23rd European Photovoltaic Solar Energy Conference, 1-5 September 2008, Valencia, Spain

OPTIMIZATION OF THE SECONDARY OPTICS FOR PHOTOVOLTAIC UNITS WITH FRESNEL LENSES

V.M.Andreev1, V.A.Grilikhes1, A.A.Soluyanov2, E.V.Vlasova2, M.Z.Shvarts1

1
Ioffe Physico-Technical Institute, 26 Polytechnicheskaya str., St.-Petersburg, 194021, Russia
Tel.: +7(812) 292 7394; fax: +7(812) 297 1017; e-mail: shvarts@scell.ioffe.rssi.ru
2
Technoexan LTD, 26 Polytechnicheskaya str., St.-Petersburg, 194021, Russia

ABSTRACT: The aim of the presented investigations was choosing the optimal parameters of secondary optics for
PV modules with flat Fresnel lens concentrators. The following types of secondary optics are under consideration: an open
truncated tetrahedral equilateral pyramid with specular walls and a kaleidoscope with a flat or convex top surface, which
ensure achieving high optical efficiency of a two-element optical system, lowering its sensitivity to the Sun tracking
inaccuracy and increasing uniformity of concentrated radiation distribution on a solar cell. To solve the problem raised,
simulation mathematical model for calculating the optical-power characteristics (OPC) of the “Fresnel lens-secondary
concentrator” system have been developed. As a result of the investigations carried out, the optimum parameters of the
secondary optics have been determined (tilt angle of walls and height of the pyramid or kaleidoscope, curvature radius of the
semispherical input surface), and the best optimal version has been selected for the developed high concentration flat Fresnel
lens.
Keywords: Fresnel Lens, Secondary Optics, Multijunction Solar Cell

1 INTRODUCTION secondary optical elements are a specular pyramidal

The economical estimations carried out in recent
years show that the installations with solar concentrators
and multijunction solar cells (SC) can ensure the
minimum cost of “solar” electrical power at high (up to
1000 X) average radiation concentration ratio [1, 2]. The
main obstacle to achieve the required radiation
concentration ratio in using Fresnel lenses (FL) as
concentrators is the chromatic aberration, which “smears
out” the concentrated radiation and decreases the average
level of the SC irradiance [3-8]. A corresponding choice
of the FL optimal parameters allows decreasing the
negative effect of the chromatic aberration, but to
eliminate it completely in a system comprising only the
primary concentrator is practically impossible [5-9].
Accounting for the nonuniformity of the concentrated
radiation distribution on the SC surface, one can obtain,
due to reducing the cell size, the higher average
concentration without essential losses in the total power
of radiation passed though the lens. However, in
choosing the SC dimensions, it is necessary accounting
for two contradicting tendencies: in increasing the
average concentration ratio the “FL - SC” system optical
efficiency (ηopt) decreases and, on the contrary, the
average concentration ratio decreases with the rise of ηopt.
The final choice is determined by the economical
reasoning with allowing for the design specific features
of a photovoltaic module and requirements to the
accuracy of mutual location of all its elements, and also
with allowing for requirements to the precise orientation
to the Sun of the photovoltaic system, which is especially
important for keeping the power production in real
operation conditions under action of wind and vibrations.
One of the ways for lowering the effect of the
chromatic aberration and inaccurate orientation on the
concentrator system power efficiency and for rising the
average level of the radiation concentration without
reducing the “FL - SC” system optical efficiency is the
application of the secondary concentrating optics located
directly before the SC surface.
As the secondary optics, specular or refracting
elements of different form and also their combinations
can be used. Among the most often met representative
concentrator, a compound concentrator and a glass
kaleidoscope [10-14].
In this work, the results of optimizing parameters of
the secondary concentrator for the following design
versions are presented:
- open equilateral pyramid with specular walls (further
– specular pyramid);
- kaleidoscope (equilateral truncated glass pyramid
with a flat or convex top surface).
Calculation of the secondary optics parameters was
carried out for a system based on a square Fresnel lens of
urethane with side dimension of 40 mm, focal distance
f = 70 mm and profile step of 0.3 mm. Investigations
were carried out for the AM 1.5D LAOD spectrum in the
range of 340 – 920 mm. The lens profile was preliminary
optimized by the procedure presented in [15].
A square solar cell of 1.4 mm x 1.4 mm in size
located in the primary lens focal plane was considered as
a receiver of the radiation concentrated by the “FL-
secondary concentrator” system. At the constructional-
geometrical parameters of a FL enumerated above,
accurate system orientation to the Sun and absence of the
secondary optics, the average concentration ratio Cav on
such a receiver is 700 X, the system optical efficiency is
85.8 %. The values mentioned are obtained with
accounting for the mean-statistical value of the FL tooth
tilt angle deviation (σ) equal to 6 min. of arc and served
as the references in comparing optical systems having the
secondary optics of different types both between each
other and with a system without the secondary optics.
2 MATHEMATICAL MODEL
For each type of the secondary optics, new
mathematical models and algorithms for calculating the
optical-power characteristics (OPC) of the “FL-
secondary concentrator” system have been developed.
The mathematical models and algorithms allow tracing
the path of beams, emerging from the primary lens,
which are passing through all elements of the secondary
optical system up to their getting a receiver. In this case,
dispersion in passing beams through the system

