Molecular and morphological influences on the open circuit voltages of organic photovoltaic devices

Article abstract of DOI:10.1021/ja9007722

We explore the dependence of the dark current of C₆₀-based organic photovoltaic (OPV) cells on molecular composition and the degree of intermolecular interaction of several molecular donor materials. The saturation dark current density, J_S, is an important factor in determining the open circuit voltage, V_oc. The V_oc values of OPVs show a strong inverse correlation with J_S. Donor materials that show evidence for aggregation in their thin-film absorption spectra and polycrystallinity in thin film X-ray diffraction result in a high dark current, and thus a low V_oc. In contrast, donor materials with structures that hinder intermolecular π-interaction give amorphous thin films and reduced values of J_S, relative to donors with strong intermolecular π-interactions, leading to a high V_oc. This work provides guidance for the design of materials and device architectures that maximize OPV cell power conversion efficiency.

Full text of DOI:10.1021/ja9007722

Published on Web 06/11/2009
Molecular and Morphological Influences on the Open Circuit
Voltages of Organic Photovoltaic Devices
M. Dolores Perez,† Carsten Borek,† Stephen R. Forrest,*,‡ and
Mark E. Thompson*,†
Department of Chemistry, UniVersity of Southern California, Los Angeles, California 90089, and
Departments of Electrical Engineering and Computer Science, Physics and Materials Science
and Engineering, UniVersity of Michigan, Ann Arbor, Michigan 48109
Received January 31, 2009; E-mail: met@usc.edu; stevefor@umich.edu
Abstract: We explore the dependence of the dark current of C60-based organic photovoltaic (OPV) cells
on molecular composition and the degree of intermolecular interaction of several molecular donor materials.
The saturation dark current density, JS, is an important factor in determining the open circuit voltage, Voc.
The Voc values of OPVs show a strong inverse correlation with JS. Donor materials that show evidence for
aggregation in their thin-film absorption spectra and polycrystallinity in thin film X-ray diffraction result in a
high dark current, and thus a low Voc. In contrast, donor materials with structures that hinder intermolecular
π-interaction give amorphous thin films and reduced values of JS, relative to donors with strong intermolecular
π-interactions, leading to a high Voc. This work provides guidance for the design of materials and device
architectures that maximize OPV cell power conversion efficiency.
Scheme 1
Introduction
The recent growing demand for cost-effective renewable
energy has generated considerable interest in the field of organic
solar cells. Organic photovoltaic (OPV) cells have the potential
to substantially lower costs for solar energy conversion, given
their expected low materials and fabrication costs. For OPVs
to make a meaningful contribution to meeting our overall energy
needs, however, their efficiencies need to be significantly
improved from current values that are in the range of 5-6%.1,2
This is compared to inorganic solar cell efficiencies of >20%.
A basic understanding of the processes that control and limit
the operation of solar cells is crucial to their improvement and
can serve as a guide for identification of new materials for high
performance devices. A key limitation in OPVs to date is their
low open circuit voltages (Voc), which is typically less than half
of the incident photon energy. Here we explore the relationship
between materials properties and the open circuit voltage.
OPVs harvest energy on the basis of exciton generation and
transport. Photogenerated excitons can form in either the electron
donor (D) or acceptor (A) layer in a bilayer cell, as shown in
Scheme 1. They subsequently diffuse to the D/A interface where
they dissociate to give a free hole and electron that are eventually
collected at the contacts. The solar power conversion efficiency
(η) is the ratio of the electrical power produced by the cell to
the incident power of the solar irradiation (Po). The power
conversion efficiency is the product of the short circuit current
(JSC), open circuit voltage (Voc), and the fill factor (FF),
following η ) JSCVocFF/Po. A number of factors affect the
photocurrent, including the light intensity, optical density of the
cell, the degree of overlap between the D and A layer absorption
spectra with that of the emission spectrum of the sun, the exciton
diffusion lengths in the D and A layers, and the charge
separation efficiency at the D/A interface. The FF is related to
the cell series resistance (Rs), while Voc depnds on the choice
of the D/A material combination, electrode material, light
intensity, and temperature of operation.3-7 Additionally, dif-
ferent donor and acceptor materials combinations influence Voc.
