Semiconductor technology apparatus and method

Abstract

The invention relates to an apparatus (1) for semiconductor lithography for measuring and / or registering structures on a photomask (7), comprising an illumination system having a beam source (2) for illuminating the photomask (7) with measurement light (12) and comprising an objective (8) for imaging the structures onto a recording device (10) In this context, the apparatus is configured to generate an image representation of the photomask (7) with a resolution of better than 74 nm, preferably less than 68 nm and particularly preferably less than 60 nm half-pitch. The invention also relates to a method for correcting illumination errors of the illumination when measuring and / or registering at least one structure on a photomask (7) for semiconductor lithography, including the following method steps: - measuring the illumination, - ascertaining the illumination error of the illumination, - correcting the ascertained illumination error when registering at least one structure.

Description

[0001] Semiconductor technology apparatus and method

[0002] It is hereby declared that the content of the priority-establishing application DE 10 2025 102280.1 dated January 22, 2025 and the parallel patent applications with the official application numbers 102025 102245.3, 102025102252.6 and

[0003] 10 2025 102260.7, filed by the same applicant on 22 January 2025, is incorporated into the present application in full. This comprises not only the description, claims and drawings but also all technical features and exemplary embodiments disclosed in these applications.

[0004] In particular, all features of the aforementioned applications are intended to be regarded as disclosed in the present application, both on an individual basis and in any desired combinations. This also applies to combinations of features from different applications.

[0005] The invention relates to a semiconductor technology apparatus, in particular a mask inspection apparatus. The invention also relates to a method for improving the performance of such an apparatus.

[0006] In the production of semiconductor components, it is customary to optically expose semiconductor substrates 30 to 50 times in what is known as a projection exposure apparatus. What are known as photomasks, whose structures are imaged onto a photoresist on the substrate generally with a reduction by a factor of 4, serve as exposure templates. These structures are introduced into the photoresist during each exposure and transferred into the substrate by subsequent chemical processes. As a result, multiple layers of electrical lines, transistors or other elements lying one above the other are formed in the substrate.

[0007] A particular requirement in this case is that conductor tracks in different layers that are intended to be connected to each other are assigned in a spatially correct manner; this is referred to as an overlay, i.e. an overlap accuracy. In advanced components, the overlay should be in the range from approx. 1 to 3 nm. It should be observed that overlay errors add up over the entire manufacturing process, and sothe overlay error originating from the mask is allowed to use up only a small part of this 1 to 3 nm error budget. Mask metrology, i.e. in particular mask inspection systems for ascertaining the pose and shape of structures on masks, may in turn use up only a small part thereof.

[0008] The structures on the photomasks must be arranged and formed correctly in order to attain the required overlay. The structures must thus observe a predetermined line width or structure size (within a given tolerance of generally 7% or 10%). In addition, all structures must be precisely where they should be. A distortion or position deviation of the structures from the target position must be less than e.g. 1 nm. This required alignment accuracy is called registration.

[0009] Photomask registration is typically measured using a mask inspection system that is designed as a registration measuring system. Current systems measure the position of selected structures on the mask relative to other structures or to reference structures on the mask, the so-called alignment markers, with an accuracy of better than 0.2 nm. Considering that a typical photomask has an edge length of 152 mm, it becomes clear how challenging it is to measure distance with an accuracy of 0.2 nm at a distance of 152 mm.

[0010] The resolving power of the mask inspection systems is usually specified by a value referred to as half-pitch, which corresponds to half the periodicity of the smallest resolvable structures.

[0011] In the past, it was conventional for the registration to be measured at so-called registration markers. That is to say, specific registration markers are applied to the photomask, e.g. in a grid of 5 or 10 mm, and the registration tool measures the positional accuracy of these markers.

[0012] The demands on the overlay are becoming more stringent in current applications. The registration is therefore measured using a denser grid or even extensively over a continuous area of e.g. 1 x 1 mm. However, it is not possible to place an unlimited number of markers on the mask - this would take up too much space required for the actual used structures.For this reason, markers are increasingly dispensed with, and the actual used structures are used for the registration measurement instead; this is also referred to as indie registration.

[0013] However, the structures on the mask are becoming ever smaller in the case of EUV masks, down to a half-pitch of 50 nm or even 32 nm. The resolution of known systems is not sufficient for this purpose. A higher resolution is required in order to be able to perform in-die registrations for such small structures as well.

[0014] The problem addressed by the present invention is that of specifying an apparatus and a method which can be used to effectively address the aforementioned challenges.

[0015] This problem is solved by a device and a method having the features of the independent claims. The dependent claims relate to advantageous developments and variants of the invention.

[0016] An apparatus according to the invention for semiconductor lithography for measuring and / or registering structures on a photomask comprises a beam source for illuminating the photomask with measurement light and an objective for imaging the structures onto a recording device. According to the invention, the apparatus is configured to generate an image representation of the photomask with a resolution of better than 74 nm, preferably less than 68 nm and particularly preferably less than 60 nm halfpitch.

[0017] In particular, the photomask can be an EUV mask, i.e. a mask intended to be imaged onto a wafer with EUV light in a projection exposure apparatus.

[0018] In this case, the measurement light may have a wavelength shorter than 200 nm, in particular 193 nm, 157 nm or 13.5 nm (EUV). In the latter case, the use of mirror optics makes sense; conventional optical elements such as lens elements lose their effect or become opaque in that range.

[0019] In an advantageous embodiment of the invention, an immersion element is arranged between the last optical element of the objective and the structures to be imaged. Within the meaning of the present application, an immersion element is an elementwith a refractive index greater than 1 , by means of which the resolving power of the associated optical system is increased on account of the accompanying increase in the numerical aperture (NA). In this case, the immersion element may be in direct contact with the last optical element in the objective and a structure to be imaged. In particular, the immersion element may be a liquid volume of an immersion liquid. The immersion liquid can be water, in particular high-purity water. At a wavelength of 193 nm, water has a refractive index of 1.44, and so this can increase the NA - e.g. to 1.2, 1.3 or 1.35.

[0020] In a variant of the invention, the liquid volume may be formed in such a spatially limited manner in the region between the last optical element of the objective and the structures to be imaged that the photomask is covered only in regions. In other words, only a droplet of immersion liquid may be arranged between the last optical element of the objective and the mask.

[0021] For better conveyance of this droplet in the case of a relative movement between the objective and the photomask, the surface of the photomask may be of a hydrophobic form at least in regions.

[0022] For this purpose, the photomask may be provided with a hydrophobic coating at least in regions.

[0023] As a result of a last optical element of the objective in the direction of the photomask being of a hydrophilic form at least in regions on the surface which faces the photomask, the droplet of immersion liquid can be formed with improved stability. In particular, this variant - especially in conjunction with a hydrophobic surface of the photomask - facilitates conveyance of the droplet when approaching a new measurement position on the photomask. As an alternative to conveyance, the droplet may also be prepared at an envisaged measurement position by way of suitable means. Once the preparation of the droplet has been completed, the objective may be moved to the measurement position, and contact with the immersion liquid may be established.ln an embodiment of the invention, a supply and a drain for the immersion liquid may be present. This can ensure a constant flow of the immersion liquid such that potential particles do not dwell for a relatively long period of time in the region passed by the measurement radiation. Likewise, local heating of the immersion liquid and the associated problem of undesired evaporative cooling can be avoided in this manner. The drain may at least partially surround the liquid volume; to this end, the drain may be embodied in the form of a ring segment.

[0024] The liquid volume may completely cover the photomask in a further variant of the invention. To this end, the liquid volume may be formed for example in a liquid located in a tank.

[0025] Means for moving the tank in the vertical z-direction may be present in order to be able to immerse the photomask into the immersion liquid.

[0026] Furthermore, means for moving the tank horizontally in the x- and / or y-directions may be present.

[0027] In addition to that or in an alternative, a pumping device for filling and an outlet for emptying the tank may be present.

[0028] Wave-damping elements for damping liquid movements in the tank may be present in order to prevent unwanted image errors due to a moving immersion liquid.

[0029] It is advantageous for a last optical element of the objective in the direction of the photomask to be formed flat or with a comparatively large radius of curvature, in particular with a radius of curvature of greater than 500 mm, preferably of greater than 1000 m, on the side which faces the mask. It is also advantageous to form the last optical element from quartz glass; the use of CaF2should be avoided as far as possible on account of the solubility of the latter.

