Effect of Interface Modification on Mechanoluminescence-Inorganic Perovskite Impact Sensors

It is becoming increasingly important to develop innovative self-powered, low-cost, and flexible sensors with the potential for structural health monitoring (SHM) applications. The mechanoluminescence (ML)-perovskite sensor is a potential candidate that combines the light-emitting principles of mechanoluminescence with the light-absorbing properties of perovskite materials. Continuous in-situ SHM with embedded sensors necessitates long-term stability. A highly stable cesium lead bromide photodetector with a carbon-based electrode and a zinc sulfide (ZnS): copper (Cu) ML layer was described in this article. The addition of a magnesium iodide (MgI2) interfacial modifier layer between the electron transport layer (ETL) and the Perovskite interface improved the sensor’s performance. Devices with the modified structure outperformed devices without the addition of MgI2 in terms of response time and impact-sensing applications.


Introduction
Structural health monitoring (SHM) aims to ensure the diagnostics of the host structure's conditions, as well as its safety and integrity, on a continuous basis. The advancement of SHM technologies is inextricably linked to the overall safety of structures. Real-time SHM systems for advanced composite structures are in high demand [1][2][3]. To meet the demand, a flexible mechanoluminescent (ML)-perovskite impact sensor is proposed [4][5][6][7][8].
The device consists of a photodiode constructed on top of an ML layer. ML materials emit light in response to mechanical stimuli. The active absorber layer of a photodiode is in charge of light absorption and charge transport in order to generate electric signals. The electrical signals can then be analyzed to determine the condition of the structure. In practice, the ML layer would respond to load or impact events by emitting light into the photodetector device, which would then convert the light into electric signals that could be conditioned and interpreted for health monitoring [4][5][6]. Due to their exceptional optoelectronic properties, such as high charge carrier mobility, low exciton binding energy, tunable bandgap, and long carrier diffusion lengths, perovskite-based photodetectors appeared to be the most promising candidate for the active layer. Organic-inorganic hybrid perovskite solar cells have a power conversion efficiency (PCE) that exceeds 25.2%, outperforming most polycrystalline silicon solar cells [9][10][11][12][13].
Unfortunately, the organic-inorganic lead halide perovskites decompose due to chemical instability and susceptibility to moisture, heat, oxygen, and other environmental conditions. To address the aforementioned instability, an all-inorganic perovskite active layer is an alternative and promising candidate. CsPbX 3 (X = Br, I) all-inorganic cesium-lead halide perovskites are gaining popularity due to their ability to increase a device's inherent stability by replacing volatile organic cations with an inorganic cesium cation (Cs + ) [14,15]. All inorganic CsPbBr 3 materials are promising for optoelectronics due to their excellent light absorption, high charge carrier mobility, and long carrier diffusion length. Aside

Sensor Device Fabrication
Hydrochloric acid and zinc powder were used to etch ITO-coated polyethylene terephthalate (PET) substrates. The substrates were then cleaned with nano pure water, acetone, and isopropanol. After cleaning, the flexible substrates were treated for 5 min with oxygen plasma. SnO2 colloid precursor was diluted (1:6) with deionized water and stirred overnight. The SnO2 solution was spin-coated for 30 s at 3000 rpm onto PET/ITO substrates before being annealed in ambient at 100 °C for 60 min, followed by 5 min of plasma treatment. MgI2 was dissolved in methanol in various concentrations and spincoated onto the PET/ITO/SnO2 substrates for 30 s at 3000 rpm before being annealed at 100 °C for 15 min. CsBr (0.3 M) and PbBr2 (0.3 M) were mixed in 1 mL DMSO to produce the CsPbBr3 precursor. The perovskite precursor was spin-coated on the substrates for 45 s at 1500 rpm before being annealed at 70 °C for 3 min and then at 105 °C for 20 min. Figure 1 depicts the one-step CsPbBr3 thin film fabrication procedure. Carbon paste electrodes were applied to the substrates with a doctor blade and dried for 15 min at 80 °C. The active area of the devices was set at 0.06 cm 2 . The device was built using our previously described method for ML layer integration [4,6]. Using a Thinky mixer, the ZnS:Cu material was mixed into a polydimethylsiloxane (PDMS) elastomer. The ZnS:Cu/PDMS composite was deposited on the other side of the PET substrate using a doctor blade and heated until a total cure was obtained.

