Preprint to be published in the proceedings of the
31st European Photovoltaic Solar Energy Conference, 14-18 September 2015, Hamburg, Germany
A NEW LIGHT INDUCED VOLUME DEGRADATION EFFECT OF MC-SI SOLAR CELLS AND MODULES
F. Kersten1,2*, P. Engelhart1, H.-C. Ploigt1, F. Stenzel1, K. Petter1, T. Lindner1,
A. Szpeth1, M. Bartzsch1, A. Stekolnikov1, M. Scherff1, J. Heitmann2 and J.W. Müller1
1Hanwha Q CELLS, OT Thalheim, Sonnenallee 17-21, D-06766 Bitterfeld-Wolfen, Germany
2TU Bergakademie Freiberg, Institute of Applied Physics, Leipziger Straße 23, D-09599 Freiberg, Germany
*Phone: +49 (0)3494 66 99-52134; e-mail: f.kersten@q-cells.com
ABSTRACT: In this work the performance stability of rear side passivated mc-Si solar cells and modules under carrier
injection at different temperatures is investigated. Severe degradation levels of above 10% can be detected which cannot
be explained by B-O complex formation or FeB pair dissociation. A high statistic of cells and modules degraded in lab
and outdoor using material from different suppliers confirm the relevance of this new effect. LeTID (Light and elevated
Temperature Induced Degradation) is a mc-Si bulk phenomena leading to a highly injection dependent degradation and
features a regeneration phase after degradation. Characteristics of LeTID as a function of temperature and injection
level are presented and a comparison between laboratory and outdoor tests is drawn. The time constant of this
degradation mechanism accelerates with increasing temperature, however, the time span for degradation and
regeneration of thousands of hours at relevant temperatures between 60-85°C demands for a solution on wafer material
or processing side. LeTID can be significantly reduced by adapting the cell process and processing sequence.
Keywords: PERC, Degradation, Multicrystalline Silicon, Solar Cell, PV Module, Lifetime
1
3 CHARACTERISTICS AND OBSERVATIONS OF
LeTID ON MULTICRYSTALLINE SILICON
INTRODUCTION
Multicrystalline silicon (mc-Si) material achieved a
market share within the c-Si PV of about 60% in 2014 [1]
and will remain the dominant material within the next 10
years. Additionally, the International Technology
Roadmap for Photovoltaic (ITRPV) predicts a mainstream
market for multi PERC (Passivated Emitter and Rear Cell).
Thus, to support the PV industry, more research is needed
on mc-Si material especially on next generation cell
concepts. Recently some research on the light induced
degradation of mc-Si PERC cells show an unexpected
strong degradation level [2-5]. Unfortunately, there is no
test standard in c-Si PV for the temperature range during
LID and no agreed LID test parameters at elevated
temperatures (50-80°C) and time (> 48 h) to cover field
relevant conditions. In this work the degradation at
elevated temperature is emphasized and a new mc-Si
“long-term” volume degradation and regeneration
mechanism called LeTID (Light and elevated Temperature
Induced Degradation) is characterized, which can cause
degradations of above 10%.
2
3.1 Results on PERC cells and mc-Si lifetime samples
Fig. 1 shows in-situ measured VOC data of four mc-Si
(neighboring sorted material) LeTID-sensitive PERC solar
cells at two different elevated temperatures (50°C and
95°C) and at two operational modes (VOC and ISC mode)
during illumination at 300 Wm-2 [4]. The measured VOC
values are illumination and temperature corrected to STC
(25°C, 1000 Wm-2) and normalised to the initial VOC value.
A significant VOC degradation at 95°C of ~9% after
approximately 150 h can be observed by operating in VOC
mode. The PERC cell which was hold at the same
conditions but operated in ISC mode (thus lower injection
level) shows a slower degradation behavior. The same
experiment is carried out at a lower temperature with two
other cells at 50°C. The degradation obviously happens
much slower than at higher temperature and again for the
lower injection density at ISC mode the degradation
happens even slower. From Fig. 1 we conclude that LeTID
is accelerated by either a higher temperature or higher
injection level. After reaching the maximum degradation
level LeTID features also a regeneration effect. After
~1000 h at 95°C in VOC mode, the PERC cells are almost
completely recovered.
Fig. 2 displays the degradation of mc-PERC cells
representing a whole brick distribution (from bottom to
top) after illumination at 1000 Wm-2 and 60°C for 24 h [4].
Please note that the degradation values after 24 h represent
a “snapshot” only, since the degradation maximum is
achieved not before several hundreds of hours (see Fig. 1).
