JP6114952B2 - Method for transferring a pattern onto a photosensitive surface by lithography and illumination system of a microlithographic projection exposure apparatus - Google Patents

Description

  The present invention relates generally to microlithographic exposure apparatus, and more particularly to a method for transferring a pattern onto a photosensitive surface by lithography and an illumination system for such an apparatus.

  Microlithography (also called photolithography or simply lithography) is a technique for manufacturing integrated circuits, liquid crystal displays and other microstructured devices. Microlithographic processes, along with etching processes, are used to pattern features in a thin film stack formed on a substrate, such as a silicon wafer. In each layer of manufacture, the wafer is first coated with a photoresist, which is a material that is sensitive to radiation, such as deep ultraviolet (DUV) light. Next, a wafer having a photoresist thereon is exposed to projection light in a projection exposure apparatus. This apparatus projects a mask containing a pattern onto the photoresist, so that the photoresist is exposed only at specific locations determined by the mask pattern. When the exposure is completed, the photoresist is developed, and an image corresponding to the mask pattern is generated. The pattern is then transferred into the thin film stack on the wafer by an etching process. Finally, the photoresist is removed. This process is repeated using different masks, resulting in a multilayer microstructured component.

A projection exposure apparatus typically includes an illumination system for illuminating the mask, a mask stage for aligning the mask, a projection objective, and a wafer alignment stage for aligning the photoresist-coated wafer. The illumination system illuminates a field on the mask that may have the shape of, for example, a rectangular or curved slit.
In current projection exposure apparatus, two different types of apparatus can be distinguished. In one type, each target portion on the wafer is irradiated by exposing the entire mask pattern to the target portion at once. Such an apparatus is generally called a wafer stepper. In other types of devices, commonly referred to as step and scan devices or scanners, the mask pattern is continuously scanned along the scan direction under the projection beam, while the substrate is synchronized in this direction. Each target part is irradiated by moving in parallel or antiparallel. The ratio of the wafer speed to the mask speed is equal to the magnification of the projection objective, and is usually less than 1, for example 1: 4.

It should be understood that the term “mask” (or reticle) should be interpreted broadly as patterning means. Widely used masks include transmissive or reflective patterns, and may be, for example, binary, alternating phase shifted, halftone (attenuated) phase shifted masks, or various hybrids It may be a mask type. However, there are also active masks, for example masks realized as programmable mirror arrays. It is also possible to use a programmable LCD array as an active mask.
With advances in technology for manufacturing microstructured devices, the demand for lighting systems has never been greater. Ideally, the illumination system illuminates each point of the illuminated field on the mask with projection light having a clear irradiance and angular distribution. The term angular distribution describes how the total light energy of a light beam that converges towards a particular point in the mask plane is distributed among the various directions of the light rays that make up that light beam.
The angular distribution of projection light incident on the mask is usually adapted to the type of pattern projected onto the photoresist. For example, features that are relatively large in size may require a different angular distribution than features that are small in size. The most widely used angular distribution of projection light is called the conventional annular dipole illumination setting and the quadrupole illumination setting. These terms mean the irradiance distribution in the pupil plane of the lighting system. In the case of an annular illumination setting, for example, only the annular area is illuminated in the pupil plane. Therefore, only a small range of angles are present in the angular distribution of the projection light, so that all light rays are incident obliquely on the mask at similar angles.

  If the mask pattern to be imaged on the photoresist contains portions that appear to be very different with respect to structural orientation, density and / or pitch, it is usually necessary to image all mask pattern areas with optimum image quality. There is no specific angular light distribution that is perfectly suitable. For example, if there is a first mask pattern region that includes features extending along the X direction and a second mask pattern region that includes features extending in the orthogonal Y direction, a quadrupole illumination setting or C-quad The illumination setting images all features with the same sub-optimal image quality. This is because only two poles contribute to the image formation of each feature, and the other two poles contribute to the background illumination, which lowers the contrast. Ideally, features extending along the X direction will be illuminated exclusively using the Y dipole illumination setting, and features extending along the Y direction will have an X dipole illumination setting. Will be used and illuminated.

Such improved illumination, and thus improved imaging, allows the pattern to be exposed to the photosensitive surface in two (or more) separate exposures using a different photomask rather than in a single exposure. This can be achieved when transferred. The first mask includes all features extending along the X direction and is illuminated using the Y dipole illumination setting. The second mask includes features extending along the Y direction and is illuminated in subsequent exposures using an X dipole illumination setting. This technique of distributing a complex pattern of imaging between two separate exposures is commonly referred to as “double exposure”. If this approach is extended to more than two subsequent exposures, it is called multiple exposure.
Since the mask can be changed quickly, the illumination system must also be able to quickly change the angular light distribution at the mask level.

  Different means are known in the art for modifying the angular distribution of the projected light in the mask plane, thereby achieving the desired illumination setting. In the simplest case, a diaphragm with one or more apertures is arranged in the pupil plane of the illumination system. Since the position in the pupil plane converts to an angle in a Fourier related field plane such as the mask plane, the size, shape and position of the aperture in the pupil plane determines the angular light distribution at the mask plane level. However, in order to arbitrarily change the illumination setting, it is necessary to replace the diaphragm. This makes it difficult to fine-tune the lighting settings, since changing the stop requires a large number of stops with slightly different sizes, shapes or positions of apertures. Furthermore, the use of an aperture unavoidably impairs the light and reduces the throughput of the device.

Thus, many widely used illumination systems include adjustable elements that can continuously change the illumination of the pupil plane at least to some extent.
The most common approach is to use a diffractive optical element that defines an irradiance distribution in the pupil plane. Diffractive optical elements can be manufactured so that almost any arbitrary far-field irradiance distribution can be generated. Such diffractive optical elements are often realized as computer generated holograms (CGH). In order to be able to change the irradiance distribution in the pupil plane, a plurality of different diffractive optical elements can be accommodated in the exchange holder, so that one of these diffractive optical elements can be It can be quickly inserted into the optical path of the projection light. Such an exchange holder may be, for example, a linear exchange holder that can be rotated around a rotation axis, or a turret holder.
The irradiance distribution in the pupil plane defined by the diffractive optical element can be supplementarily corrected with the aid of zoom optics and a pair of axicon elements. By displacing the lens and / or axicon element of the zoom optical system along the optical axis, the irradiance distribution can be continuously corrected within a certain limit.

However, exchanging diffractive optical elements and displacements along the optical axis of lenses and other optical elements is a relatively slow process and may therefore limit the throughput of the apparatus.
A more flexible and quicker approach to generating different irradiance distributions in the pupil plane of the illumination system uses mirror arrays instead of diffractive optical elements.
In EP 1262836 (A1), the mirror array is realized as a microelectromechanical system (MEMS) with more than 1000 microscopic mirrors. Each of the mirrors can be tilted in two different planes that are orthogonal to each other. Thus, radiation incident on such a mirror device can be reflected in (desirably) any desired direction of the hemisphere. A condensing lens placed between the mirror array and the pupil plane converts the reflection angle produced by the mirror into a position in the pupil plane. According to this known illumination system, multiple spots can be used to illuminate the pupil plane, each spot being coupled to one specific microscopic mirror and by tilting this mirror The entire pupil plane can be moved freely.

Similar lighting systems are known from US 2006/0087634, US 7061582 (B2) and WO 2005/026843.
The illumination system disclosed in WO 2010/006687 uses a mirror array not only to determine the angular irradiance distribution but also to determine the spatial irradiance distribution in the mask plane. To that end, the mirror produces a very small spot and therefore a different light pattern can be produced on the light entrance facet of the light integrator. Since these light entrance facets are optically conjugate with the mask plane, the specific light pattern generated on the light entrance facets is imaged directly onto the mask plane. By using this ability to change the geometry of the illuminated field in the mask plane, the need for adjustable blades in the illumination system can be eliminated. Such blades are used at the beginning and end of each scanning cycle in a scanner type device to enlarge or reduce the illuminated field along the scan direction. It is also described that in a given example, different light entrance facets can be illuminated with different light patterns, thus creating a spatial irradiance distribution with a smooth profile.

  The unpublished International Patent Application No. PCT / EP2011 / 000416 also discloses an illumination system that, in a given example, can produce different angular light distributions in different parts on the mask during the scanning process. doing. This is also achieved by generating different light patterns on the light entrance facets of the light integrator.

According to the unpublished international patent application No. PCT / EP2010 / 005317, different angular light distributions are generated in different parts of the mask with the aid of a plurality of modulator units arranged in the vicinity of the optical integrator. can do. Each modulator unit is coupled to a single light channel of the light integrator and redistributes the spatial and / or angular light distribution of each light channel in a variable manner without blocking any light.
WO 2006/097135 discloses an illumination system in which the angular light distribution at the mask level can be changed very quickly, for example between successive pulses of a pulse train emitted by a light source. This can be achieved, for example, by alternately shielding specific portions of the pupil plane by rotating diaphragms, by using LCD transmission filters in the pupil plane, or by using acousto-optic effects in diffractive optical elements. can do. The state of polarization can be changed simultaneously with the angular light distribution with the aid of a polarizing element rotating at high speed.

US Patent Application Publication No. 2008/0013065 also discloses an illumination system that can very quickly change the angular light distribution at the mask level. To that end, the illumination system includes first and second pupil defining units each including a diffractive optical element, a zoom optical system, and a pair of axicon elements. The projection light is guided through either the first or second pupil defining unit using a suitable light switching component before entering the light integrator.
Overlay errors are a serious problem when double (or generally multiple) exposures are performed. The term overlay error originally relates to the positioning of adjacent pattern layers within a microstructure device. If the features that are to be placed on top of each other are displaced laterally, this offset is called overlay error. On the other hand, the term overlay error is still used to describe the relative displacement of features within a single layer, since this overlay error can occur in a layer structured by a double exposure process. Has been. Many microstructure devices are more sensitive to overlay errors within a single layer than to overlay errors between adjacent layers. As such, the overlay error budget within a single layer produced by a double exposure process is often particularly small. Too strict specifications drastically reduce the yield of the manufacturing process. Thus, despite the obvious advantages associated with illuminating features using projection light with optimal angular light distribution, double exposure is still not widely used.

  One significant source of overlay error in the double exposure process is related to the superposition of two different mask images. The next most significant cause is not only the alignment errors of the two masks that contribute to the overlay error budget, but also manufacturing tolerances that cannot be avoided when the two masks are manufactured. Mask manufacturing is a very elaborate process, so it is impossible to fully reproduce the conditions that were dominant when one mask was manufactured while other masks were manufactured. In other words, even if the alignment error during the second exposure can be reduced to zero, a significant overlay error still exists because the two masks do not completely correspond to each other.

  Accordingly, it is an object of the present invention to provide a method for transferring a pattern to a photosensitive surface by lithography in a multiple exposure process, wherein the overlay error is reduced.

According to the invention, this object is
a) providing a mask comprising a first mask pattern region and a second mask pattern region;
b) Projecting light is guided onto the mask, whereby the first exposure pattern area, which is an image of the first mask pattern area, and the second exposure pattern area, which is an image of the second mask pattern area, are exposed to the photosensitive surface. Generating steps,
c) repeating step b) using the same mask such that an image of the first mask pattern area is superimposed on the second exposure pattern area, and the first mask pattern The projection light that illuminates the region has a first angular light distribution, and the projection light that illuminates the second mask pattern region has a second angular light distribution different from the first angular light distribution.
The present invention is based on the concept that overlay errors can be significantly reduced when the multiple exposure process is performed using only a single mask, rather than a plurality of different masks. This requires that different mask pattern areas are illuminated with projection light having different angular light distributions that are adapted to a particular mask pattern area. In other words, two different masks conventionally used in the double exposure process are combined into a single mask, and thus this single mask is used for both exposures. This method eliminates one significant contribution to overlay error, i.e., manufacturing tolerances while manufacturing two masks, because all mask pattern regions are manufactured in a single manufacturing process.

For example, if a microscale change occurs while manufacturing a mask, two different masks will be slightly different with respect to their scale, so that the two masks are used in a double exposure process In such a case, an overlay error will inevitably occur. According to the present invention, different mask pattern regions will inevitably have the same scale, so a change in scale cannot impair feature overlay within a single layer.
Again, according to the present invention, each mask pattern region can be illuminated using an optimal illumination setting. In other words, the first angular light distribution is usually adapted to the first mask pattern region, and the second angular light distribution is adapted to the second mask pattern region. Thus, the first and second mask pattern areas are actually designed in the same way as two different masks in a conventional double (or generally multiple) exposure process.

  The first and second angular light distributions can be combined with the illumination settings included in the group consisting of conventional illumination settings, annular illumination settings and multipole illumination settings. Multipole illumination settings can be included in a group consisting of scan direction dipole illumination settings, cross scan direction dipole illumination settings, quadrupole illumination settings, and C-quad illumination settings.

  During step b), in the former case, the illumination system is able to direct projection light having different angular light distributions to the first mask pattern region and the second mask pattern region simultaneously or sequentially. It implies that it must be possible to simultaneously generate projection light with different angular light distributions in different parts of the field illuminated above. Illumination systems that can generate a field-dependent angular light distribution are known in the art as such. In this context, reference is made to the unpublished international patent applications PCT / EP2011 / 000416 and PCT / EP2010 / 005317, both of which are briefly described above.

During step b), if the first and second mask pattern regions are generated simultaneously using projection light having different angular light distributions, and the illuminated field is the first mask pattern region and the second When the mask is moved along the scan direction during step b) so as to scan the mask pattern area of the first mask pattern area, the first mask pattern area and the second mask pattern area are in the scan direction and the normal line on the mask. Can be arranged next to each other along a cross scan direction orthogonal to both.
The simplest technique for superimposing the image of the first mask pattern area on the second exposure pattern area is then between steps b) and c) with an appropriate amount of displacement, eg cross scan direction. Displacing the photosensitive surface laterally by half the length of the illuminated field along. Similar considerations apply to wafer stepper type devices.
A more complicated approach would be to rotate the mask by 180 ° around the normal on the mask prior to performing the second exposure. Thereby, during step c), not only can the image of the first mask pattern area be superimposed on the second exposure pattern area, but at the same time the image of the second mask pattern area is superimposed on the first exposure pattern. It can be superimposed on the area.
During step b), if the projection light is first directed to the first mask pattern area and subsequently to the second mask pattern area, the illumination system will make the angular light distribution very rapidly during the exposure process. Must be able to change. Illumination systems that can do this are known from WO 2006/097135 and US Patent Application Publication No. 2008/0013065, also briefly described above.

This technique can also be implemented using a wafer stepper. In that case, for example, a first row in which the first and second exposure pattern regions are alternately arranged can be generated. Next, a second exposure is performed, and an image of another mask pattern region is superimposed on each exposure pattern region.
If the mask is moved along the scan direction during step b) so that the illuminated field scans over the first mask pattern region and the second mask pattern region, the first mask pattern region and the second mask pattern region The mask pattern regions can be arranged vertically along the scanning direction on the mask. Thereby, during step c), not only can the image of the first mask pattern area be superimposed on the second exposure pattern area, but also the image of the second mask pattern area can be superimposed on the first exposure pattern area. Can be superimposed on top.

