Delayed stress memory by CaAl2O4:Eu2+ mechanoluminescent phosphor with defect engineering regulation

Real-time stress sensing based on mechanoluminescence materials has been widely studied for structural health monitoring of bridges, buildings, high-pressure vessels, and other infrastructure surfaces. However, this approach is difficult to detect the stress information of closed mechanical structures. Here, we propose a delayed stress memory strategy to record the stress information of closed mechanical structure by the flexible film composed with CaAl2O4:Eu2+,Sm3+ phosphor. After the force is applied, the optical information on the film can be read out by the near-infrared laser after a period of time without real-time monitoring, and the stress distribution information of bearings and gears in the engine can be obtained. Furthermore, the regulation of trap depth from 0.662 to 1.042 eV allows the captured carriers to remain in the traps for a long time without being released as long persistent luminescence, which is beneficial to the delayed stress memory. Therefore, this work promotes the application prospect of mechanoluminescence materials in stress sensing, and provides a new idea to record the stress information of closed mechanical structures.


Introduction 
Mechanoluminescence (ML) is the phenomenon that † Yiyu Cai and Sibo Liu contributed equally to this work. destroys ML material, ML can be divided into two categories: destructive ML including fractoluminescence (FML) and non-destructive ML including elasticoluminescence (EML) and triboluminescence (TML) [3,12]. Compared with FML and TML, EML has drawn extensive attention in stress detection and visualizing stress distribution due to the advantages of recoverable emission and the linear relation between ML intensity and the magnitude of the loading force. In particular, Xu et al. firstly reported the EML from ZnS:Mn [13] and SrAl 2 O 4 :Eu 2+ [14], and examined the structural damage of the bridge by attaching several ML films embedded with SrAl 2 O 4 :Eu 2+ particles to the cracks being visible on a concrete girder of the bridge [8]. Klein et al. [15,16] also proposed to determine the stress field by coating the aircraft surface with ML sensor materials, and the sensor materials generate a sensitive spectral response to the applied pressure in a wind tunnel. Subsequently, a large number of EML materials have also been developed as depicted in Table 1, which exhibits distinctive ML performance of intense brightness, high mechanical sensitivity, abundant emission wavelength, and other properties. Although stress sensing based on these ML materials has realized remote, real-time, and large-area monitoring of stress information for macroscopic infrastructure surfaces, it is still difficult to detect the stress information of closed mechanical structures, and it should be noticed that the ML properties are always accompanied with a long persistent luminescence (LPL) behavior.
Typically, the excited carriers are partially captured by the trap centers in the ML material once exposed to ultraviolet or visible light. These captured carriers are released under appropriate mechanical stimuli and eventually participate in the light emission process, producing the ML [3,10]. Nevertheless, the carriers captured by the trap centers would be thermally unstable if the trap distribution in the matrix is relative shallower, and then the de-trapping process occurs under thermal disturbance at room temperature, eventually resulting in LPL [53], which is fundamentally detrimental to the ML output. Herein, Rahimi et al. [54] presented an approach of increasing the delay time to reduce the LPL intensity and improve the signal-to-noise ratio (SNR) of ML accordingly. The ML material firstly was exposed to the excitation light source and then deposited in the dark for some time before applying the mechanical stimulation, which endows the response to the stress information more effectively. Recently, Zhang et al. [33,55] also proposed a strategy of sacrificing the trap density to achieve high SNR of ML materials for stress imaging. Although these results provide a possibility for obtaining high contrast ML performance, the ML behavior is inevitable to be impacted by increasing the delay time or sacrificing trap density. Therefore, it is of great importance for the ML materials to achieve  [51,52] www.springer.com/journal/40145 intense stress information free from the LPL interference, thereby improving the SNR to satisfy the requirements of advanced practical applications.
In this work, we regulate the trap depth from 0.662 to 1.042 eV by introducing foreign Ln 3+ ions in CaAl 2 O 4 :Eu 2+ (CAO:Eu 2+ ), which allows the carriers in the regulated deep traps to remain for a long time and hardly be affected by the thermal disturbance at room temperature. Furthermore, it is also found that, the carriers captured by the deep traps not only can be released by the mechanical stimulation for achieving ML but also be induced by infrared (980 nm) irradiation to produce the photo-stimulated luminescence (PSL). Therefore, we propose a delayed stress memory strategy to record the stress information of closed mechanical structures. The results indicate that the stress distribution information of bearings and gears in the engine can be obtained by recording the PSL intensity difference of the flexible films at the positions with and without the mechanical stimulation after applying force. This non-real-time stress sensing opens a new path for detecting and recording the stress information of closed mechanical structures. . After cooling to room temperature, the ML phosphors were ground for further use.

