Photoluminescence and electroluminescence imaging of PV devices
- Authors: Roodt, Roelof Petrus
- Date: 2024-04
- Subjects: Photoluminescence , Biosensors
- Language: English
- Type: Master's theses , text
- Identifier: http://hdl.handle.net/10948/64333 , vital:73676
- Description: Luminescence imaging has become a particularly useful and valuable tool for the characterisation of photovoltaic devices. This study entailed the design, construction, and optimisation of a system for the electroluminescence (EL) and photoluminescence (PL) imaging of various solar cell devices. The system can perform EL and PL imaging of solar cells of different cell technologies and materials systems, including Si, perovskite, and triple-junction concentrator solar cells. This required appropriate electrical power supplies for carrier injection for EL imaging and optical excitation for PL imaging. The different materials systems also required wavelength appropriate filters for PL imaging. In addition, the system utilized a temperature-controlled sample stage and was placed in a chamber for environmental control and isolation of UV radiation from laboratory. In addition to optimization of imaging conditions, luminescence images need to be optimized to facilitate detailed analysis and the application of appropriate algorithms to extract device parameters and hence generate device parameter images of the devices under investigation. For EL imaging, two power supplies were used to inject current into the solar cells. The reason for the two power supplies is that the first power supply had a current range of ± 1 A and an applied voltage capability of ± 21 V. This was used for the smaller solar cells. It was also convenient to use as the power supply could also measure the injected current and applied voltage and digitally store it with the images. For the larger solar cells, a second power supply was utilized, which could inject current into the samples in the range of ± 12 A at an applied voltage of ± 40 V. To measure the current and voltage of the power supply provided, two digital multimeters were utilized. For acquiring images, the same camera was used for EL and PL imaging. The sensor used in the camera is a silicon CMOS sensor. For PL imaging, four light emitting diode (LED) boards, consisting out of sixty-four LED’s, per board, of four different wavelengths, were used to optically excite the solar cells. The four wavelengths emitted by the LED’s were chosen to match the bandgaps of the different solar cell devices investigated. The LEDs were powered with a multi-channel constant voltage power supply, where the current could be varied. The Si solar cell is a 156 x 156 mm commercial solar cell. The perovskite solar module is a 40 x 40 mm module, which consists out of six cells connected in series. The triple-junction concentrator solar cell has a dimension of 10 x 10 mm which consists of three junctions staked on top of one another. These three layers consist of indium gallium phosphate (InGaP), indium gallium arsenide (InGaAs) and germanium (Ge). To capture EL and PL images of these various solar cell devices, filters of specific wavelengths were placed in front of the camera to isolate the light generated by the different devices. In addition to isolating the luminescence observed from the solar cells, an image correction procedure was adapted from literature, to be applicable to acquiring luminescence images of these various solar cells. As there are a range of factors which influence the quality and clarity of the luminescence images, i.e., chromatic aberration, diffraction, and absorption depth, to name a few, the wavelength dependency of these factors was investigated. This was done by acquiring a point spread function (PSF) for each of these devices and then using these PSF's together with a deconvolution algorithm to correct the luminescence images. The PSF was acquired by fitting a point source emission image to a function that includes exponential and Gaussian terms. The point source image was obtained by placing a black piece of vinyl with a pinhole in it over the solar cell. To communicate with all the various devices and to acquire images at various intensities a LABVIEW program was written. This was used then used to control the power supplies, digital multimeters, camera, and the LED's. This allowed for the user to specify at what points along the current-voltage (I-V) curve data points needed to be measured together with the luminescence images captured. For PL imaging the intensity of the LED's was then also adjusted according to user specified values. The system was utilised to acquire EL images of the Si solar cell, EL and PL images of the perovskite solar cell and EL images of the InGaP and InGaAs layers in the triple-junction concentrator solar cell. With the correction procedure utilised in this study, it was seen that the image quality and clarity improved, compared to the conventional way of capturing luminescence images. These statements are supported by the results obtained for the series resistance maps of the Si solar cell and the perovskite solar module, as the series resistance maps obtained from the corrected luminescence images have less noise and more detail compared to the results from the raw luminescence images. From the EL images captured for the two layers of the triple junction concentrator, it was clear that the intensity profile of the two layers is