Light Delivery, Acoustic Read‐Out, and Optical Thermometry Using Ultrasound‐Induced Mechanoluminescence and the Near‐Infrared Persistent Luminescence of CaZnOS:Nd3+

Light emission from CaZnOS:Nd3+ by high‐intensity ultrasound excitation is demonstrated. Acoustic power and duty ratio enable simultaneous control of the degree of (local) thermo‐acoustic heating in isothermal or nonisothermal conditions. The Nd‐related photoemission provides direct thermometric feedback at an absolute sensitivity on the order of 10−4 K−1, within the physiological temperature range. From the individual temperature dependence of persistent luminescence and mechanoluminescence, both effects are attributed to a heterologous de‐trapping mechanism. In addition to acoustic heating and optical thermometry, this enables optical information storage and smart labeling with optoacoustic read‐out.


Introduction
Mechanoluminescent (ML) materials emit light under mechanical load. The mechanical stimulus can be anything from compression to impact, friction, or ultrasound (US) exposure; materials exhibiting ML have a broad range of applications. For example, ML enables optical stress sensing and self-reporting structures, smart product labeling and data encryption, or remotely controlled light emission for local photo-physical or DOI: 10.1002/adom.202300331 photo-chemical stimulation, optical thermometry, and optogenetics. [1][2][3][4] As the most widespread ML phenomenon, fractoluminescence was predicted to accompany bond rupture in brittle, inorganic solids, and may even occur in some ductile materials. [5] Alternatively, ML induced by optical detrapping of electronic states requires prior excitation (charging), for example, through irradiation with ultraviolet (UV) light. As a selection criterion, materials already exhibiting long-lasting (persistent) luminescence (pPL) are of potential interest for ML applications, too. [5] In this case mechanical stimulation may accelerate the thermal de-trapping reaction causing pPL, [6] leading to simultaneous ML functionality.
Optical materials with pPL have been explored for real-time optical bio-imaging, benefitting from their energy storage capability and the avoidance of autofluorescence. In this application, biocompatible particles exhibiting pPL are first optically activated using UV or visible (Vis) light irradiation, and then injected into biological tissue for subsequent monitoring. The available observation time depends on the decay kinetics of the underlying pPL process: a major disadvantage of this approach, posing restrictions on in situ observations and targeted activation. Aside from adapting spectral properties, pPL material discovery has therefore been devoted to extending the emission decay time. [7][8][9][10] Alternatively, in vivo (re-)activation of pPL materials remains a challenge: it requires an optical stimulus which is -in addition to effectively charge pPL trap states in the employed materialcapable to penetrate biological tissue. For example, Cr 3+ -doped ZnGa 2 O 4 was proposed for this purpose, using low-energy red photoexcitation for in vivo pPL activation to observe vascularization, tumor dynamics, and grafted cells. [11] The exploration of ML starts on very similar grounds. While pPL materials offer a rich source for the design of ML materials, in turn, ML could overcome the problem of ex situ or untargeted activation of pPL. For example, the mechanical stimulus can be used to switch on a pPL process and ML emission can act as a local light source. On this line, US excitation is of particular interest: in comparison to other ways of mechanical stimulation, it avoids large displacement and the risk of matrix damage. Furthermore, it does not require haptic contact and can penetrate solid as well as liquid matrices: USML can be triggered by high-intensity focused ultrasound (HIFU), which is a noninvasive stimulus with the ability to penetrate hydrous media, biological tissue, or other materials. The acoustic power levels required for USML lie in the range of therapeutic HIFU, [12] making bio-imaging tasks accessible for this technique.