[0030] The last optical element may be designed to be interchangeable for various applications, especially also for changing from immersion to dry operation of the apparatus.The temperature of the immersion liquid is advantageously chosen in such a way that potential temperature differences between the immersion liquid and the photomask are kept as low as possible in order to reduce or prevent thermally induced deformations of the photomask. In particular, the effect whereby evaporating immersion liquid leads to unwanted evaporative cooling in the region of the mask under consideration must be taken into account. To this end, the presence of means for setting the temperature of the immersion liquid in a mK range is advantageous. In particular, the temperature of the immersion liquid may be chosen to be somewhat higher than the mask temperature with a slight offset so as to compensate for the described effect of evaporative cooling. In addition to that or in an alternative, a somewhat elevated pressure may also be set in the region of the immersion liquid in order to reduce or prevent evaporation of some of the immersion liquid. In this context, this somewhat elevated pressure may be set only locally, but in an alternative it may also be set over the entire apparatus. In addition, a desired temperature of the mask and / or of the immersion liquid may also be set by providing a suitably chosen inflowing gas stream.

[0031] It is also advantageous if means for assessing the quality of the immersion liquid or the quality of cleaning of the mask surface of immersion liquid after the measurement are present. In particular, it may be expedient to take an image of an autofocus grid on the mask surface after the measurement in order to check the cleaning quality. Such autofocus grids are obliquely positioned image representations of grid structures which are produced in the plane of the mask and which contain information relating to the pose of the mask relative to a focal plane of the apparatus on account of their contrast profile. Remaining residues of immersion liquid in the region under consideration would lead to deviations from the expected images here. The assessment of the quality of the immersion liquid may make use of systems capable of taking optical measurements, for example scattered light measurements by means of a laser beam or else image representations, from outside of the actual measurement beam path, for example from the side (parallel to the mask surface); in particular, the use of a line-scan camera is also conceivable.Furthermore, it may be advantageous if means are present for drawing inferences about the state of the immersion liquid by means of imaging at a wavelength significantly different from the customary measurement wavelength, typically at a significantly longer wavelength. At longer wavelengths, the influence of the mask structure on the obtained image is reduced, and so the image primarily contains information about the state of the immersion liquid present between the objective and the mask surface.

[0032] The raised structures located on the photomask and the recesses connected therewith also entail the risk of the immersion liquid not penetrating completely into the recesses and hence of pockets of gas or cavities which falsify the obtained image remaining between the immersion liquid. Then again, it is also conceivable that the additional gas / liquid interfaces arising as a result lead to an improvement in the imaging, for example as a result of the total-internal reflection associated therewith. One option for avoiding such gas inclusions between the mask surface and immersion liquid consists in creating a gas atmosphere made of a gas which is readily soluble in the immersion liquid such that the gas present in the gas inclusion completely dissolves in the immersion liquid, and the inclusion disappears as a result. For example, CO2lends itself as a gas if water is used as an immersion liquid.

[0033] The creation of mechanical contact between the objective and the mask when an immersion liquid is used may lead to the mask, which of course is held in its lateral position purely on account of its weight and the static friction at the contact points associated therewith, suffering from a certain amount of lateral offset in the case of a lateral displacement between the individual measurement positions. Firstly, this offset may be measured; then again, it may also be suppressed. In particular contactless methods, i.e. , for example, capacitive or optical methods such as interferometry, come into consideration for a measurement. There is also the option of preventing lateral slipping of the mask by means of weak opposing forces (e.g. due to soft elastic elements) or by means of clamping the mask with a small clamping force. In this case, the forces should be chosen such that deformation of the mask is avoided as far as possible.The immersion element may be a solid-state element in a further embodiment of the invention. In this case, the effect that is exploited is that the evanescent field of the immersion element interacts contactlessly with the structures on the mask given a sufficiently small distance between the immersion element and the structures on the photomask. Moreover, increasing the numerical aperture beyond the options afforded by water is conceivable when solid-state elements are used.

[0034] It is advantageous if the solid-state element is a refractive element with a refractive index greater than 1.5, preferably greater than 1.8.

[0035] It is further advantageous if the apparatus is configured to bring the immersion element up to a distance of less than 50 nm, preferably of less than 20 nm and further preferably of less than 10 nm, from a surface of the photomask.

[0036] In this case, it is advantageous for the apparatus to comprise a distance control device for setting the distance between the immersion element and the surface of the photomask. Distance control is essential because, firstly, the photomask is not perfectly flat, and, moreover, it comprises structures with a structure height in the range from 50 nm to 100 nm. It may be necessary to measure the distance with an accuracy of 1 nm. For example, this may be achieved by virtue of the distance control device comprising electrical sensors, in particular capacitive sensors.

[0037] In addition to that or in an alternative, the distance control device may be configured to determine the distance between the immersion element and the surface of the photomask from the reflection behaviour of the immersion element. In this case, the effect that is exploited is that total-internal reflection at the side of the immersion element which faces the photomask collapses as soon as the evanescent field interacts with the mask structures.

[0038] What can be achieved by virtue of the immersion element being uneven on the side which faces the photomask is that the risk of contact and hence of mechanical damage is reduced. For example, the immersion element may have a projection on the side which faces the photomask; in this case, the projection may in particular have a lateral extent of approx. 100 pm x 100 pm to 1 mm x 1 mm, which includes the recorded region including a certain safety margin. The aforementioned approach of the im-mersion element to the photomask up to small distances in the low two-digit nanometre range then only occurs in the region of the projection.

[0039] Typically, a measurement position would initially be approached to a safe distance of the immersion element of approx. 1 mm. Then, the immersion element is made to approach the mask surface, typically by a movement of the mask holder in the z-direc-tion. This is followed by the measurement and a return to the safe distance in order to prepare for the approach of the next measurement position.

[0040] It is also conceivable to design the immersion element in the form of a planar element of a highly viscous medium. In this case, the planar element would cover several or all of the regions of the photomask to be examined. In that case, the objective would be brought into contact with the immersion element at the respective measurement positions to be examined. Given a suitable choice of the nature of the immersion element or of the immersion element / photomask material pairing, said immersion element could also be placed onto the photomask in the manner of a soft mat and lifted off again after the measurement.

[0041] In cases in which the apparatus can be switched from immersion operation to operation without an immersion element, the removal of the immersion element causes a change in the refractive index in the region of the beam path previously occupied by the immersion element. This changes the properties of the optics, in particular the spherical aberration and higher-order spherical aberrations. In order to compensate for this effect, an additional optical element can be introduced into or removed from the beam path at a different location in the beam path. In addition to that or in an alternative, it is also possible to modify air-gap distances between the optical elements (lens elements) to a predetermined extent by way of actuators that operate parallel to the beam direction, and so the aberrations are compensated for again.

[0042] It goes without saying that the embodiments described above are also advantageous for systems that provide a lower resolution than that mentioned above.

[0043] In addition to that or in an alternative, the optical performance of the apparatus may be increased by virtue of the illumination system being configured to generate a structured illumination.A structured illumination can be structured in an angular space or in a spatial space. Structured in the angular space means that the measuring light of the illumination in the pupil has a spatial distribution which, in particular, masks the central area of the pupil, for example. This area in the center of the pupil, also known as the on-axis area, can be at least 50% of the pupil radius, preferably at least 60% of the pupil radius, and particularly preferably at least 70% of the pupil radius. The outer 50%, 40%, and 30% of the pupil thus correspond to the off-axis illumination area. At least partial illumination of the pupil in the peripheral area causes the measuring light to strike the object at an angle in the field. The closer the measuring light is to the edge of the pupil, the greater the angle in the field.

[0044] In particular, the illumination can be configured as an off-axis illumination. An off-axis illumination allows the contrast to be increased by increasing the proportion of the + / -1st order of diffraction that enters the entrance pupil of the imaging optics unit. This effect can be traced back to the change in the angle of the 1 st or -1 st order of diffraction on account of the oblique incidence of the illumination light. As a result, more contrast can be generated for the same half-pitch of the structures to be measured since the smaller diffraction angle increases the component of the + / -1 st order of diffraction that enters the imaging optics unit.

