1 Introduction
Nanomaterials exhibit unique physical properties mainly due to the large surface/volume ratio and the confinement effects. Among the various nanomaterials, one-dimensional (1-D) semiconductor nanostructures are considered to be promising building blocks for future electronic and optoelectronic nanodevices owning to their abilities to effectively transport electrons and photons [
1–
4]. A variety of miniaturized optoelectronic devices, e.g., light emitting diodes, nanolasers, detectors, and optical waveguides have been fabricated successfully based on individual 1-D nanostructures [
5–
17]. The above recent developments in nano-optoelectronics require not only nanoscale manipulation and processing of semiconducting nanostructures, but also characterization of optical, electrical and structural properties with resolution and sensitivity of individual nanostructures. In addition, the electrical and optical properties of optoelectronic nanodevices are closely correlated and are often required to be measured simultaneously. If we further consider the considerable inhomogeneity of as-grown nanomaterials, co-measurements of different aspects of the nanostructures, e.g., microstructure, composition, and electrical and optical properties, based on the same individual nanostructures are often essential for a direct “structure-properties” correlation [
18,
19].
The recent instrumentation developments of electron microscopy, a major tool of microstructural and compositional characterization of nanomaterials, have incorporated nanomanipulation and electrical measurement [
20–
25]. This trend is represented by the scanning probe stage inside transmission electron microscope (TEM) [
20,
21] and the nanoprobe technique inside scanning electron microscope (SEM) [
22–
25]. For example, by controlling nanometer-sized metal tips with nano-manipulators, electrical contacts can be made to individual nanostructures inside SEM, thus
in situ electrical measurements can be carried out [
22–
24]. For optical characterization of individual nanostructures, micro-photoluminescence (micro-PL), [
26–
28] scanning near field optical microscopy [
28–
30] and cathodoluminescence (CL) inside SEM [
30–
32] are often employed. The integration of CL and SEM or a scanning transmission electron microscope can be traced back to several decades ago [
33,
34]. Compared to the optical apparatuses attached TEM or STEM [
33,
35], the SEM based CL spectroscopy and imaging have been used more widely in characterizing semiconductors due to the advantages of larger space, easier integration of cold stage and relatively lower damages. A typical modern SEM-based CL attachment includes a paraboloidal mirror just above the sample. The electron beam goes through a small hole on top of the mirror and excites the sample, while the generated luminescence is collected by the mirror and directed to the analyzer [
36,
37]. A drawback of this type of optical attachment is that the sizeable mirror may block most other detectors and the nanoprobe attachment, thus hindering the co-measurement of electrical and optical properties of nanodevices and nanomaterials.
Partly driven by the challenges raised by the 1-D optoelectronic nanostructures, we have adopted two strategies to combine optical techniques and in situ scanning electron microscopy to achieve comprehensive characterization of the same individual nanostructure. In this review, we summarize our recent technical developments and their applications in studying the deep level emission and the waveguiding behavior in 1-D ZnO nanostructures, and in situ assembly and characterization of optoelectronic nanostructures for device purposes. Using these examples, we demonstrate that the combination of optical techniques and electron microscopy can be useful for nanomaterial and nanodevice studies.
2 Strategies to integrate optical characterization and in situ electron microscopy
Two straightforward technical approaches have been adapted: (1)
in situ electron microscopy and micro-photoluminescence are used separately to locate and measure the same individual nanostructure in a compatible sample [
18]; (2) a comprehensive characterization system is assembled by combining space-saving fiber optics and
in situ nanomanipulation technique inside a scanning electron microscope [
19].
