Method and optical system for processing semiconductor materials

Abstract

Disclosed are a method and an optical system for processing a semiconductor material layer, in particular for generating a crystalline semiconductor layer. The method comprises the following steps: providing a first laser beam (74) having a first laser pulse (76) and a second laser beam (84) having a second laser pulse (86); shaping the first laser pulse (76) and the second laser pulse (86) into linear laser pulses having a short axis and a long axis by means of a beam shaping device; imaging the shaped linear laser pulses into an irradiation line having a short axis and a long axis on the semiconductor material layer by means of an imaging device, wherein the method further comprises the following steps: setting the polarization direction of the first laser pulse (76) in the direction of the short axis of the irradiation line (36); setting the polarization direction of the second laser pulse (86) in the direction of the long axis of the irradiation line; and causing the second laser pulse (86) to be delayed in time relative to the first laser pulse (76) by a predetermined time interval Δt, wherein the predetermined time interval Δt is selected such that the irradiation line imaged on the semiconductor material layer has a combined time intensity variation curve (96) of the pulse shape of a first maximum value (M1) and a second maximum value (M2).

Description

Technical Field

[0001] The present disclosure relates to a method for processing semiconductor material, in particular for growing a crystalline semiconductor layer, and an optical system for processing semiconductor material, in particular for growing a crystalline semiconductor layer. Background Art

[0002] Lasers are used to crystallize thin-film layers, such as those used to produce thin-film transistors (TFTs). The semiconductor to be processed is, in particular, silicon (Si), more specifically, amorphous silicon (a-Si). The semiconductor layer has a thickness of, for example, 50 nm and is typically located on a substrate, such as a glass substrate, or another carrier.

[0003] The layer is irradiated with laser light, such as light from a pulsed solid-state laser. Light having a wavelength of, for example, 343 nm is formed into an irradiation line and imaged onto an imaging plane of the semiconductor material. The irradiation line has a short (narrow) axis and a uniform long beam axis. The short or narrow axis has a Gaussian or flat intensity distribution.

[0004] The irradiation beam is moved across the semiconductor layer in the direction of its short axis at a feed rate of typically about 5 to 50 mm / s. The power density of the beam (in the case of a continuous-wave laser) or the pulse energy density (in the case of a pulsed laser) is set to a level that, in the case of amorphous silicon, partially melts it and subsequently solidifies the molten silicon in a polycrystalline structure, starting with unmelted solid silicon, on the glass substrate. Melting and solidification typically occur in the range of 10 to 100 nanoseconds, with subsequent cooling of the film to room temperature typically taking several hundred microseconds.

[0005] When irradiating an amorphous silicon layer and converting it into a polycrystalline silicon layer, uniform radiation intensity—that is, the uniformity of the spatial intensity distribution integrated along the minor and / or major axes—is particularly important. The more homogeneous or uniform the intensity distribution of the radiation, the more homogeneous or uniform the crystal structure of the thin-film layer (e.g., the grain size of the polycrystalline layer), and the better the electrical properties of the final product made of the thin-film layer, such as a thin-film transistor. A uniform crystal structure can lead to high electrical conductivity, for example, due to the high mobility of electrons and positively charged holes.

[0006] As the irradiation line moves along the short axis of the semiconductor layer, non-uniformities may appear particularly along the long beam axis and the short beam axis, which is perpendicular to the long beam axis. This non-uniformity is called "mura." So-called "scanning mura" originates from non-uniformities along the beam axis and appears as stripes extending in the scanning or feed direction. Perpendicular to this is so-called "shot mura," which is caused by fluctuations in intensity and energy density from pulse to pulse.

[0007] To minimize "photographic mura," fluctuations in the energy density and intensity variation over time from laser pulse to laser pulse should be minimized, for example by using lasers with very good pulse stability and by superimposing the laser beams from multiple laser sources. To minimize "scanning mura," the intensity of the irradiated line should be as uniform as possible along the long axis. Furthermore, it is known to reduce "scanning mura" by reciprocating the irradiated line by 1 to 2 mm along the long axis during scanning using a mirror that oscillates at a frequency of 10 to 200 Hz about its rotation axis, thereby "smearing out" any unevenness along the long axis.

[0008] So-called "Nitrogen (Stickstoff) Mura" is also known, which originates from the fact that the exposed surface of the irradiated substrate Specifically, during exposure, the surface of a material layer to be processed (e.g., a semiconductor layer) is flushed with nitrogen to reduce the oxygen concentration there to between 10 ppm and 20 ppm, thereby preventing oxidation of the material (e.g., silicon). To this end, a laminar flow of nitrogen flows directly over the material layer to be exposed. Inhomogeneities in the laminar flow can lead to inhomogeneities in the crystal structure, so-called "nitrogen mura."

[0009] In order to produce a regular polycrystalline grain structure during the crystallization process, it is known to utilize the surface interference effect, which produces a modulated intensity distribution during exposure and strengthens the particle structure of about the size of the wavelength of light through repeated exposure during the feed. This effect is called "Laser Induced Periodic Pattern Structure (LIPPS)". Therefore, at a wavelength of 343nm, for example, a particle structure of about 0.3μm to 0.4μm is given. For linearly polarized light, the modulated intensity distribution is only formed in the polarization direction, that is, in the direction of the electric field vector. Experimental studies have shown that when light is polarized on the long axis, a regular structure along the long beam axis is formed, and correspondingly, when the light is polarized in the feed direction, the effect can be observed in the feed direction. Summary of the Invention

[0010] The present disclosure provides an improved method for processing semiconductor materials, particularly a method for producing a uniformly crystallized semiconductor layer. The present disclosure also provides an improved apparatus for processing semiconductor materials, particularly an apparatus for producing a uniformly crystallized semiconductor layer. A uniformly crystallized semiconductor layer is particularly one having a uniform grain size.

[0011] A method for processing a semiconductor material layer, in particular for producing a crystalline semiconductor layer, is disclosed, comprising the following steps:

[0012] - providing a first laser beam having first laser pulses and a second laser beam having second laser pulses,

[0013] - shaping the first laser pulse and the second laser pulse into a linear laser pulse having a short axis and a long axis by means of a beam shaping device,

[0014] - imaging the thus shaped linear laser pulse onto the semiconductor material layer by means of an imaging device as an illumination line having a short axis and a long axis,

[0015] - setting the polarization direction of the first laser pulse in the short axis direction of the irradiation line,

[0016] - setting the polarization direction of the second laser pulse in the long axis direction of the irradiation line, and

[0017] - Delaying the second laser pulse in time relative to the first laser pulse by a predetermined time interval Δt, wherein the predetermined time interval Δt is selected so that the radiation line imaged on the semiconductor material layer has a pulse-shaped combined time intensity variation curve, which has a first maximum and a second maximum.

[0018] The order of the method steps given above does not necessarily reflect the chronological order in which the steps are to be implemented. Generally, the order of the method steps given above does not necessarily reflect the chronological order in which the steps are to be implemented. Typically, for example, the polarization direction in the beam path is set at the point where the individual beams, i.e., the first and second laser beams, are separated. After providing the first and second laser beams, the individual beams need to be shaped, homogenized, and superimposed. The time delay step can be performed before or after the polarization alignment of the first and second laser beams. In principle, the chronological order of the steps given is also possible.

[0019] The semiconductor material to be processed may be, for example, a thin layer of amorphous silicon with a thickness of approximately 50 nm applied to a carrier, such as a glass substrate.

[0020] The first and second laser beams are provided by at least one laser, as described in further detail below. The laser can be, for example, a UV solid-state laser emitting light at a wavelength of 343 nm. The full width at half maximum (FWHM) of the first and second laser pulses ranges from 15 ns to 20 ns. The first and second laser pulses are typically linearly polarized.

[0021] In another method step, the polarization of the first and second pulses is set in a specific, predefined direction, for example, using a polarization device. Thus, the first laser pulse is linearly polarized in the direction of the minor axis of the illumination line, and the second laser pulse is linearly polarized in the direction of the major axis of the illumination line. Pulses emitted by solid-state lasers are typically linearly polarized. If the emitted pulses are linearly polarized, according to the disclosed method, the polarization directions of the emitted first and second pulses are each rotated in a predetermined, defined direction. This can be achieved using a polarization device, such as a λ / 2 plate, which has a corresponding orientation relative to the light beam or pulse impinging on the λ / 2 plate. In particular, the polarization direction of the first pulse is rotated in the direction of the minor axis, as will be described below. As a result, the polarization of the first pulse is aligned in the direction of the minor axis, that is, aligned almost exclusively in the direction of the minor axis, so that, for example, the linearly polarized light has a component perpendicular to the minor axis of only 1% (polarization ratio of 100:1), or, for example, a polarization ratio of 95:5. The polarization direction of the second pulse is rotated in the direction of the major axis, perpendicular to the minor axis, as will be described below. Therefore, the polarization of the second pulse is aligned in the long axis direction, that is, the polarization is aligned almost only in the long axis direction, for example, the component of the linearly polarized light in the direction perpendicular to the long axis accounts for only 1% (polarization ratio is 100:1), or for example, the polarization ratio is 95:5.

