High resolution imaging

[0001] The present invention relates to a method and apparatus for providing image data on the basis of which an image of a target object can be created. Image data provide high resolution information about the structure of the target object and allow high-resolution image created on its basis. In particular and inter alia, the present invention relates to a method and apparatus for providing images at a resolution limited by the length of the light wavelength without the need for accurate positioning of the incident radiation relative to the probed target.

[0002] Many types of imaging techniques are known for obtaining spatial information about a target object (or object). In conventional transmission imaging, the object is irradiated with flat wave light 10. The waves scattered by the object are subjected to repeated interference by the lens 12 to form the image shown in Figure 1A. For imaging using very short wavelengths of radiation (x-rays or electrons), this technique creates many known difficulties associated with aberrations and instability caused by the lens, which limit the resolution and the interpretation of the image obtained. Typical, achievable resolution is many times greater than the theoretical radiation wavelength. Conventional scanning transmission imaging is another example of such an imaging technique in which the lens is used to focus the point of radiation passing through the target object. On the side behind the object there are one or more detectors ensuring detection of scattered radiation. Various types of detector strategies are known, such as angular detectors, quadrant detectors and / or detectors excluding access. However, these methods rely on scanning a focused radiation point at all points for which it is desirable to create an image. There are a number of problems associated with such techniques, such as the requirement for very accurate spot control, because if a 1000 x 1000 pixel image is desired, it is necessary to use one million accurate position probes. Another problem is the need for very high quality lenses. This requirement results not only from the fact that the resolution of the final image is only as good as the sharpness and location of the spot allows, but also because of various forms of radiation, such as electrons or x-rays, there are many problems, such as effects aberration, chromatic scattering and current instability of the lens, which can affect the image creation process and drastically limit the resolution. This process is schematically shown in Figure 1B, in which incident radiation 15, such as an electron beam or X-ray radiation, falls on sample 16 being the target. The radiation diffused by the object leaves the target object and propagates to the detector surface 17.

[0003] A known problem associated with conventional scanning transmission imaging is the considerable time it takes to obtain the images as a result of the number of points probed by the incident radiation point. If the target moves while data is being collected, this may lead to inaccurate data collection and thus to inaccurate images. In addition, conventional methods of transmission scanning imaging do not allow measurement of information about the radiation phase leaving the target object. It is only possible to measure the total intensity of scattered radiation reaching the detectors. It is not possible to collect phase information as such regarding an outgoing wave that reaches beyond the target object.

[0004] Modification of conventional scanning transmission imaging is deconvolutionary four-dimensional imaging. It uses a device similar to that of figure 1, but the entire diffraction pattern is recorded for each probe position. This solution provides a method of determining the structure of the target object with a resolution better than the size of the radiation point or the function of the lens used, but also has a number of problems. The most important problem is the need to collect huge amounts of data, which means that in order to provide an adequate field of view, it is necessary to collect data for several hours. This makes performing the experiment very difficult in practice, because very accurate control of the probe radiation and its exact movement to scan each (million) pixel for reconstruction of the final image is crucial. The huge doses of incident radiation required for large image acquisition times can also cause serious damage to the target object.

[0005] Another well-known imaging technique is pure diffraction imaging. With this alternative strategy, it is possible to omit the lens and the target object is exposed to a simple, flat wave of probe radiation. The radiation scattering pattern measured in the far field creates a Fourier plane diffraction pattern, the intensity of which can be recorded. Then an iterative method is used, which uses information obtained from the measured intensity values ​​to calculate the estimated wave field leaving the object. In order to determine the actual information about the target object based on the estimated wave field, it is necessary to provide an area in real space that is known to be in this area or is masked in a certain way. It is only on the basis of knowledge of this fact that the current estimation of the wave field representing the object can be changed in an iterative way. However, there are many problems associated with purely diffractive imaging. The most important problem is to suspend or isolate the target object in a certain way, in a fixed place.

This solution is very difficult to implement in practice. It is also impossible to extend the solution to new or other parts of the object or to obtain a large image with good resolution for the entire image. It is possible to expose and obtain a solution only for one, isolated area of ​​the object. The target object must also have one value. This means that it must be represented by a single real number. This number can be a representation of absorption or phase change, but it cannot represent both. In fact, most of the real waves of target objects (i.e., wave functions leaving the target object) appear as complex numbers that have both phase and amplitude components.

[0006] Another important problem associated with purely diffractive imaging is the fact that the edge of the target must be sharply defined and must therefore have a clear edge. This makes it possible to accurately determine the area where it is known that the object is not in it or is masked in it in a certain way. In practice, it is difficult to create an object or a gap with such a defined edge.

