efficient blue and white perovskite light-emitting diodes ...efﬁcient blue and white perovskite...

Article

Efficient Blue and White Perovskite Light-Emitting Diodes via Manganese Doping

Shaocong Hou, Mahesh K.

Gangishetty, Qimin Quan,

Daniel N. Congreve

[email protected] (Q.Q.)

[email protected] (D.N.C.)

HIGHLIGHTS

Adding a small amount of Mn to

perovskite nanocrystals increases

their brightness

Mn doping leads to blue LEDs

with high efficiency and

brightness

Putting downconverters at the

front of the blue LED gives white

light

For modern color-rendering applications, efficient blue, green, and red LEDs are

required. While efficient green and red perovskite LEDs have been demonstrated,

blue devices lag significantly behind due to the poor quality of chloride-based

perovskites. Here, we show that doping manganese into blue perovskite

nanocrystals increases their brightness and efficiency. By putting doped

nanocrystals into LEDs, we see a 4-fold increase in efficiency, demonstrating that

blue LEDs can be as efficient as their red and green cousins.

Hou et al., Joule 2, 1–13

November 21, 2018 ª 2018 Elsevier Inc.

https://doi.org/10.1016/j.joule.2018.08.005

mailto:[email protected]



Please cite this article in press as: Hou et al., Efficient Blue and White Perovskite Light-Emitting Diodes via Manganese Doping, Joule (2018),https://doi.org/10.1016/j.joule.2018.08.005

Article

Efficient Blue and WhitePerovskite Light-EmittingDiodes via Manganese DopingShaocong Hou,1,2 Mahesh K. Gangishetty,1,2 Qimin Quan,1,* and Daniel N. Congreve1,3,*

Context & Scale

Light-emitting devices have

become an integral part of our

society, and new technologies

such as organic light-emitting

diodes (OLEDs) have come to

dominate the market, yet their

high cost has held them back from

truly ubiquitous deployment.

Perovskites have the potential to

fill that gap, as they are

inexpensive and have a

straightforward synthesis that

leads to bright materials with

sharp linewidths. To compete with

SUMMARY

Despite heavy research, blue perovskite nanocrystal LEDs have struggled to

match the high efficiencies of their red and green cousins, particularly at wave-

lengths blue enough to meet the NTSC blue standard. One of the most critical

problems is the low photoluminescence yield of the nanocrystals in thin films.

Recently, manganese doping has been shown to increase the photolumines-

cence yield of the perovskite even as a decay pathway to a long-lived emissive

state on the manganese ion is introduced. Here, we employ a two-step synthetic

approach to carefully tune the manganese doping to increase the blue photolu-

minescence while preventing significant manganese emission, allowing for blue

perovskite LEDs with quantum efficiencies over 2% that meet the NTSC stan-

dard. Manganese doping increases the photoluminescence yield and lifetime,

reduces trap states, andmakes the dots moremonodisperse, reducing the emis-

sion bandwidth. Finally, we utilize perovskite nanocrystal downconverters to

build an all-perovskite white LED.

OLEDs, efficient blue, green, and

red LEDs are required. Green and

red LEDs have been shown to be

efficient, but blue efficiencies

have failed to keep up. Here, we

add manganese to the perovskite

crystal lattice, increasing

brightness and allowing for

efficient blue LEDs. By adding

perovskite downconverters to the

front, we are further able to build

an all-perovskite white LED that

could eventually be used for room

lighting.

INTRODUCTION

Perovskites have become one of the most exciting material systems over the

past several years due to their low cost, solution processability, and high-quality

electronic properties.1,2 Rapid research has led to high-efficiency light-emitting

diodes (LEDs)3,4 and solar cells,5,6 demonstrating that these promising materials

can translate directly into high-performance device applications. Lead-halide

perovskite materials take the form ABX3, where A is typically the organic cation

methylammonium, formamidinium, or the inorganic cation cesium, B is lead, and

X is chloride, bromide, iodide, or a combination of the three. One of the unique

and attractive properties of this system is that the bandgap can be tuned by

adjusting the halide composition, from the deep blue chloride to the near infrared

iodide. When considered in the context of LEDs, this means that multi-color

systems such as white should be achievable by a combination of chemical

compositions.

Of particular recent interest are perovskite nanocrystals, which are nanocrystalline

grains of perovskite material surrounded by organic ligands. These materials are

easily synthesized and demonstrate high photoluminescence quantum yield

(PLQY).7 Red and green LEDs based on these nanocrystals have shown strong per-

formance, with external quantum efficiencies reaching over 5% for red and 10%

for green.8–14 For the achievement of white light and accurate color renderings in

real-world applications, however, efficient blue LEDs are also required, but the effi-

ciency of blue materials has lagged far behind.

Joule 2, 1–13, November 21, 2018 ª 2018 Elsevier Inc. 1

1Rowland Institute at Harvard University,


It is well understood that the PLQY of perovskite materials is dramatically lessened as

the energetics blue-shift.15 Great progress has been made in the sky-blue region

around 480–490 nm emission, with quantum efficiencies as high as 2.6%.8,16 To

meet the National Television System Committee (NTSC) standard, however, blue

LEDs must emit at approximately 470 nm or bluer. Only a handful of inorganic perov-

skite nanocrystal LEDs have been shown to work in this regime, with highly reduced

efficiencies relative to greener devices.11 Song et al. fabricated a bright-blue LED

with a 455-nm emission peak and an external quantum efficiency (EQE) as high as

0.07%.17 Yao et al. utilized solvent engineering techniques to achieve a bright

LED with a 470-nm emission peak and maximum EQE of 0.07%.18 Recently, we

discovered that the device structure plays an important role in the low performance

of the materials and designed transport layers to rectify the issue.19With these layers

in place, we saw an increase in EQE from 0.03% to 0.50%.19 After further evaluation,

we found that this 0.50% value was fundamentally limited by the photoluminescence

quantum yield of thematerial, which wemeasured in thin films at 9%, likely limited by

non-radiative defects present in higher-bandgap perovskites.20 Of the possible

20%–25% EQE theoretically available from a vertically structured LED,21 a huge frac-

tion is lost from this low photoluminescence yield in the nanocrystals. Clearly, a so-

lution to this efficiency ceiling must be found in order for these materials to become

commercially relevant.

