Introduction

Nitrides and in particular, group III-Nitride (III-N) compound semiconductors such as GaN, InN and AlN have been investigated intensively over the past decade due to their applications in electronic and optoelectronic devices like field effect transistors, light emitting diodes and lasers [1–3]. III-N semiconductors are especially attractive because their band-gap can be tailored from 0.7 eV in InN [4] up to 6.2 eV in AlN [5] but also due to the strained induced charges that provide an extra degree of freedom which can be used to tailor the band-profile and consequently the properties of devices [6].

However, in contrast to III-N compound semiconductors there are few investigations on group IV-Nitride (IV-N) compounds such as Ge3N4 [7, 8] and even less on Sn3N4 [9–22].

Tin nitride Sn3N4 is a relatively unknown semiconductor and the first investigation of Sn3N4 was carried out by F. Fisher et al.[9] as early as 1908. It has an energy band-gap that was estimated to be ≈1.5 eV [11] and so far Sn3N4 thin films have been grown by a variety of methods [12–19], including, atmospheric pressure chemical vapour deposition (APCVD) using halides [12, 13], metal organic chemical vapour deposition (MOCVD) [14], sputtering [15–18] and ammonothermal synthesis [19, 20]. Thin films of Sn3N4 have also been proposed as materials for optical storage [21, 22] since it was demonstrated that it dissociates into β-Sn upon exposure to a focused beam of light but also as a material for batteries [23].

Not surprisingly there are very few investigations on Sn3N4 nanostructured materials and it appears that the only study carried out so far concerns the synthesis of Sn3N4 nanoparticles (NPs) on Au coated Si(001) via chemical vapour deposition (CVD) using SnCl4 · 5H2O as a solid precursor, by Nand et al.[24]. To date, there is no report on the synthesis of Sn x N y nanowires (NWs) despite the fact that nanowires constitute a fundamental building block for the development of nanoscale devices such as third generation solar cells which require low cost, nanostructured materials.

Therefore in order to complement our earlier investigations on the synthesis and properties of InN NWs and related oxides such as In2O3 and SnO2 NWs [25–27], a preliminary investigation on the growth of Sn x N y NWs was undertaken.

Here the synthesis of the first Sn x N y NWs on Au/Si(111) is described and it is shown that Sn x N y nanowires (NWs) can not be grown via the direct nitridation of Sn over a broad range of temperatures i.e. between 300–800 °C due to the formation of Sn droplets on the surface of the Si(111). While there is evidence of one dimensional (1D) growth occurring at 500 °C via the direct nitridation of Sn, the yield is extremely poor. One-dimensional growth was promoted and significantly enhanced via the incorporation of NH4Cl in the Sn and its sublimation which acts as a dispersant thereby enhancing the vapour pressure of Sn. Sn x N y NWs with an average diameter of 200 nm were obtained at 450 °C while intricate assemblies of NWs have also been obtained at lower temperatures.

Experimental Procedure

The Sn x N y NWs were grown using an APCVD reactor which consists of four mass flow controllers (MFC’s) and a horizontal quartz tube furnace, capable of reaching a maximum temperature of 1,100 °C. Initially, fine Sn powder (Aldrich, < 150 μm, 99.5%) was loaded into a quartz boat together with a square piece of Si(111) approximately 7 mm × 7 mm in size, which was coated with a few nm’s of Au. The Au layer was deposited via sputtering at a slow rate using an Ar plasma under a pressure <10−4mBar. The Au/Si(111) sample was positioned ≈ 5 mm downstream from the Sn and subsequently the boat was loaded into the APCVD reactor and positioned directly above the thermocouple used to measure the heater temperature at the centre of the quartz tube. After loading the boat at room temperature (RT), Ar (99.999%) was introduced at a flow rate of 500 standard cubic centimetres per minute (sccm) for 5 min in order to purge the tube and eliminate O2and H2O. Following this the temperature was ramped to the desired growth temperature in a NH3flow of 250 sccm at a rate of 30 °C/min. Upon reaching the growth temperature (TG) the flow of NH3was maintained at 250 sccm for a further 60 min after which the tube was allowed to cool down over at least an hour, in a flow of 50 sccm NH3. The sample was removed only when the temperature was lower than 100 °C.

In order to enhance the one dimensional growth of Sn x N y NWs an equal amount of anhydrous NH4Cl (VWR Int 99.9%) was added to the Sn and mixed thoroughly in the boat. Then the same gas flows and temperature-time profile described above was employed. A summary of the temperatures and conditions is given in Table 1. The morphology of the Sn x N y NWs was examined with a TESCAN scanning electron microscope (SEM) while the crystal structure and the phase purity of the Sn x N y NWs were investigated using a SHIMADZU, XRD-6000, X-ray diffractometer with a Cu Ka source while a scan ofθ–2θ in the range between 10° and 80° was performed. Finally optical spectroscopy was carried out using a standard spectrophotometer UV/V (Perkin–Elmer Lambda 950) in the reflection mode at near normal incidence to the surface of the sample.

