Group III-Nitride (III-N) semiconductors have been investigated extensively over the past decades due to their applications as electronic and optoelectronic devices. In addition, they are promising for the realization of high efficiency, multi-junction solar cells since their band-gaps vary from 0.7 eV in InN through to 3.4 eV in GaN up to 6.2 eV in AlN; thereby, allowing the band gaps of the ternaries In _x Ga_1-x N and Al _x Ga_1-x N to be tailored in between by varying x. Nanowires solar cells (NWSCs) are also receiving increasing attention but so far they have been fabricated from Si and metal-oxide (MO) NWs. Nitride NWs such as InN 1 , GaN 2 and AlN 3 are, therefore, promising for the realization of full-spectrum third generation NWSCs. However, their growth and properties must be understood beforehand in order to make nanoscale devices. So far we have grown InN 1 and GaN NWs 2 using the direct reaction of In or Ga with NH₃, while more recently we showed that Ga₂O₃ NWs may be converted to GaN by post-growth nitridation using NH₃ and H₂ 4 . Here, we have undertaken a systematic investigation into the conversion of In₂O₃ to InN NWs, which has not been carried out previously by others, thereby complementing our earlier work on the conversion of Ga₂O₃ to GaN NWs.

Therefore, we have grown straight In₂O₃ NWs with diameters of 50 nm and a high yield and uniformity. We find that the post-growth nitridation of In₂O₃ NWs using NH₃ leads to the elimination of the NWs above 600°C. The In₂O₃ NWs are preserved for temperatures less than 700°C but are not converted into InN even after long nitridation times of 6 h. However, the nitridation process was enhanced significantly via the use of H₂ or by employing a two-step temperature nitridation process, which also lead to a suppression of the photoluminescence (PL) peak at 550 nm similar to the nitridation of Ga₂O₃ NWs 4 .

Initially In₂O₃ NWs were grown using an atmospheric pressure chemical vapour deposition (APCVD) reactor described elsewhere 5 . For the growth of In₂O₃ NWs, 0.2 g of fine In powder (Aldrich, Cyprus, Mesh 100, 99.99%) was weighed and loaded in a quartz boat, while square pieces of n ⁺ Si(001) ≈ 7 mm × 7 mm, coated with ≈1.0 nm of Au, were loaded at various distances from the In. The Au layer was deposited via sputtering using Ar under a pressure of ≈10^-2 mBar. The boat was positioned directly above the thermocouple used to measure the heater temperature at the centre of the 1" quartz tube (QT). Another quartz boat with ≈5 ml of de-ionised (DI) H₂O was positioned at the inlet of the tube. After loading the boats at room temperature (RT), Ar (99.999%) was introduced at a flow rate of 500 standard cubic centimetres per minute (sccm) for 10 min. Following this, the temperature was ramped to 850°C under a flow of 50 sccm Ar using a ramp rate of 30°C/min. Upon reaching the growth temperature (T _G), the flow of Ar was maintained at 50 sccm for 30 min in order to grow the In₂O₃ NWs after which the reactor was allowed to cool down in a flow of 50 sccm of Ar for at least 30 min. The sample was always removed only when the temperature was lower than 100°C.

The nitridation of the In₂O₃ NWs was carried out in a new 1" QT without any solid precursors. After loading each sample with In₂O₃ NWs from the downstream side, a flow of 500 sccm Ar was introduced for 10 min after which the temperature was ramped to the nitridation temperature (T _N) under a flow of NH₃ that varied between 125 and 250 sccm using a ramp rate of 30°C/min. Upon reaching T _N, the same flow of NH₃ was maintained for various times between 1 and 6 h after which the reactor was allowed to cool down to RT under the same flow of NH₃. A list of the different temperatures, nitridation times and NH₃ gas flows used for the nitridation of the In₂O₃ NWs are shown in Table 1. Similarly nitridation was carried out using NH₃ and H₂. In this case, the temperature was ramped to 500°C under a flow of NH₃ and H₂ whose relative flows varied using a ramp rate of 30°C/min. Upon reaching T _N, the same flow of NH₃ and H₂ was maintained for 1 h. The total flow of NH₃ and H₂ was kept constant at 200 sccm and a list of the different flows of H₂ is listed in Table 1. Finally, we carried out a two-step temperature process. In this case, the temperature was ramped to 500°C under 125 sccm of NH₃ using a ramp rate of 10°C/min. Upon reaching T _N, the same flow of NH₃ was maintained for 1 h. Then, the temperature was ramped to 700°C and the same flow of NH₃ was maintained for 30 min after which the reactor was allowed to cool down to RT.