Over the past year, manganese (Mn) doping of perovskite nanocrystals has been

demonstrated to provide several key benefits to the material. Initially developed

as a magnetic dopant, these Mn-doping sites proved to be highly efficient emitters

in their own right.22,23 Excitons generated in the nanocrystal quickly energy transfer

to the Mn ions,24 where they efficiently emit with orange-red color and high yield,

with an extremely long emissive lifetime.25 Later, it was demonstrated that the addi-

tion of Mn ions enhances the stability of CsPbBr3 nanocrystals in a green LED.26

Interestingly, despite focusing on the Mn energy transfer and emission, most papers

noted a significant increase in the blue perovskite PLQY, a surprising effect indi-

cating that the benefits of Mn doping might extend beyond magnetic nanocrystal

applications.

To take advantage of these materials, we needed first to find a way to increase the

PLQY of the perovskite without inducing a Mn-emission peak, which would ruin the

color purity of the LED. To do so, we carefully tune the Mn doping inside the perov-

skite nanocrystal. By adding only a small amount of dopant, we increase the PLQY of

the nanocrystal thin films by over 3-fold while crucially maintaining robust and pure

blue electroluminescence. The doping increases the photoluminescence lifetime

and brightness while slightly reducing the Urbach energy, pointing toward a reduc-

tion in trap states. We then construct pure blue LEDs and observe an increase in EQE

from 0.50% to 2.12%, neatly following the PLQY gains. The devices demonstrate

higher brightness and narrower emission bandwidth than the undoped control

devices, with a pure blue spectrum that meets the NTSC requirement. Finally,

we demonstrate the first all-perovskite white LEDs with a wide color gamut by utiliz-

ing red and green perovskite nanocrystals as optical downconverters at the front of

the LED.
Cambridge, MA 02142, USA
2These authors contributed equally

3Lead Contact

*Correspondence:[email protected] (Q.Q.),[email protected] (D.N.C.)


RESULTS AND DISCUSSION

We synthesize nanocrystals following Protesescu et al.7 with some modifications for

theMn addition.22,23 DopingMn into perovskite nanocrystals has only been success-

ful with MnCl2 precursors thus far. In our synthesis, the Mn emission is significantly

2 Joule 2, 1–13, November 21, 2018




A

B

C

Figure 1. Mn Doping in Perovskite Nanocrystals

(A) Perovskite crystal structure with Mn doping.

(B) Energy diagram of the doped nanocrystal system. Band-edge emission (1) competes with non-

radiative recombination (2) and energy transfer to the Mn ion (3), which then can emit on long

timescales (4). The perovskite emission occurs on the nanosecond timescale, while the long-lived

Mn emission is on the micro- to millisecond timescale.

(C) Photoluminescence spectra of the nanocrystals in solution as a function of Mn doping. Inset

shows a picture of the vials under UV irradiation.


suppressed when attempting a bromide-rich synthesis, indicating a synthetic or

thermodynamic limit of doping Mn in a Br-rich perovskite. In a typical one-step

hot-injection synthesis, the Mn-doping concentration was saturated at 0.09%, and

the Mn emission was barely observed in a Cl/Br-alloyed perovskite with the emission

peak at 468 nm (Figure 1C, curve of 0.09%), which is consistent with previous re-

ports.22 To overcome this limitation, we adopt a two-step sequential synthesis

approach utilizing both thermodynamic and kinetic processes. That is, we first syn-

thesize Cl-rich perovskite nanocrystals with high Mn concentration by hot injection

and then kinetically exchange Cl� with Br� at room temperature to align the emis-

sion peak to our targeted perovskite emission of 467–470 nm. We note that

exchanging from Cl/Br-alloyed nanocrystals, instead of pure Cl-based nanocrys-

tals,27 avoids excess anion exchange and purification processes, minimizing damage

to nanocrystals. This two-step sequential synthesis approach allows us to increase

Mn-doping concentration up to 16-fold, from 0.09% to 1.50%, while maintaining

an emission peak of approximately 468 nm and strong optoelectronic properties.

The emergence of a new broad photoluminescence peak due to exciton-to-dopant

Joule 2, 1–13, November 21, 2018 3

A B C

Figure 2. Optical Characterization of the Mn-Doped Nanocrystals

(A) PLQY and 1/e lifetime of the perovskite emission of thin films of nanocrystals, demonstrating

that mild doping is the key to efficient blue emission. Both values are for the perovskite peak only.

(B) Time-resolved perovskite luminescence of the different nanocrystal species.

(C) Urbach energy measurements on the absorption spectra, showing that Mn addition slightly

reduces the Urbach energy.


energy transfer confirmed the successful incorporation of additional Mn ions into the

perovskite nanocrystals. The Mn-doping concentrations quoted in this work were

determined by inductively coupled plasma emission spectroscopy analysis. See Fig-

ures S1 and S2 for a more detailed discussion of the doping process.

The photoluminescence of our final set of nanocrystals in solution can be seen in Fig-

ure 1C, normalized to the exciton peak around 470 nm. Unsurprisingly, an increase in

Mn concentration leads directly to an increase in Mn emission around 600 nm, reach-

ing a maximum of approximately 26% of the perovskite peak for our materials. This

Mn peak is less evident in thin films of nanocrystals such that it is only apparent in the

most heavily doped samples (see Figure S3).

The benefits of Mn doping can be clearly seen in Figure 2. We quickly noticed that

the brightness of spuncast films of thesematerials varied immensely withMn doping.

In Figure 2A, we measure the PLQY of the 470-nm perovskite peak of thin films of

nanocrystals. The undoped films gave a yield of only 9%, but the moderately doped

films increased to as high as 28% for the 0.19% doped sample. We hypothesize that

the roll-off in PLQY for higher doping levels is due to energy transfer to the Mn ions,

followed by eventual long-lived emission. Indeed, significant Mn luminescence is

observed for the most heavily doped thin film sample. The Mn-doping level is there-

fore a crucial balance for these blue materials: too little and the gains are limited, too

much and the Mn ions themselves act as a decay pathway.

