A Facile Centrifuge Coating Method for High-Performance CsPbBr₃ Compact and Crack-Free Nanocrystal Thin Film Photodetector

Abstract

All-inorganic perovskite quantum dots (QDs), a promising semiconductor material, is suitable for new generation optoelectronic application. While there are many kinds of coating procedures for producing perovskite QDs peorovskite film, those methods require post-treatments and an additional dispersion support agent while still retaining pinholes and cracks. In this work, we report a facile method to produce CsPbBr₃ film on a pre-patterned Pt electrode using a centrifuge coating method for photodetector (PD) application. Compact and crack-free films with ~500 nm thick from various particle sizes of 8 nm, 12 nm, and >30 nm were achieved with a suitable ratio of toluene/ethyl acetate solvent for visible light photodetector application. The optimized device has an on/off ratio of 10³, detectivity of 3 × 10¹² Jones, and responsivity of 6 A/W. In comparison, the on/off ratio of the device fabricated by the centrifuge coating method was 10² times higher than by the drop-coating method. The PD performance exhibited considerable moisture stability at mild high ambient temperature with no encapsulation for more than two weeks. The results suggest that this is a potential method for fabricating all inorganic perovskite nano-semiconductor films for further optoelectronic application in photodetectors, LEDs, and solar cells.

1. Introduction

The rapid improvement of the fabricating methods of all-inorganic pervovskite has led to a remarkable improvement of its optoelectronic applications such as solar cell, light emitting diode, photodetector, and so on [1,2,3]. For example, in solar cell application, a variety of photovoltaic technologies for light harvesting have been employed to date. They are classified based on photoactive layers, including silicon solar cell, thin film solar cell (e.g., gallium arsenide and copper indium gallium selenide), and some emerging PV tecnologies, such as organic photovoltaic (OPV) [4,5], sensitization-based solar cells [6], and perovskite solar cells [7]. Among these, perovskite solar cell is a promising candidate due to the high light-absorption coefficient, long carrier diffusion length, and solution processibility of metal halide perovskite materials [7]. Because the bulk perovskite crystals exhibit poor phase stability due to the weak ion bonding nature easily leading to phase separation [8,9], by reducing the size of crystals, the performance of perovskite nanocrystal (PVK NC)-based devices was largely improved due to reduced crystal strain resulting in their good phase stability, narrow PL emission, tunable light absorption spectrum, and high photoluminescence quantum yield (PLQY) in comparison with bulk counterparts [10,11,12]. The all-inorganic perovskite quantum dots are one of next generation semiconductor materials which are currently of great interest for solar cell application because they are more durable than other hybrid perovskite materials. The power conversion efficiency of PQDSCs was significantly increased from 10.77% to 17.39% (certified 16.6%) by the control fine surface chemistry of PQDs and the device physics of PQDSCs [13,14]. Despite extensive research efforts, the efficiency of the PSCQSC device is still below that of other hybrid perovskite materials with the highest conversion efficiency of 25.17% [15]. Therefore, there is an important interest in further boosting the performance of fully inorganic PQDSCs through optimizing the synergies of films and device interfaces. Generally, the as-prepared PVK NCs using a solution process are usually coated with an insulating nature of long chain-ligand which makes it difficult to directly apply in optoelectronic devices. Thus, after fabrication, it requires a purification procedure for removing excess ligands and enhancing the opto-electronic performance [16,17]. Various PVK NC films have been successfully prepared by diverse coating methods, including spin coating, spray coating, dip-coating, and so on [18,19,20,21]. Several studies have reported on CsPbBr₃ NC film-based optoelectronic devices [21,22,23,24,25]. A large amount of residual solvent, problems adjusting viscosity, retaining pinholes, and cracks are major drawbacks of these methods. Lack of electrical connectivity within NC film, poor adhesion of the NC film to the substrate, and the requirement for post-treatment methods to provide conductive networks are further limitations of the above-mentioned approaches.

Recently, a high-speed centrifuge coating process was introduced for CNT, Ag nanowires, as well as CsPbBr₃ PVK NC film with long-term stability optoelectronic devices [25,26,27,28]. This technique relies on centrifugal force to form a uniform layer on the substrate surface perpendicular to the force, resulting in a closely packed nanomaterial film. Here, we demonstrate the fabrication of various particle size films of CsPbBr₃ NCs via the centrifuge coating method with a suitable ratio of toluene/ethyl acetate solvent for visible light photodetector application. The structural, photoluminescence properties and I–V performance of the CsPbBr₃ NC film photodetectors are also presented.

