1Department of Physics, Chemistry and Biology (IFM), Linköping University, Linköping, Sweden. 2Key Laboratory of Flexible Electronics (KLOFE) and Institute of Advanced Materials (IAM), Jiangsu National Synergetic Innovation Center for Advanced Materials (SICAM), Nanjing Tech University (NanjingTech), Nanjing, China. 3Beijing Computational Science Research Center, Beijing, China. 4International Collaborative Laboratory of 2D Materials for Optoelectronics Science and Technology, Shenzhen University, Shenzhen, China. 5Center for Nano Science and Technology @Polimi, Istituto Italiano di Tecnologia, Milan, Italy. 6School of Materials Science and Engineering, Nanyang Technological University (NTU), Singapore, Singapore. 7Laboratory for Nanoscale Materials Science, Empa, Dubendorf, Switzerland. 8Department of Physics, University of Basel, Basel, Switzerland. 9Laboratory of Solid State Microstructures and Innovation Center of Advanced Microstructures, Nanjing University, Nanjing, P. R. China. 10School of Physics, Beihang University, Beijing, China. 11Shaanxi Institute of Flexible Electronics (SIFE), Northwestern Polytechnical University (NPU), Xi’an, China. *e-mail: [email protected];
[email protected]; [email protected]
S
olution-processed metal halide perovskites (MHPs) have received significant interest for cost-effective, high-perfor- mance optoelectronic devices1–4. In addition to the great suc- cesses in photovoltaics (PVs), their excellent luminescence and charge transport properties also make them promising for light- emitting diodes (LEDs)5. To achieve high-efficiency perovskite LEDs (PeLEDs), extensive efforts have been carried out to enhance radiative recombination rates by confining the electrons and holes6. These confinement efforts include the use of ultra-thin emissive layers7, the fabrication of nanoscaled polycrystalline features8, the design of low-dimensional or multiple quantum well structures9,10, and the synthesis of perovskite quantum dots11. As a result, the external quantum efficiency (EQE) values of PeLEDs have improved from less than 1% to 14%7–11, a value that is still much lower than that predicted by optical simulations4.In addition to enhancing radiative recombination rates, it is equally important to decrease the non-radiative recombination for improving the device performance. Unfortunately, state-of-the-art solution-processed perovskite semiconductors suffer from severe trap-mediated non-radiative losses12–14, which have been identified as a major efficiency-limiting factor for both PVs and LEDs15,16. The trap states are generally believed to be associated with ionic defects, such as halide vacancies17. Defect passivation through a molecular passivation agent (PA), which can chemically bond with the defects,
is an attractive methodology to tackle this issue18. A few functional groups (for example –NH2, P = O) have been identified to passiv- ate perovskite semiconductors for PV applications19–21. It is found that these PAs show strong structure-dependent performance, even though they share identical functional groups to interact with the perovskite defects18–21. A lack of deep understanding of how the PA chemical structures influence the passivation effects prevents rational design of PAs to minimize the non-radiative recombination losses. These functional groups have also been borrowed to improve the efficiency of LEDs, resulting in limited success so far. For exam- ple, the use of trioctylphosphine oxide (TOPO) treatment in green PeLEDs can result in only moderate EQE enhancement from 12%
to 14%22.
Here, we demonstrate high efficiencies for PeLEDs through the rational design of passivation molecules. We demonstrate that the candidate amino-functionalized PAs which form stronger hydro- gen bonds with organic cations in perovskites are less effective in healing defect sites. Firmly based on our findings, we design new passivation molecules with decreased hydrogen-bonding abil- ity, and hence improve their interaction with defects. In particu- lar, we exploit O atoms within the PAs to polarize the passivating amino groups through the inductive effect, reducing their electron- donating ability and hence relevant hydrogen-bonding ability. This results in enhanced coordination of the PA functional groups with
Rational molecular passivation for high-
performance perovskite light-emitting diodes
Weidong Xu
1,2, Qi Hu
3, Sai Bai
1, Chunxiong Bao
1,4, Yanfeng Miao
2, Zhongcheng Yuan
1, Tetiana Borzda
5, Alex J. Barker
5, Elizaveta Tyukalova
6, Zhangjun Hu
1, Maciej Kawecki
7,8,
Heyong Wang
1, Zhibo Yan
1,9, Xianjie Liu
1, Xiaobo Shi
1, Kajsa Uvdal
1, Mats Fahlman
1, Wenjing Zhang
4, Martial Duchamp
6, Jun-Ming Liu
9, Annamaria Petrozza
5, Jianpu Wang
2, Li-Min Liu
3,10*, Wei Huang
2,11* and Feng Gao
1*
A major efficiency limit for solution-processed perovskite optoelectronic devices, for example light-emitting diodes, is trap- mediated non-radiative losses. Defect passivation using organic molecules has been identified as an attractive approach to tackle this issue. However, implementation of this approach has been hindered by a lack of deep understanding of how the molecular structures influence the effectiveness of passivation. We show that the so far largely ignored hydrogen bonds play a critical role in affecting the passivation. By weakening the hydrogen bonding between the passivating functional moieties and the organic cation featuring in the perovskite, we significantly enhance the interaction with defect sites and minimize non-radi- ative recombination losses. Consequently, we achieve exceptionally high-performance near-infrared perovskite light-emitting diodes with a record external quantum efficiency of 21.6%. In addition, our passivated perovskite light-emitting diodes main- tain a high external quantum efficiency of 20.1% and a wall-plug efficiency of 11.0% at a high current density of 200 mA cm−2, making them more attractive than the most efficient organic and quantum-dot light-emitting diodes at high excitations.
the perovskite defect sites and hence much improved passivation efficiency. As a result, we are able to substantially decrease the trap- mediated non-radiative recombination and boost the electrolumi- nescence (EL) performance of PeLEDs, giving an average EQE of 19.0 ± 0.8% and a record value of 21.6%.
Results and discussion
Perovskite film characterization and device performance. Amino groups have been frequently employed to passivate perovskite semiconductors due to their coordination bonding to unsaturated PbI6-octahedral20. Here, we select two similar amino-functionalized PAs—that is, 2,2′-(ethylenedioxy)diethylamine (EDEA) and hexa- methylenediamine (HMDA) (Fig. 1a), which have identical length of alkyl chains; the difference is that EDEA has two additional O atoms within the chain. We perform first-principles calculations to dem- onstrate that both of them can help to passivate the surface iodide vacancy (VI) through Pb–N coordination bonding, and thus show the potential to improve the EL performance (Supplementary Figs.