To understand the source of this luminescence gain, we first measured the time-

resolved photoluminescence decay. The samples were excited by a 379-nm 81-ps

laser and measured with a streak camera; traces were integrated across the nano-

crystal emission wavelength around 470 nm to give the curves in Figure 2B (see Sup-

plemental Information for more information). The traces neatly match the PLQY in

Figure 2A: light doping shows a significant increase in emissive state lifetime before

additional Mn introduces a new decay pathway, reducing the lifetime at higher

doping levels as the energy transfers to the Mn ions. At the lower doping levels,

the brightness increase combined with the longer lifetime paints the picture of a

reduction in trap-state densities.

To investigate further, we measure the Urbach energy of the absorption tail using

UV-visible absorption spectroscopy.28 Typically, one would expect a larger Urbach

energy as dopants are added and contribute to the tail absorption.29 With these

materials, however, we observe a roughly constant or even slightly reduced Urbach

energy, with the undoped control giving 17.4 meV and the Mn-doped samples

4 Joule 2, 1–13, November 21, 2018

A

B

C

Figure 3. Structural Characterization of the

Mn-Doped Nanocrystals

(A) XRD of the nanocrystal samples shows a

slight lattice contraction that saturates at

moderate Mn densities.

(B) TEM of the 0.19% nanocrystals shows

excellent monodispersity. The crystal planes

noted agree with orthorhombic structure. Scale

bars, 50 nm (left) and 5 nm (right).

(C) EDX measurements of a single nanocrystal

show that the Mn is uniformly doped throughout

the nanocrystal. The dashed box indicates a

single nanocrystal. Scale bars, 5 nm.


reaching as low as 14.7 meV. Based on the slightly reduced Urbach energy, higher

lifetime, and greater PLQY in the presence of Mn, we conclude that the addition

of a small amount of Mn to these nanocrystals passivates trap states, reducing the

non-radiative decay.

To further understand the improvements in the material, we use X-ray diffraction

(XRD) as shown in Figure 3A. The XRD patterns are similar for all Mn dopings, indi-

cating an orthorhombic crystal structure, consistent with previous results in the liter-

ature.22,26 We observe a slight shift in the diffraction peak along the (202) plane

at 31�, indicating a lattice contraction as the smaller Mn replaces the larger Pb. Inter-

estingly, this effect saturates at the same Mn concentration (0.19%) that gives us the

maximum PLQY and above which Mn emission becomes much more dominant.

We next turn to device fabrication to identify the engineering benefits of

this material improvement. Previously, we demonstrated that a poly(3,4-ethylene-

dioxythiophene):poly(styrenesulfonate) (PEDOT:PSS)/TFB (poly(9,9-di-n-octyl-

fluorene-alt-(1,4-phenylene-(4-sec-butylphenyl)imino)-1,4-phenylene))/PFI (Nafion,

perfluorinated ionomer) hole transport layer (HTL) allowed for much higher quan-

tum efficiencies at these blue wavelengths.19 We fabricate devices in a similar

way, which can be briefly summarized as follows. PEDOT:PSS, TFB, and PFI are

sequentially spun onto a cleaned indium tin oxide (ITO)-on-glass slide, with an

anneal between each spin step. The synthesized nanocrystals are then spuncast

onto the HTL. The slides are transferred to a thermal evaporator where 40 nm of

TPBi, 1.1 nm of LiF, and 60 nm of aluminum are evaporated on top, defining a

2-mm diameter device. The device structure can be found in Figure 4A; the details

of the fabrication are given in Supplemental Information.

The electroluminescence spectra of the devices are presented in Figure 4B. All sam-

ples show pure blue emission from the perovskite emitter layer only (see Figure S6).

Joule 2, 1–13, November 21, 2018 5

A B

DC

Figure 4. Structure and Performance of Blue LEDs

(A) Device structure of the LEDs in this work. Energy levels of the transport layers are from Li et al.11

and Yao et al.18 PFI is represented as bandbending to a deeper work function, as it has been

previously demonstrated to increase the surface work function.30,31 The energy levels of the

nanocrystals are determined from XPS.19

(B) Electroluminescence spectra of the fabricated LEDs with nanocrystals of varying Mn content.

(C) J-V-L characteristics show increased brightness across all doping concentrations. Inset shows an

image of the 0.19% Mn device.

(D) The EQE of the fabricated devices show up to a 4-fold enhancement over control nanocrystals.

The 0.19% Mn device achieves a maximum EQE of 2.12%. Color coding for the data in (C) and (D)

follow the legend in (B). 0% Mn curves are from Gangishetty et al.19


Emission from the Mn peak was not observed from any samples; we hypothesize

that the very long emissive lifetime saturates the emission in such a thin layer (see

Supplemental Information for a further discussion of the Mn-emission quenching).

The TFB underlayer seems to reduce Mn emission as well (Figure S5). Thus, we are

able to use mild doping to achieve performance enhancement without sacrificing

color purity. Increasing Mn content leads to a sharp reduction in the full width at

half maximum (FWHM) of the emission peak to as low as 17.1 nm. This is likely

due to the increasedmonodispersity of the nanocrystals, as can be seen in Figure 3B.

Indeed, by monitoring the emission spectrum as a function of time with a streak cam-

era, we can clearly see a red shift in emission peak in the first few nanoseconds in the

undoped samples, which disappears in the Mn-doped nanocrystals (see Figure S4),

correlating with the reduced FWHM. This red shift in undoped nanocrystals is greatly

reduced in solution, indicating that it is likely inter-nanocrystal energy transfer from

polydispersity,32 although further study of this phenomenon is required. See Table 1

for the FWHM of all synthesized samples.

