Content uploaded by Zhengguo Xiao
Author content
All content in this area was uploaded by Zhengguo Xiao on Oct 09, 2017
Content may be subject to copyright.
Efficient perovskite light-emitting diodes featuring
nanometre-sized crystallites
Zhengguo Xiao1, Ross A. Kerner1, Lianfeng Zhao1,NhuL.Tran
2,KyungMinLee
1, Tae-Wook Koh1,
Gregory D. Scholes2and Barry P. Rand1,3*
Organic–inorganic hybrid perovskite materials are emerging as highly attractive semiconductors for use in optoelectronics.
In addition to their use in photovoltaics, perovskites are promising for realizing light-emitting diodes (LEDs) due to their
high colour purity, low non-radiative recombination rates and tunable bandgap. Here, we report highly efficient perovskite
LEDs enabled through the formation of self-assembled, nanometre-sized crystallites. Large-group ammonium halides
added to the perovskite precursor solution act as a surfactant that dramatically constrains the growth of 3D perovskite
grains during film forming, producing crystallites with dimensions as small as 10 nm and film roughness of less than 1 nm.
Coating these nanometre-sized perovskite grains with longer-chain organic cations yields highly efficient emitters,
resulting in LEDs that operate with external quantum efficiencies of 10.4% for the methylammonium lead iodide system
and 9.3% for the methylammonium lead bromide system, with significantly improved shelf and operational stability.
Hybrid organic–inorganic perovskites are emerging as a new
generation of low-cost, solution-processed semiconducting
materials with favourable optoelectronic properties such as
strong absorption coefficients, tunable bandgap1,2, large and
balanced electron and hole mobilities, long carrier diffusion
lengths, small exciton binding energy3,4 and unique defect proper-
ties with only shallow point defects formed5. These characteristics
have allowed for consistent and rapid progress in solar cell effi-
ciency, which has risen dramatically from 3.8% to over 22% in
less than five years due to efforts on perovskite film morphology
optimization, interface engineering, and so on6–17.
Owing to their facile solution processing, high colour purity and
tunable bandgap18–22, hybrid perovskites are also promising for
LEDs. Nevertheless, the small exciton binding energy in 3D perovs-
kites (methylammonium lead iodide, MAPbI
3
, and methylammo-
nium lead bromide, MAPbBr
3
) result in small electron–hole
capture rates for radiative recombination18. Therefore, ultrathin per-
ovskite layers and/or small perovskite grain size have been employed
to spatially confine electrons and holes to promote bimolecular
radiative recombination18–20. Recently, an external quantum effi-
ciency (EQE) of 8.5% for MAPbBr
3
-based LEDs has been demon-
strated through the formation of MAPbBr
3
nanograins with an
average size of 100 nm, and reduction of metallic lead impurities19.
However, due to the rapid crystallization speed of 3D perovskites,
grain sizes are generally hundreds of nanometres in size from
either one-step or two-step solution processes13,23,24, with a resulting
large surface roughness.
Here, we report a solution process to form highly uniform and
ultra-flat perovskite films with nanometre-sized grains that allow
us to demonstrate highly efficient LEDs. The addition of long-
chain ammonium halides (for example, n-butylammonium
halides (BAX, X = I, Br)) in the 3D perovskite precursor solution
impedes the growth of 3D perovskite grains and dramatically
decreases film roughness to 1 nm. The nanometre-sized grains
feature reduced dimensionality, starting a transition from 3D to
layered (so-called Ruddlesden–Popper) perovskite structures,
manifesting in stronger and blue-shifted photoluminescence (PL)
and electroluminescence (EL) compared with the emission from
the 3D perovskite. Notably, the EQE of iodide perovskite (I-perovs-
kite, BAI:MAPbI
3
) LEDs increased from 1.0% to 10.4% due to the
incorporation of BAI. Similarly, the EQE of bromide perovskite
(Br-perovskite, BABr:MAPbBr
3
) LEDs increased from 0.03% to
9.3% following the incorporation of BABr. The power efficiency
(PE) and current efficiency (CE) reached 13.0 lm W
–1
and
17.1 cd A
–1
, respectively, for Br-perovskite LEDs. Furthermore,
the shelf stability of unencapsulated I- and Br-perovskite LEDs in
N
2
was dramatically improved after adding long-chain halides.
Specifically, LEDs without BAX degrade to 60–70% of their initial
value within 2 days whereas LEDs with BAX (X = I, Br) show no
degradation after storage in the glove box for more than 8 months.
Perovskite film formation and characterization
To deposit uniform pin-hole-free perovskite films with small grain
size, toluene was dropped on the spinning film at an appropriate
time to extract the processing solvent (dimethylformamide,
DMF), thus halting the morphological evolution and maintaining
small crystallite size25,26. To ensure consistency with the device
structure, the optical, morphological and structural properties of
I- and Br-perovskite films were probed on poly[N,N′-bis(4-butyl-
phenyl)-N,N′-bis(phenyl)-benzidine] (poly-TPD) and poly(N-
vinylcarbazole) (PVK), respectively. The perovskite thickness is
approximately 80 nm for I-perovskite films and 70 nm for Br-per-
ovskite films. Atomic force microscope (AFM) images of MAPbI
3
(Fig. 1a) and MAPbBr
3
(Fig. 1g) films without annealing show
that both are very uniform. The root mean square (r.m.s.) roughness
is 4.9 nm for the MAPbI
3
film and 3.4 nm for the MAPbBr
3
film,
among the most flat and uniform perovskite films reported.
Long-chain ammonium ions cannot fill the corner of PbX
4
(X = I, Br, Cl) octahedral layers and therefore induce the formation
of layered perovskites27. Notably, when BAX (X = I, Br) is added into
the MAPbX
3
precursor solution, the growth of 3D perovskite grains
during the film formation process is dramatically impeded.
