Solar Energy Materials & Solar Cells 95 (2011) 538545

Contents lists available at ScienceDirect

Solar Energy Materials & Solar Cells

journal homepage: www.elsevier.com/locate/solmat

Water immersion cooling of PV cells in a high concentration system

Li Zhu a,n, Robert F Boehm b, Yiping Wang b,c, Christopher Halford b, Yong Sun c
a
b
c

School of Architecture, Tianjin University, Tianjin 300072, China

Center for Energy Research, University of Nevada, Las Vegas, NV 89154-4027, USA
School of Chemical Engineering and Technology, Tianjin University, Tianjin 300072, China

a r t i c l e in f o

a b s t r a c t

Article history:
Received 23 December 2009
Received in revised form
2 July 2010
Accepted 24 August 2010
Available online 14 October 2010

Temperature control of solar cells at high concentrations is a key issue. Short-term efciency drop and
long-term degradation should be avoided by effective cooling methods. Liquid immersion cooling
eliminates the contact thermal resistance of back cooling and should improve cell performance. A 250X
dish concentrator with two-axis tracking was utilized to evaluate a new CPV system using de-ionized
water for immersion cooling. Time-dependent temperature distributions of the PV module of high
power back point-contact cells were measured, as well as the IV curves. The cooling capacities of the
liquid immersion approach are very favorable. The module temperature can be cooled to 45 1C at a
940 W/m2 direct normal irradiance, 17 1C ambient temperature and 30 1C water inlet temperature. The
temperature distribution of the module is quite uniform, but the electrical performance of the cell
module degrades after a fairly long time immersion in the de-ionized water.
& 2010 Elsevier B.V. All rights reserved.

Keywords:
Liquid immersion cooling
Concentrator photovoltaic
Dish system
De-ionized water
BPC solar cell

1. Introduction
Cooling of the cells is a critical issue when designing
concentrating photovoltaic (CPV) systems. In these systems, solar
cells under high illuminations will have increases in temperature
with the attendant cell efciency drops, so a reliable heat
dissipation system is needed to cool the cells effectively. For
dishes or heliostat elds that have large point-focus areas, the
receiver generally consists of a multitude of densely packed cells.
Densely packed modules present great problems for cooling
compared to the single cells and linear geometry type CPV
systems. Under high concentrations ( 4150 suns) the densely
packed modules require an active cooling system with a low
thermal resistance. A comprehensive review has been published
by Royne et al. in 2005 [1], which summarized the cooling
technologies for concentrator photovoltaic systems including
those for the densely packed modules. Also, in the review it is
said that only impinging jets and micro-channels have been
reported to achieve low enough thermal resistances for satisfactory cooling performance. Soon after the review Royne et al. [2]
published another paper on impingment cooling and led one
patent on an active cooling approach using this method [3].
Some researchers put forward the idea of submerging bare
cells directly inside a dielectric liquid. Originally the efforts on
liquid immersion concentrated on the possible effects of the
liquid on the cells, whether in the optical or electrical aspects.

Corresponding author. Tel./fax: +86 2227404771.

E-mail address: [email protected] (L. Zhu).

0927-0248/$ - see front matter & 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.solmat.2010.08.037

Besides the patents and papers reviewed in [4], recently more and
more patents in this area have been led. Tanaka [5,6] suggested
using a shallow liquid layer or a gel layer [7] surrounding solar
cells for light trapping and also cell surface wetting. Carcangiu
et al. [8] patented a liquid immersion photovoltaic panel. The
panel uses a liquid-tight chamber to house solar cells immersed in
a poly-dimethylsilicone liquid, which circulates through the inlet
and outlet passages. Ignacio et al. [9] used curved optically
transparent covers to enhance the concentrating effect of the
immersion dielectric liquid. Falbel [10] patented a surrounding
reective surface for a solar cell, which reects back the light rays
which are not absorbed by the solar cell. Besides the liquid and
gel, Cherney et al. [11] even extended the refractive medium to a
solid. Using water as the immersion liquid in [12], panels are
congured with liquid super-concentrators having outwardly
disposed liquid imaging lenses.
Aside from the optical and surface wetting advantages of liquid
immersion, the direct contact between cells and their surrounding
liquid make it possible for a new cooling method to be used for
cells, especially for cells at high concentrations. Eliminating the
thermal resistance of the contact wall between solar cell and uid
of conventional active cooling approaches, which controls the rate
of heat dissipation, the cells could be effectively cooled down for a
desirable sunlight-to-electricity conversion efciency. Wang [13]
has put forward this idea in his patent. Phase change liquids have
been patented by Koehler [14] and Vasilinina et al. [15]. Tange
[16] applied a patent on the sealing structure for direct water
cooling of solar cells. Detailed laboratory experimental as well as
simulation work has been performed by researchers in China, and
considerations have been given to every aspect of this new

