Solar Energy Materials & Solar Cells 132 (2015) 523–527

Contents lists available at ScienceDirect

Solar Energy Materials & Solar Cells

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

Influence of photovoltaic angle-dependence on overall power output

for fixed building integrated configurations
Yunhua Ding a, Margaret Young a, Yimu Zhao a, Christopher Traverse a, Andre Benard b,
Richard R. Lunt a,c,n
a
Department of Chemical Engineering and Materials Science, Michigan State University, East Lansing, MI 48824, USA
b
Department of Mechanical Engineering, Michigan State University, East Lansing, MI 48824, USA
c
Department of Physics and Astronomy, Michigan State University, East Lansing, MI 48824, USA

art ic l e i nf o a b s t r a c t

Article history: Building integrated photovoltaic (BIPV) systems are an emerging approach to reduce installation costs
Received 24 August 2014 while supplementing building energy needs. However, the physical constraints of the building
Accepted 24 September 2014 architecture often prevent photovoltaic systems from being installed at their optimal orientation.
Recently, it was shown that thin film photovoltaics can be designed for improved angle-dependent
Keywords: responsivity at specific angles. In this study, the complex impact of angular dependency on overall power
Thin film photovoltaics output is explored based on detailed hourly solar position, location, and flux data. These results
Building integrated photovoltaics demonstrate that reducing the angular roll-off dependence can enhance overall power outputs by 30% or
Angle-dependence more in fixed orientation configurations depending on the geographical location, orientation, and angle-
Non-normal
dependent roll-off characteristics.
Solar illumination
& 2014 Elsevier B.V. All rights reserved.

1. Introduction designed to have specific angle dependence properties [7–9] with

Building integrated photovoltaics (BIPVs) have received con-
siderable attention due to their sustainable attributes and func-
tional value [1–3]. A typical BIPV application, for example, is the
use of silicon-based or copper indium gallium selenide photo-
voltaics (PVs) on rooftops and shingles. A significant number of
studies have been conducted to identify the optimal tilt angle and
orientation to obtain longer periods of near-normal incident
illuminated sunlight [4,5]. However, considering the trajectory of
the sun, these deployed solar cells are rarely illuminated at normal
incidence. In addition, the high cost of rotational mounting
systems and environmental considerations (wind, snow, etc.) have
primarily limited practical designs to fixed configurations, and
many synergistic BIPV approaches (such as solar shingles) are
already necessarily in fixed configurations. Hence, it is critical to
design PVs that give stable performance over a range of incident
angles.
In comparison to traditional silicon-based PV applications, thin
film and organic photovoltaic technologies have increased their
market share due to their potential for light weight, flexible, and
transparent applications [6]. Importantly, thin film PVs can be
n
Corresponding author at: Michigan State University, 4135 Engineering Building,
East Lansing, MI 48824, USA. Tel.: þ1 517 432 2132.
E-mail address: rlunt@msu.edu (R.R. Lunt).
the improvement of external and internal quantum efficiencies
[10,11]. Moreover, window integrated solar cells are gaining
attention due to the development of transparent photovoltaics
(TPVs) which have exceptional low-cost potential and are enabled
by new excitonic materials [12,13]. In this case, TPVs can be
considered for siding, windows, and skylights, which normally
have fixed structures [14]. Recently, TPV designs with enhanced
angle-dependence were demonstrated, leading to the improve-
ments in responsivity by as much as 50% at particular oblique
angles with nearly identical performance at normal incidence [15–
17]; the improvements were achieved by considering and optimiz-
ing layer structures for non-normal incidence.
For all of these PV technologies, fixed configuration deploy-
ment situations naturally leads to the question: what impact can
the designs with minimized angle-dependence have? Here, the
overall power output for the TPV designs with different respon-
sivities over varying incidence angle and solar radiation was
evaluated to assess the impact of angle-dependent PV efficiency
in a range of configurations and locations.
2. Theory and modeling
The global tilted irradiance ðI GT Þ, which is needed to estimate
the power output from a solar panel, is calculated using the

http://dx.doi.org/10.1016/j.solmat.2014.09.031
0927-0248/& 2014 Elsevier B.V. All rights reserved.
524 Y. Ding et al. / Solar Energy Materials & Solar Cells 132 (2015) 523–527

