Axially multifocal metalens and photoacoustic imaging device including the same

Abstract

The present invention relates to an axial multifocal metalens and a photoacoustic imaging device including the same. By using the metalens design method of the present invention, an axial multifocal metalens can be effectively designed, and efficient volumetric imaging is possible through the metalens.

Description

Technology Field

[0001] The present invention relates to an axial multi-focus metalens and a photoacoustic imaging device including the same. Background Technology

[0003] Human brain organoids (hBOs) are three-dimensional miniature brains derived from embryonic stem cells (ESCs) or induced pluripotent stem cells (iPSCs). Because they reflect the regional characteristics of the actual human brain, they have recently garnered significant attention as a unique 3D in vitro system capable of studying brain-related diseases or drug efficacy. In particular, the defective production of neuromelanin resulting from the loss of dopaminergic neurons in the human midbrain is believed to be associated with the pathogenesis of Parkinson's disease (PD). However, organoids present many challenges to researchers due to their high heterogeneity and complexity, caused by variations in cell density, the presence of necrotic centers, and the distribution of protein aggregation. While optical imaging is a representative method for observing cross-sections of traditional organoids, the depth of field of a standard bright-field microscope (BF) is limited to approximately 100 μm due to the opacity of biological tissues. This necessitates time-consuming processes such as 2D tissue staining before imaging or 3D tissue staining after tissue clearing. Therefore, there is an urgent need for the development of advanced non-invasive imaging technology that can visualize living organoids without fluorescent dyes or tags.

[0004] Volumetric photoacoustic (PA) imaging technology, capable of providing deep tissue imaging without fluorescent labeling, is a highly suitable method for imaging organoids containing neuromelanin. PA imaging provides structural and functional insights into biological tissues by detecting ultrasound generated as tissue rapidly expands and contracts upon absorbing pulsed laser light, thereby producing images based on light absorption contrast. Neuromelanin exhibits a high light absorption coefficient for visible light, making it highly advantageous for observation via PA imaging. In particular, optical-resolution photoacoustic microscopy (OR-PAM) utilizes precisely focused light to acquire high-resolution images up to a depth of approximately 1 mm within biological tissues. Since the axial resolution of OR-PAM is determined by time-resolved ultrasound detection, increasing the depth of field (DOF) can significantly improve the performance of OR-PAM in volumetric imaging of samples with uneven surfaces or those requiring high penetration depth. Unlike traditional Gaussian beam-based PAM, OR-PAM, which utilizes an elongated beam, enables high-resolution imaging in a single image plane without the need for slides and eliminates the axial multi-scanning process. Various techniques have been attempted to extend the Degrees of Field (DOF) in PA imaging, including structured illumination, contour scanning PAM, and synthetic aperture focusing. However, these techniques entail reduced imaging speeds and complex post-processing steps. Alternatively, Bessel beam-based PAM offers near-diffraction-limited resolution and a long DOF similar to conventional PA imaging; however, Bessel beams involve strong side lobes, and the significant variation in intensity within the DOF makes uniform excitation difficult in non-uniform samples.

[0005] Accordingly, forming an incident wavefront to generate an elongated beam with a high aspect ratio provides an opportunity to minimize side lobes and maintain uniform excitation intensity within a defined axial distance. Recently, PAM methods have been attempted using diffractive optical elements (DOEs) that extend the depth of field (DOF) by controlling the optical path length within individual unit cells, either in combination with an objective lens or alone. However, one of these methods failed to meet miniaturization requirements due to the need for an objective lens, while the others have limitations in application as they have not been verified in biological tissues.

[0006] In contrast, metalens function as ultra-small planar optical devices capable of controlling the characteristics of incident light by controlling the phase of light through interaction with nanoscale meta-atoms deposited on a substrate. Furthermore, metalens technology already surpasses diffraction lens technology in several aspects, including high-efficiency high-numerical-number (high-NA) applications, miniaturization, tunability, and polarization sensitivity. Although metalens theoretically possess characteristics advantageous for high-resolution OR-PAMs because sub-wavelength meta-atoms enable precise phase control, no experimental attempts regarding metalens-based DOF-extended PAMs (MeD-PAMs) had been reported prior to this study.

[0007] In this study, we propose a titanium dioxide nanopillar-based transmissive metalens device that generates an axial multifocal beam at a wavelength of 532 nm without an objective lens (Fig. 1). We adaptively assign phase values ​​to multiple focal points distributed along the optical axis and correct the diameter and focal positions of the elongated beam by adding the phase pattern of a condensing lens with a numerical aperture of 0.15 to the phase map. By adjusting the phase value corresponding to the shift of each focal point, we secure lateral resolution of the diffraction limit, and by precisely controlling the focal positions, we generate an elongated beam with high uniformity of intensity, lateral resolution of the diffraction limit, and low side lobes. Generally, approaches for designing phase and amplitude maps include extended Nijboer-Zernike wavefront formation theory and inverse engineering techniques, but these are computationally intensive and have limitations in beam shape control. Prior art literature

[0009] Republic of Korea Published Patent No. 10-2020-0047612 The problem to be solved

[0010] The inventors have made diligent research efforts to develop a metalens capable of performing optical resolution photoacoustic imaging on samples with non-uniform surfaces without z-scanning. As a result, it was discovered that a metalens composed of titanium dioxide nanopillars with a thickness of less than a micrometer can generate an elongated beam with an aspect ratio of 286 and maintain uniform intensity across the DOF while having a diameter at the diffraction limit level. Furthermore, the present invention was completed by discovering that the metalens improves the visualization of phantom samples with tilted surfaces compared to conventional lenses and enables volumetric imaging of neuromelanin up to a depth of 500 micrometers in human midbrain and forebrain organoids.

[0011] Accordingly, the object of the present invention is to provide a method for designing an axial multifocal metalens in which a plurality of focal points are arranged along an optical axis.

[0012] Another objective of the present invention is to provide an axial multifocal metalens in which a plurality of focal points are arranged along an optical axis.

[0013] Another objective of the present invention is to provide a photoacoustic detection device comprising the metalens.

[0014] Another objective of the present invention is to provide a photoacoustic detection method comprising the steps of: transmitting light through the metalens to irradiate a sample; and collecting ultrasound emitted from the sample.

[0015] Other objects and advantages of the present invention will become more apparent from the following detailed description of the invention, claims, and drawings. means of solving the problem

[0017] In one aspect of the present invention, the present invention provides a method for designing an axial multifocal metalens in which a plurality of focal points are arranged along an optical axis, comprising the following steps:

[0018] (a) Set the target operating wavelength λ and numerical aperture (NA), and the relationship

[0019]

[0020] Focal length f corresponding to multiple focal positions on the optical axis according to k The step of determining,

[0021] The above V is a constant as a focal interval optimization parameter, and d interfoci Is It is the focal interval;

[0022] (b) Depending on the (x, y) coordinates representing the two-dimensional position on the metalens surface, the condensing lens phase function φ is as follows lens (x,y) and focus shift phase function φ fociStep of calculating (x,y) respectively,

[0023] ,

[0024] ,

[0025] Above, k is the focal number, K is the total number of focal points, n is the refractive index of the medium, λ is the target operating wavelength of the metalens, A is a coefficient for controlling the output light diameter and side lobe intensity, and Loc k (x, y) is a binary matrix for assigning each phase function to a meta-atom at a specific location; and

[0026] (c) The two phase functions above are combined to form the final phase map φ as shown below. map Determine (x,y), and

[0027] ,

[0028] For each meta-atom of the metalens lattice, Loc k A step of assigning a phase value such that if (x,y)=1, the phase value of the k-th focus is assigned, and if (x,y)=0, the corresponding phase value is not assigned.

[0029] The inventors have made diligent research efforts to develop a metalens capable of performing optical resolution photoacoustic imaging on samples with non-uniform surfaces without z-scanning. As a result, it was discovered that a metalens composed of titanium dioxide nanopillars with a thickness of less than a micrometer can generate an elongated beam with an aspect ratio of 286 and maintain uniform intensity across the DOF while having a diameter at the diffraction limit level. Furthermore, it was discovered that the metalens improves the visualization of phantom samples with tilted surfaces compared to conventional lenses and enables volumetric imaging of neuromelanin up to a depth of 500 micrometers in human midbrain and forebrain organoids.

[0030] The term "metasurface" in this specification refers to an artificially designed, extremely thin surface, generally composed of nanoscale structures. These nanostructures can manipulate interactions with electromagnetic waves, particularly light, to control optical properties such as refraction, reflection, polarization, and phase change. Metasurfaces are much smaller than traditional optical elements and can manipulate light more efficiently; thanks to these characteristics, they are applicable in various fields such as optics, sensors, communications, and imaging. Furthermore, metasurfaces can also be applied to lenses that allow light to pass through while adjusting focus based on optical properties (amplitude, phase, etc.), which are referred to as metalenses.

[0031] The present invention relates to a method for designing an axial multifocal metalens having a structure in which a plurality of focal points are arranged along an optical axis. In this specification, the term "axial multifocal" means that a plurality of focal points are arranged along a single optical axis (z-axis), thereby enabling the metalens to generate a light beam with a significantly longer depth of focus (DOF) compared to a single-focus lens. This structure is particularly advantageous for high-depth imaging of three-dimensional specimens, such as biological tissues.

[0032] In the present invention, the focal length f k is defined to increase progressively according to the focus number k, and its mathematical expression is as follows:

[0033]

[0034] Here, f1 is the first focal length, d interfoci is the focal interval, and V is a constant parameter for optimizing the focal interval. Through this formula, the design allows the focal interval to increase steadily, as well as optimize intensity uniformity through V.

[0035] Designed focal length f kThe phase function at each focal point is calculated according to this, and the phase map of the entire metalens is the base lens phase function φ lens (x,y) and the focus dispersion phase function φ foci It is defined as the sum of (x,y).

[0036]

[0037] φ lens (x,y) is the general spherical wavefront phase term used in single-focus lenses, which functions to focus incident light to the first focal point (f1). On the other hand, φ foci (x,y) is each focal point f k A phase term for forming multiple foci by adding phases corresponding to , defined by the following equation:

[0038]

[0039] Here, K is the total number of focal points, n is the refractive index of the medium, λ is the operating wavelength, A is a coefficient controlling the output beam diameter and side lobe intensity, and Lock(x,y) is a binary matrix that determines whether the meta-atom at position (x,y) is given a phase term corresponding to the focal point k.

