Intrinsic Halide Segregation at Nanometer Scale Determines the
High Efficiency of Mixed Cation/Mixed Halide Perovskite Solar Cells
Paul Gratia,
†
Giulia Grancini,*
,†
Jean-Nicolas Audinot,
‡
Xavier Jeanbourquin,
§
Edoardo Mosconi,
∥
Iwan Zimmermann,
†
David Dowsett,
‡
Yonghui Lee,
†
Michael Gra
tzel,
⊥
Filippo De Angelis,
∥,#
Kevin Sivula,
§
Tom Wirtz,
‡
and Mohammad Khaja Nazeeruddin*
,†
†
Group for Molecular Engineering of Functional Materials (GMF), Institute of Chemical Sciences and Engineering, Swiss Federal
Institute of Technology, CH-1951 Sion, Switzerland
‡
Advanced Instrumentation for Ion Nano-Analytics (AINA), Materials Research and Technology Department, Luxembourg Institute
of Science and Technology (LIST), L-4422 Belvaux, Luxembourg
§
Molecular Engineering of Optoelectronic Nanomaterials Lab (LIMNO), Swiss Federal Institute of Technology, CH 1015 Lausanne,
Switzerland
∥
Computational Laboratory for Hybrid/Organic Photovoltaics (CLHYO), CNR-ISTM, Via Elce di Sotto 8, I-06123 Perugia, Italy
⊥
Laboratory for Photonics and Interfaces (LPI), Institute of Chemical Sciences and Engineering, Swiss Federal Institute of
Technology, CH-1015 Lausanne, Switzerland
#
Computet, Istituto Italiano di Tecnologia, Via Morego 30, I-16163 Genova, Italy
*
S
Supporting Information
ABSTRACT: Compositional engineering of a mixed
cation/mixed halide perovskite in the form of (FAP-
bI
3
)
0.85
(MAPbBr
3
)
0.15
is one of the most effective strategies
to obtain record-efficiency perovskite solar cells. However,
the perovskite self-organization upon crystallization and
the final elemental distribution, which are paramount for
device optimization, are still poorly understood. Here we
map the nanoscale charge carrier and elemental distribu-
tion of mixed perovskite films yielding 20% effi cient
devices. Combining a novel in-house-developed high-
resolution helium ion microscope coupled with a
secondary ion mass spectrometer (HIM-SIMS) with
Kelvin probe force microscopy (KPFM), we demonstrate
that part of the mixed perovskite film intrinsically
segregates into iodide-rich perovskite nanodomains on a
length scale of up to a few hundred nanometers. Thus, the
homogeneity of the film is disrupted, leading to a variation
in the optical properties at the micrometer scale. Our
results pr ovide unprecedented understandin g of th e
nanoscale perovskite composition.
P
erovskite photovoltaics (PVs), which appear in the list of
the “Top 10 Emerging Technologies of 2016”,
1
promise to
be a major player in the near-future carbon-free energy
landscape. Among the large variety of perovskite compositions,
the mixed cation/mixed halide (FAPbI
3
)
0.85
(MAPbBr
3
)
0.15
perovskites (MA = methylammonium; FA = formamidinium)
currently hold the lead since they have repeatedly proven to
yield high power conversion efficiencies (PCEs) beyond 20%,
2,3
competing with established thin-film PV technologies. This has
been achieved with exceptional ease by tuning the material’s
optoelectronic properties and processing upon simple chemical
substitution.
4−6
Efforts to optimize pure MAPbI
3
7
by composi-
tional engineering of cations (e.g., substitution of MA by FA)
and anions (e.g., introducing a small amount of Br) along with
the addition of excess lead iodide have indeed induced a
breakthrough in device efficiency and reproducibility.
8,5
Recent
developments even include triple cation structures containing
cesium, MA, and FA.
9
The higher efficiency compared with
single MAPbI
3
perovskite structures has been generically
attributed to the improved crystal quality of the film,
3
but the
exact reasons remain under intense debate. Previous reports
have found that light-induced halide segregation in
MAPbBr
x
I
3−x
(no FA cation),
10,11
iodine migration in
MAPbI
3
perovskites,
12
and electric-field-induced halide migra-
tion
13
have enormous impacts o n the device behavior.