[0022] Furthermore, the second pulse is delayed in time relative to the first pulse by a predetermined time interval Δt. The time delay is typically 5 to 20 ns. Time interval Δt is typically selected such that, in the imaging of the radiation on the semiconductor material layer, as will be described below, the temporal superposition of the two pulses results in a single pulse. This temporal superposition results in a combined temporal intensity variation having a first maximum and a second maximum.

[0023] In a further method step, the first and second pulses are shaped into a laser line, i.e., a linear laser pulse, by means of a beam shaper. The beam shaper can be configured as an anamorphic optical element. For example, the beam shaper can comprise a lens array homogenizer, which is based on the principle of splitting one or more incident laser beams into four sub-beams and subsequently spatially superimposing them. The laser line has a short axis and a long axis.

[0024] The laser line thus formed is imaged onto an imaging plane of the semiconductor material using an imaging device as an illumination line. The illumination line also has a minor axis and a major axis, the directions of which predetermine the polarization directions of the first and second pulses. Typically, the directions of the minor and major axes of the illumination line coincide with those of the linear laser pulses. The length of the illumination line, i.e., its geometrical extent in the direction of the major axis, is typically between 100 mm and 1000 mm, for example, 100 mm, 250 mm, 750 mm, or 1000 mm. Longer illumination lines can also be achieved by appropriately designing the beam shaping and / or imaging device. The width of the illumination line, i.e., its geometrical extent in the direction of the minor axis, is given as the half-maximum width (FWHM) for a Gaussian distribution and is typically between 20 μm and 200 μm. For a flat distribution, this width is measured at the 90% intensity position (90% full width) and is typically between 20 μm and 200 μm.

[0025] In another method step, the irradiation line can be moved in a feed direction relative to the semiconductor material layer. In this way, since the feed direction corresponds to the minor axis direction, the first pulse is linearly polarized in the feed direction. By sweeping the irradiation line across the semiconductor material layer, the entire semiconductor material layer, or at least a large region of the semiconductor material layer, can be exposed and thus processed. The carrier with the semiconductor material can, for example, be placed on a table that moves in the feed direction and thus can be moved relative to the irradiation line. The feed speed is typically 5 mm / s to 50 mm / s.

[0026] According to another aspect, the relative intensities of the first and second laser pulses are selected such that the ratio of the first maximum to the second maximum in the combined temporal intensity profile is in the range of 0.8 to 1.4, in particular in the range of 0.9 to 1.2, and in particular 1.0. Since the combined temporal intensity profile is obtained by superimposing the temporal intensity profiles of the first and second pulses, the ratio of the first maximum to the second maximum in the combined temporal intensity profile can be set by the intensities of the first and second pulses, also taking into account the time interval Δt. It has been found that, in combination with a defined polarization direction of the first and second pulses, a very uniform grain structure can be achieved for a given range of the ratio of the first maximum to the second maximum in the resulting (poly)crystalline semiconductor layer, as described above.

[0027] The combined temporal intensity profile of the radiation can have a temporal half-value width associated with a first maximum of the combined temporal intensity profile of between 40 ns and 50 ns. The relatively long pulse duration influences the crystallization process by tens of ns and promotes the formation of a homogeneous grain structure.

[0028] According to another aspect, a first laser and a second laser are provided, each configured to emit a first laser beam and a second laser beam, respectively, and controlled so that the emission of the second laser pulse is delayed by a time interval Δt relative to the emission of the first laser pulse. This delay can be achieved, for example, by electronically delaying a trigger signal of the second laser relative to a trigger signal of the first laser.

[0029] Alternatively, a first laser is provided that is designed to provide a pulsed laser beam, and the laser beam of the first laser is split into a first laser beam component and a second laser beam component, wherein the first laser beam component forms a first laser beam having first laser pulses, and the second laser beam component forms a second laser beam having second laser pulses. Thus, in this alternative, a laser is provided that operates in pulsed mode and splits the laser beam it emits into the first laser beam and the second laser beam by means of a beam splitter.

[0030] In this alternative, the time delay of the second pulse relative to the first pulse can be achieved by making the optical path length of the second laser beam from the splitting point to the imaging plane of the semiconductor material greater than the optical path length of the first laser beam from the splitting point to the imaging plane of the semiconductor material, thereby causing a phase shift in the first pulse relative to the second pulse.

[0031] In principle, of course, in a variant it is also possible to use two lasers, the time delay of the second pulse relative to the first pulse being provided by the longer optical path of the second pulse.

[0032] According to another aspect, the first laser pulse can be one of a plurality of first laser pulses of the first laser beam, and the second laser pulse can be one of a plurality of second laser pulses of the second laser beam, wherein each of the plurality of laser pulses of the pulsed second laser beam is delayed in time by a predetermined time interval Δt relative to another of the plurality of laser pulses of the pulsed first laser beam. Thus, one or more lasers are operated in a pulsed manner and emit a plurality of laser pulses at a specific pulse repetition rate, for example, 10 kHz. The laser pulses of the second laser beam are delayed in time relative to the first laser pulses such that the first and second laser pulses always overlap to form a pulsed irradiation line having first and second maxima, which is imaged onto the semiconductor material. In other words, the semiconductor material is exposed to the irradiation line in pulsed form at the pulse repetition rate of the laser.

[0033] The feed speed, the pulse repetition rate of the first laser beam and the second laser beam, and the geometric half-width of the irradiation line in the minor axis direction are selected so that a location of the semiconductor material is exposed multiple times by the irradiation line. In other words, the semiconductor material is moved slowly relative to the irradiation line so that the geometric half-width of the irradiation line in the minor axis direction is large enough to correspond to the pulse repetition rate of the pulse beam or a multiple of the pulse repetition rate after a period of time. This results in a very short distance traveled by the semiconductor material, allowing previously exposed locations to be exposed again or multiple times.

[0034] According to another aspect of the method disclosed herein, a third laser beam having a third laser pulse and a fourth laser beam having a fourth laser pulse are provided, the first laser pulse, the second laser pulse, the third laser pulse, and the fourth laser pulse are shaped into linear laser pulses having a short axis and a long axis by a beam shaping device, and the linear laser pulses thus shaped are imaged into an illumination line on an imaging plane of the semiconductor material layer by an imaging device. Furthermore, the polarization direction of the third laser pulse is set in the direction of the short axis of the illumination line, the polarization direction of the fourth laser pulse is set in the direction of the long axis of the illumination line, and the fourth laser pulse is time-delayed by a predetermined time interval Δt relative to the third laser pulse, the predetermined time interval Δt being selected such that the illumination line imaged on the semiconductor material layer has a pulse-shaped combined temporal intensity variation curve having a first maximum value and a second maximum value.

[0035] Therefore, according to this aspect of the method, four pulse beams are uniformly superimposed and imaged, wherein two of the four pulse beams have pulses that are linearly polarized in the short axis direction of the irradiation line and are synchronized in time, and the other two of the four pulse beams have pulses that are linearly polarized in the long axis direction of the irradiation line and are delayed in time relative to the pulses of the initial two pulse beams.

[0036] According to another aspect of the present disclosure, an optical system for processing a semiconductor material layer, in particular for growing a crystalline semiconductor layer, is provided, comprising:

[0037] a beam shaping device configured to shape the first laser pulse of the first laser beam and the second laser pulse of the second laser beam into linear laser pulses having a short axis and a long axis,

[0038] - an imaging device arranged to image the thus shaped linear laser pulse onto the semiconductor material layer as an illumination line having a short axis and a long axis,

[0039] - polarization means arranged to align the polarization direction of the first laser pulse in the direction of the short axis of the irradiation line and the polarization direction of the second laser pulse in the direction of the long axis of the irradiation line, and

[0040] - a delay device, which is configured to delay the second laser pulse relative to the first laser pulse by a predetermined time interval Δt, wherein the predetermined time interval Δt is selected so that the irradiation line imaged on the semiconductor material layer has a pulse-shaped combined time intensity variation curve, which has a first maximum value and a second maximum value.

[0041] Therefore, the polarization device is designed and arranged so that the first laser pulse includes almost exclusively a polarization component in the direction of the short axis of the illumination line (polarization ratio of, for example, 1:100), and the second laser pulse includes almost exclusively a polarization component in the direction of the long axis of the illumination line (polarization ratio of, for example, 1:100). The polarization device may include a first polarization device for the first laser beam including the first laser pulse, and a second polarization device for the second laser beam including the second laser pulse. In this manner, the first polarization device is particularly arranged in the beam path of the first laser beam, and the second polarization device is particularly arranged in the beam path of the second laser beam.

