Spray Coated Colloidal Quantum Dot Films for Broadband Photodetectors

A technique for scalable spray coating of colloidal CdSeTe quantum dots (QDs) for photovoltaics and photodetector applications is presented. A mixture solvent with water and ethanol was introduced to enhance the adhesive force between QDs and the substrate interface. The performance of the detector reached the highest values with 40 spray coating cycles of QD deposition. The photodetectors without bias voltage showed broadband response in the wavelength range of 300–800 nm, and high responsivity of 15 mA/W, detectivity of more than 1011 Jones and rise time of 0.04 s. A large size QD-logo pattern film (10 × 10 cm2) prepared by the spray coating process displayed excellent uniformity of thickness and absorbance. The large area detectors (the active area 1 cm2) showed almost the same performance as the typical laboratory-size ones (the active area 0.1 cm2). Our study demonstrates that the spray coating is a very promising film fabrication technology for the industrial-scale production of optoelectronic devices.


Introduction
Semiconductor nanocrystals with characteristic size below the Bohr radius of the charge carriers are called quantum dots (QDs). Application of the QDs in optoelectronic devices have attracted great attention because of their distinctive physical characteristics, such as size-tunable bandgap, high extinction coefficient, narrow emission spectra, high luminescent efficiency, and multiple exciton generation [1][2][3][4]. Colloidal QDs have been used to construct various optoelectronic devices like solar cell, light-emitting diodes, and photodetectors [5][6][7][8][9]. For example, detectors based on colloidal QDs have advantages like superior response range, high detectivity, low dark current density, high operating temperatures, excellent stability, and weak light response [10]. Photodetectors, which can capture optical signals and convert them into electrical signals, are playing increasingly significant role in many fields like optical communication, environmental monitoring, night vision and biomedical imaging [11][12][13][14]. For example, PbS quantum dots and 2D nonlayered CdS x Se 1−x nanosheet hybrid photodetectors with a detectivity of up to 1.05 × 10 15 Jones have been constructed [15]. Zhang et al. fabricated a vertical field-effect phototransistor based on PbS quantum dots, which displayed a high detectivity of 2 × 10 13 Jones [16]. Peng et al. assembled a flexible photodetector based on the PbS QDs/ZnO nanoparticles with a detectivity of 3.98 × 10 12 Jones [17]. Both rigid and flexible broadband photodetectors based on CdSeTe colloidal QDs have been made by our group [3,18]. These detectors reached remarkably high detectivity of more than 2 × 10 12 Jones in the range of 300-800 nm without

CdSeTe QD Synthesis
CdSeTe QDs were synthesized according to the previous work with some minor modifications [23]. 0.3084 g CdO, 18 mL paraffin liquid, and 6 mL OA were placed into a 200 mL three-necked flask. 0.0316 g Se, 1 mL TOP and 3 mL paraffin liquid were placed into a 50 mL three-necked flask. 0.026 g Te, 0.5 mL TOP and 1.5 mL paraffin liquid were placed into a 50 mL three-necked flask. These precursors were stirred at room temperature for 30 min while vacuuming, and then the temperature was raised to 60 • C. Ar 2 was pumped into the flask for two minutes, then the flasks were degassed for 30 min before raising the temperature to 90 • C. If there was no gas in the flasks, Ar 2 was injected into the flasks until the end of the experiment. The temperature of Cd precursor was eventually set at 260 • C until dissolved. The corresponding temperature for Se precursor is 150 • C, and for Te precursor 300 • C. Finally, each precursor was stored at a temperature of 50 • C and was ready for the next step. The mixed precursors (milliliter ratio, Cd:Se:Te = 5:0.7:0.3) were put into a 100 mL three-necked flask. The mixture was stirred at 50 • C for 5 min while vacuuming. The temperature of the flask was set to 90 • C and Ar 2 was pumped for 10 min. Then the flask was dried until there was no gas. Ar 2 was again pumped into the flask and the temperature was raised, the precursor was stirred at a 310 • C for 15 min. Then the flask was cooled to 260 • C, and 4 mL of OAm was rapidly injected into the mixture. Finally, the flask was cooled to room temperature. The QDs capped with OAm were obtained by centrifugation and dispersed in dichloromethane.