The family of crystalline oxysulfides of the CaZnOS type is among the best-studied ML materials. For example, CaZnOS:Mn 2+ , [13] CaZnOS:Cu + , [14] CaZnOS:Bi 3+ , [15] CaZnOS:Sb 3+ [16] and a range of rare-earth doped variants of CaZnOS [17] have been reported to possess pPL, and CaZnOS:Er 3+ , [12] CaZnOS:Mn 2+ [18] and CaZnOS:Bi 3+ ,Mn 2+ [19] are currently known to also exhibit USML. Previously, ML from CaZnOS:Er 3+ was attributed to piezotronic interaction, [20] but HIFU studies indicated a more complex origin of USML. [12] Furthermore, CaZnOS:Nd 3+ was identified to exhibit significantly stronger ML than any other CaZnOS material, similar even to the most efficient ML phosphor currently reported, SrAl 2 O 4 :Eu 2+ , Dy 3+ , but covering the near-infrared (NIR) spectral range in which biological tissue has its windows of transparency. [21] Here, we demonstrate intense NIR pPL and USML of CaZnOS:Nd 3+ , and verify the ability of remote activation in water and biological tissue. Differing from impact ML, acoustic heating by HIFU accelerates pPL detrapping, whereby the spectral emission characteristics of Nd 3+ allow for simultaneous, highly sensitive optical thermometry. In addition to imaging, force monitoring or light delivery, we show that the high thermometric sensitivity and USML performance enable information encryption and US read-out, for example, in smart labeling, safety, or security applications.

Crystal Structure and Dopant Distribution
CaZnOS:xNd 3+ was synthesized with a dopant concentration of x = 0.2 mol% to 4 mol%. An additional 2 wt.% of LiF was added as a fluxing agent to facilitate the incorporation of Nd 3+ ions into CaZnOS. [21] X-ray diffraction (XRD) patterns of the materials obtained after solid-state reaction indicate hexagonal CaZnOS (JCPDS #01-076-3819) as the predominant phase, with negligible traces of CaS (JCPDS #08-0464), and unreacted CaO (JCPDS #37-1497), ZnS (JCPDS #036-1450) and Nd 2 O 3 (JCPDS #01-083-1347) (Figure 1a). The latter three were found only for a dopant concentration of x > 1.0%, which is why the following spectroscopic characterization is focused on the doping range of x ≤ 1.0%. The relative intensity of the (004) diffraction peak decreases while its position shifts to slightly higher angles with increasing addition of NdF 3 , indicating crystal orientation and lattice contraction upon the introduction of F − . (Figure 1b). [22][23][24] Traces of crystalline Nd 2 O 3 for x = 4.0% indicate the rare earth solubility limit of this material. CaZnOS exhibits a hexagonal structure (P6 3 mc) with tetrahedral Zn 2+ (CN = 4, r Zn = 0.60 Å) and octahedral Ca 2+ (CN = 6, r Ca = 1.00 Å) ( Figure 1c); [25] both cation sites are available for the incorporation of transition metal or lanthanide ions. [21,26,27] The sulfur and oxygen anions are arranged in alternating layers, leading to mixed coordination in the cation polyhedra, [ZnO 3 S] and [CaO 3 S 3 ]. High-resolution transmission electron microscopy (HRTEM) reveals numerous platy crystallites with a hexagonal shape, which yield spotty selected area electron diffraction patterns (Figure 1d). Energy dispersive X-ray spectroscopy (EDX) on these crystals provides clear evidence for the incorporation of Nd 3+ (r Nd = 0.98 Å [28] if in octahedral coordination). By way of example, the spectrum shown in Figure 1f exhibits all three L X-ray emission lines of Nd 3+ . While this does not fully exclude traces of Nd 3+ being present in other forms (such as Nd 2 O 3 found for x = 4.0%), it confirms earlier observations by X-ray absorption spectroscopy which concluded that rare earth dopants would be incorporated on the octahedral sites of Ca 2+ . [29] Selected-area electron diffraction yielded clear lattice fringes, implying high crystallinity of the synthesized CaZnOS:Nd 3+ material in accordance with powder XRD. The largest d-spacing is 0.28 nm, which corresponds to the lattice repeat of the (102) plane in CaZnOS. Due to the platelet crystal orientation, it is not possible to obtain a diffraction ring from the family of planes with indices {(002), (004), …}, however, this does not affect fitting of the lattice constants. Fit data of 11 interplanar spacings are provided in Table S1, Supporting Information. They correspond closely to those calculated from literature reports of crystalline CaZnOS. [25] Based on the HRTEM data, we obtain the lattice constants a = 3.74 Å and c = 11.62 Å (compared to a = 3.75726 Å, c = 11.4013 Å from Sambrook et al. [25] ).