126
 23rd European Photovoltaic Solar Energy Conference, 1-5 September 2008, Valencia, Spain

refracting elements is simulated, and all main types of convenient for analysis is the dimensionless parameter
energy losses on their surface and in the bulk are taken R (relative radius), which is the ratio of Rs to the radius
into account: beams going by the top (input) surface of of a circle circumscribed about the pyramid top base R:
the secondary concentrator, Fresnel losses on the R =Rs /R. At R =1, the covering lens is a semisphere, at
kaleidoscope top and bottom surfaces both with allowing
R >1 - a sphere segment, thickness of which decreases
for an antireflective coating and without it,
counterreflection of beams (not towards a SC but towards with R and at R →∞ - the top surface becomes flat and
the secondary optics element top surface), absorption in kaleidoscope is shaped in a form of truncated tetrahedral
the material bulk of the refracting elements, losses at equilateral pyramid.
specular single and multiple reflection from the walls of
the pyramid and kaleidoscope. a b
With the use of the developed models and algorithms R
θ Rs
for each type of the secondary optics, calculations of
concentrated radiation distribution over the SC surface, h
of dependencies of the average concentration ratio Cav
for a SC and of the system optical efficiency ηopt on the
acceptance angle ν have been carried out.
It should be noted that, at predetermined parameters
Figure 1: Kaleidoscope secondary concentrator: a – front
of the primary lens and SC dimensions, changes in the
view; b – view from the top convex surface
Cav and ηopt values result from the effect on them of only
the secondary optics parameters, and both these values
Thus, the optimized parameters for specular pyramid
(i.e. Cav and ηopt) may equally be used for comparative are the pyramid height h and the inclination angle of
estimation of the efficiency of systems with different
walls θ, and those for the kaleidoscope - h, θ and R ,
type secondary concentrators. In taking this into account,
correspondingly.
as an indicator of the system efficiency in choosing
Other initial data used in simulating are:
optimum values of the secondary optics parameters, the
- reflectance of the pyramid specular walls taken equal
system optical efficiency ηopt was used, since this value
to 0.93;
was the most convenient in plotting and understanding
- material of the kaleidoscope – optical glass with
the dependencies.
corresponding dependence of its refractive index on
Since the secondary concentrator parameters’
wavelength;
optimum values obtained for the conditions of accurate
- antireflection coating on the kaleidoscope input
and inaccurate orientations do not coincide in most cases,
surface.
it was necessary to formulate the criterion of their choice
Below presented are the results of parameters’
in such a way that this contradiction could be resolved.
optimization for the enumerated above secondary
For this purpose, the concession principle was used. This
concentrators.
principle implies that the system optimization is carried
out independently over all criteria available, and the
choice of optimal parameters is done according to one of
3 RESULTS OF OPTIMIZATION OF SECONDARY
them, which is accepted as main one. For the rest of the
OPTICS PARAMETERS
criteria such concession values are taken that determine
the acceptable level in reducing the system efficiency by
3.1 Specular pyramid
these criteria.
Figure 2 presents dependencies of the optical
According to this principle, optimization of the
efficiency ηopt on the angle of inclination of the pyramid
systems with secondary concentrators of different types
walls θ for different values of h at ν = 0° and at ν = 1°.
was carried out twice: by the maximum ηopt criterion at
accurate orientation and by that at a typical value of the
ν = 0o
h=18 mm
acceptance angle ν =1°. As the main criterion, the optical 12 mm
efficiency maximum at ν =1° was chosen, and the 89 10 mm
6 mm
concession value was taken equal to 10%. Optimization 4 mm
Optical efficiency, %