[0045] The apparatus may comprise a device for setting an illumination setting in a further embodiment of the invention. In particular, setting refers to the shape of the illumination of the illumination pupil, which makes up a substantial part of the illumination setting in addition to the intensity distribution.

[0046] In addition to the intensity of the radiation, an illumination setting in particular also comprises the shape of the illumination of the illumination pupil, which may also comprise only parts of the entire possible illumination pupil. In case of an off-axis illumination, for example, only the peripheral area and even than is not necessarily completely illuminated.

[0047] Furthermore, the device can be arranged in a pupil plane or a conjugated plane thereto.

[0048] Besides the device can be arranged in a field plane or a conjugated plane thereto.In particular the device can be arranged as well in a pupil plane as in a field plane or in a respective conjugated plane.

[0049] The device can contain an amplitude element, like for example a plate with a partial chrome layer and / or a phase element, like for example a quartz plate with binary respectively discrete and / or continuous elevation elements, which are also known as computer generated holograms. An alternative possibility for an amplitude element can also be a multi-mirror-array.

[0050] The illumination of the illumination pupil may be in the form of a dipole in a further embodiment. A dipole is a type of illumination of the illumination pupil which is known from projection exposure apparatuses and which generates an off-axis illumination. The dipole may usually comprise two opposite sections, referred to as poles, of a ring illumination, with the two sections being arranged for maximum contrast in the direction of the periodicity of the structures. For example, a structure with perpendicular lines in the y-direction would be illuminated using a dipole with two sections on the x-axis.

[0051] In addition, the illumination of the illumination pupil may be in the form of a quadrupole. A quadrupole is also a type of off-axis illumination which is known from projection exposure apparatuses. The quadrupole comprises in each case two opposite poles of a ring illumination, with the two pole pairs being arranged offset by 90 degrees to each other. For example, the position of the individual poles may be on the x- and y-axes, i.e. at 0°, 90°, 180°, 270°, or for example arranged with a 45° offset, i.e. at 45°, 135°, 225° and 315°. For example, a quadrupole illumination of the illumination pupil is advantageous in the case of a structure formed as an array of contact holes which have a periodicity both in the x-direction and in the y-direction.

[0052] Furthermore, at least one section of an illumination of the illumination pupil may be embodied in a leaf form. The leaf form is a special shape of the sections cut out of an annular gap as explained further above. The leaf form may be inherently mirror-symmetrical and in particular find use in a dipole, quadrupole or hexapole illumination or in further illumination arrangements with further poles and in each case 2 poles arranged opposite each other. As a result, a complete overlap of the Oth order and the+ / -1 st order of diffraction is possible. As the angle of incidence (or pitch) increases, the + / -1 st order of diffraction migrates mirror-invertedly from the outside into the NA stop of the imaging optics unit (see Figure 11 b / 11 c). As a result of the mirror symmetry, the + / -1 st order of diffraction is 100% within the NA stop of the imaging optics unit in the case of small angles vis-a-vis the usually utilized ring segment section, and both orders of diffraction may overlap 100%.

[0053] The illumination of the illumination pupil may be embodied in the form of sunspots in a further embodiment. A sunspot illumination within the meaning of the invention is an illumination of the illumination pupil in which only individual regions in the illumination pupil are illuminated, in a manner comparable with sunspots, hence the name given to such illumination. This arrangement of the illuminated regions, which initially appears to be random, is determined by the structures of the object to be measured. Especially in the case of structures which have no or only a small number of multiple structure elements that occur periodically next to one another, as may be the case in logic circuits and / or in the case of structures from what is known as inverse lithography, the off-axis illumination is locally adapted to the corresponding structures in order to improve the contrast of the image representation.

[0054] Furthermore, the illumination of the illumination pupil may be embodied in the form of a hexagonal form. This kind of illumination can also be designated as a quadropol illumination corresponding the dipole or quadrupole illumination. In particular, this kind of illumination can be advantageously used for structures to be measured also containing hexagonal geometries.

[0055] In addition, the illumination of the illumination pupil may be embodied in the form of a ring structure. The ring structure can for example be arranged from the edge of the pupil radius in direction to the middle of the pupil up to 50% or 70% or 80% or 90% of the radius of the pupil. A ring structured illumination can in particular be advantageous for structures containing small isolated dots and / or structures oriented in many different directions.

[0056] The best illumination setting for the structure to be measured or registered may be selected on the basis of (rigorous) optical simulations and / or numerical optimization.This may optimize the reproducibility of the registration for the mask structure in the measurement region. For example, the optimal illumination setting may also be selected on the basis of a trained machine learning model.

[0057] In addition, a stop in the apparatus may be designed to be movable in order to set the illumination of the illumination pupil. A stop designed to be movable firstly allows the positioning of the illumination on the mask and / or is also capable of setting the centroid of the illumination, at least in small regions, byway of a suitable selection of the position of the stop.

[0058] In an alternative the stop can be composed of at least to partial stops. In this case the two partial stops can be arranged in a way, that they together generate an illumination setting in the pupil. A first partial stop can thereby for example generate a ring illumination of a designated size. The second partial stop can generate a wind mill wing like illumination in the pupil, for example with two (four or six) wings. In combination a dipole (quadrupole, hexapole) with a height defined by the ring illumination can result. This enables to adjust a greater number of illumination settings with a fewer number of stops.

[0059] Furthermore, the stop (partial stop) may also be removed from the illumination system in a stop (partial stop) swapping system configured to this end and may be replaced by another stop (partial stop) which generates a different illumination of the illumination pupil, i.e. a different illumination setting. The positioning accuracy must be significantly better than the mechanical stop diameter, i.e. preferably better than 0.2 mm, particularly preferably better than 0.02 mm. The open-loop and / or closed-loop control of the position may firstly use the image representations captured during the measurement or registration as position detection. In an alternative, the stop (partial stop) may have a sensor calibrated with respect to a reference in the apparatus.

[0060] The optimum position may also be ascertained by a calibration measurement for each stop (partial stop). This position may be stored and read out and approached again in the case of a renewed change to the already calibrated stop (partial stop). In particular, the position of the stop (partial stop) may be measured directly using the main beam path of the apparatus with the aid of a Bertrand optics unit.It is also possible for multiple stops (partial stop) to be arranged on a component which for example is designed to be positioned in rotating or linear fashion under the illumination system. In this case, a change of stop (partial stop) and stop (partial stop) fine positioning may be effected by the same actuator.

[0061] Furthermore, a zoom optic can be arranged after the stop (partial stops) in the illumination system. As a result, the illumination setting generated by the stop (partial stops) can be varied in size, whereby a greater variability of the adjustable illumination settings can be generated.

[0062] In a further embodiment of the invention, the illumination system may comprise a device with an arrangement of individual illumination surfaces which can be added or switched off. Within the meaning of the invention, adding or switching off may also be realized by blocking or deflecting individual radiation components. Individual radiation components may also be merely partially blocked. Different brightness levels may be realized as a result. In addition, this allows correction / compensation of an inhomogeneity present in the illumination pupil. Ideally, it is always possible to use all of the light provided by the radiation source for the purpose of illuminating the illumination pupil.

[0063] In particular, the illumination system may comprise a multi-mirror array. By tilting the individual mirrors, a multi-mirror array can steer the light from the radiation source onto the regions of the illumination pupil within the scope of the resolution of the array. Thus, there is no loss of light power.

[0064] Alternatively, for example, a fibre bundle capable of fully illuminating the illumination pupil may also be controlled, or the individual fibres thereof may be supplied with light, such that a desired illumination of the illumination pupil, for example a sunspot illumination, can be set. This may be realized firstly by selectively radiating light into the individual fibres or else by way of selectively blocking individual fibres by switching switches in the course of the fibres.

[0065] In the case of a mask inspection system, it may be the case that despite a loss of light that occurs as a result of blocking individual fibres, said loss of light has no ef-fects on the measurement and / or registration of the structure on account of the radiation source that is over-dimensioned in comparison with the required power.

[0066] The apparatus may comprise a control device with a module for machine learning in a further embodiment of the invention. Machine learning is an ever more widespread method which allows a prediction for future measurements with increasingly higher accuracy by making use of measurements already taken. In the present case, it may even be possible to ascertain a very good correction of occurring illumination errors on the basis of input data that are comparatively easy to ascertain. Further details are explained further below.