The morphology of the as grown In₂O₃ NWs and those treated with NH₃ were examined with a TESCAN scanning electron microscope (SEM), while their crystal structure and phase purity were investigated using a SHIMADZU, X-ray diffraction (XRD-6000), with Cu-Ka source, by performing a scan of θ - 2θ in the range between 10° and 80°. Finally, PL measurements were carried using above bandgap (approx. 3.75 eV 6 ) excitation at 267 nm. The pulse excitation was the second harmonic of a beam from an optical parametric amplifier pumped with a mode-locked TiSapphire laser. The pulses were 100 fs FWHM at a repetition rate of 250 kHz. The energy per pulse incident on the samples was 40 pJ over a spot of 2 mm in diameter.

Previously, we obtained In₂O₃ NWs by dry oxidation at 700°C 7 . A high yield of In₂O₃ NWs with an average diameter of ≈100 nm and lengths of ≈1 μm was obtained on Si(111) and quartz. However, these In₂O₃ NWs were slightly tapered; their diameters were larger and lengths were shorter compared to the In₂O₃ NWs obtained here by wet oxidation. Moreover, the distribution of the In₂O₃ NWs obtained by wet oxidation was far superior and much more uniform compared to those obtained by dry oxidation. A typical image of In₂O₃ NWs that were obtained at T _G = 850°C by wet oxidation is shown in Figure 1. It should be pointed out that a high yield and uniform distribution of In₂O₃ NWs extending over 1 cm² was obtained when the distance between the In and the Au/n⁺Si (001) was ≥15 mm, which led to a light blue-like deposit. The In₂O₃ NWs have diameters of ≈50 nm, lengths ≥2 μm and exhibited clear peaks in the XRD as shown in Figure 2 by the top curve, corresponding to the body centred cubic (bcc) crystal structure of In₂O₃ with a = 10.12 Å, in agreement with Dai et al. who obtained twisted In₂O₃ NWs by wet oxidation 8 . The In₂O₃ NWs shown in Figure 1 are straight 9 10 and in our case In₂O₃ NWs grow by a simple chemical route involving the following reaction: 2In + 3H₂O → In₂O₃ + 3H₂ 8 . Wet oxidation is a facile method and generally occurs faster than dry oxidation. No NWs were obtained on plain Si(001), suggesting the growth of In₂O₃ NWs occurs via the vapour-liquid-solid (VLS) mechanism with Au acting as the catalyst. In this case, Au NPs absorb In until they become supersaturated after which In₂O₃ NW growth commences via the reaction of In with H₂O as outlined above.

The PL spectrum following excitation at 267 nm at 300 K consisted of two broad peaks, centred at 400 and 550 nm as shown in Figure 3 Similar peaks in the PL have been observed by Yan et al. 11 who obtained a broad luminescence band centred at 395 nm from In₂O₃ nanorods, Liang et al. 12 who found a peak at 470 nm from In₂O₃ nanofibres and Wu et al. 13 who observed two distinct peaks at 416 and 435 nm from In₂O₃ nanowires. It is important to point out that these peaks are commonly attributed to the presence of oxygen vacancies.

Next, we will describe the conversion of In₂O₃ NWs into InN and in particular consider the nitridation of In₂O₃ NWs at different temperatures. To begin with In₂O₃ NWs were subjected to 250 sccm of NH₃ for 1 h at various temperatures between 500 and 900°C as listed in Table 1.