Nomenclature I input intensity (kW/m2)

JSC photocurrent density (A)
FF fill factor METSTATmeteorological–statistical solar model
I DFH METSTAT-modeled diffuse horizontal irradiance (kW/ Q yearly power output
m2 ) R responsivity (A/W)
I DFT Global tilt irradiance (kW/m2) VOC open-circuit voltage (V)
I DRN METSTAT-modeled direct normal irradiance (kW/m2) Α Azimuth angle of the PV module (degrees)
I DRT direct tilt irradiance (kW/m2) Β tilt angle of the PV module (degree)
I EH extraterrestrial horizontal irradiance (kW/m2) δ improvement of yearly power output
I GH METSTAT-modeled global horizontal irradiance (kW/ η efficiency of a solar cell
m2 ) Θ incidence angle of solar ray to PV modules (degrees)
I GR ground reflected irradiance (kW/m2) φ hourly mean zenith angle of the sun (degrees)
I GT global tilted irradiance (kW/m2) ω hourly mean azimuth angle of the sun (degrees)

average values of the hourly data of sun position and solar illumination flux (I) by the following equation:
irradiance published in the National Solar Radiation Database over





Q i θ; I ¼ η θ; I I ¼ FF θ; I V oc θ; I R θ; I I ð4Þ
a ten-year period [18,19]. The calculation follows the separation


method described in the Hay model [20]: where R is the responsivity calculated from R θ; I ¼ J sc ðθ; IÞ=I; JSC is
I GT ¼ I DRN cos θ þ I DFT þI GR ð1Þ the short-circuit photocurrent density; VOC is the open-circuit
voltage; and FF is the fill factor.
and To determine the effect of θ and I, angular dependency and


ðI GH  I DFH Þ cos θ cos φ ð1  ðI GH  I DFH ÞÞ 1 þ cos β intensity are assumed to be essentially independent so that
I DFT ¼ I DFH þ



I EH 2 η θ; I ¼ η θ; Iref η θref ; I ð5Þ
ð2Þ
where the reference intensity is 1 mW/mm2 (1-sun), and the
where I DRN is the direct normal irradiance; I DFT is the diffuse tilted
reference incident angle is 01. While the impact of angular
irradiance; I GR is the ground reflected irradiance; I GH is the global
dependency on FF and VOC is typically negligible [13], the intensity
horizontal irradiance; I GR is the ground reflected irradiance; I DFH is
dependencies of FF, VOC, and R are nonetheless captured in the
the diffuse horizontal irradiance; I EH is the extraterrestrial hor-
intensity dependent component of experimentally determined
izontal irradiance; θ is the incident angle of direct solar irradiance
efficiencies. Therefore, the angle dependent component of η is
with respect to the PV module; φ is the zenith angle of the sun
proportional to the angle dependent R at a fixed I [15].
with respect to the horizon; ω is the azimuth angle of the sun with
In the simulation of yearly power output, the average value of
respect to north; β is the tilt angle of the PV module with respect
responsivity ðRÞ over 0–901 incident angle is used for the diffuse
to the horizon; and α is the azimuth angle of the PV module
irradiance components. Then, the yearly power output (Q) is
(Fig. 1). The solar irradiance is modeled by the meteorological–
evaluated for a conventional thin film PV (device A), a thin film
statistical solar method (METSTAT) [19]. Here, I GR can be neglected
PV architecture designed with improved angle dependence
due to the insignificant impact on I GT comparing to the I DT and I DFT
(device B) [15], and devices with idealized angle-dependent cut-
[21]. Then, θ is calculated by the following equation:
offs shown in Fig. 2 as follows:
! !
θ ¼ cos  1 S U M ð3Þ
τ 