2.1 Combination of micro-PL and in situ electron microscopy
Figure 1 illustrates the comprehensive characterization on an individual suspended ZnO nanowire [Fig. 1(a)] by the combination of
in situ SEM and micro-PL. We prepared the suspended nanowire samples by attaching an individual nanowire to nanometer-sized tungsten (W) tip using
in situ SEM nanomanipulation and localized carbon deposition, which has been described in Refs. [
19,
21]. Such “nanostructure-on-a-tip” configuration can also be used for micro-PL measurement. Using a chain of ZnO nanowires with special features, we show in Fig. 1(b) that the same nanowire can be located in the same orientation using SEM and a micro-PL optical system. Employment of such suspended individual nanowires provides a technique compatible with established electrical measurement using the
in situ nanoprobe technique inside SEM [
24]. Figure 1(c) is a typical
in situ I–
V curve of the individual ZnO nanowires with the W tip and Pt wire as electrodes. Figure 1(d) shows a typical PL spectrum of suspended individual ZnO nanowires. Similar to some of the previous reports [
38], the PL spectra of the as-grown nanowires consist of a UV near band edge (NBE) emission peak and a broad defect-related deep level emission at
520 nm. Thus, microstructural (SEM), electrical (nanoprobe technique) and optical (PL) characterization can be carried out on the same nanowire.
2.2 Integrated system based on a SEM
Figure 2(a) shows a schematic configuration of the comprehensive characterization system assembled based on SEM. To incorporate optical characterization in the SEM, one or two optical fiber probes are attached to nano-manipulators, while the rest of the four manipulators still operate nanometer-sized W tips. Figure 2(b) is a digital photograph showing an optical fiber probe and two sharp metal tips operated by nanomanipulators. The outside end of the fiber probe is coupled to either a spectrometer for light analysis or a laser for optical excitation. The inside end of the fiber probe is either flat to enhance the angular resolving power or with a tip micro-lens to increase the light collection efficiency. The nano-manipulators enable 3-dimensional and nanometer accuracy movement of the fiber probes. Thus, the fiber probe can be routinely placed several microns away from the interesting areas for high efficiency optical detection or excitation. In contrast to the paraboloidal mirror light collector which blocks most other detectors, the space-saving design of the fiber probe component permits simultaneous action of the fiber probe and other SEM attachments (e.g., nanoprobe and EDS). By combining the above fiber optical attachment and the well-established in situ SEM techniques, a variety of characterization functions can be performed on the same individual nanostructures.
3 Characterization of surface defects and deep level emission in ZnO nanowires
Because of the much increased surface/volume ratio compared to the bulk materials, the surface defects often play an important role in the physical properties of 1-D nanostructures. For example, the so-called “green” emission in ZnO nanostructures has been commonly considered to be related to surface defect and have complicated influences on the optical and electrical properties of the ZnO nanostructures [
38–
42]. Such defect-related deep level emission has been widely investigated; however, its origin is still under debate [
38–
42]. Using
in situ burnt-out ZnO nanowires, we show that the comprehensive characterization yields insights on the green emission of ZnO nanowires.
A~3 μA current was applied to a
20-μm-long nanowire suspended between two tungsten (W) tips. The accumulated heating eventually resulted in a “burnt-out” near the center of the nanowires. Figure 3(a) and (b) shows SEM images of the nanowires before and after the burnt-out. The variations of O/Zn ratio and electrical conductance were monitored by energy dispersive X-ray spectroscopy (EDS) and electrical transport measurements using the W tips as electrodes [Fig. 3(c) and (d)]. The EDS spectra [Fig. 3(c)] taken from the same position before and after heating, indicates a remarkable oxygen loss (>5% in the areas close to the “burnt-out” point) during the
in situ annealing process. On the other hand, the electrical conductance [Fig. 2(d)] first increases and then decreases even before the observable change in the nanowire shape. This is probably corresponding to the increase of carrier density and decrease of carrier mobility during the
in situ annealing process in vacuum. The upper part of the “burnt-out” nanowire was then taken out of the SEM for micro-PL measurement. Figure 3(e) compares micro-PL spectra of the unheated nanowire and along different locations of the nanowire after the “burnt-out”, showing a dramatically increased green emission and a noticeable UV emission redshift with the temperature increase. The correlation of PL, electrical transport and EDS measurements indicates that the UV emission redshift, green emission, and carrier density are closely related to the oxygen deficiency, which supports native defect complexes and/or hydrogen at oxygen sites serving as stable and effective shallow donor [
43,
44].