Table 1 Maximum diameter (Ø) of Sn droplets (DPs) and Sn x N y NWs obtained from heating up Sn under NH3and from the reaction of Sn + NH4Cl respectively, at different temperatures. In all cases a ramp rate of 30 °C/min was used to reachTGunder 250 sccm’s of NH3which was then maintained for a further 60 min at the growth temperature and after which the reactor was allowed to cool down to RT in a reduced flow of 50 sccm’s NH3
Full size table

Results and Discussion

As stated above the only investigation on the synthesis of nanostructured Sn3N4 is that of single phase, cubic tin nitride nanoparticles grown via atmospheric pressure-halide vapour phase epitaxy by Nand et al.[24]. In particular Nand et al. employed SnCl4 · 5H2O as a source of Sn and used 10 nm Au/Si(001) p-type substrates that were positioned at different distances from the SnCl4 · 5H2O, along the reactor. The SnCl4 · 5H2O was heated up to 500 °C under a flow of NH3 and N2 and the temperatures of the samples along the reactor were 400, 300 and 150 °C, respectively. However, Nand et al.[21] did not obtain any NWs. Before discussing the synthesis of the Sn x N y NWs obtained here it is instructive to consider first the synthesis of SnO2 NWs on 0.5 nm Au/Si(111) that were previously obtained by heating up Sn in an inert gas flow of 100 sccm Ar at 30 °C/min up to 800 °C and then maintaining the flow of Ar at this temperature for a further 90 min before cool down [27]. A typical SEM image of the SnO2 NWs obtained in this way is shown in Fig. 1a from which it is clear that a large yield of SnO2 NWs was obtained with an average diameter of 50 nm due to the reaction of Sn with residual O2 in the APCVD reactor. Performing the reaction under a direct flow of O2 leads to the formation of SnO2 around the molten Sn which limits the vapour pressure significantly and hence the growth of NWs. As a consequence the molten Sn upstream always had a grey like, non reflective appearance at the end of the process, while no droplets were observed among the SnO2 NWs. A similar process was also used recently for the growth of In2O3 NWs at 700 °C [26].

Figure 1
figure 1

SEM images ofa SnO2NWs with an average diameter of 50 nm grown on Au/Si(111) at 800 °C in an inert gas flow of 100 sccm Ar. No droplets exist among the NWsb Sn droplets on Au/Si(111) deposited using the same temperature-time profile ina but in a steady gas flow of 250 sccm of NH3as opposed to Ar. A broad distribution in the sizes of the Sn droplets was obtained with diameters ≤5 μmc similar tob but smaller Sn droplets with diameters ≤0.5 μm were obtained at 500 °C many of which are also elongated

Full size image

At first sight it would seem that the synthesis of Sn x N y NWs by direct nitridation of Sn with NH3 is feasible by changing from Ar to NH3 since Sn, like In, has a low melting point [28] and InN NWs have been obtained by direct nitridation of In with NH3 at a heater temperature of 600 °C [25]. However, Sn x N y NWs were not obtained by the direct nitridation of Sn with NH3. Instead many Sn droplets appeared on the Si(111) surface and a typical SEM image of such Sn droplets after the attempted nitridation of Sn with NH3 at 800 °C is shown in Fig. 1b. The Sn droplets cover the entire surface and have a density of ≈107 cm−2 and diameters ≤5 μm. Furthermore the size of the Sn droplets decreased as the temperature was reduced to 600 °C and many of then became elongated as shown in Fig. 1c.

The formation of large droplets on the Au/Si(111) surface during the direct nitridation of Sn with NH3 is a direct consequence of the reducing action of NH3 which eliminates the background O2 in the APCVD reactor. This in turn prevents the formation of an oxide around the Sn and so the molten drop always had a highly reflective, metallic like surface. In contrast when Sn is heated up in a flow of pure, inert Ar, the surface is grey like and not reflective due to the O2 background which is responsible for the formation of SnO2 NWs that were grown optimally on 0.5 nm Au/Si(111) at 800 °C using the same temperature-time profile and Ar as opposed to NH3[27].

