Unusual photoluminescence behavior of CH3NH3PbI3 thin-films

International Conference on Hybrid and Organic Photovoltaics
Proceedings of International Conference on Hybrid and Organic Photovoltaics 2015 (HOPV15)

Poster, Vipul Kheraj, 438
Publication date: 5th February 2015

In spite of rapid progress in power conversion efficiency from mere 3.8% [1] to 20.1% [2] in last five years, reliability and reproducibility issues with the CH3NH3PbI3 are still the major roadblocks in development of commercially viable solar PV devices. Moreover, the unusual characteristics, in particular, the anomalous hysteresis in the I-V curve [3] and slow dynamic processes [4] are fundamental issues that lack comprehensive explanations till date, making this an interesting material-system from research point of view. Here, we report a systematic investigation on the illumination and air-exposure induced degradation of CH3NH3PbI3 films. We used two samples of CH3NH3PbI3 films, one immediately encapsulated in the glove-box after deposition and one kept un-encapsulated in ambient air. Strong XRD peaks corresponding to CH3NH3PbI3 and no peak related to PbI2 phase in the XRD profiles measured immediately after removing the film from glove-box, revealed that the conversion of CH3NH3PbI3 from the precursors was complete at time-zero. However, with time the XRD profiles shown shrinking volume of CH3NH3PbI3 with increasing peak of PbI2. Interestingly, this was accompanied by the increase in the photoluminescence (PL) yield of peak corresponding to CH3NH3PbI3 and the carrier life time as measured by the time-resolved photoluminescence. These phenomena can be explained by the decomposition of CH3NH3PbI3 back to PbI2, especially at the grain boundaries, resulting in self-passivation of the grain-boundaries [5]. In addition to these variations in PL yield and carrier life time on longer time scales of few days, we also observed that the intensity of PL peak keeps changing for several seconds after opening the shutter in front of the excitation laser before getting stabilized. To investigate this further, we carried out PL measurements by periodically opening the shutter for 300 second after closing it for varying amount of time. We observed in both the samples that the variation of PL peak intensity was quite slower during periodic illumination as compared to that during continuous illumination. This indicates that the laser-illumination speeds up the decomposition process of CH3NH3PbI3. Interestingly, on a shorter time scale during the laser-illumination, the PL peak intensity showed gradual increment for a few seconds after opening the laser-shutter irrespective of shutter-closed time in case of encapsulated samples. However, the un-encapsulated sample behaved strangely when illuminated with laser periodically. When the shutter was opened first time after long breaks, it showed gradual increment but on closing and again opening the shutter after short time (5-10 minutes), it exhibited spike followed by gradual decrements in the PL peak-height. We increased the shutter-closed time and found that it takes about a couple of hours of darkness to change the behavior from spike-decrement to gradual-increment. Although the exact reason for these strange variations of PL intensity is not yet known to us, the following hypothesis could possibly explain these behaviors. The light-soaking effect caused during the laser illumination accelerates the decomposition of CH3NH3PbI3 into PbI2 and Ch3NH3I [6]. Out of these, the PbI2 is known to passivate the grain boundaries, thereby reducing the non-radiative recombination centres, which in turn improves the PL yield initially. The CH3NH3I, on the other hand, may diffuse to the surface and further degrade into CH3NH2 and HI. This finally leaves the halide ions on the surface if the sample is exposed to air as in case of un-encapsulated sample. The halide ions are relatively stable to air exposure at the surface and they may introduce trap states which enhance non-radiative recombination [7]. Thus, once this process begins, every time when the shutter is opened, the decomposition of CH3NH3I into CH3NH2 and HI is prevailing reaction and hence the PL intensity starts decaying on laser illumination. Also, since the film keep losing one of these precursors, CH3NH3I, continuously due to further decomposition, the rate of degradation is much higher in case of un-encapsulated sample. However, when kept unilluminated for sufficient time, the CH3NH3PbI3 decomposition slows down, making the surface depleted of CH3NH3I. When the shutter is opened after such a long period, again the primary decomposition of CH3NH3PbI3 dominates making more PbI2 which improves the PL yield initially for some time. In case of the encapsulated sample, the volatile precursor cannot easily escape the film and hence the precursors' concentration is more or less maintained even after the primary decomposition of the CH3NH3PbI3. Hence, on closing the laser-shutter, the co-existing precursors in the film again converts back to CH3NH3PbI3 to some extent. Thus, every time when the shutter is opened, the formation of PbI2 at the grain boundaries is the primary reaction taking place. This causes the PL yield to increase gradually on illumination in case of encapsulated sample. A series of controlled experiments, however, are necessary to further investigate this unusual PL behavior of CH3NH3PbI3.

[1] Kojima et al, J. Am. Chem. Society 131 (2009) p. 6050 [2] N J Jeon et al, Nature 517 (2015) p. 476 [3] Snaith et al, J. Phys. Chem. Lett. 5 (2014) p. 1511 [4] Sanchez et al, J. Phys. Chem. Lett. 5 (2014) p .2357 [5] Chen et al, Nano-Letters 14 (2014) p. 4158 [6] Niu et al, J Mater Chem A (2014) DOI: 10.1039/c4ta04994b [7] Abate et al, Nano-Letters 14 (2014) p. 3247

© FUNDACIO DE LA COMUNITAT VALENCIANA SCITO