Ultrasound‐Induced Luminescence from Cr3+‐Doped ZnGa2O4 Glass–Ceramic Composites

An intense ultrasound‐induced luminescence (USL) from Cr3+‐activated glass–ceramic composites, obtained by direct precipitation of nanoscale ZnGa2O4 from silicate melts upon cooling, is reported. The USL band overlaps with the first near‐infrared transmission window of biological tissue and spans further into the visible red spectral range, generating interest for visible data encryption and labeling as well as for photophysical stimulation with a remote, non‐optical energy source. Time‐resolved observations of luminescence build‐up and decay, US heating, and persistent luminescence reveal thermal de‐trapping as the origin of the observed US sensitivity. The spectroscopic performance is very similar to that of phase‐pure ZnGa2O4, the fabrication process leads to a robust, dense, and biocompatible composite without requiring secondary encapsulation.


Introduction
Mechanically stimulated light-emission, termed mechanoluminescence (ML), enables the detection and optical visualization of mechanical strain across a wide variety of applications, from biophysical stress monitoring to material testing, labeling, and data encryption. [1][2][3][4][5][6] Traditional ML materials are designed according to their response to haptic contact, for example, monitoring states of compression, friction, or destructive fracture. [7][8][9][10] Nondestructive ML is of interest for the continuous detection of stresses over extended periods of use. [11][12][13] Using ultrasound, in particular, high-intensity focused ultrasound (HIFU), contactfree, remote stimulation can be achieved, which might even penetrate aqueous media, biological tissue, or other kinds of materials; HIFU is used extensively in biomedical therapy and imaging. [14][15][16] When luminescence is induced by ultrasound (termed ultrasoundinduced luminescence, USL), it enables a full new range of applications. [17] For example, USL can be used for acoustic field mapping, for remotely inducing light emission, or for simultaneous ultrasonic heating and optical thermometry. [18][19][20][21] Polymer membranes with embedded USL particles were used for fast visualization of the cross-section of ultrasound pressure fields with a spatial resolution below 200 μm for acoustic pressures in the range of 150-4,500 kPa, and frequencies of 1-25 MHz. [22] Such applications could notably benefit from the enhanced mechanical stiffness, thermal stability, and spatial homogeneity offered by fully inorganic ceramics and ceramic composites.
Current ML materials development is focused on polycrystalline ceramics, ceramic powders, and particles embedded in organic matrices in the form of semitransparent membranes. [1][2][3] However, depending on the desired application, organic matrix materials (such as epoxies or thermoplastic polymers) can be thermally, chemically, or mechanically unstable, and can also be susceptible to ultrasound heating and irradiation damage. In addition to the choice of the matrix material, homogeneous dispersion of the ML phosphor presents another challenge. As an alternative for overcoming the former issues, inorganic glass matrices and phosphor-in-glass composites have been considered, using low-melting glasses into which ML particles were embedded through conventional melting or sintering techniques. [23] Manufacture of such composites requires the phosphor to be thermally and chemically stable while being immersed within the inorganic glass melt. Furthermore, the problem of spatial homogeneity is not addressed in this way. As a third alternative, the ML-active crystalline phase could therefore be directly precipitated from the glass melt at elevated temperature, either by controlled crystallization of a precursor glass during reheating (leading to an ML glass ceramic), [17] or by devitrification of the melt during cooling (leading to a glass-ceramic composite material). [24,25] For both cases, however, USL has not yet been demonstrated.