[0040] Loc used in the above equation k (x,y) is a binary matrix having values ​​of 0 or 1 indicating whether to assign the phase of focus k to each (x,y) position, and is configured so that each meta-atom corresponds to only one of the foci k. Loc k Matrices can be assigned randomly with fair randomness and can be generated in a reproducible manner using the same random seed.

[0041] Phase values ​​are assigned to meta-atoms corresponding to the surface lattice of the metalens, and meta-atoms are typically implemented in the form of circular nanopillars hundreds of nanometers in size formed from high-refractive-index materials (e.g., TiO2).

[0042] In the present invention, the term "meta-atom" may be used interchangeably with "nanostructure," and each meta-atom is formed with a different diameter depending on the phase value at the corresponding position to realize the entire phase map.

[0043] In one embodiment of the present invention, the metalens generated through the design method provides a depth of field (DOF) improved by at least 13 times compared to a single-focus metalens and can generate a uniform beam in which the intensity variation is maintained at 5% or less across the entire DOF area (e.g., 500 μm). This beam is a high aspect ratio beam with an aspect ratio of approximately 286, and while having a diameter at the diffraction limit level, the side lobe intensity is maintained at a low level of approximately 9% or less, making it highly suitable for high-quality photoacoustic imaging and 3D imaging.

[0044] In one embodiment of the present invention, the target operating wavelength is 100 to 1500 nm.

[0045] In this specification, the term "target operating wavelength" refers to the center wavelength of the light beam generated by the metalens. The present invention may be designed to support wavelengths in the range of 100 nm to 1500 nm. This range includes a broadband spectrum extending from ultraviolet (UV) to near-infrared (NIR) and can be utilized in various optical and bioimaging applications. The target operating wavelength may be, for example, 100 nm to 1500 nm, 100 nm to 1300 nm, 100 nm to 1100 nm, 100 nm to 900 nm, 100 nm to 700 nm, 100 nm to 500 nm, 100 nm to 300 nm, 300 nm to 1500 nm, 500 nm to 1500 nm, 700 nm to 1500 nm, 900 nm to 1500 nm, 1100 nm to 1500 nm, 1300 nm to 1500 nm, 300 nm to 1200 nm, 300 nm to 900 nm, or 300 nm to 700 nm, but is not limited thereto.

[0046] In one embodiment of the present invention, the target numerical aperture is 0.1 to 1.0.

[0047] The term "target numeric aperture (NA)" in this specification is a parameter representing the light-gathering ability of a metalens, which has a direct effect on the diffraction limit and resolution of the beam. In the present invention, by setting the NA to a range of 0.1 to 1.0, it can be designed to be applicable to various optical systems ranging from low resolution to high resolution.

[0048] In one embodiment of the present invention, the focal length optimization parameter V is selected in the range of 5 nm to 30 nm so that the standard deviation of the axial intensity distribution of the output light is 10% or less.

[0049] In this specification, the term "focal interval optimization parameter V" refers to a constant that controls the gradual change in the spacing between each focal point in an axial multifocal structure of a metalens, and is used to ensure that multiple focal points distributed along the optical axis are stably arranged without interference with one another. V is a factor that causes the spacing to increase linearly as the focal number increases, and is introduced to ensure uniformity of the axial intensity distribution.

[0050] An appropriate V value is determined through numerical simulation and optimized so that the standard deviation of the intensity distribution is maintained below a certain threshold. In this invention, the standard was set as maintaining the standard deviation of the axial intensity distribution of the output light at 10% or less, and the range of V values ​​satisfying this condition was derived to be approximately 5 nm to 30 nm. By selecting V within this range, foci are evenly distributed without constant interference, intensity is maintained uniformly without ripple within the entire DOF, and the multifocal beam can maintain the desired resolution and signal contrast. The determination of such V values ​​was verified through alignment with experimental PSF measurement results and numerical Rayleigh-Sommerfeld propagation simulations.

[0051] The V value may be, for example, 5 nm to 30 nm, 5 nm to 25 nm, 5 nm to 20 nm, 5 nm to 15 nm, 5 nm to 10 nm, 10 nm to 30 nm, 15 nm to 30 nm, 20 nm to 30 nm, 25 nm to 30 nm, 10 nm to 25 nm, or 10 nm to 20 nm, but is not limited thereto.

[0052] In one embodiment of the present invention, the focal interval d interfoci It is 1 to 50㎛.

[0053] The term "focal interval (d)" in this specification interfoci"" refers to the distance between adjacent focal points arranged along the optical axis of an axial multifocal metalens, and is an important design factor that affects the uniformity of the axial intensity distribution and DOF (Depth of Field) of the generated beam. The focal spacing is appropriately adjusted considering the number of focal points K and the total DOF, and by maintaining a spacing sufficiently smaller than the Rayleigh distance, multiple focal points overlap each other to stably realize an elongated and uniform beam shape.

[0054] The focal interval may be, for example, 1 to 50 µm, 1 to 40 µm, 1 to 30 µm, 1 to 20 µm, 1 to 10 µm, 10 to 50 µm, 20 to 50 µm, 30 to 50 µm, 40 to 50 µm, 5 to 30 µm, 5 to 20 µm, or 5 to 10 µm, but is not limited thereto.

[0055] In one embodiment of the present invention, the coefficient A for controlling the output light diameter and side lobe intensity is selected from the range of 0.1 to 0.5.

[0056] The above coefficient A is the focus shift phase function φ foci As a correction term included in (x, y), it is a key parameter that determines the Full Width at Half Maximum (FWHM) of the output beam formed by the metalens and the relative intensity of the side lobes formed around the beam. The output light diameter refers to the Full Width at Half Maximum in the central intensity distribution of the axial beam transmitted from the metalens, which is ultimately directly related to lateral resolution. Side lobes are unwanted secondary intensity peaks generated at locations other than the main focal point and can act as noise elements in the image signal. Increasing the value of coefficient A reduces the output light's width and improves resolution, but there is a trade-off in that the relative intensity of the side lobes increases simultaneously. Therefore, the value of A is set considering this balance, and in this embodiment, A It was set to 0.14 to achieve optimal performance with a diameter at the diffraction limit level and low side lobe intensity. If the value of A is less than 0.1, the beam diameter expands and resolution may be reduced, and if it exceeds 0.5, there is a concern about image quality degradation due to side lobes. The above A may be 0.1 to 0.5, 0.1 to 0.4, 0.1 to 0.3, 0.1 to 0.2, 0.2 to 0.5, 0.3 to 0.5, 0.4 to 0.5, or 0.2 to 0.3, but is not limited thereto.

[0057] In one embodiment of the present invention, the number of foci K is 10 to 100.

[0058] In this specification, "number of focal points K" refers to the number of individual focal points generated by the metalens along the axial direction, and this is an important design element that determines the uniformity of light intensity and beam extension within the depth-of-field (DOF). If the number of focal points is too small, the intensity distribution within the DOF may become non-uniform, while conversely, if there are too many, the burden on metalens design complexity and manufacturing precision may increase. Therefore, in this invention, a balance between beam uniformity and manufacturability is secured by setting the number of focal points K to an optimal range of 10 to 100, derived through simulation and experiment. The number of foci may be, for example, 10 to 100, 10 to 90, 10 to 80, 10 to 70, 10 to 60, 10 to 50, 10 to 40, 10 to 30, 10 to 20, 20 to 100, 30 to 100, 40 to 100, 50 to 100, 60 to 100, 70 to 100, 80 to 100, 90 to 100, 20 to 90, 40 to 90, or 50 to 70, but is not limited thereto.

[0059] In one embodiment of the present invention, the Loc k The (x,y) matrix is ​​generated by assigning each nanostructure on the metalens surface to a fair random with a probability of 1 / K.

[0060] The Lock(x,y) matrix is ​​a binary matrix that determines whether to assign a phase value corresponding to a specific focus to each nanostructure (or meta-atom) on the surface of the metalens, and is generated such that each nanostructure is randomly assigned with an equal probability (1 / K) for the total number of focuses K. This random number-based assignment method contributes to minimizing optical interference during beamforming and realizing a uniform intensity distribution across the axial depth of field (DOF). In particular, such equal random assignment suppresses structural interference ripple, and the random number seed may be fixed to ensure the reproducibility of Lock(x,y) generation.

[0061] In one embodiment of the present invention, the Loc k The (x,y) matrix is ​​generated by fixing the random number seed so that random number assignment can be reproducible.

[0063] In one embodiment of the present invention, the present invention provides an axial multifocal metalens in which a plurality of focal points are arranged along an optical axis.

[0064] In one embodiment of the present invention, the metalens may be formed by any one of processing methods such as high-speed laser direct lighting, lithography, and nanoimprinting, but is not limited thereto.

[0065] In one embodiment of the present invention, a metalens may be manufactured using an electron beam (e-beam) lithography technique. In the e-beam lithography technique for manufacturing a metalens, a focused e-beam may be scanned across the surface of a substrate to create a pattern corresponding to a desired meta-surface structure. In some cases, the surface of the substrate may be coated with a resist material that changes properties when exposed to e-beam energy. Depending on the type of resist material used, either the exposed resist material or the unexposed resist material may be selectively removed, while the other portion remains on the surface of the substrate. When the resist material is selectively removed, the substrate may be exposed and etched (e.g., by wet etching, dry etching, reactive-ion etching (RIE), etc.) to remove a portion of the substrate material. In some cases, the etching process may create geometric features of a meta-surface on the surface of the substrate material to form a metalens. In some cases, the manufacturing process can be time-consuming and costly because geometric features of the meta-surface must be patterned onto the resist material by directing a focused e-beam onto the resist material.

[0066] Systems, devices, processes (also referred to as methods), and computer-readable media (collectively referred to as systems and techniques) are described herein for manufacturing metalens and optical systems including metalens in a scalable manner. For example, semiconductor manufacturing technology is used to simultaneously produce multiple devices (e.g., microprocessors, application-specific integrated circuits, etc.) on a single silicon wafer. In contrast to the e-beam lithography technique described above, features manufactured on the surface of the silicon wafer are not drawn individually.

