Spectrally resolved, high-resolution 3D localization microscopy

Abstract

Localization microscope comprising an imaging device for imaging a sample light-emitting sample (2) from a focal plane (3) into an image plane (9), - wherein the imaging device has an optical manipulation device (18) which is designed to influence a point image blurring function of the imaging in a depth-dependent manner and influences the point image blurring function of the imaging such that a point emitter in the image plane is mapped into a point emitter image (30, 32) which is rotationally asymmetrically distorted, - wherein a shape and / or orientation of the rotationally asymmetric distortion of the point emitter image (30, 32) depends on the position of the point emitter relative to the focal plane (3) and a wavelength of the sample light and - wherein the optical manipulation device (18) has a first and a second anisotropy element (22, 24), each of which anisotropically influences the point image blurring function of the imaging in order to generate the rotational asymmetry of the point emitter image (30, 32), and each of which has an anisotropy axis (23, 25), characterized in that - the two anisotropy elements (22, 24) are arranged one behind the other in the direction of the image and their anisotropy axes (23, 25) are at an angle to each other - and both anisotropy elements (22, 24) each have a neutral wavelength (λ1, λ2) at which they do not anisotropically influence the point image blurring function of the imaging, - where the neutral wavelengths (λ1, λ2) are different.

Description

[0001] The invention relates to a localization microscope comprising an imaging device for imaging a sample light-emitting sample from a focal plane into an image plane, wherein the imaging device comprises an optical manipulation device configured for depth-dependent influencing a point image blurring function of the imaging and influencing the point image blurring function of the imaging in such a way that a point emitter in the image plane is imaged into a point emitter image that is at leasthas two image flaps, wherein the relative positions of the image flaps of the point-emitter image depend on the position of the point-emitter to the focal plane and a wavelength of the sample light, and wherein the optical manipulation device has a first and a second anisotropy element, which each anisotropically influence the point-image blurring function of the imaging in order to generate the image flaps of the point-emitter image, and each have an anisotropy axis.

[0002] The invention further relates to a method for localization microscopy, wherein a sample light-emitting sample is imaged from a focal plane into an image plane, wherein a point image blurring function of the imaging is influenced depth-dependently by means of an optical manipulation device such that a point emitter in the image plane is imaged into a point emitter image which has at least two image lobes, wherein the relative positions of the image lobes of the point emitter image depend on the position of the point emitter to the focal plane and a wavelength of the sample light, and wherein a first and a second anisotropy element are used in the optical manipulation device, which each anisotropically influence the point image blurring function of the imaging in order to generate the image lobes of the point emitter image, and each have an anisotropy axis.

[0003] Various methods for overcoming the diffraction limit in microscopy have been developed in the prior art. From WO 2006 / 127 692 A2 or DE 102006021317 B3, a method abbreviated as PALM (photo-activated light microscopy) is known that uses a labeling agent to image a sample, which can be activated, for example, by optical radiation. Only in its activated state can the labeling agent emit specific fluorescence radiation. Non-activated molecules of the labeling agent emit no fluorescence radiation, or at least no noticeable fluorescence, even after being irradiated with excitation radiation. The activation radiation is therefore generally referred to as the switching signal.In the PALM method, the switching signal is applied in such a way that at least a certain proportion of the activated labeling molecules are spaced from neighboring activated molecules at a distance sufficient to be separated by the optical resolution of the microscopy or subsequently separable using image processing techniques. This is referred to as isolating at least a subset of the fluorescent emitters. The term "isolated" therefore refers to the optical resolution of the imaging. After the sample light is captured, the center of the resolution-limited radiation distribution of these isolated emitters is determined. From this, the position of the emitters can be calculated with higher accuracy than the optical resolution would otherwise permit. This process is called the localization step.The increased resolution achieved through computational determination of the diffraction distribution's center of gravity is also referred to as "super resolution" in English-language literature. This requires that at least a subset of the activated emitters in the sample be distinguishable, i.e., isolated, with optical resolution. Their positions can then be determined with higher accuracy; they can be localized. Since the high resolution is obtained through computational localization of the locations of emitters emitting sample light, the term "localization microscopy" has become established for this type of microscopy, which has been further developed in various directions and also for emitters that can be isolated without a switching signal.

[0004] Several advanced training programs address achieving three-dimensional resolution, i.e., resolution that is not only given in the lateral direction, as explained in WO 2006 / 127692 A2, but also in the depth direction. Examples include DE 10 2009 060 490 A1, DE 10 2012 200 344 A1, and DE 10 2012 224 306 A1. For three-dimensional localization microscopy, the point image blurring function must be modified to encode the position of a sample light emitter in the depth direction within the image using a lateral property of the emitter image.