In Figure 4C we present the J-V-L (current density-voltage-luminance) characteristics

for these devices. The Mn again provides a number of benefits, showing a reduction

in the turn-on voltage of around 1 V and stronger brightness under all voltage

conditions. The devices demonstrate a maximum brightness of 389 cd/m2 for the

0.25% Mn content device, over three times brighter than the best undoped control.

6 Joule 2, 1–13, November 21, 2018

Table 1. Material and Device Parameters of the Mn Doped Nanocrystals

Mn Content 1/e Lifetime* PLQY* Maximum EQE EL FWHM Maximum Brightness Emission Peak Maximum EQE above100 cd/m2

0% 2.08 ns 9% 0.50% 25.1 nm 111 cd/m2 469 nm 0.05%

0.09% 3.03 ns 12% 0.73% 18.4 nm 245 cd/m2 466 nm 0.25%

0.19% 3.23 ns 28% 2.12% 17.9 nm 245 cd/m2 466 nm 0.71%

0.25% 2.50 ns 16% 1.46% 17.1 nm 389 cd/m2 470 nm 1.15%

1.50% 1.77 ns 9% 0.61% 17.7 nm 250 cd/m2 470 nm 0.22%

*Of the perovskite peak only.


The benefits of the Mn doping can be even more clearly recognized by examining

the EQE (Figure 4D). We observe up to a 4-fold improvement in EQE with the

0.19% Mn doping, reaching a maximum of 2.12%, which, to the best of our knowl-

edge, is the highest value at this wavelength range for perovskite nanocrystals.

The 0.25% Mn sample also showed strong EQE, reaching as high as 1.46%. These

high quantum efficiency values closely track the overall improvement in PLQY and

lifetime (see Table 1), demonstrating that these material improvements translate

directly to strong engineering value. Furthermore, the devices maintain good

efficiency at application-level brightness, with the 0.25% Mn device demonstrating

an EQE of 1.15% at brightness above 100 cd/m2.

These high-performance devices allow for previously inaccessible technologies such

as white perovskite LEDs to be achieved. To further demonstrate this, we construct

white LEDs as shown in Figure 5. On the other side of the glass substrate of our

efficient blue perovskite LEDs, we place red and green downconverters made

from perovskite nanocrystals. These absorb a portion of the blue light and re-radiate

it at the longer wavelength, allowing for red, green, and the partially transmitted

blue light to reach the eye. When these are properly calibrated, the eye sees white

light.

To build our downconverters, we synthesize red and green cesium lead-halide

perovskite nanocrystals. The nanocrystals are then suspended in a 200 mg/mL

polyisobutylene-in-toluene mixture and poured onto a glass slide. The film is

allowed to dry overnight, yielding a rubber film that downconverts to the appro-

priate color (see inset in Figure 5B). Further details of the synthesis can be found

in Supplemental Information. This approach allows for easy fabrication and tuning

as well as the prevention of halide exchange between nanocrystals, a common

problem in these materials.15

Once appropriately tuned, we can measure the properties of the entire structure

as a single LED, yielding the spectrum shown in Figure 5B. The transmitted blue

electroluminescence is observed at 469 nm, consistent with the devices in Figure 4.

The downconverted peaks are evident at 511 and 627 nm, closely matching the

photoluminescence measured on the downconverters individually (dashed lines).

The device gives off white emission by eye and camera (see inset in Figure 5E).

With the spectrum in hand, we can calculate the color coordinates of the device. The

CIE spectrum represents the colors seen by the human eye (see Figure 5C). White

occupies the center of this plot at a value of (0.33, 0.33). This white device is close

to this true white value, with a CIE coordinate of (0.311, 0.326). Furthermore, we

examine the coordinates of the blue device and show that it achieves a coordinate

of (0.127, 0.077), meeting the crucial blue NTSC standard of (0.14, 0.08). When

Joule 2, 1–13, November 21, 2018 7

A B C

ED

Figure 5. Structure and Performance of White LEDs

(A) Schematic of our white LED design. Blue photons are absorbed by red and green downconverters which, when properly mixed, give white light.

(B) Photoluminescence from the downconverter layers (dashed red and green) and electroluminescence from the blue and white LEDs (solid lines). Inset

shows the downconverters under room lights and UV illumination.

(C) The CIE color coordinates of our devices. The photoluminescence of the downconverters are represented by squares; the blue and white

electroluminescence is represented by circles. The white triangle represents the color space achievable with these three sets of nanocrystals. The white

LED has a CIE coordinate of (0.311, 0.326).

(D and E) The J-V-L (D) and EQE (E) curves of the white LED, peaking at a brightness of 102 cd/m2 and a quantum efficiency of 0.25%. Inset shows an

image of the white LED.


combined with the downconverters, we can cover a large percentage of the CIE plot,

represented by the white triangle. The white LED has a color-rendering index of 10.

We note that this low value can trivially be improved by inclusion of a higher number

of downconversion nanocrystal wavelengths.

Finally, we can measure the efficiency of the overall white LED structure. The J-V-L

curves are presented in Figure 5D, and the EQE curve in Figure 5E. The device peaks

at a luminance of 102 cd/m2 and an efficiency of 0.25%. When compared with the

base blue device, we see a reduction of approximately 10-fold in efficiency. This

value agrees well with the measured PLQY of the downconverters (67% and 85%

for green and red, respectively) and the significant waveguide losses in the rubber

films. Future work must focus on increasing the outcoupling of the downconverter

films, whether by roughening the surface or adding scatterers to the rubber.

These films are fairly stable to illumination (see Figure S9). Overall, this LED demon-

strates the versatility of these inexpensive, bright, and easily synthesized perovskite

materials.

8 Joule 2, 1–13, November 21, 2018


Stability remains a significant concern for perovskite LEDs, as most decay within

seconds or minutes.33 In particular, blue devices have demonstrated poor stability,

as their higher energetics and mixed halide character has led to both rapid

degradation and color shifting under bias.34 We observe both effects on our control

device (see Figures S7 and S8). The Mn doping, in addition to the high opto-electronic

benefits demonstrated here, allows for moderate improvements in operational

stability with a significant increase in spectral stability, as seen in Video S1. Recent

work has shown color shifts due to ion migration to be tied to halide vacancies and

defects,35 and thus we hypothesize that the better color stability in the Mn materials

arises from the same defect reduction observed in Figure 2. Although a step forward,

much work remains to be done before these materials are stable enough for

commercialization.