Q θ; I ¼ ∑ ðR θ; 1 I DRN;i þ RI DFT;i Þ1 h ð6Þ
!
where S ¼ ðω; φ  901; 1Þ is the unit vector of incoming radiation, i¼1
!
and M is the normal vector of the PV module defined in the left- where τ is the number of hours in one year. The yearly power
handed spherical coordinate system (α; β,1), which is calculated output for A and B were normalized to the power output of device
!
from a unit vector within the ground plane, G ¼ ðα 901; 01; 1Þ, A at β ¼ 01, and the yearly power output for the cutoff devices
!
and a second unit vector perpendicular to G within the tilted were normalized to the ideal device response at each tilt. The ideal
! ! ! !
plane, T ¼ ðα þ1801; β; 1Þ, of the PV module as M ¼ T  G .
Hourly instantaneous power output (Qi) is calculated from the
power conversion efficiency of the PV module (η) and the

Fig. 1. Schematic showing all angles related to the position of the sun in the Fig. 2. Normalized angle-dependent responsivity for a conventional thin film PV
northern hemisphere and orientation of the PV module. Note that the dotted lines (device A), thin film PV designed for improved angle-dependence (device B), and
are in the horizontal plane and the solid lines are out-of-plane for clarification. selected cutoff angle of angle-dependence designed thin film PVs.
 Y. Ding et al. / Solar Energy Materials & Solar Cells 132 (2015) 523–527 525

integrated flux data are typically found to be consistent with

monthly and yearly averaged values within 10–15% of values
reported elsewhere across a range of locations and orientations
[22].

3. Experiment

The current density versus voltage (JV) characteristics were

measured for an archetypal a mono-Si PV (Narec Solar) and a thin
film solar cell composed of a chloroaluminium phthalocyanine
(ClAlPc) – C60 planar heterojunction [15] under various incident
angles and overall light intensities. The former was chosen for its
high efficiency and the latter was chosen because of its potential
use in transparent PVs for building integrated applications and its
use in PVs with high photovoltage approaching the Shockley–
Quiesser limit [23]. To determine the intensity dependence at
normal incidence, a Newport 67005 Arc Lamp with an AM1.5 Filter
and a series of neutral density filters (Thorlabs) were used to
provide different light intensities, which were corrected by the
solar spectrum mismatch factor at each intensity [24]. To char-
acterize the incident angle dependence, the external quantum
efficiencies (EQE) of the solar cells were measured with a lock-in
amplifier (Standford Research System SR830 DSP) and a picoam-
meter (Keithley 485) at various angles with respect to a fixed
monochromatic light source controlled by a substrate rotation
stage (Newport). The light source underfilled all devices with a
chopped beam at 77 Hz. The EQE was then integrated with the
product of the solar flux to give the angle-dependent responsivity.

Fig. 4. Normalized yearly power output of device A at various orientations in

(a) Lansing, MI, (b) Phoenix, AZ, and (c) Meru, Kenya. Note that the overall output
4. Results and discussion yearly powers are normalized to the non-tilted (horizontal) configuration.

