The Motivation And Structure of This Dissertation

Chapter 1 Introduction

1.5 The Motivation And Structure of This Dissertation

Nano films have been widely applied in the numerous research fields because of their potential properties. In this dissertation, we use three kinds of ultrathin oxide layers (oxide nano films) to realize two OBDs and one solar cell for giving alternative solutions to the physical limit in the future scaling technology and for extending the functions of solar cells, respectively.

In the beginning, we fabricate and analyze an OBD with an n-Si/Alq₃/Al structure, which an ultrathin Al-O compound layer forms unintentionally at the Alq3/Al interface.

According to published results, the properties of interface oxide layers between the organic

layer and the electrode make influences on the characteristics of OBDs. In order to understand the effect of the ultrathin Al-O compound layer on our OBD, we investigate the characteristics of OBDs with the same structure made at different Alq3 deposition rates.

Moreover, we study an OBD using a nanostructured MoO_x layer interposed by two Alq3 layers. The nanostructured MoOx layer can be trap sites to produce the resistance switching of the OBD under electrical stress due to the energy barrier between MoO_x and Alq3. In organic electronics, MoOx has been extensively used to improve their performance. It suggests that the OBD can be easily integrated with other organic electronics.

In addition to the two OBDs, ultrathin oxide layers also play an important role on the photovoltaic features of Schottky solar cells. Inserting ultrathin oxide layers in the Schottky solar cells, metal-insulator-semiconductor (MIS) solar cells, can provide larger Voc under sunlight. MIS solar cells with high V_oc appear a promising application, converting solar energy for hydrogen production, and can reduce the contact loss of photovoltaic modules.

However, the V_oc of MIS solar cells is not larger than the dissociating potential of water (1.23 V). A stacking MIS solar cell is proposed to further enhance Voc, and has potential to provide a V_oc beyond 1.23 V.

This dissertation comprises six chapters. Chapter 1 gives the background and motivation of the studies. Then, the fabrication and characterization methods of samples are presented in Chapter 2. The following two chapters pay attention to the characteristics of an OBD with an interfacial oxide layer, and then involve the electrical features of an OBD with a nanostructured oxide layer. Materials in the Chapter 5 demonstrate a high Voc stacking MIS solar cell for a future photovoltaic application in water splitting. Finally, the conclusions and future works of this dissertation are summarized in Chapter 6.

Chapter 2 Fabrication And Characterization Methods

2.1 The Fabrication of Organic Bistable Devices With Interfacial Oxide Layers

The OBDs consists of an Alq3 thin film interposed between n-Si and an Al electrode, as shown in Fig. 2.1. The fabrication procedures of the OBDs are described as follows and are illustrated in Fig. 2.2. First, 1 Ω-cm resistivity n-type silicon substrates were clean through the standard RCA cleaning process, as shown in Fig. 2.3. Then, 150-nm-thick Alq3 thin films were deposited on the cleaned substrates kept at room temperature by thermal evaporation in a vacuum below 3 x 10^-6 Torr. Finally, 80-nm-thick Al top-electrodes were deposited on the Alq3 thin films through shadow masks. The area of each Al electrode is 0.64 mm². The deposition rates of the Alq3 thin films are 0.05 nm/s, 0.15 nm/s, 0.2 nm/s, and 0.3 nm/s. The

Fig. 2.2 The fabrication flow of OBDs with interfacial oxide layers.

Fig. 2.1 The structure of OBDs with interfacial oxide layers

deposition rates are controlled by setting the temperature of crucibles and are monitored by the quartz crystal microbalance. The corresponding setting temperature for each deposition rate is listed in Table 2.1.

2.2 The Fabrication of Organic Bistable Devices With Nanostructured Oxide Layers

The OBDs consist of an Alq3/MoOx/Alq3 tri-layer structure interposed between top and bottom electrodes, as shown in Fig. 2.4, and fabrication procedures are illustrated in Fig. 2.5.