  The technique described above for performing double exposure using one and the same mask can be easily extended to multiple exposure. For example, the mask may supplementarily include a third mask pattern region, and an image of the third mask pattern region may be superimposed on the first exposure pattern region and the second exposure pattern region. The projection light that illuminates the third mask pattern region has a third angular light distribution different from the first angular light distribution and the second angular light distribution. Similar to the embodiment described above, the projection light can be directed simultaneously or continuously to the first, second and third mask pattern regions during step b).

  Similarly, the mask may additionally include a fourth mask pattern region, and an image of the fourth mask pattern region may include the first exposure pattern region, the second exposure pattern region, and the third exposure pattern region. The projection light that can be superimposed on the exposure pattern area and that illuminates the fourth mask pattern area is a fourth light different from the first angular light distribution, the second angular light distribution, and the third angular light distribution. The angular light distribution is as follows.

  One technique for superimposing the images of all four mask pattern areas on the photosensitive surface is to generate the first exposure pattern area and simultaneously apply the second exposure pattern area to the photosensitive surface during step b). May be implied to generate. Simultaneously with the generation of the third exposure pattern area, which is an image of the third mask pattern area, a fourth exposure pattern area, which is an image of the fourth mask pattern area, is generated on the photosensitive surface. During step c), an image of the third mask pattern area is superimposed on the first exposure pattern area, thereby generating a combination of the first mask pattern area plus the third mask pattern area. At the same time, the image of the fourth mask pattern region is superimposed on the second exposure pattern region, thereby generating a combination of the second mask pattern region plus the fourth mask pattern region. In another step d), an image of the first mask pattern area is alternately superimposed on the combination of the second mask pattern area and the fourth mask pattern area, whereby the first mask pattern area plus the second A combination of the second mask pattern area and the fourth mask pattern area, and an image of the third mask pattern area is superimposed on the combination of the second mask pattern area and the fourth mask pattern area. Thus, a combination of the second mask pattern area plus the third mask pattern area plus the fourth mask pattern area is generated. In another step d), images of the first mask pattern area are alternately superimposed on the combination of the second mask pattern area, the third mask pattern area and the fourth mask pattern area, thereby Of the second mask pattern area plus the second mask pattern area plus the third mask pattern area plus the fourth mask pattern area, and the image of the third mask pattern area is the first mask pattern area plus the first mask pattern area. Two mask pattern areas plus a fourth mask pattern area, so that the other first mask pattern area plus the second mask pattern area plus the third mask pattern area plus the fourth mask pattern Create a region.

  The method according to the present invention can be used equally well in an exposure apparatus using a light source that generates deep ultraviolet light (DUV), vacuum ultraviolet light (VUV), or extreme ultraviolet light (EUV). Please note that.

  An illumination system capable of changing the angular light distribution very quickly may comprise a light source configured according to the invention to emit a train of continuous projected light pulses, the pulse train being f> 1 kHz. It has an average pulse frequency. The illumination system further comprises a pupil plane and a diffractive optical element that includes a first portion and a second portion. The first portion, when illuminated by the light pulse, produces a first irradiance distribution in the pupil plane. When illuminated by the light pulse, the second portion produces a second irradiance distribution in the pupil plane that is different from the first irradiance distribution. A drive device is provided that is configured to rotate the diffractive optical element about an axis of rotation. At the first rotational position of the diffractive optical element, the first part is illuminated by the light pulse, and at a second rotational position of the diffractive optical element different from the first rotational position, the second part is Illuminated by a light pulse. Change between the first rotational position and the second rotational position between two light pulses of the pulse train, preferably between five consecutive light pulses of the pulse train, more preferably between two successive light pulses. A control unit is provided that is configured to control the light source such that occurs.

  The illumination system of the present invention is based on the perception that a diffractive optical element, which can for example have the shape of a disk or a ring, can be rotated at very high speeds around the axis of rotation, but nevertheless has high angular accuracy Therefore, a specific portion of the diffractive optical element can be accurately placed at a desired rotational position at a predetermined time. Secondly, different parts of the diffractive optical element can be placed very frequently in the beam path of the light pulse, and thus desired by simply properly synchronizing the emission of the light pulse and the rotation of the diffractive optical element. Part can be selected.

In principle, during a scan cycle it can be considered that the light pulses are emitted perfectly regularly so that the time interval between two successive light pulses is always equal. However, in that case, it is necessary to change the angular velocity of the diffractive optical element when the illumination setting changes. The angular velocity of the diffractive optical element remains constant according to the invention, since it is difficult to achieve this, especially because it is difficult to achieve this between two successive light pulses of a pulse train. It is preferable. Thus, the control unit illuminates either the first part or the second part depending on the time shift, while the diffractive optical element rotates around the axis of rotation at a constant angular velocity. Thus, the emission time of each light pulse in a single pulse train is configured to be corrected by a time shift of at least 1 / 10f, preferably 1 / 5f, more preferably 1 / 2f. According to this method, the light pulse in the time domain is changed so that the light pulse is incident on the diffractive optical element at the very moment when the part of the diffractive optical element is inserted into the optical path that generates the desired irradiance distribution. By simply shifting by a certain amount, the desired irradiance distribution in the pupil plane can be selected.
By assuming the typical dimensions of the projection light beam that illuminates the rotating diffractive optical element, and a pulse frequency of f> 1 kHz, the drive system allows the circumference of the diffractive optical element to exceed 10 m / sec. The diffractive optical element can be configured to rotate so as to move at a tangential velocity, preferably greater than 100 m / sec, more preferably greater than 200 m / sec. If the pulse frequency is high and the crossing section of the projection light beam when entering the diffractive optical element is large, even supersonic speed may be required.

In some embodiments, the rotational frequency of the diffractive optical element is equal to the average pulse frequency f. In this case, the time between two successive light pulses is sufficient (at least on average) for complete rotation of the diffractive optical element.
As mentioned above, in general, the change between the first rotational position and the second rotational position preferably occurs between two successive light pulses of the pulse train. However, it is also possible to waste a small number of light pulses during the change of rotational position. The wasted light pulse can be incident, for example, on the absorption region of the rotating support that supports the diffractive optical element. By allowing a certain number of light pulses to be wasted, additional time is provided for changing the rotational position.

  In addition, the diffractive optical element includes a plurality of first portions and a plurality of second portions. When each first portion is illuminated by a light pulse, the first irradiance distribution is obtained as a pupil plane. And each second portion, when illuminated by a light pulse, produces a second irradiance distribution in the pupil plane that is different from the first irradiance distribution. In this case, the angular frequency of the diffractive optical element can be lowered by alternately arranging the first part and the second part along the circumference of the diffractive optical element. For example, when the diffractive optical element includes n> 2 first portions and n> 2 second portions, the rotational frequency of the diffractive optical element can be lowered to f / n.

As a matter of course, the diffractive optical element has m = 3, 4, 5,. . . It is also possible for each of the m parts to be different from the irradiance distribution produced in the pupil plane by all other parts when illuminated by a light pulse. An illuminance distribution is generated in the pupil plane.
In other embodiments, the control unit may include a drive and / or a light source such that the first portion and the second portion are illuminated by light pulses during a single 360 ° rotation of the diffractive optical element. Configured to control. In this case, two different irradiance distributions are alternately generated in the pupil plane, so that finally the combination of the two illumination settings coupled to the first and second parts of the diffractive optical element. Scan integrated lighting settings are obtained.

The first portion and the second portion can be included in a circular ring that is concentric with the axis of rotation. If illuminated by a light pulse, the diffractive effect produced by the ring changes continuously while the diffractive optical element is rotating, by carefully changing the light pulse emission time, the pupil plane The irradiance distribution inside can be finely adjusted.
In one embodiment, the illumination system comprises a static diffractive optical element that emits radiation at least substantially generated by the static diffractive optical element in the pupil plane. The illumination distribution is fixedly arranged so as to be a convolution of the first or second irradiance distribution generated by the first part and the second part, respectively.
This is particularly advantageous when the static diffractive optical element generates a basic configuration of the irradiance distribution in the pupil plane. In this case, different portions of the rotating diffractive optical element can be used to correct this irradiance distribution. This modification can include changing the position and / or size of the irradiance distribution, or portions thereof in the pupil plane. According to this method, for example, the telecentricity or pole balance of the illumination setting defined by the static diffractive optical element can be adjusted variably.
Also, by using a static diffractive optical element, a first and a second light source are used that are each configured to emit a sequence of consecutive projected light pulses having an average pulse frequency of f> 1 kHz. be able to. The light pulses generated by the first and second light sources are incident on either a stationary diffractive optical element or a diffractive optical element rotating at high speed from different directions. In this case, the first and second portions of the rotating diffractive optical element can be configured to tilt the light beam, the tilted light beam being one of the diffractive optical elements. Incident at an angle. Therefore, the tilt associated with the oblique incidence is compensated by rotating the diffractive optical element at high speed.

If the first and second light sources are configured to emit light pulses alternately, an effective pulse train having an average pulse frequency of 2f can be obtained.
It is also possible to use light sources with different average pulse frequencies. In this case, first and second parts of other structures of the diffractive optical element are necessary.

The subject of the present invention is
a) generating a sequence of consecutive projected light pulses having an average pulse frequency of f> 1 kHz;
b) rotating the diffractive optical element simultaneously around a rotation axis (preferably at a constant angular velocity), the diffractive optical element comprising a first part and a second part, wherein the first part comprises When illuminated by a light pulse, it produces a first irradiance distribution in the pupil plane of the illumination system, and when the second part is illuminated by a light pulse, the first irradiance distribution is Generating a different second irradiance distribution in the pupil plane; and
c) of the light pulse in step a) so that one light pulse illuminates the first part and another light pulse, preferably a light pulse immediately following the same pulse train, illuminates the second part. Controlling the rotation of the diffractive optical element in emission and / or step b) to operate a microlithographic projection exposure apparatus.

The above remarks regarding the lighting system of the present invention apply here, mutatis mutandis.
Specifically, the diffractive optical element has a specific light pulse of the first part or the second part depending on the time shift while the diffractive optical element rotates around the rotation axis at a constant angular velocity. It can be rotated around the rotation axis at a constant angular velocity to illuminate either, and the emission time of the light pulse is at least 1 / 10f, preferably at least 1 / 5f, more preferably at least 1 / 2f Only the time shift can be corrected within a single pulse train.

The circumference of the diffractive optical element can be moved at a tangential speed of greater than 10 m / sec, preferably greater than 100 m / sec, more preferably greater than 200 m / sec. The emission time can be controlled so that the first and second portions are alternately illuminated during a single pulse train.
The first part and the second part shift laterally along the opposite direction the irradiance distribution generated in the pupil plane by the static diffractive optical element, for example to compensate for telecentricity errors Can do.
The light pulse is generated by the static diffractive optical element in the first or second irradiance distribution, wherein the irradiance distribution in the pupil plane is at least substantially generated by the first and second portions, respectively. It can additionally pass through a static diffractive optical element that is fixedly arranged to be a convolution of the generated irradiance distribution.
The train of light pulses can be generated by a first light source that emits a first train of consecutive projection light pulses and a second light source that emits a second train of consecutive projection light pulses, The one pulse train and the second pulse train are arranged alternately.
The first part can be illuminated only by light pulses of the first pulse train, and the second part is illuminated only by light pulses of the second pulse train. In this case, the first part and the second part compensate for the inclination of the projection light pulse incident from different directions into the diffractive optical element from the first and second light sources. One diffractive optical element is preferably a static diffractive optical element.

Furthermore, the first and second light sources can be configured to emit light pulses alternately so as to obtain an effective pulse train having an average pulse frequency of 2f.
According to another aspect of the invention, there is provided an illumination system for a microlithographic projection exposure apparatus comprising an optical integrator that includes an array of optical raster elements. A light beam is coupled to each light raster element. While the illumination system is operating, superimpose a combined light beam on a light raster element in a common field plane that is exactly the same as the mask plane in which the mask to be illuminated is located or is optically conjugate. A capacitor is provided. The illumination system further comprises a polarization modulator configured to modify the field dependence of the state of polarization of the projection light in the illuminated field illuminated in the mask plane by the illumination system.
By using such a polarization modulator, projection light having different polarization states can be used to illuminate different parts of the illuminated field at a given time. These polarization states are carefully adapted to the size and orientation of the features to be imaged on the photosensitive surface, resulting in improved imaging quality.

  In one embodiment, the polarization modulator comprises a plurality of modulator units. Each modulator unit is arranged in a raster field plane that includes a plurality of raster field plane elements. Each raster field plane element is imaged over the common field plane and combined in a one-to-one correspondence with one of the optical raster elements. The modulator unit is configured to variably adjust the polarization state of the light beam that is otherwise coupled to the optical raster element in at least two portions of each raster field plane element in response to the control signal. .

In one embodiment, the illumination system comprises a control unit that can control the modulator unit in such a way that different polarization states exist in the illuminated field.
In one embodiment, each modulator unit comprises two birefringent optical components having a thickness along the optical axis of the illumination system, which thickness can be varied individually. In this case, the light effect of the birefringent optical component can be varied, for example, between the light effect produced by a quarter wave plate and the light effect produced by a half wave plate.
If each optical component comprises two birefringent wedges and a drive configured to displace at least one wedge, it is easy to change the thickness of the optical component by changing the relative position of the birefringent wedge be able to.

The subject of the present invention is
a) illuminating the first mask pattern region using a first projection light having a first polarization state while the mask is moving along the scan direction;
b) A second mask pattern region disposed on a mask adjacent to the first mask pattern region along a cross scan direction orthogonal to the scan direction is different from the first polarization state. Illuminating with a second projection light having a state of polarization of: illuminating a mask in a microlithographic projection exposure apparatus.
The first projection light may have a first angular light distribution, and the second projection light may have a second angular light distribution that is different from the first angular light distribution.
The first polarization state can be changed to a third polarization state different from the second polarization state, for example, after the movement of the mask along the scan direction is completed.
Definitions The term “light” is used herein to denote any electromagnetic radiation, in particular visible light, UV light, DUV light, VUV light and EUV light and X-rays.
The term “ray” is used herein to describe light whose propagation path can be described by a line.
The term “beam” is used herein to denote a plurality of rays having a common origin in the field plane.
The term “light beam” is used herein to denote light that passes through a particular lens or other optical element.

The term “position” is used herein to represent the position of a reference point of an object in three-dimensional space. This position is usually indicated by a set of three orthogonal coordinates. Thus, orientation and position completely represent the displacement of the object in three-dimensional space.
The term “surface” is used herein to represent any planar or curved surface in three-dimensional space. This surface may be part of the object or completely independent, as is common in the case of field or pupil planes.

The term “field plane” is used herein to denote a plane that is optically conjugate to the mask plane.
The term “pupil plane” is used herein to denote the plane where peripheral rays that pass through different points in the mask plane meet. As is common in the art, the term “pupil plane” is used in the mathematical sense even if it is not actually a plane, but is slightly curved, and thus in the strict sense the pupil surface. Should be called.
The term “uniform” is used herein to denote a position independent property.

