1. Introduction

Solution-processed organic-inorganic hybrid perovskites (OIHPs) have gained unprecedented attention in photonic and electronic applications for developing high performance solar cells, light emitting diodes (LEDs), field effect transistors (FETs) and solid state lasers. These are primarily due to their excellent optoelectronic properties, such as broad photon absorption spectra, direct band-gap, long charge carrier diffusion lengths and relatively small effective masses for electron and hole [1–3]. By far, the polycrystalline thin film form of the OIHPs still plays the dominating role in photonic devices. Many efforts have been devoted to synthesize and optimize high quality polycrystalline films via various methods such as post-treatment [4,5], newly synthesized perovskite precursors [6,7], organic additives [8,9], different fabrication methods [10,11] and organic solvent engineering [12,13], in order to extensively elaborate device performance. Nevertheless, it is generally agreed that the polycrystalline film form of the OIHPs may contain a large number of traps/defects existing at surfaces and grain boundaries; consequently, charge transport scattering, trap-assist recombination and ion migration via the grain boundaries are expected to be responsible for the device degradation [14–16].

Fortunately, some experimental studies have shown large single crystalline OIHPs can be synthesized via the solution method even in the ambient temperature. It brings us a new opportunity to investigate electronic transport behaviors in single crystalline OIHPs based optoelectronic devices in order to explore some novel and multi-functional properties of the devices. For example, a CH(NH₂)₂PbI₃ single crystal based photodetector was made showing a low noise current at low frequencies [17]. A high gain of more than 10⁴ electrons per photon, and large MAPbBr₃ single crystals based photodetector was demonstrated [18]. The CH₃NH₃PbI₃ single crystal was also reported for a self-powered photodetector [19]. A perovskite of the cubic structure was made in order to boost performance and stability of a photodiode [20]. By bulk defect reduction and surface passivation, a high mobility of 1.2 × 10⁻² cm²V⁻¹ and small surface charge recombination velocity of 64 cms⁻¹ were achieved [21]. In comparison to polycrystalline OIHPs in photovoltaic and light emitting devices, the single crystalline counterparts are particularly favorable in the applications for photodetector and photodiode due to their low bulk defect density, strong visible light absorption abilities, fast photo-electrical responses, high mobility and excellent stability in the air.

In this work, we have successfully synthesized and fabricated a sizable ($>5 mm$) CH₃NH₃PbI₃ single crystal with low trap density and high mobility based white-light photodetector comprising a simple vertical structure glass/ITO/Ga/CH₃NH₃PbI₃/Au. A large tunable on-off ratio can be achieved from 65 to 2250 at zero bias. The photodetector exhibits a relatively high responsivity (R) with the best record of 36.2 mA/W under weak white-light excitation. Despite this, the detectivity (D*) was measured to be approximately 2.68 × 10¹¹ Jones at the same condition. Both of them start to decrease with the increase of the illumination intensity. Based on ac-modulated impedance spectroscopic measurements and illumination intensity dependence of $J_{sc}$, the behavior has been attributed to the presence of surface trap-related electronic transport process.

2. Experiments

2.1 CH₃NH₃PbI₃ single crystal synthesis

The CH₃NH₃PbI₃ single crystal was grown by the inverse temperature crystallization method [22]. During the synthesis, a CH₃NH₃PbI₃ precursor was prepared by mixing CH₃NH₃I and PbI₂ with a molar rate of 1:1 in the GBL organic solvent. The solution was stirred at 60 °C for approximately 2 hours, and then it was filtered by a PTFE filter with a 0.2 μm pore size. A 2 ml filtrate was placed in a vial and it was then kept in an oil bath at 110 °C. It took approximately 6 hours to complete the growth of the single crystal. After this, the diethyl ether (DEE) and isopropyl alcohol (IPA) were used to clean the crystal in order to remove the remaining precursors.

2.2 Photodetector fabrication

During the device fabrication, the CH₃NH₃PbI₃ single crystal was mounted on a lapping fixture using a paraffin wax. Then, it was polished by SiC abrasive papers (2000, 5000, and 7000 grit in order). After this, some residual paraffin waxes were thoroughly washed away by a warm (80 °C) isooctane. Finally, on top of the ITO coated glass, the Ga and Au were made as the bottom and top electrodes respectively for the photodetector.

2.3 Material and device characterizations

Steady state photoluminescence (SSPL) spectra were recorded by a fluorescence spectrometer (Jobinyvon Horiba Fluorolog-3). Time-resolved photoluminescence (TRPL) spectra were measured by a single photon counting system, and a 405 nm pulsed laser diode was utilized as the photo-excitation source. X-ray diffraction (XRD) was performed by an X-ray diffracted spectrometer (Bruker D-8 Advance). Photocurrent responses of the photodetector were recorded by a source meter unit (Keysight B2912A) and a standard solar simulator was used as the photo-excitation source. Different illumination intensities were adjusted by optical density (OD) filters. Impedance spectroscopic measurements were performed by an LCR impedance analyzer (Agilent E4990A).

