Suppressed ion migration in powder-based perovskite thick films using an ionic liquid

Article abstract of DOI:10.1039/d1tc01554k

While solution-processed halide perovskite thin films caused enormous attention when used in solar cells, thick films prepared by compressing perovskite powders are considered promising candidates for the next generation of X-ray detectors. However, X-ray detectors based on such powder-pressed perovskites typically suffer from relatively high dark currents, which were attributed to be caused by ion migration. Here we show that the dark current in 800 μm thick powder-pressed MAPbI₃-pellets can be reduced by a factor of 25 when using a passivated powder. The passivation was achieved by adding 1 mol% of the ionic liquid (IL) BMIMBF₄to the precursors MAI and PbI₂during the mechanochemical synthesis of the MAPbI₃powder. NMR verified the presence of the IL, and its impact on the excited state recombination dynamics was manifested in an increase in the photoluminescence (PL) intensity and a decrease in the monomolecular (trap-assisted) recombination rate, both by about one order of magnitude. By measuring the migration of a PL quenching front upon application of an electric field in a microscope, we determine an ionic diffusivity in the typical range of iodide vacancies in the non-passivated pellet. At the same time, we observe no such PL quenching front in the passivated pellet. Concomitantly, darkI-Vcurves are hysteresis-free, and light-soaking effects are absent, in contrast to non-passivated pellets. Thus, our work demonstrates the effect on the optical and electrical properties when passivating mechanochemically synthesized halide perovskite powders using an IL, which will facilitate the further development of powder-based perovskite X-ray detectors.

Full text of DOI:10.1039/d1tc01554k

Journal of
Materials Chemistry C
View Article Online
View Journal
PAPER
Suppressed ion migration in powder-based
perovskite thick films using an ionic liquid†
Cite this: DOI: 10.1039/d1tc01554k
b
a
Philipp Ramming,a Nico Leupold,
Konstantin Schotz,a Anna Kohler,
¨
¨
b
c
a
Ralf Moos,
Helen Gruninger
and Fabian Panzer
*
¨
While solution-processed halide perovskite thin films caused enormous attention when used in solar
cells, thick films prepared by compressing perovskite powders are considered promising candidates for
the next generation of X-ray detectors. However, X-ray detectors based on such powder-pressed
perovskites typically suffer from relatively high dark currents, which were attributed to be caused by ion
migration. Here we show that the dark current in 800 mm thick powder-pressed MAPbI3-pellets can be
reduced by a factor of 25 when using a passivated powder. The passivation was achieved by adding
1 mol% of the ionic liquid (IL) BMIMBF4 to the precursors MAI and PbI2 during the mechanochemical
synthesis of the MAPbI3 powder. NMR verified the presence of the IL, and its impact on the excited state
recombination dynamics was manifested in an increase in the photoluminescence (PL) intensity and a
decrease in the monomolecular (trap-assisted) recombination rate, both by about one order of
magnitude. By measuring the migration of a PL quenching front upon application of an electric field in a
microscope, we determine an ionic diffusivity in the typical range of iodide vacancies in the non-
passivated pellet. At the same time, we observe no such PL quenching front in the passivated pellet.
Concomitantly, dark I–V curves are hysteresis-free, and light-soaking effects are absent, in contrast to
non-passivated pellets. Thus, our work demonstrates the effect on the optical and electrical properties
when passivating mechanochemically synthesized halide perovskite powders using an IL, which will
facilitate the further development of powder-based perovskite X-ray detectors.
Received 3rd April 2021,
Accepted 7th July 2021
DOI: 10.1039/d1tc01554k
rsc.li/materials-c
in the laboratory using high-quality single crystals.9,10 However,
only sophisticated methods allow the growth of perovskite
Introduction
Halide perovskites have undergone a spectacular development in single crystals in a thick-film geometry that is required for
recent years so that perovskite solar cells are now on the edge of commercialization.11
commercialization.1 However, in the shadow of the solar cells,
An attractive alternative in terms of process technology is to
also other perovskite-based optoelectronic applications have transfer perovskites in powder form into thick layers, e.g., by
emerged, such as LEDs,2 memristors, and especially detectors compressing the powder, and to use these thick layers in X-ray
for high-energy radiation, like X-ray or gamma-rays.3–6
detectors.12 It is possible to produce halide perovskite powders,
The developments of the latter are particularly noteworthy. even on the kilogram scale per batch, for example by mechan-
For solar cells, lab efficiencies are now on par with the ochemical synthesis via ball milling.13–15 However, detectors
commercially established competitor, i.e., silicon.7 However, based on perovskite powder layers typically exhibit significantly
when used for direct-conversion X-ray detectors, halide perovs- higher dark currents than their single crystal counterpart
kites actually outperform commercially used semiconductor devices.16,17 These high currents limit the detector sensitivity
materials such as a-Se regarding their sensitivity.6,8 Highly and they often exhibit instability over time, which is a current
efficient perovskite X-ray detectors have so far been realized hurdle for commercialization of perovskite X-ray detectors.16 It
has been suggested that mobile ions are responsible for the
high dark current,12,16 and, in the past, attempts have been
made to reduce and temporally stabilize the dark current, e.g.,
by using low-dimensional perovskite materials.18,19
In the field of perovskite solar cells, reducing ion migration
has been a focus of interest for quite some time.20–23 For this
purpose, various strategies were applied, such as compositional
engineering,24 the use of low dimensional perovskite materials,25
a Soft Matter Optoelectronics, University of Bayreuth, Bayreuth, 95440, Germany.
E-mail: fabian.panzer@uni-bayreuth.de
b Department of Functional Materials, University of Bayreuth, Bayreuth, 95440,
Germany
c Institute for Molecules and Materials, Radboud University, Heyendaalseweg 135,
6525 AJ Nijmegen, The Netherlands
† Electronic supplementary information (ESI) available. See DOI: 10.1039/
d1tc01554k
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. C

View Article Online
Paper
Journal of Materials Chemistry C
interface modifications,26 and in particular the introduction of iodide (MAPbI3) during the mechanochemical synthesis. The
additives such as potassium. While some of the additives, such resulting passivated powder is then pressed into a thick film.
as choline chloride, mainly showed a passivating effect at the We systematically investigate the luminescence and charge trans-
interface between the perovskite and the transport layer,27,28 ionic port properties of non-passivated and passivated powder pellets
liquids (IL) have recently been shown to passivate the bulk of the using various optical and electrical measurement methods. With
perovskite layer.29 In this context, the term ‘‘passivation’’ refers to the addition of the IL, we find a suppressed light soaking effect
chemical passivation leading to a reduction in defect trap density.30 and a significant reduction of the non-radiative recombination
This resulted in improved power conversion efficiencies and rate in the powder pellets, corresponding to a reduced trap
excellent stabilities of solar cells fabricated from such passivated density. By photoluminescence (PL) microscopy investigations,
perovskite thin films.31–33 For mechanochemically synthesized we can identify a suppression of ion migration leading to
perovskite powders, we recently demonstrated successful chemical hysteresis-free IV curves of the passivated powder pellets. In
passivation by adding potassium iodide to the milling process.34 particular, the suppressed ion migration results in a significant
This reduced the non-radiative decay and led to improved power reduction and substantially increased long-term stability of the
conversion efficiency of solar cells, when the passivated powder was dark currents. Our work thus demonstrates an important strategy
dissolved and used as stock solution in the solution processing.34 to better exploit the optoelectronic potential of powder-pressed
However, it is not yet clear whether the defect-passivating pellets in the future.
effects resulting from adding an organic compound, as it is a
promising strategy in case of perovskite thin films for solar
cells, also apply to powder-pressed thick films. This is, however,
Results and discussion
crucial for highly-efficient powder-based X-ray detectors.
