High-Temperature Photoluminescence of CsPbX3 (X = Cl, Br, I) Nanocrystals

Article abstract of DOI:10.1002/adfm.201606750

Recent synthetic developments have generated intense interest in the use of cesium lead halide perovskite nanocrystals for light-emitting applications. This work presents the photoluminescence (PL) of cesium lead halide perovskite nanocrystals with tunable halide composition recorded as function of temperature from 80 to 550 K. CsPbBr₃ nanocrystals show the highest resilience to temperature while chloride-containing samples show relatively poorer preservation of photoluminescence at elevated temperatures. Thermal cycling experiments show that PL loss of CsPbBr₃ is largely reversible at temperatures below 450 K, but shows irreversible degradation at higher temperatures. Time-resolved measurements of CsPbX₃ samples show an increase in the PL lifetime with temperature elevation, consistent with exciton fission to form free carriers, followed by a decrease in the apparent PL lifetime due to trapping. PL persistence measurements and time-resolved spectroscopies implicate thermally assisted trapping, most likely to halogen vacancy traps, as the mechanism of reversible PL loss.

Full text of DOI:10.1002/adfm.201606750

FULL PAPER
Photoluminescence
www.afm-journal.de
High-Temperature Photoluminescence of CsPbX3
(X = Cl, Br, I) Nanocrystals
Benjamin T. Diroll, Georgian Nedelcu, Maksym V. Kovalenko, and Richard D. Schaller*
light-emitting devices such as low-threshold
lasers,[9,10] light-emitting diodes (LEDs),[11–15]
and downconverting layers[16–18] similar to
those used in quantum dot display tech-
nology. Survival of advantageous prop-
erties at high temperature, which is a
significant problem for organolead halide
perovskites,[19,20] is one challenge to the use
of such NCs in many of these technologies,
which require stable materials at tempera-
tures often exceeding 100 °C.
Recent synthetic developments have generated intense interest in the use of
cesium lead halide perovskite nanocrystals for light-emitting applications. This
work presents the photoluminescence (PL) of cesium lead halide perovskite
nanocrystals with tunable halide composition recorded as function of tempera-
ture from 80 to 550 K. CsPbBr3 nanocrystals show the highest resilience to tem-
perature while chloride-containing samples show relatively poorer preservation
of photoluminescence at elevated temperatures. Thermal cycling experiments
show that PL loss of CsPbBr3 is largely reversible at temperatures below 450 K,
but shows irreversible degradation at higher temperatures. Time-resolved meas-
urements of CsPbX3 samples show an increase in the PL lifetime with tempera-
ture elevation, consistent with exciton fission to form free carriers, followed by
a decrease in the apparent PL lifetime due to trapping. PL persistence measure-
ments and time-resolved spectroscopies implicate thermally assisted trapping,
most likely to halogen vacancy traps, as the mechanism of reversible PL loss.
This work studies the photolumines-
cence (PL) of CsPbX3 NCs of variable
halide composition from 80 K to up to
550 K. The PL intensity of the NC sam-
ples decreases with temperature elevation,
decreasing to intensities less than 0.1% of
the initial PL as measured at 80 K. How-
ever, the resilience of PL at temperatures
below 450 K is strongly dependent on the halide composition
of the NCs. Chloride-containing NCs show consistently greater
susceptibility to thermal quenching than pure bromide, iodide,
or iodide-bromide-mixed halides. Pure CsPbBr3 NCs show the
greatest thermal stability among the samples measured, with
bromide-iodide and pure iodide compounds substantially more
stable than chloride-containing NCs. However, PL persistence
is considerably lower than other reported NCs—typically core/
shells—proposed for optoelectronic technologies. Most CsPbX3
NCs display small blueshifts with increasing temperature and
each display increased sensitivity of thermal broadening com-
pared to other visible NC emitters.
Experiments on reversibility of PL loss of CsPbBr3 NCs
implicate both reversible (to 450 K) and irreversible (>450 K)
loss pathways. Irreversible loss corresponds to ligand desorp-
tion, supported by thermogravimetric analysis (TGA) meas-
urements, and underscores the substantial advantages of any
potential core/shell structure. Examination of several sizes of
CsPbBr3 NCs suggests that reversible PL loss most likely arises
from thermally activated trapping into halogen vacancies. Time-
resolved PL data show that for CsPbX3 NCs, the intrinsic radia-
tive rate decreases with temperature elevation, due to exciton
fission, up to a threshold value at which the apparent lifetime
of PL and transient absorption (TA) decreases due to thermally
assisted carrier trapping at halogen vacancies.