2. Materials and Methods

Materials: All reagents were used without further purification. Lead bromide (PbBr₂, 99.9% metal basis), cesium carbonate (Cs₂CO₃, 99.9% metal basis), oleylamine (OAm, 80–90%), oleic acid (OA, analytical grade), octandecene, ethyl acetate, and toluene (>99%, analytical grade) were bought from Macklin Chemical.

CsPbBr₃ NCs fabrication by hot injection method: Firstly, the Cs Oleate precusor was prepared by mixing 120 mg of Cs₂CO₃, 0.5 mL OA, and 5 mL ODE in a 50 mL three-necked flask, and heated at 120 °C until Cs₂CO₃ powder was completely dissolved. The precursor solution was maintained at 120 °C before use. Next, 70 mg PbBr₂ with 0.5 mL OA, 0.5 mL OLA, and 5 mL ODE was heated at 140 °C with magnetic stirring until completely dissolved. For CsPbBr₃ fabrication, there was rapid injection of 0.5 mL of Cs-oleate precursor remains at 120 °C into the PbBr₂ precursor at 120, 140, and 160 °C and stirring continued for 1 min. The solution after stirring was immediately soaked in ice water.

Centrifuge coating CsPbBr₃ NC film on Pt electodes: The planar integrated Ti/Pt (50/200 nm) electrode is fabricated using standard photolithography to pattern the photoresist for subsequent metallization and lift-off. The electrode included 40 fingers with the channel width and length of 20 μm and 1.60 mm, respectively. The total active area was 0.128 mm².

Prior to coating, pre-patterned Pt electrodes on the SiO₂/Si substrate were ultrasonically cleaned in acetone and ethanol for 10 min and dried in an oven at 100 °C for 30 min. Samples were placed vertically into a 0.5 cm diameter centrifuge tube. Before coating, 1 mL crude CsPbBr₃ NCs was washed by adding 1 mL ethyl acetate and centrifuged at 15,000 rpm for the removal of all excess organic agents such as ODE, OA, OLA. The supernatant was discarded and the particles were redispersed in toluene/ethyl acetate mixed solvent. Then, centrifuge coating was carried out with pre-patterned Pt electrodes faced with the direction of applied centrifuge force with CsPBr₃ NCs dispersed in toluene/ethyl acetate mixed solvent at a rotational speed of 15,000 rpm for 15 min. The resulting substrate was gently removed from the tube and dried at ambient air temperature to completely evaporate any remaining solvent.

3. Results and Discussion

CsPbBr₃ nanoparticles are synthesized using the hot-injection method which is similar to a previous method proposed by M. Kovalenco and colleagues [29]. It is well known that reaction temperature is the important parameter for controlling the size of nanoparticles. As shown in Figure 1, the size of nanoparticles stays in the range of 8 to 12 nm with reaction temperature of around 120 to 140 °C. By increasing the reaction temperature up to 160 °C, the particle size sharply increased up to 30–40 nm. Usually after fabrication, a purification procedure of the QD crude solution is required. Typically, a polar solvent such as ethyl acetate is used to promote the precipitation during centrifugal purification [30]. After that, the QD was redispersed in toluene nonpolar solvent for long-time storage. Moreover, the washing process can effectively remove excess ligands of CsPbBr₃ QDs and enhance the opto-electronic performance. Here, we realized that by adding a small amount of toluene (T) nonpolar solvent into ethyl acetate (EA) polar solvent during the centrifugal process could control the precipitation process of CsPbBr₃ NC film.

We first demonstrate the mixed EA/T solvent effect on the precipitation of the crude CsPbBr₃ NC solution. A schematic illustration for fabrication of CsPbBr₃ NC film via the centrifuge coating method is shown in Figure 2a. For this, a mixture of crude NC solution with various EA/T volume ratios was centrifuged at 15,000 rpm for 3 min. The photos, as shown in Figure 2b, indicated that the CsPbBr₃ NCs in crude solution were formed a yellow-green film on the wall of the centrifugal tubes when centrifuged with the volume ratio of EA/T from 1:0 to 1:0.4. Without toluene addition, The NCs are agglomerated in non-uniform CsPbBr₃ NC precipitation onto the wall of the tube. Conversely, with the excess of toluene content of 1:0.5 and 1:0.7, the particles can only settle at the bottom of the tube. As can be seen, the EA/T volume ratio of 1:0.2 shows a thick and uniform yellow film on the tube wall in comparison to the rest, suggesting that it was the optimal condition for the centrifugal coating film′s purpose. Then, by placing a Pt electrode surface so that it faces the centrifugal force in the tube, a green-yellow film was coated onto the electrode surface (see Figure 2b).