We mechanochemically synthesized MAPbI3 powder by ball
In this work, we investigate whether a successful reduction milling, as described in more detail in ref. 34. In brief, the
of the defect trap density and a subsequent reduction of the reactant powders PbI2 and methylammonium iodide (MAI) are
dark current can be achieved in powder-based perovskite thick placed in a milling jar, where milling balls and a small amount of
films. We approach this by directly adding the ionic liquid cyclohexane, used as milling agent, are present. During planetary
1-butyl-3-methylimidazolium tetrafluoroborate (BMIMBF4) to the ball milling, MAPbI3 synthesis is enabled by the mechanical
precursors of the model halide perovskite methylammonium lead energy release when the milling balls are impinging on the
Fig. 1 (a) Schematic illustration of the mechanochemical synthesis of MAPbI3 powder by ball milling and the subsequent pressing step using a
mechanical press. For the passivated MAPbI3 powder, 1 mol% BMIMBF4 is added to the MAI and PbI2 before milling. (b) XRD patterns of the reference and
the passivated powder containing the ionic liquid. (c) Chemical structure of BMIMBF4 together with 1H MAS NMR spectra of the passivated powder. The
corresponding H-atoms of the BMIMBF4 in the spectrum are highlighted as colored bars. (d) SEM images of the reference and passivated powder and
corresponding powder-pressed pellet surfaces. The scale bar applies to all four images.
J. Mater. Chem. C
This journal is © The Royal Society of Chemistry 2021

View Article Online
Journal of Materials Chemistry C
Paper
reactant powders at the inner walls of the milling jar (Fig. 1a). We homemade mechanical press (Fig. 1a), which is described in
prepared one batch of non-passivated MAPbI3 powder (referred to more detail in ref. 38. As the NMR measurement has already
as ‘‘reference’’) and in parallel, under the same process conditions, confirmed the presence of the IL, the lack of a detectable signal
yet with 1 mol% BMIMBF4 added to the milling jar, a second batch in EDX indicates that there are no local accumulations of IL in
of MAPbI3 powder (referred to as ‘‘passivated’’). We used 1 mol% of the pellet but that the IL is homogeneously distributed within
BMIMBF4, as similar amounts were found to lead to a prominent the pellet (Fig. S4, ESI†). Fig. 1d shows top view scanning
passivation effect in solution-processed thin film solar cells.29
electron microscopy (SEM) images of both powder and pellet
The X-ray diffraction (XRD) pattern for both the reference of the reference and the IL passivated MAPbI3 samples (see
and the passivated powder show only peaks associated with Fig. S5, ESI† for additional images). From this, it is evident that
MAPbI3 in the tetragonal phase (Fig. 1b).35 In particular, we both the morphology of the powder and of the pellet surface do
cannot observe any additional peaks, e.g., at 12.61, due to the not change significantly by adding IL in the milling process.
presence of crystalline PbI2.36 This indicates that even when the Rather, both powders exhibit the typical morphology known
IL is added, the MAPbI3 prepared by the mechanochemical from previous work, with powder particle sizes in the range of
synthesis remains completely phase pure, at least with respect 300–800 nm,34 and mainly compact surfaces after pressing at
to the sensitivity achievable by XRD.
100 MPa at room temperature for about 30 minutes.38 Thus, we
To investigate the role of the liquid form of the IL in the conclude that the IL added during ball milling is present in the
mechanochemical synthesis, we prepared another batch of MAPbI3 powder in the expected amount and leaves the synth-
MAPbI3 powder using the same synthesis parameters as for esis and the morphology of the MAPbI3 powder unchanged.
the MAPbI3 powders described before, but added 1 mol% of the
To investigate whether the IL reduces the non-radiative
solid salt choline chloride during the ball milling. In the past, decay in the mechanochemically synthesized MAPbI3 powder,
choline chloride was added to perovskite precursor solutions we characterized the reference and passivated pellets using PL
and lead to a passivation of defect states in the solution- measurements. PL is well suited for this purpose,39 since
processed thin films.28 However, while the XRD pattern of the charge carrier recombination in lead halide perovskites can
powder containing 1 mol% BMIMBF4 only showed the desired be described by the rate equation,40
peaks associated to MAPbI3, in the XRD pattern of the powder
dn
¼ ꢀk1n ꢀ k2n2 ꢀ k3n3;
(1)
containing 1 mol% choline chloride two additional peaks at
10.21 and 12.61 are visible (see Fig. S1, ESI†). This demonstrates
that for the synthesis parameters used here, the MAPbI3 synth-
esis of the powder containing choline chloride is not phase-
pure, while adding IL leads to a full MAPbI3 synthesis. Here, the
liquid form of the IL appears to be helpful to ensure a sufficient
distribution of the additive during the ball milling, so that a
complete perovskite synthesis is maintained.
Complementary to the XRD experiments, we performed
high-resolution, solid-state 1H magic angle spinning (MAS)
nuclear magnetic resonance (NMR) experiments to identify
and to quantify the ionic liquid in our passivated powder
sample (Fig. 1c). While the full 1H MAS NMR spectrum is
shown in Fig. S2 (ESI†), a magnified section on the low intensity
1H signals is depicted in Fig. 1c. The 1H chemical shifts of these
signals, indicated by colored vertical bars in Fig. 1c, are in good
agreement with reported values for BMIMBF4.37
dt
with n being the photoinduced charge carrier density, k1 is the
rate constant for monomolecular (trap-assisted) recombina-
tion, k2 is the rate constant for bimolecular recombination,
and k3 that for Auger-recombination. In lead halide perovskites,
k2 and k3 can be considered to be material constants and are
thus sample-independent, while k1 depends on the trap density
of the investigated sample. Note that in eqn (1), electron and
hole density are assumed to be equal, and charge carrier
diffusion and photon recycling are neglected (see ESI,† Section
S1 and Fig. S6 for details). Further, k1 and k3 are non-radiative,
while k2 is radiative in halide perovskites. The PL quantum
yield at a given carrier density is therefore determined by,41
k2n
FPLðnÞ /
:
(2)
k1 þ k2n þ k3n2
The only obvious difference between the liquid spectrum of Thus, for low to medium charge carrier densities (i.e., k2n + k3n2
the pure component and our experiment is the splitting of is not significantly larger than k1), an increase in PL intensity
the 1H signals of the imidazole ring (Fig. 1c) between 7 and under identical experimental conditions is due to a reduction
9 ppm.37 This splitting indicates that the IL interacts with the in k1.41 The PL measurements presented in this work are
perovskite surface via the imidazole moiety, changing the carried out in such a low to medium charge carrier density
electronic environment of the corresponding H atoms.
Furthermore, a quantitative analysis of the low intensity 1H to a charge carrier density of about 1016 cmꢀ3 42
signals of BMIMBF4 relative to the two strong resonances for the Fig. 2a shows the steady-state PL of the reference (black line)
regime, where we used an excitation fluence that corresponds
.