[0042] The beam shaping device can be configured as an anamorphic optical device. The beam shaping device can, for example, comprise a lens array homogenizer, which is based on the principle of splitting one or more incident laser beams into four sub-beams and subsequently spatially superimposing them. The laser line has a short axis and a long axis.

[0043] The laser line thus formed is imaged onto an imaging plane of the semiconductor material as an illumination line by means of an imaging device. The illumination line also has a minor axis and a major axis, the directions of which predetermine the polarization directions of the first and second pulses. Typically, the minor and major axes of the illumination line coincide with the minor and major axes of the linear laser pulse.

[0044] The polarization device of the optical system may in particular include a first λ / 2 plate, which is arranged in the beam path of the first laser beam, in particular before the beam shaping device, and oriented relative to the first laser pulse incident on the λ / 2 plate, so that the first laser pulse is linearly polarized in the short axis direction after passing through the λ / 2 plate; and a second λ / 2 plate, which is arranged in the beam path of the second laser beam, in particular before the beam shaping device, and oriented relative to the second laser pulse incident on the λ / 2 plate, so that the second laser pulse is linearly polarized in the long axis direction after passing through the λ / 2 plate.

[0045] The first λ / 2 plate is oriented so that it rotates the polarization direction of the first laser beam having linear polarization of the first laser pulse in the direction of the minor axis. The second λ / 2 plate is oriented so that it rotates the polarization direction of the second laser beam having the second laser pulse in the direction of the major axis.

[0046] According to one variant, the delay device comprises a delay circuit for setting a trigger signal for the second laser to emit a second laser beam having a second laser pulse delayed by a time interval Δt relative to a trigger signal for the first laser, wherein the first laser is configured to emit a first laser beam having a first laser pulse. The second trigger signal can be electronically delayed relative to the first trigger signal.

[0047] According to one variant, the delay device has a beam detour, so that the optical path length of the second laser beam to the imaging plane of the semiconductor material layer is greater than the optical path length of the first laser beam to the imaging plane of the semiconductor material layer. According to this variant, the time delay is caused by the path difference by the second laser pulse having a longer optical path length to the superposition than the first pulse.

[0048] The first and second laser beams of the optical system may be provided by first and second laser light sources, or alternatively by one laser light source, the laser beam emitted by which is then split into the first laser beam and the second laser beam by a beam splitter.

[0049] According to another aspect, the beam shaping device can be configured to shape the first laser pulse of the first laser beam, the second laser pulse of the second laser beam, the third laser pulse of the third laser beam, and the fourth laser pulse of the fourth laser beam into linear laser pulses having a short axis and a long axis; the imaging device can be configured to image the linear laser pulses thus shaped into an irradiation line having a short axis and a long axis on the semiconductor material layer; the polarization device can be configured to linearly polarize the third laser pulse in the short axis direction of the irradiation line and the fourth laser pulse in the long axis direction of the irradiation line; and the delay device can be configured to delay the fourth laser pulse by a predetermined time interval Δt relative to the third laser pulse, and the predetermined time interval Δt is selected so that the irradiation line imaged on the semiconductor material layer has a pulse-shaped combined time intensity variation curve, and the combined time intensity variation curve has a first maximum value and a second maximum value.

[0050] According to this aspect, the optical system includes a four-beam arrangement, wherein four laser beams are uniformly superimposed and imaged onto a semiconductor material layer as an illumination line, wherein a polarization device is arranged and configured so that two of the four pulse beams respectively have pulses that are linearly polarized in the short axis direction of the illumination line and are synchronized in time, and so that the other two of the four pulse beams respectively have pulses that are linearly polarized in the long axis direction of the illumination line and are delayed in time relative to the pulses of the initial two pulse beams.

[0051] The present disclosure also includes an apparatus for processing a semiconductor material layer, in particular for producing a crystalline semiconductor layer, comprising an optical system according to the above aspects, wherein the apparatus is configured to move the semiconductor material layer in a feed direction relative to an irradiation line, wherein the feed direction corresponds to the direction of the minor axis of the irradiation line. The semiconductor material layer can be moved relative to the irradiation line, for example, by means of a feed device, such as a worktable that moves in the feed direction, so that a large area, up to the entire semiconductor layer, is exposed to the irradiation line and thereby processed, wherein a carrier having the semiconductor material layer is placed on the movable worktable. The feed direction corresponds to the direction of the minor axis, so that the polarization orientation of the first pulse and / or the third pulse corresponds to the feed direction. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] The present invention will be described in detail below with the aid of the accompanying drawings.

[0053] Figure 1 A schematic diagram of a semiconductor material layer is shown, which, for processing the semiconductor material layer, is exposed to an irradiation line that moves in a feed direction relative to the semiconductor material layer;

[0054] Figures 2a to 2c The linear geometry of the imaged illumination line is shown;

[0055] Figure 3a and Figure 3b A schematic illustration of an optical system of a device for processing semiconductor layers is shown, with the aid of which an irradiation line can be formed and imaged onto the semiconductor material;

[0056] Figure 4 shows a schematic diagram of an embodiment of an optical system, wherein the first laser beam and the second laser beam are provided via a beam splitter of the laser beams;

[0057] Figure 5 A schematic diagram of an embodiment is shown, in which four laser beams are provided by four laser light sources, wherein the pulses of two laser beams are respectively emitted with a time delay relative to the pulses of the other two laser beams;

[0058] Figure 6 A schematic diagram of an embodiment is shown, in which two laser beams are provided by two laser sources, wherein the pulses of one laser beam are delayed in time with respect to the pulses of the other laser beam;

[0059] Figure 7 Schematic representation of the combined intensity distribution of the radiation lines over time, which is given by the uniform superposition of the individual pulses;

[0060] Figure 8a shows a scanning electron microscope image of a crystalline silicon layer produced according to the disclosed method, and

[0061] Figure 8b Shown is a scanning electron microscope image of a crystalline silicon layer processed according to a comparative method. DETAILED DESCRIPTION

[0062] Figure 1 FIG2 schematically illustrates how, according to the disclosed method, a semiconductor material is irradiated with a laser beam to produce a uniformly crystalline layer. A layer 12 of the semiconductor material to be processed is applied to a carrier 10, such as a glass substrate. In the example shown here, the semiconductor material to be processed is amorphous silicon. The thickness of semiconductor material layer 12 is typically approximately 50 nm.

[0063] A linear laser beam 14 is imaged onto the semiconductor material and moved relative to the semiconductor material in a feed direction X, such that the laser beam 14 sweeps across and irradiates at least a portion of the semiconductor material layer 12. In the example shown here, the carrier 10 and the semiconductor material layer 12 are spatially movable relative to the fixed laser beam 14. The laser beam 14 can be moved relative to the semiconductor material layer 12 so that the entire semiconductor material layer 12 is irradiated by the laser beam 14. Typically, the laser beam 14 can be moved relative to the semiconductor material layer 12 so that a specific area is irradiated multiple times by the laser beam 14. Typical feed speeds range from 5 mm / s to 50 mm / s.

[0064] In the embodiment shown here, the propagation direction of the laser beam 14 is perpendicular to the surface of the semiconductor material layer 12 , that is, the laser beam 14 strikes the surface of the semiconductor material layer 12 perpendicularly at an incident angle of 0°.

[0065] Figures 2a to 2c The linear geometry of the laser beam 14 is shown. Figures 2a to 2c The intensities associated with specific directions are shown in FIG.

[0066] Figure 2a The intensity of the laser line along the major axis is shown, that is, the intensity distribution 16 integrated along the minor axis (along the x-axis). This integrated intensity distribution 16 is shown here along the major axis (along the y-axis). Conventionally, in figures, the minor axis extends parallel to the x-axis and the major axis extends parallel to the y-axis. As can be seen in this figure, distribution 16 is approximately rectangular, i.e., ideally uniform along the major axis. The length of the irradiation line in the y-direction can typically be between 100 mm and 1000 mm, for example, 100 mm, 250 mm, 750 mm, or 1000 mm, or greater than 1000 mm.

[0067] exist Figure 2b and Figure 2cIn FIG, the intensity of the laser line in the short axis direction is shown, that is, the intensity distribution 18, 20 integrated along the long axis (ie, along the y-axis), and the intensity distribution integrated in this way is displayed along the short axis (ie, along the x-axis). Figure 2b The intensity in has a Gaussian curve 18. Alternatively, the intensity can also be as follows Figure 2c The curve 20 shown is flat (“flat top”), ie, it is approximately rectangular.

[0068] Typical intensity widths in the x-direction are between 20 μm and 200 μm. Figure 2b The width of the Gaussian curve 18 in is given as the Full Width at Half Maximum (FWHM). Figure 2c For the flat curve 20 in FIG. 1 , the width is the width of the curve at an intensity equivalent to 90% of the maximum intensity (FW90%: Full Width at 90%).

