Sei sulla pagina 1di 6

See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/258106483

Solar cell efficiency mapping by LBIC

Conference Paper January 2009

CITATIONS READS

4 214

3 authors:

Matev Bokali Marko Jankovec


University of Ljubljana University of Ljubljana
39 PUBLICATIONS 153 CITATIONS 37 PUBLICATIONS 390 CITATIONS

SEE PROFILE SEE PROFILE

Marko Topic
University of Ljubljana
281 PUBLICATIONS 3,178 CITATIONS

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Spatially resolved characterisation in thin film photovoltaics View project

ETIP-PV / Grid integration View project

All content following this page was uploaded by Matev Bokali on 26 July 2017.

The user has requested enhancement of the downloaded file.


45th International Conference on Microelectronics, Devices and Materials
& The Workshop on Advanced Photovoltaic Devices and Technologies
9 11 September 2009, Postojna, Slovenia

Solar Cell Efficiency Mapping by LBIC


Matev Bokali, Marko Jankovec, Marko Topi
University of Ljubljana, Faculty of Electrical Engineering
Traka 25, SI-1000 Ljubljana, Slovenia
Phone: +386 1 4768 276 Fax: +386 1 4264 630
(E-mail: Matevz.Bokalic@fe.uni-lj.si)

Abstract Light beam induced current (LBIC) measurement is performed by illuminating small spots throughout
the whole solar cell area, while measuring the photo-induced current. Results are presented as a map of local
performance distribution of the cell and provide information about various defects, which compromise the overall
cell performance.
The construction of the LBIC apparatus and measuring methodology are explained. Measurement results of two
different solar cells and two mini-modules are presented and discussed. Polycrystalline silicon solar cell local
efficiency map clearly shows crystal boundaries while results for dye-sensitized solar cell indentify local
imperfections and inhomogeneities of hand-made TiO2 layers. LBIC results of PV modules show differences
between performance of cells connected in series in polycrystalline silicon mini-module whereas in the case of the
amorphous silicon PV module an important role of the TCO resistance and the effect of built-in resistance between
the adjacent bottom TCO layers in the monolithic contact are revealed.

different laser sources for their high spatial resolution


1 INTRODUCTION
ability over large area, since the beam can be very
Energy conversion performance is the main precisely focused. In addition, it can also be
characteristic of solar cells and modules. Among upgraded to LBIV and CELLO methods, if required.
other factors (e.g. material purity, light management, The basic principle of the LBIC method, shown on
current collection efficiency) that influence the Fig. 1, is to measure the induced short circuit current,
overall performance of solar cells, localised defects while exposing the device to a focused light beam.
are of great importance, especially in the thin film By selectively illuminating the device surface we
devices. Several measurement methods have been obtain a map of local responses. LBIC measurement
developed for solar cell efficiency characterisation. system is basically comprised of a light source,
The most commonly used method is IV scanning of focusing optics, a positioning system, and measuring
the cell that is homogenously illuminated by the solar equipment. Selection of the elements depends on the
simulator. While this method gives information only required resolution, spot size, specimen type and
for generalised cell performance, which is a sum of size, excitation light wavelength and specimen
local efficiencies, a spatially resolved methods give orientation.
much more detailed insight in the cells performance.
To obtain the information about the local efficiency
distribution, several methods are used, such as
thermography [1], photoluminescence [2],
electroluminescence [3] and several methods based
on light beam excitation: LBIC (Light Beam Induced
Current) [4], LBIV (Light Beam Induced
Voltage) [5] and CELLO (Solar Cell Local
Characterization) [6].
These methods provide the means to identify
localised areas with poorer efficiency, local defects,
weak interconnections and crystal grain boundaries.
In combination with simulations the cause for weaker
Figure 1: Basic LBIC Setup.
areas can be deducted, leading to production process
optimization. Several LBIC setups have been reported, e.g.: small
This article describes the LBIC apparatus and area, high resolution apparatus at Colorado State
presents the measurement results of different types of University [7], large area facilities at Loughborough
solar cells and modules. Although it is relatively time University [8] and at Institute for Environment and
consuming, we have chosen the LBIC method with Sustainability in Ispra, Italy [9].