pPL, ML, and USML
For pPL characterization, cylindrical samples were prepared by embedding the phosphor powder within an epoxy resin (see Section 4.2 for details). After 60 s of irradiation at 365 nm, samples exhibited pPL at the main PL band (880 -960 nm, Figure 3a) for at least 180 s. The decay time for pPL to reduce to 1/e of its initial level was ≈15 s (Figure 3b). While the origin of pPL is still debated in detail, we take that it originates from slow, thermally activated depletion of charge carriers captured in trap states according to a Boltzmann probability distribution. This relaxation process involves energy transfer to the luminescent centers, causing light emission. [31,32] Mechanoluminescence was initially observed in ball-drop experiments (Supporting Information, Figure S2, Supporting Information) conducted in accordance with previous protocols. [12] Similar to CaZnOS:Er 3+ , [12] ML saturation was observed when the impact energy exceeded 40 mJ ( Figure S2c, Supporting Information). This indicates that the saturation effect could be an intrinsic property of the matrix material, related to its piezoelectricity (affected by the specific dopant only in as far as it affects the matrix crystal structure and its defect states): de-trapping is induced by a local piezoelectric field at a rate which decreases with reducing number of trapped states, leading to luminescence saturation. [33] USML signals were obtained with a setup reported previously. [12] In the range of 850-1000 nm (Figure 3a), PL, pPL, and USML exhibit very similar features, characteristic for the electronic properties of Nd 3+ and indicating their similar origin. As a first difference, the USML intensity profile is affected by acoustic heating. In Figure 3b, pPL and USML decay curves are compared for variable acoustic power. Higher acoustic power leads to more intense USML and faster decay. When the sonication time is reduced from 60 s to 10 s, similar peak intensities of USML and similar carrier depletion rates are reached, however, depletion is incomplete (as seen from the residual USML during prolonged sonication, highlighted in Figure 3c): while the emission intensity represents the release rate of charge carriers, integration of the decay curve provides a measure of the total number of released carriers. [34,35] Obviously, the number of released carriers cannot exceed the number of carriers stored carriers by UV charging in the first place, whereby pPL and USML draw from the same reservoir. Figure 3d,e shows the time-dependence of USML intensity and simultaneous heating (in terms of achieved specimen temperature) as a function of applied acoustic power. In order to minimize signal interference from pPL, samples were kept in a dark environment for 5 min directly after UV charging. In this period, the pPL afterglow reached the noise level of the employed CMOS detector. USML experiments were done only immediately after this dark delay. Four parameters were used to evaluate the USML performance, i.e., the time to reach maximum emission intensity (from the start of sonication), the maximum intensity itself, the released energy (obtained through integration of the decay curve), and the maximum specimen temperature achieved during sonication. Figure 3f,g shows the dependence of these parameters on applied acoustic power. All four parameters exhibit a weak onset in the power range of 0.1 -1 W. For sonication at 2 -10 W, the maximum USML intensity increased approximately linearly with acoustic power, with gradually reducing time to a maximum. The total released energy is saturated for excitation above 3 W, similar to US heating. In Figure 3h, USML cycles are shown for subsequent sonication periods at an acoustic power of 1 W (below the saturation limit of total energy release, Figure 3f). In this experiment, the USML intensity reached per cycle decreases because the remaining density of trapped states reduces with each sonication period, in accordance with the timeresolved energy release (Figure 3i). On the other hand, the temperature profiles are highly repeatable within the five observation periods.