The term “light raster element” is used herein to denote any optical element, such as a lens, prism or diffractive optical element, that is arranged with other identical or similar light raster elements. Thus, each optical raster element is coupled to one of a plurality of adjacent optical channels.
The term “optical integrator” is used herein to denote an optical system that increases the product NA−a, where NA is the numerical aperture and a is the illuminated field area.

The term “condenser” is used herein to describe an optical element or optical system that establishes (at least approximately) a Fourier relationship between two planes, eg, a field plane and a pupil plane. ing.
The term “conjugate plane” is used herein to denote a plane in which an imaging relationship is established between them. For more information related to the concept of conjugate planes, see the essay E. Delano entitled: "First-order Design and the y, y Diagram", Applied Optics, 1963, vol. 2, no. 12, pages 1251-1256. Has been.
The term “field dependency” is used herein to denote some functional dependency of a physical quantity from a position in the field plane.
The term “spatial irradiance distribution” is used herein to describe how the total irradiance changes on a real or virtual surface upon which light is incident. Usually, the spatial irradiance distribution can be described by a function I _s (x, y), where x and y are the spatial coordinates of a point in the surface. When applied to the field plane, the spatial irradiance distribution inevitably integrates the irradiance generated by the multiple light fluxes.
The term “angular irradiance distribution” is used herein to describe how the irradiance of a light beam varies depending on the angle of the light rays that make up the light beam. Normally, the angular irradiance distribution can be described by _a function I _a (α, β), where α and β are angular coordinates describing the direction of the light beam. If the angular irradiance distribution is field dependent, I _a is also a function of field coordinates, ie I _a = I _a (α, β, x, y).

  The term “scan cycle” is used herein to define the process of imaging the entire mask onto the photosensitive surface in one go.

  The various features and advantages of the present invention will be more readily understood by reference to the following detailed description, taken in conjunction with the accompanying drawings, in which:

1 is a schematic perspective view of a projection exposure apparatus according to an embodiment of the present invention. It is an expansion perspective view of the 1st mask projected by the projection exposure apparatus shown by FIG. FIG. 3 is a schematic top view of the photosensitive surface at different stages during the projection of the first mask shown in FIG. 2, illustrating a first double exposure mode according to the invention. It is an expansion perspective view of the 2nd mask projected by the projection exposure apparatus shown by FIG. FIG. 5 is a schematic top view of a photosensitive surface at different stages during projection of the second mask shown in FIG. 4 illustrating a second double exposure mode according to the present invention. It is an expansion perspective view of the 3rd mask projected by the projection exposure apparatus shown by FIG. 6 is a schematic representation of the photosensitive surface at different stages during the projection of the mask shown in FIG. 6, showing a third double exposure mode, which is a combination of the two exposure modes shown in FIGS. It is a top view. FIG. 4 is a meridian cross-sectional view of an illumination system that is part of the apparatus shown in FIG. 1 and suitable for carrying out the exposure method shown in FIGS. FIG. 9 is a perspective view of a mirror array included in the illumination system shown in FIG. FIG. 9 is a perspective view of an optical integrator included in the illumination system shown in FIG. FIG. 11 is a top view of an array of optical raster elements that can be used in place of the optical integrator shown in FIG. FIG. 12 is a cross-sectional view of the array shown in FIG. 11 taken along line XII-XII. FIG. 9 is a schematic meridional section of a portion of the illumination system shown in FIG. 8, with only a mirror array, a condenser and an array of light raster elements shown. FIG. 9 is an enlarged cutaway view of FIG. 8 showing the first and second microlenses and condenser lens of the optical raster element. It is a top view of the spatial irradiance distribution on the light entrance facet of the light integrator. It is sectional drawing of the polarization manipulator in a YZ plane. FIG. 17 is a perspective view of a polarization manipulator unit included in the polarization manipulator illustrated in FIG. 16. It is a top view of an optical integrator and a polarization manipulator. FIG. 4 is a meridian cross-sectional view of an illumination system according to an alternative embodiment that is also suitable for carrying out the exposure method shown in FIGS. 4 is a meridian cross-sectional view of an illumination system according to another alternative embodiment that is also suitable for carrying out the exposure method shown in FIGS. 2 and 3. FIG. An illumination system that is part of the apparatus shown in FIG. 1 and suitable for carrying out the exposure method shown in FIGS. 4 and 5 by using a diffractive optical element that rotates at high speed. It is a meridian cross-sectional view. FIG. 5 is a top view of a photosensitive surface, a diffractive optical element rotating at high speed, a polarization manipulating element rotating at high speed, and a pupil plane at different stages during a scanning operation. FIG. 6 is a top view of a diffractive optical element rotating at high speed, a polarization manipulating element rotating at high speed, and a pupil plane according to an alternative embodiment. FIG. 5 is a top view of a fast rotating diffractive optical element, a fast rotating polarization manipulating element and a pupil plane according to yet another alternative embodiment. It is a top view of a part of a diffractive optical element rotating at high speed, including a plurality of first and second portions arranged alternately. FIG. 4 is a top view of a diffractive optical element rotating at high speed, including a plurality of portions each producing a different irradiance distribution in the pupil plane. FIG. 27 is a diagram showing an irradiance distribution when one of the portions of the diffractive optical element shown in FIG. 26 is illuminated by a projection light pulse. FIG. 4 is a top view of a diffractive optical element rotating at high speed, including a plurality of portions each producing a different irradiance distribution in the pupil plane. FIG. 29 is a diagram showing an irradiance distribution when three portions of the diffractive optical element shown in FIG. 28 are subsequently illuminated by a projection light pulse. FIG. 3 is a top view of a portion of a diffractive optical element rotating at high speed, including a circular ring that produces a continuously varying irradiance distribution in the pupil plane. FIG. 2 is a meridian cross-sectional view of an illumination system that is part of the apparatus shown in FIG. 1 and includes a stationary diffractive optical element and a diffractive optical element rotating at high speed. It is a figure which shows convolution with the irradiance distribution produced | generated by a static diffractive optical element, and the irradiance distribution produced | generated by the diffractive optical element rotated at high speed in the 1st angular position. It is a figure which shows convolution with the irradiance distribution produced | generated by a static diffractive optical element, and the irradiance distribution produced | generated by the diffractive optical element rotated at high speed in a 2nd angular position. FIG. 6 illustrates a convolution of an irradiance distribution generated by a stationary diffractive optical element with an irradiance distribution generated by a fast rotating diffractive optical element at a first angular position according to another embodiment. is there. FIG. 6 is a diagram illustrating a convolution of an irradiance distribution generated by a stationary diffractive optical element and a irradiance distribution generated by a fast rotating diffractive optical element at a second angular position according to another embodiment. is there. FIG. 2 is a meridional section of an illumination system that is part of the apparatus shown in FIG. 1 and comprises a stationary diffractive optical element and a diffractive optical element rotating at high speed and two light sources. A tatami between the irradiance distribution generated by a stationary diffractive optical element and the irradiance distribution generated by a fast rotating diffractive optical element when a light pulse is emitted by one of two light sources. FIG. The irradiance distribution generated by the stationary diffractive optical element and the irradiance distribution generated by the fast rotating diffractive optical element when the light pulse is emitted by one of the two light sources; It is a figure which shows the convolution of. 3 is a flow chart showing the important method steps of the present invention.

I. FIG. 1 is a highly simplified perspective view of a projection exposure apparatus 10 according to the present invention. The apparatus 10 comprises an illumination system 12 that generates a projection light beam. The projection light beam illuminates a field 14 on the mask 16 that includes a pattern 18 formed by a plurality of microfeatures 19 schematically shown as thin lines in FIG. In this embodiment, the illuminated field 14 has a slit-like rectangular shape and is centered on the optical axis OA of the device 10. However, other shaped illuminated fields 14 for the ring segments and placement of illuminated field 14 outside the optical axis OA are also contemplated.

  The projection objective 20 images the pattern 18 in the illuminated field 14 onto a photosensitive layer 22, for example a photoresist, supported by a substrate 24. A substrate 24, which can be formed by a silicon wafer, is arranged on a wafer stage (not shown) so that the top surface of the photosensitive layer 22 is accurately located in the image plane of the projection objective 20. The mask 16 is positioned by a mask stage (not shown) in the object plane of the projection objective lens 20. Since the projection objective 20 has a magnification β where | β | <1, a reduced image 18 ′ of the pattern 18 in the illuminated field 14 is projected onto the photosensitive layer 22.

During projection, the mask 16 and the substrate 24 move along a scan direction corresponding to the Y direction shown in FIG. The illuminated field 14 then scans over the mask 16 so that a patterned area wider than the illuminated field 14 can be continuously imaged. The ratio of the speed of the substrate 24 and the speed of the mask 16 is equal to the magnification β of the projection objective lens 20. When the projection objective 20 reverses the image (β <0), the mask 16 and the substrate 24 move in the opposite direction as indicated by arrows A1 and A2 in FIG. However, the present invention can also be used with a stepper tool where the mask 16 and substrate 24 do not move while the mask is being projected.
II. Double exposure-first embodiment

FIG. 2 is an enlarged perspective view of the mask 16. In this embodiment, the pattern 18 on the mask 16 includes three first identical mask pattern regions 181 arranged vertically along the Y direction. For simplicity, it is assumed that the features 19 of the first mask pattern region 181 are straight lines extending along the Y direction.
The pattern 18 further includes three identical second mask pattern regions 182 that are also arranged vertically along the Y direction, and the first mask pattern region. The first mask pattern region 181 is spaced laterally so that the 181 and second mask pattern regions 182 do not have a common X coordinate. It is assumed that the second mask pattern region 182 includes only features 19 that extend along the X direction.

Within one complete scanning cycle of the mask 16, three first exposure pattern areas associated with the first mask pattern area 181 and three second exposure pattern areas associated with the second mask pattern area 182. Can be generated. The scanning direction is then reversed or the mask 16 is returned to its original position without any illumination and another scanning cycle is performed.
It is further assumed that the die produced by the apparatus 10 has a size corresponding to one of the exposure pattern areas. According to this method, two rows of dies can be simultaneously exposed on the substrate 24.
When maximum image quality is desired, different patterns usually require different angular irradiance distributions at the mask level. In this embodiment, it is assumed that the feature 19 extending along the Y direction is best imaged on the photosensitive layer 22 using the X dipole illumination setting. In FIG. 2, the pupil 261 associated with the light beam that converges toward a field point located in one of the first mask pattern regions 181 is indicated by a dashed circle. Within the pupil 261, two poles 27 spaced apart along the X direction represent the direction in which light propagates toward the field point. Since the pattern is assumed to be uniform throughout the first mask pattern region 181, this X dipole illumination setting can be generated at each field point in the first mask pattern region 181. is necessary.

The second mask pattern region 182 includes only features extending along the X direction. For these features 19, it is assumed that the best image quality is obtained with the Y dipole illumination setting. In FIG. 2, the pupil 262 associated with the light beam converging toward a field point located in one of the second pattern areas 182 is also indicated by a dashed circle. Within the pupil 262, two poles 28 spaced along the Y direction represent the direction in which light propagates toward the field point. Since the pattern is assumed to be uniform across the second pattern region 182, this Y dipole illumination setting needs to be generated at each field point in the first pattern region 182. is there.
This implies that the lighting system 12 must be able to generate two different lighting settings in the illuminated field 14 simultaneously and side by side. The structure of the lighting system 12 that can perform this task is described in more detail in Section V with reference to FIGS.

  The mask 16 shown in FIG. 2 is assumed to be used in a double exposure process according to the first embodiment of the present invention, which will be described below with reference to FIGS. .

In a double exposure process, each die is exposed twice prior to performing subsequent manufacturing steps such as etching. This usually requires the use of two different masks. In the first scanning process, the pattern contained in the first mask is transferred to the photosensitive surface 22. The first mask is then replaced with the second mask. In the second scanning process, the pattern contained in the second mask is transferred to the photosensitive surface 22.
Double exposure is advantageous because the patterns contained in the first and second masks can be individually illuminated using optimal illumination settings. This promotes an improvement in the contrast of the photosensitive surface since all projected light contributes to image formation. However, when using two different masks in a double exposure process, it is difficult to reduce the overlay error to an acceptable range. Since masks are manufactured and tested in different processes, it is challenging to ensure that the patterns contained in the two masks correspond completely to each other.

3a-3c illustrate the process steps in which double exposure is performed using only the mask 16. FIG. In FIG. 3 a, which is a schematic top view of the photosensitive surface 22 during a scanning operation, the reference number display 301 represents a first exposure pattern region that is arranged vertically along the scanning direction Y. Each first exposure pattern area 301 is an image in the first mask pattern area 181. The reference number display 302 represents a second exposure pattern area that is also arranged vertically along the scanning direction Y and is arranged beside the first exposure pattern area 301. Each second exposure pattern region 302 is an image of the second mask pattern region 182.
While the illuminated field 14 is simultaneously scanning over the first mask pattern region 181 and the second mask pattern region 182, a new pair of first and second exposure pattern regions 301, 302 is formed on the photosensitive layer 22. Generated on top of. During the scanning operation, the substrate 24 supporting the photosensitive layer 22 moves along the Y direction, as indicated by the arrows in FIG. 3a.
When the three first and second mask pattern areas 181, 182 are transferred to the photosensitive surface 22, a new series of three pairs of first and second exposure pattern areas are generated on the photosensitive surface 22. So that the direction of movement of the mask 16 can be reversed, or the mask 16 is quickly returned to its starting position.
When the first scanning movement of the substrate 24 is completed, the first row R1 of the first exposure pattern region 301 and the second row R2 of the second exposure pattern region 302 are completed as shown in FIG. 3b. Produced on the photosensitive surface 22.

  Next, the substrate 24 supporting the photosensitive surface 22 is displaced laterally by one row along the X direction, and the scanning operation is repeated using the same mask 16. However, since the image is displaced in the horizontal direction by one row, the image of the first mask pattern region 181 is superimposed on the second exposure pattern region 302 forming the second row R2. Therefore, the dies in the second row R2 are exposed twice, that is, the first time using the image of the second mask pattern region 182 and the second time using the image of the first mask pattern region 181. Is done. In FIG. 3b, 301 + 302 indicates a superimposed image. Each image contributing to the superimposed image is represented by the optimal illumination setting shown in FIG. 2, namely the X dipole for the first mask pattern region 181 and the Y dipole for the second mask pattern region 182. Generated using child lighting settings.