L. Zhu et al. / Solar Energy Materials & Solar Cells 95 (2011) 538545

cooling method, from cooling agent selection, cell performance

variation, cooling capacity and temperature distributions, heat
transfer enhancement methods, etc. [4,1719]. Experimental
results indicate that cell temperature under high concentrations
( 4200 suns) can be cooled to lower than 45 1C and the convective
heat transfer coefcient can be higher than 3000 W/m2 1C [4,17].
Also the dimethyl silicon oil is quite stable under high intensity
UV light and high working temperature compared to other
candidate immersion liquids [18,19].
Before claiming the liquid immersion cooling method the most
favorable choice for concentrator photovoltaic systems, its
effectiveness is to be further veried in the present paper using
a dish concentrator system at the University of Nevada, Las Vegas
(UNLV). An innovative receiver based on a liquid immersion
method was designed and built using Amonix Back Point-Contact
(BPC) silicon cells. Because of its good uid and thermal properties, de-ionized water was chosen as cooling agent. Under a 250
suns concentration and a two-axis tracking mode, the real-time
temperature distribution of the cell module was measured as
were the IV curves.

2. Experimental
The liquid immersion CPV dish system mainly consists of a
two-axis tracking dish concentrator, a liquid-immersed receiver
and other auxiliary parts. In the following sections detailed
descriptions are given for each main component.
2.1. Dish concentrator
The dish installed at UNLV originated from a National Renewable Energy Laboratory (NREL) project, which aimed at building
the rst dish concentrator photovoltaic system in the United
States. The project started in 2003 and ended in 2007 with
valuable outcomes for similar systems. But the key issue
throughout the whole project is the heat dissipation for the
monolithic cells at high concentrations. The dish was designed
and built by the Science Applications International Corporation
(SAIC), which was improved to incorporate the xed focus mirror
facets, DC variable-speed drive motors, a fail-safe fast off-track

539

system and other features. Details of the design, test and

installation can be found in reference [20]. As shown in Fig. 1a,
the concentrator has 16 mirrored facets, each facet having 76
mirror tiles, for a total of 1,216 mirror tiles and 113 square meters
of mirror area for the system. The facets are outtted with small
(30 cm by 30 cm), at mirror tiles that are precisely aimed during
manufacture to produce a uniform ux prole at the receiver. The
mirrors of the dish have a reectance of approximately 90% when
clean, and the highest expected direct normal irradiance is about
1100 W/m2. For the receiver size of 60 cm by 60 cm, the solar ux
distribution is given in Fig. 1b and the whole focus area has a
concentration ratio of approximately 250 suns. The dish consists
of a steel truss structure that tracks the sun with a high precision,
a tubular steel pedestal supports the structure above the ground,
and a mechanical gearbox supplies the drive torque to the
structure.
2.2. Liquid immersion receiver
2.2.1. Cell modules
The PV module made for the SAIC project was modied for use
here. The original module was made from Amonix BPC silicon
cells, as shown in Fig. 2. The cells are mono-crystalline silicon
with back contacts, with an active area of 1 cm by 1 cm, designed
for 250X nominal solar concentration and having an efciency of
about 24% at 25 1C. The temperature coefcients of these cells are
about 0.25% per 1C. The cells were mounted on copper-clad
electrically insulating substrates to form modules. The front
copper surface was etched to form a circuit pattern that provided
the series connections to the cells when they were soldered onto
the substrate. Each module has 88 cells connected in series,
and the rated output power is 543 W at 250 suns with a Voc of
72.2 VDC. The Vmax and Imax are, respectively, 55.44 VDC and
9.8 A, giving the conversion efciency of 16.4% under rated
conditions using the module aperture area for the efciency
denition. Due to the internal diameter of the glass pipe container
(described below), the side edges for bolts mounting for the old
module were machined off. Several small holes were drilled at the
back surface for installing the thermocouples, and the rigid
electrical connection wires were changed to soft wires for easily
inserted into the liquid container.