[0007] Further problems occur with weakly scattering objects, which are a type of objects commonly found in the case of X-ray scattering and electrons, where most of the radiation passing through the object reaches the center of the diffraction pattern. Information is lost in this zone because it does not help in the process of creating the image, and in addition radiation passing through the object can damage it. It is also necessary to use parallel exposure. However, this means that for a source of a given intensity in the object plane, we get very low counting values. In conjunction with the fact that a significant part of the radiation passing through the lightly scattering objects reaches the middle zone, as mentioned above, it means that in practice the whole experiment requires a long time to obtain the right number of counts. If the object or other imaging device moves or moves during the exposure stage at the data collection stage, it may result in data loss. [0008] A method of searching for a solution that has enjoyed significant interest in recent years is an iterative method proposed for the first time by Gerchberg and Saxton [RW Gerchberg and W. O. Saxton. Optik, 35 (2): 237-246, 1972]. Such iterative methods have been used in recent years to the geometry shown in Figure 2, both for X-rays and electrons. In this configuration, incident radiation 20 is directed to sample 21 being the target object. The target object scatters the incident radiation over a wide range of angle values, creating a diffraction pattern in the diffraction plane 22. Such a diffraction pattern in the diffraction plane can be recorded by any suitable method, such as a photographic film or CCD detector. The experimental advantage of diffraction is that the interference condition is determined only by the scattering of radiation inside the target itself, thus avoiding the serious difficulties of using lenses at a short wavelength.

[0009] US 6,049,074 is a patent describing a photoelectric conversion device, intended for use in a digital x-ray imaging device, capable of converting the energy of incident x-rays into visible light.

[0010] US 2004/000630 is a document describing an imaging system with direct conversion of a high energy scan, having a sensor capable of converting incident X-rays into an electronic signal.

[0011] EP1120086 is a patent for a mammography device that detects coherently scattered radiation at a reduced angle value resulting from irradiation of an object with penetrating radiation to determine certain characteristics of the object being exposed. FAULKNER HML ET AL: "Moveable aperture lensless microscopy: a novel phase retrieval algorithm", ELECTRON

MICROSCOPY AND ANALYSIS 2003. INSTITUTE OF PHYSICS ELECTRON MICROSCOPY AND ANALYSIS GROUP CONFERENCE 3-5 SEPT. 2003 OXFORD, UK; [INSTITUTE OF PHYSICS CONFERENCE SERIES, ISSN 1755-1315; VOL.179], INSTITUTE OF PHYSICS PUBL, BRISTOL, vol. 179, 18 February 2004 (2004-02-18), pages 337-340 is a paper describing a method that provides image data for a target object area by providing radiation incident from the radiation source to the target object, detecting the intensity of radiation scattered by the target object using a detector and using a movable, sharp gap and an iterative process. The purpose of embodiments of the present invention is to at least partially alleviate said problems. The purpose of another embodiment of the present invention is to provide a method and apparatus for providing image data that can be used to construct a high resolution image of an object using the advantages of iterative techniques but not burdened with the disadvantages of iterative methods known in the art.

[0012] The purpose of embodiments of the present invention is to provide a method and device providing image data that can be used to construct a high resolution target image without the need for high resolution positioning techniques to set the incident radiation relative to the target object or to embed the target in strictly specific position.

[0013] The purpose of embodiments of the present invention is to provide a transmission microscopy method suitable for use with all forms of radiation having a resolution independent of the use of the lens and / or holographic reference wave and / or any form of far field interferometry.

[0014] The purpose of embodiments of the present invention is to provide a method and apparatus providing image data that can be used to generate an image with a wavelength limited resolution. The image may include part of the target or, in the case of a well-defined target structure, radiation at a selected location on the experimental track.

[0015] According to a first aspect of the present invention, there is provided a method of providing image data for the purposes of constructing a high resolution image of a target area, comprising the steps of: providing radiation incident from a radiation source onto the target; detecting the intensity of scattered radiation by the target using the slot behind the target in the first position relative to the target object, displacement of the gap relative to the target and re-detecting the intensity of scattered radiation by the target using the slot in the second position relative to the target using at least one detector, or using at least one detector, detection of the intensity of radiation scattered by the target object with incident radiation in the first position relative to the target object, displacement of the incident radiation relative to the target object and detection of the intensity of radiation scattered by the target object with incident radiation in the second position relative to the target object and selection of the second position as such, that the first surface of the target area selected in the first position at least partially overlaps with the second surface of the target area determined in the second position, and corresponding at least to the intensity detected in the first and second positions, providing image data in the iterative process and the use of a slightly variable function transmittance or function of the mobile exposure relative to the target object, where the slightly variable function of transmittance or exposure is a function limited by bandwidth, not containing high spatial frequencies.