Although the image of the second mask pattern region 182 is not superimposed, the next row R3 of the second exposure pattern region 302 is formed.
This process of displacing the substrate 24 by one row each time a new row of the second exposure pattern area 302 is completed is repeated until the entire available photosensitive surface is exposed. FIG. 3 c shows the completion of the last row R 7 of the second exposure pattern area 302.
The first row R1 and the last row R7 are not exposed twice. These rows R1, R7 can therefore be placed outside the part of the photosensitive surface used to produce the die.
III. Double exposure-second embodiment

FIG. 4 is an enlarged perspective view of the mask 216 according to the second embodiment. The pattern 218 on the mask 216 includes two first identical mask pattern regions 2181 arranged side by side along the X direction, ie, the cross scan direction. For simplicity, it is assumed that the feature 19 of the first mask pattern region 2181 is a straight line extending along the Y direction.
The pattern 218 further includes two identical second mask pattern regions 2182, which are also arranged adjacent to each other along the X direction, and the first mask pattern region 2181. The second mask pattern region 2182 is spaced apart from the first mask pattern region 2181 along the scanning direction Y so that the second mask pattern region 2182 does not have a common Y coordinate. It is assumed that the second mask pattern region 2182 includes only the features 19 extending along the X direction.

Within one complete scanning cycle of mask 216, two first exposure pattern areas associated with first mask pattern area 2181 and two second exposure pattern areas associated with second mask pattern area 2182 Can be exposed. The scanning direction is then reversed or the mask 216 is returned to its original position without any illumination and another scanning cycle is performed.
Also in this embodiment, the feature 19 extending along the Y direction is best sensitive using the X dipole illumination setting, as shown in FIG. 4 by the two poles 27 in the pupil 261. It is assumed that an image is formed on the layer 22. The feature 19 of the second mask pattern region 2182 extends along the X direction and is therefore best using the Y dipole setting, as shown in FIG. 4 by the two poles 28 in the pupil 262. To be imaged.
Since the pattern is assumed to be uniform across the first and second pattern regions 2181, 2182, the X and Y dipole illumination settings are within the first and second pattern regions 2181, 2182, respectively. Must be generated at each field point.
This implies that the lighting system 12 must be able to change extremely quickly between two different lighting settings. Assuming typical values of scan speed and distance between dies along the scan direction Y, the time required to change the illumination setting should be less than 5 milliseconds, preferably less than 1 millisecond. The structure of the lighting system 12 that can perform this task is described in more detail in Section VI with reference to FIGS.

The mask 216 shown in FIG. 4 is assumed to be used in a double exposure process according to the second embodiment, which will be described below with reference to FIGS.
5a-5c illustrate a series of process steps in which double exposure is performed using only mask 216. FIG. In FIG. 5 a, which is a schematic top view of the photosensitive surface 22 during a scanning operation, the reference number display 2301 represents a first exposure pattern region that is arranged side by side along the cross-scan direction X. Each pair of first exposure pattern areas 2301 is an image of a pair of first mask pattern areas 2181.

The reference number display 2302 represents a second exposure pattern region that is also arranged adjacently along the cross scan direction X. Each second exposure pattern area 2302 is an image in the second mask pattern area 2182.
While the illuminated field 14 simultaneously scans first over the first mask pattern area 2181 and then over the second mask pattern area 2182, the first and second exposure pattern areas 2301, 2302 A new pair is created on the photosensitive layer 22. During the scanning operation, the substrate 24 supporting the photosensitive layer 22 moves along the Y direction, as indicated by the arrows in FIG. 5a. When the pattern 219 on the mask 16 is completely transferred to the photosensitive surface 22, the direction of movement of the mask 16 is reversed in this embodiment. Therefore, when ignoring the first exposure pattern region and the last exposure pattern region in each row, the two first exposure pattern regions and the two second exposure pattern regions 2301 and 2302 are along the scan direction Y. Arranged vertically. Alternatively, the mask 216 can be quickly returned to its starting position so that a new series of two pairs of first and second exposure pattern regions 2301, 2302 can be generated on the photosensitive surface 22. .

  When the first scanning movement of the substrate 24 is completed, two identical rows R1, R2 of the exposure pattern areas 2301, 2302 are generated on the photosensitive surface 22. In each row R1, R2, except for the first region and the last region in each row, two first exposure pattern regions and two second exposure pattern regions 2301, 2302 are alternately arranged along the scanning direction Y. Is done.

Next, the mask 216 is displaced longitudinally by one length, and the scanning operation is repeated using the same mask 216. However, due to the displacement of the mask 216, the image of the first mask pattern area 2181 is superimposed on the second exposure pattern area 2302, and the image of the second mask pattern area 2182 is superimposed on the first exposure pattern area 2301. Is superimposed on the top. In FIG. 5b, the region with the superimposed image is indicated by 2301 + 2302. Therefore, the dies in the rows R1 and R2 are used twice, that is, the first time using the image of the first mask pattern area 2181 and the second time using the image of the second mask pattern area 2182. The same applies to exposure and vice versa. Each image is formed using the optimal illumination settings shown in FIG. 4, ie, the X dipole for the first mask pattern region 2181 and the Y dipole illumination setting for the second mask pattern region 2182. Is done.
When the two rows R1, R2 are completely exposed twice, the substrate 26 is displaced laterally by two rows along the cross scan direction X and the same process is repeated. These steps are performed until the entire available photosensitive surface 22 is exposed. FIG. 5c shows the completion of the last two rows R5, R6.
IV. Double exposure-third embodiment

FIG. 6 is an enlarged perspective view of the mask 316 according to the third embodiment. The pattern 318 on the mask 316 includes a first mask pattern region 3181 that includes only features 19 that extend along the Y direction, and a second mask pattern region 3182 that includes only features 19 that extend along the X direction. The third mask pattern region 3183 includes only the features 19 extending in parallel to the −XY diagonal, and the fourth mask pattern region 3184 includes only the features 19 extending in parallel to the XY diagonal. A pupil 326, each containing poles 27, 28, 31 and 32, represents an illumination setting that ensures optimal imaging of the features 19 in each mask pattern region 3181-3184.
The four mask pattern areas 3181 to 3184 are arranged in pairs along the X direction and the Y direction so that each mask pattern area is directly adjacent to the X direction and the Y direction.

  The mask 316 shown in FIG. 6 is actually a quadruple exposure process that is a combination of the first embodiment shown in FIGS. 2 and 3 and the second embodiment shown in FIGS. It is assumed that

  Hereinafter, this quadruple exposure process will be described with reference to FIGS. 7a to 7d showing a series of process steps in which quadruple exposure is performed using only the mask 316. FIG. In FIG. 7 a, which is a schematic top view of the photosensitive surface 22 during the scanning operation, the reference number display 3301 represents a first exposure pattern region that is an image of the first mask pattern region 3181, and the reference number display 3302 A second pattern area that is an image of the second mask pattern area 3182 is represented, and a reference number display 3303 represents a third exposure pattern area that is an image of the third mask pattern area 3183, and a reference number display. Reference numeral 3304 denotes a fourth exposure pattern region 3304 which is an image of the fourth mask pattern region 3184. In FIG. 7a, it can be seen that by repeating a plurality of scan cycles, groups of four exposed pattern regions 3301-3304 corresponding to the two rows R1, R2 of the die are repeatedly generated on the photosensitive surface 22. In the row R1, the first exposure pattern region 3301 and the third exposure pattern region 3303 are alternately arranged. In the row R2, the second exposure region 3302 and the fourth exposure region 3034 are along the scanning direction Y. Alternately arranged. Of course, also in this embodiment, each of the four mask pattern regions 3181-3184 is illuminated using the optimal illumination settings shown in FIG. This also applies to subsequent exposures described below.

FIG. 7b shows the next process step in which a second exposure is performed in the same way as the second embodiment shown in FIGS. 5a-5b, using the same mask 316. FIG. While the substrate 24 supporting the photosensitive surface 22 is moving along the opposite direction here, the image of the third mask pattern region 3183 is superimposed on the first exposure pattern region 3301, At the same time, the image of the fourth mask pattern area 3184 is superimposed on the second exposure pattern area 3302. Similarly, the image of the first mask pattern area 3181 is superimposed on the third exposure pattern area 3303, and at the same time, the image of the second mask pattern area 3182 is superimposed on the fourth exposure pattern area 3304. Superimposed. When this scanning operation is finished, a first row R1 including an overlap 3301 + 3303 of the image of the first mask pattern region 3181 and the image of the third mask pattern region 3183 is obtained, and the second mask pattern region 3182 A second row R2 is obtained that includes a superimposition 3302 + 3304 of the image and the image of the fourth mask pattern region 3184.
The process so far can be regarded as a double exposure process according to the second embodiment described with reference to FIGS. This process is then combined with the process described above with reference to FIGS. Thus, prior to the next scan cycle, the substrate 24 supporting the photosensitive surface 22 is displaced by one row along the cross scan direction X, as shown in FIG. 7c. During the subsequent scan cycle, the images of the first and third mask pattern areas 3181 and 3183 are alternately superimposed on the combined exposure pattern areas 3302 + 3304. As a result, a row R2 which is a series of exposure pattern regions 3302 + 3304 + 3301 and 3302 + 3304 + 3303 is obtained. Further, a third row R3 in which the second exposure pattern region 3302 and the fourth exposure pattern region 3304 are alternately arranged is generated.

Next, the moving direction of the substrate 24 supporting the photosensitive surface 22 is reversed once again, and the image of the first mask pattern area 3181 is superimposed on the combined exposure area 3302 + 3304 + 3303, and alternatively, the third mask. The image of the pattern area 3183 is superimposed on the combined exposure pattern area 3303 + 3304 + 3301. In any case, this results in a complete overlay 3303 + 3304 + 3301 + 3302 of all four images of the mask pattern regions 3181-3184 as can be seen from row R2 in FIG. 7d.
V. Illumination system for the first embodiment

Hereinafter, an illumination system that can simultaneously generate two different angular light distributions on the mask 16 side by side will be described. Such an illumination system is required to implement the first embodiment of the double exposure scheme described above with reference to FIGS. 2 and 3.
1. General structure

FIG. 8 is a meridian cross-sectional view of the illumination system 12 shown in FIG. For clarity, the illustration of FIG. 8 is significantly simplified and not to scale. This implies, inter alia, that different optical units are represented by only one or very few optical elements. In practice, these units can comprise a large number of lenses and other optical elements.
The illumination system 12 includes a housing 29 and a light source 33, which in the illustrated embodiment is implemented as an excimer laser. The light source 33 emits projection light having a wavelength of about 193 nm. Other types of light sources 33 and other wavelengths such as 248 nm or 157 nm are also contemplated.
In the embodiment shown, the projection light emitted by the light source 33 is incident on a beam expansion unit 34 that outputs an expanded and substantially collimated light beam 35. To that end, the beam expansion unit 34 can comprise a plurality of lenses or can be realized, for example, as a mirror structure.

The projection light beam 35 then enters a pupil defining unit 36 that is used to generate a variable spatial irradiance distribution in the subsequent pupil plane. To that end, the pupil defining unit 36 comprises a mirror array 38 containing a very large number of very small mirrors 40 that can be individually tilted around two orthogonal axes with the aid of an actuator. FIG. 9 is a perspective view of the mirror array 38, showing how the two parallel light beams 42 and 44 are reflected in different directions according to the tilt angle of the mirror 40 on which the light beams 42 and 44 are incident. 8 and 9, the mirror array 38 includes only 6 × 6 mirrors 40, but in practice the mirror array 38 can include hundreds or even thousands of mirrors 40.
The pupil defining unit 36 further comprises a prism 46 having a first flat surface 48a and a second flat surface 48b, both inclined with respect to the optical axis OA of the illumination system 12. Light incident on these inclined surfaces 48a and 48b is reflected by total internal reflection. The first surface 48a reflects incident light toward the mirror 40 of the mirror array 38, and the second surface 48b directs light reflected by the mirror 40 toward the flat exit surface 49 of the prism 46. Lead. Accordingly, the angular irradiance distribution of light emerging from the exit surface 49 can be individually varied by tilting the mirror 40 of the mirror array 38. For further details regarding the pupil defining unit 38, see US Patent Application Publication No. 2009/0116093.

  The angular irradiance distribution generated by the pupil defining unit 36 is converted into a spatial irradiance distribution with the aid of the first capacitor 50. The capacitor 50, which may be omitted in other embodiments, guides incident light toward the optical integrator 60. The light integrator 60 comprises a first array 70 and a second array 72 of light raster elements 74 in this embodiment. FIG. 10 is a perspective view of the two arrays 70 and 72. Each of the arrays 70 and 72 includes a pair of cylindrical lenses 71 and 73 extending along the X direction and the Y direction, respectively, on both sides of the support plate. The order of the cylindrical lenses 71 and 73 can be reversed on at least one side of the arrays 70, 72. Each volume defined by the intersection of two orthogonal cylindrical lenses 71, 73 forms an optical raster element 74. Thus, each light raster element 74 can be viewed as a microlens having a cylindrical curved surface. The use of a cylindrical lens is particularly advantageous when the refractive power of the optical raster element 74 must be different along the X and Y directions.

FIGS. 11 and 12 show a top view and a cross-sectional view along line XII-XII, respectively, of a first array 70 ′ of light raster elements 74 ′ according to alternative embodiments. Here, the optical raster element 74 ′ is formed by a plano-convex lens having a rectangular outline. The second array differs from the first array 70 ′ only in the curvature of the convex surface of the optical raster element 74 ′.
Referring once again to FIG. 8, the light raster elements 74 of the first and second arrays 70, 72 are respectively one light raster element 74 of the first array 70 and one light raster element 74 of the second array 72. Are arranged side by side in such a way that they are combined in a one-to-one correspondence. Two optical raster elements 74 that are associated with each other are aligned along a common axis and define an optical channel. In the optical integrator 60, a light beam propagating in one optical channel does not cross or overlap with a light beam propagating in another optical channel. In other words, the optical channels coupled to the optical raster element 74 are optically isolated from each other.

  The front focal plane of the condenser 78 forms a pupil plane 76 of the illumination system 12. In this embodiment, the pupil plane 76 is located behind the second array 72, but could equally be located in front of the second array 72. The second capacitor 78 thus establishes a Fourier relationship between the pupil plane 76 and the field stop plane 80 where the adjustable field stop 82 is located.

The field stop plane 80 is optically conjugate to the raster field plane 84 that is located in or near the first array 70 of the light integrator 60. This is because each raster field plane element, i.e., the region of the raster field plane 84 within the optical channel, is imaged across the field stop plane 80 by the associated optical raster element 74 and the second capacitor 78 of the second array 72. Means that. The image of the irradiance distribution in the raster field plane element is superimposed in the field stop plane 80, so that its very homogeneous illumination is obtained. This process is often described by identifying illuminated areas in the light channel using secondary light sources that commonly illuminate the field stop plane 80.
The polarization modulator 77 is placed between the two arrays 70, 72, or at or near another location in the raster field plane 84. The polarization modulator 77 is configured to correct the field dependence of the state of polarization of the projection light within the illuminated field 14. To that end, the polarization modulator 77 variably modifies the polarization state of the light beam that is otherwise coupled to the optical raster element 74 in at least two portions of each raster field plane element in response to the control signal. A plurality of individual modulator units 79 are configured. This is described in more detail below with reference to FIGS. 14 and 16-18.
The field stop plane 80 is imaged by the field stop objective 86 on a mask plane 88 on which the mask 16 is placed with the aid of a mask stage (not shown). An adjustable field stop 82 is also imaged in the mask plane 88 and defines at least a short lateral side of the illuminated field 14 extending along the scan direction Y.