Specifically, several studies have demonstrated a dependence
of Voc on the energy difference between the highest occupied
molecular orbital (HOMO) energy of the donor and the lowest
unoccupied molecular orbital (LUMO) energy of the acceptor;
a difference called the D/A interface energy gap (∆EDA, Scheme
(3) Mihailetchi, V. D.; Blom, P. W. M.; Hummelen, J. C.; Rispens, M. T.
J. Appl. Phys. 2003, 94, 6849–6854.
† University of Southern California.
(4) Koster, L. J. A.; Mihailetchi, V. D.; Ramaker, R.; Blom, P. W. M.
Appl. Phys. Lett. 2005, 86, 123509.
‡ University of Michigan.
(1) Xue, J. G.; Uchida, S.; Rand, B. P.; Forrest, S. R. Appl. Phys. Lett.
2004, 84, 3013–3015.
(5) Mandoc, M. M.; Koster, L. J. A.; Blom, P. W. M. Appl. Phys. Lett.
2007, 90, 133504.
(2) Peet, J.; Kim, J. Y.; Coates, N. E.; Ma, W. L.; Moses, D.; Heeger,
A. J.; Bazan, G. C. Nat. Mater. 2007, 6, 497–500.
(6) Liu, J.; Y., S.; Yang, Y. AdV. Funct. Mater. 2001, 11, 420–424.
(7) Rand, B. P.; Burk, D. P.; Forrest, S. R. Phys. ReV. B 2007, 75, 115327.
9
10.1021/ja9007722 CCC: $40.75  2009 American Chemical Society
J. AM. CHEM. SOC. 2009, 131, 9281–9286 9281

A R T I C L E S
Perez et al.
1).7-14 However, under normal operating conditions (i.e., room
temperature and 1 sun intensity illumination) the experimental
values for Voc can differ from those inferred from ∆EDA for
some materials systems. This disparity is a result of the
electronic properties of the donor, acceptor and the D/A
interface. A more thorough understanding of molecular materials
properties that influence Voc is important to develop new OPV
materials that lead to a high Voc.
Here, we identify a correspondence between molecular
composition and crystal morphology and the open circuit voltage
in small-molecule-based OPVs. Data are presented for several
OPVs in which different donor materials are used along with a
common acceptor (i.e., C60) to highlight the most important
factors governing this relationship.
JSC
JS
nkT
q
Voc
≈
ln
(3)
(
)
Hence, at a given JSC a low dark current (JS) results in a
high Voc. Insight can be gained by examining the origin of JS,
which is the current resulting from carriers generated thermally
at the D/A interface, or originating within the film bulk. The
saturation current due to interface charge generation has been
shown to vary exponentially with ∆EDA, which is represented
by eq 4 for systems where JS is dominated by recombination
(n ≈ 2), as observed for most OPVs.7,15,16,19
-∆EDA
2nkT
JS ) JSO exp
(4)
(
)
Physical Origin of the Open Circuit Voltage
The factor of 2 accounts for the thermal generation of both an
electron and hole at the D/A heterointerface, giving an activation
energy of ∆EDA/2 for the process. The magnitude of JSO depends
on a number of materials properties that determine the carrier
generation/recombination rate, independent of the energy barrier,
∆EDA. Materials properties affecting the magnitude of JSO
include the reorganization energy for DfA electron transfer,
the intermolecular overlap at the D/A interface, the layer
electrical conductivities, the area of the D/A interface, and the
density of states at the HOMO and LUMO energies of the D
and A materials.
The generalized Shockley equation, eq 1,15,16 describes the
current density (J) vs voltage characteristics of organic solar
cells:
Rp
q(V - JRs)
V
Rp
J )
J exp
- 1 +
- Jph(V)
(
)
s
[
]
{
}
Rs + Rp
nkT
(1)
Here, Rp is the parallel resistance, JS is the saturation current
density, q is the fundamental charge, n is the diode ideality
factor, and Jph(V) is the voltage-dependent photocurrent density.
