sputtering and fragmentation at 45° incidence for 3 different molecular weights (282, 1388 and 14002 amu). The computed statistics of sputtering are compared and an excellent agreement is found with the available experimental data of Ar bombardment obtained at 45° incidence [5]. The variance of the sputtering and polymer fragmentation results with changing beam parameters is explained via the microscopic analysis of the interaction in the simulations, for instance, the energy deposition in the impact region.

[1] J. F. Mahoney, J. Perel, S. A. Ruatta, P. A. Martino, S. Husain, T. D. Lee, Rapid. Commun. Mass Spectrom. 5 (1991) 441-445.

[2] S. Ninomiya, K. Ichiki, H. Yamada, Y. Nakata, T. Seki, T. Aoki, J. Matsuo, Rapid Commun. Mass Spectrom. 23 (2009) 1601–1606.

[3] T. Mouhib, C. Poleunis, N. Wehbe, J. J. Michels, Y. Galagan, L. Houssiau, P. Bertrand , A. Delcorte, Analyst, 138 (2013) 6801–6810.

[4] A. Delcorte, V. Cristaudo, V. Lebec, B. Czerwinski, Int. J. Mass. Spectrom. 370 (2014) 29-38.

[5] M. P. Seah, J. Phys. Chem. C 117 (2013) 12622-12632.

Because of its high surface sensitivity and low sputtering rate, secondary ion mass spectrometry (SIMS) is one of the powerful tools used to obtain surface information from organic samples. However, there are two kinds of samples that are difficult to evaluate by conventional SIMS using a keV-energy ion probe: samples with large molecules that have a low molecular ion yield, and volatile molecular samples that evaporate under high vacuum. In recent years, we have developed a novel SIMS method using an MeV-energy ion probe, called MeV-SIMS. The MeV-energy ion probe enables obtaining intact large molecular ions by electronic excitation. In addition, MeV-SIMS can work under low vacuum conditions, under which the evaporation of volatile samples is suppressed, because the MeV-energy heavy ion has high transmission capability under these conditions.

In this study, a new instrument was produced for ambient analysis by using MeV-energy ion probe. This instrument enables analysis of water containing samples that have a quite high vapor pressure (~3000 Pa at 25ºC), under ambient conditions. Moreover, this novel method of ambient analysis has high surface sensitivity, and these features imply that the new method can be applied to evaluate liquid—solid interface.A solution sample of benzoic acid was measured using this method under near-ambient conditions, and the transition of surface conditions was observed as a dry process. Benzoic acid has the unique characteristic of molecular assembly. A dimer of benzoic acid is formed in the solid phase by hydrogen bonds, whereas in water the molecules of benzoic acid are almost isolated. As a result, in the solid sample spectrum, the dimer molecular ion ([2M+H]⁺) was detected with high intensity. In contrast, the solution sample spectrum showed quite a low intensity of [2M+H]⁺, and the molecular ion with water attached ([M+H₂O+H]⁺) was the main peak. Subsequently, we obtained a change of spectrum during the drying process of the solution sample. The surface condition was monitored by a CCD camera, and it was found that the solution dried completely after 10 min. During this interval, the intensity of [M+H₂O+H]⁺ increased gradually through the volatilization of water, meaning that this ion intensity was related to the concentration of benzoic acid in the water solution. Moreover, the intensity of [2M+H]⁺ increased drastically after 10 min, indicating that the sample condition changed sharply to solid at about this time. These results suggest that this novel ambient analysis method could be used to evaluate the component information of the liquid—solid interface in situ.

Since the first PDMS experiment performed by R.D. Macfarlane [1] a great amount of data has been provided thanks to the availability of high energy atomic ion beams. These allow to vary the energy density deposited near the solid surface by changing the energy and the charge state of these ions. I shall start my talk by showing results of secondary ion emission (SI) obtained with high energy atomic ions at the surface and from deeper layers. These studies demonstrate that the energy density deposited in the track is the main parameter. The question that arises is the feasibilty of the development of a MeV-SIMS instrument complementary to other methods already available as IBA techniques by using small single stage or tandem accelerators [2]. The possibility of delivering heavy ion (oxygen to copper) micro-beams permits to develop Mass Spectrometry (MS) ion imaging [3] correlated with µ-PIXE and/or µ-RBS techniques. A second promising step is SIMS imaging at pressure close to ambient [4].

The use of clusters as projectiles leads to an increase of the energy density deposit and therefore to the enhancement of SI yields. The second part of my presentation will concern cluster ion beams. More than fifteen years ago, the fundamental interest of the cluster ions was clearly shown and within a few years we had a whole set of sources and accelerator facilities to cover a large energy and mass range: from keV to MeV and clusters to molecules and nanoparticles. I will point out certain results and illustrate the cluster-solid interaction by presenting results about the modifications induced in the material by cluster impacts.

I shall synthesize the information which can be extracted from a large projectile energy and mass range. Then I shall try to deduce, from this synthesis, the advantages for the SIMS surface analysis, of the various beams which can be accelerated by small electrostatic accelerators in the energy range from 1 to 4 qMeV.

Development of the modern industry in the last century caused appearance of many new paint materials, especially synthetic organic binders and pigments, which started to be extensively used by the artists. Behavior of such materials in contact with other materials, as well as their degradation under the influence of different environmental conditions is not well understood. Therefore it is important to study them by using different chemical and physical characterization methods, especially those methods that can provide information about the molecular structure such as Gas Chromatography Mass Spectrometry (GC/MS), Pyrolysis-GC/MS, FTIR or Raman spectroscopy.

In the present work we explore for the first time, potential of the MeV Secondary Ion Mass Spectrometry (MeV SIMS) for the identification of synthetic organic materials used for paint materials. Self-made paint mock-up samples were prepared by mixing different pigment powders with alkyd and acrylic binders on glass slides. Also, samples of some commertially available paint formulas were also prepared. The mock-ups were artificially aged using enhanced conditions of UV radiation and temperature. Three sets of samples were analysed using MeV SIMS setup at the heavy ion microprobe in Zagreb [1]: unaged, aged for two (UV1) and aged for four months (UV2). Prior to the MeV SIMS measurements, different ways of sample mounting were tested due to the fact that sample holder is kept on +5 kV and homogeneous extraction field should be provided in order to get highest yield of high mass secondary molecular ions. For the measurements 5 MeV Si⁴⁺ primary ions were used. We will demonstrate that different synthetic organic pigments and binders can be easily identified by their molecular masses with MeV SIMS. Also, chemical imaging of mock-up cross sections using focused ion beam to study how deep ageing effects can be seen in the sample was tested and first results will be reported.

This work is supported by the Unity through Knowledge Fund (UKF) project: “Study of modern paint materials and their stability using MeV SIMS and other analytical techniques”.

[1] T. Tadić, I. Bogdanović Radović, Z. Siketić, D. D. Cosic, N. Skukan, M. Jakšić, J. Matsuo, Development of a TOF SIMS setup at the Zagreb heavy ion microbeam facility, Nuclear Instruments and Methods in Physics Research B 332 (2014) 234-237.

* Present address: Center for Nuclear Technologies, Technical University of Denmark, Risø Campus, DK-4000 Roskilde, Denmark