2 Fabrication of ML composite films
To quantitatively analyze the ML properties, the synthesized phosphors were ground and sieved (300 mesh/in). The phosphors were mixed with polydimethylsiloxane (PDMS, Sylgard 184, Dow Corning) in a weight ratio of 2 : 1. The mixture was poured into a rectangle mold (45 mm × 25 mm × 1 mm) and then put in a vacuum drying oven at 60 ℃ for 4 h. After the curing process, an ML film with an average thickness of 1-2 mm was obtained.

3 Characterizations
A D8 ADVANCE/Germany Bruker X-ray diffractometer (XRD) with Cu Kα radiation (λ = 1.5405 Å) operating at 40 kV and 30 mA was employed to characterize the crystalline phases of the obtained powders, and the raw XRD data were further analyzed by the Rietveld refinement method (GSAS software package). The surface morphology and elemental mapping were characterized by a field-emission SEM instrument (SU8010, Hitachi) with an energy dispersive X-ray (EDX) spectroscope system. Photoluminescence (PL) and photoluminescence excitation (PLE) spectra were recorded by a Hitachi F7000 fluorescence spectrophotometer with an excitation source of 150 W Xe lamp. The thermoluminescence (TL) curves were measured with an FJ-427A TL meter (Beijing Nuclear Instrument Factory). The weight of the measured powders was constant (0.002 g). Before the TL measurement, the samples were firstly exposed to the radiation from sunlight for about 15 min, and then heated from room temperature to 550 K at a rate of 60 K/min. The ML intensity of the films was investigated with a lab-built system that comprises a universal testing machine (AG-X, Shimadzu) (Fig. S1 in the Electronic Supplementary Material (ESM)) and a photo-counting system, which consisted of a photomultiplier tube (55777, Oriel Instruments) and a photo-counter (C9692, Hamamatsu). A fiber spectrometer with a wavelength range from 200 to 1000 nm (QE Pro, Ocean Optics) was applied to record the PSL and ML spectra. All optical photos were taken by a digital camera (D7100, Nikon).  1(c) display the Rietveld refinement of XRD patterns of CAO:Eu 2+ ,Dy 3+ and CAO:Eu 2+ ,Sm 3+ , respectively. The refinement parameter values are calculated to be R p = 6.1%, R wp = 7.82% of CAO:Eu 2+ ,Dy 3+ , and R p = 6.66%, R wp = 8.89% of CAO:Eu 2+ ,Sm 3+ , which further confirms the pure phase of the as-obtained samples. According to the crystallographic data, CAO belongs to the monoclinic system with a space group of P21/n (14), and there are three different sites of Ca 2+ ions. Two sites of Ca1 and Ca2 are coordinated with six oxygen ions having different bond lengths, whereas the third site of Ca3 is coordinated with nine oxygen ions [55][56][57]. All calcium ions array in a line along the b-axis, and Ca3 ions align more regularly compared with Ca1 and Ca2 ( Fig. 1(d)). Considering the ionic radii and valence, the Ca3 (r = 1.00 Å, coordination number (CN) = 6, and r = 1.18 Å, CN = 9) sites are preferentially replaced by Eu 2+ (r = 1.17 Å), Dy 3+ (r = 0.91 Å), Tm 3+ (r = 0.88 Å), Nd 3+ (r = 0.98 Å), Y 3+ (r = 0.88 Å), and Sm 3+ (r = 0.95 Å) ions [58]. The crystal morphology and X-ray energy dispersive spectrum (EDS) analysis with elemental mappings (Figs. 1(e)-1(g) and Fig. S2 in the ESM) illustrates that all elements are homogeneously distributed within the corresponding particle, and further confirms the successful incorporation of Eu 2+ and Ln 3+ ions.

Results and discussion
As illustrated in Fig. 2(a), the as-prepared CAO:Eu 2+ phosphor is further composited with the transparent PDMS matrix to form a CAO:Eu 2+ /PDMS composite film with a rectangle shape. Since the CAO:Eu 2+ -based materials are the typical long-persistent luminescence systems, the CAO:Eu 2+ /PDMS composite film firstly is exposed to the excitation light source, and then deposited in the dark for some time to eliminate the interference of LPL before applying the mechanical stimulation.