3. Results and discussion

We started with basic characterizations of the single crystal CH₃NH₃PbI₃. Figure 1(a) provides the photographic image of the solution-processed single crystal with the planar size of more than 5 mm. Owing to the single crystallinity, both the XRD and SSPL spectra are distinctly different from its polycrystalline form in Figs. 1(b) and 1(c) respectively. The XRD spectrum of the single crystal CH₃NH₃PbI₃ (i.e. the bottom one) reveals less diffracted peaks by comparing to its polycrystalline film (i.e. the top one) indicating the formation of less impurity crystalline phases [23,24]. In Fig. 1(c), the normalized SSPL spectrum locates at 764 nm (red) for the CH₃NH₃PbI₃ single crystal with respect to the one of the polycrystalline film at 760 nm (blue). The relative red-shift of the spectrum is ascribed to the increase of the grain size since a relatively weaker lattice distortion appears in the crystal with large grain size; as a consequence, it results in a smaller bandgap [24,25]. Figure 1(d) shows the TRPL spectra for both the polycrystalline film and the single crystal respectively. Their corresponding lifetimes are extracted by fitting the decay curves using the bi-exponential model $Y =A_1\exp \left(-\frac{t}{{\tau }_1}\right)+A_2\exp \left(-\frac{t}{{\tau }_2}\right)+y_0$, in which, A₁ and A₂ denote fractions for a fast $({\tau }_1)$ and slow $({\tau }_2)$ decay times respectively. As we can see from Table 1, the CH₃NH₃PbI₃ single crystal has the lifetime of approximately 67 ns, in contrast, the one of the polycrystalline film is decreased to 11 ns. It is, thus, possible to obtain a much longer electron-hole lifetime in the CH₃NH₃PbI₃ single crystal.

 

Fig. 1 (a) The photographic image of the solution-processed CH₃NH₃PbI₃ single crystal. (b) XRD spectra for the CH₃NH₃PbI₃ single crystal (bottom) and polycrystalline film (top) respectively. (c) Steady-state and (d) time-resolved photoluminescence spectra for the CH₃NH₃PbI₃ single crystal and polycrystalline film.

Download Full Size | PDF

Table 1. Fitting Parameters of the TRPL Spectra in Fig. 1(d) for the CH₃NH₃PbI₃ Single Crystal and Polycrystalline Film.

View Table | View all tables in this article

The trap density (n_trap) of the CH₃NH₃PbI₃ single crystal can be evaluated by the J-V characterization method for a hole-only device such as Au/CH₃NH₃PbI₃/Au [Fig. 2(a)] in the dark condition. The corresponding J-V characteristic curve is shown in Fig. 2(b), and it can be distinguished by three different regions (I, II and III). At the low electric field strength (Region I), $J_{sc}$ increases linearly with V indicating the Ohmic response for the device. The conductivity (σ) of the single crystal was estimated to be approximately $4.6\times {10}^{-4}\text{ S}\cdot {\text{cm}}^{-1}$ for n = 1, which is higher than the one of a typical CH₃NH₃PbI₃ polycrystalline film ($1.5\times {10}^{-5} \text{S}\cdot {\text{cm}}^{-1}$) [25]. At the moderate bias range (Region II), the rate of the increase of the $J_{sc}$ becomes fast for n > 2. When V > 3.7 V at the kink point, it is expected that all traps in the single crystal are completely filled by injected charge carriers. The applied voltage at the point is defined as the trap-filled limit voltage (V_TFL), which can be found by the expression $V_{TFL}=\frac{e{n}_{trap}L}{2\epsilon {\epsilon }_0}$ [26], where e is the elementary charge of the electron, L is the thickness of the CH₃NH₃PbI₃ single crystal, ε is the relative dielectric constant for CH₃NH₃PbI₃ (ε = 32.1) [27], ε₀ is the dielectric constant in the free space. In this case, the $V_{TFL}$ was estimated to be approximately 3.7 V for the CH₃NH₃PbI₃ single crystal, and n_trap was calculated to be approximately 1.3 × 10¹⁰ cm⁻³. It is much lower than the one of the CH₃NH₃PbI₃ polycrystalline film (8.9 × 10¹⁶ cm⁻³) [27]. At higher bias voltage (Region III), the J-V characteristic curve is quadratic dependent (n = 2), $\mu$ can be calculated by the Mott-Gurney Law $\mu =\frac{8J_D{L}^3}{9\epsilon {\epsilon }_0{V}^2}$ after fitting [28], and J_D is the dark current density. In this case, μ is approximately equal to 37.6 cm²V⁻¹s⁻¹ which is 4 orders of magnitude higher than the one of the polycrystalline CH₃NH₃PbI₃ film (3.9 × 10⁻³ cm²V⁻¹s⁻¹) [29]. In spite of this, the properties of some CH₃NH₃PbI₃ single crystals that have been reported in literatures are summarized in Table 2. In comparisons, our single crystal exhibits the lowest $n_{trap}$ using the ITC method. Reference [30] shows even higher $\mu$ than ours; however, $\sigma$ is known from that work. Other methods, such as AVC and TSSG have been reported for the syntheses of the single crystal as well. As the data shown in Table 2, the CH₃NH₃PbI₃ single crystals made by these two methods contain higher $n_{trap}$ than ours.