Table 2.1 The setting temperatures corresponding to the four deposition rates of Alq₃ thin films, and the properties of the Alq₃ thin films obtained from XPS measurements for the four deposition rates.

Fig. 2.3 The RCA cleaning process

First, 50-nm-thick Alq3, XX-nm-thick MoOx, and 50-nm-thick Alq3 thin films were evaporated in sequence onto cleaned p⁺-type silicon substrates. The XX is equal to 3, 5, and 8.

The average deposition rates of the Alq3 thin films and that of the MoOx layers are about 0.1 and 0.01 nm/s, which are monitored by the quartz crystal microbalance, respectively. Finally, 100-nm-thick Al thin films as the top electrodes were evaporated through metal masks with 2 mm x 2 mm square patterns. Note that all materials were evaporated at a vacuum bellow 3x10⁻⁶ Torr and the substrates were kept at room temperature during fabrication.

2.3 The Fabrication of Metal-Insulator-Semiconductor Solar Cells

The thickness of the insulating layers of MIS solar cells affects the blocking efficiency of their majority carriers and the tunneling probability of their excess minority carriers; hence, it influences their potency. Although thermal SiO₂, [151]-[157] chemical SiO₂, [158]

evaporated SiOx, [159] and SiO2 using an anodization technique [160] have been used as their Fig. 2.5 The fabrication flow of OBDs with nanostructured MoO_x layers.

Fig. 2.4 The structure of OBDs with nanostructured MoOx layers.

ultrathin insulating layers, controlling the thickness of the insulating layers well is still difficult to achieve. To easily control the thickness and to produce the insulating layers at the low temperature, radio frequency (RF) magnetron sputtering is adopted to deposit the ultrathin SiO₂ layers.

The n/p-type MIS solar cells were fabricated using phosphor/boron-doped (100) silicon substrates with resistivity ranging from 1 to 10 Ω-cm. The fabrication procedures are shown step by step in Fig. 2.6. First, ultrathin SiO2 layers were sputtered on n/p-type silicon substrates, which had cleaned through the standard RCA cleaning process, by RF magnetron sputtering under different working pressures. The thickness of the ultrathin SiO2 layers is controlled by sputtering duration. The optimized thicknesses for the n-type and p-type silicon substrates are about 2 and 1 nm, which the ultrathin SiO2 layers were sputtered at 20 mTorr, respectively. Thicker sputtering SiO₂ insulating layers are required for the n-type MIS solar cells to suppress the larger tunneling probability of the majority carriers.

Then, most samples were annealed in H₂ atmosphere at 500℃ for 1 hour, but the others are not for being reference samples. Finally, Al/Au(-Ni) films as the back electrodes, semi-transparent 20/10-nm-thick Au/Al thin film layers on the top of the ultrathin SiO₂ layers, and Au(-Ni)/Al front finger electrodes were introduced by thermal evaporation. The area of all MIS solar cells is 0.2 cm².

Fig. 2.6 The fabrication flow of MIS solar cells

2.4 The Fabrication of the Stacking Metal-Insulator-Semiconductor Solar Cell

The proposed stacking MIS solar cell is composed of an n-type MIS solar cell as the top cell and a p-type MIS solar cell as the bottom cell with a tunneling junction in between. The structure is shown in Fig. 2.7, and its fabrication procedures are illustrated in Fig. 2.8. First, an n-n⁺ sample and a p-p⁺ one were fabricated using ion implantation and rapid thermal annealing (Sec. 2.4.1). Then, an n-n⁺-p⁺-p sample was prepared by bonding the n-n⁺ sample with the p-p⁺ one (Sec. 2.4.2). After the bonding process, both sides of the n-n⁺-p⁺-p sample were thinned by wet etching (Sec. 2.4.3). Since the n-n⁺-p⁺-p sample is too fragile for thickness measurement, the thicknesses are first roughly estimated by reference Si substrates using scanning electron microscopy. The thickness of the n-type Si substrate is about 90 μm and that of the p-type Si substrate is around 400 μm. The thickness of the n-type Si is designed to be much thinner than that of the p-type Si for current matching consideration. The optimal thickness for the n-type Si is about 4 μm. Finally, the sample was processed based on the optimized conditions obtained from the individual n-type and p-type MIS solar cells. The area of the stacking MIS solar cell is 0.2 cm².