Nevertheless, a rational investigation of the intrinsic mixed
cation/mixed halide (FAPbI
3
)
0.85
(MAPbBr
3
)
0.15
perovskite
composition is still missing. In this work, we provide in-depth
insight into the local elemental composition and structural−
optical properties of high-efficiency mixed perovskite devices by
using advanced nanoscale mapping techniques.
A novel in-house-developed helium ion microscope coupled
to a secondary ion mass spectrometer (HIM-SIMS) revealed
that the mixed perovskite fi lm self-organizes to form large
domains of pure iodide perovskite (up to a few hundred
nanometers in size) spatially segregated from the mixed iodide/
bromide phases. Such nanoscale segregation is responsible for
driving local charge distribution and charge accumulation in the
iodid e rich domains, as revealed by Kelvin probe force
microscopy (KPFM) mapping.
High-efficiency (FAPbI
3
)
0.85
(MAPbBr
3
)
0.15
solar cells were
fabricated using a previously described protocol.
3
Device
current−voltage (I−V) characteristics showing a champion
device efficiency of 20.4% and device statistics are plotted in
Received: September 24, 2016
Published: November 29, 2016
Communication
pubs.acs.org/JACS
© 2016 American Chemical Society 15821 DOI: 10.1021/jacs.6b10049
J. Am. Chem. Soc. 2016, 138, 15821−15824
Figure S1. The measurements presented in this work were
carried out on twin samples fabricated within the same batches
as the high-efficiency devices. Figure 1a shows the scanning
electron microscopy (SEM) image of the top surface of the
mixed (FAPbI
3
)
0.85
(MAPbBr
3
)
0.15
device consisting of crystal
grains ranging from 100 to 500 nm in size. Figure 1b−d shows
the elemental distributions of
127
I,
79
Br, and
12
C across a 10 μm
× 10 μm area of the perovskite surface obtained using the
unique HIM-SIMS setup developed in 2015 by Wirtz et al.
from the Luxembourg Institute of Science and Technology.
14,15
The instrument allows high-sensitivity imaging of surfaces
(surface sensitivity is 10−20 nm) with lateral resolution down
to 10 nm, representing a factor of 5 enhancement with respect
to the best commercially available SIMS instrument (Cameca
NanoSIMS 50). Details can be found in the literature
16
and the
Supporting Information.
By mapping three isotopes, namely,
127
I,
79
Br and
12
C, we
observe that the
79
Br content varies from null/very low to high
intensity, indicating compositional nonhomogeneity at the
nanoscale. In the case of
79
Br, several local minima extend to
regions of up to 300 nm (Figure 1c). Local carbon hotspots are
clearly revealed (Figure 1d), although a background
12
C signal
due to vacuum contamination is also present. Combining the
elemental maps of
79
Br and
127
IinFigure 1e reveals a clearer
correlation. Most of the regions with null/very low
79
Br content
correspond to regions with a high amount of
127
I(“blue
spots”). A zoomed view of a region containing such a blue spot
as well as the profiles of the
79
Br and
127
I signals along the
dashed white lines are shown in Figure 1f−h. Although the
SIMS technique is not quantitative, it clearly illustrates that
127
I
counts totally prevail over
79
Br counts for at least 100 nm. We
exclude that such regions consist of pure PbI
2
because they
mostly correspond to the
12
C hotspots. This strongly
underlines the presence of FA/MA cations, ruling out the
presence of PbI
2
. Since both MA and FA contain one C atom,
the variation of the
12
C signal can be assigned to the difference
of the corresponding SIMS signal obtained from FA
x
MA
y
PbI
3
phase compared with the one that also contains
79
Br. We
suggest that the segregation of (FAPbI
3
)
0.85
(MAPbBr
3
)
0.15
results in aFA
x
MA
y
PbI
3
+ bFA
x
MA
y
PbI
z
Br
3‑z
+ cFA
x
MA
y
PbBr
3
,
where a, b, and c represent the relative weights of the phases.
While b dominates, HIM-SIMS shows that a ≠ 0.
