The air-stability of vapour-phase-deposited methylammonium lead triiodide (CH3NH3PbI3) perovskite thin films has been studied using X-ray diffraction. It is found that the perovskite structure without organic coating decomposes completely within a short period of time (∼two days) upon exposure to ambient environment. The degradation of the perovskite structure is drastically reduced when the perovskite films are capped with thin N,N′-Di(1-naphthyl)-N,N′-diphenyl-(1,1′-biphenyl)-4,4′-diamine (NPB) films. We discovered that the amount of lead iodide (PbI2), a product of the degradation, grows as a function of time in a sigmoidal manner. Further mathematical modeling analysis shows that the perovskite degradation follows the Avrami equation, a kinetics theory developed for quantifying phase transformations in solid-state materials.

The high demand for energy and increasing concerns regarding sustainability have driven the development of light-harvesting technologies. While there has been research over the years on a variety of solar energy technologies, silicon photovoltaics (PV) currently dominate the solar electricity market. Si PV can only be replaced with alternative materials that show significant potential in greater energy conversion efficiency or lower processing cost. Recently, hybrid organic-inorganic halide perovskite compounds (CH3NH3PbX3, X = I, Br, Cl) have been identified as promising candidates for thin-film solar cells with prospects in both domains.1 Extremely rapid progress has been made since 2013 with confirmed energy conversion efficiency of 20.1% by the end of 2014,2 compared to only a maximum of 25.6% seen thus far with crystalline silicon.2 

While a lot of focus so far has been on increasing device efficiency, there are other factors that affect the viability of perovskites in becoming practical solar materials. Currently, a major obstacle is their stability when exposed to air, which triggers chemical decomposition and consequently, loss of their function for solar energy conversion. Previous research has found it essential that perovskite film fabrication be performed in controlled atmospheric conditions including a low humidity environment.3,4 Burschka et al. claimed as low as <1% in relative humidity.4 To imagine the magnitude of impact, Yang et al. had extrapolated that, for solution-processed methylammonium lead triiodide (CH3NH3PbI3) films fabricated under 20% relative humidity condition, the degradation rate was low enough so that the efficiency half life (τ1/2) was about 10 000 hours, compared to 1 000 hours with 50% relative humidity.5 

Understanding the degradation behaviour of these perovskite films is thus of interest in order to be able to effectively impede this process. Thus far, degradation studies have been limited to exploring the underlying chemical pathways. Niu et al. made the initial claim that the perovskite decomposition pathway began with the deprotonation of the methylammonium cation by water to form methylamine, hydrated hydrogen iodide (HI), and lead iodide (PbI2).3 Yang et al. suggest an alternative pathway wherein hydration of the perovskite first occurs, forming the hydrate (CH3NH3)4PbI6⋅2H25 followed by subsequent decomposition into PbI2. Other attempts have aimed at increasing the material’s hydrophobicity thereby increasing moisture resistance by altering composition of the perovskite film itself.6,7

Also notable is the previous work involving perovskite encapsulation films. If CH3NH3PbI3 films can tolerate <20% relative humidity, lower quality - thus less expensive - encapsulation strategies would be sufficient to stabilize the films to prevent degradation.5 Niu et al. have shown the degradation prevention using an Al2O3 overlayer.3 Researchers Zheng et al. have experimented with oligothiophene, a hole transport layer (HTL) material which would be useful in the solar device structure with perovskite, but as a dual function, also demonstrated improved stability of the perovskite resulting from the moisture resistance of the oligothiophene.8 Yang et al. also experimented with a number of HTL films and also saw the impact of doping these layers with lithium bis(trifluoromethanesulfonyl)imide (Li-TFSI) salt.4,5

Our research attempts to understand degradation kinetics using X-ray diffractometry (XRD). Vapour-deposited samples were prepared as it has been demonstrated by Liu et al. that there is advantage over solution-processed films for increased layer uniformity.9 For our investigation, we use N, N′-Di(1-naphthyl)-N, N′-diphenyl-(1,1′-biphenyl)-4,4′-diamine (NPB), a common HTL in organic photovoltaics and organic light-emitting diodes,10,11 as an overlayer. NPB is known to have excellent chemical stability and film forming capability.10 The crystal structure of NPB is found in a twisted conformation which results from the steric effect of the hydrogen atoms in the ortho-positions of the phenyl and naphthyl rings of the chemical.12 This conformation can thus protect the nitrogen atoms from chemical attack, contributing to the material having an excellent lifetime.12 Different thicknesses of overlayer were also studied to observe its impact on delaying perovskite degradation. The effectiveness of a material as an encapsulation method depends to a great extent on its ability to form a uniform, thick, dense and pinhole-free layer. Studies of NPB on indium tin oxide (ITO)13,14 found that the NPB forms a continuous coating. Furthermore, as can be found in the Supporting Information,15 SEM micrographs reveal continuous coverage of the NPB on perovskite.