For solar cells with minimal leakage current (Rp . Rs), eq 1
can be simplified to
Substitution of eq 4 into eq 3 yields
JSC
JSO
∆EDA
2q
nkT
q
Voc
)
ln
+
(5)
(
)
q(V - JRs)
J ) J exp
- 1 - J (V)
(2)
(
)
Equation 5 suggests a linear dependence of Voc on the
interface energy gap and a logarithmic dependence on JSC and
light intensity (JSC ∝ P0), as experimentally observed for both
bulk heterojunctions3,20,21 and lamellar OPVs.7,22,23 A linear
correlation of Voc to ∆EDA has been reported previously for
several materials systems.7,13,14 Here, ∆EDA is the thermal
activation energy for charge separation at the D/A interface,
such that at low temperatures the measured Voc approaches
∆EDA/2, as recently reported by Rand, et al. for vapor deposited
OPVs.7 Note that eq 5 is similar to, but with important physical
implications different from that recently reported by Potscavage,
et al.24 The equation reported in their publication has two
different ideality factors (n and n′) and suggests a scaling of
Voc by ∆EDA/q, rather than ∆EDA/2q, in contrast to what has
been observed experimentally. In addition, Potscavage, et al.,
have assumed that JSO is constant for different D/A pairs, which
is inconsistent with the data presented below.
S
[
]
ph
nkT
The first term describes the thermally generated current that is
typically dominated by recombination at the D/A interface, and
the second term accounts for photogenerated carriers, Jph(V).
Under open circuit conditions (J ) 0, V ) Voc), the rate of carrier
recombination equals the rate of optical carrier generation.
Assuming open circuit conditions, a short circuit current of JSC
) Jph(0) . JS, and a low series resistance, eq 2 can be solved
to give eq 3, in which Voc is given by7,17,18
(8) Brabec, C. J.; Cravino, A.; Meissner, D.; Sariciftci, N. S.; Fromherz,
T.; Rispens, M. T.; Sanchez, L.; Hummelen, J. C. AdV. Funct. Mater.
2001, 11, 374–380.
(9) Gadisa, A.; Svensson, M.; Andersson, M. R.; Inganas, O. Appl. Phys.
Lett. 2004, 84, 1609–1611.
(10) Kooistra, F. B.; Knol, J.; Kastenberg, F.; Popescu, L. M.; Verhees,
W. J. H.; Kroon, J. M.; Hummelen, J. C. Org. Lett. 2007, 9, 551–
554.
In this work, we study the relationship of various materials
combinations to Voc as expressed by eq 5. In particular, we
explore the role of JSO in controlling the Voc, and what molecular
properties affect the magnitude of JSO. The intermolecular
overlap in the donor and acceptor layers and at the D/A interface,
(11) Mutolo, K. L.; Mayo, E. I.; Rand, B. P.; Forrest, S. R.; Thompson,
M. E. J. Am. Chem. Soc. 2006, 128, 8108–8109.
(12) Sarangerel, K.; Ganzorig, C.; Fujihira, M.; Sakomura, M.; Ueda, K.
Chem. Lett. 2008, 37, 778–779.
(13) Scharber, M. C.; Wuhlbacher, D.; Koppe, M.; Denk, P.; Waldauf, C.;
Heeger, A. J.; Brabec, C. L. AdV. Mater. 2006, 18, 789–xxx.
(14) Vandewal, K.; Gadisa, A.; Oosterbaan, W. D.; Bertho, S.; Banishoeib,
F.; Van Severen, I.; Lutsen, L.; Cleij, T. J.; Vanderzande, D.; V., M. J.
AdV. Funct. Mater. 2008, 18, 2064–2070.
(19) Wu¨rfel, P. Physics of Solar Cells: From Principles to New Concepts;
Wiley-VCH: Weinhem, 2005.
(15) Fahrenbruch, A. L.; Aranovich, J. Solar Energy ConVersionsSolid-
State Physics Aspects; Springer-Verlag: Berlin, Heidelberg, New York,
1979; Vol. 31.
(16) Bube, H. R.; Fahrenbruch, A. L. Book Title Academic: New York,
1981, p 163.
(20) Schilinsky, P.; Waldauf, C.; Brabec, C. J. Appl. Phys. Lett. 2002, 81,
3885.
(21) Riedel, I.; Parisi, J.; Dyakonov, V.; Lutsen, L.; Vanderzande, D.;
Hummelen, J. C. AdV. Funct. Mater. 2004, 14, 38.
(22) Peumans, P.; Yakimov, A.; Forrest, S. R. J. Appl. Phys. 2003, 93,
3693–3723.
(17) Li, N.; Lassiter, B. E.; Lunt, R. R.; Wei, G.; Forrest, S. R. Appl. Phys.
Lett. 2009, 94, 023307–3.
(23) Signerski, R. J. Non-Cryst. Sol. 2008, 354, 4465.
(24) Potscavage, J. W. J.; Yoo, S.; Kippelen, B. Appl. Phys. Lett. 2008,
93, 193308–3.
