1
Colloidal Cs
1-x
FA
x
PbI
3
quantum dots derived from ligand-assisted cation exchange for
record-efficiency and segregation-free solar cells
Mengmeng Hao
1
, Yang Bai
1
*, Stefan Zeiske
2
, Long Ren
3
, Junxian Liu
4
, Yongbo Yuan
5
, Nasim
Zarrabi
2
, Ningyan Cheng
3
, Mehri Ghasemi
1
, Peng Chen
1
, Shanshan Ding
1
, Miaoqiang Lyu
1
,
Dongxu He
1
, Jung Ho Yun
1
, Yi Du
3
, Yun Wang
4
, Ardalan Armin
2
, Paul Meredith
2
, Gang Liu
6,7
,
Hui-Ming Cheng
6,8,9
, Lianzhou Wang
1
*
1
Nanomaterials Centre, Australian Institute for Bioengineering and Nanotechnology and
School of Chemical Engineering, The University of Queensland, St Lucia, Queensland 4072,
Australia.
2
Department of Physics, Swansea University, Swansea, SA2 8PP, Wales, UK.
3
Institute for Superconducting and Electronic Materials (ISEM), Australian Institute for
Innovative Materials (AIIM), University of Wollongong, Wollongong, New South Wales
2500, Australia.
4
School of Environment and Science, Centre for Clean Environment and Energy, Griffith
University, Gold Coast, Queensland 4222, Australia.
5
Institute of Super-Microstructure and Ultrafast Process in Advanced Materials, School of
Physics and Electronics, Central South University, Changsha, 410083, PR China.
6
Shenyang National Laboratory for Materials Science, Institute of Metal Research, Chinese
Academy of Sciences, 72 Wenhua Road, Shenyang 110016, China
7
School of Materials Science and Engineering, University of Science and Technology of
China, 72 Wenhua Road, Shenyang 110016, China
8
Shenzhen Geim Graphene Center, Tsinghua-Berkeley Shenzhen Institute (TBSI), Tsinghua
University, 1001 Xueyuan Road, Shenzhen 518055, China
9
Advanced Technology Institute (ATI), University of Surrey, Guildford, Surrey GU2 7XH,
UK
* Corresponding authors. Emails: l.wang@uq.edu.au (L.W.); y.bai@uq.edu.au (Y.B.).
2
Abstract
The mixed caesium and formamidinium lead triiodide perovskite system (Cs
1-x
FA
x
PbI
3
) in the
form of quantum dots (QDs) offers a new pathway towards stable perovskite-based
photovoltaics and optoelectronics. However, it remains challenging to synthesize such
multinary QDs with desirable properties for high-performance QD solar cells (QDSCs). Here
we report an effective ligand-assisted cation exchange strategy that enables controllable
synthesis of Cs
1-x
FA
x
PbI
3
QDs across the whole composition range (x: 0-1), which is
inaccessible in large-grain polycrystalline thin films. The surface ligands play a key role in
driving the cross-exchange of cations for the rapid formation of Cs
1-x
FA
x
PbI
3
QDs with
suppressed defect density. The hero Cs
0.5
FA
0.5
PbI
3
QDSC achieves a certified record power
conversion efficiency (PCE) of 16.6% with negligible hysteresis. We further demonstrate that
QD devices exhibit substantially enhanced photostability compared to their thin film
counterparts because of the suppressed phase segregation, retaining 94% of the original PCE
under continuous 1-sun illumination for 600 hours.
Main
Solution processed organic-inorganic lead halide perovskite materials with a common
formulation of ABX
3
(where A is an organic cation or Cs; B is commonly Pb or Sn; and X is a
halide) have shown promise in many applications including photovoltaics (PVs)
1, 2, 3, 4
, light-
emitting diodes (LEDs)
5, 6
, and lasers
7, 8
. This potential is in part due to a range of desirable
optoelectronic properties, such as large absorption coefficient, long carrier diffusion length and
low exciton dissociation energy. The rapidly increasing power conversion efficiencies (PCEs)
for perovskite solar cells (PSCs), which have been achieved with mixed methylammonium
(MA) and formamidinium (FA) A-site cations and mixed Br and I X-site anions, have matched
or surpassed other more established technologies such as multicrystalline silicon, cadmium
telluride (CdTe) and copper indium gallium selenide (CIGS) photovoltaics
9, 10, 11
. However,
device degradation due to MA degassing and photoinduced halide segregation remains a
critical hurdle for commercialization
12, 13, 14
. Cs
1-x
FA
x
PbI
3
without Br and volatile MA was
recently demonstrated as a promising perovskite formulation, showing enhanced moisture-,
thermal-, and photo-stability
12, 13, 14
. Unfortunately, solar cells using Cs
1-x
FA
x
PbI
3
perovskites
suffer from large open-circuit voltage (V
OC
) losses despite rubidium (Rb) or polymer
passivation
12, 14, 15
, and thus their PCEs still lag behind those of the state-of-the-art perovskites
containing MA and Br. Furthermore, it is still challenging to continuously tune the
3
compositions of Cs
x
FA
1-x
PbI
3
at room temperature
16
, and a considerably high percentage
(x≥0.15) of Cs incorporation can cause cation segregation under continuous light illumination
or with an applied bias
13, 17
. This hinders optical bandgap optimization and long-term
operational stability for tandems and colour-tuning of LEDs
13, 18, 19
.
