1. Introduction

Mechanical-responsive luminescent materials have attracted special attention as sensing system owing to the remote observability of light emission [1–7]. One of the typical interests is trap-controlled mechanoluminescence (TC-ML), which is a form of reproducible, quantitative mechano-optical conversion resulting from trap-mediated storage and release of excitation energy [8–10]. This kind of reproducible ML enables ones to directly observe the stress concentration of load-bearing objects through imaging the distribution of ML, showing broad prospects in fields as diverse as stress sensors, structural healthcare, bio-imaging and lighting displays [11–17]. The first report of the robust TC-ML dates back to 1999 when Chao-Nan Xu and associates discovered the intense, reproducible elasticoluminescence from ZnS:Mn²⁺ and SrAl₂O₄:Eu²⁺ [18,19]. Xu’s studies are driving a renewed interest in ML research, and are creating opportunities to develop new and improved TC-ML materials to solve real-world problems. It is known that TC-ML materials developed during the past 20-years are universally activated by the doped luminescent ions, such as ZnS:Mn²⁺/Cu/Cu,Mn²⁺ [18,20], SrAl₂O₄:Eu²⁺/Ce³⁺/Eu²⁺,Dy³⁺/Eu²⁺,Er³⁺ [19,21,22], CaZnOS:Mn²⁺/Cu/Er³⁺/Sm³⁺ [23–26], ZrO₂:Ti⁴⁺ [27], and CaZr(PO₄)₂:Eu²⁺ [28]. Among the developed ML materials, there exists a series of isomorphic materials which show a similarity of crystal form including the same chemical formulation and space group, and containing similar atoms which have corresponding chemical properties and similar sizes. The typical instances are Eu²⁺-activated CaAl₂Si₂O₈ and SrAl₂Si₂O₈ (space group P1) [29], Eu²⁺-activated Sr₂MgSi₂O₇ and Ca₂MgSi₂O₇ (space group P$\overline{4}$2₁m) [30], and Mn²⁺-activated ZnGa₂O₄, MgGa₂O₄ and ZnAl₂O₄ (space group Fd$\overline{3}$m) [31,32]. Interestingly, these isomorphic materials activated by the same luminescent ions show differentiated ML properties (e.g. intensity and wavelength) due to the slight discrepancy in unit cell dimensions, thereby supplying selectivity for diverse application needs. This clue hints us to develop new TC-ML materials and regulate their characteristics via constructing isomorphic crystals of the existing TC-ML materials.

Recently, we created a series of novel TC-ML materials through doping Pr³⁺ activators to introduce luminescent centers and carrier traps in piezoelectric calcium niobates [8]. We showed that Ca₂Nb₂O₇:Pr³⁺ could be used as a sensitive optical probe for real-time observation of dynamic stress concentration because of the intense emission and high reproducibility of ML. However, Ca₂Nb₂O₇:Pr³⁺ emits a bright long-lasting afterglow after UV excitation. Such strong afterglow noise would mask the ML signal, reducing the sharpness of ML images. In order to achieve high-contrast ML images, Ca₂Nb₂O₇:Pr³⁺ requires a long delay time to fade the strong noise of afterglow. Obviously, long-time delay complicates the imaging process, thereby proposing the requirement of reducing delay time. Inspired by the isomorphic strategy, in this work, we report a new TC-ML material Sr₂Nb₂O₇:Pr³⁺ displaying intense ML under the mechanical actions of compression and friction. More importantly, Sr₂Nb₂O₇:Pr³⁺ shows a weak and short-lasting afterglow and enables us to achieve high-contrast ML in a short-time delay, which greatly simplifies the imaging process to meet application needs. Furthermore, we investigate the unit cell dimensions, trap distribution and optical bandgap of Sr₂Nb₂O₇ and Ca₂Nb₂O₇, and finally propose a possible mechanism underlying their different properties of afterglow and ML.