1 and 2). The formamidinium lead tri-iodide (FAPbI3) perovskite layers are deposited by spin-casting the precursors with a molar ratio of PbI2: formamidinium iodide (FAI): PA = 1: 2: x (x = 0~0.3), where an FAI excess is used to eliminate the non-perovskite δ-phase (Supplementary Fig. 3)23. We fabricate PeLEDs with the device archi- tecture of indium tin oxide (ITO)/polyethylenimine ethoxylated
(PEIE): modified zinc oxide nanocrystals (ZnO:PEIE)/perovskite/
poly(9,9-dioctyl-fluorene-co-N-(4-butylphenyl)diphenyl-amine) (TFB)/molybdenum oxide (MoO3)/Au, as depicted in the high- angle annular dark-field scanning transmission electron micro- scope (HAADF-STEM) cross-sectional images in Fig. 1b. Both HAADF-STEM and scanning electron microscope (SEM) images (Supplementary Fig. 4) show the formation of separated nano- island features in the perovskite emissive layer. These nano-island features have not resulted in strong leakage currents (Fig. 1e), pos- sibly due to the different TFB thickness on perovskite nano-islands and on ZnO:PEIE, as well as unfavourable charge injection from ZnO to TFB (Supplementary Fig. 5). All the devices show EL peaks at 800 nm/1.55 eV (Fig. 1c) and low turn-on voltages around 1.25 V (see Supplementary Fig. 6 for the characterization set-up), where the measurements were performed in a N2-filled glovebox. In spite of the small difference between the chemical structures of EDEA and HMDA, we notice a significant difference in the EQE values of the devices treated with these two PAs (Fig. 1d). The peak EQE is 10.9% for the HMDA-treated devices and 17.9% for the EDEA- treated ones. The EDEA-treated devices show high efficiencies even if we make significant changes to the emissive layer morphologies (Supplementary Figs. 7 and 8), indicating that the difference in the device performance of HMDA- and EDEA-treated devices results from the different intrinsic passivation effects of these two PAs.
H2N NH2
H2N O
O NH2 HMDA
EDEA
500 550 600 650 700 750 800 850 900 Wavelength (nm)
Normalized EL intensity (a.u.)
Control HMDA EDEA
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 10–6
10–5 10–4 10–3 10–2 10–1 100 101 102 103
V (V)
10–3 10–2 10–1 100 101 102 103
0 200 400 600 800
100 101 102 103 104
M+ = [EDEA]+ (148.12u)
Intensity (counts)
Sputter time (s) Pb+ (206.96u)
[M–H]+ (147.12u)
[M–NH3]+ (131.10u) In2+ (229.79u)
0 50 100 150 200 250 300 350 400 0
2 4 6 8 10 12 14 16 18 20
Control HMDA EDEA
EQE (%)
5 10 15 20 25 30 35 40
Control HMDA EDEA
2θ (°)
# # #
3,500 3,400 3,300 3,200 3,100 3,000
Transmittance (a.u.) PbI2:EDEA
EDEA
α # #
α α
c d e
f g h
Control HMDA EDEA
a
J (mA cm–2)
J (mA cm–2)
R (W sr–1 m–2)
Wavenumber (cm–1)
b
ITO ZnO:PEIE FAPbI3 TFB MoOx/Au
Normalized intensity (a.u.)
Fig. 1 | PeLED architecture, performance and perovskite film characteristics. a, The molecular structures of HMDA and EDEA. b, HAADF-STEM cross- sectional image of an EDEA-treated device (left, scale bar 500 nm) and a zoom-in image (right, scale bar 100 nm). c–e, Representative characteristics for the optimized control, EDEA- and HMDA-treated devices: EL spectra at 2.5 V (c); EQE–current density (J) curves (d); current density–voltage–radiance (J–V–R) characteristics (e). f, ToF-SIMS dual-beam depth profiling conducted on the 0.25 EDEA-treated perovskite film on the ITO/ZnO:PEIE substrate, showing depth profiles of unfragmented EDEA (M+ = EDEA+) and two characteristic fragment ions ([M–NH3]+ and [M–H]+). u, atomic mass unit. g, ATR- FTIR (N–H stretching) for EDEA and PbI2:EDEA mixture. h, XRD patterns for the control, EDEA- (x = 0.25) and HMDA-treated (x = 0.25) films on the ITO/
ZnO:PEIE substrates. α and # denote the diffraction peaks corresponding to α-FAPbI3 and ITO, respectively. a.u., arbitrary units.
To elucidate the different passivation effects between EDEA and HMDA, we first gather information on the molecular interactions of these moieties with perovskites. The first question that arises is whether these molecules are retained in the perovskite films after the annealing process. We performed time-of-flight secondary ion mass spectrometry (ToF-SIMS) depth profiling (Fig. 1f) and X-ray pho- toelectron spectroscopy (XPS) characterizations (Supplementary Fig. 9) on perovskite films treated with EDEA, which has the lower boiling point of 105 °C and thus represents the most criti- cal sample. From ToF-SIMS we observe the depth distribution of EDEA across the perovskite film by monitoring the unfragmented positive molecular ion (C6H16N2O2+; m = 148.1 u), and from XPS we observe changes in line shape of C1s, O1s, N1s core-level spectra in the resulting perovskite films compared to the control ones. Both results confirm the adsorption of EDEA molecules in the perovskite films, thus providing the opportunities for passivation. Passivation through coordination bonding is first evidenced by the strong inter- action between PbI2 and EDEA, leading to a change of the solution colour in their mixture, followed by the formation of a white pre- cipitate (Supplementary Fig. 10). We further performed attenuated total reflectance-Fourier transform infrared (ATR-FTIR) spectros- copy on the EDEA:PbI2 mixture. As shown in Fig. 1g, the stretching absorption (ν) bands from the –NH2 of the mixture shift to a lower wavenumber with respect to those from the pure EDEA, indicating the formation of coordination bonds between Pb2+ and –NH2.
Notably, the molecules used as potential PAs could also be used as templating molecules to synthesize low-dimensional perovskites24,25. Thus, it is worth investigating whether these PAs affect the three-dimensional (3D) crystal structure of FAPbI3. X-ray diffraction (XRD) measurements indicate no additional diffraction peaks other than those from 3D FAPbI3 in the treated perovskite films (Fig. 1h). In addition, no features indicating the presence of Ruddlesden–Popper phases in the ground-state adsorption and photoluminescence (PL) spectra are observed (Supplementary Fig. 11). To further confirm the 3D structure in the treated films, we performed transient absorption (TA) spectroscopy (Supplementary Fig. 12) on the control and EDEA-
treated films to detect any possible charge transfer kinetics from layered perovskite to the bulk 3D phases, which in principle would present a cascade system in terms of energy levels26,27. No additional spectral feature which could be related to low-dimen- sional phases (that is, of short-lived photobleach) is observed in the treated films, confirming that their photophysics is funda- mentally similar to the control samples in this regard.