  The pupil defining unit 36 and the polarization modulator 77 are connected to a control unit 90, which is connected to an overall system control 92, shown as a personal computer. The control unit 90 provides the modulator 40 of the pupil defining unit 36 and the modulator of the polarization modulator 77 in such a way that the desired field dependence of the angular irradiance distribution and also the field dependence of the state of polarization in the mask plane 88 are obtained. The unit 79 is configured to be controlled.

The manner in which this is achieved will now be described with reference to FIGS.
2. Lighting system functions and controls a. Spot formation

FIG. 13 is a schematic cut-away view from FIG. 8 showing several optical raster elements 74 of the first array 70 of mirror array 38, first capacitor 50 and optical integrator 60. In this embodiment, the light raster element 74 is shown as a rotationally symmetric biconvex lens with a rectangular border for simplicity.
Each mirror 40 of the mirror array 38 generates a light beam that illuminates a small spot 98 of diameter D on the light entrance facet 100 of one of the light raster elements 74. The position of the spot 98 can be changed by tilting the mirror 40. The geometric structure of the spot 98 is determined, inter alia, by the optical properties of the mirrors 40 of the mirror array 38. If there is an additional microlens array between the beam expanding unit 34 and the pupil defining unit 36 that avoids projection light being absorbed in the gap between the mirrors 40, such microlens array is also a spot 98 geometry. Will affect the structure. In some embodiments, the geometry of the spot 98 is circular, and in other embodiments, the geometry can be generally rectangular, and in particular rectangular.
As can be seen from FIG. 13, the diameter D of the spot 98 is smaller than the diameter of the light entrance facet 100 of the illuminated light raster element 74. Typically, the total area of each spot 98 illuminated on the light entrance facet 100 of the light raster element 74 should be significantly smaller than the area of each light entrance facet 100, for example at least one fifth, preferably at least It should be 1/10, more preferably at least 20 times. If the light entrance facets 100 have different areas and each spot 98 can be generated on any of these facets, the maximum area of the light entrance facets 100 can be referenced. it can.
If the spot 98 is sufficiently small compared to the light entrance facet 100 of the light raster element 74, a different light pattern can be generated on the light entrance facet 100. The light pattern can be easily changed by appropriately controlling the mirror element 40 with the assistance of the control unit 90.
b. Effect of changing the light pattern

Since the light entrance facet 100 is located in or near the raster field plane 84, the light pattern on the light entrance facet 100 is routed through the light raster elements 74 and the second capacitor 78 of the second array 72. Then, an image is formed on the intermediate field plane 80.
This will be described below with reference to FIG. 14, which is an enlarged view of the cutout of FIG. 8 showing a portion of the optical integrator 60, the second capacitor 78, and the intermediate field plane 80, not on scale. For simplicity, only two pairs of light raster elements 74 of light integrator 60 are shown in FIG. The two optical raster elements 74 coupled to a single optical channel are referred to below as the first microlens 101 and the second microlens 102. As mentioned above, the microlenses 101, 102, also referred to as field and pupil honeycomb lenses, can be configured as each microlens having, for example, a rotationally symmetric refractive surface and a rectangular boundary, or is shown in FIG. It can be configured as an intersecting cylindrical microlens. All that is required is that the microlenses 101, 102 have a non-zero refractive power along at least one direction orthogonal to the optical axis OA of the illumination system 12.

  Each pair of microlenses 101, 102 coupled to a particular light channel creates a secondary light source 106 in the pupil plane 76. In the upper half of FIG. 14, it is assumed that the convergent light beams L1a, L2a, and L3a indicated by the solid line, the dotted line, and the broken line are incident on different points of the light entrance facet 100 of the first microlens 101. . When passing through the two microlenses 101 and 102 and the condenser 78, the light beams L1a, L2a and L3a converge to the focal points F1, F2 and F3, respectively. Thus, from the top half of FIG. 14, the positions where the light rays enter the light entrance facet 100 and the positions where they pass through the intermediate field plane 80 (or any other conjugate field plane) are optically conjugate. Becomes clear.

  The lower half of FIG. 14 shows a case where the parallel light beams L1b, L2b, and L3b are incident on different regions of the light entrance facet 100 of the first microlens 101. This is a more realistic case because the light incident on the light integrator 60 is typically substantially truly collimated. The light beams L1b, L2b and L3b are focused on a common focal point F located in the second microlens 102, and then collimated once again and pass through the intermediate field plane 80. Again, as a result of optical conjugation, it can be seen that the regions where the light beams L1b, L2b and L3 are incident on the light entrance facet 100 correspond to the regions illuminated in the intermediate field plane 80.

  Thus, the size of the field illuminated in the intermediate field plane 80 (and thus the field 14 illuminated in the mask plane 88) is changed by changing the area illuminated on the light entrance facet 100 of the first microlens 101. Can be changed. The size and geometry of this region can be changed very effectively by reassembling the spot 98 with the aid of the mirror array 38 of the pupil defining unit 36, as described above with reference to FIG. it can.

Of course, these considerations apply individually to the X and Y directions as well. Thus, the geometric structure of the illuminated field 14 can be changed independently for the X and Y directions by individually changing the illumination of the light entrance facet 100 for the X and Y directions, respectively. In other words, if the area illuminated on the light entrance facet 100 of the first microlens 101 is properly determined, almost any arbitrary geometry of the field illuminated in the intermediate field plane 80 can be obtained. it can.
For example, as shown in the lower half of FIG. 14, when one of the first microlenses 101 is illuminated with the light beam L1b and the other first microlens 101 is illuminated with the light beam L2b, 2 Secondary light sources associated with the two different first microlenses 101 will illuminate different portions in the intermediate field plane 80. However, a very important point in this regard is that two different first microlenses 101 can be placed at different positions, so that the associated secondary light source illuminates the intermediate field plane 80 from different directions. . In other words, in this case, different parts in the intermediate field plane 80 are illuminated with different angular light distributions.
FIG. 15 is an exemplary top view of the light entrance facet 100 of the plurality of first microlenses 101 of the optical integrator 60. The light entrance facet 100 is shown in an enlarged size along the scan direction Y for illustrative purposes. In practice, the geometric structure of the light entrance facet substantially corresponds to the geometric structure of the illuminated field 14.

Some of the light entrance facets 100 are illuminated using the first light pattern 108 with the assistance of the pupil defining unit 36. Each first light pattern 108 includes a transverse strip 110 that extends along the Y direction and covers half of the light entrance facet 100 (when viewed along the X direction). The light entrance facet 100 illuminated using the first light pattern 108 forms a light integrator 60 such that they form two generally circular or elliptical poles PX1, PX2 spaced along the X direction. Located on the entrance surface of the.
The other light entrance facet 100 is illuminated using a second light pattern 114 that also includes only transverse elongated stripes 116 extending along the Y direction. However, the elongated stripe 116 of the second light pattern 114 is above the other half of the light entrance facet 100 (when viewed along the X direction) as compared to the elongated stripe 110 of the first light pattern 108. positioned. The light entrance facets 100 illuminated using the second light pattern 114 are arranged such that they form two generally circular poles PY1, PY2 spaced along the Y direction.

  The first microlens 101 whose light entrance facet 100 is illuminated using the first light pattern 108 is lateral to the first light pattern 108 in the intermediate field plane 80 and thus in the mask plane 88. A portion having the same geometric structure as the elongated filament 110 is illuminated in common. Since these light entrance facets 100 are arranged on the entrance side of the light integrator 60 in the form of two poles PX1, PX2, the secondary light source associated with these light entrance facets 100 is an X dipole illumination. Because of the setting characteristics, these portions (only) are illuminated obliquely from the X direction. Accordingly, the first portion 181 on the mask 16 is illuminated using an X dipole illumination setting as shown in FIG.

The light entrance facet 100 illuminated using the second light pattern 114 contributes to the illumination of the other half of the elongated strip in the illuminated field 14. Since these light entrance facets 100 are arranged on the entrance side of the light integrator 60 so that they form two poles PY1, PY2, the other half in the illuminated field 14 is also shown in FIG. Illuminated using the Y dipole illumination setting as is.
Thus, two different lighting settings can be generated simultaneously in different parts of the illuminated field 14.
Thus, with the assistance of the control unit 90, by appropriately repositioning the spot 98 on the light entrance facet 100 of the light integrator 60, almost any field dependence of the angular light distribution is in the mask plane 88. Can be generated. The time required to change the field dependence of the angular light distribution depends primarily on how quickly and precisely the mirror 40 of the mirror array 38 can be tilted by the required tilt angle.
c. Polarization control

As already mentioned, the illumination system 12 can illuminate the mask pattern area 181 on the one hand and the mask pattern area 182 on the other hand using projection light having different polarization states. . It is known in the art that not only the angular light distribution, but also the state of polarization of the projection light at the mask level has an effect on the quality of the image formed on the photosensitive surface 22, especially on the contrast. Therefore, it is desirable to be able to illuminate different portions on the mask 16 using projection light having different polarization states.
This can be achieved in the illumination system 12 with the aid of the polarization modulator 77 briefly described above. The polarization modulator 77 comprises a plurality of modulator units 79 arranged in or near the raster field plane 84.

The modulator unit 79 will first be described with reference to FIG. 14 where two modulator units 79 are shown. Each modulator unit 79 is located in the immediate vicinity of the raster field plane 84. Assuming typical dimensions, this implies that the distance from the raster field plane 84 is less than 3 mm. Each modulator unit 79 is coupled to a single optical channel, in other words, to a single raster field plane element defined as the region of the optical channel in raster field plane 84.
Each modulator unit 79 can variably modify the state of polarization in at least two portions of each raster field plane element in response to a control signal provided by the control unit 90. Since each raster field plane element is optically conjugate to the entire field 14 illuminated on the mask, the modulator unit 97 may produce different polarization states in different parts of the illuminated field 14. it can.

In order to modify the state of polarization differently in the two parts of each raster field plane element, each modulator unit 79 comprises two birefringent components 79a and 79b with variable thickness. Therefore, by appropriately arranging the optical axis of the birefringent material, the state of polarization of the projection light that passes through the first component 79a on the one hand and the second component 79b on the other hand is modified differently. be able to.
For example, the thickness of the birefringent components 79a, 79b is selected so that circularly polarized light is converted to s-polarized light by the first birefringent component 79a and also converted to p-polarized light by the second birefringent component 79b. can do. As best seen in the lower part of FIG. 14, the projection light emerging as s-polarized light from the first birefringent component 79a illuminates exclusively half of the illuminated field and also the second birefringent component. The light emerging from the refractive component 79b exclusively illuminates the other half of the illuminated field. According to this method, different portions on the illuminated field can be illuminated using projection light having different polarization states.

  In order to be able to change the thickness of the birefringent components 79a, 79b, each of the birefringent components is displaced in a direction orthogonal to the optical axis of the illumination system with the aid of a drive device. Two birefringent wedges that can be included. This is shown schematically in FIGS. FIG. 16 shows two wedges 81 a, 81 b arranged between two arrays 70, 72 of the optical integrator 60 in meridian section, and FIG. 17 shows two pairs of wedges forming a modulator unit 79. FIG. 18 is a top view of the optical integrator 60 and the plurality of modulator units 91 as viewed along the optical axis from the light source side. As a matter of course, the wedges 81 a and 81 b can be arranged in front of the array 70.

The wedge 81 associated with the birefringent components 79a, 79b is displaced along the direction shown in FIG. 16 by a double arrow, and the thickness along the direction of the optical axis is under the assistance of the drive 83. Can be changed.
17 and 18, it can be seen that each modulator unit 79 with four wedges 81 extends along the scan direction Y over the entire length of the optical integrator 60. FIG. Thus, the thickness of the birefringent components 97a, 97b changes only along the cross scan direction X and does not change along the scan direction Y. This ensures that the modulator unit 79 has the same effect on the state of polarization within all the optical channels shown in dashed lines in FIG.
3. Alternative embodiments

In the following, two alternative embodiments of the illumination system 12 will be described with reference to FIGS. 19 and 20, which are equally suitable for simultaneously generating different angular light distributions in different mask pattern regions.
a. Diffractive optical element

In the illumination system 12 shown in FIG. 8, the different angular light distributions on the first and second mask pattern regions 181, 182 are on the light entrance facet 100 of the light integrator 60 with the aid of the pupil definition unit 36. This is due to the different irradiance distributions generated. These irradiance distributions are corrected by repositioning the light spot 98 generated by the mirror 40 of the mirror array 38.
FIG. 19 shows an embodiment in which the pupil defining unit 36 includes a diffractive optical element 96 rather than a mirror array 38. As can be seen from the enlarged cutout 98, the diffractive optical element 96 includes a plurality of minute diffractive structures 100. The diffractive optical element 96 can be implemented as a computer generated hologram (CGH), for example, as is known in the art as such.

In the illustrated embodiment, the diffractive optical element 96 comprises a plurality of portions 96a, 96b having different diffractive properties. The diffractive optical element 96 is received in a turret holder 102 that is configured to rotate the diffractive optical element 96 about a rotational axis with the aid of a drive device 104. Thus, one of the portions 96a, 96b can be inserted into the optical path of the projection light at a given time. Of course, other types of replacement holders can be used for this purpose.
Depending on the portion 96a or 96b inserted in the light path, a different irradiance distribution is generated on each light entrance facet 100 of the light integrator 60, as described above with reference to FIGS. Thus, the illumination system 12 can also generate different field-dependent angular light distributions at the mask level and modify these angular light distributions by inserting other parts of the diffractive optical element 96. Can do.
The use of a pupil defining unit 36 that includes a diffractive optical element 96 instead of the mirror array 38 provides less flexibility with respect to simultaneously generating different angular light distributions on the mask 16. On the other hand, the use of a diffractive optical element 96 to define the irradiance distribution in the pupil plane 76 is an established technique that requires less complex control schemes than the mirror array 38.
b. Combining the pupil irradiance distribution

FIG. 20 illustrates different techniques that can be used to simultaneously generate two different angular light distributions in different mask pattern regions.
The illumination system 12 shown in FIG. 20 includes two conventional pupil defining units 38a, 38b that each include diffractive optical elements 96a and 96b, respectively, and zoom optics shown by lenses 50a and 50b, respectively. , Respectively, a pair of axicon elements 108a and 110a and a pair of 108b and 110b, optical integrators 60a and 60b, and polarization manipulators 77a and 77b, respectively. As is known in the art, axicon elements 108a, 108b and 110a, 110b have complementary conical surfaces. By changing the distance between these surfaces, the light can be shifted radially by a variable amount, including no shift at all when the conical surfaces are in contact with each other.
The two condenser lenses 78a, 78b establish on the one hand a Fourier relationship between the two pupil planes 76a, 76b and on the other hand establish a common intermediate field plane 80. The prisms 112a, 112b, which are formed by flat parallel plates or formed as rhomboid prisms, allow the fields illuminated by the two pupil planes 76a, 76b to be joined almost seamlessly in the intermediate field plane 80. The projection light beams emerging from the light integrators 60a and 60b are used to offset in the lateral direction. Thus, half of the field in the intermediate field plane 80 is illuminated by the first pupil plane 76a and the other half is illuminated by the second pupil plane 76b.

