Thin Film Deposition - Fabrication Process and Analyzing Method

Chapter 2 Fabrication Process and Analyzing Method

2.1.2 Thin Film Deposition

To analyze the characteristics of the undoped ZnO, Al-doped ZnO (AZO), and p-type ZnO thin films, we deposited single layer thin films. To analyze the characteristics of nano-crystalline silicon (nc-Si) quantum dot (QD) thin films, we deposited multilayer thin films. All of them were deposited by radio-frequency (RF) magnetron sputtering method.

For undoped ZnO and AZO thin films, we used two targets of pure ZnO and ZnO doped with 2 wt% Al₂O₃ and deposited them on p-type Si(100) wafers to form p-n junctions, shown in Fig. 2.3(a) and (b). During deposition, argon (Ar) gas was chosen as the working gas, and the working pressure was kept at 5.0x10^-3 Torr. The power of ZnO and AZO targets is fixed at 75 W. The thickness was about 100 nm. The

sputtering parameters in detail are shown in Table 2.1.

Fig. 2.3 The structure of (a) undoped p-type ZnO, (b) Al-doped ZnO, and (c) p-type ZnO thin films deposited on Si wafers.

Table 2.1 The sputtering parameters of undoped ZnO and AZO thin films.

For the multilayer thin films, the targets of pure ZnO, ZnO doped with 2 wt%

Al2O3, and Si were used. During deposition, Ar gas was chosen as the working gas, and the working pressure was kept at 5.0x10^-3 Torr. To tune the concentration of Al as matrix materials, we used the ZnO and AZO targets at the same time, and changed the power of them, shown in Fig. 2.4. We used the Si target to deposit amorphous Si and fixed the power at 110 W. The thicknesses of each AZO and Si layers were fixed at 5 nm and 3 nm, respectively. We deposited 20 pairs of AZO and Si layers and added 5 nm of AZO as a capping layer. Fig. 2.5 shows the structure of the multilayer thin films. Table 2.2 lists the sputtering parameters for different [AZO/Si]₂₀multilayer thin films. The co-sputtering AZO samples, such as AZ5075 and AZ5025, are named by the power of the ZnO and AZO targets. For example, the sample AZ5075 means the ZnO sputtering power is 50 W and the AZO sputtering power is 75 W. In order to separate the multilayer structure from the single-layer structure, the names of [AZO/Si]20 multilayer thin films are added ML.

Fig. 2.4 Operation of sputtering during co-sputtering ZnO and AZO targets.

Fig. 2.5 The structure of the multilayer thin films.

Table 2.2 The sputtering parameters of the [AZO/Si]20 multilayer thin films.

For the p-type ZnO thin films, we used the pure ZnO target and deposited them on n-type Si(100) wafers to form p-n junctions, shown in Fig. 2.3(c). High purity Ar and nitrogen (N₂) gas were introduced into the chamber. We used the N₂ as acceptor dopants to contribute to p-type properties. During Deposition, the total gas flow was fixed at 30 sccm. The Ar-to-N₂ flow was varied from 30:0 sccm to 10:20 sccm. The working pressure was kept at 5.0x10^-3 Torr. We also tried to change the working

pressure to 1.5x10^-3 Torr. At this pressure, due to limitations of the sputtering machine, we changed the total gas flow to 15 sccm. All p-type ZnO sputtering power was fixed at 100 W. The thickness was about 200 nm. The sputtering parameters in detail are shown in Table 2.3. The samples are named by the ratio of the N₂ and the working pressure. For example, the sample named as N17%-5 means the ratio of the N₂ is 17 % and the working pressure is 5 mTorr. In other words, the flow of the N₂ is 7 sccm (30×17%=7).

Table 2.3 The sputtering parameters of p-type ZnO thin films.

2.1.3 Annealing Process

After deposition of the multilayer thin films, these samples were annealed by the quartz furnace in a N2 atmosphere to form nc-Si QD thin films. We fixed annealing time at 5 minutes and changed annealing temperature between 700 to 1000 ℃. An illustration of the formation of nc-Si QDs in the AZO matrix is shown in Fig. 2.6.

Fig. 2.6 Illustration of the formation of nc-Si QDs embedded in the AZO matrix.

