Optical focusing inside scattering media with time-reversed ultrasound microbubble encoded light

Focusing light inside scattering media in a freely addressable fashion is challenging, as the wavefront of the scattered light is highly disordered. Recently developed ultrasound-guided wavefront shaping methods are addressing this challenge, albeit with relatively low modulation efficiency and resolution limitations. In this paper, we present a new technique, time-reversed ultrasound microbubble encoded (TRUME) optical focusing, which can focus light with improved efficiency and sub-ultrasound wavelength resolution. This method ultrasonically destroys microbubbles, and measures the wavefront change to compute and render a suitable time-reversed wavefront solution for focusing. We demonstrate that the TRUME technique can create an optical focus at the site of bubble destruction with a size of ∼2 μm. We further demonstrate a twofold enhancement in addressable focus resolution in a microbubble aggregate target by exploiting the nonlinear pressure-to-destruction response of the microbubbles. The reported technique provides a deep tissue-focusing solution with high efficiency, resolution, and specificity.


Supplementary Note 1 | Focusing on size-decreased microbubble.
The mechanisms of microbubble destruction include fragmentation and diffusion. While the fragmentation occurs within the timescale of microseconds, the acoustically driven diffusion effect takes up to tens of milliseconds 1 . Fundamentally, lower ultrasound pressure and larger microbubble diameter result in a longer time for the gas dissolution in the surrounding medium. In our experiment, we found that acoustically driven diffusion of microbubble gas takes longer in gel than in saline (PBS). Ideally for TRUME, the microbubbles are destroyed within the ultrasound duration (~29 ms in our experiment) as shown in Supplementary Movie 1 (scale bar: 10 µm). However, we also observed incomplete destruction of microbubbles within that period (Supplementary Movie 2), which typically occurs with low ultrasound pressure and large microbubble size. Nevertheless, this effect also enables an optical focus ( Supplementary Fig. 7) to be formed. Alternatively, a longer insonation time allows the microbubble gas to completely dissolute into the surrounding medium (Supplementary Movie 3).

Optical setup
The optical setup schematic diagram is shown in Supplementary Fig. 1. The setup can be partitioned into four major modules: light source regulation, DOPC, sample operation, and observation.
The light source regulation module provides three laser beams: a sample beam, reference beam, and playback beam. The intensity of each laser beam is regulated by a waveplate and a polarizing beam splitter. Both the sample beam and the reference beam are frequencyshifted by 50 MHz using acousto-optic modulators (AOMs). During hologram recording, the signal to AOM1 is phase-shifted following the trigger signals to the camera of the DOPC system. A path length matching arm is applied to the reference beam, as our laser has a relative short coherent length (~7 mm). All beams are spatially filtered using single mode fibers and collimated for DOPC application. The polarization of all beams is set to the horizontal direction, matching the polarization of the SLM. During DOPC recording, the shutter in the playback beam path (SH1) is closed and shutter in the sample beam path (SH2) is open. Both shutters flip after recording.
The DOPC module consists of a sCMOS camera (CAMS) and a SLM. These two components are precisely aligned 2 through a plate beam splitter (BSP) so that the optical field recorded by the camera can be reconstructed faithfully in space by the SLM. A polarizer is used to match the polarization of the sample beam and SLM. The sample beam and reference beam are combined through a 90% beam splitter (BST). The recorded phase of the sample beam is conjugated and sent to the SLM, which modulates the collimated playback beam. A beam compensator is used to compensate for the wavefront distortion introduced by the plate beam splitter.
In the sample operation module, light transmitted through the sample is collected by a 50 mm lens (L4) whose focus is positioned in the middle of the sample. An aperture (AP3) is used to control the speckle area, which covers ~9 pixels in our case. A stage positioned ultrasound transducer is placed above the sample (normal to the diagram) but it is flattened to the left hand side in this two-dimensional diagram. The mirror (M4) for fluorescence signal detection is placed below the sample but also flattened in this diagram.
The observation module has an imaging system and a fluorescence detection system. In the imaging system, a ×20 objective (OB) and a 100 mm tube lens (L6) are arranged in a 4-f configuration through which the microbubbles in the sample are imaged onto a camera (CAM). In the experiment that demonstrates cytometry, the fluorescence emitted from the sample is collected by a lens (L5) and filtered by a fluorescence filter (FF). A single photon counting avalanche photodiode (APD) is used to detect the fluorescence signals.