The surface and cross-section morphologies of the as-prepared CsPbBr₃ films were obtained via scanning electron microscopy (SEM) as shown in Figure 3. Both surface and cross-section images showed a dense packaging as well as compact, crack-free CsPbBr₃ films. The average particle sizes of 8, 12, and >30 nm for each sample are consistent with TEM images described above (Figure 3a,c,d), and an approximately 500 nm-thick film adhered directly to the SiO₂ substrate (Figure 3b). For comparison, the SEM image of drop-coated film was measured and is shown in Figure 3e. It can be seen that the image shows nanoparticles assembled discretely when compared to the sample made by centrifugal coating method. It is worth noting that isolated ligand-coated nanoparticles usually prevent the film′s SEM image from achieving high magnification. In this work, all CsPbBr₃ NCs prepared at 120 °C (S1), 140 °C (S2), and 160 °C (S3) with particle sizes ranging from 8 nm to >30 nm could be clearly seen in SEM images, implying that nanoparticles were not coated with organic long chain ligand; this allows us to hope that all films could have good opto-electrical properties for further application. The sample analyzed by EDX measurements showed a Cs:Pb:Br ratio of ~1:1:3 for CsPbBr₃ NCs (Figure 3f).

The PL spectra of three CsPbBr₃ NC coated films are shown in Figure 4a. The PL peaks are red-shifted from 515 nm to 526 nm, corresponding to a decrease in the bandgap energy from 2.4 to 2.35 eV for the CsPbBr₃ NCs with various particle sizes from 8 to ~30 nm. The Tauc plot curves calculated from UV-Vis spectra of CsPbBr₃ NCs solutions prepared at 120, 140, and 160 °C also confirm the change of the bandgap energy (Figure 4b).

Time-resolved photoluminescence (TRPL) was measured to clarify the change in the carrier charge transfer and recombination process for crude CsPbBr₃ NCs and centrifugal coated film, as shown in Figure 4c,d. The three exponential decays are described in the following equation [31]:

$$\mathrm{I}(t)={\mathrm{I}}_0+{\mathrm{A}}_1\exp (-\frac{t}{t_1})+{\mathrm{A}}_2\exp (-\frac{t}{t_2})+{\mathrm{A}}_3\exp (-\frac{t}{t_3})$$

(1)

where I_(t) is the luminescence intensity at time t; I₀ is the luminescence intensity at time t = 0; $t_1,t_2$, and $t_3$ are short and long decay components, respectively; and A₁, A₂, and A₃ are their pre-exponential factors, respectively. The average lifetimes (s) were evaluated using the following equation:

$$t_{\mathrm{avg}}=({\mathrm{A}}_1{t}_1^2+{\mathrm{A}}_2{t}_2^2+{\mathrm{A}}_3{t}_3^2)/({\mathrm{A}}_1{t}_1+{\mathrm{A}}_2{t}_2+{\mathrm{A}}_3{t}_3)$$

(2)

As expected, the decay time was increased from 60 ns (crude NCs) to 9.7 ms (dense film). It could be said that the NCs that formed a dense film with carrier charge transfer within the film are larger than isolated crude NCs, causing the recombination probability of the electron-hole declines. This is an advantage for opto-electric applications such as solar cell or photodetector. The qualifications of CsPbBr₃ NC centrifuge coated films were also confirmed by XRD analyses (Figure 5). All three samples exhibited a pure cubic phase (PDF# 54-0752), confirming the integrity of the perovskite structure.

The performances of three photodetector devices were under tungsten-halogen light source and are presented in Figure 6. Figure 6a shows the current–voltage (I–V) logarithm curves of devices with different particle size films. It can be observed that both the dark current and photocurrent (P_light = 1 mWcm⁻²) of the I–V curves are symmetrical under bias voltage applied from −2 V to +2 V with a gap of ~10³ times for all three samples. The response/recovery behavior in a single on/off cycle for three samples is presented in Figure 6b. The corresponding photocurrents at a constant applied voltage were 1.9 × 10⁻⁶ (A), 3.4 × 10⁻⁶ (A), and 2.3 × 10⁻⁶ (A) for S1, S2, and S3 samples, respectively. Even so, the difference is negligible compared to the gap of 10³ times between dark current and photocurrent. For comparison, the gap between the dark current and the photocurrent of the drop-coated film device is ~10 (from 2 × 10⁻⁹ to 1.2 × 10⁻⁸ A, see Figure 6a).

