Engineering, 2013, 5, 185-190
doi:10.4236/eng.2013.51b034 Published Online January 2013 (http://www.SciRP.org/journal/eng)
Testing Bifacial PV Cells in Symmetric and Asymmetric
Concentrating CPC Collectors
Gomes João1*, Henrik Davidsson2, Gruffman Christian3, Maston Stefan 3, Karlsson Björn4
1R&D Department,Solarus Sunpower, Älvkarleby, Sweden
2Energy and Building Design, Lund University, Lund, Sweden
3Finsun AB, Älvkarleby, Sweden
4Division of Energy Engineering, Mälardalen University, Västerås, Sweden
Received 2013
ABSTRACT
Bifacial PV cells have the capacity to produce solar electricity from both sides and, thus, amongst other advantages,
allow a significantly increase both in peak and annual power output while utilizing the same amount of silicone. Ac-
cording to the manufacturer, the bifacial cells are around 1.3 times more expensive than the single-sided cells. This
way, bifacial PV cells can effectively reduce the cost of solar power for certain applications. Today, the most common
application for these cells is in stationary vertical collectors which are exposed to sunlight from both sides, as the rela-
tive position of the sun changes throughout the day. Another possible application is to utilize these cells in concentrat-
ing collectors. Three test prototypes utilizing bifacial PV cells were built. The initial two prototypes were built for in-
door testing and differed only in geometry of the reflector, one being asymmetric and the other symmetric. Both proto-
types were evaluated in an indoor solar simulator. Both reflector designs yielded positive electrical performance results
and similar efficiencies from both sides of the cell were achieved. However, lower fill factor than expected was
achieved for both designs when compared to the single cell tests. The results are discussed and suggestions for further
testing are presented. A third prototype was built in order to perform outdoor evaluations. This prototype utilized a bi-
facial PV cells string laminated in silicone enclosed between 2 standard glass panes and a collector box with an asym-
metric CPC concentrator. The prototype peak electrical efficiency and temperature dependence were evaluated. A
comparison between the performance of the bottom and top sides of the asymmetric collector is also presented. Addi-
tionally, the incidence modifier angle (IAM) is also briefly discussed.
Keywords: Bifacial PV Cells; Symmetric and Asymmetric Concentrating Concentrators; Reflector Geometry;
Prototype; NOCT
1. Introduction
1.1. The Use of Bifacial PV Cells in Collectors
Photovoltaic (PV) cells are normally the most expensive
component in solar panels [1].
Bifacial PV cells have the capacity to produce solar
electricity from both sides. This way, these cells offer the
possibility of significantly increasing both the peak and
annual power output while utilizing the same amount of
silicone during the cell production. This can lead to a
lower material consumption per peak power produced
and in turn a lower kWh cost.
Recent developments have allowed significant ad-
vances both in the technology of bifacial PV cells and on
the production methods [2-4], thus allowing mass pro-
duction and substantially lowering their production cost.
According to the cell manufacturer, the price of one bifa-
cial cell is around 1.3 times more expensive than the sin-
gle- sided cells but, if each side receives the same annual
solar radiation as a single cell, the output will be dou-
bled.
Another important advantage is found during the col-
lector manufacturing process. Bifacial PV cells can allow
halving the number of cells that need to be soldered, thus
doubling the capacity of the cell soldering machines
which normally represent a big investment and a produc-
tion capacity bottleneck for many collector manufactur-
ers. For these reasons, bifacial PV cells can effectively
reduce the cost of solar power for certain applications.
However, the amount of radiation that each side of the
bifacial cell sees differs considerably from application to
application. Today, the most common application for
these cells is in stationary vertical collectors which are
exposed to sunlight from both sides, as the relative posi-
*Corresponding author.

