Salvatore Grasso oral presentation (SN1-Mon2-3-5)
Increasing Nickel and Phosphorus secondary ion yield in Copper in ToF-SIMS depth profile
1 STMicroelectronics, Via C. Olivetti N2, 20864 Agrate, Italy
2 Laboratorio MDM IMM-CNR, Via C. Olivetti N2, 20864 Agrate, Italy
Secondary Ion Mass Spectrometry (SIMS) is widely used to get compositional depth profiles of elements in a specific matrix. Depth profile gives insights on many physical and chemical quantities, such as element concentration and position, useful information to understand element diffusion mechanism and eventually other physical effects that could occur during a thermal treatment, a voltage bias application or different processes.
Detection Limit (D.L.) mainly depends on the element ionization rate in a specific matrix. To increase ionization rate different methods can be exploited. A possible approach is the use Oxygen flux to create molecular fragments (or clusters) with the element to be analyzed. In this way, the formed cluster could have an increased ionization probability with respect to the value of the originating elemental species.
In this work, we studied the dramatic change in D.L. when using Oxygen flux in either Ni or P depth profiles in Cu matrix. By using a ToF-SIMS instrument, we could reveal the existence of an on/off mechanism that allows detecting both P and Ni only in the presence of Oxygen in the main chamber during the analysis.
Several artifacts happen during the analysis with Oxygen flux. For instance, both POx and NiOx negative secondary ions have a high noise level; moreover PO normalized profile is different depending on the run of analysis.
Interestingly, we observed that a slight change of either the flood gun parameters for charge effect neutralization, or the Oxygen pressure in the analysis chamber, or the geometry of the analysis beams, dramatically affects POx and NiOx collected counts while Cu intensity profile (i.e. the matrix element) remains at the same level. This means that cluster formation and cluster ionization rates show a significant dependence on analysis setting other than on the elemental ionization rate. Our hypothesis is that the control of the parameters during the analysis is not enough accurate to precisely sustain the process of clustering formation and collection at the sample surface. Even if these effects reduce reproducibility, diffusion mechanism can be well recognized as well.