In this work, we have synthesized Mn-doped blue perovskite nanocrystals.

We have demonstrated that the Mn dopant provides strong benefits to the

material, increasing the emissive lifetime and photoluminescence quantum

yield while reducing the trap states, as reflected in a slight reduction of

Urbach energy. Finally, we have fabricated NTSC blue and white LEDs, clearly

demonstrating the commercial value of these materials. Mn doping provides a

straightforward way to increase the performance of blue-emitting perovskite

nanocrystals, allowing for a variety of new and exciting applications for these

materials.

EXPERIMENTAL PROCEDURES

All chemicals were purchased from Sigma-Aldrich unless otherwise specified.

For preparation of cesium-oleate (Cs-oleate) precursors, 0.814 g of cesium carbon-

ate (Cs2CO3, purity 99.9%), 40 mL of octadecene (purity 90%), and 2.5 mL of oleic

acid (purity 90%) were loaded into a 100-mL flask, dried under vacuum at 120�Cfor 1 hr, and heated to 150�C under nitrogen (N2) protection, yielding a clear

solution. The solution was cooled to room temperature for storage, and reheated

up to 100�C under vacuum before use.

For synthesis of undoped perovskite nanocrystals, 165 mg of PbBr2 (0.450 mmol,

purity 98%), 83.6 mg of PbCl2 (0.301 mmol, purity 98%), 20 mL of

octadecene, 2 mL of oleylamine (purity 98%), 2 mL of oleic acid, and 2 mL of

trioctylphosphine (purity 97%) were loaded into a 100-mL three-neck flask,

dried under vacuum at 130�C for 45 min, and heated to 150�C under stirring.

The yielded clear solution was heated to 165�C under N2 protection, after

which 1.7 mL of pre-heated Cs-oleate precursors was swiftly injected into the

solution. After reacting for 10 s, the product was cooled to room temperature

in an ice/water bath.

Synthesis of low Mn2+-doped perovskite nanocrystals (0.09% nomenclature)

involved loading 210 mg (0.572 mmol) of lead(II) bromide (PbBr2, purity 98%),

50 mg of manganese(II) chloride (MnCl2, purity 99.999%), 20 mL of octadecene,

2 mL of oleylamine (purity 98%), 2 mL of oleic acid, and 2 mL of trioctylphosphine

(purity 97%) into a 100-mL three-neck flask. This was dried at 130�C for 45 min

and heated to 150�C under vacuum. The yielded solution was then heated to

165�C under N2 protection, after which 1.7 mL of pre-heated Cs-oleate precur-

sors was rapidly injected into the solution. After having reacted for 10 s, a crude

product was cooled to room temperature in an ice/water bath.

Joule 2, 1–13, November 21, 2018 9


Synthesis of high Mn2+-doped perovskite nanocrystals (0.19%, 0.25%, 1.50%

nomenclature) was the same as for the low-doped synthesis, except more MnCl2was used (70, 80, or 90 mg of MnCl2 precursors for 0.19%, 0.25%, and 1.50% doping

concentrations, respectively). A blue shift of the perovskite emission peak was

observed, which required alignment back to 468 nm by Br� post-exchanging as

described below.

The blue emission shift of nanocrystals was aligned by exchanging the as-synthesized

colloidal nanocrystal solution with PbBr2 stock solution. PbBr2 stock solution was pre-

pared bymixing 2.2 g of PbBr2 (6mmol), 50mL of octadecene, 2mL of oleylamine, and

2mL of oleic acid in a 100-mL flask at 130�Cunder vacuum for 40min, after which it was

cooled down to room temperature. The amount of PbBr2 stock solution used was care-

fully adjusted by tracking the fluorescent emission peak with a spectrometer (Ocean

Optics QEPro). As an example, 0.7 mL of PbBr2 stock solution was used to align the

emission of the nanocrystals prepared with 70 mg of MnCl2 at 468 nm. The 0.19%

Mn nanocrystals achieved a PLQY of 44% in solution. See Table S1 for composition

of the different nanocrystal series.

For purification of the perovskite nanocrystals, an equal volume of anhydrous ethyl

acetate (purity 99.8%) was added to an aforementioned crude product to precipitate

the nanocrystals. After centrifuging at 4 krpm (kilorevolutions per minute) for 5 min,

the precipitate was dissolved in 10 mL of anhydrous hexane (purity 95%). The nano-

crystals were washed again in anhydrous ethyl acetate (volume ratio of ethyl acetate/

hexane 2.2:1), centrifuged at 7 krpm for 5 min, and redispersed in 8 mL of octane or

hexane. This yielded a solution of approximately 16 mg/mL, which was diluted to

11 mg/mL before use in devices. The solution was filtered by polytetrafluoroethy-

lene (PTFE) filter (0.2 mm) before use.

For synthesis of downconversion green perovskite nanocrystals, 276 mg of PbBr2(0.752 mmol, purity 98%), 20 mL of octadecene, 2 mL of oleylamine (purity 98%),

and 2 mL of oleic acid were loaded into a 100-mL three-neck flask, dried under

vacuum at 130�C for 40 min, and heated to 150�C under stirring. The yielded clear

solution was heated to 184�C under N2 protection, after which 1.7 mL of pre-heated

Cs-oleate precursors was swiftly injected into the solution, which turned to a yellow

color immediately. After reacting for 5 s, the product was cooled to room tempera-

ture in an ice/water bath. Before preparing the green downconversion film, the

crude nanocrystal products were purified by centrifuging at 7 krpm for 10 min and

then dispersed in 5 mL of hexane.

