FULL PAPER
©
2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1
wileyonlinelibrary.com
1. Introduction
The growing demand for renewable
energy has brought solar cells to the
forefront as a potential energy harvester.
Consequently, organic–inorganic metal
halide perovskite solar cells have gar-
nered profound research interest in recent
years as next-generation photovoltaics.
[ 1 ]
This importance stems largely from their
exceptional charge carrier mobilities, and
optical and electronic properties that are
tunable by varying the chemical compo-
sition.
[ 2,3 ]
Recently, perovskite-based solar
cells have demonstrated an impressive
power conversion effi ciency (PCE) of over
20%.
[ 4 ]
In most of the practical applica-
tions of solar cells, besides the impor-
tance of short-circuit current density,
high open-circuit voltage ( V
oc
) is an indis-
pensable parameter to attain both high
power conversion effi ciency with reduced
series resistance losses and to drive elec-
trochemical reactions including water-
splitting reactions and CO
2
reduction.
[ 5,6 ]
In principle, a high open-circuit voltage
in a perovskite solar cell can be attained
by aligning the energy levels of absorber
and charge extraction layers.
[ 7 ]
Despite having enticing benefi t
of higher bandgap for pure bromide-based perovskite material
Photovoltaic and Amplifi ed Spontaneous Emission
Studies of High-Quality Formamidinium Lead Bromide
Perovskite Films
Neha Arora , M. Ibrahim Dar , * Mahmoud Hezam , Wolfgang Tress , Gwénolé Jacopin ,
Thomas Moehl , Peng Gao , Abdulah Saleh Aldwayyan , Benoit Deveaud , Michael Grätzel ,
and Mohammad Khaja Nazeeruddin *
This study demonstrates the formation of extremely smooth and uniform
formamidinium lead bromide (CH(NH
2
)
2
PbBr
3
= FAPbBr
3
) fi lms using
an optimum mixture of dimethyl sulfoxide and N , N -dimethylformamide
solvents. Surface morphology and phase purity of the FAPbBr
3
fi lms are
thoroughly examined by fi eld emission scanning electron microscopy and
powder X-ray diffraction, respectively. To unravel the photophysical properties
of these fi lms, systematic investigation based on time-integrated and time-
dependent photoluminescence studies are carried out which, respectively,
bring out relatively lower nonradiative recombination rates and long lasting
photogenerated charge carriers in FAPbBr
3
perovskite fi lms. The devices
based on FTO/TiO
2
/FAPbBr
3
/spiro-OMeTAD/Au show highly reproducible
open-circuit voltage ( V
oc
) of 1.42 V, a record for FAPbBr
3
-based perovskite
solar cells. V
oc
as a function of illumination intensity indicates that the
contacts are very selective and higher V
oc
values are expected to be achieved
when the quality of the FAPbBr
3
fi lm is further improved. Overall, the devices
based on these fi lms reveal appreciable power conversion effi ciency of
7% under standard illumination conditions with negligible hysteresis. Finally,
the amplifi ed spontaneous emission (ASE) behavior explored in a cavity-free
confi guration for FAPbBr
3
perovskite fi lms shows a sharp ASE threshold at a
fl uence of 190 µJ cm
−2
with high quantum effi ciency further confi rming the
high quality of the fi lms.
