**1. Introduction**

Thin film deposition techniques such as atomic layer deposition (ALD), for three-dimensional microscopic structures and the development of large-scale integrated structures like trenches and cavities that are of importance, especially in the semiconductor industry, introduce a true challenge for thin film characterization. These 3D-micro-structured substrates have typically vertical high aspect ratio (VHAR) structures. The study and characterization of specific regions of interest like local defects, doping concentration, conformality, and interfaces, rely predominantly on cross-sectional sample preparation and characterization by electron microscopy or X-ray techniques. This approach faces the specific di fficulties of single lamella preparation of each trench, low spatial resolution to characterize interfacial di ffusion, or lack of detection of light elements in transmission electron microscopy (TEM) and cleavage plane inaccuracy for scanning electron microscopy (SEM).

To overcome the di fficulties of the abovementioned characterization techniques, we used time-of-flight secondary ion mass spectrometry (ToF-SIMS) on HAR structure analysis. The SIMS technique is a powerful tool which provides information about a given material from both its surface, with an overall sensitivity of parts per million (ppm), and bulk, with a sensitivity of part per billion (ppb) [1,2]. The in-depth analysis of a thin film's chemical composition, as well as interfacial characterization, yields information about the elemental distribution in the range of a few monolayers

(1–3 monolayers). Furthermore, by using advanced liquid metal ion guns (LMIG), one can reach a lateral resolution of about 100 nm for the elemental detection (ToF-SIMS imaging) [3–5]. In addition to both remarkable in depth and lateral resolution, numerous parameters can affect the ToF-SIMS measurement technique that make data acquisition and interpretation challenging. For instance, the achievable spatial resolution in ToF-SIMS imaging is a function of the sample matrix, material concentrations, surface geometry, primary ion intensity, instrument transmission, and spot size of the primary ion beam [3,6,7].

In this paper, the first part of the study focuses on a non-copper-plated area at the bottom of large through-silicon-vias (TSVs). The TSV interconnects provide the shortest electrical pathway, lower power consumption, lower noise, smaller form factor, and yield better performance and more functionality in comparison with the conventional chip multi-layers stacking (CMLS) [8,9]. The most common TSV metallization stack is composed of a copper (Cu) diffusion barrier and a seed layer followed by the bottom-up being electroplated [10]. One of the main difficulties in the characterization of TSV metallization is the elemental analysis of thin barrier/seed layers through the depth of TSVs before and after Cu electroplating. The thin film characterization in the range of nanometers deep down into TSVs (within the range of hundred-micrometers) make it crucial to determine conformality, step coverage, purity and corrosion. With ToF-SIMS, we are able to check the corrosion under the influence of the Cu electrolyte [11] and map detailed elemental composition information onto the TSV wall.

In the second part, as an illustration, we choose to monitor the uniformity of silicon dopant concentration in the HfO2 thin films deposition as atomic layer. The recent discovery of ferroelectric properties in this binary oxide material [12], commonly established as a dielectric for high-k metal gate technology (HKMG), handover a vast quantity of applications beyond the standard dielectric material. These applications span from non-volatile memories [13,14], steep slope devices [15] energy storage applications [16,17] and infrared sensors [18]. In this part of study, we focus on improving the analysis of the HAR structures by bypassing the obstacles of TSV characterization by ToF-SIMS and using LHAR test structures. ToF-SIMS in dual-beam mode (dynamic mode) is a well-established technique due to its high detection sensitivity for concentration ratio measurements of the dopant materials [19,20]. However, the analysis of these VHAR structures for quantifying dopant concentration is hardly possible. To solve this issue, we analyzed layer composition in a different system while being consistent with the results of the initial structures. In this system, instead of using the vertically oriented structures, lateral high aspect ratio (LHAR) [21–23] structures are used which make it possible to analyze them in the form of 2D-structures.
