2D Homologous Perovskites as Light-Absorbing Materials for Solar
Cell Applications
Duyen H. Cao,
†
Constantinos C. Stoumpos,
†
Omar K. Farha,
†,‡
Joseph T. Hupp,
†
and Mercouri G. Kanatzidis*
,†
†
Department of Chemistry and Argonne-Northwestern Solar Energy Research (ANSER) Center, Northwestern University, 2145
Sheridan Road, Evanston, Illinois 60208, United States
‡
Department of Chemistry, Faculty of Science, King Abdulaziz University, Jeddah 21589, Saudi Arabia
*
S
Supporting Information
ABSTRACT: We report on the fabrication and properties of the
semiconducting 2D (CH
3
(CH
2
)
3
NH
3
)
2
(CH
3
NH
3
)
n−1
Pb
n
I
3n+1
(n =1,
2, 3, and 4) perovskite thin films. The band gaps of the series decrease
with increasing n values, from 2.24 eV (CH
3
(CH
2
)
3
NH
3
)
2
PbI
4
(n =1)
to 1.52 eV CH
3
NH
3
PbI
3
(n = ∞). The compounds exhibit strong light
absorption in the visible region, accompanied by strong photo-
luminescence at room temperature, rendering them promising light
absorbers for photovoltaic applications. Moreover, we find that thin
films of the semi-2D perovskites display an ultrahigh surface coverage
as a result of the unusual film self-assembly that orients the [Pb
n
I
3n+1
]
−
layers perpendicular to the substrates. We have successfully
implemented this 2D perovskite family in solid-state solar cells, and
obtained an initial power conversion efficiency of 4.02%, featuring an
open-circuit voltage (V
oc
) of 929 mV and a short-circuit current density
(J
sc
) of 9.42 mA/cm
2
from the n = 3 compound. This result is even more encouraging considering that the device retains its
performance after long exposure to a high-humidity environment. Overall, the homologous 2D halide perovskites define a
promising class of stable and efficient light-absorbing materials for solid-state photovoltaics and other applications.
■
INTRODUCTION
The emergence of hybrid halide perovskite compounds, AMX
3
(A = Cs
+
,CH
3
NH
3
+
, or HC(NH
2
)
2
+
;M=Sn
2+
and Pb
2+
; and
X=Cl
−
,Br
−
, and I
−
), in solid-state solar cells has triggered a
phenomenal advance in the photovoltaic efficiency in the last
three years.
1− 6
Perovskite compounds, in the form of
CH
3
NH
3
PbX
3
, were originally employed as light-absorbing
materials in liquid dye-sensitized solar cells by the Miyasaka
group.
7
However, they did not engender a lot of attention
because of short device lifetime resulting from the fast
dissolution of perovskites in the redox electrolyte solution.
Three years later, we have witnessed the return of perovskite-
based solar cells in a different fashion: solid-state.
2,8
The
photovoltaic performance race has been remarkable ever since,
and an e fficiency of 20.1% has been certified by NREL.
9,5
Among the light absorber candidates, 3D methylammonium
(MA) lead iodide (MAPbI
3
) is the most prominent choice
owing to its outstanding properties for a solar cell absorber,
including a high extinction coefficient,
10
a medium band gap,
11
a small exciton binding energy, and long exciton and charge
diffusion lengths.
12
From a commercialization point of view, the
large-scale implementation of perovskite solar cells requires
toxicity and stability issues to be resolved. Recently, works on
tin-based perovskites have been reported, demonstrating a
promising efficiency of ca. 5% for the CH
3
NH
3
SnI
3−x
Br
x
system
13,14
as well as in the mixed-metal CH
3
NH
3
Sn
1−x
Pb
x
I
3
system.
15−17
The moisture instability of MAPbI
3
however has
been poorly addressed. Recently, Smith et al. reported the solar
cell application of a layered (PhC
2
H
5
NH
3
)
2
(CH
3
NH
3
)
2
Pb
3
I
10
perovskite light absorber with enhanced moisture stability.
18
From a fundamental point of view, efficient external
luminescence is an indirect indication of accessing the highest
possible open-circuit voltage, a major factor in the total power
output aside from the short-circuit current and fill factor.