DOI: 10.1002/adfm.201504977
Dr. N. Arora, Dr. M. I. Dar, Dr. P. Gao,
Prof. M. K. Nazeeruddin
Group for Molecular Engineering of Functional
Materials
Institute of Chemical Sciences and Engineering
École Polytechnique Fédérale de Lausanne
Lausanne CH-1015 , Switzerland
E-mail: ibrahim.dar@epfl .ch; mdkhaja.nazeeruddin@epfl .ch
Dr. M. I. Dar, Dr. W. Tress, Dr. T. Moehl,
Prof. M. Grätzel
Laboratory of Photonics and Interfaces
Institute of Chemical Sciences and Engineering
École Polytechnique Fédérale de Lausanne
Lausanne CH-1015 , Switzerland
M. Hezam
King Abdullah Institute for Nanotechnology
King Saud University
Riyadh 11451 , Saudi Arabia
M. Hezam, Dr. G. Jacopin, Prof. B. Deveaud
Laboratory of Quantum Optoelectronics
Institute of Physics
École Polytechnique Fédérale de Lausanne
Lausanne CH-1015 , Switzerland
Prof. A. S. Aldwayyan
Physics and Astronomy Dept, Photonics Lab
College of Science
King Saud University
Riyadh 11451 , Saudi Arabia
Adv. Funct. Mater. 2016,
DOI: 10.1002/adfm.201504977
www.afm-journal.de
www.MaterialsViews.com
FULL PAPER
2
wileyonlinelibrary.com
©
2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
which indeed can also be used as a top cell in tandem solar
cells,
[ 8 ]
much of the research mainly focused on pure iodide
(CH
3
NH
3
PbI
3
), mixed halide (CH
3
NH
3
PbI
3−
x
Br
x
) or double-
mixed ((CH(NH
2
)
2
PbI
3
)
1−
x
(CH
3
NH
3
PbBr
3
)
x
) perovskites.
[ 4 ]
Nevertheless, promising studies on effi cient methylammo-
nium lead bromide (CH
3
NH
3
PbBr
3
)-based device architecture
showing enhanced photovoltage in comparison to pure iodide-
based perovskite have been documented. Kojima et al. reported
on employing CH
3
NH
3
PbBr
3
as visible-light sensitizers in pho-
toelectrochemical cells exhibiting a promising photovoltage of
0.96 V with a power conversion effi ciency of 3.1%.
[ 9 ]
Later on,
Edri et al. demonstrated a high V
oc
of 1.3 V based on alumina\
CH
3
NH
3
PbBr
3
\ N , N ′-dialkyl perylenediimide (PDI) solar cell
structure, however the effi ciency of the devices was quite low
(< 1%).
[ 10 ]
Recently, by controlling the crystallization process
of CH
3
NH
3
PbBr
3
as well as by tailoring the highest occupied
molecular orbital (HOMO) levels of different hole-transporting
materials (HTMs), a V
oc
of 1.51 V has been reported.
[ 11 ]
So far, most of the studies on bromide perovskites are
devoted to CH
3
NH
3
PbBr
3
, however, the focus of this work
constitutes formamidinium lead bromide (CH(NH
2
)
2
PbBr
3
=
FAPbBr
3
) perovskite which has been studied scantly. Earlier,
Hanusch et al. described the fabrication of planar heterojunc-
tion FAPbBr
3
perovskite solar cells with effi ciency of over
6%,
[ 12 ]
however, the device exhibited severe hysteresis which
leads to an overestimation of power conversion effi ciency.
Mostly the presence of mesoporous TiO
2
scaffold has proven
to be advantageous in reducing the scan speed dependent hys-
teresis observed in current–voltage ( J – V ) curves in perovskite
solar cells.
[ 13,14 ]
Tuning the growth of perovskite structures has
been one of the critical parameters to fabricate high-effi ciency
solar cells. To that end, we investigated the role of solvent on
the growth and morphology of FAPbBr
3
perovskite fi lms, which
eventually led to the enhancement in the photovoltage to a
remarkable extent. We fabricated FAPbBr
3
fi lms employing the
sequential deposition method involving a mixture of solvents,
N , N -dimethylformamide and dimethyl sulfoxide(DMF+DMSO)
for the PbBr
2
precursor solution. Employing spiro-OMeTAD as
a hole-transporting material led us to fabricate FAPbBr
3
per-
ovskite devices with a J
sc
of 6.8 mA cm
−2
, FF of 72%, and V
oc
of 1.42 V resulting in an overall effi ciency of 7.0%. It is to be
noted that these devices exhibited negligible hysteresis. The
impact of the solvent on structural, morphological, and various
photophysical properties which dictate the photovoltaic perfor-
mance of a device was explored by X-ray diffraction (XRD), fi eld
emission scanning electron microscopy (FESEM), UV–visible
absorption, low temperature photoluminescence (PL), and
charge carrier lifetime studies.
