KR101227803B1 - Homogenizer with reduced interference - Google Patents

Abstract

According to the present invention, the first input light beam 4 and the second input light beam 5 are cylindrical from the first homogenizer from the first input light beam 4 and the second input light beam 5 traveling in the first travel direction. A method of forming the target intensity distribution 20 in the target field 3 by making it enter the light incident surfaces 6a and 6b of the lens array 1a. According to the invention, the second input light beam 5 travels in a second travel direction different from the first travel direction. The invention also provides a first light source for generating a first input light beam 4 and a second light source for generating a second input light beam 5, or a single light source for generating a light beam and a first input of the light source. A beam splitter for dividing the light beam and the second input light beam, a first guiding element for guiding the first input light beam 4 in a first travel direction, and the first input light beam 4 and the second input light beam ( An optical system for forming a target intensity distribution 20 in a target field 3, comprising a homogenizer having a cylindrical lens array 1a having light incident surfaces 6a, 6b on which 5) is incident. will be. According to the present invention, there is a second guide element for guiding the second input light beam 5 to travel in a second travel direction different from the first travel direction when it enters the light incident surfaces 6a, 6b. .

Description

Homogenizer with reduced interference

The present invention aims to cause the first input light beam and the second input light beam to enter the light incident surface of the cylindrical lens array of the homogenizer from the first input light beam and the second input light beam traveling in the first travel direction. A method for forming a target intensity distribution within a field.

The present invention relates to an optical system for forming a target intensity distribution in a target field from one or two light beams having an intensity distribution.

The invention also relates to a material processing apparatus, in particular a laser annealing apparatus comprising the above-described optical system.

In addition, the present invention relates to a pulse stretcher capable of generating the aforementioned first and second laser beams.

The present invention is for example in the field of flat panel displays such as liquid crystal displays (e.g. thin film transistor displays (TFTs), etc.) or light emitting displays (inorganic or organic light emitting diodes (LEDs, OLEDs), or electroluminescence (EL)). In the field of light (eg, laser) induced crystallization of a substrate, it is useful for annealing large substrates. The present invention can also be used in the manufacture of thin film photovoltaic devices.

In particular, the present invention is useful for producing devices capable of forming polycrystalline silicon (p-Si) by crystallizing amorphous silicon (a-Si) thin films. Such polycrystalline silicon thin films are widely used in microelectronics and display technology as mentioned above. P-Si has higher charge carrier mobility than a-Si, which is useful for fabricating higher speed switches or integrating higher quality drive electronics on display substrates. In addition, p-Si has a lower absorption coefficient for light in the visible spectral region, making it possible to use p-Si as a back electrode for LCD-products that transmits the backlight. Finally, the defect density of p-Si is lower than a-Si, which is a prerequisite for the fabrication of high efficiency solar cells.

The conversion of a-Si to p-Si may be performed by heat treatment at about 1000 ° C. Such a process can only be used for a-Si on heat resistant substrates such as quartz. Such materials are expensive compared to ordinary float glass for display purposes. Light induced crystallization of a-Si enables the formation of p-Si from a-Si without damaging the substrate by thermal loading during the crystallization process. Amorphous silicon can be formed on substrates such as glass, quartz or composites by inexpensive processes such as sputtering or chemical vapor deposition (CVD). Crystallization processes are known as excimer laser crystallization (ELC), sequential lateral solidification (SLS) or thin beam crystallization procedure (TDX). An overview of these different manufacturing processes is given, for example, in D.S. Knowles et al. "Thin Beam Crystallization Method: A New Laser Annealing Tool with Lower Cost and Higher Yield for LTPS Panels"; SID 05 Digest, 1-3 in the brochure of TCZ GmbH Company entitled "LCD Panel Manufacturing Moves to the next Level-Thin-Beam Directional X'tallization (TDX) Improves Yield, Quality and Throughput for Processing Poly -Silicon LCDs". Ji-Yong Park et al., "P-60: Thin Laser Beam Crystallization Method for SOP and OLED Application."

A line beam with a typical size and homogeneous intensity distribution, for example of 0.5 mm x 300 mm, is applied to silicon annealing on large substrates, for example using excimer laser crystallization (ELC). State-of-the-art optical systems use refractive optical illumination systems that include cross cylindrical lens arrays to form the desired intensity distribution. These arrays whose functionality is described, for example, in US 2003/0202251 A1, are examples of more general groups of homogenization schemes that split the input beam into multiple beams using suitably shaped sub apertures. The overlap of these multiple beams in the field plane averages the intensity variations to make the beam homogeneous. Typically, two perpendicular directions, the so-called short axis direction and the so-called long axis direction, are homogeneously separated using cylindrical optics.

WO 2006/066706 A2 discloses the production of line beams having a uniaxial half width (FWHM) size of 4 to 7 μm and a long axis size of 700 mm. Homogenizers for major axis size include cylindrical fly-eye transducers and rods.

In a laser homogenization scheme that includes a fly's eye as an integrator, a uniform intensity distribution is achieved by superimposing different portions of the laser profile at any one field point. This method is described, for example, in the article "Cylindrical fly's eye lens for intensity redistribution of an excimer laser beam" by Yoshihary Ozaki and Kiichi Takamoto, published in APPLIED OPTICS / Vol. 28, No. 1/1 January 1989, pages 106-110. Is disclosed.

Due to this overlap at any one field point of the different parts of the laser profile, interference may occur. The contrast of the interference pattern depends on the number of independent spatial coherence cells of the laser. The spatially coherent cell of a laser refers to the area when two beams are coherent with each other.