[0005] WO 2012 / 039636 A2 discloses a localization microscope or a method for localization microscopy of the type mentioned above. Here, phase manipulation is performed in the pupil plane by using two opposing phase ramps to ensure that each point-emitter image, i.e., the image of an isolated point center, has two image lobes. Depending on the depth of field, i.e., the z-position of the point center, these image lobes shift relative to each other.

[0006] This approach becomes problematic when spectral resolution is also required. It is often essential in fluorescence microscopy, as a single fluorescence stain typically does not provide functional information in cell biology, the primary application of high resolution. Classically, i.e., with normal resolution, multicolor imaging is achieved either through rapid sequential imaging in different color channels, for example, by using a suitable, fast filter change, or with a color splitter in the detection beam path and a separate detector for each color channel. However, both of these approaches are disadvantageous or unusable for localization microscopy. Since localization microscopy requires time series of up to 100,000 individual images, the sequential approach would be extremely time-consuming.Furthermore, dynamic processes could not be investigated because, by its very nature, the recordings in the color channels cannot be simultaneous. While color splitting in the detection channel would avoid this problem, it requires a very complex detector for each color channel. In addition, the detectors must be aligned with pixel-level precision.

[0007] For two-dimensional localization microscopy, the detection of multiple colors is known. For this purpose, the point image blurring function is modified so that, unlike in the aforementioned 3D localization microscopy, it encodes color information rather than depth. This is implemented, for example, in DE 102012201286 A1.

[0008] To achieve three-dimensional and spectral resolution, the point image blurring function must be influenced, i.e., encoded, with respect to both the depth position of the emitters and the spectral information. The aforementioned WO 2012 / 039636 A2 provides an approach for this purpose, employing pupil division. One half of the pupil is occupied by a phase-manipulating element for depth resolution, i.e., 3D localization microscopy, while the other half of the pupil contains an emission filter. This embodiment, described in WO 2012 / 039636 A2, Fig. However, as shown in Figure 14, this has two disadvantages. The emission filter is a spectrally selective element. By its very nature, such an element always leads to the suppression of photons, and thus, in localization microscopy, which is particularly photon-sensitive, to either lower resolution or a much longer measurement time. Furthermore, the intensity between the image flaps becomes very uneven for different wavelengths. Equal intensity across the image flaps can only be achieved for a single wavelength. This further complicates three-dimensional localization and ultimately degrades the result.

[0009] The invention is therefore based on the objective of providing a localization microscope and a method for localization microscopy in which no photons are spectrally suppressed and yet both spectral resolution and depth resolution are achievable.

[0010] The invention is characterized in claims 1 and 7. The dependent claims relate to preferred embodiments of the invention.

[0011] The localization microscope has an imaging device for imaging a sample that emits light. This could, for example, be a fluorescent sample that has been appropriately excited. The sample is imaged onto an image plane at a focal plane.

[0012] The imaging device has an optical manipulation unit. This unit is configured to influence the point image blur function of the imaging process in a depth- and wavelength-dependent manner. It affects the point image blur function of the imaging process (hereinafter also referred to as "PSF") such that a point emitter is imaged onto a point emitter image that is rotationally asymmetrically distorted, e.g., has at least two image lobes. The anisotropic distortion of the point emitter image depends on the position of the point emitter relative to the focal plane and on a wavelength of the sample light.

[0013] In one embodiment of the invention, the rotational asymmetry is represented by image lobes, and the configuration refers to their relative positions. The image lobes then have relative positions to each other that depend both on their position relative to the focal plane and on the wavelength of the sample light. Another embodiment of the rotational asymmetry is an anisotropic distortion of the point-emitter image.

[0014] The optical manipulation device has a first and a second anisotropy element. Each anisotropy element anisotropically influences the point-image blurring function of the imaging process and generates the anisotropic distortion. It is therefore designed as an anisotropic distortion element. During imaging, as the term "localization microscopy" implies, isolated emitters are present in the sample to emit the sample light.

[0015] Each anisotropy element is characterized by an anisotropy axis. The two anisotropy elements are arranged one behind the other in the imaging direction. Their anisotropy axes are not parallel to each other. The angle formed by the anisotropy axes depends on the implementation of the anisotropy elements in various embodiments; however, the anisotropy axes are generally at an angle to each other.

[0016] Each first anisotropy element anisotropically affects the point image blurring function, but is neutral for a first wavelength. The opposite is true for the second anisotropy element. It also affects the point image blurring function anisotropically, but is neutral for a second wavelength. This can be interpreted as each anisotropy element being neutral for its assigned, individual wavelength. The anisotropy elements thus differ in the position of their anisotropy axes and in their neutral wavelengths. In this way, the orientation and / or shape of the distortion is not only depth-dependent but also wavelength-dependent.