 

Fig. 2 (a) A schematic drawing for the hole-only device structure. (b) A J-V characteristic curve for the hole-only device, $V_{TFL}$ stands for trap-filled limit voltage.

Download Full Size | PDF

Table 2. Parameters were Obtained by Fitting the J-V Curve of Fig. 2(b). ITC, AVC and TSSG Denote Inverted Temperature Crystallization, Antisolvent Vapor-assisted Crystallization and Top-Seeded Solution Growth, Respectively.

View Table | View all tables in this article

Based on the as-prepared single crystal, the photodetector was fabricated and Fig. 3(a) schematically shows the device structure consisting of glass/ITO/Ga/CH₃NH₃PbI₃/Au. The illumination intensity dependence of the J-V characteristic curves for the device is given in Fig. 3(b). Apparently, it produces a very low J_D of approximately 1.9 × 10⁻⁴ mA/cm² at 0 V (inset). Such property is the essential prerequisite for a photodetector to achieve an outstanding performance and reliability [34]. With the increase of the illumination intensity, J increases dramatically with the increase of V, for example J is about $1\times {10}^{-1} \text{mA}/{\text{cm}}^2$ under 100 mW/cm² illumination at 0 V. Figure 3(c) provides light intensity dependence of $J_{sc}$ and their linear relationship can be fitted by the formula $J_{sc}=P^{\alpha }$ [35]. The deviation of the $\alpha -$ parameter from 0.5 indicates the presence of trap-assist recombination process. In this case, we found $\alpha =0.62$. Because of the lack of bulk trap states in the CH₃NH₃PbI₃ single crystal, we have assumed that surface trap states may dominate.

 

Fig. 3 (a) A schematic drawing of the photodetector glass/ITO/Ga/CH₃NH₃PbI₃/Au. (b) Illumination intensity dependence of J-V measurements, the inset shows the zoom-in on V close to 0. (c) Illumination intensity dependence of $J_{sc}$.

Download Full Size | PDF

The photodetective switching behavior is a unique feature to evaluate the performance of the photodetector. Here, we have measured the time-dependent photocurrent at the $J_{sc}$ condition under different illumination intensities using the white-light. In Fig. 4(a), the $J_{sc}$ increases with the rising illumination intensity from 0.1 mW/cm² to 100 mW/cm². Here, it undergoes the switching on-off for 4 cycles. An average $J_{sc}$ of 132 μA/cm² was achieved under 100 mW/cm² illumination intensity, showing an excellent on-off ratio of 2250. It is worthwhile to mention that the photodetector did not show any noticeable degradation during the measurements and all the $J_D$ coincide at the same level. In order to quantitatively evaluate the optoelectronic performance of the CH₃NH₃PbI₃ single crystal based photodetector, the three critical parameters such as responsivity (R), detectivity (D*) and on-off ratio were calculated at different illumination intensities ranging from 0.1 mW/cm² to 100 mW/cm² under white-light at 0 V bias, and the results are given in Figs. 4(b)-4(d) respectively [36]. R is defined as $=\frac{J_{light}-J_D}{P_{in} }$, in which P_in is incident light intensity. D* can be expressed as $D^{*}= \frac{R}{\sqrt{2e{J}_D}}$. Clearly, both R and D* decrease with the increase of $J_D$. The maximum R and D* are approximately equal to 36.2 mA/W and 2.68 × 10¹¹ Jones, respectively at 0.1 mW/cm². In comparison, the on-off ratio keeps rising until a maximum value of 2250 at 100 mW/cm².

 

Fig. 4 (a) Illumination intensity dependence of the switching behavior at 0 V for the CH₃NH₃PbI₃ single crystal based photodetector. Extracted values are for (b) responsivity R, (c) detectivity D* and (d) on-off ratios.