Passivation mechanism. Assured about the presence of molecu- lar interactions and a lack of structural changes in the thin films, we have investigated the defect physics of the samples. We reveal that the remarkable performance improvement of EDEA-treated PeLEDs origins from the significantly reduced defects in the perovskite emissive layer. Thermal admittance spectroscopy (TAS) was performed to probe the trap density and the energy depth of trap states in the PeLEDs (Fig. 2a). The control and HMDA-treated devices show typical temperature-dependent capacitance versus frequency (C–f) plots28,29. The sub-gap energy deduced from the temperature dependent C–f plots shows a trap energy depth of 0.40 eV and 0.16 eV for the control and HMDA-treated devices, respectively (Supplementary Fig. 13). Figure 2b,c shows the trap density deduced from the room-temperature C–f plots, giving a peak trap density of 7.8 × 1015 cm−3 eV-1 and 5.9 × 1015 cm−3 eV−1 for the control and HMDA-treated devices, respectively. These results indicate a moderate passivation effect of HMDA. In contrast, the EDEA-treated devices show almost temperature-independent C–f plots, indicating a negligible influence from trap states, and hence excellent passivation.
Excellent defect passivation, which results in significantly reduced trap states in EDEA-treated perovskites, eventually also results in much enhanced external photoluminescence quantum yields (PLQYs) across a large range of excitation fluence, showing a peak PLQY of 56% (Fig. 2d). Even at a low fluence of 0.02 mW cm−2, the EDEA-treated films maintain a high PLQY of 40%, consistent with a low defect density. In contrast, the PLQYs of the control and HMDA-treated films show a strong intensity dependence due to trap-mediated non-radiative recombination. Low trap-mediated
1 2 3 4 1 2 3 4 1 2 3
4 1016
102 103
102
101 101
100 8×1015 6×1015
4×1015
10–1 100 101 102 10–2 10–1 100 101 102 103
Frequency (kHz)
Capacitance (nF)Capacitance (nF)Capacitance (nF)
320 K
240 K 240 K
320 K 240 K
Control
HMDA
EDEA 320 K
0.34 0.36 0.38 0.40 0.42 0.44 0.46 Energy (eV)
Control
0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 HMDA
Energy (eV)
Control HMDA EDEA
PLQY (%)
Light intensity (mW cm–2)
0 1,000 2,000 3,000 4,000 5,000
Intensity (counts)
Time (ns)
Control HMDA EDEA
a b c
d e
NT (cm–3 eV–1)
1016 8×1015 6×1015
4×1015 NT (cm–3 eV–1)
Fig. 2 | Passivation effects of EDEA treatment. a, Temperature dependence of C–f plots for control, HMDA- and EDEA-treated devices (in the range 320–
240 K). b,c, Trap density deduced from the room-temperature C–f plots for the control (b) and the HMDA-treated (c) samples. d, Fluence-dependent PLQY.
e, TCSPC probed PL lifetime. The excitation density for the TCSPC measurement is around 1015 cm−3.
recombination in the EDEA-treated samples is also confirmed by the time-correlated single-photon counting (TCSPC) measure- ments (Fig. 2e), which show a prolonged PL lifetime of 1330 ns compared to the control (130 ns) and HMDA-treated films (690 ns).
The remarkable difference in the passivation effects between EDEA and HMDA indicates the significance of O atoms in EDEA for efficient passivation. A straightforward possibility for the O atoms to affect the passivation is that the two pairs of lone pair electrons at the O atom in EDEA can coordinate with Pb2+ and hence passivate the defects. To examine this possibility, we employed ethylene glycol diethyl ether (EGDE) as the PA molecule (Supplementary Fig. 14).
The only difference between EGDE and EDEA is that the former does not contain amino groups. Compared with the control samples, we observe no improvement on the PL properties as well as the device performance in the EGDE-treated ones (Supplementary Fig. 14), indicating no passivation effects of EGDE. Our first-principles calculations confirm that the O atom in EGDE cannot effectively passivate FAPbI3 for the following two reasons: the O atoms lie in the middle of the molecule chain of EGDE, causing a strong lat- tice distortion energy when coordinated with Pb2+; the molecular dynamics shows that even if the O atom initially bonds with Pb2+, the coordination bonds could be quickly broken to form hydrogen bonds between O and FA+, destroying the passivation effects.
Having excluded direct passivation mechanisms through the O atom, we proceed to investigate other reasons for the different pas- sivation effects between EDEA and HMDA. We notice that there are strong intermolecular interactions between FAI and PAs, resulting in the formation of gel-like mixtures (Supplementary Fig. 15a). The ATR-FTIR spectra (Fig. 3a) show obvious broadened absorption from ν(N–H) and ν(C–H) of the EDEA/FAI mixture compared to those in pure FAI and EDEA. In addition, the N–H scissoring vibra- tion (δ(N–H)) absorption bands greatly weaken in the mixture, sug- gesting the restriction of N–H bending caused by intermolecular interactions. A similar phenomenon is also observed in the ATR- FTIR spectra of the HMDA/FAI mixture (Supplementary Fig. 15b).
All these results confirm the formation of hydrogen bonds between PAs and FAI30,31.
Since both hydrogen bonds and passivating coordination bonds result from the lone pair electrons at the N atoms in the amino groups, changes in the hydrogen-bonding ability will influence the passivation effects. Importantly, the hydrogen bonds between the amino groups and FA+ can be affected by the O atom because of the electron-withdrawing inductive effect of O atoms. Figure 3b shows the electron distributions of the EDEA and HMDA. Compared with HMDA, the electrons at the N atoms of EDEA polarize toward the O atoms due to the inductive effects, which hence reduce the
4,000 3,500 3,000 2,500 1,800 1,700 1,600 1,500
Absorbance (a.u.)