different, as the intensity for the InGaP layers was that the device had bright edges and darker intensity on the interior where exactly the opposite was observed for the InGaAs layer, having a bright interior and darker edges. This is most likely due to the opto-electric coupling of these layers withing the triple junction solar cell. For the series resistance images obtained for the Si solar cell, it is observed that at lower carrier injection, the series resistance is lower compared to higher carrier injection levels. This result can also be influenced by the increase in cell temperature with the increase in injected carriers. The series resistance maps obtained from the perovskite EL images shows an interesting result. As the perovskite solar cell has degraded, three of the six cells have optically inactive regions, showing lower luminescence intensities. The series resistance of the other three cells are much lower compared to these cells that have inactive regions under low injection conditions. As the injection level increases, it is seen that the series resistance values of five of the six cells become comparable to one another. With regards to the PSF, it was found that using a bandpass filter in front of the lens reduced to amount of spreading observed from a single point source across the detector. Furthermore, there is a strong wavelength dependency in the PSF as the severity increased with increase in the emission wavelength of the solar cells under investigation.In this study an opto-electrical characterisation system was constructed to acquire PL and EL images of various solar cell technologies. In addition to this, a range of factors that influence the quality of these images were investigated and used in the image correction procedure to correct the images for all these cell technologies. It was shown that the correction procedure works for all three of the technologies investigated in this study, and all these factors showed a strong wavelength dependency. These corrected luminescence images together with current-voltage (I-V) data was then used to determine characteristic parameters of a one-diode model of the various PV devices. This was not only achieved, but it also clearly indicated that all the correction procedures need to be considered to obtain a clear and accurate representation of the actual PV device. This has a major influence on the understanding and improvement of these PV devices. , Thesis (MSc) -- Faculty of Science, School of Computer Science, Mathematics, Physics and Statistics, 2024
- Full Text:
- Date Issued: 2024-04
Growth and characterization of ZnO nanorods using chemical bath deposition
- Authors: Urgessa, Zelalem Nigussa
- Date: 2012
- Subjects: Zinc oxide , Photoluminescence , Semiconductor nanocrystals , Semiconductors -- Materials , Chemical reactions
- Language: English
- Type: Thesis , Doctoral , PhD
- Identifier: vital:10559 , http://hdl.handle.net/10948/d1021124 , http://hdl.handle.net/10948/d1021119
- Description: Semiconductor devices are commonplace in every household. One application of semiconductors in particular, namely solid state lighting technology, is destined for a bright future. To this end, ZnO nanostructures have gained substantial interest in the research community, in part because of its requisite large direct band gap. Furthermore, the stability of the exciton (binding energy 60 meV) in this material, can lead to lasing action based on exciton recombination and possibly exciton interaction, even above room temperature. Therefore, it is very important to realize controllable growth of ZnO nanostructures and investigate their properties. The main motivation for this thesis is not only to successfully realize the controllable growth of ZnO nanorods, but also to investigate the structure, optical and electrical properties in detail by means of scanning electron microscopy (SEM), transmission electron microscopy (TEM), photoluminescence (PL) spectroscopy (steady state and time resolved) and X-ray diffraction (XRD). Furthermore, strong rectification in the ZnO/p-Si heterojunction is demonstrated. Nanorods have been successfully synthesized on silicon by a two-step process, involving the pre-coating of the substrate by a seed layer, followed by the chemical bath deposition of the nanorods. ZnO seed layers with particle sizes of about 5 nm are achieved by the thermal decomposition of zinc acetate dihydrate dissolved in ethanol. The effects of the seed layer density on the distribution, alignment and uniformity of subsequently grown nanorods were studied. The aspect ratio, orientation and distribution of nanorods are shown to be well controlled through adjusting the density of the ZnO nanoparticles pre-coated onto the substrates. It is shown that the seed layer is a prerequisite for the growth of well aligned ZnO nanorods on lattice mismatched Si substrate. The influence of various nanorod growth parameters on the morphology, optical and electrical properties of the nanorods were also systematically studied. These include the oxygen to zinc molar ratio, the pH of the growth solution, the concentration of the reactants, the growth temperature and growth time, different hydroxide precursors and the addition of surface passivating agents to the growth solution. By controlling these xii parameters different architectures of nanostructures, like spherical particles, well aligned nanorods, nanoflowers and thin films