of the systems with different types of secondary 80

concentrators by both criteria was carried out by the pull- h=18 mm ν = 1o
down method. 75
The calculations carried out have shown that for all 12 mm
versions of the secondary optics the presence of an air 70 10 mm

gap between its bottom surface and a SC results in 65 6 mm

deterioration of the concentrating system efficiency.
Taking into account this circumstance, the versions of the 60 4 mm
secondary optics with in an air gap were excluded from
16 18 20 22 24 26
consideration at this stage of the work.
Inclination angle θ, degrees
To describe the set of the parameters being optimized
in Fig. 1, as an example, the kaleidoscope with a convex Figure 2: Theoretical dependencies of the optical
top surface is presented. Parameters of this type efficiency (ηopt) on the inclination angle of walls (θ) and
concentrator being optimized are: pyramid height h, height (h) of a specular pyramid at the acceptance angles
inclination angle of its walls θ and radius of curvature of ν = 0° (top graph) and ν =1° (bottom graph). h values
the covering input surface Rs. In practice, more are indicated on a plot

127
 23rd European Photovoltaic Solar Energy Conference, 1-5 September 2008, Valencia, Spain
It follows from the analysis of the dependencies that
in both cases the optimum value of θ lies in the range of
16 - 17°. Existence of the optimum by θ results from the
contradicting effect of this parameter on the value of the
radiation flow of the primary lens intercepted by the
secondary concentrator and on the value of losses at
counterreflection from the pyramid walls. Increasing the
pyramid height results in rising ηopt, which is caused by
the increase in the coefficient of interception of the
primary lens radiation by the secondary concentrator
input surface. At the same time, with increasing the
pyramid height, the increase in ηopt becomes less
noticeable, and at h > 18 mm, it practically stops, which
results from the rise of the amount of specular reflections
from the pyramid walls and energy losses associated with
this.
3.2 Kaleidoscope with flat top surface
Fig. 3 presents dependences of the optical efficiency
ηopt on a SC on the angle of inclination of the
kaleidoscope walls θ for different h values at ν = 0° and
ν = 1°. It follows from the dependencies that more
substantial effect on this type secondary concentrator
efficiency comes from the kaleidoscope height h. This
influence is explained by two factors. First, the
kaleidoscope height affects directly the value of losses at
absorption of the sunlight in it. Second, with increasing
the kaleidoscope height, the number of reflections from
its walls rises, in most cases only the first reflection
being total (due to the total internal reflection effect), and
all following ones being accompanied by a drastic rise of
the energy losses.
87
86
85
84
Optical efficiency, %
83
o dependencies (see Fig. 4 and Fig. 5) that, to achieve
77 h=10 mm ν =0
70 8 mm
6 mm
63
4 mm
2 mm
56
14 16 18 20 22 24
Inclination angle θ, degrees
Figure 3: Theoretical dependencies of the optical
efficiency ηopt on the inclination angle of walls (θ) and
height (h) of a kaleidoscope with flat top surface at the
acceptance angles ν = 0° and ν = 1°. h values are
indicated on a plot
The increase of these losses is associated with that, at
the second and following reflections, the angle of the
beam incidence on the “glass-air” demarcation line
becomes, as a rule, smaller than the limiting one, and
more and more rising part of its energy passes into the
refracted beam, and, due to this, is lost.