Apart from droplets, no nanostructures were obtained via the attempted nitridation of Sn with NH3in the temperature range 600 °C < TG < 800 °C. Turning on the flow of NH3, after ramping up the temperature in an inert gas flow of Ar, did not lead to the growth of Sn x N y NWs either but again resulted into the deposition of Sn droplets. However, there was some evidence of one-dimensional growth atTG = 500 °C. Literally a few NWs with diameters >500 nm’s and lengths up to 3 μm appeared at a few locations on the Si(111) surface, hence the yield was extremely poor. Nevertheless, a further reduction of the growth temperature down to 300 °C did not lead to any significant deposition on the Si, no NWs were obtained and moreover, the Sn upstream lost its metallic shine due to the build up of a black deposit on the molten Sn which limited the vapour transport. In addition, no differences were observed upon changing the flow rate of NH3during the growth while keeping everything else equal at all temperatures so the direct nitridation of Sn alone under a flow of NH3is not effective and leads to the deposition of Sn droplets on the Au/Si(111) surface which impedes one-dimensional growth over a wide temperature range i.e. 300–800 °C as shown below in Table 1.

The XRD spectrum of the Sn droplets deposited at 800 °C is shown in Fig. 3and is characterized by an intense peak corresponding to the (2 0 0) orientation of Sn and less intense but well resolved peaks corresponding to the (1 0 1), (3 0 1), (4 0 0) and (3 2 1) orientations. In addition to the Sn droplets the Al holder peaks have also been identified but no peaks associated with Sn x N y were found.

These findings are in direct contrast with the case of InN where NWs can be grown by direct nitridation of In with NH3via a self-catalytic mechanism. The optimum heater temperature for the growth of InN NWs was found to be 600 °C where its vapour pressure is equal to 10−6 Torr. Large In droplets comparable in size to those in Fig 1b started appearing only at temperatures ≥800 °C in contrast to the Sn droplets whose density was large and persisted even down to 600 °C where its vapour pressure is <10−11 Torr. It appears therefore that the Sn droplets are born out from the melt upstream and are transferred to the Si(111) surface where they coalesce to form larger droplets.

The tendency for one-dimensional growth observed atTG = 500 °C during the direct nitridation of Sn with NH3was promoted by the addition of NH4Cl into the Sn at a ratio of 1:1 by weight. The reaction of NH4Cl with Sn was carried out under a flow of NH3keeping the flow rate, ramp rate and temperature profile identical to those used in the case of ‘direct nitridation’ of Sn with NH3. A typical SEM image of Sn x N y NWs obtained by the reaction of Sn with NH4Cl under NH3at 450 °C is shown in Fig 2a. The Sn x N y NWs have an average diameter of 200 nm’s and lengths up to 5 μm while the reaction of Sn with NH4Cl always lead to the deposition of a white powder downstream, near the cool end of the reactor, in contrast to the direct nitridation of Sn with NH3where no by products occurred.

Figure 2
figure 2

a Sn x N y NWs grown on Au/Si(111) at the optimum temperature of 450 °C using a Sn:NH4Cl mixture under a flow of 250 sccm’s NH3b high magnification SEM image of Sn x N y NWs

Full size image

The XRD spectrum of the Sn x N y NWs grown at 450 °C is shown in Fig. 4 and is characterized by the (2 2 0), (3 1 1), (5 1 1) and (4 4 0) peaks, which can be indexed to the hexagonal structure of Sn3N4[18]. The intense peak of Sn(200) observed in Fig. 3 has disappeared and once more the Al peaks appearing in the XRD spectrum of Fig. 4 due to the sample holder have been identified. Furthermore there are no peaks associated with SnO2[27].

Figure 3
figure 3

XRD spectrum of Sn droplets deposited on Si(111) at 800 °C under a flow of 250 sccm NH3

Full size image
Figure 4
figure 4

XRD spectrum of Sn x N y NWs grown on Au/Si(111) at 450 °C via the reaction of Sn and NH4Cl

Full size image

The promotion of one dimensional growth is attributed to the dissociation of NH4Cl. Upon increasing the temperature NH4Cl undergoes sublimation at 338 °C and therefore dissociates into NH3and HCl according to the following reaction.

(1)

As explained by Chaiken et al.[29] the sublimation rate of NH4Cl increases by a factor of 104 when changing the temperature from T = 100 to 600 °C and the typical sublimation weight loss of NH4Cl is over 90% when heated for ≈60 min. It is interesting to point out here that this sublimation process is endothermic and the temperature is expected to be reduced only by a few tens °C in the case of NH4Cl [29]. The decomposition of NH4Cl enhances the porosity of the Sn melt and more importantly acts as a dispersant increasing the amount of Sn that is transferred into the gas stream. In fact the sublimation of NH4Cl and the generation of NH3 and HCl gasses which act to disperse the molten Sn occurs abruptly and leads to strong dispersion of the molten Sn inside the boat for the ramp rate used here i.e. 30 °C/min suggesting that lower ramp rates would be more suitable. In addition to acting as a dispersant, the sublimation of NH4Cl yields HCl which reacts with Sn leading to the formation of SnCl2 according to the reaction,

(2)

Note that SnCl2melts at 38 °C and decomposes above 600 °C. Subsequently the SnCl2reacts with NH3according to,

(3)

Consequently the role of the NH4Cl is two fold. First, it prevents the Sn from melting up into one single drop and second it supplies the necessary HCl for the formation of SnCl2. As stated above heating up Sn alone in NH3did not lead to the deposition of any products near the cool end of the reactor so the deposition that occurs from heating up Sn and NH4Cl in NH3is due to the reaction of Sn with HCl since XRD of the deposit showed no peaks related to NH4Cl.