The XRD spectra of the In₂O₃ NWs treated at different temperatures is shown in Figure 2. As can be seen most of the oxide peaks disappear at temperatures >600°C. However, a new peak appears, which corresponds to the (101) crystallographic direction of InN 1 . Furthermore, SEM images reveal that the In₂O₃ NWs have been eliminated above 600°C, but a thin layer of InN remains on the Si(001). Evidently, the nitridation of the In₂O₃ NWs is destructive above 600°C due to the fast decomposition of In₂O₃ to In₂O, which is a gas. We should also point out that in addition to the temperature we also varied the nitridation time. In particular, we carried out nitridations of In₂O₃ NWs at 500 and 600°C under a flow of 125 sccm NH₃ for different times as described in Table 1.

Again the conversion of In₂O₃ NWs to InN appears to be incomplete as can be clearly seen from the XRD spectra in Figure 4 where one can observe the presence of In₂O₃ peaks and just one peak at (101) corresponding to InN. In order to achieve the efficient conversion of In₂O₃ NWs to InN without eliminating them, we used two different approaches. In the first one, we have carried out post-growth nitridation, which included H₂ as shown in Table 1 and in the second approach, we have utilised a two-step temperature nitridation process. The corresponding XRD spectra are shown in Figure 5. As can be seen from the XRD spectra, H₂ plays a significant role in the removal of the oxygen and thus all major oxide peaks are eliminated and the conversion to InN is achieved with 40% H₂. As already described above, NH₃ alone does not promote the efficient conversion of In₂O₃ NWs into InN at temperatures between 500 and 600°C. This is likely due to the formation of an InN shell around the In₂O₃, which prevents the diffusion of N into the In₂O₃ core. However, H₂ appears to promote the conversion of In₂O₃into InN 14 .

In addition, the two-step process lead to the effective conversion of In₂O₃ NWs to InN using just NH₃. In this case, the temperature was ramped at 10°C/min up to 500°C and held constant over a period of 1 h, after which the temperature was ramped again slowly to 700°C in order to promote the nitridation. Recall that the In₂O₃ NWs were eliminated during a single-step nitridation process at 700°C using a fast ramp rate of 30°C/min. However, it should be noted that the NWs treated by this two-step temperature nitridation process were bent probably due to the fact that the crystal structure changes from bcc to the hexagonal wurtzite structure, and there is a non-uniform strain distribution between the core and shell. The effect of the post-growth nitridations on the PL of the In₂O₃ NWs is shown in Figure 3.

In the case of the nitridation using just NH₃ for 3 h at 500°C, one may observe that there is no substantial change in the shape of the PL of the In₂O₃ NWs except from the fact that the PL intensity has been reduced. However, the nitridation of the In₂O₃ NWs using NH₃ and H₂ leads to a clear suppression of the peak at 550 nm, which is attributed to oxygen consistent with previous investigations on Ga₂O₃ 4 . The peak around 400 nm maybe attributed to In vacancies 15 , but not O₂ as commonly suggested 11 12 13 . However, further work is required to clarify the origin of the PL peak around 400 nm.

Straight In₂O₃ NWs with diameters of 50 nm, lengths ≥2 μm and a bcc crystal structure have been grown on Au/Si(001) via the wet oxidation of In at 850°C. These exhibited two broad peaks in the PL, centred around 400 and 550 nm. The post-growth nitridation of In₂O₃ NWs was found to be effective by using NH₃ and H₂ at 500 and 600°C or a two-step temperature, nitridation process at 500 and 700°C. This lead to a suppression of the PL peak around 550 nm related to O₂ consistent with previous investigations on Ga₂O₃. In contrast, single-step temperature, nitridations using just NH₃, carried out with fast ramp rates above 600°C lead to the complete elimination of the In₂O₃ NWs, while they were not effective at 500 and 600°C.

APCVD: atmospheric pressure chemical vapour deposition; bcc: body centred cubic; DI: de-ionised; MO: metal-oxide; NWs: nanowires; NWSCs: nanowires solar cells; PL: photoluminescence; QT: quartz tube; RT: room temperature; SEM: scanning electron microscope; VLS: vapour-liquid-solid; XRD: X-ray diffraction.

The authors declare that they have no competing interests.

MZ and PP carried out the growth, scanning electron microscopy and x-ray diffraction measurements. AO carried optical characterization. All authors read and approved the final manuscript.