Single-shot linear dichroism optical-resolution photoacoustic microscopy.

Dichroism is a material property that causes anisotropic light-matter interactions for different optical polarizations. Dichroism relates to molecular types and material morphology and thus can be used to distinguish different dichroic tissues. In this paper, we present single-shot dichroism photoacoustic microscopy that can image tissue structure, linear dichroism, and polarization angle with a single raster scanning. We develop a fiber-based laser system to split one laser pulse into three with different polarization angles, sub-microseconds time delay, and identical pulse energy. A dual-fiber optical-resolution photoacoustic microscopy system is developed to acquire three A-lines per scanning step. In such a way, dichroism imaging can achieve the same speed as single-wavelength photoacoustic microscopy. Moreover, the three polarized pulses originate from one laser pulse, which decreases pulse energy fluctuations and reduces dichroism measurement noise by ∼35 %. The new dichroism photoacoustic imaging technique can be used to image endogenous or exogenous polarization-dependent absorption contrasts, such as dichroic tumor or molecule-labeled tissue.

Introduction

Dichroism refers to anisotropic optical absorption property for different light polarizations. This property closely relates to molecular type and alignment. Dichroism can be determined via measuring absorption of different polarized irradiations with respect to the molecular or structural axis [1,2]. Dichroism detection can probe molecular types and structures of some biopolymers (such as protein and DNA) and synthetic polymers [[3], [4], [5], [6]].

As an emerging optical imaging technique, photoacoustic (PA) tomography has seen a wide range of applications, including imaging dichroism in biological tissue. PA imaging also shows great potentials to provide multiple optical properties by being integrated with other optical modalities [[7], [8], [9]]. PA signal is proportional to absorbed optical energy [[10], [11], [12], [13], [14], [15]] and, thus, provides a sensitive approach to measure dichroism. Dichroism-sensitive optical-resolution photoacoustic microscopy (OR-PAM) was first reported by Hu et al. [16]. In Hu’s work, the OR-PAM system detects linear dichroism of amyloid plaques. An electro-optical modulator (EOM) alternates two perpendicularly polarized excitation beams pulse by pulse. While promising, two polarized beam is not enough to fully quantify the linear dichroism. In addition, the delay time between the two perpendicularly-polarized pulses is the laser pulse repetition time. The misalignment between the two polarizations is half step size, which may cause errors in dichroism imaging [17,18]. Qu et al. realized dichroism imaging in photoacoustic computed tomography (PACT) [19]. The polarization directions of dichroic materials, i.e., linear polarizers and bovine tendons, were successfully identified at depths of several transport mean free paths. Zhang et al. [20] developed a quantitative method to determine the dichroism from the anisotropy value. Both methods rotate a half-wave plate (HWP) to control the laser polarization and are time-consuming. This may be problematic in high-speed imaging.

Here, we develop a new OR-PAM technique that can measure linear dichroism in a single shot. The fast polarization measurement reduces misalignment among different polarized excitations, especially in fast scanning OR-PAM. In addition, the three different polarized laser pulses are generated from a single laser pulse, significantly reducing the energy fluctuation between the three pulses.