G. JOÃO ET AL.
186
tion of the sun changes throughout the day. These types
of vertical installations take advantage of having a sim-
plified cleaning process since most of the dirt falls to the
ground by gravity. This can be a considerable advantage
in desert areas that are prone to desert storms and lack
access to water. However, in vertical applications, each
side of the bifacial cell is far from yield its maximum
potential, producing annually significantly less than a
standard solar cell at tilt inclination.
Another possible application for bifacial PV cells is to
utilize these cells in concentrating collectors where you
can extract more annual power from each side of the cell
while taking advantage of the cost reduction that concen-
tration can provide. Still, concentrating PV collectors
present several difficulties. PV cell efficiency can be
reduced both by non-uniform light and due to the in-
crease of temperature from concentration, although oper-
ating at slightly higher concentrations that 1C increases
slightly the cell efficiency. This way, reflector shapes
and normal operating cell temperature (NOCT) of the
collector must be carefully considered during the design
of these types of solar collectors.
Additionally, the annual output of a concentrating PV
collector is strongly affected by the IAM.
2. Method
2.1. Prototype Design
Three test prototypes utilizing bifacial PV cells were
built. All prototypes used the same string of bifacial PV
cells. This string consisted in 37 cells connected in series.
Each cell has a dimension of 26*148mm. Thus, the total
cell area of the string is 0.14m2.The string was enclosed
between two glasses panes with standard glass transpar-
ency (often 85%) for the initial prototype testing pre-
sented in this first paper. Future prototypes should be
built using anti transparency glass like the glass cover of
the box of third prototype. This glass cover is made of
low iron glass with solar transmittance 0.95 at normal
incidence angle according to the datasheets of the manu-
facturer.The reflectance of the aluminum reflectors used
for all prototypes is 94% for the visible light spectra
[5],according to the manufacturer datasheets.
The initial two prototypes for indoor measurements
The initial two prototypes were built for indoor testing
and differed only in the collector box where the bifacial
solar string was inserted. Both collector boxes had com-
pound parabolic collectors (CPC) reflector but while the
first was symmetric reflector the second was asymmetric.
Figure 1 shows the reflector shapes for both proto-
types.
Reflector area was equal for all prototypes. The sym-
metric reflector concentrates 1,53 suns in each cell side
while the asymmetric reflector receives 1 sun in the top
side and 2,06 suns in the bottom side. Table 1 describes
the areas and the concentration factors.
One of the reasons for testing both reflector shapes
was to verify that the total cell output and other relevant
parameters were identical. Even if the final concentration
factor is the same for both prototypes, the light distribu-
tion on the cells is very different. These two prototypes
were not inserted into a collector box and had no box
glass cover.
Asymmetric prototype for outdoor measurements
The third prototype was inserted into a more robust
collector box in order to be possible to perform outdoor
measurements. Figure 2 shows the third prototype. This
reflector design is patented by Solarus. A more detailed
analysis geometry presented is present in the paper by
Bernardo R. et al[6].
This prototype has bifacial PV cells laminated in sili-
cone and enclosed between 2 glass panes. The average
transparency of the silicone is 97% for the visible light,
according to the manufacturer specifications. After test-
ing the initial two prototypes, an asymmetric shape was
selected for the reflector of the third prototype.
A
B
Figure 1. Reflector shape and optical axis of both asymmet-
ric (B) and symmetric (A) prototypes.
Table 1. Areas and concentration factors of 2 bifacial con-
centrating collector design.
Prototype
Areas (m2)
Concentration Factor
Cell
A
B
Total
A
B
Total
Symmetric
0,14
0,22
0,22
0,44
1,53
1,53
3,06
Asymmetric
0,14
0,14
0,29
0,44
1,00
2,06
3,06
148mm
292mm
Figure 2. The geometry of the investigated bifacial PV pro-
totype.

G. JOÃO ET AL.
187
In order to monitor cell temperature, a K-type sensor
was installed inside the silicone in between the two glass
panes in a location very close to the cells but without
creating shading. Picture 1 shows the location of the
K-type sensor.
