Ions and X-rays: Unique partners for accurate stopping power determination of ¹⁵N ions in Si for hydrogen depth profiling

In many areas of material sciences hydrogen analysis is of particular importance. For example, hydrogen is most abundant as impurity in thin-film materials - depending on the deposition process - and has great influence on the chemical, physical and electrical properties of many materials. Therefore, it is necessary to monitor the H-concentration by depth profiling. Best-suited methods for depth-resolved hydrogen analysis are ion beam techniques such as elastic recoil detection analysis (ERDA) and nuclear reaction analysis (NRA). In principle, both methods can be performed as primary - reference material free - methods.
The most common method, NRA, makes use of the 6385 keV resonance of the 1H(15N,αγ)12C nuclear reaction. The correct quantification of the depth scale in the measured hydrogen profiles essentially relies on accurate stopping power values, i.e. any imperfection in the stopping power values is influencing all H-values provided by NRA. For the determination of the accurate stopping power of ~6.4 MeV 15N ions in hydrogen-containing amorphous Si-layers (a-Si:H), we have, therefore, combined NRA with X-ray reflectometry (XRR), also a primary method.
The samples are prepared by magnetron sputtering of a-Si in an Ar/H2-atmosphere on a Cr-layer, which is needed as contrast material for XRR. The energy loss in the layers is measured by NRA at FZD. The layer thickness, density and roughness are determined by XRR using synchrotron radiation. XRR measurements were performed at the electron storage ring BESSY at the hard X-ray beamline BAMline. The beam was monochromatised to 10 keV using a Si [111]-double-crystal monochromator. The reflected photons of the θ-2θ-scans from the 6-circle goniometer are counted by a scintillation detector and a photodiode, respectively. Data analysis is performed by the IMD 4.1 software package.
The unique combination of results from NRA and XRR allows the accurate calculation of the mass stopping power independent of the density of the material. Our preliminary results show significant discrepancies in the order of 15-20% to the commonly used stopping powers calculated by the well accepted SRIM program (version 2006). These discrepancies have to be considered, if transforming the energy loss scale to a depth scale in hydrogen profiles. Thus, further work for confirmation is absolutely needed.
Acknowledgments: The help of D. Grambole and S. Merchel (FZD) is greatly appreciated.