Photoresponse features have been plotted to infer the rise time (τr: Time taken for the photocurrent to rise from 10 to 90% of its amplitude) and decay time (τf: Time taken for the photocurrent to drop from 90 to 10% of its amplitude). The rise and decay times of devices were evaluated at about 4 s and 1 s for all three samples. Normally, the speed of response results involved using a light chopper. Because of limitations of the apparatus employed to acquire the photocurrent during the transient measurements, the rise and fall times should not be regarded as the absolute speed of response of our photodetectors. In fact, the actual photodetector response times are expected on the scale of the carrier lifetimes of the CsPbBr₃ NC film, which is described in Figure 4c.

A detailed study of the as-prepared photodetector was performed by assessing their photocurrent as a function of light intensity and with different wavelength cutoff filters (for the S2 sample only). As shown in Figure 6c,d, the photocurrent values were effectively changed. As expected, the photocurrent increased with the increasing of illumination intensity (see the (I) zone in Figure 6c) from 0 to 1 mW/cm⁻². Moreover, the photocurrent decreased with different wavelength cutoff filters (see (II)–(IV) zones in Figure 6c) and almost disappeared with 540 nm cutoff filter (IV) corresponding to the energy value of 2.3 eV (smaller than CsPbBr₃ NCs bandgap of ~2.35 eV).

Two important parameters of the device, the sensitivity (R) and detection index (D*), can be obtained from Equation (3):

R = I_ph/P_light

(3)

where I_ph is the photocurrent, P_light is the light power intensity (1 mWcm⁻²). D = R/$\sqrt{2{\mathrm{qI}}_{\mathrm{dark}}}$, where ${\mathrm{I}}_{\mathrm{dark}}$ is the dark current, q is the elementary charge (Coulomb). The highest values achieved for responsivity and specific detectivity were 6 A/W, 3.7 × 10¹² Jones, respectively, when the devices were exposed to tungsten-halogen lamp illumination with an intensity at 0.1 mW cm⁻². The analytical parameters of CsPbBr₃-based photodetectors are summarized in

Table 1. Comparison of device performance of CsPbBr₃-based photodetectors.

Active Materials	Wavelength (nm)	Responsivity (A/W)	Rise/Fall (s)	On/Off Ratio	Detectivity (Jones)	Ref
CsPbBr3 QDs	0.40 mWcm−2 @ 532 nm	4.7 × 10−3	0.2/1.3	1.6 × 105	16.84 × 108	[24]
CsPbBr3 micro-particles	1.01 mWcm−2 @ 442 nm	0.18	1.8/1	8 × 103	6.1 × 1010	[32]
CsPbBr3 nanosheets	0.35 mWcm−2 @ 442 nm	0.25	0.019/0.025	103	-	[33]
Single crystal CsPbBr3	1 mW @ 450 nm	0.028	90.7/57	102	1.8 × 1011	[34]
CsPbBr3 nano crystals film	Halogen lamp 0.1 mW/cm2	6.0	4/1	103	3.6 × 1012	this work

. The superior performance obtained from photodetectors via centrifuge deposited perovskite NC films might be attributed to the high intrinsic quality of the perovskite films (high quality compact films and with no pinholes).

For practical application, operational stability in air ambient as well as a mild-high temperature (up to 75 °C) are evaluated. Figure 7a shows that the photocurrent has no change after two weeks stored in air atmosphere without encapsulation (RH ~ 60%). The effect of ambient temperature on the operation of the component is observed by directly putting the sample on a heater holder at 75 °C and cooling to a temperature of 20 °C for a period of one hour. The photocurrent change according to temperature is recorded and presented in Figure 7b. As shown in the figure, the photocurrent started at ~50% value at 75 °C and increased with the decrease of ambient temperature; the photocurrents almost recovered to their original value. Thus, our photodetectors exhibited considerable moisture stability and mild temperature with no encapsulation for more than two weeks. The highly photodetector performance described in this work is expected to encourage the further development of all-inorganic perovskite nanocrystals films for optoelectronics applications.

4. Conclusions

We successfully fabricated CsPbBr₃ NC film on a pre-patterned Pt/SiO₂/Si substrate as a planar photodetector with metal/semiconductor/metal structure via a centrifuge coating method. A ~500 nm dense, pinhole- and crack-free CsPbBr₃ NC film coating was achieved using a mixed toluene/ethyl acetate solvent. As a consequence, the photodetector device performance showed several figures of merit such as an on/off ratio of 10³ times, responsivity of 6 A.W⁻¹, and detectivity of 3.5 × 10¹² Jones (by using 0.1 mWcm² tungsten-halogen light source). In comparison, the on/off ratio of the device fabricated by the centrifuge coating method was 10² times higher than by the drop-coating method. The results suggest that this is a potential method for fabricating high quality all-inorganic perovskite films for opto-electronic thin film devices such as photodetectors, LEDs, and solar cells.