2. EXPERIMENTAL SECTION
2.1. Cu 2 O thin film growth
In this work, the precursor solution was prepared using copper nitrate trihydrate salt (Cu (NO3)2 · 3H2O) as a copper source, monoethanolamine (MEA) as a stabilizer and isopropanol (IPA) or water as a solvent. The precursor solution was prepared with 0.6 moles of copper ion:
MEA with a ratio of 1:0.5. After the addition of the solvent, the solution was stirred at 55oC for 1 hour with magnetic stirring at 450 rpm. Then, the solution was filtered through Whatman 52 filter paper and aged at room temperature for up to 7 days.
Aged precursor solution was deposited on sapphire (0001) substrate (cleaned in acetone, methanol, and water consecutively in ultrasonic bath for 10 min and dried in ambient atmosphere) by spin coating precursor solution for up to 5 layers. Preheating steps in range of 200-750oC were employed between each successive layer to evaporate the remaining organic residues. Finally, after the deposition, the samples were subjected to post annealing step at various temperatures in range of 400-900oC in vacuum atmosphere (with or without air-flow) in quartz tube furnace with heating ramp of 10oC/min.
Figure 2.1. Schematic of Cu2O deposition process.
14
Detailed growth parameters of the samples are summarized in Table 2.1. and Table 2.2.
Table 2.1. Growth parameter of Cu2O thin film samples annealed in vacuum
Cu2+: MEA Aging Spinning
Table 2.2. Growth parameter of Cu2O thin film samples annealed under air flow
Cu2+: MEA Aging Spinning speed Preheating Layer Post annealing
Air flow
2500rpm, 60s 300: 10min 5 600,700,800,900: 30 min PO2= (5.60.1)10-3 Torr
15 2.2. X-ray diffraction analysis
X-ray diffraction technique is a non-destructive technique that is used to identify and analyze the structure of minerals, as well as other crystalline materials. When X-ray is diffracted by the ordered atoms in the lattice planes, diffracted rays form a constructive interference. As illustrated in Figure 2.2. X-ray beam with same wavelength and phase are scattered by the atoms with interplanar spacing d the second beam travels extra length of 2dsinθ. We can measure the distances between the planes of the atoms that constitute the sample by applying the Bragg's Law.
nλ = 2dhklsinθ
Where the integer n is the order of the diffracted beam, λ is the wavelength of the incident X-ray beam, d is the distance between adjacent planes of atoms (the d-spacings), and θ is the angle of incidence of the X-ray beam. The diffracted peaks at these incident angles determine the each crystal planes present in the sample [28].
Figure 2.2. Schematic illustration of Braggs law.
The crystal structure of the Cu2O sample was identified using XRD at room temperature.
The XRD analysis was done with a diffractometer operating at 40kV and 40mA with Cu Kα (=1.54184Å) radiation. Sample were analyzed with 3 degrees/min with spin and diffraction angle range of 20-80 degrees. The results were analyzed with MDI Jade 6 with JCPDS file database.
The results were analyzed with MDI Jade 6 with JCPDS file database and crystal phases are assigned with CuO (PDF 80-0076 ), Cu2O (PDF 78-2076 ) and Cu (PDF 85-1326) cards.
16 2.3. Raman spectroscopy
When light is incident on sample when the scattered light has no loss of energy or no frequency change then this type of scattering is Rayleigh scattering which is elastic scatting process. In the contrary, Raman scattering is inelastic scattering process, where the frequency and the wavenumber of the incident beam is changed during interaction with the phonons in the sample. When incident photon transfers its energy to the crystal lattice causing lattice vibration in the sample and the photon energy is reduced with higher frequency of phonon, the more energy is transferred according to the momentum and energy conversion and scattered in different direction. This phonon creation process is called Stokes process and reversibly if there is energy of phonon transferred to the incident photon resulting in photon with higher frequency then the process is called anti-Stokes process [29].
Figure 2.3. Schematic of Rayleigh, Stokes, and anti-Stokes scattering process [29].
17
Not all phonon modes in solids are Raman active and its due to the selection rule.
Selection rule of Raman scattering is determined by polarizability change during vibration.
Raman analysis can be effective tool for phase analysis as materials can be characterized by their unique vibrational modes [29].
For the purpose of confirming the copper oxide phase, Micro-Raman spectra were recorded with at 300 K. For sample excitation, a diode laser with 532 nm wavelength was used.
Characteristic phonon modes related to Cu2O were identified based on Table 2.3.
Table 2.3. Symmetry and characteristic Raman shift observed experiment and theoretical study of Cu2O [26, 27].