19,20
The 2D A
2
MI
4
-based perovskite compounds, where M is a
divalent group 14
21
or lanthanide
22
metal, have been reported
to display high photoluminescence (PL) at room temperature,
and up until now, they have been employed in field-effect
transistor (FET)
23
and light-emitting diode (LED) devices.
24,25
To move from the 3D to the 2D perovskites, the small MA
+
cation is replaced by a much bulkier organic primary
ammonium cation, thus confining the perovskite in two
dimensions because of steric effects. With these considerations
in mind, we turned our attention toward the multilayered
perovskite compounds, (A)
2
(CH
3
NH
3
)
n−1
M
n
I
3n+1
, and their
potential as light-absorbing materials, where the bulky
Received: April 13, 2015
Published: May 28, 2015
Article
pubs.acs.org/JACS
© 2015 American Chemical Society 7843 DOI: 10.1021/jacs.5b03796
J. Am. Chem. Soc. 2015, 137, 7843−7850
ammonium (the spacer) and methylammonium (the “perovski-
tizer”) cations are employed simultaneously.
2D multilayered halide perovskites take the generic structural
formula of (A)
2
(CH
3
NH
3
)
n−1
MX
3n+1
(n is an integer), where A
is a primary aliphatic or aromatic alkylammonium cation, M is a
divalent metal, and X is a halide anion. The 2D network
consists of in organic layers o f corner-sh aring [MX
6
]
4−
octahedra confined between interdigitating bilayers of interca-
lated bulky alkylammonium cations.
26
The unit layers are
stacked toget her by a combination of Coulombic and
hydrophobic forces to maintain the structure integrity. These
2D compounds could be regarded as natural multiple-quantum-
well structures in which the semiconducting inorganic layers act
as “ wells” and the insulating organic layers act as
“barriers”.
24,27,28
In this article, we report on the fabrication and properties of
thin fi lms of the 2D lead iodide perovskite
(CH
3
(CH
2
)
3
NH
3
)
2
(CH
3
NH
3
)
n−1
Pb
n
I
3n+1
series, which com-
bines the structural features of the simple 2D (n = 1) and the
3D (n = ∞) perovskites. We then show that the thin fi lms
remarkably, grow with the [Pb
n
I
3n+1
] slabs perpendicular to the
substrates and as a result can be used as the light-absorbing
layer to fabricate functional solar cells. We establish here that
unlike 3D MAPbI
3
which requires more complex film
fabrication methods to achieve high-quality films,
3,29,30
the
2D analogues yield smooth, ultrahigh surface coverage films
from a simple one-step spin-coating approach. In addition, 2D
perovskite-based films are notably moisture-resistant. In this
work, our best power conversion efficiency of 4.02% was
obtained by using (CH
3
(CH
2
)
3
NH
3
)
2
(CH
3
NH
3
)
2
Pb
3
I
10
as a
light absorber, with an open-circuit voltage (V
oc
) of 929 mV
and a short-circuit current (J
sc
) of 9.42 mA/cm
2
.
■
EXPERIMENTAL SECTION
Materials. 2,2,7,7′-Tetrakis(N,N-di-p-methoxyphenylamine)9, 9′-
spirobifluorene (spiro-OMeTAD) was p urchased from Feiming
Chemical Limited. Transparent titania (TiO
2
) paste (Dyesol 18NR-
T) was purchased from DyeSol. All other chemicals were purchased
from Sigma-Aldrich. Unless otherwise stated, all were used as received.
Methylammonium iodide (MAI) was synthesized by neutralizing
equimolar amounts of a 57% w/w aqueous hydriodic acid (HI) and
40% w/w aqueous methylamine (CH
3
NH
2
,pH∼7). The white
precipitate was collected by evaporation of the solvent using rotary
evaporation at 60 °C under reduced pressure.
Synthesis. PbO powder was dissolved in a mixture of 57% w/w
aqueous HI solution and 50% aqueous H
3
PO
2
by heating to boiling
under constant magnetic stirring for about 5 min, forming a bright-
yellow solution. Subsequent addition of solid CH
3
NH
3
I to the hot
yellow solution initially caused the precipitation of a black powder that
rapidly redissolved under stirring to afford a clear bright-yellow
solution. n-CH
3
(CH
2
)
3
NH
2
was then added dropwise under vigorous
stirring over a period of 1 min without any changes in the solution.
