6 Plasma treatment on ZnO nanorods
6.3 Rectifying behavior of ZnO nanorods by ammonia-plasma
The ZnO nanorods present single crystalline structure and a well-defined hexagonal plane with a homogeneous diameter of approximately 60-70 nm. The cross-sectional scanning electron microcopy (SEM) image in Figure 6.6(a) shows that the highly oriented ZnO nanorods with a uniform length of 500-520 nm are perpendicularly grown to the substrate. In addition, it was found that by controlling the experimental conditions, highly arrayed ZnO nanorods or nanowires with different aspect ratios can be grown from the chemical aqueous solution [116]. A close observation on the
microstructure of the ZnO nanorods, as show in Figure 6.6(b), reveals that the surface morphology of ZnO nanorods shows a waved shape structure with 0.5 ~ 1 nm roughnesses. In addition, some stacking faults are also observed (marked with arrows).
However, these small surface roughness and stacking faults seem not to affect the crystal quality of ZnO nanorods because the selected-area electron diffraction pattern in the inset of Figure 6.6(b) reveals that the ZnO nanorods still exhibit a single crystalline structure. Figure 6.6(b) clearly describes the perpendicular directional growth of the ZnO nanorods where the singular fringes spacing is about 0.51nm, which is nearly consistent with the c-axis parameter in hexagonal ZnO structure (c = 0.521 nm in ZnO wurtzite structure), indicating that <001> is the preferred growth direction for the ZnO nanorods, in consistence with XRD patterns that shows a single strong ZnO (002) peak at 2θ =34.4o (not shown here).
Figure 6.6: Cross-sectional scanning electron microcopy (SEM) micrograph of (a) ZnO nanorods/ZnO film/Si, (b) surface microstructure and stacking faults (marked with arrows) of ZnO nanorods. The inset image of (b) displays SAED pattern.
Figure 6.7(a) illustrates the photoluminescence (PL) property of ZnO nanorods at room temperature. The PL spectrum of non-plasma ZnO nanorods presents a weak ultraviolet (UV) emission peak at 3.28 eV and a relatively strong deep-level emission peak at 2.10 eV. The UV emission peak of ZnO is generally attributed to an exciton-related activity [1], and the deep level emission may be due to the transitions of native defects such as oxygen vacancies and zinc interstitials. In addition, the imperfect boundaries in ZnO nanorods would cause the unstable surface status to trap impurities and further damage the optical property, especially for the nano-scale ZnO nanorods. In contrast, it was found that the peak intensity of the UV emission increases with the plasma duration up to 90 sec but the deep level emission in all plasma-treated samples tends to disappear, indicating that the native defects or impurities, contributing to visible transition, can be much reduced by NH3 plasma treatment. As shown in Figure 6.7(b), a maximum relative PL ratio (peak intensity of ultraviolet emission (IUV) to that of deep level emission (IDLE)) shows up around 90-120 sec, implying that the plasma-treated ZnO nanorods exhibit much better PL properties than that of non-plasma sample. The improvement in the optical properties of ZnO nanorods with the plasma treatment may be attributed to the reduction of the defect concentration due to the occupation of N ions on the defect sites, i.e., oxygen vacancies, which will be further elucidated later. However, a further increase in plasma duration such as 180 sec would result in the decrease in the relative PL ratio that is probably related to the plasma etching as revealed by the SEM image in Figure 6.7(c) where the ZnO nanorods were deteriorated that is in contrast to the ZnO nanorods with the plasma-treated duration less than 90 sec, well-aligned ZnO nanorods with well-defined hexagonal plane remains unchanged with an average length of 500 nm, similar to that without NH3 plasma treatment.