(18) Sze, S. M. Physics of Semiconductor DeVices, 2nd ed.; Wiley-
Interscience: New York, 1981.
9
9282 J. AM. CHEM. SOC. VOL. 131, NO. 26, 2009

Influences on Voltages of Organic Photovoltaic Devices
A R T I C L E S
Figure 1. Molecular structures of donors and the corresponding HOMO (blue) and LUMO (red) electron density distributions for tetracene, rubrene, CuPc,
and PtTPBP. Phenyl rings are shown as yellow space-filling surfaces.
Table 1. Dark and Light Characteristic Parameters for All Devices with the Structure ITO/Donor/C60 (400 Å)/BCP (100 Å)/Al (1000 Å)
donor (Thickness, Å)
JS (µA/cm2)
Voc (V)
JSC (mA/cm2)
FF
η (%)
n
SubPc (130)
rubrene (200)
NPD (100)
PtTPBP (150)
tetracene (600)
PtOEP (150)
PdTPBPa (150)
CuPc (400)
3.2 × 10-6
2.5 × 10-3
5.17 × 10-3
0.020
0.077
0.081
0.24
1.1
18
0.97
0.92
0.85
0.69
0.55
0.52
0.61
0.48
0.31
3.3611
2.35
2.35
4.5
1.8
1.6
2.41
5.5
4.2
0.5711
0.43
0.47
0.63
0.54
0.59
0.62
0.60
0.51
2.111
0.92
0.93
1.9
0.53
0.49
1.8
1.811
2.1
2.5
2.1
2.2
2.0
2.4
2.0
2.2
1.6
0.66
PtTPNP (120)
a Illumination intensity ) 0.5 suns.
as well as film morphology influence JSC by limiting the exciton
diffusion length, as well as by impacting the energetics and
kinetics of charge transfer at the interface, thus affecting JSO.
Clearly, to achieve the maximum possible Voc for a given D/A
pair, JSO must be minimized.
to those used here, utilizing rubrene,25 tetracene,26 NPD,7 and
CuPc22 donors have been reported. The Voc, JSC, FF, and η
values presented in Table 1 are consistent with the literature
reports (JS values were not reported). Values of JSO calculated
using eq 4 are listed in Table 2. The interface gap, ∆EDA, was
calculated using the C60 LUMO value (3.5 eV) measured by
inverse photoelectron spectroscopy (IPES),27 and the HOMOs
of the donors were obtained both from ultraviolet photoelectron
spectroscopy (UPS) and electrochemistry.28 The calculated Voc
values according to eq 5 are also provided in Table 2, along
Results and Discussion
A number of photovoltaic devices with the structure ITO/
donor/C60 (400 Å)/bathocuproine (BCP) (100 Å)/Al (1000 Å),
using the donor materials in Figure 1, were fabricated, and both
the light and dark performance characteristics were studied (see
Experimental Section).
The dark current J-V characteristics were fit to eq 1 to extract
n and JS (Table 1). The parameters, JSC, Voc, FF, and η, were
obtained under simulated 1 sun, AM1.5 illumination conditions,
with results provided in Table 1, combined with values taken
from the literature (for SubPc11). OPVs with the structure similar
(25) Taima, T.; Sakai, J.; Yamanari, T.; Saito, K. Jpn. J. Appl. Phys., Part
2: Lett. Express Lett. 2006, 45, L995–L997.
(26) Chu, C. W.; Shao, Y.; Shrotriya, V.; Yang, Y. Appl. Phys. Lett. 2005,
86, 243506.
(27) Schwedhelm, R.; Kipp, L.; Dallmeyer, A.; Skibowski, M. Phys. ReV.
B 1998, 58, 13176.
(28) D’Andrade, B. W.; Datta, S.; Forrest, S. R.; Djurovich, P.; Polikarpov,
E.; Thompson, M. E. Org. Electron. 2005, 6, 11–20.
(29) Sato, N.; Seki, K.; Inokuchi, H. J. Chem. Soc., Faraday Trans. II 1981,
77, 1621–1633.
9
J. AM. CHEM. SOC. VOL. 131, NO. 26, 2009 9283

A R T I C L E S
Perez et al.