Heterologous Trap States in pPL and USML
ML detrapping is often related to piezo-electricity. [36][37][38][39] The associated trap depth can be probed through the temperature depen-dence of emission decay under constant or zero load (the latter termed thermoluminescence, TL). [40] TL trap energy levels of 0.5 -0.8 eV have been reported for a range of pPL materials. [41] For pPL involving Nd 3+ , values of 0.38 -0.73 eV (Sr 2 SnO 4 ), [42] 0.92 eV (CaMgSi 2 O 6 :Mn 2+ ,Nd 3+ ) [43] and 0.76 eV (ML, CaZnOS:Nd 3+ [21] ) were previously found; these energy levels are several times higher than the thermal energy at room temperature (≈0.025 eV), what is the cause of the sluggish carrier relaxation time. However, TL observations do not distinguish between different types of traps such as might be responsible for either pPL or ML. [44] In the following, we, therefore, analyze the isothermal decay . Thermo-acoustic heating, persistent luminescence, ultrasound-induced mechanoluminescence from CaZnOS:Nd 3+ . a) Normalized PL, pPL, and USML spectra of CaZnOS:xNd 3+ (x = 1.0%). b,c) Time-dependent pPL and USML intensity for variable acoustic power and USML excitation time, respectively; excitation time in b) was 60 s; acoustic power in c) was 2 W. Power-dependence of USML intensity (d) and specimen temperature (e) for sonication after pPL decay with a delay time of 5 min. f,g) Total released energy, maximum specimen temperature, time to reach maximum USML emission intensity, and maximum emission intensity as functions of applied acoustic power. h) USML intensity and thermos-acoustic heating during cyclic US stimulation (five sonication periods) at 1 W, without intermediate UV recharging. Energy release during a sequence of cyclic stimulations (i) for variable acoustic power.
kinetics of USML and pPL in order to obtain separate insight at the underlying trap levels.
Thermal depopulation of the pPL or ML trap levels occurs with a Boltzmann probability distribution, therefore, the de-trapping rate is governed by an Arrhenius equation with the energy barrier ΔE, [41,42,45,46] where p is the decay constant, is a frequency factor, ΔE is the trap depth, k B is the Boltzmann constant, and T is the absolute temperature. For a single trap species and in the absence of re-trapping, the observed decay in emission intensity I(t) is dependent solely on the number of trapped carriers (first-order kinetics), leading to a simple exponential decay function with the lifetime (T), ∼ 1/p. [45][46][47][48] However, re-trapping can normally not be ignored in slow TL and pPL, in particular, when multiple trap states are expected. Therefore, general order kinetics were employed for the present case, using the empirical relation of Chen and Kristianpoller for a fixed temperature, [49,50] with the parameter, b denoting the effective kinetic order of the observed decay in emission intensity. Using this approach, only those traps are covered which indeed participate in the radiative relaxation processes visible in pPL or USML.   Table S2, Supporting Information. In pPL, isothermal conditions were achieved directly using a thermostat (see Methods Section). For USML, data fits were restricted to the isothermal plateau following acoustic heating (see Figure 4d). In this way, the decay constant p(T) and, finally, the trap depth ΔE were obtained. [51] Data are shown in Figure 4c,f. For pPL, a trap depth of ΔE pPL ∼ 0.170 eV is obtained, significantly lower than previously reported values estimated using general order kinetics (i.e., 0.623 eV and 0.652 eV for pPL of CaZnOS:Mn 2+ ,Pr 3+ and CaZnOS:Mn 2+ ,Ce 3+ ; [13] note that significantly higher emission lifetime was reported for these materials, too, around ≈45 s as compared to the current 15 s. For USML, the decay process appears to occur in two regimes, depending on applied acoustic power, i.e., with ΔE low ∼ 0.203 eV when the acoustic power is below 3 W, and ΔE high ∼ 1.349 eV for higher US power. ΔE low is very similar to ΔE pPL , indicating that in this power range, pPL and the observed USML signal have the same origin. In the high-power regime, ΔE high is well above the reported value for ML of CaZnOS:Nd 3+ , [21] ΔE ML ∼ 0.762 eV (obtained by assuming simple first-order kinetics). Assuming that the ultrasound-induced mechanical force [52,53] results in a local piezoelectric field accelerating the release of charge carriers, the observed trap depth lets us expect efficient energy transfer into the excited states of Nd 3+ (located at similar energy levels). This could also explain the higher intensity of USML as compared to pPL. [11,17,21] In addition, these results are consistent with the previous conclusions that the contributing trap states could vary depending on the kind of luminescence and/or excitation process, even if they follow the same mechanism. [12,54,55]

USML Labeling
The simultaneous response of pPL and USML from CaZnOS:Nd 3+ to temperature, light, and acoustic loading enables multimodal data storage and read out, for example, in a smart label. This is illustrated in Figure 5a-b: Heating a labeled specimen to 120°C thermally cleans the material from all trap states. In step 2, using a photomask, UV charging can be restricted to a specific surface region, in which traps are repopulated. Initial afterglow decays within a specific time, e.g., 300 s in the current example. Further heating or ultrasonic treatment can then be used to read out the stored information.