[0017] The localization microscopy method is implemented analogously. A sample emitting light is imaged from a focal plane onto an image plane. A point-image blur function of this image is manipulated depth-dependently by an optical manipulation device such that a point emitter is imaged onto a point-emitter image that is anisotropically distorted. The nature of the anisotropic distortion depends on the position of the point emitter relative to the focal plane and on the wavelength of the sample light. The optical manipulation device comprises a first and a second anisotropy element, each of which anisotropically influences the point-image blur function of the image to generate the anisotropic distortion. Each anisotropy element is characterized by its anisotropy axis. The two anisotropy elements are positioned one behind the other in the imaging direction, and their anisotropy axes form an angle with each other.The two anisotropy elements do not affect the first or second wavelength.

[0018] By combining the anisotropy elements in series and considering the angle enclosing their anisotropy axes, and in combination with the fact that the anisotropy elements are neutral for the dot image distortion function at different wavelengths, the anisotropic distortions in the image plane for the first and second wavelengths exhibit a structure that extends, for example, along different structural axes, since only one of the anisotropy elements acts in each axe. In this way, the wavelength can be decoded from the structural axis direction. The relative position of elements of the anisotropically distorted dot-mitter image encodes the depth for this color channel.

[0019] The generation of anisotropic distortions would be highly disruptive in a conventional wide-field microscope; a usable image would be impossible. However, this is not a problem in localization microscopy, as at least some of the point emitters emitting into the sample light are isolated, as the term "localization microscopy" implies. The generation of anisotropic distortion, which would manifest as unacceptable image smearing in conventional wide-field microscopy, is now uncritical and can be evaluated in the localization step not only with regard to the lateral position of the corresponding point emitter but, more importantly, also with regard to the depth and color channel of the point emitter. For this purpose, the shape and / or orientation of the distortion, e.g., the relative position of elements of the point emitter image (e.g., image lobes), is determined. The center of the distortion indicates the lateral coordinate of the point emitter.The orientation and / or shape of the z-coordinate encodes the color channel in which the sample light was emitted by the corresponding point emitter.

[0020] For this evaluation, the microscope typically has a corresponding evaluation unit, which can be implemented as a computer, for example. The evaluation can be performed online, i.e., directly at the time of acquisition of each individual image, or alternatively after a large number of individual images have been recorded. This is generally known from localization microscopy, but not for the combination of lateral position determination in the image, depth information, and color channel.

[0021] Anisotropic distortion of the point-emitter image typically occurs when an emitter emits the sample light at exactly one of the wavelengths. However, cases are also possible where emitters emit the sample light at a wavelength between the two neutral wavelengths. In such cases, the point-emitter image represents a mixture or transitional form of the point-emitter image at the first wavelength, which has a first structural axis, and the point-emitter image at the second wavelength, which has a second structural axis. This form can be distinguished from the forms (so-called pure forms) at the first and second wavelengths, so that, in general terms, the shape of the point-emitter image encodes both the depth and the wavelength. In this way, more than two color channels can be realized with just two anisotropy elements.

[0022] In a simple configuration, the anisotropy elements can each be designed as phase elements. The anisotropic distortion then takes the form of at least two image flaps. Each phase element has an effect on the PSF (phase-flash distortion), but not at its individual wavelength, where it is neutral with respect to the PSF. In this way, each anisotropy element does not produce image flaps when emitted at the neutral wavelength. Typically, the anisotropy element has a PSF-influencing function, i.e., image-flag-generating function, at all wavelengths outside its neutral wavelength. This is permissible and, as explained, even allows for an increase in the number of color channels, as long as the anisotropy element has no phase-influencing effect at one wavelength, i.e., is neutral, whereas the other anisotropy element distorts at that wavelength.It is therefore useful to effectively characterize the anisotropy element by its (lack of) effect at the neutral wavelength, which can then be understood as the wavelength of the other anisotropy element that is not neutral there.

[0023] A particularly simple embodiment of a phase element for realizing one of the anisotropy elements consists of a wedge pair of two glasses with different dispersion properties, chosen such that the total refractive index of the wedge pair vanishes at the neutral wavelength. Thus, the two glasses cancel each other out with respect to their dispersive effect at the neutral wavelength. Otherwise, and especially at the neutral wavelength assigned to each wedge pair, they exhibit a dispersive effect and thus produce the image lobes.

[0024] The wedge angles and refractive indices of the wedge pair are chosen such that the neutral wavelength is transmitted without deflection. Shorter and longer wavelengths, respectively, are deflected. Because the wedge pairs are crossed, the desired differentiation is achieved with regard to both depth and color channel. The advantage of this approach is that technologically very simple glasses can be used, since only two parameters for realizing the wedge pair can be varied: the wedge angle and the refractive index of each wedge within the pair. This results in a total of four parameters: two refractive indices and two wedge angles. In a particularly simple embodiment, the wedge angles are chosen so that the wedge pair has parallel outer surfaces, meaning the entrance surface of the phase element is parallel to the exit surface. The wedge angles are then opposite or complementary to zero.