The intermediate field plane 80 is imaged on the mask 16 via the objective lens 86 as in the previously described embodiment.
Because the pupil defining units 38a, 38b can be individually controlled, different irradiance components can be generated in the pupil planes 76a, 76b, and thus the mask as shown in FIG. Different angular light distributions are produced in two parts, half each in the field 14 illuminated on 16.

In the embodiment shown, the diffractive optical elements 96 a, 96 b are illuminated by a common light source 33. The light emitted from the light source 33 is reflected toward the beam splitting mirror 115 by the first beam splitting mirror 117a. The beam splitting mirror 115 guides half of the light intensity toward the second flat folding mirror 117b that illuminates the diffractive optical element 96a of the first pupil defining unit 36a, and directly directs the other half of the light intensity to the first. Guide to the diffractive optical element 96b of the second pupil defining unit 36b. However, it is also conceivable to use two light sources 33 so that each light source 33 actually illuminates only half of the field 14 on the mask 16. This has the advantage that the photosensitive surface 22 can be exposed using twice the light energy and thus the output of the projection exposure apparatus 10 can also be doubled.
VI. Illumination system for the second embodiment

In the following, an illumination system that can generate different angular light distributions on the mask 16 that change very rapidly will be described. Such an illumination system is required to implement the second embodiment of the double exposure scheme described above with reference to FIGS. 4 and 5.
1. General structure

FIG. 21 is a meridional section through an illumination system 12 that is also suitable for use in the projection exposure apparatus 10 shown in FIG. The lighting system 12 has the same structure as the lighting system shown in FIG.
In this embodiment, it is assumed that the light source 33 is configured to emit a train of consecutive projection light pulses during the scan cycle. The pulse train has an average pulse frequency f above 1 kHz, preferably between 4 kHz and 8 kHz. The term “average frequency” represents a frequency obtained by averaging the entire pulse train. The light source 33 is connected to the control unit 60 so that the control unit 60 can correct the emission time of each light pulse within a single pulse train by a time shift of at least 1 / 2f. In other words, the emission time of each light pulse can be shortened or delayed within certain limits so that the light pulse can be emitted slightly earlier or slightly later than completely regular.
The term “time shift” refers to the regular emission time obtained when the pulse train is completely periodic. For example, if the average pulse frequency f is 5 kHz, the average time interval between two consecutive light pulses is 200 microseconds. Thus, the regular release time is again divided by 200 microseconds. In this case, a delay of 1 / 2.5 kHz corresponds to 100 microseconds. This means that a specific light pulse can be triggered within 100 microseconds after the regular emission time has elapsed. The possible time interval between two successive light pulses is therefore in the range between 100 and 300 microseconds.
Similar to the embodiment shown in FIG. 19, the pupil defining unit 36 of this embodiment also comprises a diffractive optical element 96 mounted on the turret holder 102 and is rotated with the aid of the turret driver 104. It can be rotated about axis 118. The rotation axis 118 is aligned parallel to the optical axis OA of the subsequent optical component of the illumination system 12.

However, in contrast to the embodiment shown in FIG. 19, the turret driver 104 is configured to rotate the diffractive optical element 96 about the rotational axis 118 at a very high speed. In the embodiment shown, the angular velocity is constant and the rotation frequency is equal to or an integer multiple of the average pulse frequency f of the light source 33. Assuming that the average pulse frequency f is in the range of several kHz and the diameter of the diffractive optical element 96 is in the range of several centimeters, the circumferential speed of the turret holder 102 is close to the sound speed, or even the sound speed. Become faster.
The diffractive optical element 96 includes a first portion 96a and a second portion 96b having different diffractive characteristics, so that the first portion 96a has a first irradiance distribution when illuminated by a light pulse. Is generated in the pupil plane 76. When illuminated by the light pulse, the second portion 96b is completely different from the first irradiance distribution, even though the arrangement of the optical components between the diffractive optical element 96 and the pupil plane 76 is exactly the same. A second irradiance distribution is generated in the pupil plane 76.

  These optical elements include a zoom optical system represented by a single lens 50 and a pair of axicon elements 108, 110, the distance of the axicon elements 108, 110 described above with reference to FIG. Can be changed as follows. By using these optical elements, the irradiance distribution in the pupil plane 76 produced by one of the portions 96a, 96b of the diffractive optical element 96 can be further modified.

The illumination system 12 further includes a polarization manipulator 77 disposed in the beam path between the pair of axicon elements 108, 110 and the subsequent light integrator 60. The polarization manipulator 77 includes a polarization manipulating element 126 that includes a first portion 126a and a second portion 126b that affect the state of polarization differently. The polarization manipulating element 126 is mounted on a turret holder 122 that can be rotated around a rotation axis 128 with the aid of a turret drive 124, which is also on the optical axis OA of the illumination system 12. They are aligned parallel to each other. Similar to the turret driver 104 used for the diffractive optical element 96, the turret driver 124 of the polarization manipulator 77 moves the polarization manipulator around the rotation axis 128 and the average pulse frequency f of the pulse train emitted by the light source 33. It is configured to rotate at an extremely high speed at a constant angular velocity equal to or an integral multiple of.
2. Lighting system functions and controls

Next, the function of the illumination system 12 shown in FIG. 21 will be described with reference to FIGS.
The left side of each of these figures is a scanning cycle in which the illuminated field 14 scans over the first and second mask pattern regions 2181, 2182 in the manner described above with reference to FIGS. Showing progress.
The portion on the right side of the figure is a top view of the diffractive optical element 96 at a different rotational position with respect to the rotation axis 118. A rectangular field 130 indicates a position where the projection light pulse is incident on the rotating diffractive optical element 96. The rotation at a constant angular velocity is indicated by an arrow, and the two different portions 96a, 96b of the diffractive optical element 96 having different diffractive properties are represented by different hatching.
The illustration to the right of the diffractive optical element 96 is a top view of the rotating polarization manipulating element 126 including portions 126a, 126b that have different effects on the state of polarization. Here, the first portion 126a converts the incident circularly polarized light into light that is linearly polarized along the Y direction, and the second portion 126b converts the incident circularly polarized light into the linearly polarized light along the X direction. It is assumed that it is converted to light. To that end, portions 126a, 126b can comprise birefringent optical materials having different thicknesses and appropriately aligned optical axes, as is known in the art. A circular region 132 indicates a position where the projection light pulse enters the polarization operation element 126.
The right part of each figure shows the irradiance distribution in the pupil plane 76. The hatched area shows the pole 27 illuminated by the projection light pulse during the scanning operation. A double arrow 134 represents the linear state of the polarization of the projection light passing through the illuminated area in the pupil plane 76.

  FIG. 22 a shows a state that prevails at some stage when the first mask pattern portion 2181 is illuminated by the field 14 on the mask 16. The control unit 60 controls the light source 33 and the driving device 104 so that the light pulse is completely synchronized with the rotation of the diffractive optical element 96. Therefore, every time the light source 33 emits a light pulse, the diffractive optical element 96 is located at the same rotational position. Here, this rotational position is selected such that the field 130 illuminated by the light pulse is approximately located at the circumferential center of the first portion 96 a of the diffractive optical element 96. When illuminated by a light pulse, this first portion 96a is first in a pupil plane 76 that includes two poles 27 spaced along the X direction, as shown on the right side of FIG. 22a. Is configured to generate an irradiance distribution.

  The polarization manipulating element 126 also rotates at a rotation frequency equal to the average frequency f of the pulse train. In the first rotational position shown in FIG. 22a, the region 132 illuminated by the light pulse is located in the first portion 126a. This portion is configured in this embodiment so that the projection light is linearly polarized along the Y direction, as shown on the right by the double arrow 134. This is the optimum state of polarization for imaging structures that are also aligned along the Y direction.

  FIG. 22 b shows the state at a certain stage during the time period that elapses before the next light pulse is emitted by the light source 33. Accordingly, there are no regions 130 and 132 illuminated on the diffractive optical element 96 and the polarization manipulating element 126, respectively. The diffractive optical element 96 and also the polarization manipulating element 126 have rotated half a turn, so that the positions of the first and second portions 96a, 96b and 126a, 126b, respectively, have changed. However, this has no effect on the illumination of the mask 14 since no light pulses are emitted.

FIG. 22 c shows the prevailing state when the next light pulse is emitted by the light source 33. The mask 14 is moved a little further, the diffractive optical element 96 is again positioned at the first rotational position shown in FIG. 22a, and so is the polarization manipulating element 26. Accordingly, the irradiance distribution and polarization distribution in the pupil plane 76 are the same as in FIG. 22a.
This synchronization of the light source 33 with the diffractive optical element 96 and the polarization manipulating element 126 is maintained as long as the first mask pattern area 2181 is illuminated.
When the second mask pattern region 2182 is imaged on the photosensitive layer 22, the angular light distribution must change very quickly to the Y dipole illumination setting shown in FIG. In this case, the two poles 28 must be spaced along the Y direction, as shown on the right side of FIG. 22d. Furthermore, the projection light must be linearly polarized along the X direction, as shown in this figure by the double arrow 135.
This rapid change in illumination setting is achieved by delaying all subsequent light pulses by a half pulse period, ie, 1 / 2f. This implies that the emission time between two successive light pulses is increased by 1 / 2f only once. All remaining light pulses are emitted again at the average frequency f. Once the pulse period is increased by ½f, the diffractive optical element 96 has its first illumination region 130 positioned above the second portion 96b of the diffractive optical element 96 each time a light pulse is emitted. 2 position. The second portion generates a second irradiance distribution in the pupil plane 76 that is completely different from the first irradiance distribution in that the poles 28 are spaced along the Y direction rather than the X direction. Composed.

The emission time delay has an effect similar to the effect on the polarization state. Accordingly, the polarization manipulating element 126 is positioned at the second rotational position every time a light pulse is emitted. Thus, the region 132 illuminated by the light pulse is a second configured such that the projection light emerging from this portion is linearly polarized along the X direction, as shown on the right by the double arrow 134. Located in portion 126a. This is the optimal state of polarization for imaging structures that are also aligned along the X direction.
When the illumination of the second mask pattern area 2182 ends and the illumination setting is changed back to the X dipole illumination setting, the light pulse is no longer delayed. This implies that all remaining light pulses are emitted again at the average frequency f.

  By replacing diffractive optical element 96 and polarization manipulating element 126 with elements having different optical properties, it is possible to change very quickly between almost any arbitrary angular and polarization distribution at the mask level. This is illustrated in FIGS. 23 and 24, which show two different stages in the scan cycle in a similar view to FIGS. 22a-22d. Here, it is assumed that the first portion 96a of the diffractive optical element 96 produces a conventional illumination setting where the illuminated portion in the pupil plane 76 has the shape of a disc 136 (see FIG. 23). . The second irradiance distribution generated when the second portion 96b of the diffractive optical element 96 is illuminated by the light pulse corresponds to an annular illumination setting in which the annular region 137 is illuminated in the pupil plane 76. (FIG. 24). Thus, by using such a diffractive optical element 96, the conventional setup shown on the right side of FIG. 23 between the two successive pulses of a single pulse train emitted by the light source 33, It can be changed very quickly between the annular illumination settings shown on the right side of 24.

The time shift required for the change between the illumination of the first part 96a and the illumination of the second part 96b is, in this embodiment, the first and second rotational positions of the diffractive optical element 96 instead of 180 °. By arranging them so as to be separated by 60 °, it is shortened from 1 / 2f to 1 / 6f.
In this embodiment, the polarization manipulating element 126 comprises a first portion 126a that is configured to produce circularly polarized light (see arrow 139a) when an incoming light pulse is linearly polarized along a given direction. . In this case, the first portion 126a can be formed by, for example, a quarter wave plate. The second portion 196b is configured to generate a tangential polarization distribution, as indicated by the double arrow 139b on the right side of FIG. Such a tangential polarization distribution is achieved when the second portion 126b is assembled from a plurality of birefringent elements oriented in different directions, as described in U.S. Patent Application Publication No. 2002/0176166. can do. If the polarization manipulating element 126 is rotating at the same angular frequency as the diffractive optical element 96, the irradiance distribution in the pupil plane 76 and again between successive light pulses and again if the light pulse is delayed by 1 / 6f. The polarization distribution can be changed simultaneously.
3. Alternative embodiments

In the embodiment shown in FIGS. 21 to 24, the angular frequency of the diffractive optical element 96 and the polarization manipulating element 126 is equal to or an integral multiple of the pulse frequency f. This requires fast angular velocities of the diffractive optical element 96 and the polarization manipulating element 126 on the one hand when changing the irradiance distribution and polarization distribution in the pupil plane 76 and also requires a relatively long time shift on the other hand. It is.
a. Alternating parts

FIG. 25 shows an embodiment of a diffractive optical element 96 in which a plurality of first portions 96 a and the same number of second portions 96 b are alternately arranged along the circumference of the diffractive optical element 96. It is. Again, each first portion 96a generates an X dipole illumination setting with poles 27 in the pupil plane 76 shown above the diffractive optical element 96, as shown in FIG. And each second portion 96b is assumed to generate a Y-dipole illumination setting with poles 28.
When there are n first portions 96 a and n second portions 96 b, the angular frequency of the diffractive optical element 96 can be reduced to 1 / n with respect to the average pulse frequency f of the light source 33. it can. Further, the first angular position where one of the first portions 96a is illuminated by a light pulse and the second rotational position where one of the second portions 96b is illuminated by a light pulse are here: They are separated by 180 ° / n instead of 180 °. Thus, a change between the first and second rotational positions can be achieved by a time shift of only 1/2 nf. Assuming that the pulse frequency is f = 5 kHz and that the diffractive optical element 96 has 16 first portions 96a and 16 second portions 96b, the first and second portions 96a, 96b. The time shift required to change between the two illumination settings associated with is only 6.25 microseconds, which is a regular period between two consecutive light pulses, 200 microseconds Of about 3%. When the light source 33 is formed by an excimer laser as is typical, a shorter time shift of the emission time of each light pulse has a beneficial effect on the performance of the light source 33.
b. Multiple lighting settings

  FIG. 26 is a top view of a diffractive optical element 96 according to another embodiment. Unlike the embodiment shown in FIGS. 22-24, the diffractive optical element 96 comprises ten different portions 96a-96j rather than just two. Each portion produces a different irradiance distribution in the pupil plane 76 when illuminated by the light pulse. Thus, portions 96a-96j are represented in FIG. 26 by the irradiance distribution generated in the pupil plane 76 by illuminating each portion.

