SIMS2015 Session DI1-TuA: Depth Profiling/Inorganics

SIMS analysis of arrays of repeatable structures can provide quantitative compositional data for nm-size transistor features. 1.5D SIMS methodology includes the formation of periodic (2D and 3D) arrays with features of interest, topography planarization (cap layers), SIMS depth profiling and data quantification. 1.5D refers to ”dimensionality” of the measurements – only averaged lateral concentration can directly be measured, but not lateral distribution. Thus, this method provides additional information beyond conventional 1D depth profiles, but insufficient for 2D (or even 3D) compositional mapping. For known geometries and dimensions, 1.5D SIMS data can be accurately quantified with conventional compositional standards. On the other hand, if features have uniform and known compositions, features size and geometry can be accurately reconstructed (compositional microscopy).

The 1.5D SIMS method is capable of providing data for a large (~1E6) statistical population of transistors. The method is invariant of feature pitch or dimensions – pitch reduction only increases the intensity of SIMS signals, so 1.5D SIMS method can support development of future transistor scaling. Limitations of the method include degradation of depth resolution due to residual topography and ion yield artifacts related to lateral cascade mixing involving non-homogenous materials. A variety of SIMS analytical conditions and instrumental platforms were tested to minimize artifacts. Low energy (<1kV) O2+or Cs+ sputtering beams at normal incidence enables sufficient resolution and accuracy of the data.

Demonstration of the analytical capability of 1.5D SIMS method will include analysis of conformal Fin doping by ion implantation, composition of SiGe source/drains, contamination and doping of Cu lines.

The CAMECA NanoSIMS 50L is a unique ion microprobe designed for trace element and isotope analysis at sub-micron resolution. Standard operating conditions for Cs primary ion bombardment are 8 keV acceleration of the primary ion beam and -8 keV sample potential for acceleration of the negative secondary ion species resulting in a net impact energy of 16 keV. This combination provides the ability to achieve high spatial resolution, high mass resolution and high secondary ion transmission. However, these primary beam impact conditions also result in significant sputter knock-on effects and the depth resolution obtainable in depth profile applications.

In this work, the standard operating condition of 16 keV net Cs beam impact energy of the NanoSIMS 50L was reduced to 8 keV, 4 keV and 2 keV. In addition to examining the improvement in depth resolution versus impact energy, data will be presented on the effects on mass resolution, transmission, obtainable Cs beam current and beam spot size.

Oxygen-induced segregation during depth profiling was first demonstrated by Williams and Baker [1] as a dramatic increase (factor of ~5) in the decay length of the mixing tail of a silver delta layer in silicon sputtered with Ar⁺ when oxygen gas was blown onto the surface. In contrast, sputtering with O₂⁺ showed essentially no change in the mixing tail. The authors attributed the oxygen backfill effect to antisegregation of Ag at the surface due to the steep oxygen chemical gradient under backfill conditions, in contrast to the almost flat near-surface oxygen profile produced by O₂⁺ sputtering. However, subsequent observers of this effect explained it instead in terms of formation of a bulk oxide and exclusion of the impurity to the subsurface oxide-silicon interface, as can be observed in thermal oxidation of silicon. The non-observation of any segregation effect under O₂⁺ sputtering was rationalized by arguing that the O₂⁺ sputtering in [1] did not create a stoichiometric oxide layer in the silicon target. We have revisited this issue using implanted ¹⁸O as an internal standard to monitor the surface oxygen level quantitatively during sputtering. The results show no evidence that SiO₂ stoichiometry at the surface, alone, can produce the observed segregation effects. However, oxygen sputtering, to SiO₂ stoichiometry, combined withoxygen backfill does cause segregation. Alternatively, oxygen sputtering at near-normal impact angles, which produces an excess of adsorbed oxygen at the oxide surface and thus a steep gradient, similarly elongates impurity decay lengths. Species with high oxygen affinities such as Ca and Mg show either no antisegregation or even positive segregation -- a shortening of the decay length under oxygen backfill conditions.

[1] "Implantation and ion beam mixing in thin film analysis", Peter Williams, Judith Baker, Nucl. Instrum. Methods 182/183 (1981) 15-24

Hydrogen is the most abundant element in the universe but cannot be detected by most commonly available analytical techniques. Hydrogen is known to be important for passivation of silicon interfaces and knowledge of the amount and distribution of hydrogen is of significant interest for many technologies. SIMS can provide analysis for hydrogen and the hydrogen isotopes, deuterium and tritium. It is of interest to compare the detection limit that can be routinely obtained for various SIMS instrument configurations under typical operating conditions. Time of flight, magnetic sector, and quadrupole analyzers were used to analyze hydrogen ion implanted silicon and aluminum substrates and deuterium implanted silicon.