1Department of Electrical Engineering, Princeton University, Princeton, New Jersey 08544, USA. 2Department of Chemistry, Princeton University, Princeton,
New Jersey 08544, USA. 3Andlinger Center for Energy and the Environment, Princeton University, Princeton, New Jersey 08544, USA.
*e-mail: brand@princeton.edu
ARTICLES
PUBLISHED ONLINE: 16 JANUARY 2017 | DOI: 10.1038/NPHOTON.2016.269
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics 1
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Figure 1b–f and h–l show AFM images of I- and Br-perovskite films
prepared from different molar ratios of BAI:MAPbI
3
and BABr:
MAPbBr
3
in the precursor solution, respectively. As shown in
Fig. 1, when the molar ratio of BAX:MAPbX
3
(X = I, Br) increases
to 10:100, films feature smaller grain size and therefore reduced
roughness. As the BAX concentration increases, grain size continues
to decrease and films become smoother, decreasing to 0.6 nm r.m.s.
for I-perovskite films and 1.0 nm r.m.s. for Br-perovskite films, for a
molar ratio of BAX:MAPbX
3
(X = I, Br) of 40:100. In fact, the mor-
phologies of these films are similar to solution-processed polymers,
where grain sizes are too small to be resolved by the AFM tip. When
the BAX:MAPbX
3
molar ratio is increased to larger than 60:100,
larger grains start to form with the roughness of the film
simultaneously increasing, due to the formation of layered
perovskite phases.
X-ray diffraction (XRD) was conducted to determine the compo-
sition and crystallinity of the films. As shown in Fig. 1m,n, no dif-
fraction peaks corresponding to layered perovskite grains are
observed when the BAX molar ratio is smaller than 40:100 for
I-perovskite and 20:100 for Br-perovskite films, indicating that the
as-formed films are composed of grains that are nearly indistin-
guishable from the 3D perovskite crystal structure, with BA
cations self-assembled at the grain surfaces. It should also be
noted that the full-width at half-maximum (FWHM) of the diffrac-
tion peaks after adding BAX increases, indicating that crystallite size
decreases. The crystallite size of MAPbI
3
calculated from the XRD
I-perovskite Br-perovskite
r.m.s. = 4.9 nm
4.2 nm
0.7 nm
0.6 nm
1.7 nm
r.m.s. = 3.4 nm
2.3 nm
2.4 nm
1.0 nm
3.3 nm
0
40 nm
0
25 nm
5 nm
0
0
15 nm
0
20 nm
0
40 nm
0
80 nm
0
25 nm
0:100
10:100
20:100
40:100
60:100
0:100
10:100
20:100
40:100
60:100
6.5 nm 9.3 nm100:100 100:100
a
b
c
d
e
g
h
i
j
k
fl
mn
6
5
4
Normalized intensity (a.u.)
3
2
1
0
10 15 20
2θ (°)
25
0:100
10:100
20:100
40:100
60:100
Br-perovskite **
*
*
*
*
*
30 35
6
5
4
Normalized intensity (a.u.)
3
2
1
0
10 15 20
2θ (°)
25
0:100
10:100
20:100
40:100
60:100
100:100 100:100
I-perovskite *
*
*
*
30 35
Figure 1 | Morphology and XRD study of I- and Br-perovskite films. a–f, AFM images of the I-perovskite films with different molar ratio of BAI:MAPbI
3
in
the precursor solution. g–l, AFM images of the Br-perovskite films with different molar ratio of BABr:MAPbBr
3
in the precursor solution. The molar ratios
and r.m.s. roughness of the films are marked on the left and right sides of the images, respectively. All images use a zaxis range of 0–40 nm to best allow
for direct comparison, whereas the topographic colour-map is changed according to the colour scale bars. The scan area of all images is 2 µm×2µm.
m,n, Normalized XRD data of the I- and Br-perovskite films with different molar ratio of BAI:MAPbI
3
(m) and BABr:MAPbBr
3
(n). The peaks marked with
* indicate the peaks corresponding to layered perovskite.
ARTICLES NATURE PHOTONICS DOI: 10.1038/NPHOTON.2016.269
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics2
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
data applying a Scherrer analysis (Supplementary Information) is 16
and 13 nm, respectively, when the BAI:MAPbI
3
ratio is increased
from 10:100 to 20:100, which is approximately 25 and 20 PbI
4
layers (considering the lattice constants of a=b= 8.86 Å, c=
12.65 Å for the tetragonal structure)6. Similarly, the crystallite size
of the MAPbBr
3
is calculated to be 16 and 11 nm for BABr:
MAPbBr
3
loadings of 10:100 to 20:100, which is approximately 27
and 19 PbBr
4
layers (considering the lattice constants of cubic
MAPbBr
3
:a=b=c= 5.9 Å). In agreement with the AFM and
XRD results, the small grain size of the perovskite film after incor-
poration of BAX is confirmed by scanning electron microscopy
(SEM) as shown in Supplementary Fig. 1. The crystallite of
MAPbX
3
with BAX self-assembled at the surface can also be con-
sidered within the unit cell of layered BA
2
MA
n–1
Pb
n
X
3n+1
Ruddlesden–Popper perovskite phases where nequals the number
of PbX
4
octahedral layers in the crystallite. The discrepancy
between the nvalues calculated above and the nvalues expected
from the BAX and MAX molar ratios in the precursor solution
may result from the different reaction speed between MAX and
BAX with PbX
2
(X = I, Br). Herein, we use the term ‘3D perovskite
crystallites’in the discussion of the abovementioned films for sim-
plicity, even though we recognize that these particles possess
reduced dimensionality from n=∞. For loadings beyond these
threshold values, BAX is no longer playing the role of surfactant
400 500 600 700 800
0.0
0.2
0.4
0.6
0.8
1.0 I-perovskite
Absorbance
Wavelength (nm)
0:100
10:100
20:100
40:100
60:100
100:100
300 400 500
0.0
0.2
0.4
0.6
0.8 Br-perovskite
Absorbance
Wavelength (nm)
0:100
10:100
20:100
40:100
60:100
100:100
600 700 800 900
0.0
0.2
0.4
0.6
0.8
1.0
I-perovskite
Normalized PL intensity (a.u.)
