Method and device for interferometric microscopy

Abstract

An interferometric microscope (200, 300) comprising: a sample holder (207) configured to hold a sample; a light source (201) arranged to illuminate the sample in the sample holder; an optical detector (210) arranged to capture images of light scattered by a feature of the sample; an optical system comprising a plurality of optical elements (202, 203, 204, 206) arranged to direct an illumination beam of light from the light source to the sample holder and to collect light reflected by the sample and by the sample holder to form an image beam directed to the detector; an optical component (212, 313) arranged in the path of the image beam and configured to divide the light in the image beam into a plurality of beam portions, wherein the beam portions are propagating along a common optical beam path, wherein the beam portions interfere at the detector; and a processing unit (211) configured to determine, from a plurality of images captured by the detector, a parameter indicative of the absolute value of the optical field of the light scattered by the feature

Description

[0001] METHOD AND DEVICE FOR INTERFEROMETRIC MICROSCOPY

[0002] Field of invention

[0003] The invention relates to a method and a device for characterizing submicron- and nanoparticles by the use of interferometric microscopy. In particular, the invention relates to a combination of interferometric scattering microscopy with holographic methods.

[0004] Background of the invention

[0005] The invention relates to a method and a device for characterizing submicron- and nanoparticles by the use of interferometric microscopy.

[0006] Interferometric scattering microscopy (iSCAT) is a method where light scattered by the particle interferes with unscattered light which has travelled the same optical path from the sample to the detector. The most common implementation of the method utilizes light back-reflected by the interface between the sample holder (which often is a simple glass cover slip) and the liquid medium, as reference for the scattered light to interfere with. Instead of a coverslip, a chamber can be used, which can be in the shape of a micro- or nanofluidic channel. The method can quantify the scattered light from a particle, but does not quantify the full optical field. It is also possible to implement the method by directing light from a different angle than from the direction of the detector (150-180 degrees) to interact with the sample and sample holder.

[0007] Summary

[0008] It is an object of the present invention to provide improvements in Interferometric scattering microscopy.

[0009] According to first and second aspects of the invention, it is therefore provided an Interferometric scattering microscope and a method for characterizing a feature or particle in an Interferometric scattering microscope according to the independent claims.

[0010] Interferometric scattering microscopy includes the use of a light source for illuminating the liquid sample. The coherence length of light is the propagation distance over which the coherence of the light wave significantly decays, resulting in 1 / e less interference when a part of the light is combined with another part of the same original light beam where the propagation difference between the beams is the coherence length. The coherence length depends on the bandwidth in frequency or wavelength of the light, the greater the bandwidth over which the light temporally varies, the shorter the coherence length. The coherence length can therefore be estimated based on the bandwidth of a light source. Different types of laser light sources can have coherence lengths with magnitudes from micrometers to kilometers. Laser light sources have the advantage of high power concentrated on a small surface, but also less coherent light sources such as LED can be used for iSCAT imaging. Lower coherence decreases the problem of speckly background in the image due to coherent reflections, but also limits the available focus depth.

[0011] The method also includes the use of a detector such as a camera arranged to record images of light backscattered by particles in the sample. Backscattering is meant to include light which is scattered by the particle at an angle of 90 - 180 degrees relative to the optical axis.

[0012] When light is emitted from the light source in a direction towards a sample in the sample holder, there will be reflected light from one or several interfaces in close vicinity to the sample. Most commonly the lower or first (meaning closest to the microscope objective) interface between the sample holder and the liquid sample provides a first reflection due to difference of refractive index between the floor and the sample. In the case of a chamber there may also be reflected light from the upper or second interface. Reflected light from these interfaces will be reflected back to the sensor (camera) and interfere with light backscattered by the optical features in the sample. The reflected light can be used to interfere with light scattered by the optical feature at the detector. Variations can be made in whether reflected light from one or several interfaces is used, whether these reflections are in phase or not, and also other interfaces in close proximity of the sample may be used.

[0013] Nanoparticles are typically defined as particles that range in size from approximately 1 to 100 nanometers in diameter while submicron particles typically are defined as particles ranging from 10 nanometers to 1000 nanometers, i.e. up to 1 micrometer. Hence, the size of the particles to be characterized are in the size range from 1 nanometer to 1000 nanometers. In particular, the method is suitably used for particles being in the range from 20 to 500 nanometers. The term optically determined diameter / radius / size is used to describe a method which use the properties of detected light from one or several captured images to determine the size of a particle, in contrast to methods which make use of the Brownian motion of a particle to determine its size. The term hydrodynamic diameter / radius / size is used to describe a diameter / radius / size from a size estimation which is based on measuring the Brownian motion of a particle.

[0014] Concerning the liquid medium, particles dispersed in water are most commonly analyzed, but particles dispersed in a wide variety of different liquids can be analyzed. The method may for example be used for characterizing particles within a cell wherein the cytosol is the liquid in which different particles of interest may be studied. Cytosol is the gelatinous liquid that fills the inside of a cell which is composed of water, salts, and various organic molecules. The cytoplasm comprises, in addition to cytosol, organelles and other particles which in many cases are of interest to be studied. In addition to analyzing particles, embodiments of the invention may be utilized to generate three-dimensional images and image sequences of larger complex objects such as cells or tissue samples which comprise different types of optical features. By optical feature is broadly meant physical structures or inhomogeneities which give an optical contrast in intensity and I or phase. An optical feature could be the extended structures like outer membrane of a cell or actin filaments inside, it can also be the ensemble of many internal inhomogeneities and density gradients of a cell.

[0015] By particles is meant to include a wide variety of different substances and the term is meant to include clusters or agglomerates comprised of same or different molecules or subparts to form the particle. Gaseous bubbles can also be considered as particles in a liquid media. In a broad interpretation of the term particle, any concentrated substance comprised in the liquid having a Rl differing from the Rl of the liquid can be considered to be a particle, e.g. oil droplets in an aqueous solution. However, in most cases the particles consist of or at least comprise a rigid supporting structure.

[0016] The scattering cross section of a particle is a measure of the probability of light to be scattered by it. It is defined as the effective area presented by the particle that interacts with the incoming light wave. The scattering cross section is usually denoted by the symbol o and has units of area. The scattering cross section depends on various factors, such as the energy of the incoming photons and the mass and mass distribution of the particles. The scattering cross section can be calculated using the following formula: where k = 2 * IT / A , A is the wavelength of light, a is the polarizability of the particle, oo is the scattering cross section at 0 degrees, and need to be adjusted using the form factor to get the scattering cross section for other scattering angles. The polarizability if approximately proportional to the volume V of the particle and the difference between its Rl and the Rl of the surrounding medium, An. The polarizability is defined as: where V= particle volume, np=refractive index of the particle, nm=refractive index of the liquid medium. This expression is often approximately similar to 7 An, where An = np— nm.

[0017] In the context of light scattering, the form factor is a measure of how the scattering amplitude, which describes the interaction between the incident light and the particle, varies with the momentum transfer between the light and the particle. The form factor is related to the internal structure of the particle and provides information about the distribution of mass within the particle. The scattering cross section for a specific scattering angle, o, can be used to calculate form factor f. The scattering cross section at a specific angle is defined as the product of the scattering cross section oo and the square of the form factor f. It is given by the formula: o = Oo * If I2

[0018] In the following, the term scattering cross section is used with the meaning of scattering cross section for the specific scattering angle (angle between incoming and scattered light). More specifically, it should be mentioned that when using an image captured by a microscope objective, it is light scattered at a range of scattering angles which is quantified and used to determine a scattering cross section. The determined scattering cross section is therefore an average for a range of scattering angles.