Materials and methods

System setup

A linear dichroism OR-PAM system is shown in Fig. 1. A 532-nm pulsed laser (7 ns pulse width, linear polarization, VPFL-G-20, Spectra-Physics) is used as the photoacoustic excitation source. The output laser beam is split into three paths using two sets of polarizing beam splitters (PBS1 and PBS2) and halfwave plates (HWP1 and HWP3). The halfwave plates adjust the energy ratio among the paths. One laser path is delivered to an OR-PAM probe via a 2-m polarization-maintaining fiber (PMF, single mode, P1-488PM-FC-2, Thorlabs). The other two paths go through two polarization maintaining fibers (PMF, 30-m and 60-m long, respectively, single mode, HB450-SC, Fibercore) to generate different time delays. Then the two delayed paths are combined at a polarizing beamsplitter (PBS3) and delivered to the OR-PAM probe via another 2-m PMF (single mode, P1-488PM-FC-2, Thorlabs). Four halfwave plates (HWP2, HWP4, HWP5 and HWP8) are used to make sure the laser polarizations are aligned with the fast or slow axis of the polarization-maintaining fibers. Two halfwave plates (HWP6 and HWP7) are used to ensure that the laser beams are reflected or transmitted through PBS3 with high efficiency. The polarization directions of the two delayed paths are aligned with the fast and slow axes of the 2-m PMF, respectively. At the output end, the two 2-m PMFs are connected to the probe with two fiber adapters, and the angle between the fast axes of the two 2-m PMFs is 45°, so that we obtain three laser pulses with three different polarizations: 0°, 45°, and 90°. The three laser pulses have ∼150-ns time interval and unchanged pulse energy ratio among them. In the imaging probe, the two laser beams from the fibers are collimated with an achromatic lens (AC064-013-A, Edmund) and then combined with a beamsplitter. Another achromatic lens (AC064-013-A, Edmund) focuses the combined beam. Then the focused beam is reflected in an optical/acoustic beam combiner (two prisms, PS910 and MRA10-F01 from Thorlabs Inc, glued together with NOA 61 from Norland Products, Inc.) [21], transmits through an ultrasonic lens, and illuminates the sample. Optically induced ultrasonic waves are collected by the ultrasonic lens, transmit through the optical/acoustic beam combiner, and are received by a 50-MHz piezoelectric transducer (V214-BC-RM, Olympus-NDT). Optical excitations from the two fibers are both aligned with the acoustic detection coaxially and confocally to maximize the detection sensitivity. This OR-PAM system acquires three A-lines with different polarized excitations at every scanning step. Raster scanning the PA probe allows for acquiring volumetric images.

Fig. 1

Schematic of the dual-fiber single-shot dichroism OR-PAM system. BS, beamsplitter; HWP, halfwave plate; L, lens; PBS, polarizing beamsplitter; PMF, polarization-maintaining fiber; UL, ultrasound lens; UT, ultrasound transducer; WT, water tank.

Principle

Principle of the linear dichroism measurement is presented as follows. In linear photoacoustic range, the initial pressure can be expressed as [[22], [23], [24]]where is the Grueneisen parameter, is the heat conversion efficiency, is the absorption coefficient, and F is the optical fluence. If the sample is dichroic, varies with the excitation polarization angle. For dichroic samples, assuming the optical axis orientation is , the polarization angle of light is . The depends on and as [16]where is the absorption coefficient when optical axis for target is and the polarized angle for light is . and are the optical absorption coefficient of the direction parallel with and perpendicular to the optical axis, respectively.

If we measure PA signals with three linear polarized pulses at 0°, 45° and 90°, we obtainFor simplicity, we let the three pulses have the same fluence. The dichroism is defined as [16]. Solving Eq. (3), we can determine the dichroism and polarization angle from the three PA measurements.

Results and discussion

Comparison between a linear polarizer and a black ink sample

Fig. 2 shows dichroism OR-PAM of a linear polarizer (Jiateng Video equipment Ltd China) and a black ink sample (unpolarized). Fig. 2a shows the PA images at three polarization angles. The upper row is the results from the linear polarizer sample, whose polarization angle is 135°. The lower row is the black ink sample. The linear polarizer results show obvious different PA intensities at different polarized excitations. The PA intensity at 45° excitation is lower than the other two, and the PA images at 0° and 90° excitations are nearly the same. Because the black ink has no dichroism, the PA images shows no obvious differences at different polarized light excitations. Dichroisms are calculated and shown in Fig. 2b. The averaged dichroism value for the linear polarizer is 0.96 ± 0.003, and the value for the black ink sample is nearly 0. We validate the dichroism of the linear polarizer in transmission mode using a power meter. The extinction ratio of the polarizer equals to 0.99, which is close to the PA imaging result.

Fig. 2

(a) Linear polarizer (upper) and black ink (lower) images at 0°, 45°, and 90°. (b) Calculated dichroism of the linear polarizer (upper) and the black ink (lower) samples.

Phantom experiment

To verify the dichroism imaging, three linear polarizers are positioned in different angles, as shown in Fig. 3. Fig. 3a shows three PA images acquired with three polarized light excitations. The polarizer gives strongest PA signals when its polarization angle aligns with the polarization angle of the excitation light. Fig. 3b shows the computed dichroism of the three linear polarizers. As expected, their dichroism is the same regardless their orientations. Fig. 3c shows the computed polarization angles according to Eq. (4). These results demonstrate that our single-shot OR-PAM can image dichroism correctly. An alternative approach is to measure the dichroism from independent pulses, such as using an EOM to switch between two different polarized laser beams [14]. Compared with the EOM-based method, our approach measures three PA signals originating from one laser pulse and thus can fully quantify the linear dichroism. In addition, the three laser pulses have reduced pulse energy fluctuations and misalignment. The standard deviation of the dichroism results from our approach is 0.0455, 35 % less than the other approach.