A promising approach towards the mitigation of these challenges lies in the synthesis of
perovskites in the form of nanometre-sized quantum dots (QDs) or nanocrystals. Perovskite
QDs have shown notable advantages over their large-grain perovskite thin film (often called
bulk) counterparts. These advantages include: additional bandgap-tuning by utilising the
quantum-confinement effect
20
; multi-exciton generation (MEG)
21
; near-unity
photoluminescence quantum yield (PLQY)
20, 22
; and photon up- and down-conversion
23, 24
.
Thus, QDs offer great promise to surpass the Shockley-Queisser limit
25
and further boost the
efficiency of tandem devices. Another advantage of perovskite QDs is the colloidal synthesis
and processability using industrially friendly solvents at room temperature, which decouples
grain crystallization from film deposition and enables rapid low-cost manufacturing
26
. More
importantly, both colloidal CsPbI
3
and FAPbI
3
QDs exhibit significantly enhanced phase
stability at room temperature
26, 27
in comparison with their bulk counterparts, providing a new
strategy for improving the lifetime of PSCs. Perovskite QDs that are defect-tolerant have
shown great promise over traditional chalcogenide QDs when incorporated into PV devices
20
.
CsPbI
3
QD solar cells (QDSC) have delivered a larger open-circuit voltage (V
OC
)
26, 28
than thin-
film counterparts yielding a certified PCE of 13.4%
29
, which breaks the record efficiency of
PbS QDSCs. However, both CsPbI
3
and FAPbI
3
QDs are highly susceptible to moisture and
polar solvents during post-synthesis purification and surface treatment
26, 27, 30, 31
. This poses
significant challenges in effective ligand engineering and causes a significant increase in
surface defects, which limits the carrier transport and thus PCE enhancement.
Mixed-cation Cs
1-x
FA
x
PbI
3
QDs are more desirable than pure CsPbI
3
and FAPbI
3
QDs in terms
of stability and charge transport properties. This is believed to be due to the concomitant
incorporation of FA and Cs inducing entropic stabilization of the perovskite structure under
ambient condition
16
, and the fast rotation of FA resulting in enhanced orbital overlap and easier
polaron formation that lead to longer carrier lifetimes and reduced non-radiative
recombination
32, 33
. Unfortunately, the rational synthesis of such multinary QDs with desirable
optoelectronic properties remains a significant challenge considering the complex interplay
between the bulk thermodynamics of the solid solutions, crystal surface energies, and dynamics
of capping ligands in solution
34, 35
. Hence, only a few studies have been reported so far (e.g.,
4
droplet-based microfluidics, cation exchange) to explore the synthesis of Cs
1-x
FA
x
PbI
3
QDs
34,
35, 36, 37
, but their use as a single light absorbing layer in QDSCs resulted in deteriorated PCEs
36
and the PCEs of state-of-the-art QDSCs
38
still lag far behind that of the respective thin film
counterparts.
Here we report a ligand-assisted cation exchange strategy to synthesize high-quality Cs
1-
x
FA
x
PbI
3
QDs + in a time- and cost-efficient manner. We show that the surface ligands are
important for the formation and diffusion of A-site cation vacancies, which play a critical role
in driving the cross-exchange of cations. Compared to pure CsPbI
3
or FAPbI
3
QDs, the resultant
Cs
1-x
FA
x
PbI
3
QDs are far more stable in ambient air or polar solvents and exhibit significantly
lower trap density and longer carrier lifetime. Solar cells fabricated with colloidal
Cs
0.5
FA
0.5
PbI
3
QDs under ambient conditions deliver a certified steady-state PCE of 16.6%, a
new record for the QDSC category
11
. We also find a larger PLQY of 10% and thus reduced
non-radiative recombination in perovskite QD devices relative to their bulk counterparts, which
explains the smaller V
OC
losses observed in perovskite QDSCs. We further demonstrate that
the use of colloidal Cs
x
FA
1-x
PbI
3
QDs opens a new avenue towards segregation-free PSCs
showing enhanced stability under continuous light illumination beyond the reach of large-grain
perovskite thin films of identical compositions.
Ligand-assisted synthesis of Cs
1-x
FA
x
PbI
3
QDs
We first calculated the energies (ΔE
ex
) required for the formation of Cs
1-x
FA
x
PbI
3
(x=0.25, 0.50
and 0.75) (Supplementary Fig. 1) via A-site cation exchange, and the negative values suggest
that this reaction is thermodynamically favourable. To prepare the alloy Cs
1-x
FA
x
PbI
3
QDs,
CsPbI
3
and FAPbI
3
QDs were first synthesized using a modified hot-injection method and then
mixed in controlled molar ratios. We took the synthesis of Cs
0.5
FA
0.5
PbI
3
QDs as an example
to explore the feasibility and closely monitored the A-site cation exchange reaction between
the parent CsPbI
3
and FAPbI
3
QDs. When the parent colloidal solutions were purified twice
with methyl acetate before mixing, a large fraction of OA ligands were removed as indicated
by the significant reduction of the vibration bands at 1640 cm
-1
(–C=C– group) in the Fourier-
Transform Infrared (FTIR) spectra (Supplementary Fig. 2), forming an OA-less environment.
