In this study, octaethylporphine platinum (Pt(II) octaethylporphine, PtOEP), a transition metal complex, is used as the phosphorescent material, and the molecular structure is shown in Figure 2. It is in powder form at room temperature and is insoluble in water. It is easily soluble in organic solvents such as toluene (Ashkenazi et al., 2008). Research has found that PtOEP is an oxygen-sensitive material, and its luminescence will be quenched by oxygen. The peak absorption wavelengths of the singlet and triplet states of PtOEP are 532 and 740 nm, respectively. We use 10 ml of toluene solution to dissolve 7.38 mg of the PtOEP powder to obtain a mixed solution with a final concentration of 1 mM.

As mentioned before, black ink was used to imitate the melanin in-body model to test the ability of TTD to remove strong background signals.

The semicircular ring array has 128 elements, the center frequency is 3 MHz, and the average relative bandwidth is 60%, but 32 elements of the top are hollowed out for wiring. 128 elements are equally spaced and evenly distributed on a semicircular arc with a radius of 60 mm and an angle of 180°, while ensuring the focus in the z-axis (shown in Figure 3). The advantage of such an array design is that for each target object, a piezoelectric element is perpendicular to its edge direction, and all the information of the target object can be obtained at one time, and the image quality can be optimized by calculating the sound velocity. The 64-channel signal passes through the port and is input to the amplifier with a multiple of 40 dB. Each amplifier contains eight input channels and four output channels. The selection of input and output channels is controlled by a 2 1-channel selection multiplexer. 128 channels can output 64 channels at a time and can input 8 8-channel capture cards. The sampling rate of each capture card is 50 MS/s with a 24-bit resolution.

A DG645 digital delay/pulse generator (Stanford Research Systems, CA, United States) is used to control the synchronization between the two lasers. It can provide a precisely defined pulse repetition rate up to 10 MHz. The input is the synchronization trigger signal of the Nd:YAG laser. After a time delay controlled by the DG645, the OPO laser is triggered. We set different time delays ranging from −2 to 8 μs between the two lasers with ±3.5 V input signal amplitude.

The schematic of the system is shown in Figure 4. A Q-switched Nd:YAG laser (Beamtech Optronics Co., Ltd., Beijing, China) is used to generate excitation pulse with a wavelength of 532 nm, a repetition frequency of 10 Hz, a pulse width of 10 ns, an energy of 10 mJ, and a beam width of 6 mm. The SpitLight 600 OPO laser (Innolas Laser GmbH, Munich, Germany) is used to generate probe pulse with 740 nm wavelength, 20 Hz repetition frequency, 5 ns pulse width, 13 mJ energy, and 2 mm beam width. The beam is focused by the lens and coupled into a quartz fiber. After the YAG laser externally triggers the DG645, DG645 generates a set of delay signals (−2, 0, 0.5, 1, 2, 4, and 8 µs) to delay the synchronous output trigger signal that controls the OPO laser to emit probe pulse. The two laser beams are directed to overlap on the target area. The ultrasonic transducer converts the acoustic signal into an electrical signal, and a digital oscilloscope is used to simultaneously observe the delays of various trigger signals amplified by the signal amplifier and the triplet state photoacoustic signals that vary with different delay times. The data are collected by a data acquisition card and analyzed by a computer.

Figure 5A shows the PA signals of PtOEP excited by different wavelengths. The red line is the PA signal of the glass tube with PtOEP stimulated by 740 nm beam (PA₇₄₀). The green line is the PA signal of the glass tube with PtOEP stimulated by 532 nm beam (PA₅₃₂). The black line is the PA signal of the glass tube with PtOEP stimulated by 532 and 740 nm combined beam. The blue line is the PA signal using 532 and740 nm combined beam subtracts the PA signal stimulated by 740 nm beam and then subtracts the PA signal using 532 nm beam. This last differential PA signal is called the TTD signal (PA_transient), described as follows: PA t r a n s i e n t , τ = PA 740 + 532 , τ − PA 532 , τ − PA 740 , τ , (1) where τ is the pump-probe delay time:

Because the triplet state photoacoustic signal only can be generated by the phosphorescent material, after the TTD process, the area with the phosphorescent material keeps the obvious signal (Figure 5B). Therefore, the obtained triplet state photoacoustic signal can be used for image reconstruction to remove the background signal.

