ADVANCED TESTING OF PV MODULE ENCAPSULANTS
S. Pingel1-2*, S. Janke2, B. Stannowski2, S. Fechner1 and L. Podlowski1
1: PI Photovoltaik-Institut Berlin AG (PIB), Wrangelstr. 100, 10997 Berlin, Germany
2: PVComB, HZB Helmholtz Zentrum Berlin
*: now with Fraunhofer ISE
ABSTRACT: Encapsulation materials are of critical importance for long-term reliability and safety of PV modules.
Only months in the ﬁeld may lead to drastic power output degradation for example due to PID and also in the longer
run adhesion and discoloration issues (Hot Spot) can reduce power output or even may lead to critical electrical safety
issues. These ﬁndings are not sufficiently covered by IEC 61215 and other standard testing. Additional and different
types of tests are required for an in-depth understanding of the encapsulation material’s impact on PV module
performance. PI Berlin has developed a test sequence that covers many ﬁeld relevant aspects for PV encapsulations.
For example: i) Potential Induced Degradation (PID), ii) Hot Spot Durability and iii) High Temperature Adhesion
(delamination). Additionally, material properties which impact the processability in module manufacturing are tested.
Also, very new aspects like the effect of encapsulation aging on PID are evaluated.
Keywords: Reliability, PID, PV Module
Close to 100% of the modules in the PV market are
laminated using encapsulation material which is based on
a certain chemical raisin. The manufacturing companies
have different recipes for different products. These vary in
the base material and a product specific composition of
different additives which are essential e.g. for the chemical
cross linking, UV stabilization, glass adhesion, etc.
There are many different encapsulation materials used in
the PV industry. According to ITRPV  the most
dominant material will remain to be ethylene-vinyl-acetate
(EVA) with more than 70% predicted market share in 2025
but also polyolefin (POE) materials will become more
relevant, here more than 20% share in 2025 is predicted.
Ten industry relevant encapsulation materials are
compared in the test procedure presented in this paper. The
material types range from EVA to POE, but also more
exotic materials like polyvinyl butyral (PVB) and ionomer
The test procedure covers different aspects relevant for
module operation in harsh field conditions. In the Fig.1
below the test sequence is schematically shown.
Fig.1: The test sequence for encapsulation materials.
Highlighted in black are the tests addressed in this work
while the tests in grey are not shown within this paper.
Central are the following four aspects:
- Potential Induced Degradation (PID) 
- Hot Spot Durability [3,4]
- Water interaction
- High Temperature Adhesion (Delamination) 
In the last few years many module manufacturers have
diversified their product range from glass-backsheet (GB)
modules to glass-glass (GG) modules. Glass-glass
modules are expected to be more stable in the long run,
product warranty is typically extended for this type of
modules. Since the two types of modules are relevant and
the share of GG modules is expected to rise in this study
both module configurations are tested in parallel.
2 MATERIALS AND TEST SEQUENCE
In the following table (Tab.1) the ten encapsulation
materials compared in this project are listed. Besides EVA
and POE also an ionomer and a PVB material is included.
This makes the benchmark testing even more interesting
since the variability of material’s characteristics is much
broader and also extreme cases are covered. Most of the
tested materials are chemically crosslinking, besides the
ionomer that is crosslinked via ionic bonds and the non-
The materials vary in their UV transmission, where two
types are separated:
i) materials with high transmission in the UV range (so
called UV-transmissive - UVT) which is usually used in
front of the solar cells to allow also current generation by
ii) materials with low UV transmission which is achieved
by adding UV absorbers that are meant to protect the
layers below from UV aging. These materials are typically
called UV-close (UVC) and are used behind the solar cells
to reduce the UV transmitted to the backsheet which might
degrade from the UV stress.
Tab.1: Encapsulation materials compared in the test
sequence. Materials marked with (c) are combined
UVT/UVC materials and (s) stands for single materials.
The test samples designed for the different tests are
described at the beginning of the particular chapters.
Typically, the test samples were made from all ten types
of encapsulation materials. Only in case of single solar cell
coupons only seven variations were laminated and tested.
These consisted of four single materials (Tab.1: (s)) and
three UVT/UVC combinations (c).
