Abstract

Controlling the growth of semiconducting nanowires with desired properties on a reproducible basis is of particular importance in realizing the next-generation electronic and optoelectronic devices. Here, we investigate the growth of cupric oxide (CuO) nanowires by direct oxidation of copper-containing substrates at for 150 minutes at various oxygen partial pressures. The substrates considered include a low-purity copper gasket, a high-purity copper foil, compacted CuO and thin layers, and layered Cu/CuO and Cu/ substrates. The morphology, composition, and structure of the product CuO nanowires were analyzed using scanning electron microscopy, energy-dispersive X-ray spectroscopy, transmission electron microscopy, selected area electron diffraction, X-ray diffraction, and UV-Visible absorption. Selected oxidation processes have been monitored using a thermogravimetric analyzer. The layering structure of the substrate after oxidation was analyzed to elucidate the growth mechanism of CuO nanowires.

1. Introduction

One-dimensional nanostructures that can efficiently transport electrical carriers have been an exciting area of research due to their potential applications in next-generation optoelectronic [1], electronic [1], bio- [2], and gas- [3] sensing devices. Research in the controlled growth of semiconducting nanowires (NWs) is of particular importance in achieving the desired properties on a reproducible basis. Various growth techniques have been employed, including the vapor-liquid-solid growth [4], epitaxial growth [5], vapor-solid growth [6], wet chemical methods [7], and electrospinning [8]. For vapor-solid growth, it has been known since the 1950s that the oxidation of various metals, including copper, iron, zinc, and tantalum, at intermediate temperatures, results in a parallel oxide layering with high-density oxide wires/whiskers at the surface [9]. However, only recently with the advent of advanced characterization and assembly techniques has this phenomenon revived interest. The simplicity of the vapor-solid technique, together with its high NW yield and potential for controllability, makes it an attractive avenue for NW growth.

Cupric oxide (CuO) has been extensively studied due to its application in high Tc superconductors [10] and as a heterogeneous catalyst [11]. In addition, bulk CuO with a known band gap of 1.2 eV has an interesting monoclinic crystal structure belonging to the Mott insulator material class whose electronic structure cannot be described by conventional band theory [12]. CuO NW arrays have been recently applied to various applications including gas sensing [3], field emission [13, 14], and photovoltaic devices [7]. The vapor-solid oxidation technique has been widely used to grow CuO NWs [6, 1417] from copper substrates. Recent studies on CuO NWs have focused on the influence of temperature, growth time, and oxidative environment on the growth rate, size distribution, and areal density of CuO nanowires [6, 14, 15]. It has been shown that the growth time can be used to control NW length distribution, the growth temperature can be used to control the NW diameter distribution [6], and the O2 and H2O partial pressures strongly affect the NW areal density [15, 16]. The surface condition of the copper substrate was also shown to significantly affect the nanowire growth [17].

A common problem with the vapor-solid oxidation growth of CuO NWs from copper substrates is that the mismatching stress between the substrate and the oxide layer leads to the flaking of the oxide layer and hence the poor electrical contact between the NW and the substrate. Herein, we investigate the growth of CuO NWs directly from copper-containing substrates, such as CuO and Cu2O layers with and without a Cu substrate. In addition, the effect of O2 partial pressure on NW areal density is investigated. The layering of the substrate after oxidation is analyzed to elucidate the growth mechanism. The resulting CuO NWs on a conducting/semiconducting substrate can be directly used for electronic applications.

2. Experimental Details

Both copper substrates and copper-containing substrates were used for the growth of CuO NWs. Copper substrates used include a low-purity Cu gasket with high surface roughness and a high-purity Cu foil (5Ns, 0.2 mm thick) with low surface roughness. Before the NW growth, the Cu substrates were cleaned in 1 M HCl solution for seconds, then rinsed in deionized water, and finally dried in a dry airflow. Copper-containing substrates used include CuO and Cu2O thin layers that were created by compacting commercial CuO (3Ns, ) and Cu2O (2Ns, ) powders under 103 MPa for minutes. Likewise, CuO and Cu2O powders were deposited onto high-purity Cu foils and compacted to create Cu/CuO and Cu/Cu2O layered substrates.

CuO NWs were grown by direct oxidation of the above substrates in pure O2 or Ar–O2 mixtures. The oxidation processes were normally carried out inside a Lindberg/Blue Mini-Mite tube furnace. However, to monitor the oxidation process, a few substrates were oxidized using a TA Instrument SDT 2960 horizontal thermogravimetric analyzer (TGA). During a typical oxidation procedure, a substrate was exposed to an Ar–O2 gas mixture of 5 lpm with the desired O2 partial pressure, then heated at to an isotherm of 500°C and held for 150 minutes. Upon cooling down, a pure inert gas flow was introduced to prevent the further oxidation. The O2 partial pressures were controlled at 10%, 15%, 20%, and 100% (pure O2) by mixing O2 with Ar at appropriate flow rates using mass flow controllers.