Figure 6.8 shows the relative intensity ratio of internal N (Ni) to surface N (Nf)
dependent on etching time for the ZnO nanorods with NH3 plasma treatment. No nitrogen signal was detected in the reference sample [see the inset of Figure 6.8(a)]
but the N1s peak could be detected in plasma-treated ZnO nanorods from surface to internal region for the ZnO nanorods (duration at 90 sec). (Note: The N1s (~401 eV) internal signal was obtained by following the measurement steps. Firstly, ZnO nanorods were etched by Ar ion source and then the N1s signal of the etched ZnO nanorods can be detected. Next, both intensities of N1s signal (401 eV) for etched and non-etched nanorods were compared and considered as the internal and surface signals, respectively.) In addition, Figure 6.8 (b) illustrates the peak shift of 0.2-0.6 eV toward higher binding energies in Zn 2p spectrum for NH3 plasma-treated ZnO nanorods as compared to that of the reference sample. This peak shift indicates the reduction of the surface band bending [125] and this may be related to the doping or incorporation of N ions into ZnO nanorods. The doped N atom could have the probability to occupy O sites to form the acceptor (NO) in ZnO nanorods. Thus, the NH3 plasma treatment seems to reduce the densities of the surface defects (Figure 6.6(b)) on ZnO nanorods. However, it was also observed in the inset of Figure 6.8(b) that as the ZnO nanorods were NH3-plasma treated more than 180 sec, the Zn 2p peak is shifted toward a higher binding energy around 4-5 eV. This phenomenon implies that the surface band bending of the ZnO nanorods was strongly influenced by NH3
plasma treatment and therefore, the structure morphology of ZnO nanorods would be probably changed as shown in Figure 6.7(c).
Figure 6.7: (a) Room-temperature PL spectra and (b) PL ratio of IUV/IDEL for ZnO nanorods with and without NH3 plasma treatment. (c) The FESEM surface morphology of ZnO nanorods with NH3 plasma treatment of t =180 sec.
Based on the above discussion, a mechanism for the formation of p-type ZnO nanorods under NH3-plasma treatment is tentatively proposed as follows. As we can see in Figure 6.6(b), imperfect boundaries (surface roughness) and many defects (stacking faults) were observed in the ZnO nanorods where N and H radicals (ions) can be easily doped into ZnO nanorods by thermal diffusion via the nature defect routes. In general, the H ions are more easily diffused into ZnO than N ions in the plasma-treated process and can combine with other defects to form a shallow donor in ZnO. However, according to the PL analysis, it was found that the luminescence properties were enhanced with the increase of plasma-treated duration up to 90 sec,
indicating that the incorporation of H into ZnO nanorods does not show negative influence on the optical properties of ZnO nanorods in such short plasma duration. In other words, the improvement in optical property of ZnO nanorods can be primarily attributed to the reduction in surface defects and defect concentration of ZnO nanorods due to the occupation of N ions on the defect sites [127]. However, a long NH3-Plasma treatment over 180 sec not only causes the plasma etching but also promotes the diffusion of H ions into the ZnO nanorods as well as increases the concentration of shallow donors to restrict the formation of p-type ZnO.
Figure 6.9 shows the I-V curve for a homojunction of p-type ZnO nanorods (treated by NH3 plasma duration of 90 sec) on n-type ZnO films. The graph in the inset of Figure 6.9 shows surface I-V characteristics using In and Au/In electrodes on both n- and p-type ZnO. For the linear dependences of I-V characteristics, the ohmic contacts are fairly confirmed. Rectification by p-n junction is clearly displayed. The threshold voltage appears at 1.5 V under forward bias and it is almost half of the band gap energy of ZnO. It was also found that the I-V curve presents a little leakage current in reverse bias and this could be due to incomplete contact between some nanorods and contact metal electrode. Further experiments are required to understand the possible origin of the rectifying behavior with such a smaller threshold voltage.
Nevertheless, the I-V characteristics of nitrogen-doped ZnO nanorods supports that the current approach is a valuable one to fabricate ZnO nanorods for p-n junction device application.
Figure 6.8: (a) Dependence of relative intensity ratio of internal N (Ni) to surface N (Nf) on etching time for the ZnO nanorods with the inset showing the N1s spectra obtained from the ZnO nanorods with and without NH3 plasma treatment (t= 90 sec).
(b) Zn 2p spectra obtained from the ZnO nanorods with and without NH3 plasma treatment. The inset shows the Zn 2p3/2 spectra of ZnO nanorods with NH3 plasma treatment of t=180 sec.