Multifunctional Composite Fabrication
Multifunctional composites with embedded impact sensors were produced using plain weave carbon fiber fabric as reinforcing fibers and vinyl-ester resin as the system's matrix. As shown in Figure 2, the sensors were embedded between the third and fourth carbon fiber plies. The data from the sensors were collected using embedded copper electrodes. Six layers of reinforcing fibers were used to create the final composite. Vacuum-assisted resin transfer molding (VARTM) infusion was used as usual. The device was built using our previously described method for ML layer integration [4,6]. Using a Thinky mixer, the ZnS:Cu material was mixed into a polydimethylsiloxane (PDMS) elastomer. The ZnS:Cu/PDMS composite was deposited on the other side of the PET substrate using a doctor blade and heated until a total cure was obtained.

Multifunctional Composite Fabrication
Multifunctional composites with embedded impact sensors were produced using plain weave carbon fiber fabric as reinforcing fibers and vinyl-ester resin as the system's matrix. As shown in Figure 2, the sensors were embedded between the third and fourth carbon fiber plies. The data from the sensors were collected using embedded copper electrodes. Six layers of reinforcing fibers were used to create the final composite. Vacuum-assisted resin transfer molding (VARTM) infusion was used as usual.

Materials and Device Characterization
An Agilent Cary 5000 was used to obtain the ultraviolet-visible (UV-Vis) absorption spectra. The current-voltage (I-V) parameters were measured with a Keithley 2410 and LabView under a white light-emitting diode (LED) lamp with a 100 mW/cm 2 intensity. The time-dependent response was obtained with a NI-6210 DAQ and boosted with a Hamamatsu C7319 on a low bandwidth setting and 10 5 gain. The data were processed using MATLAB. The impact testing was carried out with the help of a customized droptower setup, and the cyclic 3-point bending test was carried out with the help of a   The time-dependent response was obtained with a NI-6210 DAQ and boosted with a Hamamatsu C7319 on a low bandwidth setting and 10 5 gain. The data were processed using MATLAB. The impact testing was carried out with the help of a customized droptower setup, and the cyclic 3-point bending test was carried out with the help of a Shimadzu mechanical testing system. The sensor's response signal was collected using the same configuration for I-V measurements.