2. Experimental

2.1 Material synthesis

Pr³⁺-activated strontium niobate and calcium niobate with compositions of Sr_2-xPr_xNb₂O₇ (0 ≤ x ≤ 0.03, abbreviated as Sr₂Nb₂O₇:Pr³⁺ (0-3 mol%)) and Ca_2-xPr_xNb₂O₇ (0 ≤ x ≤ 0.02, abbreviated as Ca₂Nb₂O₇:Pr³⁺ (0-2 mol%)) were synthesized by solid-state reaction among CaCO₃ (Aladdin, 99.99%), SrCO₃ (Aladdin, 99.99%), Nb₂O₅ (Sinopharm, 99.99%), and Pr₆O₁₁ (Alfa Aesar, 99.99%). Stoichiometric raw materials were ground and pre-calcined at 900 °C for 4 h in air, and re-ground. The powders were finally sintered at 1400 °C for 4 h in air, and pulverized after natural cooling.

2.2 Physical characterization

Powder X-ray diffraction (XRD) patterns were collected on a Rigaku SmartLab X-ray diffractometer (Cu Kα, λ = 1.54056 Å) with a continuous scanning rate of 5° min⁻¹ for phase identification and a step scanning rate of 8 s step⁻¹ (step size: 0.01°) for Rietveld analysis. The refinement of crystal structure was performed using the General Structure Analysis System program. Diffuse reflectance spectra were recorded by a UV/vis/NIR spectrophotometer (V570, Jasco). Spectra and decay curves of afterglow were measured using a fluorescence spectrometer (F4600, Hitachi). Thermoluminescence (ThL) was measured using a ThL meter (FJ427A1, Beijing Nuclear Instrument Factory). To characterize ML triggered by compression and friction, cylinder-shaped composite disks (25 mm in diameter and 15 mm in thickness) were prepared by mixing the as-synthesized particles with an optical epoxy resin. Compressive loading (0-1000 N) at a deformation rate of 3 min min⁻¹ was applied on the composite disks using a lab-built electronic universal testing machine. Mechanical friction was applied by an in-house made friction machine with a metal tip of 1 mm diameter (load of 10 N, friction rate of 120 round min⁻¹). Afterglow decay and ML response of composite disks were recorded using a lab-made photon-counting system. Spectra of ML were captured with a photon multi-channel analyser system (QE65000, Ocean Optics). Optical photographs of afterglow and ML were recorded using a Canon 7D camera. The imaging parameters are manual/ISO 6400/0.5 s. Before the measurements of afterglow, ML and ThL, if not otherwise specified, the samples were irradiated by a hand-held UV lamp (8 W, 308 nm) for 1 min to charge traps. All measurements except ThL were performed at room temperature.

3. Results and discussion

The XRD results confirm that the as-synthesized Sr₂Nb₂O₇:Pr³⁺ and Ca₂Nb₂O₇:Pr³⁺ both crystallize in phase-pure orthorhombic structure with space group Pna2₁ [33,34]. Figures 1(a) and 1(b) show the Rietveld refined XRD patterns of Sr₂Nb₂O₇ and Ca₂Nb₂O₇ host lattices, respectively. In comparison with Ca₂Nb₂O₇, Sr₂Nb₂O₇ shows a lattice expansion due to larger radius of Sr²⁺ than Ca²⁺. The refined lattice parameters are a = 26.7803 Å, b = 7.9101 Å, c = 5.7015 Å for Sr₂Nb₂O₇, and a = 26.4559 Å, b = 7.6901 Å, c = 5.4964 Å for Ca₂Nb₂O₇, as summarized in Table 1. Figures 1(c) and 1(d) show the crystal structures and Sr/Ca-O coordination environments of Sr₂Nb₂O₇ and Ca₂Nb₂O₇, respectively. Sr₂Nb₂O₇ and Ca₂Nb₂O₇ have extremely similar perovskite-derivative structures, in which the perovskite slabs consisting of vertex sharing NbO₆ octahedra are separated by additional O atoms. Sr²⁺ and Ca²⁺ ions both possess two coordination numbers (CN = 8 and 12), but the average bond distances of Sr-O (2.7913 Å for CN = 8 and 2.9069 Å for CN = 12) are longer than those of Ca-O (2.6394 Å for CN = 8 and 2.7275 Å for CN = 12). The average bond distance of Nb-O (2.085 Å) in Sr₂Nb₂O₇ is also a little longer than that (2.000 Å) in Ca₂Nb₂O₇. The differences in these structural parameters together result in Ca₂Nb₂O₇ exhibiting a more distorted crystal structure than Sr₂Nb₂O₇ [Figs. 1(c) and 1(d)].

 

Fig. 1 Rietveld refinement of XRD profiles, crystal structures and Sr/Ca-O coordination environment (including bond distances): (a) and (c) for Sr₂Nb₂O₇; (b) and (d) for Ca₂Nb₂O₇.