[0019] The intensity of the scattered light in iSCAT consists of two different terms, one which is proportional to the square root of the backscattering cross section and one which is proportional to the backscattering cross section. The proportional term is due to the scattered light directly and the square root term is due to interference. For small nanoparticles, it is common to neglect the proportional term as it is much smaller than the interference term. However, in the size range 0.1-1 micron, both terms can be significant contributors to the measured intensity / contrast. For small enough particles that only the interference contribution is significant, the image brightness contrast of the particle is proportional to o , which for a non-absorbing (dielectric) particle is given by where V= particle volume, An = refractive index difference and f = form factor. The last factor is due to the interaction of light scattered by the particle and light scattered by the glass surface, which serves as reference. Az is the distance between the particle and the glass surface, is the wavelength.

[0020] In the present invention an optical component is introduced in the path of the light beam traveling from the sample and sample holder to the detector, in order to make different portions of the reference light interfere with each other at the camera sensor.

[0021] The interferometric microscope comprises: a sample holder configured to hold a sample; a light source arranged to illuminate the sample in the sample holder; an optical detector arranged to capture images of light scattered by a feature of the sample; an optical system comprising a plurality of optical elements arranged to direct an illumination beam of light from the light source to the sample holder and to collect light reflected by the sample and by the sample holder to form an image beam directed to the detector; an optical component arranged in the path of the image beam and configured to divide the light in the image beam into a plurality of beam portions, wherein the beam portions are propagating along a common optical beam path, wherein the beam portions interfere at the detector; and a processing unit configured to determine, from a plurality of images captured by the detector, a parameter indicative of the absolute value of the optical field of the light scattered by the feature.

[0022] One important purpose of the invention is to provide a means to digitally refocus particles in a volume of liquid. Small particles outside of the optical focus will appear in a microscope image as a large pattern of concentric circles (diffraction rings) with a more extended central spot of lower contrast, the further from focus, the larger and lower contrast. Close to focus the central spot of the ring pattern is smaller and has higher contrast. To digitally refocus means to digitally propagate the image or optical field so that it looks like it is in the actual optical focus, in the sense that a maximal contrast and minimal extension is achieved. Refocusing can be made to any distance to the optical focus plane, but if nothing else is stated, digitally refocused means that the image or optical field is propagated to attain maximal contrast and minimal extension.

[0023] This may include fine-tuning the focus within a wavelength of light, but more importantly to be able to analyze particles over a depth range of up to 1 , 5, 10, 20 or even up to 50 or 100 micrometers. There are several good reasons to analyze particles suspended in a relatively deep liquid volume. Analyzing particles in the bulk of a liquid enables analyzing a dispersion in its native and unperturbed state and also to resolve rapid changes in the dispersion such as aggregation or growth of particles. Some prior art teaches to adsorb or bind particles to a flat surface before analyzing them, or confining them in a narrow volume of liquid in a very thin channel. An advantage of such approaches is to be able to optically focus the particle very precisely, and to unambiguously determine its axial position (distance from optical focus along the optical axis) if the particle is confined to within one wavelength of light in the axial direction.

[0024] Furthermore, in the case of adsorption also it becomes possible to image the particle at the same position for an extended time and thereby get a clearer image by averaging many images of the same particle. However, it can be challenging to make particles adsorb or bind to a surface and to make sure that the particles which bind are representative of the particles residing in the bulk of the liquid. Furthermore, once the particles are at the surface there is risk that they interact with each other or the surface and change their composition and properties. Furthermore, it takes time for particles to diffuse to the surface to adsorb or bind, making the analysis slower. A further example of prior art teaches to adsorb particles on the surface of the sample holder or the coverslip and subsequently move the stage a few micrometers along the optical axis to physically sweep the particle in and out of the optical focus while capturing multiple images. This enables to precisely determine the distance of the particles center to the surface and thereby indirectly its radius. It can however take time to perform a sweep and the method places special demands on the sample stage and its mechanics. Furthermore, such methods place strong requirements on controlling the cleanliness and properties of the surface, adding substantial technical challenges. Similarly, for very narrow channels relative to the size of the particle to be analyzed, there is a risk that the particles which enter the channel are not representative of the particles in the complete sample, the channel itself effectively functioning as a filter which may not necessarily have a very sharp cutoff, instead giving a size bias over an extended size range. Furthermore, in a narrow channel it is also very important to control the surface properties, to avoid adsorption of the particles. Many such surface treatments have limited shelf life and need to be made soon before the analysis, adding substantial labour and logistics for the end user. Furthermore, the narrower the channel, the greater the risk of clogging of the channel. Furthermore, there is risk of the analysis taking long time since very little sample volume can pass through a very narrow channel.

[0025] It is therefore preferable that the sample holder comprise a chamber with an internal floor and a ceiling, where the illumination beam illuminates the sample through the floor. The floor thus needs to be transparent for the illumination light, but this is not necessary for the ceiling / roof of the chamber. The internal height between the floor and the ceiling is at least 1 micrometers, more preferably at least 5 micrometers and most preferably more than 10 micrometers. In other cases, it may be preferable to use even higher sample chambers, for example if imaging or analyzing very large particles or structures such as cells. In such case the internal height between the floor and the ceiling is at least 20 micrometers, more preferably at least 50 micrometers. With very high chambers it may be preferable to have only a reflection from either the floor or the ceiling, for example achieved by providing either surface with anti-reflection coating.

[0026] Using a sample chamber with a depth of at least 5 micrometers decreases problems with adsorption of particles to the chamber surfaces, and if the chamber is in the form of a channel, decreases risk of clogging of the channel and obstructing the flow. It also enables imaging a larger volume of liquid simultaneously, enabling analyzing lower particle concentrations. Refocusing of particles far from focus increases their contrast and thereby improve their automatic detection. Furthermore, by detecting all particles between the ceiling and floor of the sample chamber or channel, the concentration of particles can be determined in a volume with well- defined boundaries. Another advantage is that if the particles are not detected in the entire depth of the chamber, if their 3D position is known, the imaged volume is known, and the particle concentration can still be reliably determined.

[0027] In prior art, determining the axial position of moving particles in iSCAT microscopes is often troublesome. Even if the distance to optical focus can be estimated, it can be difficult to determine on which side of the optical focus a particle is located. This comes from the periodicity of the optical signal, where the relative phase between the particle and the background repeat itself when the particle moves approximately half a wavelength in depth. Due to the rapidly shifting phase of the particle optical field, it is difficult to interpret the diffraction ring patterns in the image correctly. Sudden (unrealistic) jumps in detected axial position along a particle trace is a commonplace problem. Imaging at extremely high frame rates, such that several images of the particle are captured between every time it switches phase, can facilitate determination of axial position in regular iSCAT images. In this case, a multitude of images are used for this purpose. However, for small particles undergoing Brownian motion this requires many thousand image frames per second and such high imaging speeds requires using very few camera pixels. In contrast, the present invention enables imaging at 100-1000 frames per second and determining axial positions for each particle image individually. This enables imaging a larger area and therefore many particles or optical features simultaneously.

[0028] A further purpose of the described microscope is to provide a design with low sensitivity to vibrations. One way to provide interference and digital focusability could be by dividing the image beam into two separate beams and reunite them only at the camera sensor where they would interfere due to a small angle between the two beams. This would be a set-up similar to off-axis digital holographic microscopy. Such set-ups do however suffer from vibrations in the image which makes detection and analysis of particles more difficult. This is since vibrations may affect the two beams differently and when they interfere the effect of this difference is amplified. Furthermore, when using two separate beams it is difficult to use light sources of low coherence length since the length of the two beams need to be similar to within one coherence length. This forces such optical arrangements to use laser light sources of longer coherence length, which further deteriorates the image quality due to coherent scattering and reflections. For example, a diode laser can have coherence length substantially lower than 1 mm, which gives substantially less background speckle due to coherent scattering and reflections, than using laser of higher coherence length. Other possible light sources of low coherence length include supercontinuum lasers, light emitting diodes (LEDs) and more.