Device Fabrication
FTO covered glasses were successively put into ionized water, acetone, and ethanol for ultrasonic cleaning for 20 min, and then treated with UV-ozone for 20 min. A compact TiO 2 film was deposited on the FTO substrate by spray pyrolysis at 450 • C. When the temperature of FTO was reduced to room temperature, a layer of mesoporous TiO 2 was spin-coated with 4000 rpm for 30 s, and then heated to 450 • C for 30 min. The ligands of QDs were exchanged by mercapto acetic acid (TGA) to make them water solvable [3]. Then the TGA-QDs were dissolved in a mixture of deionized water and ethanol with a volume ratio of 2:1. QDs Spray Deposition: the distance between the nozzle and the substrate was set to 18 cm, the pressure value of the spray was 0.3 MPa, and the substrate heating temperature was 100 • C. The automatic motor controlled the cycling motion of the spray gun by setting different spray cycle times. The QD film was covered by a layer of Spiro-OMeTAD formed by spin coating with 4000 rpm for 30 s. Finally, thermal evaporation was used to make a layer of the Ag electrode under a vacuum.

Characterization
The scanning electron microscopy (SEM) images were measured by using a scanning electron microscope (SU8020, Hitachi Co., Tokyo, Japan). The UV-vis absorption of QDs and films were measured by an ultraviolet (UV) near-infrared (NIR) spectrophotometer (Puxi-T10, PERSEE, Beijing, China). Photoluminescence (PL) of QDs was measured by the Guangdong luminescence spectrometer FL-380 (GANGDONG, Tianjin, China). The transition electron microscopy (TEM) images were measured by a TEM instrument (JEM-2010, Hitachi Co., Tokyo, Japan). The current density-voltage curves were measured by a Keithley 2400 multimeter (TE Connectivity, Schaffhausen, Switzerland) under AM 1.5 G illumination simulated sunlight (100 mW/cm 2 ) (7-SS1503A). The structure of the materials was measured by X-ray diffraction (XRD, PANalytical B.V., Almelo, Netherlands) with Cu Kα radiation. The responsivity was measured by a Keithley 2400 multimeter, a 150 W xenon lamp (SOFN, Beijing, China) and a Spectral Product DK240 monochromator (Spectral Products, Putnam, CT, USA). The current-time response was measured by an electrochemical workstation (Chenhua, Shanghai, China). Figure 1a shows the schematic diagram of the spray coating platform. The air pressure, the distance between nozzle and substrate can be adjusted for processing QDs film. QDs solution will be atomized into liquid mist by a pressurized N 2 nozzle. When the liquid mist drops onto the substrate while being heated, the solvent will evaporate rapidly. The residual solute QDs condense on the substrate to form a compact film. The surface of colloidal CdSeTe QDs is initially coated with oleylamine (OAm) ligands. Here, in order to improve the charge carrier mobility of QDs film, the OAm ligands are replaced with short-chain mercapto acetic acid (TGA) ligand according to the previous work [3]. Dichloromethane (CH 2 Cl 2 ) as the solvent of the QD solution is also exchanged to deionized water. To balance the solvent volatilization and the adhesive force between QD and substrate, ethanol is added into the TGA-QDs aqueous solution. Our study demonstrated the optimal volume ratio of water to ethanol is 2:1. Figure 1b displays the energy band structure of the photodetector. The QDs layer will harvest photons and then generate excitons (electron-hole pairs). The photogenerated excitons dissociate in the QD layer and then electrons transfer to the electron transport layer (ETL) while holes correspondingly to the hole transport layer (HTL). Figure 1c,d show the corresponding structure and scanning electron microscopy (SEM) cross-section image of the photodetector consisting of ETL (compact and mesoporous TiO 2 ), active layer (TGA-QDs), HTL (Spiro-OMeTAD) and Ag electrode. Figure 1e shows the ultraviolet-visible (UV-vis) absorption and photoluminescence (PL) spectra of CdSeTe QDs. The absorption spectrum covers broad range from ultraviolet (UV) to near-infrared (NIR), which provides wide spectral response of the photodetector. Figure S1a shows the X-ray diffraction (XRD) pattern of CdSeTe QDs, in which there are two sharp diffraction peaks between the standard zinc blende CdTe (JCPDS Card No. 65-1081) and CdSe (JCPDS Card No. 19-0191). The transition electron microscopy (TEM) in Figure 1f displays the size uniformity of colloidal CdSeTe QDs. Figure S1b is the histogram of particle size distribution of the QDs, showing that the average diameter of the CdSeTe QDs is estimated to be 5 nm. The inset of high-resolution TEM (HRTEM) image exhibits a 0.37 nm lattice spacing according to the (111) plane of CdSe (lattice spacing: 0.351 nm) and CdTe (lattice spacing: 0.374 nm). These demonstrate that the QDs have good crystallinity. Figure 2 shows the SEM images of QDs films deposited by spray coating with different deposition cycles. The substrates are mesoporous TiO 2 (m-TiO 2 ) films as shown in Figure 2a. After 20 cycles of spray coating, the morphology of the m-TiO 2 substrate cannot be completely covered by QDs as shown in Figure 2b (marked by red circles). The m-TiO 2 substrate is completely covered by QDs after spray coating 30 cycles (Figure 2c). However, there are many pinholes (marked by white circles) in the QDs film. When the spray coating cycles are up to 40, the pinholes are diminished greatly (Figure 2d). The size of QDs aggregation is about 400 nm (marked by black circle). When the cycles of spray coating are more than 40, the morphology of the QDs film is no longer change. Figure S2a shows the UV-vis absorption spectra of QDs films fabricated by spray coating with different deposition cycles. The absorption value of the film gradually increases with increasing deposition cycles, indicating the increase of the thickness of the QD film. The fluctuations in the short wavelength of the absorption spectra are caused by the light scattering of TiO 2 substrates. The gradual deepening of the color of the QD film in Figure S2b also represents the increase of thickness.