For the p-type ZnO thin films, we also annealed samples. We changed the annealing temperatures from 600 to 800 ℃ and times from 5 minutes to 1 hour in order to get p-type properties.

2.1.4 Electrode Deposition

Finally, contact electrodes were evaporated through the metal mask with 0.8 mm × 0.8 mm square pattern for the electrical properties measurement. For the p-type ZnO thin films, we deposited Ni electrodes (~120nm) on the top of samples and Al electrodes (~120nm) on the bottom of samples in order to reduce the influence of contact. For other samples, such as the undoped ZnO, AZO and multilayer thin films, opposite electrodes were deposited on the samples.

2.2 Analyzing Method 2.2.1 Four-Point Probe

Sheet resistance is one of the important characteristics to conductive materials.

Four-point probe is the most commonly used tool to measure the sheet resistance. The current between the two probes is fixed, while the voltage difference between the other two probes is measured, then the sheet resistance can be calculated.

For a conductor, the resistance can be written as [23]:

R = ρ^L_A= ρ_Wt^L (2.1)

where ρ is the resistivity, A is the cross-sectional area and L is the length. The cross-sectional area can be split into the width W and the thickness t. We can define the sheet resistance R_s as:

Rs =^ρ_t (2.2)

Thus, the resistance can be rewritten as:

R = Rs L

W (2.3) We use four-point probe (NAPSON, RT-80/RG-80) to measure the sheet resistance of thin films. And then, we check the thickness of thin films by N&K analyzer or alpha-step. Finally, we can understand the resistivity of the thin films from Eq. (2.2).

2.2.2 Current-Voltage Measurements

The current-voltage (I-V) curves were measured by using the E5270B 8-slot precision measurement mainframe (Agilent Technologies) and a halogen lamp was used for photo-response measurements. We can understand the formation of p-n junction between the thin films and the substrates by I-V characteristics.

In addition, we can use I-V characteristics to estimate the resistivity. The resistance is measured by the linear trend line of I-V characteristics. We can calculate the resistivity by combining the resistance with the eq. (2.1).

2.2.3 High Resolution X-Ray Diffractometer

X-ray diffractometer is a non-destructive analytical technique which reveals information about the crystallographic structure, chemical composition, and physical properties of materials and thin films. When x-ray irradiates crystalline materials,

constructive interference produces in some directions, determined by Bragg’s law:

2dsinθ = nλ

(2.5) where d is the spacing between diffracting planes, θ is the incident angle, n is any integer, and λ is the wavelength of the beam.

We can estimate the size of nc-Si QDs by using the Scherrer formula [12]:

R =Δ2θ(rad)∙cosθ^Kλ (2.6)

where L is the average nanocrystal size, λ is the x-ray wavelength, 2θ is the Bragg diffraction angle in radians, and K is a constant correction factor depending on the shape and size of the crystalline clusters, and on the direction of the diffracting planes.

We analyzed our samples by high resolution x-ray diffractometer (Bede, D1) with θ-2θ mode. The wavelength of the x-ray was 1.54 Å.

For pure ZnO peaks, (002), (101) and (102), appear at 2θ = 34.68°, 36.46° and 47.50°, respectively [24]. For ZnO:N films, two peaks at 2θ = 34° and 43° are attributed to the zinc nitride (321) and (332), respectively [16]. For nc-Si QD thin films, three peaks at 2θ = 28.4°, 47.4° and 56.3° are attributable to Si(111), (220) and (311), respectively [12].

2.2.4 High Resolution Confocal Raman Microscope

Raman microscope is a technique used to study vibrational modes in a material.

When the incident photons interact with the molecules, the electrons transit from the ground state to a virtual state. If the energy isn’t absorbed by the molecules, it is released through the scattering method. The released energy is the same as the energy of the incident photons, called Rayleigh scattering; otherwise, it is called Raman scattering [25]. Therefore, Raman microscope is a powerful and non-destructive technique to understand the physical and chemical properties of a material.

We analyzed our samples by high resolution confocal Raman microscope (HOROBA, Lab RAM HR), and used a 488 nm diode-pumped solid-state (DPSS) laser. Si wafers were used to calibrate the crystalline Si signal at 520 cm^-1 before measuring our samples.