[0067] Instead, features of the device (or negative representations of the features) can be patterned onto a mask. Features of a single device can be repeated in an array to fill an area (or part of an area) of the surface of a silicon wafer having multiple devices. By a single exposure to light, the pattern on the mask can be transferred to a photosensitive resist (photoresist) material. In semiconductor manufacturing, multiple masks may be used to manufacture different features of the device, such as metal layers, transistors, passivation layers, mechanical structures, etc. Therefore, it would be advantageous if the photolithography process used to manufacture semiconductors could also be used to manufacture meta-lenses.

[0068] In some embodiments, silicon materials used in many semiconductor manufacturing applications are transparent to certain wavelengths of light. In some cases, optical applications can detect light at wavelengths to which silicon is transparent. Therefore, silicon can be a suitable substrate material for manufacturing metalens for image sensing applications, where silicon is transparent to the wavelengths of light being detected. For example, applications use short-wave infrared (SWIR). In some cases, SWIR-sensitive image sensors can be manufactured using semiconductor manufacturing techniques. For example, SWIR-sensitive imagers can be manufactured on silicon wafers using germanium-silicon (GeSi)-based complementary metal-oxide-semiconductor (CMOS) technology. In some cases, the semiconductor manufacturing techniques described above can be used to manufacture metalens on silicon wafers.

[0069] For some optical applications, silicon may not be a suitable substrate for fabricating metalens because the wavelengths of light relevant to the application may not be able to pass through silicon. For example, silicon is opaque at visible light wavelengths. Many optical applications detect light at visible wavelengths. In such cases, materials that are transparent at visible light wavelengths may be suitable substrates for fabricating metalens. In one exemplary example, metalens can be fabricated on a glass substrate. In some cases, nanoimprinting lithography techniques can be used to fabricate metalens on a glass substrate.

[0070] Meanwhile, in one embodiment of the present invention, the method for manufacturing the metalens involves forming a nanostructure on a metalens substrate, and this can be done using an electron beam lithography (E-beam lithography) process. As an example of the method for manufacturing the metalens of the present invention, first, a silica (SiO2) layer is deposited by plasma enhanced chemical vapor deposition (PECVD), and then an electron beam is focused onto PMMA, which is a coated positive tone photoresist, to process the desired structural shape. Next, a chromium layer is deposited to be used as an etching mask, and after a dry etching process, the remaining etching mask is removed, thereby finally completing the nanostructure of the metalens.

[0071] In one embodiment of the present invention, the metalens is designed according to the axial multifocal metalens design method.

[0072] In one embodiment of the present invention, the metalens is made of HfO2, TiO2, ZnO, SiO2, Al2O3, ZrO2, Si3N4, or a combination thereof.

[0073] In one embodiment of the present invention, the metalens is composed of HfO2, TiO2, ZnO, SiO2, Al2O3, ZrO2, Si3N4, or a combination thereof. However, it is not limited thereto and may be one or more of photoresist, quartz glass, silicon nitride, titanium oxide, alumina (sapphire), crystalline silicon (including crystalline and amorphous silicon), gallium nitride, crystalline germanium, selenium sulfide, selenium sulfide, and chalcogenide glass.

[0074] In one embodiment of the present invention, the diameter of the metalens is 1 x 10 2 μm to 1 X 10 4 It is μm.

[0075] In one embodiment of the present invention, the metalens has a depth of focus (DOF) of 100 to 1000 μm. The depth of focus may be, for example, 100 to 1000 μm, 100 to 800 μm, 100 to 600 μm, 100 to 400 μm, 100 to 200 μm, 200 to 1000 μm, 400 to 1000 μm, 600 to 1000 μm, 800 to 1000 μm, or 400 to 600 μm, but is not limited thereto.

[0076] The above DOF is extended by approximately 3 to 13 times compared to conventional diffractive lenses with the same numerical aperture (NA), and this metalens can form a beam evenly focused across the DOF through an axial multifocal structure and phase optimization. This extension of the depth of focus provides the advantage of being able to image a wide area with high resolution in a single scan for biological specimens with non-uniform or three-dimensional surfaces.

[0077] In one embodiment of the present invention, the metalens comprises a plurality of cylindrical meta-atoms having a diameter of 50 nm to 300 nm, a height of 100 to 1000 nm, and a combination thereof.

[0078] In one embodiment of the present invention, the meta-atom is in the form of a cylinder or a prismatic column.

[0079] The column structure of the meta-atom above, the nano-column structure, may be, for example, a positive cylinder, a negative cylinder, a hollow nano-column structure, a square nano-column structure, or a topological nano-column structure. Preferably, it may be a positive cylinder structure.

[0080] In one embodiment of the present invention, the metalens generates a beam with an aspect ratio of 100 to 500, having a width within the range of ± 10% of the diffraction limit of the lens.

[0081] The metalens described above is designed to achieve performance close to the diffraction limit of the lens. In particular, the metalens is adjusted so that the beam diameter of the output light falls within a range of ±10% relative to the diffraction limit of an ideal lens having the same numerical aperture (NA), and simultaneously, it can generate a high aspect ratio beam with an axially elongated aspect ratio in the range of 100 to 500. The aspect ratio of the beam may be 100 to 500, 100 to 400, 100 to 300, 100 to 200, 200 to 500, 300 to 500, 400 to 500, or 200 to 300, but is not limited thereto. These characteristics enable high-resolution and deep-depth images to be realized with only a single plane scan in a metalens-based photoacoustic microscope, and provide significantly superior performance compared to conventional single-focus optical systems.

[0082] In one embodiment of the present invention, the metalens is configured such that the standard deviation of the output light intensity is maintained at 5% or less throughout the entire depth of focus (DOF) range.

[0083] The metalens described above forms an elongated axial multifocal beam through multiple focal points distributed along the optical axis, and is designed to maintain high uniformity in the intensity distribution of the output light across the entire depth of focus (DOF). Specifically, by optimizing the values ​​of the focal spacing, the number of focal points, the phase function, and the coefficient A so that the standard deviation of the output light intensity is maintained at 5% or less across the entire DOF, uniform optical stimulation is enabled without variations in the intensity of the beam generated by the metalens. This characteristic provides favorable conditions for acquiring stable and precise photoacoustic images regardless of changes in the depth of the specimen.

[0085] In one aspect of the present invention, the present invention provides a photoacoustic detection device comprising the metalens.

[0086] The photoacoustic detection device of the present invention may be a photoacoustic-based sensor device. According to one example, the photoacoustic detection device may include light sources, a detector, and a control unit. The light sources may generate ultrasound by providing source light to a subject. Each of the light sources may include a laser, a laser diode, or a light-emitting diode.

[0087] A photoacoustic detection device includes a transducer configured to detect acoustic waves generated by the heating of a medium induced by light waves.

[0088] The above absorption spectrum can be measured through a device or method that spectroscopically analyzes the absorption spectrum, for example, using a spectrophotometer.

[0089] In one embodiment of the present invention, the photoacoustic detection device can acquire a three-dimensional volumetric image without focus shift.

[0090] A photoacoustic detection device according to one embodiment of the present invention includes a metalens having multiple focal points in the axial direction, thereby providing uniform optical stimulation across the entire sample without separate focal scanning, and thus enabling the acquisition of a three-dimensional volumetric image with only a single scan. Accordingly, image acquisition time is shortened and system configuration is simplified, and high-speed and high-precision photoacoustic imaging is possible even for specimens such as biological tissues or organoids with complex shapes. These characteristics can replace the axial mechanical scanning process required by conventional photoacoustic systems, making it advantageous for real-time image realization.

[0092] In one embodiment of the present invention, the present invention provides a photoacoustic detection method comprising the steps of: transmitting light through the metalens to irradiate a sample; and collecting ultrasound emitted from the sample.

[0093] The target molecule of the above photoacoustic detection method may be, for example, biomolecules such as proteins, DNA, RNA, peptides, viruses, cellular components, chemical molecules, and nanoparticles, but is not limited thereto, and may include any molecule having light-absorbing properties. For example, the molecule may be nucleic acid molecules such as DNA and RNA, hemoglobin, melanin, etc., but is not limited thereto.

[0094] In one embodiment of the present invention, the sample is a brain organoid. Effects of the invention

[0096] The features and advantages of the present invention are summarized as follows:

[0097] (a) The present invention provides a method for designing an axial multifocal metalens in which a plurality of focal points are arranged along an optical axis.

[0098] (b) The present invention provides an axial multifocal metalens in which a plurality of focal points are arranged along an optical axis.

[0099] (c) The present invention provides a photoacoustic detection device comprising the metalens.

[0100] (d) Another object of the present invention provides a photoacoustic detection method comprising the steps of transmitting light through the metalens to irradiate a sample; and collecting ultrasound emitted from the sample.

[0101] (e) When using the metalens design method of the present invention, an axial multifocal metalens can be effectively designed, and efficient volumetric imaging is possible through the metalens. Brief explanation of the drawing