ν(C–H)
Wavenumber (cm–1) Wavenumber (cm–1)
EDEA FAI:EDEA=1:1 FAI
EDEA FAI:EDEA=1:1 FAI
1,595
1,636 1,601 1,694
1,690
10 9 8 7 6 5 4
15% EDEA
15% HMDA
Chemical shift (ppm) FAI HMDA
EDEA ν(N–H)
ν(N–H) ν(C–H) ν(N–H)
δ(N–H) νas(C–N)
νas(C–N) δ(N–H)
Ead,P = –1.66 eV Ead,V = –1.21 eV
Ead,V = –1.65 eV Ead,P = –1.42 eV
a b c
d e
f g
b b b
a a
a NH2
HC + I– NH2
a b
Fig. 3 | The influence of hydrogen bonds on passivation effects. a, ATR-FTIR spectra of FAI, EDEA and FAI:EDEA (1:1) mixture. νas(C–N) denotes the C–N antisymmetric stretching vibration from FAI32. b, Calculated electron distribution (valence band maximum) of HMDA and EDEA. c, 1H NMR spectra of FAI in DMSO and FAI with 15% EDEA or HMDA. Peaks a and b are characteristic of H (a) and H (b) as labelled in the molecular structure of FAI in the inset. d–g, Adsorption configurations of HMDA (d,e) and EDEA (f,g) on a perfect FAPbI3 110 slab (d,f) and on a defect-containing structure (e,g). The red, blue, cyan and white spheres represent O, N, C and H atoms, respectively. The hydrogen and coordination bonds are denoted by red dotted lines and blue squares, respectively.
electron-donating ability of the amino groups and the relevant hydrogen-bonding ability33. The stronger hydrogen-bonding ability of HMDA is first evidenced by its higher melting point (43 °C) com- pared to EDEA (liquid at room temperature) due to the enhanced intermolecular interactions. We further confirm that hydrogen- bonding abilities of EDEA and HMDA with FAI are different by using 1H nuclear magnetic resonance (1H NMR) measurements.
Generally, strong hydrogen bonding causes large chemical shifts to low-field (high ppm) because of de-shielding. As shown in Fig. 3c, the proton chemical shift at 8.9 ppm from pure FAI corresponds to the resonance from the active protons at its amino groups20. This peak is broadened and moves to 7.7 ppm and 5.9 ppm with the introduction of 15% HMDA and EDEA, respectively. These results not only confirm the formation of new hydrogen bonds for both PA molecules, but also demonstrate weaker hydrogen-bonding ability of EDEA with respect to HMDA.
To quantify how the hydrogen bonds affect the passivation mechanism, we investigate the adsorption configurations of the PAs on a perfect 110 FAPbI3 slab (Figs. 3d,f) and on a defect-containing surface (iodide vacancy, VI) (Figs. 3e,g) by first-principles calcula- tions. The adsorption energy difference (ΔEad) between these two surfaces is calculated to probe the preferred PA adsorption location, and hence to quantify the competition between the passivation and hydrogen bonding. The adsorption energy on the perfect surface (Ead,P) is determined by van der Waals interactions and the hydrogen bonds (between FA+ and all the electron-rich groups (N, O)). As a result of the inductive effect, we find a decrease of the absolute Ead,P value of EDEA, even considering that its O atoms can provide an additional hydrogen bond with FA+ (Fig. 3f). This result is con- sistent with the 1H NMR measurements in which HMDA shows stronger hydrogen bonding with FA+. For the surface with VI, coor- dination bonds also contribute to the adsorption energy (Ead,V) in
addition to hydrogen bonding and van der Waals interaction. The ΔEad is then determined by Δ =Ead Ead,V−Ead,P. Negative ΔEad
values indicate a preferred interaction with the defect-contained structure through coordination bonds; while positive ΔEad values indicate a preferred interaction with the perfect perovskite surface through hydrogen bonds. The ΔEad for EDEA is −0.23 eV and that for HMDA is 0.45 eV, indicating that it is much easier for EDEA to break down the hydrogen bonds and turn to work with defects. The significant difference of the ΔEad values between EDEA and HMDA explains the remarkable effect of O atoms in manipulating the pas- sivation effects. An additional advantage of the O atom in EDEA is that it might help to improve the stability. The extra hydrogen bonds provided by the O atoms can stabilize the rotating FA+ (Fig. 3f) and hence mitigate the negative effects from thermal vibration and the relevant lattice distortion (as shown in the HMDA adsorbed structure, Fig. 3d). In reality, the PA–perovskite interactions could be much more complex, involving van der Waals and hydrogen bonding between neighbouring surface modifiers; regardless, the experimental data confirm that EDEA is more effective at reducing non-radiative recombination.
The dependence of device performance on the passivation.
Having established the role that the O atom plays in influenc- ing the passivation effects, we proceed to explore new PAs, aim- ing to both further validate our conclusions and improve the device performance. We designed three PAs (Fig. 4a) with dif- ferent strengths of inductive effects, which are expected to result in different hydrogen-bonding abilities and hence different pas- sivation effects. Compared with EDEA, the inductive effect can be increased by introducing one additional O atom (as in 2,2′-[oxybis(ethylenoxy)]diethylamine (ODEA)), and reduced by increasing the length of the alkyl chain between the N and
H2N O
O O
NH2
H2N O O
O NH2
H2N O O NH2
ODEA
TTDDA
DDDA a
b
a b
c
a b
c
0 50 100 150 200 250 300 350 400 0
2 4 6 108 12 14 16 18 20 22 24
J (mA cm–2)
EQE (%)
ODEA champion device 0 2 4 6 8 10 12 14 16 18 20 22
Wall-plug efficiency (%)
4 6 8 10 12 14 16 18 20 22 0
5 10 15 20 25
Counts
Peak EQE (%) Control
ODEA
–0.6 –0.4 –0.2 0.0 0.2 0.4 0.6 10
12 14 16 18 20
HMDA DDDA
TTDDA EDEA
Peak EQE (%)
Ead (eV) ODEA
0 2 4 6 8 10 12 14 16 18 20 22 0
5 10 15 200 2 4 6 8
ODEA
EQE (%)
Time (h)
Control
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 10–4
10–3 10–2 10–1 100 101 102 103
V (V)
10–3 10–2 10–1 100 101 102 103 ODEA champion device
J (mA cm–2) R (W sr–1m–2)
T50
T50
a b c
d e f
Fig. 4 | The dependence of EL performance on passivation effects determined by the hydrogen bonds. a, Molecular structures of selected PAs (ODEA, TTDDA, DDDA). The letters a, b and c highlight the different lengths of the carbon chain between the N and O atoms. b, Dependence of the average peak EQE values from various PA-treated PeLEDs on ΔEad. Each value is an average of 60 devices. The error bars represent the standard deviation. c, Histograms of the peak EQEs for control and ODEA-treated devices. d–f, Device characteristics for the best-performing ODEA-treated device: J–V–R characteristics (d); EQE and wall-plug efficiency as a function of the current density (e); steady-state EQE for the control and ODEA-treated devices at 25 mA cm−2 (f). We select the maximum EQE as the initial value for the calculation of T50. The device shown in d–f was fabricated using a different batch of ZnO nanocrystals, and hence is not included in the statistics of b,c.