Good vacuum is known to improve hydrogen detection limit.¹ Vacuum conditions can be optimized by methods such as overnight pumping of samples and sample holder heating. Adsorption of hydrogen from the vacuum environment (H₂, H₂O) can be minimized with use of high sputtering rate, which may be achieved using either higher impact energy and/or reduced raster size.^2,3 The species monitored is also important and may be atomic or molecular, such as H^- or Cs₂H⁺. The latter species provides a practical means for H-profiling in dielectric (i.e. insulating) films in magnetic sector instruments with conventional charge compensation.

The detection limits varied for the different analyzers used. The superior vacuum on the quadrupole instrument used has an impact on the results.

¹ C. W. Magee, J. Vac. Sci. Technol. A1, 901 (1983)

² R. G. Wilson, F. A. Stevie, C. W. Magee, Secondary Ion Mass Spectrometry, Wiley, New York (1989) p. 2.8-1

³ A. L. Pivovarov. F. A. Stevie, D. P. Griffis, G. M. Guryanov, J. Vac. Sci. Technol. A21, 1649 (2003)

Ion exchange is recognized as an integral mechanism influencing the corrosion of nuclear waste glasses. However, due to the formation of alteration layers in aqueous conditions, it is difficult to conclusively deconvolute the mechanism of ion exchange from other processes. Therefore, alkali diffusion was isolated by conducting a series of experiments in which glass coupons were contacted with a solution of ⁶LiCl dissolved in DMSO at temperatures ranging from 25 to 125 °C for various durations with several high-level waste simulant glass. ToF-SIMS using argon cluster ion sputtering source is one of the superior methods to perform the depth profiling of insulating materials, such as glasses, to get diffusion profiles [1] . The results show that Li from the non-aqueous solution exchanges with both Li and Na from the glass. An interdiffusion model was developed to model the alkali depth profiles measured using data obtained with ToF-SIMS and to derive diffusion coefficients and their activation energies. The diffusion coefficients were found to vary significantly with depth due to changes in alkali concentrations and structural changes that occur during the exchange process.

[1] Z. Wang, B. Liu, E. Zhao, et al., J. Am. Soc. Mass Spectrom., 1 (2015)

[2] This work was supported by the Department of Energy, Office of Nuclear Energy. This research was performed using EMSL, a national scientific user facility sponsored by the Department of Energy's Office of Biological and Environmental Research located at Pacific Northwest National Laboratory.

Abstract

Combining 2D imaging in lateral with sputtering in depth, layered SIMS imaging can be reconstructed in 3D. Such 3D-SIMS technique become a useful and popular tool to generate full-view images in 3D space for various specimens with layered or complex structure from inorganic, organic and biological research. It provides a wealth of composition information and gives deep insights that cannot easily be attained in other analysis. Normal approach in 3D-SIMS can displays the species distribution in layered or complex structure. However, detailed analysis of interfaces between phases, such as concentration change and inhomogeneity of diffused species near interface, attract more attention of researchers that need to be explored.

In the case of protection Zircaloy from severe oxidation and mechanical degradation during overheating, ceramic can be coated on Zircaloy to effectively improve the accident-tolerance of the nuclear system. Diffused interface is expected to be a sign of solid boning for the safety of service life. With the help of 3D-SIMS base on the data processing from Tof-SIMS, both depth profiling and 3D-imaging distributions of different species are displayed. Besides that, new 3D imaging approach is developed to separate the diffused species through the interface. In terms of the ion imaging of CsAl+ in diffusion region, diffused Al at the interface between the ceramic coating and Zircaloy substrate were investigated in both as-deposited and annealed states. Diffused Al with concentration even lower than 2% (relative to it in the bulk ceramic) can be observed in the diffusion region of 300 nm thick. The new imaging toolkit to resolve diffused species is a key evaluation to develop coating layer on alloys.

3D-SIMS is expected to provide diffusion information between coating and substrate with new view. It will continue to advance and gives more information about the formation of diffusion interface which facilitate the interface analysis from variety of research interest.

References

[1] Erin H. Seeley et al. Anal. Chem., 2012, 84 (5), 2105–2110

[2] James Bailey et al. ACS Appl. Mater. Interfaces 2015, 7, 2654−2659

[3] Alam et al. Nucl. Eng. Des. 241(2011)3658-3677

[4] Pint et al. J. Nucl. Mater. 440(2013)420-427

[5] Houssiau et al. Appl. Surf. Sci. 231-232(2004)585-589