269
45th International Conference on Microelectronics, Devices and Materials
& The Workshop on Advanced Photovoltaic Devices and Technologies
9 11 September 2009, Postojna, Slovenia

we have chosen an additional computer controlled


2 METHODOLOGY
fibre pigtailed calibrated attenuator (4) with 60 dB
We have designed LBIC setup to measure solar attenuation for this purpose. At the end of the optical
cells, mini modules and optical sensors using a single fibre, the collimator (5) is placed, which prepares the
setup. On one hand, mini modules and larger solar beam for further optical management.
cells measurements require large scan area, while on Collimated laser beam, approximately 3 mm in
the other hand, optical sensors measurements and diameter, then hits the polarization insensitive beam
defect studies require high resolution. Specific splitter (6). 50 % of the beam is steered to the
requirements arise from these goals: reference photodiode (7), the other 50 % passes
scan area 15 by 15 cm, through towards the specimen. Prior to the specimen,
variable illuminated spot size 1 mm2 to 1 m2 a focusing lens (8) is mounted onto a z-axis
maintaining constant light intensity, translation stage (9), which enables computer
comparative measurement for quantum controlled focusing. The specimen is placed
efficiency calculations, vertically into the holder, which is mounted onto x
selectable excitation light wavelength, and and y-axis translation stages (10). Controller (11)
full computer control. manipulated translation stages with 15 cm maximal
The developed system is shown in Fig. 2. travel and typical uni-directional repeatability under
Initially, we have chosen a cost effective 638 nm 1 m have been chosen for consistent specimen
single-mode fibre pigtailed laser diode (1) with positioning.
power of 20 mW to cover estimated 13 dB losses A simple resistor is used as a load for specimen to
along optical path. When appropriately driven, it measure the current. The resistor value has to be
produces sine wave modulated laser beam. Single chosen carefully. It must be low enough to keep the
mode fibre maintains Gaussian beam and allows an specimen very close to the short-circuit conditions,
easy replacement with compatible laser sources. The but at the same time it has to be high enough, so that
Laser diode controller (2) operates in one of the two the voltage on the resistor can be accurately
modes: open loop current control mode, or closed measured. In case of the smallest spot size
loop power control mode. Both modes support sine illuminated with AM 1.5 intensity, the maximum
wave modulation up to 20 kHz that is obtained from induced current is lower than 1 nA, therefore a lock-
the waveform generator (3), which is also used as a in measuring technique is used. Lock-in
reference signal source for lock-in amplifiers. While amplifiers (12) have the ability to measure the
modulating the laser diode supply current, we must component of known frequency in a noisy
ensure that the current does not fall under the diodes environment because they exploit the orthogonal
threshold current or exceeds the diodes peak power. properties of the sine function.
In order to comply with high laser intensity Finally, all the components are connected to the
dynamics requirements (constant laser spot light computer (13), where a LabVIEW application
intensity with area spanning from 1 mm2 to 1 m2) controls the measurements.

Figure 2: Designed LBIC Setup.

270
45th International Conference on Microelectronics, Devices and Materials
& The Workshop on Advanced Photovoltaic Devices and Technologies
9 11 September 2009, Postojna, Slovenia

The application features a modular design. Each reveals that the axes of the collimator, the z-axis
computer controlled component of the setup translation stage and the lens are not sufficiently
communicates with its own software module. This aligned. Further improvements are necessary and are
design approach allows simple setup modifications. already under way. The collimator is being replaced
At the start-up, the application initializes all the with an adjustable one and the focusing lens is being
modules and sets all the positioning stages in their mounted onto a z-axis translation stage via 6-axis
home position. In order to keep the spot size the same adjustable mount. This will allow us to adjust the
throughout the whole specimen area, the exact collimator and the lens mount precisely, so we can
position of specimen plane has to be defined. This ensure accurate optical axis alignment and reach a
allows correction of the focusing lens position during diffraction limited performance of the optical system
the measurement and keeps the spot size constant. and a smaller spot size.
Beam diameter is measured with a knife-edge
technique [10]. Light beam is focused onto a photo-
detector. The response current is measured while the
knife blocks the beam in small steps. The distance
between the positions with 10 % and 90 % of the
maximal response current value is defined as a beam
width. Instead of the actual knife, a contact finger, a
sharp cell edge or a defect can be used to block the
beam, thus measuring the beam width exactly on the
surface plane of the specimen. The beam diameter
has to be approximately the same as the measurement
step in order to uniformly cover the whole cell area.
1 D and 2 D measurement methods are Figure 4: Beam width vs. lens position.
implemented. The results of the first one are
presented as a magnitude vs. position graph, while 3.2. Microcrystalline solar cell
the results of the second one are presented on a y vs. The first measurement was carried out over the
x graph where the magnitude is presented either on whole area of microcrystalline silicon (Si) solar cell
the z axis or in a colour intensity scale. (100 mm by 52 mm). For that we chose a resolution
To enable repeatability of the measurements and to of 500 m, which resulted in 20,800 measurement
allow for additional calculations, measured values are points. The results are shown in Fig. 5.
stored raw together with all the configuration Typical structure of the solar cell with metal
parameters. contact fingers is observed. The horizontal contact
3 RESULTS AND DISCUSSION fingers are not perfectly aligned with the scanning
x-axis, therefore aliasing effect is noticeable along
3.1. Beam focusing, spot size them. The boundaries between adjacent crystals with
With this equipment we achieved the minimal spot lower performance in short-circuit current are clearly
width of 27 m, as shown in Fig. 3. seen, and differences in efficiency between crystals
are also evident. The zero current area at the bottom
is not the defect, but a shadow of measuring
equipment. There is another defect visible at
(x = 138000 m, y = 20000 m) which cannot be
attributed to a single crystal or optical defect.