Table 2. Comparison of Measured and Calculated Performance Parametersa
donor
JSO
film
(nkT)/(q) ln ((Jsc)/(Jso))
(∆EDA)/(2q)
Voc calcd
Voc exptl
SubPc
rubrene
NPD
PtTPBP
tetracene
PtOEP
PdTPBPb
CuPc
5.5
0.43
11
12
Am
Am
Am
Am
PC
PC
-
-0.02
0.09
1.0
0.9
0.95
0.7
0.80
0.9
0.65
0.85
0.45
0.98
0.99
0.85
0.65
0.56
0.50
0.57
0.44
0.30
0.97
0.92
0.85
0.69
0.55
0.52
0.61
0.48
0.31
-0.10
-0.05
-0.24
-0.40
-0.082
-0.41
-0.15
150
5.1 × 103
9.1
1.5 × 104
62
PC
-
PtTPNP
a JSO are in mA/cm2, and the remaining data are in V. “Film” refers to the morphology of the donor thin film, as determined by X-ray diffraction,
Am ) amorphous, PC ) polycrystalline.17,30-36 b Illumination intensity ) 0.5 suns.
film spectra, strong intermolecular π interactions in tetracene
thin films yield broadened and red-shifted spectrum relative to
solution. Weak intermolecular interactions in rubrene decrease
both the D/D and particularly D/A interactions relative to
tetracene. The weaker D/A interaction in the rubrene-based OPV
leads to a correspondingly lower JSO.
The Voc measured for copper phthalocyanine (CuPc)- and
platinum tetraphenylbenzoporphyrin (PtTPBP)-containing de-
vices also shows a strong dependence on intermolecular
interactions, and thus JSO. The two compounds have comparable
π-systems, both with 38 π-electrons. While a larger Voc would
be expected for the CuPc-based device due to its higher ∆EDA
(∆EDA ) 1.7 eV, vs ∆EDA ) 1.4 eV for PtTPBP/C60), the
experimental CuPc/C60 Voc is actually less than that of PtTPBP
by 0.21 V (cf. Table 1). Note that CuPc is a flat molecule with
strong intermolecular interactions in the solid state. In contrast,
PtTPBP has a highly distorted saddle-shaped conformation with
four orthogonal phenyl rings that reduce the availability of the
π-system to intermolecular interactions and electronic delocal-
ization.37 CuPc shows marked broadening of its low-energy
absorption in its thin film spectra,38,39 while the solution and
thin film spectra of PtTPBP are the same line width.30 Added
steric bulk, through appended phenyl rings, decreases the degree
of intermolecular interaction for PtTPBP relative to CuPc. The
high Voc ) 0.69 V for the PtTPBP device results from
diminished recombination (Figure 3 and Table 2), and hence a
lower JSO than for the CuPc device (JSO ) 12 and 1.5 × 104
mA/cm2, respectively).
We have also examined PdTPBP as a donor material. The
structure of the Pd analogue is close to that of PtTPBP, where
the Pd substitution shifts the HOMO energy slightly, although
it does not alter the strengths of the intermolecular interactions.
Hence, the Pd and Pt complexes give similar JSO of 9 and 12
mA/cm2, respectively.
Electronic interactions, or coupling between donor and
acceptor molecules, can be modified by varying the strength of
the π-conjugation. To explore these effects, we prepared Pt
tetraphenyl naphtholporphyrin (PtTPNP, Figure 1), an analogue
of PtTPBP whose π-system is extended by benzanulation of
four rings onto the benzo moieties of PtTPBP. The lowest energy
conformation of the naphthol analog has a saddle shape similar
to PtTPBP, resulting from steric repulsions of the meso and
pyrolle substituents. The current-voltage characteristics of the
Figure 2. Current density vs voltage characteristics in the dark (closed
circles) and illumination under 1 sun, AM1.5 (open circles) conditions for
tetracene- and rubrene-based OPVs.
with the measured values for comparison. The calculated Voc
closely match those obtained experimentally.
To determine the molecular properties that lead to minimizing
JSO for a given ∆EDA, we first consider the donors, tetracene,
and rubrene, with results shown in Figure 2 and Table 1. Despite
the similarity of the HOMO energies of 5.1 eV for tetracene
and 5.3 eV for rubrene,29 the Voc for the two cells are con-
siderably different at 0.55 and 0.92 V, respectively. The
difference in Voc is due to the large difference in the current-
dependent terms of eq 5. That is, JSO is 2 orders of magnitude
larger for tetracene than for rubrene (see Table 2).
