Photothermal bleaching in time-lapse photoacoustic microscopy

Abstract

We studied the phenomenon of photothermal bleaching — a gradual reduction of contrast agent particles during repeated scans in photoacoustic microscopy. The dependence of the photothermal bleaching rate on the excitation pulse energy, pulse duration, and the absorber’s size was determined while the laser focal diameter was held constant. Our results showed that, the dependence of the photothermal bleaching rate on the excitation pulse energy differed before and after the absorbers were raised to their melting point by the deposited laser energy. Based on this finding, we suggested an optimal excitation pulse energy, which balances the photothermal bleaching rate and signal amplitude, for time-lapse imaging applications.

Introduction

Photobleaching, a common phenomenon in fluorescence microscopy, occurs when a fluorophore permanently loses its ability to fluoresce due to long exposure to intense excitation light [1, 2]. On the one hand, photobleaching undesirably limits the observation time for monitoring a dynamic process; on the other hand, photobleaching can be beneficially employed in molecular diffusion or motion studies via techniques such as fluorescence-recovery after photobleaching (FRAP) [3, 4] or fluorescence loss in photobleaching (FLIP) [5]. Properly exploiting photobleaching requires a thorough understanding of its dependence on the excitation light power. Such a dependence in fluorescence microscopy has been carefully characterized [1, 2, 6].

However, photobleaching is not exclusively a feature of fluorescence microscopy: It also exists in other imaging modalities. One example is photoacoustic microscopy (PAM) [7, 8], in which the sample is irradiated by a short laser pulse (~ns duration). The absorbed light energy is converted into heat and produces an ultrasonic wave via thermoelastic expansion. The ultrasonic wave is detected by an ultrasound transducer. Contrasts in PAM stem from light absorption by endogenous or exogenous absorbers, such as hemoglobin [9] or nanoparticles [10, 11], respectively. These absorbers can be intentionally photo-destructed using strong excitation light, resulting in a gradual reduction of photoacoustic signals during repeated scans.

The photobleaching mechanism in PAM differs from that in fluorescence microscopy. Upon absorbing a photon, an electron transits from the ground state to an excited state. The electron’s energy can be released via primarily two paths: radiative decay, i.e., fluorescence, or non-radiative decay, i.e., thermal dissipation. In fluorescence microscopy, photobleaching is the photochemical destruction of a fluorophore – during radiative decay the excited electrons are trapped in a relatively long-lived triplet state (lifetime ~ 1 μs), thus allowing the fluorophore a much longer time to undergo irreversible chemical reactions with the environment than would be allowed by the singlet-singlet transition [12]. In PAM, photobleaching is the photothermal destruction of an absorber – the heat generated during non-radiative decay alters the structure of the absorber, thus reducing sample’s absorption coefficient at the original excitation wavelength. As an example, Fig. 1 shows the scanning electron microscopic images of 100-nm gold nanoparticles (GNPs) before and after they were imaged by PAM. The GNPs were fragmented by the deposited laser energy during photoacoustic imaging and thus vanished. To distinguish the photobleaching in PAM from its fluorescent counterpart, we term this photothermally induced absorption change as “photothermal bleaching”.

Fig. 1.

Photoacoustic destruction of contrast agents in PAM. Scanning electron microscopic images of a sample with 100 nm GNPs fixed on a cover glass (a) before and (b) after it was imaged by PAM at an intentionally high fluence of 82.6 mJ/cm².

To begin to exploit the photothermal bleaching in PAM, here we studied its dependence on the laser pulse energy, the absorber’s diameter, and the laser pulse duration while the laser focal diameter was held constant. GNPs were chosen as the targets because they were widely used as contrast agents in photoacoustic imaging applications [10, 11, 13]. Our results revealed that, within the linear excitation range, photothermal bleaching behaved differently before and after the GNPs were raised to their melting point. Below this critical point, the photothermal bleaching rate had weak dependence on the laser pulse energy; while above the melting point, the photothermal bleaching rate increased rapidly. Based on this finding, we suggest a method to determine the optimal excitation laser pulse energy for time-lapse photoacoustic imaging.