[0029] The microscope therefore comprises an optical component in the image beam which divides the light reflected by the sample holder in two or in many portions or sub-beams, to cause them to interfere with each other at the camera. These different portions follow a common optical beam path, which means that they pass the same lenses, mirrors and other optical components. The greatest extension of the complete image beam, comprising all light portions, perpendicularly to the optical axis, is therefore not more than 50mm, preferably less than 25 mm. Interference at the camera sensor is achieved if two or more light portions are directed to the same area on the camera sensor at slightly different angles. There are multiple types of optical components which can divide and guide the light in the desirable way. To divide the image beam in two portions, a beam splitter may be used. This type of component typically is a cube which consists of two cemented triangular prisms, but can also consist of a flat or wedge-shaped plate, with a special surface coating. Beam splitters typically generate two beams with 90 degrees angle to each other and are therefore somewhat difficult to use in a common-path arrangement. Another option is to use shear plate, which may consist of glass and is typically slightly wedge-shaped. A thin glass plate can be used. Light is guided at typically around 45 degrees angle to the plate, and light is reflected from both the front and back surface of the plate and thereby form two light portions. However, in such case most of the light is lost since only a small fraction of light is reflected from a clean glass surface in air. Suitable surface coatings to modify the reflectivity of each surface can used to decrease the light losses, but may give additional unwanted reflections.

[0030] Another component which can be used is a beam displacer, which can be made from one single block of transparent material, and which split light into two beams with orthogonal polarization and parallel propagation. Also 1 -dimensional diffraction gratings can be used to split light into two or a small number of portions, consisting of the 0thand 1stdiffraction orders and possibly more. Diffraction gratings are available in different types which can be used in transmission or in reflection. Also, digital spatial light modulators can function as diffraction gratings.

[0031] For dividing light into a large number of portions, (for example more than 100), several component types can be used; diffraction gratings (which can be transmissive or reflective), microlens arrays (including Shack-Hartmann wavefront sensors), or thin diffusers. A novel type of optical component is meta surfaces, which can mimic several types of optical components, including diffraction gratings, and can be used for dividing light into different portions.

[0032] When dividing the light into a large number of portions (typically more than 100), it is advantageous to use a component which divides the light in two dimensions, generating a 2-dimensional interference pattern. In one implementation this is achieved by using a grating which comprise a 2-dimensional pattern where different fields provide a different phase shift of the light, these fields may be separated by opaque lines. One type of such a 2D grating pattern is a chessboard pattern on a glass or quartz substrate where half of the squares are etched down to provide a phase shift compared to light having passed the other squares. Light portions having passed different fields interfere with each other at the detector, generating an interference pattern based on four different diffraction orders. The grating is typically placed 0.5-2 mm from the camera sensor, or from an image plane in the optical system. Since the light is divided only over a very small space, such an arrangement is very resistant against the effect of vibrations and can be used with light of very short coherence length.

[0033] When adding an optical component to generate interference, some of the light in the image beam is usually lost by being blocked or diffracted. Furthermore, following fourier transformation of the raw image, most of the image data in the fourier image is discarded before reconstructing phase and intensity contrast images or optical field. This decreases the effective numerical aperture of the microscope and thereby the range of scattering angles accessible. Furthermore, a feature of iSCAT is that most of the illumination light is being lost at the sample and only small fraction of the light is reflected. This enhances the contrast of features and particles in the sample, but also enhances image noise coming from reflections and optical imperfections and which may vary from image to image. This noise is further amplified in the optical field reconstruction making it more challenging to utilize holographic approaches for iSCAT imaging than for transmitted light. Despite these shortcomings, it has been found that there are significant advantages in that particles and features can be reliably refocused and optical data for a feature or particle can be statistically averaged to substantially improve signal-to-noise ratio. In following is described a suitable computational pipeline for analysis of moving particles. For image analysis for more stationary and complex objects such as live cells, it may be advantageous to do the image analysis slightly differently. By dividing the reference light into at least two portions which are directed to the camera sensor at slightly different angles and thus cause interference between them, regular patterns are created in the raw image. When Fourier transforming such an image, three or more circular Fourier peaks will appear in the Fourier image, whereas in a normal image there would be only one central Fourier peak. The peaks which are not in the center may contain information about the phase of light in the image. By separating one or several of these Fourier peaks and transforming them back to real space, intensity as well as phase images may be created and a complex optical field may be generated, comprising both intensity and phase information.

[0034] In an example implementation, a two-dimensional grating is used, comprising a chessboard pattern. Half of the squares are etched down to provide a path difference and thereby a phase shift of the passing light compared to light passing the other squares. Furthermore, optionally there are opaque lines covering the border between the squares. Such a grating and grid-mask will give at least 4 additional Fourier peaks, where said additional peaks comprise a phase gradient in two perpendicular directions. The phase gradient from two of said peaks will then need to be integrated to get an entire phase contrast image. Intensity is extracted from the central peak, and may be integrated with the phase contrast image into an optical field.

[0035] In another implementation, the light is divided into only two sub-beams or portions which interfere at the camera sensor. This gives rise to stripes in the image which when Fourier transformed generates only two additional peaks in addition to the central peak. In this case the side peaks comprise both phase and intensity information and can be directly transformed into an optical field.

[0036] The reconstructed fields or images normally have a background pattern due to dirt, scratches and reflections in the optics and the phase may not be uniform and flat due to imperfections in the optical system. The background pattern obscures the particles and makes analysis difficult, it is therefore customary to apply various algorithms to improve the image quality and to make the particles visible and to have higher contrast than the background noise. There are many such algorithms and methods which can be used to flatten the phase (make it uniform) of the image and to remove the background pattern. The more the particles move, the easier it is to subtract the background pattern without subtracting the particle itself. In particular for larger particles which undergo slower Brownian movement, it is therefore advantageous to let the sample flow through a channel while images are recorded. A background image or field to be subtracted from each image or field can be captured before the sample is placed in the sample holder, or it can be generated from one or multiple other images in the same recording sequence of the same sample and can be dynamically adapted to achieve an optimal background subtraction from each individual image. One method which has been previously described by the inventors is to iterate through a selection of 20-30 previous or later image frames and select those for subtraction which best minimize the background of the present image frame upon subtraction. An average of the selected frames is then subtracted from the present frame. Instead of, or in combination with, subtracting background from the reconstructed field, background removal and flattening of the phase can be made already in the reconstruction stage by multiplying the optical field with a conjugate reference field.

[0037] When using the invention for imaging of larger and more complex as well as more stationary structures, such as cells or tissue samples, some changes compared to the previously described image processing may be suitable. Background subtraction is easier when particles move due to flow and Brownian motion, as the background is more easily separated from the objects to be imaged and analyzed. When imaging cells it may be suitable to instead move the stage a few micrometers back and forth in order to separate out background patterns from reflections and scattering from dirt and imperfections in the optical system. Alternatively, a reference field or image can be captured before imaging the sample, which can be used for background subtraction.

[0038] The next step is to detect the particles in each image, which can be made with classical or machine learning based algorithms. Detection is most often made by identifying small regions which have an image parameter value above a certain threshold. Said parameter could be the absolute value of the optical field, the light intensity or Real part of the optical field, or the phase or imaginary part of the optical field. Detection can be made directly in the image or for example based on a total projection. A total projection can be made by refocusing / propagating the image / field to multiple planes and subsequently projecting the maximum value at each pixel to the total projection image. This method simplifies particle detection since particles have higher contrast and are easier to detect when they are in focus and the total projection will comprise the maximum contrast of each particle. Furthermore, it enables to simultaneously determine the particle’s position along the optical axis by simply finding at which focal plane the particle has the highest contrast which contributes the most to its contrast in the total projection.