The perovskite thin film layers were prepared by thermal co-evaporation in a Kurt J. Lesker system with a base pressure of 10−8 Torr. The evaporation temperature of PbI2 (Alfa Aesar) was 302°C, and CH3NH3I (abbreviated as MAI, from Wuhan Arike Technology) was at 125°C. The overlayer was also deposited via thermal evaporation.

Deposition rate was monitored using a quartz crystal microbalance sensor (INFICON). Each film deposition resulted in 4 rectangular 5mm x 7mm samples. The perovskite films’ 60 nm thickness was chosen due to limitations of the vapour deposition system. Samples without any encapsulation (0nm), and with 10, 20, and 30 nm NPB overlayers were deposited. The 30 nm NPB is a typical thickness used for perovskite-organic heterojunction solar cells.

Once prepared, thin film samples were stored under ambient lab conditions; temperature =22±2°C and humidity=35±9%. XRD analysis was performed using a Rigaku Mini Flex 600 X-Ray Diffractometer. The operating parameters are: Cu Kα line, 40 kV anode bias voltage and 15 mA filament current.

All XRD spectra collected over time for each of the different samples were normalized to the (110) perovskite peak, as illustrated in Figure 1. The XRD data shows the presence of perovskite and PbI2 phases in the as-grown films, consistent with literature spectra reported by Jeon et al.16 The following analysis was focused on the relative changes in peak intensity of (001) PbI2 peak and (110) perovskite peak within the 10° to 20° 2θ range, as peaks outside this range had fairly low intensities.

FIG. 1.

Normalized XRD spectra acquired from perovskite samples with or without NPB overlayer, as labelled.

FIG. 1.

Normalized XRD spectra acquired from perovskite samples with or without NPB overlayer, as labelled.

Close modal

Figure 1(a) clearly shows that without encapsulation, the perovskite peak disappeared after 1 day, signifying rapid degradation of the film. The normalized XRD curves of the samples with varied encapsulation thicknesses (10nm, 20nm and 30nm) over time, are shown in Figures 1(b) to 1(d), respectively. As shown in Figures 1(b) and 1(c), the perovskite diffraction peak for samples with 10nm and 20nm NPB encapsulation both disappeared after 99 days. Perovskite crystal was still detected in the sample with 30nm NPB encapsulation even after 153 days (approximately 5 months) as shown in Figure 1(d).

The perovskite film degradation is characterized using an intensity ratio, the fraction of the perovskite XRD intensity over the sum of perovskite and PbI2 intensities. Figure 2 shows the normalized intensity ratios as a function of air exposure time. As shown in Figure 2, the 0nm NPB curve decreases sharply after 22 hours (about 1 day). By 46 hours (about 2 days), the perovskite peak is no longer visible. All the samples with NPB encapsulation degrade at a much slower rate compared to the 0nm NPB curve. The 30nm NPB sample yields a less steep curve compared to the 20nm and 10nm NPB sample curves. Evidently, it is seen that all NPB-encapsulated samples lasted much longer under experimental conditions than perovskite film without encapsulation. The sample with 30nm NPB encapsulation (the thickest encapsulation used in this investigation) shows the best stability. Even after 3527 hours (about 153 days, or 5 months), the XRD intensity ratio remains above 15%.

FIG. 2.

Normalized perovskite XRD intensity ratios vs. air exposure time for samples with different encapsulation thicknesses. Root-mean-square error is magnified 10 times.

FIG. 2.

Normalized perovskite XRD intensity ratios vs. air exposure time for samples with different encapsulation thicknesses. Root-mean-square error is magnified 10 times.

Close modal

The exposure time corresponding to an intensity ratio that was 50% of the initial intensity ratio was used to quantify the effect of overlayer thickness on the degradation of the samples. The exposure time at 50% of the initial integrated intensity was graphically determined via linear interpolation, based on the experimental data given in Figure 2. The exposure time at 50% initial intensity was plotted as a function of overlayer thickness. This plot is given in Figure 3 and demonstrates a strong linear relationship between the half lifetime of the perovskite and the thickness of the NPB overlayer.