The effect of the angle of inclination of the
kaleidoscope walls θ on its efficiency, as in the case of
the specular pyramid, is associated with contradicting
effect of this parameter on the primary lens intercepted
radiation flow value and on the value of losses at
reflection from the “glass -air” demarcation line, which
was discussed above. At a small kaleidoscope height, a
large part of beams is reflected once (i.e., without losses)
practically independently on the angle θ in a wide range
of its variation. This explains a comparatively weak
dependence of the kaleidoscope efficiency on this
parameter in the region of small h value.
The maximum ηopt value at accurate orientation is
achieved at h = 3 mm and the quite wide range of
inclination angles of walls θ = 18-30°.
At the presence of misorientation, the contradicting
influence of different factors on the kaleidoscope
efficiency becomes to be more complicated. This results
in appearing the optical efficiency ηopt maximum in the
region of h = 8-10 mm and θ = 17-18° (see Fig. 3).
3.3 Kaleidoscope with convex top surface
The process of optimizing the kaleidoscope with
convex top surface parameters by the pull-down method
has been organized in the form of three home loops: by
the specular pyramid height h (outer loop), by the angle
of inclination of its walls θ and by the curvature relative
radius of the input surface R (inner loop). Iterations by
variable parameters inside each cycle were performed up
to achievement of the local maximum of the chosen
system efficiency. The optimum values of the secondary
concentrator parameters correspond to the largest value
of the system efficiency among obtained local maxima.
Fig. 4 presents dependencies of ηopt on the inclination
angle of the kaleidoscope walls θ for its heights h from 2
to 10 mm at four values of the curvature relative radius
of the input surface R and at the acceptance angle
2 mm
4 mm ν = 0°, and Fig. 5 presents the same dependencies for
6 mm ν = 1º.
It follows from the analysis of the presented
8 mm
maximum values of the optical efficiency and,
correspondingly, maximum values of the average
concentration ratio in the misorientation conditions, the
curvature relative radius of the kaleidoscope top surface
must be in the limits of R ≈1.5 - 2 at inclination angle of
ν =1
o walls of 22 - 28º and height of 4 - 8 mm.
26 28 30 32 4 FLUX DENSITY DISTRIBUTION ON A SC
Beside capability of the secondary optics to raise the
average concentration ratio and optical efficiency, its
important property is a possibility to change the flux
density distribution character and to create uniform
irradiance of the SC surface with the aim to compensate
the negative effect of the radiation redistribution on the
multijunction SC characteristics. For this reason, to
choose an optimal “FL – secondary concentrator” optical
system, it is necessary to compare the optical-power
characteristics (OPC) of these system with the aim to
choose a version ensuring the most uniform irradiance
distribution on a SC. In the given case, of great interest is
the analysis of OPCs obtained in a system with a
kaleidoscope at different orientation conditions, since
possibilities to change the irradiance character by a
specular pyramid appear to be insignificant compared
with distribution produced by a primary lens.
128
 23rd European Photovoltaic Solar Energy Conference, 1-5 September 2008, Valencia, Spain

a 88
h=2 mm a 88

86 86
4 mm
84 84
Optical efficiency, %

Optical efficiency, %
6 mm h=4 mm
82 82

80 80 2 mm
8 mm
78 78
76 76
74 74
6 mm
72 72 8 mm

70 70
18 20 22 24 26 28 30 32 18 20 22 24 26 28 30 32
b 88
h=2 mm
b 88
86 4 mm 86