1. Introduction
The tremendous success of organolead halide perovskites for
photovoltaic cells[1] and other optoelectronic applications[2,3]
demonstrated in the last five years has led to intense exploration
of related chemistries.[4,5] Among the most promising reported
materials for light-emitting applications in the visible spectrum
are the cesium lead halide perovskite analogues (CsPbX3), in
which the organic cation is replaced by cesium.[6,7] By tuning the
halide composition of these materials, the photoluminescence
can be tuned through virtually the whole visible spectrum. As col-
loidal nanocrystals (NCs), these materials have emission with high
quantum yield and high color purity[6,8] which have been used in
Dr. B. T. Diroll, Prof. R. D. Schaller
Center for Nanoscale Materials
Argonne National Laboratory
Lemont, IL 60439, USA
E-mail: schaller@anl.gov, schaller@northwestern.edu
G. Nedelcu, Prof. M. V. Kovalenko
Department of Chemistry and Applied Bioscience
ETH Zürich
Vladimir Prelog Weg 1
CH-8093 Zürich, Switzerland
Prof. M. V. Kovalenko
Empa – Swiss Federal Laboratories for Materials Science
and Technology
Überlandstrasse 129
CH-8600 Dübendorf, Switzerland
Prof. R. D. Schaller
2. Temperature Dependence of Steady-State
Photoluminescence of CsPbX3 Nanocrystals
Department of Chemistry
Northwestern University
Evanston, IL 60208, USA
Figure 1 shows transmission electron microscopy (TEM) images
DOI: 10.1002/adfm.201606750
and optical spectra of the samples used in this work. Unlike
©
Adv. Funct. Mater. 2017, 1606750
1606750 (1 of 7)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
Figure 1. TEM images of CsPbX3 perovskite NCs: a) CsPbCl3, b) CsPbCl2Br, c) CsPbClBr2, d) CsPbBr3, e) CsPbBr2I, f) CsPbBrI2, and g) CsPbI3. Scale
bars in images (a)–(g) represent 50 nm. h) Absorption spectra (thin lines) and PL spectra (thick lines) of CsPbX3 samples. i) 2D map showing the PL
intensity of a CsPbBr3 NC sample plotted against temperature and energy. The color map is on a logarithmic scale.
quantum dots such as CdSe, tunable emission throughout the
visible spectrum does not accompany major changes in the par-
ticle size. The samples range in size from 9 to 13 nm and all
formed predominantly into NCs of cubic morphology. The com-
position of the samples is varied by changing the lead halide
precursors used in the synthesis (e.g., synthesis of CsPbBrI2
used 67 mol% PbI2 and 33 mol% PbBr2) and tuning the reac-
tion temperature to obtain similar sized NCs. Each of the seven
samples progressively changes the composition by one halide
ion, varying the PL energy of the samples at room temperature
from 3.05 eV for pure CsPbCl3 to 1.75 eV for pure CsPbI3.
A typical example of the high-temperature PL measurement
is plotted in Figure 1i, showing the emission of a CsPbBr3
sample as a function of emission energy and temperature. Like
all other examples in this work, the NC sample was embedded
into a polymer host to reduce inter-NC interactions which may
diminish PL intensity and mimic the form of use in NC-based
displays (see the Experimental Section and Figure S1, Sup-
porting Information, for more details). Certain qualitative fea-
tures are noteworthy in the PL data shown in Figure 1i. First,
the emission intensity, which is plotted on a logarithmic scale,
decreases subtly to 450 K and then drops dramatically lower for
temperatures exceeding 450 K. Second, the emission energy of
the sample blueshifts with higher temperatures, indicating that
the band gap of the CsPbBr3 increases, behavior which is also
observed in related halide perovskites,[21–23] but the opposite of
behavior of IV, III–V, and II–VI semiconductors.[24] Third, the
bandwidth of the PL, a critical parameter for enhanced color
gamut, broadens at high temperature.
A quantitative analysis of the temperature-dependent PL
of each sample appears in Figure 2. Figure 2a plots the inte-
grated PL of the samples normalized to the measured emis-
sion at 80 K. Although PL of each of the samples decreases
to ≈10−3 or lower for >500 K, the chloride-containing samples
Figure 2. a) Integrated PL of CsPbX3 NC samples plotted versus measurement temperature. The inset bar plot shows the fraction of emission (relative
to 80 K) of each sample at 400 K, in corresponding colors. b) Emission peak center energy of the CsPbX3 samples versus temperature. c) Temperature-
dependent PL bandwidth measured from the full-width at half maximum (FWHM) of the CsPbX3 samples.
©
Adv. Funct. Mater. 2017, 1606750
1606750 (2 of 7)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
showed markedly lower preservation of PL at intermediate
temperatures, a fact emphasized in the inset in Figure 2a.