FA
x
MA
y
PbBr
3
is not revealed either because c ≈ 0 or because
those phases segregate below our resolution limit. In the first
case, for mass conservation, the excess of Br can be contained in
the FA
x
MA
y
PbI
z
Br
3−z
, where z is slightly less than 0.85. Overall,
the analysis proves the nonhomogeneity in the elemental
composition of the mixed perovskite surface and indicates that
multiple interconnected perovskite phases with different halide
contents are formed at nanometer scale. In order to determine
whether the
79
Br and
127
I amounts vary through the depth of
the capping layer, we also performed conventional SIMS depth
profiling (without lateral resolution as in HIM-SIMS) of the
capping layer (Figure S2). Interestingly, no significant gradient
is observed.
Figure 1. Elemental nanoscale HIM-SIMS mapping. (a) SEM surface image of (FAPbI
3
)
0.85
(MAPbBr
3
)
0.15
perovskite deposited on a mesoporous
TiO
2
scaffold. (b−d) HIM-SIMS elemental mapping of
127
I (blue),
79
Br (red), and
12
C (green) across a 10 μm × 10 μm area. (e) Overlap of the
79
Br
(red) and
127
I (blue) signals. The “blue” spots have low
79
Br and high
127
I intensities. (f, g) Zoomed views of (f) combined
79
Br and
127
I and (g)
79
Br
alone. (h) Profiles of the
79
Br and
127
I signals across the dashed lines in (f) and (g). (i) Overlap of
79
Br and
12
C signals revealing that the carbon
hotspots in (d) correspond to
79
Br signal minima.
Figure 2. KPFM measurements in the dark and under illumination. (a) A 5 μm × 5 μm surface potential map of the capping layer of a mixed
perovskite solar cell (no cathode) in the dark. (b) A 1.5 μm × 1.5 μm map of the area marked by the gray box in (a). (c) Corresponding
photovoltage buildup that links bright regions to dark regions in (b), indicating possible hole accumulation.
Journal of the American Chemical Society Communication
DOI: 10.1021/jacs.6b10049
J. Am. Chem. Soc. 2016, 138, 15821−15824
15822
To confirm the surface nonhomogeneity and to investigate
the influence of the composition on the charge carrier
distribution, we carried out KPFM measurements. Figure 2
shows the results over the mixed cation/mixed halide
perovskite surface (capping layer). A 5 μm × 5 μm KPFM
surface potential map of the perovskite capping layer in the
dark shows characteristics that can be immediately linked to the
grain morphology (Figure S3a and Figure 1a). The map shows
an average contact potential difference (CPD) of around 600
mV. Interestingly, the map is not homogeneous but shows
several darker spots with lower CPD (as indicated by the arrow
in Figure 2a). A lower CPD value is related to a higher work
function and therefore a deeper Fermi level. A 1.5 μm × 1.5 μm
map (Figure 2b) reveals that these lower-CPD regions overlap
with well-defined crystal grains. In addition, darker stripes on
the crystal grain facets also exhibit lower CPD values. In Figure
2c, corresponding to the difference of the map under
illumination and in the dark, a clear trend is observed: high-
CPD regions in the dark lead to lower photovoltage, whereas
low-CPD areas (e.g., the dark grain indicated by the arrow) are
connected to higher photovoltage. These findings can be
rationalized in the following way: the average CPD value of the
mixed perovskite (600 mV) is about 120 mV higher than in the
darker spots, where the CPD equals 480 mV. This value exactly
matches the results for the pure MAPbI
3
perovskite surface (see
Figure S3f for comparison) and indicates the presence of
nanodomains constituted of FA
x
MA
y
PbI
3
perovskite phases on
the surface of the mixed perovskite. In agreement, the higher
photovoltage is caused by a slight downshift in the Fermi level
under illumination, which could be due to hole accumulation
within the I-rich perovskite grain. This correlates with the HIM-
SIMS measurements and demonstrates that the segregation of
FA
x
MA
y
PbI
3
, extending up to about 300 nm, is a fundamental
property of the (FAPbI
3
)
0.85
(MAPbBr
3
)
0.15
films (excluding the
possibility of ion-beam-induced segregation). It is also worth
noting that the chemical composition of the grain might vary
with the crystal facet present at the surface, which may be
responsible for the local change in CPD. However, this seems
unlikely, as several flat crystals do not exhibit low CPD.