O atoms (as in 4,9-dioxa-1,12-dodecanediamine (DDDA) and 4,7,10-trioxa-1,13-tridecanediamine (TTDDA))33. Among all the PAs with O atoms, the inductive effect in DDDA is the least effec- tive since its N and O atoms are almost isolated from each other, resulting in the strongest hydrogen-bonding ability of the amino groups. The morphological and optical characterizations for each PA-passivated perovskite film are shown in Supplementary Figs. 16 and 17. The adsorption configurations and Ead values are depicted in Supplementary Fig. 18 and summarized in Supplementary Table 1. Figure 4b shows the average peak EQE values for all the passivated systems as a function of ΔEad (see representative device characteristics in Supplementary Figs. 19 and 20). It clearly shows that the EL performance is strongly dependent on the ΔEad and hence the hydrogen-bonding ability of amino groups. As expected, ODEA, which shows a ΔEad value of −0.42 eV, delivers the high- est average peak EQE of 19.0 ± 0.8% (Fig. 4b,c and Supplementary Fig. 21). The excellent passivation effects of the ODEA are also confirmed by the fluence-dependent PLQY measurements and TAS measurements (Supplementary Fig. 22).
We show the characteristics for the best-performing ODEA- treated device, which gives a peak EQE up to 21.6% (Fig. 4e), approaching the best organic and quantum dot LEDs34,35. The radi- ance rises rapidly after the device turns on, reaching a high radiance of 308 W sr−1 m−2 at 3.3 V (Fig. 4d). The high EQE and low driving voltage result in an exceptionally high peak wall-plug efficiency up to 15.8% (Fig. 4e). High efficiencies at high current densities have been challenging in other low-temperature processed LED techniques (for example, organic LEDs) due to low charge carrier mobilities and strong exciton-induced quenching effects. Our device exhibits a low efficiency roll-off, maintaining a high EQE of 20.1 % and a wall- plug efficiency of 11.0% at a high current density of 200 mA cm−2, which makes them much more efficient than OLEDs and QLEDs at high excitations34,35. We also tested the operation lifetime (T50, time to half of the initial radiance) of these devices in the glove- box without encapsulation. The ODEA-treated devices are among the most stable PeLEDs to date8–11,22,36,37, showing a long lifetime of 20 h at 25 mA cm−2 compared with the control devices (T50= 1.5 h at 25 mA cm−2) (Fig. 4f). The improved lifetime may result from the reduced Joule heating due to the high efficiency, or the sup- pression of ion migration due to the low defect density12,38. We also notice a rapid degradation at high current densities (T50 = 18 min at 200 mA cm−2) (Supplementary Fig. 23). Future research on the deg- radation mechanisms, especially at high current densities, will be of key importance to practical applications of PeLEDs.
Conclusions
In summary, we have demonstrated high-efficiency near-infrared PeLEDs with a peak EQE of 21.6%, which represents the most effi- cient PeLEDs to date. Our devices also show low-efficiency roll- off, maintaining a high EQE of 20.1% and a wall-plug efficiency of 11.0% at a high current density of 200 mA cm−2. Our results indicate a unique opportunity for PeLEDs to achieve solution-processed large-scale LEDs with high efficiencies at high brightness. The high efficiencies stem from our deep understanding of the passiv- ation mechanisms of perovskites by organic molecules. We reveal the critical role of hydrogen bonds in influencing the passivation effects. By weakening the hydrogen bonding between the passiv- ating functional groups and the organic cations of perovskites, we significantly reduce the non-radiative recombination. Our findings provide a broad avenue to explore the potential of molecular pas- sivation for improving various perovskite applications where non- radiative losses should be minimized.
online content
Any methods, additional references, Nature Research reporting summaries, source data, statements of data availability and asso-
ciated accession codes are available at https://doi.org/10.1038/
s41566-019-0390-x.
Received: 22 September 2018; Accepted: 11 February 2019;
Published: xx xx xxxx References
1. Tan, Z.-K. et al. Bright light-emitting diodes based on organometal halide perovskite. Nat. Nanotechnol. 9, 687–692 (2014).
2. Akkerman, Q. A. et al. Strongly emissive perovskite nanocrystal inks for high-voltage solar cells. Nat. Energy 2, 16194 (2017).
3. Dou, L. et al. Solution-processed hybrid perovskite photodetectors with high detectivity. Nat. Commun. 5, 5404 (2014).
4. Shi, X. et al. Optical energy losses in organic–inorganic hybrid perovskite light-emitting diodes. Adv. Opt. Mater. 6, 1800667 (2018).
5. Stranks, S. D. & Snaith, H. J. Metal-halide perovskites for photovoltaic and light-emitting devices. Nat. Nanotechnol. 10, 391–402 (2015).
6. Xing, G. et al. Transcending the slow bimolecular recombination in lead-halide perovskites for electroluminescence. Nat. Commun. 8, 14558 (2017).
7. Li, G. et al. Efficient light-emitting diodes based on nanocrystalline perovskite in a dielectric polymer matrix. Nano Lett. 15, 2640–2644 (2015).
8. Xiao, Z. et al. Efficient perovskite light-emitting diodes featuring nanometre- sized crystallites. Nat. Photon. 11, 108–115 (2017).
9. Wang, N. et al. Perovskite light-emitting diodes based on solution-processed self-organized multiple quantum wells. Nat. Photon. 10, 699–704 (2016).
10. Zou, W. et al. Minimising efficiency roll-off in high-brightness perovskite light-emitting diodes. Nat. Commun. 9, 608 (2018).
11. Li, G. et al. Surface ligand engineering for near-unity quantum yield inorganic halide perovskite QDs and high-performance QLEDs. Chem. Mater.
30, 6099–6107 (2018).
12. Abdi-Jalebi, M. et al. Maximizing and stabilizing luminescence from halide perovskites with potassium passivation. Nature 555, 497–501 (2018).
13. Stranks, S. D. Nonradiative losses in metal halide perovskites. ACS Energy Lett. 2, 1515–1525 (2017).
14. Tress, W. Perovskite solar cells on the way to their radiative efficiency limit—insights into a success story of high open-circuit voltage and low recombination. Adv. Energy Mater. 7, 1602358 (2017).