[0103] Figure 1 shows a schematic design of a metalens for PAM with an elongated DOF. (A) Schematic diagram of a MeD-PAM designed for imaging brain organoids with non-uniform surfaces. UST, ultrasound transducer. (B) Phase map design method. A focus-shift phase is superimposed on the phase of the converging lens, which is selectively assigned to specific meta-atoms via a binary matrix (Lock) to generate the final phase map. F1, F10, F35, and F60 are focus-shift phase maps corresponding to the focal numbers, respectively. (C) Numerical simulation of the axial multifocal beam intensity distribution (top) and single-focal beam intensity distribution (bottom) resulting at an NA of 0.15. (D) Scanning electron microscope images of the fabricated metalens structure taken at various magnifications and field of view. Figure 2 shows the results of the experimental characterization analysis of the metalens PSF. (AC) Transverse focal spot intensity distribution at axial positions z = 250, 0, and 250 μm, respectively. z = 0 μm represents the center plane of the elongated beam. (D) Horizontal cutoff line of the intensity distribution shown in (A)-(C). (E) Axial intensity distribution of the elongated beam. (F) Axial intensity distribution cutoff line obtained through numerical simulation and experimental data. Figure 3 shows the elongated DOF PA imaging results of the phantom sample. (AC) Depth-encoded PA images of two axially tilted carbon fibers acquired using a metalens (A), an aspherical lens (B), and an objective lens (C), respectively. The characters indicated in (A) correspond to the line profile presented in Figure 18. The distance values ​​represent the visible length of the left fiber observed in the PA image. (DF) Depth-encoded PA images of an axially tilted carbon leaf acquired using a metalens (D), an aspherical lens (E), and an objective lens (F). Scale bar: 500 μm. Figure 4 shows the volumetric PA imaging results of neuromelanin in melanin-containing forebrain and midbrain organoids using an axial multifocal metalens. (A) PA MAP image of a forebrain organoid cultured for 170 days. (B) Depth-encoded PA image of a forebrain organoid showing the axial position of neuromelanin in color. (C) PA cross-sectional image of a forebrain organoid. The top cross-sectional image corresponds to the red dashed line in (A), and the bottom cross-sectional image was acquired along the yellow dashed line in (A). (D) BF image of a forebrain organoid. The four arrows in (A) and (D) point to the same location. (E) 3D-rendered PA image of a forebrain organoid. (F) Fontana-Masson staining of a forebrain organoid. The arrow indicates the black pigment. (G) PA MAP image of a midbrain organoid cultured for 132 days. (H) Depth-encoded PA image of a midbrain organoid showing the axial position of neuromelanin in color. (I) PA cross-sectional image of the midbrain organoid. The top cross-sectional image corresponds to the red dashed line in (A), and the bottom cross-sectional image was acquired along the yellow dashed line in (A). (J) BF image of the midbrain organoid. (K) 3D rendered PA image of the midbrain organoid. (L) Fontana-Masson staining of the midbrain organoid. Arrows indicate black pigment. Scale bar: 500 μm. au, arbitrary units. Figure 5 shows the results of the analysis of neuromelanin content characteristics in various melanin-containing forebrain and midbrain organoids using MeD-PAM. (A) Superimposed BF and PA images of a forebrain organoid at day 150. (B) Superimposed BF and PA images of a midbrain organoid at day 132. (C) Normalized neuromelanin content in the brain organoids of (A) and (B). Neuromelanin content is evaluated as the sum of PA signal pixels exceeding background noise. Error bars indicate the change that occurs when the threshold is lowered within the standard deviation of background noise. (D, E) Superimposed BF and PA images of forebrain organoids at days 170 and 216, respectively. (F) Normalized neuromelanin content in the brain organoids of (A), (D), and (E). Scale bar: 500 μm. FIG. 6 illustrates the assignment of focal numbers to meta-atoms within a metalens lattice. (A) When meta-atoms are uniformly distributed at each focal point. (B) Distribution of meta-atoms having the same ring area within the metalens. The metalens area corresponding to each focal point is the same, and the outer radius is · Confined within the ring of R1. Here, k is the focal number, and R1 is the radius of the first focal ring. (C) Case where meta-atoms are adaptively or randomly distributed at the focals. Colors correspond to the focal numbers and are limited to 7 focals for visual clarity. (DF) Axial intensity distribution over a length of 600 μm according to the methods of (A)–(C), respectively. Figure 7 illustrates the focal position assignment for generating a uniform axial field profile. (A) Schematic diagram of focal redistribution for controlling field intensity along the axial length of the beam. The optimization parameter V controls the spacing between focal points as the focal number increases. (B) Axial intensity distribution of a multifocal beam consisting of 60 focal points with different V values. Figure 8 illustrates the effect of randomization on the PSF. (A) Axial intensity distribution of a multifocal metalens having a typical meta-atom distribution according to the focal number. (B) Probability distribution of focal assignments with a mean of 30 and a standard deviation of 10. All 60 focals are placed within a 3σ range. (C) Transverse intensity distribution of a multifocal beam based on a random distribution generated by the normal distribution of (B) and various fixed random number functions. The random number functions are fixed through the seed setting of the Python Numpy library. Figure 9 shows the simulation results regarding the effect of the A factor on image performance. (A) Normalized transverse intensity distribution at the needle beam center (z = 3.35 mm) according to the A factor value. (B) Beam diameter (FWHM) measured in three planes (z = 3.15, 3.35, 3.55 mm). The dashed line is the diffraction limit diameter of the conventional lens with NA = 0.15. (C) Ratio of side lobe intensity to main lobe intensity measured in the same plane. Figure 10 shows the results of a comparison of the intensity profiles of the Bessel beam and the multifocal beam. (A) Intensity distribution of the Bessel beam and the multifocal beam generated by a phase-modulated annular aperture under conditions where the NA and axial length are the same. (B) Axial intensity distribution of the two beams. (C) Transverse intensity distribution in the case with the smallest side lobe. Figure 11 shows the results of the analysis of the phase and transmittance characteristics of meta-atoms. (A) Meta-atom nanopillar shape. (B) Phase change according to nanopillar diameter. Visualization of the entire 2π phase control range by setting the phase to 0 at a diameter of 90 nm. (C) Transmittance change according to nanopillar diameter. Figure 12 shows the results of the analysis of phase and transmittance changes according to unit cell size and nanopillar diameter. (A) Phase map according to unit cell size and nanopillar diameter. (B) Transmittance map for the same parameters. The translucent white lines represent the pillar shapes selected in the meta-atom library. FIG. 13 illustrates the fabrication process of an axial multifocal metalens. (A) Master mold, (B) Soft mold fabrication, TiO2 nanoparticle-containing resin curing, and mold removal process. Figure 14 shows an optical microscope setup for analyzing lens PSF characteristics. Fig. 15 shows a detailed view of an experimental photoacoustic imaging system. US: Ultrasound, λ / 2: Half-wave plate, DAQ: Data acquisition system. Figure 16 shows the PSF of a conventional lens. (A) Axial intensity distribution of an aspherical lens. (B) Vertical cutting line at z = 0 in the intensity distribution of (A). (C) Horizontal cutting line along the optical axis. (D~F) Results measured in the same manner for the objective lens. Figure 17 shows the results of optical and PA characteristic analysis for the DOF of the needle-beam. Evaluation through PSF measurement (optical) and carbon fiber PA signal measurement (PA) in multiple planes. Figure 18 shows the transverse line profile and FWHM as a function of depth in the carbon fiber sample. The FWHM measured at three lines is 11.5 ± 0.27 μm, which is consistent with the known thickness of the carbon fiber (approx. 5–10 μm). Figure 19 shows the results of brain organoid generation. (A) Forebrain organoid generation protocol. (B) SOX2 and MAP2 immunostaining for 30 and 85 days. Scale bar: 50 μm. (C) Neuromelanin PA images in the midbrain for 30 and 100 days. Scale bar: 500 μm. (D) Midbrain organoid generation protocol. (E) SOX2 and DAT immunostaining for 33 and 83 days. Scale bar: 50 μm. (F) Neuromelanin PA images in the midbrain. Scale bar: 500 μm. Figure 20 shows the results of brain organoid tissue staining. (A) Immunostaining for verification of neuronal gene expression in forebrain organoids. (B) Detection of neuromelanin (black pigment) via Fontana-Masson staining. Scale bar: 30 μm. (C–D) The same method applied to midbrain organoids. Scale bars: 70 μm and 30 μm. Figure 21 shows the results of neuromelanin production. (A) Melanin production from both neural crest cells (SOX10+) and dopaminergic neurons (DAT+) in forebrain and (B) midbrain organoids. Red arrow: SOX10+DAT-, green arrow: SOX10-DAT+ signal. Scale bar: 100 μm. Figure 22 shows the results of the correlation image analysis of PA and Fontana-Masson staining. (A) BF image of the section, (B) PA MAP, (C) Fontana-Masson image, (D) melanin-highlighted area. Figure 23 shows BF and PA MAP images of brain organoids. (A) 170-day forebrain BF image. (BD) PA image using metalens, an aspherical lens, and an objective lens. The bottom panel shows the axial position profile based on the dotted line. (E) 132-day midbrain BF image. (FH) PA image using the same method. The red arrow indicates neuromelanin visualized with metalens. Scale bar: 500 μm. Figure 24 shows the transducer frequency dependence in the PA image. (A) Intercept BF image. (B) PA MAP images of the same intercept obtained with 5, 20, and 40 MHz transducers and the bottom B-scan. Figure 25 shows the pulse-echo test results of an ultrasonic transducer. (A) Pulse-echo response in the time domain, (B) Pulse-echo response in the frequency domain. Figure 26 shows the resolution of metalens-based photoacoustic imaging. (A) Blade PA MAP image and (B) edge spread function (ESF), line spread function (LSF). Lateral resolution: 2.30 μm. (C) Carbon fiber PA MAP and (D) axial intensity profile. Axial resolution: 89.9 μm. Figure 27 shows an image processing flowchart of a photoacoustic image. (A) Data acquisition flowchart, (B) Image reconstruction flowchart. Specific details for implementing the invention

[0104] The present invention will be described in more detail below through examples. These examples are intended solely to explain the invention more specifically, and it will be obvious to those skilled in the art that the scope of the invention is not limited by these examples according to the gist of the invention.

[0106] Examples

[0107] Example 1: Materials and Method

[0108] 1-1. Design of metalens

[0109] Titanium dioxide (TiO2) was selected as a meta-atom material with no loss and a high refractive index at a wavelength of 532 nm. Simulation data for phase and transmittance were generated using Ansys Lumerical FDTD and were based on the refractive index measured by ellipsometry of a resin containing titanium dioxide nanoparticles. The dimensions of the meta-atom were set within a fabrication-feasible range, and the unit cell size was 300 nm. Target phase data were generated through numerical simulation using a Rayleigh-Sommerfeld ray propagator, and a plane wave was used as the incident light.

[0110] After the optimal target phase profile was generated, the optimal diameter of the titanium dioxide nanopillars suitable for the final metalens device was determined through a brute force search. The resulting actual phase profile was input into a Rayleigh-Sommerfeld propagator to simulate the expected point diffusion function (PSF). The PSF of an ideal single-focus lens with the same working distance as the metalens was also simulated in the same manner using the following lens phase equation:

[0111]

[0112] Here, λ is the laser wavelength and f is the focal length of the lens.

[0113] The meta-axicon was simulated using a Rayleigh-Sommerfeld propagator set to a unit cell size of 300 nm, and the phase equation of the axicon is as follows:

[0114] Here, NA is the numerical aperture, which is 0.15 in this study.