The stirring was then discontinued, and the solution was left to cool to
room temperature during which time deep-red rectangular-shaped
plates started to crystallize. The precipitation was deemed to be
complete after ∼2 h. The crystals were isolated by suction filtration
and thoroughly dried under reduced pressure.
Device Fabrication. FTO-coated glass (1.5 cm × 2.0 cm, TEC 7,
2.2 mm, Hartford Glass Co., Inc.) was patterned by etching away a 5
mm strip with zinc powder and 4 M HCl. Then, substrates were
cleaned by sonication in detergent, isopropanol, acetone, and dried
under an air flow before use. A 20 nm thick TiO
2
compact layer was
deposited onto the substrates by atomic layer deposition (ALD;
Savannah S300, Cambridge Nanotech Inc.) using titanium isoprop-
oxide (0.15 s pulse time, 8 s exposure time, and 20 s purge time) and
water (0.015 s pulse time, 8 s exposure time, 20 s purge time) as
precursors. For planar structure, the ALDed TiO
2
substrates were
soaked in a 0.1 M aqueous solution of TiCl
4
for 30 min at 70 °C,
rinsed with deionized water, and dried at 500 °C for 20 min. For
sensitized structure, a mesoporous TiO
2
layer composed of 20 nm
particles was deposited on the ALD-treated TiO
2
substrates by spin-
coating at 4000 rpm for 30 s using a commercial TiO
2
paste (Dyesol
18NR-T) diluted in anhydrous ethanol (2:7 weight ratio).
Mesoporous TiO
2
substrates were then gradually annealed by heating
from room temperature to 500 °C(8°C/min) for 15 min, followed by
post-treating in a 0.1 M aqueous solution of TiCl
4
for 30 min at 70 °C.
The TiO
2
substrates were finally rinsed with deionized water and dried
at 500 °C for 20 min. The light-absorbing layers were deposited by
spin-coating 1.8 M Pb
2+
perovskite precursor solutions at 3000 rpm for
30 s. The 1.8 M Pb
2+
precursor solutions of MAPbI
3
,
(BA)
2
(MA)
n−1
Pb
n
I
3n+1
(n = 4, 3, 2, and 1) were prepared by
dissolving the corresponding amount of perovskite powders in
anhydrous dimethylformamide (DMF) solvent with stirring at 70 °C
for 30 min prior to film deposition. MAPbI
3
film was formed after
annealing in air at 100 °C for 15 min, whereas other 2D perovskite
films were formed immediately after spin-coating at room temperature.
The spiro-OMeTAD hole-transporting material (HTM) solution,
comprised of 65.3 mM spiro-OMeTAD, 9.1 mM lithium bis-
(trifluoromethanesulfonyl)imide, and 93.8 mM 4-tert-butylpyridine in
chlorobenzene solvent, was then deposited on the perovskite layer by
spin-coating at 4000 rpm for 30 s. Films were dried under vacuum
overnight before completing the device fabrication process by thermal
evaporating 80 nm of gold on top of the HTM layer.
■
RESULTS AND DISCUSSION
Film Fabrication and Film Growth Characteristics. The
2D (CH
3
(CH
2
)
3
NH
3
)
2
(CH
3
NH
3
)
n−1
Pb
n
I
3n+1
family of perov-
skite compounds (n =1−4 ) was syn thes ized from a
stoichiometric reaction between PbI
2
, MAI, and n-butylammine
(BA). Structurally, the 2D perovskites are the product of slicing
the 3D perovskite along the (110) plane, in such a way that
some of the oriented MA cations are partially (n = 2, 3, and 4)
or fully (n = 1) substituted by BA cations.
24
More details of the
synthesis and an in-depth study of crystal structures, physical,
and optical properties of the 2D perovskite family will be
reported separately.
The incorpor ation of perov skite light absorber into a
functional solid-state solar device requires the transformation
of perovskite powder materials into thin films. In this work, all
perovskite films were fabricated by a one-step deposition
method, by means of spin-coating the DMF precursor solutions
of perovskite on mesoporous TiO
2
substrates. The well-studied
MAPbI
3−x
Cl
x
film was also prepared for comparison purposes.