All samples were laminated with a 20min lamination
process at 150°C. The process was qualified by the Soxhelt
method: all chemically cross-linking materials had a gel
content level in the target range given by the
manufacturers. Only the PVB material needed an adaption
of the process due to bubble formation.
For characterization IV curves were recorded with a
Halm flasher, EL images were taken with a high-resolution
Si camera from Sensovation. For visual inspection photos
were taken and a Lambda 1050 spectrometer from Perkin
Elmer was utilized for transmission measurements.
Before the results are presented here an important side
note: for all results the type of material is named but to
protect the intellectual property of the participating
encapsulation producers the companies and product name
are not disclosed. Additionally, the results are randomly
mixed. This means that for example EVA.1 in DH must
not correspond to EVA.1 in PID.
3 DAMP HEAT TESTING
3.1 Long-term Damp Heat test
Damp Heat (DH 85°C / 85% RH) testing is one of the
frequently applied tests in PV industry for certification and
quality control. According to IEC61215  a DH test for
1000 hours is necessary for module qualification. The
modules are inspected visually, electrical insulation is
measured and output power degradation is determined
(power degradation criterium <5%). In PV industry DH
tests are typically extended to 2000 or even 3000 hours .
In this project testing was further extended to DH5000 to
identify quality differences between the materials. For this
test single cell coupons were made from p-type multi
crystalline Al-BSF cells with SiNx AR coating and screen
printed & fired contacts. For module making a standard
PV ribbon was soldered on the busbars of the cells. For
lamination encapsulation material was placed on the front
and rear side of the cells. As mentioned above a total of
seven combinations were used, four single materials and
three encapsulation combination. In the latter case the
UVT material was placed in front of the solar cell and the
UVC material on the rear side. Per combination two
module designs GG and GB were laminated. For each
variation two coupons were made. The materials are
compared concerning the long-time reliability of the
output power. The power degradation versus DH duration
is shown in Fig. 2 below. In most cases up to DH3000
degradation is negligible small.
In GG configuration modules are stable for 3000 h
since water ingress is limited by the front and rear glass to
the edges of the module. The samples made with the two
types of EVA material are among the first to degrade in
the interval DH3000-5000, most likely the formed acetic
acid leads to corrosion of the cell metallization. POE.2 also
shows a borderline result for DH5000 with about 4%
degradation while the other two POE, the PVB and
ionomer samples are with <2% power degradation stable
Fig. 2: Power degradation in DH5000 for GB (top) and
GB (bottom) single cell modules.
For GB modules the result changes due to faster and
stronger water ingress via the backsheet. POE.2 seems to
lead to DH susceptible modules since the samples degrade
by 8% already after DH1000, a clear fail. The EVA
materials again show earlier and even stronger degradation
compared to GG around DH3000-4000. For GB also PVB
is showing degradation after DH4000. As found already
for GG the ionomer and two POE materials achieve stable
results in DH5000. Water indicator tests (not shown in this
paper) inspired by  show that GB samples are
influenced by humidity already before being placed in the
climate chamber (color change at 8% after storage in the
laboratory). For GG samples in a distance of 4 cm from the
sample edge a level of about ~8% RH is reached after
~DH500 (latest DH1000). This comparison shows that DH
stability varies with the encapsulation material and also the
module layup has a significant influence. It is important to
test all relevant combinations.
3.2 DH comparison SHJ versus multi BSF cells
Also for DH some GG coupons were made from n-
type bifacial silicon hetero-junction (SHJ) cells. In these
cells the amorphous silicon passivation layers are covered
by a transparent conductive oxide (TCO: ITO) on front and
rear side. Due to the temperature sensitivity of the thin
passivation layers a low temperature Ag-paste screen
printing metallization (dried and cured at low temperature)
was applied. The PV ribbons were glued with a conductive
adhesive (ECA) to the busbars of the cell. These type of
modules or cells is expected to be more susceptible to DH
degradation due to the low temperature metallization
process and the TCO that is known to degrade when
exposed to humidity .
Fig. 3: Comparison of power degradation in Damp Heat
for GG modules made from SHJ and multi solar cells.
The results of the comparison SHJ versus standard multi
BSF cells is shown above in Fig. 3. In combination with
EVA, SHJ cell degrades much faster (SHJ DH2000 -29%
vs multi -4% DH4000). EL images in Fig. 4 below show
the degradation of the SHJ cell (corrosion of TCO and or
metallization) after DH2000 while the multi cell is stable.