[0025] For phase elements, it is preferable to arrange the anisotropy axes at an angle between 80 and 90°, preferably at 90° to each other.

[0026] Furthermore, phase elements for the purposes of the invention are most effective when arranged in a pupil plane of the detection beam path through which the sample light is imaged into the point-emitter images. An arrangement in or near a pupil plane, e.g., the objective pupil plane, is therefore preferred.

[0027] Another alternative for the anisotropy elements are anisotropy lenses that are neutral at one wavelength and otherwise exhibit anisotropy over the remaining spectral range. The anisotropic distortion is then a longitudinal distortion of the point-emitter image. In this case, the anisotropy axes are characterized by the cylinder axis of the anisotropy lenses. Preferably, they lie at an angle between 35° and 55°, particularly preferably at 45° to each other.

[0028] It is understood that the features mentioned above and those to be explained below can be used not only in the combinations given, but also in other combinations or on their own, without leaving the scope of the present invention.

[0029] The invention is explained in more detail below with reference to exemplary embodiments and the accompanying drawings, which also disclose essential features of the invention. These exemplary embodiments serve only for illustration and are not to be interpreted as limiting. For example, a description of an exemplary embodiment with a plurality of elements or components is not to be interpreted as meaning that all of these elements or components are necessary for implementation. Rather, other exemplary embodiments may also contain alternative elements and components, fewer elements or components, or additional elements or components. Elements or components from different exemplary embodiments may be combined with one another unless otherwise specified. Modifications and variations described for one of the exemplary embodiments may also be applicable to other exemplary embodiments.To avoid repetition, identical or corresponding elements in different figures are designated with the same reference symbols and are not explained multiple times. The figures show: Fig. 1. A schematic representation of a localization microscope, Fig. 2 and Fig. 3 Cross-sectional views of various optical manipulation elements used in the microscope of the Fig. 1 be used, Fig. 4A to 4D illustrations to explain the effect of the optical manipulation element, Fig. 5 dispersion curves for glasses used in the manipulation element of the Fig. 4A to 4D can be used, Fig. 6A to 6J distortions that produce a point emitter image in the microscope of the Fig. 1 for different wavelengths and depth positions of the point center in localization microscopy, Fig. 7 a modified microscope of the Fig. 1, wherein the modification concerns the design of an optical manipulation element, and Fig. 8A to 8C representations according to the Fig. 4A, Fig. 4B and Fig. 4D, but now for the microscope of the Fig. 7.

[0030] Fig. Figure 1 shows a microscope 1 designed for the aforementioned localization microscopy. It images a sample 2, located in a focal plane 3, with an objective 4 via a tube lens 6 onto a detector 8, for example, a camera, which lies in an image plane 9. The entire imaging process is performed in wide-field mode. The sample 2 is illuminated by a light source 10, which directs the corresponding illumination radiation onto the sample 2 via a beam splitter 11 and the objective 4. An illumination optic 12, 14, the implementation of which is not relevant here, conditions the illumination radiation from the light source 10. The entire microscope is controlled for localization microscopy by a control unit 16, which can, for example, be a computer. In localization microscopy, the sample 2 is prepared such that isolated emitters emit sample light.The term "isolation" refers to the fact that there are two emitters in the sample which can be distinguished in the image plane 9 with the optical resolution ultimately determined by objective 4 and tube lens 6. Regarding localization microscopy, reference is made to the introductory part of the description and the general description of the invention, which is cited above.

[0031] Microscope 1 differs from conventional localization microscopes in that it incorporates an optical manipulation element 18, which in the embodiment of Fig. 1 lies in a pupil plane 20. In the illustrated embodiment, it is the rear pupil plane of the objective lens 4. However, this is not mandatory. If necessary, the pupil plane 20 could also be generated by intermediate imaging optics. It would then be a pupil plane conjugate to the rear pupil plane of the objective lens 4. It can also lie only in the imaging beam path.