6.4 Summary
We have developed well-aligned arrays ZnO nanorods on four-inch PC organic substrates buffered with ZnO film by combining a simple chemical solution at low temperatures with H2 plasma treatment. The photoluminescence spectra indicate that the optical quality of the ZnO nanorods can be much improved by H2 plasma treatment as evidenced by the remarkable increase in the ultraviolet emission intensity.
The I-V results suggest that n-type ZnO nanorods with a higher conductivity can be obtained by combining the chemical solution process with hydrogen plasma treatment.
The XPS analysis shows that the nitrogen ions can be doped into ZnO nanorods through surface adsorption or defect routes under NH3 plasma treatment. The native
defects of the ZnO nanorods can be effectively reduced and the optical property can be much improved. Moreover, the electrical transport data reveal that p-type ZnO nanorods with a smaller threshold voltage of 1.5 V can be obtained by combining the chemical solution process with NH3 plasma treatment.
Figure 6.9: I-V curves for a p-n homojunction formed by p-type ZnO nanorods grown on n-type ZnO films. Inset show the Ohmic contacts on n-type and p-type ZnO.
Chapter 7
One dimensional ZnO and carbon nanotube (CNT) composite
7.1 Introduction
Semiconductor nanoparticles have been widely investigated from fundamental and applied research viewpoints, owing to their potential in photocatalysis, solar energy conversion, and optoelectronic applications [128-135]. The assembly of isotropic nanoparticles onto one-dimensional architecture represents an important step towards the integration of nanoparticles into nanodevices [136]. Particularly, nanoparticles can be assembled on 1D nanostructure of a different material to form novel and interesting composite nanomaterial system [137]. Single-walled carbon nanotubes (SWNTs), due to their excellent structural and electronic properties, is the one of the most potential 1D nanostructures in the past few years [138,139]. To achieve multifunctional nanodevices, many research groups have developed several methods to coat metal and semiconductor nanoparticles, such as Au, Ag, SnO2, TiO2, SiO2, and Al2O3, on carbon nanotubes (CNTs) [140-145]. Moreover, some of coated CNT composite system have been demonstrated that their original physical properties would be improved after coating specific nanoparticles [146,147]. Zinc oxide (ZnO), wide band-gap semiconductor (3.4 eV) with a large exciton binding energy (60 meV), has been extensively investigated because of its excellent performance in sensors and optoelectronics systems [15,72]. There are several methods to fabricate ZnO clusters
and nanoparticles on nanostructures [148,149]. To date, there are few reports on assembling the composite from CNTs and ZnO. Zhu et al [150]. and Ravindran et al [151]. have reported the coating of ZnO nanoparticles on multi-walled carbon nanotubes (MWCNTs) by sputtering and electrostatic coordination approach, respectively. However, form above results showed it is difficult to form a continuous layer on CNTs, even in MWCNTs. Hence, the coating behaviors and mechanisms on CNTs need to be further investigated.
In this chapter, we will present the fabrication of ZnO-coated SWNTs either as buckypaper or as individual tubes at low temperature via ultrasound irradiation of oxidative SWNTs, zinc acetate, and triethanolamin in aqueous solution. The successful formation of ZnO layer on SWNTs has been confirmed using atomic force microscopy (AFM), scanning transmission electron microscopy (STEM), and electron transport measurement. Finally, we will show the photoconductivity properties of individual ZnO nanorod device to compare with composite materials.