If the rotational frequency of the diffractive optical element 96 is equal to the average pulse frequency f of the light source 33 (or an integer multiple thereof), any irradiance distribution associated with the portions 96a-96j can be generated in the pupil plane 76. In this case, all that is necessary is to control the emission time of each light pulse so that each light pulse illuminates the same desired portion each time while the diffractive optical element 96 is rotating. For example, if it is desired to generate the irradiance distribution shown in FIG. 27 in the pupil plane 76, this particular irradiance distribution, and hence the irradiance distribution shown in FIG. Each time the portion 96 h configured as such intersects the path of the light pulse in the illumination system 12, a light pulse must be emitted by the light source 33.
It will be appreciated that the pole 28 in the pupil plane 76 must be made smaller or larger in order to allow optimal transfer of the mask pattern 18 to the photosensitive surface 22. The emission time can be modified so that one of the other portions 96d, 96i or 96j is illuminated by the light pulse while is rotating at a constant angular velocity.
c. Scan integrated pupil filling

The number of different parts that can be utilized on the diffractive optical element 96 is necessarily limited. In particular, when generating a more complex irradiance distribution in the pupil plane 76, it may be necessary to provide a number of different diffractive optical elements that individually include a number of portions having different diffractive properties.
Another way in which a wide variety of different irradiance distributions can be generated in the pupil plane 76 is based on the consideration that the entire irradiance distribution in the pupil plane 76 need not be generated by each light pulse. . Instead, it is sufficient that the desired irradiance distribution in the pupil plane 76 is generated during the entire time interval when a particular point on the mask 16 is illuminated by the light pulse. In other words, the light pulse produces only a portion of the desired irradiance distribution in the pupil plane 76, and the entire irradiance distribution is assembled from these portions during the scanning process. This is sometimes referred to as “scan integrated pupil filling”.

FIG. 28 is a top view of a diffractive optical element 96 according to one embodiment in which there are also ten portions 96a-96j that again produce different irradiance distributions in the pupil plane 76 as shown. is there. This diffractive optical element 96 is, among other things, configured to generate different C-scud illumination settings. Such an illumination setting is a small conventional with an X dipole setting with two poles 27, a Y dipole setting with two poles 28, and a center pole 25, as exemplarily shown in FIG. This is a combination of settings.
While the diffractive optical element 96 is rotated once by 360 ° (preferably at a constant angular velocity), the three shaded portions 96b, 96f and 96i are alternately illuminated by light pulses. Each of the three portions 96b, 96f, 96i produces a different irradiance distribution in the pupil plane 76, so that the scan integrated irradiance distribution shown in FIG.

When changing the size of the area within this irradiance distribution, all that is required is to modify the emission times of some light pulses. For example, if the size of pole 27 spaced along the X direction is reduced, the emission time of the light pulse must be modified so that part 96c or 96d is illuminated, not part 96b.
d. Continuous correction of lighting settings

FIG. 30 illustrates one embodiment of a diffractive optical element 96 where portions that produce different irradiance distributions in the pupil plane 76 form a continuous circular ring 140 centered with respect to the axis of rotation 118 of the diffractive optical element 96. Is shown. The diffraction effect produced by the ring 140 varies continuously along at least a portion of the circumference. The ring 140 thus has a virtually infinite number of rotational positions where different irradiance distributions are generated in the pupil plane 76.
In FIG. 30, the three portions 96a, 96b and 96c are indicated by dashed lines. The first portion 96a, when illuminated with a light pulse, produces a Y dipole illumination setting with poles 27, as shown in FIG. The second portion 96b, when illuminated with a light pulse, produces an X dipole illumination setting with poles 28. A third portion 96c disposed between the first portion 96a and the second portion 96b generates a mixture of X and Y dipole illumination settings and has four poles 27,28. Brings a pole lighting setting. Accordingly, the irradiated area in the pupil plane is twice as large as the dipole illumination setting, so the intensity of each pole 27, 28 is halved. This is illustrated in FIG. 30 by the gray shading of the poles 27, 28 generated by the third portion 96c.

The portion between the first portion 96a and the third portion 96c, and again the portion between the second portion 96b and the third portion 96c, is indicated by the irradiance distribution M in the pupil plane 76. Thus, an irradiance distribution is generated in the pupil plane 76 in which the irradiance of the pole in the pupil plane 76 continuously shifts from the pole 27 to the pole 28 and vice versa.
Accordingly, the diffractive optical element 96 shown in FIG. 30 generates different irradiance distributions only for the irradiance of regions having the same shape as well as different shapes of the irradiated region in the pupil plane 76. Can do. This can be used, for example, to adjust the telecentricity and pole balance of the lighting settings.
e. Combination with fixed diffractive optical element

In the embodiment shown in FIGS. 26-30, the diffractive optical element 96 rotating at high speed can not only produce different irradiance distributions in the pupil plane 76 but also the shape of the poles 27, 28, for example. A given irradiance distribution can be slightly modified by changing the irradiance and / or.
It is also possible to modify or correct the basic irradiance distribution in the pupil plane exclusively using the diffractive optical element 96 rotating at high speed. In this case, the basic irradiance distribution can be generated by conventional pupil defining means, for example a fixed diffractive optical element combined with a zoom optical system and a pair of axicon elements.

  FIG. 31 shows a lighting system 12 that is substantially identical to the lighting system 12 shown in FIG. 21 in a meridian sectional view similar to FIG. The only difference is that the additional static diffractive optical element 96 'is located in the immediate vicinity of the diffractive optical element 96 which rotates at high speed. Alternatively, the two diffractive optical elements 96, 96 'can be placed in a plane that is optically conjugated, for example in the object plane and image plane of the imaging optics. It is. Using such a structure, or using the structure shown in FIG. 31, the irradiance distribution generated in the pupil plane 76 by the two diffractive optical elements 96, 96 ′ can be expressed as a static diffractive type. It can be mathematically described as a convolution of the irradiance distribution generated individually by the optical element 96 ′ with the irradiance distribution generated by different parts of the diffractive optical element 96 rotating at high speed.

This is illustrated in FIGS. 32 and 33. FIG. 32a shows an exemplary irradiance distribution including a single pole P generated in the pupil plane 76 by a static diffractive optical element 96 ′. FIG. 32b schematically illustrates different irradiance distributions that can be generated by portions 96a-96i of the diffractive optical element 96 rotating at high speed. Portions 96a-96i have an irradiance distribution in the pupil plane 76 that consists essentially of two microspots having two different available spot sizes, which are basically spaced apart by different distances along the Y direction. Are different from each other in that they are generated.
For example, if the emission time of the light pulse is controlled so that only the portion 96d is illuminated by the light pulse, the irradiance distribution shown in FIG. 32c is obtained in the pupil plane 76. This irradiance distribution is mathematically a convolution (indicated by symbol 142) of the irradiance distribution shown in FIG. 32a with the irradiance distribution associated with the portion 96d shown in FIG. 32b. Can be described. It can be seen that the pole P produced by the static diffractive optical element 96 ′ is doubled, so that the two poles 28 d appear at different positions in the pupil plane 76.

For example, when the light pulse emission time is controlled so that the light pulse illuminates a portion 96g that produces two larger spots spaced a maximum distance along the Y direction, as shown in FIG. 31a. By convolution with the basic irradiance distribution, the irradiance distribution shown in FIG. 33c is obtained. Compared to FIG. 32c, the two poles 28g are slightly larger and have a longer distance along the Y direction.
This approach can also be used to correct for telecentricity errors or polar equilibrium errors for a given lighting setting. These terms relate to the spatial energy distribution in the pupil plane 76. If the energy distribution in the pupil plane 76 does not meet certain symmetry requirements, a telecentricity error or a polar equilibrium error occurs. For example, in a quadrupole illumination setting, if the poles spaced along the X direction are brighter than the poles spaced along the Y direction, this asymmetry is called a pole balance error. Violations of such symmetry requirements in the pupil plane 76 can cause various undesirable effects. For example, images of feature 19 having exactly the same shape and size in different orientations will have different sizes.
The left side of FIG. 34a shows the irradiance distribution generated in the pupil plane 76 by the stationary diffractive optical element 96 'when no diffractive optical element 96 rotating at high speed is present. The two poles P have a shape that is particularly adapted to the feature to be imaged.

If there is no diffractive optical element 96 rotating at high speed, or if a portion 96a that does not modify the angular distribution of the passing light is illuminated, the irradiance distribution shown in FIG. 34a is obtained in the pupil plane 76. It is expected. However, this is not the case for various reasons. For these reasons, imperfections in the static diffractive optical element 96 ′, variations in the direction of the projection light beam 35 incident on the static diffractive optical element 96 ′, and also the static diffractive optical element 96 ′ and the pupil plane 76 Including aberrations associated with optical elements disposed between the two.
Here, it is assumed that the irradiance distribution finally obtained in the pupil plane 76 will be a distribution as shown in FIG. 34c. It can be seen that the poles 27a have the desired shape and size, but they are slightly displaced along the + X direction. In this case, since more light is incident on the point on the mask 16 from the + X direction than from the −X direction, this displacement causes a telecentricity error.

  Such telecentricity errors can be detected, for example, by measuring the angular light distribution in the mask plane 88 or the spatial irradiance distribution in the pupil plane 76. Next, the emission time of the light pulse can be modified so that the portion 96j of the diffractive optical element 96 rotating at high speed is illuminated by the light pulse instead of the portion 96a. When illuminated by the projection light, the portion 96j generates an irradiance distribution in the pupil plane 76 consisting of minute spots that are slightly displaced along the −X direction. The convolution of such irradiance distribution with the basic irradiance distribution shown in FIGS. 34a and 35a is such that the basic irradiance distribution is slightly along the −X direction, as shown by the arrow in FIG. 35c. Will shift. In this case, the pole 27i is perfectly aligned along the Y direction, so that the same amount of projection light enters the mask 14 from the + X direction and the-direction.

The diffractive optical element 96 that rotates at high speed includes a large number of portions 96a to 96i, and these portions have a minute amount of irradiance distribution along the + X direction and the −X direction, and the + Y direction and the −Y direction. Only by shifting the light pulse emission time while the diffractive optical element 96 is rotating at a constant angular velocity, allowing for telecentricity and pole equilibrium. Lighting settings can be optimized. Furthermore, due to the high speed rotation of the diffractive optical element 96, it is clear that the illumination setting can be adjusted extremely quickly. Therefore, if different mask pattern areas 181 and 182 require different amounts of adjustment, the illumination settings can be adjusted quickly.
f. Alternate arrangement of pulse trains

It is also possible to alternately arrange two optical pulse trains so that the effective pulse rate is doubled by using a combination of a diffractive optical element 96 rotating at high speed and a static diffractive optical element 96 ′. is there.
This includes not only one light source 33 but also two light sources 33a and 33b, two beam expanding units 34a and 34b, two first plane folding mirrors 117a and 117b, and two second plane folding mirrors 121 and 121b. FIG. 36 shows a lighting system 12 that differs from the lighting system shown in FIG. The second folding mirrors 121a and 121b guide the projection light beams 35a and 35b toward the stationary diffractive optical element 96 ′. Therefore, as a result of the oblique incidence, the irradiance distribution generated in the pupil plane 76 by the static diffractive optical element 96 ′ is in the + X direction depending on whether the light pulse is generated by the light source 33a or whether it is generated by the light source 33b. Or it will shift in the -X direction. Since each light pulse produces the same irradiance distribution regardless of its origin, a fast rotating diffractive optical element 96 is used to compensate for the irradiance distribution shift caused by oblique incidence. Hereinafter, this will be described with reference to FIGS.
FIG. 37a shows that the static diffractive optical element 96 is illuminated by a light pulse emitted by the first light source 33a and there is no diffractive optical element 96 rotating at high speed. The irradiance distribution to be generated in the pupil plane 76 is shown. The irradiance distribution includes two poles P that are shifted toward the + X direction as a result of the oblique incidence. This lateral offset is compensated by a diffractive optical element 96 that rotates at high speed. When the light pulse is incident on the portion 96g of the diffractive optical element 96 obliquely, this portion tilts the light beam 35a toward the -X direction, and this offsets the tilt generated by the oblique incidence.

Expressed mathematically, the irradiance distribution shown in FIG. 37a and the pupil plane when the rapidly rotating portion 96f of the diffractive optical element 96 is illuminated by a light pulse generated by the first light source 33a. Convolution with the irradiance distribution generated in 76 results in the irradiance distribution shown in FIG. 37c, in which the two poles 28g are centered with respect to the X direction.
The emission time of the next light pulse emitted by the second light source 33b is controlled by the control unit 90 to illuminate the portion 96a of the diffractive optical element 96 where the light pulse rotates at high speed. The portion 96b shifts the irradiance distribution in the reverse direction, that is, in the + X direction. This offsets the shift in irradiance distribution, which is the result of the oblique incidence of the light pulse on the static diffractive optical element 96 ', shown in FIG. 38a.

The fast rotating diffractive optical element 96 can include portions that tilt incident light to different degrees along the + X and −X directions, and thus as described above with reference to FIGS. 34 and 35. In addition, the telecentricity error and the pole balance error can be corrected.
VII. Illumination system for the third embodiment

In the following, on the one hand, two different illumination settings can be generated in different parts of the illuminated field 14 and again between exposures of mask pattern areas arranged vertically along the scanning direction. An illumination system that can change the illumination setting very quickly will be briefly described. Such an illumination system is necessary to implement the third embodiment of the double exposure scheme described above with reference to FIGS. 6 and 7.
The illumination system that can perform this task is almost identical to the illumination system shown in FIG. The only modification is to allow a more complex irradiance distribution to be generated on the light entrance facet 100, as described above with reference to the embodiment shown in FIGS. That is, the portions 96a and 96b of the diffractive optical element 96 rotating in the above must be designed. These irradiance distributions result in different illumination settings at different positions on the mask, while the diffractive optical element 96, which rotates at high speed, allows these illumination settings to be changed very quickly.
VIII. Important method steps

In the following, the important method steps according to the invention will be described with reference to the flowchart shown in FIG.
In a first step S1, a mask comprising first and second mask pattern areas is provided.
In the second step S2, the first exposure pattern area is exposed as an image of the first mask pattern area, and the second exposure pattern area is generated as an image of the second mask pattern area.
In the third step S3, step S2 is repeated using the same mask so that the image of the first mask pattern area is superimposed on the second exposure pattern area. The projection light for illuminating the first mask pattern region has a first angular light distribution, and the projection light for illuminating the second mask pattern region is a second different from the first angular light distribution. The angular light distribution is as follows.
IX. Important features of the present invention

The following text summarizes important features of the present invention.
1. A method of transferring a pattern (18) to a photosensitive surface (22) by lithography in a multiple exposure process, the method comprising:
a) providing a mask (16, 216, 316) comprising a first mask pattern region (181, 1821, 3181) and a second mask pattern region (182, 2182, 3182);
b) Projected light is guided onto the mask, whereby the first exposure pattern area (301, 2301, 3301) which is an image of the first mask pattern area (181, 1821, 3181), and the second mask pattern area Generating a second exposure pattern region (302, 232, 3302) which is an image of (182, 2182, 3182) on the photosensitive surface;
c) Use the same mask (16, 216, 316) so that the image of the first mask pattern area (181, 1821, 3181) is superimposed on the second exposure pattern area (302, 232, 3302). And the step b) is repeated, and the projection light for illuminating the first mask pattern region (181, 1821, 3181) has a first angular light distribution, and the second mask pattern region ( 182, 2182, 3182) has a second angular light distribution different from the first angular light distribution.