Download Full Size | PDF

The illumination intensity dependence of $J_{sc}$ [Fig. 3(c)] can be used to evaluate the trap-assist recombination process in an optoelectronic device. As it has been discussed above, the white-light intensity dependence of the on-off ratio, D* and R, are of critically importance for the determinations of the outputs of the single crystal photodetector. A complete understanding of the physical phenomena behind requires the non-destructive ac-modulation impedance spectroscopic characterization for the photodetector at working condition. Figure 5(a) displays the illumination intensity dependence of the impedance spectra for the photodetector consisting of glass/ITO/Ga/CH₃NH₃PbI₃/Au, and all the spectra were fitted by the equivalent electronic circuit in Fig. 5(b). In general, it contains two sets of the R-C circuits connected in parallel, $R_s$ denotes the series resistance, the two capacitors stand for the surface ($C_{surf}$) and bulk capacitance ($C_{bulk}$), respectively [37]. In this case, $C_{bulk}$ is represented by a constant phase elements (CPE) and it is usually determined by two major components, CPE-T and CPE-P. The value of CPE-P that varies from 0.5 to 1 decides whether the CPE-T behaves like a capacitor or a resistor. $R_{bulk}$ and $R_{surf}$ are the bulk and surface resistances respectively. Despite these, $R_{surf\_trap}$ and $C_{surf\_trap}$ are the surface trap-related resistance and capacitance respectively [38]. All the fitting parameters have been summarized in Table 3. The surface charge and bulk charge recombination lifetimes $({\tau }_{surf}, {\tau }_{bulk})$ that could be extracted from the impedance spectra were plotted in Fig. 5(c) [39]. As we can see, both of them tend to decrease with the increase of the white-light intensity; nevertheless, ${\tau }_{surf}$ exhibits two distinct features by comparing with ${\tau }_{bulk}$, (i) ${\tau }_{surf}$ is much shorter than ${\tau }_{bulk}$ within the complete white-light intensity range; (ii) the reduction of ${\tau }_{surf}$ is faster than ${\tau }_{bulk}$ at higher illumination intensities. The results elucidate that the surface recombination is expected to play dominating role for the performance of the photodetector at the J_sc condition. Moreover, $R_{surf\_trap}$ and $C_{surf\_trap}$ of Fig. 5(d) evidently reveals that the surface-trap associated recombination resistance decreases with the rising white-light intensity. Owing to the occupancy of the trap-states by photo-excited charge carriers at higher illumination intensities, $C_{surf\_trap}$ is increased. In the comparison with Figs. 4(b) and 4(c), the decrease of the photodeteictive parameters R and D* at higher illumination intensities can be explained by considering the effect of surface trap-assist recombination process at higher white-light intensities causing a significant reduction of the photo-generated current. In fact, the effect can be also reflected from the aforementioned result in Fig. 3(c) since the $\sigma$-parameter is in-between 0.5 and 1.0 indicating the presence of trap-related recombination process.

 

Fig. 5 (a) White-light illumination intensity dependence of Nyquist plots for the photodetector consisting of glass/ITO/Ga/ CH₃NH₃PbI₃/Au. (b) An equivalent electronic circuit for fitting the spectra. (c) Light intensity dependence of ${\tau }_{bulk}$ and ${\tau }_{surf}$. (d) Light intensity dependent of $R_{surf\_trap}$ and $C_{surf\_trap}$.

Download Full Size | PDF

Table 3. A Summary of Impedance Parameters Fitted from Fig. 5(a).

View Table | View all tables in this article

4. Conclusion

In summary, we have synthesized the CH₃NH₃PbI₃ single crystal with relatively lower trap density $n_{trap}$ of approximately 1.3 × 10¹⁰ cm⁻³, higher conductivity $\text{σ}=4.6\times {10}^{-4}\text{ S}\cdot {\text{cm}}^{-1}$ and mobility $\mu =37.6{\text{ cm}}^2{\text{V}}^{-1}{\text{s}}^{-1}$ by the ITC method. Based on this, the white-light photodetector of the structure glass/ITO/Ga/CH₃NH₃PbI₃/Au was fabricated. It exhibits remarkably large R (36.2 mA/W) and D* ($2.68\times {10}^{11}$ Jones) under weak white-light photo-excitation intensity such as 0.1 mW/cm² at zero bias. With the assist of impedance spectroscopic measurements, the reductions of the R and D* are mainly due to the enhancement surface trap-related recombination process. We believe this work serves as a fundamental study for the surface trap associated recombination mechanism in the CH₃NH₃PbI₃ single crystal based photodetector. The device is applicable for sensitive and weak white-light photo-detection.