The cation exchange reaction was very slow, taking more than 24 hours as shown in Figure
1A and the as-prepared alloy sample was denoted as QD-OL. The slow reaction is in
accordance with a previous report
36
, which indicates the presence of a high kinetic barriers for
the cations to leave the starting QDs and diffuse into other QDs. We proposed that the OA
ligands could be the key to eliminating such barriers by solvating A-site cations of QDs as
5
highly mobile Cs-oleate and FA-oleate in the colloidal solution. To examine this hypothesis,
we intentionally kept more OA ligands in the parent colloidal solutions by reducing the
purification to once, forming an OA-rich environment. Encouragingly, we observed a rapid
completion (30 min) of the cation exchange reaction (Fig. 1B). We also observed a rapid cation
exchange reaction (colour change) by adding Cs-oleate in FAPbI
3
QD solution or adding FA-
oleate in CsPbI
3
QD solution as shown in Supplementary Fig. 3, while adding other salts such
as caesium (formamidinium) acetate and caesium (formamidinium) iodide did not lead to
cation exchange, which suggests the critical role of OA in solvating A-site cations of QDs and
validates the existence of mobile Cs-oleate and FA-oleate reaching an equilibrium in the
colloidal solution. The ligand density in each colloidal QD solution was estimated and
summarized in Supplementary Table 1. The as-synthesized alloy sample in the OA-rich
environment was named as QD-OR. Part of this sample was purified again and denoted as
QD*-OR. Interestingly, we note that the alloy QDs synthesized in the OA-rich environment
showed much higher PL intensity (Fig. 1C) and the corresponding PLQYs (Supplementary Fig.
4) are 2-3 times higher than that prepared in the OA-less environment. We further checked the
influence of different solvents on the cation exchange reaction. Supplementary Fig. 5 shows
the evolution of the PL emission peaks with time for both solvents (toluene and octane), which
both exhibit rapid cation exchange comparable to that in hexane. These results verify our
hypothesis that the noteworthy difference in ligand engineering is the key, which not only
promotes the cation exchange reaction, but also preserves high radiative efficiency by
suppressing the formation of surface defects as illustrated in the scheme (Fig. 1D).
By using this ligand-assisted cation exchange method, we demonstrate the fine-tuning of the
absorption onset and PL emission peak (Supplementary Fig. 6) of the alloy QDs from pure
CsPbI
3
to pure FAPbI
3
. The bandgaps of each QD composition are summarized in
Supplementary Table 2. All the alloy Cs
1-x
FA
x
PbI
3
QDs retain their perovskite structure, as
shown in the XRD patterns (Supplementary Fig. 7a). The zoomed-in view of the XRD patterns
(Supplementary Fig. 7b) indicates that the substitution of Cs
+
with FA
+
induces a gradual shift
in the diffraction peaks such as (200) towards a lower angle.
Scanning transmission electron microscopy (STEM) images (Supplementary Fig. 8a-e) reveal
that all the Cs
1-x
FA
x
PbI
3
QDs are uniform nanocubes with size around 14 ± 2 nm. To identify
the distribution of Cs and FA cations and understand the structural evolution in atomic
resolution among these samples, high angle annular dark-field STEM (HAADF-STEM)
characterizations were conducted, as shown in Supplementary Fig. 8f-j. The CsPbI
3
QDs in
![Figure 2. Morphology and crystal structure of Cs1-xFAxPbI3 QD. (A, B) HAADF-STEM images of Cs0.5FA0.5PbI3 QDs. (C, D) Representative atomic-resolution HAADF-STEM image and the derived color-coded image of Cs0.5FA0.5PbI3 QD. The image is viewed along the cubic [001] zone axis established from the lattice arrangement similar with the original CsPbI3 (or FAPbI3) cubic crystal structure; the green dots represent the atom column in the area that are uniform contrast distribution, the red dots and purple dots (red color represent the high contrast atoms and purple color represent the low contrast atoms) represent the atom columns in the area that are distinct contrast distribution. (E) The zoom-in view, (F) the atomic models, and (G) the simulated atomic resolution HAADF-STEM image for the region marked with a red square. All the experimental STEM images of the samples containing FA are distorted due to the fast decomposition of local crystal structure from FAPbI3 to PbI2 under e-beam illumination. (H) The simulated atomic resolution HAADF-STEM image and (I) the atomic models for the original view of Cs0.5FA0.5PbI3 QD along the [001] zone axis after re-construction. (J) The established crystal structure of the Cs1-xFAxPbI3 QD. Scale bar in E, G, H, 10 Å.](/figures/figure-2-morphology-and-crystal-structure-of-cs1-xfaxpbi3-qd-3vd8i7uc.webp)