Wavelength (nm)
Time delay (ns) Time delay (ns)
0:100
10:100
20:100
40:100
60:100
100:100
450 500 550 600
0.0
0.2
0.4
0.6
0.8
1.0
Br-perovskite
Normalized PL intensity (a.u.)
Wavelength (nm)
0:100
10:100
20:100
40:100
60:100
100:100
ab
cd
f
gh
e
MAPbI
3
Wavelength (nm)
700
800
750
650
MAPbBr
3
Wavelength (nm)
460
540
580
500
600
BAI:MAPbI
3 = 20:100
Wavelength (nm)
600
700
800
750
650
BABr:MAPbBr
3 = 20:100
Wavelength (nm)
02468100246810
Time delay (ns) Time delay (ns)
02468100246810
460
540
580
500
Figure 2 | Absorption and PL of I- and Br-perovskite films. a,c, Absorption and PL spectra of I-perovskite films with different molar ratios of BAI:MAPbI
3
in the precursor solution. b,d, Absorption and PL spectra of Br-perovskite films with different molar ratios of BABr:MAPbBr
3
in the precursor solution.
e–h, Time-resolved PL spectra of MAPbI
3
(e), MAPbBr
3
(f), BAI:MAPbI
3
(20:100) (g) and BABr:MAPbBr
3
(20:100) (h)films.
NATURE PHOTONICS DOI: 10.1038/NPHOTON.2016.269 ARTICLES
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics 3
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
and crystallite terminator, but rather participates in crystallization.
When this transition occurs, grain size begins to increase again
and films become rougher. For instance, when the molar ratio is
further increased to 60:100 for BAI:MAPbI
3
or 40:100 for BABr:
MAPbBr
3
, layered perovskite diffraction peaks start to appear.
Ultimately, when the BAX ratio reaches 100:100, films become
dominated by layered perovskite grains, which resemble the
BA
2
MA
n–1
Pb
n
X
3n+1
perovskite, with n= 2 (Supplementary Fig. 2).
The absorption and PL spectra of I- and Br-perovskite films with
different molar ratios of BAX:MAPbX
3
(X = I, Br) in the precursor
solution are shown in Fig. 2a–d. With increased BAX loading, the
absorption edge blue-shifts, and excitonic absorption peaks corre-
sponding to layered perovskite grains emerge when the BAX ratio
is increased to 60:100 for I-perovskite or 40:100 for Br-perovskite,
in line with our observations from XRD as discussed above. In
turn, the PL peaks of BAX:MAPbX
3
films gradually blue-shift with
increasing BAX content from 0:100 to 40:100. The PL quantum
yield (QY) concomitantly increases from 0.2 to 0.9, 1.9 and 6.6%
for I-perovskite, and from 0.2 to 0.5, 7.0 and 40.1% for Br-perovskite
films with BAX molar ratio increasing from 0:100 to 10:100, 20:100
and 40:100, respectively. The time-resolved PL measurement showed
that the average decay time is increased from 8.8 ns (14.3 ns) to 11.4
ns (26.2 ns) with the BAI (BABr) molar ratio increasing from 0:100
to 20:100 (Fig. 2e–h and Supplementary Table 1). The increased and
blue-shifted PL emission is due to the reduced 3D perovskite
crystallite dimensionality to n< 30 surrounded by BA cations, indu-
cing a disorder effect when the lattice periodicity is interrupted28.
When the BAX (X = I, Br) ratio is further increased to 60:100,
layered perovskite grains with excitonic absorption form, inducing
significant blue-shifts of absorption and PL29, as shown in Fig. 2b,d.
The blue-shift accompanying the reduction in crystallite size is
also suggested by the evolution of the absorption and PL spectra
due to thermal annealing (Supplementary Fig. 3). Thermal anneal-
ing drives grain growth and increases grain size as confirmed by
both AFM (Supplementary Fig. 4) and XRD (Supplementary
Fig. 5) for both I- and Br-perovskite films. The FWHM of the
(110) and (220) diffraction peaks for I-perovskite and (100) and
(200) diffraction peaks for Br-perovskite reduce after thermal
annealing for various molar ratios of BAX:MAPbX
3
in the
precursor solution.
Hybrid perovskite LEDs
The LED structures are: ITO (150 nm)/HTL (30 nm)/I-perovskite
(80 nm) (or Br-perovskite, 70 nm)/TPBi (60 nm)/LiF (1.2 nm)/Al
(100 nm) (ITO, indium tin oxide; HTL, hole-transporting layer;
TPBi, 2,2′,2″-(1,3,5-benzinetriyl)-tris(1-phenyl-1-H-benzimida-
zole)). This structure is shown schematically in Fig. 3a, along with
energy diagrams in Fig. 3b,c. Supplementary Fig. 6 shows cross-sec-
tional SEM images of I-perovskite devices, which confirm film
thicknesses and the small grain size after BAI incorporation. The
Poly
TPD
ITO
I-perovskite
2.7
5.6
2.0
5.4
3.8
4.7
6.2
eV
eV
Br-perovskite
5.9
3.6
PVK
ITO
6.2
2.7
4.7
5.8
2.0
ITO
HTL
Al
LiF
V
BA cation
MAPbX3
a
bc
TPBi
LiF/Al
TPBi
LiF/Al
Perovskite
TPBi
Figure 3 | Struct ure and energy diagram of perovskite LEDs. a, Device structure of the perovskite LEDs. A schematic of a nanometre-sized grain with BA
cations decorating its surface is shown on the right. b,c, Energy diagram of I-perovskite (b) and Br-perovskite (c) LEDs. The energy levels of the 3D
perovskite (MAPbI
3
and MAPbBr
3
) and organic semiconductor layers are taken from literature37.