When the excitation light is removed, the time required for the phosphorescence intensity to drop to 1/e of the maximum phosphorescence intensity during excitation is called phosphorescence lifetime; we fit the photoacoustic signals after multiple acquisitions and differentials and obtain the phosphorescence lifetime photoacoustic signal as shown in Figure 6. As the delay time increases, the TTD signal gradually decreases. By calculating the phosphorescence lifetime according to the reference by Shaoqi (Shao et al., 2013), we collect a series of transient power amplifier signals under different pump-probe delays (τ) and fit them to an exponential decay function, as described in Eq. 2, and we get that the phosphorescence lifetime of PtOEP is 2.204 µs: Ρ A transient = Ρ A transient , 0 e - 1 T τ , (2) where T is the lifetime, PA_transient is the transient photoacoustic signal, and PA_{transient, 0} is the maximum photoacoustic signal.

Figure 7 shows the TTD imaging of PtOEP. The time delay is the period between the excitation pulse (532 nm) and probe pulse (740 nm). Minus time means the probe pulse stimulated before excitation pulse. When the time delay is −2 µs, no TTD signal can be obtained. When the time delay is 0 µs, the glass tube with PtOEP can be imaged using the semicircular ring array. As the time delay increases, the amplitude of the TTD signal decreases. The results indicate that the system and method in this study are effective for TTD imaging.

In order to test the background removal ability of TTD imaging, we set a glass tube with black ink close to a tube with PtOEP (shown in Figure 8), the black ink with strong light absorption is widely used as the phantom of melanoma, and its absorbance at 532 nm wavelength is 3.052 and at 740 nm wavelength is 2.489. Figure 9A shows the PA images stimulated by 532 and 740 nm combined beam. In Figure 9B, we can observe the image of PtOEP because the absorption peak of PtOEP is 532 nm. In Figure 9C, the PA images stimulated by 740 nm are shown. Compared with the signal from black ink, the signal of PtOEP is too weak to observe. Table 1 shows the signal amplitude of Figure 9A. Although the amplitude of PtOEP is 30.7718 mV, it is weak compared with the amplitude of black ink. These results imply that, even using the exogenous contrast such as PtOEP, we cannot obtain the specific molecular images with strong background like melanin. Figure 9D shows the image after the TTD process according to Eq. 1, the background signal is removed successfully, and the signal of PtOEP survived. As Table 1 shows, although the amplitude of the PtOEP signal is reduced from 30.7718 to 10.1934 mV (a reduction of about 67%), the amplitude of the strong black ink signal is reduced from 98.9145 to 3.2878 mV (a reduction of about 96%), and the ratio between PtOEP signal and black ink signal is increased from 0.3111 to 3.1003, which increases to about 10 times the previous one. Therefore, TTD imaging can successfully remove the strong background signal like black ink.

We then study the effect from signal averages. Figure 10 shows that as the average number increases, the image intensity becomes weaker and the image quality becomes better. Table 2 lists the signal amplitude and ratio between black ink and PtOEP. The results indicate that as the number of averages increases, the signal amplitude continues to decrease, but the ratio to the background signal continues to increase, indicating that the image signal continues to stabilize, the jitter decreases, and the signal and image are more realistic.

However, if the number of averages increases, the acquisition time will also increase. This means that more time is needed for signal acquisition, which will increase the uncertainty of the system, and due to the extension of time, the laser crystal energy will continue to decrease, which will affect the accuracy of the experiment.

Based on the obtained results, we can draw the following conclusion: as the number of averaging increases, although the signal ratio continues to increase, the PtOEP signal continues to decrease, which will affect the observation of the image. In summary, choosing 50 times the average signal can not only meet the request of the imaging but also provide high contrast.