18 2.4 Photoluminescence
Photoluminescence is nondestructive spectroscopic technique applied for material characterization. Photoluminescence (PL) is a process absorption high frequency (hν > Eg) light by semiconductor exciting electrons and subsequent emission of photons as relaxation of electrons to ground state. Many localized defect states in crystal lattice can serve as radiative recombination centers. Consequently these imperfections can be detected by PL spectroscopy and resulting PL spectra can be used to identify the specific type of semiconductor defect [32].
The samples were mounted on closed cycle cryostat (CTI-Cryogenics) with copper plate.
The temperature of the sample chamber is monitored and controlled by heater in range of 25K to 300K. The excitation source was Nd:YAG laser with 355 nm wavelength (photon energy of 3.49 eV). The light beam was focused on a sample surface in horizontal direction by combination of reflective mirrors. The PL signal was detected by air cooled charge coupled device camera after dispersion by monochromator /Horiba iHR320.
Figure 2.4. Schematic of Photoluminescence measurement setup.
19
3. RESULTS AND DISCUSSION
3.1. Effects of annealing temperature and duration on post annealing in vacuum
First, series of XRD measurements were done to determine the state of the film before the post annealing step to realize the influence of structure of intermediate state on the formation of the Cu2O thin film. For this purpose, different preheating temperatures in the range of 200 to 750oC were used to preheat the samples to evaporate the organic residues between each successive layer. The results are shown in Figure 3.1. The sample preheated at 200oC does not show any diffraction peaks and is in amorphous state. Therefore, preheating in 200oC for 10 min does not provide enough thermal energy to drive formation any crystalline phase in air.
However, starting from the samples preheated at 250oC further to 750oC, the presence of CuO characteristic peaks starts to appear therefore the film is starting to transform in to crystalline state and the intensity of the CuO diffraction peaks become more intense as the temperature increases.
Figure 3.1. Crystalline phase of the film after the preheating at 200oC to 750oC for 10 minutes.
20
Subsequently the preheated samples were all subjected to post annealing step in vacuum at 750oC, temperature was chosen according to the phase diagram of the copper oxide in Figure 1.5. The intensity of the XRD results are normalized to Cu (200) peaks, to further investigate the influence of the state of the films during preheating step on the final product, the XRD measurement was done after the post annealing step. As seen from Figure 3.2 the resulting films all transformed to the Cu2O and metallic Cu phase after annealing in a vacuum at 750oC for two hours. But film is mostly dominated by the Cu phase over Cu2O.
Figure 3.2. XRD patterns of film post-annealed at 750oC for 2 hours in vacuum.
Since the state of the film before the post annealing step is CuO phase, then the post annealing in vacuum atmosphere should promote the reduction of the CuO due to lack of oxygen in vacuum atmosphere as indicated in the equation below.
4CuO 2Cu2O + O2
21
Furthermore, with higher preheating temperature the presence Cu2O phase is more intense as observed from the XRD patterns normalized to Cu (200) peak intensity. This may indicate the formation of Cu2O phase from CuO phase through oxide reduction in oxygen poor atmosphere is more efficient than formation of Cu2O phase from non-crystalline Cu-O chains.
This is not surprising since under the annealing ambient, the oxygen reduction process 2Cu2O 4Cu + O2 is efficient and to obtain the Cu2O phase, oxygens supplied by the pre-existed CuO is necessary. In short conclusion, at post annealing temperature of 750oC, there is strong presence of Cu phase and we cannot obtain the pure Cu2O films because of the efficient oxygen-reduction process.
Therefore, we studied the post-annealing temperature effect on the oxygen-reduction process. 400oC to 800oC are used for the post annealing step in vacuum to study the effects of temperature. The samples are preheated at 200oC. XRD results of the samples subjected to post annealing at 400oC, 500oC, 600oC, and 800oC in vacuum atmosphere with pressure of (2.40.3)10-2 Torr for 2 hours are shown in Figure 3.3. It clear that all the samples almost fully composed of Cu phase. Even after annealing at lower temperature of 400oC the film is still fully reduced to Cu phase, despite the phase diagram shown in Figure1.5 indicates the reduction to Cu phase happens at higher temperature with oxygen partial pressure of ~10-3 Torr. These results show that the post-annealing temperature is not a crucial factor that can control the oxygen-reduction process.
22
Figure 3.3. XRD patterns of film annealed at 400oC - 800oC two hours under vacuum (P= (2.40.3)10-2 Torr
Since the post annealing at lower temperature in vacuum still resulted in Cu film, the influence of the annealing duration was studied at temperature of 600oC. The target of this investigation is to examine whether the oxygen-reduction process can be controlled by annealing-duration or not. Figure 3.4 shows the XRD patterns of the films deposited at preheating temperature of 200oC and 600oC and post annealed at 600oC in vacuum atmosphere with base pressure of (2.70.5)10-2 Torr for 10, 20 and 30 minutes.