[0116] 1-2. Metalens fabrication

[0117] Metalens were fabricated using a nanoimprint lithography process. A master mold for the multifocal metalens was formed on a silicon (Si) substrate using a high-resolution electron beam lithography (EBL) system. First, a photoresist (Microchem, 495 PMMA A6, MicroChem) was coated to a thickness of 400 nm and fixed to the Si substrate by baking at 180°C for 5 minutes. Subsequently, an optimized metalens pattern was exposed to the photoresist layer using an electron beam lithography system (ELS-BODEN, Elionix), and the exposed area was developed using a 1:3 methyl isobutyl ketone (MIBK) / isopropyl alcohol solution.

[0118] Next, a 40 nm thick chromium (Cr) thin film was vertically deposited on the patterned area using an electron beam deposition system (KVE-ENS4004, Korea Vacuum Tech Co. Ltd.), which is a preparatory step for successful lift-off. After removing the Cr-patterned photoresist area using acetone, the remaining Cr thin film served as a mask layer for selective silicon etching. Subsequently, a 900 nm thick silicon nanostructure was formed using a dry etching device (silicon / metal hybrid etcher, DMS Co. Ltd.). Afterward, the residual Cr mask was removed using a Cr etchant (CR-7), thereby completing the master mold for the multifocal metalens.

[0119] For the nanoimprint process using a high-resolution TiO2 nanoparticle-based resin (nano-PER), a double-layer mold of hard polydimethylsiloxane (hard PDMS, hereinafter h-PDMS) and standard PDMS was used. First, the h-PDMS solution was prepared as follows: 3.4 g of vinylmethyl copolymer (VDT-731, Gelest), 18 μl of platinum catalyst (SIP6831.2, Gelest), 0.1 g of modifier (2,4,6,8-tetramethyl-2,4,6,8-tetravinylcyclotetrasiloxane, Sigma-Aldrich), 2 g of toluene, and 1 g of siloxane-based reducing agent (HMS-301, Gelest) were mixed. This solution was spin-coated onto a master mold at 1000 rpm for 60 seconds and then cured at 70°C for 2 hours.

[0120] Subsequently, PDMS mixed with a curing agent (Sylgard 184 A, Dow Corning) in a weight ratio of 10:1 was poured onto the h-PDMS layer and further cured at 80°C for 2 hours. The resulting h-PDMS / PDMS bilayer mold was separated from the master mold, and a fluorinated surfactant [(tridecafluoro-1,1,2,2-tetrahydrooctyl)trichlorosilane, Sigma-Aldrich] was vapor-deposited at 130°C for 5 minutes to reduce the surface energy.

[0121] The TiO2 nanoparticle resin (nano-PER) was prepared as follows: TiO2 nanoparticles dispersed in MIBK (DT-TIOA-30MIBK(N30), Ditto Technology), a monomer (dipentaerythritol penta- / hexaacrylate, Sigma-Aldrich), a photoinitiator (1-hydroxycyclohexyl phenyl ketone, Sigma-Aldrich), and an MIBK solvent (Duksan General Science) were mixed and adjusted to a final weight ratio of 4 wt% TiO2 nanoparticles, 0.7 wt% monomer, and 0.3 wt% photoinitiator.

[0122] 0.4 ml of this TiO2 nano-PER solution was spin-coated onto a fabricated soft mold at 2000 rpm for 60 seconds. Subsequently, after allowing the mold to absorb the solvent, a pressure of 2 bar was applied between the mold and the SiO2 substrate for 60 seconds, followed by pressing at 8 mW / cm² 2 It was cured by irradiating with ultraviolet light for 15 minutes. Finally, by removing the soft mold, the fully cured multifocal metalens was successfully transferred onto the SiO2 substrate.

[0124] 1-3. Analysis of Metalens Optical Characterization

[0125] The point diffusion function (PSF) of a metalens generating an elongated beam was evaluated using a microscopy system. A collimated laser beam with a wavelength of 532 nm was directed toward the metalens with a beam diameter slightly larger than that of the metalens. The incident beam must be well parallel, and if the incident light is focused, the DOF and diameter of the elongated beam generated by the metalens can be significantly altered.

[0126] The beam focused by the metalens was visualized through an imaging system consisting of an Olympus UPLFLN 20x objective lens (NA = 0.5) and a Lumenera Infinity 2-1R CCD camera. This objective lens was mounted on a motor-driven stage to acquire images in multiple planes. By performing a focus profile in the z-axis direction, we accurately measured the shape of the elongated beam and the DOF.

[0127] To measure the PSF over the entire beam length, a total of 360 transverse beam images were captured at 2 μm intervals along the optical axis (z-axis). PSF measurements for an aspherical lens and an objective lens with an NA of 0.6 were also performed using the same microscope system. In this case, the effective numerical aperture (NA_eff) of the existing refractive lenses was calculated using the following formula:

[0128]

[0129] Here, r beam is the radius of the incident beam (fixed at 1 mm), and f is the focal length of the corresponding lens.

[0130] Since the size of the metalens was slightly smaller than the incident beam, the effective NA matched the nominal value well.

[0132] 1-4. PA Imaging Settings

[0133] The excitation light source used here is a nanosecond laser with a wavelength of 532 nm (Photonics Industries International Inc., DX-Air cooled nanosecond laser, DX-532-2), with an operating pulse repetition frequency of 20 kHz and a pulse width of 6 ns. The laser energy was controlled via a half-wave plate polarizing beam splitter (VA5-532 / M, Thorlabs Inc.). The laser beam was parallelized using a plano-convex lens (LA1908-A-ML, Thorlabs Inc.) with a focal length of 500 mm, and the diameter of the parallelized laser beam was 2 mm. For the OR-PAM using a metalens, the beam diameter was reduced to 1 mm using a ring-adjustable iris diaphragm (SM1D12D, Thorlabs Inc.).

[0134] The axial positions of the metalens, aspherical lenses, and objective lenses were adjusted using a laterally adjustable 6.5 mm moving translation stage (MS1S / M, Thorlabs Inc.). The laser beam was focused and irradiated onto a sample immersed in a glass deionized water bath. The pulsed laser irradiation energy at a wavelength of 532 nm was approximately 15 mJ / cm². 2 This falls below the biological safety standards of the American National Standards Institute (ANSI).

[0135] PA signals were acquired using a planar high-frequency ultrasonic transducer with a center frequency of -6 dB at 19 MHz, a focal length of 11.6 mm (15.7 μs), and a bandwidth of 69% (Fig. 25). The transverse resolution obtained from the blade's edge spread function was 2.3 μm (Figs. 26 A and B). The axial resolution measured based on the PA B-scan images of the carbon fiber was 90 μm (Figs. 26 C and D).

[0136] An xyz translation stage (PT3 / M, Thorlabs Inc.) was used to align the ultrasonic transducer to deionized water, and a motor-driven xy scanning stage (8MTF, Standa) was used to move a glass Petri dish and scan images in the x and y directions. PA signals were acquired via the ultrasonic transducer in transmission mode, and a dual pulser / receiver (DPR500, JSR Corporation) with a bandwidth of 500 MHz was used. The received analog signals were converted to digital via a digitizer (ATS9352, Alazar Technologies Inc.) with a 12-bit waveform and a real-time sampling rate of 500 MS / s. Dedicated software was used to control the motor-driven stage and store A- and B-scan data.

[0137] The conventional refractive lens used for PA image acquisition was an aspherical lens (A110TM-A, Thorlabs Inc.) with an effective focal length of 6.24 mm and an objective lens LUCPlanFL 40X / 0.60 Ph2 (Olympus). B-scan images were post-processed using dedicated Python code to extract PA intensity images and interpret depth / axis position information. The speed of sound was set to 1540 m / s to convert signal time delay into axis position. A median filter with a window size of 2 x 2 pixels was applied to the depth image map to ensure smooth RGB (red, green, blue) colors.

[0139] 1-5. PA Image Acquisition and Reconstruction

[0140] PA signals acquired data at each location while performing scans in the x and y directions (Fig. 27A). A B-scan dataset was acquired and stored by scanning rapidly in the x-axis direction, while 3D volume data was acquired and stored by scanning slowly in the y-axis direction. After the entire 3D volume data was acquired, PA image reconstruction was performed (Fig. 27B).

[0141] Wiener filters and lowpass filters were applied to remove noise and high-frequency components included in each B-scan dataset. The filtered 3D data was visualized as 2D grayscale PA images through maximum amplitude projection in the z-axis direction and as 2D depth-encoded PA images through depth encoding.

[0143] 1-6. Generation and Characterization of Brain Organoids

[0144] To generate a human whole-brain organoid containing melanin, human induced pluripotent stem cells (hiPSCs; BIONi010-C, K3P53; The European Bank for induced pluripotent stem cells) were cultured using a guided self-organizing method modified based on previous studies.

[0145] To briefly explain, to generate melanin-containing whole-brain organoids, high-quality undifferentiated hiPSCs were cultured in cell culture plates coated with ESC-certified Matrigel and then isolated using ReLeSR. The isolated colonies were dissociated into single cells and harvested into AggreWell embryoid body (EB) forming medium containing Y-27632. On Day 0, the cells were seeded into AggreWell 800 24-well plates, and on Day 1, the medium was replaced with EB forming medium to induce self-organization.

[0146] From day 2 to day 5, the medium was replaced every two days with Dulbecco's modified Eagle's medium (DMEM) / F12-based medium containing a SMAD signaling inhibitor (0.1 mM 2-mercaptoethanol, dorsomorphin, SB-431542). From day 6 onwards, the cells were cultured in a neuro-medium consisting of neurobasal-A medium, vitamin A-removed B-27 supplement, penicillin (100 U / ml), streptomycin (100 μg / ml), GlutaMAX, and a 0.5% basement Matrigel membrane matrix.

[0147] From day 7 to day 24, the medium was replaced using neurodifferentiation medium, and from day 25 to day 42, neuromaturing medium was replaced every two days for neuronal maturation. From day 43 onwards, only neuronal medium was used for maintenance culture and replaced every four days, while minimizing plate vibration.

[0148] For midbrain organoids, a protocol described in existing literature was modified and applied. First, EBs were generated in a manner similar to that of forebrain organoids and inoculated into ultra-low adhesion 96-well plates (7007, Corning) filled with 50 μl of neuroinduction medium (IM). The IM is a mixed medium of DMEM / F12 (Thermo Fisher Scientific) and neurobasal medium, containing N2 supplement, vitamin A-depleted B27, GlutaMAX, MEM NEAA, β-mercaptoethanol, heparin, SB431542, Noggin, CHIR99021, ROCK inhibitor Y27632, and GFR Matrigel.