These samples show the onset of a steep decline in PL begin-
ning near 300 K, whereas samples that do not contain chloride
ions decrease dramatically closer to 375 K and above, with the
CsPbBr3 sample showing the greatest preservation of PL at
nearly all measured temperatures.
coupling and nonlinearly based on longitudinal optical (LO)
phonon–exciton coupling.[32,33] The nonlinear increase in the
PL linewidth with elevated temperature is evidence for the con-
tribution of both acoustic and optical phonons in the thermal
broadening. Raman spectra of CsPbX3 samples indicate a
complex manifold of LO phonons which may contribute to the
broadening of PL spectra at higher temperatures.[34,35] Data in
Figure 2c indicate that PL broadening in CsPbX3 is significantly
more temperature-sensitive than for II–VI NCs, especially
above 300 K.[32,36–39] For example, the CsPbBr3 sample broadens
from 82 meV at 300 K to 139 meV at 500 K (+70%), whereas
CdSe/CdS NCs show an increase of full width at half maximum
(FWHM) from 87 to 96 meV (+10%) over the same tempera-
ture range.[36] As a practical matter, this means that tempera-
ture elevation of CsPbX3 reduces achievable color gamut more
than II–VI materials, although the wavelength of the emission
appears relatively more stable.
It is also worth comparing these results to some of the
results previously published in the literature on similar mate-
rials, primarily at lower temperature ranges. Compared to
organolead halide perovskites, Cs-based lead trihalide perovs-
kites are known to substantially higher stability under most
environmental conditions.[25–27] Single crystals of Cs-based
plumbotrihalides have, for example, demonstrated thermal sta-
bility to 600 °C, whereas organolead halide perovskites become
unstable near 300 °C.[25] Indeed, bulk organolead halide per-
ovskites show near complete loss of PL when heating from
300 to 400 K, although quantum dots of the same preserve
≈30% of PL over the same temperature range comparable to
our samples.[22] CsPbBr3 NCs synthesized at room tempera-
ture in DMF/toluene mixtures show substantial (≈85%) PL loss
upon heating from 80 to 273 K,[11] compared to losses reported
here of less than 10% for the same composition over a similar
temperature range. Similar materials also made at room tem-
perature showed PL losses of up to 80% upon heating above
room temperature to 373 K.[28] In a related work, CsPbBr3 and
CsPbCl3 single crystals grown by the Bridgman technique are
also reported to show relatively poorer PL persistence at low
temperature, with PL decreasing by more than 90% when
heating samples from 10 to 180 K.[29] Later in this work, we will
address a series of experiments which explore the origins of PL
loss in CsPbX3 NC samples.
To obtain the emission energy and linewidth, Voight fits
were applied to the PL data for each of the samples. As the PL
becomes very weak for the highest measured temperatures,
fitting data from those points with less than 0.1% of the orig-
inal PL intensity were excluded. Figure 2b shows that like the
CsPbBr3 samples in Figure 1i, PL energy blueshifts slightly (less
than 25 meV) with elevated temperature for most samples.
This behavior is unusual; for comparison, CdSe NCs redshift by
100 meV over same measured temperature range. On the other
hand, in some cases in Figure 2b, the PL energy barely changes
at all and reports on CsPbX3 materials have yielded similarly
mixed results.[30] A Varshni relation of the form Egap(T) =
Egap(0) − αT2/(β + T)[24] was used to fit the PL data with the
fit parameters reported in Table S1 (Supporting Information).
Varshni fits of the PL data yield an α parameter which ranges
from −1.9 × 10−4 to 3.7 × 10−5 K−1 for the samples. In addition
to being negative for most samples indicating a blueshift, this
represents a smaller wavelength shift with temperature than
found for most semiconductors[24] and consistent with previous
reports which claim small shift of PL energy of CsPbBr3 NCs
from 25 to 100 °C.[31] The β parameter obtained from Varshni
fits, which in principle corresponds to the Debye tempera-
ture,[24] is ≈300 K for most of the CsPbX3 NCs.
3. Mechanisms of Photoluminescence Loss
To determine the cause of PL loss with temperature elevation,
several experiments were performed. Cyclic heating experi-
ments disambiguate loss of PL due to reversible and irrevers-
ible processes. Figure 3 shows the results of multiple heating
and cooling steps for a CsPbBr3 sample normalized to the PL
intensity at 300 K. Cooling samples below 300 K increased the
PL, as shown in Figure 2a, but the integrated intensity was the
same when returning to 300 K, indicating full reversibility of
PL loss up to room temperature. Figure 3 shows that PL loss
upon heating to 400 K is reversible to >95% of the intensity at
300 K and PL loss to 450 K is >80% reversible. Previous works
have also shown substantially reversible PL losses in samples
heated from room temperature to 373 K.[28] Only upon heating
the sample above 450 K does PL loss becomes mostly irrevers-
ible. The onset of substantial irreversible PL loss occurs at the
same temperatures where the integrated PL intensity drops
rapidly with increasing temperature for CsPbBr3 samples.