Figure 3: Beam width measurement with knife-edge


method.
We performed additional measurements to
investigate why the spot size does not meet the Figure 5: LBIC measurement results of the
requirements. Beam width vs. z-axis graph in Fig. 4 microcrystalline Si solar cell (Relative response).

271
45th International Conference on Microelectronics, Devices and Materials
& The Workshop on Advanced Photovoltaic Devices and Technologies
9 11 September 2009, Postojna, Slovenia

3.3. Dye-sensitized solar cell lower cell in Fig. 7 with positive response was
directly connected into the measuring circuit, as
A measurement of a small 55 mm2 dye-sensitized
marked with (1) in Fig. 8. Surprisingly, the upper
solar cell was performed, where a 100 m resolution
cell, marked with (2) in Fig. 8, was not directly
was chosen. A cell that gave insufficient results was
connected into the measuring circuit, but nonetheless
chosen among a series of cells, fabricated in our
generated a response. This can be explained by the
laboratory [11].
nature of monolithic contact, as shown in Fig. 8. The
Results in Fig. 6 show large variations in local
resistances (3, 4) across the isolation cut (where the
performance distribution and even some almost dead
front TCO on glass substrate is removed) are low
areas. One of the reasons for such inhomogeneous
enough to provide an auxiliary path for the current
response lays in the hand-made production.
and the back TCO to front TCO resistance (5) is high
enough to form a current divider. Then, the current is
measured with a negative sign.

Figure 6: LBIC measurement results of the DSSC Figure 8: Schematic cross-section of three adjacent
(Relative response). cells.
3.4. Cell in amorphous silicon module The influence of the TCO and the contact resistance
One cell of the amorphous silicon module is evident in the results. On the top of the lower cell,
represents typical conventional thin-film solar cells. from x = 90000 m to x = 75000 m the contact to
Although one cell is only 5 mm long, a larger area of measuring wire supports good current collection and
16x35 mm2 was chosen for the scan, which included consequently good performance. Next 10 mm to the
a neighbouring cell as well. The cell was contacted left, the silver paste and the TCO collect the current,
via a layer of silver paste 25 mm long, but only the therefore the performance is decreasing. Further to
first 15 mm of the layer from the right side was the left, only TCO collects the current, but the
contacted by a measuring wire. This contact resistance is too high and the performance is very
configuration allows observation of how the contact poor.
quality influences cell performance. The measured 3.5. Crystalline mini module
area was scanned with a resolution of 500 m and the For observing the local efficiency measurement on
results are shown in Fig. 7. a module of cells connected in series the crystalline
silicon mini module combined of 8 cells was
measured with the resolution of 500 m. The results
on Fig. 9 clearly show all 8 cells. Single cell is
observed as a thick vertical bar of approximately the
same local efficiency; in addition, between the cells a
line of lower local efficiency is visible, because the
cells are not accurately positioned close to each
other. One might assume that the cell which generates
the biggest LBIC response contributes the most to the
overall module efficiency. However, because of the
Figure 7: LBIC measurement results of the cell in
influence of other cells connected in series, this is not
amorphous silicon module (Relative response).
necessarily true. Further measurements and
Interesting effects, caused by the monolithic contact simulations must be conducted to give the correct
structure, were noticed on the neighbouring cell. The interpretation of the results.