[0067] A further improvement in the performance of the apparatus according to the invention can be achieved by virtue of the apparatus comprising means for setting the polarization of the measurement light, whereby the image contrast is likewise increased. Especially in cases in which the E-field vectors of the measurement light are aligned parallel to the prevailing object structures, a higher image contrast can be generated than in the case of a polarization orthogonal thereto. This effect becomes ever more pronounced the greater the numerical aperture is, i.e. the greater the angles of incidence are. Therefore, an object-side polarized illumination is worthwhile, especially in combination with an immersion system which uses an immersion element, in particular a solid-state element.

[0068] In particular, it is appropriate in this case to choose the polarization of the measurement light in such a way that the E-field vectors are aligned parallel to the most critical object structures that occur on the mask.

[0069] Furthermore, the use of a tangential polarization is expedient, e.g. in the case of a quadrupole illumination, annular illumination or 4-leaf illumination, i.e. the use of a polarization in which the E-field vector is substantially tangential to the optical axis at each pupil location.

[0070] The means for setting the polarization can contain a manipulator for inserting a polarisation element into the beam path of the apparatus.The means for setting the polarization can additionally contain a manipulator for rotating the polarisation element around an axis parallel to the beam direction.

[0071] The manipulator can either be arranged in the illumination beam path or additionally also in the projection beam path.

[0072] The polarisation elements can contain at least one variable anisotropy. This kind of polarisation elements can be for example designed as a segmented half wave plate, where applicable or necessary with a spatially variable sub wavelength structure, spatially structured polarizing elements or as a so-called wire-grid polarizer. Elements using structures smaller than the wavelength for polarisation of light are also referred to as meta surfaces.

[0073] A method according to the invention for correcting illumination errors of the illumination when measuring and / or registering at least one structure on a photomask for semiconductor lithography comprises the following method steps:

[0074] - measuring the illumination,

[0075] - ascertaining the illumination error of the illumination,

[0076] - ascertaining the influence of the illumination error on the result of the measurement or registration. For example, this may be implemented by a computational simulation or by a comparison of two measurements, one with illumination errors and one without.

[0077] - correcting the measured registration by the ascertained influence of the illumination error.

[0078] Measuring the illumination, in the specific case the position and the centroid of the illumination pupil, allows the registration that has already been carried out by a measurement to be corrected. Alternatively, the ascertained illumination error may already be corrected in the illumination prior to the measurement such that the ascertained illumination errors are already corrected when the structure is measured and / or registered.Furthermore, the illumination can be measured using calibration masks.

[0079] In particular, the calibration masks may comprise calibration structures and reference markers.

[0080] In a further embodiment of the invention, the calibration structures may be measured or registered using different, for example the off-axis illumination and / or using an on-axis illumination. The calibration structures of the calibration masks may have at least a similar embodiment to the structures used during the actual registration measurement, said similarity relating to the half-pitch and the type of structures, e.g. contact holes or lines.

[0081] The calibration structures may also be measured twice, whereby the structures are measured with at least two illumination settings, for example once with the off-axis illumination and once with the full illumination.

[0082] Furthermore, the deviation of the ascertained positions of the measurements or registrations may be determined using at least one of the measurements with off-axis and with full illumination and may be stored. The deviation may be taken into account by calculation in subsequent measurements, i.e. for example subtracted from the ascertained registration. In this case, as already explained, it is advantageous if the calibration structures for the comparison measurement have at least a similar embodiment to the actual used structures to be measured later. In particular a calibration structure can contain a size of smaller than 100 nm and greater than 74 nm, in particular greater 68 nm and particularly preferably greater than 60 nm half-pitch. As a result, a scale connection between the size differences of the so far structures to be measured and the in future to be measured structures, therefore from smaller than 100 nm down to the in the introduction formulated target dimension of 32 nm in particular, can be achieved.

[0083] Furthermore, the reference markers may be imaged using a fully illuminated illumination pupil since the reference markers usually have a sufficient size to generate sufficient contrast even in the case of a fully illuminated illumination pupil.In a further embodiment of the invention, a further measurement of the illumination may be performed in order to correct the registration following the registration measurement for at least one structure on the photomask. In the case where the registration is intended to be corrected subsequently, it is appropriate to take a further measurement of the illumination and ascertain the illumination error on the basis thereof after the registration measurement has been taken.

[0084] Furthermore, a correction value for the position and / or the illumination of the illumination pupil may be ascertained from the measurements of the illumination before and after the registration measurement by way of a temporal interpolation of said two measurements.

[0085] The illumination of the illumination pupil comprises the intensity over the location and the centroid of the illumination ascertained therefrom. In the ideal case, the latter may be embodied such that the intensity distribution of the light incident on the mask is symmetric, in particular rotationally symmetric.

[0086] Furthermore, the registration may be corrected with the aid of machine learning.

[0087] In particular, a module for machine learning may be trained on the basis of an algorithm ascertained with the aid of calibration masks.

[0088] In addition, the module for machine learning may be trained on the basis of an algorithm ascertained on the basis of previous registration measurements ascertained with the apparatus used.

[0089] In an advantageous embodiment of the invention, the method as invented can be carried out after an already performed identification and correction of an illumination error using means existing in the illuminations system. This means, that the inventive method above can be used for further improvement of a registration measurement after an already performed mechanical and / or optical correction, called mechanical-optical correction further on. The inventive method therefore reduces the remaining error after a mechanical-optical correction of the illumination setting.

[0090] Furthermore, a subsequent mathematical correction of the registration can also be used for correcting errors fundamentally also correctable with a mechanical-opticalcorrection. As a result, the time span between time consuming mechanical-optical corrections of the illumination system can be extended advantageously. This has a positive impact on the availability of the apparatus.

[0091] The measures for mechanical-optical correction of illumination errors presented below can be used independently or in combination with a mathematically correction to achieve beneficial results.

[0092] In a further embodiment of the invention the correction of the illumination error in the illumination system can be performed by at least one spatially variable and movable attenuator.

[0093] Furthermore, an illumination error in the illumination can be corrected by pole balancing.

[0094] In particular, the degree of polarization between at least two sub apertures of the pupil can be aligned by pole balancing. This enables identical or at least almost identical degrees of polarisation in all sub apertures of the pupil.

[0095] In a further embodiment of the invention, the illumination error of the illumination may be corrected by a movably embodied stop, as already explained further above.

[0096] Furthermore, the illumination error of the illumination may be corrected by adapting the centroid of the illumination pupil.

[0097] This can be achieved by a flat plate adjustable to an angle or a mirror adjustable to an angle. The flat plate can have a thickness corresponding to the desired effect of a beam offset.

[0098] Furthermore, the pupil position can be corrected spatially and angularly. Both the flat plate and the mirror can be arranged in front of the stop of the illumination system.

[0099] In particular, the adjustment of the spatial pupil position and the angle of the pupil can take place in two non-conjugated planes.Furthermore, the apodisation as one of the illumination errors can be measured in regular intervals, as for example daily, weekly or monthly. In case of a deviation of the apodisation from a predetermined set point the apodisation can be corrected.

[0100] The correction can be achieved by a further down explained multi-mirror-array. If correction means are not available or are not sufficient for correction, the apodization can be restored to its original or at least improved condition by so-called freeburning. By freeburning and simultaneous addition of oxygen and / or water contamination, in particular organic contamination, can be burned, so to say removed, on surfaces of optical elements, in particular a lens element. As a result, the nominal transmission of the lens can be restored, whereby also the apodisation can be restored in the desired nominal range.

[0101] In particular, the illumination error of the illumination may be corrected by a multi-mirror array.

[0102] Furthermore, the individual mirrors in the multi-mirror array may be movable and controllable.

[0103] The illumination error of the illumination may thus be corrected by a region-by-region adaptation of the intensities of the illumination pupil. In other words, if the illumination pupil has an intensity profile, in particular a non-symmetric intensity profile about the optical axis, then a transmission of the multi-mirror array, for example, is set such that the intensity profile is compensated for (at least the non-rotationally symmetric component of the intensity profile). Hence, an illumination pupil is obtained, the illumination intensity of which is homogeneous (or at least rotationally symmetric).