[0149] On day 7, 100 μl of IM containing SHH-C25II (464-SH, R&D Systems) and FGF8 (423-F8, R&D Systems) was added per well to perform neural patterning. On day 9, the medium was replaced with tissue growth medium containing neurobasal medium, N2 supplement, vitamin A-depleted B27, GlutaMAX, MEM NEAA, β-mercaptoethanol, insulin, laminin, SHH-C25II, FGF8, and GFR Matrigel, and cultured for 24 hours.

[0150] On day 10, the organoids were transferred to a very low-adhesion 24-well plate (3473, Corning) and cultured in 500 μl of organoid medium per well. This medium contained neurobasal medium, N2 supplement, vitamin A-depleted B27, GlutaMAX, MEM NEAA, β-mercaptoethanol, brain-derived neurotrophic factor (BDNF), glial-derived neurotrophic factor (GDNF), ascorbic acid, dibutyryl cyclic AMP (db-cAMP), and GFR Matrigel, and the medium was replaced every 3 days starting from day 10.

[0152] 1-7. Immunohistochemistry and image acquisition

[0153] Immunohistochemical analysis and image acquisition were performed based on previous studies. Briefly, organoids were fixed overnight in 4% paraformaldehyde at 4°C, then washed and infiltrated in 30% sucrose at 4°C for 48–72 hours. Subsequently, they were transferred to a crymold using FSC 22 compound (3801480, Leica) and frozen.

[0154] Frozen organoids were cut using a cryosectioning machine, and the sections were washed and permeabilized with 0.3% Triton X-100 (X100, Merck) in phosphate-buffered saline (PBS) at room temperature for 30 minutes. The sections were blocked in 5% normal horse serum (S-2000, Vector Laboratories) in PBS for 1 hour at room temperature. The primary antibody was diluted in the same solution and incubated overnight at 4°C. The secondary antibody was diluted 1:500 in 3% bovine serum albumin in PBS and incubated for 1 hour at room temperature.

[0155] Finally, after washing, the sections were stained with DAPI (4', 6-diamidino-2-phenylindole) for nuclear staining and mounted on slide glasses. The following antibodies were used on sections mounted on coverslips (0101222, Marienfeld):

[0156] anti-NeuN (neuronal nucleus protein, 1:500; 24307T, Cell Signaling Technology), anti-MAP2 (microtubule-binding protein 2, 1:500; ab254143, Abcam), anti-DAT-Nt (dopamine transporter N-terminal antibody, 1:500; MAB369, Sigma-Aldrich), anti-GIRK2 (G-protein-gated potassium channel Kir3.2, 1:500; APC-006, Alomone Labs), anti-SOX10 (1:500; AF2864, R&D Systems), anti-rat IgG 488 (1:500; A11006, Thermo Fisher Scientific), anti-rabbit IgG 594 (1:500; A11037, Thermo Fisher Scientific), anti-mouse IgG 647 (1:500; A31571, Thermo Fisher Scientific), DAPI (1:5000; D9542, Sigma-Aldrich)

[0157] For the histological visualization of neuromelanin granules, Fontana-Masson staining was performed using a commercial kit (Abcam, ab150669) according to the manufacturer's manual. Images were acquired at maximum resolution using a Leica Thunder DMi8 microscope (Leica) in Lightening or Thunder LVCC (High-Volume Computational Removal) mode. Cell counting and visualization were performed using Imaris software (Oxford Instruments).

[0159] Example 2: Axial multifocal metalens design

[0160] Figure 1B illustrates the method of generating a phase map of a metalens through a forward design. The metalens was designed with a numerical aperture (NA) of 0.15 and a diameter of 1 mm. The axial multifocal beams are spaced approximately 10 μm apart (d) along the optical axis. interfociIt consists of 60 focal points distributed as ). Each focal distance is defined according to the following relationship:

[0161]

[0162] Here f k f1 and f1 are the distances of the k-th and first foci, respectively, where k is the foci number and V is an optimization parameter that adjusts the change in distance between foci thereafter. The maximum distance between foci is the Rayleigh distance (Z_R) of each foci. It is maintained sufficiently smaller than 13.5 μm to ensure an elongated beam shape and uniform intensity within the DOF.

[0163] Phase of the condensing lens to provide a light-gathering function within the same device It is included as the basis for the phase map design. Subsequently, a focal shift phase is added to distribute the focal positions within the designed spatial range. This phase (φ_foci) is defined as follows:

[0164]

[0165] Here, k is the focal number, K is the total number of focal points, n is the refractive index of the medium (1 in air), λ is the operating wavelength of the metalens (532 nm), A is a coefficient for controlling the beam diameter and side lobe intensity, and Loc k is a binary matrix that has 1 if the meta-atom is assigned to focus k, and 0 otherwise.

[0166] By adding a focus splitting phase increment to a phase map that generates a single focus and using a phase correction term πkA, the following output field is generated:

[0167]

[0168] Here, F is the focus function for a specific focal length f, and FT is the Fourier transform. Since FT[Lock(x, y)] can be approximated by δ(x, y) / K, the entire output field is for each focal length f k It is proportional to the sum of the corresponding focus functions. The selection of the Lock matrix has a significant impact on the shape and intensity uniformity of the resulting beam.

[0169] We compared various Lock matrix generation methods and explored ways to induce a uniform beam with low side lobes through numerical optical propagation simulations (see Materials and Methods). Conventional approaches that divide the metalens surface into equal area regions and assign each to a focal point (e.g., ring or fan shapes) resulted in low field uniformity within the DOF, which is due to lattice effects within the metasurface and interference caused by a large number of focal points (Fig. 6). On the other hand, assigning each meta-atom to a specific focal point with a uniform random number distribution reduces ripple or intensity fluctuations across the entire DOF.

[0170] The V parameter controls the change in the spacing between adjacent foci, and by appropriately selecting it, the spacing can be adjusted for farther foci to control the axial intensity distribution within the DOF (Fig. 7). We optimized the V value to 15 nm to account for the natural expansion of the focal waist at long focal lengths, thereby obtaining a uniform intensity distribution across the entire DOF. Since several random functions with fixed random states showed virtually identical results in side lobe intensity or beam diameter, this method does not rely heavily on the random number algorithm (Fig. 8).

[0171] The parameter A plays a key role in controlling the beam diameter by adjusting optical interference between foci. Increasing the value of A allows the beam diameter to be drastically reduced to the diffraction limit level, but at the cost of increased side lobes (Fig. 9). We confirmed that relatively low side lobe intensity is achieved along with performance close to the diffraction limit when A = 0.14. When the value of A is 0.1 or less, the side lobes decrease somewhat but the beam diameter widens; therefore, A The 0.1 level is consistent with observations in existing literature and provides performance close to the diffraction limit while maintaining a moderate side lobe.

[0172] Analysis of the axial intensity profile revealed that the depth of field (DOF) of the beam generated by this metalens is approximately 13.5 times longer than that of a traditional lens with the same NA (Fig. 1C). Although the point spread function (PSF) of the metalens may be similar to that of a Bessel beam, the Bessel beam refracts the incident wave at a fixed angle of refraction, whereas the metalens assigns different meta-atoms according to the focal number, so the wavefront angle is not fixed. A Bessel beam of similar length can be numerically generated using a donut-shaped phase mask (Fig. 10), but its axial intensity profile is accompanied by large oscillations, leading to non-uniform excitation of molecules within the DOF. In contrast, the axial multifocal method exhibits fewer such oscillations and allows for fine-tuning of the axial intensity. Furthermore, while the numerically simulated side lobe intensity of the Bessel beam was always over 16%, it remained at approximately 9% for this multifocal beam.

[0174] Example 3: Fabrication of Metalens and Analysis of Optical Properties

[0175] We used circular nanopillars composed of titanium dioxide particle composites to match the target phase pattern. Titanium dioxide exhibits low loss at an operating wavelength of 532 nm, and nanopillars with a height of 900 nm and a pitch of 300 nm allow for phase control over the entire 2π range by adjusting the diameter from 90 nm to 255 nm (Fig. 11). Nanopillars of the selected diameter and height satisfy the Nyquist sampling criterion (unit cell size < NA·λ / 2) and exhibit high light transmittance in the near field. Additionally, precise phase control is possible without abrupt phase changes when the unit cell size is less than 330 nm (Fig. 12).

[0176] After deriving a meta-atom array that reproduces the target phase map, an axial multifocal metalens was fabricated. First, a silicon master mold was fabricated using electron beam lithography (EBL), and subsequently, a soft mold having a structure complementary to it was fabricated (Fig. 13). The titanium dioxide medium for meta-atom formation is prepared by dispersing titanium dioxide particles in a polymer resin, coating the soft mold, and curing it by applying pressure to a glass slide. After removing the soft mold, the fabrication of a metalens with a diameter of 1 mm is completed. The columnar meta-atoms maintain a precise circular nanocolumn shape with a fixed pitch spacing (Fig. 1D).

[0177] Experimental characterization of the point diffusion function (PSF) was performed using a transmission microscope system with a 532 nm continuous-wave laser source (Fig. 14). After incidenting the focused laser beam onto the metalens, the PSF was imaged in multiple axial planes (see Materials and Methods for details). As a result, it was confirmed that the beam diameter remained constant throughout the entire axially elongated beam range (Fig. 2, AC). The average full width at half maximum (FWHM) was measured to be 2.03 μm and remained nearly constant over a minimum range of 500 μm (Fig. 2, D).

[0178] Figure 2E shows the shape of the elongated beam formed by the metalens, exhibiting high axial intensity uniformity and weak side lobes. The standard deviation (SD) of intensity across an axial distance of 500 μm within the DOF did not exceed 5%, which is consistent with existing methods based on the extended Nijboer-Zernike theory. The axial intensity full width at half maximum was measured to be 580 μm, which is in very good agreement with the numerical simulation results (Figure 2F).