Light Delivery, Acoustic Heating, and Static Thermometry
Aside simple read-out, the stored energy, and its corresponding NIR emission characteristics can also be used for biomedical imaging and temperature sensing. In particular, this  could allow for acoustic heating using HIFU with simultaneous thermometry, [12] for example, for minimally invasive HIFU therapy. The Nd 3+ energy levels of 4 F 5/2 and 4 F 3/2 are thermally coupled with an energy difference ΔE of ≈1000 cm −1 in most host materials. [56] Their transitions to 4 I 9/2 cover the NIR-I window available for photodynamic therapy or biomedical imaging (834 nm and 909 nm in CaZnOS:Nd 3+ ). [21] The intensity ratio between these two transitions increases with increasing temper-ature, what allows for optical thermometry. [57] In comparison to conventional techniques, using USML for this purpose does not require an optical stimulus, thus, the matrix in which thermometry is to be conducted does not have to be transparent for the excitation light.
As shown in Figure 6a, an isothermal plateau (within ±0.7°C, recorded directly on a thermographic camera) was reached when US sonication was conducted at a variable duty ratio after ≈30 s www.advancedsciencenews.com www.advopticalmat.de of treatment. In this experiment, samples were simultaneously charged by UV light during heating; UV irradiation was stopped after 30 s (when isothermal conditions were achieved) and USML was monitored together with pPL afterglow. In order to increase the signal-to-noise ratio, an integration time of 10 s was applied for spectra acquisition in this case; the spectra are shown in Figure 6c. Fixing the output acoustic power at 3 W, setting the duty ratio from 0% to 100% provided control of the level of the isothermal plateau within the range of 21.5°C to 148.7°C (Figure 6b; the temperature curve addressed in this way exhibits two regimes, separated by the glass transition temperature of the epoxy resin employed for sample encapsulation, 70°C). Therefore, datasets obtained in the range above ∼ 70°C are from a composite with a viscoelastic matrix. As shown in the following, we did not find that this had any notable effect on the observed USML characteristics up to a temperature of 143°C (with a duty ratio of 90%). In this way, the temperature-induced variation in the luminescence intensity ratio (LIR) was obtained (Figure 6d), [56] where I 834nm , and I 909nm are the luminescence intensities of the corresponding thermally coupled energy levels of Nd 3+ and ΔE LIR is the effective energy gap between 4 F 5/2 and 4 F 3/2 . C is an empirical correction factor representing state degeneracy and the spontaneous emission probability. The experimental LIR was well-fit using Equation 3 in linearized form (R 2 = 0.992) with C = 1.02 (Figure 6e). For the energy gap ΔE LIR , a value of ∼ 876 cm −1 (1.08 eV) was obtained, close to the value 749 cm −1 (0.929 eV) found from the PL excitation spectra (Figure 2c). The temperature sensing ability was evaluated through the absolute and the relative sensitivity, S A and S R , defined as the absolute and relative variation of the LIR with respect to temperature, respectively, [58] Data obtained for S A and S R are shown in Figure 6f. S A monotonically increases with temperature from 2.17 × 10 −4 K −1 at 293 K to the maximum of 3.90 × 10 −4 K −1 at 423 K, whereas S R has its maximum value of 1.47% K −1 at 293 K and decreases with rising temperature. The value of S R is comparable to that of benchmark materials for high-performance optical thermometry, e.g., 1.68% K −1 at 300 K for SrF 2 :Nd 3+ [59] and 1.75% K −1 at 288 K for Gd 2 O 3 :Nd 3+ : [60] the current results demonstrate similar sensing capability for optical thermometry using USML. This offers a novel optical thermometric method which does not require photoexcitation. Given the emission characteristics and the material transparency window, the approach enables temperature monitoring of biological tissue during remote HIFU stimulation (Figure S3, Supporting Information).