15. Correa-Baena, J.-P. et al. Promises and challenges of perovskite solar cells.
Science 358, 739–744 (2017).
16. Braly, I. L. et al. Hybrid perovskite films approaching the radiative limit with over 90% photoluminescence quantum efficiency. Nat. Photon. 12, 355–361 (2018).
17. Agiorgousis, M. L., Sun, Y. Y., Zeng, H. & Zhang, S. Strong covalency- induced recombination centers in perovskite solar cell material CH3NH3PbI3. J. Am. Chem. Soc. 136, 14570–14575 (2014).
18. Noel, N. K. et al. Enhanced photoluminescence and solar cell performance via Lewis base passivation of organic–inorganic lead halide perovskites. ACS Nano 8, 9815–9821 (2014).
19. Dequilettes, D. W. et al. Photoluminescence lifetimes exceeding 8 μs and quantum yields exceeding 30% in hybrid perovskite thin films by ligand passivation. ACS Energy Lett. 1, 438–444 (2016).
20. Wang, F. et al. Phenylalkylamine passivation of organolead halide perovskites enabling high-efficiency and air-stable photovoltaic cells. Adv. Mater. 28, 9986–9992 (2016).
21. Zheng, X. et al. Defect passivation in hybrid perovskite solar cells using quaternary ammonium halide anions and cations. Nat. Energy 2, 17102 (2017).
22. Yang, X. et al. Efficient green light-emitting diodes based on quasi-two- dimensional composition and phase engineered perovskite with surface passivation. Nat. Commun. 9, 570 (2018).
23. Ma, F. et al. Stable α/δ phase junction of formamidinium lead iodide perovskites for enhanced near-infrared emission. Chem. Sci. 8, 800–805 (2016).
24. Cortecchia, D. et al. Broadband emission in two-dimensional hybrid perovskites: the role of structural deformation. J. Am. Chem. Soc. 139, 39–42 (2017).
25. Chen, Z. et al. High-performance color-tunable perovskite light emitting devices through structural modulation from bulk to layered film. Adv. Mater.
29, 1603157 (2017).
26. Neutzner, S., Srimath Kandada, A. R., Lanzani, G. & Petrozza, A. A dual-phase architecture for efficient amplified spontaneous emission in lead iodide perovskites. J. Mater. Chem. C 4, 4630–4633 (2016).
27. Yuan, M. et al. Perovskite energy funnels for efficient light-emitting diodes.
Nat. Nanotechnol. 11, 872–877 (2016).
28. Duan, H. S. et al. The identification and characterization of defect states in hybrid organic–inorganic perovskite photovoltaics. Phys. Chem. Chem. Phys.
17, 112–116 (2015).
29. Bi, C. et al. Non-wetting surface-driven high-aspect-ratio crystalline grain growth for efficient hybrid perovskite solar cells. Nat. Commun. 6, 7747 (2015).
30. Dai, H. et al. A temperature-responsive copolymer hydrogel in controlled drug delivery. Macromolecules 39, 6584–6589 (2006).
31. Chen, H. et al. A solvent-and vacuum-free route to large-area perovskite films for efficient solar modules. Nature 550, 92–95 (2017).
32. Taylor, V. C. A. et al. Investigating the role of the organic cation in formamidinium lead iodide perovskite using ultrafast spectroscopy. J. Phys.
Chem. Lett. 9, 895–901 (2018).
33. Gero, A. Inductive effect and hydrogen bonding as factors in the base strength of polymethylenediamines. J. Am. Chem. Soc. 76, 5159–5160 (1954).
34. Dai, X. et al. Solution-processed, high-performance light-emitting diodes based on quantum dots. Nature 515, 96–99 (2014).
35. Ly, K. T. et al. Near-infrared organic light-emitting diodes with very high external quantum efficiency and radiance. Nat. Photon. 11, 63–68 (2016).
36. Cao, Y. et al. Perovskite light-emitting diodes based on spontaneously formed submicrometre-scale structures. Nature 562, 249–253 (2018).
37. Lin, K. et al. Perovskite light-emitting diodes with external quantum efficiency exceeding 20 per cent. Nature 562, 245–248 (2018).
38. Xiao, X. et al. Suppressed ion migration along the in-plane. ACS Energy Lett.
3, 684–688 (2018).
Acknowledgements
We thank O. Inganäs, T.C. Sum, S.S. Lim, J. Zhang, W. Tress, W. Chen, Y. Puttisong, Y.T. Gong, C.Y. Kuang and C. Deibel for useful discussions. This work is supported by the ERC Starting Grant (717026), the National Basic Research Program of China (973 Program, grant number 2015CB932200), the National Natural Science Foundation of China (61704077, 51572016, 51721001, 61634001, 61725502, 91733302 and U1530401), the Joint Research Program between China and the European Union (2016YFE0112000), the Natural Science Foundation of Jiangsu Province (BK20171007), the National Key Research and Development Program of China (grant number 2016YFB0700700), the European Commission Marie Skłodowska-Curie Actions (691210), the Swiss National Science Foundation (CR23I2-162828), Nanyang Technological University start-up grant M4081924, and the Swedish Government Strategic Research Area in Materials Science on Functional Materials at Linköping University (Faculty Grant SFO-Mat-LiU no. 2009-00971). The TEM measurements were performed at the Facility for Analysis, Characterization, Testing and Simulation (FACTS) in Nanyang Technological University, Singapore. A.P. and T.B. acknowledge financial support from the ERC Consolidator grant SOPHY (grant agreement number 771528). A.P. and A.J.B. acknowledge the
project PERSEO-‘Perovskite-based solar cells: towards high efficiency and long-term stability’ (Bando PRIN 2015-Italian Ministry of University and Scientific Research (MIUR) Decreto Direttoriale 4 novembre 2015 n. 2488, project number 20155LECAJ) for funding. W.X. is a Wenner-Gren Postdoc Fellow; F.G. is a Wallenberg Academy Fellow.