Results and discussion

To confine our studies within the linear photoacoustic imaging range, where the photoacoustic signal amplitude is proportional to the laser pulse energy, we first identified the pulse energy range for linear excitation. GNPs of 400 nm in diameter were suspended in water solution and imaged by a PAM system [14] using different pulse energies. Since GNPs diffuse freely in the water solution, the photothermally bleached particles were rapidly replaced, and thus negligible. The measured photoacoustic amplitude versus laser pulse energy is shown in Fig. 2. Below ~120 nJ (fluence: 0.33 J/cm²), photoacoustic amplitude is linearly related to the pulse energy, so laser pulse energies within this linear range were used in the following experiments. The focal diameter of the laser beam was set to 6.8 μm for all experiments, and the laser wavelength was 532 nm.

Fig. 2.

Photoacoustic amplitude vs. delivered excitation laser pulse energy. The 400-nm-diameter GNPs were suspended in water solution and imaged by PAM. The coefficient of determination, R², was 0.990.

Next, the 400-nm-diameter GNPs were fixed on a cover glass to eliminate the effect of diffusion on the measurement of photothermal bleaching rate (see Materials and methods). The photothermal bleaching dynamics were measured under different laser pulse energies, and the results are shown in Fig. 3. The photothermal bleaching rate k was calculated by fitting a double exponential curve to the acquired data (see Materials and methods).

Fig. 3.

Photothermal bleaching of 400-nm-diameter GNPs under different delivered laser pulse energies. The samples were continuously imaged by PAM at 0.5 Hz C-scan rate.

Fig. 4 is a log-log plot of photothermal bleaching rate versus laser pulse energy. Since the photothermal bleaching in PAM originates from instantaneous optical heating, we estimated the temperature rise of a GNP upon a pulse laser excitation and correlated it to the measured photothermal bleaching rate. The light absorption of a spherical GNP obeys Beer’s law, which gives

$$\mathrm{\Delta }E=\int F(1-e^{-{\mu }_{a}l})\mathit{d\; \sigma }=\underset{0}{\overset{\pi /2}{\int }}\frac{I}{A}(1-e^{-{\mu }_{a}D\cos \theta })\pi \frac{D^2}{4}\sin 2\theta d\theta ,$$ (1)

where F is the optical fluence (J/cm²), σ is the projected cross-section perpendicular to the incident light direction, l is the optical path length, μ_a is the absorption coefficient ( ${\mu }_a^{\mathit{Au}}=5.6867\times {10}^5{\mathrm{cm}}^{-1}$), D is the absorber’s diameter, I is the laser pulse energy, A is the focal spot area, and θ is the polar angle in the spherical coordinates. Since the thermal confinement time for a 400 nm GNP in water solution is

$${\tau }_{\mathit{th}}=\frac{d_c^2}{{\alpha }_{\mathit{th}}}=\frac{{(0.4\times {10}^{-4}\mathrm{cm})}^2}{1.3{\mathrm{cm}}^2/\mathrm{s}}=1.2\mathrm{n}s,$$ (2)

(d_c is the characteristic dimension of the heated region, and α_th is the thermal diffusivity of gold), which is longer than the laser pulse duration (1 ns), the optical heating process was regarded as thermally confined, i.e., thermal conduction was negligible during the pulse laser excitation. Consequently, the temperature rise of a GNP approximates

$$\mathrm{\Delta }T=\frac{\mathrm{\Delta }E}{C_V\mathit{\rho V}},$$ (3)

where C_V is the specific heat ( $C_V^{\mathit{Au}}=0.13\mathrm{J}/\mathrm{g}.\mathrm{k}$), ρ is the mass density (ρ^Au =19.3 g.cm⁻³), and V is the volume. By substituting ΔE in Eq. 1 into Eq. 3, the temperature rise ΔT under the excitation of different pulse energies was calculated. The results are shown in Table 1 and labeled next to the corresponding laser pulse energies in Fig. 4.

Fig. 4.

Dependence of photothermal bleaching rate k on the delivered laser pulse energy and temperature rise. I is the laser pulse energy (nJ), and k is the photothermal bleaching rate (s^-1). The slope for the data points with pulse energies > 30 nJ is 3.6 ± 0.3 (95% confidence bounds). The coefficient of determination, R², is 0.996.

Table 1.

Calculated temperature rise of a 400 nm diameter GNP upon excitation by different laser pulse energies.