For synthesis of downconversion red perovskite nanocrystals, 108.8 mg of PbBr2(0.296 mmol, purity 98%), 208.2 mg of PbI2 (0.452 mmol), 20 mL of octadecene,

2 mL of oleylamine (purity 98%), and 2 mL of oleic acid were loaded into a 100-mL

three-neck flask, dried under vacuum at 130�C for 40 min, and heated to 150�Cunder stirring. The yielded clear solution was heated to 184.5�C under N2 protec-

tion, after which 1.7 mL of pre-heated Cs-oleate precursors was swiftly injected

into the solution. After reacting for 5 s, the product was cooled to room temperature

in an ice/water bath. The crude nanocrystal products were purified by centrifuging

at 7 krpm for 10 min and then dispersed in 5 mL of hexane.

For fabrication of the downconverting layers, polyisobutylene (MW �500,000) was

mixed at 200mg/mL in toluene and heated at 90�C overnight until a viscous, uniform

solution was obtained, 2 mL of which was added to a 7-mL glass vial, wherein 300 mL

of the green nanocrystal solution or 100 mL of the red nanocrystal solution was

10 Joule 2, 1–13, November 21, 2018


added. The solution was stirred thoroughly with a spatula until uniformly mixed. The

vial was left to sit for 2 hr until the bubbles disappeared and was then poured slowly

onto a glass slide. This was then allowed to dry overnight in a N2 environment. The

rubber films could then be cut and stacked to achieve a specific color goal. For the

white LED, three red and seven green films were stacked to achieve the appropriate

amount of downconversion.

For fabrication of the white LED, the previously mentioned downconverting layers

were placed on a 0.7-mm glass slide. The edges were wrapped with black tape to

prevent any wave-guided light from being collected. The slide was then combined

with the blue LED such that the glass slide was in contact with the green

downconverters, with the red downconverter on top of the green. The red

downconverters were placed directly against the glass face of the blue LED.

Measurements of efficiency then proceeded as usual.

For fabrication of the blue LEDs, we followed our previously established proced-

ure, which we briefly repeat here. Nafion perfluorinated resin solution 5 wt% in

lower aliphatic alcohols and water (PFI), LiF (evaporation grade), and aluminum

were purchased from Sigma-Aldrich and used as received. ITO substrates, TPBi,

and TFB were purchased from Luminescence Technology and used as received.

PEDOT:PSS was purchased from Heraeus (Clevios P VP AI 4083) and used as

received.

ITO patterned glass was solvent cleaned by sonication in Micron-90 detergent,

23 water, 23 acetone, and soaking in boiling isopropanol for 10 min per step.

They were dried under blowing air and treated with O2 plasma at 200 W using

0.5 Torr O2 gas for 5 min. PEDOT:PSS was spuncast through a 0.4-mm polyvinyli-

dene fluoride filter at 4,000 rpm for 45 s (ramp of 2,000 rpm/s) in air, and annealed

at 140�C for 30 min in a N2 environment. After cooling, a TFB (4 mg/mL in chloro-

benzene) layer was spuncast at 3,000 rpm for 45 s (ramp of 2,000 rpm/s), and

annealed at 140�C for 20 min. A thin layer of PFI (0.05 wt% in isopropanol) was

then spuncast at 3,000 rpm for 45 s and dried at 145�C for 10 min. On top of these

layers, perovskite nanocrystals in octane were spuncast. The nanocrystal layer was

approximately 20–25 nm thick. These films were then transferred to an evaporation

chamber, where 40 nm TPBi, 1.1 nm LiF, and 60 nm Al were deposited at 6 3

10�6 mbar at 2 A/s, 0.2 A, and 3 A/s, respectively. Devices were unpackaged

and measured in air.

For transient measurements, samples were excited by a 379-nm laser from Hama-

matsu (C10196) with an 81-ps pulse width. Optical density filters were used to

reduce the excitation intensity. The excitation was incident at a 45� angle to the glass

face. The photoluminescence was collected with a 25.4-mm focal length lens normal

to the glass face. It was focused through a 400-nm longpass filter into an SP2150i

spectrograph coupled to a Hamamatsu C10627 streak unit and C9300 digital

camera. Data were integrated across the nanocrystal emission wavelength to obtain

the curves presented in Figure 2.

For device characterization, EL spectra were taken with an Ocean Optics QE Pro

with 1 mA or 500 mA sourced to the device from a Keithley 2400. Current-voltage

and EQE characteristics were measured with an HP 4145A with a calibrated

half-inch ThorLabs photodetector physically held just above the face of the device,

negating the need for a geometric correction. The device (1 mm radius) was much

smaller than the photodetector (9.7 mm). The photodetector was smaller than the

Joule 2, 1–13, November 21, 2018 11


glass slide (12.2 mm), which, combined with the black material construction of the

EQE holder, blocks the collection of wave-guided light, preventing overestimation

of the EQE.36 Luminance was calculated from the J-V-L curves and the spectra of

the device.

Materials Characterization

Morphologies of the perovskite nanocrystals were characterized by a JEOL ARM

200F STEM (80 kV). Elemental mapping was characterized by a transmission electron

microscope (JEOL 2100, 120 kV). UV-visible absorption spectra were recorded by

a PerkinElmer Lambda 25, and the photoluminescence spectra were collected by

an Ocean Optics spectrometer (QE Pro) pumped by a 365-nm LED source. XRD

was acquired by a Bruker D2 Phaser.

For measurement of the PLQY of the samples, we followed themethod pioneered by

de Mello et al.37 In brief, the luminescence with the sample in the beam, with the

sample out of the beam, and the beam without the sample are all measured in an

integrating sphere. From these measurements the PLQY can be accurately calcu-

lated. The PLQY was measured on thin films of the same nanocrystals used in the de-

vices. All optical components, including the sphere, were calibrated against a

calibrated Newport photodetector.

DATA AND SOFTWARE AVAILABILITY

The data presented in this work are available from the corresponding authors upon

reasonable request.

SUPPLEMENTAL INFORMATION

Supplemental Information includes Supplemental Experimental Procedures, nine

figures, one table, and one video and can be found with this article online at

https://doi.org/10.1016/j.joule.2018.08.005.