Download Full Size | PDF

Table 1. Refinement, crystallographic, and structural parameters of Sr₂Nb₂O₇ and Ca₂Nb₂O₇

View Table

Figure 2 shows the diffuse reflectance spectra of Sr₂Nb₂O₇ and Ca₂Nb₂O₇. According to the Kubelka-Munk function [35] and the Tauc relation [36], the optical band gaps are estimated [insets of Figs. 2(a) and (b)], i.e., 3.97 eV for Sr₂Nb₂O₇ and 4.17 eV for Ca₂Nb₂O₇, which are really close to the previously reported values of 3.92 eV [37] and 4.18 eV [38], respectively. After the introduction of Pr³⁺ ions, the absorption edges of both compounds show an insignificant shift, indicating a negligible change of bandgaps. It indicates that the bandgap of Sr₂Nb₂O₇:Pr³⁺ is ca. 0.2 eV narrower than that of Ca₂Nb₂O₇:Pr³⁺.

 

Fig. 2 Diffuse reflectance spectra of (a) Sr₂Nb₂O₇:Pr³⁺ (0 and 0.4 mol%) and (b) Ca₂Nb₂O₇:Pr³⁺ (0 and 0.2 mol%).

Download Full Size | PDF

Figure 3 shows the characteristics of red-emitting long-lasting afterglow from Sr₂Nb₂O₇:Pr³⁺ and Ca₂Nb₂O₇:Pr³⁺. The spectra of afterglow indicate that the red emission of Sr₂Nb₂O₇:Pr³⁺ and Ca₂Nb₂O₇:Pr³⁺ both mainly originates from the ¹D₂-³H₄ transition of Pr³⁺ activators [upper insets of Figs. 3(a) and 3(b)]. Furthermore, the concentration of Pr³⁺ ions strongly influences the intensity and duration of afterglow, suggesting that the Pr³⁺ ions, in addition to serving as luminescent centers, could also regulate trap distribution. Overall, the afterglow of the Sr₂Nb₂O₇:Pr³⁺ series is much weaker and shorter than that of the Ca₂Nb₂O₇:Pr³⁺ series. Take their optimization components for comparison, after irradiation of a hand-held UV lamp (308 nm, 8 W) for 1 min, the afterglow of Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) powders lasts less than 2 min [lower insets of Fig. 3(a)], whereas that of Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%) powders can be observed by naked eyes more than 20 min [lower insets of Fig. 3(b)].

 

Fig. 3 Dependence of afterglow decay curves on Pr³⁺ concentration: (a) Sr₂Nb₂O₇:Pr³⁺ (0.2-3 mol%) (315 nm irradiation, 1 min) and (b) Ca₂Nb₂O₇:Pr³⁺ (0.1-2 mol%) (295 nm irradiation, 1 min). The upper insets show the spectra of afterglow. The lower inset show the photographs of afterglow captured at different delay times.

Download Full Size | PDF

In a further set of experiments, we incorporated the Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) and Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%) powders into phosphor/resin composition disks, and executed compression- and friction-triggered ML tests under different delay times after UV irradiation [Figs. 4 and 5]. We firstly measured the compression-triggered ML of two disks without any delay after irradiation. For the Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) disk showing a weak afterglow background [Fig. 4(a-i)], we can clearly observe the fusiform distribution of ML [Fig. 4(a-ii)]. The profile and location of ML spectra (not shown here) are consistent with those of afterglow spectra [insets of Fig. 3] [8,10,28], indicating that ML also mainly originates from the ¹D₂-³H₄ transition of Pr³⁺. However, the compression-triggered ML from Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%) disk recorded under the same compression condition is unrecognized [Fig. 4(b-ii)] due to the cover of strong afterglow noise [Fig. 4(b-i)]. In order to eliminate the negative effect of afterglow, we set the irradiated disks in dark for some time, i.e., 90 s delay for Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) disk [Fig. 4(a-iii)] and 900 s delay for Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%) disk [Fig. 4(b-iii)], which fades the afterglow background to an unobservable level for naked eyes. After such a delay, compression-triggered ML from Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) and Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%) disks all present a linear growth with increasing compressive load, and at peak compression ML reach a close intensity [Figs. 4(a) and 4(b)] and a similar image contrast [Figs. 4(a-iv) and 4(b-iv)]. Figure 5(a) lists the photographs of compression-triggered ML of two disks captured at different delay times. It is evident that Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) disk shows high-contrast images of compression-triggered ML after a short delay time (≤ 90 s) due to a weak and short afterglow, whereas Ca₂Nb₂O₇:Pr³⁺ disk showing a strong and long afterglow takes a long-time delay (≥ 600 s) to obtain a similar image of compression-triggered ML. Moreover, as the delay time increases, the intensity of compression-triggered ML shows a similar attenuation as that of afterglow, indicating that afterglow and ML originate from the evacuation process of carriers from the same traps. In addition to the characteristic of compression-triggered ML, the Sr₂Nb₂O₇:Pr³⁺ and Ca₂Nb₂O₇:Pr³⁺ disks also show friction-triggered ML [Fig. 5(b)], which is reported for the first time. Similar with compression-triggered ML, the Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) disk enables to capture a high-contrast image of friction-triggered ML within a shorter delay compared to the Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%) disk. These results show that the short-delay, high-contrast ML of Sr₂Nb₂O₇:Pr³⁺ has a more attractive prospect in the practical application of stress imaging.