Additionally, amplifi ed spontaneous emission (ASE) was
recorded from the FAPbBr
3
perovskite fi lms. Recently, there
has been a report in which formamidinium lead bromide per-
ovskite nanoparticles were incorporated into light-emitting
electrochemical cells.
[ 15 ]
In the literature various studies have
demonstrated the enormous potential of perovskite materials
as optical gain mediums even with a cavity-free confi gura-
tion.
[ 16,17 ]
The low nonradiative recombination rate, long dif-
fusion length, and high mobility of charge carriers reported
for perovskites even with high defect trap densities, are key
factors for their unprecedented progress in photovoltaics.
[ 18–20 ]
In addition to slow Auger recombination, the various possibili-
ties of fabricating them in different cavity confi gurations make
them suitable for an effi cient semiconductor laser.
[ 21–23 ]
In this
work, a sharp transition from spontaneous emission (SE) to
ASE at a pump fl uence of 190 µJ cm
−2
per pulse was observed
in FAPbBr
3
perovskite fi lms which corresponds to an ASE
threshold carrier density of ≈2.3 × 10
18
cm
−3
. Although ASE
has been reported for both CH
3
NH
3
PbI
3
and CH
3
NH
3
PbBr
3
perovskites in different resonating and cavity-free confi gura-
tions as well as in random networks of nanocrystals,
[ 16,17,21–23 ]
to the best of our knowledge, there has been no report till date
demonstrating the light gain applications of such FAPbBr
3
perovskite fi lms.
2. Results and Discussion
Two different FAPbBr
3
perovskite fi lms were deposited using
the sequential deposition approach. Typically, a solution of
PbBr
2
(DMF) or PbBr
2
(DMF+DMSO) was spin casted onto
a mesoporous TiO
2
scaffold. Subsequently the resulting
fi lms were converted into FAPbBr
3
perovskite fi lms by dip-
ping into isopropanol solution of formamidinium bromide
(CH(NH
2
)
2
Br), which were labeled as FA(1) (obtained from
PbBr
2
(DMF)) and FA(2) (obtained from PbBr
2
(DMF+DMSO)).
(See the Experimental Section for more details.)
2.1. Structural and Morphological Characterization
To evaluate the crystallinity of FA(1) and FA(2) perovskite fi lms,
X-ray diffraction was performed at room temperature. The
XRD patterns (Figure S1, Supporting Information) obtained
from the fi lms could be indexed to the cubic phase of FAPbBr
3
with a space group Pm– 3 m , which is consistent with previous
reports.
[ 12,15 ]
The morphology of lead bromide and FAPbBr
3
fi lms was
probed by FESEM. From SEM analysis, it is evident that cov-
erage of the lead bromide fi lms obtained from PbBr
2
(DMF)
( Figure 1 a) solution is relatively poor whereas the deposi-
tion of PbBr
2
(DMF+DMSO) solution led to the formation of
a uniform lead bromide fi lm (Figure 1 b). Furthermore, top
view SEM micrograph of FA(1) perovskite sample reveals the
formation of an inhomogeneous fi lm composed of aggregated
FAPbBr
3
structures (Figure 1 c). On the contrary, extremely uni-
form and continuous FAPbBr
3
fi lms with full surface coverage
were obtained in case of FA(2) , as shown in Figure 1 d. From a
marked difference in the surface morphology of lead bromide
and FAPbBr
3
fi lms for FA(1) and FA(2) samples, we establish
that the nature of the solvent employed for PbBr
2
precursor
solution can considerably improve the surface coverage of the
fi lms, which ostensibly could be benefi cial for the fabrication of
effi cient devices.
