Tuning the orientation of the top-facets of GaN nanowires in molecular beam epitaxy by thermal decomposition

T. Auzelle, G. Calabrese, and S. Fernández-Garrido

Phys. Rev. Materials 3, 013402 – Published 9 January 2019

Abstract

In this work, we demonstrate the possibility of controlling the orientation of the top-facets of GaN nanowires grown by plasma-assisted molecular beam epitaxy by means of an in situ thermal treatment at temperatures in the 800–910 $^{\circ} C$ range, at which the material decomposes with a non-negligible rate. Depending on whether the process is carried out under vacuum or active nitrogen exposure, the nanowires can develop either ${1 \bar{1} 0 \bar{2}}$ or ${1 \bar{1} 0 \bar{8}}$ semipolar top-facets. The shape transformation is reversible; namely, the original $(000 \bar{1})$ polar facets of as-grown nanowires can be recovered after further GaN growth. Using reflection high-energy electron diffraction, the reshaping process is monitored in vivo by tracking the formation and subsequent evolution of chevrons in the diffraction pattern. This analysis reveals that the reshaping takes place after a certain delay time that can last up to several tens of minutes. It also evidences that the shape transformation is not abrupt; instead there is a continuous evolution from high- to low-index semipolar facets. We observe energy barriers higher than 3 eV along the reshaping path toward the ${1 \bar{1} 0 \bar{2}}$ top-facets. The formation of ${1 \bar{1} 0 \bar{8}}$ top-facets is assigned to an intermediate energy minimum along the reshaping path where the system freezes in the presence of active nitrogen. A change in the Ga chemical potential at the nanowire tip is proposed as the trigger for the reshaping, a process that could be either thermodynamically or kinetically driven. The formation of semipolar facets after thermal decomposition is also observed for $GaN (000 \bar{1})$ layers, demonstrating that the reshaping process is not related to the peculiar nanowire morphology. This unprecedented ability to form semipolar facets at the tip of N-polar GaN nanowires may pave the way for the fabrication of semipolar axial nanowire heterostructures along the $[000 \bar{1}]$ direction.

Received 24 September 2018

DOI:https://doi.org/10.1103/PhysRevMaterials.3.013402

Physics Subject Headings (PhySH)

Crystal forms

Nanowires

Annealing Molecular beam epitaxy

Condensed Matter, Materials & Applied Physics

Authors & Affiliations

T. Auzelle^1,*, G. Calabrese¹, and S. Fernández-Garrido^1,2,†

¹Paul-Drude-Institut für Festkörperelektronik, Leibniz-Institut im Forschungsverbund Berlin e. V., Hausvogteiplatz 5–7, 10117 Berlin, Germany
²Grupo de Electrónica y Semiconductores, Departamento de Física Aplicada, Universidad Autónoma de Madrid, Calle Francisco Tomás y Valiente 7, 28049 Madrid, Spain

^*auzelle@pdi-berlin.de
^†sergio.fernandezg@uam.es

Article Text (Subscription Required)

Click to Expand

References (Subscription Required)

Click to Expand

Issue

Vol. 3, Iss. 1 — January 2019

Reuse & Permissions

Access Options

Author publication services for translation and copyediting assistance advertisement