NH3 and the CH3 group of MAPbI3 indicates that an amount of and the passivated pellet (red line), both measured in the same
about 0.8 mol% IL is present in the sample. This is in good experimental geometry, i.e., in the focal plane of a PL micro-
agreement with the amount of 1 mol% that was added in the scopy setup (see Experimental section for details). The PL
milling process.
intensity of the passivated pellet is increased by more than
We pressed B0.8 mm thick pellets (Fig. S3, ESI† for images) one order of magnitude compared to the reference pellet.
by filling the powders in a mold and compacted them using a Fig. 2b shows the normalized time-resolved PL (TRPL) curves
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. C

View Article Online
Paper
Journal of Materials Chemistry C
passivated sample, i.e., the monomolecular decay rate is nearly
one order of magnitude lower. This ten-fold reduced value of
k1,pass indicates a concomitantly reduced trap density in the
MAPbI3 powder due to the addition of the IL.
To conduct the TRPL measurements, both samples were
exposed to the excitation pulse only briefly (about 10 s). How-
ever, when the samples were exposed to the exciting laser for a
longer period, we observed that the PL intensity of the reference
sample changed. To investigate this further, we excited the
pellets with the excitation laser of our TRPL setup, i.e., with
68 ps laser pulses at a frequency of 1 MHz under ambient
conditions. Fig. 2c shows the evolution of PL intensity within
the first 10 minutes of illumination for the reference and the
passivated pellet. The PL intensity of the reference sample
increases by about 25% in the first 90 seconds and decreases
to 55% of the initial value thereafter. In contrast, no initial PL
increase occurs in the passivated sample, and the PL intensity
drops only slowly during the 10 minutes of illumination to
about 96% of the initial value.
A temporal increase in PL intensity during illumination of
halide perovskites is well known and has been referred to as
‘‘light soaking effect’’.44 This light soaking effect was found to
be more pronounced in samples with relatively high non-
radiative decay rates, i.e., relatively high trap densities. Motti
et al. systematically investigated the light soaking effect and
attributed an initial PL increase to light-induced annihilation
of iodide Frenkel defect pairs.45 Furthermore, they associate a
PL intensity drop over time with the formation of I2, which is
formed when two iodide ions come together in the perovskite.
For both the Frenkel Pair annihilation and the I2 formation
process, ion migration was found to play a crucial role, so that
accordingly, light soaking was linked to ion migration in the
perovskite.44 We consider that this process might also account
for the PL decrease in the reference pellet in Fig. 2c. The time
scale of the light soaking effect, i.e., seconds to minutes, is
consistent with the time scale of ion migration. It is also
significantly larger than the time scale of other effects such
as electronic transport and recombination or trap filling, which
occur in the range of 100 ns–100 ms.45
Fig. 2 (a) Steady-state PL spectra of the reference (black) and passivated
(red) pellets normalized to the maximum intensity of the passivated pellet.
(b) Time-resolved photoluminescence decay curves with fits based on
eqn (1) (blue lines). A pulsed laser illuminated the pellets with 1 MHz
repetition rate, a wavelength of 560 nm, and a fluence of 0.1 mJ cmꢀ2
for a total of 10 seconds. An alternative version of (b) to fully represent the
baselines and the instrument response function is provided in Fig. S6 (ESI†).
(c) Time-trace of the PL intensity for illumination conditions as in (b).
DeQuillets et al. associate the migration of ions out of the
illuminated region with the light soaking effect, and by time-
dependent PL experiments at different temperatures, they
found a corresponding activation energy of about 0.2 eV.46
This value agrees well with the typical activation energies of
0.1–0.3 eV for iodide ion migration in MAPbI3.47–49 Further-
more, a significantly reduced light soaking effect was found
of the reference (black) and the passivated pellet (red), when passivating mixed perovskites with potassium, consistent
together with fits (blue line) of simulated transient PL with our results.39 To summarize the findings of Fig. 2, we
decay curves using eqn (1). The PL decay is significantly slower observe an increase in PL intensity and lifetime upon inclusion
in the passivated pellet than in the reference pellet. To fit the of the IL in the powder, as well as an absence of any light-
measured transients, we set k2 = 6.8 ꢁ 10ꢀ10 cm3 sꢀ1 and k3 = soaking effect. We interpret this as an indication for the
10ꢀ28 cm6 sꢀ1, in accordance with literature,43 and estimated chemical passivation of defect traps by the IL.
the initial charge carrier density n(t = 0) based on the laser
We note that we have introduced solid potassium into the
fluence used in the experiments (see Experimental section and milling process of MAPbI3 in a previous work leading to a
ESI,† Section S1). From the fit, we obtain k1,ref = 2.2 ꢁ 107 sꢀ1 passivation of the perovskite powder.34 In addition, Abdi-Jalebi
for the reference sample and k1,pass = 3.0 ꢁ 106 sꢀ1 for the et al. used the same potassium salts for the passivation of
J. Mater. Chem. C
This journal is © The Royal Society of Chemistry 2021

View Article Online
Journal of Materials Chemistry C
Paper
halide perovskite thin films, which was also accompanied by an towards the negative electrode. This dark front reaches the
inhibition of halide migration.39 Therefore, the liquid form of center of the channel after approx. 50 s. After 165 s, the area of
BMIMBF4 appears not to be necessary to achieve passivation of the reduced PL extends over the entire width of the channel.
the perovskite.
Concomitantly the PL intensity in the vicinity of the positive
To further clarify whether the temporal PL changes in the electrode appears to become brighter again. PL brightening in
reference pellet in Fig. 2 are due to ion migration and whether the vicinity of the electrode, was also identified in some
it is reduced by passivation, we investigate our pellets by PL samples by Li et al.52 They suggested that the additional
microscope measurements under an applied electric field. PL vacancies annihilate with iodide interstitials and thus reduce
imaging microscopy has proven in recent years to be a valuable non-radiative recombination of electrons and holes.
method to directly monitor and study ion migration in halide
Fig. 3b shows the temporal evolution of the PL intensities
perovskites.50,51 Thus, we evaporated Au-electrodes, laterally averaged over 10 mm thick horizontal stripes parallel to the
arranged on top of the reference and IL passivated pellets. electrodes, along the positive electrode, in the middle of the
Fig. 3a shows PL microscopy images from the top surface of the channel, and along the negative electrode (as indicated in
pellet between the electrodes at the beginning, 50 s, and 165 s Fig. 3a by the colored stripes for the 0 s passivated sample).
after applying an electric field of 67 V cmꢀ1
.
For each position, the PL intensity is normalized to the value at
The reference pellet shows a relatively homogeneous PL t = 0 s. For the reference sample, it is evident that the PL
intensity throughout the 150 mm wide channel at t = 0 s. After intensity drops intensively at the positive electrode when apply-
applying the electric field, the PL intensity reduces in the area ing the electric field, to only 40% of the initial PL intensity after
along the positive electrode. Over time, the area with the about 10 s. After that, the PL intensity recovers continuously,
reduced PL intensity expands so that a dark front moves reaching values just above the initial value after 165 s. In the
Fig. 3 (a) PL microscope images of the reference (top) and passivated (bottom) perovskite pellets (channel length 150 mm), 0 s (left), 50 s (middle) and
165 s (right) after applying an electric field of 67 V cmꢀ1 and with an illumination intensity of 42 mW cmꢀ2. The plus and minus signs indicate the
polarization of the gold electrodes (dark area at the top and bottom). In the reference sample, a dark front advances from the positive to the negative
electrode, with its position is indicated with orange dashed lines. To the left of each PL image of the reference pellet, a vertical cross-section of PL
intensity is indicated (yellow). (b) Evolution of PL intensity at different channel positions (horizontal cross-sections), for reference (top) and passivated
(bottom) pellet, normalized to their value at t = 0 s. The accumulated intensity is calculated from the green (close to the positive electrode), red (mid-
channel), and blue (negative electrode) boxes shown in (a). The same area was used for the passivated and the reference sample. (c) The temporal
evolution of position x of the dark front within the channel, extracted by horizontal cuts (blue circles) or directly from the PL images (black squares),
together with a fit (red line) using eqn (4). The graph on the right shows the relative drop of the intensity after the dark front moves through a position x.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. C

View Article Online
Paper
Journal of Materials Chemistry C
middle of the channel, the PL intensity remains relatively intensity contrast between the two sides of the front fades as
constant within the first 50 s and then decreases by about the front advances, in accordance with a diffusive behavior.