[0016] According to a second aspect of the present invention, there is provided a device providing image data for generating high resolution image of a target area having: solution locating the target object in a fixed environment, a source of radiation providing radiation incident to the target object with the help of a locating solution, at least one detector detecting the intensity of radiation scattered by the target object, a solution locating the incident radiation in the first and second position relative to the target object or a solution locating a gap behind the target in the first and second positions relative to the target and selecting the first and second positions such that the first surface of the target area defined in the first position overlaps at least partially with the second surface of the target area defined in the second position, and a processing solution providing image data, responding to the detected intensity of scattered radiation in the first and second positions by means of the iterative method and using a slightly variable transmittance function or exposure function, while the slightly variable transmittance or exposure function is a function of limited bandwidth, not containing high spatial frequencies.

[0017] The methods preferably provide image data with a wavelength limited resolution.

[0018] A convenient solution is to provide image data for the target area in real time.

[0019] Incident radiation is preferably a localized wave field. [0020] According to a third aspect of the present invention, a computer program according to claim 31 is provided.

[0021] Embodiments of the present invention use an iterative method to provide image data that can then be used to construct an image of a target area with a wavelength limited resolution. A slightly focused beam of radiation falls on the target object or a movable gap is placed behind the object. One, two or more diffraction patterns are collected either by moving the target object or by moving the gap, which allows detection of another area of ​​the wave function leaving the object. Preferably, the gap may travel relatively long distances (equal to half the gap width) or larger before registering the scattering patterns in positions. This means that the method according to the present invention makes it possible to scan large fields of view and obtain images with very high resolution in real time. An alternative to moving the target or slot is to change the selected probe or object characteristics.

[0022] Embodiments of the present invention are described below, by way of example only, by reference to the accompanying drawings, in which:

Figures 1A and 1B illustrate the use of conventional transmission imaging and conventional scanning transmission imaging, respectively;

Figure 2 shows how diffraction does not limit the range of angular values;

Figure 3 shows how the configuration of the movable gap allows measurement for a large field of view;

Figure 4 shows the movable slot positioned behind the target;

Figure 5 shows how the mobile focusing probe allows measurement for a large field of view;

Figure 6 shows a probe incident on a target;

Figure 7 shows a phase acquisition algorithm;

Figure 8 shows the results for intensity and for phase;

Figure 9 shows additional results for intensity and for phase with added noise;

Figures 10A, 10B and 10C illustrate alternative methods of providing radiation on a target;

Figures 11A and 11B show how embodiments of the present invention can provide a surface profilometer; and

Figure 12 shows devices providing image data.

[0023] Analogous reference numbers in the drawings relate to analogous parts.

[0024] Figure 3 shows how the scattering pattern can be developed and how it can be used to determine high resolution information about the structure of the target. It should be understood that the term "target" means any sample or object placed in the path of incident radiation that causes the scattering of this radiation. It should be understood that the target should be at least partially transparent to incident radiation. The target may or may not have a repeating structure.

[0025] Incident radiation 30 is directed so that it hits the target 31. It should be understood that the term "radiation" broadly means energy from a radiation source. It will include electromagnetic radiation, including x-rays and emitted particles such as electrons and / or sound waves. Such radiation should be represented by the wave function Ψ (γ). This wave function has a real part and an imaginary part, which will be understood by people with skills in the field. It can be represented by means of a wave and phase function module. Ψ (γ) * means complex stress Ψ (^ and Ψ (γ) Ψ (γ) * = | Ψ (γ) |² where | Ψ (γ) |² is the intensity that can be measured with the wave function.

[0026] Incident radiation 30 is scattered as it passes through and out of the sample 31. The wave function of the incident radiation leaving the sample will be changed in both amplitude and phase compared to the wave function of the incident radiation before passing through the sample. The scattering that occurs may include Fourier diffraction, deflection and / or Fresnel diffraction and any other form of scattering due to which the incident radiation pattern is changed due to radiation propagation behind the sample. If a series of detectors such as the CCD 32 detector are placed at a large distance from the sample, a diffraction pattern will be created in the diffraction plane 33. The Fourier diffraction pattern will arise when the detectors 32 are at a distance D from the sample, where D is large enough to be able to effectively create a diffraction pattern from a point source. If the diffraction plane is formed closer to the sample by placing the detector closer to it, a Fresnel diffraction pattern will be created. Slot 34 is behind the target, allowing the target area to be tested. The gap is formed in the mask, so that it defines a "bracket". A bracket is a surface of a function within which the function has non-zero values. In other words, the function assumes zero values ​​outside the bracket. In the space outside the bracket, the mask blocks radiation transmittance. Unlike techniques known in the art, the gaps used in the present invention need not be finite or sharply defined. They can be movable and slowly change at the edges. Thanks to this solution, the gently changing exposure function or transmittance does not contain high spatial frequencies. In other words, it is a function limited by bandwidth. Because this system does not use a lens, the detectors 32 can measure for a large field of view. The term "slot" describes the localized function of radiation transmission. It can be represented by means of a two-dimensional complex variable having a module value from 0 to 1. An example may be a mask having a physical gap area with variable transmittance.