Following the oxidation process, each sample was analyzed using a Topcon ABT-32 scanning electron microscope (SEM) equipped with an energy dispersive X-ray (EDX) spectrometer. SEM images were used to analyze the NW morphology and oxide layering and the EDX was used to determine the elemental composition of the sample. The as-produced NW samples were first scratched off their growing substrates, and then transferred onto copper grids and analyzed using a Hitachi H 9000 NAR transmission electron microscope (TEM) for bright field (BF) TEM and high-resolution TEM (HRTEM) imaging and selected area electron diffraction (SAED). A Scintag XDS 2000 X-ray diffractometer (XRD) was used for further analyzing the crystal structure of the samples. An Ocean Optics 2000 UV-Visible spectrometer was used to obtain the absorption spectrum (reflection mode) of an as-grown high-density NW array. The reflection from a high-quality mirror was used as the reference.

3. Results and Discussion

Figures 1(a) and 1(c) show low-magnification TEM images of nanowires grown in pure O2 from a high-purity Cu foil and a low-purity Cu gasket, respectively. The corresponding HRTEM images of the NWs are shown in Figures 1(b) and 1(d), respectively. The visible stripes identified with different contrasts in Figures 1(a) and 1(c) indicate that as-grown nanonwires bear faceted instead of circular cross-sections. The SAED pattern of the NW in the inset of Figure 1(a) confirms the crystallinity of product nanowires and the direction of diffraction spots is consistent with the fringe direction shown in the HRTEM image of Figure 1(b). The lattice spacing of the NWs was measured as 0.252 nm, corresponding to the ( ) plane of CuO. The direction of the fringes in Figure 1(b) is in parallel with the nanowire axis, while the fringes shown in Figure 1(d) are not parallel to the wire axis, with a similar angle of 50° reported in [9]. The two observed fringe directions may be attributed to the difference in surface orientation or surface roughness of the corresponding substrates [18, 19]. The direction in which a crystal grows on a substrate is dependent on the orientation of the seed crystal. As CuO NWs grow from a Cu foil, which is polycrystalline and consists of grains with varying crystallographic orientations, their growth directions would be different from each other, leading to dissimilar fringes of the NWs in HRTEM images.

The low-purity copper substrates were oxidized at four different O2 partial pressures of 10%, 15%, 20%, and 100%. Upon inspection after oxidation, the surfaces of the copper substrates were tarnished from light to dark grey with respect to the 10% to 20% O2 partial pressures, whereas the substrate oxidized in 100% O2 was tarnished almost completely black with only several small grey areas. The SEM images of the oxidized copper substrates are shown in Figures 2(a)2(d). Since the NW density varies throughout the surface of each individual substrate, the images shown are those representative areas with more uniformly dense NWs on each substrate. The NW density and the vertical alignment increase with increasing O2 partial pressures, while the average length and diameter of the NWs seem to decrease with decreasing O2 partial pressures. Most NWs appear to be tapered like a needle or whisker, which may be due to the high roughness of the substrate and the subsequent merging of multiple NWs into one NW during the growth. In the case of pure O2, the areas of highest NW density were those that tarnished black. And the light to dark grey samples with 10 to 20% O2 partial pressures correspond to the low to medium NW density. Since bulk CuO is known to be grey to black in color and Cu2O is red, it can be roughly inferred that the oxide surface and possibly the NWs are CuO. However, in many cases the color of nanostructures is very different from that of their corresponding bulk structures due to morphology-induced interference and scattering. Thus, the observed color can only be used as a rough guide and the actual composition of the surface and the NWs must be obtained with the aid of advanced characterization techniques. A UV-Visible absorption spectrum of a high-density NW array similar to the sample shown in Figure 2(d) is presented in Figure 3. The band gap of the as-grown NWs was estimated from the absorption spectrum by fitting a tangential line passing the midpoint of the first absorption edge (750–940 nm) and reading out the -intercept. In this way, a band gap of approximately 1.2 eV (1030 nm) corresponding to bulk CuO can be extracted.

To further analyze the oxide layering structure, an oxide flake was carefully removed from a copper substrate oxidized in pure O2. The SEM image of the flake is shown in Figure 4(a), with an amplified figure shown in the inset. Three layers can be distinguished: a bottom layer that lies directly above the copper substrate, an intermediate layer that the NWs grow from, and a top NW layer. The XRD spectrum of an oxidized Cu substrate is shown in Figure 4(b), which indicates the presence of both Cu2O and CuO. However, no information on the individual layer composition can be extracted. The theoretical layering composition (labeled in Figure 4(a)) has been discussed in [8, 9] with CuO as the NWs and the intermediate layer and Cu2O as the bottom layer. The theoretical hypothesis on the NWs has been proved by our experiments.