[0179] The focusing efficiency of the metalens is 20%, which is defined as the ratio of light focused at the central focal point out of the laser beam incident on the metalens. This relatively low efficiency is attributed to the following factors: (i) manufacturing defects, (ii) a reduction in the absolute output of the main lobe due to the presence of the phase correction term (A), (iii) Rayleigh scattering caused by non-uniform fluctuations in the refractive index within the meta-atom, and (iv) a decrease in efficiency compared to a single-focus lens with increasing axial length of the beam.

[0180] Nevertheless, this method enables a highly uniform field profile within the DOF, precise focusing, and focusing efficiency comparable to that of a meta-axicon. The numerical optimization process is also relatively simple, allowing for the easy implementation of various beam aspect ratios.

[0181] Background noise originating from unmodulated or scattered light can have a negative effect on PA measurements, but the effect is significant only near the metalens interface and is negligible at the distance where the elongated beam is generated, as confirmed by optical PSF measurements (Fig. 2F). In particular, it is also noteworthy that while the output reaching the main lobe in the case of the Bessel beam is only about 5%, the focusing efficiency of the multifocal metalens reported in this study significantly exceeds this.

[0183] Example 4: Performance Verification of MeD-PAM

[0184] We developed a self-made transmission-mode MeD-PAM system (Fig. 15). To demonstrate the advantages of the MeD-PAM, we acquired photoacoustic (PA) images using a metalens and two types of conventional lenses (an aspherical lens with an effective NA of 0.12 and an objective lens with an NA of 0.19). The power density incident on the sample was kept the same for all three lenses to ensure a fair comparison of performance.

[0185] According to the experimental results (Fig. 16), the DOF (580 μm) of the metalens is about 3.8 times longer than that of the aspherical lens (152 μm) and about 27.1 times longer than that of the objective lens (21.4 μm). Since DOF is inversely proportional to the square of NA (43), it is normal for the objective lens to theoretically have a short DOF close to the diffraction limit. In the case of the aspherical lens, the theoretical DOF is 49 μm, but it is possible that the DOF has become longer geometrically due to aberration. It is noteworthy that the lateral resolution of the metalens is superior to that of the aspherical lens, which suggests that it is a very promising device for high-resolution PAM imaging.

[0186] Carbon fibers were used as phantom samples with light absorption contrast. Tensioned carbon fibers were fixed at different angles and heights to create axial surface inclination, a condition that is difficult to observe with conventional PAMs having a shallow DOF. A simple method for acquiring and reconstructing PA data is described in Materials and Methods.

[0187] First, the axial profile of the PA signal was evaluated while sliding the metalens beam axially through a horizontally fixed carbon fiber within the water surface. As a result of recording the PA signal between the metalens and the carbon fiber at various distances, it was confirmed that the PA signal profile matched the optical measurements well (Fig. 17).

[0188] Next, to demonstrate the metalens-based compact PA imaging capabilities, two carbon fiber phantom samples were fabricated tilted at 4° 45' and 9° 30' relative to the substrate plane, respectively. We confirmed that the MeD-PAM clearly images axially tilted carbon fibers at significantly longer imaging distances than a single-focus refractive lens (Fig. 3, AC). The power density incident on the sample from each lens was adjusted to be equal. Through time-resolved ultrasonic sensing, the depth information of the carbon fibers could be decoded (see Materials and Methods for details), and the color-encoded depth information confirms the axial tilt of the carbon fibers.

[0189] PA images with an elongated DOF were able to detect carbon fiber signals at a distance at least 2.6 times longer than that of an aspherical lens and at least 6.3 times longer than that of an objective lens. Additionally, while MeD-PAM imaged both carbon fibers with a single plane scan, the conventional PAM failed to observe them separately due to the large axial positional difference between the two carbon fibers. The carbon fiber PA image acquired with an aspherical lens showed a non-negligible signal lobe at the top of the axis (Fig. 3B), which appears to be due to light intensity ripple in the out-of-focal region (Fig. 16).

[0190] We calculated line profiles along the three lines (A, B, C) of Fig. 3A and demonstrated that MeD-PAM provides well-maintained lateral resolution at different depths (Fig. 18). The FWHM of the three lines was measured to be 11.5 ± 0.27 μm, which is in good agreement with the known diameter of the carbon fiber (approx. 5-10 μm).

[0191] Additionally, we used a carbon leaf with axial positional variation as a second phantom sample, positioned diagonally (Fig. 8, DF). MeD-PAM was able to clearly image the fibrous structure within the carbon leaf within a field of view of approximately 3 mm and precisely verify axial positional information. In contrast, PA images acquired with a conventional lens showed clear leaf images only in a limited area within the DOF, while leaf areas outside the DOF were blurry or had weakened signals and were not imaged.

[0193] Example 5: 3D High-Resolution and Deep Tissue Neuromelanin PA Imaging in Living Brain Organoids

[0194] Based on, but with some modifications from, previous research, we used a guided self-organizing method to generate brain organoids for culturing human-induced pluripotent stem cells (hiPSCs) (see Materials and Methods and Fig. 19 for details). We photoacoustically acquired PA images of neuromelanin distribution within three-dimensional forebrain and midbrain organoids using MeD-PAM and analyzed them to verify the effective applicability of this technology to biological samples (Fig. 4).

[0195] First, we prepared forebrain organoids containing melanin to visualize the internal shape of neuromelanin signaling. This was because it was hypothesized that while most mature forebrain neurons are glutamatergic neurons, some expressed dopaminergic neurons (Fig. 20) could also produce internal neuromelanin. We confirmed that melanin-producing cell types—the neural crest progenitor cell marker SRY-box transcription factor 10 (SOX10) and the dopaminergic neuron marker dopamine transporter (DAT)—were expressed in both forebrain and midbrain organoids (Fig. 21).

[0196] The PA maximum amplitude projection image (PA MAP) of Fig. 4A was obtained from a forebrain organoid, and the four arrows in Fig. 4A and D point to the same location. In the brightfield (BF) image of Fig. 4D, the black neuromelanin pigment within the forebrain organoid is observed, but most of the interior, excluding the top left corner, appears blurry, and is particularly pronounced at the location indicated by the four arrows. It is difficult to observe neuromelanin formed mainly within the forebrain organoid using conventional optical techniques. On the other hand, the PA MAP image shows the detailed structure of the neuromelanin signal throughout the organoid. In particular, the location indicated by the four arrows shows a distinct difference from the BF image, which demonstrates that the 3D structure within the organoid can be effectively distinguished through the extended DOF of MeD-PAM.

[0197] Figure 4B shows a PA image in which neuromelanin signals are color-encoded according to depth information. Blue represents signals near the surface, and red represents signals deeper within the tissue. Although these depth-encoded images may be affected by background signals, they are useful as an auxiliary means for intuitively visualizing the depth distribution of melanin.

[0198] For a more detailed analysis, cross-sectional images at different depths of the 3D PA were examined in Fig. 4C. The first cross-section was measured along the red dashed line in Fig. 4A, and the red arrow indicates the neuromelanin PA signal at a depth of 0.5 mm inside the organoid (#1 Fig. 4C). This signal is observed faintly in the image, making it difficult to estimate the depth. The second cross-sectional image was acquired along the yellow dashed line and shows the distribution of neuromelanin deep within the organoid (#2 Fig. 4C). The top left mainly represents neuromelanin near the surface, and the two yellow arrows indicate a signal corresponding to a depth of approximately 0.31 mm, but this also appears faintly in the image.

[0199] Figure 4E shows a 3D-rendered forebrain organoid visualizing neuromelanin signals at various depths simultaneously. We performed pathological tissue staining experiments to confirm that the signals observed in PA and BF images in the forebrain organoid were actual neuromelanin. As a result of Fontana-Masson staining, partial accumulation of neuromelanin was observed in the form of black pigment (Figure 4F, Figure 20B). Through correlational image analysis of PA images and Fontana-Masson staining, we demonstrated that PA contrast originated from neuromelanin and confirmed the neuromelanin specificity of MeD-PAM (Figure 22). In addition, the expression of dopaminergic neuron markers (DAT) and neural crest markers (SOX10) within the forebrain organoid was verified using immunohistochemical methods (Figures 20, 21).

[0200] Next, PA imaging was performed on midbrain organoids, which are known to accumulate a rich amount of neuromelanin. Since dopaminergic neurons constitute the major cell population in midbrain organoids, neuromelanin production is active. The PA MAP image in Fig. 4G was obtained from the midbrain organoid in Fig. 4J, and the PA and BF images confirm that neuromelanin is more widely distributed across the entire surface compared to forebrain organoids. Additionally, the depth-encoded midbrain PA image in Fig. 4H shows that neuromelanin is primarily produced near the surface of the organoid (blurred light blue spectrum).

[0201] For a more precise analysis, the 3D PA image cross-section of Fig. 4I was examined. The first cross-section was acquired along the red dashed line of Fig. 4G, and the red arrow indicates the neuromelanin PA signal at a depth of 0.44 mm inside the organoid (#1 Fig. 4I). This cross-section shows a uniform distribution of neuromelanin across the entire surface, and the signal at the deep location indicated by the red arrow is not clearly visible in the image. The second cross-section was measured along the yellow dashed line and also shows that neuromelanin is uniformly distributed near the surface (#2 Fig. 4I). The signal at a depth of approximately 0.37 mm indicated by the yellow arrow is also not clearly observed in the image.

[0202] Figure 4K shows a 3D-rendered midbrain organoid capable of simultaneously visualizing neuromelanin signals at various depths. We confirmed that the majority of mature neurons within the midbrain organoid were dopaminergic neurons (DAT+ / MAP2+) (Figure 20), and abundant dark pigment accumulation was also observed in Fontana-Masson staining images (Figure 4L).

[0203] We performed PA imaging on forebrain and midbrain organoids using an axial multifocal metalens and a conventional lens (Fig. 23). The aspherical lens and objective lens showed similar lateral resolution and image quality, but failed to effectively visualize neuromelanin signals at depths exceeding 200 μm from the glass-organoid interface.

[0205] Example 6: Quantitative analysis of neuromelanin accumulation in living brain organoids using MeD-PAM

[0206] We extended the capabilities of MeD-PAM to quantitatively analyze the neuromelanin content in brain organoids. Visual clarity was enhanced by superimposing brightfield (BF) images onto images acquired with MeD-PAM (Fig. 5). First, forebrain and midbrain organoids with similar culture periods were compared, and the results confirmed that midbrain organoids contained a significantly higher amount of neuromelanin than forebrain organoids (Fig. 5, A and B).

