Exposure apparatus, exposure method, and method for manufacturing articles

Abstract

To provide a technique advantageous for suppressing a line width change in a resist image due to a temperature change of an LED device, in an exposure apparatus having a light source including the LED device.SOLUTION: An exposure apparatus includes: an illumination optical system configured to illuminate an original plate with synthetic light composed of light from a first solid light emitting device having a first wavelength feature and light from a second solid light emitting device having a second wavelength feature; and a projection optical system configured to form a projection image on a substrate with the light from the original plate. The first wavelength feature and the second wavelength feature are determined so that a first line width change being a line width change of a pattern transferred onto the substrate caused by a change in the first wavelength feature and a second line width change being a line width change of the pattern caused by a change in the second wavelength feature are in different codes.SELECTED DRAWING: Figure 4

Description

Technical Field

[0001] The present invention relates to an exposure apparatus, an exposure method, and an article manufacturing method.

Background Art

[0002] An exposure apparatus is a device that transfers a pattern formed on a mask (original plate) to a plate (substrate). It illuminates the mask, which is the irradiated surface, with light through an illumination optical system, and projects an image of the pattern of the mask onto the plate through a projection optical system. As a light source of the exposure apparatus, generally, one having a spectrum in the ultraviolet region is used. An example thereof is a mercury lamp, which emits illumination light in the ultraviolet region including g-line (435 nm), h-line (405 nm), i-line (365 nm), etc. as emission lines.

[0003] In recent years, as an alternative to the mercury lamp, replacement with a light-emitting diode (LED: Light Emitting Diode), which is a solid-state light-emitting element, is expected. The LED element has a merit of energy saving and long life because the time from when a current is passed through the substrate circuit that controls light emission until the light output becomes stable is short, and it is not necessary to emit light constantly like a mercury lamp.

[0004] Since the light emission amount per LED element is small compared to a mercury lamp, a single LED element cannot provide sufficient productivity as a light source of an exposure apparatus. Therefore, it has been considered to use an LED array in which a plurality of LED elements are arranged as a light source. However, even when the LED elements are arrayed, in order to further improve productivity, it is required to increase the power input to the LEDs and enhance the light output.

[0005] The higher the integration degree of the LED array and the larger the input power, the greater the total amount of heat generated from the LED array light source. A cooler is provided to suppress the temperature rise of the LED elements due to heat, but depending on the design, there may be cases where the temperature change of the LED elements cannot be completely suppressed.

[0006] As the temperature of the LED element rises, a phenomenon occurs where the center wavelength (or peak wavelength) of the emission wavelength characteristics shifts to the longer wavelength side. This wavelength shift amount can be about 1 nm for a temperature change of 20 to 30 degrees in the case of a UV-LED element.

[0007] For example, it is known that changes in wavelength characteristics also occur in a laser light source. In the case of an exposure apparatus or an image recording apparatus using a laser light source, this change in wavelength characteristics affects the variation in the line width of the generated pattern and the unevenness of the recorded image. According to the inventors' study, it has been clarified that even in the case of an exposure apparatus using an LED as a light source, non-uniformity in the line width of the resist image occurs due to changes in wavelength characteristics.

[0008] Patent Document 1 describes a technique for suppressing sensitivity unevenness that occurs due to non-uniformity of a resist film and changes in the wavelength of laser light in an image recording apparatus that records an image on a recording medium using a laser light source. The document discloses that by combining a plurality of wavelengths over the variation period of the light transmittance with respect to the wavelength and performing exposure in order to average the variation in the light transmittance of the resist film with respect to the wavelength, sensitivity unevenness is reduced.

Prior Art Documents

Patent Documents

[0009]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0010] However, when suppressing line width errors and image unevenness in patterns using the conventional techniques described above, it is necessary to distribute wavelengths widely over the period in which line width errors and sensitivity fluctuate. In that case, multiple light sources with different emission wavelengths are required, resulting in a complex configuration. Furthermore, when LED elements with such closely spaced wavelength characteristics are used as light sources, the beam combining efficiency by beam combiners such as dichroic mirrors is significantly reduced, leading to problems in achieving sufficient productivity as an exposure device.