For many applications, the residual interference pattern should be reduced. Different methods and apparatuses for reducing residual interference patterns are known.

US 6,594,090 B2, for example, discloses a homogenizer consisting of a fly-eye integrator and a condenser lens. A fly's eye integrator comprising two cylindrical lens arrays is arranged in front of the condenser lens. In order to reduce the interference pattern, a rotary diffuser is used in front of the fly's eye. The effect of the rotary spreader is to change the relative phase between the interfering beams in the image plane. Rotating diffusers enhance etendue, ie the extent to which light "spreads" in areas and angles in two perpendicular directions. There are many applications where the increase in etendue does not present disadvantages, but especially in laser annealing apparatus having a very high aspect ratio homogenized beam in the substrate, the increase in etendue in the short axis direction is disadvantageous.

US 6,650,480 B2 discloses another possibility of reducing interference in homogenizers comprising flyeye. The example described here refers to the use of a distributed delay device in front of the homogenized cylindrical lens array. The distribution delay element is a quartz block having a stepped shape so that light path differences occur continuously. The size of one step should be equal to the size of one lens of the cylindrical lens array. If the difference in path length introduced is greater than the coherence length of the laser beam, the interference of light rays from different positions of the lens array, which are focused at any one field point in the image plane, can be reduced. In order to be independent of the variation of the laser beam, a large number of cylindrical lenses in the fly's eye are advantageous. In this case, a small step size distribution delay element is required. This makes the fabrication of this device very difficult.

A reduction in speckle contrast using a pulse expander is disclosed in US Pat. No. 6,693,930 B1. The reduction in contrast is performed based on the fact that laser beams entering the diffuser at different angles or positions leave the diffuser and form a changed speckle pattern. With the use of a pulse stretcher, a number of incoherence beams are formed that form a number of speckle patterns that reduce residual contrast.

US 6,191,887 discloses another system for speckle reduction. The system divides the pulse into small pulses that are temporarily separated and spatially aberrations. Spatial aberrations are located along the delay lines and deform the wave front shape. Due to aberrations, different speckle patterns occur in the image plane.

In addition to conventional optical devices having distribution delay elements, all other cited optical devices according to the prior art have the disadvantage of modifying the angle and wavefront distribution due to diffusers and aberrations. Especially in the laser annealing apparatus, this effect in the short axis direction is unacceptable.

It is an object of the present invention to provide a method for reducing interference in a laser beam homogenizer using a cylindrical lens array.

Another object of the present invention is to provide an optical system / apparatus for forming a homogenized target intensity distribution in a target field from an input light beam having an intensity distribution.

It is a further object of the present invention to provide a material processing apparatus having a highly homogenized target intensity distribution, preferably with a high aspect ratio of at least 30000, due to the reduction of the interference effect.

It is a further object of the present invention to provide a pulse expander capable of forming two incoherent laser beams traveling in different directions from one coherent incident laser beam.

Advantageous embodiments are disclosed in the dependent claims.

The method of forming the target intensity distribution in the target field is based on the first (preferably not diverging) input light beam and the second (preferably not diverging) input light beam traveling in the first travel direction. The first input light beam and the second input light beam are incident on the light incident plane of the cylindrical lens array of the homogenizer. According to the present invention, the second input light beam travels in a second travel direction different from the first travel direction. Interference patterns of both the first and second input light beams overlap in the target field resulting in a reduction of residual interference. This effect can be induced assuming that the first input light beam includes a plurality of coherent subbeams and the second input light beam also includes a plurality of coherent subbeams. The interference analysis and the respective calculation results will be described in the following paragraphs based on FIGS. 1-3.

Advantageously, said cylindrical lens array comprises a plurality of cylindrical microlenslets each having a straight apex line running in parallel, said first and second input light beams. Encloses an angle in a plane perpendicular to the vertex lines. This superimposition of the oriented input light beams results in an increased reduction of unwanted residual interferences, while angular orientation in other planes generally results in a lower reduction.

According to a preferred variant of the method according to the invention, the angle holds the following relation (1):

[Equation 1]

Where P is the pitch of the cylindrical lens array and i, m are integers.

The peak intensity and / or the peak intensity of the first input light beam such that the first input light beam and the second input light beam are incoherent with respect to each other and the peak intensities of the first and second input light beams are about the same. Assuming further that the peak intensity of the second input light beam is adjusted, in the ideal case the residual interference pattern in the target field is minimized.

As a homogenizer in reality, the superposition of two input light beams of the type mentioned above and in particular of the same peak intensity does not necessarily result in a target intensity distribution in the target field with a flat top hat profile, According to the invention the peak intensity of the first and / or second input light beam may be adjusted to minimize the residual interference pattern in the target field. This can be achieved, for example, by measuring the beam profile in the target field and combining the measured value back as a control variable that adjusts the peak intensity of the first and / or second input light beam.

It is possible that the first and second input light beams are generated by different light sources. In this case, in general, there is no phase relationship between the different light beams (unless executed by each clock pulse). Nevertheless, it is preferred that the first and second input light beams are generated by a single light source. This arrangement has the advantage that in most cases it will be known the phase relationship between the first and second light beams and it will be possible to calculate the phase relationship taking into account the design parameters of the optical system.

