In the last years, much attention has been paid to the quantitative measurement of the HAADF STEM intensity wherefrom thickness and/or sample composition values may be obtained. The use of averaged values of this quantity allowing a useful filtering of unwanted signals, led to the determination of doping profiles in ultra shallow junctions in Si at subnanometer resolution (1). In general, there are still some open issues concerning the origin of the contrast in HAADF STEM images that mainly regard the dependence of the intensity on sample thickness and orientation. One of these issue concerns the suggestion (2) that for thin specimen, out of an axial sample orientation, elastic scattering is enhanced at the expense of thermal diffuse scattering. In this work, we focus on two main problems: i) we test this suggestion by seeking its maximum thickness limit and ii) we investigate the precise sample orientation conditions leading to maximize the Z-contrast contribution. The former problem is studied by a comparison of the experimentally determined average normalized intensity depth profiles with those obtained by Monte Carlo simulations. In MC simulations, scattering events are completely independent so that elastic scattering is temperature independent. Moreover, as in these simulations each single electron trajectory is followed, the HAADF detector response may be easily introduced. The second problem is faced by a comparison with Multislice simulations in which full account of sample orientation and thermal diffuse scattering is taken (3). We do this