The idea with the construction of the third prototype
was to provide initial testing results before the construc-
tion of a full collector with the dimensions shown in
Figure 3. This collector will have 38 series connected
cut cells per string. The strings will be connected in par-
allel between them and also between receiver sides, as
shown in Figure 3.
2.2. Experimental Testing Setup
The three collector prototypes were measured both in-
doors and outdoors. The electricity was measured using
an IV tracer connected to a laptop. Data was retrieved
from the IV tracer using software designed by Christian
Gruffman at Finsun Inresol. This software is able to per-
form both single and continuous measurements. In the
indoor measurement only the single measurement func-
tion was used while for the outdoor measurements both
functions were utilized. The IV tracer draws an IV-curve
based on 100 values at different currents and voltages,
while the software retrieves this data and records the
values of Imp, Isc, Vmp, Voc, Pmax, FF and cell effi-
ciency. The electricity produced by the cells was not be-
ing continuously extracted during the measurements
which may lead to cell overheating and consequent pow-
er reduction due to cell temperature dependence.
Picture 1.Location of the K-type sensor and aspect of the
bifacial cell prototype.
Figure 3.Top view of the future bifacial collector. The red
lines show the electrical connections.
Indoor test set-up
The indoor solar simulator was designed to simulate
solar radiation and consists of two rows of 8halogen light
bulbs each with 1000W of power. As in many solar sim-
ulators, the light distribution is the drawback, since it is
far from perfect and it depends strongly on the position
of the cells within the simulator. (Picture 2)
Outdoor test set-up
The outdoor set-up consisted in a tilt adjustable wooden
stand where the collector was facing the sun. The select-
ed tilt was the best for maximizing output for the location
and the time of the year in which the measurements took
place, March, May, June and July of 2011. Location was
Älvkarleby, Sweden. Picture 3 shows the test set-up.
The solar radiation was measured utilizing 3 sensors type
SRS1000produced by the company Finsun Inresol. This
pyranometer uses a sensing element made in a single
crystal Si-cell. The output voltage is 100 mV when ex-
posed to 1000 W/m2 solar radiation [8]. The 3 pyranome-
ters mounted on the stand were calibrated against two
other pyranometers. One of the 3 pyranometers was cov-
ered by a diffuser ring in order to obtain the diffuse radi-
ation, as shown in Picture 3. Beam radiation is obtained
by subtracting diffuse to global radiation.
In Picture 3, only the top reflector trough is related to
the bifacial PV prototype. The bottom reflector trough
contains a collector prototype of a PVT for another ex-
periment and is not related to this paper.
Picture 2.The solar simulator for the indoor measurements.
Picture 3.The measured bifacial collector (top trough) and
the solar radiation sensors.

G. JOÃO ET AL.
188
Picture 4 shows the multimeter (Biltema Art. 15-275)
that was prepared to read the cell temperature by con-
necting to the K-type wire shown in Picture 1.
3. Results
The two initial prototypes were evaluated in an indoor
solar simulator.
3.1. Indoor Results
Cell Evaluation
The bifacial cells evaluated had a dimension of
156mm*156mm which is also the standard market size
for single side cells. These cells were tested and the re-
sults are presented in Table 2.
Since, concentrating the sunlight can cause current
capability constrains that limit the collector output, ac-
tions were taken to reduce the current on the collector.
This way, the bifacial cells were cut using a laser cell
cutting machine that divided the cells into 6. Additionally,
the cells were also cut 4mm on each side to match the
size of Solarus PVT collector cells and patented design.
Thus, the cut cells have the dimensions of
148mm*26mm.In both cases, only one side of the bifa-
cial cells was measured.
Reflector design evaluation
37 cut cells (dimension 26*148mm) were soldered to-
getherin series to form a string with an area of0,142m2.
This cell string was then inserted and tested under two
different reflector configurations, one symmetric and one
asymmetric. Table 3 describes the measuring results
obtained.