[0032] The manipulation element 18 influences the point image blur function, according to which the sample 2 is imaged onto the image plane 9, such that a point-emitter image of a point-like emitter is anisotropically distorted. This distortion results in the point-emitter image being rotationally asymmetric, i.e., generally having an axis of anisotropy and, for example, image lobes that are rotationally asymmetric. To realize at least two color channels in addition to depth resolution, i.e., to distinguish at least two spectral ranges for the sample light, the manipulation element is composed of two parts: anisotropy elements 22 and 24. These are arranged one behind the other in the imaging direction, i.e., from the objective 4 to the detector 8. Fig. 1. The thickness of these anisotropy elements 22, 24 is greatly exaggerated. In fact, they lie de facto in the pupil plane 20. Each anisotropy element produces the anisotropic image distortion. However, each anisotropy element 22, 24 has a neutral wavelength at which it does not produce image distortion. The neutral wavelengths of the anisotropy elements 22, 24 are different. Furthermore, each anisotropy element 22, 24 has an individual anisotropy axis. The anisotropy axes of the anisotropy elements 22, 24 are at an angle to each other, i.e., they are not parallel. The effect of these two anisotropy elements is determined by the Fig. 4A to 4D will be explained later.

[0033] One possible structure is in the Fig. 2 and Fig. Figure 3 shows an exemplary representation of the anisotropy element 22. The anisotropy element 22 here consists of two glass wedges 26, 28, which in the embodiment of Fig. The glass wedges are designed to form parallel end faces of the anisotropy element 22. The glass wedges differ in terms of the type of glass used. They are selected such that the refractive indices of the glasses are the same at the neutral wavelength and different at other wavelengths. In this way, such a glass wedge has a phase-influencing effect at every wavelength outside the neutral wavelength, resulting in the desired PSF distortion for depth encoding and generating the rotationally asymmetric, i.e., anisotropic, distortion in the point-emitter image. If two glass wedges are used, inverted with respect to their wedge directions and located in the halves of the pupil 20, a point-emitter image consisting of two image lobes is obtained for all wavelengths outside the neutral wavelength. If the emitter is located in the focal plane 3, the image lobes lie on the same axis. They are offset perpendicular to this axis if the point emitter is located above or below the focal plane 3.The offset is unambiguous in that it allows a distinction to be made as to whether an emitter lies below or above the focal plane. The direction of the axis in the image is unambiguously defined by the direction along which the slopes of the wedges are oriented.

[0034] The anisotropy element 24 is similar to the anisotropy element 22 with two differences. Firstly, the glasses are different and selected such that the neutral wavelength differs. Secondly, the wedge direction of the glass wedge pair 22, and thus the axis of anisotropy, is rotated relative to the wedge direction of the glass wedge pair 24, preferably by 90°. Due to adjustment, this value can vary between 85° and 105°. The wedge direction is understood to be the direction along which the wedge surfaces rise or fall.

[0035] Fig. Figure 2 shows the case of a plane-parallel double wedge, here purely as an example for the anisotropy element 22. Except for the neutral wavelength (and of course the arrangement in the beam path), the anisotropy element 24 corresponds to element 22. Fig. In section 3, element 22 / 24 is modified such that the entrance and exit surfaces of the wedge pair are not parallel to each other. Here, the neutral wavelength is not obtained at the value where the refractive indices of the glass wedges 26 and 28 are equal, but at a different wavelength, which is precisely such that the difference in refractive index between the glass wedges 26 and 28 ensures that such sample light is only displaced, but not propagated in a different direction. The approach of Fig. 3 offers greater freedom in the choice of wedge angles and glass materials, i.e., refractive indices.

[0036] The Fig. Figures 4A to 4D show the effect of the two anisotropy elements 22, 24, here in the form of the glass wedge pair according to Fig. 2. Fig. Figure 4A shows the two pairs of glass wedges in cross-sectional view, indicating that the neutral wavelengths differ. For glass wedge pair 22, λ1 is the neutral wavelength. Here, the refractive indices are the same, and the sample light is not deflected in a different direction. For glass wedge pair 24, the wavelength λ2 is neutral.

[0037] Fig. Figure 4B shows how the glass wedges 22 and 24 are positioned in a top view relative to the pupil plane. The anisotropy element 22 is shown in the upper view, and the anisotropy element 24 in the lower view. Fig. Figure 4B shows that two pairs of glass wedges are placed in the halves of the pupil, with the wedge direction inverted relative to each other in the pupil halves. Both glass wedges extend along an axis 23, which is the anisotropy axis. The anisotropy element 24, on the other hand, is formed by the two pairs of glass wedges, which are arranged in the same way, but with their anisotropy axis 25 rotated by 90°. Fig. Figure 4C shows how the anisotropy elements 22, 24 with crossed anisotropy axes 23, 25 are ultimately connected one after the other in the direction of the image.