[0039] Subsequently the detections a linked together to form a trace for a specific particle. In preparation for further analysis, the 3D position of each particle in each image is saved, together with a cropped image / field (Region of interest, i) of each particle in focus, for each recorded 3D position.

[0040] It is advantageous to filter out incorrect detections at this stage to facilitate the further analysis. Firstly, it is preferable to exclude traces consisting of few positions, for example less than 10 positions, as such short traces often consist of incorrect detections of other features in the image than individual particles. Furthermore, it is preferable to remove ROIs, and their corresponding positions, where the particle is for example not well centered, or not correctly focused, or something else than a particle has been detected by mistake.

[0041] From the 3D positions of a particle along a trace, the Brownian motion of a particle can be analyzed in order to determine its hydrodynamic diameter or radius. The Brownian motion is typically analyzed by determining the position of a particle in several consecutive image frames and based on the distance it has moved between these frames, estimating its diffusivity. The mean square displacement (MSD) can be calculated in different ways, the most common is the following:

[0042] , where N is the number of frames, rtis the position in frame i, and ri+nthe position n frames later. From the mean square displacement (MSD), the diffusivity (D) can be determined according to:

[0043] , where t = time, and d is the dimension. For MSD in one dimension, d=1 , two dimensions, d=2, and three dimensions, d=3.

[0044] From diffusivity, temperature and liquid viscosity, the hydrodynamic diameter of a particle can then be calculated with the help of the Stokes-Einstein relation. The MSD is thus indicative of diffusivity which is indicative of hydrodynamic diameter or radius.

[0045] The MSD can be determined based on the motion in one, two or three dimensions. For particles in flow, the diffusivity is most commonly determined based on movement in one or two dimensions perpendicular to the flow direction. It is however also possible to analyze Brownian motion in the flow direction, although the uncertainty is often worse. Utilizing three dimensions to determine MSD provides better statistics in principle, but only if the position uncertainty is similar in all three dimensions. Often the position along the optical axis is more difficult to determine accurately, therefore often only motion lateral to the optical axis is utilized to determine MSD.

[0046] One advantage of the invention is that it facilitates determination of the axial position of a particle. This can be done in different ways and in one or two steps. One possible method is to propagate (digitally refocusing) the optical field of a particle to many different axial positions and to find at which of these axial positions the particle has the highest contrast. This can advantageously be done when performing a total projection for particle detection. This method is most suitable for a first, less accurate, determination of axial position. Subsequently each particle can be propagated to a larger number of more narrowly spaced axial planes, to fine-tune the axial position. At this stage it is possible to use a different focus criteria than in the first stage, if suitable.

[0047] When the above operations have been performed, the optical properties of the particles can be analyzed in different ways.

[0048] Similarly as described earlier by the inventors for light passing a particle in transmission and generating forward scattering of light, light interacting with a particle by backscattering will also be affected in two ways; phase shift and extinction. For small dielectric particles the phase shift (integrated over the projection area of the particle) is proportional to the polarizability or approximately to Particle Volume multiplied with the Rl difference between particle and surrounding medium.

[0049] If the optical field of the background light (in this case a fraction of the illumination light reflected from the sample holder) is determined, the optical field of the background light is normalized to 1 and subtracted, the optical field of the particle is isolated. If the optical field of a very small particle is expressed as a complex number, the Imaginary part for a small dielectric particle is approximately equal to the phase shift of the interacting light:

[0050] <p(Ep) = arctan( Im(Ep)l ('\ +Re(Ep))) ~ arctan( / m(£p)) « Im(Ep), where is the phase shift, Ep is the optical field of the particle, Im(Ep) is the imaginary part of the field of the particle and Re(Ep) is the real part of the field of the particle. This will be further explained in the detailed description, including some figures explaining the relationships, but the general reasoning is as follows: In a complex number plane depicted with a Real axis and Imaginary axis, the light which has interacted with a particle achieves a phase shift and a smaller absolute signal due to extinction than the light which has not interacted with any particle. If the background light is subtracted, the field of the particle is extracted, which has a negative Real part and a positive Imaginary part.

[0051] However, this reasoning is only valid for a particle imaged in the optical focus of the microscope and at a negligible distance from the surface where the reflection of background / reference light takes place. When the particle is at a distance from the reflecting surface, there is an additional phase difference between light backscattered by the particle and the light reflected by the interface. For a suspended particle undergoing Brownian motion, this phase shift will rapidly vary. The result is that the optical field of the particle will have an Imaginary part and a Real part which are fluctuating between two extreme values and can thus not be easily related to particle properties. The absolute value (also referred to as the amplitude) of the optical field (integrated over the projected area of the particle) will however be consistent and is proportional to the particle’s polarizability and to the square root of the particle’s scattering cross section.

[0052] It is known to determine the intensity contrast of particles in iSCAT images and selecting only a small percentage of the images of each particle, which have the highest contrast (which corresponds to the Real part of the optical field). The present microscope and method instead enables to use the absolute value of the optical field of the particle derived from each image of a particle. Averaging a larger number of images gives less spread (for example quantified as standard deviation) in the determined parameter. Averaging can be made either by determining the integrated absolute value for each particle ROI and then averaging the values, or by generating an averaged particle optical field ROI based on all the individual ROIs and subsequently determine the integrated absolute value from this averaged particle image. In a noisy image the latter method is often preferable since the individual images may be too noisy to make a correct analysis, and also because it is faster to compute. It is thus advantageous to use a plurality of images of the same particle to determine the integrated absolute value of the optical field, and it is advantageous to use a large fraction of the captured images, at least 25% of the captured images and preferably more than 50% of the captured images of a particle to quantify its properties.

[0053] Combined with the hydrodynamic size, the integrated amplitude (absolute value) or therefrom derived scattering cross section can be used to estimate the particle Rl, specifically by dividing the scattering cross section with the cube of the hydrodynamic radius, more specifically particle volume.

[0054] If the Rl of the particle and medium is known, the integrated amplitude or scattering cross section can be used to estimate the particle size. By plotting the amplitude or scattering cross section or any parameter indicative of these against hydrodynamic size, different particle populations can be identified in the same dispersion. The strong ability to differentiate different particle populations in the same sample is a key advantage of the interferometric microscope. Other pairs or groups of parameters may also be used for this purpose. Since hydrodynamic size and scattering cross section are determined as independent parameters, this combination is particularly useful.

[0055] A further possibility of embodiments of the present invention is to determine the particle size directly from its optical scattering. This can be achieved by evaluating how the scattering of light varies with scattering angle. The angle dependence of the scattering amplitude can be extracted by Fourier transforming the particle ROI, provided that the ROI is in the form of a complex optical field. The resulting Fourier image is a complex field where the distance to the center of the image is proportional to sin (0), where 0 = scatter angle. By radially averaging (averaging all pixels on a certain distance from the center) the Fourier image and taking the absolute value of the field along this radius, the scattering amplitude as a function of the scattering angle is extracted. This output needs to be related to what is expected for particles of different sizes and refractive indices. This data can be extracted from optical theory, where Mie theory is the most complete theory. Mie theory is used to make simulations rather than calculations and thus it can be used to build a reference library to which the measurement is compared. A reference library could also in principle build on a large amount of experimental data. The function between scattering amplitude and scattering angle can either be fitted to data in the reference library by classical algorithms, or the reference library can be used to train a machine learning model which then is used to predict particle size from experimental data.