ARTICLES NATURE PHOTONICS DOI: 10.1038/NPHOTON.2016.269
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics4
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
HTL utilized for I-perovskite-based LEDs was poly-TPD with a
highest occupied molecular orbital (HOMO) of 5.6 eV whereas
PVK with a deeper HOMO level of 5.8 eV was used for Br-perovs-
kite-based LEDs owing to better energy level alignment for hole
injection and electron blocking. The TPBi layer serves as an elec-
tron-injecting and hole-blocking layer for both I-perovskite and
Br-perovskite LEDs. It should be noted that the energy levels of
the perovskites may change as a result of the BAX incorporation.
However, this change is expected to be small (approximately 60
meV for I-perovskite and 80 meV for Br-perovskite calculated
from the PL peak shift) such that it has a minimal effect on
charge injection and blocking.
Figure 4a,d show the normalized EL spectra of I- and Br-perovs-
kite LEDs, respectively. Similar to the PL, EL is blue-shifted with
increasing molar ratio of BAX:MAPbX
3
(X = I, Br). The small
blue-shift of EL peak position compared with PL (solid lines) for
some devices may result from a distribution of crystallite sizes and
corresponding nvalues, with the smaller crystallites the more
easily electrically populated30. The EL of I-perovskite LEDs with
molar ratio of 100:100 and Br-perovskite LEDs with molar ratio
of 60:100 and 100:100 are too weak to be measured, possibly the
result of the increasing population of layered perovskite phases
that have been reported to have strong electron–phonon coupling
that inhibits efficient EL31. The current density–voltage–luminance
(J–V–L) curves of I- and Br-perovskite LEDs and corresponding
EQE curves are shown in Fig. 4b,c and Fig. 4e,f, respectively. The
angular spectra and intensity profiles are shown in Supplementary
Fig. 7. The EQE of LEDs without the addition of BAX are low, possi-
bly a result of larger grain size and consequently reduced electron–
hole capture rate. With increasing molar ratio of BAX:MAPbX
3
to
20:100 in the precursor, the EQE of the device dramatically
increases. The highest EQEs reach 10.4% for 20:100 I-perovskite
and 9.3% for 20:100 Br-perovskite LEDs. Notably, the EQE of the
perovskite LEDs are relatively flat over the range of the measure-
ment, in contrast to recent work showing that EQE can increase
substantially with current density20. We believe this to be primarily
due to the low leakage currents in our devices, at approximately
10
−6
mA cm
–2
, indicating very few parasitic leakage pathways exist
between electrodes. Also, the fact that the EQE versus current
density slope reduces with the addition of BAX indicates that,
even at low current density, non-radiative recombination is slower
than radiative bimolecular recombination, in agreement with the
increased QY and longer PL lifetimes of films with BAX. It
should be noted that the EQE has a higher value than PLQY for
some devices, an aspect due to the nearly two-orders-of-magnitude
higher charge density associated to a current of 10–20 mA cm
–2
comparedwith the charge density generated by the optical generation
of the PLQY measurement. The PE and CE curves of both I- and
Br-perovskite LEDs are shown in Supplementary Fig. 8. The peak
PE and CE for Br-perovskite LEDs reach 13.0 lm W
–1
at 1,200 cd m
–2
and 17.1 cd A
–1
at 2,900 cd m
–2
, respectively, while PE and CE of
the I-perovskite devices are relatively low due to the near-infrared
light emission (Fig. 4a). While EQE reaches a maximum within
the range of 5.0–6.5 V, corresponding to 15–50 mA cm
–2
, there is
not a strong roll-off at higher voltages, although devices consistently
600 700 800
0.0
0.2
0.4
0.6
0.8
1.0
0:100
10:100
20:100
40:100
60:100
I-perovskite
Normalized EL/PL intensity (a.u.)
Wavelength (nm)
10
−
1
10
0
10
1
I-perovskite
EQE (%)
10 20 30 40 50
450 500 550
0.0
0.2
0.4
0.6
0.8
1.0 Br-perovskite
0:100
10:100
20:100
40:100
Normalized EL/PL intensity (a.u.)
Wavelength (nm)
10
−
2
10
−
1
10
0
10
1
Br-perovskite
EQE (%)
10 20 30 40 50
a c
d
b
ef
1356
10−7
10−5
10−3
10−1
101
103
0:100
10:100
20:100
40:100
60:100 0:100
10:100
20:100
40:100
60:100
I-perovskite
024
0
20
40
60
80
Luminance (cd m
−
2
)
Voltage (V)
0123456
10−7
10
−5
10
−3
10
−1
10
1
10
3
0:100
10:100
20:100
40:100
0:100
10:100
20:100
40:100
Br-perovskite
Current density (mA cm
−
2
)Current density (mA cm
−
2
)
Current density (mA cm
−
2
)
Current density (mA cm
−
2
)
Voltage (V)
2
4
6
8
Luminance (×1,000 cd m
−
2
)
Figure 4 | Perovskite LED performance characterization. a–c, I-perovskite LEDs with different BAI:MAPbI
3
molar ratios, showing EL (open circles) and PL
(solid lines) spectra (a), J–V–Lcurves (b) and EQE curves (c). It should be noted that BAI:MAPbI
3
(60:100) LED has higher luminance than 20:100 due to
the EL spectrum that has shifted closer to peak eye response. d–f, Br-perovskite LEDs with different BABr:MAPbBr
3
molar ratios, showing EL (open circles)
and PL (solid lines) spectra (d), J–V–Lcurves (e) and EQE curves (f). The scanning rate for all devices is 0.1 V s
–1
.
