Morphology optimization of MOCVD-grown GaN nucleation layers by the multistep technique

Teemu Lang, M. Odnoblyudov, V. Bougrov, Sami Suihkonen, Markku Sopanen, Harri Lipsanen

Research output: Contribution to journalArticleScientificpeer-review

Abstract

The heteroepitaxial growth of gallium nitride (GaN) on sapphire substrates by metal-organic chemical vapor deposition is most commonly carried out using the two-step growth process. This process involves the deposition of a thin GaN nucleation layer (NL) at a temperature of approximately 450-600 °C. The morphology of this low-temperature film after annealing is known to have a crucial effect on the quality of GaN buffer layers. In this paper, we report on efficient control of the GaN NL morphology using a multistep technique. The technique is used to control the size and reduce the density of nucleation islands (NIs) on the NL in order to optimize the surface morphology for a subsequent higher temperature overgrowth step. Together with process parameter optimization a density as low as 1×107 cm-2 for the NIs is obtained. The NL morphology is analyzed by atomic force microscopy. The dislocation density of GaN buffer layers grown on multistep NLs is evaluated by etch-pit density measurements and X-ray diffraction is used to support and elaborate the results. The crystalline quality of individual NIs is studied by transmission electron microscopy. Measurements indicate that the multistep technique is successfully used to significantly reduce the threading dislocation density in GaN films. A threading dislocation density of 1.0×108 cm-2 is demonstrated with the method.
Original languageEnglish
Pages (from-to)26-32
JournalJournal of Crystal Growth
Volume292
Issue number1
DOIs
Publication statusPublished - 15 Jun 2006
MoE publication typeA1 Journal article-refereed

Keywords

  • Atomic force microscopy
  • Etch-pit density
  • Nucleation
  • Recrystallization
  • MOCVD
  • GaN

Fingerprint

Dive into the research topics of 'Morphology optimization of MOCVD-grown GaN nucleation layers by the multistep technique'. Together they form a unique fingerprint.

Cite this