When mechanical stimuli such as stretching, rubbing, and folding are applied, intense blue light emission of CAO:Eu 2+ /PDMS film is observed with naked eyes, which proves the ML behavior originated from the corresponding film ( Fig. 2(b)). Figure 2(c) presents the PL, LPL, and ML spectra of CAO:Eu 2+ /PDMS film. The ML spectrum is similar to the PL and LPL spectra, showing a broadband ranging from 410 to 500 nm with a maximum emission peak located at 440 nm ascribed to the transition from 4f 6 5d to 4f 7 level of Eu 2+ ions [59][60][61]. Figure 2(d) shows the ML response of CAO:Eu 2+ /PDMS film under the stretching force, and the peak intensity of ML has a significant decay under the consecutive stretching-releasing cycles, which is attributed to the gradual release of carriers in the trap centers. Moreover, the ML and TL intensities of film for sunlight filling and thermal bleaching after 5 cycles are recorded, as shown in Fig. 2(e). The results show that there is obvious ML after the trap centers are filled with carriers, while the ML disappears after the carriers in the trap centers are thermally released (at 200 ℃ for 30 min), which further proves that the ML behavior is derived from the release of carriers in the trap centers under the mechanical stimulation. Furthermore, the intensity of ML gradually decreases with the increase of the number of cycles, for the PDMS matrix is inclined to be damaged during the process of thermal treatment and hardly to completely recovered. Generally, the concentration and distribution of trap centers have been identified as critical factors affecting the ML performance [62,63]. Hence, the ML performance of CAO:0.0025Eu 2+ ,yDy 3+ /PDMS film (y = 0.0005, 0.001, 0.00125, 0.0025, and 0.0075) is evaluated to study the effect of trap concentration on the ML properties. Figure 3(a) gives the ML spectra of CAO:0.0025Eu 2+ ,yDy 3+ /PDMS film, which has similar broadbands, indicating the same emitting center for those films and the introduced Dy 3+ ions act as the trap centers. Figure 3(b) shows the change of the ML intensity with the increasing dopants, which reaches the maximum at y = 0.0025 and then decreases. In addition, as shown in Fig. 3(c), the TL curves of CAO:0.0025Eu 2+ ,yDy 3+ /PDMS film are recorded, and the TL intensity also reaches the maximum value at y = 0.0025, which is consistent with the behavior of ML in Figs. 3(a) and 3(b). The results indicate that the intensity of ML is positively correlated with the concentration of trap centers. Figure 3(d) displays the ML response of CAO:0.0025Eu 2+ ,yDy 3+ /PDMS film under the consecutive stretching-releasing cycles, and the intensity of ML exhibits a significant decay with the increase of the number of cycles, which is caused by the gradual release of carriers in the trap centers under the mechanical stimulation. However, it is noted that, the ML intensity can be fundamentally recovered after the sunlight irradiation, as shown in Fig. 3(e), indicating the excellent reproducibility of ML in CAO:0.0025Eu 2+ ,0.0025Dy 3+ /PDMS film. Figure 3(f) plots the dependence of ML intensity on the stretching forces of CAO:0.0025Eu 2+ ,yDy 3+ /PDMS film, which demonstrates that the ML intensity increases monotonously with the increase of the stretching force. Moreover, the typical ML images of CAO:0.0025Eu 2+ , yDy 3+ /PDMS film under different stretching forces are shown in Fig. S4 in the ESM.
To further verify the reliability of the above conclusions, the trap concentration is also adjusted by introducing different Ln 3+ ions (i.e., Dy 3+ , Tm 3+ , and Nd 3+ ) into CAO:Eu 2+ Fig. S5(a) in the ESM. The trap depth does not change in a certain range, while the trap concentration changes greatly, wherein the trap concentration of CAO:Eu 2+ ,Dy 3+ /PDMS film is the highest. Figure S5( S5(c) in the ESM). Therefore, the ML intensity is confirmed to be linearly correlated with the trap concentration.