Another obtained result was that the top part produced
44% of the total power, on the asymmetric design.
Picture 4.Multimeter measuring a cell temperature of 64ºC
in a sunny day with 28ºC of outside temperature.
Table 2. Data for the 2 measured cells with different areas.
Type
Voc
Isc
Pmax
Vmp
Imp
FF
Area
Power In
Eff.
A
0,62
1,12
0,57
0,54
1,06
81,5
0,004
1050
14,13
B
0,63
6,15
3,04
0,53
5,71
78,0
0,024
900
13,86
Units
V
A
W
V
A
%
m2
W/m2
%
Notes: Type A – Cut cell; Type B – Standard Bifacial cell; Power In refers
to the radiation that is hitting the cells; Eff. refers to efficiency.
Table 3. Results of 2 bifacial concentrating collector de-
signs.
Asymmetric
Side
Voc
Isc
Pmax
Vmp
Imp
FF
P. In
Area
Eff.
Both
22,9
4,1
60,3
17,6
3,4
64,2
1200
0,44
11,5
Bottom
22,3
2,4
33,9
17,7
1,9
63,0
1100
0,29
10,5
Top
22,5
1,9
27,0
18,2
1,6
68,0
1600
0,14
11,9
Symmetric
Side
Voc
Isc
Pmax
Vmp
Imp
FF
P. In
Area
Eff.
Left
22,3
1,8
25,5
18,5
1,4
63,6
1100
0,22
10,7
Right
22,4
1,7
24,9
18,5
1,3
65,1
1100
0,22
10,4
Both
22,9
3,4
49,5
18,8
2,6
63,6
1100
0,44
10,3
Units
V
A
W
V
A
%
W/m2
m2
%
3.2. Outdoor Results
The collector was measured outside in 2 different occa-
sions.
Preliminary Testing - March 2012:
For a standard flat plate collector, the best tilt corre-
sponds to 90º minus the projected solar altitude for the
specific location and time of the year. As detailed in the
paper from Bernardo R. et al [6,7], for this specific
asymmetric geometry the best tilt suffers a deviation
from the best tilt of a standard flat plate by 5º to 10º in
the direction. The best entrance angle for the radiation is
also shown by the radiation arrows in Figure 2.
The collector was measured outside for about one hour
between 11h20 and 12h20. This period had extremely
stable weather conditions during which regular radiation
measurements were recorded manually.
At this time, no equipment was available to monitor
the changes in cell temperature and the only available
information was that the initial cell temperature was
equal to outdoor temperature of 0ºC. Assuming standard
temperature dependence from literature of 0,5%/ºK [9]
and assuming that Pmax has to remain stable when ad-
justed by the above coefficient to the reference tempera-
ture of 25ºC, a temperature curve was linearly estimated.
Figure 4 shows the measured values of Pmax and effi-
ciency over time as well as an estimation of cell temper-
ature, Pmax at 25ºC and efficiency at 25ºC.
Table 4 presents a comparison between the bottom
and both sides of the prototype. All data is measured with
the exception of the temperature that is estimated. The 2
measurement (bottom and both) were done within 7 sec-
onds. Table 5 presents the same comparison for the es-
timated values at 25ºC. Bottom side represents 55% of
the total power of the string.
June 2012 Testing
New measurements on the bifacial prototype were
done when Solarus solar laboratory gained the capacity
to log the radiation and measure cell temperature. Figure

G. JOÃO ET AL.
189
5 plots the global radiation in Älvkarleby on the 22nd of
June. The tilt was the most adequate for maximizing the
output for this collector at this latitude and time of year.
Vmp, Isc, Pmax and cell temperature are shown in the
Figure 6.
The cell temperature was taken at 6 different times and
then extrapolated using a 3rd order polynomial equation
matching the measured data with an R2 of 0,999. Highest
measured cell temperature was 88ºC at 14h30 after con-
stant perfect radiation day with an ambient temperature
of 32ºC.