Electrical signal flow for TRUE
The TRUE optical focus shown in Figure 2e was created using the same setup but with a modified signal flow (not shown in Supplementary Fig. 2). Instead of modulating the sample beam with an AOM, TRUE used single cycle ultrasound to shift the frequency of the sample beam so that the ultrasound encoded light was measured. The reference beam was modulated at 50.010 MHz rather than 50 MHz so that the camera averaged out the pattern due to interference between unmodulated light and reference beam, which was otherwise locked by the 20 kHz laser pulses. In this case, each ultrasound pulse was phase-inversed from the preceding pulse so that only ultrasound modulated light is locked by the reference beam 3 . The field measurement for TRUE was the same as that in TRUME.

PBR measurement
PBR was calculated from a region of interest (ROI, 600 pixels (x) by 40 pixels (y)) centred on the tube for both TRUME and TRUE foci. First, a one dimensional focus profile was extracted from the ROIs for peak calculation. For the TRUME focus, the row across the center of the focus was used as the one-dimensional focus. For the TRUE focus (Figure 2e), the onedimensional focus was calculated by averaging the ROI in the y direction. In the second step, a one-dimensional Gaussian profile was fitted to each one-dimensional focus profile. The amplitude of the fitted Gaussian profile was considered as the peak intensity. Finally, to calculate the background light intensity, the SLM was shifted by 10 pixels in both x and y directions after the focus was made and the background image (e.g. Figure 3d in the main article) captured. The background intensity was estimated by averaging the ROI on the background image. The PBR was then calculated by taking the ratio between peak intensity and background intensity.

Modulation efficiency measurement
To compare the modulation efficiency of guidestar used in TRUE and TRUME, we measured the light fields modulated by the ultrasound focus and microbubble in clear media respectively and calculated the modulation efficiency. The field images of the guidestars were captured using the camera of the DOPC system.

a. Ultrasound modulation
We measured the ultrasound modulated light field based on the lock-in scheme and 4-phase stepping holography used for the phase recording of a TRUE process (see Electrical signal flow). Mathematically, the optical field on the sensor plane can be decomposed into a reference field r () where A and  denote the amplitude and phase of each complex field with associated subscripts, 0  and a  are the frequency of the light and ultrasound respectively. It should be noted that the ultrasound is assumed to be continuous wave here for simplicity. The light intensity on the sensor plane at the k th stepping phase can then be expressed as By substituting the field terms with the right side of Supplementary Equation (1) The AC term With the four measured interference patterns (at 0, 1, 2 k  and 3 ), we are able to compute the modulated field which is optically amplified by the reference light field: r I and s I were measured separately to calculate the modulation efficiency. The modulation efficiency is then averaged over the ROI, whose diameter is the full width at half maximum (FWHM) of the one-dimensional modulation efficiency profile in the horizontal direction ( Supplementary Fig. 3). In this experiment, the ultrasound peak pressure is ~2 MPa, and the calculated modulation efficiency is ~0.5%.

b. Microbubble modulation
We measured the microbubble modulated light field with phase stepping holography and field subtraction method used for the phase recording of a TRUME process.
where the prime ( ' ) symbol denotes the signal field before modulation. Here we assume that the microbubble modulates the amplitude and phase of signal field by m a (ranging from 0 to 1) and   respectively. Then, the light intensity at the th k stepping phase is given by