Fig. 1. Dish concentrator picture and the solar ux prole at the focus point. (a) Mirror facets and (b) solar ux prole.

540

L. Zhu et al. / Solar Energy Materials & Solar Cells 95 (2011) 538545

2.2.2. Receiver assembly

As shown in Fig. 3, the module was inserted into a beaded glass
pipe (8828-411, ACE Glass Incorporated), 50.8 mm internal
diameter and 457.2 mm long. The electrode leads and thermocouple wires were xed before inserting the receiver into the
pipe, and all the connection wires went through the threaded
nipple and into the stainless steel pipe, where a leak-free
connection was used for pulling the wires out to the monitoring
equipment. A stainless rod was fastened to the back of the PV
module, which is used to x the module position during
installation. The two ends of the pipe were connected to stainless
steel tees using victaulic coupling assemblies (8856-11, ACE Glass
Incorporated). A 8870 male threaded adaptor of 304 stainless
steel (ACE Glass Incorporated) was used to regulate the diameter
to 25.4 mm.

2.3. Cooling system

The circulation loop for cooling the PV module by DI water is
shown in Fig. 4a. Heat is rejected from the DI water to tap water
through a shell and tube heat exchanger (SSF-603-ER-1P, Young
Touchstone Company). The DI water is in the tubes and the

cooling water is inside the shell. The two uids run in a parallel
mode so the warmest DI water at the receiver outlet could
exchange heat with the coldest tap water. The shell has a
diameter of 155.4 mm and the outside diameter of the tube
is 9.5 mm. The pump for circulating the DI water is the model
TE-4-MD-HC 582604 (Little Giant Pump Company) and is a totally
enclosed and thermally protected type. The tap water is ltered
before running into the shell to avoid fouling.
DI water is a good option to use as an immersion medium
because of low viscosity and high thermal capacity, provided that
the cell performance remains stable during water immersion.
Defects will be caused on the cell from varying surface
recombination or polarization when the ion concentration is
high. The DI water was produced by a water purier (OPTION-S
15BP, ELGA LabWater Global Operations) with an outlet resistivity
of 15 MO cm. The resistivity of DI water was controlled using an
ion-exchange column to remove the ions accumulated during
circulation. The resistivity of the DI water is kept at a desired
setpoint during operation by opening a water treatment piping
branch when the resistivity becomes too small. This is done by
owing a portion of the water through an ion-exchange element.
This process continues until the measured resistivity is higher
than the setpoint.
Also, it is very important to protect the cooling equipment and
connections from being exposed directly to the intensied
illumination. Rigid ceramic insulation with binder and metal
shields were installed for this purpose as can be seen in Figs. 3
and 4b.

3. Measurement and control

3.1. Measurements

Fig. 2. Cell module used for the receiver assembly.

Measurement locations were indicated in Fig. 4a, in which the

circles with symbols give the information of the parameters
measured. Four quick disconnect thermocouples (CHROMEGAALOMEGA 304 SS Sheath, Omega Company) were inserted inside
the tubes and xed with compression ttings (OMEGLOK, Omega
Company). The mass ow rate at the inlet of the cooling water
loop and the immersion liquid loop is measured by an electronic
digital meter (A109GMN100NA1, accuracy 71.5%, repeatability
70.2%, Great Plains Industries, Inc.). The resistivity of the DI
water inside the loop was measured by an electrochemical
analyzer (Model 873RS-AIPFGZ, Invensys Process Systems Inc.)