Time-resolved emission and transient absorption measure-
ments collected at different heating temperatures provide addi-
tional insight into PL loss pathways and support a thermally
assisted trapping model. Figure 4a shows the time-resolved
photoluminescence (TRPL) of a CsPbBr3 sample (≈200 ps
time resolution), which shows an increase in the PL lifetime
from 80 to 450 K, whereupon the lifetime decreases rapidly,
The bandwidth of PL for all of the samples varies with
temperature in a similar manner, as shown in Figure 2c, in
each case presenting <75 meV linewidth at 80 K. Thermal
broadening of PL increases linearly on acoustic phonon–exciton
Figure 3. Thermal cycling measurements of CsPbBr3 PL. High set point
temperatures are highlighted in red for clarity.
©
Adv. Funct. Mater. 2017, 1606750
1606750 (3 of 7)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
Figure 4. a) 2D map of PL of CsPbBr3 NCs plotted against temperature and time collected using an avalanche photodiode. The color scale is loga-
rithmic and data are normalized to the maximum at each temperature. b) Normalized streak camera TRPL traces as a function of temperature for the
same CsPbBr3 NCs. c) Normalized transient absorption traces of the band-edge bleaching feature of the same CsPbBr3 NCs.
corresponding to the precipitous loss of PL intensity. A similar
hump-shaped TRPL signal has been observed previously in
CdSe/CdS heterostructures,[36,40] but represents another unu-
sual feature of CsPbX3 NCs. Core-only CdSe, Si, and InP NCs
all show a monotonic decrease in PL lifetime with heating.
To confirm this trend and analyze temperature-dependent
recombination of excitons with higher time resolution, we per-
formed ultrafast TRPL measurements and transient absorption
(TA). Careful comparison of TRPL and TA data of III–V and
II–VI NCs allows discrimination of electron and hole trapping
behavior.[41,42] However, disambiguating the cause of decreased
PL with temperature in CsPbX3 NCs is a challenge, because the
effective masses of electrons and holes in this composition are
very similar.[6,43] Because both carriers have potential to gen-
erate a similar bleach feature in TA, the ultrafast TRPL and TA
signals are expected to be similar, which is what we indeed find
in Figure 4b,c. This means that we cannot distinguish if trap-
ping is unipolar or which carrier is trapped. However, ultrafast
measurements recapitulate the pattern observed in Figure 4a:
as the temperature is raised, the TRPL and TA signals show an
increase in the decay time up to a threshold temperature (here
400 K), after which the decay time decreases. The discrepancy
in the apparent threshold temperature is most likely due to dif-
ferences in the time resolution of the experiments.
materials,[22,44,45] exciton separation is not, by itself, an expla-
nation for decreased PL, which requires nonradiative decay
pathways.
To identify possible nonradiative decay pathways responsible
for reversible and irreversible PL loss, a series of size-
dependent measurements were performed using CsPbBr3 NCs.
Although the NC size does not substantially change the emis-
sion energy at the size range measured for the CsPbX3 NCs
described above, it may still yield effects on thermal stability.
Several sizes of CsPbBr3 NCs were measured with the results
shown in Figure 5a. Among the samples which range from
7.1 to 11.0 nm (400% change in volume), there is no system-
atic difference in the PL properties with size, with each sample
All sets of time-resolved data on CsPbBr3 show an initial
increase in the lifetime followed by a substantial decrease in
carrier lifetime above a threshold temperature, with the excited-
state lifetime becoming comparatively brief (<1 ns) at the
highest temperatures measured. Although the focus here is
on a CsPbBr3 sample, CsPbCl3 and CsPbI3 show substantially
similar behavior in time-resolved measurements (Figure S3,
Supporting Information). Time-resolved data are interpreted
to show a decrease in the intrinsic radiative rate for CsPbBr3
with increasing temperature, followed, above a threshold tem-
perature, by the increasing presence of nonradiative pathways
for recombination, first to reversibly occupied trap states and
then to irreversible PL loss. Previous work has identified fission
of bound excitons, with an energy of 40–50 meV as the cause
for weaker PL at higher temperatures in CsPbBr3 NCs.[7,11,39]
The increased energy available for separation of excitons is
consistent with the increase in the apparent lifetime, wherein
free carriers recombine more slowly than bound excitons
resulting in delayed fluorescence. Although it has been impli-
cated in the higher PL of perovskite quantum dots versus bulk
Figure 5. a) Integrated PL of several sizes of CsPbBr3 NCs measured as
a function of temperature. b) Fitting of PL decay of CsPbBr3 samples to a
single reversible trap state model.