[0056] This output is fitted (by classical algorithm or by Machine Learning) to a theoretical model for light scattering of particles, such as Mie theory. It could also be evaluated by comparing to experimental reference data. It is advantageous to perform this operation on an averaged ROI rather than to analyze each individual ROI of a particle. For such an averaging to be useful, the complex field will need to be adjusted so that all the complex ROIs to be averaged have the same phase. This phase can be arbitrarily chosen, for example the phase of the first ROI in a trace can be used.

[0057] As mentioned, the measured optical field will shift phase depending on the distance to the optical focus and the reflecting surface. In principle, if the optical field at one axial position is known, the phase will be predictable as a function of axial distance to focus and the wavelength of the light. As the particles undergo Brownian motion as well as due to any flow in the sample, the distance to optical focus will vary. This means that the full optical field of a particle can be determined if the axial position can be determined with high accuracy. If the axial position at each detection of a particle is known, the phase at each detection can be compensated and the optical field extrapolated to the optical focus or to any arbitrary axial position. A further possibility in this context is to fit a function of a field parameter (for example Phase, Real part, Imaginary part) of axial position to experimental data, which may enhance the accuracy of the determination of the optical field.

[0058] A key advantage of example embodiments of the present invention is the ability to detect and analyze particles freely diffusing or moving in a liquid volume, the use of the invention is not limited to this scenario. Earlier, iSCAT is commonly used to analyze particles adsorbed at the surface from which the background light is also reflected, typically being imaged while they are adsorbing at this surface since the change in optical signal is easier to detect and quantify. Various embodiments of the invention can be used also in this situation with some advantages as follows.

[0059] The described microscope allows the full optical field of a particle adsorbed at the floor of the sample holder or chamber to be determined, providing additional information on the properties of the particle compared to prior art where only the intensity contrast is determined and related to particle properties. In this case, the optical focus is advantageously close to the surface. Images are captured while the particles are adsorbing, facilitating removal of background patterns.

[0060] One further advantage of the described microscope and method is that when the axial positions of multiple particles in the sample have been determined it is possible to estimate the position of the sample holder or chamber relative to optical focus. This enables to, manually or automatically, adjust the optical focus to the desired axial position.

[0061] When analyzing biological particles, it is often desirable to identify these with the help of fluorescent markers to confirm their identity and composition, in addition to the purely optical data extracted by interferometric microscopy. The combination of fluorescence data and optical field data may be particularly useful.

[0062] It is thus an attractive possibility to combine the invention with fluorescent microscopy and simultaneously analyze both properties of individual particles. Implementation of fluorescence microscopy in the same optical system can for example be made by splitting one beam with the help of a dichroic mirror which separates out the wavelength of the fluorescence. This beam is directed at a separate camera, or optionally to a different part of the same camera sensor used for interferometric microscopy. Optionally an additional color filter is added before the camera sensor to further remove background light. In case the fluorescence light is weak relative to the background noise, the invention can help by tracking the particle with the optical scattering signal and subsequently average all the fluorescence images of a particle in order to statistically improve the signal to noise ratio and extract a quantified fluorescence signal. To facilitate quantification of the fluorescence light from a particle or optical feature, the fluorescence images can be digitally refocused using the already known axial position of the optical feature. Under some circumstances this can be done already on a regular fluorescence image. It can also be advantageous to add a component to generate interference also in the fluorescence image in order to facilitate accurate refocusing. For excitation of fluorescence, the same light source as used for interferometric microscopy could be used, or a separate light source with a different wavelength could be added to excite the chosen fluorescent dye. Other possibilities exist in order to extend the size range and the capabilities to extract more data. One option is to combine the here described interferometric microscope with and additional interferometric microscopy method using light transmitted through the sample. Thereby two different scattering angle ranges are accessed and more accurate data can be extracted as well as a wider size range of particles better accessed. This principle has been described previously by the inventors. Since the contrast of an optical feature in transmitted light is normally much lower than for interferometric backscattering, it is advantageous to track a particle using the contrast of the backscattering interferometric images or optical field and then average many corresponding optical field ROI from the corresponding 3D position based on transmitted light to generate a ROI of high contrast or signal-to- noise ratio. To access the corresponding 3D position, the transmission-based image is thus propagated to the corresponding depth along the optical axis. This way, particles which are substantially fainter than the image noise in individual images can be viewed clearly, and their optical field be quantitatively determined based on transmitted light.

[0063] Yet another possibility is to use two light sources of significantly different wavelength. This would allow to better access and characterize particles in the size range 0.1 - 1 pm, where backscattering as a function of particle size exhibit several minima, the position of which depends on the wavelength.

[0064] Further features of, and advantages with, the present invention will become apparent when studying the appended claims and the following description. The skilled person realize that different features of the present invention may be combined to create embodiments other than those described in the following, without departing from the scope of the present invention.

[0065] Brief description of the drawings

[0066] These and other aspects of the present invention will now be described in more detail, with reference to the appended drawings showing an example embodiment of the invention, wherein:

[0067] FIG. 1 is a schematic illustration of an interferometric scattering microscope according to prior art.

[0068] FIG. 2 is a schematic illustration of an interferometric microscope according to one embodiment of the invention, utilizing a 2D diffraction grating. FIG. 3 is a schematic illustration of an interferometric microscope according to one embodiment of the invention, utilizing a 1 D diffraction grating and a mask.

[0069] FIG. 4 is a schematic illustration of a simple sample holder and a sample chamber, and corresponding reflections, for interferometric scattering microscopy.

[0070] FIG. 5 is a schematic illustration of the relationship between light reflected by the sample holder and light scattered by the particle, and the effect of particle distance from the sample holder surface.

[0071] FIG. 6 is a schematic illustration of the complex optical field of illumination light, its interaction with a particle and the corresponding optical field of a particle.

[0072] FIG. 7 is a schematic illustration of the complex optical field of one particle at different distances from the sample holder surface.

[0073] FIG. 8 is a schematic illustration of a calculation process to perform the invention.

[0074] FIG. 9 is a schematic illustration of an interferometric microscope according to one embodiment of the invention.

[0075] Detailed description of the invention

[0076] The present invention will now be described more fully hereinafter with reference to the accompanying drawings, in which currently preferred embodiments of the invention are shown. This invention may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein; rather, these embodiments are provided for thoroughness and completeness, and fully convey the scope of the invention to the skilled person. Like reference characters refer to like elements throughout.

[0077] Fig. 1 shows a microscopy set-up 100 for interferometric backscattering microscopy according to prior art. The set-up comprises a light source 101 . The setup further comprises an optical system comprising multiple optical elements; lenses / beam expander 102, beam splitter 103, microscope objective 106, tube lens 104 and optionally quarter wave plate 105, lenses 108, 109) and one or several mirrors, to guide the light to the sample, and light reflected and scattered by the sample to the optical detector 110. The light passes one or several lenses 102 in order to expand the illumination beam and provide a desirable degree of collimation of the light at the sample held by a sample holder 107. The light is further reflected by a beam splitter 103, which is preferably polarization-dependent, and enters the microscope, following the optical path “backwards”. The light is focused at the focal plane of the microscope objective 106 and is collimated when reaching the sample from below. Light reflected by the sample holder 107, together with light scattered by particles in the sample, together makes out the image beam, which travels through the objective and tube lens 104, towards an optical detector such as a camera, 110. Optionally, there is a quarter-wave plate 105 below / behind the objective 106 in the case of using a polarization selective beam splitter. As light first passes this plate 105 in one direction and then on the way back after having been reflected, its polarization is in total rotated 90 degrees. This enables the very most of the light to pass through the beam splitter and this arrangement limits the light losses in the set-up and improves signal-to noise ratio. It is however not strictly necessary. One further option is to use two lenses 108,109 to create a focal plane where collimated light in the image plane is focused to a small spot. In the focal plane spatial filters may be placed for selective removal of background light. Lenses 108, 109 can also be utilised to magnify the image further, if they have different focal length.