PLUG
TEE

CELL MODULE
GLASS PIPE
VICTAULIC COUPLING
ADAPTER
BUSHING
STAINLESS STEEL PIPE

Fig. 3. Liquid immersion CPV receiver assembly.

L. Zhu et al. / Solar Energy Materials & Solar Cells 95 (2011) 538545

V2

541

F1
T1

F2 V4 Filter

T2

Cooling Water

Cooling Water
T4

Heat Exchanger

T3
R

metal shield

datalogger
V7

V9

V8

ion-exchange
column

Drain
Ion-exchange Column
switchbox
V6

HEX

V5

receiver

V1
t1

I-V curve tracer

tx
V3

I-V

Pump
D.I. Water

ceramic insulation shield

Fig. 4. (a) Liquid immersion cooling system layout and (b) photo.

3.2. Control methods

Fig. 5. The locations of the thermocouples are shown.

with a contacting resistivity sensor (871CC-A2). The tested

resistivity is temperature-corrected and the accuracy of the
analyzer is 70.5% of FSC range at 25 1C.
The temperature distribution on the cell module receiver is
considered to be the most important measurement for the present
test. Aside from the desired low working temperature for the
module, a uniform temperature distribution among the cells is
very desirable. Since all the cells are connected in series, the
smallest cell current will degrade the whole modules performance. The locations where temperatures were measured are
shown in Fig. 5. It can be seen that both axial and lateral
temperature distributions were measured. Thermocouples
(5TC-TT-K-36-72 or 5TC-TT-K-36-36, Omega Company) were
used. These were glued into the pre-drilled holes of the copper
substrate at the back using thermally conductive epoxy adhesive
(OB-101-16, Omega Company) with a thermal conductivity of
0.32 W/(m K).
The IV curve of the cell module was monitored by an IV
Curve Tracer (DC-100C, Daystar, Inc.). The accuracy for the voltage
measurements was 70.5% of reading (resolution 18 mV) and for
the current was 70.5% of reading ( 745 mA with resolution of
0.3 mA). To minimize cable runs, the IV curve tracer was xed
onto the rack as shown in Fig. 4b. Also, a wireless connector was
used to transfer the signals from the RS-232 Serial Port to the USB
port on a remote computer.

A solenoid valve was installed at the inlet of the cooling water at

the test section to turn on/off the water in response to the cell module
temperature. If the temperature at any point of the module rises
above the maximum desired value, the valve is kept on. The highest
temperature tolerable in the present test was set at 80 1C, and if any
of the measured temperatures exceeds that point, the alarm will
sound and the concentrator alignment will be adjusted. Another
solenoid valve was located in line with a needle valve in the
ion-exchange column loop, and the valve is turned on/off according to
the resistivity of the DI water. The needle valve controls the side ow
rate, and the solenoid valve controls the on/off status of the branch.
The control wires of the two solenoid valves were connected to
the data-logger and also to a switch box. The two valves can be
switched by remote control or manually on-site. All the temperature,
ow rate and resistivity signals coming from the sensors are recorded
using a Relay Multiplexer (CS-AM 16/32, Campbell Scientic, Inc.)
with an interval of 15 s. Mechanical relays in the CS-AM 16/32
connect each of the sensor channels in turn to a common output that
is sent to the data-logger. The data-logger is a CS-CR10X measurement and control module that is connected to the CS-AM 16/32
multiplexer and a CS-SDM-SW8A pulse counter. A control panel
enclosure (RJ1816HPL, Stahlin Non-Metallic Enclosures) was used to
house all the measurement and control equipment. All the data
recorded is saved in a le created in the computer. Moreover, a
data collection and analysis program was edited using Campbell
Scientics PC 208 data-logger support software. Fig. 6 shows the
interface of the data collection and status control.

4. Results and discussion

4.1. Test conditions
Fig. 7 shows the system on sun. To achieve a nearly-constant
DNI, testing was performed + /  1 h around solar noon. Fig. 8

542

L. Zhu et al. / Solar Energy Materials & Solar Cells 95 (2011) 538545

Fig. 6. Interface shown in the computer for data collection and status control.