84 6 mm 84
Optical efficiency, %

Optical efficiency, %
8 mm
82 10 mm 82 4 mm
6 mm
80 80
8 mm
78 78 10 mm

76 76
h=2 mm
74 74

72 72

70 70
18 20 22 24 26 28 30 32 18 20 22 24 26 28 30 32
c 88
h=2 mm
c 88

86 4 mm 86
6 mm
84 84
Optical efficiency, %
Optical efficiency, %

8 mm
6 mm
82 10 mm 82

80 80

78 78 8 mm
10 mm
76 76

74 74 4 mm
72 72 h=2 mm
70 70
18 20 22 24 26 28 30 32 18 20 22 24 26 28 30 32
d 88
h=2 mm d 88
86 4 mm 86
6 mm
84
Optical efficiency, %

84
Optical efficiency, %

8 mm
82 10 mm
82 h=10 mm
80 80 6 mm
78 78
8 mm
76 76
74 74
72 72
4 mm
70 70
18 20 22 24 26 28 30 32 18 20 22 24 26 28 30 32
Inclination angle θ, degrees Inclination angle θ, degrees

Figure 4: Theoretical dependencies of optical efficiency
(ηopt) on the height (h) and the inclination angle of walls
(θ) of a kaleidoscope with convex top surface at the
acceptance angle ν=0°. The curvature relative radius R
is equal to: a – 1.1 , b –1.5 , c – 2 , d – 4
Figure 5: Theoretical dependencies of optical efficiency
(ηopt) on the height (h) and the inclination angle of walls
(θ) of a kaleidoscope with convex top surface at the
acceptance angle ν=1°. The curvature relative radius R
is equal to: a – 1.1 , b –1.5 , c – 2 , d – 4

129
 23rd European Photovoltaic Solar Energy Conference, 1-5 September 2008, Valencia, Spain

Combined data on the optimum values of the 2000

secondary concentrator parameters obtained by a
simulation are presented in Table 1. It should be taken 1500
ν =1
o

into account that the optimum secondary concentrator 1000 o

parameters correspond to the criterion of the maximum ν =0
ηopt at the acceptance angle ν = 1º for a system with 500

preassigned parameters of the primary lens and SC 0

dimensions.
3500
b ν =0
o

Table I: Data for the optimum types of the secondary

Concentration ratio, X
optics: 1-7 - kaleidoscope with convex (1-6) and flat (7) 3000
top surface, 8 - specular pyramid 2500
o
ν =1
2000
R h, θ, Cloc, X Cloc, X ηopt , % ηopt , %
Type

mm 1500
degree ν = 0º ν = 1º ν = 0º ν = 1º
1000
1 1.1 4 27 5620 8230 85.3 82.5
2 1.1 6 22 870 1220 84.0 79.2 500
3 1.5 6 25 5560 7880 84.8 83.1
0
4 2 6 25 3760 3880 84.9 83.3 2000