[0078] Any angle between reflected light and the optical axis, in the direction the illumination light travels, of more than 90 degrees and up to 180 degrees may be considered as backscattering and may be considered an iSCAT set-up. It is however simple and useful to implement the invention with close to straight back reflection, with between 160 and 180 degrees between reflected light and the optical axis in the illumination direction. A certain angle is useful to remove unwanted reflections from the objective lens or top of the chamber roof or from other interfaces, from the image. Such an angle is most appropriately between 5 and 20 degrees relative to the optical axis. If using smaller angles, other measures may have to be taken to diminish unwanted reflections in the image. By optical axis is meant an imaginary line through the geometrical center of the optical system, and here specifically a line through the center of the microscope objective.

[0079] The processing unit 111 may include a microprocessor, microcontroller, programmable digital signal processor or another programmable device. The processing unit may also, or instead, include an application specific integrated circuit, a programmable gate array or programmable array logic, a programmable logic device, or a digital signal processor. Where the processing unit includes a programmable device such as the microprocessor, microcontroller or programmable digital signal processor mentioned above, the processor may further include memory holding computer executable code that controls operation of the programmable device.

[0080] Fig. 2 shows an interferometric microscope 200 for interferometric backscattering microscopy according to an example embodiment of the invention. The interferometric microscope 200 comprises: a sample holder 207 configured to hold a sample; a light source 201 arranged to illuminate the sample in the sample holder 207; an optical detector 210 arranged to capture images of light scattered by a feature of the sample; an optical system comprising a plurality of optical elements lenses / beam expander 202, beam splitter 203, microscope objective 206, tube lens 204 and optionally quarter wave plate 205, lenses 208, 209) and one or several mirrors, arranged to direct an illumination beam of light from the light source to the sample holder and to collect light reflected by the sample and by the sample holder 207 to form an image beam directed to the optical detector 210.

[0081] The interferometric microscope 200 further comprises an optical component 212 arranged in the path of the image beam and configured to divide the light in the image beam into a plurality of beam portions, wherein the beam portions are propagating along a common optical beam path, wherein the beam portions interfere at the detector; and a processing unit 211 configured to determine, from a plurality of images captured by the optical detector 210, a parameter indicative of the absolute value of the optical field of the light scattered by the feature.

[0082] The set-up is thus similar to that in Fig. 1 , except for the addition of a grating 212 in front of the optical detector 210 (camera). The grating 212 may comprise a 2- dimensional pattern where different fields provide a different phase shift of the light, these fields may be separated by opaque lines. One type of such a 2D grating pattern is a chessboard pattern on a glass or quartz substrate where half of the squares are etched down to provide a phase shift compared to light having passed the other squares. Light portions having passed different fields interfere with each other at the detector, generating an interference pattern based on four different diffraction orders. The grating is typically placed 0.5-2 mm from the camera sensor. Alternatively, the grating can also be placed on a similar distance from an image plane earlier in the optical path, for example in an intermediate image plane between the tube lens 204 and lens 208. The processing unit 211 captures images from the camera 210 and process the images to extract the optical information as described. Fig. 3 shows an interferometric microscope 300 for interferometric backscattering microscopy according to an example embodiment of the invention. The microscope of Fig. 3 is thus in large parts similar to the microscope 200 in Fig 2. The main difference is that in Fig. 2 the light is divided in many portions whereas in Fig. 3 the light is divided into few portions of which only two are utilized. Another difference is that that the optical element dividing light is close to an image plane in Fig. 2 whereas it is close to an earlier image plane in Fig 3. It is however not strictly necessary to divide the light near an image plane, it can also be done anywhere else in the beam including at a focal plane. A 1 -dimensional grating 313 is splitting the beam in several different parts, where the 0thand one of the 1stdiffraction orders are used. As these beams reach the focal plane between lens 308 and 309 plane, they are passed through a mask 312 with two apertures of different sizes. The small aperture allows only light to pass which has not been scattered by any particles or other features, with other words uniform illumination. This light portion will serve as reference light. The larger aperture allows light in the other light portion to pass, including also light scattered by the sample. These two light portions then interfere at the camera sensor. One alternative implementation of this method is to instead of a 1 D grating 313, using a beam displacer, which may be placed downstream lens 308. Using a beam displacer may also require components to control the beam polarization since the beam displacer will split the beam in two orthogonally polarized portions. The processing unit 311 captures images from the camera 310 and process the images to extract the optical information as described.

[0083] Fig. 4A shows a common type of sample holder 400 for interferometric backscattering microscopy. A glass cover slip used as a floor 401 is placed above the microscope objective. A drop of liquid sample 402 is placed on top of the glass cover slip, said sample 402 comprising a particle 403. Incident light 404 is illuminating the sample 412 and sample holder 400 from below. A minor part of the light is partially reflected backwards from the liquid / glass boundary, i.e. floor interface light reflection 405. Incident light 404 is scattered by the particle 403, some of the scattered light is directed backwards as particle light reflection 406.

[0084] Due to the difficulties of performing interferometric backscattering microscopy on enclosed samples, due to the many reflections, this simple sample holder is the most common. Fig. 4B shows a sample holder 410 where the sample 402 is placed in a chamber, having roof 408 and a floor 401 . The sample comprises the particle 403. Incident light 404 illuminating the sample holder from below is reflected by the ceiling / upper interface of the chamber as ceiling interface light reflection 407, by the floor / lower interface 405 as floor interface light reflection 405, and scattered by the particle as particle light reflection 406. Depending on the wavelength, the Rl of the liquid and the depth of the chamber, the reflections 405 and 407 may be in phase or out of phase and thus have constructive or destructive interference with each other. Both cases may be used and have their advantages and disadvantages.

[0085] Fig. 5 shows two particles different distances from the floor 501 of the sample holder or chamber. This means that the particles also have different axial positions, which means they are at different positions (in different planes) along the optical axis. Each particle reflects / backscatters light 503 and 504. The interface between the cover slip (typically made of glass) and the liquid medium reflects light 502. Depending on the distance between the interface and the particle, the scattered light from the particles may interfere constructively (503 and 502) or destructively (504 and 502). The particles in the image will have diffraction rings around the center and if the interference is constructive the center 505 will be brighter than the image average, whereas with destructive interference the center 506 will be darker than the image average. This results in that particles are rapidly blinking as they undergo Brownian motion, rapidly changing their distance to the reflecting interface. This makes detection more difficult to consistently detect and track particles in a pure intensity contrast image, and it also makes it more difficult to quantitatively relate the optical information from the image to particle properties.