23
Figure 3.4. XRD patterns of film annealed at 600oC for 10–30 minutes under vacuum (2.70.5)10-2 Torr
After annealing at 600oC for 10 minutes, almost all of CuO phase has already disappeared and there is small peak of Cu2O (111) phase. Cu phase starts to emerge as seen with presence of Cu (111) and (200) peaks. As the annealing duration is increased to 20 minutes the Cu2O (111) peak is still present and traces of CuO phase is not seen. Furthermore, the intensity of the Cu phase is increased and additional Cu (220) peak has been appeared. After 30 minutes the Cu2O phase has disappeared and the films are fully reduced Cu phase. This result shows that even after annealing for short durations the results are same with Cu phase. Thus, we cannot obtain Cu2O film under vacuum annealing without supplying air or oxygen into the annealing tube.
The results from this section all show that annealing CuO film in vacuum atmosphere results in Cu phase due to lack of oxygen. This probably result of the oxygen in CuO film is pumped out of the annealing tube and CuO reduced to Cu phase during annealing. These results
24
are quite contradictory to previously reported result by Yu et al mentioned in the literature review section where Cu2O film was successively obtained using sol-gel technique with post annealing at 500-700oC in vacuum atmosphere with base pressure in range of 2 x 10-6 for 2 hours. But this difference in results is probably due to the different annealing conditions where the vacuum base pressure is significantly higher than discussed here. Also the crystalline form and thickness of the film before post annealing are not discussed by Yu et al , which could affect the formation of the Cu2O phase [26].
After several attempts, it was found that to obtain pure Cu2O phase in vacuum atmosphere may require certain pre-conditions, such as crystalline structures before the post-annealing, thickness of films and so on. Therefore, to directly get the pure Cu2O films without attempting abundant growth-parameters, we tried to control the oxygen-reduction process by supplying airflow and controlling the oxygen partial pressure by using a diaphragm-pump in the post-annealing process.
25
3.2. Effects of post annealing temperature and oxygen partial pressure on Cu2O film
Since the results from post annealing the film in vacuum resulted with Cu phase rather than Cu2O phase, this section describes the results of post annealing step performed with air flow to supply the oxygen needed in oxygen poor vacuum environment to obtain Cu2O film.
Figure 3.5 shows the crystalline phase of the film deposited at preheating temperature of 300oC and post annealed at 400oC to 750oC with air flow for 30 minutes at oxygen partial pressure PO2 of (1.90.2)10-3 Torr. The intensities are normalized at Cu (200) peak intensity.
After introducing oxygen by flowing air during the post annealing step, the results show that the films annealed at 400, 500, 600 and 750oC now contain both Cu2O and Cu phase. The film annealed at the 600oC shows most intense Cu2O (111) peak compared to other samples in this result. But the films still have Cu phase present, so the low oxygen partial pressure during the annealing step is insufficient to obtain pure Cu2O phase.
Figure 3.5. XRD patterns of film annealed at 400oC - 750oC for 30 minutes with air flow (PO2 = (1.90.2)10-3 Torr)
26
Therefore, the preformed films with CuO phase deposited at preheating temperature of 300oC were subjected to post annealing step at 600oC, which is chosen based on the intensity of the Cu2O phase from this result. The effect of oxygen partial pressure and longer duration at post annealing step were studied further.
Figure 3.6. XRD patterns of film annealed at 600oC for 30 minutes with air flow
At higher degree of oxygen partial pressure PO2 = (4.81.1)10-3Torr, the film consisted mainly of CuO phase and very small Cu2O phase when annealed for 30 minutes. Furthermore, when the film annealed for longer duration of 2 hours, the intensity of the Cu2O phase became more pronounced. These results indicate that when annealing at 600oC, at oxygen partial pressure around (4.81.1)10-3 Torr and with longer annealing durations, we might be able to obtain pure Cu2O film.
Subsequently, to study the effect of the annealing duration at the post annealing step at 600oC with oxygen partial pressure of (3.20.4)10-3 Torr, films deposited at preheating
27
temperature of 300oC were annealed for 2, 4, 6 and 8 hours. The results are shown in Figure 3.7.
As the annealing duration is increased the more CuO phase is transformed into Cu2O phase.
After annealing for 2 and 4 hours the films both show similar results with both Cu2O and CuO phase. But the Cu2O peak intensity is clearly stronger than that of CuO phase. Furthermore after 6 hours the CuO phase is almost fully transformed into Cu2O phase. Finally, after 8 hours of annealing, pure Cu2O film is obtained with no phase transformation into Cu phase.
Figure 3.7. XRD patterns of film annealed at 600oC for various duration at oxygen partial pressure of (3.20.4)10-3 Torr
Since we can obtain pure Cu2O film by annealing CuO film at 600oC with oxygen partial pressure of (3.20.4)10-3 Torr. Afterwards, we have studied the effect of the higher annealing temperature on the reduction of CuO film at slightly higher degrees of oxygen partial pressure.