 

Fig. 4 Afterglow decay and compression-triggered ML response during applying dynamic triangle loading on phosphor/resin composition disks: (a) Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) and (b) Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%). Insets show photographs of afterglow (i, iii) and ML (ii, iv) captured at typical delay times.

Download Full Size | PDF

 

Fig. 5 (a) Photographs of compression-triggered ML (captured at the peak compression of 1000 N) after different delay times: Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) (upper) and Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%) (lower). (b) Friction exertion device (left), and photographs of afterglow and friction-triggered ML recorded at different delay times: Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) (top right) and Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%) (bottom right).

Download Full Size | PDF

In order to understand the differences of trap-controlled afterglow and ML described above, we characterized ThL to clarify the distribution of traps in Sr₂Nb₂O₇:Pr³⁺ and Ca₂Nb₂O₇:Pr³⁺ [Fig. 6]. Both Sr and Ca compounds show ThL after Pr³⁺ doping, indicating the formation of traps by Pr³⁺ dopants [Figs. 6(a) and 6(c)]. It is known that the intensity and location of ThL curves reveal two important parameters of trap levels, i.e., the former reflects information of trap density and the latter that of trap depth [39]. The concentration of Pr³⁺ activators significantly changes the trap density of both compounds. However, the increase of Pr³⁺ concentration shifts the ThL profiles of Sr₂Nb₂O₇:Pr³⁺ to higher temperature, while hardly shifts that of Ca₂Nb₂O₇:Pr³⁺. According to the discrepancy of ion radius between Ca²⁺, Sr²⁺ and Pr³⁺ [40], we argue that the non-shift in Ca₂Nb₂O₇:Pr³⁺ should be ascribed to the extremely close radii of Pr³⁺ and Ca²⁺, while the shift in Sr₂Nb₂O₇:Pr³⁺ might result from the less radius of Pr³⁺ than that of Sr²⁺. Despite the underlying reason still needs further study, it is reasonable to infer that more Pr³⁺ substitution for Sr²⁺ would induce more lattice disturbance, thereby shifting the ThL curves. Furthermore, with the increase of Pr³⁺ concentration, the variation trend of ThL integrated intensity [insets of Figs. 6(a) and 6(c)] is consistent with that of afterglow intensity [Figs. 3(a) and 3(b)] in two compounds, indicating a strong dependence of afterglow intensity on the trap density. Typically, the intensity of ThL curve of Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) is much less than that of Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%) [Figs. 6(a) and 6(c)]. Accordingly, the much weaker afterglow of Sr₂Nb₂O₇:Pr³⁺ [lower insets of Fig. 3] is ascribed to its greatly lower trap density than that of Ca₂Nb₂O₇:Pr³⁺. The lower trap density in Sr₂Nb₂O₇:Pr³⁺ may result from the more non-radiative lattice defects due to the larger size difference of Sr²⁺ relative to Pr³⁺. On the other hand, each ThL curve can be further deconvoluted into two components which should result from two kinds of traps (Trap 1 and Trap 2) in materials. Figures 6(b) and 6(d) shows the Gaussian deconvolution of ThL curves of Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) and Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%), respectively. According to the above structural analysis, these two traps should originate from the substitution of Pr³⁺ for Sr²⁺/Ca²⁺ (i.e., electron traps [Pr_Sr/Ca]^o) in two different coordination environments (CN = 8 and 12) [Fig. 1(c)]. We estimate the depths of two traps by the Hoogenstraaten’s method [41], i.e., 0.46 eV (Trap 1) and 0.59 eV (Trap 2) in Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%), while 0.52 eV (Trap 1) and 0.63 eV (Trap 2) in Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%). The correspondingly shallower depths in Sr₂Nb₂O₇:Pr³⁺ results in the shorter duration of afterglow [Fig. 3(a)], conversely in Ca₂Nb₂O₇:Pr³⁺ [Fig. 3(b)]. The shallower trap depth in Sr₂Nb₂O₇:Pr³⁺ may be related to the narrower bandgap of Sr₂Nb₂O₇ [Fig. 2]. A narrow bandgap narrows the trap level to the bottom of the conduction band, making the trap shallow.