Only two input light beams, ie the first and second input light beams above, can be involved in the homogenization scheme according to the invention, as well as one additional or as referred to below as at least one additional input light beam. Note that multiple input light beams may be associated. Each additional input light beam not only differs in the travel directions of the first and second beams-if one or more additional input light beams are involved-but also travels in another travel direction that is also different from the travel directions of the other additional input light beams. Can be. All input light beams, ie the first and second input light beams and one or more additional input light beams, can be adjusted with respective peak intensities such that the residual interference pattern in the target field is minimized.

In a preferred embodiment, the light beam emitted by the single light source is split into a first light beam and a second light beam and may be split into one or more additional light beams, if applicable. The first light beam travels through the pulse expander without delay. The second (and applicable additional) light beam (s) is delayed and deflected with respect to the first light beam by a pulse expander. The first light beam may be used as the first input light beam and the delayed and deflected second light beam may be used as the second input light beam and if applicable one or more additional light beam (s) may be additional input light beam (s). Can be used as This embodiment has the advantage that two or more of the plurality of input light beams can be formed by an optical beam shaping device, ie a pulse expander which is mainly part of an optical system.

Deflection of the second beam (or the one or more additional light beam (s)) with respect to the first beam is wedged in the beam path of the second (or the one or more additional) light beam (s). It can be easily formed by disposing a wedge. Instead, a wedge may be disposed in the beam path of the first light beam.

Another preferred embodiment consists in that the first travel direction and / or the second travel direction and / or the additional travel direction (s) are adjusted automatically or manually. The adjustment allows to minimize the residual interference pattern in the target field.

If the light beam is split by at least one beamsplitter, the reflectivity of the at least one beamsplitter is preferably selected such that the residual interference pattern in the target field is minimized. Minimization of the residual interference pattern can be achieved by selecting the appropriate polarization dependence of the beamsplitter on light incident on the beamsplitter. Another possibility is achieved by selecting the appropriate overall reflectivity of the beamsplitter to incident light. As will be explained below with reference to the drawings, the reflectivity of the beamsplitter is preferably between 60% and 70% when allowing a tolerance of 2/3, or some degree of reflectivity. This is particularly true for beamsplitters having reflectivity independent of polarization.

In a preferred embodiment of the invention, two or more pulse expanders are arranged in succession to split the light beam into the first, second and the one or more additional light beam (s). Successive use of two or multiple pulse expanders not only allows the formation of multiple beams with different propagation directions, but also to meet certain interference conditions in the target field such as those required to minimize the residual interference pattern. The advancing directions of the beams of may be predetermined.

An optical system for forming a target intensity distribution in a target field according to the invention is:

A first light source for forming a first input light beam and a second light source for forming a second input light beam, or a single light source for forming a light beam and a beam splitter for dividing the light beam into a first input light beam and a second input light beam ,

A homogenizer having a cylindrical lens array having a light incidence surface to which the first input light beam and the second input light beam are incident,

A first guide element for guiding the first input light beam in a first travel direction upon incidence upon the light incident surface;

A second guide element for guiding the second input light beam to travel in a second travel direction different from the first travel direction when it enters the light incident surface.

The above mentioned problem is completely solved by this optical system according to the invention.

Advantageously, said cylindrical lens array comprises a plurality of cylindrical microlenses, each having a straight vertex line running in parallel, wherein said first and second input light beams are said vertex lines. Surround an angle in a plane perpendicular to the field. Overlapping of input light beams with such angular orientations with respect to each other results in increased reduction of unwanted residual interferences, while angular orientations in other planes generally result in a lesser degree of reduction.

A preferred embodiment of the optical system according to the invention is characterized by the cylindrical lens array having a pitch and an angle that maintains the relation (1) given above.

Preferably, the first input light beam and the second input light beam are incoherent with respect to each other and the peak intensities of the two input light beams are the same.

In general, the first and / or second light sources or the single light source are one or more lasers.

In a preferred embodiment the optical system according to the invention comprises a pulse expander which transmits the first light beam with little delay and delays and deflects the second light beam with respect to the transmitted first light beam. On the one hand the transmitted first light beam forms a first input light beam and on the other hand the delayed and deflected second light beam forms a second input light beam.

Preferably, a wedge is disposed in the beam path of the second light beam to deflect the second light beam.

The optical system according to the invention preferably comprises a travel direction adjustment element for automatically and / or manually adjusting the first travel direction and / or the second travel direction. The advancing direction control element may be a movable, in particular rotatable and / or linearly slidable mirror.

Alternatively or additionally, the optical system according to the invention may comprise a peak intensity adjusting element which automatically and / or manually adjusts the peak intensity of the first input light beam and / or the peak intensity of the second input light beam. have. The peak intensity adjusting element may serve to adjust peak intensities of the first and second input light beams to minimize the residual interference pattern in the target field. In an ideal case, the peak intensity adjusting element may serve to equally adjust peak intensities of the first and second input light beams. As the peak intensity adjusting element, for example, an attenuator can be used.

As already mentioned at the beginning of the present application, the present invention may be useful when constructing an apparatus for laser annealing of large substrates. The invention thus also relates to a material processing apparatus, in particular a laser annealing apparatus, which forms a target intensity distribution in a target field from an input light beam having an intensity distribution, wherein the intensity distribution in the target field extends in the first dimension and in the second dimension. It is also possible to focus on a material processing apparatus having an extension of and allowing the extension to the first dimension to exceed the extension to the second dimension by 30000 times. According to the invention, the material processing apparatus comprises an optical system according to one of the types described above.