Optical Characterization
MgI 2 solution ratios of 1, 5, and 10 mg·mL −1 were used to investigate the ETL/ perovskite interface modification. Mg 2+ and I − ions effectively inhibit the formation of deep trap states at the ETL/perovskite interface, promoting surface passivation and decreasing device carrier recombination. Mg 2+ ions can also diffuse into the interstitial regions of the perovskite lattice, resulting in an active passivation action [21,24,27]. The UV-Vis absorption technique was used to investigate the optical properties of CsPbBr 3 films with and without MgI 2 treatment. As shown in Figure 3a, all samples had a sharp absorption edge of around 530 nm. Because of the absorption property of the CsPbBr 3 films, they can be used as an active layer for visible photo-detection, particularly in the green light region. The MgI 2 layer improved the absorption of the perovskite films, implying that the interface between the ETL and the perovskite was effectively modified. Further research was carried out to determine the effect of the interfacial modifier on the crystal quality of the inorganic CsPbBr 3 perovskite. The crystallinity of the perovskite films was investigated using X-ray diffraction (XRD). Figure 3b depicts the XRD patterns of CsPbBr 3 perovskite films on PET substrates. The XRD pattern's break region attempts to remove the strong diffraction from the PET substrate. Because PET flexible substrates were used, the inorganic perovskite annealing temperature was kept constant at 105 • C. The intensities of the peaks increased as MgI 2 concentration increased, indicating that the films were more crystalline. The XRD pattern of CsPbBr 3 perovskite shows strong and prominent peaks with a high degree of crystallinity, which is beneficial for efficient charge transfer. Higher crystallinity, as well as larger grain size with fewer grain boundaries, are both indicated by a narrower and stronger X-ray diffraction peak, which is directly related to photovoltaic performance. Crystallinity was increased in thin films with 1 and 5 mg·mL −1 additions.  The crystallite size of the perovskite film was calculated using Scherrer's equation, shown in Equation (1)   The crystallite size of the perovskite film was calculated using Scherrer's equation, as shown in Equation (1) [42]. where D, K, λ, B, and θ are the crystallite size (nm), Scherrer constant, X-ray wavelength (nm), Full Width at Half Maximum (FWHM) (radian), and XRD peak position (degree), in that order. Table 1 shows the calculated crystallite size for films with various MgI 2 concentrations. The film containing 5 mg·mL −1 MgI 2 had an average crystallite size of 74.54 nm, whereas the pristine films had 55.35 nm. These findings also suggested that the interface modification improved crystallinity and reduced defects, thereby improving the photophysical capabilities of the device. SEM images of the obtained CsPbBr 3 films with and without the addition of MgI 2 at various concentrations were compared. The pristine CsPbBr 3 film has some discontinuities and numerous pinholes, as shown in Figure 4. The film morphology significantly improves as the MgI 2 concentration increases, with greater coverage uniformity and fewer minor pinholes. The thin film with a MgI 2 concentration of 5 mg·mL −1 appears to have the best film coverage, with very compact grain boundaries and few pinholes. The films with a concentration of 10 mg·mL −1 showed signs of degradation, including poor film coverage and dissolved grain boundaries. The degradation of the films with a higher concentration of MgI 2 may indicate saturation of the Mg layer, which prevents the perovskite layer from crystallizing properly. The disproportional addition of the compound can disbalance the perovskite crystallization, leading to an incomplete reaction and the formation of PbI 2 , preventing the full development of perovskite crystals on the film [23,[43][44][45]. The SEM results agree with the UV-Vis and XRD characterization, indicating that adding MgI 2 can help improve the morphology and crystallization of the CsPbBr 3 film.

Electrical Characterization
A vertical planar PET/ITO/SnO2/MgI2/CsPbBr3/carbon structure was used to investigate the photo-detection response of the CsPbBr3 photodiode. When light is absorbed by the perovskite layer, electron and hole pairs are separated. Electrons from the light-absorbing layer of the perovskite are carried into the conduction band of the perovskite, where they are injected into the SnO2 electron transport layer (ETL) and collected by the ITO electrode. To complete the electrical circuit, the carbon contact electrode collects the holes. Under a 0 V bias, the devices were tested with a pulsing white light LED at 1 Hz. Figure 5 depicts the devices' consistent and stable response over time (a-d). The device's response to the LED is consistent with the UV-vis and XRD data, which show that adding 5 mg·mL −1 of MgI2 improves sensor performance (Figure 5d).