7.2 Synthesis and characterization of ZnO-SWNT composite
SWNTs produced by Hipco method were obtained from Rice University. These nanotubes were sonicated in nitride acid (HNO3) for 10 min. This treatment is known to produce shortened nanotubes and to also create defects on side-walls, both of which are capable of being terminated by carboxyl and hydroxyl groups in the presence of oxidizing acid. The schematic representation for the fabrication of SWNTs and ZnO composite is shown in Figure 7.1. After oxidizing treatment, the SWNTs were dispersed into 20 mL aqueous solution; then the solution containing SWNTs was heat
to 95oC with intense ultrasonic waves. And then, 0.005 mol/L zinc acetate and 1.5 mol/L triethanolamine (TEOH) were dropped smoothly into the SWNTs aqueous solution. In typical experiment, the system was continuously stirred with ultrasonic waves for 2 h at 95oC. The thickness of the coating can be adjusted by changing the reaction time and the concentration of zinc acetate. For the transport measurement, individual SWNT and SWNT@ZnO were positioned on top to bridge two noble metal electrodes (Au-Pd), which acts as the source and drain. The electrodes were deposited on top of a SiO2 layer on a silicon substrate, and the silicon was used as the gate.
According to our pervious study, the average diameter of pristine SWNTs was about 1.2 nm. After coating process, the average height of SWNTs increased to 8 nm, which means there could be a ZnO layer about 7 nm thick on SWNTs. To confirm the ZnO layer was around on SWNTs, the chip was heated at 500 and 600oC in air for 30 min. Generally, pristine SWNTs will be burned out and disappeared. In AFM image (Figure 7.2), it was found that the same area reveals linear features formed at the positions of SWNTs after the chip was burned. The insert image shows that surface morphology of coating SWNTs after 500oC heating in air. We have observed some crystalline structure at the same region of SWNTs, and we proposed this crystalline behavior is due to the higher heating temperature transfer amorphous ZnO layer to poly-crystalline ZnO.
Figure 7.1 The schematic representation for the fabrication of SWNTs and ZnO composite.
Figure 7.2 AFM images of SWNT@ZnO. (a) Before heating, (b) heating at 500oC,
The ZnO coating studies have been complemented by micro Raman spectroscopy on the individual SWNTs and bucky-paper after coating process, as shown in Figure 7.3. The main features in the Raman spectra of SWNTs are the radial breathing mode (RBM) which is in the frequency range up to 200 cm-1, G-line around 1600 cm-1 from the tangential vibrations of the carbon atoms, and the D-line at 1280 cm-1. The intensity of the D-line is related to the amount of sp3 defects on carbon nanotubes.
Therefore, we suggested that the intensity of D-line increased is due to SWNTs were treated with nitric acid to create surface function group. Moreover, we found the Raman spectrum of SWNT@ZnO shows slow decrease and slight shift in the RBM intensity in individual nanotubes. In addition, it was found all Raman signals shifted toward low frequency about 2-5 cm-1, in the Raman spectra of SWNT@ZnO bucky-paper materials, RBM peak especially (~5 cm-1). We proposed the signals decrease was related to possible scattering and these peaks shift was contributed by the chemical interaction between carbon atom and TEOH-Zn, respectively. In order to demonstrate our contention, we evaporated 20 nm silicon monoxides to cover the individual SWNTs on chip and than measured by Raman spectroscopy. A similar quenching of the Raman signals occurs if a thick layer covered around SWNTs.
However, within this SWNTs-SiOx experiment, no shift in the Raman frequencies was observable. It means that only van der waal interaction between silicon monoxide and SWNTs, therefore, only scattering effect could be detected.
Figure 7.3 (a) Micro Raman spectroscopy on the individual SWNTs with and without ZnO coated. Laser spot was focused coated SWNTs as shown in AFM images (b) and (c). (d) Bucky-paper material was also measured by the same Raman system.
Figure 7.4 shows a transmission electron microscopy (TEM) image of coated SWNTs. The image presents the coating of nanotubes with a thin layer of non-crystalline material. In order to confirm this thin non-crystalline layer, energy-dispersive x-ray measurements made on a VG501 scanning TEM was used to analyze the element composition. The EDS spectrum shown in Figure 7.4b was obtained with an electron beam spot size of 1.0 nm and the beam with this spot size is focused directly in the coated SWNTs to get the spectrum. The Zn, O, and a significant carbon peaks could be clearly obtained by long time detecting.
Figure 7.4 (a) and (b) show the low magnification TEM images of SWNT@ZnO.
The high magnification TEM image with EDS elements mapping were shown in (c) and (d).