[0207] For quantitative analysis, neuromelanin content was defined as the total number of pixels in PA images with signals exceeding three times the standard deviation (SD) of the background noise. Neuromelanin content was normalized based on the midbrain organoid that showed the highest signal in this study. The midbrain organoid at day 132 of culture contained approximately 3.4 times more neuromelanin than the forebrain organoid at day 150, which is at a similar maturation stage (Fig. 5C).

[0208] As confirmed in the superimposed images, neuromelanin was uniformly distributed throughout the midbrain organoid, whereas in the forebrain organoid, it was observed only in the form of some dense clusters. Immunohistochemical analysis and Fontana-Masson staining results confirmed that the formation of dopaminergic neurons and neuromelanin granules occurred in a similar manner (Fig. 20), and this neuromelanin granule pattern is presumed to be the source of the signal observed in the MeD-PAM images.

[0209] Finally, a series of whole-brain organoids with different culture durations were analyzed to track the progression of melanogenesis (Fig. 5, A, D, and E). Neuromelanin granules were observed to gradually increase during the maturation period from 150 to 216 days (Fig. 5, F). Thus, MeD-PAM presents potential as a compact imaging system capable of quantitatively monitoring melanogenesis within whole-living brain organoids.

[0211] In this study, a titanium dioxide-based metalens was designed to generate an elongated beam with a high aspect ratio, featuring a uniform intensity distribution and weak side lobes within the depth of field (DOF). Notably, this metalens demonstrated a DOF approximately 3.8 times longer than that of commercial aspherical lenses and improved lateral resolution. By applying this optical control technology to the MeD-PAM, a new possibility was opened to image the three-dimensional distribution of analytes with absorption contrast at a wavelength of 532 nm.

[0212] The proposed metalens features a true planar structure with a sub-micrometer thickness and a precisely controlled unit structure geometry, generating an elongated beam solely through phase modulation. The MeD-PAM acquires high-resolution images using a point-based scanning method, and the miniaturization of the metalens is a major advantage, enabling a much more compact system configuration compared to axicons or objective lenses used in conjunction with the DOE. Intensity fluctuations across the DOF are maintained at less than 5%, allowing for uniform axial excitation even on non-uniform specimens. Furthermore, the ability to adjust the polarization, wavelength, and orbital angular momentum of the metasurface will enable future expansion into a multi-functional MeD-PAM system.

[0213] MeD-PAM is a transmissive structure designed to sufficiently detect PA signals in an in vitro environment, and based on OR-PAM, it can acquire high-resolution PA signals from biological tissues up to a depth of about 1 mm. Therefore, while this system is suitable for the current size and characteristics of brain organoids, its performance can be improved for future organoids with a size of 2-3 mm by adjusting the optical focal size and light intensity.

[0214] The selection of a wavelength of 532 nm is based on the unique optical properties of melanin in the visible and near-infrared regions. In this wavelength range, the absorption coefficient of melanin is significantly higher than that of other biomolecules, including hemoglobin. While melanin signals can overlap with hemoglobin signals in general tissues, such interference does not exist in the hBO used in this study because it lacks blood. According to a study by Berezhnoi et al., melanin exhibits a strong signal without background signal across the 488–900 nm range, and this study also confirmed that the 532 nm wavelength is highly effective for clearly and specifically imaging neuromelanin.

[0215] In normal tissues, the maximum penetration depth of green light is limited to about 1.5 mm due to hemoglobin absorption, but deeper penetration is possible in hBO, which lacks blood. For example, deeper penetration is possible at 650 nm, and in the near-infrared region, it can reach up to 3 mm depth without loss of clear contrast and resolution.

[0216] While this study focused on the detection of neuromelanin, the following considerations can be made to expand the practicality of MeD-PAM. First, multispectral PA imaging can improve the accuracy of distinguishing between neuromelanin and hemoglobin by generating multiple wavelengths through stimulated Raman shifting. This enables differentiation even in cases where it is difficult to distinguish between high concentrations of low-absorbing substances and low concentrations of high-absorbing substances using a single wavelength. Second, real-time imaging and analysis techniques are useful for observing dynamic responses in organoids, such as drug delivery reactions, and are feasible based on high-speed PA microscopy.

[0217] Traditional Gaussian beams have a very shallow depth of focus, making it difficult to image a wide area at once when there are significant variations in specimen height. However, the extended DOF of MeD-PAM enables a relatively wide field of view with a single scan, even on irregular specimen surfaces, and improves imaging efficiency by reducing the need for refocusing. In particular, this metalens demonstrated superior performance compared to existing Gaussian beam-based PAMs, as it can effectively visualize neuromelanin with cell-level resolution in the axial direction up to a range of 500 μm. The axial resolution of MeD-PAM was confirmed to be approximately 90 μm (see Materials and Methods), and the introduction of a high-frequency ultrasound transducer is required for more precise analysis.

[0218] We acquired PA B-scan images of neuromelanin in organoids using ultrasound transducers at 5, 20, and 40 MHz and compared them (Fig. 24). At 5 MHz, axial resolution was degraded due to strong reverberation, whereas this problem did not occur at 20 and 40 MHz. The signal-to-noise ratios (SNR) obtained from the organoid PA images in Figs. 4 (A and G) were 43.9 dB and 45.8 dB, respectively, and the SNR calculation followed the following formula:

[0219]

[0220] The noise standard deviations (SD) of the measurements are 63.7 and 84.1, respectively. This result demonstrates that MeD-PAM provides high signal contrast.

[0221] Theoretically, 3D distribution imaging of neuromelanin is also possible via axicons, but additional deconvolution is required to reduce the influence of large side lobes and perform accurate quantitative analysis. In particular, the partial presence of neuromelanin pigment in melanin-containing forebrain organoids may not be solely due to partially expressed dopaminergic neurons; since the expression of SOX10+ neural crest-derived cells has also been confirmed, and considering that both neurons and melanocytes originate from neural crest cells, it may be an effect of aging.

[0222] We successfully monitored the melanogenesis process in living organoids at various stages of maturation and quantitatively confirmed the distribution and location of neuromelanin. Neuromelanin is closely associated with the pathology of Parkinson's disease (PD), and the loss of dopaminergic neurons within the substantia nigra is closely linked to a decrease in neuromelanin. Therefore, this system can be utilized as a tool to effectively monitor the progressive neurodegeneration of PD and can contribute to the development of rapid and portable diagnostic tools and drug efficacy screening platforms based on slide-free biological samples.

Claims

Claim 1 A method for designing an axial multifocal metalens in which multiple focal points are arranged along an optical axis, comprising the following steps: (a) setting a target operating wavelength λ and a numerical aperture (NA), and a relationship Focal length f corresponding to multiple focal positions on the optical axis according to k A step of determining, wherein V is a constant as a focal interval optimization parameter, and d interfoci Is (b) focal interval; depending on the (x, y) coordinates representing the two-dimensional position on the metalens surface, the condensing lens phase function φ as follows lens (x,y) and focus shift phase function φ foci Step of calculating (x,y) respectively, , , above, k is the focal number, K is the total number of focal points, n is the refractive index of the medium, λ is the target operating wavelength of the metalens, A is a coefficient for controlling the output light diameter and side lobe intensity, and Loc k (x, y) is a binary matrix for assigning each phase function to a meta-atom at a specific location; and (c) combining the two phase functions to obtain a final phase map φ as follows map Determine (x,y), and , for each meta-atom of the metalens lattice, Loc k A step of assigning a phase value such that if (x,y)=1, the phase value of the k-th focus is assigned, and if (x,y)=0, the corresponding phase value is not assigned. Claim 2 A method for designing an axial multifocal metalens according to claim 1, wherein the target operating wavelength is 100 to 1500 nm. Claim 3 A method for designing an axial multifocal metalens according to claim 1, wherein the target numerical aperture is 0.1 to 1.

0. Claim 4 A method for designing an axial multifocal metalens, wherein, in claim 1, the focal interval optimization parameter V is selected from a range of 5 nm to 30 nm such that the standard deviation of the axial intensity distribution of the output light is 10% or less. Claim 5 In claim 1, the focal interval d interfoci A method for designing an axial multifocal metalens in which the length is 1 to 50 μm. Claim 6 A method for designing an axial multifocal metalens according to claim 1, wherein the coefficient A for controlling the output light diameter and side lobe intensity is selected from the range of 0.1 to 0.

5. Claim 7 A method for designing an axial multifocal metalens according to claim 1, wherein the number of focal points K is 10 to 100. Claim 8 In paragraph 1, the above Loc k An axial multifocal metalens design method in which the (x,y) matrix is ​​generated by fairly randomly assigning each nanostructure on the metalens surface with a probability of 1 / K. Claim 9 An axial multifocal metalens having multiple focal points arranged along an optical axis, designed according to the design method of any one of claims 1 to 8. Claim 10 In claim 9, the metalens is an axial multifocal metalens composed of HfO2, TiO2, ZnO, SiO2, Al2O3, ZrO2, Si3N4, or a combination thereof. Claim 11 In claim 9, the diameter of the metalens is 1 x 10 2 μm to 1 X 10 4 Axial multifocal metalens that is µm in size. Claim 12 In claim 9, the metalens is an axial multifocal metalens having a depth of focus (DOF) of 100 to 1000 μm. Claim 13 An axial multifocal metalens according to claim 9, wherein the metalens comprises a plurality of cylindrical meta-atoms having a diameter of 50 nm to 300 nm, a height of 100 to 1000 nm, and a combination thereof. Claim 14 In claim 9, the metalens is an axial multifocal metalens that generates a beam with an aspect ratio of 100 to 500 while having a width within ± 10% of the lens's diffraction limit. Claim 15 In claim 9, the axial multifocal metalens is configured such that the standard deviation of the output light intensity is maintained at 5% or less throughout the entire depth of focus (DOF) range. Claim 16 A photoacoustic detection device comprising the metalens of claim 9. Claim 17 In claim 16, the photoacoustic detection device is capable of acquiring a three-dimensional volumetric image without focus shift. Claim 18 A photoacoustic detection method comprising the steps of: transmitting light through a metalens of claim 9 to irradiate a sample; and collecting ultrasound emitted from the sample. Claim 19 A photoacoustic detection method according to claim 18, wherein the sample is a brain organoid. Claim 20 delete

Images