[0086] Fig. 6 discloses a 2-dimensional coordinate system having a Real axis 601 and an Imaginary axis 602 so as to form a complex coordinate system. Hence, a complex number can be represented in this coordinate system and, thus, an optical field expressed as a complex number can be presented in such a coordinate system. The complex number can be either expressed by its Imaginary and Real value, or by the length of a vector (such as 604) and the angle 603 between said vector and the Real axis 601 . The length of the vector is called the absolute value or amplitude of the complex number (or optical field) and the angle 603 is called the phase. In Fig. 6 are disclosed examples of detected and calculated values of different optical field as complex numbers. The optical field of light which has not interacted with any particle, also commonly referred to as background light, is depicted as background optical field vector 605 and stretches along the Real axis 601 . The optical field of light which has passed interfered with a particle is depicted as total optical field vector 604. The light which has interacted with a particle may achieve a phase shift causing the total optical field vector 604 to deviate from the extension direction of the background light vector 605. The phase shift is represented by a phase shift angle 603 between the background optical field vector 605 and the total optical field vector 604. Furthermore, the light which has interacted with the particle will have a shorter amplitude than the background light if the light scattered by the particle undergo destructive interference with the background light reflected by the sample holder (as depicted in the image) and will have a higher amplitude if the interference is instead constructive. If the optical field of the background light is subtracted from the optical field of light passing and interfering with the particle, the optical field of the particle is extracted. In Fig. 6, the optical field of the particle expressed as a complex number may be found by subtracting the background optical field vector 605 from the total optical field vector 604 so as to form a particle optical field vector 606, which has a negative Real part and a positive Imaginary part. The particle optical field vector 606 thus represent a complex value of the optical field of the particle.

[0087] Optical extinction causes a decrease in the absolute value of the optical field, also referred to as optical signal, of light having passed the particle, represented by the total optical field vector 604, compared to the illumination background light represented in the figure by background optical field vector 605, as is obvious from Fig 6. The absolute value of the optical field corresponds to the length of a vector and the absolute value of the optical field from light having passed the particle corresponds to the length of the total optical field vector 604 and the absolute value of the optical field of the background light corresponds to the length of the background optical field vector 605. The particle optical field vector 606, which is created by subtracting the background light vector 605 from the total light vector 604 will translate into a negative Real part of the particle’s optical field, i.e. a negative Real part of the particle optical field vector 606.

[0088] Fig. 7 Illustrates how the optical field of a particle will change during a measurement in a 2-dimensional coordinate system having a Real axis 701 and an Imaginary axis 702. As the particle is changing position along the optical axis from image to image, the scattered light will undergo constructive or destructive interference and the phase shift of the interfering light will change between positive and negative values. The result is that the determined optical field 706 for each image 707 will be found at a different position along a circular pattern in the coordinate system. For this reason, it is difficult to determine the optical properties of a particle from field parameters such as phase shift or intensity contrast. Prior art teaches therefore to determine the intensity contrast for all particle images and use a small percentage with the highest values to determine particle properties. However, in the present invention the absolute value of the field amplitude can be determined and will be similar for each image. This means that considerably more image data can be averaged to find a more accurate value of the particles (back)scattering cross section. Averaging a larger number of data points gives less standard deviation in the determined parameter. Averaging can be made either by determining the amplitude (integrated over the particle projection) for each particle ROI and then averaging the values, or by generating an averaged particle ROI based on all the individual ROIs and subsequently determine the integrated amplitude.

[0089] The integrated amplitude (absolute value) of the optical field can be used to estimate the scattering cross section, which is primarily proportional to the particle volume and the difference between the Rl of the particle and the surrounding medium. Using the hydrodynamic size of the particle and the scattering cross section, the refractive index can therefore be estimated from these two independent parameters. If the Rl of the particle and medium is known, the integrated amplitude or scattering cross section can instead be used to estimate the particle size.

[0090] Plotting the amplitude or scattering cross section of each particle against the hydrodynamic size allows different particle populations to be differentiated.

[0091] A method for characterizing a feature of submicron- or nanoparticles in an interferometric microscope comprises: by a light source, illuminating a sample in a sample holder; by an optical system comprising a plurality of optical elements, directing an illumination beam of light from the light source to the sample holder, and collecting light reflected by a feature of the sample and by the sample holder to form an image beam directed to an optical detector; by an optical component arranged in the path of the image beam, dividing the light in the image beam into a plurality of beam portions, wherein the beam portions are propagating along a common optical beam path, and wherein the beam portions interfere at the optical detector; by the optical detector, capturing a plurality of images of light scattered by a feature of the sample; and by a processing unit, determining, from a plurality of images captured by the detector, a parameter indicative of the absolute value of the optical field of the light scattered by the feature.

[0092] Fig. 8 illustrates a calculation process to perform the invention. 801 : Images are captured from the camera. If computation is more time consuming than the capture rate, the images are preferable stored in a video file or image stack and subsequently processed after the capturing process is stopped. 802: Images are Fourier transformed, masked to select information of interest, and the optical field is reconstructed. 803: Images are processed to remove background noise / patterns and separate out the field contrast for the features / particles of interest. 805: If image features other than particles are imaged, the absolute value of the optical field can now be taken and absolute field contrast images be generated and analyzed. 804: Particle detection and trace generation. Images with absolute value contrast are generated at one or preferably several propagated image planes. One option to facilitate detection is to generate a total projection of these planes to extract the maximum contrast of each particle. Following detection, including position determination, the detections in different images are linked together to form a trace. A focused cropped image of the traced particle in each image is stored. Optionally a post-cleanup is performed where traces with few images are discarded and individual particle images (ROI) of poor quality are removed. 806: Generating amplitude average for each particle / trace. The first option is to take the absolute value of each field ROI and thus generate an absolute value contrast image of each particle image (ROI) An average image is then generated by averaging all the ROIs in the trace for the particle. This option is more insensitive to accurate position determination in the axial direction. 807 Another way to generate the absolute field for a particle is to generate an average field ROI. Since the phase of the field fluctuates as a function of the axial position of the particle, the phase as a function of the axial position will need to be adjusted for each ROI. This is preferably done to extrapolate the field to the optical focus, but can be done to any arbitrary z-position. In one embodiment it can even be done without knowing the axial position, only be adjusting the phase to match that of an arbitrarily chosen ROI. In such case only the amplitude information is of use for further analysis. Following averaging of all the field ROIs, the absolute value is taken to generate an absolute value contrast ROI. 808: The amplitude contrast is quantified by integrating the contrast over the projected area of the particle. This can be facilitated for faster processing by fitting a gauss function to the particle ROI and subsequently integrating this gauss function.

[0093] Fig. 9 shows an interferometric microscope 900 which is similar to the one described in Fig. 2, but with addition of components to enable imaging of the sample also with transmitted light. A second light source 913 is directing light towards the sample 907, light transmitted through the sample passes through the microscope objective 906 and further through the beam splitter 903 and the tube lens 904. The light from light source 913 is of a different wavelength than light from the first light source 901. Downstream a first lens 908, a dichroic mirror 914 splits the transmitted light originating from second light source 913 from the light originating from the first light source 901 . After passing a second lens 915, preferably similar to lens 909, the transmitted light is directed to a second optical detector 916 with a grating 917 of similar type as 912. This microscope allows to capture images from backscattered light and transmitted light simultaneously. Using suitable software algorithms, the overlay of the two captured images can be fine-tuned. The dichroic mirror 914 can be placed in different positions, such as upstream lens 908 or even upstream tube lens 904. Furthermore, the lenses 908, 909 and 915 are used for additional magnification and are optional to include depending on the objective used and magnification desired. They can also be utilized for spatial filtering, by placing a spatial filter in the focal point between them, as described earlier by the inventors. A further alternative embodiment is to direct both beams to different areas of the same detector instead of using two separate detectors.

[0094] Even though the invention has been described with reference to specific exemplifying embodiments thereof, many different alterations, modifications and the like will become apparent for those skilled in the art.

[0095] Additionally, variations to the disclosed embodiments can be understood and effected by the skilled person in practicing the claimed invention, from a study of the drawings, the disclosure, and the appended claims. In the claims, the word "comprising" does not exclude other elements or steps, and the indefinite article "a" or "an" does not exclude a plurality. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measured cannot be used to advantage.

