Letter pubs.acs.org/NanoLett
Three-fold Symmetric Doping Mechanism in GaAs Nanowires M. H. T. Dastjerdi,† E. M. Fiordaliso,‡ E. D. Leshchenko,§ A. Akhtari-Zavareh,† T. Kasama,‡ M. Aagesen,∥,⊥ V. G. Dubrovskii,§,#,∇ and R. R. LaPierre*,†,§
Nano Lett. 2017.17:5875-5882. Downloaded from pubs.acs.org by UNIV OF NEW ENGLAND on 01/15/19. For personal use only.
†
Department of Engineering Physics, Centre for Emerging Device Technologies, McMaster University, Hamilton, Ontario Canada, L8S 4L7 ‡ Center for Electron Nanoscopy, Technical University of Denmark, DK-2800 Kongens Lyngby, Denmark § ITMO University, Kronverkskiy pr. 49, 197101 St. Petersburg, Russia ∥ Gasp Solar ApS, Gregersensvej 7, DK-2630 Taastrup, Denmark ⊥ Center for Quantum Devices, Niels Bohr Institute, University of Copenhagen, Universitetsparken 5, 2100 Copenhagen, Denmark # St. Petersburg Academic University, Khlopina 8/3, 194021 St. Petersburg, Russia ∇ Ioffe Physical Technical Institute of the Russian Academy of Sciences, Politekhnicheskaya 26, 194021 St. Petersburg, Russia S Supporting Information *
ABSTRACT: A new dopant incorporation mechanism in Ga-assisted GaAs nanowires grown by molecular beam epitaxy is reported. Off-axis electron holography revealed that p-type Be dopants introduced in situ during molecular beam epitaxy growth of the nanowires were distributed inhomogeneously in the nanowire cross-section, perpendicular to the growth direction. The active dopants showed a remarkable azimuthal distribution along the (111)B flat top of the nanowires, which is attributed to preferred incorporation along 3-fold symmetric truncated facets under the Ga droplet. A diffusion model is presented to explain the unique radial and azimuthal variation of the active dopants in the GaAs nanowires. KEYWORDS: GaAs, gallium arsenide, nanowires, self-assisted, molecular beam epitaxy, doping, beryllium
S
effect,4−6 terahertz photoconductivity,7 four-point probe resistivity,8 field effect transistors,9 off-axis electron holography,10−13 Raman spectroscopy,14−16 X-ray photoemission spectroscopy,17 secondary ion mass spectrometry,18 Kelvin probe force microscopy (KPFM),19,20 capacitance−voltage measurements,21 and atom probe tomography (APT).22−24 A review on the doping of semiconductor NWs is available in ref 25. Among these techniques, off-axis electron holography is perhaps one of the most powerful methods for dopant assessment in nanostructures.10−13 Electron holography is a transmission electron microscopy (TEM) technique that measures a spatially resolved phase difference, Δφ, by interference between electrons that pass through the specimen (object wave) and electrons that pass through vacuum (reference wave). Δφ is related to the crystal potential, or mean inner potential (MIP) V(x,y,z), according to
emiconductor nanowires (NWs) are structures with length typically on the order of microns and diameter on the order of tens to hundreds of nanometers. The III−V NWs impart various benefits to optoelectronic and electronic devices, as compared to their thin film counterparts, including the ability to accommodate large lattice mismatch strain that enables new heterostructure material combinations and monolithic growth on Si.1,2 Group III−V semiconductor NWs are therefore being investigated for various device applications including solar cells, photodetectors, and sensors.1 The most common method of NW growth involves the vapor−liquid−solid (VLS) method.1,2 In VLS growth, a seed particle such as Au, or a metallic element of the NW itself, acts as a collector of the growth species during material deposition (for example, during molecular beam epitaxy (MBE) or metalorganic vapor phase epitaxy (MOVPE)). The seed particle becomes supersaturated in the growth species, resulting in localized nucleation of the NWs. One of the critical aspects of NW growth is the intentional incorporation of impurity dopants, which is critical for the functionality of optoelectronic and electronic devices, especially for the formation of p−n junctions. However, the controlled incorporation of impurity dopants into NWs is not straightforward due to the unique growth mechanism of NWs. The doping in semiconductor NWs has been assessed by a variety of techniques, including photoluminescence,3 Hall © 2017 American Chemical Society
t
Δφ = C E ∫ V (x , y , z)dz 0
(1)
where CE is a microscope acceleration voltage-dependent constant and t is the specimen thickness. Received: February 23, 2017 Revised: September 2, 2017 Published: September 13, 2017 5875
DOI: 10.1021/acs.nanolett.7b00794 Nano Lett. 2017, 17, 5875−5882
Letter
Nano Letters
