Molecular beam epitaxy of InN nanorods on Si- and C-faces of SiC substrates

1. Introduction

Group III-nitrides are now actively studied worldwide for solar cell applications. InGaN alloys have a direct band gap from 0.7 to 3.4 eV, which covers most of the visible spectrum. The growth of nanorods promises a possible improvement in the quality of the InGaN material and therefore over recent years there has been increasing interest in InGaN nanorods research for LEDs and solar energy conversion.

Nanorods of GaN, InN and InGaN have been successfully grown on different substrates including Si and sapphire by Plasma-Assisted Molecular Beam Epitaxy (PA-MBE) and Metal Organic Chemical Vapour Deposition (MOCVD). It has been established that GaN and InN nanorods can be grown by PA-MBE under strongly N-rich growth conditions.

Several groups have demonstrated the growth of GaN nanorods by PA-MBE on (0 0 0 1) sapphire. The GaN nanorods were found to be mostly free of threading dislocations. We have shown that by changing from N-rich to Ga-rich MBE conditions, the growth mode can be changed from vertical to lateral, leading to growth of a continuous GaN overlayer. As a result of growing on the top of the nanorods, many of the threading defects are eliminated, leading to coalesced GaN overlayers with average threading dislocation densities up to two orders of magnitude lower than in continuous GaN epilayers.

InN nanorods were also successfully achieved by MBE and MOCVD on Si substrates by several research groups. However, it was evident that the band alignment is not suitable for the formation of the p–n junctions between InN and Si. The conduction band of InN is found to be below the valence band for Si, and that makes this material combination not ideal for potential device applications.

The problem of the band alignment can be overcome by growing InN on SiC substrates. There is only one report so far on the growth of InN nanowires on 4H–SiCsubstrates by chemical vapor deposition method. However, to the best of our knowledge, there are no known publications on InN nanorods grown by MBE on SiC substrates.

The aim of our research was to study PA-MBE growth of InN nanorods on 6H–SiC for potential solar cell applications. We have also investigated coalescence of InN nanorods as a way to improve the quality of InN layers grown on 6H–SiC substrates.

2. Experimental details

The InN samples were grown in a custom-designed PA-MBE system. The chamber is equipped with an Oxford Applied Research CARS25 RF plasma nitrogen source and a conventional K-cell for the In. The In and N fluxes (beam equivalent pressure (BEP)) were measured by inserting a Bayard-Alpert ion gauge in the place of the substrate and the growth was monitored in-situ using Reflection High Energy Electron Diffraction (RHEED). The expected growth temperatures are below 550 °C, and therefore we were not able to use a pyrometer for the growth temperature measurements. All growth temperatures are presented here as estimated from substrate thermocouple readings. All InN films in this study were grown on 10×10 mm² 6H–SiC substrates from CREE.

SiC substrates were heated to ~700 °C for 20 min to allow outgassing and removal of any residual surface contamination, and then cooled to the desired growth temperature. After that the nitrogen flux was introduced to the growth chamber and the nitrogen plasma was initiated. The MBE growth started by simultaneous opening of the In and N shutters. The growth of the nanorods was carried out under strongly N-rich conditions for 5 h.

After growth, InN samples were analysed ex-situ using Scanning Electron Microscopy (SEM), Transmission Electron Microscopy (TEM), Convergent Beam Electron Diffraction (CBED) and Photoluminscence (PL). The existence of InN nanorods on the surface and their densities were studied by SEM using a JEOL JSM 6330F, operating at 12 kV. TEM was used to examine the structure of the InN nanorods and InN layers using a Philips EM430 with an accelerating voltage of 200 kV. CBED was utilised in order to establish the polarity of the nanorods by comparing the experimental image with a simulated image. The optical properties of the InN nanorods and layers were studied using PL in a wide temperature range.

3. Results and discussion

In order to establish the optimal temperature conditions for InN nanorod growth on the Si-face of 6H–SiC, a fixed In:N ratio was chosen under N-rich MBE conditions. We have investigated the growth of InN as function of the MBE growth temperature in the range between 350 and 450 °C. Fig. 1a shows an image of the surface of the InN layer grown at ~350 °C, which demonstrates that there are practically no InN nanorods present at low MBE growth temperatures. Fig. 1b shows an image of the surface of the InN sample grown at ~400 °C. At that temperature we observe a high density of hexagonal InN nanorods on the surface. We have a dense array of short nanorods and a lower density of longer nanorods. With the further increase of the growth temperatures up to ~450 °C (Fig. 1c), the density of the longer InN nanorods has decreased. Fig. 1 clearly demonstrates that we are able to achieve the growth of InN nanorods on 6H–SiC substrates and that the optimum growth temperatures resulting in the highest density of the InN nanorods is about 400 °C.

 Fig. 1. SEM images taken at ×60,000 for InN grown on Si-face SiC with In BEP of ~2.0×10⁻⁷ Torr at (a) ~350 °C, (b) ~400 °C and (c) ~450 °C.

Fig. 2 shows the TEM image for the hexagonal InN nanorod array grown at about 400 °C. These nanorods are predominantly free of threading defects, but some contain basal plane stacking faults. The main contrast features observed are bend contours.

 Fig. 2. Cross section TEM of InN nanorods grown on SiC at ~400 °C.

Each 6H–SiC substrate has 2 faces—Si-face and C-face. In our studies we have investigated the growth of InN nanorod on both faces of SiC.