Dynamic Thermometry
For verifying potential applications, we performed USML thermometry during the ultrasonic heating process, i.e., in nonisothermal conditions (Figure 7a-b). The temperature is extracted from the intensity ratio of the thermally coupled emission levels of I 834 /I 909 (using the scaling obtained from Figure 6) or, alternatively, I 905 /I 935 ( Figure S4, Supporting Information). For comparison, an IR camera is used to monitor the global (effective) temperature at the sample surface. Expectedly, the observed USML intensity increases with global temperature, as does the LIR. However, other than in the above isothermal thermometry, this experiment is affected by a range of practical issues. In particular, thermometry is done after pPL decay, that is, drawing primarily from the deeper traps of ΔE high (data in Figure 7 are for 3 W acoustic excitation), what may affect the reaction kinetics of the detrapping process. Application-wise, this presents a case where the material is UV-charged prior to its actual use, for example, in biomedical imaging or as a temperature sensor in HIFU therapy. Furthermore, the delay between experimental data acquisition and heating dynamics affects the observed LIR; in the current case, the spectral acquisition time was 1.0 s with an average heating rate of 10 K s −1 (see Figures 3e and S5, Supporting Information). For higher temperature, this rate was lower so that the high-T data are less affected. The time to reach maximum emission intensity is only 30 s (for 3 W excitation power, Figure 3g). The IR camera used for simultaneous monitoring of the specimen temperature systematically underestimates the local temperature obtained by spectroscopy (Figure 7d). For one, this reflects the delay in global heating of the sample relative to the immediately monitored spectroscopic characteristics (which we associate with local temperature); it indicates that the local response to HIFU is much faster than the global heating process monitored by camera observation. In accordance with the data presented in Figure 4 and Figure 6, this highlights the simultaneous action of ML and thermally-accelerated pPL in HIFU excitation. The method provides remote and dynamic access to local temperatures not otherwise accessible by outside camera monitoring.

Conclusions
CaZnOS:Nd 3+ exhibits pronounced near-infrared USML as well as pPL. For low acoustic power, pPL and USML are driven by a similar mechanism of thermal de-trapping, whereby acoustic heating accelerates pPL relaxation. A heterologous de-trapping mechanism is observed at higher excitation power, with the trap depth ranging from ≈0.2 eV to ≈1.4 eV. Thereby, acoustic power and duty ratio provide control over the simultaneous thermoacoustic heating effect, which can be directly monitored using the thermally coupled energy levels of 4 F 5/2 and 4 F 3/2 in Nd 3+ . In this way, thermometric feedback is obtained at an absolute sensitivity on the order of 10 −4 K −1 , within the physiological temperature range. Such a feedback loop could enable significantly less-invasive protocols for HIFU treatment and therapy. The relation between pPL and USML could be used in the future design of further USML materials. Application wise and in addition to combining acoustic heating and optical thermometry, it also enables optical information storage and smart labeling with optoacoustic read-out.  Figure S5, Supporting Information). c) Observed total USML intensity as a function of global sample temperature. d) Optical thermometry using LIR (I( 4 F 5/2 -4 I 9/2 )/I( 4 F 3/2 -4 I 9/2 )), and comparison to the global specimen temperature recorded with an IR camera.

Experimental Section
Phosphor Synthesis: A series of Nd 3+ -doped CaZnOS phosphors were synthesized by high-temperature solid-state reaction. [12] For this, batches were prepared with nominal chemical composition of Ca 1-x ZnOS:xNdF 3 , using an additional 2 wt.% LiF as fluxing agent (x = 0.2, 0.5, 1.0, 2.0, and 4.0 mol). Raw materials of CaCO 3 (99.5%, Carl Roth), ZnS (99.99%, Acros Organics), NdF 3 (99.9%, Projector GmbH), and LiF (99.99%, Carl Roth) were mixed and ground by wet-milling for 24 h in ethanol using a rotational ball mill with ZrO 2 ceramic balls. The obtained slurries were dried and sintered in covered alumina crucibles at 1050°C for 4 h in a horizontal tube furnace under N 2 atmosphere with a flow rate of 1.5 l min −1 . Samples obtained in this way were ground again in an agate mortar and stored in sealed plastic tubes for physical characterization and further use.