2
To fabricate MAPbI
3−x
Cl
x
film, the precursor solution was
prepared by mixing a 3:1 molar ratio of MAI and PbCl
2
in
DMF. Deposition of the MAPbI
3
film by the one-step approach
has been shown to produce low-quality films that suffer from
low su rface coverage and large, nonuniform crystal size,
resulting in low conversion efficiency. Therefore, many different
approaches of film deposition have been examined to improve
the film quality, including high-vacuum vapor deposition,
sequential deposition, vapor-assisted deposition, solvent en-
gineering, etc.
29,3,30,4
Conversely, in this work we observe high-
quality 2D perovskite films can be easily formed using the one-
step method. The 2D perovskites self-assemble to form well-
defined films on the substrates with nearly perfect surface
coverage and a fine texture. The growth of the films is clearly
driven by the 2D nature of the compounds forming highly
oriented crystals with only a few grain boundaries. Interestingly,
the 2D films are readily formed immediately after the spin-
coating process without requiring annealing steps, demonstrat-
Journal of the American Chemical Society Article
DOI: 10.1021/jacs.5b03796
J. Am. Chem. Soc. 2015, 137, 7843−7850
7844
ing a facile preparatory method and rendering it compatible
with flexible substrates.
The thin films of (BA)
2
(MA)
n−1
Pb
n
I
3n+1
perovskites show a
highly remarkable tendency when the orientation of the
structure on the substrate is considered (Figures 1 and 2).
Generally, the orientation of 2D materials on flat substrates
strongly favors growth where the layers orient parallel to the
substrate, a trend that has also been observed in the case of
single-layer halide perovskites.
31
In the case of
(BA)
2
(MA)
n−1
Pb
n
I
3n+1
, this trend appears to be true only for
the n = 1 compound, where preferential growth along the (110)
direction occurs, thus exclusively revealing the (00l) reflections.
As soon as the layers become thicker (n > 1), a competition
arises between the BA ions, which try to confine the growth
within the planar layer, and the MA ions, which try to expand
the perovskite growth outside the layer. Already for the n =2
compound, the (0k0) reflections are “contaminated” with the
(111) and (202) reflections, which reveal the vertical growth of
the compound with respect to the substrate plane. (Note that
the (0k0) reflections for n =2−4 correspond to the (00l)
reflection for n = 1 and ∞.) The n = 3 and 4 compounds
continue the trend by showing exclusively the (111) and (202)
reflections and lacking (0k0) reflections, clearly indicating the
vertical growth of the perovskite compounds. This effect
becomes even more pronounced when one compares it with
thepreferentialgrowthofthebulk(BA)
2
(MA)
n−1
Pb
n
I
3n+1
compounds (Figures 1 and 2). These appear to follow the
standard norms (crystallizing along the (h0l) plane), thereby
showing only the (0k0) reflections. We reason that the (111)
reflection dominates the diffraction patterns because the (101)
reflection is not allowed (h =2n and h, l =2n reflection
conditions); therefore, the closest plane describing the vertical
growth becomes the (111) plane.
Figure 1. XRD of thin films vs bulk materials of (a) BA
2
PbI
4
and (b)
MAPbI
3
perovskites, with the illustration of their respective diffraction
planes. In b, films of MAPbI
3
correspond to the unique and ideal case
of perfect orientation obtained from the 3D MAPbI
3−x
Cl
x
system.
Figure 2. XRDs of thin films vs bulk materials of (a) (BA)
2
(MA)Pb
2
I
7
, (b) (BA)
2
(MA)
2
Pb
3
I
10
, and (c) (BA)
2
(MA)
3
Pb
4
I
13
perovskites, with the
illustration of their respective diffraction planes. Note that the Miller indices are different from those of (BA)
2
PbI
4
and MAPbI
3
because of the
different assignment of the orthogonal unit cell axes.
Journal of the American Chemical Society Article
DOI: 10.1021/jacs.5b03796
J. Am. Chem. Soc. 2015, 137, 7843−7850
7845
The extremely well oriented (BA)
2
(MA)
n−1
Pb
n
I
3n+1
perov-
skite films are already superior to MAPbI
3
films deposited by
the crude one-step method, and their quality is already
comparable to the best-oriented MAPbI
3
films obtained by
means of the superior “mixed-halide” MAPbI
3−x
Cl
x
method
(Figure 1b).