PV performance characteristics of the archetypal thin film PV
and mono-Si PVs are normalized to those devices at normal
incidence and one sun illumination, and shown in Fig. 3 as a
function of intensity. Smaller intensity dependence is observed in
FF, R, and VOC for the thin film PV compared to the Si PV. For the Si
PV, this intensity dependence primarily stems from larger reduc-
tions in the VOC. For the thin film PV, the R decreases only slightly
with decreasing intensity, while FF increases slightly. The opposite
trend is seen for the Si PV. Consequently, the η for the Si PV shows
Fig. 3. Normalized properties of a representative thin film PV (a) and Si PV
(b) versus intensity.
a logarithmic intensity dependence while the η for the thin film PV
is essentially constant across all intensities above 0.1 mW/mm2,
below which the uncertainties in the measurement begin to
dominate. The trend of the η shown in Fig. 3 suggests that the
thin film PV would perform well throughout the day because
sunlight illumination is often above 0.1 mW/mm2, and its effi-
ciency is constant in this regime. Since the overall intensity
dependence is not significant for the thin film PV, it does not
significantly impact the overall power generation.
The output power of the two thin film PV device designs with
significant variation in the angle dependence (Fig. 2) are compared
in Figs. 4,5. Device B, a thin film architecture designed with angle
dependence in mind, exhibits limited efficiency roll-off with
incidence angle compared to the conventional structure (device
A). Fig. 4 shows the normalized yearly power output of device A
(QA) at various PV orientations for Lansing, MI, Phoenix, AZ, and
equator represented by Meru, Kenya. The PV orientation covers the
full 3601 range of azimuth angle with 301 increment and selected
tilt angles of 01 (horizontal), 301, 601, and 901 (vertical). The overall
power output is clearly the same at all orientations when the tilt
angle is 01 since the PV is in a horizontal configuration. When the
tilt angle (β) is 4 01, a symmetric response appears with a
maximum at the south-facing direction (α ¼1801) for Lansing
and Phoenix (Fig. 4a and b). This is because the irradiance and
incident angle are maximized for the south-facing orientation and
are similar for the east-facing and west-facing orientations. The
power output is diminished for Lansing and Phoenix when facing
north because the sun mostly appears in the southern sky for
latitudes above 1.5 1N and the photoresponse stems primarily from
diffuse irradiance [4]. However, the variation in power output is
greater for lower latitude locations due to a greater variation in
incident angle. In addition, the optimal tilt angles which yield the
largest overall power output are close to 301 for Lansing and
526 Y. Ding et al. / Solar Energy Materials & Solar Cells 132 (2015) 523–527
Fig. 5. Normalized yearly power output improvement made by device B in
(a) Lansing, MI, (b) Phoenix, AZ, and (c) Meru, Kenya.
Phoenix, and 01for Meru, which are consistent with previous
reports [4,21,25] and closely related to the latitude. As the tilt
angle diverges from the optimal tilt angle, the irradiance decreases
in all directions, which reduces the overall power output.
The trends are different for tilted devices in Meru (Fig. 4c): the
data are asymmetrical and reveal peaks at the east-facing and
west-facing directions with a higher value for the west-facing
direction; the values are the same at the north-facing and south-
facing directions. The divergence in the east-west symmetry for
Meru stems from two phenomena: (1) the time that the sun is
positioned in the southern and northern sky is nearly the same at
equator, so the power output of the modules facing either south or
north are similar; and (2) the irradiance is slightly stronger and the
illumination period is longer in the afternoon than in the morning,
which results in a greater power output for the west-facing
direction than the east-facing direction.
Fig. 5 shows the improvement in power output by using device
B ðδB Þ, which has minimal angle dependence, at various PV
orientations for the three locations. An optimum enhancement
appears at the north-facing and south-facing directions, which
becomes more pronounced as the tilt angle increases from 301 to
901 for Lansing and Phoenix, and from 01 to 901 for Meru. Fig. 5
clearly illustrates that the improvement in device B's angular
dependence (highlighted with the blue area from Fig. 2) signifi-
cantly increases power output. For Lansing and Phoenix, θ is