©
Adv. Funct. Mater. 2017, 1606750
1606750 (4 of 7)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
exhibiting substantially similar thermal stability. This strongly
suggests that both reversible and irreversible PL loss mecha-
nisms are independent of the size of the NCs over the range
examined in this work.
ture regime yield trap activation energies which are mostly
consistent with previous DFT calculations (0.16–0.23 eV).[29,51]
Estimated barrier heights for CsPbBr3 samples range from
90 to 200 meV (Table S2, Supporting Information). The agree-
ment between predicted trap energies and our experimental
data leads us to identify the mechanism of reversible PL loss in
CsPbX3 NCs as thermal electron occupation of halogen vacancy
centers. Thus, although these materials are remarkably tolerant
of halide vacancies, in part because they do not form traps
within the band gap, they will still diminish PL intensity with
temperature elevation.
From thermal cycling experiments described above and
shown in Figure 3, irreversible PL losses dominate when the
CsPbBr3 samples are heated above 450 K. Thermal analysis of
CsPbBr3 NCs (Figure S4, Supporting Information), which show
a large weight loss around 425 K, supports the argument that
irreversible PL loss is related to loss of organic ligands on the
NC surface.[40] In the case of CsPbX3 NCs, the ligand shell is
composed of both carboxylate and amine ligands bound to cati-
onic and anionic sites, in contrast to the metal-rich surfaces of
most semiconductor nanocrystals,[46,47] and desorption or alter-
ation of either could yield irreversible PL loss. This may be due
to both the loss of electronic passivation and surface traps, as
has been found in CdSe[40] and InP,[48] but can also arise from
structural changes which also occur at elevated temperatures
and are enabled by dynamic ligand coverage. X-ray diffraction
data (Figure S5, Supporting Information) show that heating to
500 K induces sintering in dense films of NCs, but not in films
of well-separated NCs embedded in a polymer, such as those
used in our studies.[16,47] Therefore, consistent with thermal
analysis data, X-ray diffraction argues for increased mobility of
inorganic ions accompanying surface ligand desorption. Given
previous thermal analysis of CsPbBr3, in which decomposition
occurs in a nitrogen atmosphere at temperatures >600 °C,[25]
and our own X-ray diffraction data, wholesale decomposition of
the inorganic NC material is also not the likely source of irre-
versible PL loss, although subtle changes in the NC composi-
tion due to the increased mobility of ions are possible.
Chemical identification of the thermally reversible trap state
is not straightforward from the presented measurements. As
excellent ionic conductors with exceptionally fast transport of
halides, CsPbX3 materials often have many vacancies.[49,50]
The excellent optoelectronic performance of perovskite hal-
ides, compared to other halide compounds, has been linked
to halide vacancy tolerance arising from the shallow nature
of halide vacancy states.[51] CsPbBr3 is the best studied of the
samples examined in this work: previous density functional
theory (DFT) calculations claim that, within CsPbBr3, bromide
vacancies are 0.23[29] or 0.16 eV[51] above the conduction band
edge. Nonetheless, these well-known vacancies may be respon-
sible for reversible PL loss at elevated temperatures despite
being electronically passive at lower temperatures at which they
are not significantly occupied. Reversible, thermally assisted
4. Conclusion
The temperature dependence of PL was measured for the
family of CsPbX3 perovskite NCs with great promise in opto-
electronic applications. PL persistence at high temperatures
was lower than that of core/shell NCs which have been studied
before, but comparable to core-only NCs. Halide composition
significantly alters PL intensity above 400 K, with chloride-con-
taining halides showing particularly diminished stability. The
samples demonstrate smaller temperature-dependent changes
in wavelength than most other semiconductors, including an
unusual blueshift with increasing temperature, but also larger
temperature-dependent changes in PL broadening, which
undermines color purity. Measurements of reversibility, life-
time, and size-dependence implicate thermally assisted trap-
ping, most likely to halogen vacancy centers, as the mechanism
of reversible PL loss. These results underscore that although
much of the appeal of CsPbX3 NCs rests on their high room-
temperature quantum yields due to resilience to defects in the
crystal, these defects may play an important role in diminished
performance at higher operating temperatures of optoelectronic
devices. Relatively low-temperature irreversible loss observed
for these samples also indicates the durability of the strategy of
using core/shell structures for enhanced stability.
5. Experimental Section
Materials: Polyvinylbutyral Butvar B-98, oleic acid (90%), oleylamine
(70%), octadecene (ODE, 90%), tricotylphosphine (97%), lead bromide
(>98%), lead chloride (98%), lead iodide (99%), and cesium carbonate
(99.9%) were purchased from Sigma-Aldrich. Oleic acid and oleylamine
were dried and stored in a nitrogen filled glovebox. All solvents were ACS
grade or higher.
Synthesis: Synthesis of perovskite NCs followed the procedure of
Protesescu et al.[6] Typically, a reaction flask containing 4 mL ODE and
0.19 mmol PbX2 was heated to 120 °C under vacuum and held 1 h.