Fig. 4: EL images of SHJ (left) and multi (right) solar cells
in GG-EVA coupons after DH2000.
In this comparison the combination of ionomer with this
type of SHJ solar cell degraded quite fast, maybe the root
cause of this strong degradation is the susceptible cell
metallization combined with the stiffness of the ionomer
or the chemistry of the ionomer leads to accelerated cell
degradation. Ionomer materials are known to have a low
water vapor transmission rate (WVTR) what makes the
material group interesting for solar cells that are
susceptible to humidity, fast water ingress should not be
the problem. Laminated with one type of POE material
both type of cells pass DH2000 (SHJ) DH5000 (multi)
with power degradation rate <2%. This comparison shows
that a material may perform good with one type of cell but
in case of another cell technology results may be different.
4 WATER INTERACTION
4.1 Water Uptake
Samples for water uptake were made by laminating
two or three sheets from one type of material to achieve
roughly one mm thick samples. This allows comparable
samples for all materials that vary in delivered sheet
thickness (350µm – 750µm). The laminated samples were
dried and weighted, then exposed to 60°C warm water.
The water uptake was determined by the weight difference
before and after final drying (storage at 50°C for 24 hours)
after the finished water bath test.
In Fig. 5 below the results of the water bath test are shown.
After 63 hours in the water bath the EVA materials gained
about 0.2% in weight. Surprisingly the ionomer on a
similar level, while the POE materials gained with about
0.05% substantially less. The PVB showed the highest
water uptake in this test, >4%.
Fig. 5: Weight change during 60°C water bath exposure
(left) and relative weight gain due to water uptake (right).
This test shows that by the selection of encapsulation
material the amount of water that might diffuse into the
module can be influenced. How fast water actually
diffuses into a GG module is a different question since the
water ingress is limited to the module edge. This could be
shown in another test with water indicator , here the
ionomer showed the best water barrier characteristic while
in the water uptake test the material gained significant in
weight by water absorption.
4.2 Copper samples in Damp Heat
For this test 0.1 mm thin sheets of bare copper were
laminated between encapsulation material in GG and GB
samples. In four alternating DH500 UV 25 kWh/m² test
steps an image of the front and rear side of the samples was
Depending on encapsulation chemistry and increasing
water ingress chemical reactions with the copper occurred.
Clearest color change was detected for the ionomer
samples as shown in Fig. 6 below for the GB sample.
Fig. 6: Images of the GB copper sample in alternating
Damp Heat UV test.
For the other encapsulation materials water ingress is
typically clearly visible and also color change due to
chemical reaction e.g. the formation of copper oxides was
clearly visible. The ionomer shown above produced the
most colorful result.
5 HIGH TEMPERATURE PERFORMANCE
5.1 Hot Spot testing
Solar cells in a solar module that generate less current
compared to the system current turn from power producer
to consumer and dissipate power. The reason for lower
current might be shadowing (partial or full) or cell
degradation (cell cracking, PID or others). If the solar cell
current under reverse bias flows locally (e.g. due to shunts
or breakdown areas) high temperatures may occur in
laboratory hot spot testing or in the field [3,4] where
temperatures in the range of 180°C and above are reported.
Focus of this investigation is degradation of the
transmission in a GG samples with an area of 10 x 10 cm²
exposed to high temperatures. Gas diffusion (e.g. oxygen)
into the sample is limited to the edges by the glass-glass
configuration. This is to some extend comparable to a GB
solar module where the front glass and the solar cell act as
barrier for diffusion. The area in front of the solar cell is
accessed only via module backsheet and the edges of the
silicon wafer. Before starting the test, initial state was
recorded by images and transmission measurements.
Per encapsulation one of two samples was exposed to i.)
high temperature in the dark and the other to ii.) high
temperature with additional irradiation by an HRI lamp.
i) Dark test: 500 h @ 180°C + 1000 h @ 190°C
ii) Light test (0.6 suns): 500 h @ 180°C + 500 h @ 190°C
The test temperature in the second test step was
increased from 180°C to 190°C since degradation was
small after the first step. The spectral transmission was
measured over a wide spectral range. A change in the
transmission spectrum was found especially in the short
wavelength range. From this part of the spectrum the
yellowness index (YI) was determined. In Fig.7 below is
shown the change of the YI versus test time.