272
45th International Conference on Microelectronics, Devices and Materials
& The Workshop on Advanced Photovoltaic Devices and Technologies
9 11 September 2009, Postojna, Slovenia

References
[1] O. Breitenstein, M. Langenkamp, K. Itoh, et al.,
"Lock-in Thermography", Advanced
Microelectronics, Springer, Berlin, 2003.
[2] T. Trupke, R.A. Bardos, "Photoluminescence: a
surprisingly sensitive life time technique",
Proceedings of the 31st IEEE Photovoltaic
Specialists Conference and Exhibition, 37
January 2005, Orlando, FL, USA, pp. 903906.
[3] T. Fuyuki, H. Kondo, Y. Kaji, A. Ogane,
Y. Takahashi, "Analytic findings in the
electroluminescence characterization of
crystalline silicon solar cells", Journal of Applied
Physics, Vol. 101, Issue 2, 2007.
Figure 9: LBIC measurement results of the [4] M. Rinio, H. J. Mller, M. Werner, LBIC
microcrystalline mini-module (Relative response). Investigations of the Lifetime Degradation by
Extended Defects in Multicrystalline Solar
Similarly to the measurements of the Silicon, Solid State Phenomena, vol. 63-64,
microcrystalline Si solar cell, the front side metal pp. 115-122, 1998.
finger contacts are shown; however, their [5] J. Salinger, V. Benda, Z. Machacek, A Note on
representation is unexpected. They are shown as Solar Cell Diagnostics Using LBIC and LBIV
regions with higher performance, which is Methods, Proceedings of 26th International
contradictory to the results observed in Fig. 5. There, Conference on Microelectronics 2008, 11-14
the contact finger blocks the beam, so that it does not May 2008, Ni, Serbia.
reach the active area of the cell. The reason for this [6] J. Carstensen, G. Popkirov, J. Bahr, H. Fll,
unusual behaviour in Fig. 9 is the protective coating CELLO: an Advanced LBIC Measurement
of the module which traps the reflected laser beam at Technique for Solar Cell Local Characterization,
the front metal finger contacts and scatters it over a Solar Energy Materials & Solar Cells, vol. 76,
larger cell area and surrounding cells. This causes the pp. 599-611, 2003.
current generation in other cells, which results in a [7] J. R. Sites, T. J. Nagle, LBIC Analysis of Thin-
relatively higher measured response. Film Polycrystalline Solar Cells, Record of
As is already the case with this simple Photovoltaic Specialist Conference, 3-7 January,
measurement, the interpretation of the LBIC results 2005, pp. 199-204.
of the modules is very demanding, but it generates [8] T. Michel, R. Gottschalg, D. G. Infield, Large
interesting results [12]. Certainly, a need for absolute Area Laser Scanning of Amorphous Silicon
quantification of measurement results is needed. We Modules, Proceedings of the 19th European
plan to calibrate the LBIC system for measurement of Photovoltaic Solar Energy Conference, 711 June
wavelength- and voltage-dependent collection 2004, Paris, France, pp. 25222524.
efficiency. [9] H. Mllejans, F. Merli, E. D. Dunlop, Large
Area Photocurrent Maps for Routine PV Module
5 CONCLUSIONS Inspection, Proceedings of PV for Europe,
The LBIC setup has been explained in detail. The Rome, Italy, 2002.
setup has been successfully used for measuring [10] W. Plass, R. Maestle, K. Wittig, A. Voss,
specimens ranging from single solar cells and A. Giesen, High-Resolution Knife-Edge Laser
modules to optical sensors. Measurements of four Beam Profiling, Optics Communications,
different types of PV devices have been presented, vol. 134, issues 1-6, pp. 21.24, 1997.
explanation for defects and behaviour has been [11] U. Opara Kraovec, M. Berginc, M. Hoevar,
enclosed. M. Topi, Unique TiO2 Paste for High
More work will be done on improving the setup and Efficiency Dye-Sensitized Solar Cells, Solar
on development of the simulation equipment, which Energy Materials & Solar Cells, vol. 93, pp. 379-
will help us to better understand the measurements. 381, 2009.
[12] I. L. Eisgruber, J. R. Sites, Extraction of
Acknowledgments Individual-Cell Photocurrents and Shunt
The development of this setup was funded by Resistances in Encapsulated Modules using
Slovenian Research Agency. The authors would like Large-Scale Laser Scanning, Progress in
to thank the DSSC team of LPVO for providing a Photovoltaics: Research and Application, vol. 4,
DSSC solar cell for measurements. pp. 63-75, 1996.

273
View publication stats

Potrebbero piacerti anche