Electrical Characterization
A vertical planar PET/ITO/SnO 2 /MgI 2 /CsPbBr 3 /carbon structure was used to investigate the photo-detection response of the CsPbBr 3 photodiode. When light is absorbed by the perovskite layer, electron and hole pairs are separated. Electrons from the lightabsorbing layer of the perovskite are carried into the conduction band of the perovskite, where they are injected into the SnO 2 electron transport layer (ETL) and collected by the ITO electrode. To complete the electrical circuit, the carbon contact electrode collects the holes. Under a 0 V bias, the devices were tested with a pulsing white light LED at 1 Hz. Figure 5 depicts the devices' consistent and stable response over time (a-d). The device's The response speed of a photodiode is an important metric that reflects its ability to detect a rapidly changing optical signal [10,11,46,47]. High-performance photodiodes used in SHM must respond quickly and consistently to light illumination. The response time was evaluated using the previously described method. Under ambient conditions, a 470 nm pulse light source from an LED was controlled by a function generator with square waves at a frequency of 50 Hz to measure response time [11,46,47]. The reaction times of pristine CsPbBr3 photodiodes and devices incorporating 5 mg·mL −1 MgI2 are shown in Figure 6. The devices demonstrated a consistent and repeatable response with excellent performance. Figure 6 depicts the response rise (τrise) and fall (τfall) times, which are the times it takes for a photodiode to reach 90% and 10% of steady-state values, respectively (a, b). The device containing 5 mg·mL −1 MgI2 has a rise and fall time of 0.65 ms and 0.69 ms (Figure 6a), which is comparable to previously reported devices [47][48][49][50] but significantly faster than the pristine devices (Figure 6b). Table 2   The response speed of a photodiode is an important metric that reflects its ability to detect a rapidly changing optical signal [10,11,46,47]. High-performance photodiodes used in SHM must respond quickly and consistently to light illumination. The response time was evaluated using the previously described method. Under ambient conditions, a 470 nm pulse light source from an LED was controlled by a function generator with square waves at a frequency of 50 Hz to measure response time [11,46,47]. The reaction times of pristine CsPbBr 3 photodiodes and devices incorporating 5 mg·mL −1 MgI 2 are shown in Figure 6. The devices demonstrated a consistent and repeatable response with excellent performance. Figure 6 depicts the response rise (τ rise ) and fall (τ fall ) times, which are the times it takes for a photodiode to reach 90% and 10% of steady-state values, respectively (a, b). The device containing 5 mg·mL −1 MgI 2 has a rise and fall time of 0.65 ms and 0.69 ms (Figure 6a), which is comparable to previously reported devices [47][48][49][50] but significantly faster than the pristine devices (Figure 6b).

Impact Sensing
A perovskite photodiode with a MgI2 layer (5 mg·mL −1 ) was combined with ZnS:Cu-PDMS and mechanically tested to evaluate the possibility of the modified CsPbBr3 devices for constructing ML-perovskite impact sensors. Light emission in ML materials is caused by mechanical stress. As a result, photon emission is predicted as the ML layer is stressed. The perovskite photodiode then captures the photons and converts them into an electrical current. Upon photoexcitation, electron-hole pairs are generated in the perovskite layer. Then, photogenerated carrier pairs, in the presence of the inherent electric field, are extracted and collected by the electron-transport layer and electrodes, generating electrical current [39,40,58]. The current fluctuation can be measured and correlated to the load applied to the device for sensing purposes. The change in the electrical current was expressed as follows: where I represents the electrical current measured during the test and I0 represents the sensor's baseline electrical current, also known as the dark current. The mechanical energy applied to the composite is transmitted to the sensor, causing ML emission. The perovskite layer absorbs light and converts it into an electrical current. The intensity of an impact has a linear effect on the sensor signal [4,6]. A mechanical three-point bending test was performed to evaluate the capability of the modified ML-Perovskite sensor for SHM applications. Figure 7 depicts the three-point bending test configuration (a). The sensor was inserted into a carbon fiber composite sample and bent 250 times. Each bending cycle used a constant displacement of 1.5 mm.