Laser pulse energy (nJ)	Temperature rise (K)
15	483
25	806
44	1418
58	1869
69	2223
78	2513
90	2900
99	3190

The data in Fig. 4 gives two slopes : below I_c =30 nJ, the photothermal bleaching rate k has a weak dependence on the laser pulse energy; while above it, the slope equals 3.6 ± 0.3, indicating a nonlinear relation between photothermal bleaching rate and laser pulse energy −k ∝ I^3.6. The corresponding temperature rise for the critical point I_c =30 nJ was calculated by Eq. 1 and Eq. 3, yielding

$$\mathrm{\Delta }{T}_c=967K.$$ (4)

Assuming the room temperature was 23 °C, the temperature of the GNP after a pulse laser excitation approximated

$$T=T_{\mathit{room}}+\mathrm{\Delta }{T}_c=990°C,$$ (5)

close to the melting point of gold ( $T_{\mathit{melting}}^{\mathit{Au}}=1,064°C$). This implies that the phase change of the absorber (from solid to liquid state) may contribute to the two different photothermal bleaching behaviors observed within the linear PA excitation range.

Since photothermal bleaching behaves differently around an absorber’s melting point under linear excitation, this dependence suggests an optimal excitation pulse energy for the time-lapse photoacoustic imaging: I_c, the critical value that raises the GNPs to their melting point. Below I_c, raising the excitation pulse energy increases the photoacoustic signal linearly but does not noticeably speed up the photothermal bleaching rate. While above I_c, although increasing the excitation pulse energy can linearly improve the signal, the photothermal bleaching accelerates nonlinearly, significantly shortening the time window for imaging a dynamic event.

Additionally, the dependence of the photothermal bleaching rate on the absorber’s diameter was also investigated. GNPs with diameters of 50 nm, 70 nm, 100 nm, and 400 nm fixed on a glass slide were excited under the same laser pulse energy (I =15 nJ). The sample was continuously scanned at 0.5 Hz C-scan rate until it became totally photothermally bleached. Fig. 5 shows the photoacoustic amplitude versus time for GNPs of different diameters. Because μ_aD > 1 for gold, the energy deposition ΔE is primarily related to D². By contrast, the temperature rise is inversely proportional to D³. As a result, smaller particles sustained higher temperature rises and they had faster photothermal bleaching rates than GNPs with larger diameters.

Fig. 5.

Photothermal bleaching of different diameter GNPs under the same delivered excitation pulse energy (15 nJ). The samples were continuously imaged by PAM at 0.5 Hz C-scan rate.

Different laser pulse durations also affect the photothermal bleaching rate. We excited 100-nm-diameter fixed GNPs with two lasers having the same pulse energy (28 nJ) but delivered over different pulse durations (1 ns and 10 ns). As shown in Fig. 6, the laser pulses of 1 ns duration bleached the GNPs much faster than the 10-ns ones (k = 4.40 s^-1 at 1-ns pulse duration; k = 0.0775 s^-1 at 10-ns pulse duration), indicating that pulse duration plays an important role in the photothermal bleaching process. Although the total deposited energy ΔE was the same for both cases, the heat delivered in 10-ns duration diffuses into a larger volume than that delivered in 1-ns duration. Thus the ultimate temperature of GNPs raised by 10 ns laser pulses is lower than that by 1-ns ones. This fact may explain the accelerated photothermal bleaching under illumination by shorter duration laser pulses.

Fig. 6.

Photothermal bleaching of 100 nm diameter GNPs under two different laser pulse durations. The laser pulse energy delivered was 28 nJ for both pulse durations. The sample was imaged by PAM at 0.5 Hz C-scan rate.

Conclusion

In summary, we studied the phenomenon of photothermal bleaching in PAM. The dependence of the photothermal bleaching rate on the excitation laser pulse energy, duration, and absorber’s diameter was carefully measured and analyzed. Our results showed that, under linear photoacoustic excitation, two different photothermal bleaching behaviors exist around the absorber’s melting point. Based on this finding, an optimal excitation laser pulse energy, which balances the strength of the photoacoustic signal and the bleaching rate, was suggested for time-lapse imaging applications. Additionally, increasing the absorber’s diameter or prolonging the laser pulse duration also beneficially reduces the photothermal bleaching rate.

To the best of our knowledge, this is the first time that the phenomenon of photothermal bleaching was characterized in PAM. This knowledge will be valuable not only for the GNP-based PA imaging, but also for establishing new photothermal-bleaching-assisted techniques such as photoacoustic-recovery after photothermal bleaching or photoacoustic-loss in photothermal bleaching.