of different thicknesses are demonstrated. A possible growth mechanism for ZnO nanostructures in solution is proposed. XRD indicated that all the as-grown nanostructures produced above 45 C crystallize in the wurtzite structure and post growth annealing does not significantly enhance the crystalline quality of the material. In material grown at lower temperature, traces of zinc hydroxide were observed. The optical quality of the nanostructures was investigated using both steady-state PL and time-resolved (TR) PL from 4 K to room temperature. In the case of as-grown samples, both UV and defect related emissions have been observed for all nanostructures. The effect of post-growth annealing on the optical quality of the nanostructures was carefully examined. The effect of annealing in different atmospheres was also investigated. Regardless of the annealing environment annealing at a temperature as low as 300 C enhances the UV emission and suppresses defect related deep level emission. However, annealing above 500 C is required to out-diffuse hydrogen, the presence of which is deduced from the I4 line in the low temperature PL spectra of ZnO. TRPL was utilized to investigate lifetime decay profiles of nanorods upon different post growth treatments. The bound exciton lifetime strongly depends on the post-growth annealing temperature: the PL decay time is much faster for as grown rods, confirming the domination of surface assisted recombination. In general, the PL analysis showed that the PL of nanorods have the same characteristics as that of bulk ZnO, except for the stronger contribution from surface related bound excitons in the former case. Surface adsorbed impurities causing depletion and band bending in the near surface region is implied from both time resolved and steady state PL. Finally, although strong rectification in the ZnO/p-Si heterojunction is illustrated, no electroluminescence has been achieved. This is explained in terms of the band offset between ZnO and Si and interfacial states. Different schemes are proposed to improve the performance of ZnO/Si heterojunction light emitting devices.
- Full Text:
- Date Issued: 2012
Surface modifications of InAs: effect of chemical processing on electronic structure and photoluminescent properties
- Authors: Eassa, Nahswa Abo Alhassan Eassa
- Date: 2012
- Subjects: Indium arsenide , Chemical processes , Photoluminescence
- Language: English
- Type: Thesis , Doctoral , PhD
- Identifier: http://hdl.handle.net/10948/8714 , vital:26423
- Description: In this thesis, the effects of various chemical treatments on the surface modification of bulk InAs are investigated. The study focuses on the chemical processes that occur upon the exposure of the surface to sulphur-, chlorine- and bromine-containing solutions and oxygen, and the resulting changes to the electronic structure of the surface, as deduced from photoluminescence (PL) measurements, X-ray photoelectron spectroscopy (XPS), Auger electron spectroscopy (AES), Raman scattering and scanning electron microscopy (SEM). Three processing treatments were evaluated: i) treatment with sulphur-based solutions (Na2S:9H2O, (NH4)2S + S, [(NH4)2S / (NH4)2SO4] + S); ii) etching in halogen-based solutions (bromine-methanol and HCl: H2O); and iii) thermal oxidation. A significant overall enhancement in PL response was observed after chemical treatment or thermal oxidation, which is associated with a reduction in surface band bending. These changes correlate with the removal of the native oxide, in addition to the formation of well-ordered layers of In-S (or In-As)O as a passivating layer, indicating that electronic passivation occurs at the surface. The passivating effect on sulphide treated surfaces is unstable, however, with an increase in band bending, due to reoxidation, observed over periods of a few days. The lowest re-oxidation rate was observed for ([(NH4)2S / (NH4)2SO4] + S). Etching in HCl:H2O and Br-methanol solutions of appropriate concentrations and for moderate times (1 min) resulted in smooth and defect-free InAs surfaces. Etching completely removed the native oxides from the surface and enhanced the PL response. The adsorption of bromine and chlorine onto the InAs surface led to the formation of As-Brx , In-Brx, As-Clx and In-Clxcompounds (x = 1, 2, 3), as inferred from changes in the In 3d3/2; 5/2 and As 3d core level binding energies. The etch rate was found to decrease because of strong anisotropic effects. The improvements in surface properties were reversed, however, if the concentrations of the etchants increased or the etch time was too long. In the worst cases, pit formation and inverted pyramids with {111} side facets were observed. Surface treatments or thermal oxidisation significantly enhanced the PL intensity relative to that of the as-received samples. This was due to a reduction in the surface state density upon de-oxidation, or in some cases, to the formation of a well ordered oxide layer on the surface. The overall increase in PL intensity after treatment is ascribed to a reduction in band bending near the surface. This allows several welldefined peaks not observed or reported previously for bulk InAs (with a carrier concentration n~2x1016 cm-3), to be studied. A combination of PL and XPS measurements before and after the various treatments was used to identify the chemical nature of the impurities giving rise to bound exciton recombination in InAs (111).