Fig. 2.7 The structure of the stacking MIS solar cell.

2.4.1 The Fabrication of the n-n⁺ and p-p⁺ Samples

The stacking MIS solar cell needs a tunneling junction which photoexcited majority carriers in the top cell can recombine with those in the bottom cell. Before forming the tunneling junction, an n-n⁺ sample and a p-p⁺ one are required. Following is the fabrication procedures of them, as shown in Fig. 2.9. First, n-type and p-type Si substrates were cleaned through the standard RCA process. Then the polished surface of the n-type Si substrate was doped with arsenic ions by ion implantation with doses of 5x10¹⁵ cm^-2. The energy of ion implantation is 80 KeV and 20 KeV in sequence. Similarly, the polished surface of the p-type Si substrate was implanted with 24 KeV boron ions with a dose of 5x10¹⁵ cm^-2 and then with 10 KeV boron ions with a dose of 5x10¹⁵ cm^-2. The activation of the implanted ions, by rapid thermal annealing at 1050℃ for 10s, was executed after the bonding process.

2.4.2 The Bonding Process

After obtaining the implanted samples, the n-n⁺-p⁺-p sample with a tunneling junction can be generated through a bonding process. The bonding procedures are disclosed below, as shown in Fig. 2.10. First, the implanted samples were dipped with a diluted HF solution and Fig. 2.8 The fabrication flow of the stacking MIS solar cell.

were rinsed with deionized (DI) water. Then, the samples were cleaned with aceton (ACE) and isopropanol (IPA) solutions in an ultrasonic cleaning tank in sequence. After the cleaning processes, the cleaned n-type sample was contacted directly with the cleaned p-type one in an IPA solution at room temperature, and then the contacted sample was fixed in the fixtures.

Finally, the fixture was put into a heating chamber, and was heated at 500℃ for 1 hr in hydrogen atmosphere. The hydrogen can prevent the bonding interface from oxidation, which attributes to impurities in the chamber, during the heating process.

Fig. 2.10 The flow of the bonding process.

Fig. 2.9 The fabrication flows of the n-n⁺ and p-p⁺ samples.

2.4.3 The Thinning Process

For balancing the photon fluxes of the top and bottom cells, the thickness of the top cell must be thinned toward the optimized thickness (4 m). The thinning process, as illustrated in Fig. 2.11, is revealed as follows. First, the well bonded sample was immersed in a etchant (HF : CH3COOH : HNO3 = 6 : 7 : 20) for 12 to 13 minutes. Finally, the etched sample was rinsed using DI water.

2.5 Material And Interface Analyses

2.5.1 High Resolution X-Ray Photoelectron Spectrometer

High resolution X-ray photoelectron spectroscopy (HRXPS) can help deduce the chemical composition of materials on the surfaces. When monoenergetic soft X-rays with photon energy hν irradiated by solids in vacuum illuminate on the surface of a material, photoelectrons with kinetic energy Ekin will emit out of the surface and are collected by the analyzer, as shown in Fig. 2.12. The spectrum is a plot of the number of emitted photoelectrons per energy interval versus their binding energy Eb, as expressed by Eq. (2.1).

(2.1) The composition of the Alq₃/Al interfaces and the atomic concentration of the Alq₃ thin films were analyzed by measuring the Al-2p, and N-1s and O-1s photoelectron spectra, respectively, using HRXPS (ULVAC-PHI Quantera SXM) after Ar⁺ milling on the sample surfaces. The HRXPS equips with a monochromatic Al Kα radiation X-ray source and

Fig. 2.11 The flow of the thinning process of the bonded n-n⁺-p⁺-p sample.

a hemispherical energy analyzer. The power of the X-ray source is 25 W at 15 kV accelerating voltage, and the analyzed area of the samples is about 100 μm².