[0027] Figure 4 schematically shows wave propagation through the system shown in Figure 3. Incident radiation 30 falls on the upper side of the sample 31 and is dispersed by the sample as it passes through it. The sample wave O (r) is the initial wave function of the radiation after its interaction with the object 31. In this way, O (r) represents a two-dimensional complex function, so that each point of the O (r) function, where r is a two-dimensional coordinate, has a complex value assigned to it. O (r) will physically represent a wave function that would emanate from an object exposed to a flat wave. For example, in the case of electron scattering, the O (r) function would represent phase and amplitude changes introduced into the incident wave as it passes through the object of interest. The slot 34 provides a probe function P (r) (also called a filter function) that selects a portion of the wave function that leaves the object for analysis. It should be understood that behind the object's function, instead of choosing a slot, there may be a transmission grid or other filtering function of this type. The P (rR) probe function is a slot transmission function where the slot is in position R. The probe function can be presented in the form of a complex function whose complex value is given by a module and phase, representing modifications of the module and phase introduced by the probe to the ideal plane wave above it.

[0028] The output wave function ip (r, R) 43 is the output wave function of the radiation leaving the gap. This wave function Ψ (γ ^) 43 creates a diffraction pattern Ψ (^) 44 in the diffraction plane 33. In this case, Γ is a vector coordinate in real space, and k is a vector coordinate in diffraction space.

[0029] Figure 5 shows another embodiment of the present invention without requiring the presence of a gap. In this case, incident radiation 50 falls on the first surface of the target 51. Incident radiation is scattered in the sample and the transmitted radiation propagates to the diffraction plane 52 on which the diffraction pattern is formed.

[0030] Figure 6 illustrates this process in more detail. Radiation 50 is coarsely focused, for example with a weak lens, thereby illuminating the area of ​​the first surface of the target. The weak lens may of course be any suitable focusing device, such as a set of plates and a power source generating an electron beam or an X-ray reflecting surface. Poor focusing is sufficient to limit the probe beam. Thus, it is not necessary to focus the radiation sharply, although of course it is possible to use highly focused radiation. In this case, the target object provides the O (r) function representing the phase and amplitude changes introduced to the incident wave as a result of passing through the object of interest. The irradiation incident on the target object represents the probe function P (r) creating the irradiation function such as that generated by the caustic profile or the irradiation profile created by the lens or other optical element. P (r) is the complex stationary value of this wave field calculated in the plane of the object. The output wave function Ψ (^ R) 63 determines the scattered radiation when it leaves the distal surface of the target. During propagation in space, the output wave will create a diffraction pattern Ψ (ΚΚ) 64 in the diffraction plane 33. It should be understood that in the case of the embodiment with the molded gap shown in Figure 4 and the gapless embodiment described with reference to Figure 6, if the diffraction plane in which the scattered radiation is detected moves closer to the sample, then instead of Fourier diffraction patterns detected there will be Fresnel diffraction patterns. In this case, the propagation function from the output function Ψ (^ Κ) to the diffraction pattern Ψ (ΚΚ) will be a Fresnel transform rather than a Fourier transform.

[0031] Figure 7 shows an algorithm for obtaining the wave function of an object and thus obtaining image data, which can then be used to create high resolution images of the object. Figure 7 shows one possible way of using the first embodiment of the present invention illustrated in Figures 3 and 4 and moving the gap, after measuring the diffraction pattern, from the first position to the second position in which a second, suitable diffraction pattern can be measured. It should be understood that embodiments of the present invention may use one or more slot positions. The embodiments of figures 5 and 6 can also be used, instead of shifting the gap, a location can be chosen where the poorly focused radiation falls on the sample.

[0032] As noted above, O (r) and P (r) represent two-dimensional complex functions, meaning that each point of the function O (r) or P (r), where r is a two-dimensional coordinate, has a complex value assigned . In the further part of the text O (r) will physically represent the output wave function that would emanate from the function of an object exposed by a flat wave. For example, in the case of electron scattering, the O (r) function would represent phase and amplitude changes introduced into the incident wave as it passes through the object of interest.

[0033] Hereinafter, P (r) represents either an exposure function, such as a function generated by a caustic profile or an exposure created by a lens or other optical element (such as shown in Figures 5 and 6, it being understood that P (r) is the complex stationary value of this wave field calculated in the plane of the object's function) or a filtering function such as a slot or transmission grid mounted after the object's function (as shown in figures 3 and 4).