Results and Discussion
Card No. 19-0191). The transition electron microscopy (TEM) in Figure 1f displays the size uniformity of colloidal CdSeTe QDs. Figure S1b is the histogram of particle size distribution of the QDs, showing that the average diameter of the CdSeTe QDs is estimated to be 5 nm. The inset of high-resolution TEM (HRTEM) image exhibits a 0.37 nm lattice spacing according to the (111) plane of CdSe (lattice spacing: 0.351 nm) and CdTe (lattice spacing: 0.374 nm). These demonstrate that the QDs have good crystallinity.     Figure 3a shows the current density-voltage curves of the photodetectors under light conditions. With the increase of spray coating cycles, the current density first increases and reaches the maximum when the deposition cycles are up to 40. It is attributed to both the increase of thickness and improvement of the coverage of QD films. When the deposition cycles are more than 40, the current density decreases gradually. The possible reason is that the film thickness exceeds the diffusion length of the charges, resulting in a decrease in the charge collection efficiency [22,23]. Figure 3b shows the current density-voltage curves of the devices under dark conditions. When the number of spray coating cycle is less than 30, the substrates cannot be completely covered by QDs. The much many pinholes of QD film caused serious carrier recombination, which results in high dark current density according to previous works [34,35]. When the substrates are uniformly covered by QDs, the devices show a lower dark current density due to the decrease of carrier recombination. However, when spray coating cycles are more than 40, the decrease of carrier separation efficiency and collection efficiency of QDs films result in the increase of the carrier recombination. Thus, the dark current density increases accordingly. The dark current density of the device is an indicator of noise degree of the photodetector, which reflects the ability to resist external interference. The lower dark  Figure 3a shows the current density-voltage curves of the photodetectors under light conditions. With the increase of spray coating cycles, the current density first increases and reaches the maximum when the deposition cycles are up to 40. It is attributed to both the increase of thickness and improvement of the coverage of QD films. When the deposition cycles are more than 40, the current density decreases gradually. The possible reason is that the film thickness exceeds the diffusion length of the charges, resulting in a decrease in the charge collection efficiency [22,23]. Figure 3b shows the current density-voltage curves of the devices under dark conditions. When the number of spray coating cycle is less than 30, the substrates cannot be completely covered by QDs. The much many pinholes of QD film caused serious carrier recombination, which results in high dark current density according to previous works [34,35]. When the substrates are uniformly covered by QDs, the devices show a lower dark current density due to the decrease of carrier recombination. However, when spray coating cycles are more than 40, the decrease of carrier separation efficiency and collection efficiency of QDs films result in the increase of the carrier recombination. Thus, the dark current density increases accordingly. The dark current density of the device is an indicator of noise degree of the photodetector, which reflects the ability to resist external interference. The lower dark current density means less background noise, indicating the better performance of the photodetector.