[ 13 ]
The device fabrication was completed by spin coating spiro-
OMeTAD as the hole-transporting material on FAPbBr
3
perovs-
kite layer, followed by thermal evaporation of a 70 nm thick gold
layer as a back contact. The cross-section SEM micrograph of
FA(1) (Figure 1 e) shows the formation of a nonuniform FAPbBr
3
fi lm of ≈500 nm thickness (including mesoporous TiO
2
).
Adv. Funct. Mater. 2016,
DOI: 10.1002/adfm.201504977
www.afm-journal.de
www.MaterialsViews.com
FULL PAPER
3
wileyonlinelibrary.com
©
2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
From the top view and cross-section SEM analysis, parasitic con-
tacts between organic conductor (HTM) and the mesoporous
TiO
2
layer as well as between pillared perovskite layer and top
gold contact are observed. Expectedly, a mixture of solvents
(DMF+DMSO) yields smoother and continuous FAPbBr
3
fi lm
with a uniform thickness of ≈300 nm, which is homogeneously
covered with spiro-OMeTAD (Figure 1 f).
From a detailed SEM analysis, we noticed that using a mix-
ture of solvents (DMF+DMSO) for lead bromide precursor
solution resulted in the formation of smoother fi lms.
[ 24–26 ]
In
addition to enhancement in solubility, mixed solvents possibly
slow down the growth of lead halide structures during spin
coating, yielding a uniform fi lm (Figure 1 b).
[ 27 ]
Low boiling
solvent (DMF) evaporates rapidly and eventually enhances the
loading (Figure 1 e) whereas because of slower evaporation of
high boiling solvent (DMSO), discontinuous and less loaded
lead bromide and FAPbBr
3
fi lms are formed, as evident from
SEM and UV–visible spectrum (Figures S2,S3, Supporting
Information). Smoother FAPbBr
3
fi lm with an optimum thick-
ness could warrant the fabrication of effi cient perovskite solar
cells. Therefore, to tailor the formation and loading of FAPbBr
3
structures, it became imperative to use a judicious ratio of
low and high boiling solvents. As is known DMSO solvent
has a tendency to form an adduct with lead iodide or it could
pronounce the amorphous nature of lead
iodide fi lms.
[ 27,28 ]
Using DMSO–lead iodide
adduct, the fabrication of high-effi ciency
solar cells has been reported recently.
[ 4,28 ]
To
understand the role of DMSO further, we
recorded XRD patterns of FA(1) and FA(2)
lead bromide fi lms. Variation in the crystal
structure of lead bromide fi lms deposited
from a single and a stoichiometric mixture of
DMF+DMSO solvent (Figure S4, Supporting
Information) was marginal which is in con-
trast to what has been observed in lead iodide
fi lms.
[ 27 ]
2.2. Photovoltaic Properties of Formami-
dinium Perovskite Devices
Tailoring the growth and morphology of per-
ovskite layer is one of the critical factors that
led to the evolution of high-effi ciency per-
ovskite solar cells. Herein, we investigated
the effect of morphology of FAPbBr
3
fi lms
on the performance of the resulting devices.
The current–voltage curves of the FA(1) and
FA(2) bromide perovskite devices are shown
in Figure 2 and the extracted photovoltaic
parameters are summarized in Table 1 .