7.3 Electrical transport in individual core-shell nanotube
For the electrical transport measurement, the coating SWNTs were deposited on a degenerately doped silicon substrate with a thermally grown SiO2 layer. Afterward Au/Pd electrodes were applied in two probe configuration by standard e-beam lithography. In our transport studies on SWNTs@ZnO, we have measured many individual tubes with ultraviolet (UV) light on them at room temperature. Figure 7.5a presents a set of Isd-Vg curves obtained from a typical device made by pristine
O Zn
SWNTs and these pristine SWNTs showed p-type FET characteristics. During the measurement of Isd, while we shined the UV light on the device, the current would slight decrease from 5x10-8 to 3x10-8 A. This phenomenon was called “photoinduced conductivity” and has been discussed by two research groups [152,153]. According to former articles, the photoinduced conductivity in the pristine SWNTs maybe due to (i) hot electrons attach to adsorbed molecules to induce desorption or (ii) contact electrodes were oxidized to change the barrier between electrode and SWNTs by UV light. In the (i) case, it should be difficult observed on metallic nanotubes because of large Isd current could not easy compensate by photoinduced effect. Additionally, if photoinduced conductivity was from contact electrodes oxidized, the device should be not reproducible because of electrodes were hard to recover as original one. However, we also could observe the similar transport characteristics on the coated SWNTs with UV light. For the coated SWNTs, firstly we have measured the Isd as the Vg changes from -5.0 to 5.0 V. As a result, some of the coated SWNTs exhibit distinct p-type characteristics and other tubes show metallic behavior. No matter ZnO coated on semiconductor or metallic SWNTs, we can easy observe the Isd current decrease drastically while shining the UV light on device. We suggest that current reduction behavior in coated nanotubes is a non-thermal process and is attributed by electronic compensation. As the UV excited the electron to conduction band, it will create some holes carrier at the ground state. At the same time, the electrons transported through the SWNTs will be compensated by the ground state hole carrier to reduce current. If we turned off the UV light, the e-h compensation behavior would disappear and than the current will increase again.
Figure 7.5 (a) Conductance vs gate voltage (Vg) of pristine SWNT recorded under a bias voltage of 1 V (scalar bar is 1 µm) and (b) SWNT@ZnO under a Vsd at 1500 mV in air and via UV exposure.
7.4 Ultraviolet photoresponse of single ZnO nanorod
ZnO nanorods were fabricated by soft chemical process and vapor-liquid-solid method from our lab and ITRI in Taiwan, respectively. The typical length of the resulting nanowires was 1 µm, with typical diameters in the range of 30–50 nm.
Selected-area diffraction patterns showed the nanorods to be single crystal. They were
released from the substrate and then transferred to SiO2 coated metallic Si substrates.
E-beam lithography was used to pattern sputtered Pd-Au electrodes contacting both ends of a single nanorod. The separation of the electrodes was 500 nm. Atomic force microscopy (AFM) of the completed device is shown in Figure 7.6. Au wires were bonded to the contact pad for current voltage (I–V) measurements performed at 25 oC in air. In some cases, the nanorods were illuminated with above band gap light mixed with 254 or 366 nm from Hg lamp. Figure 7.7 shows the I–V characteristics of the nanorod in the dark. The conductivity is nearly non-changed as applying voltage, even at high bias. It suggests that terrible ohmic contacts were formed between the nanorod and electrodes. However, in photoresponse measurement, the conductivity is greatly
E-beam lithography was used to pattern sputtered Pd-Au electrodes contacting both ends of a single nanorod. The separation of the electrodes was 500 nm. Atomic force microscopy (AFM) of the completed device is shown in Figure 7.6. Au wires were bonded to the contact pad for current voltage (I–V) measurements performed at 25 oC in air. In some cases, the nanorods were illuminated with above band gap light mixed with 254 or 366 nm from Hg lamp. Figure 7.7 shows the I–V characteristics of the nanorod in the dark. The conductivity is nearly non-changed as applying voltage, even at high bias. It suggests that terrible ohmic contacts were formed between the nanorod and electrodes. However, in photoresponse measurement, the conductivity is greatly