Browsing by Author "Roduner, Emil (Prof. Dr. phil.)"

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • Thumbnail Image
    ItemOpen Access
    The role of the heterointerfaces in the Cu(In,Ga)Se 2 thin film solar cell with chemical bath deposited buffer layers
    (2004) Nguyen, Hong Quang; Roduner, Emil (Prof. Dr. phil.)
    The replacement of the toxic chemical bath deposited CdS buffer layer in the heterostructure device ZnO/CdS/Cu(In,Ga)Se2 is a challenge for research and development of the thin film solar cell based on the Cu(In,Ga)Se2 (CIGS) material. Though a variety of alternative Cd-free buffers has been tested with considerable success, their solar cells in general are unstable and less efficient compared to the CdS buffer device. The aim of this work is to elucidate the superiority of the CdS buffer as well as the inferiority of an alternative In(OH,S) buffer by comparing the chemical and electronic properties at the heterointerfaces in the solar cells with those buffers. The first part of this thesis studies the formation of the CdS/CIGS interface, focusing on the interaction between the CIGS absorber layer and a standard chemical bath solution for the CdS deposition. The second part investigates the role of the heterointerfaces in the device with chemical bath deposited In(OH,S) buffer in comparison to that in the CdS buffer device. The study of the interface formation in the CdS buffer device first presents the growth characteristics of the CdS buffer from a standard chemical bath onto the CIGS substrate. X-ray photoelectron spectroscopy (XPS) reveals a complete coverage of the CIGS absorber layer after 4 min of CdS deposition. Taking into account that damage by ZnO sputtering can reach 5 nm depth in the substrate layer, a CdS layer deposited in 5 min is required to separate successfully the ZnO and CIGS layers. This result explains why the performance of the CdS buffer solar cells is best when the CdS layer is deposited in 5 min. A reduction of the CdS deposition duration, aiming at a decrease of the buffer thickness, results in a drastic decrease of the cell performance due to an incomplete coverage of the CIGS absorber surface by CdS, which is in turn leads to shunting between the ZnO and Cu(In,Ga)Se2 layers. The incomplete coverage also results in a deviation of the S/Cd ratio from unity during the CdS deposition. X-ray photoelectron spectroscopy investigations reveal a large excess of Cd compared to S (S/Cd ≈ 0.1) in the early stage of the CdS deposition. This strong deviation from stoichiometry is mainly due to an incorporation of Cd from the chemical bath into the Cu(In,Ga)Se2 surface layer. The S/Cd ratio increases as long as the absorber surface stays in contact with the chemical bath solution. As the absorber surface is completely covered by CdS, the S/Cd ratio saturates at a value of 0.75. Another factor that influences the S/Cd ratio is a co-deposition of an impurity containing Cd in the form of Cd(OH)2. During the chemical bath process, Cd(OH)2 can be converted into CdS due to a metathesis of thiourea on the surface of Cd(OH)2. However, this process plays a minor role in the trend of increasing the S/Cd ratio. The chemical mechanism of the incorporation of Cd into the absorber layer and the formation of Cd(OH)2 are next targeted by investigations of the treatment of the CIGS absorber in the Cd-NH3 solutions. The combination of secondary ion mass spectroscopy and XPS measurements shows an intimate tie of both these processes to the reduction-oxidation of the CIGS surface, which is controlled by the NH3 concentration in the chemical bath solution. With high NH3 concentration, a Cd diffusion into the absorber layer goes hand in hand with an adsorption of hydroxide ions onto the CIGS surface. When the NH3 concentration is low, the CIGS surface is oxidized under formation of Se atoms and hydroxide. The latter induces the deposition of Cd(OH)2 onto the CIGS surface, which prevents a further Cd-diffusion. Both the formation of hydroxides on the CIGS surface and the Cd-diffusion are unfavorable for the solar cell performance. In contrast to the polycrystalline CdS buffer layer deposited at a high pH value (11.6), the In(OH,S) buffer deposited at a much lower pH value (3.3) has quite different properties. In particular, it has an amorphous structure and contains a large fraction of hydroxide. To separate the role of bulk and interface properties in the electrical performance of the In(OH,S) buffer solar cell, I introduce a simple method by comparing the performance of this device to that of the devices with CdS and combinations of CdS/In(OH,S) buffers. The comparison leads to the following important conclusions: i) The interfaces of the In(OH,S) buffer to both the CIGS absorber and the ZnO window layers contain more defects than in the CdS buffer device. ii) However, the performance and the stability of the solar cell with the In(OH,S) buffer is predominantly controlled by the defects at the In(OH,S)/ZnO interface, but not by those at the In(OH,S)/CIGS interface. This result indicates that the advantage of the CdS buffer over the In(OH,S) buffer is due to the formation of a more favorable interface to the ZnO window layer. A high density of acceptor-like defects at the In(OH,S)/ZnO interface creates a barrier across this interface, which hinders photogenerated electrons to leave the CIGS absorber and results in a poor fill factor, a low current collection and low open circuit voltage. The reversible effect of air annealing, light soaking is mainly due to a temporal decrease of the defect density at this interface, whereas the reverse voltage bias increases it. Treatments of the In(OH,S) buffer grown on the CIGS absorber layer in the solution containing Zn2+ passivate these defects, and thus improve the cell performance. In this way, I obtain the Cd-free In(OH,S) buffer device with efficiencies comparable to those of the standard CdS device without any need of annealing or light soaking steps. iii) The results suggest the presence of a thin inverted surface defect layer (SDL) on the top of the CIGS absorber with a bandgap larger than in the bulk CIGS. A lowering of the valence band energy in the SDL results in a large barrier for holes at the buffer/CIGS interface, and hence makes the interface recombination more tolerant to the defect density at this interface. Thus, the role of the chemical bath deposited buffers is rather to save this SDL than to modify it.