[0038] Fig. Figure 4D shows the resulting effect, with a left-hand representation for wavelength λ1 and a right-hand representation for wavelength λ2. At λ1, only the anisotropy element 24 is active, since element 22 is neutral at this wavelength. At λ2, the opposite is true. The "pure" form of the distortion consists of the point emitter image being split into two image lobes 30, 32 and 36, 38, respectively, which shift relative to each other along structure axes 34 and 40, respectively, if one imagines moving a point emitter along the depth direction, i.e., the z-axis. In reality, of course, no shift occurs, as the point emitters are fixed in position. Furthermore, the pure forms only result when the sample light from the point emitter is emitted at the corresponding wavelengths λ1 or λ2. At wavelength λ1, the structure axis 34 is therefore determined by the orientation of the anisotropy axis 25.At wavelength λ2, however, the anisotropy element 24 is neutral and only the anisotropy element 22 is effective - with the effect that the structure axis 40 is defined by the anisotropy axis 23.

[0039] In this way, not only the depth, which is encoded by the offset along the structural axis 34 or 38, but also the wavelength, which is encoded by the direction of the structural axis 34, 38, can be decoded from the rotationally asymmetrically distorted point-emitter image, here consisting of the image lobes 30 and 32. This is effected by the control unit 16.

[0040] Fig. Figure 5 shows an example of how the glasses can be selected. Dispersion curves for three different glasses, labeled in the legend, are shown. It is easy to see that it is simple to find glass pair combinations that have the same refractive index at a specific wavelength (i.e., can be combined neutrally), but differ at another wavelength. One possible combination would be, for example, the glass pair OHA-S BAL 11 and OHA-S BAL 14. They have the same refractive index at a wavelength of approximately 500 nm, but a significantly different refractive index at a wavelength of approximately 610 nm. This could, for example, be used in the anisotropy element 22. A suitable configuration for the anisotropy element 24 would be the glass pair OHA-S BAL 14 and OHA-PDL 26Y. This glass pair would have the same refractive index at a wavelength of 610 nm, but a different refractive index at approximately 500 nm.This would mean that the neutral wavelengths of the wedge pairs would be different with this choice of glass, and the distortion could be exactly as described. Fig. 4D can be achieved, with wavelengths λ1 and λ2 then being approximately 500 nm and 610 nm.

[0041] The Fig. Images 6A to 6J show distortions that appear as two image flaps 30, 32 for different situations. In the Fig. 6D, Fig. 6E and Fig. In 6F, the point emitters are located in focal plane 3, in the Fig. 6A, Fig. 6B and Fig. 6C above, in the Fig. 6G, Fig. 6H and Fig. 6 years below. In the Fig. 6A, Fig. 6D and Fig. 6G emits sample light exclusively at wavelength λ1. In the Fig. 6C, Fig. 6F and Fig. 6J, however, only at λ2. In the Fig. 6B, Fig. 6E and Fig. 6H the wavelength of the sample light emitted by the point emitter lies between λ1 and λ2.

[0042] As can be seen, the distortion unfolds, i.e., the image flaps 30, 32 are oriented in Fig. 6D and Fig. 6F exclusively along the structural axes 34, 38, which are defined by the anisotropy axes 23, 25. In the Fig. 6A, Fig. 6C, Fig. 6G and Fig. In image 6J, the image lobes are shifted orthogonally to these structural axes 34 and 38. Thus, the color channel can be encoded from the direction of the structural axes 34 and 38, and the depth can be encoded from the offset relative to these axes. The same applies to a point emitter that has neither the pure wavelength λ1 nor λ2, resulting in the point emitter images 6B, 6E, and 6H. These can be distinguished from the "pure" forms and thus classified with respect to color channel and depth.

[0043] In Fig. In Figure 1, the anisotropy element is arranged as a phase-manipulating element in a pupil plane. However, there are other anisotropic distortions that can also be exploited. Fig. Figure 7 shows an embodiment of a microscope 1 in which two pairs of anamorphic lenses 42 and 44 are used as optical manipulation elements 18. They can be positioned before or after the tube lens 6 or combined with it. Their design and function are described in the Fig. Figures 8A to 8C show this. As can be seen, the lens pairs 42 and 44 do not appear anamorphic for different neutral wavelengths λ1 and λ2, respectively. Furthermore, when the anamorphic lens pairs 42 and 44 are connected in series, their cylindrical axis is rotated 45° relative to each other. The cylindrical axis thus corresponds to the anisotropy axis 23 or 25. Fig. Figure 8C shows how the distortions differ. The left column of the Fig. Figure 8C shows the distortion of the point center at wavelength λ1 for different depths. In the upper half, the point center lies above, and in the lower half below, the focal plane. The right half shows the distortion at wavelength λ2.

[0044] The evaluation of the recorded dot emitter images takes place in control unit 16.

[0045] In the general case of two prisms 26, 28, which form a straight-vision prism 22, 24 according to Fig. 2 or Fig. When combining the three elements, for small wedge angles ε, the total deflection δ = (n1-1)*ε1 + (n2-1)*ε2 = 0. This results in the ratio of the prism wedge angles ε2 / ε1 = -(n1-1) / (n2-1) for any refractive indices n1 and n2 > 1 at the neutral wavelength, yielding a total wedge angle ε1 + ε2 = ε1*(1 - (n1-1) / (n2-1)). In the special case of equal refractive indices n1 = n2 at the neutral wavelength, ε2 / ε1 = -1, and thus the "planar plate" ε1 + ε2 = 0.