2
The successful vertical crystal growth of the 2D
perovskites on the substrate is further confirmed by the
scanning electron microscopy (SEM) images, exemplified in
the form of Pb
3
I
10
films. Figure 3d shows extremely well-
packed, verticall y orient ed crystallites displaying excellent
surface coverage. Vertical growth is highly desirable in
photovoltaic devices because it facilitates charge transport
along the Pb−I−Pb pathway to the electron- and hole-
accepting contacts. Additionally, the evolution of continuous
crystallites with few crystal boundaries promises significant
improvement in the carrier transport mobility.
Another favorable property of the 2D perovskite films that
also benefits their potential technological exploitation is their
extremely high moisture stability. As an example, a film of
MAPbI
3
gradually decomposed to yellow PbI
2
after a short time
in moist atmosphere because of the gradual loss of the MA
+
cation (Figure 4b). However, a film of (BA)
2
(MA)
2
Pb
3
I
10
remained unchanged after 2 months exposure under a 40%
humidity condition. The stability of the film was additionally
confirmed by XRD (Figure 4a), in which no PbI
2
peak was
observed in the 2 month old (BA)
2
(MA)
2
Pb
3
I
10
film. The
moisture-resistant property of the 2D perovskites may be
attributed to the hydrophobicity of the long BA cation chain
and the highly oriented and dense nature of the perovskite
films, which prevent direct contact of adventitious water with
the perovskite.
Optical Properties. For convenience, in the present and
forthcoming sections, we will abbreviate MAPbI
3
as PbI
3
, and
the 2D series as Pb
4
I
13
,Pb
3
I
10
,Pb
2
I
7
, and PbI
4
when n is equal
to 4, 3, 2, and 1, respectively. Previously, in our work on two-
step sequentially deposited films of PbI
3
, we observed some
unusually broad absorption edges lying between the band edges
of PbI
2
(2.4 eV) and PbI
3
(1.5 eV).
32
We speculated these
absorption characteristics were due to incomplete conversion of
PbI
2
, leading to the formation of intermediate layered
compounds that were stabilized by surface effects on the thin
films. We investigated the optical properties of the series in
both bulk and thin film samples. The optical band gaps (E
g
)in
the (BA)
2
(MA)
n−1
Pb
n
I
3n+1
series increase with decreased
thickness of the inorganic slabs from 1.52 eV (n = ∞)to
2.24 eV (n = 1) because of quantum confinement effects from
the dimensional reduction of the perovskite chromophore.
33−35
In addition to the primary absorption edge, we observed
another peak above the absorption edge region appearing in the
2D perovskites. The intensity of the second peak is strongest
for the n = 1 compound and progressively subsides as the
number of the inorganic slabs increases; it practically disappears
for the n = 4 compound (Figure 5a). This secondary absorption
is attributed to a long-lived excitonic state trapped in the strong
electrostatic field generated by the localized positively charged
BA ions around the negatively charged ( MA)
n− 1
Pb
n
I
3n+1
layers.
28
This local electric field provides sufficient charge-
screening that inhibits the long-range separation of the
photogenerated electron−hole pair, thus increasing its
recombination probability. The excitonic binding energies of
the layered perovskite series are thus expected to decrease with
increasing n (from 1 to ∞). The same behavior is also observed
in the absorption spectra of the spin-coated TiO
2
−perovskite
films of the compounds (Figure 5b), thus further validating the
successful deposition of the target compound onto the
substrate. Interestingly, the absorption spectra of the pure
compounds are in good agreement with those of the
intermediate absorption spectra observed in the two-step-
deposited MAPbI
3
,
32
an observation that hints toward the
possibility of the formation of layered-like intermediates during
the formation of the 3D perovskite films that are stabilized by
the TiO
2
surfaces.
As is well-demonstrated, thin films of both the 2D PbI
4
and
3D PbI
3
compounds display PL at room temperature
(RT).