Fig. 8. Climate conditions of the test days.

Fig. 7. The test system shown on sun.

gives the climate conditions on the test days, which comes from
the site meteorological station and was downloaded from http://
www.nrel.gov/midc/unlv/. As shown, the test days are mostly
sunny and clear, and the DNI during PST 10:30 am and 12:30 pm
is nearly constant and the value is above 900 W/m2. The ambient
temperature and wind speed also does not change signicantly
during test and averages 17.5 1C and 2.1 m/s, respectively.

Under these test conditions, temperature distributions of the

cell module were recorded as was the time-dependent electrical
performance. Also, the stability of electrical performance with
resistivity of DI water was examined. Fig. 9 shows the change of
the DI water resistivity during the test. Firstly the ion-removal
branch was shut off to let the resistivity drop until above
0.5 MO cm, and this process was repeated the next day. Afterwards, the resistivity was kept constant at 5.0 MO cm.

L. Zhu et al. / Solar Energy Materials & Solar Cells 95 (2011) 538545

4.2. Temperature distributions

The main purpose of the present test is to determine whether
or not the liquid immersion method is capable of cooling the cell
module down to a desirable temperature and whether a uniform
temperature distribution of the whole module could be achieved.
Consider Fig. 10, which gives the temperature of the cell module
and its distributions along the ow direction and the radial
direction. The results are quite favorable. With an inlet temperature of about 30 1C for the DI water, the maximum module
temperature is below 50 1C and the maximum difference among
all the measuring points is less than 4 1C except the bottom point,
which receives some shadow from the shield. For the axial
distribution given in Fig. 10a, it is shown that the temperature
increases with the increasing DI water temperature in the ow
direction. While for the lateral distribution, it is assumed that the
difference was caused by the tilted position of the cell module.
This was caused by complications in the installation procedure for
the current conguration that will be addressed in future designs.
For the clearness of the cooling capacity of the liquid
immersion method, Table 1 gives the temperature values of the
cell module under various testing conditions. The symbols Ts,i, Ts,o

Fig. 9. Resistivity variation curve during test.

543

Tr,i and Tr,o used in Table 1 are, respectively, the inlet and outlet
temperatures for the cooling water and DI water. Ta is the ambient
temperature and v is the wind velocity. Vs and Vr are the ow
rates for the cooling source and receiver, respectively. The
uniformity was characterized by the standard deviation of the
temperature values, with All representing all the measured data
and X(T5) representing T5 excluded due to the shadow error.
For the direct cooling method, it is seen that with stronger DNI the
cell module temperature increases but not signicantly. At a DNI
of 985 W/m2 the averaged module temperature is only 46.8 1C
with a uniformity of less than 1.0, clearly showing the favorable
cooling performance of the proposed method.

4.3. Cooling capacity of liquid immersion method

The test was conducted at a constant ow rate 2.22 m3/h for
the DI water and 1.05 m3/h for the cooling water. The inlet temperature of the cooling water varies with the tap water
temperature, which is mostly stable around 20 1C, and the inlet
temperature of the circulating DI water is stabilized at 30 1C once
the system reaches the steady status. As given in Fig. 11a, the
temperature difference between inlet and outlet of DI water is
2.4 1C and for the cooling water this value is 5.9 1C for its small
ow rate. Accordingly, Fig. 11b gives the calculated heat removal
capacity q of the DI water and cooling water, also with the
calculated overall convective heat transfer coefcient using the
experimental data. The equation hq/(Tm,avg-Tw,avg)/Ac was used
for calculation, where the Tm,avg and Tw,avg represent the average
temperature of the module and the average of the DI water inlet
and outlet temperature, and Ac is the sum of all the surface areas
contacting with cooling medium. As shown, the heat removal rate
from the cell module reaches around 6 kW and the convective
heat transfer coefcient can be higher than 6000 W/m2 K.
Comparatively, for the conventional wall-contact active cooling method the convective heat transfer coefcient of the owing
water inside the back tube could reach as high as 6000 W/m2 K.
However, the overall heat transfer coefcient will be greatly
reduced by the thermal contact resistance of the gluing layer. For
example, a 1 mm adhesive layer with a thermal conductivity of
3.5 W/m K will cause about 70 1C temperature difference between
the cell back surface and the metal substrate under 250 suns. For
the liquid immersion method, the measured cell module temperature directly represents the metal ngers temperature and
the cell top surfaces temperature. Obviously, the direct cooling

Fig. 10. Temperature distributions of the cell module.(a) Axial direction and (b) lateral direction.