Fig. 7: Change of YI in the dark hot spot test (left) and in
the irradiated hot spot test (right). UVC materials (marked
in orange) showed more yellowing.
In the dark hot spot test the only thing to mention is the YI
increase for PVB.1. All other encapsulation materials are
stable even when exposed to 180 or 190°C for 1000 h. The
ionomer even increases in transmission what results in a
decrease of the YI.
With additional light (HRI lamp with ~0.6 suns) the
situation changes, the YI increased faster. Especially some
(not all) samples with UV absorber (UVC) degrade
(marked in orange in Fig.7). For these YI increased in the
range of 5 to 20. Again PVB.1 shows the strongest reaction
with YI>50 after 1000 h in the irradiated hot spot test.
The color change in the 1000 h shorter irradiated test is
more severe compared to the 1500 h dark test where
YI<20 was determined for PVB.1.
Below in Fig. 8 the images of an EVA sample in the
dark hot spot test are shown. On one side the glass of this
sample cracked and due to the increased oxygen ingress
the browning due to oxidation is happing in the crack area
and is not limited to the edge of the samples.
Fig 8: Images of dark hot spot test for EVA.2. In Fig. 9 the
upper left quarter indicated by the blue square is shown.
The final images at the end of the two tests are shown in
the Fig. 9 below. In the hot spot test with light for five of
six UVC materials the yellowing is clearly visible and also
the non-uniform distribution of the UV absorbers
(EVA.1&2 and PVB.1) is eye catching. The PVB material
shows in both tests a strong degradation starting from the
sample edge, potentially the material is in comparison
more open for oxygen diffusion.
Fig. 9: Final images of one quarter (2.5 cm x 2.5 cm) of
the samples at the end of the dark hot spot test (top) and
irradiated hot spot test (bottom). UVC materials are
marked in orange.
IEC 61215  includes only UV preconditioning with 15
kWh/m². This is to be seen in contrast to measurements in
 where already within one year in the desert Negev in
Israel about 120 kWh/m² are collected in the UV range.
This motivated additional long-term testing at 60°C with
100 W/m² UV radiation (alternating DH500 + UV 25
kWh/m²) and a total dose of 100 kWh/m² (+ DH2000).
Although the total duration is with 1000 h the same as in
the irradiated hot spot test this resulted in negligible
transmission change. As discussed in  UV aging
should be accelerated by temperature increase and if at all
only to a small extend by increase of the UV irradiation.
The results show that testing hot spot performance of
encapsulation materials comparable to field conditions
needs to be carried out at elevated temperatures and light
with sufficient UV component is regarded to be crucial.
5.2 Mechanical Glass adhesion
Temperature dependent delamination was tested by
recording peel curves at three to four different
temperatures for different materials in . Adhesion
decreased with increasing temperature. Since
delamination at elevated temperature occurring in the field
might lead to degradation or even safety issues a simplified
test for glass adhesion was developed for the encapsulation
test sequence. For this test glass adhesion GB samples
were laminated from glass, two layers of encapsulation
and a white backsheet. Half of the samples were aged in
alternating DH 2000 UV 100 kWh/m². Before the
adhesion test 1 cm wide stripes were cut. The samples
were fixed in a climatic chamber and each GB sample was
loaded with 1,2,3,4 and 5 kg weights in parallel. Then the
temperature in the climatic chamber was slowly ramped
from about 20°C to 120°C with a rate of approx.
~0.8K/min. For all loaded stripes the temperature of
failure meaning the beginning of delamination was
recorded. Images of the test chamber and a sample after
test is shown in Fig. 10 below. For the shown sample four
stripes are completely peel while the stripe with the lowest
force of about 1 N/mm is not completely peeled.
Fig.10: Photo of the glass adhesion test in the climatic
chamber (left) with different weights per samples. On the
right a photo of a sample after the peel test.
The result of the test is shown in Fig. 11, as expected the
temperature of failure decreased with increasing load. On
the left side the data of unaged samples is shown while the
data of the measurements after aging are shown on the
Fig. 11: Temperature of failure for different loads and ten
encapsulation materials for unaged samples (left) and
samples after aging in DH2000 UV 100 kWh/m² (right).