[0011] The present invention provides an advantageous technique for suppressing linewidth changes in the resist image caused by temperature changes of the LED element in an exposure apparatus equipped with a light source including an LED element. [Means for solving the problem]

[0012] According to one aspect of the present invention, the present invention comprises an illumination optical system that illuminates a master plate with composite light of light from a first solid-state light-emitting element having a first wavelength characteristic and light from a second solid-state light-emitting element having a second wavelength characteristic, and a projection optical system that forms a projected image on a substrate with light from the master plate. Based on the relationship between the film thickness of the resist coated on the substrate, the exposure wavelength, and the line width change of the pattern transferred onto the substrate, The first linewidth change, which is a change in the linewidth of the pattern transferred onto the substrate caused by the change in the first wavelength characteristic, and the second linewidth change, which is a change in the linewidth of the pattern caused by the change in the second wavelength characteristic, have opposite signs. Furthermore, the first exposure amount, which is the amount of light exposure from the first solid-state light-emitting element, and the second exposure amount, which is the amount of light exposure from the second solid-state light-emitting element, are set such that the sum of the first line width change and the second line width change is smaller than the allowable value. An exposure apparatus characterized by the above is provided. [Effects of the Invention]

[0013] According to the present invention, in an exposure apparatus equipped with a light source including an LED element, it is possible to provide an advantageous technique for suppressing linewidth changes in the resist image caused by temperature changes of the LED element. [Brief explanation of the drawing]

[0014] [Figure 1] A diagram showing the schematic configuration of an exposure apparatus. [Figure 2] A diagram showing the configuration of a typical illumination optical system. [Figure 3]A diagram showing the configuration of an LED array light source. [Figure 4] A diagram showing the configuration of the illumination optical system in the embodiment. [Figure 5] A diagram showing examples of the wavelength characteristics of an LED element and the transmittance characteristics of a dichroic mirror. [Figure 6] A diagram showing how the wavelength characteristics of an LED element change with temperature. [Figure 7] This figure shows the simulation results of the resist linewidth when the exposure wavelength is changed. [Figure 8] This figure shows the calculated linewidth change when a wavelength change occurs from the ground state. [Figure 9] A schematic diagram showing the relationship between the change in linewidth and the change in wavelength. [Figure 10] A figure showing the simulation results of Example 1. [Figure 11] A figure showing the simulation results for Example 2. [Figure 12] A figure showing the simulation results for Example 3. [Figure 13] Flowchart of the exposure method. [Figure 14] A flowchart of the process for determining exposure conditions. [Figure 15] A flowchart of the process for determining the exposure amount. [Figure 16] A diagram showing the simulation conditions. [Modes for carrying out the invention]

[0015] The embodiments will be described in detail below with reference to the attached drawings. Note that the following embodiments do not limit the invention as defined in the claims. While the embodiments describe multiple features, not all of these features are essential to the invention, and the features may be combined in any way. Furthermore, in the attached drawings, identical or similar configurations are given the same reference numerals, and redundant descriptions are omitted.

[0016] <First Embodiment> Figure 1 shows a schematic configuration of the exposure apparatus 100 in an embodiment. The exposure apparatus 100 is a lithography apparatus that illuminates a mask 2 (master plate) with light containing multiple wavelength ranges and transfers the pattern of the mask 2 onto a plate 8 (substrate). The exposure apparatus 100 is an apparatus for manufacturing flat panel displays, semiconductor devices, MEMS (Micro Electro Mechanical Systems), etc.

[0017] The projection optical system PO is, for example, a reflective optical system and includes planar mirrors 4 and 7, a concave mirror 5, and a convex mirror 6. The projection optical system PO reflects light from the mask 2 in the order of planar mirror 4, concave mirror 5, convex mirror 6, concave mirror 5, and planar mirror 7, forming a projected image of the mask 2 on the plate 8. When the projection optical system PO is composed of a reflective optical system, the chromatic aberration of the light from the light source is smaller than that of a refractive optical system. Such a configuration is suitable when using broadband light (broadband illumination light) that includes multiple wavelength ranges. However, the present invention is not limited to a mirror optical system for the projection optical system PO, and the projection optical system PO may be a refractive optical system consisting of refractive optical elements.