Table 1 | Perovskite LED performance metrics.
BAX:MAPbX
3
Peak PL/
EL
position
(nm)
PE (lm W
–1
)
(average/
max.)
EQE (%)
(average/
max.)
CE (cd A
–1
)
(average/
max.)
BAI:MAPbI
3
0:100 763/763 0.01/0.02 0.8/1.0 0.02/0.03
10:100 757/756 0.05/0.06 4.3/5.2 0.05/0.06
20:100 750/748 0.08/0.10 9.7/10.4 0.08/0.09
40:100 737/732 0.02/0.03 1.8/2.4 0.02/0.03
60:100 708/695 0.13/0.18 1.7/2.1 0.17/0.22
BABr:MAPbBr
3
0:100 526/522 0.03/0.04 0.02/0.03 0.06/0.09
10:100 520/519 3.5/4.3 1.8/2.4 4.2/4.9
20:100 516/513 11.2/13.0 8.8/9.3 15.8/17.1
40:100 509/503 3.5/4.0 4.2/4.5 3.9/4.5
The device statistics are based on ten devices of each composition from two batches.
NATURE PHOTONICS DOI: 10.1038/NPHOTON.2016.269 ARTICLES
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics 5
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
fail at voltages of 8–10 V (0.2–1Acm
–2
) due to Joule heating
(Supplementary Fig. 9). For either the I- or Br-perovskite LEDs,
when the BAX ratio increases beyond 40:100, the device perform-
ance decreases. Detailed performance metrics of both I- and
Br-perovskite devices are provided in Table 1.
It has recently been reported that excess MABr in MAPbBr
3
can
reduce the content of metallic lead and increase the EQE of
MAPbBr
3
-based LEDs19. As a comparison, we fabricated both
I- and Br-perovskite LEDs with MAX:MAPbX
3
(X = I, Br) molar
ratio of 5:100 and 20:100 (MAX:PbX
2
molar ratio of 105:100 and
120:100), with EQE curves shown in Supplementary Fig. 10.
While excess MAX is effective in increasing EQE for both I- and
Br-perovskite LEDs (to 3.6 and 2.1%, respectively), we find that
the incorporation of bulky ammonium halide ligands provide
greater efficiency gains because they are effective in allowing for
smaller grain size, and thin films with low roughness. We note
that excess MAX can also effectively decrease crystallite size by
forming many stacking faults within a grain32. A high density of
stacking faults can serve the same crystal termination effect as the
long-chain alkylammonium halides used herein, although less effi-
ciently. As further evidence of the versatility of this approach,
another type of large-group cation, phenylethylammonium iodide
−1 0 1 2 3 4 5
10
−8
10
−6
10
−4
10
−2
10
0
10
2
10
−6
10
−4
10
−2
10
0
10
2
I-perovskite
Current density (mA cm
−2
)
Voltage (V)
0:100
Forward scan
Reverse scan
20:100
Forward scan
Reverse scan
−
10 1 2 3 4 5
10
−
4
10
−
3
10
−
2
10
−
1
10
0
10
1
10
2
I-perovskite
0:100
Forward scan
Reverse scan
20:100
Forward scan
Reverse scan
Luminance (cd m
−2
)
Voltage (V)
345
10
−
1
10
0
10
1
I-perovskite
0:100Forward scan
Reverse scan
20:100
Forward scan
Reverse scan
EQE (%)
Voltage (V)
Br-perovskite
0:100
Forward scan
Reverse scan
20:100
Forward scan
Reverse scan
Current density (mA cm
−2
)
Voltage (V)
−
10 1 2 3 4 5 6
−
10 1 2 3 4 5 6
10
−
1
10
0
10
1
10
2
10
3
Br-perovskite
0:100
Forward scan
Reverse scan
20:100
Forward scan
Reverse scan
Luminance (cd m
−2
)
Voltage (V)
3.5 4.0 4.5 5.0 5.5
10
−
3
10
−
2
10
−
1
10
0
10
1
Br-perovskite
0:100
Forward scan
Reverse scan
20:100
Forward scan
Reverse scan
EQE (%)
Voltage (V)
0.6
0102030405060708090
0.8
1.0
1.2 I-perovskite
Normalized EQE (a.u.)
0:100
20:100
0.4
0.2
0.6
0.8
1.0
1.2
1.4 Br-perovskite
Normalized EQE (a.u.)
0:100
20:100
abc
def
gh
ij
0.1 1 10 100
10
−
4
10
−
3
10
−
2
10
−
1
10
0
10
1
Br-perovskite
0:100
20:100
EQE (%)
Time (d)
0.1 1 10 100
10
−
2
10
−
1
10
0
10
1
I-perovskite
0:100
20:100
EQE (%)
Time (d)
Time (s)
0102030405060708090
Time (s)
Figure 5 | Hysteresis and steady-state output of perovskite LEDs. a–c, Current density (a), luminance (b)andEQE(c) of I-perovskite LEDs with BAI:MAPbI
3
molar ratios of 0:100 and 20:100. d–f, Current density (d), luminance (e)andEQE(f) of Br-perovskite LEDs with BABr:MAPbBr
3
molar ratios of 0:100 and
20:100. The scanning rate is 0.1 V s
–1
and scanning range is −0.5 V to +5.5 V for all devices. g,h, Steady-state EQE measurements of I-perovskite (g)and
Br-perovskite (h) LEDs at a constant voltage of 5.0 V. The EQE of the LEDs are normalized to their initial value. i,j, Shelf stability of unencapsulated
I-perovskite (i) and Br-perovskite (j) LEDs as a function of storage time in N
2
.