Under illumination of 100 mW cm
−2
, the
device based on FA(1) fi lm (FAPbBr
3
fi lm
involving PbBr
2
(DMF) precursor solution),
exhibits a high short-circuit current density
( J
sc
) of 7.1 mA cm
−2
, a modest open-circuit
voltage V
oc
of 1.14 V, and a fi ll factor (FF) of
0.67, resulting in an overall power conver-
sion effi ciency (
η
) of 5.4% (Figure 2 a). By contrast, FA(2) device
exhibits similar short-circuit current density ( J
sc
) of 6.8 mA
cm
−2
, a record V
oc
for formamidinium perovskite of 1.42 V, and
a FF of 0.72, resulting in an overall conversion effi ciency (
η
) of
7.0% (Figure 2 b). Ostensibly, a higher J
sc
value of 7.1 mA cm
−2
obtained from FA(1) device could be due to the formation of
thicker FAPbBr
3
fi lms as evident from cross-sectional SEM
analysis (Figure 1 e). By changing the solvent from DMF to a
stoichiometric mixture of DMF+DMSO solvents for the PbBr
2
precursor solution, V
oc
increases signifi cantly from 1.14 to
1.42 V which demonstrates that the morphology of the fi lm has
a considerable impact on the photovoltage.
2.2.1. Insight into the Origin of High V
oc
The V
oc
> 1.4 V for FA(2) devices is extraordinarily high con-
sidering the contact materials FTO/TiO
2
and spiro-OMeTAD.
Assuming that the work function of the FTO/TiO
2
is ≈ −4.0 to
−4.2 eV and the HOMO in the doped spiro-OMeTAD at −5.4 eV
and, thus, the work function close but smaller than this value,
V
oc
is larger than the built-in potential.
[ 29 ]
Therefore, the ener-
getics of the contact materials do not merely defi ne the V
oc
as
commonly reported in the literature.
[ 30,31 ]
Such a high V
oc
is
Adv. Funct. Mater. 2016,
DOI: 10.1002/adfm.201504977
www.afm-journal.de
www.MaterialsViews.com
Figure 1. Scanning electron microscopy analysis. Top view of a) PbBr
2
(DMF), b) PbBr
2
(DMF+DMSO) fi lms, c) FA(1) : FAPbBr
3
perovskite fi lms prepared from PbBr
2
(DMF), d) FA(2) :
FAPbBr
3
perovskite fi lms prepared from PbBr
2
(DMF+DMSO). Cross-sectional SEM image of
devices based on e) FA(1) and f) FA(2) fi lms.
FULL PAPER
4
wileyonlinelibrary.com
©
2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
only possible when free charge carrier generation occurs within
the perovskite itself, i.e., interfaces to TiO
2
or spiro-OMeTAD are
not required to split excitons. Additionally, recombination of pho-
togenerated holes at FTO and electrons at spiro-OMeTAD has to
be suppressed, i.e., contacts have to be selective. In that case, V
oc
is determined by the equilibrium between carrier generation and
recombination in the absorber itself, which defi nes the splitting of
electron and hole quasi Fermi level. Given a suffi cient diffusivity,
charges are driven by diffusion towards their respective electrodes
to generate photocurrent for voltages close to but smaller than V
oc
.
To get a closer insight into the recombination mechanisms
limiting V
oc
, we measured V
oc
as a function of illumination inten-
sity.
[ 32 ]
The data shown in Figure 3 with a slope of 120 mV/decade
indicate that V
oc
is determined by recombination via defects.
Even at 1 sun, the slope is not decreased, which would be
expected for limitations by surface recombination in case of a
compensated built-in potential for large V
oc
.
[ 32 ]
This indicates
that the contacts are very selective and V
oc
continues increasing
for higher light intensities. In particular, spiro-OMeTAD does
not seem to limit V
oc
and higher V
oc
values may be reached
when the quality of the FAPbBr
3
/TiO
2
fi lm is further improved,
i.e., the defect density reduced, while maintaining the contact
materials. The high selectivity of the device structure is given
by offsets in the valence band of TiO
2
and the lowest occupied
molecular orbital (LUMO) of spiro-OMeTAD referred to the
energy levels of FAPbBr
3
and might be enhanced by dipoles
formed at the interfaces. A further indication for the high selec-
tivity is given by the fact that the J – V curves for FA(2) devices
(Figure 2 ) under illumination and in the dark do not show a
pronounced point of intersection.