Fig. 3

(a) PA images of three linear polarizers excited with polarized light at 0°, 45°, and 90°. (b) Calculated dichroism image of the three linear polarizers. (c) Calculated polarization angle of the three linear polarizers.

Biological tissue experiment

We further demonstrate single-shot dichroism OR-PAM with biological tissue. Congo Red dye is a dichroic molecule. In experiment, 0.01 g Congo Red (C8450, Solarbio) was dissolved in 0.2 ml deionized water, and was used to dye a ∼200-μm-thick chicken breast tissue sample. First, we tested photo bleaching rate of the labelled sample in OR-PAM imaging. The laser pulse energy in experiment was 30 nJ. The averaged PA amplitude decreases ∼58 % per one million pulsed excitations. For raster scanning, the laser spot size is 5 μm and the step size is 1.25 μm. Thus, the sample bleaching at one spot is 0.002784 % for three pulsed excitations. Thus, photo bleaching causes negligible influence to the single-shot dichroism imaging.

Fig. 4a shows a maximum-amplitude-projected PA image of the chicken breast sample. The polarization angle of the excitation light for this image is 0°. The laser repetition rate is 8 kHz for each polarization angle. The scanning step size is 1.25 μm, and the field of view is 1.25 × 1.25 mm2. Fig. 4b shows the spatial profile of averaged PA amplitudes along the B-scan direction at three excitation polarization angles. Dichroism of the sample causes dramatic differences among the three images. Because the time delay among the three pulses is sub-microseconds, the Grueneisen parameter might be changed due to local heating [[22], [23], [24]]. To exclude this possibility, we use two non-polarized laser pulses to excite the sample. The two pulses have the same pulse energy of 30 nJ. The time delay between them is ∼150 ns. As seen in Fig. 4c, the PA amplitude variation associated with the Grueneisen effect does exist, as the second PA amplitude increases ∼3 % due to local heating. Nevertheless, compared with polarization difference-induced difference (∼25 % variation for each different polarization) shown in Fig. 4b, the Grueneisen effect causes ∼12 % error in the dichroism measurement. In the future, this error may be reduced via increasing the time delay among the three pulses or using less pulse energy. Fig. 4d and e show the dichroism and polarization angle distributions of the Congo Red-dyed chicken breast tissue sample. The dichroism of the sample is relatively uniform due to the use of the same dye. The polarization angle distribution, however, shows more variations due to the inhomogeneities of muscle fiber directions.

Fig. 4

(a) OR-PAM of a Congo-Red-dyed chicken breast tissue sample. (b) Averaged PA amplitudes excited with different linear polarization light at 0°, 45°, and 90°. (c) Test of Grueneisen relaxation effect. Averaged PA amplitudes excited with two non-polarized pulses. Time delay between the two pulses is 150 ns. (d) Calculated dichroism image. (e) Calculated polarization angle image. (scale bar: 150 μm).

Conclusion

We present the development of a single-shot linear dichroism OR-PAM. To quantify the linear polarization, the OR-PAM system excites the sample with three laser pulses at different polarization angles. One laser pulse output of a fiber-based laser source is split into three with different polarization angles and different time delays. This approach provides stable energy ratio among the three pulses and thus can reduce the noise in dichroism imaging. Polarization switch is implemented via fiber delay. This offers sub-microseconds polarization switching and thus can reduce misalignment among different polarizations, even in fast scanning. The three laser pulses are delivered to a dual-fiber OR-PAM probe for PA imaging. Experimental results show that dichroism OR-PAM can quantify dichroism and polarization angle in a single raster scanning. This technical advancement offers a new polarization contrast to optical-resolution photoacoustic microscopy and may extend it to broader preclinical and clinical applications.

Declaration of Competing Interest

C.L. and L.W. have a financial interest in PATech Limited, which, however, did not support this work.