2.5.2 Grazing Incident X-Ray Diffraction

X-ray diffraction is a nondestructive approach to identify the crystallization structures of solids. For a sample with regular arrays of atoms, the atoms will scatter the waves when X-ray waves illuminate into the sample, as shown in Fig. 2.13. Diffraction patterns can be obtained as long as the phases of not less than two scattering waves superposition to meet Bragg‟s law, as expressed by Eq. (2.2), where d is the spacing between diffracting planes,

θ

^is

the incident angle, n is any integer, and λ is the wavelength of the X- ray beam. According to a plot of diffraction intensity verse diffraction angles, the crystallization structures of the sample can be identified.

λ (2.2) The structural information of the Alq3 thin films deposited at different deposition rates was analyzed via XRD (Rigaku, RU-H3R) using a grazing incident X-ray diffraction (GIXRD) mode. The incident X-ray was fixed at a tiny angle with respect to the surfaces of the Alq3

thin films, and the detector did 2θ scan.

Fig. 2.12 The X-ray photoelectron process of HRXPS.

2.5.3 Atomic Force Microscope

Atomic force microscope (AFM) is an apparatus whose primary component is an extremely sharp tip mounted or integrated on the end of an extreme small cantilever spring which is moved by a mechanical scanner over the surface of a sample, as shown in Fig.

2.14(a). AFM can provide the topography of bare and monolayer coated solids. When the tip is moved close to a sample surface, forces between the tip and the sample surface can result in a deflection of the cantilever owing to Hooke's law. According to the deflection detected by a laser beam, the morphology of the sample surface can be obtained. For conventional AFM, two operation modes are available, the contact mode (the tip onto the sample surface, repulsive force) and the tapping mode (the tip in close vicinity of the sample surface, attractive force), as shown in Fig. 2.14(b).

The surface morphology of the Alq₃ thin films on the Si substrates and that of the MoO_x layers on the Alq3 thin films were obtained by using AFM (DI-Veeco Instruments, D3100).

We chose the tapping mode to scan the sample surfaces because of the softness of organic thin films.

Fig. 2.13 The X-ray diffraction process.

2.5.4 Transmitted Infrared Imaging System

The bonding interface of the stacking MIS solar cell plays an important role in the performance of the stacking MIS solar cell because the interface is where the majority carriers of the top cell recombine with those of the bottom cell. If the interface cannot be a good tunneling junction and/or is not well bonded, the Voc enhancement of the stacking MIS solar cell will be blocked. In order to observe the interface, a transmitted infrared (IR) imaging system, which consists of a light bulb, a sample stage, a focus lens, an IR charge coupled device (CCD), and a screen, as shown in Fig. 2.15, was established.

The operation of the system is presented as follows. When photon fluxes, emitted from the light bulb, are incident into the bonded sample, photons with energy larger the bandgap of Si (1.124 eV) are absorbed by the bonded sample. The rest photons, whose energy less than 1.124 eV, are detected by the IR CCD. If any unbonded area (interface bubbles or voids) generates in the interface, fringe patterns, which result from the interference of the passing IR photons, can be observed on the screen.

Fig. 2.14 (a) The operation of AFM. (b) The forces between the AFM tip and the sample surface.

2.6 Device Characterization

2.6.1 Current-Voltage Measurement

The I–V curves of OBDs (the writing process, the erasing process, the reading process, the retention test, and the write-read-erase-read cycle) were measured in an ambient environment using a measurement system which is composed of a personal computer, a Hewlett Packard 4156A semiconductor parameter analyzer, a switching box, and a probe station. The writing bias, the erasing bias, and the reading bias are sweeping voltages and/or (quesi-)pulse voltages.

2.6.2 Capacitance-Voltage Measurement

A measurement system which includes a personal computer, an Agilent 4284A Precision LCR meter, a switching box, and a probe station was adopted to record the high frequency capacitance-voltage (C-V) curves of OBDs and MIS solar cells.

Fig. 2.15 The setup of the transmitted IR imaging system.