[0046] On the other hand, with different dispersions of the two glasses 26, 28, which can be described by different Abbe numbers v1 and v2, a deflection angle δn=ε1*(n) can be determined for the non-neutral wavelength. 1λ -1-(n 2λ -1)*(n1-1) / (n2-1)) are set, which determine the wedge angle(s) ε. Here, n denotes 1λ and n 2λ the refractive indices of the two materials at the deflection wavelength.

[0047] In this way, cemented prisms can be designed for any media, which transmit at a neutral wavelength and generate a specific target deflection at another wavelength.

[0048] It is also possible to arrange such prisms with different, in particular interchanged, neutral wavelengths in the z-direction.

[0049] How Fig. 6B, Fig. 6E and Fig. As shown in 6H, a third color channel, which lies between the two design wavelengths λ1, λ2, becomes usable through the “mixing” of the phase effect of the two orthogonally aligned anisotropy elements 22, 24.

[0050] The principle can also be applied to three design wavelengths by sequentially arranging three anisotropy elements with different neutral wavelengths λ1, λ2, λ3. It should be noted that the resulting beam deflections result from the overall combination of the remaining anisotropy elements. Analogous to the above procedure, at least three anisotropy elements made of different materials can be combined, for which the following conditions apply: δ(λ1)=0=(n1(λ1)−1)*ε1+(n2(λ1)−1)*ε2+(n3(λ1)−1)*ε3 δ(λ2)=0=(n1(λ2)−1)*ε1+(n2(λ2)−1)*ε2+(n3(λ2)−1)*ε3 δ(λ3)=δ=(n1(λ3)−1)*ε1+(n2(λ3)−1)*ε2+(n3(λ3)−1)*ε3 where δ denotes the splitting at the effective wavelength and λ1 and λ2 denote the neutral wavelengths. To achieve a desired effect from the splitting, the three prism angles for suitable media must then be determined.

[0051] Based on this model, for example, a neutral effect at 546 nm and 643 nm can be achieved with a SYGH51 wedge angle of -0.328°, an S-FTM16 wedge angle of +33.158°, and an S-NBH51 wedge angle of -25.891°, while the wavelength at 480 nm is deflected by 0.2'. A neutral effect at 480 nm and 643 nm can be achieved with the same 0.2' splitting of the 546 nm wavelength using an S-YGH51 wedge of -0.303°, an S-FTM16 wedge of -72.384°, and an S-NBH51 wedge of 57.548°. Similarly, the 0.2' splitting of 643 nm at neutral wavelengths of 546 nm and 480 nm is conceivable with an S-YGH51 wedge of 0.644°, an S-FTM16 wedge of 39.269°, and an S-NBH51 wedge of -31.701°. These abbreviations denote common glass types from the manufacturer Ohara.

[0052] Furthermore, the azimuthal rotation of the prism units must be reduced from 90° to 60° and taken into account in the PSF calculation.

[0053] Similar to how the inclined planes and spot displacements of the wedge pairs are interconnected, these distortions can also be achieved radially, leading to the fabrication of a single element and technologically eliminating the need to combine different wedge arrangements in the (divided) pupil. This can initially be understood analogously to Fig. 2. This leads to a monochromatic spiral mask on a substrate. The combination of two such structures on different materials, cemented together on a common plane, can then produce a similar neutral wavelength in the overall effect as the partial prisms cemented together to form a straight-view prism. The sequential arrangement of such units with different neutral wavelengths then allows for wavelength-separated effects of the phase structures. In principle, a wide variety of chromatic phase masks for different applications can be realized in this way.

[0054] The 3D localization of the isolated emitters is based on experimentally determined or simulated PSFs. In principle, algorithms that evaluate specific PSF parameters can also be used (see S. Pavani et al., “Three-dimensional, single-molecule fluorescence imaging beyond the diffraction limit by using a double-helix point spread function”, Proceedings of the National Academy of Sciences, 106.9 (2009): 2995-2999). Using the previously measured 3D PSF has the advantage that system-induced aberrations are also taken into account and that the PSF can still be used in the more defocused areas, where individual parameters can no longer be clearly distinguished between the different color PSFs.

[0055] The 3D localization of the isolated emitters can be advantageously achieved by fitting an experimentally determined or simulated 3D-PSF, typically using a maximum likelihood estimation algorithm as published in M. Juette, et al., Nature Methods, Vol. 5, No. 6, p. 527, June 2008, and the accompanying supplementary material, or in Li, Yiming, et al. “Fast, robust and precise 3D localization for arbitrary point spread functions.” bioRxiv (2017): 172643.