To confirm the rest of the layering composition, EDX spectra were obtained from the top and the bottom surfaces of an oxide flake similar to that shown in Figure 4. In addition, EDX standards were produced from compacted CuO (3Ns) and Cu2O (2Ns) powders. The resulting spectra and region analysis are presented in Figure 5. The spectra were made up of counts, taken at identical parameters of 5 keV and magnification with a 60 degree sample tilt. Regions of 0.400–0.600 keV and 0.620–1.080 keV were created about the OKα and CuLα peaks, respectively, and the number and the percentage of gross counts in each region was determined. The relatively small fraction of oxygen counts detected is due to the low energy X-rays produced and the subsequent reabsorption upon exiting the sample material. The results indicate that the percent of gross counts for oxygen and copper taken from the bottom of the flake matched that of the standard Cu2O. Likewise, the percent of gross counts for oxygen and copper taken from the top of the flake matched that of the standard CuO. Furthermore, a simple calibration has been carried out to correlate the gross count ratio of oxygen to copper with the molar ratio of O to Cu (or value of , where equals 0, 0.5 and 1 for Cu, Cu2O, and CuO, resp.) in standard reference samples. The resulting value is 0.99 for the top layer and 0.48 for the bottom layer in our sample. Our experimental results indicate distinct oxide layers and directly agree with the theoretical speculation [8, 9] of the bottom layer of Cu2O and the intermediate layer of CuO. The layering structures can be used to interpret the CuO NW growth mechanism as elaborated in [6]: Cu2O acts as a precursor for the second oxidation step to CuO where the slow production rate of CuO results in a localized CuO vapor pressure and a continuous growth mode for CuO NWs by means of the vapor-solid growth mechanism.

Compacted CuO and Cu2O layers with and without the Cu foil support were also used to grow CuO NWs in pure O2 at 500°C for 150 minutes. The SEM images of the oxidized substrates are presented in Figure 6. High-density NW growth on the Cu/Cu2O substrate, medium NW growth on the Cu/CuO substrate, low NW growth on the Cu2O substrate, and no NW growth on the CuO substrate have been observed. The results indicate the importance of Cu2O to CuO conversion for the growth of CuO NWs as well as the presence of a Cu substrate for high-density NW growth. The importance of the Cu substrate is possibly in its role as a Cu ion source and in establishing a chemical potential gradient for the diffusion of Cu ions to the substrate surface. In Figure 6(a), the relatively thick compacted Cu2O layer with high-density NWs at its surface is seen at the top side of the Cu substrate and the familiar layering scheme is observed at the bottom of the substrate. Also note that the oxide layers of Figures 6(a)6(c) are uniformly structured with less or without pores, whereas the presumably unchanged CuO sample in Figure 6(d) has a porous structure consistent with compacted powders.

To determine if the oxidation took place, the Cu2O and CuO thin layers were processed separately in a TGA. From the TGA results shown in Figure 7, it is clear that the Cu2O thin layer oxidized whereas the CuO thin layer did not. The small decrease in the weight for the CuO thin layer is a consequence of the thermal expansion of the TGA and can be ignored. XRD spectra were also obtained from each sample and compared to those of unprocessed Cu2O and CuO in Figure 8. For those samples which oxidized to form CuO NWs (Cu2O, Cu/Cu2O , Cu/CuO), the XRD spectra indicate the presence of both CuO and Cu2O, suggesting the importance of the oxide layering on NW growth.

4. Conclusion

CuO NWs have been successfully grown from both copper and copper-containing substrates through direct oxidation of the substrates in oxidative environments. The as-grown NWs have a band gap of (1030 nm) corresponding to bulk CuO. Oxygen partial pressure was found to significantly affect NW growth. The vertical alignment of the NW increases with increasing O2 partial pressures, while the average length and diameter of the NWs seem to decrease with decreasing O2 partial pressures. Analyses on the oxidized substrates indicate CuO as the NWs and the intermediate layer and Cu2O as the bottom layer. The oxidation of compacted CuO and Cu2O layers with and without a Cu substrate suggests the importance of Cu2O to CuO conversion for the growth of CuO NWs, as well as the importance of a Cu substrate for high-density NW growth.

Acknowledgments

This work was supported by National Science Foundation (DMI-0609059), WiSys Technology Foundation, and the University of Wisconsin-Milwaukee (UWM) Advanced Analysis Facility (AAF) Summer Internship Program. The authors thank Steven Hardcastle for assistance with SEM, EDX, and XRD analyses, which were carried out at the AAF of UWM. TEM analyses were performed at the UWM HRTEM Laboratory. The authors also thank the anonymous reviewers for their valuable comments and suggestions.