FIG. 3.

The 50% perovskite XRD intensity ratio as a function of NPB overlayer thickness.

FIG. 3.

The 50% perovskite XRD intensity ratio as a function of NPB overlayer thickness.

Close modal

It is evident that the presence of moisture in contact with perovskite has a significant impact on the overall degradation process. Several reaction mechanisms have been proposed. Niu et al.3 proposed a 4-step mechanism by which CH3NH3PbI3 first dissociates into CH3NH3I and PbI2 in the presence of water during the first step.

CH 3 NH 3 PbI 3 CH 3 NH 3 I + PbI 2
(1)

Yang et al. proposed a different mechanism whereby during the early stages of degradation, the reaction between perovskite and water takes the form:

4CH 3 NH 3 PbI 3 + 2H 2 O ( CH 3 NH 3 ) 4 PbI 6 2H 2 O + 3PbI 2
(2)

Common to both proposed reaction schemes, however, is the role of water as a crucial component in the formation of PbI2.

Increasing fraction of solid PbI2 corresponds to a decrease in perovskite fraction and thus, degradation. A plot of the fraction of new PbI2 phase formed as a function of time (ln(t), where t is in hours) is given in Figure 4. The increase in PbI2 fraction is observed to follow the sigmoidal trend characteristic of Avrami kinetics.17 As the amount of PbI2 formed during deposition was not the result of degradation, it has been subtracted from the total amount of measured PbI2. During the early stages, the fraction of PbI2 is observed to increase slowly, wherein nucleation of second phase PbI2 particles begins, followed by a period of rapid increase signifying second phase growth, finally reaching a plateau.

FIG. 4.

Fraction of PbI2 phase as a function of air exposure time obtained from samples with various NPB capping layer thickness, as labelled. The solid lines are the theoretical data computed using Avrami equation with fitting parameters shown in the inset.

FIG. 4.

Fraction of PbI2 phase as a function of air exposure time obtained from samples with various NPB capping layer thickness, as labelled. The solid lines are the theoretical data computed using Avrami equation with fitting parameters shown in the inset.

Close modal

The Avrami equation has the following form18 

C t = 1 e K t n
(3)

This equation is also known as the Johnson-Mehl-Avrami-Kolmorgorov equation or simply the Avrami equation.17 Here K, with units of hr−n, is the Avrami constant, which is proportional to nucleation and growth rate of second phase, and n is the Avrami exponent, a material specific parameter dependent on nucleation site and growth mechanics.17,18 These parameters are generally experimentally derived and thus the values were determined using curve fitting. It is proposed that this kinetic pathway begins upon contact between the perovskite and moisture, which has diffused through the NPB layer.

As shown in Figure 4, the theoretical data (in solid lines) generated based on the Avrami equation fit well with all experimental data obtained from the perovskite samples with and without NPB overlayers. This clearly indicates PbI2 as a second phase that nucleates and grows out of the parent phase, following the typical process in a phase transformation. Parameters K and n provide insight into the observed reaction characteristics. As discussed previously, the value for n is material specific and depends on microstructural characteristics of the parent and second phases. The n value of 1.5 is indicative of nucleation and growth taking place primarily at grain boundaries and other interfaces as opposed to occurring homogenously in the parent matrix.19 As expected, this value is invariant across the samples. Furthermore, this parameter determines the slope of the plot during the linear portion in the middle of the curve wherein the fraction of second phase is rapidly increasing. The value for K, shown in the inset table of Figure 4, is highly dependent on nucleation and growth rates of the second phase. We observed three orders of magnitude decrease in K from the first, unprotected sample, to that with a 30nm capping layer. This trend can be well understood within the Avrami model. As the presence of water is essential for the nucleation and growth of the PbI2 phase, increasing NPB layer thickness thereby limiting moisture penetration into the perovskite film, causes the nucleation and growth rate of PbI2 to decrease accordingly. This is evidenced by the decreasing value for K and the rightward shift of the fitted curves.