The deep level emission was also found to be sensitive to the surface condition controlled by growth or sample processing. Figure 4 shows the influence of the surface morphology on the deep level emission in ZnO nanowires with varying diameters. For those nanowires possessing circular cross section and smooth surface [corresponding to straight boundary and uniform contrast in the side view, Fig. 4(a)], the green emission is normally 1–2 orders lower than the UV emission and does not show clear dependence on the nanowires diameters [Fig. 4(b)]. On the other hand, the nanowires with irregular cross section and rough surface (slightly wavy boundary and varying contrast in the side view) usually exhibit much stronger “green” emission intensities accompanied by weakened and redshifted UV emission [Fig. 4(c) and (d)]. Unlike those studies showing anomalous blueshift in UV emission with decreasing nanowire diameter [
41,
42], our results show that the UV emission peak energy has little dependence on the diameter but exhibits a strong tendency to redshift with the increase of the “green” defect emission, indicating defect induced modification of the electronic structure or an increased exciton-phonon interaction.
For micro-PL studies on individual nanowires, the nanowires are usually sparsely dispersed on a flat substrate [
45–
47]. The employment of individual suspended nanostructures can effectively avoid the complex influences of the substrates and the often necessary sonication process in solutions. We have observed that the sonication process may cause dramatically decreased UV emission and increased defect emission. Figure 1(c) compares the PL spectra of a sonication processed ZnO nanowire lying on SiO
2/Si substrate and an individual suspended as-grown ZnO nanowire. Much stronger green emission and redshifted NBE emission are observed for the sonication processed ZnO nanowire, likely being resulted from the surface change of the nanowire after solution sonication.
4 Characterization of optical confinement in 1-D ZnO nanostructures
1-D nanostructures may exhibit confinement on the photons or excitons, especially in the radial direction, which results in intriguing anisotropic optical properties. The optical confinement plays important roles in the reported optoelectronic nanodevices based on 1-D nanostructures, for example, light emitting diodes [
48], nanolasers [
6,
9,
16,
17] and waveguides [
8,
49]. Photoluminescence (PL) imaging, and spatially resolved PL spectroscopy using confocal and scanning near field optical microscopes have been employed to study the emission from the ends and the sidewall positions of 1-D nanostructures [
29,
50–
55]. For such measurements on individual 1-D nanostructures, the direction of the luminescence collection was fixed to be perpendicular to the axis of the 1-D nanostructures. Such experimental condition limits the excitation and detection angles relative to the individual nanorods/nanowires so that the direction perpendicular to the nanorod/nanowire axis is usually the only practical direction for both excitation and detection. Here we show that by using the techniques described in this paper, the angular dependent PL measurement of an individual nanostructure can be achieved.
First, we employ micro-PL spectroscopy on the “nanostructure-on-a-tip” samples prepared by
in situ nanoprobe technique. The flexibility of the individual suspended nanostructures attached to sharp metal tips allows large angle rotation and tilting, which makes it possible to carry out angular dependent micro-PL measurements on individual nanorods. In Fig. 5 we show micro-PL spectra acquired along the perpendicular and parallel to the axis of the same individual suspended nanorod. The emission from the parallel direction can be two orders of magnitude higher than that from the perpendicular direction, which demonstrates the waveguiding efficiency quantitatively [
56].
For a standard micro-PL setup, an objective lens is used for both excitation and signal collection, thus the excitation and collection directions cannot be separated for a large angle. In contrast, the integrated system (Fig. 2) hires separated excitation (the electron beam) and collection (the fiber probe). In addition, the fiber probe has an adjustable collection angle. Thus, more flexible and quantitative characterization on the anisotropic light emitting properties of nanostructures can be achieved. In Fig. 6, we show angular resolved CL spectra of a single ZnO nanorod at different orientations with respect to the fiber probe. By rotating the sample stage, the angle between the nanorod axis and the projection of the fiber axis on the substrate can be varied [Fig. 6(a) and (b)]. Figure 6(c) shows CL spectra taken with the electron beam focused on the same position near one end of the nanorod for the orientation angles of 0° [Fig. 6(a)], 90° [Fig. 6(b)], and 180°, respectively. It is found that the CL intensity reaches a minimum at 90° (
17 times lower than that at 0°), while a redshifted NBE emission peak is observed at 180°. These results are in good agreement with the waveguiding behavior of the ZnO nanorods [
8,
56]. The generated light mainly propagates along the nanorod axis (
c axis) and exits from the two end facets of the nanorod. To further understand the redshift caused by longer propagation distance, similar measurements have been carried out at 81 K, and the results are shown in Fig. 7. As the propagation distance increases, the LO phonon replicas of the free excitons (FX-1LO) are enhanced dramatically, indicating that the redshift of the NBE emission at room temperature is caused by the increased spectral weight of the LO phonon replicas.