[0104] In a further embodiment, a selection of an illumination setting suitable for the structure to be measured may be ascertained with the aid of an algorithm.

[0105] In particular, the algorithm may use image simulations, for example with different illumination settings. The simulated images may be assessed according to suitable criteria (e.g. contrast or edge steepness or behaviour in the case of a slight defocus), and a suitable illumination setting may be selected thereafter.In a further embodiment, a selection of an illumination setting suitable for the structure to be measured may be created on the basis of machine learning. For example, images may be generated by means of machine learning. Alternatively, image generation may also be dispensed with, and the assessment criteria are generated by the machine learning itself.

[0106] Exemplary embodiments and variants of the invention will be explained in more detail below on the basis of the drawing, in which:

[0107] Figure 1 shows a schematic illustration of an apparatus in the form of a mask inspection microscope,

[0108] Figure 2 shows a variant of the invention,

[0109] Figure 3 shows a further variant of the invention,

[0110] Figure 4 shows a variant of the solution shown in Figure 3,

[0111] Figure 5 shows a further variant of the arrangements shown in Figures 3 and 4, Figure 6 shows an embodiment of the invention in which a tank is used, Figure 7 shows one possibility for determining the position of the photomask, Figure 8 shows an embodiment of the invention in which a solid-state element is used as the immersion element,

[0112] Figure 9 shows a variant of the embodiment shown in Figure 8,

[0113] Figures 10a,b show a detail of the invention,

[0114] Figures 11 a-c show a further detail from the prior art and the invention,

[0115] Figure 12 shows a diagram for explaining the invention,

[0116] Figure 13 shows a flowchart of a method according to the invention and Figure 14 shows a diagram for explaining the invention.Figure 1 shows, in a first embodiment of the invention, a schematic illustration of an apparatus designed as a mask inspection microscope 1 for measuring an object for semiconductor lithography in the form of a photomask or mask 7. The mask inspection microscope 1 comprises a beam source 2 for measuring the mask 7 in reflection. The mask 7 is arranged on a mask holder 6 capable of positioning the mask 7 laterally (xy-direction) and vertically (z-direction) in the sub-nanometre range. In particular, the positioning accuracy or position measurement accuracy may be of the order of better than 500 pm, in particular better than 250 pm, in this case.

[0117] During a measurement, the measurement light 12 emitted by the beam source 2 initially passes through an aperture stop 3 and is reflected off a beam splitter 4, and subsequently passes through an objective 8, which acts as an illumination optics unit at the same time, and so a part of the mask surface 16 is illuminated.

[0118] The objective 8 typically comprises multiple optical elements and is only depicted schematically in the figure.

[0119] The illuminated mask surface 16 is then imaged onto the recording device 10 through the beam splitter 4 and the imaging optics unit 9, in a manner enlarged by the objective 8.

[0120] The recording device 10, the stop 3, the mask holder 6, the objective 8, the imaging optics unit 9 and the beam source 2 are connected to a controller (not depicted separately in the figure) which controls the interplay of the individual components 2, 3, 6, 8, 9 and 10 in an open- or closed-loop manner and which is also designed to process the captured image representations.

[0121] An immersion element 11 is also shown in Figure 1 ; in the example shown, it is arranged as a spatially delimited liquid volume in the form of a droplet between the objective 8 and the surface of the mask 7. The liquid may be water in particular. In this case, the liquid volume is held by the surface tension between the objective 8 and the surface of the mask 7 in the example shown. The nature of the droplet, i.e. its geometry in particular, but also the presence of possibly interfering particles or gas bubbles, is ascertained continuously or at discrete time intervals by way of a monitoring system 13 in the example shown.The optical switching element 14, likewise shown in the figure, creates the possibility of switching between operation using an immersion element and operation without using an immersion element.

[0122] In the example shown, the lateral mask position relative to the mask holder 6 is monitored by the interferometer 15, while the interferometer 17 measures the position of the mask holder relative to the optical axis of the system (not indicated separately in the figure), and so the relative position of the mask relative to the optical axis of the system can be ascertained as a result.

[0123] Figure 2 shows a variant of the invention in which the immersion liquid 11 is arranged in the region between a supply 19 and a drain 20.

[0124] In this case, the positions of the supply 19, the drain 20 and the objective 8 remain unchanged relative to one another such that in the event of a relative movement between the mask 7 and objective 8, the same conditions prevail for the immersion liquid 11 , and the "droplet" migrates over the mask 7.

[0125] The supply 19 and the drain 20 may continuously supply new liquid or perform aspiration, and so possible contaminations are flushed away by the immersion liquid 11. Alternatively, the liquid 11 may be introduced once at a start of a measuring procedure, and a small amount of liquid may be supplemented for compensation purposes only in the event of a loss of liquid (for example evaporation). The droplet is held on the objective 8 by surface tension, and so said droplet is moved relative to the mask 7 in the event of a change in the measurement position.

[0126] The supply 19 is aligned in a direction parallel to the mask surface 16, and so no forces act in the z-direction as a result of the flow. The liquid is preferably discharged with a laminar flow. It is introduced between the objective 8 and the mask 7 in such a way that the necessary region is completely filled, i.e. the outermost possible rays in the beam path pass through at least a portion of the liquid.

[0127] Figure 3 shows a sectional plan view of the region of the supply 19 and the drain 20, the region of the immersion liquid 11 and the region 21 of the mask (not shown separately in the figure) that is captured by the objective 8. In this case, the supply 19 ex-hibits a widened and straight end piece that serves to reduce the flow rate of the immersion liquid 11 and provide the latter with a direction such that as far as possible said immersion liquid is situated only directly in the region 21 captured by the objective.

[0128] Figure 4 shows a variant of the solution shown in Figure 3, in which the drain 20 has a wider embodiment than the supply 19 such that possible liquid residues can be discharged in an improved manner.

[0129] Figure 5 shows a further variant of the arrangements shown in Figures 3 and 4, in which the drain 20 is in the form of a ring segment that at least partially surrounds the captured region 21, whereby the delimitation of the immersion liquid can be further improved. In principle, implementing a complete enclosure of the immersion liquid 11 is also possible.

[0130] Figure 6 illustrates an embodiment of the invention in which the photomask 7 can be completely submerged in a tank 22. In this case, the photomask 7 can move into the tank 22, and / or the tank 22 moves in such a way that the photomask 7 and the objective 8 are immersed in the immersion liquid 11.

[0131] A lowered position of the tank 22 is depicted in the figure using dashed lines.

[0132] In addition to that or in an alternative, the liquid level in the tank 22 may be controlled in such a way that immersion functionality is provided when measurements should be taken. The immersion liquid 11 can be pumped out again thereafter.

[0133] The mask holder 6 may be horizontally displaceable relative to the tank 22 or jointly with the latter.

[0134] Since pressure differences on account of liquid waves generate image errors, wavedamping elements 23 may be arranged in the tank 22 in such a way that liquid waves are damped. Such elements may be arranged on at least one side of the tank 22. They are arranged on both sides of the tank 22 in the example shown.Figure 7 shows one option for determining the position of the photomask 7 relative to the optical axis 24 of the system. Knowledge of this position in the order of sub-nano-metres is essential.

[0135] The end face of the photomask 7 is not perfectly flat. Hence the photomask 7 lies on a mask holder 6. The latter is equipped with precisely manufactured reference mirrors 29, which are measured by laser interferometers 25. By way of example, one of these reference mirrors 29 is shown in the figure. However, the reference mirrors 29 might not be perfectly flat even in the case of high manufacturing quality. It may therefore be necessary to perform a calibration step.

[0136] The figure also shows an immersed etalon 26, which can be used to determine the properties (e.g. refractive index) of the immersion liquid 11. It is also possible to use 2 etalons, one in the immersion liquid 11 for determining the properties of the immersion liquid 11 and one outside the immersion liquid 11 for determining the properties of the surrounding medium (e.g. air).

[0137] Figure 8 shows an embodiment of the invention in which a solid-state element 11.1 is used as the immersion element. In this case, the immersion element is flat on its side facing the photomask 7 and convex on the side which faces the last lens element 8.1 of the objective 8. It is also evident from the figure that the last lens element 8.1 is concave on its side facing the immersion element and convex on the other side.