Figure 7 shows the measured electrical efficiency de-
pendence on cell temperature expressed both in total area
and cell area. The dependence of electrical efficiency on
temperature is -0.51 %/K.
Figure 4. Pmax and efficiency over time (measured at a
temperature represented by T curve); Pmax and efficiency
at 25ºC (estimated)and cell temperature.
Table 4. Comparison bottom vs both sides at 36ºC.
Side
Voc
Isc
Vmp
Imp
FF
Bottom
18,1 V
1,7 A
13,3 V
1,4 A
57,8%
Both
18,9 V
2,9 A
13,9 V
2,3 A
58,5%
Side
Cell T.
Pmax
Eff/Total area
Eff/Cell area
Bottom
36ºC
18,0W
7,2%
14,9%
Both
36ºC
32,5W
8,8%
13,4%
Table 5. Comparison bottom vs both sides at 25ºC.
Side
Cell T.
Pmax
Eff/Total area
Eff/Cell area
Bottom
25ºC
19,0W
7,6%
15,6%
Both
25ºC
34,3W
9,2%
14,1%
Figure 5. Global radiation on 22th of June of 2012.
4. Conclusions
4.1. Indoor Testing
Cell Evaluation
Two sizes of cells were evaluated. As expected, a
smaller cell size presented a higher FF even when ex-
posed to a high radiation level and thus a slightly higher
efficiency.
The value of 14% efficiency from the manufacturer
was confirmed. The cell manufacturer stated that these
efficiencies will be raised in the near future to signifi-
cantly higher levels.
Reflector design evaluation
Both geometries yielded similar results in terms of FF,
efficiency and Voc. The difference in proximity to the
light source due to the asymmetric reflector led to dif-
ferent light intensities which in turn lead to different Isc,
Imp and Pmax. Nevertheless, these values remain con-
sistent. However, a significant decrease in FF and cell
efficiency was observed for both geometries when com-
pared to the single side cell testing. While the single cell
shows a FF of 80% and an efficiency of 14%, the string
measurements for both geometries shows a FF of 65%
and around 11% of efficiency. This decrease is due to the
losses from the reflector reflectance and possibly a cur-
rent capability constrains may also occur due to the con-
centration of light. Additionally, at least a part of the
decrease can also be attributed to non-uniform cell dis-
tribution from the solar simulator and thus further out-
door testing is required.
Figure 6. Pmax, Imp, Vmp and cell temperature for the day of
22th of June. The (R) and (L) symbolizes right or left axis.
Figure 7. Electrical efficiency dependence on cell tempera-
ture - measured

G. JOÃO ET AL.
190
The Solarus PVT reflector geometry was selected for
outdoor testing both for practical reasons and since there
were no relevant constrains verified during this step.
Current capability seems to be equal for both designs
since FF is similar. Thus another conclusion is that the
side of the solar cell from where the current is flowing
into the busbar does not impact the current capability.
In the asymmetric reflector geometry, the top part was
producing 44% of the total power. Since top part repre-
sents 33% of the area, it should produce also 33% of the
power, if there were no losses due to reflector reflection.
This decrease is again due to the reflector losses and
non-uniform light distribution. Additionally, it is im-
portant to notice that the solar simulator gives more than
1000w/m2 and that its radiation spectra is not equal to the
solar radiation which further motivates performing out-
door testing.
4.2. Outdoor Testing
Bottom side represents 55% of the total power of the
string which is consistent with the indoor testing. This is
below the theoretical maximum and can be improved.
The obtained FF values seem to imply that there is a cur-
rent capability constrain. 58% outdoor and 64% indoors
are low values in comparison to standard PV modules,
according to literature [9]. At 36ºC, Imp /Isc is 0,83 while
Vmp/Voc is 0,72 signaling that shunt resistance is more
significant than series resistance. As expected, the bot-
tom side presents a higher efficiency per cell area due to
the concentration factor of 2 while a lower efficiency per
total area due to the extra losses with reflector reflec-
tance.