always very large for titled modules at the north-facing directions,
so the improvement in power output exceeds that in the south-
facing directions. However, the most beneficial orientations are
still south-facing where the most direct incidence occurs. For
modules with small tilt angle β at midday, the incident angle is
also small such that the advantage of device B is less significant
(Fig. 2). In contrast, for a large module tilt (e.g. vertical orienta-
tion), the incident angle θ would be greater than 401 at midday
[26], where the difference of responsivity between devices B and A
Fig. 6. Normalized yearly power output for idealized responsivities with sharp
incident-angle dependent cutoff performance with modules at various tilt angles in
Phoenix, AZ. Note that the overall output power in each panel is normalized to the
corresponding ideal output power with no angle-dependence.
begins to emerge and the illumination is the strongest, so the
overall power output is greatly enhanced with improved angle
dependence. Since the solar irradiation is larger at a lower
latitudes, the improvement ðδB Þ is more substantial in locations
such as Phoenix and Meru. This data ultimately highlights that the
angle dependent performance plays a large role for a range of
orientations in BIPVs that can translate into substantially
enhanced power outputs or equivalently to higher efficiencies.
To further generalize the impact of cutoff angle in the angle-
dependent PV design, additional simulations for hypothetical
modules with sharp responsivity cutoff angles (401, 551, 701, 801,
and 901) with several selected tilt angles were performed using
the solar irradiance data from Phoenix. Fig. 6 shows that: (1) for
horizontal and low tilt angle (o 301), the critical cutoff angle
(where total output drops by more than 20%) is in the range of θ
around 55–701 so that good responsivity at very oblique incident
angles is less significant; and (2) for highly tilted modules, the key
cutoff angle becomes greater (θ around 70–801) so that this effect
is more substantial, and good responsivity across all incident
angles is highly desirable. It is also clear that larger cutoff angle
result in more uniform power output for deployment over a full
range of azimuthal rotations than smaller cutoff angle at tilted
 Y. Ding et al. / Solar Energy Materials & Solar Cells 132 (2015) 523–527
configurations. For cutoff angles below 551, the reduction is
particularly severe for south-facing directions (α ¼ 1801), actually
resulting in less power output than from east-facing and west-
facing directions. Hence, reducing or eliminating any angle-
dependent roll off with θ up to at least 701 is a reasonable target
for maximizing power output in a range of deployment conditions.
5. Conclusions
This work explores the impact of angle dependence on the
overall power output for thin film and traditional PVs in fixed
angle configurations. The power output of angle-dependent thin
film PVs was simulated at various tilt angles and orientations
based on integration of hourly illumination data and based on
demonstrated and idealized angle dependences for three repre-
sentative locations: Lansing, MI; Phoenix, AZ; and Meru, Kenya.
While this work does not explore these effects exhaustively, it is
shown that for several archetypal PVs with different angle-
dependent efficiencies, the overall power output can be enhanced
by 30% or more for designs with a reduced angle-dependent roll-
off, with the most substantial impact occurring for south-facing
modules with 901 tilt angle and lower latitude locations where
flux is greater. This work also shows that designing PVs with
limited drop in response up to at least 701 is desirable for
maximizing power output in a range of deployment conditions.
Considering a wide range of angular performance have been
demonstrated, this work highlights that improving the angle
dependence of traditional and thin film PVs is therefore a key
consideration in the design of the PVs for fixed orientation and
BIPV applications to maximize power outputs that is equivalent to
enhancing the PV efficiency directly.
Acknowledgment
The authors would like to thank Andrew Grossman for useful
discussions. Financial support for this work was provided in part
by the Infrastructure Planning and Facilities (IPF) through an Office
of Sustainability Seed Grant at Michigan State University and the
National Science Foundation (CAREER award, CBET-1254662).
References
[1] M. Oliver, T. Jackson, Energy and economic evaluation of building-integrated