544

L. Zhu et al. / Solar Energy Materials & Solar Cells 95 (2011) 538545

Table 1
Temperature distribution of the cell module under various testing conditions.
Type

DNI

Ta

Vs

Ts,i

Ts,o

Vr

Tr,o

Tr,i

(W/m2)

(oC)

(m/s)

(m3/h)

(1C)

(1C)

(m3/h)

(1C)

(1C)

Testing conditions

801
841
904
959
985

14.62
14.49
15.50
16.40
14.38

1.05
1.06
1.04
1.02
1.02

18.7
19.6
18.7
18.6
18.9

23.9
25.4
24.2
24.9
25.1

2.23
2.23
2.21
2.21
2.23

29.9
32.3
30.5
32.3
32.0

27.8
29.9
28.1
29.8
29.7

Temperature distribution

Lateral direction
T3
42.8
44.0
46.4
47.2
48.3

1.73
1.32
0.40
0.90
2.85

Axial direction

Uniformity

T6

T7

T2

T3

T4

T5

43.0
44.2
48.2
48.5
48.8

41.4
42.6
45.5
46.2
47.6

43.8
45.1
47.6
48.7
49.2

42.8
44.0
46.4
47.2
48.3

41.6
42.7
44.8
46.1
46.4

36.1
36.8
39.0
40.0
39.4

X(T5)

Fig. 11. Cooling capacity of the DI water immersion arrangement. (a) Inlet and outlet temperature and (b) cooling capacity.

9
8
7

Amps (A)

6
5
4
3
2
1
0
0

10

20

30

40
Volts

50

60

70

80

Fig. 12. Cell module performance during tests. (a) IV curve and (b) other module parameters variation.

0.9
1.0
1.3
1.1
1.0

All
2.6
2.8
3.1
2.9
3.4

L. Zhu et al. / Solar Energy Materials & Solar Cells 95 (2011) 538545

method by liquid immersion has great capacity for heat dissipation for the solar cells under high concentrations.

545

reaction takes the responsibility for this degradation. More detailed

work needs to be carried out for mechanism probing. Furthermore,
tests on silicon oil immersion will be conducted in the near future.

4.4. Cell module performance

A one minute sweep was used to trace the IV curve and to
determine the other parameters of the cell module. A typical full
IV curve is given in Fig. 12a. The short-circuit current (Isc), opencircuit voltage (Voc), maximum power point (Pmax) and also the ll
factor were collected and plotted with time, and are given in
Fig. 12b. Taking the change of DNI given in Fig. 8 into account, it
can be seen that the module performance is stable during the rst
two days test and this stability does not seem to be affected by
the water resistivity variation shown in Fig. 9. But the obvious
degradation was detected after three days test even though the
resistivity of DI water was kept constant. As shown in Fig. 12b,
the curves of the Isc, FF and Pmax have overall decrease trends as
well as sharp uctuation points. It is then deduced that the ion
concentration is not the only fact to be considered when choosing
DI water as the cooling agent, the electrolytic reaction at the
electrodes or connections of the solar cells might cause more
problems. Another deduction is that the resistivity used is still too
low to prevent the degradation. To determine which of them is
right or whether there are still other possible factors, more
detailed work is needed.