15% within approximately 10 s. Similarly, after approximately
120 seconds the intensity at the negative electrode exhibits a
step-like drop by approximately 5% within 10 seconds. These
distinct intensity drops after B50 and B120 s in the middle
and at the bottom of the channel, respectively, correspond to
the dark front reaching the respective channel position (also
see Fig. S7, ESI†).
While the reasons for this are not fully understood yet, it
appears that the measurement conditions, such as illumination
intensity or humidity, as well as the morphology and polycrys-
tallinity of the perovskite influence the profile of the dark
front.53 Based on the qualitative similarity of the field-induced
PL reduction that we observe compared to that reported by Li
et al., we consider that the same mechanism applies.
In previous work, the occurrence of a reduced PL intensity
and the propagation of a dark front was attributed to the
migration of positively charged iodide vacancies or neutral
iodide Frenkel defects towards the negatively charged
electrode.52 The properties of the front reportedly depend on
parameters such as the morphology of the sample, applied
field, humidity, and laser intensity.53 Li et al. analyzed the PL
quenching dynamics in detail by PL microscopy and associated
current measurements.52 They consider a MAPbI3 perovskite
film between two electrodes that contains a certain density
of positive iodide vacancies and a larger density of negative
iodide interstitials, so that the perovskite can be considered as
p-doped. The central element of their model is the notion that
the applied voltage causes positively charged iodide vacancies
to drift towards the negative electrode. This would start a local
solid-state reaction that creates abundant non-radiative recom-
bination centers.52,54 In this way, the drift of iodide vacancies
would be associated with a reduction in radiative electron–hole
recombination, and thus a reduced PL intensity. Hence, a
continuously advancing dark front is associated with the drift
of the iodide vacancies. In the model presented by Li et al., the
velocity v of the front reduces as it propagates through the
channel according to52
Thus, we use the expression derived by Li et al. to analyze the
data of Fig. 3a. We identify the position of the dark front using
two independent approaches. First, we quantify x(t) by optically
localizing the position of the dark front in the channel directly
from the microscopy images at different times after the electric
field was applied (e.g., see orange dashed lines in Fig. 3a and
Fig. S7, ESI†). At longer times and thus larger distances from
the positive electrode, the precise determination of the front
position x becomes more and more difficult as the stepwise
reduction in PL intensity becomes less pronounced (Fig. S8,
ESI†). To improve the accuracy in determining the front posi-
tion from our data, we exploit the signature of the front, i.e., the
steep intensity reduction in the edge of the front. Therefore, we
consider the integrated PL intensity in horizontal slices at the
different channel positions x as a function of time and extract
the time at which the front arrives from the inflection point of
the PL intensity transient. This quantitative analysis further
allows us to determine the change in relative intensity directly
before and after the front passes through (see Fig. S8, ESI† for
details about the analysis approach).
Fig. 3c shows the evolution of x(t), as extracted using both
analysis approaches, together with the strength of relative
intensity drop at a certain channel position. We fit x(t) in
Fig. 3c using eqn (4) in the time range between 15 s and
100 s and obtain v0 = 4.6 mm sꢀ1 and g = 8.2. Using
mCV
ꢀ
ꢁꢀ1=2
2g
d
vðtÞ ¼ 1 þ v0t
v0
(3)
v0 ¼ ðg þ 1Þ
, we determine the mobility of the ions as
d
Here d and t are the channel length and the propagation time.
p0 mCV
mC = 7.5 ꢁ 10ꢀ8 cm2 Vꢀ1
s
ꢀ1. From the Einstein relation
The initial velocity v0 ¼
depends on the ionic mobility
(electrical mobility equation) D = mkBT/e, the diffusion constant
p1
d
of the ions results as Dion = 1.9 ꢁ 10ꢀ9 cm2
s
ꢀ1. This value
mC, the voltage V and the initial hole density p0 and the hole
density in the dark region, p1. The quantity g is defined as
agrees well with the typical values derived by PL microscopy
methods that are in the range of 10ꢀ8–10ꢀ11 cm2 sꢀ1 57
.
p0
g ¼ ꢀ 1. By integrating eqn (3) over time, the temporal
In contrast to the reference pellet, the passivated pellet
shows no temporal change in PL intensity along the entire
width of the channel and over the entire measurement period
of 165 s. In particular, we do not observe any PL quenching,
formation of a dark PL front (Fig. 3a), or intensity dynamics
directly at the electrode (Fig. 3b). This is consistent with
observations from Bai et al., where passivation of solution-
p1
evolution of the distance x of the dark front from the positive
electrode is given by:
!
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
d
2g
d
xðtÞ ¼
1 þ v0t ꢀ 1
(4)
g
The evolution of the PL intensity we observe in the powder- processed halide perovskite thin films using an IL led to
pressed MAPbI3 in Fig. 3 qualitatively resembles that described temporally stable PL microscopy images.29 Thus, based on
by Li et al. for a solution-processed film, though we find a less our findings from the PL microscopy investigations, it becomes
pronounced reduction of the PL intensity.
A moderate clear that the addition of IL in the mechanochemical synthesis
reduction rather than a full quenching of the PL intensity upon of MAPbI3 passivates the perovskite, which significantly
applying an electric field has also been found in other reduces ion migration in the powder pellet.
studies,49,55,56 and it was reported that the dark PL front does
Regarding the passivating mechanism, we recall that the PL
not always advance over the whole channel.52 Rather, the dynamics from Fig. 3c could be described well in a quantitative
J. Mater. Chem. C
This journal is © The Royal Society of Chemistry 2021

View Article Online
Journal of Materials Chemistry C
Paper
way based on the model of Li et al., which suggests that ion depend on the properties of the counter anion and on the
migration in our powder pellet takes place primarily through hydrophilicity of the IL.29
iodide vacancies.52 This is also supported by a recent work from
We now address the question of how the ion migration
our group, where we investigated the electrical conductivity of impacts the electrical properties of the pellet, and in particular
MAPbI3 films based on the same powder as used here.58 its dark current. Therefore, we measured the IV characteristics
Analyzing the electrical conductivity as a function of iodine of the reference and the passivated samples in the dark
partial pressure by defect chemical modeling, we find a (Fig. 4a). We first measured the current when sweeping
relatively high iodide vacancy concentration in the order of the voltage in forward direction, i.e., from ꢀ10 V to 10 V,
1018 cmꢀ3 58
.