The above results are consistent with our micro-PL measurements on the nanorod Fabry–Pérot cavities, which demonstrate the important role of LO phonons in the exciton polariton relaxation [
55]. In addition, the waveguiding behavior and the redshift due to the enhanced exciton-phonon interaction can also be used to explain some observed anomalous blueshift of the NBE emission at reduced nanowires/nanorod diameters. For example, Fig. 8 shows macro-PL spectra from arrays of ZnO nanorods with average diameters ranging from 450 nm to 80 nm [
57]. The room temperature NBE emissions show gradual blueshift with the decreasing diameter. However, the temperature dependent PL show that the peak positions of the FX and FX-
nLO emissions are the same for different samples at 10 K, while the blueshift of the smaller diameter samples are caused by the reduced exciton-phonon interaction due to the degraded waveguiding properties for small diameter nanorods [
58].
5 In situ assembly and characterization of optoelectronic nanostructures for device purposes
The above integrated system also allows
in situ assembly and characterization of nanostructures with simple optoelectronic functions. Figure 9 shows the assembly and characterization of an
in situ light emitter based on a single
n-ZnO nanowire dispersed on p
+-Si substrate. Figure 9(a) and (b) are the top-view SEM image and the side-view schematic of the assembled structure, respectively. A W tip is driven by a nano-manipulator to make contact with the nanorod. An electrical voltage is applied between the W tip and the substrate.
I–
V measurement indicated a typical p–n junction characteristic [
19]. Because of the unfavorable band alignment [
59], no luminescence has been detected in the contact mode even at an injection current of tens of μA. However, as the W tip was moved upward to form a nanometer-sized gap between the nanorod and the tip, strong tunneling-induced luminescence was observed at
30 V positive voltage between the substrate and the W tip. The corresponding tunneling current was about several μA. The tunneling luminescence shows dominating narrow NBE emission, indicating a potential application in high efficiency narrowband UV light emitting nanodevices [Fig. 9(d)].
In Fig. 10, we demonstrate the assembly and characterization of an
in situ photodetector based on an individual suspended Bi
2S
3 nanowire placed between two W tips [Fig. 10(a)] via
in situ nano-manipulation [
19,
60]. A laser beam (633-nm He–Ne laser) was used to ill-uminate the nanowire through a fiber probe. A constant 5-V voltage was applied to the nanowire. Figure 10(b) shows that varied excitation intensity leads to a current variation between 23 nA and 113 nA due to the generated extra carriers.
Such in situ technique with high flexibility and efficiency can be developed into an effective route to explore optoelectronic nanodevices, especially in building prototype devices and selecting suitable nanostructures for device purposes. Compared to the typical nanodevice fabrication process, the above in situ assembly and characterization technique can avoid some complications, e.g., nanostructure dispersion, lithography, vacuum plating, and lift-off due to its simple process.
6 Summary
In this paper, we review our recent progress of in situ characterization of optoelectronic nanostructure and nanodevices. By combining optical techniques and in situ scanning electron microscopy, we have developed two approaches allowing comprehensive microstructural, compositional, electrical and optical characterization of the same 1-D nanostructure. The first approach is to locate and characterize the same individual nanostructure in a compatible sample using a variety of characterization techniques. The second is to assemble an integrated system by combining optical fiber probe and in situ SEM techniques. By using these two techniques, the origin of the deep level emission in ZnO nanowires and the waveguiding behavior of the ZnO nanorods were studied. In addition, the integration approach allows in situ assembly and characterization of optoelectronic nanostructures for device purposes. Using the above examples, we demonstrate that the combination of optical techniques and in situ electron microscopy can be a powerful tool for nanomaterial and nanodevice studies.
©Higher Education Press and Springer-Verlag Berlin Heidelberg 2010
PDF
(504KB)