The reaction outlined above is in a way similar to that put forward by Nand et al.[21] whereby SnCl4 reacts with NH3 according to,

(4)

leading to the growth of Sn3N4 NPs on 10 nmAu/Si(111) that were placed at various positions along the reactor but also similar to the growth of Sn3N4 thin films by APCVD using halides, by Gordon et al. and Takahashi et al. [12, 13].

A similar kind of reaction was also used to grow InN nanocrystals on Si(111) whereby the incorporation of NH4Cl into the In lead to the complete elimination and transfer of the In into the gas stream where it formed primarily InCl which in turn reacted with the NH3 leading to the formation of InN nanocrystals with diameters of 300 nm [30].

While the addition of NH4Cl in Sn did not result into its complete transfer in the gas stream like with In, it provided nonetheless the necessary HCl for the formation of SnCl2which subsequently reacts with NH3on the Au/Si(111) leading to the one dimensional growth of Sn x N y . Interestingly the distance of the sample from the Sn:NH4Cl mixture was found to be critical and for distances >10 mm the reaction led to the formation of closely packed NPs with sizes <100 nm on the Au/Si(111) most probably due to the lower vapour pressure of the SnCl2.

Although the details of the growth mechanism are not understood thoroughly at present it is suggested that the Sn x N y NWs grow self catalytically from Sn x N y NPs although the role of the Au which appears to enhance the one dimensional growth still needs to be clarified [32, 33].

A first estimate of the band-gap of the Sn x N y nanowires grown on Si(111) was obtained from optical reflection measurements using a UV–IR spectrometer at near normal incidence on both the NW sample and the Si(111) substrate for comparison, shown in Fig. 5. Clearly evident is the distinct difference in the reflection spectra from the substrate and the NWs. Also evident is the band edge of the Sn x N y NWs which is estimated to be approximately 2.6 eV [31].

Figure 5
figure 5

Optical reflection from the Sn x N y NWs (Left) and plainn-type Si(111) substrate (Right)

Full size image

In addition to Sn x N y NWs that were obtained at TG = 450 °C there is also evidence for the formation of more complex nanostructures obtained for TG < 450 °C as shown in Fig. 6a and b. However, their density was smaller compared to that in Fig. 2a due to the lower growth temperature which limits the amount of Sn transferred over to the Si(111). The radial growth of NWs from the droplet shown in Fig 6a is very similar to the case of InN [25] whereby nucleation centres form on the surface of droplets which then facilitate radial growth [32]. Moreover, the circular arrangement of NWs shown in Fig. 6b is due to the formation of droplets that accumulate near the periphery of well defined circles similar to the growth of In2O3 nano pyramids that self assemble in the form of wreaths due to the reaction of In with NH4Cl in a flow of N2[30].

Figure 6
figure 6

Nanostructures obtained atT = 400 °C on 0.7 nm Au/Si(111)a NWs emanating from a droplet andb NWs which have grown from droplets in a circular configuration

Full size image

Conclusions

The first tin nitride, Sn x N y nanowires have been grown by CVD on Au coated Si(111) via the reaction of Sn with NH4Cl at 450 °C under a steady flow of NH3. Attempting direct nitridation of Sn with NH3leads to the formation of Sn droplets due to the reducing action of the NH3which eliminates O2in the reactor and which in turn inhibits one-dimensional growth over a wide temperature range between 300–800 °C. The formation of large Sn droplets was suppressed by adding NH4Cl which dissociates into NH3and HCl by sublimation at 338 °C and acts as a dispersant thereby enhancing the vapour pressure of Sn. Furthermore the Sn reacts with HCl and yields SnCl2which subsequently reacts with NH3leading to the formation of Sn x N y nanowires which have diameters of 200 nm and lengths up to 5 μm. Finally nanowires protruding from droplets and intricate assemblies of NWs arranged in the shape of wreaths were also obtained at temperatures <450 °C. The synthesis of metal (M)-nitride i.e. M x N y nanowires where the metal component is readily available and has a low cost, is expected to be important for third generation solar cells based on nanostructured semiconductor materials.