- Full Text:
- Date Issued: 2012
Development of MgZnO-grown MOCVD for UV Photonic applications
- Authors: Talla, Kharouna
- Date: 2011
- Subjects: Photoluminescence , Photonics , Zinc oxide
- Language: English
- Type: Thesis , Doctoral , PhD
- Identifier: vital:10537 , http://hdl.handle.net/10948/d1012585 , Photoluminescence , Photonics , Zinc oxide
- Description: MgxZn1-xO has emerged as a material of great technological importance. Having a direct energy band gap that is tunable throughout much of the ultraviolet (UV) region of the spectrum from the near-UV (~370 nm) to the deep-UV (~176 nm), this compound is of interest for a variety of optoelectronic devices operating in this part of the electromagnetic spectrum. MgxZn1-xO offers advantages over the more mature compound semiconductor AlGaN which stem mainly from the unusually high exciton binding energy (60 meV in ZnO). In this study the growth of ZnO and MgxZn1-xO thin films using metal organic chemical vapour deposition (MOCVD) is systematically investigated. The films are mainly grown on c-Al2O3 and Si (100) and characterized using various techniques, such as photoluminescence (PL), x-ray diffraction (XRD), scanning electron microscopy (SEM), transmission electron microscopy (TEM) and auger electron spectroscopy (AES). The optical and the structural properties are essentially inspected in order to improve their quality. In this thesis the optimisation of ZnO grown using oxygen gas as a new oxidant in our reactor is investigated. The growth temperature and VI/II ratio are varied in order to find optimum parameters giving high quality layers. The effects of Si (100), Si (111), c- and r-sapphire, glass, GaAs and ZnO substrates on the optical, structural and morphological properties of ZnO thin films grown with tert-butanol (TBOH) is examined. Similar morphologies are observed for all substrates, with the films comprising hexagonal columns having cone shaped ends. The photoluminescence spectra are similar, but the various transitions have different relative intensities. It is clear that the different substrates influence neither the orientation of the films, nor the surface morphology, significantly. The photoluminescence hints at larger stacking fault densities in films grown on silicon and glass, however, as evidenced by stronger basal plane stacking fault-related luminescence at ~3.319 eV in the relevant low temperature photoluminescence spectra. The morphology changes with Mg incorporation, from hexagonal columnar structures to cubic faceted columns. From PL, the full with at half maximum is found to gradually increase with Mg content due to alloy broadening. The deep level emission (DLE) is observed to shift with Mg content. By changing the Mg content, the band gap of MgxZn1-xO film is tuned by ~450 meV, which provides an excellent opportunity for band gap engineering for optoelectronic applications. The c-lattice constant of ZnO (5.205 Å) decreases by only 0.6% when the Mg content reaches x=0.39. The introduction of Mg into ZnO is shown to increase the relative PL intensity of stacking fault-related transitions (at 3.314 eV for ZnO). This becomes the dominant near band edge emission. Using TEM a thin Mg rich layer is observed at the interface between the film and the Si or Al2O3. Temperature dependent PL measurements on layers with low Mg concentration (x=0.05 and 0.1) show that the main bound exciton peak exhibits an “s-shaped” temperature dependence, characteristic of localization in a disordered alloy. The origin of the PL line broadening of MgxZn1-xO (x≤0.04) is also analyzed with respect to alloy broadening, taking into account a random cation distribution and alloy clustering. The influence of various MOCVD growth parameters such as growth temperature and VI/II ratio is studied. Varying the temperature from 280 ˚C to 580 ˚C reveals strong morphological changes and optical degradation of the films. Low (<280 ˚C) and high (>580 ˚C) growth temperatures reduce the Mg incorporation. High VI/II ratios also decrease the Mg incorporation, as evidenced by the red-shift of the donor bound exciton (D°X) line. This is ascribed to a stronger premature reaction between (MeCp)2Mg and the oxidant or a preferential heterogeneous interaction between the Mg and oxygen species on the growth front. For both oxidizing agents (O2 and TBOH), the growth at 420 ˚C and a VI-II ratio of 60 on c-Al2O3 gave optimal quality layers in terms of their optical and structural quality. A comparison of films grown using TBOH and O2 gas as oxidizing agent shows no major difference in terms of Mg incorporation. The effect of annealing, the inclusion of a buffer layer and the influence of growth rate on the properties MgxZn1-xO thin films are also reported.