Images

Figure 1
(a)–(f) Angular RHEED intensity profiles around selected GaN diffraction spots of patterns recorded along the $[1 \bar{1} 20]$ and $[1 \bar{1} 00]$ azimuths (as indicated at the top of the figures) of (a) and (b) as-grown, (c) and (d) N-exposed, and (e) and (f) vacuum-exposed nanowire ensembles. The raw pictures of the corresponding diffraction spots are shown in the respective insets. The 0 deg direction is parallel to the [0001] direction in the GaN reciprocal space (perpendicular to the shadow edge and vertical in the images) and the rotation direction is counterclockwise, as sketched in the inset of (a). GaN planes, for which the observed chevrons are perpendicular, are indicated on the graphs by dashed lines (the indexes for those facets are shown at the top). (g)–(i) Schematic of the average nanowire top-facets and their projection on the $(11 \bar{2} 0)$ and $(1 \bar{1} 00)$ azimuthal planes as deduced from the chevron orientation of (g) as-grown, (h) N-exposed, and (i) vacuum-exposed nanowire ensembles. Arrows in the azimuthal planes are normal to the labeled planes.
Reuse & Permissions
Figure 2
Histograms of the nanowire top-facet angles projected on the $(1 \bar{1} 00)$ plane for the (a) as-grown, (b) N-exposed. and (c) vacuum-exposed nanowire ensembles. For each ensemble, we analyze about 100 nanowires. The arrows indicate, for a selection of GaN facets, the theoretical projected facet angles. The insets (i)–(vi) are $[1 \bar{1} 00]$ side-view scanning electron micrographs of typical nanowires from the corresponding ensembles. The extraction of the projected top-facet angles from the scanning electron micrographs is exemplified in (v). The scale bars in (i)–(vi) represent 50 nm.
Reuse & Permissions
Figure 3
(a) Intensity map illustrating the temporal evolution of angular RHEED intensity profiles recorded along the $[11 \bar{2} 0]$ azimuth after various Ga and N shutter sequences for two different experiments labeled as i and ii. The opening/closing events of the Ga and N shutters are shown by arrows at the top of the time line. The nanowire ensembles are different in i and ii. (b) Average chevron angles extracted from the N-exposure step highlighted by a red rectangle in (a) as a function of the time (the Ga shutter is shuttered at $t = 0$ s). (c) Logarithmic plot of the delay time observed prior to the reshaping of GaN nanowires under N exposure as a function of the GaN growth duration prior to shuttering the Ga cell. (d) Arrhenius representation of the temperature dependence of the characteristic nanowire reshaping time toward ${1 \bar{1} 0 \bar{8}}$ top-facets measured for N exposure and toward ${1 \bar{1} 0 \bar{8}}$ and ${1 \bar{1} 0 \bar{2}}$ top-facets measured for vacuum exposure. The dotted lines are fits yielding an activation energy of $3.3 \pm 0.3$ and $6.6 \pm 0.6$ eV for the respective formation of ${1 \bar{1} 0 \bar{8}}$ and ${1 \bar{1} 0 \bar{2}}$ top-facets. (e) Average chevron angles for two N-exposure steps performed at a substrate temperature of 860 and 890 $^{\circ} C$ for the same nanowire ensemble. The time origin is set once the reshaping sets in. The N exposure at 890 $^{\circ} C$ is interrupted after 250 s. This is the reason why no more data points are shown beyond this time for the experiment performed at 890 $^{\circ} C$ .
Reuse & Permissions
Figure 4
(a), (b) GaN decomposition rate measured by QMS during the annealing of GaN nanowire ensembles at different temperatures. For each curve, the vertical dashed lines delimit the time frame for which the substrate is maintained at the indicated temperature. The dotted line in (a) illustrates a monoexponential decay as reported by Zettler et al. [26] and the gray area depicts the transient state. The polygons in (a) refer to additional nanowire ensembles annealed at 880 $^{\circ} C$ for shorter times. (c), (d) Histograms of the nanowire top-facet angles projected on the $(1 \bar{1} 00)$ plane for ensembles annealed at (c) 880 $^{\circ} C$ for different durations, and (d) different temperatures for $11 \pm 1$ min. To construct each histogram, we analyze about $\sim 60$ nanowires per sample. The insets show side-view scanning electron micrographs of typical nanowires from the respective ensembles. The scale bars in the insets of (c) and (d) represent 50 nm. (e) Average nanowire length of ensembles annealed at 880 $^{\circ} C$ as a function of the annealing time.
Reuse & Permissions
Figure 5
(a), (b) Angular RHEED intensity profiles around selected GaN diffraction spots recorded along the (a) $[11 \bar{2} 0]$ and (b) $[1 \bar{1} 00]$ azimuths upon the onset of the decomposition process of a $GaN (000 \bar{1})$ layer above 700 $^{\circ} C$ . (c) Bird's eye view scanning electron micrograph of a $GaN (000 \bar{1})$ layer thermally decomposed above 700 $^{\circ} C$ for 180 min. (d), (e) Angular RHEED intensity profiles around selected GaN diffraction spots recorded along the (d) $[11 \bar{2} 0]$ and (e) $[1 \bar{1} 00]$ azimuths after the N-rich growth of a $GaN (000 \bar{1})$ layer. (f) Corresponding bird's eye view scanning electron micrograph acquired after the N-rich growth process.
Reuse & Permissions

Physical Review Materials