5 2 4 28 3450 2690 85.8 81.7 c ν =1

o
1500
6 4 10 20 1500 1660 83.3 81.3 o
ν =0
7 ∞ 10 18 720 970 83.5 77.0 1000

8 - 18 17 2340 1690 89.3 76.6 500

Comparison of OPCs show that the uniform 0

-0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4 0.6 0.8
irradiance distribution can be achieved at the following Radius, mm
kaleidoscope parameters: h = 6 mm, θ = 22º, R = 1.1
(Fig. 6). However, in this case, the optical losses at Figure 6: Distribution of the concentrated radiation over
misorientation become significant, which result in the solar cell radius for the kaleidoscope systems of
efficiency drop down to ~79% (Table I). The optical following configurations (see Table I): a – 2, b – 5, c - 6
losses can be to some extent decreased, if the version
For the kaleidoscope with the convex top surface
with h = 10 mm, θ = 20º, R =4 (Table I and Fig. 6) is
accepted, the illumination distribution nonuniformity (h = 6 mm, θ = 22º, R = 1.1 ), in which, at a precise
rising. orientation, a uniform irradiance distribution is ensured
Maximum optical efficiencies are achieved in the (see Fig. 6), the optical efficiency lowers from 83.4% at
ν = 0º down to 80% at ν = 0.95º with some increase in
system with R ≈1.5 - 2 (see Table I), but in these cases
illumination difference over the SC surface (see Fig. 7).
it is impossible to compensate the pronounced
A similar situation takes place also in using kaleidoscope
nonuniformity of the irradiance distribution at any
with a flat top surface at h = 10 mm and θ = 18º.
combination of h and θ parameters. The most acceptable
However, in the present case, the acceptable values of the
version with a maximum optical efficiency at the level of
optical efficiency (more than 80%) are conserved at the
85.8% at precise orientation, that comparable with
acceptance angle less than 0.85º.
optical efficiency for a system without the secondary
optics, is achieved at the kaleidoscope parameters h = 4
mm, θ = 28º, R = 2 (the local concentration in the center 90
is 3450X). Passing to a position with ν = 1º for the
88
mentioned kaleidoscope configuration leads to forming
8
Optical efficiency, %

the more uniform light distribution (see Fig.6) in 86

conserving the efficiency at the level of 81.7 %.
84
4
5 COMPARISON OF THE SYSTEM EFFICIENCY 82 5
WITH THE OPTIMUM SECONDARY OPTICS. 6
80
2
Fig. 7 presents dependencies of the optical efficiency 78
of the concentrating system of the considered above type 7
and configuration (see Table I) on the orientation 76
accuracy. 0.0 0.2 0.4 0.6 0.8 1.0
Comparison shows that in a system with a specular Acceptance angle ν , degrees
pyramid, due to low optical losses at refraction from the
specular walls, an advantage compared with Figure 7: Theoretical dependencies of the optical
kaleidoscope of any configuration is ensured only at the efficiency on the acceptance angle ν for the systems with
acceptance angles less than 0.5º. At greater angles the the secondary optics of different configurations indicated
optical efficiency of such a system drops down to 76.6%. in Table I