Whereas tetracene is a planar molecule consisting of fused
conjugated aromatic rings, the four nearly orthogonal pendant
phenyl rings of rubrene prevent close association of the tetracene
cores of adjacent rubrene molecules or between the tetracene
core and the C60 acceptor (see Figure 1), as is evident from
comparisons between the solution and thin-film spectra of the
two compounds. While rubrene gives similar solution and thin-
(30) Perez, M. D.; Borek, C.; Djurovich, P. I.; Mayo, E. I.; Lunt, R. R.;
Forrest, S. R.; Thompson, M. E. Adv. Mater. 2008.
(31) Gundlach, D. J.; Nichols, J. A.; Zhou, L.; Jackson, T. N. Appl. Phys.
Lett. 2002, 80, 2925–2927.
(32) Mattheus, C. C.; Michaelis, W.; Kelting, C.; Durfee, W. S.; Wo¨hrle,
D.; Schlettwein, D. Synth. Met. 2004, 146, 335–339.
(33) Mori, T.; Oda, S.; Ooishi, N.; Masumoto, Y. Jpn. J. Appl. Phys., Part
1: Regul. Pap. Brief Commun. ReV. Pap. 2007, 46, 5954–5959.
(34) Seo, S.; Park, B.-N.; Evans, P. G. Appl. Phys. Lett. 2006, 88, 232114–
3.
(35) Weinberg-Wolf, J. R.; McNeil, L. E.; Liu, S.; Kloc, C. J. Phys.:
Condens. Matter 2007, 19, 276204.
(36) Leznoff, C. C.; Lever, A. B. P. Phthalocyanines: properties and
applications; VCH: New York, 1989.
(37) Borek, C.; Hanson, K.; Djurovich, P. I.; Thompson, M. E.; Aznavour,
K.; Bau, R.; Sun, Y.; Forrest, S. R.; Brooks, J.; Michalski, L.; Brown,
J. Angew. Chem., Int. Ed. 2007, 46, 1109–1112.
(38) Brown, R. J. C.; Kucernak, A. R.; Long, N. J.; Mongay-Batalla, C.
New J. Chem. 2004, 28, 676–680.
(39) Ferreira, J. A.; Barral, R.; Baptista, J. D.; Ferreira, M. I. C. J. Lumin.
1991, 48-49, 385–390.
9
9284 J. AM. CHEM. SOC. VOL. 131, NO. 26, 2009

Influences on Voltages of Organic Photovoltaic Devices
A R T I C L E S
recombination at the D/A interface. This gives JSO ) 5.5 mA/
cm2, which is sufficiently small that it does not significantly
influence Voc, as determined from the interface energy gap.
SubPc is an example of a donor molecule that comprises features
that maximize Voc, a strong absorptivity in the visible, and a
high ∆EDA resulting in a low JSO.
The device architecture for the OPVs studied here, i.e., ITO/
Donor/C60/BCP/Al, was designed to focus on changes in the
donor and at the D/A interface. The acceptor, blocking layer,
and cathode materials are kept constant throughout our study.
However, as different donors are inserted into the structure, both
the D/A and ITO/donor interfaces are affected. While both
interfaces could contribute to the observed Voc, the D/A interface
dominates in the devices studied here. This was shown by
preparing two different devices with a PtTPBP donor, one with
the donor in direct contact with ITO and the other with a thin
NPD film between ITO and PtTPBP (i.e., ITO/NPD(x Å)/
PtTPBP(150 Å)/C60/BCP/Al, x ) 0, 50). The HOMO level of
NPD is 0.5 V below that of PtTPBP. If the ITO/donor interface
energetics contributed to the Voc in these devices, a marked shift
in the Voc values of the two device is expected. The two devices,
with and without the NPD layer, give the same Voc of 0.69 V,
supporting our assumption that dark current is limited by
recombination at the D/A interface.
X-ray diffraction studies of the donor materials investigated
here indicate that both amorphous and polycrystalline thin films
are formed, Table 2.17,30-36 Materials that show evidence of
aggregation in thin film absorption spectra, i.e., tetracene,
PtOEP, and CuPc, result in polycrystalline thin films. Materials
that show little or no evidence of aggregation in their thin film
absorption spectra, i.e., SubPc, rubrene, NPD, and PtTPBP, form
amorphous thin films. While the strength of intermolecular
interactions, via π-system overlap, is only one factor in
controlling whether a given material will form an amorphous
or polycrystalline film, it is useful to examine the correlation
between thin film morphology and JSO. There is a clear
distinction in JSO values between the amorphous (JSO ) 1.1 -
12 mA/cm2) and polycrystalline materials (JSO ) 150-1.5 ×
104 mA/cm2). The π-π stacking forces that encourage crystal
growth likely also leads to strong donor/C60 interactions,
contributing to the high JSO values observed for the polycrys-
talline materials.