Moreover theoretical works also proposed that followed by a backward scan to ꢀ10 V with the same scan rate
iodide ions migrate in MAPbI3 via iodide vacancies.48,59 Overall, (400 mV sꢀ1.) For the reference sample, we observed a hyster-
these different findings thus allow to conclude that ion migra- esis loop in both the negative and the positive voltage range.
tion within the reference powder pellet takes place through We repeated the IV scans with scan rates between 0.4 V sꢀ1 and
iodide vacancies.
10 V sꢀ1, and found the hysteresis to become more pronounced
The results of the 1H NMR analysis (Fig. 1c) indicate that the for decreasing scan rates (Fig. S9, ESI†). This is in contrast to
IL interacts with the perovskite surface via the imidazole the passivated pellet, where we could not observe any hysteresis
moiety, most likely at positively charged defects, such as iodide behavior in the IV curves for all scan rates used in our experi-
vacancies, as also discussed in literature.39,60 This is further ments (Fig. 4a and Fig. S9, ESI†).
supported by computer simulations61 as well as X-ray photo-
The hysteresis in the IV scans of perovskite pellets with
electron spectroscopy and Fourier-transform infrared spectro- symmetrical contacts was, e.g. also found by Yang et al., who
scopy measurements.60 The interaction between the IL and the assigned it to the influence of migrating ions, notably in the
iodide defects immobilizes the latter, causing the observed form of iodine vacancies, in the pellet.62,63 An absence of
inhibition of iodide ion migration.
hysteresis in the IV scans in Fig. 4a thus indicates slower ion
In passing, we note that it is reasonable to assume that the migration in the passivated pellet. A reduced dark current also
type of interaction and the strength of the passivating effect occurs in the time-dependent current measurements, where we
Fig. 4 (a) Current–voltage characteristics of the reference (black) and the passivated (red) pellet. The inset shows a photograph of a perovskite pellet
with the lateral gold electrodes (width of 1 mm and channel length of 150 mm). (b) Temporal evolution of dark current within the first 12 s (b), and within a
longer time range (c) of 36 000 s (i.e., 10 hours) upon applying a voltage of 10 V. The arrows indicate the time when the current increases considerably. (d)
Same as in (c), but the reference dark current was divided by a factor of 25, while the time axis of the passivated sample was scaled by dividing by 29.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. C

View Article Online
Paper
Journal of Materials Chemistry C
applied 10 V to the reference and passivated pellets (Fig. 4b). one electrode through the perovskite channel towards the
The passivated pellet shows a current of 2.5 nA directly after opposite electrode. This hypothesis is strengthened by the
applying the voltage, which remains quasi constant during the observation in Fig. 4c that the time scale, after which the main
measurement period (Fig. 4b bottom). In contrast, when the rise of the dark current occurs, i.e., about 500 s (E8 minutes),
voltage is applied to the reference pellet, we observe a signifi- agrees roughly with the time the ions need to migrate through
cantly higher initial current of 55 nA, which further rises steeply the channel, identified as 165 s (E3 minutes, Fig. 3). This
and stabilizes after about 8 s at 73 nA (Fig. 4b top). Fitting the hypothesis also agrees with the significantly delayed increase in
current rise using an exponential gives a rise time of about 1 s current in the passivated pellet, where a reduced ion diffusion
(Fig. S10, ESI†). This time scale agrees with the typical values due to passivation with the IL would result in a correspondingly
found in potentiostatic measurements of MAPbI3 at room delayed reaction at the electrode. It should be noted here that
temperature,27 and it was associated with ion migration pro- similar behavior was observed years ago for classical inorganic
cesses in the perovskites.64 The absence of this rise dynamic in titanate perovskites that are applied for capacitors. Under
the passivated sample thus further supports the deduction that applied voltages, the electrical field causes the positively
the addition of the IL leads to a strong reduction of ion charged oxygen vacancies to move from one electrode to the
migration in MAPbI3 and to a passivation.
other, leading to a deterioration of the capacitors.70 Higher
We also measured the evolution of dark current for both anion vacancy concentrations, e.g., obtained by iron acceptor
pellets in a nitrogen atmosphere over a longer time range of doping, and higher temperatures that lead to higher oxygen
10 hours, which is shown in a double logarithmic representa- (vacancy) diffusion coefficients support the oxygen movement.
tion in Fig. 4c. The current of the reference pellet increases only This effect was carefully studied optically or by electrocolora-
moderately in the first 500 s under bias, then increases by tion by the groups of Waser71 and Maier,72 respectively.
almost one order of magnitude between approx. 500 s
and 10 000 s, and finally approaches a constant value of approx.
6.4 ꢁ 10ꢀ7 A for the longest measured times. In contrast, the
current of the passivated pellet remains almost constant at
3 ꢁ 10ꢀ9 A within the first 10 000 s, and then starts to increase
continuously to 4.4 ꢁ 10ꢀ9 A after 36 000 s. It is noteworthy
that, for t o 500 s, the current in the passivated pellet is
reduced by a factor of 25 compared to the current in the
reference pellet. Similarly, the onset of a current rise in
the passivated pellet is delayed by a factor of 29 compared to
the reference pellet. In fact, the two current profiles in Fig. 4c
have a remarkably similar qualitative shape. This becomes
evident when the current profiles are scaled accordingly in
terms of both current and time (Fig. 4d). This agreement
suggests that the increase in current in the passivated pellet
at long times has the same physical origin as in the reference
pellet. Thus, we conclude that the passivation of the MAPbI3
powder with the IL does not lead to a full suppression but to a
significant delay of the process causing the current drift by a
factor of roughly 27 ꢂ 2.
Conclusion
In summary, we investigated in detail the impact of adding the
ionic liquid BMIMBF4 to the mechanochemical synthesis of the
halide perovskite MAPbI3. We find that adding 1 mol% of the IL
does not change the crystal structure and the morphology of the
mechanochemical synthesized perovskite powder. Compared to a
reference MAPbI3 powder pellet, the passivated powder pellet
shows, however, a reduction of the non-radiative decay rate by
nearly one order of magnitude and exhibits no light soaking and
PL degradation during illumination. Similarly, rise dynamics of
the dark current on the time scale of 1 s, which occur in the
reference pellet when applying an electrical field, are absent in the
passivated pellet. This behavior is also reflected in hysteresis-
free IV curves of the passivated pellet when sweeping between
ꢀ10 and 10 V. PL microscopy under an electric field demonstrates
that no PL darkening fronts are present within the passivated
powder pellet, suggesting a significantly reduced ion migration.
This reduced ion migration leads to a reduction of dark current by
a factor of 25 compared to the dark current of the reference
sample. Due to the reduction of ion migration, an increase of dark
current on the hours-time scale, most likely due to an electro-
chemical reaction of the Au electrode with iodide ions, is delayed
by a factor of 29 in the passivated powder pellet. Our work thus
contributes to gain a more fundamental understanding regarding
how the optical and the electrical properties of powder-based
perovskite thick films change by the addition of an ionic liquid.