[0069] If the irradiation line 14 is directed over a semiconductor material layer (eg, a-Si) 12 to be processed, the semiconductor material layer 12 is melted and solidified into a crystalline layer with improved electrical properties in a short period of time.

[0070] Figure 3a and Figure 3b The optical system 30 of an apparatus for processing semiconductor layers is schematically shown, with the aid of which a combination of Figure 1 2 and images the irradiation line 14 onto the semiconductor material.

[0071] The optical system 30 includes a beam shaping device 32, which will be described in more detail below. The beam shaping device 32 is configured to shape a laser beam so that its beam profile has a major axis and a minor axis; and an imaging device 34, which is arranged downstream of the beam shaping device 32 in the beam path of the laser beam and is configured to image the thus shaped laser beam into an illumination line 36. In the example shown here, four laser beams impinge on the beam shaping device 32: a first laser beam 38, a second laser beam 40, a third laser beam 42, and a fourth laser beam 44. However, according to the present disclosure, two laser beams can also impinge on the beam shaping device 32, as will be described below by way of example. In principle, the number of laser beams is not limited to four or two; rather, any other number of laser beams is possible and falls within the scope of the present disclosure.

[0072] In the example shown here, the laser radiation is laser radiation with a wavelength of 343 nm emitted by a plurality of UV solid-state lasers. However, in principle, other laser sources, in particular other solid-state laser sources, such as solid-state lasers emitting in the green spectral range, can also be used.

[0073] Here, as in Figure 1 As shown in Figure 2, Figure 3a and Figure 3b As also shown in FIG, the minor axis is parallel to the x-axis, the major axis is parallel to the y-axis, and the optical axis of the optical system extends parallel to the z-axis.

[0074] Figure 3a shows the imaging characteristics of the optical system 30 in the y direction, that is, the imaging characteristics along the long axis of the shaped laser beam and the irradiation line, Figure 3b The imaging characteristics of the optical system 30 in the x-direction, ie, along the short axis of the shaped laser beam and the irradiation line, are shown.

[0075] Figure 3a and Figure 3b The beam shaping device 32 of the optical system 30 in the embodiment has an anamorphic homogenizing optical element 46, which homogenizes the intensity of the incident laser beam in the y-axis direction. The anamorphic homogenizing optical element 46 comprises, for example, two cylindrical lens arrays arranged parallel to each other. These cylindrical lens arrays divide the incident light beam into individual sub-beams and superimpose them over the entire plane, thereby homogenizing the laser radiation to the greatest extent possible. In the case of multiple incident laser beams, each laser beam is divided into individual sub-beams and superimposed in a homogenized manner. Such homogenizing optical elements are described in detail, for example, in the prior art DE 42 20 705 A1, DE 3829728 A1 or DE 102 25 674 A1 cited in this application.

[0076] The beam shaping device 32 of the optical system 30 further includes a focusing cylindrical lens 48, which is arranged in the beam path downstream of the anamorphic homogenizing optical element 46. The focusing cylindrical lens 48 is configured to telecentrically deflect the laser beam, which has been divided and homogenized by the anamorphic homogenizing optical element 46, onto the illumination line 36 and to superimpose the beams there about the major axis, i.e., in the y-direction. The combination of the anamorphic homogenizing optical element 46 and the focusing cylindrical lens 48 thus ensures that the incident laser radiation is uniformly imaged into the illumination line 36 on the imaging plane.

[0077] An imaging device 34 is provided in the beam path downstream of the focusing cylindrical lens 48. The imaging device 34 is configured to focus the laser beam about its minor axis, i.e., in the x-direction, onto an illumination line 36. In other words, the imaging device 34 images the laser beam onto the illumination line 36, homogenizing only the minor axis of the beam profile, without homogenizing the already homogenized major axis of the beam profile. The imaging device 34 may, for example, be a focusing cylindrical lens.

[0078] The combination of the anamorphic homogenizing optics 46 and the focusing cylindrical lens 48 can be an anamorphic optics or part of such an optics. In particular, they can be part of an anamorphic optics, as described in the document DE 10 2012 007 601 A1 cited in the present application. Figures 4 to 6 As described with respect to the anamorphic optical device 42 in FIG.

[0079] Therefore, the beam shaping device 32 may also include one or more of the following optical elements:

[0080] a first collimating cylindrical lens, designated by reference numeral 54 in DE 10 2012 007 601 A1, for collimating the laser beam emitted about the x-axis,

[0081] a second collimating cylindrical lens, designated by reference numeral 56 in DE 10 2012 007 601 A1, for collimating the laser beam emitted about the y-axis,

[0082] a cylindrical lens arranged in the beam path after the first collimating cylindrical lens, which is designated by reference numeral 58 in DE 10 2012 007 601 A1 and serves to focus the laser beam with respect to the x-axis onto an intermediate image, which is designated by reference numeral 60 in DE 10 2012 007 601 A1,

[0083] an intermediate collimating cylindrical lens, which is provided in the beam path after the first collimating cylindrical lens 54 and is designated by reference numeral 58 ′ in DE 10 2012 007 601 A1, for collimating the laser beam of the first intermediate image, and / or

[0084] - a further cylindrical lens arranged in the beam path after the first intermediate image, in particular after the intermediate collimating cylindrical lens, which is marked with reference numeral 62 in DE 10 2012 007 601 A1, for focusing the laser beam with respect to the x-axis onto a second intermediate image, which is marked with reference numeral 64 in DE 10 2012 007 601 A1.

[0085] The above-described deformation homogenizing optical element 46 can be, for example, the one described in DE 10 2012 007 601 A1. Figures 4 to 6 Component 68 shown in, or including such a component.

[0086] The focusing cylindrical lens 48 described above may be, for example, the one described in DE 10 2012 007 601 A1. Figures 4 to 6 The focusing cylindrical lens 74 shown in FIG. 1 , or a component including such a component.

[0087] Finally, the imaging device 34 described above may be, for example, the imaging device described in DE 10 2012 007 601 A1. Figures 4 to 6 Component 66 shown in, or including such a component.

[0088] Furthermore, the optical system may include a polarization device 50 for each laser beam 38, 40, 42, 44 incident on the anamorphic optical element. The polarization device is an optical element for adjusting the polarization device 50, such as a λ / 2 plate in the beam path of each incident laser beam 38, 40, 42, 44. The optical element 50 is disposed in the beam path before the anamorphic optical element or the anamorphic homogenizing optical element 46. Each incident laser beam passes through the optical element 50, such that the laser beam 38, 40, 42, 44 passing through the optical element 50 is linearly polarized in a defined direction. Specifically, the laser beam emitted by the laser is already linearly polarized, such as in the example of a UV solid-state laser shown here, and the polarization orientation is rotated in a defined direction by the optical element 50. The optical element 50, such as a λ / 2 plate, is oriented relative to the polarization direction of the incident linearly polarized light so that two of the four laser beams are linearly polarized in the long axis direction after passing through the optical element 50, and the remaining two of the four laser beams are linearly polarized in the short axis direction after passing through the optical element 50. Here, the feed direction corresponds to the direction of the minor axis, so that the remaining two of the four laser beams are linearly polarized in the feed direction. More specifically, according to the present disclosure, an optical device 50 (e.g., a λ / 2 plate) disposed in the beam paths of first laser beam 38 and second laser beam 40 is oriented so that first laser beam 38 and second laser beam 40 are polarized in the feed direction, respectively, in the minor axis direction, i.e., the x-direction. Furthermore, an optical device 50 (e.g., a λ / 2 plate) disposed in the beam paths of third laser beam 42 and fourth laser beam 44 is oriented so that third laser beam 42 and fourth laser beam 44 are polarized in the major axis direction, i.e., the y-direction. In the example shown here, the laser is also pulsed, so that each pulse of the respective laser beam has the aforementioned polarization direction of the respective laser beam.

[0089] The four laser beams 38 , 40 , 42 , 44 may be laser radiation emitted by four laser sources, ie each laser beam corresponds to a separate laser source.

[0090] Alternatively, laser beams 38, 40, 42, and 44 can also be realized by splitting a laser beam emitted by a single laser source into a first sub-beam and a second sub-beam using a beam splitter. The beam splitter can be configured so that the first sub-beam produced by the splitting is a transmitted beam, and the second sub-beam is a transmitted beam, for example, approximately 50% each. For this purpose, polarization optical devices, such as so-called thin-film polarizers, can be used. A thin-film polarizer is an optical substrate with a special coating that allows light with p-polarization (the oscillation plane of the electric vector is parallel to the plane of the incident beam and the perpendicular line on the substrate surface) to pass through while reflecting light with s-polarization (the oscillation plane of the electric vector is perpendicular to the plane of the incident beam and the perpendicular line on the substrate surface). The polarization direction of the laser beam emitted by the laser source can be rotated using a λ / 2 plate placed before the thin-film polarizer in the beam path, so that the laser beam has equal components of p-polarization and s-polarization before the thin-film polarizer, achieving an approximately 50% split. However, it is also possible to rotate the λ / 2 plate before the thin-film polarizer so that the laser beam is polarized in different components, p- and s-polarized, before the thin-film polarizer, to achieve a split other than 50%. In principle, the relative intensity of the first and second sub-beams can be adjusted by the orientation of the λ / 2 plate.

