Characterization of Hafnia Powder Prepared from an Oxychloride Sol–Gel

Abstract

Hafnium‐containing compounds are of great importance to the semiconductor industry as a high‐κ gate dielectric to replace silicon oxynitrides. Here, the crystallization processes and chemistry of bulk hafnia powders are investigated, which will aid in interpretation of reactions and crystallization events occurring in thin films used as gate dielectrics. Amorphous hafnia powder was prepared via a sol–gel route using the precursor HfOCl2·H2O. The powders were subjected to various heat treatments and analyzed using X‐ray diffraction and thermal analysis techniques. A large change in the crystallization pathway was found to occur when the sample was heated in an inert environment compared with air. Instead of the expected monoclinic phase, tetragonal hafnia also formed under these conditions and was observed up to temperatures of ∼760°C. The tetragonal particles eventually transform into monoclinic hafnia on further heating. Possible mechanisms for the crystallization of tetragonal hafnia are discussed. It is proposed that, in an inert environment, tetragonal hafnia is stabilized due to the presence of oxygen vacancies, formed by the reduction of HfIV to HfIII. As the temperature increases the crystal grows until there are too few oxygen vacancies left in the structure to continue stabilizing the tetragonal phase, and hence transformation to monoclinic hafnia occurs.

Investigation of Crystallization Processes from Hafnium Silicate Powders Prepared from an Oxychloride Sol‐Gel

“…In our previous study, looking at HfO 2 prepared from hafnium oxychloride precursor, it was found that the HfO 2 precursor crystallized into m ‐HfO 2 as expected upon heating. However, after heating in vacuum or in an inert environment, a different crystallization route was taken where tetragonal HfO 2 ( t ‐HfO 2 ) formed simultaneously with m ‐HfO 2 at temperatures below 1000°C, before transforming into m ‐HfO 2 as the crystallite size increased.…”

“…This difference in hydration state resulted in slight changes to the crystallization temperatures obtained for different phases observed when heating sample prepared from the two precursors. For this reason and for consistency and comparison with previous study, all the samples used in this study have been prepared with the Fluorochem precursor.…”

“…No chlorine was detected after the onset of crystallization and so no further comment is made on it in this study. For more detailed discussion see previous study …”

“…To this, one mole equivalent of tetraethoxyorthosilane (TEOS, Si(OC 2 H 5 ) 4 , Sigma‐Aldrich, Gillingham, UK, 98% solution) was added and the solution was left stirring at room temperature until the viscosity increased. It was then left to dry at 45°C–50°C for 2–7 d. Three amorphous precursor samples were made: Hf 0.3 Si 0.7 O 2 , Hf 0.5 Si 0.5 O 2 , and Hf 0.7 Si 0.3 O 2 in addition to the previously reported HfO 2 sample . This study will mainly focus on the Hf 0.5 Si 0.5 O 2 sample with reference to results from the other compositions.…”

“…A number of different experiments were carried out to look at the crystallization of the sample. These have been described elsewhere and only a summary is provided herein.…”