A detailed analysis and interpretation of the current drifts
observed here is beyond the scope of this work and will have to
be addressed in the future. Nevertheless, we would like to
comment on possible mechanisms that might explain our
observations. Wu et al. suggested that an increase in dark
current on long time scales could be associated with reactions
of the migrating ions at the electrode. Such reactions would
modify the electrode interface,65 such as to facilitate charge
extraction, thus increasing the current flow through the chan-
nel. Different electrochemical reactions of MAPbI3 with gold
that involve mobile iodide ions have been reported for voltages
above 0.8 V,66–69 independent of the perovskite thickness.66
Considering that we applied a voltage of 10 V in our experi-
Experimental section
Mechanochemical synthesis
ments, such a reaction could easily occur in our sample. The The precursors MAI and PbI2 (TCI Deutschland GmbH) were
scenario for our reference pellet would then be that mobile weighted in a stoichiometry of 1 : 1 (1.59 MAI and 4.61 g PbI2)
iodine ions, traveling by a vacancy diffusion mechanism from and put in an 80 mL ZrO2 milling jar containing ZrO2 milling
J. Mater. Chem. C
This journal is © The Royal Society of Chemistry 2021

View Article Online
Journal of Materials Chemistry C
Paper
balls with a 10 mm diameter. Cyclohexane was added as the chemical shift was referenced using adamantane for 1H
milling agent. For the passivated powder, additionally 1 mol% (1.85 ppm) as secondary reference. The 1D 1H MAS NMR
with respect to Pb atoms of BMIMBF4 was added to the spectrum was recorded at a MAS frequency of 35 kHz. The
reactants. The milling was performed in a Fritsch Pulverisette recycle delay was set to 85 s to reach full relaxation (1H T1 B 17 s
5/4 planetary ball mill at 400 rpm. After 5 min of milling, the determined by 1H saturation recovery experiments).
rotation was paused for 20 min to cool down the jar. The
PL measurements
procedure was repeated 10 times. The cyclohexane was
removed by evaporation before the powder was sieved with a
63 mm mesh size at ambient conditions.
The time-correlated single-photon counting (TCSPC) measure-
ments were performed with a PicoQuant MT200 system on an
inverted microscope. The samples were excited with a 560 nm
diode Laser (PicoQuant LDH-D-TA-560) with a frequency of
1.0 MHz and a pulse width of 68 ps. The laser beam was
focused on the sample with an Olympus Objective with
4ꢁ magnification and a numerical aperture of 0.1, leading to
a spot diameter of about 6 microns in the focal plane. The
excitation fluence was set to 100 nJ cmꢀ2. The emission was
filtered with a 561 nm long-pass filter and collected by a PMA
Hybrid PMT 40 photomultiplier tube (PicoQuant). The signal
processing was carried out by a TimeHarp 260 Pico TCSPC
board (PicoQuant). On the same setup, the steady-state PL
spectra were collected with a Shamrock SR163 spectrograph
in combination with a Newton 970 EMCCD camera (Andor).
MAI synthesis
28.8 mL methylamine (33 wt% in ethanol, Sigma Aldrich) were
dissolved in 170 mL of ethanol under an argon atmosphere.
16 mL HI (57 wt% in H2O, Alfa Aesar) is added dropwise within
30 min while being cooled in an ice bath and the solution is
stirred overnight. The solvent is removed by a rotary evaporator
and the powder is washed with ether and ether/ethanol mixture
until the product is white.
Pellet pressing and gold electrodes
For pressing the pellets, 0.4 g perovskite powder was weighed
and put in a 1.3 cm press tool (Maassen) between two polished
hard metal press pins. The pressing was carried out in a
homemade mechanical screw press. By turning a threaded
rod with a handle, a pressure of 100 MPa was applied for
30 minutes, yielding free-standing thick films with a thickness
of 0.8 mm. For the PL Microscopy with lateral field and the
electrical measurements, four pairs of gold electrodes were
evaporated with a shadow mask on the pellets. Each electrode
pair forms a 150 mm long channel with a width of 1.0 mm. Gold
was evaporated with a rate of 0.02 nm sꢀ1 until an electrode
thickness of 100 nm was reached.
PL microscope measurements with lateral electrical field
The time-dependent PL measurements with an applied electric
field were carried out on a Zeiss Axio Imager.A2m microscope.
The excitation in the range of 380 nm to 427 nm was applied by
the microscope LED light, which was filtered with a 440/SP HC
short-pass filter and an HC BS 484 imaging beam splitter
(both AHF Analysentechnik AG). The light was focused with a
10ꢁ Objective (Zeiss). The excitation intensity was about
42 mW cmꢀ2. The PL emission was filtered by a 458 LP Edge
Basic long-pass filter (AHF) and collected with a pco.pixelfly usb
camera (PCO AG). A bias of 10 V at the gold electrodes was
applied by a Keithley 236 source meter.
XRD characterization
The structural characterization of the perovskite powders was
carried out by reflection mode XRD using a Bruker D8 Discover
A25 with Cu KaI radiation (l = 0.15406 nm), whereby the KaII
-
Electrical measurements
rays are removed by a Ge-KaI-monochromator. The device is
operated at 40 kV and 40 mA. The diffractograms were recorded
in the 2y range from 101 to 451 with a 2y step size of 0.0161.
The current–voltage characteristics, as well as the current versus
time curves, were recorded in the dark and under a nitrogen
atmosphere using a Keithley 2450 source meter. For the I–V
characteristics a voltage-sweep from ꢀ10 V to 10 V to ꢀ10 V was
used with a voltage step of 200 mV and a holding time of
500 ms before each current measurement. For the current
versus time measurements, a constant voltage of 10 V was
applied.
SEM and EDX characterization
Powders and pellets were mounted on a standard sample
holder by conductive adhesion graphite-pads (Plano) and sput-
tered with 1.5 nm platinum. The morphology was characterized
by SEM using a Zeiss Leo 1530 instrument (FE-SEM with
Schottky-field-emission cathode, In-lens detector and SE2
detector) using an accelerating voltage between 3.0 kV and Conflicts of interest
20 kV. For the EDX characterization, the pellets were vapor-
coated with carbon and an EDS UltraDry SDD Thermo Fisher
There are no conflicts to declare.
Noran detector was used.
Acknowledgements
Solid-state NMR spectroscopy
NMR spectra were recorded on a Varian VNMRS system operat- The authors acknowledge financial support by the German
ing at a magnetic field strength of 20.0 T (850 MHz) using a National Science Foundation DFG via the projects PA 3373/3-1
Varian 1.6 mm T3 HXY probe at room temperature. The and MO 1060/32-1 and GR 5505/1-1. We further thank Irene
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. C

View Article Online
Paper
Journal of Materials Chemistry C
16 B. Yang, W. C. Pan, H. D. Wu, G. D. Niu, J. H. Yuan,
K. H. Xue, L. X. Yin, X. Y. Du, X. S. Miao, X. Q. Yang,
Q. G. Xie and J. Tang, Nat. Commun., 2019, 10, 1989.
17 N. Leupold and F. Panzer, Adv. Funct. Mater., 2021,
31, 2007350.
18 S. Tie, W. Zhao, D. Xin, M. Zhang, J. Long, Q. Chen,
X. Zheng, J. Zhu and W. H. Zhang, Adv. Mater., 2020,
32, 2001981.
19 R. Zhuang, X. Wang, W. Ma, Y. Wu, X. Chen, L. Tang,
H. Zhu, J. Liu, L. Wu, W. Zhou, X. Liu and Y. Yang,
Nat. Photonics, 2019, 13, 602–608.
20 Y. Zhao, W. K. Zhou, Z. Y. Han, D. P. Yu and Q. Zhao,
Phys. Chem. Chem. Phys., 2021, 23, 94–106.
Bauer for MAI synthesis, Monika Daubinger for help with the
powder synthesis, the Department of Metal and Alloys (Prof.