[0091] Figure 4 Such an arrangement is schematically shown. A linearly polarized laser beam 52 emitted by a laser source strikes a λ / 2 plate 54, which is arranged in the beam path before a beam splitter 56, where a thin-film polarizer is arranged. The λ / 2 plate 54 is oriented so that the relative components of s- and p-polarization in the laser beam after the λ / 2 plate correspond to the desired relative intensities of the two sub-beams 58, 60 after the beam splitter 56, as described above. Thus, the first sub-beam 58 can be, for example, Figure 3a and 3b In the arrangement of the first laser beam 38, the second sub-beam 60 can be, for example, Figure 3a and 3b The third laser beam 42 is shown. The second sub-beam 60 is deflected by a reflective element 62 so that it runs parallel to the first sub-beam 58. In addition, a λ / 2 plate 64 is provided in the beam path after the beam splitter 56 in the beam path of the first and second sub-beams, which corresponds to Figure 3a and 3b λ / 2 plates 50 are shown in the beam paths of the first laser beam 38 and the third laser beam 42. Specifically, the λ / 2 plate 64 provided downstream in the beam path of the first sub-beam 58 is used for polarization in the short-axis direction. The λ / 2 plate 64 provided downstream in the beam path of the second sub-beam 60 is used for polarization in the long-axis direction.

[0092] The second laser beam 40 and the fourth laser beam 44 can be used in conjunction with Figure 4A further arrangement corresponding to the arrangement shown, with beam splitter 56, is provided by beam splitting. Thus, the downstream λ / 2 plates 64 are again oriented such that they polarize the third and fourth sub-beams in the direction of the minor axis and the major axis, respectively. Figure 3a and 3b The four laser beams in can alternatively be provided by two laser sources, the laser beams emitted by which can be divided into the first and second sub-beams 58, 60 or the third and fourth sub-beams, respectively, by beam splitting.

[0093] As described above, Figure 3a and 3b The optical system 30 in FIG. 3 can also be used to superimpose a number of laser beams other than four, for example two laser beams. Thus, corresponding to the four laser beams, the two laser beams can be provided by two laser sources, or by one laser source and then by means of the corresponding Figure 4 The laser beam emitted by the laser source is divided into a first sub-beam and a second sub-beam by beam splitting using the arrangement shown or an arrangement identical thereto.

[0094] You can also Figure 3a and 3b The optical system 30 in the embodiment is configured to polarize the pulses differently with a predetermined time interval Δt, offset in time from one another. This can be achieved by electronically delaying the trigger signal of the laser source for each laser beam when using separate laser sources. When the laser beams are provided by means of sub-beams, the time delay can be achieved by beam detours. Figure 4 As shown, the second sub-beam 60 has a path length Δs that is approximately backward compared to the first sub-beam 58. The path Δs can be selected so that the second sub-beam 60 has a time delay of a predetermined time interval Δt compared to the first sub-beam 58. Here, the predetermined time interval is preferably 10 ns to 20 ns.

[0095] Thus, with the aid of the aforementioned optical system 30, at least one laser pulse polarized in the minor axis direction and one laser pulse polarized in the major axis direction can be imaged in a uniformly superimposed manner on the irradiation line, wherein the pulse polarized in the major axis direction is delayed by a time interval Δt relative to the pulse polarized in the minor axis direction. When superimposing four laser beams, two laser beams are polarized in the minor axis direction and two laser beams are polarized in the major axis direction, wherein the laser beams polarized in the minor axis direction are synchronized in time, but the laser beam polarized in the major axis direction is delayed by the same time interval Δt relative to the laser beam polarized in the minor axis direction. By superimposing two (or more) temporally synchronized laser beams from two (or more) laser sources, possible fluctuations in energy density from pulse to pulse can be reduced. Such fluctuations can lead to different crystallization results and inhomogeneities in the crystal structure along the direction of displacement ("imaging mura") when the irradiation line shifts from pulse to pulse in the minor axis direction. It should be noted that the combined intensity distribution of two (or more) laser beams (laser pulses) superimposed in this manner can vary in time (from one combined intensity distribution to another) due to the existing temporal jitter (i.e., short-term deviations of the intensity from the ideal value), thus causing inhomogeneities in the crystal structure. In particular, these inhomogeneities arise from the fact that when combining two (or more) laser sources, each laser source has a temporal jitter that is independent of one another. Therefore, it is advantageous to use (pulsed) laser sources with as little temporal jitter as possible, in the nanosecond range.

[0096] Anamorphic homogenizing optical element 46 is designed to split each incident laser beam into a plurality of sub-beams and superimpose them in a homogenized manner in the long-axis direction. That is, each individual beam produces a uniform line. Therefore, in the aforementioned pulsing configuration with two laser beams, there are not only laser pulses polarized in the short-axis direction and temporally advanced, superimposed and imaged as a uniform line, but also laser pulses polarized in the long-axis direction and temporally delayed relative to the first pulse. Therefore, in a configuration with four laser beams, each laser pulse polarized in the short-axis direction and temporally advanced is superimposed and imaged as a uniform line, and each laser pulse polarized in the long-axis direction is temporally delayed relative to the first pulse.

[0097] This will be described again in detail with reference to the disclosed method.

[0098] Figure 5, the disclosed method is described with reference to an arrangement having four laser sources, namely a first laser source 66, a second laser source 68, a third laser source 70, and a fourth laser source 72. The first laser source 66 and the second laser source 68 are each configured to provide a first laser beam 74 having a first laser pulse 76 and a second laser beam 78 having a second laser pulse 80, wherein the first and second laser pulses 76, 80 are emitted simultaneously by a synchronized trigger signal 82 of the first and second laser sources 66, 68. The third and fourth laser sources 70, 72 are configured to provide a third laser beam 84 having a third laser pulse 86 and a fourth laser beam 88 having a fourth laser pulse 90, wherein the trigger signals 92 of the third and fourth laser sources 70, 72 are each electronically delayed by a time interval Δt, for example, by an electronic delay circuit 94, so that the third and fourth laser pulses 86, 90 are emitted and propagate in a time-delayed manner, respectively, with respect to the first pulse 76 and the second pulse 80. According to the disclosed method, the first laser pulse 76 and the second laser pulse 80 are also linearly polarized in the feed direction, i.e., in the x-direction, i.e., the polarization is oriented in the feed direction, and the third laser pulse 86 and the fourth laser pulse 90 are linearly polarized in the long axis direction, i.e., the polarization is oriented in the long axis direction, e.g., by means of a reference Figure 3a and 3b The λ / 2 plate 50. The first to fourth laser pulses 76, 80, 86, 90 typically have a full width at half maximum (FWHM) in the range of 15 ns to 20 ns. The time interval Δt is typically between 10 ns and 20 ns.

[0099] The four laser beams 74, 78, 84, 88 with four laser pulses 76, 80, 86, 90 are generated, for example, by means of Figure 3a and Figure 3b The beam shaping device 32 shown is shaped into a linear laser pulse with a short axis and a long axis. Figure 3a and Figure 3b The imaging device 34 shown images the laser pulse shaped in this way as an illumination line 36 on an imaging plane of the semiconductor material.

[0100] Figure 6 The disclosed method is described in the example of a setup with two laser sources. Here, the first laser source 66 corresponds to Figure 5 The first laser source in the second laser source 70 corresponds to Figure 5 The third laser source in the embodiment. Accordingly, the trigger signal 92 of the second laser source 70 is electronically delayed by a time interval Δt relative to the trigger signal 82 of the first laser source 66, so that the second laser pulse 86 propagates with a time interval Δt relative to the first laser pulse 76. In addition, for example, by means of a reference Figure 3a and 3bWith the λ / 2 plate 80 described, the first laser pulse 76 is linearly polarized in the direction of the minor axis of the subsequently formed laser pulse or linear irradiation line, and the second laser pulse 86 is linearly polarized perpendicular thereto in the direction of the major axis. The minor axis direction corresponds to the feed direction in which the subsequently formed irradiation line 36 moves relative to the semiconductor material 12 to be processed. Thus, similar to Figure 5 The method shown, two laser beams 74, 84 with two laser pulses 76, 86, for example by means of Figure 3a and Figure 3b The beam shaping device 32 shown is shaped into a linear laser pulse having a short axis and a long axis. Figure 3a and Figure 3b The imaging device 34 shown images the laser pulse shaped in this way as an illumination line 36 on an imaging plane of the semiconductor material 12 .