For unaged material all EVA and POE materials fail at
typical operating temperatures of 60°C and a load of 5
N/mm. For a reduced load of 1 N/mm the temperature of
failure increases to 70-80°C. The ionomer performed best
(80-90°C) for all loads while PVB showed the earliest
adhesion failure for all five loads (30°C).
Besides the ionomer that showed improved adhesion after
aging, all material failed earlier. Typical failure for a load
of 2 N/mm was found in a temperature range of 20-50 °C.
Even for a load of only 1N/mm the temperature of failure
is well below typical operating temperatures (50-60°C).
This result indicates that adhesion test at operating
temperature or above are important to qualify a material
for a hot climate (highest reported module operating
temperatures are above 90°C). Further testing is needed to
identify if the combined DH/UV aging is causing the
drastic degradation of the glass adhesion or if DH or UV
stress alone leads also to earlier failure.
6 PID testing
6.1 Ultra-harsh PID test
For PID the same set of samples as for long term DH
testing was laminated for the test (the layup is described in
chapter 3.1). Half of the GG and GB coupons with PID
sensitive multi solar cells were aged in DH500 followed
by a UV25 kWh/m² step. Per variation two samples were
tested in PID. The total of 56 PID samples were tested in
a DH chamber operated at 85°C and 85% RH. The front
glass of the coupons was covered by a grounded aluminum
plate and the cells were connected to -1500V. Compared
to standard PID testing these conditions are much more
severe. This makes the test highly accelerated because of
the high temperature of 85°C, complete grounding of the
glass surface and the humidity in the chamber. Last but not
least the test voltage of 1500V is the upper limited that
might occur in systems with high system voltages and
worst-case grounding/inverter layout.
The samples were exposed to PID stress with readouts
after a total of 12, 48, 168, 500 and 1000 h. Due to the
severe nature of the test it is expected to be selective
concerning the PID performance of the encapsulation. The
result of the test is shown in Fig.12 (a & b) below.
As expected a strong variability in the long term PID test
is found. Ageing (DH 500 UV 25 kWh/m²) leads to earlier
occurrence of PID, but not in all cases. Comparing the two
tested module layups GB and GG, the second tend to
degrade faster, potentially due to the glass on the rear side
of the module that acts as diffusion barrier for PID
supporting species. For example, EVA forms acetic acid
in DH that might accelerate PID.
Fig. 12 a: Power degradation in the PID test for unaged
GG (left) and GB (right) samples in 1000 h PID test.
Fig. 12 b: Power degradation in the PID test for aged
GG (left) and GB (right) samples in 1000 h PID test.
Looking the different encapsulation materials, samples
with PVB generally showed extremely fast degradation.
For EVA results are extremely mixed, samples made with
EVA.2 degraded fast while EVA.1 showed PID only after
prolonged testing. Only EVA.1 showed a tendency for
recovery in the long-term test. Samples with POE
encapsulation typically showed improved PID stability.
On one side POE.3 performed similar compared to EVA.1,
POE.1 shows good stability. Top runners are POE.2 and
the ionomer that had extraordinary PID stability and
passed in almost all cases the 1000h test with substantially
less than 5% power degradation. The result shows the
importance of PID testing before an encapsulation material
is selected for a product (GG or GB) that might be exposed
to high system voltages or other field conditions that lead
to potential induced degradation.
6.2 Volume Resistivity
Since PID is accompanied by a leakage current
flowing from the grounded glass to the solar cell one of
basic parameters that is obvious to test is the volume
resistivity of the encapsulation material. To measure
volume resistivity single layers of encapsulation were
laminated flat between two sheets of fluorinated ethylene-
propylene copolymer (FEP). Half of the samples were
aged in the FEP sandwich in a combined step of DH 500
and UV 25 kWh/m². The thickness of the samples was
measured, then the samples were placed between two
electrodes in a Faraday cage and 1500 V was applied from
one electrode. The leakage current was recorded for 20
min. For evaluation the median value in the range of 15-
20 min was used to calculate the volume resistivity. In Fig.
13 below the resistivity for 25°C and also 60°C for fresh
and aged materials are presented.
Fig. 13: Measured resistivity unaged/aged for 25°C/60°C.