Author contributions
F.G. and W.X. conceived the idea and designed the experiments; W.X. performed the experiments and analysed the data under the supervision of F.G. and W.H.;
Q.H. performed first-principles calculations on the molecular passivation under the supervision of L.-M.L.; Y.M., Z.C.Y., H.W., X.S. and Z.B.Y. contributed to device fabrication and measurements; Y.M. performed fluence-dependent PLQY and TCSPC measurements and analysed the data under the supervision of J.W. and W.H.; Y.M.
and J.W. cross-checked the device performance at Nanjing Tech University; S.B. and Z.C.Y. synthesized and modified the ZnO nanocrystals and contributed to the device development; C.B. performed the TAS measurements and analysed the data; Z.H.
performed the FTIR measurements and analysed the data; X.L. performed XPS tests and analysed the data; E.T. prepared the STEM specimen using FIB and performed the STEM imaging under the supervision of M.D.; T.B. and A.J.B. performed the transient absorption measurements and analysed the data under the supervision of A.P.; M.K.
performed the ToF-SIMS measurements and analysed the data; J.-M.L., M.F., K.U. and W.Z. contributed to the data analysis; W.X. and F.G. wrote the manuscript; S.B., J.W.
and A.P. provided revisions to the manuscript; F.G. supervised the project. All authors discussed the results and commented on the manuscript.
Competing interests
F.G. and W.X. have filed a patent application related to this work (application no. SE1950272-3).
Additional information
Supplementary information is available for this paper at https://doi.org/10.1038/
s41566-019-0390-x.
Reprints and permissions information is available at www.nature.com/reprints.
Correspondence and requests for materials should be addressed to L.-M.L., W.H.
or F.G.
Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
© The Author(s), under exclusive licence to Springer Nature Limited 2019
Methods
Materials. The passivation agents (PAs), including hexamethylenediamine (HMDA), 2,2′-(ethylenedioxy)diethylamine (EDEA), 4,9-dioxa-1,12- dodecanediamine (DDDA), 2,2′-[oxybis(ethylenoxy)]diethylamine (ODEA), 4,7,10-trioxa-1,13-tridecanediamine (TTDDA), ethylene glycol diethyl ether (EGDE) were purchased from Sigma-Aldrich. Formamidinium iodide (FAI) was purchased from Dyesol. PbI2 (beads, 99.999%) was purchased from Alfa Aesar. Poly(9,9-dioctyl-fluorene-co-N-(4-butylphenyl)diphenylamine) (TFB) was purchased from Ossila. Other materials for device fabrication were all purchased from Sigma-Aldrich.
Preparation of the perovskite solution. Perovskite precursors (FAI: PbI2: PA molar ratio of 2: 1: x, x = 0–0.3) were prepared with dimethylformamide (DMF) as the solvent. A 10 mg ml−1 PA solution was prepared at first, and then was diluted according to the required molar ratio to Pb2+. The optimal concentration for PbI2
was 0.13 M. The solution precursors were stirred at 50˚C for 12 h before spin- coating. Colloidal ZnO nanocrystal was synthesized by a solution–precipitation process, the details of which can be found in the literature39.
PeLED fabrication. The indium tin oxide (ITO) glass substrates were sequentially cleaned by detergent and TL-1 (a mixture of water, ammonia (25%) and hydrogen peroxide (28%) (5:1:1 by volume)). The clean substrates were then treated by ultraviolet-ozone for 10 min. The ZnO nanocrystal solutions were spin-cast onto the substrates at 4,000 r.p.m. for 30 s in air. Then the substrates were moved into a N2- filled glovebox. Next, a layer of polyethylenimine ethoxylated (PEIE) was deposited at 5,000 r.p.m. (0.05 wt%, in IPA), followed by annealing at 100 °C for 10 min. After cooling down to room temperature, the perovskite films were deposited from the precursors with various PA contents and Pb2+ concentrations at a spin-coating speed of 3,000 r.p.m., followed by annealing at 100 °C for 10 min. For the control perovskite films prepared by anti-solvent treatment, the spin-casting rate is 5,000 r.p.m. In addition, 150 μl chlorobenzene (CB) was dropped after 5 s spinning. The TFB layer was deposited from its CB solution (12 mg ml−1) at 3,000 r.p.m. Finally, the MoOx/Au electrode was deposited by a thermal evaporation system through a shadow mask under a base pressure of ∼1 × 10−7 torr. The device area was 7.25 mm−2, as defined by the overlapping area of the ITO films and top electrodes.
PeLED characterization. All PeLED device characterizations were carried out at room temperature in a nitrogen-filled glovebox. A Keithley 2400 sourcemeter and a fibre integration sphere (FOIS-1) coupled with a QE Pro spectrometer (Ocean Optics) were used for the measurements. The PeLED devices are tested on top of the integration sphere and only forward light emission can be collected, consistent with the standard OLED characterization method. The absolute radiance was calibrated by a standard Vis–NIR light source (HL-3P-INT-CAL plus, Ocean Optics). The devices were swept from zero bias to forward bias. The time evolution of the EQE was measured using the same testing system. To verify the accuracy of our characterization set-up, we crosschecked our results at Nanjing Tech University. The performances obtained in the two laboratories are in good agreement (Supplementary Table 2).
First-principles calculations. We used the CP2K/Quickstep package to carry out the first-principles calculations. The exchange correlation energy was described with the generalized gradient approximation (GGA) of Perdew–Burke–Ernzerhof (PBE)40. The norm-conserving Goedecker–Teter–Hutter (GTH) pseudopotentials were used to describe the core electrons41. Gaussian functions with molecularly optimized double-zeta polarized basis sets (m-DZVP) were adopted for expanding the wavefunction of Pb 6s26p2, I 5s25p5, H 1s2, C 2s22p2, O 2s22p4 and N 2s22p3 electrons42. The auxiliary basis set of plane waves was set as a 500 Ry cut-off energy.
We modelled the FAPbI3 (110) surface as a two-layer slab with a 2 × 2 surface supercell (25.74 × 25.13 Å). We used a vacuum of 15 Å to separate images along the surface normal direction.
The adsorption energies were calculated with the following equation:
= −
Ead Eadsorbed Edecoupled (1)
Here, Eadsorbed is the energy of the adsorption configuration, while Edecoupled is the energy of the decoupled system, including the clean system and the single molecule.
Perovskite film characterizations. X-ray diffraction patterns were obtained from an X-ray diffractometer (Panalytical X’Pert Pro) with an X-ray tube (Cu Kα, λ = 1.5406 Å). Steady-state PL spectra of the perovskite films were recorded by means of a fluorescent spectrophotometer (F-4600, HITACHI) with a 200 W Xe lamp as an excitation source. Absorption spectra were measured with a PerkinElmer model Lambda 900. The XRD patterns, UV and PL spectra for DDDA-, ODEA- and TTDDA-treated perovskite films are shown in Supplementary Figs. 16 and 17. Similarly, we find no evidence for the formation of low-dimensional perovskite within the resulting films from XRD, UV and PL measurements. The perovskite film morphology was characterized by a scanning electron microscope (SEM, LEO 1550 Gemini).