The value of In to N flux ratio is strongly influencing the MBE growth of InN nanorods. Fig. 3 presents a set of SEM images of the InN nanorods on the Si-face 6H–SiC substrates. For a fixed N flux the In:N was varied in the range of In BEP from 1.0×10⁻⁷ Torr to 3.0×10⁻⁷ Torr at the growth temperature at ~400 °C. We have achieved MBE growth of InN nanorods on the Si-face under all In fluxes studied. As the In flux increases we have observed an increase in the lateral size of the InN nanorods. In all cases we observed two sets of rods, tall thin hexagonal nanorods growing out of smaller wider hexagonal nanorods, which we term the under-layer. At low In:N ratios (Fig. 3a), the under-layer is made up of discrete wide nanorods with thinner taller InN nanorods. At the higher In:N ratios (to Fig. 3c), there is more coalescence of the InN nanorods in the under-layer making the islands much larger. There are also fewer tall nanorods protruding. Fig. 3b has the best coverage of nanorods, so the most favourable growth conditions for InN nanorods on Si-face SiC are—an In flux BEP of ~2.0×10⁻⁷ Torr and a growth temperature ~400 °C.

Fig.   Fig.3. SEM images of the InN nanorods grown on Si-face SiC across an In flux range of BEP (a) ~1.0×10⁻⁷ Torr, (b) ~2.0×10⁻⁷ Torr, and (c) ~3.0×10⁻⁷ Torr.

MBE growth of InN was also carried out on the C-face of 6H–SiC substrates to see if there are any differences in the InN nanorod formation. Fig. 4 presents SEM images taken for the In flux range from BEP 1.0×10⁻⁷ Torr to 3.0×10⁻⁷ Torr. Fig. 4a shows a densely packed array of InN nanorods. Similar to the growth on Si-face of SiC we observed two sets of rods, tall thin hexagonal nanorods growing out of smaller wider hexagonal nanorods, which we term the under-layer. We have observed a gradual decrease in the density of the long InN nanorods with an increase in the In flux. The effect is similar to the behaviour observed for InN nanorods on Si-face substrates (see Fig. 3), but the effect is more pronounced on the C-face.

 Fig. 4. SEM images of the InN nanorods grown on C-face SiC across an In flux range of BEP (a) ~1.0×10⁻⁷ Torr, (b) ~2.0×10⁻⁷ Torr, and (c) ~3.0×10⁻⁷ Torr.

In order to achieve coalescence of InN nanorods we have established the In flux required for growth of continuous InN layers on SiC substrates at these temperatures. An In flux of BEP ~2.0×10⁻⁶ Torr at ~400 °C allows us to grow a continuous InN layer without In droplet formation on the layer surface.

InN coalesced samples were then grown on the Si-face of 6H–SiC. First we have grown InN nanorods under N-rich conditions with an In BEP of ~2.0×10⁻⁷ Torr and later we have increased the In flux to ~2.0×10⁻⁶ Torr to change the growth conditions from N-rich MBE conditions (active N flux is higher than In flux) to In-rich MBE conditions (active N flux is lower than In flux). The aim was to achieve a continuous InN layer on the top of InN nanorods and so to change from a Volmer–Weber 3D island formation growth mode to a Frank-van der Merwe 2D layer-by-layer growth mode. Fig. 5a is a SEM image of a successfully coalesced InN nanorod layer. Fig. 5b is a TEM image of the same InN layer. Fig. 5a is a SEM image of a successfully coalesced InN nanorod layer. Fig. 5b is a TEM image of the same InN layer. The images show the complete InN layer and confirm that coalescence of the rods can be achieved to form a continuous InN layer. TEM images have allowed us to see that the coalescence of the nanorods occurs right down to the substrate. In the areas where coalescence has occurred, dislocations are generated at grain boundaries. However some grains have no dislocations present and are up to 500 nm wide.

 Fig. 5. SEM (a) and TEM (b) images of coalesced InN layer grown on top of the InN nanorods.

CBED analysis was carried out on the coalesced InN layers grown on SiC. We have found that InN layers grown on the Si-face of SiC substrates have In-polarity, as shown in Fig. 6.We are currently investigating polarity of InN nanorods and layers grown on C-face.

 Fig. 6. CBED data for InN sample grown on Si-polar 6H–SiC and the CBED simulations for In-polar InN.

Fig. 7 shows the low temperature PL spectrum from an InN coalesced layer grown on the top of the InN nanorods on Si-face 6H–SiC wafer. We have observed a clear peak at ~0.7 eV. The position of the PL peak is at a slightly higher energy than in the best reported PL for InN nanorods. This is attributed perhaps to a higher density of defects in the coalesced InN layer compared to pure InN nanorods.

 Fig. 7. Low temperature (~4 K) PL spectrum from an InN coalesced layer on Si-face 6H–SiC.

The next step will be investigation of the MBE of InN nanorods on p⁺ SiC in order to produce heterostructures with p–n junctions and we will present that data in the due course.

4. Conclusions

We have investigated the MBE growth of InN nanorods on 6H–SiC substrates. We have demonstrated successful growth of InN nanorods on both the Si- and C-faces of 6H–SiC wafers. The most suitable PA-MBE growth conditions for InN nanorods were with an In flux BEP of ~2.0×10⁻⁷ Torr and a growth temperature ~400 °C. We were able to coalesce InN nanorods to form continuous InN layers. We have demonstrated that InN layers grown by MBE on Si-face 6H–SiC have In-polarity. Epitaxy of InN nanorod on p+ SiC may a have a high potential for solar energy conversion devices.

Source:Journal of Crystal Growth

If you need more information about Molecular beam epitaxy of InN nanorods on Si- and C-faces of SiC substrates, please visit our website:http://www.qualitymaterial.net , send us email at powerwaymaterial@gmail.com.