Since generally volume resistivity measurements of high
resistance material (>1016 Ωcm) showed poor
reproducibility the result shown in the left graph is capped
at this resistivity level. At 25°C additional aging leads to
an increased or similar resistivity. Going to 60°C leads to
a drastic drop in resistivity (factor of 5 to 25). At 60°C
aging seems to lead to lower resistivities. Results for the
EVAs goes from >1015 Ωcm for 25°C to >1014 Ωcm for
60°C, while ionomer and POEs start with >1016 Ωcm and
end at 60°C typical in the range >1015 Ωcm or >1016 Ωcm,
where no difference is accessible by the this type of
measurement. Lowest resistivity was found for PVB that
goes from ~1012 Ωcm to ~1011 Ωcm for 25°C/60°C.
From the volume resistivity (aged measured at 60°C)
and material thickness the effective resistivity is estimated
for the covered cell area (243 cm²). This rough estimate
for each front encapsulation material is compared to
average PID results after 168h test (unaged / aged GG and
GB). The result is also shown in Fig. 14 below.
Fig. 14: Correlation of calculated effective resistivity
(aged 60°C) with degradation after 168h PID test.
A tendency is clearly visible. Higher resistivity leads to
improved PID. This tendency is only valid if the result for
one EVA is neglected. This material showed for EVA
typical volume resistivity but PID degradation was almost
100% after only 168 h of testing. One parameter that might
explain the run-away result is the comparatively low
laminated thickness of the sample. Around 350µm was
measured for this sample while all the other encapsulants
lay in the range 400-750µm.
This investigation shows that volume resistivity is a
suitable quality parameter to control PID, effects of
temperature and aging must be considered.
In this paper an excerpt of the results produced in the
first round of PI Berlin’s advanced test procedure for
encapsulation material is presented. Most relevant
findings in this benchmark test are the following:
- Standard cells are well protected by most current
encapsulation materials in long-term DH. Results vary
especially by module design (GG or GB). Other cell
concepts e.g. SHJ cells or cells made with non-standard
(e.g. copper plating) need a careful selection of a suitable
encapsulation material. The investigated materials vary
significantly in their water uptake, water ingress and also
in the interaction with copper.
- Realistic transmission stability testing of an
encapsulation in extreme temperature environments
(Hot Spot) is only possible if the samples are exposed to
high temperatures and additionally exposed by light with
significant UV contribution. Here some materials with
UV absorber showed significant yellowing.
- Temperature dependent glass adhesion testing revealed
delamination at rather low loads and typical module
operating temperature (or even below). Preaging lead to
even earlier adhesion failure. These results need to be
considered especially for hot (and humid) climates.
- When does PID occurrence depends on the type of
encapsulation material, module design and aging
history. Encapsulation volume resistivity is a suitable
parameter for material selection to prevent PID.
We would like to thank all companies participating in this
project, also for their patience. Also we thank Mr. Lackner
from the Solinex GmbH for providing the white backsheet
material used for different tests.
International Technology Roadmap for
Photovoltaics (ITRPV) – 2017 Results
Berghold et al.: “PID: From material properties
to outdoor performance & quality control
counter measures” SPIE Optics 2015
Wendlandt et al.: “Thermal stress analysis at
encapsulation and backsheet materials for PV
modules” EUPVSEC 2015
Janke et al.: “Comparison of hot spot endurance
tests: Temperature behavior of bare vs.
encapsulated crystalline silicon cells”,
Oreski et al.: “Comparative study of temperature
dependent delamination behavior of four solar
cell encapsulants to glass and backsheet-
laminate”, EU PVSEC 2011
IEC 61215 Terrestrial PV modules - Design
qualification and type approval Edition 3 (2016)
Pingel et al.: “Investigation of DH degradation
mechanism and correlation to an accelerated test
procedure (HAST) “, EUPVSEC 2012
Miyashita et al.: “Measuring Method of Moisture
Ingress into Photovoltaic Modules”, Japanese
Journal of Applied Physics 2012
Morales-Vilches A. et al. „ITO-free silicon
heterojunction solar cells with ZnO:Al/SiO2
front electrodes reaching a conversion efficiency
of 23 %”, WCPEC-7 (2018)
Köhl “The challenges of testing the UV-impact
on PV-modules”, PVMRW (2012)
Shioda et al.: “UV accelerated test based on
analysis of field exposed modules” EUPVSEC