% IMPORTANT: The following is UTF-8 encoded. This means that in the presence
% of non-ASCII characters, it will not work with BibTeX 0.99 or older.
% Instead, you should use an up-to-date BibTeX implementation like “bibtex8” or
% “biber”.
@PHDTHESIS{Misic:568190,
author = {Misic, Boris},
othercontributors = {Rau, Uwe and Werner, Jürgen H.},
title = {{A}nalysis and simulation of macroscopic defects in
{C}u({I}n,{G}a){S}e$_2$ photovoltaic thin film modules},
school = {RWTH Aachen},
type = {Dissertation},
address = {Aachen},
reportid = {RWTH-2016-01132},
pages = {1 Online-Ressource (iv, 147 Seiten) : Illustrationen,
Diagramme},
year = {2015},
note = {Veröffentlicht auf dem Publikationsserver der RWTH Aachen
University 2016; Dissertation, RWTH Aachen, 2015},
abstract = {The present work deals with production induced defects in
CIGS thin film modules which can deteriorate the electrical
performance of the module. The motivation of this work is to
both find ways how these defects can be diagnosed, e.g. in a
quality control in the production site, and to gain a better
understanding of the actual defect influence on the voltage
and current in the surrounding of the defect. Thus, I
investigate the use of electroluminescence and thermography
as diagnostic tools to detect and identify common defects
occurring in CIGS production. I begin this work with a study
of the CIGS production process and list potential defect
origins for relevant production steps. In order to allow an
experimental investigation of defects, I intentionally
implement them into CIGS photovoltaic modules in a real CIGS
production site environment. The defect implementation
includes e.g. interrupted P1, P2, and P3 scribing lines for
the monolithic series connection, as they can be caused by
faulty scribing tools, and changes in the normal CIGS layer
structure, as they can be caused by local contamination.
Furthermore, I vary the geometry of the implemented defects.
I characterise the implemented defects with microscope,
electroluminescence (EL), and dark lock-in thermography
(DLIT) measurements. EL and DLIT are chosen as they are
spatially resolved measurements and therefore allow a
comparatively fast investigation of a complete module. In
addition to the defect measurements, I implement a software
that is based on the principle of network simulation model
and allows to model and simulate the implemented defect
types in a CIGS module. The software models the CIGS module
with a network, that consists of equivalent circuits of a
solar cell and resistances, and translates the network into
a non-linear system of equations that are solved. Finally, I
investigate methods to repair an incomplete insulation of
the Mo back contacts of two neighbouring cells. The
measurements on defects yield that P1, P2, and P3 scribing
defects have characteristic EL and DLIT patterns. I show
that the scribing defects can be reliably identified by
these patterns if the length of the scribing line
interruptions is sufficiently resolved in the EL and DLIT
images. The explanation of the characteristic El and DLIT
patterns of scribing defects is facilitated by the
simulation software that yields the voltage distribution,
which can explain the El images, and the current flow, which
can explain the DLIT images. With regard to the simulation
software itself, I present an alternative solving approach
for the non-linear system of equations, that differs from
the standard Newton-Raphson procedure and allows a scaling
behaviour of the simulation duration with the number of
equations close to the linear optimum. Moreover, EL
measurements of point defects with implemented abnormal
layer sequences show that the i-ZnO/CdS buffer layer
combination has a shunt mitigating effect. From comparison
of various point defect types I conclude that the CdS has
during chemical bath deposition a surface smoothing effect
and fills up cavities and holes in the CIGS absorber. I
suggest that this surface smoothing allows the sputtered
i-ZnO to form a uniform and unbroken layer, and thus prevent
shunting contact between the front and back electrode. In
contrast, when I remove the CdS layer locally, the point
defects show strong shunting in EL and DLIT measurements if
the defect implementation has caused sharp edges or flakes
at the Mo or CIGS layer. Furthermore, Cu-rich debris is a
potential contamination during CIGS co-evaporation, where it
can fall down from the Cu evaporation source onto the
module. For Cu-rich debris I find that its position within
the cell determines whether it can be reliably identified by
EL and DLIT measurements. Cu-rich debris on the P1 line
evokes among all defects a unique EL brightness pattern,
which I explain by the simulation software as an overlay of
a P1 defect with two shunting defects, one located in each
of the neighbouring cells. Eventually, I develop two repair
processes for defective Mo back contact insulation (P1
line). The first process is a thermal repair which uses
thermally induced mechanical stress to create insulating
fractures at the defect position. The second process is an
electrical repair, where an applied current melts and
evaporates remaining conductive Mo at the defect position,
which finally results in an insulation, too. To draw a
conclusion, the present work contributes to the
understanding and diagnostics of production induced defects
in Cu(In,Ga)Se$_2$ photovoltaic thin film modules, and in
case of the P1 scribing defects even shows ways how they can
be repaired.},
cin = {615610},
ddc = {621.3},
cid = {$I:(DE-82)615610_20140620$},
typ = {PUB:(DE-HGF)11},
urn = {urn:nbn:de:hbz:82-rwth-2016-011326},
url = {https://publications.rwth-aachen.de/record/568190},
}
guest ::