[0034] Hereinafter, it can be assumed that the functions O (r) or P (r) can be shifted relative to each other by relative distances R. The adopted terminology has been written as a moving function P (r), although alternatively the function can also be moved O (r) relative to P (r). In both situations, the complex value of the O (r) function is changed by creating the product of O (r) and P (rR) to obtain the total output wave function ψ (^, i.e.

ψ (r, R) = 0 {r) P (rR) j

This relationship will generally be met. It is worth noting that there are very few practical restrictions imposed on the function of an object or on the function of a probe / slit. No function can be a flat or periodic wave with a period being a multiple of the difference between the different values ​​for R. These conditions result from the fact that the algorithm requires several different measurements for proper operation. In experimental practice, these criteria are easily met.

[0035] The algorithm works in such a way that it searches for the phase and intensity of the complex function ψ (Γ ^). It requires, as input, knowledge of the P (rR) function, and one or more (preferably several) measurements of the intensity of the wave function in a plane other than the plane in which the sample is located. A convenient solution is to use the diffraction plane associated with the plane in which the sample is located by Fourier transformation. In this case, the measured input data are the intensities of the diffraction patterns at one or more probe / gap positions. The use of diffraction data has several advantages, including ease of collection, no requirements for focusing the output wave function to the image, and an increase in resolution obtained by measuring data for a high angle value.

[0036] It is also possible, however, to implement an algorithm based on a set of unfocused images measured at a distance from the sample / gap outlet surface. In this situation, the free space propagator for Fourier transformation is replaced.

[0037] This algorithm is not limited to the use of the two mentioned transformations. It is possible to use other effective transformations that enable moving from one data plane to another. Below, the general transformation T refers to the transformation that transforms the wave function from the first plane, called plane 1, to the second plane, called plane 2.

[0038] The algorithm works as follows and with reference to figure 7:

1. The algorithm begins at step S700 with the guessed function of the object Og, n (r), where the index g, n represents the guessed wave function in the nth iteration of the algorithm. These functions occur in plane 1 (which is the real plane when using Fourier transform). Preferably, the first guessed function Og, n (r) has a value equal to one for all points of r. This corresponds to a situation in which the sample is absent.

2. At step S701, a slot known in terms of location and characteristics is selected. Its selection provides the P (rR) probe function. At step 702, the currently guessed object function is multiplied by the slot or probe at the current position R, P (rR). The result of this operation gives a guessed output wave function (still in plane 1) for the position R, ψi⁼ gg_(I1 (r) P (σ'-R) ..., 2

3. Then, in step S703, the transformation of the ip function is performed_{g> n}(r, R) to obtain the appropriate wave function in plane 2 (which would be the plane of the diffraction space if Fourier transforms are used), for this position R. In this case, T represents a general transformation that would often be a Fourier transform, but it may also be a propagator of Fresnel free space, or other transformation appropriate for this particular application of the algorithm.

T_g, "(K, R) = Wg, n (r, R)] ₃ k is the coordinate in plane 2. (in the case of Fourier transformation, k would be the typical inverse space coordinate). In the case of the propagator, k would be the xy coordinate in the focal plane.) It is worth noting that Ψ_9ιΠ(Κ ^) is the "guessed" version of the actual wave function in plane 2, because it arose from the guessed function of object O_g,_n(R). Subsequent iterations of the algorithm will give increasingly accurate versions of the function Ψ_αη(^).

It should be noted that the function Ψ_{9> η}(^ R) can be written as:

Ψ_9ιΙ1 (k, R) = | Ψ_9ιη (k, R) | e '⁰⁹'^{n [k, E>} ... 4 where ^_g,_n(K, R) | is the (guessed) amplitude of the wave function, and 0_{g> n}(k, R) is the (guessed) phase in space 2 for iteration n, for position R.

Based on the measurement of the diffraction pattern intensity using known techniques, such as the detector network 32, data on the actual transformed output wave function is known. The measured intensity of the diffraction pattern at the place where the gap is in the first position is therefore the basis for estimating the complex wave function of the diffraction pattern. The measured intensity value, however, does not provide information about the phase of the wave function. The measured intensity is more comparable to the square of the module Ψ (Γ). This quantity is | Ψ (Γ) |². Once the radiation intensity has been found in the two-plane diffraction pattern, the next step can be performed in step S704.

4. Correctly, in step S705 the intensities of the guessed wave function of plane 2 are determined, to known values.

Ψ ^ Λ) = | Ψ0ί ^) | β^ίθ®·^{τ (ΜΙ)} ... 5 where ^ (k, R) | is a known module of plane 2. It is the square root of the intensity measured in the image plane.