Characterization: Powder X-ray diffraction (XRD) patterns were acquired from all samples with an X-ray diffractometer (MiniFlex, Rigaku) with Cu K radiation ( = 1.54059 Å). Powdered samples were loaded on holey carbon grids and crystallite morphology, chemical composition, and lattice parameters were then analyzed in detail by transmission electron microscopy (TEM) using a 200 kV FEI Tecnai G 2 FEG equipped with an Oxford 80 mm 2 energy-dispersive silicon drift X-ray detector and a Gatan UltraScan 2k CCD camera. Aside image data, selected area electron diffraction (SAED) patterns were collected, and energy dispersive spectroscopy (EDS) analysis was conducted to verify the presence and spatial distribution of Nd. Diffuse reflection spectra were recorded with a UV-Vis spectrophotometer (Cary 5000, Agilent) over the spectral range of 250 -800 nm. Photoluminescence emission (PL) and excitation (PLE) spectra were recorded using a high-resolution fluorescence spectrometer (Fluorolog-3, Horiba) equipped with a 450 W Xe lamp for excitation, and a set of photomultiplier tubes (PMTs; Hamamatsu R928 for UV-Vis, R2658 for UV-Vis-NIR and H10330 for NIR) for detection. The pPL and ML spectra were acquired using a compact fiber spectrometer (Shamrock 163, Andor) equipped with a CCD detector (iDus 420, Andor). Photographic pictures of pPL and ML emission were taken with a monochromatic CMOS camera (DCC1545M, Thorlabs).
Initial ML tests were conducted with a routine ball-drop experiment, [12,61] using powder samples of 0.1 g inserted into the center an epoxy sample holder. A stainless-steel ball with diameter of 16 mm and a mass of 16.705 g was dropped vertically through a guiding pipe on a piston placed at the top part of the sample holder. At the bottom part of the sample, the coupling head of a fiber spectrometer was placed to receive the ML emission signal. The impact energy was adjusted in the range of 10 to 180 mJ by changing the drop-height. Before each drop, the sample was illuminated with a UV-A lamp for 60 s, and a delay of 60 s was applied so as to reduce pPL afterglow to the noise level. Spectra acquisition was started immediately at the release of the ball impactor; the integration time of the spectrometer was set to 3 s. Ten individual experiments were conducted in this way for each sample and each impact energy value, from which average spectra and values of standard deviation of the ML emission intensity were obtained.
In order to observe the temperature effect on pPL, powder samples of 1.0 g were homogeneously dispersed in a 2.0 g of epoxy resin (Epoxy Resin L and Harderner GL 2, R&G GmbH) to form a testing specimen with a diameter of 20 mm and a thickness of ≈2.5 mm. Optical charging was done with the same UV lamp as above, for 60 s, followed by dark relaxation for 120 s at room temperature. Immediately thereafter, the sample was transferred to a heating stage, on which further experiments were conducted at a controlled, elevated temperature. The heating stage consisted of an aluminum block with a heating element in the center, a thermocouple, and a digital controller (box-3216, MRC). Emitted light was collected with the head of a multimode optical fiber placed close to the center of the specimen; the other end of the fiber was coupled to the receiving spectrometer in the same way as with the impact ML experiments mentioned above.
For ultrasound-induced mechanoluminescence experiments, the same epoxy-embedded powder sample was used. A high-intensity focused ultrasound (HIFU) transducer (H-104, Sonicconcepts), a signal generator, and an amplifier (TPO-102, Sonicconcepts) were used for the excitation source. To receive maximum acoustic power, the specimen was placed at the focal region of the HIFU transducer. The transducer was immersed at the bottom of a water tank filled with degassed water. Emitted light was collected with the head of a multimode optical fiber placed close to the center of the specimen; the other end of the fiber was coupled to the receiving spectrometer. The HIFU transducer was running at output power within 0.1 -10 W with a duty ratio between 0 to 100% at 500 kHz. A thermographic camera (VarioCAM, Jenoptik) was used to independently monitor the change of temperature accompanying US exposure.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.