CARATTERIZZAZIONE DELL' INTERFACCIA TRA OSSIDO CONDUTTIVO TRASPARENTE E SILICIO AMORFO IDROGENATO PER APPLICAZIONI IN CELLE SOLARI A FILM SOTTILE

This work is the result of three years of studies during the PhD course in “Applied Physics” at the University of Palermo. The main object of the scientific activity is the electrical characterization and simulation of heterojunctions between transparent conductive oxide (TCO) and hydrogenated amorphous silicon (a-Si:H) for applications in thin film solar cells. The goal of this research is to improve the performance of p-i-n diode hydrogenated amorphous silicon thin film solar cells, in order to define the possible improvement of these prototypes. This PhD work stems from a collaboration between the Department of Physics, University of Palermo, the Institute for Microelectronics and Microsystems of National Council of Research (IMM-CNR) in Catania and STMicroelectronics in Catania. We can ideally divide this research into three main partes. In the first part, an electrical characterization of the solar cell prototypes was carried out. In the second part, a simulation study of solar cells prototypes was carried out and the most critical element was identified in the interfaces between TCO and a-Si:H. Finally, these interfaces were electrically characterized with the aim of determining the transport mechanism through a detailed simulation and analytic study. The a-Si:H solar cells has been studied since the 80’s and now are commercially available as multijunction and also as homojunction. The p-i-n a-Si:H solar cells are realized mainly by PECVD technique at low temperature. This technique allows the use low-cost substrates, so these types of solar cells are interesting for consumer applications. The studied solar cell has a superstrate structure and it is composed by a glass substrate on which the SnO2:F (tin oxide fluorine doped) as TCO is grown . Then XXIV the p-i-n solar cell and the ZnO:Al (zinc oxide aluminum doped), produced by STMicroelectronics, are deposited. The electrical characterization of the a-Si:H solar cell prototypes were performed both in dark and under illumination. The performance of this solar cell are not optimized and still require improvements in the design and technological processes. Some accurate admittance measurements on these prototypes were also performed to extract information about the properties of these structures (admittance spectroscopy) and in particular about the materials quality and the used interfaces. It was found that the capacitance per unit area depends on the area of the solar cell. This phenomenon, due to the distributed resistance of the TCO, was explained by using an analytical transmission line model for strip and circular geometry. The sheet impedance of two used TCOs and the capacitance of the solar cells was measured. The admittance measurements had shown a capacitance dependence at high frequencies and in reverse bias with diameter. This effect was simulated by using the same transmission-line model. The series resistance of the solar cell was determined by using dark and under illumination I-V and the value obtained from admittance measurements was compared with that obtained by using other methods. In the second part of this work a simulation study has been performed in order to determine the critical parameters of the a-Si:H solar cell. The density and properties of amorphous silicon defects were investigated. From this simulation study it results that the dangling bonds concentration plays the main role on the solar cell performance and in particular the defect concentration in the intrinsic layer. The bandtails and the pdoping concentration cause smaller effects than the dangling bonds ones. The temperature, the a-Si:H energy gap, the dangling bonds and bandtails energetic position cause significant effects on the performance of the solar cell. Then, the effect of the thickness under different wavelength was studied as from p side as from n side. The low mobility of holes causes a collapse of the electric field, and then a saturation of the maximum power for the former case and a decline in the latter case. The obtained optimal thickness is about 250 nm. By using a simulation study, the effect of the potential barrier at the interface TCO/p-type a-Si:H was investigated modelling the TCO as metal with a certain resistivity and work function. It was noted that the presence of a XXV barrier cause an heavy deterioration in the performance of the solar cell and in particular on the Voc and Jsc. The most critical parts of the investigated solar cells are the heterojunctions between TCO and a-Si:H, and in particular that between SnO2:F and p-type a-Si:H. Structures composed by molybdenum (Mo), p-type a-Si:H and SnO2: F were commissioned to STMicroelectronics in order to study the properties of these heterojunctions. Admittance and current measurements (I-V) at different temperatures were performed and a capacitance model was proposed. After checking the reproducibility and the uniformity of the investigated heterostructures it was found that the circular structures are less dispersive than those strip. The I-V curves are slightly asymmetric and show two regions: a linear region (ohmic) up to voltages |V| = 0.1 V and a superlinear region (power-law) for |V|>0.1 V. Transport mechanisms such as diffusion, recombination and SCLC were tested without success and capacitance (C-V) and impedance measurements in the linear region at different temperatures were performed. The intercept method was applied to the capacitance measurements supposing that the Mo/p-type a-Si:H contact was ohmic. The contact resistance was studied to obtain the transport mechanism. The thermionic emission model fitted good to the contact resistance and a height barrier of 0.50 eV was found at the SnO2:F/p-type a-SI:H interface. This transport mechanism could only explain the I-V curves around zero bias voltage and do not explain the power-law behaviour and the I-V symmetry. These thermionic emission model inconsistencies led to the conclusion that the contact Mo/p-type a-Si:H is not ohmic and the SnO2:F cannot be modelled as a metal. Moreover, the symmetry in the I-V suggested that the transport mechanism is the same for both positive and negative voltages. From the state of the art on the transport mechanisms in heterojunction N-p type II and in a- Si:H junction it was found that a possible transport mechanism in the investigated heterojunction could be due to the thermally generation carriers assisted by tunnel from interface traps to the respective bandedge. This mechanism has been used to explain the transport in tunnel junctions in multijunction solar cells. For this reason, it was decided to simulate this transport mechanism by using the 1D simulator SCAPS 2.9.03. Firstly, the band diagrams, the effects of interface defects, of the tunnel barrier height at the Mo/p-type a-Si:H interface and of the p-type doping on the current in these XXVI heterojunctions were simulated. Later we proceeded to simulate the I-V curve of the structures Mo/p-type a-Si:H/SnO2:F. The goal was to determine the barrier height at the interface between p-type a-Si:H and SnO2: F, the transport mechanism, the doping A N and the defect density at the interface G,int N , by using a recursive method that allow to fit simultaneously the I-V curves with the simulations and the C-V curve with the analytical back to back diode model at different temperatures. The parameters obtained are an high density of defects of 5.6 ± 0.5´1018 cm-3 , a doping density of 5.6 ± 0.5´1018 cm-3 , a barrier height at the interface p-type a-Si:H/SnO2:F of 0.16 eV at room temperature by using an holes and electron effective mass of, respectively, 0.3 and 0.1. The obtained barrier at the interface p-type a-Si: H/Mo was 0.14 eV. This value is close to the value obtained for the p-type a-Si:H/SnO2:F and confirms the hypothesis that this barrier affects the measured C-V and I-V curves of the heterostructure. In conclusion, we have demonstrated that the transport mechanism in these heterostructures is dominated by the reverse bias current of the SnO2:F/p-type a-Si:H and p-type a-Si:H/Mo diodes. In addition, this work has demonstrated that the SnO2:F, despite being a good conductor with a low sheet resistance, can not be simulated as a metal, as reported in the literature, but as a semiconductor. From the I-V and C-V simulations at different temperatures, the transport mechanism of the diode SnO2:F/p-type a-Si: H was determined. In this work we also demonstrates that the thermally assisted tunnel of e/h pairs from interface defects to the bandedge is the principal transport mechanism. Eventually, we have propose a band diagram for the structure analyzed consistent with the experimental data. The high resolution TEM has confirmed the presence of a weak interface between SnO2: F and ptype a-Si: H and hence the presence of a high density of interface defects.