Claims

CLAIMS1 . An interferometric microscope (200, 300) comprising: a sample holder (207) configured to hold a sample; a light source (201 ) arranged to illuminate the sample in the sample holder; an optical detector (210) arranged to capture images of light scattered by a feature of the sample; an optical system comprising a plurality of optical elements (202, 203, 204, 206) arranged to direct an illumination beam of light from the light source to the sample holder and to collect light reflected by the sample and by the sample holder to form an image beam directed to the detector; an optical component (212, 313) arranged in the path of the image beam and configured to divide the light in the image beam into a plurality of beam portions, wherein the beam portions are propagating along a common optical beam path, wherein the beam portions interfere at the detector; and a processing unit (211) configured to determine, from a plurality of images captured by the detector, a parameter indicative of the absolute value of the optical field of the light scattered by the feature.

2. The interferometric microscope according to claim 1 , wherein the optical system is arranged so that the plurality of beam portions propagating along a common optical beam path are propagating via the same optical components or directly to the optical detector.

3. The interferometric microscope according to claim 1 or 2, wherein the processing unit is configured to: capture a plurality of images; detect a suspended moving particle in a plurality of captured images; link the positions of the detected particle together to form a trace for each detected particle; determine a digitally refocused optical field for the particle at each position, forming a plurality of digitally refocused optical field; anduse a majority of the formed digitally refocused optical fields of the particle to determine a parameter indicative of an average absolute value of the optical field of the particle.

4. The interferometric microscope according to claim 1 or 2, wherein said optical component which divides the light into a plurality of beam portions is a beam splitter, beam displacer, grating, diffraction grating, meta surface, spatial light modulator, or microlens array.

5. The interferometric microscope according to any one of the preceding claims, wherein the optical component is configured to divide the image beam into at least 100 portions.

6. The interferometric microscope according to any one of the preceding claims, wherein the optical component is a two-dimensional diffraction grating.

7. The interferometric microscope according to any one of the preceding claims, wherein the sample holder comprises a chamber with a transparent floor and a ceiling, and wherein the sample holder is arranged so that the illumination beam illuminates the sample through the transparent floor, and the internal height between the floor and the ceiling is at least 1 micrometers.

8. The interferometric microscope according to any one of the preceding claims, further comprising: a light source arranged to illuminate a particle in order to excite a fluorescent stain in the particle; a plurality of optical elements arranged to direct an illumination beam of light from the light source to the sample holder and to collect light emitted by the fluorescent stain in the sample and form a fluorescence image beam directed to a detector; and a processing unit configured to determine, from a plurality of images captured by the detector, a parameter indicative of the quantity of light emitted by the fluorescent dye in a detected particle.

9. The interferometric microscope according to any one of the preceding claims, further comprising: an additional light source (913) arranged to illuminate the sample in the sample holder such that the illumination light is transmitted through the sample; an optical component 914 placed downstream the sample and arranged to split light from the first light source 901 and the second light source 913 into separate beams an optical detector (916 or 910) arranged to capture images of light from said light source transmitted through the sample and scattered by a feature of the sample; and a processing unit configured to determine, from a plurality of images captured by the detector, a parameter indicative of the scattering cross section of the feature of the sample10 The interferometric microscope according claim 9, wherein the processing unit is further configured to: in multiple images determine a three-dimensional position of a feature in the sample based on an image based on backscattered light; digitally refocus an image area to the corresponding three-dimensional position in the corresponding multiple images based on transmitted light; generate an average of multiple images of the optical feature from the image based on transmitted light; and use said average image to determine a parameter indicative of the scattering cross section of the feature of the sample.11 The interferometric microscope according to claim 10, wherein the processing unit is further configured to use said average image to determine the optical field of the optical feature of the sample.

12. Method for characterizing a feature of submicron- or nanoparticles in an interferometric microscope, the method comprising: by a light source (201 , 301 ), illuminating a sample in a sample holder (207,by an optical system comprising a plurality of optical elements (202, 203, 204, 206), directing an illumination beam of light from the light source to the sample holder, and collecting light reflected by a feature of the sample and by the sample holder to form an image beam directed to an optical detector; by an optical component (213) arranged in the path of the image beam, dividing the light in the image beam into a plurality of beam portions, wherein the beam portions are propagating along a common optical beam path, and wherein the beam portions interfere at the optical detector; by the optical detector, capturing a plurality of images of light scattered by a feature of the sample; and by a processing unit (211 ), determining, from a plurality of images captured by the detector, a parameter indicative of the absolute value of the optical field of the light scattered by the feature.

13. The method according to claim 12, wherein the optical feature for which a parameter indicative of the absolute value of its optical field is determined is a particle suspended in a liquid medium and undergoing diffusive movement.

14. The method according to claim 12 or 13, wherein features or particles are digitally refocused within a range along the optical axis of at least 2 micrometers.

15. The method according to any one of claims 12 to 14, further comprising, by the processing unit: determining a parameter indicative of the hydrodynamic diameter of a particle; determining a parameter indicative of the scattering cross section of a particle using parameter indicative of the hydrodynamic diameter and parameter indicative of the scattering cross section; and categorizing detected particles into particle populations with different population density maxima in the parameter space of said parameters.

16. The method according to any one of claims 12 to 15, further comprising, by the processing unit: determining a parameter indicative of the hydrodynamic diameter of a particledetermining a parameter indicative of the scattering cross section of a particle using the ratio between the parameter indicative of the scattering cross section and the cube of the parameter indicative of the hydrodynamic diameter to determine the refractive index of the particle.

17. The method according to any one of claims 12 to 16, further comprising, by the processing unit: determining the scattering amplitude at multiple different scattering angles for a particle; providing a reference library of scattering amplitude at different scattering angles for particles with a range of known sizes and refractive indices; comparing said scattering amplitude at multiple different scattering angles to said reference library to estimate the size of the particle.

18. The method according to any one of claims 12 to 17, further comprising, by the processing unit: capturing multiple images of a particle at different axial positions; determining said axial positions; determining the optical field at each of said axial positions; and using said determined axial positions and said determined optical field at said axial positions, and a known relation between the determined axial positions and the determined optical fields to estimate the optical field of the particle at one specific axial position.

19. The method according to any one of claims 12 to 18, further comprising, by the processing unit: determining a position of the particle along the optical axis; and using the position of the particle along the optical axis in multiple images to estimate a parameter indicative of the hydrodynamic diameter of the particle.

20. The method according to any one of claims 10-19, further comprising: by a light source, illuminating the sample in order to excite a fluorescent stain in the sample;by an optical system comprising a plurality of optical elements, directing an illumination beam of light from the light source to the sample holder and collecting light emitted by the fluorescent stain in the sample in order to form a fluorescence image beam and directing it to a detector; and by the processing unit, determining, from a plurality of images captured by the detector, a parameter indicative of the quantity of light emitted by the fluorescent dye in a detected particle.21 . The method according to any one of claims 12 to 20, further comprising: by an additional light source (913), illuminating the sample in the sample holder such that the illumination light is transmitted through the sample; by an optical component 914 placed downstream the sample, splitting light from the first light source 901 and the second light source 913 into separate beams; by an optical detector (910, 916), capturing images of light from said light source transmitted through the sample and scattered by a feature of the sample; and by the processing unit, determining, from a plurality of images captured by the detector, a parameter indicative of the scattering cross section of the feature of the sample.

22. The method according to any one of claims 12 to 21 , further comprising, by the processing unit: in multiple images determining a three-dimensional position of a feature in the sample based on an image based on backscattered light; digitally refocusing an image area to the corresponding three-dimensional position in the corresponding multiple images based on transmitted light; generating an average of multiple images of the optical feature from the image based on transmitted light; and using said average image to determine a parameter indicative of the scattering cross section of the feature of the sample.

Images