In Figure 3c the wavelength-dependent responsivity of photodetectors after depositing different number of spray cycles is shown. The responsivity (R) changes with the number of cycles is very similar to the photocurrent density of the devices. Initially the responsivity rises with the increase of the number of spray coating cycles until 40 cycles and starts dropping after that. Figure 3d exhibits current-time response of devices assembled by different spray coating cycles without bias voltage, representing the on-off switching properties of photodetectors. Although all variables show good on-off switching, the photodetector fabricated by spray coating 40 cycles shows the highest response current, which is consistent with its photocurrent density. However, the photocurrent of the QDs detector assembled by 20 cycles increases slowly to a steady value with the time under the exposer to the light, while the photocurrent of other devices increases rapidly and then drops to a stable value. Some previous studies have explained this behavior mainly by trap filling [14,36]. When the number of cycles is 20, the QD film cannot cover the TiO 2 substrate completely, resulting in many traps in the film. The photogenerated carriers need to fill the traps first. This process takes a relatively long time, inducing a slow increase of the photocurrent. From 30 to 60 cycles, the coverage and quality of the film are improved, so there are fewer defects. Thus, the current quickly reaches the highest value. The following slight decrease can be related to possible photo-induced traps. Figure 3e shows the normalized current of the rise and decay of the detector assembled by 40 cycles. The rise time (T r ) is defined as the time required to rise from 10% to 90% of the stable value of the current, while the decay time (T d ) is defined as the opposite [37,38]. The T r and T d of the device assembled by 40 cycles are 0.04 s and 0.05 s, respectively, indicating a fast response. Nanomaterials 2019, 9, x FOR PEER REVIEW 6 of 10 current, which is consistent with its photocurrent density. However, the photocurrent of the QDs detector assembled by 20 cycles increases slowly to a steady value with the time under the exposer to the light, while the photocurrent of other devices increases rapidly and then drops to a stable value. Some previous studies have explained this behavior mainly by trap filling [14,36]. When the number of cycles is 20, the QD film cannot cover the TiO2 substrate completely, resulting in many traps in the film. The photogenerated carriers need to fill the traps first. This process takes a relatively long time, inducing a slow increase of the photocurrent. From 30 to 60 cycles, the coverage and quality of the film are improved, so there are fewer defects. Thus, the current quickly reaches the highest value.
The following slight decrease can be related to possible photo-induced traps. Figure 3e shows the normalized current of the rise and decay of the detector assembled by 40 cycles. The rise time (Tr) is defined as the time required to rise from 10% to 90% of the stable value of the current, while the decay time (Td) is defined as the opposite [37,38]. The Tr and Td of the device assembled by 40 cycles are 0.04 s and 0.05 s, respectively, indicating a fast response.
where R is the responsivity, q is the elementary charge and Jd is the dark current density [11,17]. As shown in Figure 3f, all the values of D* are more than 10 11 Jones in the wavelength range of 350 nm to 700 nm. The relatively high D* indicates that the photodetector prepared by spray coating  Figure 3f shows the detectivity (D*) of the device fabricated by 40 cycles. The detectivity, which is a key parameter of photodetector performance, can be calculated as [39] where R is the responsivity, q is the elementary charge and J d is the dark current density [11,17]. As shown in Figure 3f, all the values of D* are more than 10 11 Jones in the wavelength range of 350 nm to 700 nm. The relatively high D* indicates that the photodetector prepared by spray coating technology has good performance. We measured the external quantum efficiency (EQE) of the devices as shown in Figure S3. It is also found that the EQE of the device fabricated by 40 cycles has the highest EQE value (6%), which is consistent with the responsivity results.
The signal-to-noise ratio (SNR) is the difference between the photocurrent minus the dark current, divided by the dark current, which represents the ratio of signal to background noise. A higher ratio indicates a lower disturbing noise. The linear dynamic range (LDR) is the parameter that the detector can detect the signal linearly within the range of incident light power, a larger LDR certifies a good detector function [14]. These detector characteristics are defined by the following formulas [3,40] where I light and I dark are the photocurrent and dark current, respectively. J light is the photocurrent density under a light intensity of 1 mW/cm 2 and J dark is the dark current density [41,42]. Figure 4a shows the SNR of the devices assembled by different cycles. It can be seen that the photodetector fabricated by 40 cycles has the highest SNR demonstrating remarkable noise endurance. LDR of devices is shown in Figure 4b.