5. Conclusions
A DI water immersion cooling method was applied to a dish
high concentration PV system to keep a submerged cell module at a
low working temperature. The real-time temperature distributions
of the cell module were measured as well as its electrical
performance. The experimental results show that the module
temperature peaks at 49 1C and the uniformity of the temperature
distribution is less than 4 1C at 250 suns with the DNI above
900 W/m2, cooling medium temperature inlet temperature of
approximately 31 1C and ambient temperature around 17 1C. The
module temperature distribution is quite uniform. The calculated
cooling capacity using the experimental data is so prominent that
the overall convective heat transfer coefcient is about 6000
W/m2 K. However, the module performance is stable at rst, but
then degrades even at a constant resistivity of 5 MO cm. It is
assumed that either the low ion concentration or the electrolytic

References
[1] A. Royne, C.J. Dey, D.R. Mills, Cooling of photovoltaic cells under concentrated
illumination: a critical review, Solar Energy Materials and Solar Cells 86
(2005) 451483.
[2] A. Royne, C.J. Dey, Design of a jet impingement cooling device for densely
packed PV cells under high concentration, Solar Energy 81 (2007) 10141024.
[3] A. Royne, C.J. Dey, Active cooling device, US patent 2007/0028960A1, 2007.
[4] L. Zhu, Y.P. Wang, Z.L. Fang, Y. Sun, Q.W. Huang, An effective heat dissipation
method for densely packed solar cells under high concentrations, Solar
Energy Materials and Solar Cells 94 (2010) 133140.
[5] K. Tanaka. Solar energy converter using a solar cell in a shallow liquid layer,
Patent US6583349B2, 2003.
[6] K. Tanaka. Solar energy converter using optical concentration through a
liquid, Patent US 2003/005957A1, 2003.
[7] K. Tanaka. Solar energy converter using a solar cell in a shallow liquid-gel
layer, Patent US7244888B1, 2007.
[8] G. Carcangiu, M. Sardo, I. Carcangiu, R. Sardo, Photovoltaic panel and solarpanel unit made using photovoltaic panels of the same sort, Patent US2008/
0092876A1, 2008.
[9] A.H. Ignacio, S.P. Gabriel, D.D. Cesar y, V.P. Marta. Photovoltaic concentrator
with a high efciency immersed into a reective optical dielectric liquid,
Patent 2302656, 2008, (in Spanish).
[10] G. Falbel. Immersed photovoltaic solar power system. Patent US6035319,
2000.
[11] M. Cherney, M. Stiles, N. Williams. Solar collector for solar energy systems.
Patent US6700054B2, 2004.
[12] W.J. Mook J.R, Solar panels with liquid superconcentrators exhibiting wide
elds of view, Patent US 2006/0185713 A1, 2006.
[13] T.Y. Wang, liquid immersion method for solar photovoltaic electricity
generation and the device, Patent CN1302089A, 2001.
[14] H.C. Koehler, Cooling photovoltaic (PV) cells during concentrated solar
radiation in specied arrangement in coolant with as low electric conductivity
as possible, Patent DE19904717, 2000, (in German).
[15] O.V. Vasilinina, J.K. Kidyashev, V.N. Potapov, S.V. Ryabikov, A.M. Stepanov,
D.S. Strebkov, Solar photoelectric conversion apparatus with cooling means,
Patent US4211581, 1980.
[16] K. Tange, Coolant sealing structure for a solar cell, Patent US6005185, 1999.
[17] J.G. Kong, L. Zhu, Y.P. Wang, Y. Sun, Q.W. Huang, Heat dissipation performance
of simulated solar cells when immersed directly by dielectric liquid under
concentrations, Acta Energiae Solaris Sinica 30 (2009) 15921596.
[18] L. Zhu, Y.P. Wang, Z.Y. Zhang, Z.L. Fang, Y. Fang, Study the weather-resistance
of dimethyl polysiloxane used in liquid-immersing CPV receivers, Acta
Energiae Solaris Sinica 31 (2010) 191196.
[19] L. Zhu, Y.P. Wang, Z.Y. Zhang, Q.W. Huang, Y. Zhang, Weather-resistance
study of glycerol to be used in concentrated PV systems, Chemistry 12 (2009)
11221128.
[20] M.A. Newmarker, Development and testing of an advanced photovoltaic
receiver, Dissertation, Department of Mechanical Engineering, University of
Nevada Las Vegas, 2007.