photovoltaics, Energy 26 (2001) 431–439.
[2] G.P. Hammond, H.A. Harajli, C.I. Jones, A.B. Winnett, Whole systems appraisal
of a UK building integrated photovoltaic (BIPV) system: energy, environmen-
tal, and economic evaluations, Energy Policy 40 (2012) 219–230.
[3] M. Pagliaro, R. Ciriminna, G. Palmisano, BIPV: merging the photovoltaic with
the construction industry, Prog. Photovolt.: Res. Appl. 18 (2010) 61–72.
527
[4] T.P. Chang, The Sun's apparent position and the optimal tilt angle of a solar
collector in the northern hemisphere, Sol. Energy 83 (2009) 1274–1284.
[5] A.K. Yadav, S.S. Chandel, Tilt angle optimization to maximize incident solar
radiation: a review, Renew. Sustain. Energy Rev. 23 (2013) 503–513.
[6] B.v. der Wiel, H.-J. Egelhaaf, H. Issa, M. Roos, N. Henze, Market readiness of
organic photovoltaics for building integration, MRS Online Proc. Libr. 1639
(2014) (mrsf13-1639-y1610-1603).
[7] G. Dennler, K. Forberich, M.C. Scharber, C.J. Brabec, I. Tomiš, K. Hingerl,
T. Fromherz, Angle dependence of external and internal quantum efficiencies
in bulk-heterojunction organic solar cells, J. Appl. Phys. 102 (2007) 054516.
[8] S. Lee, I. Jeong, H.P. Kim, S.Y. Hwang, T.J. Kim, Y.D. Kim, J. Jang, J. Kim, Effect of
incidence angle and polarization on the optimized layer structure of organic
solar cells, Sol. Energy Mater. Sol. Cells 118 (2013) 9–17.
[9] N.-K. Persson, H. Arwin, O. Inganas, Optical optimization of polyfluorene-
fullerene blend photodiodes, J. Appl. Phys. 97 (2005) 034503.
[10] D. Cheyns, B.P. Rand, B. Verreet, J. Genoe, J. Poortmans, P. Heremans, The
angular response of ultrathin film organic solar cells, Appl. Phys. Lett. 92
(2008) 243310.
[11] L.A.A. Pettersson, L.S. Roman, O. Inganäs, Quantum efficiency of exciton-to-
charge generation in organic photovoltaic devices, J. Appl. Phys. 89 (2001)
5564–5569.
[12] C.-C. Chen, L. Dou, R. Zhu, C.-H. Chung, T.-B. Song, Y.B. Zheng, S. Hawks, G. Li, P.
S. Weiss, Y. Yang, Visibly transparent polymer solar cells produced by solution
processing, ACS Nano 6 (2012) 7185–7190.
[13] R.R. Lunt, V. Bulovic, Transparent, near-infrared organic photovoltaic solar
cells for window and energy-scavenging applications, Appl. Phys. Lett. 98
(2011) 113305.
[14] J.-H. Song, Y.-S. An, S.-G. Kim, S.-J. Lee, J.-H. Yoon, Y.-K. Choung, Power output
analysis of transparent thin-film module in building integrated photovoltaic
system (BIPV), Energy Build. 40 (2008) 2067–2075.
[15] M. Young, C.J. Traverse, R. Pandey, M.C. Barr, R.R. Lunt, Angle dependence of
transparent photovoltaics in conventional and optically inverted configura-
tions, Appl. Phys. Lett. 103 (2013) 133304.
[16] X. Yan, D.J. Poxson, J. Cho, R.E. Welser, A.K. Sood, J.K. Kim, E.F. Schubert,
Enhanced omnidirectional photovoltaic performance of solar cells using
multiple-discrete-layer tailored- and low-refractive index anti-reflection coat-
ings, Adv. Funct. Mater. 23 (2013) 583–590.
[17] S.-Y. Kuo, M.-Y. Hsieh, H.-V. Han, F.-I. Lai, Y.-L. Tsai, J.-F. Yang, T.-Y. Chuang, H.-
C. Kuo, Dandelion-shaped nanostructures for enhancing omnidirectional
photovoltaic performance, Nanoscale 5 (2013) 4270–4276.
[18] W. Marion, S. Wilcox, Users Manual for TMY3 Data Sets, National Renewable
Energy Laboratory, 2008, Accessible at /http://www.nrel.gov/docs/fy08osti/
43156.pdfS (accessed 04.30.13).
[19] S. Wilcox, National Solar Radiation Database 1991–2010, Update User's
Manual, 2012.
[20] J.E. Hay, Calculation of monthly mean solar radiation for horizontal and
inclined surfaces, Sol. Energy 23 (1979) 301–307.
[21] A.A. El-Sebaii, F.S. Al-Hazmi, A.A. Al-Ghamdi, S.J. Yaghmour, Global, direct and
diffuse solar radiation on horizontal and tilted surfaces in Jeddah, Saudi
Arabia, Appl. Energy 87 (2010) 568–576.
[22] National Renewable Energy Laboratory PVWatts Version 1, Accessible at:
〈http://rredc.nrel.gov/solar/calculators/pvwatts/version1/〉 (accessed 04.10.13).
[23] R.R. Lunt, T.P. Osedach, P.R. Brown, J.A. Rowehl, V. Bulović, Practical roadmap
and limits to nanostructured photovoltaics, Adv. Mater. 23 (2011) 5712–5727.
[24] V. Shrotriya, G. LiY. YaoT. Moriarty, K. Emery, Y. Yang, Accurate measurement
and characterization of organic solar cells, Adv. Funct. Mater. 16 (2006)
2016–2023.
[25] R. Tang, T. Wu, Optimal tilt-angles for solar collectors used in China, Appl.
Energy 79 (2004) 239–248.
[26] J. Kaldellis, K. Kavadias, D. Zafirakis, Experimental validation of the optimum
photovoltaic panels' tilt angle for remote consumers, Renew. Energy 46 (2012)
179–191.