Diamond Annual Review 2019/20
34 35 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 1 9 / 2 0 D I A M O N D L I G H T S O U R C E A N N U A L R E V I E W 2 0 1 9 / 2 0 Structures and Surfaces Group Beamline I07 Tuning the optoelectronic properties of semiconductor nanowires Related publication: Balaghi L., Bussone G., Grifone R., Hübner R., Grenzer J., Ghorbani-Asl M., Krasheninnikov A.V., Schneider H., HelmM. & Dimakis E.Widely tunable GaAs bandgap via strain engineering in core/shell nanowires with large latticemismatch. Nat Commun 10 , 2793 (2019). DOI: 10.1038/s41467-019-10654-7 Publication keywords: III-V semiconductors; Core/shell nanowires; Strain engineering;Telecommunication bands T he functionality of all advanced semiconductor electronic and optoelectronic devices relies on combinations of materials with different electronic properties. As these materials have different crystal lattice parameters, it is not always possible to combine them. A team of researchers investigated using nanowires to host combinations of materials with a significant mismatch in their lattice parameters, and how this affects the fundamental semiconductor properties. They carried out their experiments on the Surface and Interface Diffraction beamline (I07). Due to the combination of the diffractometer, the operating energy range and the X-ray beam dimension, I07 offered the ideal environment for the investigation of ensembles of nanowires in grazing incidence X-ray beamgeometry. The research teams findings about the widely tunable electronic properties of gallium arsenide nanowires open up new possibilities for photonics and electronics. Different devices (lasers, photodiodes, photovoltaic cells, high electron mobility transistors, etc.) could be made of the same material system, ensuring process compatibility. Fabrication costs would be minimised, due to the absence of cross-contamination issues, and no requirement for different processing temperatures or different thermal budget limits. III-V semiconductor nanowires of high structural quality can be grown epitaxially on Si substrates despite the large lattice mismatch, paving the way to the monolithic integration of various nano-devices, such as high electron mobility transistors, lasers, light-emitting diodes, photovoltaic cells, and on- demand emitters of single photons or entangled photon pairs, in CMOS platforms. The technological importance of III-V semiconductors originates from their high electron mobility and direct bandgap 1 . The nanowire geometry also provides the advantage that core/shell heterostructures with large lattice mismatch can be grown without the appearance of misfit dislocations 2 , beyond what is possible with conventional thin film heterostructures. The lattice mismatch is accommodated by the elastic deformation of not only the shell, but also the core 3 , with the actual strain profile being dependent on the relative thicknesses of the core and the shell. In this work, it is demonstrated that GaAs/In x Ga 1-x As and GaAs/In x Al 1-x As core/shell nanowires with very large lattice mismatch (up to 4%) can be grown on Si substrates without any dislocations and morphological instabilities. In appropriately designed heterostructures, a hydrostatic tensile strain as high as 7% can be achieved in the GaAs core, leading to a dramatic narrowing of its bandgap by up to 40%. These findings render GaAs nanowires a versatile material system for photonic devices across the near-infrared range, including telecom photonics at 1.3 μm, with the additional possibility of monolithic integration in Si-CMOS chips. Vertical GaAs/In x Ga 1−x As and GaAs/In x Al 1−x As core/shell nanowires were grown on Si(111) substrates by molecular beam epitaxy 4 . The GaAs core was 20 – 25 nm in diameter and 2 μm in length. Conformal growth of the shells around theGaAscoreswasachievedundergrowthconditionsthatensuredlimitedsurface diffusivity of the indium adatoms along the nanowire sidewalls. The thickness and the composition of the ternary shell were varied systematically in order to investigate their effect on the strain. Transmission electron microscopy (TEM) showed that the shell adopted the crystal structure of the core, i.e. zinc-blende structure with rotational twins only at the two ends of the nanowires, without any misfit dislocations. An example of GaAs/In 0.20 Ga 0.80 As nanowires is shown in the scanning electron microscopy (SEM) image in Fig. 1a. The corresponding compositional map perpendicular to the axis of one nanowire, as measured by energy-dispersive X-ray spectroscopy (EDXS), is shown in Fig. 1b. The hydrostatic strain in the GaAs core and the In 0.20 Ga 0.80 As shell was first measured by micro- Raman scattering spectroscopy (λ=532nm) at 300 K on single nanowires. The results (Fig. 1c) showed that the compressive strain in the shell decreased with increasing the shell thickness and became almost equal to zero for shells thicker than 40 nm. On the other hand, the tensile strain in the core increased with increasing the shell thickness and saturated at 3.2% for shells thicker than 40 nm. This indicatesthatthestrain isgraduallytransferredfromtheshelltothecorewith increasing the shell thickness. The lattice parameters of the core and the shell along the three orthogonal crystallographic directions x, y, z (z-axis is parallel to the nanowire axis, whereas x- and y- axes are perpendicular to it) and the corresponding strain components ε xx , ε yy and ε zz were measured using X-ray diffraction (XRD) at Diamond Light Source.The Surface and Interface Diffraction beamline (I07) offered an ideal non- destructiveexperimentalconfigurationforthemeasurementofthein-planestrain state of as-grown nanowire ensembles in non-coplanar grazing incidence X-ray