For bulk ZnO, the peak at ~439 cm^-1 is E₂(high) mode associated with oxygen atoms and the one at ~591 cm-1 is E1(LO) mode [26]. For ZnO:N, the peak at 582 cm^-1 is attributed to the A₁ (LO) mode of ZnO:N [15,27]. The other four peaks with frequencies of 275, 510, 643, and 856 cm^-1 may be considered local vibrational modes which are related to nitrogen concentration in ZnO:N films. For nc-Si QD thin films, three peaks are generally detected, inclusive of the amorphous Si (a-Si) at ~480 cm^-1, the nc-Si between 510 and 520 cm^-1, and the intermediate mode from a-Si to nc-Si between 500 and 510 cm^-1 [28].

2.2.5 UV/VIS/NIR Spectrophotometer

We used UV/VIS/NIR spectrophotometer (Japan, Hitachi U-4100) to measure the transmittance T and reflectance R of our thin films. We also can find the absorbance of our thin films by

Absorbance = 100 − T − R (%).

(2.7) Thus, we can understand the optical properties of the thin films.

Otherwise, we can use the transmittance and reflectance spectra to estimate the optical bandgaps of the thin films. The absorption coefficient α is evaluated from T and R according to the following relation [7,29]:

α = ¹_tln �^1−R_T � cm⁻¹ (2.8) where t is the thickness of the films. The optical bandgaps (E_gopt) of the films are determined by the extrapolation of linear part of the absorption edge to α=0 in the

22 relationship as [29]

(αhv)^γ= B(hv − Egopt)

(2.9) where h is Planck’s constant, v is the frequency of the radiation, and B is the edge width parameter. The value of γ is dependent on the Egopt behavior, such as γ=2 for direct E_gopt and γ=1/2 for indirect Egopt.

Chapter 3 Characterization of the nano-crystalline Si quantum dots embedded in the Al-doped ZnO Thin Films

Before analyzing the characterization of the nano-crystalline silicon (nc-Si) quantum dots (QDs) embedded in the Al-doped ZnO (AZO) thin films, we would like to know the characteristics of the AZO thin films and compare the properties with the ZnO thin films.

And then, we explored the nc-Si QDs embedded in AZO thin films from different annealing temperatures, different Al concentrations, and the comparison of the nc-Si QDs embedded in the ZnO and AZO matrix. We analyzed the characteristics by using Raman microscope, UV/VIS/NIR spectrophotometer, and current-voltage (I-V) characteristics. We can obtain the crystalline phase of Si for the nc-Si QD thin films by Raman spectrum. We can know the optical properties and estimate the optical bandgaps of the nc-Si QD thin films from UV/VIS/NIR spectrum. From I-V characteristics, we can understand the rectification characteristics and resistivity of the nc-Si QD thin films.

3.1 Characterization of the Al-doped ZnO Thin Films

We used the targets of pure ZnO and ZnO doped with 2 wt% Al2O3 to deposit the ZnO and AZO thin films, and the target of ZnO doped with 2 wt% Al₂O₃ to deposit AZO thin films. The sputtering parameters in detail are shown in Table 2.1.

From Hall measurements, the AZO and ZnO thin films show the n-type properties, which is resulting from the natural defects of oxygen vacancies and zinc interstitials [15].

The resistivity of the AZO and ZnO thin films annealed at 1000 ℃ for 5 minutes is 4.5×10^-2 and 1.3×10^-1Ω-cm. The resistivity decreases with the Al dopants. This can be caused by the replacement of Al³⁺ for Zn²⁺ [10].

Fig. 3.1(a) shows the transmittance of the AZO thin films annealed at 1000 ℃ for 5 minutes. The average transmittance of all the films is about 80% in the visible region. All the films have a sharp fall in the transmittance below 400nm because of the band edge absorption. The optical bandgaps of the AZO and ZnO thin films are determined by the eq. (2.9). Because ZnO is a direct bandgap material [30], the value of γ is 2. The curves (αhν)² as a function of energy (hν) of the incident radiation is shown in Fig. 3.2(b). The inset shows the optical bandgaps of the AZO and ZnO thin films annealed at 1000 ℃ for 5 minutes. The widening of the optical band gap with Al dopants follows the Burstein-Moss effect [8].