[0101] The linear geometry of the irradiation line 36 thus shaped has been Figures 2a to 2c Now with the help of Figure 7 The intensity profile of the combination of radiation lines shaped in this way is described as a function of time, ie the intensity profile of pulses superimposed and delayed in time relative to one another. Figure 7 For example, for Figure 6 The combined time intensity curve 96 of the disclosed method with two pulsed beams. Figure 7 Also shown is the combined intensity of the two laser beams 74, 84 as well as the pulse intensity of each individual laser beam as a function of time.

[0102] Here Figure 7 The intensity curve denoted by reference numeral 98 corresponds to the intensity curve of the first laser pulse 76 of the first laser beam 74, the intensity curve denoted by reference numeral 100 corresponds to the intensity curve of the second laser pulse 86 of the second laser beam 84, and the intensity curve denoted by reference numeral 96 corresponds to the combined temporal intensity curve of the first and second pulses 76, 86. The first and second laser pulses 76, 86 each have a temporal half-value width (FWHM) between 15 ns and 20 ns. As shown in Figure 7 As can also be seen in FIG, the second laser pulse 86 is delayed in time relative to the first laser pulse 76, with the delay time being between about 10ns and 20ns. Figure 7In the example, this delay time is approximately 20 ns. Consequently, the combined temporal intensity curve 96 shows a pulse profile with a first maximum value M1 and a second maximum value M2, and a pulse duration 102 that is extended relative to the individual pulse durations, with the total pulse duration being 40 ns to 50 ns. The total pulse duration 102 corresponds in turn to the temporal half-value width associated with the first maximum value ("half-value width of the first maximum value"), i.e., the pulse width at which the intensity of the first pulse is half that of the maximum value M1.

[0103] like Figure 7 As also shown in FIG, the maximum intensity M1 of the first laser pulse 76 is greater than the maximum intensity M2 of the time-delayed second laser pulse 86. Specifically, the intensity of the first laser pulse 76 is set relative to the intensity of the second laser pulse 86 so that the ratio M1 / M2 of the first maximum value M1 to the second maximum value M2 of the combined time intensity curve 96 is between 1 / 1.2 and 1 / 0.7, that is, between 0.8 and 1.4. In an embodiment in which each laser beam is provided by a separate laser source, this is achieved by tuning the intensities of the laser beams relative to each other. Figure 4 In the schematically shown arrangement in which two laser beams are provided by splitting the laser beam, the relative intensities can be varied by varying the s- and p-components of the laser beams before the thin-film polarizer 56 by correspondingly rotating the λ / 2 plate 54 .

[0104] As already detailed above, according to the present disclosure, the first laser pulse 76 is linearly polarized in the short-axis direction, ie, the feed direction, and the second laser pulse 86 is linearly polarized in the long-axis direction.

[0105] The present disclosure is based on the recognition that the linear polarization of the first and second pulses 76, 86 described above has a positive effect on the homogeneity of a semiconductor material layer 12 processed using a laser line. It has been found that polarizing the temporally first pulse in the feed direction and the temporally delayed pulse in the longitudinal direction allows for the production of a highly homogeneous crystalline silicon layer with a regular grain structure and a thickness of 50 nm to 60 nm. The ratio of the first maximum M1 to the second maximum M2 of the combined temporal intensity profile is approximately 1:1, i.e., between 0.8 and 1.4.

[0106] In contrast, for reverse polarization, ie, the first pulse is linearly polarized in the direction of the major axis and the delayed second pulse is linearly polarized in the direction of the minor axis, no such positive effect is observed.

[0107] The following describes this cognition with the help of experimental data:

[0108] The semiconductor material to be processed is a 50 nm thick layer of amorphous silicon on a glass substrate serving as a carrier. The optical setup used is a linear beam setup with four UV solid-state lasers emitting light at a wavelength of 343 nm. These lasers operate at a pulse repetition rate of 10 kHz. The pulse length of the emitted pulses, i.e., the half-value width, is between 15 ns and 20 ns. The energy of the laser pulses is up to 20 mJ. The energy density on the substrate, i.e., the silicon layer, is 220 mJ / cm 2 . Similar to Figure 5 The method illustrated herein synchronizes the first and second laser pulses from the first and second lasers of the four lasers in time using a synchronized trigger signal from the laser source, thereby emitting each first laser pulse simultaneously with the second laser pulse. The third and fourth laser pulses from the third and fourth lasers are each delayed by 10 to 20 ns relative to the first and second laser pulses. The intensities of the four laser beams are set so that the ratio (M1 / M2) of the first maximum to the second maximum of the combined temporal intensity variation curve is 1 / 1.

[0109] Utilize and Figure 3a and 3b The corresponding setup shown shapes the laser pulses of four laser beams into laser lines and images them onto amorphous silicon as irradiation lines. The irradiation lines are moved relative to the semiconductor layer, specifically in the direction of their minor axis, at a feed rate of 20 mm / s. The irradiation line length in the major axis direction is 90 mm, with a uniformity of 1.5% (2σ). The irradiation line length in the minor axis direction is 67 μm, with a uniformity of 3% (2σ). The total pulse length of the combined temporal intensity curve is 45 ns (full width at half maximum of the first maximum).

[0110] In a first experiment (Experiment a), pulses 76, 80 of the first and second laser beams 74, 78 were linearly polarized in the minor axis direction and pulses 86, 90 of the delayed third and fourth laser beams 84, 88 were polarized in the major axis direction.

[0111] In the second experiment (Experiment b), the pulses of the first and second laser beams were linearly polarized in the long-axis direction, and the pulses of the delayed third and fourth laser beams were polarized in the short-axis direction.

[0112] Figure 8a shows an image of a silicon surface taken with a scanning electron microscope according to the laser exposure of experiment a), Figure 8b The images of the silicon surface taken with a scanning electron microscope according to the laser exposure of experiment b) are shown. In both images, the feed direction is the x-axis direction (the short axis direction of the irradiation line), i.e., relative to Figure 8a and 8b In the vertical direction.

[0113] Figure 8a The results show a regular grain structure perpendicular to the feed direction, i.e., in the y-direction, along the long axis. In particular, the grains are arranged in approximately equidistant, vertically extending rows, i.e., with a spacing of approximately 0.35 μm corresponding to the wavelength of the UV laser. Alternatively, the grain structure exhibits a stripe pattern extending in the feed direction, with the stripes being equally spaced and thus exhibiting uniformity along the long axis. Consequently, the grain size exhibits a high degree of uniformity along the long axis (y-direction). However, the grain size exhibits a lower uniformity along the short axis (x-direction) than along the long axis.

[0114] On the contrary, Figure 8b It shows that there is no obvious uniformity in the short axis direction (x direction) and the long axis direction (y direction). Figure 8a Compared to the grain structure shown, the grain structure appears disordered in both grain orientation and size.

[0115] As described above, the laser crystallization process is based on the partial melting of the amorphous silicon layer and the subsequent solidification of the unmelted solid silicon on the glass substrate into a crystalline structure. The melting and solidification times range from 10 to 100 nanoseconds, with the film subsequently cooling to room temperature requiring several hundred microseconds. A pulse repetition rate of 10 kHz corresponds to a period of 100 microseconds. Because the pulse repetition rate, feed speed, and irradiation line width in the feed direction are determined so that a location on the semiconductor material is exposed multiple times during the exposure process—that is, irradiated with multiple successive pulses whose period, corresponding to the pulse repetition rate, is less than the time required to cool the film to room temperature—the semiconductor material is repeatedly exposed to UV light during the crystallization process. Furthermore, the relatively long laser pulse duration of tens of nanoseconds results in longer pulse exposures. These multiple exposures contribute to a more uniform grain structure.

[0116] As mentioned in the introduction, it is also known that polarization of the laser light, especially in combination with the aforementioned multiple exposures, has a positive effect on the regular grain structure of polycrystalline silicon. This is due to a surface interference effect ("Laser-Induced Periodic Pattern Structure," LIPSS), which results in a modulated intensity distribution. It has been shown that when light is linearly polarized in the direction of the major axis, a regular structure is formed along the major axis. Similarly, when light is linearly polarized in the direction of the minor axis (feed direction), this effect can be observed in the feed direction.

[0117] LIPPS is discussed in many publications, for example in the following references (1) to (4). It is assumed that the modulated intensity distribution is generated by the interaction of the incident light beam with the light beam diffracted on the surface and in the direction of the surface, and the resulting periodic distribution of the pulse energy density. The periodic pulse energy density distribution has the form of so-called "ripples", whose spacing is the absolute value of λ / (1±sinΘ) for a laser with a wavelength of λ and an angle of incidence of Θ. Therefore, for vertical light incidence (Θ=0°), a spacing of the order of the wavelength λ is obtained. The "ripples" extend in a direction perpendicular to the electric field vector, i.e., perpendicular to the polarization direction of the light beam or light pulse, and have a periodicity in the direction of the electric field vector. The pulse energy density is either minimum or maximum along the "ripples". The periodic pulse energy density results in a spatially periodic temperature distribution on the exposed semiconductor material layer, where the periodic temperature distribution is similar to the periodic pulse energy density distribution. For the periodic temperature distribution, thermal diffusion in the semiconductor material layer must also be taken into account. Furthermore, it is found that due to multiple reflections inside the semiconductor material layer, the periodic distribution of the pulse energy density varies with the thickness of the semiconductor material layer.