130
 23rd European Photovoltaic Solar Energy Conference, 1-5 September 2008, Valencia, Spain
In going to the systems with higher values of optical
efficiency at ν = 1º, it should be searched for a
compromise between 2 and 4% reduction of optical
losses and a possible negative effect of significant
differences in illumination over the cell surface on the
operation of a multijunction SC (local concentration in
the spot are from 1500X to 8000X).
6 CONCLUSION
Mathematical models and algorithms for calculating
optical-power characteristics of the “Fl-secondary
concentrator” system have been developed. The results of
a theoretical investigation of the effect of design
parameters of secondary concentrators on the
concentrated on the SC surface radiation distribution
character and on the average radiation concentration ratio
and the concentration system optical efficiency for
different orientation angles are presented. Optimum
parameters for the secondary concentrators of three types
have been determined: an open lateral pyramid with
specular walls and a kaleidoscope with the flat or convex
top surface. It has been shown that it is impossible in the
considered “Fl-secondary concentrator” systems to
ensure a uniform distribution of irradiance with
conserving a high optical efficiency and weak sensitivity
of the concentration system characteristics to the
inaccuracy of the system orientation to the Sun.
It has been found that, for the considered primary
Fresnel lens with preassigned design–geometrical
parameters, there exist versions of manufacturing the
secondary concentrator of the kaleidoscope type (see
Table I, version 2 and 7), at which a uniform irradiance
distribution on the SC surface takes place at precise
orientation to the Sun (see Fig. 6), and the optical
efficiency exceeds 83.4 %. In the misorientation
conditions an insignificant difference in illumination over
the SC surface arises at conserving the optical efficiency
at the level higher 80% at angles 0.7º and 0.5º for the
kaleidoscope version 2 and 7, correspondingly.
In going to the systems with higher optical
efficiency at ν = 1º, it should be searches for a
compromise between the negative effect of arising
differences in illumination of a cell (local concentrations
in the spot from 1500X to 8000X) and the 2-4% decrease
in optical losses.
It is obvious that the “Fresnel lens -
kaleidoscope” optical system version of 2 and 6 types
(see Table I and Fig. 6) creating a more uniform
irradiance distribution are, by the sum total of effects,
preferential at their matching with multijunction SCs,
even in spite of insignificant (at the level of 2%) drop of
the optical efficiency compared with maximally
achievable rated values.
131
7 ACKNOWLEDGMENT
The authors wish to thank to N.Kh.Timoshina for the
technical assistance.
This work was partly supported by Russian
Foundation on Basic Research (Grants № 05-08-33603
and № 07-08-13616) and by FULLSPECTRUM Project
(Contract SES6-CT-2003-502620).
8 REFERENCES
[1] C. Algora, chapter 6 in Next Generation
Photovoltaics: High Efficiency Througth Full Spectrum
Utilization, IOP Publishing (2004) 108.
[2] D Aiken, B Clevenger, F Newman, B .Smith,
C. Tourino, J. Wilson, J Wood, Proceedings of 33rd IEEE
Photovoltaic Specialist Conference (2008) on CD
[3] L. James, Proceedings of First World Conference on
Photovoltaic Energy Conversion (1994) 1799.
[4] S. Kurtz, D. Friedman, J. Olson, Proceedings of First
World Conference on Photovoltaic Energy Conversion,
(1994) 1791.
[5] M.J. O’Neill, Proceedings of 25th IEEE Photovoltaic
Specialist Conference (1996) 349.
[6] M.J. O’Neill, U.S. Patent 6,031,179 (2000).
[7] V. Grilikhes, E. Bobkova, A. Soluyanov, M. Shvarts,
Technical Physics Letters, 32 (2006) 1039.
[8] V. Grilikhes, E. Bobkova, A. Soluyanov, M. Shvarts,
Geliotekhnica, 3 (2006) 50 (Translated into English in
Applied Solar Energy)
[9] V.A. Grilikhes, M.Z. Shvarts, A.A. Soluyanov,
E.V. Vlasova, V.M. Andreev, Proceedings of 4th
International Conference on Solar Concentrators for the
Generation of Electricity or Hydrogen (2007) 49.
[10] D. Aiken, M. Stan, G. Girard, S. Endicter, P. Sharps,
Proceedings of 31st IEEE Photovoltaic Specialist
Conference PVSC (2005) 743.
[11] K. Araki, M. Yamaguchi, Proceedings of the 17th
European Photovoltaic Solar Energy Conference (2001)
2187.
[12] K. Araki, M. Kondo, H. Uozumi, M. Yamaguchi,
Proceedings of 3rd World Conference on Photovoltaic
Energy Conversion (1994) 634.
[13] K. Araki, R. Leutz, M. Kondo, A. Akisawa, T.
Kashiwagi, M.Yamaguchi, Proceedings of 29th IEEE
Photovoltaic Specialist Conference (2008) 1571.
[14] J. Jaus, P. Nitz, G. Peharz, G. Shiefer, T. Schult,
O. Wolf, M. Passing, T. Gandy, A. Bett, Proceedings of
33rd IEEE Photovoltaic Specialist Conference (2008) on
CD.
[15] M.Z. Shvarts, V.M. Andreev, V.S. Gorohov,
V.A. Grilikhes, A.E. Petrenko, A.A. Soluyanov,
N.Kh. Timoshina, E.V. Vlasova, E.M. Zaharevich,
Proceedings of the 33rd IEEE Photovoltaic Specialists
Conference, (2008), on CD.