geometry (GID). With a few hundreds of µm large X-ray beam, reciprocal space maps of the in-plane (20-2) and (22-4) reflections were collected in GID geometry at an energy of 9 keV, using a 100K Pilatus detector. In addition, a constant He fluxwaskeptaroundthesamplewithinaKapton®dometo limitpossibleradiation damage.AccordingtothepenetrationdepthprofileofX-raysforthematerialunder investigation, an incident angle of 0.2° close to the critical angle of total external reflection ensured a depth sensitivity of only a fewnanometers below the surface. This reduced the diffracted contribution significantly from the growth substrate. It is worthmentioning that the final beamsize in the vertical direction corresponded to a several-mm-long footprint of the X-ray beam impinging on the substrate surface. This strongly influenced the resulting resolution in reciprocal space. As an example,Fig.2ashowsatwo-dimensionalreciprocalspacemapofreflection(22-4) forasamplewith10nmshellthickness.Here,threesignalsatdifferentQ [11-2] values are visible along the vertical direction; they are attributed, from top to bottom, to theGaAscore,the In 0.20 Ga 0.80 Asshell,andaplanarpolycrystalline In 0.20 Ga 0.80 As layer that grew unintentionally on the Si substrate. From the position of the diffracted signals, the lattice parameters in x and y directions were determined. The lattice parameter in z direction was determined frommeasurements of the out-of-plane (-1-1-1) reflection that were performed at the High Resolution X-ray Diffraction beamline P08 at the PETRA III synchrotron in Hamburg. The XRD results (Fig. 2b) show that the GaAs crystal expands in all three dimensions with increasing the In 0.20 Ga 0.80 As shell thickness up to 10 nm (the XRD signal from the core could not be resolved for thicker shells). Furthermore, the shell becomes almost strain-free when its thickness is at least 40 nm. The strain components ε xx , ε yy and ε zz in the core are calculated as the relative change of the corresponding lattice parameters (i.e. ε xx =Δα x c /α x c , etc). Finally, the hydrostatic strain in the core ΔV/V=ε xx +ε yy +ε zz (star symbols in Fig. 1c) was found to be in good agreement with the Raman scatteringmeasurements. The strain in the GaAs core became even higher when 40–80 nm thick In x Ga 1-x As or In x Al 1-x As shells with higher x were employed (Fig. 3a). This had a tremendous effect on the bandgap of GaAs, which decreased by up to 40% (Fig. 3b), as measured by photoluminescence (PL) spectroscopy, making it possible to reach the 1.3 μm (O-) telecomband. References: 1. Vurgaftman I. et al. Band parameters for III-V compound semiconductors and their alloys. J. Appl. Phys. 89 , 5815 (2001). DOI: 10.1063/1.1368156 2. Glas F., Strain in nanowires and nanowire heterostructures. Semiconductors and semimetals 93 , 79-123 (2015). DOI: 10.1016/bs.semsem.2015.09.004 3. Grönqvist J., Strain in semiconductor core-shell nanowires. J. Appl. Phys. 106, 053508 (2009) . DOI: 10.1063/1.3207838 4. TauchnitzT., Decoupling the two roles of Ga droplets in the self-catalyzed growth of GaAs nanowires on SiOx/Si(111) substrates, Cryst. Growth Des. 17 , 5276-5282 (2017). DOI: 10.1021/acs.cgd.7b00797 Funding acknowledgement: We are grateful to Diamond Light Source for the allocated beamtime SI15923 and financial support, as well as to the German Synchrotron DESY, PETRA III for the allocated beamtime I-20160337. Finally, we are thankful for the support by the Structural Characterization Facilities Rossendorf at Ion BeamCenter, the computational support from the HZDR Computing Cluster, and the funding ofTEM Talos by the German Federal Ministry of Education of Research (BMBF), Grant No. 03SF0451 in the framework of HEMCP. Corresponding authors: Dr Genziana Bussone, Deutsches Elektronen-Synchrotron (DESY), genziana. firstname.lastname@example.org and Dr Emmanouil Dimakis, Institute of Ion BeamPhysics and Materials Research, Helmholtz-ZentrumDresden-Rossendorf, email@example.com Figure 1: Morphology, composition, and strain of GaAs/In 0.20 Ga 0.80 As core/shell nanowires. (a) Side-view SEM image of as-grown nanowires with a shell thickness of 40 nm. (b) EDXS compositional map perpendicular to the axis of one nanowire from the sample shown in (a). (c) Hydrostatic strain measured by Raman scattering spectroscopy in the core (blue data points) and the shell (red data points) as a function of the shell thickness. The star symbols correspond to XRD results. Figure 2: Measurements of the lattice parameters in GaAs/In 0.20 Ga 0.80 As core/shell nanowires as a function of shell thickness by high-resolution XRD. The measurements were performed on ensembles of as-grown nanowires. (a) An example of a 2D reciprocal space map of the (22-4) reflection for nanowires with a shell thickness of 10 nm. The contributions from the core, the shell and the planar polycrystalline layer are indicated. (b) XRD-measured average lattice parameters α of the core (blue data points) and the shell (red data points) along the three orthogonal directions x, y, z (z parallel to nanowire axis) as a function of the shell thickness. Figure 3: Strain and bandgap of GaAs/In x Ga 1-x As and GaAs/In x Al 1-x As core/shell nanowires. (a) Hydrostatic strain measured by Raman scattering spectroscopy in the core (blue data points) and the shell (red data points) as a function of the In x Ga 1-x As (closed symbols) or In x Al 1-x As (open symbols) shell composition x and the corresponding core/shell misfit f. (b) Bandgap energy of the strained GaAs core in In x Al 1-x As core/shell nanowires (measured by PL at 300 K) as a function of the shell composition x. The blue dashed line is a linear fit. The bandgaps of strain-free III-As ternary alloys are shown for comparison.
Made with FlippingBook