where Ilight and Idark are the photocurrent and dark current, respectively. Jlight is the photocurrent density under a light intensity of 1 mW/cm 2 and Jdark is the dark current density [41,42]. Figure 4a shows the SNR of the devices assembled by different cycles. It can be seen that the photodetector fabricated by 40 cycles has the highest SNR demonstrating remarkable noise endurance. LDR of devices is shown in Figure 4b. The LDR values of the device assembled by different QDs deposition cycles are 50.9 dB for 20 cycles, 55.7 dB for 30 cycles, 58.0 dB for 40 cycles, 54.5 dB for 50 cycles, 52.3 dB for 60 cycles, respectively. The detector prepared by 40 QDs deposition cycles has the largest SNR and LDR. This is consistent with the detection performance of the devices.  Figure 5a shows the pattern of QDs film with a large area of 10 × 10 cm 2 deposited by spray coating. We selected seven points at different positions on the QDs film pattern (in Figure 5a) and the corresponding absorption spectra are shown in Figure 5b. These seven points have the same absorption value in the range of 300-900 nm which indicates the large QDs film prepared by spray coating has a good uniformity of thickness. Figure 5c shows the picture of a large size photodetector with an active area of 1 cm 2 and we also selected five points in the active area of the device. The responsivity curves of these five points are as shown in Figure 5d. At 380 nm, the responsivity rates of these points are 25.6 mA/W, 24 Figure 5a shows the pattern of QDs film with a large area of 10 × 10 cm 2 deposited by spray coating. We selected seven points at different positions on the QDs film pattern (in Figure 5a) and the corresponding absorption spectra are shown in Figure 5b. These seven points have the same absorption value in the range of 300-900 nm which indicates the large QDs film prepared by spray coating has a good uniformity of thickness. Figure 5c shows the picture of a large size photodetector with an active area of 1 cm 2 and we also selected five points in the active area of the device. The responsivity curves of these five points are as shown in Figure 5d. At 380 nm, the responsivity rates of these points are 25.6 mA/W, 24.9 mA/W, 25.6 mA/W, 25.2 mA/W, and 25.1 mA/W, respectively. At 530 nm, the responsivity rates of these points are 16.2 mA/W, 14.2 mA/W, 15.8 mA/W, 16.3 mA/W, and 14.7 mA/W, respectively. These five points have similar responsivity curves, which can be considered as a good uniformity QD distribution of the device with a large area. In addition, we compared two devices with different areas as shown in Figure S4. The active area of a device of normal laboratory-size is 0.1 cm 2 while the larger one is 1 cm 2 . Figure S5 exhibits a serial of parameters of the detection performance of normal and large devices. The short circuit current density of small area and large area devices is 0.39 mA/ cm 2 and 0.36 mA/ cm 2 , respectively. At 380 nm, the responsivity rates of two devices are 26.6 mA/W, 25.6 mA/W. At 550 nm, the responsivity rates of two devices are 16.8 mA/W, 15.5 mA/W. The rise time of the two devices is 0.04 s and 0.05 s, and the decay time is 0.07 s and 0.05 s, respectively. Figure S5d shows the statistical responsivity distribution for the different-size detectors at the wavelength of 380 nm and 550 nm. Although the responsivity of large-area devices is slightly lower than that of small-area devices, it still maintains a good level. These parameters indicate that the large-size device has the same detection capability as the conventional-size device, which also proves that the spray coating process is suitable for preparing large-area devices.

Conclusions
We exploited an efficient scalable spray coating process of CdSeTe QDs film for photodetectors. The mixed solvent of the QD solution was proposed to balance the solvent volatilization and the adhesive force of the QDs to the substrate so as to improve the quality of the thin film. When the number of cycles of the spray-coated quantum dots reached 40, the detector showed the best performance, with a response range of 300-800 nm, the responsivity of 15 mA/W, detectivity of more than 10 11 Jones and rise time of 0.04 s. The large size (10 × 10 cm 2 ) logo pattern QDs film prepared by spray coating displayed excellent uniformity of absorbance and thickness. In addition, it is found that the photodetector based on large-size QD film (the active area 1 cm 2 ) has the same detection performance compared to the device based on typical laboratory-size QDs film (the active area 0.1 cm 2 ), indicating that spray coating process has a broad application prospect for the realization of large-scale industrial production.
Supplementary Materials: The following are available online at www.mdpi.com/xxx/s1, Figure S1: (a) XRD pattern and (b) size distribution histogram of CdSeTe QDs, Figure S2: (a) The UV-vis absorption spectrum and

Conclusions
We exploited an efficient scalable spray coating process of CdSeTe QDs film for photodetectors. The mixed solvent of the QD solution was proposed to balance the solvent volatilization and the adhesive force of the QDs to the substrate so as to improve the quality of the thin film. When the number of cycles of the spray-coated quantum dots reached 40, the detector showed the best performance, with a response range of 300-800 nm, the responsivity of 15 mA/W, detectivity of more than 10 11 Jones and rise time of 0.04 s. The large size (10 × 10 cm 2 ) logo pattern QDs film prepared by spray coating displayed excellent uniformity of absorbance and thickness. In addition, it is found that the photodetector based on large-size QD film (the active area 1 cm 2 ) has the same detection performance compared to the