Laser Terahertz Emission Microscope for Imaging Grain Heterogeneity: A Case Study of CH₃NH₃PbI₃ Perovskite Semiconductors

Abstract

Strong terahertz (THz) emission from the methylammonium lead iodide (MAPbI₃) perovskite semiconductors has been observed following above-bandgap photoexcitation, yet local THz responses of crystalline microstructures are absent. We implement laser THz emission microscope (LTEM), yet-to-be applied to the perovskite semiconductors, as a novel and complementary tool to evaluate the electronic and grain heterogeneity of MAPbI₃ thin films. Two MAPbI₃ samples with different grain sizes are studied. Using this approach, we show that the one with a larger grain size gives more uniform THz radiation. More significant spatial THz intensity fluctuation is observed for the sample with a smaller grain size.

1. Introduction

Organic-inorganic hybrid lead halide perovskite is one of the promising candidates for next-generation solar cells due to its advantages, such as low-cost manufacturing process and high photoconversion efficiency, which has attracted significant research attention over the past few years [1,2,3,4,5]. As the prototype perovskite material, methylammonium lead iodide (MAPbI₃) shows excellent photophysical properties, such as strong absorption in the visible range and extremely long carrier diffusion length [6,7]. In addition to solar cell applications, perovskites have been demonstrated as other optoelectronic devices, such as lasers [8] and light-emitting diodes (LEDs) [9]. Recent research interest extends to the potential application in THz emission [10,11,12]. When a perovskite thin film is photoexcited by an ultrafast pulse above the bandgap, the emerging transient current generated on a picosecond (ps) time scale leads to the free-space THz emission [13,14]. The strong THz emission from MAPbI₃ thin film at room temperature has been observed by several studies [10,11,12]. Possible mechanisms, such as photo-Dember, lateral photo-Dember [10], and surface depletion fields due to defects [11], have been proposed to explain the generation of THz radiation in MAPbI₃ thin films. The reported highest electric field of THz emission from MAPbI₃ thin film is one order of magnitude lower than the efficient THz emitter of p-type InAs [11].

Laser THz emission microscopy (LTEM) is a THz imaging technique that reflects the spatial variations of THz emission intensity of a thin film after photoexcitation by an ultrafast laser pulse. The most pronounced advantage of LTEM compared to other THz spectroscopy is the much-improved spatial resolution. In the LTEM, the spatial resolution is limited by the focused excitation beam size rather than emitted THz wavelength; therefore, the resolution can be sub-μm. The LTEM scheme measures photocurrent transport that complements conventional PL measurements for carrier density and recombination in photovoltaic research. The physics mechanisms of the two methods are different. As a novel characterization tool, LTEM has been demonstrated in potential applications, such as semiconductor device inspection [15] and polycrystalline silicon solar cell evaluation [16]. However, LTEM has never been applied to metal halide perovskite materials, in contrast to PL, which has been extensively used. Inspired by MAPbI₃’s strong THz radiation under photoexcitation, it is possible to apply the LTEM to characterize the quality of perovskite thin film samples.

In this letter, we report the imaging of the THz emission from two MAPbI₃ perovskite thin film samples with different grain sizes at room temperature by employing a custom-built LTEM setup. The MAPbI₃ thin films with and without 5% lead thiocyanate (Pb(SCN)₂) additive were prepared on soda-lime glass substrates following a one-step solution process [17] and capped by a 30 nm polymethyl methacrylate (PMMA) protection layer. The spatial homogeneity of THz emission intensity varies with the sample quality. The sample with stronger photoluminescence (PL) signal and larger grain size shows more uniform THz radiation. In addition, the excitation angle dependent THz emission measurements reveal that the transient photocurrent direction is normal to the sample surface.

2. Materials and Methods

The experimental setup is illustrated in Figure 1a. A noncollinear optical parametric amplifier (NOPA) centered at 800 nm with 22 fs pulse width and 2 MHz repetition rate is frequency-doubled by a beta barium borate (BBO ) crystal. Then, the horizontally-polarized 400 nm pulse is focused on the perovskite thin film sample by a 50× long working distance objective with a 0.9 μm focal spot size. The excitation power of the 400 nm pulse is 6 mW. A two-dimensional (2D) piezo XY scanner is used to raster scan the sample in order to map out 2D images. The radiated THz electric field in transmission geometry is collected and focused onto a photoconductive THz detector with a bandwidth up to 4.5 THz [18]. An optical delay stage is used to control the relative delay between a heterodyning optical beam and the received THz radiation. The THz emission intensity as a function of photoexcitation angle is measured by rotating the sample orientation. During the measurement, the setup was purged with nitrogen gas to minimize the sample degradation and THz absorption by moisture. The substrates were placed to face the excitation beam. Further experimental details for THz spectroscopy [19,20,21,22,23] and other THz microscopy [24,25] can be found elsewhere.