The effect of the distribution of trap on the ML properties is further clarified. Figure S6 in the ESM shows the normalized TL curves of Ln 3+ (Ln = Dy, Tm, Nd, Y, and Sm) co-doped in CAO:Eu 2+ . The trap depth corresponding to the TL peak is estimated with the following approximate in Eq. (1) [64]: where E T is the calculated trap depth, and T m is the maximum peak temperature of the TL curve (in Kelvin, K). Therefore, the adjustment of trap depth from 0.662 to 1.042 eV is achieved for the foreign introduced Ln 3+ ions. Figure S7 in (Fig. 4(d)), which is beneficial for the accurate stress information detection of the closed mechanical structures. The results show that the high SNR of ML performance is achieved by the defect engineering regulation. Therefore, we proposed and illustrated the mechanism for the ML as shown in Fig. 4(e). Firstly, under the irradiation of sunlight, the electrons are excited from the valence band (VB) to the conduction band (CB), and then these electrons are partially captured by the trap levels. When the mechanical stimulation is applied, the trapped electrons can be released and come back to the 4f 6 5d level of Eu 2+ . Finally, the electrons return to the ground state and emit blue emission (ML). However, when the trap depth is shallower (i.e., CAO:Eu 2+ ,Y 3+ and CAO:Eu 2+ ,Dy 3+ ), thermally-induced de-trapping can also occur at room temperature, resulting in the Based on the above results, a delayed stress memory method is proposed to record the stress distribution information for the closed mechanical structures by CAO:Eu 2+ ,Sm 3+ /PDMS film. (i) and (ii) in Fig. 5(a) show photographs of the engine and the bearing of its internal parts, respectively. Firstly, the prepared flexible CAO:Eu 2+ ,Sm 3+ /PDMS film is exposed to sunlight radiation for about 15 min, and then deposited inside the engine bearing and applied mechanical stimulation ((iii) in Fig. 5(a)). Figure 5 and the stripe between light and dark can be clearly seen, which indirectly shows the stress information inside the engine bearing. Figure 5(d) plots the schematic diagram of a homemade microscopy setup to excite and accurately measure the PSL emission characteristics of each point in the CAO:Eu 2+ ,Sm 3+ /PDMS film before and after the mechanical stimulation. A 980 nm excitation laser is focused onto a surface of the film through a 50× objective lens. Light-emitting from the film surface is collected through the same objective lens, and the collected light is then coupled to either a spectrum analyzer or a CCD camera. Figure 5(e) shows PSL images of one point of the CAO:Eu 2+ ,Sm 3+ /PDMS film before the mechanical stimulation and PSL images of representative position ⑥ marked in Fig. 5(c) of CAO:Eu 2+ ,Sm 3+ /PDMS film after the mechanical stimulation. Also, the corresponding PSL spectra are shown in Fig. 5(f). Furthermore, the PSL images of positions ①, ②, ③, ④, and ⑤ marked in Fig. 5(c) of CAO:Eu 2+ ,Sm 3+ /PDMS film after the mechanical stimulation are displayed in Fig. S9 in the ESM.
As depicted in Fig. S10 in the ESM, the positions of PSL peaks corresponding to different positions are consistent, but the PSL intensities corresponding to the light and dark fringes in Fig. 5(c) are different. Since the number of carriers stored in the trap centers is constant, a different method of subtracting the PSL intensity of the CAO:Eu 2+ ,Sm 3+ /PDMS film after the mechanical stimulation from the PSL intensity before the mechanical stimulation (ΔI = Before − After) is proposed to quantitatively calculate the stress distribution information inside the engine bearing. Figure 5(g) provides the PSL intensity and the corresponding difference of the CAO:Eu 2+ ,Sm 3+ /PDMS film at different positions marked in Fig. 5(c) before and after the mechanical stimuli. Therefore, the stress information recording of the closed mechanical structures by the ML materials with the defect engineering regulation is successfully realized. In addition, the recording and readout of a stress signal can be repeated after 980 nm irradiation as long as the film is not damaged (Fig.  5(h)), and work effectively in a variety of environments, such as aqueous solutions and machinery oil (Fig. S11 in the ESM).

Conclusions
In conclusion, a delayed stress memory strategy was proposed to record the stress distribution information of the closed mechanical structure by recording the PSL intensity difference of the flexible CAO:Eu 2+ ,Sm 3+ / PDMS film at the positions with and without the mechanical stimulation after a period of time. It was successfully applied to bearings and gears in the engine, which presents the accurate stress distribution information with long time storage. Moreover, the concentration and distribution of traps were adjusted by introducing different Ln 3+ ions (i.e., Dy 3+ , Tm 3+ , Nd 3+ , Y 3+ , and Sm 3+ ). The results show that the adjustment of trap depth greatly realizes the stable preserved energy of the captured carriers and improves accuracy of ML information.