Biasing the device in forward and collecting the emitted
photon fl ux of electrons and holes that radiatively recombine
in the FAPbBr
3
, we measure an electroluminescence external
quantum effi ciency (EQE) of ≈10
−8
… 10
−7
for currents of
5–10 mA cm
−2
. Considering the photovoltaic EQE onset
(Figure S5, Supporting Information), we expect a photovoltage
of V
oc
,
rad
− Δ V
oc, nonrad. loss
= 2.0 − 0.5 = 1.5 V.
[ 33 ]
This rough
estimation is in decent agreement with measured values of
V
oc
close to 1.45 V for this device.
Such a signifi cant improvement in photovoltage was further
investigated by impedance spectroscopy (IS). The measure-
ments based on impedance spectroscopy revealed a clear differ-
ence in the resistive response of FA(1) and FA(2) devices. With
increasing forward voltage, both the resistances (determined
at higher as well as lower frequency range) dropped faster for
FA(1) device (Figure S6, Supporting Information). Relatively
faster nonradiative recombination can explain a lower open-
circuit voltage obtained from FA(1) devices.
[ 31,34,35 ]
The reproducibility of our results is ascertained by
depicting the photovoltaic effi ciency derived from J – V meas-
urements for a batch of 20 devices (Figure S7, Supporting
Information). Around 80% of the devices exhibited a pho-
tovoltage > 1.4 V under illumination of 100 mW cm
−2
, fur-
ther confi rming excellent reproducibility of our results. It is
worth mentioning that the photovoltage of FAPbBr
3
devices
remained remarkably stable upon prolonged storage under
ambient conditions for more than 3 months (Figure S8,
Supporting Information).
Due to the accumulation and migration of ions at various
interfaces and within the perovskite solar cells, scan speed
dependent hysteresis is observed in current–voltage curves.
[ 36,37 ]
Arguably, hysteresis creates discrepancy in estimating the
real photovoltaic parameters obtained from a J – V curve.
[ 38 ]
Adv. Funct. Mater. 2016,
DOI: 10.1002/adfm.201504977
www.afm-journal.de
www.MaterialsViews.com
Figure 2. Photovoltaic performance of perovskite devices with spiro-
OMeTAD as hole transporter recorded at a scan rate of 0.1 V s
−1
under
simulated AM1.5 100 mW cm
−2
photon fl ux. a) FA(1) : FAPbBr
3
perovs-
kite device prepared from PbBr
2
(DMF) and b) FA(2) : FAPbBr
3
perovskite
device prepared from PbBr
2
(DMF+DMSO) solution.
Table 1. Summarized photovoltaic parameters derived from current–
voltage curves for the FA(1) and FA(2) devices fabricated using different
precursor solution of PbBr
2
.
Sample
J
sc
[mA cm
−2
]
V
oc
[V]
FF [%] PCE [%]
FA(1) 7.1 1.14 67 5.4
FA(2) 6.8 1.42 72 7.0
Figure 3. V
oc
as a function of illumination intensity provided by white
LEDs. A value of 1 sun delivers a short-circuit current close to the one at
100 mW cm
−2
under simulated AM1.5.
FULL PAPER
5
wileyonlinelibrary.com
©
2016 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 4 a,b shows the hysteresis curves obtained from FA(1)
and FA(2) devices recorded at a scanning rate of 0.1 V s
−1
.
FA(2) devices exhibited negligible hysteresis as compared to
FA(1) devices; in fact the latter displayed a considerable degree
of hysteresis. Various strategies have been adopted to fabricate
hysteresis-free effi cient devices.
[ 13 ]
Surface passivation or mini-
mizing the concentration of vacancies/traps has been surmised
as one of the ways to bring down the degree of the hysteresis.
[ 39 ]
Herein, we have achieved so by controlling the growth and for-
mation of FAPbBr
3
layer, which further emphasizes the impor-
tance of perovskite fi lm formation.
