Ultra-thin film structures have become increasingly important and simultaneously gained complexity with regard to the number of layers and the elemental composition. Understanding the processes occurring during deposition is crucial for improving the film quality. Especially during the first stages of film growth, analytical techniques with high surface sensitivity and good detection limits are required to study the growth process.

Low Energy Ion Scattering (LEIS) lends itself well to these tasks with its ultimate surface sensitivity of a single monolayer and detection limits of down to 10 ppm. This is accomplished by bombarding the surface with noble gas ions of a few keV and measuring the energy loss of the backscattered ions at a fixed scattering angle. The energy spectrum is converted into a mass spectrum of the elements present at the sample surface. The absence of matrix effects allows a straightforward quantification.

Besides the composition analysis of the outermost atomic layer, depth profiling is available via two distinct methods. Static depth profiling exploits the fact that ions scattered in deeper layers lose additional energy proportional to the penetration depth. As this process involves neutralization and re-ionization, intensities are lower than for the ions scattered at the surface. Thus, these ion can be distinguished, giving information about the elemental distribution in the first few nm in a non-destructive way. Alternatively, dynamic depth profiling is available by using a second, low energy sputter ion beam to erode the surface while recording surface spectra at different depths. This yields a quantitative, high depth resolution depth profile. By observing the change in the static in-depth signal during sputtering, the sputter rate can be intrinsically and continuously determined.

The unique advantages of LEIS complement established techniques like TOF-SIMS. The latter is often hampered by sputter transients at interfaces and difficult quantification especially of matrix species, but excels as far as the detection of trace elements or the gaining of chemical composition information is concerned. We applied both LEIS depth profiling modes to a number of thin film sample systems. Hereby we show the possibilities arising from each of the two modes, as well as from the combination with TOF-SIMS. Specifically, we worked on model samples relevant to the semiconductor industry (high‑k, SiGe). Some of these samples were designed for studying the response function of the in-depth signal in order to improve the understanding and to allow the application to real-world samples, e. g. to correct for varying erosion rates.

Surface treatment of polymers produces materials that exhibit a wide range of surface compositions, properties and structures. The chemical and structural properties of these novel materials can be exploited for the fabrication of devices for bio-medical and electronics applications. Additionally, the wear-resistant properties of steel can be modified by coating the surface with a diamond-like carbon (DLC) film.

The combination of a variety of complementary surface-sensitive electron spectroscopies maximises the information available to the analyst for full quantitative surface characterisation of polymer surfaces and DLC films. The silicon content of a DLC film can affect its hardness, for example, and XPS is the ideal technique for chemical quantification of the silicon. The concentration of hydrogen in a DLC film also modifies its wear properties, but XPS cannot quantify this element. It is possible, however, to detect and quantify hydrogen using Reflected Electron Energy Loss Spectroscopy (REELS). When used together, XPS and REELS can provide a total quantification for polymer surfaces and DLC films.

There are now many examples of multivariate analysis of surface-specific technique data[1,2]. These include multivariate statistical methods such as Principal Components Analysis (PCA), Partial Least Squares (PLS), or Multivariate Curve Resolution (MCR) applied to so-called “hyperspectral” mapping data, in which hundreds of channels of spectral data are collected at each pixel of a two-dimensional pixel array spanning an area of interest. The idea of trying to mathematically correlate different sets of mapping data from the same area is not new[3], and falls under the broader category of ‘image fusion’ used in conjunction with remote sensing applications[4]. However, this approach is not prevalent in the surface science literature, with the notable exception of Fulghum and Artyushkova[5,6].

There are good reasons for this, from both the experimental and modeling perspectives. This talk will discuss the challenges associated with mathematically correlating spectroscopic and mapping data from multiple surface-specific techniques. Examples from the literature and the analytical lab will be discussed.

[1]V. S. Smentkowski, J. A.Ohlhausen, P. G. Kotula, M. R. Keenan, Applied Surface Science 2004, 231, 245.

[2] M. S. Wagner, D. J. Graham, B. D. Ratner, D. G. Castner, Surface Science 2004, 570, 78.

[3] H. Hutter, M. Grasserbauer, Chemometrics and Intelligent Laboratory Systems 1994, 24, 99.

[4] C. Pohl, J. L. Van Genderen., International Journal of Remote Sensing 1998, 19, 823.

[5] K. Artyushkova, J. E. Fulghum, ‘‘XPS and Confocal Microscopy Data Fusion for Polymer Characterization,’’ talk presented at American Vacuum Society 50th International Symposium held in Baltimore, MD November 2–7, 2003.

[6] K. Artyushkova, S. Pylypenko, J. Fenton, K. Archuleta, L.Williams, J. Fulghum, Microscopy and Microanalysis 2006, 12(Suppl. 02), 1402.

Auger electron spectroscopy (AES) depth profiling has been used to characterize thin films for decades. Thin film depth profiling using x-ray photoelectron spectroscopy (XPS) has become increasingly common due to recent advances in XPS instrumentation. Often the choice of which depth -technique is best for a particular sample is not clear. In this presentation I will compare the logistics and results of AES and XPS depth profiling of insulating and metallic multicomponent thin films. Depth profiles of SiO_xC_y and CrSi_xN_y thin films acquired using both techniques will be presented. Sample preparation, analysis set-up, and analysis time will be compared. Signal to noise and interface sharpness of the resulting profiles will be compared. The ability to determine chemical state information from the acquired data will also be discussed.

Patterned forests of carbon nanotubes (CNTs) were used as a template to fabricate novel silica-based thin-layer chromatography plates (TLC). The resulting CNTs are infiltrated with elemental silicon by low pressure chemical vapor deposition of silane. Silicon coated CNTs are annealed in air to remove the CNTs and convert the silicon to silica. The resulting material is white, which is indicative and characteristic of silica. This process produces TLC plates that are porous and robust. The microfabricated TLC plates are characterized extensively by scanning electron microscopy (SEM), which shows the precise placement of the adsorbent material. Plates are also characterized by X-ray photoelectron spectroscopy, time-of-flight secondary ion mass spectrometry, and BET isotherm measurements. Unlike almost all other commercially available plates, these microfabricated structures do not require a binder to hold the adsorbent material together. Baseline separation of a CAMAG (Muttenz, Switzerland) five-component dye test mixture using toluene as the mobile phase was obtained. The chromatographic efficiencies of these microfabricated TLC plates are typically 70% higher than commercially available high-performance TLC plates, and sometimes much higher, also showing a 150% reduction in development time; these microfabricated TLC plates allow for both improved efficiency and speed of analysis.