The composition of the lead halide was chosen for a desired output
composition of the NCs. Separately, to prepare the Cs-oleate precursor,
407 mg Cs2CO3, 1.25 mL oleic acid, and 20 mL ODE were heated under
vacuum to 120 °C and held for 1 h, then heated under nitrogen to 150 °C
and held until the Cs2CO3 completely dissolved. The Cs-oleate precursor
was maintained as a homogeneous clear solution by heating prior to
injection. Once the reaction flask was under vacuum for 1 h, it was
flushed with nitrogen and then 0.5 mL of dry and degassed oleylamine
and 0.5 mL of dry and degassed oleic acid were added and the heat was
maintained until the reaction flask contained a homogeneous solution.
For CsPbCl3 reactions, 1 mL TOP was also added to dissolve PbCl2.
Then the reaction flasks were heated under nitrogen flow to 140–170 °C.
trapping, for a single trap state with energetic barrier Etrap
decreases the PL intensity according to Equation (1)
,
I0
I(T) =
(1)
trap/kT
1+ Ae−E
in which A is a parameter related to the cross section of the trap
state.[52] At temperatures with kT comparable to the trap energy
(and above), substantial PL quenching can occur. Simple ther-
mally assisted trapping of this type should be fully reversible,
as opposed to irreversible trapping due to transformation of the
material. By fitting the reversible PL loss of the samples (from
80 to 450 K) in Figure 5b, trap barrier heights may be estimated
for the CsPbBr3 samples. The results of fitting this tempera-
©
Adv. Funct. Mater. 2017, 1606750
1606750 (5 of 7)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
Upon reaching the set point temperature, 0.4 mL of the Cs-oleate stock
solution was rapidly injected. After 5 s, the reaction flasks were cooled
with a water bath. Samples were isolated by centrifugation of the NCs.
For temperature-dependent measurements, NCs were diluted into
solid films of polyvinylbutyral (Butvar B-98). As reported previously,[16,17]
embedding the NCs into a polymer enhanced the NC stability in air. For
example, embedded NCs of CsPbI3 exposed to air retained characteristic
absorption properties longer than dense thin films of the same sample.
Nonetheless, embedding and measurement were performed immediately
after synthesis (<24 h). To form dilute polymer films, a 50 mg mL−1
chloroform solution of Butvar B-98 was made and then mixed 3:1 with
chloroform solution of NCs. According to the manufacturer (Eastman),
Butvar B-98 does not show substantial thermal degradation under
air or nitrogen until 350 °C, well above the temperatures used in this
work. No changes in the polymer films, including evidence of flow or
decomposition, were observed during measurements.
Steady-State Measurements: PL measurements were collected using
a pulsed 405 nm laser diode directed on to samples of perovskite
NCs embedded in polyvinylbutyral, loaded into a variable temperature
cryostat/oven. PL was fiber-coupled through a spectrometer to a CCD.
Time-Resolved Optical Measurements: Three kinds of time-resolved
measurement were performed in this work. Low-resolution TRPL
measurements (≈200 ps) were performed using a pulsed 405 nm laser
diode with emission directed on to an avalanche photodiode. Higher
time resolution TRPL (≈20 ps) on a streak camera (Hamamatsu) was
collected using a 325 nm pulse generated through an optical parametric
amplifier using a Ti:sapphire femtosecond laser. TA measurements
were performed with the same excitation pulse using a Helios Transient
Absorption Spectrometer (Ultrafast Systems).
M. B. Johnston, A. Petrozza, L. M. Herz, H. J. Snaith, Energy
Environ. Sci. 2014, 7, 3061.
[5] T. C. Jellicoe, J. M. Richter, H. F. J. Glass, M. Tabachnyk, R. Brady,
S. E. Dutton, A. Rao, R. H. Friend, D. Credgington, N. C. Greenham,
M. L. Böhm, J. Am. Chem. Soc. 2016, 138, 2941.
[6] L. Protesescu, S. Yakunin, M. I. Bodnarchuk, F. Krieg, R. Caputo,
C. H. Hendon, R. X. Yang, A. Walsh, M. V. Kovalenko, Nano Lett.
2015, 15, 3692.
[7] D. Fröhlich, K. Heidrich, H. Künzel, G. Trendel, J. Treusch, J. Lumin.
1979, 18, 385.
[8] Y. Bekenstein, B. A. Koscher, S. W. Eaton, P. Yang, A. P. Alivisatos,
M. V. Kovalenko, J. Am. Chem. Soc. 2015, 137, 16008.
[9] S. Yakunin, L. Protesescu, F. Krieg, M. I. Bodnarchuk, G. Nedelcu,
M. Humer, G. De Luca, M. Fiebig, W. Heiss, M. V. Kovalenko, Nat.
Commun. 2015, 6, 8056.
[10] Y. Wang, X. Li, J. Song, L. Xiao, H. Zeng, H. Sun, Adv. Mater. 2015,
27, 7101.
[11] X. Li, Y. Wu, S. Zhang, B. Cai, Y. Gu, J. Song, H. Zeng, Adv. Funct.