The pulse expander according to the invention, which is capable of expanding the incident light beam and forming first and second input light beams of the kind described above:

A beamsplitter for dividing an incident light beam into a first light beam and a second light beam, wherein the first light beam passes through the beamsplitter almost without delay and proceeds in a first travel direction,

A delay line for delaying said second light beam with respect to said passed first light beam, and

A deflection element for deflecting said delayed second light beam relative to said passed first light beam such that said delayed second light beam travels in a second direction.

The pulse expander according to a preferred embodiment of the present invention may include a wedge disposed in the beam path of the second light beam to deflect the second light beam.

Preferably the pulse expander comprises a travel direction adjustment element for automatically and / or manually adjusting the first travel direction and / or the second travel direction. For example, the advancing direction control element may be a movable, in particular rotatable and / or linearly slidable mirror.

A preferred embodiment of the pulse expander according to the invention may also comprise a peak intensity adjusting element which automatically and / or manually adjusts the peak intensity of the first light beam and / or the peak intensity of the second light beam. By way of example, the peak intensities of the first and second input light beams are equally adjusted.

In a laser beam homogenizer using a cylindrical lens array, it is possible to form a homogenized target intensity distribution within a target field from an input light beam having an intensity distribution and reducing interference.

In addition, due to the reduction of the interference effect, it is possible to obtain a highly homogenized target intensity distribution, preferably with a high aspect ratio of more than 30000, and provide a pulse expander to provide two directions that travel in different directions from one coherent incident laser beam It is possible to form an incoherent laser beam.

Exemplary embodiments will be described in detail below with reference to the following attached drawings.
1 includes two cylindrical lens arrays and a condenser illuminated by two beams with different propagation directions, with only sub-beams incident on neighboring cylindrical microlenses interfering in the field plane The latest homogenizer is shown.
FIG. 2 shows the calculated interference pattern for a beam consisting of a first set of subbeams running parallel to one another before passing through the state-of-the-art homogenizer according to FIG. 1.
FIG. 3 shows that each set of sub-beams runs parallel to each other before passing the state-of-the-art homogenizer according to FIG. 1 and different sets of sub-beams run not parallel to each other before passing the state-of-the-art homogenizer according to FIG. 1, An interference pattern composed of first and second sets of sub beams is shown.
4 shows a modified pulse expander for a pulse expander according to the prior art for forming a first set of sub beams carrying an interference pattern according to FIG. 2.
FIG. 5 shows a pulse expander according to the invention forming first and second sets of sub beams carrying an interference pattern according to FIG. 3.
FIG. 6 shows another pulse expander according to the invention forming first and second sets of sub beams carrying an interference pattern according to FIG. 3.
FIG. 7 illustrates the formation of a set of delayed pulses due to multiple internal beam splitting, such that each subsequent delay pulse exits the pulse expander in a direction of travel deflected by the same angle relative to the direction of travel of the previous pulse. A pulse expander according to 5 is shown.
8 shows the intensity of out-coupled transmission pulses and delay pulses from the pulse expander according to FIG. 5 or 6 when using a beamsplitter with a reflectance of 60%.
9 shows the intensity of externally coupled transmitted and delayed pulses from the pulse expander according to FIG. 5 or 6 when using a beamsplitter with 66% reflectivity.
FIG. 10 shows a state-of-the-art homogenizer according to FIG. 1 illuminated by two beams with different propagation directions, such that sub-beams incident on three adjacent cylindrical microlenses interfere in the field plane.
FIG. 11 shows constructive interference conditions for sub-beams spaced apart by the pitch of the cylindrical lens array according to FIG. 10.
FIG. 12 shows constructive interference conditions for sub beams two times the pitch of the cylindrical lens array according to FIG. 10.
Figure 13 shows two consecutively placed pulse expanders capable of removing interference patterns in the target field when the lateral coherence length of the beam incident on the cylindrical lens array of the fly's eye homogenizer exceeds twice the array pitch. Shows.

Homogenizers with cylindrical lens arrays form spatial interference patterns in the image plane depending on the pitch of the lens array and the focal length of the condenser lens. This will be explained below.

1 shows a plan view of a typical configuration for a homogenizer in the xz-plane of a rectangular coordinate system with coordinates x, y, z. The homogenizer is in the planes 13, 14, 15 parallel to the xy-planes of the Cartesian coordinate system x, y, z and arranged along the z-direction, with the first cylindrical lens array 1a and the second cylinder. Two cylindrical lens arrays 1a and 1b having a cylindrical lens array 1b and a cylindrical condenser 2.

In the arrangement shown in FIG. 1, the first array 1a is located at the focal plane 13, ie at the focal length f _1b of the second array 1b. The collector 2 is located at a distance d from the second array 1b. The plane located at the focal length f ₂ of the light collector 2 forms a field plane. In the following, this field plane is indicated by the reference numeral 3.

The cylindrical lens arrays 1a and 1b each comprise a plurality of, preferably identical cylindrical microlenses 8a, 8b, 8c, 8d, 9a, 9b, 9c and 9d. The term cylindrical microlenses does not only include cylindrical microlenses having a circular cross section, but also includes "cylinders" having a semicircular cross section or a noncircular concave or noncircular convex cross section. For simplicity, only four identical cylindrical microlenses 8a, 8b, 8c, 8d, 9a, 9b, 9c and 9d are shown in FIG. 1 for each cylindrical lens array 1a and 1b. It is. The same cylindrical microlenses 8a, 8b, 8c, 8d, 9a, 9b, 9c and 9d of one of the cylindrical lens arrays 1a and 1b are disposed adjacent to each other in the x-axis direction. have. The individual cylindrical microlenses 8a, 8b, 8c, 8d, 9a, 9b, 9c and 9d of the cylindrical lens arrays 1a and 1b and the apex line of the cylindrical condenser 2 line) 10a, 10b, 10c, 10d, 11a, 11b, 11c, 11d, 12a, 12b follow the y-direction of the Cartesian coordinate system x, y, z.