[0118] In summary, it can be determined that based on these observations, in order to obtain periodicity or regularity in the long axis direction, the electric field vector must be in the long axis direction, that is, the light or light pulse must be linearly polarized in the long axis direction.

[0119] Accordingly, it has been assumed hitherto that each pulse must contain at least one component polarized in the direction of the long axis in order to be able to produce a regular grain structure in the direction of the long axis.

[0120] According to the present disclosure, it has now been discovered that, with regard to the uniformity of the polysilicon grain structure, it is particularly advantageous to polarize the first pulse 76 or multiple first pulses 76, 80 in time in the short axis direction, i.e., the feed direction, and to polarize the delayed second pulse 86 or the delayed multiple delayed second pulses 86, 90 in the long axis direction.

[0121] A possible explanation for the promotion of a regular grain structure when the second pulse is polarized along the long axis is that the light along this long axis produces interferometric modulation (LIPPS) with the laser wavelength, thereby supporting the formation of structured grains in this direction. If the second pulse is linearly polarized in the feed direction, no interferometric modulation is formed along the long axis.

[0122] An explanation for why the polarization of the light in the direction of the short axis (feed direction) at the first maximum, i.e. the first pulse, is favorable for the grain structure could be that during this irradiation period the film is heated more uniformly along the long axis, as if a polarization component in the direction of the long axis already existed, since no interference modulation occurs there, and the first and second pulse components structuredly produce a partial liquid phase and promote the formation of the grain structure.

[0123] Furthermore, it was observed that when the polarization of the first pulse in the feed direction, i.e., the short axis direction, and the polarization of the delayed second pulse in the long axis direction were set according to experiment a), the size of the energy density processing window was enlarged to 20 to 25 mJ / cm2 compared to the exposure in which the first pulse and the delayed second pulse had the same polarization distribution in both directions. 2 (For energy density processing window is 210 to 230mJ / cm 2 For this case, the observed energy density processing window is only about 10 mJ / cm 2 (For energy density processing window is 215 to 225mJ / cm 2 for example).

[0124] Finally, it can be determined that in e.g. Figure 4 The "optical delay" shown is compared to Figure 5 and Figure 6 The “electronic delay” shown can achieve advantageous results. In particular, it has been determined that inhomogeneities in the crystal structure in the direction of movement (“photomura”) can be reduced with “optical delay” compared to “electronic delay”.

[0125] As described above, when superimposing two (or more) time-synchronized laser beams from two (or more) different laser sources, fluctuations in the intensity distribution may result due to the temporal jitter between the laser sources.

[0126] When observing, for example, Figure 7The combined temporal intensity profile 96 for two laser pulses 76, 86 is shown, and when formed for four laser pulses, the intensity profile 98 of the first laser pulse 76 is composed of the combined intensity profile 98 of the two first laser pulses 76, 80, and the intensity profile 100 of the second laser pulse 86 is composed of the combined intensity profile 100 of the two second laser pulses 86, 90. By using a pulsed laser source with minimal temporal jitter (e.g., in the nanosecond range), the variation from combined intensity profile 98 to combined intensity profile 98 and from combined intensity profile 100 to combined intensity profile 100 can be minimized. Since the two first laser pulses 76, 80 and the two second laser pulses 86, 90 originate from different laser sources, both in an arrangement with an optical delay and in an arrangement with an electronic delay, this "smearing" due to the existing temporal jitter does not result in different behaviors for electronic and optical delays.

[0127] Thus, by superimposing (combining) the intensity curves 98 and 100, it is possible to obtain Figure 7 The combined time intensity variation curve 96 is shown in FIG. Figure 7 As shown, the pulse width and the delay are selected such that the preceding pulse 98 or the preceding pulse combination 98 is superimposed on the following pulse 100 or the following pulse combination 100. This results in a first maximum M1 and a second maximum M2, wherein the first maximum M1 is practically unaffected by the delayed pulse 100 or the delayed pulse combination 100 (because the delayed pulse 100 or the delayed pulse combination 100 contributes little or nothing to the maximum M1 of the combined intensity profile 96), and the position and behavior of the second maximum M2 are clearly influenced by the delayed pulse 100 or the delayed pulse combination 100.

[0128] The varying temporal jitter between first pulse 98 and time-delayed second pulse 100 is therefore reflected in the position and behavior of the variation in second maximum M2, in particular in the position and behavior from combined intensity profile 96 to maximum M2 of combined intensity profile 96. The crystallization process reacts to these intensity variations in the combined temporal course with deviations in the grain structure, for example, in the grain size.

[0129] In contrast to electronic delay, optical delay is achieved by beam splitting and setting an optical delay path (for example, in combination with a pulsed laser source having minimal temporal jitter), so that the preceding pulse 98 and the delayed pulse 100 have virtually the same temporal jitter. This minimizes fluctuations in intensity variations when pulses 98, 100 are superimposed or when pulses 98, 100 are combined, and a uniform conversion of an amorphous semiconductor layer into a polycrystalline semiconductor layer is achieved over a large area.

[0130] In order to achieve a typical time delay of 20 ns for the second pulse 100 relative to the first pulse 98, an additional optical path of approximately 6 meters is required. This additional optical path can be realized, for example, by means of a long-focus spherical telescope, which is arranged in the delay path (for example at Figure 4 In this way, the laser beam ( Figure 4 The sub-beams 60 in the image are imaged in a controlled manner over a large path length. For 1:1 imaging, a telescope can be provided, for example, in which the focal length of the objective lens is the same as the focal length of the eyepiece. Furthermore, the delay path can be designed so that the delay path (its length) is variable, for example, by means of movably arranged deflection elements such as mirrors. This allows the pulse length to be set optimally for the crystallization process.

[0131] Instead of obtaining time-delayed laser beams or laser pulses through beam splitting and subsequent optical delay, a UV laser source can also be used according to the present disclosure, wherein two UV laser beams are obtained from an IR source by generating a third harmonic wavelength (343 nm) from an IR laser beam (1030 nm). Such a UV laser source uses the unconverted IR pulse energy (typically 50%) in a first SHG / THG crystal to generate a second UV laser beam in a second SHG (Second Harmonic Generation) / THG (Third Harmonic Generation) crystal. The two UV beams have no varying temporal jitter relative to one another. One of the two UV beams is then optically delayed, for example, by means of a long-focus spherical telescope as described above. Therefore, beam splitting is not necessary in this solution.

[0132] Furthermore, it is shown that in delayed pulse 100, the smaller the angular distribution of the laser light along the long axis, the more successful the formation of a periodic structure along the long axis. This is believed to be due to the correlation (λ / (1±sinθ), where θ is the angle of incidence to the surface normal) explained above regarding the LIPPS effect, which results from the interference generated along the surface. The smaller the angular distribution of the incident light, the less variation there is, i.e., the sharper the image. The light beam in the delayed pulse is linearly polarized in the direction of the long axis. Therefore, it is preferred to set the delay for the laser beam imaged close to the optical axis. This means that for the light beam imaged away from the optical axis, the light beam of the preceding pulse, polarized perpendicular to the long axis, is predominantly polarized.

[0133] References (1) to (4):

[0134] (1) P. van der Wilt, “Excimer-LASER Annealing: Microstructure Evolution and a Novel Characterization Technique,” ​​SID 2014 Digest p194

[0135] (2) S.Horita, H.Kaki, K.Nishioka, “Surface modification of an amorphousSi thin film crystallized by a linear polarized Nd:YAG pulse laser beam”, Journal of Applied Physics 102, 013501 (2007)

[0136] (3) HM van Driel, JE Sipe, and JF Young, “Laser-Induced Periodic Surface Structure on Solids: A Universal Phenomenon”, Phys. Rev. Lett. 49, 1955-1958 (1982) and S.E. Lark and D.C. Emmony, “Ultraviolet laser-induced periodic surface structures”, Phys. Rev. B 40, 2031-2041 (1989)

[0137] (4) JF Young, JS Preston, HM van Driel, and JS Pipe, “Laser-induced periodic surface structure. II. Experiments on Ge, Si, Al, and brass”, Phys. Rev. B 27, 1155-1172 (1983).