Cr 3þ -doped ZnGa 2 O 4 presents intense red emission and long afterglow at biowindow, which can be excited by wavelength from UV to near-infrared. This red emission can be in vivo (re-)activated by using low-energy red photoexcitation for realtime bioimaging, [26] which presents great advantages over DOI: 10.1002/adpr.202300018 An intense ultrasound-induced luminescence (USL) from Cr 3þ -activated glassceramic composites, obtained by direct precipitation of nanoscale ZnGa 2 O 4 from silicate melts upon cooling, is reported. The USL band overlaps with the first near-infrared transmission window of biological tissue and spans further into the visible red spectral range, generating interest for visible data encryption and labeling as well as for photophysical stimulation with a remote, non-optical energy source. Time-resolved observations of luminescence build-up and decay, US heating, and persistent luminescence reveal thermal de-trapping as the origin of the observed US sensitivity. The spectroscopic performance is very similar to that of phase-pure ZnGa 2 O 4 , the fabrication process leads to a robust, dense, and biocompatible composite without requiring secondary encapsulation.
conventional persistent photoluminescence (pPL) materials and needs to be optically activated using UV or visible (Vis) light irradiation before injection into biological tissue for subsequent monitoring. Besides, Cr 3þ -doped ZnGa 2 O 4 can be precipitated in robust glasses via conventional melt-quenching method to form a highly dense composite material with good mechanical strength and chemical stability, [17,27,28] which could lead to potential applications of ML glass and glass ceramics in biotreatment and imaging.
Here, we follow the approach of devitrification of the melt during cooling, demonstrating USL from a highly transparent glass-ceramic composite material (GCC). We achieve intense near-infrared USL emission from nanocrystalline Cr 3þ -doped ZnGa 2 O 4 (ZGO) precipitated from a silicate melt upon cooling. This enables a new platform for highly homogeneous, allinorganic USL materials covering the transparency window of biological tissue, for example, in biophysical stress monitoring or hybrid HIFU and photodynamic therapy. Figure 1a shows the X-ray diffraction (XRD) patterns of Cr 3þdoped sp-ZGO crystal and the ZGO silicate GCC, together with the JCPDS standard of spinel ZGO, 38-1240. The XRD data indicate single-phase crystallinity for both materials. Significant peak broadening in the ZGO GCC results in a reduced crystal size. In addition, an amorphous halo is visible in this case, resulting from the residual sodium silicate glass phase. Additional crystal phases were not observed by XRD. The micro-Raman spectra of sp-ZGO (Figure 1b) exhibit two primary features, located at 608 cm À1 (T 2g ) and at 713 cm À1 (A 1g ). [29] Both of these features reoccur in the GCC, superimposed by the typical Raman spectrum of a binary sodium silicate glass. [30][31][32] ZGO has space group Fd 3m symmetry, with Zn 2þ ions occupying tetrahedral sites and Ga 3þ ions occupying octahedral sites. The Ga 3þ site with ionic radius [33] r Ga = 620 pm can readily accommodate Cr 3þ (r Cr = 615 pm); we used Scherrer's equation to estimate the mean crystallite size D from the XRD data [34] D ¼ kλ=βcosθ (1) using the shape factor k = 0.89. θ is the Bragg angle of the XRD peak with half-width (corrected for instrumental line broadening) β and the wavelength λ = 0.154056 nm of the employed Cu Kα radiation. Using the (220) plane, we obtained a value of %60 nm for sp-ZGO and %15 nm for the ZGO GCC. The crystal size is directly confirmed by TEM image in Figure 2a, which clearly presents uniform nanocrystals lying homogeneously on the gray background corresponding to the glass matrix. The HRTEM ( Figure 2b) presents a well-defined lattice structure and the SAED patterns in Figure 2c proved the polycrystalline feature  www.advancedsciencenews.com www.adpr-journal.com of ZnGa 2 O 4 nanocrystals. The low crystallite size is a primary condition for optical transparency ( Figure 3c). However, ZGO has a refractive index of %1.90, whereas a common zinc sodium silicate glass has %1.55. [35] This difference in refractive indices leads to significant scattering, even at relatively low crystallite size. The observed total transmittance is %56% at 550 nm. As further seen in the transmittance and reflection spectra (Figure 3c), the Cr 3þ -doped ZGO GCC exhibits the typical absorption features attributed to octahedral Cr 3þ in ZGO [36] (indicating the incorporation of Cr 3þ ions into the ZGO crystal). The respective absorption bands are peaking at %560 nm and at %400 nm; they are assigned to transitions from 4 A 2 to 4 T 2 and 4 T 1 , respectively. [37] Red photoemission from octahedral Cr 3þ was observed in sp-ZGO as well as in the GCC, as shown in Figure 3a. The photoemission spectra exhibit the characteristic NIR emission following 2 E ! 4 A 2 . [38] The zero-phonon-line (ZPL) of Cr 3þ in undistorted Ga 3þ sites of ZGO is identified at 690 nm (R-line). The ZPL of Cr 3þ ions distorted through an antisite defect as the first cationic neighbor is found at 697 nm (N2 line). [39,40] Stokes phonon side bands (PSB) occur at 709 and 715 nm, and anti-Stokes PSB at 662, 671, and 679 nm. The integrated NIR emission intensity from the ZGO GCC is about one-quarter of that of the sp-ZGO, subject to differences in sample volume and morphology. Qualitatively, sp-ZGO and the ZGO GCC exhibit identical emission characteristics. As shown in Figure 3b, the excitation spectra of the ZGO GCC (monitoring emission at 697 nm) comprise of four bands, located at 260, 320, 410, and 560 nm. These are related to Cr-O and Ga-O charge transfer, [41,42] and to transitions from 4 A 2 to 4 T 1(4 P) , 4 T 1, and 4 T 2 , respectively. [41] PL decay curves of 2 E ! 4 A 2 are shown in Figure 3d. From these decay data, the average lifetime τ of the 2 E level of Cr 3þ was obtained at 43.21 and 36.62 μs for sp-ZGO and ZGO GCC, respectively. The decay curve for 697 nm emission from the Cr 3þ -doped sp-ZGO is single-exponential, indicating a single emission center. For the GCC, the best datafit is obtained with a double-exponential function; fit parameters are provided in Table 1. Decay in the ZGO GCC is faster than in the sp-ZGO crystals. This may be due to the different effective Cr 3þ doping concentrations, and also to more distorted local environment in the ZGO GCC.

Results and Discussion
The effect of US stimulation on PL emission is presented in Figure 4. When exposed to US loading, the ZGO GCC exhibits broadband Cr 3þ -related emission in the red spectral region (see also photographs in Figure 5), qualitatively similar to pPL  in the same material, but stronger in intensity due to simultaneous pPL and USL. The very similar band shape is an indicator that pPL and USL share a common origin. When using UV light for excitation, energy is directly absorbed into the T-levels of Cr 3þ (Figure 3b), but also into more stable trap levels, from where luminescent decay starts. [43] pPL and USL decay data are . c) Time-dependent pPL and USL, monitoring the intensity maximum within the spectral range of 650-800 nm (the pPL graph represents the decay in emission when the GCC is left at rest after UV charging). The USL data were taken with the US starting at t = 120 s for a duration of 120 s. d) Time-dependent USL intensity and specimen temperature using the same conditions as in (c). All decay data were taken following excitation at 365 nm for a duration of 60 s. e) USL and sample temperature during cyclic US stimulation (14 cycles) with a period of 10 s "ON" followed by 10 s "OFF". f ) USL and sample temperature during cyclic US stimulation allowing for full recovery of the specimen temperature after each cycle (seven cycles) through applying a period of 10 s "ON" followed by 60 s "OFF".  presented in Figure 4b,c. pPL afterglow of the GCC material (without US stimulation) is depicted by the solid red line in Figure 4c. When US stimulation is started, the USL intensity increases in an exponential plateau function; the plateau is reached after %120 s of excitation. This behavior notably differs from ML generated by friction or impact loading, where the mechanical stimulation leads to an instantaneous response (provided that any load thresholds are overcome). [44] When US stimulation is stopped, the luminescence intensity decays gradually to eventually reach a level below that of the extrapolated pPL behavior (see t > 400 s in Figure 3c). This reflects that USL-in the present case-leads to accelerated depopulation of the pPL trap level. This is explained when considering the heating effect of US treatment. As shown in Figure 4d, US treatment heats the GCC sample; the monitored temperature profile and the parallel USL emission intensity exhibit qualitatively similar time-dependence, both in the heating and in the cooling phase (however, within the observation time, the temperature profile depicted in Figure 4d does not reach saturation). The parallel observation of sample heating and USL intensity indicates that USL originates from thermally stimulated luminescence (TSL), whereby the US is the thermal trigger. US heating activates the release of pPL trap states, leading to TSL emission. [22] Cyclic US stimulation data are presented in Figure 4 for rapid cycling (Figure 4e) and slow cycling (Figure 4f ). In both cases, cycles are sufficiently short to avoid any saturation effects: the temperature increase generated on the GCC by each cycle is on the order of 3 K (in comparison to about 30 K in the quasistatic experiments reported in Figure 3). The rapid cycle did not allow for thermal recovery of the sample between cycles, so the mean sample temperature increased continuously over the course of 14 cycles. In this case, the observed decay data are superimposed by the pPL afterglow (visible in the first 100 s/first four cycles of Figure 4e). In slow cycling, sufficient time was provided to allow for sample recovery to re-attain room temperature; in this case, each cycle started at approximately the same temperature (23°C, see Figure 4f ). The observed data corroborate our conclusion that USL of the present material is induced by heating: within the considered number of cycles, the luminescence intensity build-up and decay rates are highly reproducible; the decay of luminescence intensity seen after stopping US stimulation overlaps with the rate of thermal recovery.
The power dependence of the GCC USL build-up and decay is presented in Figure 6a, monitoring the effect of 60 s of continuous HIFU stimulation and subsequent thermal recovery. In these experiments, the USL signal initially increases up to a maximum value, in accordance with the heating effect reported in the previous paragraphs. After this initial phase, the signal intensity decreases, even though HIFU stimulation is maintained. This behavior reflects the trade-off between thermal activation and the gradually decreasing number of occupied trap states available for luminescence: light is "squeezed" out of the material by HIFU loading until the available trap states are discharged. When HIFU loading is stopped, the USL intensity decreases gradually, reflecting the thermal recovery as with the data presented in Figure 4. The dependence of emission intensity on HIFU acoustic power is depicted in Figure 6b. After an onset regime up to a power of about 1.0 W (in which no clear trend is observed within experimental uncertainty), we find a gradual increase in emission intensity with acoustic power. In particular, there are no saturation effects up to the observed limit in acoustic power of 2.5 W. The build-up time (taken as the maximum of the time-resolved intensity plot in Figure 6a) decreases from %16 to 5 s when increasing the acoustic power from 0.5 to 2.5 W, best fit with a compressed exponential decay function.
For qualitatively confirming the observed USL kinetics, a time series of sample photographs taken over the course of an on/off HIFU simulation experiment is shown in Figure 7. When US stimulation was started (after a delay of 1 s), consistent with the data shown in Figure 4c, the emission intensity built up gradually over a time of about 8 s (the spot shape visible in Figure 7 reflects the US intensity cross-section generated by the HIFU source). For longer US exposure (>8 s), the light emission intensity slowly decreases due to the continuous detrapping (see also Figure 6a). When the US was turned off after 60 s, residual afterglow faded rapidly.

Conclusion
We presented ultrasound-stimulated luminescence from glass ceramic composites consisting of ZGO:Cr 3þ nanocrystals embedded in a silicate glass matrix. USL occurred over a www.advancedsciencenews.com www.adpr-journal.com broadband covering the near-infrared spectral range across the first transparency window of biological tissue. ZGO:Cr nanocrystals were generated in situ during the cooling of silicate melts, leading to dense and robust composites with photoluminescence, pPL, ML, and USL emission behavior very similar to that of single-phase ZGO:Cr reference powders. USL was found to originate from mechanically activated thermoluminescence, whereby US exposure heats the sample, leading to thermal detrapping. This enables not only local heating with a remote energy source but also simultaneous light emission, for example, for optical data encryption and read-out or local light delivery and photophysical stimulation.

Experimental Section
Sample Preparation: GCCs or glass ceramics with gallate spinel embedded in a silicate glass matrix are well-documented host materials for doping with optically active ion species. [45,46] Glass samples with a nominal composition of 60SiO 2 -4.5Na 2 CO 3 -20Ga 2 O 3 -15.5ZnO-0.1Cr 2 O 3 (in mol%, adopted from previous studies [22,45]) were prepared by melt-quenching in air, using SiO 2 , Ga 2 O 3 , Na 2 CO 3 , ZnO, and CrCl 3 .(H 2 O) 6 (analytical grade) as starting materials. The batched raw materials were melted in covered alumina crucibles at 1600°C for 60 min. After that, the melt was cast onto a stainless-steel plate and rapidly quenched by pressing with another plate. The as-quenched samples were grounded with SiC paper, and polished further to optical grade using a CeO 2 suspension. Sample sizes obtained in this way were on the order of %1 Â 1 cm 2 , with a thickness of 1.0 mm.
Characterization: X-ray diffraction (XRD) patterns were collected on a Rigaku MiniFlex 600 with Cu K α radiation (λ = 1.54059 Å) over the angular range 10°≤ 2θ ≤ 80°with a step size of 0.01°. The transmittance and absorption spectra of GCC samples were determined using a spectrophotometer (Cary 5000 UV-vis, Agilent) in a double-beam configuration. Spectra were collected over the range of 200-800 nm with 1 nm slit size, in steps of 1 nm and at a rate of 300 nm min À1 . Raman spectra were collected with a dispersive confocal Raman microscope (Renishaw inVia) for 100-1,050 cm À1 at a step width of 1.5 cm À1 , using the 514 nm excitation line of an Ar-laser. The photoemission spectra were recorded with a spectrofluorimeter (Fluorolog-3, Horiba), using a Xe lamp for excitation.
The initial USL experiment was conducted on a bulk sample with a nonfocused ultrasound transducer and a power generator (E/805/T and K80-5, MEINHARDT Ultrasonics). The maximum output power applied in this case was 100 W at 850 kHz.
For power-dependent USL measurements, GCCs were studied in the form of a finely ground powder dispersed within a gelatin matrix. Luminescence under the US stimulus was recorded by using a compact fiber spectrometer (Shamrock 163, Andor) equipped with a CCD detector (iDus 420, Andor) for data acquisition. The blaze wavelength of this set-up was the same as that of the PL system, i.e., 500 nm. Ultrasonic excitation was provided by a HIFU transducer (H-104, SonicConcepts). The transducer was placed at the bottom of a tank filled with degassed water. The fiber head of the spectrometer used for collecting ML light was aligned close to the center of the cuvette. For charging, a 10 W UV lamp with an emission spectrum peaking at 365 nm was placed at 10 cm from the sample surface, irradiating the sample for 60 s. For power-dependent USL data acquisition, irradiated samples were subsequently kept in the dark for 120 s to allow for the decay of PL afterglow. After this, the HIFU transducer was switched on for a duration of 60 s, operating at 500 kHz with a duty cycle of 95% and an output power within 0.1-2.5 W. Sample photographs were taken under daylight, UV light, and monitoring persistent photoluminescence (pPL) using a digital camera (D610, Nikon). Operando photographs were taken in time series by a CMOS camera (DCC1545M, Thorlabs), for which a HIFU output power of 8 W at 500 kHz with a duty cycle of 90 %was used. A thermographic camera (VarioCAM, Jenoptik) was used to independently monitor the changes in temperature in the environment of the HIFU focus region. For time-resolved photographs of ML emission, the ZGO GCC samples were coated with a polymer film so at to improve US coupling at the sample surface. www.advancedsciencenews.com www.adpr-journal.com