XPS tests were carried out using a Scienta ESCA 200 spectrometer in ultrahigh vacuum (~1 × 10−10 mbar) with a monochromatic Al (Kɑ) X-ray source providing photons with 1,486.6 eV. The XPS experimental condition was set so that the full-width at half-maximum of the clean Au 4f7/2 line (at the binding energy of 84.00 eV) was 0.65 eV. All spectra were measured at a photoelectron take-off angle of 0° (normal emission).
ToF-SIMS tests were performed on a ToF-SIMS.5 instrument from IONTOF, operated in the spectral mode using a 25 keV Bi3+ primary ion beam with an ion current of 0.78 pA. A mass-resolving power of approximately 6,000 m/∆m was reached. For depth profiling a 1 keV Cs+ sputter beam with a current of 39.81 nA was used to remove the material layer-by-layer in interlaced mode from a raster area of 500 × 500 µm. This raster area was chosen to ensure a flat crater bottom over an area of 100 × 100 µm used for the mass spectrometry. The position of the ITO substrate interface in the sputter depth profile was defined by the half-maximum of the In2+ secondary ion count rate.
1H nuclear magnetic resonance (NMR). The 1H NMR spectra were recorded on a Bruker Ultra Shield Plus 400 MHz NMR system. All the samples were prepared by dissolving 5 mg FAI in 0.4 ml dimethyl sulfoxide-d6 (DMSO-d6). For the blend samples, 15% EDEA or HMDA (molar ratio compared to FAI) was added.
Attenuated total reflectance-Fourier transform infrared (ATR-FTIR). The ATR- FT-IR spectra were recorded from a PIKE MIRacle ATR accessory with a diamond prism in a Vertex 70 Spectrometer (Bruker) using a DLaTGS detector at room temperature. The measuring system was continuously kept in N2 atmosphere.
The spectra were acquired at 2 cm−1 resolution and 30 scans between 4,000 and 800 cm−1. The presented spectra were baseline-corrected by subtracting a linear baseline over the spectral ranges.
Aberration-corrected scanning transmission electron microscope (STEM). An FEI dual-beam FIB Helios workstation equipped with an in-situ micromanipulator and Pt gas injection system was used to prepare thin samples for STEM imaging.
The final milling was performed at 3 kV. STEM investigations were conducted using JEOL ARM200F TEM equipped with a spherical aberration corrector at the condenser plane. A semi-convergence angle of 32 mrad was used. HAADF and annular bright-field (ABF) STEM were recorded with semi-angles in the range 68–280 mrad and 7–18 mrad, respectively.
Fluence-dependent PLQY and time-correlated single-photon counting (TCSPC) measurements. The fluence-dependent PLQY was measured by a typical three-step technique with a combination of 445 nm continuous wave (CW) laser, spectrometer and an integrating sphere43. The TCSPC measurements were performed on an Edinburgh Instruments spectrometer (FLS980) with a 638 nm pulsed laser (less than 100 ps, 0.1 MHz). The total instrument response function (IRF) was less than 130 ps and the temporal resolution was less than 20 ps. All the perovskite films were deposited on ITO/ZnO:PEIE substrates under identical spin-casting conditions for the optimized devices, and encapsulated by ultraviolet- curable resin and glass slides.
Transient absorption (TA). The perovskite film samples were mounted in a chamber under dynamic vacuum ( < 10−5 mbar). TA spectroscopy was conducted in transmission geometry. An amplified Ti:sapphire laser (Quantronix Integra-C) generated ~130 fs pulses centred at 800 nm, at a repetition rate of 1 kHz. A broadband white light probe was generated by focusing the pulses into a thin CaF2
plate, and pump light at 400 nm was obtained via second-harmonic generation in a beta barium borate (BBO) crystal. After interaction with the sample, a grating spectrometer was used to disperse the probe light on to a fast charge-coupled device (CCD) array, enabling broadband shot-to-shot detection.
Trap density measurements by thermal admittance spectroscopy (TAS). For the device capacitance measurement, we used 0.4 M Pb2+ for all the cases to increase the signals. A sinusoidal voltage with a peak-to-peak value of 30 mV generated from a Tektronix AFG 3000 function generator was applied to the device. The current signal of the devices was amplified with a SR570 low-noise-current preamplifier (Stanford Research Systems) and then analysed using a SR830 lock- in amplifier (Stanford Research Systems), where the amplitude and phase of the current can be measured. Based on the amplitude and phase of the current signal, the capacitance of the device was calculated using the parallel equivalent circuit model. The capacitance spectra of the device were measured by scanning the frequency of the sinusoidal voltage from 0.01 to 100 kHz in logarithmic steps.
The temperature of the device was controlled using a DE202AE closed cycle cryocooler (Advanced Research Systems). The capacitance–voltage curve was obtained by measuring the capacitance as the applied d.c. bias voltage was scanned from −0.5 to 1.0 V. Based on the capacitance spectra measured at different temperatures, the trap density (NT) distribution in energy (Eω) was calculated with the following relations:
ω ω
ω = −
N E V
qW C
( ) d kT
d (2)
T bi
ν
= πω
Eω kTln 2 0T2 (3)
where Vbi is the built-in potential and W is the depletion width (Vbi and W are derived from capacitance–voltage measurements), C is the capacitance measured at an angular frequency ω and temperature T, k is the Boltzmann constant, and ν0 is the attempt-to-escape frequency, which can be obtained by fitting the relation of characteristic frequency with different T based on equation (3).
Data availability
The data that support the plots within this paper and other findings of this study are available from the corresponding authors upon reasonable request.
References
39. Bai, S. et al. High-performance planar heterojunction perovskite solar cells:
preserving long charge carrier diffusion lengths and interfacial engineering.
Nano Res. 7, 1749–1758 (2014).
40. Perdew, J. P., Burke, K. & Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 77, 3865–3868 (1996).
41. Goedecker, S. & Teter, M. Separable dual-space Gaussian pseudopotentials.
Phys. Rev. B 54, 1703–1710 (1996).
42. VandeVondele, J. & Hutter, J. Gaussian basis sets for accurate calculations on molecular systems in gas and condensed phases. J. Chem. Phys. 127, 114105 (2007).
43. De Mello, J. C., Wittmann, H. F. & Friend, R. H. An improved experimental determination of external photoluminescence quantum efficiency. Adv. Mater.
9, 230–232 (1997).