Moreover, the possibility that the low-CPD grain (higher
photovoltage) is composed of a precursor crystal (e.g., PbI
2
)is
excluded by the unfavorable band alignment for hole
accumulation. Overall, the measurements suggest composi-
tional nonhomogeneities across the film at nanometer scale due
to intrinsic halide segregation. These local nonhomogeneities
are not usually revealed by standard characterization tools such
as X-ray diffraction (XRD), Raman, and optical analysis, since
they usually interrogate microscopic volumes of the sample.
17
Indeed, no clear evidence for any phase segregation in the
mixed perovskite layer could be observed by XRD (Figure S4).
However, the relatively broad diffraction peaks could indicate
the presence of domains in the perovskite having different
phase compositions, as the change in the lattice parameters
upon a small amount of Br substitution is expected to be small.
On the other sid e, combined micro-Raman and micro-
photoluminescence (micro-PL) spectroscopy, although aver-
aged over a diffraction-limited spot size of 300 nm, can provide
further information about local heterogeneities.
18
The PL peak
position map across the film over a 90 μm
2
area reveals an
optical inhomog eneity (Figure 3a). R epresentati ve peak-
position wavelengths at points 1, 2, and 3 are plotted in Figure
3b along with those of the pristine MAPbI
3
and FAPbI
3
films.
PL spectra and additional information are shown in Figures S5
and S6. Within the mixed cat ion/mixed hal ide (FAP-
bI
3
)
0.85
(MAPbBr
3
)
0.15
perovskite film, the peak positions of
points 1−3 shift from 766 to 775 nm. Although these
measurements are averaged over an ∼300 nm spot, the shift
can be related to a variation of the Br content with respect to
I.
19
In particular, point 2 exhibits a red-shifted emission similar
to what is reported for FA
y
MA
x
PbI
x
Br
3−x
with a minimal Br
content. This further indicates the presence of halide
redistribution within the film. On exactly the same area, we
performed a structural investigation by means of micro-Raman
spectroscopy. We collected the signals from points 1, 2, and 3
(Figure 3c) and in parallel performed density functional theory
(DFT) analysis to support the mode assignments (Figure 3d).
It should be noted that the Raman spectrum of the mixed
perovskite film is here reported and simulated for the first time.
The measured and DFT-calculated vibrational modes are listed
in Table 1. Further information regarding the computational
setup and geometries are presented in Experimental Methods
and Figure S8. The Raman spectra of the three points consist of
four main peaks (see Table 1). Similarly to the pristine MAPbI
3
perovskite,
20,21
peaks i and ii represent combinations of the
inorganic stretching and bending along with a minor
contribution of the organic cations, while peaks iii and iv, a
double feature not present in the MAPbI
3
perovskite
20,21
(see
Figure S7), are related to a combination of FA and MA
Figure 3. Optical and vibrational characterization. (a) Micro-PL peak
shift map of the perovskite surface. Reabsorption effects are negligible
because of the homogeneity in the thickness of the film. (b) PL peak
positions of pure MAPbI
3
and pure FAPbI
3
compared with the
selected points 1, 2, and 3 indicated in (a). (c) Micro-Raman spectra at
points 1, 2, and 3. Solid lines are fits to the experimental data. The
samples were encapsulated to prevent any air/moisture effects. (d)
DFT-simulated Raman spectrum of the mixed perovskite.
Table 1. Fitted and DFT-Calculated Peaks (in cm
−1
)
Associated with the Raman Spectra in Figure 3c
peak 1 2 3 calcd mode assignment
i 62666550Pb−I/Pb−Br bending and FA/MA
libration
ii 100 119 105 100 Pb−I/Pb−Br stretching and FA/MA
libration
150 FA/MA libration
iii 231 249 240 212 FA torsion
iv 292 297 297 317 MA torsion
Journal of the American Chemical Society Communication
DOI: 10.1021/jacs.6b10049
J. Am. Chem. Soc. 2016, 138, 15821−15824
15823
torsional modes linked together. We thus believe that this
characteristic is unique to the mixed structure. The retrieved
peak positions closely resemble the calculated Raman spectrum
(Figure 3d). Comparing the three different points, we first
observe that the relative amplitudes of peaks i and ii change.