[0056] The method can optionally also be used for 2D imaging, e.g., under TIRF illumination or with very thin samples. In this case, only the color coding of the PSF shape at focus is utilized (see [reference]). Fig. 6, middle line), this is even more clear and unambiguous than in the 3D case.

Claims

[1] Localization microscope comprising an imaging device for imaging a sample light-emitting sample (2) from a focal plane (3) into an image plane (9), - wherein the imaging device has an optical manipulation device (18) which is designed to influence a point image blurring function of the imaging in a depth-dependent manner and influences the point image blurring function of the imaging such that a point emitter in the image plane is mapped into a point emitter image (30, 32) which is rotationally asymmetrically distorted, - wherein a shape and / or orientation of the rotationally asymmetric distortion of the point emitter image (30, 32) depends on the position of the point emitter relative to the focal plane (3) and a wavelength of the sample light and - wherein the optical manipulation device (18) has a first and a second anisotropy element (22, 24) which each anisotropically influence the point image blurring function of the imaging in order to generate the rotational asymmetry of the point emitter image (30, 32), and each have an anisotropy axis (23, 25), characterized by , that - the two anisotropy elements (22, 24) are arranged one behind the other in the direction of the image and their anisotropy axes (23, 25) are at an angle to each other - and both anisotropy elements (22, 24) each have a neutral wavelength (λ1, λ2) at which they do not anisotropically influence the point image blurring function of the imaging, - where the neutral wavelengths (λ1, λ2) are different. [2] Localization microscope according to claim 1, characterized by, that the anisotropy elements are each designed as a phase element which has a phase-influencing effect at the neutral wavelength of the other anisotropy element (22, 24), and not at the neutral wavelength. [3] Localization microscope according to claim 2, characterized by , that each phase element is formed as a wedge pair (26, 28) of two glasses with different dispersion properties, wherein refractive indices and wedge angles are chosen such that the wedge angle pair is straight-vision at the neutral wavelength, otherwise not. [4] Localization microscope according to one of claims 1 to 3, characterized by , that the anisotropy axes (23, 25) lie at an angle between 80 and 100 degrees, preferably 90 degrees, to each other. [5] Localization microscope according to claim 1, characterized by , that the anisotropy elements (22, 24) are each designed as an anamorphic lens which has no cylindrical effect at the neutral wavelength. [6] Localization microscope according to claim 5, characterized by , that the anisotropy axes (23, 25) lie at an angle between 35 and 55 degrees, preferably 45 degrees, to each other. [7] Method for localization microscopy, wherein a sample light-emitting sample (2) is imaged from a focal plane (3) into an image plane (9), - wherein a point image blurring function of the imaging is influenced depth-dependently by means of an optical manipulation device (18) such that a point emitter in the image plane is mapped into a point emitter image (30, 32) that is rotationally asymmetrically distorted, - wherein a shape and / or orientation of the rotationally asymmetric distortion of the point emitter image (30, 32; 42) depends on the position of the point emitter relative to the focal plane (3) and a wavelength of the sample light and - wherein in the optical manipulation device (18) a first and a second anisotropy element (22, 24) are used, which each anisotropically influence the point image blurring function of the imaging in order to generate the rotational asymmetry of the point emitter image (30, 32), and each have an anisotropy axis (23, 25), characterized by , that - the two anisotropy elements (22, 24) are arranged one behind the other in the direction of the image and with their anisotropy axes (23, 25) at an angle to each other and - both anisotropy elements (22, 24) each have a neutral wavelength (λ1, λ2) at which they do not anisotropically affect the point image blurring function of the imaging, - where the neutral wavelengths (λ1, λ2) are different. [8] Method according to claim 7, characterized by, that for each anisotropy element (22, 24) a phase element is used which, at the neutral wavelength of the other anisotropy element, the point image blurring function has a phase-influencing effect, and not at the neutral wavelength. [9] Method according to claim 8, characterized by , that for each phase element a wedge pair (26, 28) made of two glasses with different dispersion properties is used, wherein refractive indices and wedge angles are chosen such that the wedge angle pair is straight-vision at the neutral wavelength, otherwise not. [10] Method according to claim 8 or 9, characterized by , that the anisotropy axes (23, 25) are arranged at an angle between 80 and 100 degrees, preferably 90 degrees, to each other. [11] Method according to claim 7, characterized by , that for each anisotropy element (22, 24) an anamorphic lens is used which has no cylindrical effect at the neutral wavelength. [12] Method according to claim 11, characterized by , that the anisotropy axes (23, 25) are arranged at an angle between 35 and 55 degrees, preferably 45 degrees, to each other.

Images