 

Fig. 6 ThL curves and estimated trap depths of Sr₂Nb₂O₇:Pr³⁺ and Ca₂Nb₂O₇:Pr³⁺. (a) and (c) Dependence of ThL on Pr³⁺ concentration in Sr₂Nb₂O₇:Pr³⁺ and Ca₂Nb₂O₇:Pr³⁺, respectively. Insets show the relative integral intensity of ThL as a function of Pr³⁺ concentration. (b) and (d) Gaussian deconvolution of ThL curves and trap depths estimated by Hoogenstraaten plots in Sr₂Nb₂O₇:Pr³⁺ (0.4 mol%) and Ca₂Nb₂O₇:Pr³⁺ (0.2 mol%), respectively.

Download Full Size | PDF

According to the above analyses, we propose a possible process of ML to describe the differences of TC-ML between Sr₂Nb₂O₇:Pr³⁺ and Ca₂Nb₂O₇:Pr³⁺, as shown in Fig. 7. Upon the excitation of UV light, the pumped electrons of Pr³⁺ activators would be captured by traps. Under the stimulus of ambient thermal energy, the trapped electrons are slowly released back to the excited state of Pr³⁺ through conduction band or tunneling effect [8]. The population of ³P₀ level of Pr³⁺ would be non-radiatively relaxed to ¹D₂ level via the intervalence charge transfer state of Pr-O-Nb [42,43], resulting in a red-emitting afterglow corresponding to ¹D₂-³H₄ transition. Due to the lower trap concentration and shallower depth, Sr₂Nb₂O₇:Pr³⁺ displays a weaker and shorter afterglow than Ca₂Nb₂O₇:Pr³⁺. Under the mechanical action, the captured electrons from the same traps would be detrapped much faster to produce a red-emitting ML as the emission process of afterglow. In view of the simultaneous input of ambient energy, ML is accordingly accompanied by afterglow. Therefore, Sr₂Nb₂O₇:Pr³⁺ requires a shorter delay time to achieve a high-contrast ML image due to weak and short afterglow, whereas Ca₂Nb₂O₇:Pr³⁺ conversely.

 

Fig. 7 Schematic illustration of bandgap and trap-distribution co-tailored ML of Sr₂Nb₂O₇:Pr³⁺ and Ca₂Nb₂O₇:Pr³⁺ (CB: conduction band, VB: valence band).

Download Full Size | PDF

4. Conclusion

A novel TC-ML material, Sr₂Nb₂O₇:Pr³⁺, has been developed that displays intense red-emitting ML under the mechanical actions of compression and friction. Compared with its isomorphism Ca₂Nb₂O₇:Pr³⁺, Sr₂Nb₂O₇:Pr³⁺ shows a weaker and shorter-lasting afterglow. This capability of Sr₂Nb₂O₇:Pr³⁺ enables us to achieve high-contrast ML over a short-time delay, which greatly simplifies the imaging process to facilitate the further application of ML. The investigation of crystal structure, optical bandgap and trap distribution indicates that the short-delay, high-contrast ML of Sr₂Nb₂O₇:Pr³⁺ should result from its lower trap concentration and shallower trap depth than Ca₂Nb₂O₇:Pr³⁺. When one considers a library of isomorphic candidates for host lattices, we expect that a growing amount of novel TC-ML materials will become obtainable for the fundamental research and technological utilization.