[0138] Two capacitive sensors 28 which are connected to the housing (not denoted separately in the figure) of the objective and which can be used to detect an approach of the solid-state element 11.1 to the surface of the photomask 7 are also evident in Figure 8. In particular, three or four such sensors may in this case encircle the solid-state element 11.1 in a manner spaced apart by an angle of 120° or 90°.

[0139] Figure 9 shows a variant of the embodiment shown in Figure 8, in which the solid-state element 11.2 is provided with a projection 27 on its otherwise flat side facing the photomask 7. As a result of this measure, the approach of the solid-state element 11.2 attained for the desired optical effect is implemented only in the region from which the radiation required for the imaging also originates. Overall, the measureshown in Figure 9 prevents undesired contact between the solid-state element 11.2 and the photomask 7.

[0140] The underside of the solid-state element 11.1 or 11.2 may also be somewhat convex. This also prevents undesired contact between the solid-state element 11.1 or 11.2 and photomask 7.

[0141] Figure 10a shows a detail of the invention in a first embodiment and schematically illustrates beam path guidance of an off-axis illumination 100 known from the prior art. The latter may be integrated in a mask inspection system known from the prior art.

[0142] The illumination beam path 102 with an illumination in the form of a dipole in the illumination pupil (off-axis illumination) is incident on a portion of the object in the form of a mask 107. The illumination of the illumination pupil corresponds to the term illumination setting, as used further above. The embodiment of the off-axis illumination depicted in Figure 10a uses DUV light, i.e. light with a wavelength of 193 nm. In this case, the apparatus is configured to measure or register both masks for DUV lithography in transmission and reflection and masks for so-called EUV lithography in reflection. EUV lithography uses light with a wavelength of 13.5 nm as used light for imaging the structures.

[0143] The off-axis illumination may find use in a mask inspection microscope known from the prior art or in a mask inspection microscope 1 according to the invention with an immersion element 11 , as explained in Figure 1. In this case, DUV masks are also only measured or registered in reflection.

[0144] The illumination beam path 102 is thus steered by a mirror 101 onto an object in the form of a mask 107. For reasons of clarity, the portion of the illumination beam path 102 leading to the mirror 101 has not been depicted. The direction of the beam path 102 is depicted by arrows in Figure 10a.

[0145] According to the invention, the illumination beam path 102 is formed in such a way that the Oth order of diffraction imaging the structure (not depicted here) and at least the + / -1 st order of diffraction are reflected into an imaging optical unit of the apparatus in the form of a mask inspection system. Thus, the imaging beam path 103 depicted in Figure 10a at least comprises the light with the Oth and + / -1 st order of diffraction that was diffracted in the direction of the imaging optics unit off the structure features of interest, as will be explained in more detail in Figures 11a-11c. Within the meaning of the invention, structure features of interest are for example edges of a structure, for example of a line of the structure or of a cross used as a reference marker.

[0146] The diffraction light 103 is imaged onto a recording device of the apparatus and thus corresponds to the image information of the imaging used for the registration.

[0147] The point-symmetric arrangement of the illumination beam path 102 around the surface normal NF of the object 107 is advantageous in that a z-displacement of the object 107 leads only to slight blurring of the imaged structures, i.e. for example leaves the position of the edges of a cross almost unchanged. A slight z-displacement thus has almost no effect on the centre of the cross, which is relevant to the registration, and so said centre is not displaced as a result.

[0148] Figure 10b shows a construction identical to that in Figure 10a. In contrast to Figure 10a, the illumination of the illumination pupil, i.e. the illumination setting, is in the form of a quadrupole. Alternatively, the quadrupole may also be arranged rotated by 45°. The apparatus thus comprises flexible, i.e. different, illumination settings, which can also be changed according to the invention.

[0149] The illumination settings are selected on the basis of the structure to be measured or registered, wherein the selection may be based automatically on a priori knowledge about the mask structure or may be adapted dynamically on the basis of images that are taken in the machine itself.

[0150] Within the scope of the selection, the illumination setting may be selected in such a way that the best possible contrast and / or the best possible registration accuracy may be achieved for the respective mask structure.

[0151] The illumination settings may be adapted or changed, either from mask to mask or else within a mask for different specific measurement regions on a mask. As alreadyexplained further above, setting the illumination change may be realized by changing the stop, for example.

[0152] Furthermore, Figure 10b illustrates an illumination pupil with an illumination that is optimized for the structures to be measured, said illumination also being referred to as sunspot illumination. The latter depends on the arrangement, the alignment, the type and the local half-pitches of the structures. For example, the types of structures may be contact holes or lines. The lines may have different alignments, and both the contact holes and the lines may have locally different half-pitches, i.e. periods.

[0153] Figure 11a shows imaging known from the prior art, which is based on a periodic mask structure with a half-pitch that is only minimally greater than the optical resolution limit of the apparatus.

[0154] In this case, the photomask acts like an optical grating, whereby a Oth, a +1 st and a -1st order of diffraction arises in the pupil plane of the imaging.

[0155] Figure 11a shows an embodiment in which the wavelength is 193 nm, the numerical aperture is 0.8 and half-pitch is 75 nm, and the illumination pupil is completely filled with light. The scale denotes NA units. The Oth order of diffraction 111 is represented by a solid line, which at the same time represents the NA stop of the illumination 120 of the imaging optical unit. The + / -1 st orders of diffraction 112.1 , 112.2 are represented using dashed lines in Figure 11a and in the subsequent figures. It is evident that only a very small part of the +1 st 112.1 and -1 st 112.2 order of diffraction is able to pass the NA stop and interfere with the Oth order of diffraction 111 in the image plane of the imaging optical unit, which corresponds to the plane of the recording device, and thereby be able to generate contrast.

[0156] The majority of the Oth order of diffraction 111 finds no partner for interference, i.e. generates only a constant background during the imaging but no contrast. This means that this situation is so close to the theoretical resolution limit of the imaging optical unit that although the structure can be imaged, it can only be imaged with very poor contrast. The contrast is so weak that an evaluation for determining the structure and in particular for registering the structure would not be sufficiently accurate because the signal-to-noise ratio of the imaging is insufficient.Figure 11 b shows the same situation for an off-axis illumination according to the invention in the form of a dipole. The illumination pupil thus is no longer illuminated in full but is shaped, for example by a stop in the illumination system, in such a way that the illumination radiation is incident obliquely on the object, i.e. incident on the object off axis.

[0157] In this case, virtually the entire + / -1 st order of diffraction 114.1, 114.2 is situated within the NA stop of the imaging as a result of the utilized dipole, whereby - as explained further above - the contrast is significantly improved. For comparison, the + / 1 st order of diffraction 112.1, 112.2 of the on-axis illumination, as explained in Figure 11a, is depicted as a dashed line. It is clearly evident that a very small portion of the + / -1 st order of diffraction 112.1 , 112.2 is located within the NA stop, whereby the contrast, too, is not sufficient for a registration of the mask, in comparison with the off-axis illumination represented by a dipole. As explained, the Oth order of diffraction 113 is imaged completely within the NA stop 120.

[0158] Figure 11 c shows a further embodiment of the invention in the form of an illumination of the illumination pupil referred to as two leaf illumination. As can easily be identified in Figure 11c, an advantage of this case is that, given the same half-pitch as in Figure 11 b, the entire + / -1 st order of diffraction 116.1, 116.2 is already located within the NA stop of the imaging, and so each ray of the Oth order 115 finds a partner for interference, and a maximum contrast is generated as a result. Like in Figure 11b, the + / 1 st order of diffraction 112.1, 112.2 of the on-axis illumination, as explained in Figure 11a, is depicted as a dashed line in Figure 11 c as well. As explained, the Oth order of diffraction 115 is imaged completely within the NA stop 120.

[0159] Figure 12 shows a diagram for explaining the invention, in which the contrast as a function of half-pitch is depicted by two curves 121 , 122.

[0160] A first curve 121 represented by a solid line corresponds to the contrast profile as a function of half-pitch in an off-axis illumination according to the invention, with the latter being in the form of a dipole in the embodiment depicted in Figure 12.A second curve 122, represented by dashed lines, represents the contrast profile as a function of the half-pitch for a fully illuminated illumination pupil, as known from the prior art.