2. The method of sentence 1, wherein the projection light is simultaneously guided onto the first mask pattern area (181, 3181) and the second mask pattern area (182, 3182) during step b).
3. The method of sentence 2, wherein the mask (16, 316) scans the illuminated field (14) over the first mask pattern region (181, 3181) and the second mask pattern region (182, 3182). During step b), the first mask pattern region and the second mask pattern region move along the scan direction and the normal on the mask (16, 316). They are arranged side by side along the cross scan direction (X) orthogonal to both.

4). In the method of sentence 3, only the image of the first mask pattern area (181, 3181) is superimposed on the second exposure pattern area (302, 3302) during step b).
5. The method of sentence 4, wherein the photosensitive surface (22) is displaced along the cross-scan direction (X) between step b) and step c).

6). The method of sentence 1, wherein the projection light is first guided onto the first mask pattern area (2181) and subsequently onto the second mask pattern area (2182) during step b).
7). The method of sentence 6, wherein the mask (216) is scanned in step b) so that the illuminated field (14) scans over the first mask pattern region (2181) and the second mask pattern region (2182). The first mask pattern region (2181) and the second mask pattern region (2182) are arranged vertically along the scan direction (Y). .
8). In the method of sentence 7, the image of the second mask pattern region (2182) is superimposed on the first exposure pattern region (2301) during step c).
9. The method of sentence 8, wherein the photosensitive surface (22) is displaced along the scanning direction (Y) between step b) and step c).

10. Any of the methods of sentences 1 to 9,
The mask (316) is supplementarily provided with a third mask pattern region (3183).
The image of the third mask pattern area (3183) is superimposed on both the first exposure pattern area (3301) and the second exposure pattern area (3302);
The projection light that illuminates the third mask pattern region (3183) has a third angular light distribution different from the first angular light distribution and the second angular light distribution.

11. The method of sentence 10,
The mask (316) supplementarily comprises a fourth mask pattern region (3184).
The image of the fourth mask pattern area (3184) is on the first exposure pattern area (3301), on the second exposure pattern area (3302) and on the third exposure pattern area (3303); Superimposed.
The projection light that illuminates the fourth mask pattern region (3304) has a fourth angular light distribution different from the first angular light distribution, the second angular light distribution, and the third angular light distribution.

12 The method of sentence 11,
During step b), a second exposure pattern area (3302) is generated on the photosensitive surface (22) simultaneously with the generation of the first exposure pattern area (3301).
-The fourth exposure pattern area (3304) which is the image of the fourth mask pattern area (3184) simultaneously with the generation of the third exposure pattern area (3303) which is the image of the third mask pattern area (3183). Are produced on the photosensitive surface (22).
During step c), the image of the third mask pattern area (3183) is superimposed on the first exposure pattern area (3301), so that the first mask pattern area plus the third mask pattern area ( 3301 + 3303) is generated at the same time, the image of the fourth mask pattern region (3184) is superimposed on the second exposure pattern region (3302), whereby the second mask pattern region plus the fourth A combination of mask pattern areas (3302 + 3304) is generated.
This method includes the following additional steps d) and e).
d) The image of the first mask pattern region (3181) is alternately superimposed on the combination of the second mask pattern region plus the fourth mask pattern region (3302 + 3304), whereby the first mask pattern region plus the first mask pattern region 2 mask pattern areas plus a fourth mask pattern area (3301 + 3302 + 3304) is generated, and an image of the third mask pattern area (3183) is converted into a second mask pattern area plus a fourth mask pattern area (3302 + 3304). ) To generate a combination of the second mask pattern area plus the third mask pattern area plus the fourth mask pattern area (3302 + 3303 + 3304).
e) The image of the first mask pattern region (3181) is alternately superimposed on the combination of the second mask pattern region plus the third mask pattern region plus the fourth mask pattern region (3302 + 3303 + 3304), thereby A combination of one mask pattern region plus a second mask pattern region plus a third mask pattern region plus a fourth mask pattern region (3301 + 3302 + 3303 + 3304) and generating an image of the third mask pattern region (3183) 1 mask pattern region plus a second mask pattern region plus a fourth mask pattern region (3301 + 3302 + 3304) is superimposed on the first mask pattern region plus a second mask pattern region plus a third mask. Pattern area plastic Step further generates a fourth mask pattern region (3301 + 3302 + 3303 + 3304).
13. The illumination system of the microlithographic projection exposure apparatus is
a) a light source (33, 33a, 33b) configured to emit a sequence of continuous projection light pulses, the pulse train having a mean pulse frequency of f> 1 kHz;
b) the pupil plane (76);
c) A diffractive optical element (96) comprising a first part (96a) and a second part (96b), the first part (96a) being illuminated by a light pulse, the first part (96a) And the second portion (96b), when illuminated with a light pulse, produces a second irradiance distribution different from the first irradiance distribution. A diffractive optical element (96) generated in the pupil plane (76);
d) A driving device (104) configured to rotate the diffractive optical element (96) about the rotation axis (118), wherein the first rotative position of the diffractive optical element (96) is the first And the second part (96b) is illuminated by a light pulse at a second rotational position of the diffractive optical element different from the first rotational position. (104)
e) The change between the first rotational position and the second rotational position is between two light pulses of the pulse train, preferably between five consecutive light pulses of the pulse train, more preferably two consecutive light pulses. And a control unit (90) configured to control the drive (104) and / or the light source (33) to occur between.

  14 The illumination system of sentence 13, wherein the drive (104) is configured to rotate the diffractive optical element (96) about a rotational axis (118) at a constant angular velocity, and the control unit (90) While the diffractive optical element (96) is rotating around the axis of rotation (118) at a constant angular velocity, a particular light pulse is transmitted in the first part (96a) or second part ( 96b) is configured to modify the emission time of each light pulse within a single pulse train by a time shift of at least 1/10 · f.

15. The illumination system of sentence 14, wherein the rotational frequency of the diffractive optical element is equal to the average pulse frequency f.
16. The illumination system according to any one of statements 13 to 15, wherein the driving device (104) is configured such that the circumference of the diffractive optical element (96) moves at a tangential speed exceeding 10 m / sec. ) Is configured to rotate.

  17. The illumination system of any of statements 13 to 16, wherein the diffractive optical element (96) comprises a plurality of first portions (96a) and a plurality of second portions (96b), each first When illuminated by the light pulse, the first portion (96a) produces a first irradiance distribution in the pupil plane (6), and each second portion (96b) is illuminated by the light pulse. And a second irradiance distribution different from the first irradiance distribution is generated in the pupil plane (76), and the first part (96a) and the second part (96b) are diffractive optical elements ( 96) are alternately arranged along the circumference.

18. The illumination system of any of statements 13 to 17, wherein the control unit (90) includes the first portion (96a) and the second portion while the diffractive optical element (96) makes one revolution by 360 °. The drive (104) and / or the light source (33) is configured to be controlled such that the portion (96b) is illuminated by the light pulse.
19. The illumination system of any of statements 13 to 18, wherein the first portion (96a) and the second portion (96b) are contained within a circular ring (140) centered with respect to the axis of rotation (118). Thus, the diffraction effect produced by the ring (140) when illuminated by a light pulse changes continuously as the diffractive optical element (96) rotates.
20. The illumination system of any of statements 13 to 19, wherein the illumination system (12) comprises a static diffractive optical element (96 '), the static diffractive optical element (96') being a pupil plane ( 96) is generated at least substantially by the irradiance distribution produced by the static diffractive optical element (96 ′) and the first part (96a) and the second part (96b), respectively. Are arranged so as to be convolved with the first or second irradiance distribution.

21. The illumination system of sentence 20, wherein the illumination system (12) comprises a first light source (33a) and a second light source (33b) each configured to emit a train of consecutive projection light pulses. The pulse train has an average pulse frequency of f> 1 kHz, and the light pulses generated by the first and second light sources are different from one another in the diffractive optical elements (96 ′, 96a, 96b). Incident on one.
22. The illumination system of sentence 21, wherein one diffractive optical element is a static diffractive optical element (96 ').

23. The illumination system according to any one of statements 21 to 22, wherein the first light source (33a) and the second light source (33b) transmit light pulses so that an effective pulse train having an average pulse frequency of 2 · f is obtained. It is configured to emit alternately.
24. A method of operating a microlithographic projection exposure apparatus includes:
a) generating a sequence of consecutive projected light pulses having an average pulse frequency of f> 1 kHz;
b) simultaneously rotating the diffractive optical element (96) about the rotation axis (118), wherein the diffractive optical element (96) moves the first part (96a) and the second part (96b); When the first part (96a) is illuminated by the light pulse, a first irradiance distribution is generated in the pupil plane (76) of the illumination system (12) and the second part (96b) ) Is illuminated by the light pulse, generating a second irradiance distribution in the pupil plane (76) that is different from the first irradiance distribution;
c) the step so that one light pulse illuminates the first part (96a) and the other light pulse, preferably a light pulse immediately following the same pulse train, illuminates the second part (96b). controlling the emission of light pulses in a) and / or the rotation of the diffractive optical element (96) in step b).

  25. The method of sentence 24, wherein the diffractive optical element (96) rotates around the axis of rotation (118) at a constant angular velocity, and the light pulse emission time is determined by the rotation of the diffractive optical element (96). While rotating around the axis (118) at a constant angular velocity, a particular light pulse illuminates either the first part (96a) or the second part (96b) in response to a time shift. It is corrected within a single pulse train by a time shift of at least 1/10 · f.

26. The method of sentence 24 or 25, wherein the circumference of the diffractive optical element (96) moves at a tangential velocity exceeding 10 m / sec.
27. The method of any of statements 24 to 26, wherein the emission time is controlled such that the first portion (96a) and the second portion (96b) are alternately illuminated during a single pulse train. .
28. The method of any of statements 24 to 27, wherein the light pulse has an irradiance distribution in the pupil plane (74) at least substantially generated by the static diffractive optical element (96 ') and , Stationary diffractive optical elements (fixedly arranged so as to be convolution with the first or second irradiance distribution generated by the first part (96a) and the second part (96b), respectively. 96 ').
29. 28. The method of any of statements 24 to 28, wherein the train of light pulses includes a first light source (33a) that emits a first train of consecutive projection light pulses, and a second light train of consecutive projection light pulses. Generated by a second light source (33b) emitting columns, the first and second columns are interleaved.
30. The method of sentence 29, wherein the first portion is illuminated only by the light pulses of the first pulse train and the second portion is illuminated only by the light pulses of the second pulse train.

31. The method of sentence 29 or 30, wherein the first portion and the second portion are arranged along the opposite direction to the irradiance distribution generated in the pupil plane (76) by the static diffractive optical element (96 ′). Shift horizontally.
32. The illumination system of the microlithographic projection exposure apparatus (10) comprises:
a) an optical integrator (60) comprising an array (70, 72) of light raster elements (74), wherein a light beam is coupled to each light raster element (74);
b) A common field plane (80) that is identical or optically conjugate to the mask plane (88) in which the mask (16) to be illuminated is placed while the illumination system (12) is in operation. A capacitor (78) for superimposing a light beam coupled to the optical raster element (74) in the
c) a polarization modulator (77) configured to correct the field dependence of the state of polarization of the projected light in the illuminated field (14) illuminated in the mask plane (88) by the illumination system (12); Is provided.

33. The illumination system of sentence 32, wherein the polarization modulator (77) is
A plurality of raster field plane elements, each raster field plane element being imaged over the common field plane (80), and
-Disposed in a raster field plane (84) comprising a plurality of raster field plane elements coupled in a one-to-one correspondence with one of the optical raster elements (74);
-Configured to variably adjust the polarization state of the light beam that is otherwise coupled to the light raster element (74) in at least two portions of each raster field plane element in response to the control signal; Modulator unit (79).

34. The illumination system of sentence 33, comprising a control unit (90) capable of controlling the modulator unit (79) in such a way that different polarization states exist in the illuminated field.
35. The illumination system of sentence 33 or 34, wherein each modulator unit (79) comprises two birefringent optical components (79a, 79b) having a thickness along the optical axis (OA) of the illumination system (12). And the thickness can be varied individually.

36. The illumination system of sentence 35, wherein each optical component (79a, 79b) has two birefringent wedges (84) and a drive (83) configured to displace at least one wedge (84). Prepare.
37. A method for illuminating a mask in a microlithographic projection exposure apparatus (10) comprises:
a) illuminating the first mask pattern area (181) with the first projection light having the first polarization state while the mask (16) is moving along the scan direction (Y) Steps,
b) A second mask pattern region (182) arranged on the mask adjacent to the first mask pattern (181) region along the cross scan direction (X) orthogonal to the scan direction is Illuminating with a second projection light having a second polarization state different from the polarization state of.

  38. The method of sentence 37, wherein the first projection light has a first angular light distribution, and the second projection light has a second angular light distribution different from the first angular light distribution. .

  39. The method of sentence 37 or 38, wherein the first polarization state is changed to a third polarization state different from the second polarization state.

  40. The method of sentence 39, wherein the state of the first polarization is changed to the state of the third polarization after the movement of the mask along the scan direction is completed.

Claims (3)

A method of transferring a pattern (18) to a photosensitive surface (22) by lithography in a multiple exposure process, comprising:
a) providing a mask (16, 216, 316) comprising a first mask pattern region (181, 1821, 3181) and a second mask pattern region (182, 2182, 3182);
b) directing projection light onto the mask, whereby the first exposure pattern areas (301, 3011, 3301), which are images of the first mask pattern areas (181, 1821, 3181), and the second Generating a second exposure pattern region (302, 232, 3302) on the photosensitive surface, which is an image of a mask pattern region (182, 2182, 3182);
c) The same mask (16, 216, 316) so that the image of the first mask pattern area (181, 1821, 3181) is superimposed on the second exposure pattern area (302, 232, 3302). Repeating the step b) using: the projection light for illuminating the first mask pattern region (181, 1821, 3181) has a first angular light distribution, and the second the said projection light that illuminates the mask pattern region (182,2182,3182), have a second SumiHikari distribution different from the first SumiHikari distribution,
The projection light is simultaneously guided onto the first mask pattern region (181, 3181) and the second mask pattern region (182, 3182) during step b),
Step b) so that the mask (16, 316) scans the illuminated field (14) over the first mask pattern region (181, 3181) and the second mask pattern region (182, 3182). And the first mask pattern area and the second mask pattern area are both in the scan direction and the normal on the mask (16, 316). Are arranged adjacent to each other along a cross scan direction (X) orthogonal to the direction .   The method according to claim 1, wherein during step b), only the image of the first mask pattern area (181, 3181) is superimposed on the second exposure pattern area (302, 3302).   The method according to claim 2, wherein the photosensitive surface (22) is displaced along the cross-scan direction (X) between step b) and step c).

Images