Uwe Glatzel) for the possibility to conduct XRD measurements,
the Magnetic Resonance Research Center (MRRC, Radboud
University, Prof. Kentgens) for the access to the NMR spectro-
meter, NWO for the support of the ‘‘solid-state NMR facilities
for advanced materials science’’, which is part of the uNMR-NL
ROADMAP facility, the NMR facilitiy technicians, Gerrit Jans-
sen, Hans Janssen and Ruud Aspers for their support, Werner
Reichstein, Martina Heider, Dr Ulrich Mansfeld and the KeyLab
Electron and Optical Microscopy of the Bavarian Polymer
Institute (BPI) for assistance with PL microscopy, SEM and
EDX images and Sven Hu¨ttner for fruitful discussions.
21 D. Lan, Prog. Photovoltaics, 2019, 28, 533–537.
ˇ
22 S. Meloni, T. Moehl, W. Tress, M. Franckevicius, M. Saliba,
Y. H. Lee, P. Gao, M. K. Nazeeruddin, S. M. Zakeeruddin,
U. Rothlisberger and M. Graetzel, Nat. Commun., 2016,
7, 10334.
References
1 J. Y. Kim, J.-W. Lee, H. S. Jung, H. Shin and N.-G. Park,
Chem. Rev., 2020, 120, 7867–7918.
2 K. Ji, M. Anaya, A. Abfalterer and S. D. Stranks, Adv. Opt.
Mater., 2021, 2002128, DOI: 10.1002/adom.202002128.
3 Y. Su, W. Ma and Y. Yang, J. Semicond., 2020, 41,
051204.
4 G. Kakavelakis, M. Gedda, A. Panagiotopoulos, E. Kymakis,
T. D. Anthopoulos and K. Petridis, Adv. Sci., 2020,
7, 2002098.
5 Y. Zhou, J. Chen, O. M. Bakr and O. F. Mohammed,
ACS Energy Lett., 2021, 6, 739–768.
6 H. T. Wei and J. S. Huang, Nat. Commun., 2019, 10, 1066.
7 NREL, Best Research-Cell Efficiency Chart, https://www.nrel.
gov/pv/assets/pdfs/best-research-cell-efficiencies.20200104.
pdf, (accessed 30. March 2021).
8 Y. L. Song, L. Q. Li, W. H. Bi, M. W. Hao, Y. F. Kang,
A. R. Wang, Z. S. Wang, H. M. Li, X. H. Li, Y. J.
Fang, D. R. Yang and Q. F. Dong, Research, 2020,
2020, 5958243.
9 S. Yakunin, D. N. Dirin, Y. Shynkarenko, V. Morad,
I. Cherniukh, O. Nazarenko, D. Kreil, T. Nauser and
M. V. Kovalenko, Nat. Photonics, 2016, 10, 585–589.
10 H. Wei, Y. Fang, P. Mulligan, W. Chuirazzi, H.-H. Fang,
C. Wang, B. R. Ecker, Y. Gao, M. A. Loi and L. Cao,
Nat. Photonics, 2016, 10, 333–339.
23 J.-W. Lee, S.-G. Kim, J.-M. Yang, Y. Yang and N.-G. Park,
APL Mater., 2019, 7, 041111.
24 G. Y. Kim, A. Senocrate, Y.-R. Wang, D. Moia and J. Maier,
Angew. Chem., Int. Ed., 2021, 60, 820–826.
25 Z. Huang, A. H. Proppe, H. Tan, M. I. Saidaminov, F. Tan,
A. Mei, C.-S. Tan, M. Wei, Y. Hou, H. Han, S. O. Kelley and
E. H. Sargent, ACS Energy Lett., 2019, 4, 1521–1527.
26 S. Shao and M. A. Loi, Adv. Mater. Interfaces, 2020,
7, 1901469.
27 Y. Zhong, M. Hufnagel, M. Thelakkat, C. Li and S. Huettner,
Adv. Funct. Mater., 2020, 30, 1908920.
28 X. Zheng, B. Chen, J. Dai, Y. Fang, Y. Bai, Y. Lin, H. Wei,
X. C. Zeng and J. Huang, Nat. Energy, 2017, 2, 17102.
29 S. Bai, P. Da, C. Li, Z. Wang, Z. Yuan, F. Fu, M. Kawecki,
X. Liu, N. Sakai, J. T. Wang, S. Huettner, S. Buecheler,
M. Fahlman, F. Gao and H. J. Snaith, Nature, 2019, 571,
245–250.
30 P. Zhao, B. J. Kim and H. S. Jung, Mater. Today Energy, 2018,
7, 267–286.
31 X. Deng, L. Xie, S. Wang, C. Li, A. Wang, Y. Yuan, Z. Cao,
T. Li, L. Ding and F. Hao, Chem. Eng. J., 2020, 398,
125594.
32 M. Li, C. Zhao, Z.-K. Wang, C.-C. Zhang, H. K. H. Lee,
´
A. Pockett, J. Barbe, W. C. Tsoi, Y.-G. Yang, M. J. Carnie,
X.-Y. Gao, W.-X. Yang, J. R. Durrant, L.-S. Liao and
S. M. Jain, Adv. Energy Mater., 2018, 8, 1801509.
11 X. Zhang, F. Li and R. Zheng, J. Mater. Chem. C, 2020, 8,
13918–13952.
33 N. K. Noel, S. N. Habisreutinger, B. Wenger, Y. H. Lin,
F. Zhang, J. B. Patel, A. Kahn, M. B. Johnston and
H. J. Snaith, Adv. Energy Mater., 2019, 10, 1903231.
12 S. Shrestha, R. Fischer, G. J. Matt, P. Feldner, T. Michel,
A. Osvet, I. Levchuk, B. Merle, S. Golkar, H. Chen,
¨
S. F. Tedde, O. Schmidt, R. Hock, M. Ru¨hrig, M. Goken,
¨
34 N. Leupold, K. Schotz, S. Cacovich, I. Bauer, M. Schultz,
W. Heiss, G. Anton and C. J. Brabec, Nat. Photonics, 2017, 11,
436–440.
¨
M. Daubinger, L. Kaiser, A. Rebai, J. Rousset, A. Kohler,
P. Schulz, R. Moos and F. Panzer, ACS Appl. Mater. Interfaces,
2019, 11, 30259–30268.
13 Z. H. Hong, D. Tan, R. A. John, Y. K. E. Tay, Y. K. T. Ho,
´
X. Zhao, T. C. Sum, N. Mathews, F. Garcıa and H. S. Soo,
35 F. Panzer, D. Hanft, T. P. Gujar, F.-J. Kahle, M. Thelakkat,
iScience, 2019, 16, 312–325.
¨
A. Kohler and R. Moos, Materials, 2016, 9, 277.
´
14 L. Martınez-Sarti, F. Palazon, M. Sessolo and H. J. Bolink,
¨
36 T. Meier, T. P. Gujar, A. Schonleber, S. Olthof, K. Meerholz,
Adv. Opt. Mater., 2020, 8, 1901494.
15 F. Palazon, Y. El Ajjouri and H. J. Bolink, Adv. Energy Mater.,
2020, 10, 1902499.