Hereinafter, the vertex lines 10a, 10b, 10c, 10d, of adjacent cylindrical microlenses 8a, 8b, 8c, 8d, 9a, 9b, 9c, and 9d in one cylindrical lens array 1a and 1b; The distance between 11a, 11b, 11c, and 11d is called the pitches 16 and 17. In this example, the pitches 16 and 17 of the cylindrical lens arrays 1a and 1b are the same.

For the following description, it is assumed that one or more light sources (not shown) are disposed on the left side of the first cylindrical lens array 1a. The one or more light sources emit one or more light beams that are directed to the first cylindrical lens array 1a. Light in this case means electromagnetic radiation of all wavelengths.

First, assume that a first beam 4 is emitted from one of the light sources and travel in the z-direction, and then a second beam 5 is emitted from the one or another light source and in the z-axis direction. Assume that we proceed in a direction with an angle that is different from 0 ° (measured in the xz-plane). Each of the first and second beams 4, 5 comprises a plurality of sub beams 4a, 4b, 5a, 5b. Two of each of the first and second beams spaced in the x-direction by the pitches 16, 17 of the cylindrical lens arrays 1a and 1b are shown in FIG. 1.

Both beams 4, 5 are incident on the front surface of the first cylindrical lens array 1a on positions 6a, 6b and cross the field plane 3 through the homogenizer. In this case, the positions 7a, 7b of the subbeams 4a, 5a or 4b, 5b in the field plane 3 depend only on the positions 6a, 6b in the first array 1a and the first array. It is somewhat independent of the incidence angles α _4a , α _4b , α _5a , α _5b of the sub-beams 4a, 5a or 4b, 5b when incident on (1a). FIG. 1 shows the angles of incidence α _4a , α _4b , α _5a , α _5b which are angles between the advancing direction of the sub-beams 4a, 5a or 4b, 5b and the z-direction in the xz-plane. Here, for example, it is assumed that α _4a = α _4b , = 0 °, and α _5a = α _5b ≠ 0 °.

Due to the overlap of the multiple sub beams 4a, 5a or 4b, 5b, homogenization in the field plane 3 can be achieved. The sub beams 4a and 4b enter the first lens array 1a at two positions 6a and 6b which are spaced apart by the pitch 16 of the first array 1a. The two sub beams 4a and 4b are deflected in the same way and are concentrated at the same field point 7a in the field plane 3. If the two subbeams 4a and 4b are coherent with each other, interference occurs in the field plane 3. The spatial distribution of the interference pattern in the field plane 3 depends on the pitches 16 and 17 of the arrays 1a and 1b and the focal length f ₂ of the condensing lens 2.

FIG. 2 shows, for the sake of simplicity, the interference pattern of the first beam 4 consisting of only two sub-beams 4a and 4b in the field plane 3 after passing through the homogenizer shown in FIG. 18 is shown. Note that the pattern is calculated for only the sub beams 4a and 4b shown in FIG. If additional sub-beams 4c, 4d, ... (not shown) proceed in the z-direction when incident on the front of the first cylindrical lens array 1a (but the first cylindrical lens) Need not be as far apart as the pitch 16 of the array 1a), these sub-beams 4c, 4d, ... (not shown) will also contribute to the interference and the interference pattern 18 will become more complex. will be. Nevertheless, 100% contrast will be the same.

Also shown in FIG. 1, the first array is incident on the first array 1a at different angles α _5a = α _5b for the sub-beams 4a and 4b of the first beam 4 but at the same location. Two sub-beams 5a, 5b of the second beam 5 incident on 1a. Due to the operating principle of the two cylindrical lens arrays 1a and 1b, the two subbeams 5a and 5b have the same angle β _5a = β _5b = β _4a = compared to the subbeams 4a and 4b. Leave the second cylindrical lens array 1b at β _4b . The two subbeams 5a, 5b of the second beam 5 are thus condensed at the same field point 7b = 7a compared to the two subbeams 4a and 4b of the first beam 4. If the subbeams 4a and 4b of the first beam 4 are coherent to each other and the subbeams 5a and 5b of the second beam 5 are coherent to each other, (4, 5) are incoherent to each other, and the phase difference ΔΦ _{5a, 5b} between the subbeams _{5a and 5b} is the phase difference ΔΦ _{4a, 4b} and (2i + 1) * of the subbeams 4a and 4b *. If it is different by π, where i is an integer, the residual interference can be reduced to a large extent.

The phase difference ΔΦ _4,5 between the beams 4 and 5 can be achieved if the relation (1) is satisfied for the angle of incidence α = α _5a -α _4a = α _5b -α _4b . In other words,

&Quot; (2) "

Where P is the pitch of the first cylindrical lens array 1a and i is an integer.

If the amplitudes A of the beams 4 and 5 are the same and the angle α satisfies the relation (2) above, the interference patterns 18, 19 of both the beams 4 and 5 are shown in FIG. 3. Looks like. The sum 20 of the two incoherent interference patterns 18, 19 results in a constant intensity indicated by a straight line in the normalized intensity value 1 in FIG. 3.

The following embodiments form such two beams 4 and 5 from a single light source to compensate for the disadvantages of residual interference when using a flyeye homogenizer. Embodiments focus on the generation of the different beams 4 and 5 in a pulse expander.