[0161] The y-axis shows the size of the proportion of the + / -1 st order of diffraction located within the NA stop of the imaging, with the x-axis showing the increasing half-pitch. As already explained further above, the proportion of the + / -1 st order of diffraction is tantamount to the contrast of the imaging since only light which passes through the NA stop and reaches the recording device, for example in the form of a camera, is able to generate contrast.

[0162] Both types of illumination (on-axis / off-axis) have the same resolution limit (intersection of the x-axis), but the limit above which a sufficient contrast for the registration is generated, which is represented by a dash-dotted line in Figure 12, is reached at different half-pitches. This means that the off-axis illumination does not increase the resolution limit of the apparatus, but instead the off-axis illumination betters the limit above which sufficient contrast for the registration is generated. Phrased more generally, given the same half-pitch, the off-axis illumination improves the contrast in the region of the resolution limit of the overall system, in part significantly, in relation to an on-axis illumination known from the prior art.

[0163] Figure 13 describes a possible method for correcting illumination errors of the illumination when measuring and / or registering at least one structure on a photomask (7) for semiconductor lithography.

[0164] The illumination is measured in a first method step 131.

[0165] The illumination error of the illumination is ascertained in a second method step 132.

[0166] The ascertained illumination error is corrected in a third method step 133 when registering at least one structure.

[0167] Figure 14 shows an exemplary diagram illustrating the effect of the illumination error on the relationship of the registration with the displacement of the centroid of the illumination pupil.The curve 123 represented by the solid line shows the relationship of the registration with the displacement of the centroid of the illumination pupil, as established by measurements. The centroid of the illumination pupil may be caused by different illumination errors, for example an inhomogeneity in the illumination of the illumination pupil (e.g. caused by an imperfect adjustment of the illumination optics unit).

[0168] As already explained above, there are different possible corrections. The curve 124 represented by the dashed line in Figure 14 shows an already corrected relationship between centroid of the illumination pupil and the registration. By using this corrected profile 124, the illumination errors during the registration may be corrected or at least reduced.

[0169] However, it is self-evident that the solutions disclosed in the present application are largely also combinable in partial aspects. Moreover, the presented solutions can also be advantageously used in systems with lower demands on the resolution.List of reference signs

[0170] 1 Mask inspection microscope 2 Beam source

[0171] 3 Aperture stop

[0172] 4 Beam splitter

[0173] 6 Mask holder

[0174] 7 Mask

[0175] 8 Objective

[0176] 8.1 Optical element

[0177] 9 Imaging optics unit

[0178] 10 Recording device

[0179] 11 Immersion element

[0180] 11.1 Solid-state element

[0181] 11.2 Solid-state element

[0182] 13 Monitoring system

[0183] 14 Switching element

[0184] 15 Interferometer

[0185] 16 Mask surface

[0186] 17 Interferometer

[0187] 19 Supply

[0188] 20 Drain

[0189] 21 Captured region

[0190] 22 Tank

[0191] 23 Wave-damping element 24 Optical axis

[0192] 25 Laser interferometer

[0193] 26 Etalon

[0194] 27 Projection

[0195] 28 Sensor

[0196] 29 Reference mirrorMirror

[0197] Illumination radiation

[0198] Imaging beam path

[0199] Mask

[0200] Illumination pupil

[0201] Zeroth order of diffraction

[0202] 1 st order of diffraction

[0203] -1 st order of diffraction

[0204] Zeroth order of diffraction (Figure 11 b)

[0205] 1 st order of diffraction (Figure 11 b)

[0206] -1 st order of diffraction (Figure 11 b)

[0207] Zeroth order of diffraction (Figure 11c)

[0208] 1 st order of diffraction (Figure 11 c)

[0209] -1 st order of diffraction (Figure 11 c)

[0210] NA stop

[0211] Contrast profile for half-pitch off-axis

[0212] Contrast profile for half-pitch on-axis

[0213] Relationship of registration with the displacement of the pupil centroid (illumination)

[0214] Corrected profile

[0215] First method step (measuring the illumination)

[0216] Second method step (ascertaining the illumination error) Third method step (correcting the illumination error)

Claims

Claims1. Apparatus (1 ) for semiconductor lithography for measuring and / or registering structures on a photomask (7), comprising an illumination system having a beam source (2) for illuminating the photomask (7) with measurement light (12) and comprising an objective (8) for imaging the structures onto a recording device (10),characterized in thatthe apparatus (1) is configured to generate an image representation of the photomask (7) with a resolution of better than 74 nm, preferably less than 68 nm and particularly preferably less than 60 nm half-pitch.

2. Apparatus (1 ) according to Claim 1 ,characterized in thatthe photomask (7) is an EUV mask.

3. Apparatus (1 ) according to either of Claims 1 and 2,characterized in thatthe measurement light (12) has a wavelength shorter than 200 nm, in particular a wavelength of 193 nm, 157 nm or 13.5 nm.

4. Apparatus (1 ) according to any of Claims 1 to 3,characterized in thatan immersion element (11 ) is arranged between the last optical element of the objective (8) and the structures to be imaged.

5. Apparatus (1 ) according to Claim 4,characterized in thatthe immersion element (11) is a liquid volume of an immersion liquid.

6. Apparatus (1 ) according to Claim 5,characterized in thatthe immersion liquid is water, in particular high-purity water.

7. Apparatus (1 ) according to Claim 5 or 6,characterized in thatthe liquid volume is formed in a spatially limited manner in the region between the last optical element of the objective (8) and the structures to be imaged and covers the photomask (7) only in regions.- 34 -8. Apparatus (1 ) according to any of Claims 5 to 7,characterized in thatthe surface of the photomask (7) is of a hydrophobic form at least in regions.

9. Apparatus (1 ) according to Claim 8,characterized in thatthe photomask (7) is provided with a hydrophobic coating at least in regions.

10. Apparatus (1) according to any of Claims 5 to 9,characterized in thata last optical element of the objective (8) in the direction of the photomask (7) is of a hydrophilic form at least in regions on the surface which faces the photomask (7).

11. Apparatus (1 ) according to any of Claims 5 to 10,characterized in thata supply (19) and a drain (20) for the immersion liquid are present.

12. Apparatus (1 ) according to Claim 11 ,characterized in thatthe drain (20) at least partially surrounds the liquid volume.

13. Apparatus (1) according to Claim 12,characterized in thatthe drain (20) is embodied in the form of a ring segment.

14. Apparatus (1) according to Claim 5,characterized in thatthe liquid volume completely covers the photomask (7).

15. Apparatus (1) according to Claim 14,characterized in thatthe liquid volume is formed in a liquid located in a tank (22).

16. Apparatus (1) according to Claim 15,characterized in thatmeans for moving the tank (22) in the vertical z-direction are present.

17. Apparatus (1 ) according to Claim 15 or 16,characterized in that- 35 -means for moving the tank (22) horizontally in the x- and / or y-directions are present.

18. Apparatus (1 ) according to any of Claims 15 to 17,characterized in thata pumping device for filling and an outlet for emptying the tank (22) are present.

19. Apparatus (1 ) according to any of Claims 15 to 18,characterized in thatwave-damping elements (23) for damping liquid movements in the tank (23) are present.

20. Apparatus (1) according to any of Claims 5 to 19,characterized in thata last optical element of the objective (8) in the direction of the photomask (7) is formed flat or with a comparatively large radius of curvature of greater than 500 mm, preferably of greater than 1000 m, on the side which faces the photomask (7).

21. Apparatus (1 ) according to any of Claims 5 to 20,characterized in thata last optical element of the objective (8) in the direction of the photomask (7) is designed to be interchangeable.

22. Apparatus (1 ) according to any of Claims 5 to 21 ,characterized in thatmeans for controlling the temperature of, applying pressure to or applying a gas flow to the immersion liquid (11 ) and / or the photomask (7) are present.

23. Apparatus (1) according to any of Claims 5 to 22,characterized in thatmeans for assessing the quality of the immersion liquid (11 ) or the quality of cleaning of the mask surface (16) of immersion liquid (11 ) after the measurement are present.

24. Apparatus (1) according to any of Claims 5 to 23,characterized in thatmeans are present for drawing inferences about the state of the immersion liquid (11 ) by means of imaging at a wavelength significantly different from thecustomary measurement wavelength, typically at a significantly longer wavelength.