Figure 3. Current density vs voltage characteristics in the dark (closed
circles) and illumination under 1 sun, AM1.5 (open circles) conditions for
CuPc-, PtTPBP-, and PtTPNP-based OPVs.
PtTPNP donor-containing devices are shown in Figure 3, along
with data for PtTPBP- and CuPc-based devices, with their
performance provided in Tables 1 and 2. While the PtTPNP-
and PtTPBP-based cells have similar photocurrent densities, Voc
is reduced to 0.31 V for the PtTPNP device. This low Voc is
due to (i) increased π-conjugation of PtTPNP, which reduces
the HOMO energy compared to either PtTPBP or CuPc
(PtTPNP HOMO energy ) 4.4 eV). Hence, the PtTPNP/C60
cell has ∆EDA ) 0.8 eV, markedly lower than for either PtTPBP
or CuPc. (ii) JSO is increased by a factor of 5 for PtTPBP (JSO
) 12 mA/cm2, compared to 62 mA/cm2 for PtTPNP). The
benzannulated rings added to PtTPNP extend beyond the meso-
phenyl groups, increasing the π-π overlap, and thus intermo-
lecular interactions. The resulting increase in JSO, along with
its decreased ∆EDA, significantly decrease Voc for this device
relative to PtTPBP. While π-expansion leads to an increased
JSO relative to PtTPBP, the JSO for PtTPNP is more than 200
times lower than that of CuPc.
To explore planar donors with π-systems of different sizes,
Pt octaethylporphyrin (PtOEP) and CuPc were compared (Figure
1).26 The CuPc π-system is significantly larger than that of
PtOEP, which have 38 and 22 π-electrons, respectively. As
observed for CuPc, the thin film absorption spectrum of PtOEP
shows significant broadening compared to its solution spectrum,
which was attributed to aggregate formation.40 The JSO of CuPc
is correspondingly higher than that of PtOEP (JSO ) 1.5 × 104
and 5.1 × 103 mA/cm2, respectively), but the difference is not
nearly as great as those observed above, consistent with
significant steric blocking of the meso phenyls of PtTPBP.
The analysis can be applied to other previously reported donor
molecules. For example, N,N′-bis(naphthalen-1-yl)-N,N′-bis(phe-
Conclusions
The correspondence of the calculated values for Voc obtained
using eq 5 to experimental values for all D/A materials
combinations examined provides evidence for the dependence
of the saturation dark current in organic PV cells on the
molecular and thin-film structures. Specifically, we show a
correspondence between the strength of intermolecular interac-
tions and the saturation current density, reflected in the pre-
exponential dark current term, JSO. Weak intermolecular inter-
actions lead to low JSO, and hence correspondingly high open
circuit voltages. While not explored in this study, it is reasonable
to expect that a similar relationship will be observed for acceptor
materials.
It is expected that changes in the OPV architecture will also
influence the magnitude of JSO. For example, Li, et al.,
introduces an electron-blocking layer between the OPV anode
contact and donor layer to significantly reduce electron leakage,
increasing the Voc by a factor of 2 over analogous devices
without the blocking layer.17 This approach may allow the use
of materials with strong intermolecular overlap leading to
nyl)benzidine (NPD)-based devices (see Figure 1) give Voc
)
0.85 V,7 consistent with ∆EDA ) 1.9 eV and reduced recom-
bination losses as reflected by the small JSO )11 mA/cm2. This
value of JSO is similar to that of PtTPBP, indicative of reduced
π accessibility and weak interactions with the C60 acceptor, due
to lose intermolecular packing of the nonplanar NPD molecules.
Furthermore, subphthalocyanine (SubPc) exhibits a larger
HOMO energy than CuPc, resulting in ∆EDA ) 1.9 eV, leading
to a high Voc ) 0.97 V.11 SubPc has an out-of-plane cone shape
whose steric effects reduce intermolecular interactions, and thus
(40) Kalinowski, J.; Stampor, W.; Szmytkowski, J.; Cocchi, M.; Virgil,
D.; Fattori, V.; Marco, P. D. J. Chem. Phys. 2005, 122, 154710.