PL measurements have been adapted as a standard and well-accepted method to inspect the quality of perovskite thin films [26,27]. Higher PL intensity and longer decay lifetimes are indicators of better functional materials [26]. To demonstrate the quality of samples used in the experiment, we performed PL image measurements using our home-built confocal scanning microscope [5]. The resolution of the confocal scanning microscope image system is ∼0.5 μm under 400 nm excitation. The corresponding two 5 μm × 5 μm PL images of MAPbI₃ thin films with and without Pb(SCN)${}_2$ additives are shown in Figure 1b,c, respectively, at an excitation fluence close to one sun (100 mW/cm²). The PL image of MAPbI₃ thin film with Pb(SCN)${}_2$ in Figure 1b shows bright and large grain size (>1 μm). The grain interior and grain boundary can be clearly distinguished with higher PL counts inside the grain. On the contrary, the PL image of MAPbI₃ without Pb(SCN)${}_2$ shows much weaker PL emission, as in Figure 1c, in which no clear grain/grain boundary structure can be resolved. The larger bright region is a cluster of signals from multi small grains. The grain size is much smaller than our resolution, and one cannot tell them apart. The grain size observed from PL measurement is consistent with the previously reported scanning electron microscope (SEM) images where the grain size of the MAPbI₃ thin film with Pb(SCN)${}_2$ additive is ∼2 μm and the one without an additive is ∼100–400 nm [17]. Through PL measurement, we confirmed the MAPbI₃ with Pb(SCN)${}_2$ additive used in the experiment is of higher quality than the one without Pb(SCN)${}_2$ additive.

3. Results

In Figure 2, we performed LTEM on MAPbI₃ thin films with and without Pb(SCN)${}_2$ additive. The THz emission 2D images were plotted as the peak amplitude of THz radiation by fixing the time delay at ∼3.5 ps, as shown in Figure 3a. The photoexcitation angle was set at 45 degrees for a larger THz signal size, to be discussed in Figure 3. The image scan range here is 2.5 μm × 5 μm with 100 nm step size and 2 s pixel dwell time. The THz emission images clearly reflect the difference between the two samples. From the previous report [17] and our PL measurement in Figure 1, the MAPbI₃ sample with Pb(SCN)${}_2$ additive has a larger grain size than the one without it. The THz emission image from MAPbI₃ sample with Pb(SCN)${}_2$ additive in Figure 2a shows a much smaller spatial fluctuation than the one without the Pb(SCN)${}_2$ additive (Figure 2b). Figure 2c compares spatial variation within the selected areas as labeled by the dashed rectangular box in Figure 2a,b. The positions of the rectangular boxes are selected based on the area where the largest THz fluctuation has been observed for the proof of principle demonstration of the technique. It is clearly visible that the spatial fluctuation of THz peak amplitude from the sample without Pb(SCN)${}_2$ additive is more than three times larger than the sample with the additive. The data show that a larger grain size results in more uniform THz emission. By adding Pb(SCN)${}_2$ additive, the perovskite sample will enable the passivation of grain boundary [17]. The significant spatial variation in the MAPbI₃ without Pb(SCN)${}_2$ additive indicates that the grain boundary may play a detrimental role in the THz emission. From PL images and LTEM images, the sample with a larger grain size exhibits both stronger PL emission and more uniform THz radiation. Therefore, in addition to conventional PL microscopy, the LTEM can also be used to inspect the quality of perovskite samples.

Figure 3a shows a typical time-domain THz emission signal from the MAPbI₃ thin film sample with Pb(SCN)${}_2$ additive and 400 nm excitation at room temperature. Note the emitted THz amplitude is independent of the excitation laser polarization [10]. The corresponding frequency domain spectrum is shown in Figure 3b. Unlike the reflection geometry measurement, which shows multiple peaks between 0–3 THz [10,11], the THz emission spectrum from the transmission geometry shows a single broad peak centered at ∼0.3 THz. This might be related to the strong THz absorption in a 1–3 THz range [28].

In the end, we show the effect of excitation angle on the detected THz amplitude. We measured the THz radiation on the MAPbI₃ thin film with Pb(SCN)${}_2$ additive at 0, 30, and 45 degree excitation angles by rotating the sample. The excitation angle $\theta$ is defined as the angle between the incident beam and the normal direction to the sample surface (inset of Figure 3a). As shown in Figure 3c, the strongest THz radiation signal is observed at a 45 degree excitation angle. As the excitation angle becomes smaller, the amplitude of measured THz signal is reduced. The THz signal is negligible under normal incident excitation $\theta$ = 0. The excitation angle dependent measurement indicates that the emitted THz is generated by a transient photocurrent in the direction normal to the sample surface [29]. For this reason, the excitation angle was set at 45 degrees for LTEM 2D images presented in Figure 2.

4. Conclusions

In conclusion, we reveal THz emission from the photoexcited transient current on MAPbI₃ thin film samples with and without the Pb(SCN)${}_2$ additive. The photocurrent direction is normal to the sample surface. The corresponding THz emission 2D images of the two samples are obtained by LTEM. Sample grain size closely correlates with the spatial THz emission homogeneity. For perovskite samples with a larger grain size, the local THz emission fluctuation is smaller. Therefore, LTEM has shown great potential as a novel tool to inspect and characterize the MAPbI₃ sample quality.