5. Inverse transformation of S706 back into real space allows obtaining a new and better guessed output wave function (in plane 1) (T^-1is the inverse of the previously used transform T),

6. In step S707, the guessed wave function of the object is updated in the area covered by the slit or function, using the update function

<img file="PL1740975T3_D0001.tif" />

where parameters β, δ and l are properly selected, and | P_max(RR) | is the maximum amplitude value P (r). The result is a new, guessed goal function (S708).

[0039] The updating function helps to make effective, possible deconvolution and introduces a weighting factor that causes the strongest updating of the object function in the area where the probe function has the greatest amplitude. The value of the variable l can be selected, and it can be set equal to 1. Its value can be chosen freely in the range of values ​​from 0 to 3 and it does not have to be an integer. A useful solution is to choose l&gt; 1 in the case of significant noise. The value of l can be selected as l &lt;1 if, due to the scattering geometry, the detected intensity is in the form of a Gabor hologram or similar. The value of δ is used to prevent the occurrence of a division by zero error when | P (r - R) | = 0. δ is a small real number commonly used in Weiner filters, and it is generally (although not necessarily) smaller than Pmax and it can be significantly smaller in a situation where the noise present in the recorded data is low. The β constant controls the amount of feedback in the algorithm and can be advantageously changed in a range of values ​​from about 0.1 to 1. For a value of β = less than 0.5, the previous object estimate is considered more important than the new estimate. Intermediate values ​​change the relative relative importance of both estimates. β is the parameter determining the speed of achieving the solution.

[0040] δ is a parameter that can have a fixed, defined or variable value. It indicates the level of noise in the data and is used to suppress the update carried out in response to the conditions found. If there is good data collection conditions, which means high beam current (high flow rate), which in turn would mean low noise, then it is safer to use the collected results to update the guessed estimate. This in turn means that the value of δ can be a small fraction of the Pmax value (e.g. less than 1/10).

[0041] Expression:

| P (r - R) | ^ ₈ | P_{It has}x (rR) ^ maximizes the impact of updating areas where the value of | P (r - R) | is big. This solution is useful because it is these areas that receive the largest amount of incident radiation, and thus they contain information about a relatively high signal-to-noise ratio. Such information is of course much more valuable than information obtained in areas where the amount of incident radiation is small and where noise is significantly affected.

[0042] In a situation where β = 1, 1 = 0 and δ = 0 and the function P (rR) is a mask that can be represented by an area where its value is equal to one, while it takes values ​​equal to zero outside this area, or auxiliary function, the algorithm has some similarity to the well-known Fienup algorithm. If in this situation only one R position is used, then the algorithm is reduced to a mathematically identical form to the basic Fienup algorithm. When more than one R position is used, this algorithm offers significant advantages over known methods, including the fact that it does not create uniqueness problems, and a wider field of view is possible.

[0043] After updating the current estimation of the guessed function, the algorithm shown in figure 7 proceeds to the selection of a new position R, which at least partly coincides with the previous position. The degree of coverage should preferably be more than 20%, preferably 50% or more. Such coverage can be achieved either by shifting the slot in the direction of arrow A shown in figure 3 by a certain distance or by ensuring that the irradiation radiation shown in figure 5 falls on another area of ​​the target object. It should be understood that embodiments of the present invention can successfully provide image data also for one location on the target without any change in the position of the gap or incident radiation. In such embodiments, after performing step S708, the algorithm returns to step S702. Instead of loading the initial estimate of the objective function O (r), the new guessed function O (r) obtained in step S708 is loaded. In each iteration, the newly guessed objective function will provide an ever better approximation to the actual function of the object, because in each iteration information about the known intensity is added, and thus the known amplitude component of the incident radiation in order to increase the accuracy of the estimation.

[0044] However, a more preferred method is to move to a new position R which at least partially coincides with the previous position as shown in figure 7.

[0045] The known probe function P (r-R2) in the second position is determined in step S709 and then the steps are repeated as described above, whereby the newly guessed function generated in step S708 is multiplied by the new known probe function determined in step S709. This solution is illustrated in step S710. It provides efficient generation of the output wave function either after the sample or behind the gap, depending on the particular embodiment of the invention. The resulting waveform output function is propagated in step S711 to provide an estimate of the scattering pattern that should be detected at this position. The diffraction pattern is measured in step S712, providing information about the intensity and thus information about the amplitude regarding the transformed wave function. The intensity information is used to correct the amplitude of the transformed wave function, while the phase information is saved in step S713. This corrected wave function is inversely propagated by Fourier transformation (when the image is formed in a distant field), Fresnel transformation when the image is created in a place where Fresnel diffraction dominates, or by any other suitable transformation. This solution is illustrated in step S714. The current estimate O (r) is then corrected according to the update function presented above in step S715, and the result of such correction is the newly guessed object function illustrated in step S716.