- Full Text:
- Date Issued: 2011
Photoluminescence study of ZnO doped with nitrogen and arsenic
- Authors: Dangbegnon, Julien Kouadio
- Date: 2010
- Subjects: Photoluminescence , Zinc oxide , Nitrogen , Arsenic
- Language: English
- Type: Thesis , Doctoral , PhD
- Identifier: vital:10518 , http://hdl.handle.net/10948/1216 , Photoluminescence , Zinc oxide , Nitrogen , Arsenic
- Description: In this work, the optical properties of ZnO doped with arsenic and nitrogen were studied. The ZnO samples were grown by Metalorganic Chemical Vapor Deposition (MOCVD). The solubility of nitrogen in the ZnO films, as well as its activation upon annealing, was also investigated. Hydrogen is known as a major source for passivation of the acceptors in ZnO:N. Therefore, it is crucial to dissociate the complex(es) formed by nitrogen and hydrogen and diffuse out the hydrogen in order to prevent the reformation of such complexes. High temperatures (≥ 600 C) are required for these purposes. In order to effectively remove the hydrogen impurities from the sample, it is important to know the optical fingerprints of hydrogen and its thermal stability. Therefore, the effects of annealing and hydrogen plasma treatment on bulk ZnO (hydrothermally grown) were first studied. The use of bulk material for this purpose was motivated by the well-resolved photoluminescence (PL) lines observed for bulk ZnO, which allow the identification of the different lines related to hydrogen after plasma treatment. Annealing at 850 C was effective for the removal of most of the hydrogen related transitions in the near-band-edge emission. Also, additional transitions at ~3.364 eV and ~3.361 eV were observed after hydrogen plasma treatment, which were ascribed to hydrogen-Zn vacancy complexes. In this work, a comparative study of the annealing ambient and temperature on ZnO films grown on GaAs substrate, using diethyl zinc (DEZn) and tertiary butanol (TBOH), showed that arsenic diffuses in the ZnO films and gives a shallow level in the band gap, which is involved in an acceptor-bound exciton line at 3.35 eV. This shallow level is visible when annealing is performed in oxygen, but not when annealing is performed in nitrogen, and indeed only for annealing temperatures around 550 C. However, annealing in either ambient also causes zinc to diffuse from the ZnO films into the GaAs substrate, rendering the electrical properties deduced from Hall measurements ambiguous. For ZnO:N, NO was used as both oxygen and nitrogen sources. Monitoring the concentration of nitrogen, carbon and hydrogen in the ZnO films, the formation of different complexes from these impurities were deduced. Furthermore, an investigation of the effect of annealing on the concentrations of impurities showed that their out- diffusion was strongly dependent on the crystalline quality of the ZnO films. For porous ZnO films, obtained at low growth temperatures (≤310 C), the out-diffusion of impurities was efficient, whereas for films grown at higher temperatures, which have improved crystalline quality, the out-diffusion was practically nonexistent. The out-diffusion of unwanted impurities may activate the nitrogen dopant in the ZnO films, as was confirmed by the PL measurements on the different samples grown at different temperatures. PL transitions at ~3.24 eV and ~3.17 eV were related to substitutional NO. These transitions were more dominant in the spectra of samples grown at low temperatures. An additional transition at ~3.1 eV was assigned to a donor-acceptor pair transition involving VZn, instead of NO, as previously reported.