2.6.3 Photocurrent Measurement

The J-V curves of MIS solar cells were obtained using a solar simulator system which consists of a laptop, a Keithley 2400 source-measure unit, a probe station, and a solar simulator with an Air Mass 1.5 source with a power of 100 mW/cm²(Oriel Class A). The illuminated light intensity was calibrated by a polycrystalline silicon solar cell certified by the National Renewable Energy Laboratory (NREL).

Chapter 3 Organic Bistable Devices With Interfacial Oxide Layers

3.1 Organic Bistable Devices With Interfacial Layers

Interface layers between the (metal) electrodes and organic layers of OBDs play an important role in the characteristics of organic electronics. [161] For understanding the interface effects on OBDs, some have paid attention on the interface engineering of OBDs.

[35], [161]-[166] Kondo et al. demonstrate that the ON/OFF ratio of OBDs is significantly enhanced by modifying the ITO electrodes (introducing Ag nanodots between the organic layer and the indium tin oxide surface). The Ag nanodots act as trapping sites, reducing the current of the OFF state. [162] Cho et al. form a self-assembled monolayer (4-nitrophenyl dichloride phosphate) on the hydroxyl-terminated Al electrode surface with native oxide to enhance switching reproducibility with an improved current level distribution. [163]

Moreover, when an aluminum electrode is deliberately oxidized or unintentionally is passivated with native oxide, some organic devices display written, reproducible resistance switching. [35], [164]-[166] Cölle et al. show that the switching mechanism of OBDs with a (metal) electrode/organic/(metal) electrode structure (at least one electrode using Al) is due to the oxide layer on the Al electrode, and carriers transport through filaments, whereas the organic materials have only minor influences on it. [35] Verbakel et al. realize reproducible OBDs by deliberately adding thin sputtering Al2O3 layers to the surfaces of the bottom electrodes of nominal electron-only, hole-only, and bipolar organic diodes. They attribute the operation of the OBDs to the soft breakdown of the sputtering Al2O3 layers. In addition, the

OBDs exhibit negative differential resistance (NDR) and local maximum currents in the I-V curves, which are dependent on the thickness of the sputtering Al₂O₃ layers. In their case, the polymers play a role as a current limiting series resistance. [165] Cho et al. study OBDs with a interfacial AlO_x layer, using an O₂ plasma treatment method, formed on the surface of the bottom Al electrode. The OBDs give higher ON/OFF ratios than those without the interfacial AlO_x layer because of the relatively small resistance at the OFF states. Nevertheless, the duration of the O2 plasma treatment makes a great influence on the threshold voltage distribution and the switching reproducibility. [166]

In this chapter, an OBD with Alq3/Al deposited on an n-type Si substrate, as shown in Fig. 2.1, is fabricated and investigated. The OBD shows distinct bistability with an ON/OFF current ratio over 10⁶ and a wide reading voltage range for differentiating between „„ON” and

„„OFF” states. The formation of electrically bistable states is the result of electrons being trapped in the defects at the Alq3/Al interface during electrical field stressing. This study also provides a simple approach, varying the deposition rate of the Alq₃ thin film, using which the electrical characteristics of the OBD, e.g., threshold voltage, can be tuned or controlled.

HRXPS, AFM, and GIXRD measurements are performed to help us understand the properties of Alq3 thin films and the Alq3/Al interface and explain the experimental results obtained.

Besides, the simple structure of the OBD indicates that the OBD can be easily embedded into the well-developed semiconductor fabrication processes.

3.2 Results And Discussions

Figure 3.1 shows typical I-V curves of the fabricated n-type Si/Alq₃/Al structure. As can be seen, this device exhibits two different conductance states at an identical applied voltage. The silicon electrode is kept at 0 V, and all bias conditions are applied on the aluminum electrode. At the first bias (the black square curve in Fig. 3.1), the voltage sweeps from 0 V to 10 V. Initially, the device exhibits a low conductance state (OFF state). However,

在文檔中具超薄氧化層之記憶體與光伏元件 (頁 33-0)