25. Apparatus (1) according to any of Claims 5 to 24,characterized in thatmeans are present for creating a gas atmosphere from a gas which is readily soluble in the immersion liquid (11 ).

26. Apparatus (1) according to any of Claims 5 to 25,characterized in thatmeans for determining or preventing an offset of the photomask (11) are present.

27. Apparatus (1) according to Claim 4,characterized in thatthe immersion element (11 ) is a solid-state element.

28. Apparatus (1) according to Claim 27,characterized in thatthe solid-state element is a refractive element with a refractive index greater than 1.5, preferably greater than 1.8.

29. Apparatus (1) according to Claim 27 or 28,characterized in thatthe apparatus (1 ) is configured to bring the immersion element (11 ) up to a distance of less than 50 nm, preferably of less than 20 nm and further preferably of less than 10 nm, from a surface of the photomask (7).

30. Apparatus (1) according to any of Claims 27 to 29,characterized in thatthe apparatus (1) comprises a distance control device for setting the distance between the immersion element (11) and the surface of the photomask (7).

31. Apparatus (1) according to Claim 30,characterized in thatthe distance control device comprises electrical sensors (28), in particular capacitive sensors.

32. Apparatus (1) according to Claim 30 or 31,characterized in thatthe distance control device is configured to determine the distance between the immersion element (11 ) and the surface of the photomask (7) from the reflection behaviour of the immersion element (11 ).

33. Apparatus (1) according to any of Claims 27 to 32,characterized in thatthe immersion element (11.1 ) is uneven on the side which faces the photomask (7).

34. Apparatus (1) according to Claim 33,characterized in thatthe immersion element (11.1) has a projection (27) on the side which faces the photomask (7).

35. Apparatus (1) according to Claim 4,characterized in thatthe immersion element is in the form of a planar element of a highly viscous medium.

36. Apparatus (1) according to any of Claims 4 to 35,characterized in thatmeans for choosing the mode of operation of the apparatus (1 ) with or without an immersion element (11, 11b) are present.

37. Apparatus (1) according to any of the preceding claims,characterized in thatthe illumination system is configured to generate a structured illumination (100,100’).

38. Apparatus (1) according to Claim 37,characterized in thatthe illumination system is configured to generate an off-axis illumination (100, 100').

39. Apparatus (1) according to either of Claims 37 or 38,characterized in thatthe apparatus (1) comprises a device (3, 120) for setting an illumination setting.- 38 -40. Apparatus (1) according to any of Claims 37 to 39,characterized in thatthe illumination of an illumination pupil (110) is in the form of a dipole.

41. Apparatus (1) according to any of Claims 37 to 39,characterized in thatan illumination of the illumination pupil (110) is in the form of a quadrupole.

42. Apparatus (1) according to any of Claims 37 to 39,characterized in thatat least one section of an illumination of an illumination pupil (110) is embodied in a leaf form.

43. Apparatus (1) according to any of Claims 37 to 39,characterized in thatthe illumination of the illumination pupil (110) is embodied in the form of sunspots.

44. Apparatus (1) according to any of Claims 37 to 39,characterized in thatthe illumination of the illumination pupil (110) is embodied in the form of a hexagonal form.

45. Apparatus (1) according to any of Claims 37 to 39,characterized in thatthe illumination of the illumination pupil (110) is embodied in the form of a ring structure.

46. Apparatus (1) according to any of Claims 37 to 45,characterized in thata stop in the apparatus (1 ) is designed to be movable in order to set the illumination of the illumination pupil (110).

47. Apparatus (1) according to any of Claims 37 to 46,characterized in thatthe illumination system comprises a device with an arrangement of individual illumination surfaces which can be added or switched off.- 39 -48. Apparatus (1) according to Claim 47,characterized in thatan illumination system comprises a multi-mirror array.

49. Apparatus (1) according to any of the preceding claims,characterized in thatthe apparatus (1) comprises a control device with a module for machine learning.

50. Apparatus (1) according to any of the preceding claims,characterized in thatthe apparatus (1) comprises means for setting the polarization of the measurement light (12).

51. Apparatus (1) according to Claim 50,characterized in thatthe means for setting the polarization of the measurement light (12) are configured to set a tangential illumination polarization.

52. Method for correcting illumination errors of the illumination when measuring and / or registering at least one structure on a photomask (7) for semiconductor lithography, including the following method steps:- measuring the illumination,- ascertaining the illumination error of the illumination,- ascertaining the influence of the illumination error on the result of the measurement or registration,- correcting the measured registration by the ascertained influence of the illumination error.

53. Method according to Claim 52,characterized in thatthe illumination is measured using at least one calibration mask.

54. Method according to Claim 53,characterized in thatthe calibration mask comprises calibration structures and reference markers.

55. Method according to Claim 54,characterized in that- 40 -the calibration structures are measured or registered using different illuminations.

56. Method according to Claim 55,characterized in thatthe deviation of the ascertained positions of the measurements or registrations is determined from at least one measurement of a calibration structure with off-axis (100, 100') illumination and at least one with full illumination.

57. Method according to Claim 54,characterized in thatthe reference markers are imaged with a completely illuminated illumination pupil (110).

58. Method according to any of Claims 52 to 57,characterized in thata further measurement of the illumination is performed in order to correct the registration following the registration measurement for at least one structure on the photomask (7).

59. Method according to Claim 58,characterized in thata correction value for the position and / or the illumination of the illumination pupil (110) is ascertained from the measurements of the illumination before and after the registration measurement by way of a temporal interpolation of said two measurements.

60. Method according to either of Claims 58 and 59,characterized in thatthe registration is corrected on the basis of the measurement of the illumination and with the aid of numerical simulations.

61. Method according to Claims 58 to 60,characterized in thatthe registration is corrected with the aid of machine learning.

62. Method according to Claim 61 ,characterized in thata module for machine learning is trained on the basis of an algorithm ascertained with the aid of calibration masks.- 41 -63. Method according to Claim 62,characterized in thatthe module for machine learning is trained on the basis of an algorithm ascertained on the basis of previous registration measurements ascertained with the apparatus (1) used.

64. Method according to any of Claims 52 to 63,characterized in thatthe method as invented can be carried out after an already performed identification and correction of an illumination error using means existing in the illuminations system.

65. Method according to Claim 64,characterized in thata subsequent mathematical correction of the registration can also be used for correcting errors fundamentally also correctable with a mechanical-optical correction.

66. Method for correcting illumination errors of the illumination when measuring and / or registering at least one structure on a photomask (7) for semiconductor lithography, in particular according to any of claims 52 to 65, characterized in thatthe correction of the illumination error in the illumination system can be performed by at least one spatially variable and movable attenuator.

67. Method for correcting illumination errors of the illumination when measuring and / or registering at least one structure on a photomask (7) for semiconductor lithography, in particular according to any of claims 52 to 66, characterized in thatan illumination error in the illumination can be corrected by pole balancing.

68. Method according to Claim 67,characterized in thatthe degree of polarization between at least two sub apertures of the pupil can be aligned by pole balancing.

69. Method for correcting illumination errors of the illumination when measuring and / or registering at least one structure on a photomask (7) for semiconductor- 42 -lithography, in particular according to any of Claims 52 to 68, characterized in thatthe illumination error of the illumination is corrected by a movably embodied stop.

70. Method for correcting illumination errors of the illumination when measuring and / or registering at least one structure on a photomask (7) for semiconductor lithography, in particular according to any of Claims 52 to 69, characterized in thatthe illumination error of the illumination is corrected by adapting the centroid of the illumination pupil (110).

71. Method for correcting illumination errors of the illumination when measuring and / or registering at least one structure on a photomask (7) for semiconductor lithography, in particular according to any of Claims 52 to 70, characterized in thatthe illumination error of the illumination is corrected by a multi-mirror array.

72. Method according to Claim 71 ,characterized in thatthe individual mirrors in the multi-mirror array are movable and controllable.

73. Method according to any of Claims 64 to 72,characterized in thata selection of an illumination setting suitable for the structure to be measured is ascertained with the aid of an algorithm.

74. Method according to Claim 73,characterized in thatthe algorithm uses image simulations.

75. Method according to Claim 73,characterized in thata selection of an illumination setting suitable for the structure to be measured is created on the basis of machine learning.- 43 -

Images