ACKNOWLEDGMENTS

The authors are supported by the Rowland Fellowship at the Rowland Institute at

Harvard University. This work was performed in part at the Center for Nanoscale Sys-

tems, a member of the National Nanotechnology Coordinated Infrastructure

Network, which is supported by the National Science Foundation under NSF award

no. 1541959. The authors especially thank Cynthia Friend, Mike Burns, and Angela

Healy, who made this work possible. We thank Tony Wu for helpful discussions.

AUTHOR CONTRIBUTIONS

S.H. synthesized all nanocrystals and made material characterizations. M.K.G. fabri-

cated all LEDs. D.N.C. designed and fabricated the downconverters. Q.Q. and

D.N.C. guided the research. All authors contributed to the manuscript.

DECLARATION OF INTERESTS

Harvard University has filed a patent based on this work that names the authors as

inventors.

Received: June 19, 2018

Revised: July 20, 2018

Accepted: August 15, 2018

Published: September 5, 2018

12 Joule 2, 1–13, November 21, 2018



REFERENCES

1. Stranks, S.D., Eperon, G.E., Grancini, G.,Menelaou, C., Alcocer, M.J.P., Leijtens, T.,Herz, L.M., Petrozza, A., and Snaith, H.J. (2013).Electron-hole diffusion lengths exceeding 1micrometer in an organometal trihalideperovskite absorber. Science 342, 341–344.

2. Shi, D., Adinolfi, V., Comin, R., Yuan, M.,Alarousu, E., Buin, A., Chen, Y., Hoogland, S.,Rothenberger, A., Katsiev, K., et al. (2015). Lowtrap-state density and long carrier diffusion inorganolead trihalide perovskite single crystals.Science 347, 519–522.

3. Veldhuis, S.A., Boix, P.P., Yantara, N., Li, M.,Sum, T.C., Mathews, N., and Mhaisalkar, S.G.(2016). Perovskite materials for light-emittingdiodes and lasers. Adv. Mater. 28, 6804–6834.

4. Tan, Z.-K., Moghaddam, R.S., Lai, M.L.,Docampo, P., Higler, R., Deschler, F., Price, M.,Sadhanala, A., Pazos, L.M., Credgington, D.,et al. (2014). Bright light-emitting diodes basedon organometal halide perovskite. Nat.Nanotechnol. 9, 687–692.

5. Yang, W.S., Noh, J.H., Jeon, N.J., Kim, Y.C.,Ryu, S., Seo, J., and Seok, S.I. (2015). High-performance photovoltaic perovskite layersfabricated through intramolecular exchange.Science 348, 1234–1237.

6. Liu, M., Johnston, M.B., and Snaith, H.J. (2013).Efficient planar heterojunction perovskite solarcells by vapour deposition. Nature 501,395–398.

7. Protesescu, L., Yakunin, S., Bodnarchuk, M.I.,Krieg, F., Caputo, R., Hendon, C.H., Yang, R.X.,Walsh, A., and Kovalenko, M.V. (2015).Nanocrystals of cesium lead halide perovskites(CsPbX3, X = Cl, Br, and I): noveloptoelectronic materials showing brightemission with wide color gamut. Nano Lett. 15,3692–3696.

8. Pan, J., Quan, L.N., Zhao, Y., Peng, W., Murali,B., Sarmah, S.P., Yuan, M., Sinatra, L., Alyami,N.M., Liu, J., et al. (2016). Highly efficientperovskite-quantum-dot light-emitting diodesby surface engineering. Adv. Mater. 28, 8718–8725.

9. Zhang, X., Sun, C., Zhang, Y., Wu, H., Ji, C.,Chuai, Y., Wang, P., Wen, S., Zhang, C., and Yu,W.W. (2016). Bright perovskite nanocrystalfilms for efficient light-emitting devices.J. Phys. Chem. Lett. 7, 4602–4610.

10. Li, J., Xu, L., Wang, T., Song, J., Chen, J., Xue,J., Dong, Y., Cai, B., Shan, Q., Han, B., andZeng, H. (2017). 50-fold EQE improvement upto 6.27% of solution-processed all-inorganicperovskite CsPbBr3 QLEDs via surface liganddensity control. Adv. Mater. 29, 1603885.

11. Li, G., Rivarola, F.W.R., Davis, N.J.L.K., Bai, S.,Jellicoe, T.C., de la Pena, F., Hou, S., Ducati, C.,Gao, F., Friend, R.H., et al. (2016). Highlyefficient perovskite nanocrystal light-emittingdiodes enabled by a universal crosslinkingmethod. Adv. Mater. 28, 3528–3534.

12. Song, J., Li, J., Xu, L., Li, J., Zhang, F., Han, B.,Shan, Q., and Zeng, H. (2018).Room-temperature triple-ligand surfaceengineering synergistically boosts ink stability,recombination dynamics, and charge injection

toward EQE-11.6% perovskite QLEDs. Adv.Mater. 30, e1800764.

13. Yan, F., Xing, J., Xing, G., Quan, L., Tan, S.T.,Zhao, J., Su, R., Zhang, L., Chen, S., Zhao, Y.,et al. (2018). Highly efficient visible colloidallead-halide perovskite nanocrystal light-emitting diodes. Nano Lett. 18, 3157–3164.

14. Chin, X.Y., Perumal, A., Bruno, A., Yantara, N.,Veldhuis, S.A., Martınez-Sarti, L., Chandran,B.K., Chirvony, V.S., Lo, S.-Z.A., So, J., et al.(2018). Self-assembled hierarchicalnanostructured perovskites enable highlyefficient LEDs via an energy cascade. EnergyEnviron. Sci. 11, 1770–1778.

15. Nedelcu, G., Protesescu, L., Yakunin, S.,Bodnarchuk, M.I., Grotevent, M.J., andKovalenko, M.V. (2015). Fast anion-exchange inhighly luminescent nanocrystals of cesium leadhalide perovskites (CsPbX3, X = Cl, Br, I). NanoLett. 15, 5635–5640.

16. Wang, Q., Ren, J., Peng, X.-F., Ji, X.-X., andYang, X.-H. (2017). Efficient sky-blue perovskitelight-emitting devices based onethylammonium bromide induced layeredperovskites. ACS Appl. Mater. Interfaces 9,29901–29906.