Peak ii dominates and is broader at point 1, while its intensity
and broadening are reduced at point 2. An opposite trend is
observed for peaks iii and iv, which are related to FA and MA
vibrational modes, respectively. This variation can be induced
by the structural inhomogeneity throughout the film. The slight
shift in the mode frequencies can be due to the presence of
different halides in the inorganic cage, which influence the
organic motion through hydrogen bonds.
21,22
Point 3 features
characteristics intermediate between those of points 1 and 2.
Although this measurement does not aim to be exhaustive, it
further supports the structural and compositional inhomoge-
neity across the film. Such inhomogeneity is responsible for the
local PL variation, possibly induced b y locally different
perovskite structures or modified band gaps due to the
different halide compositions.
Combining HIM-SIMS, KPFM, and micro-PL/Raman
measurements, this study provides unprecedented insights
into nanoscale properties of high-efficiency mixed cation/mixed
halide perovskite solar cells. We have demonstrated that highly
efficient perovskite devices are not constituted by a perfectly
homogeneous crystalline structure, as a conventional thin-film
solar cell would require, but that, contrary to expectations,
nanoscale segregation of multiple perovskite compositions
occurs. In particular, we have identified FA
x
MA
y
PbI
3
-rich
domains within the mixed cation/mixed halide perovskite films
extending up to hundreds of nanometers that can be more
thermodynamically stable. We envisage that exploiting the
concept of nanometer phase segregation, ultimately leading to a
“bulk-heterojunction” perovskite solar cell, will be a potentially
successful strategy for a significant advance in perovskite solar
cell technology.
■
ASSOCIATED CONTENT
*
S
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.6b10049.
Details of solar cell fabrication and device characteristics;
SIMS depth profile; HIM-SIMS equipment details;
KPFM, micro- Raman/PL, and XRD measurements;
and computational details (PDF)
■
AUTHOR INFORMATION
Corresponding Authors
*giulia.grancini@epfl.ch
*mdkhaja.nazeeruddin@epfl.ch
ORCID
Giulia Grancini: 0000-0001-8704-4222
Jean-Nicolas Audinot: 0000-0002-4966-7653
Mohammad Khaja Nazeeruddin: 0000-0001-5955-4786
Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS
I.Z. was supported by Qatar Environment and Energy Research
Institute (QEERI), Hamad Bin Khalifa University (HBKU),
and Qatar Foundation (Doha, Qatar). G.G. was supported by
the cofunded Marie Skłodowska Curie Fellowship (H2020
Grant Agreement 665667, Fund 588072). The HIM-SIMS
experiments performed at LIST were cofunded by the National
Research Fund of Luxembourg (Grant C14/MS/8345352).
■
REFERENCES
(1) These are the top 10 emerging technologies of 2016. https://
www.weforum.org/agenda/2016/06/top-10-emerging-technologies-
2016 (accessed July 24, 2016).
(2) Li, X.; Bi, D.; Yi, C.; De
coppet, J.-D.; Luo, J.; Zakeeruddin, S. M.;
Hagfeldt, A.; Gra
tzel, M. Science 2016, 353, 58.
(3) Jeon, N. J.; Noh, J. H.; Yang, W. S.; Kim, Y. C.; Ryu, S.; Seo, J.;
Seok, S. I. Nature 2015, 517, 476.
(4) Jesper Jacobsson, T.; Correa-Baena, J.-P.; Pazoki, M.; Saliba, M.;
Schenk, K.; Gra
tzel, M.; Hagfeldt, A. Energy Environ. Sci. 2016, 9, 1706.
(5) Pellet, N.; Gao, P.; Gregori, G.; Yang, T.-Y.; Nazeeruddin, M. K.;
Maier, J.; Gra
tzel, M. Angew. Chem., Int. Ed. 2014, 53, 3151.
(6) Jeon, N. J.; Noh, J. H.; Kim, Y. C.; Yang, W. S.; Ryu, S.; Seok, S. I.
Nat. Mater. 2014, 13, 897.
(7) Burschka, J.; Pellet, N.; Moon, S.-J.; Humphry-Baker, R.; Gao, P.;
Nazeeruddin, M. K.; Gra
tzel, M. Nature 2013, 499, 316.