From Fig. 2 we conclude that LeTID does not correlate to
the interstitial oxygen concentration. Fig. 3 shows the
degradation and regeneration cycle of mc-PERC cells
representing a brick distribution degraded at 75°C in CID
VOC mode. The degradation values correspond to Fig. 2
and show the highest value for the cells of tree-quarter of
brick height. The top cells also show a slower degradation
rate and a lower degradation level. All cells are fully
recovered after 2000 h to initial VOC value. In our
experiments a LeTID dependency of the brick height was
carried out, which suggests a Si volume effect. This
assumption was verified by varying the surface passiva-
EXPERIMENTAL
Mc-Si wafer material from different wafer suppliers
(several ingots with complete brick distributions) are
processed to degradation susceptible lifetime samples,
PERC cells and modules. The interstitial oxygen
concentration of the Si material is measured by infrared
absorption spectroscopy. The degradation experiments in
the lab were carried out at different temperatures from
50°C to 95°C on a hotplate or in a climatic chamber on cell
or module level, respectively. We extend the procedure to
significantly longer time periods (simulating over 20 years
roof top operation in Germany). The excess carriers are
injected either by illumination at 0.3-1 suns (LID, Light
Induced Degradation) or by current injection (CID,
Current Induced Degradation). In addition to LID the
excess carrier concentration was varied to operate the cells
either in VOC or ISC mode. For CID we adjust the current to
the same excess carrier density as for operation in ISC,
MPP or VOC mode under illumination at 1 sun.
1
Preprint to be published in the proceedings of the
31st European Photovoltaic Solar Energy Conference, 14-18 September 2015, Hamburg, Germany
the low degradation rate of LeTID a further degradation is
expected during the next few hundreds of hours (as
discussed in Fig. 1-3). However, beside the reduced
lifetime also a change of the injection dependency can be
observed. LeTID causes an increased lifetime degradation
effect at lower injection levels of e.g. n = 11013 cm-3
than at higher injection of e.g. n = 11015 cm-3. This
increased injection dependency of the minority carrier
lifetime can explain the observed degradation
characteristics as displayed in Fig. 2. The increased
injection dependency causes a main loss in ISC and
additionally leads also to an increased non-ideality of the
solar cell causing a loss in FF.
Fig. 5 shows the longtime lifetime measurement at
n = 41013 cm-3 plotted as normalised defect
concentration Nt* [6] of tree different surface passivation
types after degradation (300 Wm-2, 75°C). LeTID
represents the inverse Nt* and upper lifetime limit. One
type of sample is symmetrical coated by an Al2O3/SiNxstack (negative surface charge). The other samples are
only passivated with SiNx (positive surface charge) or
Al2O3 (negative surface charge). For both SiNxpassivation types the same time-dependent degradation
behavior is observed. In contrast to the lifetime
degradation snapshot after 24 h of τeff = 18 µs, after 1100 h,
Figure 1: VOC data of four mc-PERC cells during
illumination at two temperatures and operational modes
(50°C and 95°C; VOC and ISC mode).
Figure 2: Performance loss and interstitial oxygen
concentration of mc-PERC solar cells (representing a
brick distribution) degraded at 60°C for 24 h.
Figure 4: Effective minority carrier lifetime of a
passivated mc-Si wafer before and after degradation.
Figure 3: VOC degradation and regeneration cycle of mcPERC cells representing a brick distribution degraded at
75°C in CID VOC mode. There are missing data points at
180 h and 600 h because of measuring errors.
tion layers of PERC cells and lifetime samples (Fig. 5),
showing similar results. Due to the extremly low
degradation rate compared to LID and the mismatch with
the O-concentration we exclude B-O and FeB as the root
cause of LeTID.
Fig. 4 displays the injection dependent minority carrier
lifetime characteristic of surface passivated mc-Si wafers
before and after degradation process [4]. After 24 h
degradation (1000 Wm-2, 75°C, 24 h) a drop in lifetime
from eff = 137 µs (before) to eff = 14 µs is detected,
measured at an injection level of n = 41013 cm-3 which
corresponds to MPP conditions in operation mode. Due to
Figure 5: Effective minority carrier lifetime of a
passivated mc-Si wafer after long time degradation of
different passivation layers.