ARTICLES NATURE PHOTONICS DOI: 10.1038/NPHOTON.2016.269
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics6
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
(PEAI), was incorporated with PEAI:MAPbI
3
molar ratio of 20:100.
As shown in Supplementary Fig. 10, the highest EQE reaches 9.6%
after the addition of PEAI in the MAPbI
3
precursor solution.
Recently, efficient I-perovskite LEDs using stoichiometric
PEA
2
MA
n–1
Pb
n
X
3n+1
solutions with EQE up to 8.8% for n=5
have been reported33. The non-stoichiometric solution in our
work resulted in higher EQE that might be due to the passivation
effect of the extra organic cations.
Hysteresis and stability of hybrid perovskite LEDs
The photocurrent hysteresis of perovskite solar cells is thought to be
due to ion migration and associated traps34–36, an aspect that com-
plicates accurate efficiency measurement. Hysteresis has also been
previously reported to exist in the J–V–Lresponse of perovskite
LEDs18, and can lead to inaccurate performance characterization
due to its influence on the steady-state EL emission and angular
profile measurements. And because ion migration primarily
occurs along grain boundaries34, we have investigated the influence
of BAX additives on LED hysteresis. By measuring the J–V–L
characteristics with various scanning directions and rates, as shown
in Fig. 5a–f, it can be seen that all devices show some level of hysteresis
for both Jand L. Nevertheless, the devices with BAX:MAPbX
3
(X = I, Br) molar ratio of 20:100 exhibit reduced hysteresis.
Supplementary Fig. 11 shows the scanning-rate-dependent EQE for
both I- and Br-perovskite LEDs, demonstrating that LEDs without
BAX are more scanning-rate dependent. We believe that the long-
chain BA cations that self-assemble at the crystallite surface either
impede ion motion or prevent ions from within crystallites from
participating in ion migration processes. In essence, the BA cation,
being larger than MA, decreases the cation’sthermodynamicactivity
at the surface (that is, by reducing the equilibrium vapour pressure
at the surface). To ensure that devices are able to be reliably character-
ized, the steady-state EQE was measured at a constant voltage of 5.0 V.
As shown in Fig. 5g,h, the EQE of the device without BAX showed a
strong poling effect, while the devices with BAX rapidly stabilize. The
initial increase of the EQE in films without BAX might result from
trap filling such that injected charges have a higher radiative recombi-
nation rate, while the decrease of the EQE after a few seconds might be
due to ion-migration-induced degradation of the perovskite layer.
Stability is of critical importance for perovskite LEDs, and thus
the shelf stability of unencapsulated LEDs stored in N
2
were evalu-
ated over time. As shown in Fig. 5i,j, the EQE of the control devices
without BAX in the precursor solution decreases to 60–70% of their
initial value after 2 days, whereas the EQE of an I-perovskite LED
with BAI:MAPbI
3
molar ratio of 20:100 is stable without degra-
dation after storage for more than 8 months. Similarly, the
Br-perovskite LEDs also show no degradation after storage for
4 months. The improved stability of the perovskite LEDs with
BAX can be attributed to the ultra-smooth, pinhole-free and
compact perovskite films as well as the bulky nature of the long
alkyl chains that stabilize the crystallite surfaces. The stabilized
nanocrystallite perovskite films also enable improved operational
stability, shown in Supplementary Fig. 12, an aspect that is critical
for display and lighting applications. One possible reason for the
operational instability is that ion migration induces degradation
and sites for non-radiative recombination in the perovskite layer
as the LEDs work at high voltages. This is confirmed by the
decreased PL intensity of the LED after operating at 5 V for 5 min
(Supplementary Fig. 13). Another possible reason is thermal degra-
dation caused by Joule heating. We thus propose three approaches
to continue to improve the operational stability of perovskite LEDs:
(1) use more bulky organic ligands to stabilize the surface and
impede ion migration; (2) decrease turn-on voltage through further
interface engineering such that devices operate at lower voltages;
and (3) at least partially replace MA cations with formamidinium
ammonium or Cs to increase thermal stability.
In conclusion, we have shown that the addition of long-chain
ammonium halides in the precursor solution can dramatically
impede the grain growth of 3D perovskite crystallites and decrease
film roughness. The nanometre-sized 3D perovskite grains induce
enhanced and blue-shifted PL and EL emission, with the long-
chain ammonium cations stabilizing the grain surface. As a result,
perovskite LEDs made from BAX-incorporated layers showed
greatly improved device performance, shelf stability and operational
stability. It is also worth investigating whether the incorporation of
long-chain ammonium halide ligands within perovskite films can
benefit other devices such as photodetectors, solar cells and lasers.
Methods
Methods and any associated references are available in the online
version of the paper.
Received 11 April 2016; accepted 6 December 2016;
published online 16 January 2017
References
1. Stoumpos, C. C., Malliakas, C. D. & Kanatzidis, M. G. Semiconducting tin
and lead iodide perovskites with organic cations: phase transitions, high
mobilities, and near-infrared photoluminescent properties. Inorg. Chem. 52,
9019–9038 (2013).
2. Xing, G. et al. Low-temperature solution-processed wavelength-tunable
perovskites for lasing. Nat. Mater. 13, 476–480 (2014).
3. Miyata, A. et al. Direct measurement of the exciton binding energy and effective
masses for charge carriers in organic–inorganic tri-halide perovskites. Nat. Phys.
11, 582–587 (2015).
4. Noriko, O.-Y., Takasuke, M. & Hiroshi, S. Dielectric study of CH
3
NH
3
PbX
3
(X = Cl, Br, I). J. Phys. Chem. Solids 53, 935–939 (1992).
5. Yin, W.-J., Shi, T. & Yan, Y. Unusual defect physics in CH
3
NH
3
PbI
3
perovskite
solar cell absorber. Appl. Phys. Lett. 104, 063903 (2014).