[0046] At this stage, it is possible to introduce an additional movement of the irradiation radiation or slots into the third or subsequent position. Again, a position where at least partial overlapping of previously exposed positions occurs is preferred. In this way, it becomes optional to map the entire target. An alternative is to repeat the newly guessed function generated in step S716, without further position changes, based on knowing the results for a known diffraction pattern. Figure 7 shows the iterative method repeated by returning to step S702, in which the newly guessed function generated in step S716 is used in the multiplication step instead of the initial estimate of the object function provided in step S700.

[0047] The iterative method can be repeated until a specific event occurs. For example, iterations can be repeated a certain number of times, e.g. 1000 times, or until the sum of error squares (SSE) reaches a sufficiently small value. SSE is the value measured in plane 2 as

SSE = (| ψ ^_η (k, R) [²-] ψ (k, R),²) ²₉

N where N is the number of pixels in the series representing the wave function. [0048] During the iterative process, the most current guessed object function provides a current estimate of this object function. After the iterative process is completed based on the occurrence of a specific event, the current estimation of the objective function provides image data for places that are either irradiated with incident radiation or which are selected based on the location of the crack behind the target. Such image data contains amplitude and phase information that can then be used to generate a high resolution image of a selected area of ​​the target.

[0049] The mobile probe algorithm was used to recover the phase of the wave function formed when the STEM probe strikes the object with the transmission function shown in the first line. This transmission was created by taking a picture of gold particles on a background in the form of amorphous carbon measured using a CDD camera, treating this image only as an amplitude object and propagating the result and 1000A to obtain the intensities and phases shown.

[0050] The STEM probe has a gap size of 0.25A^-1, 3000A focal length, and a total count of 1.0 x 10⁶. These parameters lead to the intensity and phase shown in Figure 8. This probe is multiplied by the object transmission function for various probe positions in a 128 x 128 pixel matrix. The obtained wave functions are subjected to Fourier transform to obtain diffraction patterns such as the pattern shown in the third row in the figure being the diffraction pattern for the position of the sample (60, 40).

[0051] The algorithm was started, using the parameters of equation 7 with values ​​β = 1.1 = 1 and δ = 0.0001, for 2000 iterations, at which point the SSE value in the diffraction space was 1.444 x 10^-7 and continued to decrease quickly. The wave function recovered at this stage is shown in the fourth line of Figure 8. It is obvious that this algorithm works very well. The same experiment was repeated twice, including the added Poisson noise, with an average value of 1.0 first and then 5.0. The β value was changed to β = 0.6, which provided better convergence of the algorithm. The result of the simulations carried out is shown in Figure 9. It is clear that the addition of noise affected the algorithm. It can be clearly seen that the object transmission function is effectively recovered only near the set of probe positions used for such recovery. This is expected behavior because very little is known about the object in areas where the probe intensity is negligible. The results presented are scaled to the same gray scale as the original transmission function of the object. It is clearly seen that the structure and some details of the object have been recovered, even when the noise level is relatively high.

[0052] Embodiments of the present invention therefore provide a new phase recovery method that can be applied to many situations in microscopy, with particular emphasis being placed on its applicability in scanning transmission electron microscopes. This method requires as input information only the intensity of measurements for a small number (one or more) of different positions of the probe or gap, thus removing the need for a lens placed behind the sample, which avoids problems associated with the aberration of such lenses. The algorithm used shows fast convergence until the phase of the object transmission function is recovered. This in turn enables the generation of high resolution images illustrating the structure of targets in real time. This algorithm is also effective in situations of noise and works well for a very wide range of different objects and probe functions. Embodiments of the present invention also allow for the calculation of probe functions when using targets having a specific structure.

[0053] Figures 10A, 10B and 10C show alternative embodiments of the present invention, and in particular show how the radiation can illuminate the target and how the data identifying aspects of the radiation scattered by the target can be detected. Figure 10A illustrates how a radiation source 1000 that can be positioned near target 1001 can be used to create a scattering pattern on a series of detectors 1002. The source should be close enough to the target to ensure that the irradiated target surface the object is small enough to meet the Nyquist sampling condition in the detector plane. This embodiment does not require the use of slits or lenses to obtain image data that can be used to generate high resolution images of the 1001 target. However, in these specific conditions, it is important to place the 1000 source close enough in front of the 1001's front surface to achieve such high resolution . The sample or source may be mobile to provide more than one location necessary for the upgrade process.

[0054] Figure 10B illustrates another embodiment of the present invention in which the source of the plane radiation 1003 falls on the focusing tube 1004. Tube 1004 selects a radiation region 1003 that can leave the tube and which becomes radiation incident on the target 1001. According to this specific example it is not necessary to use lenses or gaps behind the target.