Mater. 2016, 26, 2435.
[12] X. Zhang, C. Sun, Y. Zhang, H. Wu, C. Ji, Y. Chuai, P. Wang, S. Wen,
C. Zhang, W. W. Yu, J. Phys. Chem. Lett. 2016, 7, 4602.
[13] G. Li, H. Wang, T. Zhang, L. Mi, Y. Zhang, Z. Zhang, W. Zhang,
Y. Jiang, Adv. Funct. Mater. 2016, 26, 8478.
[14] E. Yassitepe, Z. Yang, O. Voznyy, Y. Kim, G. Walters, J. A. Castañeda,
P. Kanjanaboos, M. Yuan, X. Gong, F. Fan, J. Pan, S. Hoogland,
R. Comin, O. M. Bakr, L. A. Padilha, A. F. Nogueira, E. H. Sargent,
Adv. Funct. Mater. 2016, 26, 8757.
[15] J. Song, J. Li, X. Li, L. Xu, Y. Dong, H. Zeng, Adv. Mater. 2015, 27,
7162.
Thermal Analysis: TGA was performed using a nitrogen purge gas
over an aluminum pan containing ≈1 mg dried NCs using a Mettler
[16] M. Meyns, M. Perálvarez, A. Heuer-Jungemann, W. Hertog,
M. Ibáñez, R. Nafria, A. Genç, J. Arbiol, M. V. Kovalenko, J. Carreras,
A. Cabot, A. G. Kanaras, ACS Appl. Mater. Interfaces 2016, 8, 19579.
[17] Q. Zhou, Z. Bai, W. G. Lu, Y. Wang, B. Zou, H. Zhong, Adv. Mater.
2016, 28, 9163.
[18] C. Sun, Y. Zhang, C. Ruan, C. Yin, X. Wang, Y. Wang, W. W. Yu, Adv.
Mater. 2016, 28, 10088.
[19] A. Dualeh, P. Gao, S. I. Seok, M. K. Nazeeruddin, M. Grätzel, Chem.
Mater. 2014, 26, 6160.
Toledo TGA. Heating was performed from 25 to 550 °C at 20 °C min−1
.
Supporting Information
Supporting Information is available from the Wiley Online Library or
from the author.
[20] A. K. Baranwal, S. Kanaya, T. A. N. Peiris, G. Mizuta, T. Nishina,
H. Kanda, T. Miyasaka, H. Segawa, S. Ito, ChemSusChem 2016, 9,
2604.
[21] C. Yu, Z. Chen, J. J. Wang, W. Pfenninger, N. Vockic, J. T. Kenney,
K. Shum, J. Appl. Phys. 2011, 110, 063526.
[22] F. Zhang, H. Zhong, C. Chen, X. G. Wu, X. Hu, H. Huang, J. Han,
B. Zou, Y. Dong, ACS Nano 2015, 9, 4533.
[23] Q. Lü, H. Wei, W. Sun, K. Wang, Z. Gu, J. Li, S. Liu, S. Xiao,
Q. Song, Adv. Mater. 2016, 28, 10165.
Acknowledgements
Use of the Center for Nanoscale Materials, an Office of Science user
facility, was supported by the U.S. Department of Energy, Office of
Science, Office of Basic Energy Sciences, under Contract No. DE-AC02-
06CH11357. M.V.K. acknowledges financial support from the European
Union through the FP7 (ERC Starting Grant NANOSOLID, GA
No. 306733).
[24] Y. P. Varshni, Physica 1967, 34, 149.
[25] M. Kulbak, S. Gupta, N. Kedem, I. Levine, T. Bendikov, G. Hodes,
D. Cahen, J. Phys. Chem. Lett. 2016, 7, 167.
Keywords
[26] R. G. Niemann, L. Gouda, J. Hu, S. Tirosh, R. Gottesman,
P. J. Cameron, A. Zaban, J. Mater. Chem. A 2016, 4, 17819.
[27] M. Saliba, T. Matsui, J.-Y. Seo, K. Domanski, J.-P. Correa-Baena,
M. K. Nazeeruddin, S. M. Zakeeruddin, W. Tress, A. Abate,
A. Hagfeldt, M. Grätzel, Energy Environ. Sci. 2016, 9, 1989.
[28] S. Wei, Y. Yang, X. Kang, L. Wang, L. Huang, D. Pan, Chem.
Commun. 2016, 52, 7265.
CsPbX3, high-temperature, perovskites, photoluminescence, stability
Received: December 21, 2016
Revised: February 15, 2017
Published online:
[29] M. Sebastian, J. A. Peters, C. C. Stoumpos, J. Im, S. S. Kostina,
Z. Liu, M. G. Kanatzidis, A. J. Freeman, B. W. Wessels, Phys. Rev. B
2015, 92, 235210.
[30] S. Gonzalez-Carrero, R. E. Galian, J. Pérez-Prieto, Opt. Express 2015,
24, A285.