In summary, improvement of vapour-deposited CH3NH3PbI3 film stability against moisture in air, at ambient conditions, was achieved with the use of NPB overlayer. Assessment of the impact of encapsulation thickness on impeding decomposition shows that exposure time at 50% of initial perovskite concentration increases linearly with NPB thickness, at least up to 30 nm. The perovskite degradation process is interpreted as a form of phase transformation, well modelled by the Avrami equation. These finding provide important insight into the kinetics governing the degradation of methylammonium lead halide-based perovskites.

Funding support for this work is provided by Canada Research Chair (Z.H. Lu), Natural Sciences and Engineering Research Council of Canada, Connaught Global Challenge Fund of the University of Toronto and National Natural Science Foundation of China. The research infrastructure funding is provided by Canada Foundation for Innovation and Ontario Research Fund-Research Infrastructure.

1.
M. A.
Green
,
A.
Ho-Baillie
, and
H J.
Snaith
,
Nature Photon.
8
,
506
(
2014
).
2.
M. A.
Green
,
K.
Emery
,
Y.
Hishikawa
,
W.
Warta
, and
E.D.
Dunlop
,
Progress in Photovoltaics
23
,
1
(
2015
).
3.
G.
Niu
,
W.
Li
,
F.
Meng
,
L.
Wang
,
H.
Dong
, and
Y.
Qiu
,
J. Mater. Chem.
2
,
705
(
2014
).
4.
J.
Burschka
,
N.
Pellet
,
S. J.
Moon
,
R.
Humphry-Baker
,
P.
Gao
,
M. K.
Nazeeruddin
, and
M.
Gratzel
,
Nature
499
,
316
(
2013
).
5.
J.
Yang
,
B.
Siempelkamp
,
D.
Liu
, and
T.
Kelly
,
ACS Nano.
9
,
1955
(
2015
).
6.
I. C.
Smith
,
E. T.
Hoke
,
D.
Solis-Ibarra
,
M. D.
McGehee
, and
H. I.
Karunadasa
,
Angew Chem. Int. Ed.
126
,
11414
(
2014
).
7.
J. H.
Noh
,
S. H.
Im
,
J. H.
Heo
,
T. N.
Mandal
, and
S. I.
Seok
,
Nano Lett.
13
,
1764
(
2013
).
8.
L.
Zheng
,
Y.-H.
Chung
,
Y.
Ma
,
L.
Zhang
,
L.
Xiao
,
Z.
Chen
,
S.
Wang
,
B.
Qu
, and
Q.
Gong
,
Chem. Commun.
50
,
11196
(
2014
).
9.
M.
Liu
,
M. B.
Johnston
, and
H. J.
Snaith
,
Nature
501
,
395
(
2013
).
10.
K. L.
Tong
,
S. K.
So
,
H. F.
Ng
,
L. M.
Leung
,
M. Y.
Yeung
, and
C. F.
Lo
,
Synt. Met.
147
,
199
(
2004
).
11.
B.-S.
Kim
,
T.-M.
Kim
,
M.-S.
Choi
,
H.-S.
Shim
, and
J.-J.
Kim
,
Organic Electronics
17
,
102
(
2015
).
12.
J.-A.
Cheng
and
P.-J.
Cheng
,
J. Chem. Crystallography.
40
,
557
(
2010
).
13.
Q.-T.
Le
,
E.
Forsythe
,
F.
Nuesch
,
L.
Rothberg
,
L.
Yan
, and
Y.
Gao
,
Thin Solid Films
363
(
1
),
42
46
(
2000
).
14.
F.
Avendano
,
E.
Forsythe
,
Y.
Gao
, and
C.
Tang
,
Synthetic metals
102
(
1
),
910
911
(
1999
).
15.
See supplementary material at http://dx.doi.org/10.1063/1.4930082 for SEM micrographs.
16.
N. J.
Jeon
,
J. H.
Noh
,
Y. C.
Kim
,
W. S.
Yang
,
S.
Ryu
, and
S.
Seok
,
Nat. Mater.
13
,
897
(
2014
).
17.
D. A.
Porter
,
K. E.
Easterling
, and
M.
Sherif
,
Phase Transformations in Metals and Alloys
, 3rd ed. (
CRC Press
,
2009
), p.
287
.
18.
M.
Avrami
,
J. Chem. Phys.
8
,
212
(
1940
).
19.
W. M.
Kriven
,
Y. C.
Zhou
, and
M.
Radovic
,
Strategic Materials and Computational Design: Ceramic Engineering and Science Proceedings
(
John Wiley & Sons
,
Hoboken
,
2010
), p.
177
.

Supplementary Material