[0055] Figure 10C illustrates yet another embodiment in which the point radiation source 1000 emits radiation incident on a slightly angled mirror 1005 or other reflecting surface. Such a reflecting surface is used in particular when the point source 1000 is an x-ray source. The radiation is reflected by the 1005 mirror at an angle close to the complementary angle and it hits the target 1001. The scattered radiation is again detected by a series of detectors 1002.

[0056] Figures 11A and 11B show yet other embodiments of the present invention. In particular, they show how embodiments of the present invention can be used to provide a surface profilometer. Point source 1100 emits radiation incident to the surface of the target sample 1101. Instead of transmitting radiation through the sample as described above, in this case we are dealing with a totally or partially reflective sample. The protuberances and other surface characteristics change the phase of the incident radiation, and this radiation, reflected from the target, is dispersed towards a series of detectors 1102 on which the pattern as described above can be detected. Figure 11B shows an alternative embodiment of a surface profilometer in which radiation from surface 1100 is first focused by lens 1103 before interacting with target 1101. It should be understood that the algorithm described above can equally well be applied to the transmittance mode described in the previous part of the text, and to the embodiments using the reflection phenomenon described with reference to figure 11. For each of the various embodiments described in Figures 10 and 11, source motion 1000, 1100, target 1001, tube 1004 and / or reflecting mirror 1005 can be used to reposition the irradiation function or probe function to be used in the next iteration described above algorithm.

[0057] Figure 12 shows an image data providing device that can be used to construct a high resolution image of the target area of ​​the object according to the embodiment described in Figures 5 and 6 above. The radiation source 1200 provides irradiation radiation to lens 1201 which poorly focuses radiation on a selected target area 51. Incident radiation has an incident wave function 1202 and an outgoing wave function 1203. This outgoing wave function propagates over a distance D in which a diffraction pattern is formed on a series of detectors 1204. The distance D is preferably large enough that the propagated output wave function 1203 forms a Fourier diffraction pattern in the far field. A series of detectors provides at least one detector that can detect the intensity of radiation scattered by the target object 51. A locating device 1205 is provided, which may be a micromotor and which can locate the target object in one or more positions relative to the target object as desired. In this way, it is possible to ensure that the radiation from the source 1200 falls into various places on the front surface of target 51.

[0058] Control assembly 1206 provides micromotor control signals and further receives intensity measurement results from each pixel detector in the array of detectors 1204. Control assembly 1206 has a microprocessor 1207 and a data carrier 1208 connected to a user interface 1209, which may include a user screen and a keyboard enabling entering user data. The control unit can be connected to another data processing device, such as a 1210 laptop or personal computer, providing remote control. Alternatively, it may also be understood that the control assembly 1206 may also be provided by a laptop or personal computer. Control assembly 1026 can automatically control the process of generating image data in real time. Alternatively, the user may select the target surfaces to be imaged or enter additional user input using the 1209 user interface.

[0059] In use, the radiation source 1200 irradiates the lens 1200 with radiation. Target 1200 is selectively positioned after actuator 1205 under control of control unit 1206. Radiation creates a diffraction pattern, detected in the appropriate positions by each of the detectors forming a series of detectors 1204. The results obtained by such detectors are entered as input to the control unit and can be recorded on a data carrier 1208. If only one location is used to acquire the image data, the microprocessor uses the detected information along with program instructions containing information on the algorithm described above to acquire the image data. However, if one or more subsequent positions are required before finalizing the image data, the control unit then sends a signal to the 1205 actuator, which places the sample in the next selected position. Actuator 1205 can place the sample in one of many different positions. After the sample has been moved, another diffraction pattern created on a series of detectors is measured, and the measured result is saved in the control unit. The 1204 series can be, for example, a CCD series of 1200 x 1200 pixels. If no further intensity measurements are required, the image data can be generated at this stage by the control unit in accordance with the two newly saved result sets using the algorithm described above. On the user interface 1209 or on a remote screen of a personal computer, or other such device, raw image data or a high resolution image generated from the image data may be displayed.

[0060] Embodiments of the present invention thus provide an iterative method of acquiring target image data. The iterative method is used in an intelligent way, thanks to which it can be used in generalized exposure systems. In such systems, the slot transmittance function is poorly defined or the radiation beam may be poorly focused.

[0061] Embodiments of the present invention provide a method of acquiring image data suitable for the purposes of subsequent generation of a high resolution image or portion of a sample with a resolution limited by the wavelength of the radiation. In this way, a method and device are provided that can produce image data having a much higher resolution than the resolution required for the positioning accuracy of the device used to obtain this information. For very short wavelength radiation (on a subatomic scale), it is possible to achieve a 40-fold or greater improvement in resolution compared to techniques known in the art. In some cases, the resolution will be subject to the limiting effects of the movement of the atoms themselves.