17. Song, J., Li, J., Li, X., Xu, L., Dong, Y., and Zeng,H. (2015). Quantum dot light-emitting diodesbased on inorganic perovskite cesium leadhalides (CsPbX3). Adv. Mater. 27, 7162–7167.

18. Yao, E.P., Yang, Z., Meng, L., Sun, P., Dong, S.,Yang, Y., and Yang, Y. (2017). High-brightnessblue and white LEDs based on inorganicperovskite nanocrystals and their composites.Adv. Mater. 29, https://doi.org/10.1002/adma.201606859.

19. Gangishetty, M.K., Hou, S., Quan, Q., andCongreve, D.N. (2018). Reducing architecturelimitations for efficient blue perovskite light-emitting diodes. Adv. Mater. 30, e1706226.

20. Stranks, S.D. (2017). Nonradiative losses inmetal halide perovskites. ACS Energy Lett. 2,1515–1525.

21. Meerheim, R., Furno, M., Hofmann, S., Lussem,B., and Leo, K. (2010). Quantification of energyloss mechanisms in organic light-emittingdiodes. Appl. Phys. Lett. 97, 253305.

22. Parobek, D., Roman, B.J., Dong, Y., Jin, H., Lee,E., Sheldon, M., and Son, D.H. (2016). Exciton-to-dopant energy transfer inMn-doped cesiumlead halide perovskite nanocrystals. Nano Lett.16, 7376–7380.

23. Liu, W., Lin, Q., Li, H., Wu, K., Robel, I., Pietryga,J.M., and Klimov, V. (2016). Mn2+-doped leadhalide perovskite nanocrystals with dual-coloremission controlled by halide content. J. Am.Chem. Soc. 138, 14954–14961.

24. Rossi, D., Parobek, D., Dong, Y., and Son, D.H.(2017). Dynamics of exciton-Mn energy transferin Mn-doped CsPbCl3 perovskite nanocrystals.J. Phys. Chem. C 121, 17143–17149.

25. Guria, A.K., Dutta, S.K., Adhikari, S.D., andPradhan, N. (2017). Doping Mn2+ in lead halideperovskite nanocrystals: successes andchallenges. ACS Energy Lett. 2, 1014–1021.

26. Zou, S., Liu, Y., Li, J., Liu, C., Feng, R., Jiang, F.,Li, Y., Song, J., Zeng, H., Hong, M., and Chen,X. (2017). Stabilizing cesium lead halideperovskite lattice through Mn(II) substitutionfor air-stable light-emitting diodes. J. Am.Chem. Soc. 139, 11443–11450.

27. Chen, D., Fang, G., and Chen, X. (2017). Silica-coated Mn-doped CsPb(Cl/Br)3 inorganicperovskite quantum dots: exciton-to-Mnenergy transfer and blue-excitable solid-statelighting. ACS Appl. Mater. Interfaces 9, 40477–40487.

28. Wenger, B., Nayak, P.K., Wen, X., Kesava, S.V.,Noel, N.K., and Snaith, H.J. (2017).Consolidation of the optoelectronic propertiesof CH3NH3PbBr3 perovskite single crystals.Nat. Commun. 8, 590.

29. Mocatta, D., Cohen, G., Schattner, J., Millo, O.,Rabani, E., and Banin, U. (2011). Heavily dopedsemiconductor nanocrystal quantum dots.Science 332, 77–81.

30. Kim, Y.-H., Wolf, C., Kim, Y.-T., Cho, H., Kwon,W., Do, S., Sadhanala, A., Park, C.G., Rhee,S.W., Im, S.H., et al. (2017). Highly efficientlight-emitting diodes of colloidal metal-halideperovskite nanocrystals beyond quantum size.ACS Nano 11, 6586–6593.

31. Zhang, X., Lin, H., Huang, H., Reckmeier, C.,Zhang, Y., Choy, W.C., and Rogach, A.L. (2016).Enhancing the brightness of cesium lead halideperovskite nanocrystal based green light-emitting devices through the interfaceengineering with perfluorinated ionomer.Nano Lett. 16, 1415–1420.

32. Voznyy, O., Levina, L., Fan, F., Walters, G., Fan,J.Z., Kiani, A., Ip, A.H., Thon, S.M., Proppe,A.H., Liu, M., and Sargent, E.H. (2017). Originsof stokes shift in PbS nanocrystals. Nano Lett.17, 7191–7195.

33. Himchan, C., Young-Hoon, K., Christoph, W.,Hyeon-Dong, L., and Tae-Woo, L. (2018).Improving the stability of metal halideperovskite materials and light-emitting diodes.Adv. Mater. https://doi.org/10.1002/adma.201704587.

34. Vashishtha, P., and Halpert, J.E. (2017). Field-driven ionmigration and color instability in red-emitting mixed halide perovskite nanocrystallight-emitting diodes. Chem. Mater. 29, 5965–5973.

35. Barker, A.J., Sadhanala, A., Deschler, F.,Gandini, M., Senanayak, S.P., Pearce, P.M.,Mosconi, E., Pearson, A.J., Wu, Y., Kandada,A.R.S., et al. (2017). Defect-assistedphotoinduced halide segregation in mixed-halide perovskite thin films. ACS Energy Lett. 2,1416–1424.

36. Forrest, S.R., Bradley, D.D.C., and Thompson,M.E. (2003). Measuring the efficiency of organiclight-emitting devices. Adv. Mater. 15, 1043–1048.

37. deMello, J.C., Wittmann, H.F., and Friend, R.H.(1997). An improved experimentaldetermination of external photoluminescencequantum efficiency. Adv. Mater. 9, 230–232.

Joule 2, 1–13, November 21, 2018 13

http://refhub.elsevier.com/S2542-4351(18)30385-4/sref1

































































































https://doi.org/10.1002/adma.201606859


































































































efficient blue and white perovskite light-emitting diodes ...efﬁcient blue and white perovskite...

Documents