6. Kojima, A., Teshima, K., Shirai, Y. & Miyasaka, T. Organometal halide
perovskites as visible-light sensitizers for photovoltaic cells. J. Am. Chem. Soc.
131, 6050–6051 (2009).
7. Burschka, J. et al. Sequential deposition as a route to high-performance
perovskite-sensitized solar cells. Nature 499, 316–319 (2013).
8. Liu, M., Johnston, M. B. & Snaith, H. J. Efficient planar heterojunction
perovskite solar cells by vapour deposition. Nature 501, 395–398 (2013).
9. Lee, M. M., Teuscher, J., Miyasaka, T., Murakami, T. N. & Snaith, H. J. Efficient
hybrid solar cells based on meso-superstructured organometal halide
perovskites. Science 338, 643–647 (2012).
10. Mei, A. et al. A hole-conductor–free, fully printable mesoscopic perovskite solar
cell with high stability. Science 345, 295–298 (2014).
11. Yang, W. S. et al. High-performance photovoltaic perovskite layers fabricated
through intramolecular exchange. Science 348, 1234–1237 (2015).
12. Zhou, H. et al. Interface engineering of highly efficient perovskite solar cells.
Science 345, 542–546 (2014).
13. Xiao, Z. et al.Efficient, high yield perovskite photovoltaic devices grown by
interdiffusion of solution-processed precursor stacking layers. Energy Environ.
Sci. 7, 2619–2623 (2014).
14. Liu, D. & Kelly, T. L. Perovskite solar cells with a planar heterojunction structure
prepared using room-temperature solution processing techniques. Nat. Photon.
8, 133–138 (2014).
15. Jeon, N. J. et al. Solvent engineering for high-performance inorganic–organic
hybrid perovskite solar cells. Nat. Mater. 13, 897–903 (2014).
16. Im, J.-H., Jang, I.-H., Pellet, N., Grätzel, M. & Park, N.-G. Growth of
CH
3
NH
3
PbI
3
cuboids with controlled size for high-efficiency perovskite solar
cells. Nat. Nanotech. 9, 927–932 (2014).
17. Kim, H.-S. et al. Lead iodide perovskite sensitized all-solid-state submicron
thin film mesoscopic solar cell with efficiency exceeding 9%. Sci. Rep. 2,
591 (2012).
18. Tan, Z.-K. et al. Bright light-emitting diodes based on organometal halide
perovskite. Nat. Nanotech. 9, 687–692 (2014).
19. Cho, H. et al. Overcoming the electroluminescence efficiency limitations of
perovskite light-emitting diodes. Science 350, 1222–1225 (2015).
20. Li, G. et al.Efficient light-emitting diodes based on nanocrystalline perovskite in
a dielectric polymer matrix. Nano Lett. 15, 2640–2644 (2015).
21. Sadhanala, A. et al. Blue-green color tunable solution processable organolead
chloride–bromide mixed halide perovskites for optoelectronic applications.
Nano Lett. 15, 6095–6101 (2015).
22. Gong, X. et al. Highly efficient quantum dot near-infrared light-emitting diodes.
Nat. Photon. 10, 253–257 (2016).
23. Wang, Q. et al. Large fill-factor bilayer iodine perovskite solar cells fabricated by
a low-temperature solution-process. Energy Environ. Sci. 7, 2359–2365 (2014).
NATURE PHOTONICS DOI: 10.1038/NPHOTON.2016.269 ARTICLES
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics 7
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
24. Eperon, G. E., Burlakov, V. M., Docampo, P., Goriely, A. & Snaith, H. J.
Morphological control for high performance, solution-processed planar
heterojunction perovskite solar cells. Adv. Funct. Mater. 24, 151–157 (2014).
25. Xiao, M. et al. A fast deposition-crystallization procedure for highly efficient lead
iodide perovskite thin-film solar cells. Angew. Chem. 126, 10056–10061 (2014).
26. Kerner, R. A., Zhao, L., Xiao, Z. & Rand, B. P. Ultrasmooth metal halide perovskite
thin films via sol–gel processing. J. Mater. Chem. A 4, 8308–8315 (2016).
27. Mitzi, D., Wang, S., Feild, C., Chess, C. & Guloy, A. Conducting layered organic-
inorganic halides containing <110>-oriented perovskite sheets. Science 267,
1473–1476 (1995).
28. Scholes, G. D. & Rumbles, G. Excitons in nanoscale systems. Nat. Mater. 5,
683–696 (2006).
29. Mitzi, D. B., Chondroudis, K. & Kagan, C. R. Organic-inorganic electronics. IBM
J. Res. Dev. 45, 29–45 (2001).
30. Cheng, K.-Y., Anthony, R., Kortshagen, U. R. & Holmes, R. J. High-efficiency
silicon nanocrystal light-emitting devices. Nano Lett. 11, 1952–1956 (2011).
31. Gauthron, K. et al. Optical spectroscopy of two-dimensional layered
(C
6
H
5
C
2
H
4
-NH
3
)
2
-PbI
4
perovskite. Opt. Express 18, 5912–5919 (2010).
32. Song, Z. et al. Impact of processing temperature and composition on the
formation of methylammonium lead iodide perovskites. Chem. Mater. 27,
4612–4619 (2015).
33. Yuan, M. et al. Perovskite energy funnels for efficient light-emitting diodes. Nat.
Nanotech. 11, 872–877 (2016).
34. Xiao, Z. et al. Giant switchable photovoltaic effect in organometal trihalide
perovskite devices. Nat. Mater. 14, 193–198 (2015).
35. Shao, Y., Xiao, Z., Bi, C., Yuan, Y. & Huang, J. Origin and elimination of
photocurrent hysteresis by fullerene passivation in CH
3
NH
3
PbI
3
planar
heterojunction solar cells. Nat. Commun. 5, 5784 (2014).