Fig. 3.1 (a) Transmittance and (b) plots of (αhν)² versus (hν) for the AZO and ZnO thin films annealed at 1000 ℃ for 5 minutes. The inset in (b) shows the optical

bandgaps of the samples.

The conductivity of the AZO thin films is better than the ZnO thin films. Thus, we want to embed the nc-Si QDs in the AZO thin films. We hope the conversion efficiency of the nc-Si QD solar cells increases by using the good conductive properties of the AZO thin films.

3.2 Influence of Different Annealing Temperatures

From now on, we start to discuss the nc-Si QDs embedded in the AZO thin films.

The sputtering parameters in detail are shown in Table 2.2.

First, we explored the influence of different annealing temperatures on the nc-Si embedded in the AZO thin films. We focused on the sample AZ5075-ML to change the annealing temperatures from 700 to 1000 ℃ and fix the annealing time at 5 minutes.

Fig. 3.2 shows the Raman spectra of the sample AZ5075-ML annealed at different temperatures. As the annealing temperature increases, the silicon peak moves from the intermediate mode (500~510cm^-1) to the nano-crystalline mode (510~520cm^-1).

This means a high annealing temperature is needed to form nc-Si QDs.

Fig. 3.2 Raman spectra of the sample AZ5075-ML annealed at different temperatures.

Fig. 3.3 shows the optical transmittance and absorbance spectra for the sample AZ5075-ML annealed at different temperatures. The transmittance increases with increasing the annealing temperature, while the absorbance decreases with increasing the annealing temperature. Changing in the transmittance is resulted from the better crystal quality of the AZO matrix after annealing at high temperatures [8,31]. In other

words, the better crystal quality of the AZO matrix after annealing means the reduction of defects, which cause the absorbance decreases.

Fig. 3.3 (a) Transmittance and (b) absorbance spectra of the sample AZ5075-ML annealed at different temperatures.

The optical bandgaps are determined by the eq. (2.9). The value of γ is 1/2 since a phonon-assisted transition dominates the optical processes for a small nc-Si QD size [29]. Here, we select the sample AZ5075-ML annealed at 900 ℃ for 5 minutes as a example for how to find the optical bandbap, shown in Fig. 3.4. The optical bandgap for the sample AZ5075-ML annealed at 900 and 1000 ℃ is 1.97 and 2.08 eV, respectively. The optical band gap increases with the increase of the annealing temperature, which means the decreases in nc-Si QD size [32].

Fig. 3.4 The example for how to find the optical bandbap. The sample is AZ5075-ML annealed at 900 ℃ for 5 minutes.

We changed different annealing temperatures to search an appropriate annealing temperature. The samples AZ5075-ML annealed at 900 and 1000 ℃ form the nc-Si QDs. The optical bandgap of the sample annealed at 1000 ℃ is wider than the sample annealed at 900 ℃, which means the decrease of the nc-Si QD size.

3.3 Influence of Different Al Concentrations

Here, we focus on the influence of different Al concentrations. All the [AZO/Si]₂₀ thin films were annealed at 1000 ℃ for 5 minutes to form the nc-Si QDs in the AZO matrix. We use different powers of the AZO and ZnO targets to change the Al concentrations. In order to know the concentration of Al in the AZO thin films, we deposited the sample AZ75, AZ5075, and AZ5025 with the thickness of about 100nm and then measured these films by X-ray photoelectron spectroscopy (XPS), listed in Table 3.1. The Al concentration of the film deposited by a single AZO target is larger than the co-sputtering films. When the sputtering power of AZO is higher, the Al concentration of the films is larger.

Table 3.1 The concentration of atoms in the AZO thin films measured by XPS.