Impact Sensing
A perovskite photodiode with a MgI 2 layer (5 mg·mL −1 ) was combined with ZnS:Cu-PDMS and mechanically tested to evaluate the possibility of the modified CsPbBr 3 devices for constructing ML-perovskite impact sensors. Light emission in ML materials is caused by mechanical stress. As a result, photon emission is predicted as the ML layer is stressed. The perovskite photodiode then captures the photons and converts them into an electrical current. Upon photoexcitation, electron-hole pairs are generated in the perovskite layer. Then, photogenerated carrier pairs, in the presence of the inherent electric field, are extracted and collected by the electron-transport layer and electrodes, generating electrical current [39,40,58]. The current fluctuation can be measured and correlated to the load applied to the device for sensing purposes. The change in the electrical current was expressed as follows: where I represents the electrical current measured during the test and I 0 represents the sensor's baseline electrical current, also known as the dark current. The mechanical energy applied to the composite is transmitted to the sensor, causing ML emission. The perovskite layer absorbs light and converts it into an electrical current. The intensity of an impact has a linear effect on the sensor signal [4,6]. A mechanical three-point bending test was performed to evaluate the capability of the modified ML-Perovskite sensor for SHM applications. Figure 7 depicts the three-point bending test configuration (a). The sensor was inserted into a carbon fiber composite sample and bent 250 times. Each bending cycle used a constant displacement of 1.5 mm. The results of a 15-cycle repeated bending test are shown in Figure 7b. The sensor generated distinct and visible signals for each cycle. Furthermore, the sensor intensity signals correlated well with the composite sample displacement. The sensor responded consistently across the cycles, indicating its potential for SHM applications.
Sensors 2023, 23, x FOR PEER REVIEW 9 The results of a 15-cycle repeated bending test are shown in Figure 7b. The se generated distinct and visible signals for each cycle. Furthermore, the sensor inten signals correlated well with the composite sample displacement. The sensor respon consistently across the cycles, indicating its potential for SHM applications. To investigate the sensor's potential for impact sensing, an in-house drop to testing setup was used to impact a composite laminate with an embedded sensor varying impact energies ranging from 0.4 J to 4 J. To evaluate the sensor's low-en impact-sensing capabilities, impact samples were subjected to ten successive impac increasing energies. The strikes were delivered directly to the surface of the comp structure, right on top of the sensor. The impact testing setup is depicted in Figur Drop-weight tower impact testing is a popular and widely used method for investiga low-velocity impact in composites and ML materials. The impact energy can be altere adding mass to the impactor or adjusting the drop height. For each impact, all de configurations produced distinct signals. As shown in Figure 7d, the signal peak inten is proportional to the impact energy.
The minimum observable signal was obtained for an impact energy of 0.4 J. minimal detectable impact energy could be reduced by improving the dark current o To investigate the sensor's potential for impact sensing, an in-house drop tower testing setup was used to impact a composite laminate with an embedded sensor with varying impact energies ranging from 0.4 J to 4 J. To evaluate the sensor's low-energy impactsensing capabilities, impact samples were subjected to ten successive impacts of increasing energies. The strikes were delivered directly to the surface of the composite structure, right on top of the sensor. The impact testing setup is depicted in Figure 7c. Drop-weight tower impact testing is a popular and widely used method for investigating low-velocity impact in composites and ML materials. The impact energy can be altered by adding mass to the impactor or adjusting the drop height. For each impact, all device configurations produced distinct signals. As shown in Figure 7d, the signal peak intensity is proportional to the impact energy.
The minimum observable signal was obtained for an impact energy of 0.4 J. The minimal detectable impact energy could be reduced by improving the dark current of the perovskite photodetector. Table 3 summarizes some performance parameters of different ML-perovskite sensors. A regression model was used to analyze the experimental data and is shown in the following equation: where E is the impact energy value. The sensor's sensitivity could be estimated as the change of sensor output due to the input parameter change (impact energy). The estimated sensitivity for the optimized sensor is 0.74 J −1 . The load applied to the ML materials produces a linear relationship between pressure and light emission. The higher the energy applied to the ML crystals, the higher the expected emission of photons. The photocurrent of a perovskite photodetector increases linearly with increasing light intensity. As a result, the system's output signal should rise linearly as the applied load increases. In other words, as the impact energy increases, the signal intensity increases, demonstrating that the sensor can be used for structural health monitoring of composite structures. The addition of MgI 2 to the interface can increase the intensity of the resulting signal.

Conclusions
In conclusion, we discuss how an inorganic CsPbBr 3 perovskite with a carbon-based electrode flexible photodetector can be utilized for the ML sensing and SHM of composite structures. The overall performance of the sensor was improved by the incorporation of MgI 2 into the interfacial layer that is located between SnO 2 and perovskite. The UV-vis data demonstrate that the device's absorption spectrum has been enhanced, and as a result, it is now appropriate for monitoring ML emissions. The optimized flexible photodetector with a PET/ITO/SnO 2 /MgI 2 /CsPbBr 3 /carbon basic construction demonstrates excellent photoresponse when exposed to white light as the illumination source. According to the findings, the sensors could detect various loads throughout the composite structure. This enabled a correlation to be established between the sensor signal and the impact distance or composite displacement. The ML-perovskite sensor has demonstrated strong potential in SHM applications involving composite materials.