Claims

1. A method for growing a crystalline semiconductor layer, comprising the following steps: - providing a first laser beam (74) having first laser pulses (76) and a second laser beam (84) having second laser pulses (86), - shaping the first laser pulse (76) and the second laser pulse (86) into linear laser pulses having a short axis and a long axis by means of a beam shaping device (32), - imaging the thus shaped linear laser pulse onto the semiconductor material layer (12) by means of an imaging device (34) as an illumination line (36) having a short axis and a long axis, in, The method further comprises the following steps: - aligning the polarization direction of the first laser pulse (76) in the direction of the short axis of the irradiation line (36) by means of a first polarization device, - aligning the polarization direction of the second laser pulse (86) in the direction of the long axis of the illumination line (36) by means of a second polarization device, and - delaying the second laser pulse (86) in time relative to the first laser pulse (76) by a predetermined time interval Δt, the predetermined time interval Δt being selected such that the irradiation line (36) imaged onto the layer of semiconductor material (12) has a pulse-shaped combined time intensity variation curve (96) having a first maximum (M1) and a second maximum (M2), wherein the first laser pulse (76) and the second laser pulse (86) each have a temporal half-value width between 15 ns and 20 ns, and wherein the predetermined time interval Δt is between 10 ns and 20 ns.

2. The method according to claim 1, comprising: - moving the irradiation line (36) in a feed direction relative to the layer of semiconductor material (12), wherein The first laser pulses (76) are linearly polarized in the feed direction.

3. The method according to claim 1 or 2, wherein: The relative intensities of the first laser pulse (76) and the second laser pulse (86) are selected such that a ratio of a first maximum (M1) to a second maximum (M2) of the combined temporal intensity profile (96) lies in the range of 0.8 to 1.

4.

4. The method according to claim 3, wherein: A ratio of a first maximum value (M1) to a second maximum value (M2) of the combined time intensity curve (96) is in the range of 0.9 to 1.

2.

5. The method according to claim 4, wherein The ratio of the first maximum value (M1) to the second maximum value (M2) of the combined time intensity variation curve (96) is 1.

0.

6. The method according to claim 1, wherein The combined temporal intensity profile (96) of the radiation line (36) has a temporal half-value width (102) associated with a first maximum of the combined temporal intensity profile, which lies between 40 ns and 50 ns.

7. The method according to claim 1, comprising: A first laser (66) and a second laser (70) are provided, which are configured to emit the first laser beam (74) and the second laser beam (84), respectively, and are controlled so that the second laser pulse (86) is emitted with a delay of the time interval Δt relative to the first laser pulse (76).

8. The method according to claim 1, comprising: - providing a first laser designed to provide a laser beam (52) having pulses, - dividing the laser beam (52) into a first laser beam component (58) and a second laser beam component (60), wherein The first laser beam component (58) forms the first laser beam (74) having the first laser pulse (76), and the second laser beam component (60) forms the second laser beam (84) having the second laser pulse (86).

9. The method according to claim 8, wherein The optical path length of the second laser beam component (60) from the splitting point to the semiconductor material layer (12) is greater than the optical path length of the first laser beam component (58) from the splitting point to the semiconductor material layer (12).

10. The method according to claim 1, wherein The first laser pulse (76) is one of a plurality of first laser pulses of the first laser beam (74), the second laser pulse (86) is one of a plurality of second laser pulses of the second laser beam (84), and wherein each of the plurality of laser pulses of the second laser beam (84) is delayed in time by the predetermined time interval Δt relative to another of the plurality of laser pulses of the first laser beam (74).

11. The method according to claim 10, wherein: The feed speed, the pulse repetition rate of the first laser beam (74) and the second laser beam (84), and the geometric half-value width of the irradiation line (36) in the short axis direction are selected so that the position of the semiconductor material layer (12) is exposed multiple times by the irradiation line (36).

12. The method according to claim 1, comprising: - providing a third laser beam (78) having a third laser pulse (80) and a fourth laser beam (88) having a fourth laser pulse (90), - shaping the first laser pulse (76), the second laser pulse (86), the third laser pulse (80) and the fourth laser pulse (90) into linear laser pulses having a short axis and a long axis by means of a beam shaping device (32); - imaging the thus shaped linear laser pulse onto the semiconductor material layer (12) by means of the imaging device (34) as an illumination line (36) having a short axis and a long axis, in, The method further comprises the following steps: - aligning the polarization direction of the third laser pulse (80) in the direction of the short axis of the irradiation line (36) by means of a first polarization device, - aligning the polarization direction of the fourth laser pulse (90) in the direction of the long axis of the illumination line (36) by means of a second polarization device, and - delaying the fourth laser pulse (90) in time relative to the third laser pulse (80) by a predetermined time interval Δt, the predetermined time interval Δt being selected such that the irradiation line (36) imaged onto the semiconductor material layer (12) has a pulse-shaped combined time intensity profile (96) having a first maximum (M1) and a second maximum (M2).

13. An optical system (30) for growing a crystalline semiconductor layer, comprising: a beam shaping device (32) configured to shape a first laser pulse (76) of the first laser beam (74, 38) and a second laser pulse (86) of the second laser beam (84, 40) into linear laser pulses having a short axis and a long axis, - an imaging device (34) arranged to image the thus shaped linear laser pulse onto the semiconductor material layer (12) as an illumination line (36) having a short axis and a long axis, in, The optical system (30) further comprises: - polarization means comprising first polarization means and second polarization means, wherein the first polarization means is configured and arranged to align the polarization direction of the first laser pulse (76) in the direction of the short axis of the illumination line (36), and the second polarization means is configured and arranged to align the polarization direction of the second laser pulse (86) in the direction of the long axis of the illumination line (36), and - a delay device, arranged to delay the second laser pulse (86) relative to the first laser pulse (76) by a predetermined time interval Δt, the predetermined time interval Δt being selected such that the irradiation line (36) imaged onto the layer of semiconductor material (12) has a pulse-shaped combined time intensity variation curve (96), the combined time intensity variation curve (96) having a first maximum value (M1) and a second maximum value (M2), wherein the first laser pulse (76) and the second laser pulse (86) each have a time half-value width between 15 ns and 20 ns, and wherein the predetermined time interval Δt is between 10 ns and 20 ns.

14. The optical system (30) according to claim 13, wherein The first and second polarization means comprise: a first λ / 2 plate arranged in the beam path of the first laser beam (74, 38), disposed before the beam shaping device (32), and oriented relative to a first laser pulse (76) incident on the first λ / 2 plate such that the first laser pulse (76) is linearly polarized in the short axis direction after passing through the first λ / 2 plate, and A second λ / 2 plate is arranged in the beam path of the second laser beam (84, 40), is disposed before the beam shaping device (32), and is oriented relative to the second laser pulse (86) incident on the second λ / 2 plate so that the second laser pulse (86) is linearly polarized in the long axis direction after passing through the second λ / 2 plate.

15. The optical system (30) according to claim 13 or 14, wherein The delay device has a delay circuit for setting a trigger signal (92) of a second laser (70) for emitting a second laser beam (84, 40) having a second laser pulse (86) to be delayed by a time interval Δt relative to a trigger signal (82) of a first laser (66), wherein the first laser (66) is configured to emit a first laser beam (74) having a first laser pulse (76).

16. The optical system (30) according to claim 13 or 14, wherein The delay device has a beam detour (Δs) so that the optical path length of the second laser beam (84, 40) to the imaging plane of the semiconductor material layer (12) is greater than the optical path length of the first laser beam (74, 38) to the imaging plane of the semiconductor material layer (12).

17. The optical system (30) according to claim 13, wherein The beam shaping device (32) is configured to shape a first laser pulse (76) of the first laser beam (74), a second laser pulse (86) of the second laser beam (84), a third laser pulse (80) of the third laser beam (78), and a fourth laser pulse (90) of the fourth laser beam (88) into linear laser pulses having a short axis and a long axis, The imaging device (34) is configured to image the linear laser pulse shaped in this manner onto the semiconductor material layer (12) as an illumination line (36) having a short axis and a long axis. The polarization means comprises a third polarization means and a fourth polarization means, wherein the third polarization means is configured and arranged to align the polarization direction of the third laser pulse (80) in the short axis direction of the illumination line (36), and the fourth polarization means is configured and arranged to align the polarization direction of the fourth laser pulse (90) in the long axis direction of the illumination line (36), and The delay device is configured to delay the fourth laser pulse (90) relative to the third laser pulse (80) by a predetermined time interval Δt, wherein the predetermined time interval Δt is selected so that the irradiation line (36) imaged on the semiconductor material layer (12) has a pulse-shaped combined time intensity variation curve (96), which has a first maximum value (M1) and a second maximum value (M2).

18. An apparatus for growing a crystalline semiconductor layer, comprising The optical system (30) according to any one of claims 13 to 17, in, The device is configured to move the semiconductor material layer (12) in an advancement direction relative to the irradiation line (36), wherein the advancement direction corresponds to the direction of the minor axis of the irradiation line (36).

Images