36. van Reenen, S., Kemerink, M. & Snaith, H. J. Modeling anomalous hysteresis in
perovskite solar cells. J. Phys. Chem. Lett. 6, 3808–3814 (2015).
37. Schulz, P. et al. Interface energetics in organo-metal halide perovskite-based
photovoltaic cells. Energy Environ. Sci. 7, 1377–1381 (2014).
Acknowledgements
We acknowledge a DARPA Young Faculty Award (award no. D15AP00093) for research
funding. B.P.R. acknowledges the support of a DuPont Young Professor Awardfor research
funding. R.A.K. acknowledges support from the National Science Foundation Graduate
Research Fellowship under grant no. DGE 1148900.
Author contributions
T.-W.K. and Z.X. designed the device structure. Z.X. performed the AFM, PL and
absorption measurements, and fabricated the LEDs. R.A.K. developed the surfacted
perovskite processing protocol, synthesized the precursors and helped to calculate grain
size. L.Z. conducted the XRD and SEM measurements. N.L.T. and G.D.S conducted the
TRPL and QY measurements. K.M.L. assisted with LED characterization.B.P.R. supervised
the work. Z.X. and B.P.R. wrote the manuscript. All authors read and commented on
the manuscript.
Additional information
Supplementary information is available in the online version of the paper.
Reprints and
permissions information is available online at www.nature.com/reprints. Correspondence and
requests for materials should be addressed to B.P.R.
Competing financial interests
The authors declare no competing financial interests.
ARTICLES NATURE PHOTONICS DOI: 10.1038/NPHOTON.2016.269
NATURE PHOTONICS | ADVANCE ONLINE PUBLICATION | www.nature.com/naturephotonics8
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
Methods
Materials synthesis. MAI, BAI and PEAI were synthesized by mixing
methylamine, butylamine and phenylethylamine (Sigma Aldrich) with equimolar
amounts of aqueous HI (Sigma Aldrich, stabilized) at 0 °C with constant stirring
under N
2
. MABr and BABr were synthesized with analogous procedures. The
organoammonium halides (MAI, BAI, PEAI, MABr and BABr) were washed
with an ethanol:ether mixture and rotovaped several times to remove the
HI stabilizer.
Perovskite film deposition and characterization. Poly-TPD (or PVK) was
dissolved in chlorobenzene at a concentration of 6 mg ml
–1
. Poly-TPD (or PVK) was
spin-coated on ITO at 1,000 r.p.m. for 60 s, followed by thermal annealing at 150°C
(or 120 °C for PVK) for 20 min. The poly-TPD layer was treated by O
2
plasma for 2 s
to improve wetting. The MAPbI
3
or MAPbBr
3
precursors were dissolved in DMF
(Sigma Aldrich, 99.8% anhydrous) at concentrations of 0.43 M and 0.3 M,
respectively. Then, different molar amounts of BAI and BABr were added into the
MAPbI
3
and MAPbBr
3
solutions, respectively, before spin-coating. The spin-coating
rate for the I-perovskite was 6,000 r.p.m., and toluene was dropped on the spinning
substrate at 5 s. The spin speed for the Br-perovskite was 4,500 r.p.m., and the
toluene was dropped on the spinning substrate at 4 s.
The AFM measurements were conducted in the tapping mode in a N
2
filled
glove box (Veeco di Innova). The SEM measurements were conducted with an FEI
XHR (extreme high resolution) SEM (Verios 460) using immersion mode. XRD
measurements were performed with a Bruker D8 Discover X-ray diffractometer with
Bragg-Brentano parallel beam geometry, a diffracted beam monochromator and a
conventional Cu target X-ray tube set to 40 kV and 40 mA.Absorption spectra were
measured using a Cary 5000 UV-Vis-NIR system (Agilent). The PL spectra were
measured using an FLS980 spectrometer (Edinburgh Instruments) with an
excitation wavelength of 380 nm.
PLQY measurement and calculation. The QY of the I- and Br-perovskite films was
measured using a PTI QuantaMaster 400 Steady State Fluorometer with Petite
Integrating Sphere. The excitation wavelengths for I- and Br-perovskite are 450 and
400 nm, respectively, with excitation intensities of 2.2 and 2.1 mW cm
–2
, respectively.
Time-resolved PL measurement. Time-resolved PL measurements were taken using
a Horiba DeltaFlex time-correlated single-photon counting system. The samples
were excited by a pulsed laser diode (DeltaDiode-Horiba) with a centre wavelength
of 406 nm, an excitation intensity of ∼4mWcm
–2
and a repetition rate that is less
than the reciprocal of the measurement range. The time resolution was determined
to be ∼100 ps from the instrument response function. All decay traces were taken at
5 nm increment and with constant run time of 10 s. The emission bandpass was
16 nm for all samples except for BABr:MAPbBr
3
(20:100 and 40:100) samples
where it was set to 1 nm to protect the detector from oversaturation.
Device fabrication and characterization. All perovskite films were dried at 70 °C for
5 min following spin-coating to fully dry the film and ensure the full reaction of the
precursors without affecting the morphology and grain size substantially. TPBi, LiF
and Al layers were sequentially thermally deposited on top of the perovskite film to
thicknesses of 60, 1.2 and 100 nm, respectively. The device area was 0.1 cm
2
.
LEDs were measured in N
2
using a homemade motorized goniometer set-up
consisting of a Keithley 2400 sourcemeter unit, a calibrated Si photodiode (FDS-100-
CAL, Thorlabs), a picoammeter (4140B, Agilent) and a calibrated fiber optic
spectrophotometer (UVN-SR, StellarNet).
NATURE PHOTONICS DOI: 10.1038/NPHOTON.2016.269 ARTICLES
NATURE PHOTONICS |www.nature.com/naturephotonics
© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.