- Full Text:
- Date Issued: 2010
On the growth and characterisation of AIGaN alloys for optoelectronic applications
- Authors: James, Grant Robert
- Date: 2005
- Subjects: Gallium nitride -- Electric properties , Photoluminescence
- Language: English
- Type: Thesis , Doctoral , PhD
- Identifier: http://hdl.handle.net/10948/8824 , vital:26433
- Description: In this study the growth and characterisation of undoped and Si-doped AlxGa1-xN has been performed. The layers were grown using low-pressure metalorganic vapour phase deposition (MOCVD) on sapphire substrates. The optical and electrical properties of the AlxGa1-xN layers were studied using variable temperature Hall effect and photoluminescence measurements. AlxGa1-xN layers were grown over the entire composition range. Room temperature ultraviolet (UV) transmission measurements showed that the material quality was very good for layers with an Al content, x, of 0 _ x _ 0.5. However, the quality of layers of higher composition was seen to rapidly decrease with increasing x. The electrical and optical properties of AlxGa1-xN with x < 0.5 were also good, comparable to those reported on in literature. The study of the Si-doping of AlxGa1-xN was performed in two parts; firstly a series of Al0.23Ga0.77N samples was grown in which the doping level was increased from zero to n _ 3 × 1018 cm-3. A similar, albeit a less rigorous, study was performed for Al0.41Ga0.59N and Al0.5Ga0.5N. A second series of samples was then grown in which the doping level was kept constant, while the Al content was incrementally increased. Room temperature Hall effect measurements performed on Si-doped Al0.23Ga0.77N showed that the electron concentration did not scale linearly with the silane flow, as was the case in GaN. It was also seen that the electron mobility of the layers increased with slight Si-doping, possibly due to an improvement in the crystalline quality and/or a change in the conduction mechanism. It was also found that at higher compositions (x = 0.41 and 0.50) an increase in the doping level resulted in an increase in the mobility. Variable temperature Hall effect and photoluminescence measurements, performed on the Al0.23Ga0.77N samples, revealed a good correlation between the first PL activation energy E1 and the donor activation energy ED, prompting the conclusion that the first PL recombination channel in AlxGa1-xN is due to the delocalisation of excitons bound at neutral Si donors. Furthermore, E1 and ED were seen to decrease with n1/3, as is the case for GaN and other semiconductor materials. It was also observed that strong exciton localisation occurs in slightly Si-doped material, with the amount of localization becoming less at higher doping levels. Possible mechanisms responsible for the second PL recombination channel of activation energy E2 were also proposed. The electrical and optical properties of the second set of AlxGa1-xN samples was then studied. The PL properties of undoped AlxGa1-xN were typical of a homogeneous alloy system, with the increase in the PL FWHM and exciton localisation energies with x following the trend predicted by alloy disorder theory. The variation of the band gap energy with the Al content could not, however, be fitted over the entire composition range using a single bowing parameter. It was proposed that this was due either to an effect of the 9 7 valence band crossover, or due to exciton localisation at alloy disorder and/or impurities. As was the case for GaN and Al0.23Ga0.77N, all undoped material was highly resistive. As was mentioned earlier, the exciton localisation energies increased according to alloy disorder theory in undoped AlxGa1-xN. In the doped samples, however, a large increase in the donor localisation energy was measured for x > 0.3. The possibility that Si could become a DX-centre in AlxGa1-xN was then investigated. However, Hall effect measurements showed that the Si activation energy increased in good agreement with the model of a shallow effective mass state donor, with no sudden increase in ED being observed up to x = 0.4. It was then suggested that the increase in the E1 and E2 activation energies, as well as the exciton localisation energies, could be due to the 9 7 valence band crossover, which occurs at roughly the same composition. However, due to the scarcity of reports on the valence band structure in AlxGa1-xN no conclusions could be made at this stage as to the effect of the 9 7 valence band crossover on the PL properties of AlxGa1-xN.
- Full Text:
- Date Issued: 2005