Materials and methods

Sample preparation

To eliminate the effect of diffusion during the photothermal bleaching rate measurement, the target GNPs were fixed on top of cover glasses by the following operations [15-17]. The cover glasses were first cleaned in warm detergent (RBS 35, Sigma-Aldrich, USA), flushed with DI water, and dried with nitrogen. Then the cover glasses were put in a plasma oxidizer (Diener Electronics, Germany) for 20 min, followed by another 20 min in a 10% solution of (3-Aminopropyl)-triethoxysilane (Sigma-Aldrich, USA) in anhydrous ethanol. After silanization, the cover glasses were rinsed in anhydrous ethanol and kept at 120 °C for 3 hours. Gold particles with mean diameters of 50 nm, 70 nm, 100 nm (TedPella, USA) and 400 nm (Cytodiagnostics, Canada) were pipetted onto the silanized cover glasses and baked in a 45 °C oven overnight. Before imaging, the samples were washed with DI water on a shaker for 3 hours to remove the unfixed residual gold particles.

Time-resolved measurement of photothermal bleaching dynamics

Time-resolved photothermal bleaching experiments were performed using a previously described voice-coil scanning PAM [14] with 0.1 NA optical objective. The PAM uses a raster-scanned focused ultrasonic transducer coupled with confocal optical illumination to detect light-absorption-induced ultrasound signals. Two 532-nm-wavelength pulsed lasers (Elforlight, Ltd., UK), with 1 ns and 10 ns pulse durations, respectively, were employed in the photothermal bleaching experiments. The excitation laser pulse energy was adjusted by rotating a variable ND filter (Thorlabs Inc., USA) in front of the laser output, and monitored by a custom-made photodiode. Before the experiment, the correspondence between the laser pulse energy at the sample and the photodiode’s measured value was calibrated.

During the photothermal bleaching experiments, the sample was continuously scanned by the PAM at 0.5 Hz C-scan rate until the signals vanished from the image. The image was acquired with 1000 voice coil scanning steps and 20 mechanical stage scanning steps, and rendered as a maximum amplitude projection image. The photoacoustic amplitude of an image was calculated by summing all the signals within the field-of-view (400 × 100 μm²) after removing the background. The photothermal bleaching rate k was calculated by fitting a double exponential curve A exp(−k₁t) +B exp(−k₂ t) to the acquired data and averaging over the exponential indices k =Ak₁ +Bk₂ [6, 18].

Biographies

Liang Gao received his B.S. in physics at Tsinghua University in 2005 and his Ph.D. in applied physics at Rice University in 2011. He is currently a postdoctoral research fellow at Washington University in St. Louis in the laboratory of Lihong V. Wang. His research interests include microscopy, optical design and fabrication, and biomedical imaging.

Lidai Wang received the B.A.Sc. and M.A.Sc. degrees in instrument science and technology from the Tsinghua University, Beijing, China, and received the Ph.D. degree in mechanical engineering from the University of Toronto, Ontario, Canada. His Ph.D. work was on microrobotics. He is currently working as a postdoctoral research associate in biomedical engineering at the Washington University in St Louis, Missouri, USA. His current research work focuses on photoacoustic imaging, sensing, and their biomedical applications.

Chiye Li reveived his B.S. in Life Sciences in 2007 from University of Science and Technology of China in Hefei, China. He is currently a Ph.D. student in Biomedical Engineering at Washington University in St. Louis. His research interests involve the application of photoacoustic imaging techniques in biological and medical study.

Alejandro Garcia-Uribe received the M.Sc. and Ph.D. degrees in electrical engineering from the Department of Electrical and Computer Engineering, Texas A&M University, College Station. He is currently a Post-Doctoral Researcher at the Optical Imaging Laboratory, Department of Biomedical Engineering, Washington University in St. Louis. His research interests include biomedical optics, micro-sensors, micro-fabrication, biomedical image analysis, pattern recognition, and digital signal processing.

Lihong V. Wang earned his PhD degree at Rice University, Houston, Texas and currently holds the Gene K. Beare Distinguished Professorship of Biomedical Engineering at Washington University in St. Louis. His laboratory invented or discovered functional photoacoustic tomography, dark-field confocal photoacoustic microscopy (PAM), optical-resolution PAM, photoacoustic Doppler effect, photoacoustic reporter gene imaging, focused scanning microwave-induced thermoacoustic tomography, the universal photoacoustic or thermoacoustic reconstruction algorithm, frequency-swept ultrasound-modulated optical tomography, time-reversed ultrasonically encoded (TRUE) optical focusing, sonoluminescence tomography, Mueller-matrix optical coherence tomography, optical coherence computed tomography, and oblique-incidence reflectometry.