¨
S. van Smaalen, F. Panzer, M. Thelakkat and A. Kohler,
J. Mater. Chem. C, 2018, 6, 7512–7519.
J. Mater. Chem. C
This journal is © The Royal Society of Chemistry 2021

View Article Online
Journal of Materials Chemistry C
Paper
37 A. Deyko, S. G. Hessey, P. Licence, E. A. Chernikova, 54 M. Gerhard, B. Louis, R. Camacho, A. Merdasa, J. Li,
V. G. Krasovskiy, L. M. Kustov and R. G. Jones, Phys. Chem.
Chem. Phys., 2012, 14, 3181–3193.
A. Kiligaridis, A. Dobrovolsky, J. Hofkens and
I. G. Scheblykin, Nat. Commun., 2019, 10, 1698.
¨
38 C. Witt, A. Schmid, N. Leupold, M. Schultz, J. Hocker, 55 Y.-C. Zhao, W.-K. Zhou, X. Zhou, K.-H. Liu, D.-P. Yu and
A. Baumann, R. Moos and F. Panzer, ACS Appl. Electron.
Mater., 2020, 2, 2619–2628.
39 M. Abdi-Jalebi, Z. Andaji-Garmaroudi, S. Cacovich,
C. Stavrakas, B. Philippe, J. M. Richter, M. Alsari,
Q. Zhao, Light: Sci. Appl., 2017, 6, e16243.
56 S.-G. Kim, C. Li, A. Guerrero, J.-M. Yang, Y. Zhong,
J. Bisquert, S. Huettner and N.-G. Park, J. Mater. Chem. A,
2019, 7, 18807–18815.
E. P. Booker, E. M. Hutter, A. J. Pearson, S. Lilliu, 57 D. R. Ceratti, A. Zohar, R. Kozlov, H. Dong, G. Uraltsev,
T. J. Savenije, H. Rensmo, G. Divitini, C. Ducati,
R. H. Friend and S. D. Stranks, Nature, 2018, 555, 497–501.
40 L. M. Herz, Annu. Rev. Phys. Chem., 2016, 67, 65–89.
41 M. B. Johnston and L. M. Herz, Acc. Chem. Res., 2016, 49,
146–154.
O. Girshevitz, I. Pinkas, L. Avram, G. Hodes and D. Cahen,
Adv. Mater., 2020, 32, 2002467.
58 N. Leupold, A. L. Seibel, R. Moos and F. Panzer, Eur. J. Inorg.
Chem., 2021, DOI: 10.1002/ejic.202100381.
59 C. Eames, J. M. Frost, P. R. F. Barnes, B. C. O’Regan,
A. Walsh and M. S. Islam, Nat. Commun., 2015, 6, 7497.
42 D. W. deQuilettes, K. Frohna, D. Emin, T. Kirchartz,
V. Bulovic, D. S. Ginger and S. D. Stranks, Chem. Rev., 60 W. Zhang, X. Liu, B. He, Z. Gong, J. Zhu, Y. Ding, H. Chen
2019, 119, 11007–11019.
and Q. Tang, ACS Appl. Mater. Interfaces, 2020, 12,
43 T. W. Crothers, R. L. Milot, J. B. Patel, E. S. Parrott,
4540–4548.
J. Schlipf, P. Mu¨ller-Buschbaum, M. B. Johnston and 61 D. Yang, X. Zhou, R. Yang, Z. Yang, W. Yu, X. Wang, C. Li,
L. M. Herz, Nano Lett., 2017, 17, 5782–5789.
S. Liu and R. P. H. Chang, Energy Environ. Sci., 2016, 9,
44 Z. Andaji-Garmaroudi, M. Anaya, A. J. Pearson and
S. D. Stranks, Adv. Energy Mater., 2019, 10, 1903109.
45 S. G. Motti, D. Meggiolaro, A. J. Barker, E. Mosconi,
3071–3078.
¨
62 T. Y. Yang, G. Gregori, N. Pellet, M. Gratzel and J. Maier,
Angew. Chem., Int. Ed., 2015, 54, 7905–7910.
C. A. R. Perini, J. M. Ball, M. Gandini, M. Kim, F. De Angelis 63 W. Tress, J. Phys. Chem. Lett., 2017, 8, 3106–3114.
and A. Petrozza, Nat. Photonics, 2019, 13, 532–539.
64 H. Wang, A. Guerrero, A. Bou, A. M. Al-Mayouf and
46 D. W. deQuilettes, W. Zhang, V. M. Burlakov, D. J. Graham,
J. Bisquert, Energy Environ. Sci., 2019, 12, 2054–2079.
T. Leijtens, A. Osherov, V. Bulovic, H. J. Snaith, D. S. Ginger 65 T. Wu, M. Ahmadi and B. Hu, J. Mater. Chem. C, 2018, 6,
and S. D. Stranks, Nat. Commun., 2016, 7, 11683. 8042–8050.
47 C. Li, S. Tscheuschner, F. Paulus, P. E. Hopkinson, 66 R. A. Kerner, L. Zhao, S. P. Harvey, J. J. Berry, J. Schwartz and
¨
J. Kiessling, A. Kohler, Y. Vaynzof and S. Huettner, Adv.
B. P. Rand, ACS Energy Lett., 2020, 5, 3352–3356.
67 R. A. Kerner, P. Schulz, J. A. Christians, S. P. Dunfield,
B. Dou, L. Zhao, G. Teeter, J. J. Berry and B. P. Rand,
APL Mater., 2019, 7, 041103.
Mater., 2016, 28, 2446–2454.
48 J. M. Azpiroz, E. Mosconi, J. Bisquert and F. De Angelis,
Energy Environ. Sci., 2015, 8, 2118–2127.
49 S. T. Birkhold, J. T. Precht, R. Giridharagopal, G. E. Eperon, 68 O. R. Yamilova, A. V. Danilov, M. Mangrulkar, Y. S. Fedotov,
L. Schmidt-Mende and D. S. Ginger, J. Phys. Chem. C, 2018,
122, 12633–12639.
50 J. Zhang, C. Li, M. Chen and K. Huang, J. Phys. D: Appl.
Phys., 2021, 54, 044002.
S. Y. Luchkin, S. D. Babenko, S. I. Bredikhin, S. M. Aldoshin,
K. J. Stevenson and P. A. Troshin, J. Phys. Chem. Lett., 2020,
11, 221–228.
69 Z. Xiao, Y. Yuan, Y. Shao, Q. Wang, Q. Dong, C. Bi,
P. Sharma, A. Gruverman and J. Huang, Nat. Mater., 2015,
14, 193–198.
¨
51 C. Li, A. Guerrero, Y. Zhong, A. Graser, C. A. M. Luna,
¨
J. Kohler, J. Bisquert, R. Hildner and S. Huettner, Small,
¨
2017, 13, 1701711.
70 R. Waser, T. Baiatu and K.-H. Hardtl, J. Am. Ceram. Soc.,
52 C. Li, A. Guerrero, S. Huettner and J. Bisquert, Nat. Commun.,
2018, 9, 5113.
1990, 73, 1645–1653.
71 R. Waser, T. Baiatu and K.-H. Hardtl, J. Am. Ceram. Soc.,
¨
53 X. Deng, X. Wen, C. F. J. Lau, T. Young, J. Yun, M. A. Green,
1990, 73, 1654–1662.
S. Huang and A. W. Y. Ho-Baillie, J. Mater. Chem. C, 2016, 4, 72 I. Denk, J. Claus and J. Maier, J. Electrochem. Soc., 1997, 144,
9060–9068.
3526–3536.
This journal is © The Royal Society of Chemistry 2021
J. Mater. Chem. C