Fig. 3.5 shows the Raman spectra of the samples with different Al concentrations annealed at 1000 ℃ for 5 minutes. The peaks of all the samples are in the range of 510 to 520 cm^-1, which represents the formation of nc-Si QDs in all the samples. In order to understand the individual contribution of the Raman signal, we fitted the Raman data of all the samples by using the program of PeakFit v4. Four peaks of the

E₂(high) mode of ZnO at ~439 cm^-1, the amorphous Si (a-Si) at ~480 cm^-1, the intermediate mode between 500 and 510 cm^-1, and the nc-Si between 510 and 520 cm^-1 were considered to fit the Raman data. The Raman spectrum and correspondingly fitting curve for the sample AZ5025-ML are shown in Fig. 3.6 as an example. Different ZnO and Si modes were considered in the best fit procedure, whose peak position, intensity area, full width at half maximum (FWHM) and crystallinity are recorded in Table 3.2. The crystallinity X_C was determined from the following equation [33]

X_C =_I^I^nc^+Iⁱ

nc+I_i+Ia× 100% (3.1) where I_nc, I_a, and I_i are integrated intensities of the nano-crystalline, amorphous, and intermediate modes of Si, respectively. The FWHM increases with the increase of Al concentration. The increase of the FWHM means the decrease of the nc-Si QD size.

We deduce that the presence of Al may limit the nc-Si QD size.

Fig. 3.5 Raman spectra of the samples with different Al concentrations annealed at 1000 ℃ for 5 minutes.

Fig. 3.6 Raman spectra and fitting curve for the sample AZ5025-ML annealed at 1000 ℃ for 5 minutes. The line (dash, red) is the Raman spectra, while the line

(solid, black) is the corresponding fit.

Table 3.2 The fitting results of the Raman curves for the samples AZ5025-ML, AZ5075-ML, and AZ75-ML annealed at 1000 ℃ for 5 minutes.

The crystallinity of the samples decreases with the increase of Al concentration, except the sample AZ5075-ML. X. J. Hao et al. investigated the effects of boron (B) concentration on Si QD/SiO₂ multilayer films [34]. The addition of B suppresses Si crystallinity. In this thesis, the role of Al is equivalent to the role of B, which act as dopants to improve the electrical properties. In other words, the decrease of the Si crystallinity is resulted from the suppression of Al. The exception of the sample AZ5075-ML will be discussed later.

Fig. 3.7 shows the transmittance spectra of the samples with different Al concentrations annealed at 1000 ℃ for 5 minutes . The transmittance decreases with the increase of the Al concentrations due to the free carrier absorption [8]. Then, we

estimated the optical bandgaps of the samples with different Al concentrations annealed at 1000 ℃ for 5 minutes. The optical bandgaps of the samples AZ5025-ML, AZ5075-ML, and AZ75-ML are 1.94, 2.08 and 2.15 eV, respectively. The widening of the optical band gaps with the increase of Al concentrations means the nc-Si QD size is narrower, which is corresponding to the trend of the FWHM from the Raman results.

Fig. 3.7 Transmittance spectra of the samples with different Al concentrations annealed at 1000 ℃ for 5 minutes.

Fig. 3.8 shows the I-V curves of the samples with different Al concentrations annealed at 1000 ℃ for 5 minutes. We find the samples AZ5025 -ML, AZ5075-ML and AZ75-ML show the rectification characteristics with the rectification values of 4.9×10¹, 1.5×10² and 4.0×10⁴ at ±2 V, respectively. We used the I-V curves to estimate the resistivity. The resistivity of the samples AZ5025-ML, AZ5075-ML and AZ75-ML is 9.0×10⁶, 4.3×10⁷, and 2.2×10⁶Ω-cm, respectively. In general, as the Al concentration increases, the conductivity decreases. However, the resistivity of the sample AZ5075-ML doesn’t show the trend. We suppose it is caused by the difference of the surface morphology. Thus, we check the surface morphology of the sputtering parameters of AZ5025, AZ5075, and AZ75 with the thickness of 5 nm by using atomic force microscopy (AFM), shown in Fig. 3.9. From the root mean square

(Rms) of the AFM images, the surfaces of the samples AZ5025 and AZ75 are flatter, while the sample AZ5075 is rougher. We also check the surface morphology of the multilayer thin films annealed at 1000 ℃ for 5 minutes by using scanning electron microscope (SEM), shown in Fig. 3.10. The multilayer thin films after annealing are bending. There are many convexes on the films. The density of the convexes is showed below the corresponding SEM images. The sample AZ5075-ML has the most convexes on the films, which is consistent with the AFM results. The rough surface

在文檔中可應用於太陽能電池之奈米結晶矽量子點埋入式摻鋁氧化鋅薄膜與p型氧化鋅薄膜之特性研究 (頁 25-0)