4 shows a typical configuration of an image pulse expander for extending the traveling path of the light beam 100 in the xz plane of a rectangular coordinate system having coordinates x, y, z. Similar pulse expanders are also disclosed in US 2005/0127184 A1, US 6,928,093 B2 and WO 2005/050799 A1.

The pulse expander shown in FIG. 4 includes four spherical mirrors 103, hereinafter referred to as first spherical mirror 103, second spherical mirror 104, third spherical mirror 105, and fourth spherical mirror 106. , 104, 105, 106). The radii of curvature r1, r2, r3 and r4 of the mirrors 103, 104, 105 and 106 are the same. The first mirror 103 and the third mirror 105 are strictly arranged on a common axis of symmetry with respect to their concave surfaces arranged to face each other at a mirror distance D1 corresponding to curvature radii r1 and r3. In a similar manner, the second mirror 104 and the fourth mirror 106 are on another common axis of symmetry with respect to their concave surfaces arranged to face each other at a mirror distance D2 = D1 corresponding to the radii of curvature r2 and r4. Is placed on. The arrangement is thus a confocal or 4f arrangement of the mirrors 103, 104, 105 and 106, as a result of which the arrangement has an imaging optical characteristic of 1: -1. In addition to the mirrors 103, 104, 105 and 106, the pulse expander according to FIG. 4 is arranged in the center between the first and third mirrors 103 and 105 and tilted at 45 ° with respect to the z-axis direction. A photo beam splitter 101. In addition to the prior art, a compensation plate 102a is inserted in the delayed beam path. This compensating plate 102a is also inclined with respect to the z-axis direction but is disposed at an angle of −45 °, ie opposite to the beam splitter 101.

In the following, the principle of the pulse expander is explained.

The light beam 100 is formed by a laser, for example an excimer laser (not shown). The pulse expander has a coupling-in area 118 that serves to couple the light beam 100 in the space between the first and third mirrors 103, 105. In this case, the insertion coupling region 118 is formed by the beam splitter 101 inclined at 45 ° with respect to the incident light beam 100.

The laser beam 100 is incident on the beam splitter 101. The incident beam 100 is divided into a transmitted portion 107 and a reflected portion 108. The reflected portion 108 proceeds to and reflects on the third mirror 105. From the third mirror 105, the light beam 108 proceeds to and reflects on the second mirror 104 and to the fourth mirror 106, where the light beam 108 is then reflected. From there the light beam 108 proceeds to the first mirror 103, where the light beam 108 is reflected again. The reflected portion 108 thus proceeded through a confocal distribution configuration and imaged back into the beamsplitter 101. The "rear" of the insertion coupling region 118 of the beam splitter 101, formed in a reflective manner, is an output coupling region 120 for coupling the light beam 108 out of the space between the two mirrors 103 and 105. Serves as. Therefore, along the arrow 109, the delayed light beam 108 after four cycles leaves the pulse expander, where the use of spherical mirrors 103, 104, 105 and 106 is above the output coupling region 20. Since 18) is imaged 1: 1, the output coupled light beam 109 and the inserted coupled light beam 100 after the pulse expander are placed on the same optical axis and have the same shape and cross-sectional area. The propagation path of the light beam 100 was then lengthened in the pulse expander by approximately four times the mirror distance D.

The additional compensation plate 102a disposed in the traveling path of the delayed light beam 108 in front of the beam splitter 101 serves to compensate only the shift of the beam splitter 101. If the compensation plate 102a has the same thickness d and the refractive index n as the beam splitter 101 and the compensation plate 102a is reflected by the first mirror 103 as shown in FIG. 4. If oriented at 45 ° with respect to the later reflected portion 108, the transmitted beam 107 and the delayed beam 108 are pulsed at the same angle to form one single output coupled light beam 109 without offset. Leaving the expander.

Of course, the remaining characteristics of the delayed beam 108 depends on the adjustment of the pulse expander. In the laser annealing apparatus, the divergence and direction indication to the short axis for the delayed beam 108 should be close to the divergence of the incident light beam 100. If this condition is achieved, the divergence and direction indication in the long axis also becomes very close to the characteristics of the incident light beam 100.

Assuming the transmitted beam 107 forms the first beam 4 of FIG. 1 and the delayed beam 108 forms the second beam 5, the sub beams 5a and 5b of FIG. It will have almost the same characteristics as the beams 4a and 4b. This means that the interference patterns will not be very different. Although the beams 4 and 5 are incoherent, the sum does not lead to a significant reduction in the interference.

Therefore, in the preferred embodiment according to the present invention shown in Fig. 5, the compensation plate 102b is a wedge for introducing a direction indication in a predetermined direction. The angle γ of the wedge 102b must be sized such that the resulting angle α of the delayed beam 108 satisfies the relation (1) given above. The predetermined direction is the same as the direction in which engagement with the fly's eyes 1a, 1b and 2 occurs. The reflectivity of the beamsplitter should be designed such that the transmitted beam 107 and the delayed beam 108 have approximately the same peak intensity. In this case, the reduction of contrast is very effective.