2
Preprint to be published in the proceedings of the
31st European Photovoltaic Solar Energy Conference, 14-18 September 2015, Hamburg, Germany
(a)
(b)
Figure 6: Absolute VOC in a) of neighboring sorted solar cells (representing a brick distribution) after degradation
and subsequent regeneration cycle at different conditions (1000 Wm-2, 75°C/25°C, 24 h) and in b) in-situ
VOC measurement of one solar cell referring to the sequence in a) under regeneration (i) → (ii)
(1000 Wm-2, 25°C) and subsequent degradation conditions (ii*) → (iii) (1000 Wm-2, 60°C).
a further severe lifetime degradation up to τeff = 6 µs is
measured. The lifetime sample passivated only with Al2O3
shows a slower degradation behavior, but degrades down
to τeff = 19 µs also. From Fig. 5 one can conclude that the
degradation of mc-Si is independent of surface passivation
charge and surface type. This is another hint for a Si bulk
degradation effect. In contrast with solar cell measurement
(orange dots in Fig. 1) the lifetime samples feature no
regeneration phase after more than 1500 h degradation
process and further time for regeneration is required.
Fig. 6 shows the absolute VOC in relation to the initial
value of neighboring sorted solar cells after different
degradation and regeneration temperatures at 1000 Wm-2.
The 1st and 3rd degradation processes (Fig. 6 a) are carried
out after 24 h at 75°C with an intermediate 2nd regeneration
step at 25°C. The average VOC value after 1st degradation
is -8.1 mV, it regenerates instable after 2nd step to -3.2 mV
and degrades afterwards again to -8.6 mV. After this
procedure one solar cell was extracted and an in-situ VOC
measurement was carried out. As shown in Fig. 6 b) the
instable regeneration (i) → (ii) at 25°C taking place in the
first subsequent degradation (ii*) → (iii) at 60°C shows no
saturation in VOC signal after 20 h. That means that lower
temperatures combined with high illumination intensities
regenerate the LeTID to different instable degradation
states. This is an important fact for analyzing the LeTID
under real field conditions.
3.2 Relevance of LeTID on mc-PERC modules in field
Fig. 7 shows the degradation-regeneration cycle of a
LeTID-sensitive PERC module installed during a half year
from summer to winter on the outdoor test field (located in
Germany). The sequence starts with a homogenous EL
image and a normalised module power to 100%. To
accelerate the degradation-regeneration cycle the module
was operated in VOC mode and a thermal isolation out of
Styrofoam was placed onto the module rear side in order
to increase the module temperature comparable to
conditions recurrent in hot climates. After 203 kWh, the
module performance decreased by 7.5% and a strong
inhomogeneity appears in the EL image. Different cells
show different degradation levels and time constants.
After 320 kWh, the module started to regenerate. After
520 kWh, almost the full original module power is
recovered. However, in MPP mode or in central European
climate conditions the regeneration takes too long to take
effect during the warranty period of a commercial PV
module (see time scale of open triangles in Fig. 8).
Figure 7: EL and module power measurement (STC) sequence showing the degradation-regeneration cycle of one
outdoor mounted module with rear side insulation. The kWh and %-value corresponds to the cumulated
irradiation impinged onto the module and the relative power degradation, respectively.
3
Preprint to be published in the proceedings of the
31st European Photovoltaic Solar Energy Conference, 14-18 September 2015, Hamburg, Germany
Figure 8: 60-cell mc-PERC module degradation measurements at T = 60-85°C in VOC and MPP mode.
3.3 Avoidance of LeTID and proposal for external
LID/CID tests
Fig. 8 shows the time dependent module power
degradation for different carrier injection densities [4].
The left side of the graph shows the equivalence of LID
(▲) and CID (●) at 85°C in VOC mode with a rapid
degradation of about 8% during the first 80 h and a
recovery to 2% during the following 400 h. In MPP mode
( and o), LeTID-sensitive modules lose more power (up
to 11%) but at a much smaller rate compared to VOC mode.
The right side of the graph shows the power loss over time
of two groups of modules after CID treatment at 75°C in
MPP mode. Medium LeTID-sensitive modules show a
power loss of about 4%. The other group shows our
optimized defect engineered Q.Antum modules, with
degradation levels of < 1%. A second time scale, on the
top of the graph, shows the specific corresponding time
period for two locations, Germany and Cyprus. We
propose CID treatment at 75°C in MPP mode to estimate
LeTID based power losses to simulate long-term field
condition in a reasonable time scale. With this method,
high numbers of modules can be tested in parallel at
moderate cost. We measured real module temperatures
over several months and years at our different worldwide
test fields (Germany, Cyprus, Australia and Malaysia).
Together with a model for the temperature dependency of
LeTID deduced from experiments we calculate a test site
specific corresponding time period for the proposed test at
75°C; e.g. 10/20 years Germany roof top or 5 years Cyprus
field as plotted in Fig 8.