Figure S5 in the Supporting Information shows external
quantum effi ciency spectra of FA(1) and FA(2) photovoltaic
devices as a function of wavelength. EQE spectra reveal that
the generation of photocurrent begins at ≈550 nm, which is in
agreement with the band gap of pure FAPbBr
3
12
and the cur-
rent densities obtained from EQE data are in good agreement
with J
sc
values acquired from current–voltage measurements.
2.3. Spectroscopic Studies
2.3.1. Absorption Studies
Comparative analysis of photovoltaic studies of FA(1) and
FA(2) devices showed that the former exhibits slightly higher
current density (Figure 2 a). Such an enhancement can be
attributed to higher loading of FAPbBr
3
perovskite in FA(1)
fi lms as confi rmed by cross-sectional SEM analysis. To unravel
the difference in short-circuit current densities further; we
examined perovskite fi lms with UV–visible absorption spec-
troscopy in transmission mode. Both FA(1) and FA(2) fi lms
showed a sharp absorption onset around 550 nm ( Figure 5 a )
however, FA(1) fi lm displayed higher absorbance. Higher
absorption indicates relatively higher loading of FAPbBr
3
which clearly rationalizes higher current density obtained from
FA(1) devices.
2.3.2. Photoluminescence Studies
The formation of thicker films which expectedly exhibit
higher absorption justifies the generation of higher pho-
tocurrent in FA(1) devices. To understand the cause of
higher V
oc
in FA(2) devices, we performed time-integrated
and time-resolved photoluminescence (TRPL) spectroscopy
as the V
oc
in case of optimized contacts is determined by
charge carrier dynamics and emission characteristics of the
absorber.
[ 40,41 ]
Figure 5 b shows the time-integrated PL spectra acquired at
15 and 300 K for FA(1) and FA(2) perovskite fi lms. At 300 K,
FA(2) exhibits stronger PL intensity (>8 times) compared to
FA(1) sample. As this difference could be due to various rea-
sons, the PL signal is also measured at 15 K, where nonradia-
tive recombination channels are reduced. By comparing the
PL intensity at 15 K (inset), which shows comparable values,
we surmise that such a contrast in PL intensity at 300 K could
arise from variations in the nonradiative recombination rates
which are apparently lower in FA(2) sample. To gain further
understanding of the kinetics of charge carrier decay, we inves-
tigated the samples through time-resolved photoluminescence.
Figure 5 c,d presents the time-dependent PL decay traces for
both FA(1) and FA(2) fi lms recorded at 300 K. FA(1) sample
shows a two exponential decay with extremely short decay time
components (
τ
1
= 500 ps and
τ
2
= 2 ns) whereas the FA(2)
sample exhibits a mono-exponential decay with a relatively long
lifetime (
τ
1
= 24 ns). From PL and TRPL studies we noticed
that the FA(2) sample exhibits a lower nonradiative recombina-
tion rate and longer lasting charge carriers compared to FA(1)
which could possibly explain the difference in V
oc
as shown in
Figure 2 .
[ 40,41 ]
2.3.3. Amplifi ed Spontaneous Emission
Amplifi ed spontaneous emission measurements were car-
ried out on FA(1) and FA(2) fi lms after pumping them with
70 ps laser pulses with tunable wavelengths. ASE was fl u-
ently observed in both samples with wide range of pumping
wavelengths (430–530 nm). Figure 6 a,b shows the evolution
of emission spectra with increasing pump fl uence of 450 nm
laser excitation pulses. At low fl uence, emission spectrum cor-
responding to SE centered at ≈550 nm was observed. However
over pump fl uence threshold, SE is accompanied with the
appearance of a sharp peak (transition to ASE) on the lower
Adv. Funct. Mater. 2016,
DOI: 10.1002/adfm.201504977
www.afm-journal.de
www.MaterialsViews.com
Figure 4. Current−voltage hysteresis recorded at a scan rate of 0.1 V s
−1
under stimulated AM1.5 100 mW cm
−2
photon fl ux for a) FA(1) device
and b) FA(2) device.