The dependence of electrical efficiency on temperature
is -0.51 %/K, in good agreement with the common value
for solar cells described in literature [9].
At 25ºC and 1000w/m2, the electrical efficiency ex-
pressed per total area is 8,4%. If expressed per cell area,
this value is12,9%.
The IAM is an important parameter for any non-track-
g concentrating collector and for the asymmetric bifacial
PV as well. Figure 6 shows that for bifacial PV, an out-
put increase starts at 11h30 under steady weather condi-
tions. Comparing to the PVT measured Bernardo R. et al
[6],the tested bifacial PV has a bigger time period unaf-
fected by shading since the string is one cell smaller and
closer to the center, the prototype. The afore mentioned
figure also shows how the voltage drops in the bifacial
PV prototype with the temperature increase.
Since the crystalline solar cells are translucid, it was
expected that the solar radiation would go through the
cells and that high operating cell temperature would not
be reached. However, this did not happen. It is recom-
mended that a new prototype is built where: 1) Only anti
reflectance glasses are used. 2) Placing the absorber in
contact with the top glass should have a big increase im-
pact in NOCT of the collector by greatly increasing the
heat losses.
REFERENCES
[1] International Renewable Energy Agency, “Renewable
Energy Technologies: Cost Analysis Series - Solar Pho-
tovoltaic”, Volume 1: Power Sector Issue 4/5, June 2012.
[2] A. Kränzl, R. Kopecek, K. Peter, P. Fath “Bifacial Solar
Cells on Multi-Crystalline Silicon with Boron BSF and
Open Rear Contact”. Proceedings of the 4th World Con-
ference on Photovoltaic Energy Conversion. May 2006
Vol. 1, pp 968 – 971.
[3] V. D. Mihailetchi et al. “Screen Printed N-Type Silicon
Solar Cells for Industrial Applications”. 25th European
Photovoltaic Solar Energy Conference and Exhibition,
6-10 September 2010, Valencia, Spain
DOI: 10.4229/25thEUPVSEC2010-2DO.2.2
[4] C. Duran et al. “Bifacial Solar Cells with Selective
B-BSF by Laser Doping”. 24th European Photovoltaic
Solar Energy Conference, 21-25 September 2009, Ham-
burg, Germany.
DOI: 10.4229/24thEUPVSEC2009-2CV.5.19
[5] Alanod, reflector manufacturer, 2012. http://alanod.com.
[6] Bernardo R. et al., “Measurements of the Electrical Incidence
Angle Modifiers of an Asymmetrical Photovoltaic/
Thermal Compound Parabolic Concentrating-Collector,”
Power and Energy Engineering Conferenceand Engineer-
ing, Sanya, China, Dec 31, 2012 - Jan 2, 2013, un-
published
[7] Bernardo, L.R et al. Performance evaluation of low
concentrating photovoltaic/thermal systems: A case study
from Sweden. Solar Energy2011, 85, 1499-1510
[8] N. Gentile et al., "Construction of a small scale laboratory
for solar collectors and solar cells in a developing coun-
try", Power and Energy Engineering Conferenceand
Engineering, Sanya, China, Dec 31, 2012 - Jan 2, 2013,
unpublished
[9] S. Wenham et al, "Applied Photovoltaics", 2nd edition,
Earthscan, London, 2007, ISBN 978-1844074013.
Nomenclature
CPC
Compound Parabolic Collector
(-)
IAM
Incidence Angle Modifier
(-)
NOCT
Normal Operating Cell Temperature
(ºC)
IV Curve
Current and Voltage Curve
(-)
Imp
Maximum Power Current
(A)
Isc
Short Circuit Current
(A)
Vmp
Maximum Power Voltage
(V)
Voc
Open Circuit Voltage
(V)
Pmax
Maximum Power
(W)
FF
Fill Factor
(%)