with a 5 min Be-doped p-GaP segment which initiated Gaassisted NW growth with an 80% yield of vertical NWs, averaged across the entire NW array. A subsequent p-GaAsP segment served to smoothly grade the NW composition from GaP to GaAs over a growth duration of 5 min. A p-GaAs NW core was then grown for a duration of 40 min. The latter segments were all grown with a nominal p-doping of 5 × 1018 cm−3, a V/III flux ratio of 2, and a Ga impingement rate of 0.125 μm/h. Nominal doping levels were based on GaAs thin film doping calibrations on (100) GaAs substrates using standard Hall effect measurements. After the p-GaAs NW core growth, the Ga droplet was consumed by lowering the As2 flux for 20 min while at constant substrate temperature of 600 °C and V/III flux ratio of 0.4, resulting in complete elimination of the Ga droplet.28,30 The substrate was then cooled to 420 °C, where subsequent growth occurred via vapor−solid deposition on the NW sidewalls. A nominally undoped GaAs shell (i-GaAs) was grown for 17 min, followed by a Te-doped n-GaAs shell (n = 5 × 1018 cm−3) for 12.5 min and a further Te-doped n-GaAs shell (n = 8 × 1018 cm−3) for 3.5 min. For shell growth, the V/III flux ratio was 2 and the Ga impingement rate was 1 μm/h. After growth, a JEOL-7000F SEM was used to characterize the as-grown NW array. Cross-sectional specimens for off-axis electron holography were prepared perpendicular to the growth direction from a region approximately 500 nm from the top of the NWs using the in situ lift-out technique in an FEI Helios dual beam focused ion beam (FIB)/SEM equipped with a micromanipulator. A second specimen was obtained at a depth of 700 nm from the top of the NWs to compare with the one obtained closer to the top. A layer of Pt (thicker than 300 nm) was deposited by electron beam deposition on the surface of the NWs prior to depositing a further protective Pt layer using FIB, in order to protect the surface of NWs from FIB-induced damage. At the final step of specimen preparation, the specimen surfaces were polished by a 2 kV Ga ion beam to minimize the damage caused by the 30 kV FIB. Off-axis electron holograms were acquired at 120 kV in an FEI Titan 80-300ST field emission gun TEM, equipped with a rotatable Möllenstedt biprism. The specimen was slightly tilted away (less than 2°) from the zone axis to minimize the diffraction contrast, and a low electron beam current density of 10 pA/μm2 was used to minimize the electron beam effect.31 The thickness and polarity of the specimens were measured in TEM by the convergent beam electron diffraction (CBED) technique. CBED for polarity determination was acquired from the FIB cross-section used for the holography measurements, oriented in the [110] direction, following ref 32. The NW structure was characterized in detail elsewhere.28 The p-core of the NWs exhibited hexagonal {1̅10} sidewall facets and had a diameter of 70 nm, as obtained from SEM and scanning TEM (STEM) measurements of separate p-core growths (Figure 1a,b). The amount of lateral growth of the pGaAs core was determined by the introduction of periodic GaAs0.8P0.2 marker layers during a separate MBE growth and their observation by high-angle annular dark-field (HAADF) STEM imaging, as described elsewhere28,30 and shown in Figure 1b. All marker layers from the NW base to the top extended across the entire diameter of the p-GaAs core, indicating negligible radial growth on the NW sidewalls under our growth conditions. As described earlier, lateral (shell) growth of the i-GaAs and n-GaAs regions over the p-core of the NWs was promoted by
For example, electron holography has been used to assess the doping in axial p−n junctions in Au-assisted VLS Si NWs10,11 and GaAs NWs.12 Similarly, the doping in Au-assisted GaP core−shell NW p−n junctions has been assessed by electron holography.13 Electron holographic tomography has provided three-dimensional mapping of electrostatic potentials of Auassisted GaAs/AlGaAs core−shell structures.26,27 Until recently,28 electron holography has not been applied to NWs grown by the self-assisted VLS process, and the application of holography to III−V NWs in general is rather limited. In ref 28, the low open-circuit voltage in a Ga-assisted GaAs core−shell p−i−n NW photovoltaic device was attributed to a low built-in potential as measured by electron holography. In principle, the built-in potential, Vbi, in a p−n junction can be determined from the phase contrast in the absence of other charges and related to the electrically active dopant concentrations according to the well-known expression29 Vbi =
⎛ kT ⎞ ⎛ NA −ND+ ⎞ ⎜ ⎟ln⎜ ⎟ ⎝ e ⎠ ⎝ ni 2 ⎠
(2)
where kT/e is the thermal voltage, NA− is the ionized acceptor concentration, ND+ is the ionized dopant concentration, and ni is the intrinsic carrier concentration (ni = 2.1 × 106 cm−3 for GaAs). In the present paper, electron holography was used to further assess the p-type Be dopant incorporation in Ga-assisted GaAs NW arrays grown by gas source molecular beam epitaxy (GSMBE). We hypothesize that the Be dopant incorporation is limited by diffusion from 3-fold symmetric facets of the NWs, resulting in a remarkable radial and azimuthal distribution of active dopants. NWs were grown via the Ga-assisted VLS method using a GS-MBE system, where group V hydride sources are cracked to supply group V dimers (As2 and P2). The NW growth is similar to that reported previously for a core−shell p−i−n photovoltaic device.28 A 300 μm thick, boron-doped Si(111) substrate with