By adapting the process conditions and sequence we
are able to influence LeTID. Fig. 8 plots time dependent
degradation of PERC modules with varying LeTID
sensitivity (max. > 10% and medium in the range of 34%). Applying our optimized defect engineered mc-PERC
process, we are able to fully avoid any LeTID with a
degradation level of < 1% at 75°C after 1000 h, which
corresponds to B-O degradation level and more than 20
years of module life time in a roof top installation in
Germany, respectively. We can conclude that our
Q.ANTUM [7] module is not susceptible to LeTID.
4
CONCLUSION
This study reports on characteristics of a new mc-Si
volume degradation effect. Based on high number of cells
and modules degraded in lab and outdoor and utilizing
material from different wafer suppliers we characterized
the mc-PERC degradation effect called LeTID. Lifetime
sample experiments including variation of the surface
passivation layers and investigations on cells within brick
height dependency allowed us to conclude that LeTID is a
strong Si material effect. On mc-Si PERC cells and
modules the performance loss of above 10% is caused by
carrier injection at elevated temperatures on a time scale
of several hundreds of hours. Both, B-O and FeB can be
excluded as the root cause. Higher injection densities or
elevated temperature accelerates the degradation. Lower
temperatures combined with high illumination intensities
after higher temperature degradation steps regenerate the
LeTID to a lower instable degradation state. From surface
passivation variation experiments we conclude that the
degradation of mc-Si is independent of surface passivation
charge and surface type. After reaching the maximum
degradation level, there is also a regeneration effect.
However, the time span for degradation plus regeneration
of thousands of hours is too long to take effect during
warranty period. Our proposal for a LeTID test norm with
moderate implementation effort and costs is CID at 75°C
in MPP mode. LeTID is highly relevant in practice
(outdoor demonstration) and potentially may hinder the
transition to high-efficiency mc-PERC technology in
industry. However, as a result of intensive R&D work at
Hanwha Q CELLS on this topic we demonstrate the
capability to decrease the degradation level and the
complete avoidance of LeTID by adapting the cell process
and processing sequence and show that our Q.ANTUM
modules are not susceptible to LeTID.
5
ACKNOWLEDGEMENTS
Special thanks to the staff of the Reiner Lemoine
Research Center, Module Test Center and pilot line at
4
Preprint to be published in the proceedings of the
31st European Photovoltaic Solar Energy Conference, 14-18 September 2015, Hamburg, Germany
Hanwha Q CELLS. This work was financial supported by
German Federal Ministry for Economic Affairs and
Energy (BMWi) under contract no. 0325775 (“AdmMo”).
6
REFERENCES
[1] International Technology Roadmap for Photovoltaic
(ITRPV), 2013 Results, Revision 1, 24 March 2014.
[2] K. Ramspeck, S. Zimmermann, H. Nagel, A. Metz, Y.
Gassenbauer, B. Brikmann, A. Seidl, Light induced
degradation of rear passivated mc-Si solar cells, in:
Proceedings of the 27th EUPVSEC, Frankfurt,
Germany 2012, pp. 861–865.
[3] F. Fertig, K. Krauß, S. Rein, Light-induced degradation of PECVD aluminium oxide passivated
silicon solar cells, Phys. Status Solidi RRL (2014) 16.
[4] F. Kersten, P. Engelhart, H.-C. Ploigt, A. Stekolnikov,
T. Lindner, F. Stenzel, M. Bartzsch, A. Szpeth, K.
Petter, J. Heitmann and J.W. Müller, Degradation of
multicrystalline silicon solar cells and modules after
illumination at elevated temperature, in: Solar Energy
Materials and Solar Cells 142 (2015) 83-86.
[5] H.-C. Ploigt, F. Kersten, P. Engelhart, A. Stekolnikov,
T. Lindner, F. Stenzel, M. Bartzsch, A. Szpeth, K.
Petter, J. Heitmann and J.W. Müller, A New mc-Si
Degradation Effect called LeTID, in: Proceedings of
the 42nd IEEE PVSC, New Orleans, USA, 2015.
[6] K. Bothe, J. Schmidt, Electronically activated boronoxygen-related recombination centers in crystalline
silicon, J. Appl. Phys. 99 (2006) 013701.
[7] P. Engelhart, D. Manger, B. Klöter, S. Hermann, A.
Stekolnikov, S. Peters, H.-C. Ploigt, A. Eifler, C.
Klenke, A. Mohr, G. Zimmermann, B. Barkenfelt, K.
Suva, J. Wendt, T. Kaden, S. Rupp, D. Rycharik, M.
Fischer, J. W. Müller, P. Wawer, Q.ANTUM – QCells next generation high-power silicon cell &
module concept, in: Proceedings of the 26th
EUPVSEC, Hamburg, Germany, 2011, pp. 821–826.
5