24,36,37
We have sought to confirm this property by
performing PL measurements on the (BA)
2
(MA)
n−1
Pb
n
I
3n+1
homologous perovskite series on films deposited onto glass
substrates. Indeed, we observe RT PL from all 2D perovskites
as shown in Figure 5c. The PL spectra of all four 2D
compounds and the 3D analogue have distinct features that are
fully consistent with the experimentally determined optical
band gaps. A very strong PL emission is observed for the n =1
compound; the emission wavelength corresponds to the high-
energy absorption peak (bound exciton), and it lies above the
ground state of the band gap. Interestingly, when additional
slabs are introduced (increasing n), the PL emission energy
shifts in each case according to the low-energy absorption peak
(free exciton) and applies to the n = 2, 3, and 4 compounds.
The efficient external luminescence observed in the compounds
is a highly desirable property for photovoltaic applications
Figure 3. (a and b) Top surface and (c and d) cross-sectional SEM
images of MAPbI
3
and (BA)
2
(MA)
2
Pb
3
I
10
on TiO
2
films.
Figure 4. (a) XRDs of fresh and aged (2 months) (BA)
2
(MA)
2
Pb
3
I
10
film. (b) Images of five different perovskite films before and after
exposure to humidity.
Journal of the American Chemical Society Article
DOI: 10.1021/jacs.5b03796
J. Am. Chem. Soc. 2015, 137, 7843−7850
7846
because it is an indirect indication of efficient carrier generation,
thus granting access to the highest possible open-circuit
voltage.
19
Device Fabrication and Photovoltaic Performance. As
stated in the previous section, the p roperties o f the
(BA)
2
(MA)
n−1
Pb
n
I
3n+1
perovskite films show unique film
growth characteristics that offer tremendous potential for
successful implementation in solar cells. We thus proceeded to
assemble and characterize photovoltaic devices. Initial attempts
yielded a promising, high V
oc
from devices of the semi-
conducting 2D Pb
3
I
10
compound. Hence, we followed up by
exploring two different aspects: (i) changing the device
structure from planar (TiO
2
compact layer only) to sensitized
(250−1100 nm mesoporous TiO
2
) and (ii) changing the
perovskite precursor concentration (0.9, 1.8, 2.7, and 3.6 M,
based on the total Pb
2+
content). The photovoltaic responses of
all fabricated devices can be found in Supporting Information
section S7. The side view of typical TiO
2
-deposited 2D
perovskite films prior to the HTM deposition step are shown in
Figure 6.
Our champion first-generation 2D device was obtained from
the Pb
3
I
10
light absorber in combination with a 350 nm TiO
2
mesoporous film and 1.8 M perovskite concentration. (See
Figure S4 for cross-sectional SEM image of a complete device.)
Under AM 1.5G solar illumination, the photogenerated V
oc
was
929 mV, and J
sc
was 9.43 mA/cm
2
, ultimately yielding a
conversion efficiency η of 4.02% as shown in Figure 7 and
Table 1. The performance of the PbI
3
device, prepared under
identical conditions, was lower than that of the current
champion device.
5
We attribute this difference mainly to the
high concentration of PbI
3
precursor and the one-step
Figure 5. Optical band gaps of (a) bulk, (b) spin-coated TiO
2
−
perovskite thin films, and (c) PL spectra of spin-coated glass−
perovskite thin films of the MAPbI
3
and ( BA)
2
(MA)
n−1
Pb
n
I
3n+1
compounds.
Figure 6. Cross-sectional SEM images of TiO
2
−perovskite films prepared from 1.8 M Pb
2+
precursors, showing preferentially oriented film growth of
2D perovskite compounds.
Figure 7. J−V curves of sensitized lead iodide perovskite-based solar
cells.
Table 1. Photovoltaic Performances of Sensitized Lead
Iodide Perovskite-Based Solar Cells
device J
sc
(mA/cm
2
) V
oc
(mV) FF (%) efficiency (%)
MAPbI
3
6.15 684 57 2.41
(BA)
2
(MA)
3
Pb
4
I
13
9.09 872 30 2.39
(BA)
2
(MA)
2
Pb
3
I
10
9.42 929 46 4.02
(BA)
2
(MA)Pb
2
I
7
1.50 800 33 0.39
(BA)
2
PbI
4
0.06 580 29 0.01
Journal of the American Chemical Society Article
DOI: 10.1021/jacs.5b03796
J. Am. Chem. Soc. 2015, 137, 7843−7850
7847
