Search

The effect of P or B predoping on As diffusion in polycrystalline Si was investigated by atom probe tomography. In all samples, a high concentration of As was found at grain boundaries, indicating that such boundaries are the main diffusion path. However, As grain-boundary diffusion was suppressed in the B-doped sample and enhanced in the P-doped sample. In a sample codoped with both P and B, As diffusion was somewhat enhanced, indicating competition between the effects of the two dopants. The results suggest that As grain-boundary diffusion can be controlled by varying the local concentration of P or B.

Three-dimensional dopant distributions in actual n- and p-channel metal-oxide-semiconductor devices of 65 nm node in two kinds of commercially available products were investigated by atom probe tomography (APT). Detailed and quantitative dopant distributions in gate, gate oxide, and channel regions were successfully obtained by APT. In particular, similarities as well as differences in the dopant distributions of these two devices, which were made by different fabrication processes, were clarified in detail by estimating the dopant concentrations in the grain, at the grain boundary, and at the interface of the polycrystalline Si gate individually.

Three-dimensional dopant distributions in actual n- and p-channel metal-oxide-semiconductor devices of 65 nm node in two kinds of commercially available products were investigated by atom probe tomography (APT). Detailed and quantitative dopant distributions in gate, gate oxide, and channel regions were successfully obtained by APT. In particular, similarities as well as differences in the dopant distributions of these two devices, which were made by different fabrication processes, were clarified in detail by estimating the dopant concentrations in the grain, at the grain boundary, and at the interface of the polycrystalline Si gate individually.

Laser-assisted atom probe tomography (APT) was successfully applied to analyze the elemental distributions in a high-k/metal gate stack of a commercially available 45 nm node real device. APT revealed the multilayer structure of the high-k/metal gate stack with nearly atomic-scale resolution, and successfully detected small amounts of Zr in the thin layer of high-kHfO2dielectrics and H in the Ti layer of the metal gate. The present results demonstrate the usefulness of APT as a tool of elemental analysis in nanoscale multilayer device structures.

Laser-assisted atom probe tomography (APT) was successfully applied to analyze the elemental distributions in a high-k/metal gate stack of a commercially available 45 nm node real device. APT revealed the multilayer structure of the high-k/metal gate stack with nearly atomic-scale resolution, and successfully detected small amounts of Zr in the thin layer of high-k HfO2 dielectrics and H in the Ti layer of the metal gate. The present results demonstrate the usefulness of APT as a tool of elemental analysis in nanoscale multilayer device structures.

Randomness of channel dopant distribution in metal-oxide-semiconductor field-effect transistor (MOSFET) structures was analyzed by laser-assisted atom probe tomography. Three-dimensional dopant distributions of boron and arsenic atoms in MOSFET channels which define the threshold voltage were obtained with nearly atomic scale resolution. In order to achieve highly statistical precision, about 130 million atoms were detected in the channel region. We found that the distribution of boron atoms was consistent with a random solid solution in a matrix. Meanwhile, the arsenic atom distribution is slightly deviated from the random distribution, implying the possibility of arsenic cluster formation.

Randomness of channel dopant distribution in metal-oxide-semiconductor field-effect transistor (MOSFET) structures was analyzed by laser-assisted atom probe tomography. Three-dimensional dopant distributions of boron and arsenic atoms in MOSFET channels which define the threshold voltage were obtained with nearly atomic scale resolution. In order to achieve highly statistical precision, about 130 million atoms were detected in the channel region. We found that the distribution of boron atoms was consistent with a random solid solution in a matrix. Meanwhile, the arsenic atom distribution is slightly deviated from the random distribution, implying the possibility of arsenic cluster formation.