28
The films deposited at preheating temperature of 300oC were subjected to post annealing at 700, 800 and 900oC with oxygen partial pressure of (5.60.1)10-3 Torr for 30 minutes. The XRD results are shown in Figure 3.8. As the post annealing temperature is raised to 700, 800 and 900oC the CuO film is fully converted to Cu2O phase just after 30 minutes. Even after annealing at 900oC there is no Cu phase present. Results show that although at 600oC we could be able to obtain Cu2O phase, but the annealing duration is significantly higher compared to higher temperature annealing. So, the 600oC does not provide enough thermal energy to fully convert CuO phase into Cu2O phase in short amount of time. Therefore, if the higher temperature up to 900oC is used for post annealing step then we can get Cu2O film with much shorter annealing time.
Figure 3.8. XRD patterns of films annealed at 600 - 900oC for 30 min at oxygen partial pressure of (5.60.1)10-2 Torr
29
3.3. Effect of post annealing duration Cu2O film at various temperatures.
The effect of post annealing duration was studied at 700oC and oxygen partial pressure of (5.70.2)10-3 Torr. The CuO films preheated at 300oC were subjected to post annealing for 30 minutes, 2 hours and 4 hours and results of the XRD is shown in Figure 3.9.
Figure 3.9. XRD patterns of films annealed at 700oC for 0.5 hr, 2hrs and 4hrs with oxygen partial pressure of (5.70.2) 10-3 Torr
The results show that in all samples CuO phase is fully transformed to Cu2O phase.
Additionally, after annealing for 4 hours there is no further reduction into Cu phase. Compared with the results shown in Fig 3.6, within the oxygen partial pressure PO2 ~ 510-3 Torr, we found that the annealing duration is not crucial above the annealing temperature of 700 oC.
30
Furthermore, to determine if the crystalline phase of the film before post annealing step affects the phase of the film after annealing, films preheated at three different temperatures were analyzed. Sample preheated at 200oC, 300oC, and 700oC have subjected to post annealing at 700oC for 2 hours and normalized XRD pattern at sapphire (0001) peak is shown in Figure 3. 10.
The results show that the all three samples are composed of Cu2O phase. As shown in the results of XRD analysis after the preheating step in Fig 3.1, at 200oC the sample is in amorphous state and at 300oC and 700oC, samples are composed of CuO phase. These results indicate regardless of the initial crystalline state of the film before the annealing process, pure Cu2O film can be obtained if the temperature and oxygen partial pressure are well-controlled.
Figure 3.10. Normalized XRD patterns of films preheated at 200oC, 300oC and 700oC annealed at 700oC for 2hrs with oxygen partial pressure of (5.70.2)10-3 Torr
31
3.4. Post annealing temperature and duration on Raman scattering in Cu2O film.
In order to investigate the crystalline structures of the samples, we performed the Raman spectroscopic studies on the Cu2O films. Cu2O unit cell contains 6 atoms, therefore there are 18 vibrational modes of which 15 optical phonon modes and 3 are acoustic modes. The symmetries of the vibrational modes at k=0 are:
A2u ⊕ E2u ⊕ 3T1u ⊕ T2u ⊕ T2g
Phonons with A, E and T are one, two and threefold degenerate. Three acoustic modes possess the T1u symmetry and two remaining modes of T1u belong to the infrared active optical lattice vibrations. Phonons with A2u, E2u, T2u are silent modes. Theoretical group theory analysis suggests perfect crystal of Cu2O should exhibit only one Raman active mode belonging to the threefold degenerate T2g mode. The other modes stem from breaking of symmetry from the defects within the crystal [25, 26].
Figure 3.11. Schematic illustration of zone-center phonon modes in Cu2O [14].
All the vibrational modes described above appears in Raman spectra of the Cu2O samples independent of the growth techniques and condition in various research reports [31].
32
In Fig. 3.12, we showed the Raman spectra of films with pure Cu2O-phase, post-annealing at 600, 700, 800 and 900oC. In the figure, there are two Raman curves (black and yellow), labelled by 700oC-0.5 hr, taken respectively from two samples grown under nearly identical condition. The two Raman spectra are nearly the same indicates that the crystal-structure of Cu2O films grown can be reproducible. As seen in Figure 3.12, the Cu2O films post-annealing at 600 to 800oC exhibit Raman shifts at 91 cm-1, 106 cm-1, 145 cm-1, 219 cm-1, 309 cm
-1, 417 cm-1, 500 cm-1 and 628 cm-1. Similar Raman scattering modes were also found in the Cu2O samples grown by other methods, under the 532 nm-laser excitation at room-temperature [31].
Most intense signal of Cu2O samples is the second order overtone 2Eu (219 cm-1). Due to
Most intense signal of Cu2O samples is the second order overtone 2Eu (219 cm-1). Due to