[1] M. Liu, M. B. Johnston, H. J. Snaith, Nature 2013, 501, 395.
[2] S. A. Veldhuis, P. P. Boix, N. Yantara, M. Li, T. C. Sum, N. Mathews,
S. G. Mhaisalkar, Adv. Mater. 2016, 28, 6804.
[3] S. D. Stranks, H. J. Snaith, Nat. Nanotechnol. 2015, 10, 391.
[4] N. K. Noel, S. D. Stranks, A. Abate, C. Wehrenfennig, S. Guarnera,
A.-A. Haghighirad, A. Sadhanala, G. E. Eperon, S. K. Pathak,
[31] A. Swarnkar, R. Chulliyil, V. K. Ravi, M. Irfanullah, A. Chowdhury,
A. Nag, Angew. Chem. Int. Ed. 2015, 54, 15424.
©
Adv. Funct. Mater. 2017, 1606750
1606750 (6 of 7)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.advancedsciencenews.com
www.afm-journal.de
[32] D. Valerini, A. Cretí, M. Lomascolo, L. Manna, R. Cingolani,
M. Anni, Phys. Rev. B 2005, 71, 235409.
[33] Z. Li-Gong, S. De-Zhen, F. Xi-Wu, L. Shao-Zhe, Chin. Phys. Lett.
2002, 19, 578.
[34] C. C. Stoumpos, C. D. Malliakas, J. A. Peters, Z. Liu, M. Sebastian,
J. Im, T. C. Chasapis, A. C. Wibowo, D. Y. Chung, A. J. Freeman,
B. W. Wessels, M. G. Kanatzidis, Cryst. Growth Des. 2013, 13, 2722.
[35] L. Mihut, S. Lefrant, I. Baltog, J. Appl. Phys. 1997, 82, 5391.
[36] B. T. Diroll, C. B. Murray, ACS Nano 2014, 8, 6466.
[37] G. Raino, T. Stöferle, I. Moreels, R. Gomes, J. S. Kamal, Z. Hens,
R. F. Mahrt, ACS Nano 2011, 5, 4031.
[42] L. A. Padilha, J. T. Stewart, R. L. Sandberg, W. K. Bae, W. K. Koh,
J. M. Pietryga, V. I. Klimov, Acc. Chem. Res. 2013, 46, 1261.
[43] G. Murtaza, I. Ahmad, Phys. B: Condens. Matter 2011, 406, 3222.
[44] S. Kondo, H. Nakagawa, T. Saito, H. Asada, Curr. Appl. Phys. 2004,
4, 439.
[45] R. Aceves, V. Babin, M. B. Flores, P. Fabeni, A. Maaroos, J. Lumin.
2001, 93, 27.
[46] A. Pan, B. He, X. Fan, Z. Liu, J. J. Urban, A. P. Alivisatos, L. He,
Y. Liu, ACS Nano 2016, 10, 7943.
[47] F. Palazon, F. Di Stasio, S. Lauciello, R. Krahne, M. Prato, L. Manna,
J. Mater. Chem. C 2016, 4, 9179.
[38] P. Jing, J. Zheng, M. Ikezawa, X. Liu, S. Lv, X. Kong, J. Zhao,
Y. Masumoto, J. Phys. Chem. C 2009, 113, 13545.
[48] C. E. Rowland, W. Liu, D. C. Hannah, M. K. Y. Chan, D. V. Talapin,
R. D. Schaller, ACS Nano 2014, 8, 977.
[39] H. Diab, G. Trippé-Allard, F. Lédée, K. Jemli, C. Vilar, G. Bouchez,
V. L. R. Jacques, A. Tejeda, J. Even, J.-S. Lauret, E. Deleporte,
D. Garrot, J. Phys. Chem. Lett. 2016, 7, 5093.
[40] Y. Zhao, C. Riemersma, F. Pietra, R. Koole, C. Donegá, A. Meijerink,
ACS Nano 2012, 6, 9058.
[49] G. Nedelcu, L. Protesescu, S. Yakunin, M. I. Bodnarchuk,
M. J. Grotevent, M. V. Kovalenko, Nano Lett. 2015, 15, 5635.
[50] Q. A. Akkerman, V. D’Innocenzo, S. Accornero, A. Scarpellini,
A. Petrozza, M. Prato, L. Manna, J. Am. Chem. Soc. 2015, 137, 10276.
[51] H. Shi, M.-H. Du, Phys. Rev. B 2014, 90, 174103.
[52] D. Bimberg, M. Sondergeld, E. Grobe, Phys. Rev. B 1971, 4, 3451.
[41] C. E. Rowland, R. D. Schaller, J. Phys. Chem. C 2013, 117, 17337.
©
Adv. Funct. Mater. 2017, 1606750
1606750 (7 of 7)
2017 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim