Real‐Time Thermal Imaging based on the Simultaneous Rise and Decay Luminescence Lifetime Thermometry

The most reliable technique of remote temperature sensing considers either the rise or the decay transient of a luminescent temperature probe. Here, real‐time visible (Vis) and near‐infrared (NIR) thermal imagings based on the simultaneous measurement of the emission rise and decay of transition or lanthanide metal activator are described. A single pulse time‐gated detection method that allows the real‐time mode of the temperature measurement, the emission detection at its highest temporal intensity, and the tunability of the emission detection in terms of time delay and gate width is proposed. Cr‐ZnGaGeO4/ZnGa2O4; Er, Ho, Yb–Y2O3 and Er, Ho, Yb–β‐NaYF4 nanoparticles that display luminescence in the Vis to NIR range (450–1200 nm) with timescales varying from 10 to 100 ms are selected. Maximum relative temperature sensitivity in Vis to Vis and NIR to NIR imaging which exceeds up to a factor of two the values obtained by the standard average lifetime method is achieved. This method applies to any lifetime‐based luminescent thermometer, opening a new avenue in designing accurate and straightforward lifetime thermal imaging systems operating in the Vis to NIR range.

DOI: 10.1002/adpr.202100208 The most reliable technique of remote temperature sensing considers either the rise or the decay transient of a luminescent temperature probe. Here, real-time visible (Vis) and near-infrared (NIR) thermal imagings based on the simultaneous measurement of the emission rise and decay of transition or lanthanide metal activator are described. A single pulse time-gated detection method that allows the real-time mode of the temperature measurement, the emission detection at its highest temporal intensity, and the tunability of the emission detection in terms of time delay and gate width is proposed. Cr-ZnGaGeO 4 /ZnGa 2 O 4 ; Er, Ho, Yb-Y 2 O 3 and Er, Ho, Yb-β-NaYF 4 nanoparticles that display luminescence in the Vis to NIR range (450-1200 nm) with timescales varying from 10 to 100 ms are selected. Maximum relative temperature sensitivity in Vis to Vis and NIR to NIR imaging which exceeds up to a factor of two the values obtained by the standard average lifetime method is achieved. This method applies to any lifetime-based luminescent thermometer, opening a new avenue in designing accurate and straightforward lifetime thermal imaging systems operating in the Vis to NIR range. and decay times of the luminescent activator (either transitional metal or lanthanide) are monitored. Several nanosized optical materials are investigated, such as Cr 3þ -ZnGaGeO 4 /ZnGa 2 O 4 , Er, Ho, Yb-Y 2 O 3 and Er, Ho, Yb-β-NaYF 4 . These materials display emissions spanning the 450 to 1200 nm range under direct excitation into Cr 3þ and Yb absorptions with timescales that vary from 10 to 100 ms. The achieved maximum relative temperature sensitivity is 2% K À1 at 490 K for 0.05Cr-ZnGaGeO 4 (Vis (visible) to Vis thermal imaging) and 0.93% K À1 at 313 K for Ho, Er, Yb-Y 2 O 3 nanoparticles (NIR (near-infrared) to NIR thermal imaging), which exceed up to a factor of two those obtained using the standard average lifetime method. Our method can be applied to any lifetime-based luminescent thermometer opening, thus a new avenue in designing accurate and straightforward lifetime thermal imaging systems.

Summary of Structural and Luminescence Properties
Figure S1, Supporting Information, gather the diffraction patterns and selected TEM images of the investigated nanoparticles.

Vis-to-Vis Lifetime Thermometry
We first assess the performance of 0.05Cr-ZGO, 1Cr-ZGO, and 1Cr -ZGGO for Vis lifetime thermometry and Ho, Er, Yb-Y 2 O 3 and Ho, Er, Yb-β-NaYF 4 for NIR lifetime thermometry using the average lifetime approach. [7,18] The average lifetime represents the thermometric parameter being estimated by integrating the area of the emission decays normalized at the maximum intensity.
The relative thermal sensitivity (S r ) [1] is usually used as a figure of merit to compare different thermometers, independently of their nature www.advancedsciencenews.com www.adpr-journal.com where Δ is the thermometric parameter, and dΔ is the variation of the thermometric parameter with temperature (dT). The temperature evolution of the 0.05Cr-ZGO, 1Cr-ZGO, and 1Cr-ZGGO emission decay, average lifetime, and relative thermal sensitivity in the range of 303-533 K are illustrated in Figures 2a and S3a-c, Supporting Information. We highlight that at 450 nm excitation wavelength, 0.05Cr-ZGO, 1Cr-ZGO, and 1Cr-ZGGO do not exhibit any persistent luminescence. Persistent luminescence requires UV or X-ray excitation [20,22,25] for activating the intrinsic traps. The emission decays are monitored at 700 nm using a laser frequency of 10 Hz with 10 ms pulse width. The Cr 3þ 2 E-4 A 2 transition is spin forbidden, leading to a long emission decay in the ms range. In ZGO/ZGGO hosts, Cr 3þ 2 E level is in thermal equilibrium with the 4 T 2 level. The 4 T 2 -4 A 2 transition is spin allowed and is characterized by a shorter lifetime value. With thermal activation, the contribution of the spin allowed transition increases at the expense of 2 E-4 A 2 transition, resulting in a reduction of the effective lifetime. [26] With increasing temperature from 303 to 533 K, the average lifetime drops from 8.75 to 1 ms for 0.05Cr-ZGO due to the enhanced contribution of the temperature-dependent 4 T 2 -4 A 2 emission ( Figure 2a,b). [26] For 1Cr-ZGO and 1Cr-ZGGO, with increasing temperature from 303 to 533 K, the average lifetime drops from 3.34 and 0.87 ms to 0.24 and 0.12 ms, respectively ( Figure S2b,e, Supporting Information). From the temperature evolution of the average lifetimes, S r is calculated using Equation (1) and its temperature evolution is presented in Figures 2c and S2c,f,S4a, Supporting Information. 0.05Cr-ZGO exhibits a maximum S r of 1.25% K À1 at 493 K. This value places 0.05Cr-ZGO on the top two most sensitive lifetime thermometers based on Cr-ZGO/ZGGO. [7,18,26] 2.3. NIR to NIR Lifetime Thermometry For Ho, Er, Yb-Y 2 O 3 , and Ho, Er, Yb-β-NaYF 4 nanoparticles (Figures 2b and S3, Supporting Information), the emission decays monitored the Ho emission ( 5 I 8 -5 I 6 ) at 1200 nm upon 973 nm excitation at 50 Hz. Since the main application of the NIR lifetime thermometer is for bio-imaging, the temperature assessment range was limited to the physiological range (303-323 K). As shown in Figure 2e the average lifetime decreases from 0.5 to 0.27 ms when heating from 303 to 533 K. Ho, Er, Yb-Y 2 O 3 nanothermometer showing the best S r value of 0.2% K À1 . For Ho, Er, Yb-β-NaYF 4 case, the Ho average lifetime decreases from 4.5 to 2.8 ms, respectively ( Figure S3, Supporting Information). The maximum S r value in the same physiological temperature range is 0.17% K À1 for Ho, Er, Yb-β-NaYF 4 ( Figure S4, Supporting Information).

Real-Time Imaging Using Single-Pulse Time-Gated Detection
In single pulse time-gated detection, parameters such as the laser pulse width, frequency, as well as delay and gate parameters are optimized to enhance the thermometer performance. This method requires a relatively simple experimental setup: a laser diode, a time-gated camera synchronized with a signal generator, and appropriate optical filters (Figure 3).
The Cr 3þ nonradiative processes are temperature-dependent and lead to accelerated dynamics of the monitored emission. [12,27,28] According to ref. [28], the rise time is temperaturedependent (Arrhenius dependence) following Equation (1) shows the relationship between the rise time constant and temperature where τ r 0 is the rise time constant at 0 K, τ r T is the rise time constant at temperature T. E is the activation energy of the electronic transitions in the luminescent ions, and k is the Boltzmann's constant. The rise time in Cr 3þ luminescence is generally disregarded in the literature due to used excitation conditions (short pulse and low power density). When an excitation source initiates, emission begins to build up until, after a sufficient time, a constant level is reached. [29] Reference [6] observed that the excitation laser pulse width and power density affect the rise time. Therefore, we employed a constant laser power density for both Vis and NIR lasers.
The emission dynamics usually follows a law of exponential increase, as shown in Equation (2): where t is the time from the start of excitation, IðtÞ is the phosphorescent intensity at time t, and I e is the fully excited emission intensity. The variation with temperature of the luminescence lifetime τ can be represented by the Mott-Seitz model [12] : Figure 3. Experimental setup used for real-time thermal imaging which includes a thermal stage to control the temperature, a time-gated camera with appropriate optical filters, and a fiber-coupled excitation laser synchronized by a signal generator (not shown in the Figure).
www.advancedsciencenews.com www.adpr-journal.com  where τ 0 is the radiative lifetime, C ¼ τ 0 τ NR0 and it is assumed that the rate of radiative processes is independent of temperature (A NR ð0Þ ¼ const: ¼ 1 τ NR0 ), while the probability of nonradiative process has a temperature dependence described by a Boltzmann factor of the form A NR ð0Þ Â exp ÀΔE kT À Á , and A NR ð0Þ ¼ 1 τ NR0 is the nonradiative transition rate when T ¼ 0 and ΔE is the activation energy of the nonradiative process. [12] Similar processes occur in the Ln counterpart with different time scales. The pulse width varied from 1 to 15 ms. The 15-ms pulse width was found to be the optimum value allowing us to achieve signal saturation during the excitation pulse while measuring the complete timescale of the emission rise (Figure 4a). Since Cr 3þ emission displays a long emission timescale of up to 75 ms, the chosen laser frequency was 10 Hz. Next, to find the optimum detection gate-width in which both integrated emission rise and the decay times are most sensitive to temperature, we varied the gate-width from 1 to 15 ms (the pulse width limits the maximum gate-width) and calculated the respective relative thermal sensitivity (Figure 4). Gate 1 represents the integrated area of the emission during the excitation pulse within the selected widths (1-15 ms) at 0 ms delay from the start of the laser pulse (the dark grey rectangle labelled as G1 in Figure 5). Gate 2 represents the integrated area from 0 and 1 ms delay after the laser pulse within the 1-15 ms selected widths (the dark grey rectangle labelled as G2 in Figure 5).
Since both the emission rise and decay accelerate with temperature, the value of Gate 1 increases while the value of Gate 2 decreases. We define the thermometric parameter (Δ) as the ratio between the selected Gates 1 and 2 Figure 5. Temperature evolution of the emission decays, thermometric parameter, and the estimated relative thermal sensitivity of 0.05Cr-ZGO (a) and Ho, Er, Yb-Y 2 O 3 (b). Temperature varied from 303 to 473-573 K. Cr 3þ emission is monitored at 700 nm upon 450 nm excitation using a 15 ms pulse width or upon 973 nm excitation using a 10 ms pulse width. Ho emission is monitored at 1200 nm upon 973 nm excitation using a 4 ms pulse width. The optimum integrated areas of Gate 1 (G1) and Gate 2 (G2) are illustrated as dark grey rectangles in a-b.
www.advancedsciencenews.com www.adpr-journal.com Δ ¼ Integrated area of the emission within Gate 1 Integrated area of the emission within Gate 2 (5) Figure 4g,h compares the obtained S r evolutions for 1, 5, 10, and 15 ms widths for Gates 1 and 2 with Gate 2 starting from 0 and 1 ms delay after the laser pulse. When integrating the Gate 2 at 0 ms delay after the laser pulse, the maximum S r is 1.2% K À1 for 15 ms optimal width at 450 K. A value of 1 ms delay after the laser pulse (Gate 2) boosts the variation of the emission decay time with temperature keeping the signal to noise ratio (S/R) above 100 (Figure 4e,f ) increasing the maximum S r to 2% K À1 at 465 K. If the gate-width is too narrow, the temperature dependence of Gates 1 and 2 cannot be accurately measured. For 0.05Cr-ZGGO, the optimum width that generates the best relative thermal sensitivity is 15 ms, with Gate 1 at 0 ms delay from the start of the laser pulse and Gate 2 at 1 ms delay after the laser pulse (Figure 4h). Figure 5a illustrates the temperature evolution of Cr 3þ emission dynamics within and after the 450 nm excitation pulse. Increasing the temperature from 303 to 533 K, Gate 1 increases from 10.1 to 13.9 ms while Gate 2 diminishes from 5.7 to 0.2 ( Figure 5a). The thermometric parameter (Equation (5)) increases from 1.77 to 67 using the single-pulse time-gated detection method. The single pulse time-gated method yields the highest S r value of 2% K À1 , compared to 1.25% K À1 obtained using the average lifetime method (Figure 5a).
Using a similar procedure as described for 0.05Cr-ZGGO (Figure 4), we determined an optimum width of 0.5 ms (Gate 1 ¼ Gate 2 ¼ 0.5 ms) with 0 ms delay from the start and after the laser pulse when monitoring the Ho emission dynamics at 1200 nm. For Ho, Er, Yb-Y 2 O 3, the excitation source was a 973 nm laser with 4 ms pulse width at 50 Hz frequency, allowing us to achieve signal saturation during the excitation pulse while measuring the complete timescale of the emission rise. The maximum S r is 0.5% K À1 which is twice the S r value of 0.25% K À1 obtained using the average lifetime method as the thermometric parameter ( Figure 5b).
As illustrated in Scheme 1, in real-time thermal imaging using 0.05Cr-ZGO and Ho, Er, Yb-Y 2 O 3 , the time-gated camera records the images denoted as Frames 1 and 2 for each excitation laser pulse using the optimum pulse-width, frequency, gate-widths, and delay after the laser pulse determined previously.
For thermal imaging, the thermometric parameter is estimated as the ratio of Frames 1 and 2 (Scheme 1) values in the 303-533 K range

Δ ¼
Integrated pixel values from the Frame 1 image recoded within Gate 1 Integrated pixel values from the Frame 2 image recoded within Gate 2 (6) Figure S5, Supporting Information, shows the calculated Δ value and the calibration curve (ΔðTÞ for 0.05Cr-ZGO, Ho, Er, and Yb-Y 2 O 3 . The obtained Δ offers excellent stability over 1000 pulses excitation, and the method repeatability is around 99.2% for 0.05Cr-ZGO and 98.5% for Ho, Er, Yb-Y 2 O 3 . The acquisition system does not affect the thermometer performance; since either a PMT or a time-gated camera presents similar thermometer performance ( Figure S6, Supporting Information). Figure 6 shows the real-time thermal images monitored using the single-pulse time-gated detection for 0.05Cr-ZGO and Ho, Er, Yb-Y 2 O 3 samples (video files are included in Supplementary Information). The thermal stage temperature varied from 300 to 500 K, and a selection of three thermal images was used to show the change of temperature across the sample.
The best temperature resolution (or temperature determination uncertainty) obtained for 0.05Cr-ZGO is %0.1 K for temperature between 350 and 450 K. For Ho, Er, Yb-Y 2 O 3 , the best temperature resolution is obtained at 373 K and is %0.2 K. ( Figure S7, Supporting Information). The accuracy of the method was tested by placing water media in the excitation and emission path of 0.05Cr-ZGO and Ho, Er, Yb-Y 2 O 3 at 333 K ( Figure S8, Supporting Information).
When the water cuvette is placed in either the excitation or emission optical path, we observe that the temperature readout over 1000 pulses varies by a maximum of 2 K for 0.05Cr-ZGO and 4 K for Ho, Er, Yb-Y 2 O 3 ( Figure S8, Supporting Information). The temperature readout differences are generated by the light scattering and absorption in water that increases the background and thus reduces the signal to noise ratio. The proposed single-pulse time-gated thermal imaging is an accurate Scheme 1. Schematic representation of the real-time thermal imaging method using single pulse time-gated detection. Frames 1 and 2 are acquired during and after the laser excitation pulse. The thermal image is obtained by the ratio of Frames 1 and 2 for every excitation pulse. Both emission rise and decay are monitored at their highest temporal intensity.
www.advancedsciencenews.com www.adpr-journal.com temperature readout method as the absorption of the medium does not seem to distort the temperature readout [4,5] significantly. A comparison of thermal imaging approaches reported in the literature based on luminescence lifetime thermometry and their thermometers performance are summarized in Table S1, Supporting Information. Previously reported methods for lifetime thermal imaging present a few drawbacks: 1) long acquisition (0.6 s in ref. [14]) time, 2) long processing time (35 s for evaluation time using decay slope analysis in ref. [17]), and 3) low excitation frequency (from 1 [5] to 4 Hz [5,30] with no details regarding the acquisition time). So far, the dual-frame lifetime method was employed to monitor the luminescence decay to get a single pulse thermal image. [15] Upon excitation at 355 nm with 10 ns pulse, the authors recorded two frames of the Bi 3þ -ScVO 4 thermographic phosphor at 610 nm emission using an interline CCD camera, achieving a maximum S r of 6% K À1 at 333 K. [15] The proposed single-pulse time-gated thermal imaging presents a short processing time (depending on the image resolution, pixel depth, and processing power, it should be less than 1-10 ms) and reduces the equipment cost necessary due to the lower frame rate (<100 Hz). The method is not limited to thermal imaging as it can be adapted to any lifetime dependent luminescent sensors or even for multiplexing and anticounterfeiting applications. [31] In short, these advantages refer to the real-time mode of the temperature measurement, the emission detection at its highest temporal intensity and the tunability of the emission detection in terms of time delay and gate width.

Conclusions
We present a new real-time Vis and NIR thermal imaging method based on the simultaneous measurement of the emission rise and decay of a transition metal or lanthanide activator. A single pulse time-gated detection method is described, allowing a real-time mode of the temperature measurement, the measurement of emission at its highest temporal intensity, and the tunability of the emission detection in terms of time delay and gate width. The method is validated using Cr-ZnGaGeO 4 / ZnGa 2 O 4 , Er, Ho, Yb-Y 2 O 3 and Er, Ho, Yb-β-NaYF 4 nanoparticles, which display a significant variation of luminescence timescales (10-100 ms) in the Vis to NIR range (450-1200 nm). To the best of our knowledge, this is the first report on real-time thermal NIR imaging using lifetime thermometry, presenting clear advantages over the existing thermal imaging approaches.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.  Supplementary Information). The Cr 3þ emission is monitored at 700 nm upon 450 nm excitation using a 15 ms pulse width. The Ho emission is monitored at 1200 nm upon 973 nm excitation using a 4 ms pulse width.