6 shows another embodiment according to the invention for dividing the laser beam 100 into a transmission beam 107 and a retardation beam 108 within the xz-plane of a rectangular coordinate system with coordinates x, y, z. . The difference from the embodiment of FIG. 4 is that instead of the spherical mirrors 103, 104, 105, and 106, the planar mirrors 112, 113, 114, 115, that is, the first planar mirror 112, the second planar mirror ( 113, the third planar mirror 115 and the fourth planar mirror 114 are used. Planar mirrors 112 and 114 are inclined at -45 ° with respect to the z-axis and planar mirrors 113 and 115 are inclined at 45 ° with respect to the z-axis. The planar mirrors 112, 113, 114, and 115 are arranged in pairs with reflecting surfaces of rectangles facing each other. The beam splitter 111 is positioned at the center between the first and third mirrors 112 and 115. The beam splitter 111 has the same orientation as that of FIG. 4, ie, tilted at 45 ° with respect to the z-axis.

Assume again that the laser beam 100 travels in the negative x-axis direction. The laser beam 100 is incident on the beam splitter 111. The incident beam 100 is divided into a transmission portion 107 and a reflection portion 108. The transmissive portion 107 is deflected twice, ie at the front 118 and rear 121 of the beam splitter 111. Along arrow 110, transmissive portion 107 leaves the pulse expander. The output coupled transmission light beam 110 and the insertion coupled light beam 100 are again placed on the same optical axis and have the same shape and cross-sectional area. The reflective portion 108 proceeds to and reflects on the third mirror 115. From the third mirror 115, the light beam 108 proceeds to the fourth mirror 114 where it is reflected and proceeds to the second mirror 113, where the light beam 108 is then reflected. From there the light beam 108 proceeds to the first mirror 112 where the light beam 108 is reflected again. Thus, the reflective portion 108 travels through the distributed configuration and is reflected from the first mirror 112 back to the beamsplitter 101. The "rear" of the insertion coupling region 118 of the beam splitter 101 formed reflectively serves as the output coupling region 120 for coupling the light beam 108 out of the space between the two mirrors 112 and 115. do. Therefore, along arrow 109, after being reflected at the four planar mirrors 112, 113, 114, 115, the light beam 108 leaves the pulse expander, and the output coupled light beam 109 and the insert coupled light beam ( 100) lies on different optical axes but has the same shape and cross-sectional area. Thus, the propagation path of the light beam 100 extends in the pulse expander by approximately the sum of the distances between the mirrors 112, 113, 114, 115. The difference from using planar mirrors 112, 113, 114, and 115 instead of using spherical mirrors 103, 104, 105, and 106 is that planar mirrors 112, 113, 114, and 115 are incident laser beams. It does not phase (100).

In the case of a delay line which does not form an image but is smaller than the coherence length of the laser, further expansion of the delayed beam 108 may be allowed.

An advantageous variant of this embodiment is to have at least one rotatable mirror (here mirror 112) of the planar mirrors. The rotatable mirror 112 allows the angle α of the beams 5a and 5b in FIG. 1 to be continuously adjusted due to the rotation of the mirror 112. If the intensity distribution in FIG. 3 is monitored, the angle of the mirror 112 can be adjusted to minimize the residual interference pattern.

A further advantageous variant of this embodiment is to have at least one of the planar mirrors shiftable (here also the mirror 112). An advantage of the possibility of shifting the mirror is to adjust the beam position of the delayed beam 108.

To ensure that the delayed beam is fully reflected, the beamsplitter 101 as in FIG. 5 or the beamsplitter 111 as in FIG. 6 transmits a portion of the input beam 100 and the remaining portion ( A polarizing beamsplitter that reflects delay beam 108, and then the remaining portion is fully reflected upon a second incidence of the beamsplitter leaving the pulse expander.

In general, however, the beamsplitter 101 or 111 will not fully reflect the delayed beam 108 and will repartition the delayed beam 108 to form another beam 122 that once again circulates within the pulse expander. The another beam is then partially reflected again, output coupled and partially transmitted. As a result, a series of pulses are formed by the pulse expander, with each pulse further delayed by a certain amount of time. FIG. 7 shows each situation in the case of the beam expander shown in FIG. 5 with respect to the center ray.

The relative intensity of the pulses is determined by the reflectivity of the beamsplitter. If the optical losses (eg by absorption) in the pulse expander can be neglected, the intensity of the pulses (relative to the input beam) is given by the formula below.

&Quot; (3) "

Intensity of transmitted pulse (0th pulse) t

Intensity of first delayed pulse (first pulse) r ²

Intensity of second delayed pulse (second pulse) tr ²

...

Intensity of the nth delayed pulse (nth pulse) t ^n-1 r ²

...

Total 1

Where r is the reflectivity of the beamsplitter and t = 1-r is the transmittance.

For the case of r = 60%, the relative intensities of the different pulses are shown by way of example in FIG. 9.

If the wedge 102b is inserted in the pulse expander instead of the compensating plate 102a, or if the mirror 112 is adjusted to introduce the angular shift α of the first delayed beam 108, each successive pulse is also n It will be deflected by the angle α _n = _n α for the -th pulse (see FIG. 7).

If the cylindrical lens array is now illuminated with these multiple beams, for example the cylindrical lens array 1a in the arrangement according to FIG. 1, the phase difference of the n th beam relative to the direct beam is approximately first. N times the phase difference between the beam and the direct beam (if the angle is small). Thus, if angle α is adjusted for a phase difference of (2i + 1) π according to a formula for minimizing interference (i.e., relation (1) for the first delayed beam out of phase), then the nth beam is n (2i + 1) pi will have a phase difference and even n beams will be in phase with the direct beam, odd n beams will not be out of phase and all will be added to form the final interference pattern. Similar considerations apply when α is chosen differently from (2i + 1) π, but in general this is less preferred.