(8) Rolda
n-Carmona, C.; Gratia, P.; Zimmermann, I.; Grancini, G.;
Gao, P.; Graetzel, M.; Nazeeruddin, M. K. Energy Environ. Sci. 2015, 8,
3550.
(9) Saliba, M.; Matsui, T.; Seo, J.-Y.; Domanski, K.; Correa-Baena, J.-
P.; Nazeeruddin, M. K.; Zakeeruddin, S. M.; Tress, W.; Abate, A.;
Hagfeldt, A.; Gra
tzel, M. Energy Environ. Sci. 2016, 9, 1989.
(10) Hoke, E. T.; Slotcavage, D. J.; Dohner, E. R.; Bowring, A. R.;
Karunadasa, H. I.; McGehee, M. D. Chem. Sci. 2015, 6, 613.
(11) Yoon, S. J.; Draguta, S.; Manser, J. S.; Sharia, O.; Schneider, W.
F.; Kuno, M.; Kamat, P. V. ACS Energy Lett. 2016, 1, 290.
(12) deQuilettes, D. W.; Zhang, W.; Burlakov, V. M.; Graham, D. J.;
Leijtens, T.; Osherov, A.; Bulovic
, V.; Snaith, H. J.; Ginger, D. S.;
Stranks, S. D. Nat. Commun. 2016, 7, 11683.
(13) Leijtens, T.; Hoke, E. T.; Grancini, G.; Slotcavage, D. J.; Eperon,
G. E.; Ball, J. M.; De Bastiani, M.; Bowring, A. R.; Martino, N.;
Wojciechowski, K.; McGehee, M. D.; Snaith, H. J.; Petrozza, A. Adv.
Energy Mater. 2015, 5, 1500962.
(14) Wirtz, T.; Philipp, P.; Audinot, J.-N.; Dowsett, D.; Eswara, S.
Nanotechnology 2015, 26, 434001.
(15) Wirtz, T.; Vanhove, N.; Pillatsch, L.; Dowsett, D.; Sijbrandij, S.;
Notte, J. Appl. Phys. Lett. 2012, 101, 041601.
(16) Wirtz, T.; Dowsett, D.; Philipp, P. In Helium Ion Microscopy;
Hlawacek, G., Go
lzha
user, A., Eds.; Springer: New York, 2016.
(17) Bi, D.; Tress, W.; Dar, M. I.; Gao, P.; Luo, J.; Renevier, C.;
Schenk, K.; Abate, A.; Giordano, F.; Correa Baena, J.-P.; Decoppet, J.-
D.; Zakeeruddin, S. M.; Nazeeruddin, M. K.; Gra
tzel, M.; Hagfeldt, A.
Sci. Adv. 2016, 2, e1501170.
(18) Draguta, S.; Thakur, S.; Morozov, Y. V.; Wang, Y.; Manser, J. S.;
Kamat, P. V.; Kuno, M. J. Phys. Chem. Lett. 2016, 7, 715.
(19) Eperon, G. E.; Stranks, S. D.; Menelaou, C.; Johnston, M. B.;
Herz, L. M.; Snaith, H. J. Energy Environ. Sci. 2014, 7, 982.
(20) Quarti, C.; Grancini, G.; Mosconi, E.; Bruno, P.; Ball, J. M.; Lee,
M. M.; Snaith, H. J.; Petrozza, A.; Angelis, F. D. J. Phys. Chem. Lett.
2014, 5, 279.
(21) Grancini, G.; Marras, S.; Prato, M.; Giannini, C.; Quarti, C.; De
Angelis, F.; De Bastiani, M.; Eperon, G. E.; Snaith, H. J.; Manna, L.;
Petrozza, A. J. Phys. Chem. Lett. 2014, 5, 3836.
(22) Mosconi, E.; Amat, A.; Nazeeruddin, M. K.; Gra
tzel, M.; De
Angelis, F. J. Phys. Chem. C 2013, 117, 13902.
Journal of the American Chemical Society Communication
DOI: 10.1021/jacs.6b10049
J. Am. Chem. Soc. 2016, 138, 15821−15824
15824