9
J. AM. CHEM. SOC. VOL. 131, NO. 26, 2009 9285

A R T I C L E S
Perez et al.
by sublimation prior to use. The cathode metal, Al (99.999%), was
used as received (from Alrich Chem. Co.).
comparatively high JSO, without a marked reduction in open
circuit voltage.
Photovoltaic cells were grown on solvent-cleaned 150 nm ITO-
coated glass substrates with a sheet resistance of 20 ( 5 Ω/sqare,
as described elsewhere.41 The substrates were exposed to UV-ozone
for 10 min immediately prior to loading into a high vacuum (1-3
× 10-6 Torr) chamber. Materials were sequentially grown by
vacuum thermal evaporation at the following rates: metal-TPBP
and PtTPNP (1 Å/s); CuPc, PtOEP, rubrene, NPD, C60, BCP (2
Å/s); and tetracene (15 Å/s). The metal cathodes were evaporated
throughashadowmaskwith1mmdiameteropenings.Current-voltage
characteristics of the cells were measured in the dark and under
simulated AM1.5G solar illumination using a Keithley 2420 3A
Source Meter. Incident power was adjusted using a calibrated Si
photodiode to match 1 sun intensity (100 mW/cm2). Spectral
mismatch was calculated to correct the measured efficiencies
following standard procedures.42 Dark current characteristics were
fit to eq 2 to obtain JS, n, and Rs. JSO was calculated using the
values for ∆EDA from the literature.
To achieve a high open circuit voltage, we find that materials
combinations with both a high ∆EDA and low JSO are required.
Materials that show weak intermolecular interactions in thin
films often result in poor carrier and exciton transport. Since
weak intermolecular interactions are also found to result in a
low JSO, materials optimized for low dark currents will generally
also have short exciton diffusion lengths and poor carrier
conductivity, leading to a correspondingly low JSC and FF. These
contradictory characteristics would appear to provide a funda-
mental limit to the ability for a single materials pair to achieve
an optimally high Voc, as well as high JSC and FF, needed for
high power efficiency. Fortunately, a correlation between low
JSO and low JSC and FF is not universaly observed. High power
conversion efficiency is possible for OPV materials with
sterically inaccessible π-systems or reduced π-extension (low
JSO), which have significant overlap between intense absorption
bands of the materials and the solar spectrum (high JSC) and
good hole conductivity (high FF). Rubrene and PtTPBP donors
are examples of such materials, giving relatively high JSC and
FF and low JSO. Our work, therefore, can serve as a guide for
the molecular design of both D and A materials with improved
spectral overlap, while achieving both a low JSO and a high
∆EDA, giving a high Voc.
Acknowledgment. This work was supported by Global Photonic
Energy Corporation, the Air Force Office of Scientific Research,
and the Department of Energy. The authors acknowledge Thin Film
Devices, Inc (www.tfdinc.com) for providing ITO.
Supporting Information Available: Details of the PtTPNP
synthetic procedure, optical properties (absorption and emission
spectra) and electrochemistry of PtTPNP. Spectral response of
devices using PtTPNP as donor. This material is available free
of charge via the Internet at http://pubs.acs.org.
Experimental Section
The synthesis and characterization of PtTPNP are given in the
Supporting Information. PtTPBP and PdTPBP were synthesized
according to literature procedures, and purified by vacuum thermal
gradient sublimation.11,30 Copper phthalocyanine (CuPc, 99% pure),
PtOEP, tetracene (98%), rubrene (98%), C60 (99.5%), and 2,9-
dimethyl-4,7-diphenyl-1,10-phenanthroline (bathocuproine, BCP,
96%) were purchased from the Aldrich Chemical Co. and purified
JA9007722
(41) Burrows, P. E.; Shen, Z.; Bulovic, V.; McCarty, D. M.; Forrest, S. R.;
Cronin, J. A.; Thompson, M. E. J. Appl. Phys. 1996, 79, 7991–8006.
(42) Shrotriya, V.; Li, G.; Yao, Y.; Moriarty, T.; Emery, K.; Yang, Y.
AdV. Funct. Mater. 2006, 16, 2016–2023.
9
9286 J. AM. CHEM. SOC. VOL. 131, NO. 26, 2009