This means that if multiple beams are present from the pulse expander, the relative intensities of the multiple beams must be adjusted so that the overall interference pattern disappears. Since there is only one easy usable means for adjusting the intensity, namely the reflectivity of the beamsplitter, the reflectivity of the beamsplitter should be chosen so that the overall interference is minimized. In the above case with odd-numbered beams out of phase and odd-numbered beams out of phase, the condition for complete cancellation is that the added intensities of these two groups are equal (i.e. 50% of input beam intensity, respectively). to be. The following formula from above can be used to determine the optimal reflectivity:

&Quot; (4) "

Intensity of transmitted pulse (0 pulse) t

Intensity of first delayed pulse (1 pulse) r ²

Intensity of second delayed pulse (2 pulses) tr ²

Intensity of third delayed pulse (3 pulses) t ² r ²

Intensity of the fourth delayed pulse (4 pulses) t ³ r ²

* Intensity of the fifth delayed pulse (5 pulses) t ⁴ r ²

...

Total t + tr ² / (1-t ² ) r ² / (1-t ² )

The condition of equal intensity is then r = 2/3 = 66.67%. The resulting relative intensities are shown in FIG.

If the reflectivity r deviates from this optimal value,

[Equation 5]

(I _max -I _min ) / (I _max + I _min )

It can be seen that the residual interference contrast defined as | (2-3r) / (2-r) |, where I _max is the maximum intensity of the residual interference pattern and I _min is the minimum intensity. For example, when r = 50%, the interference contrast is 33.33%.

The above calculation is based on the assumption that only interference of sub-beams passing through neighboring microlenses of the lens array occurs, for example, microlenses 8a and 8b or 8b and 8c of lens array 1a of FIG. I did it. However, if the lateral coherence length of the input beam exceeds twice the array pitch P, then microlenses with greater spacing as well as from neighboring microlenses (eg microlenses 8a and 8c in FIG. Or interference exists between the sub beams from 8a and 8d). For these multiple interference paths with different distances of interference rays at the starting surface, one pulse expander as described above generally cannot completely eliminate the interference pattern. For example, if there is interference between rays having a distance P (as shown in FIG. 11) and 2P (as shown in FIG. 12), the following conditions (from relation (1)) are:

&Quot; (6) "

sin (α) = λ (2i + 1) / 2P

And

[Equation 7]

sin (α) = λ (2j + 1) / 4P

It must be satisfied at the same time, which is not possible for all integers i and j.

In this case, as shown in FIG. 13, interference can be eliminated by using two pulse expanders 123, 124. Starting from the input beam 125, the first pulse expander 123 forms pulses 126, 127 having a traveling direction surrounding an angle α ₁ that satisfies the condition 6 and the second pulse expander 124. Form pulses 128, 129, 130, 131, 132, and 133 with advancing directions surrounding different angles α ₂ satisfying condition (7). The beams 126, 127 leaving the first pulse expander 123 then serve as input beams for the second pulse expander 124.

For each interval p _k of interfering rays the following conditions:

[Equation 8]

sin (α _k ) = λ (2jk + 1) / 4p _k

One pulse expander or any other angle-forming element may be used that is more general for each interference pattern (i.e. each spacing of the interference rays) that needs to be removed, each forming an angle α _k that satisfies .

The total number of pulse expanders can be reduced if one or more of the formed angles a _k satisfy the formula (8) simultaneously for one or more intervals p _k of light rays. For example, for a constant interval p _k = kP (multiple of the array pitch, k is an integer), the number of pulse expanders can be reduced to m = ceiling [log ₂ k], where ceiling [] is the following integer Is a mathematical operator that increments to. The angle α _m is then:

&Quot; (9) "

sin (α _m ) = λ (2j _m +1) / 2 ^m P

Must be satisfied.

Claims (6)

In the pulse expander for expanding the incident light beam 100,
Splitting the incident light beam 100 into a first light beam 107 and a second light beam 108 and into two or more additional light beams 122, wherein the first light beam 107 transmits without delay and undergoes a first travel Beamsplitters 101 and 111 for advancing in a direction,
Retard the second light beam 108 with respect to the transmitted first light beam 107 and one of the two or more additional light beams 122 with respect to the second light beam and other of the two or more additional light beams. Delay lines 103, 104, 105, 106 for delaying
Deflecting the delayed second light beam 108 relative to the transmitted first light beam 107 so that the delayed second light beam 108 travels in a second direction and also with respect to the transmitted first light beam 107 And deflecting elements (101, 102b; 111, 112) for deflecting one of said delayed two or more additional light beams 122 with respect to said delayed second light beam and with respect to other of said two or more additional light beams. Featured pulse expander. The method of claim 1,
And a wedge (102b) disposed in the beam path of the second light beam (108) to deflect the second light beam (108) and the two or more additional light beams (122). 3. The method according to claim 1 or 2,
And a travel direction adjustment element (112) for automatically or manually adjusting one or more travel directions from among the first travel direction, the second travel direction, and the two or more additional travel directions. The method of claim 3, wherein
And the adjusting element is a movable mirror (112). 3. The method according to claim 1 or 2,
The first light beam 107 of the peak intensity (A _4), the second light beam (108) of the peak intensity (A _5), one or more of the peak intensity from the peak intensity of the two or more further light beam (122) automatically or The pulse expander further comprises a manually adjusted peak intensity control element. The method of claim 5, wherein
And the peak intensities (A ₄ , A ₅ ) of the first and second input light beams (107, 108) and the two or more additional light beams (122) are equally adjusted.

Images