[0018] A photosensitive resist (photosensitive agent) made of photosensitive material is applied to plate 8. After the substrate coated with the resist is exposed, processing such as development and baking is performed to form a resist image on the resist corresponding to the pattern drawn on mask 2. Typically, the same mask pattern is exposed to multiple locations on a single plate.

[0019] The control unit CNT comprehensively controls each part of the exposure apparatus 100, namely the illumination optical system 1, the projection optical system PO, the mask stage 3, the plate stage 9, etc., to operate the exposure apparatus 100. The control unit CNT is composed of, for example, a PLD (Programmable Logic Device) such as an FPGA (Field Programmable Gate Array), or an ASIC (Application Specific Integrated Circuit), or a general-purpose or dedicated computer with a program built in, or a combination of all or part of these.

[0020] Figure 2 shows an illumination optical system 1' with a typical configuration. The illumination optical system 1' includes, for example, condenser lenses 12 and 15, an optical integrator 13, and an aperture diaphragm 14, which shape the light emitted from the light source 11 into illumination light and irradiate the mask 2.

[0021] The light source 11 is, for example, an LED array light source containing multiple solid-state light-emitting elements (LEDs), and its detailed configuration will be described later. The light source 11 may be understood as a light source unit as a component of the exposure apparatus, or as a light source unit located outside the exposure apparatus. In either configuration, the control unit CNT is configured to control the light source 11. The light emitted from the light source 11 is focused by the condenser lens 12 and irradiates the optical integrator 13. In Figure 2, the emission surface of the light source 11 is shown to be located near the front focal position of the condenser lens 12, and the incident surface of the optical integrator 13 is located near the rear focal position of the condenser lens 12, but the configuration is not necessarily limited to this.

[0022] The optical integrator 13 is, for example, a fly-eye lens and is composed of a large number of identical lens elements. The optical integrator 13 wavefront-splits the light on the incident surface and forms a multiple light source image of the light source 11 at its exit surface. In other words, an optical image of the light source 11 is formed on the exit surface of each of the many lens elements that make up the optical integrator 13, and this becomes a secondary light source. The aperture diaphragm 14, which is positioned near the exit surface of the optical integrator 13, controls the angular distribution of the light that illuminates the mask 2. The aperture diaphragm 14 can form deformed illumination such as annular or quadrupole by, for example, providing an annular-shaped or quadrupole-shaped transmission region. The light that has passed through the aperture diaphragm 14 is focused by the condenser lens 15 and illuminates the mask 2 as illumination light.

[0023] Here, the angular distribution of light illuminating mask 2 is called the effective light source distribution. In Figure 2, the effective light source distribution corresponds to the light intensity distribution after passing through the aperture diaphragm 14. Modified illumination, such as annular illumination and quadrupole illumination, is widely used as a super-resolution technique to improve resolution and depth of field in lithography processes with small k1 factor values. Modified illumination is characterized by a coherence factor σ value, which normalizes the angle illuminating the mask with the numerical aperture (NA) of the projection optical system. For example, in annular illumination, it may be described by a σ value corresponding to the minimum angle illuminating the mask (e.g., σ0.45) and a σ value corresponding to the maximum angle illuminating the mask (e.g., σ0.90), such as σ0.45-0.90.

[0024] The detailed configuration of the light source 11 will be described with reference to Figure 3. Figure 3(a) is a schematic cross-sectional view of the light source as seen from the x direction, and Figure 3(b) is a schematic view of the light source as seen from the z direction. The light source 11 includes a plurality of LED elements 22 mounted on a base 21. Since the radiant energy per LED element is smaller than that of a high-pressure mercury lamp, it is necessary to use a plurality of LED elements. For example, the plurality of LED elements 22 may be about 1000 LED elements. The plurality of LED elements 22 are arranged two-dimensionally in the XY direction on the base 21. In Figure 3(b), the plurality of LED elements 22 are arranged in a square grid, but are not limited to this, and may be arranged in a staggered pattern, for example.

[0025] Each of the multiple LED elements 22 is a UV-LED element that emits ultraviolet light, and its emission peak wavelengths are, for example, 365 nm, 385 nm, and 405 nm. The multiple LED elements 22 may have LEDs with different emission peak wavelengths mixed together on a single base. Alternatively, the multiple LED elements 22 may be classified into groups according to their emission peak wavelengths, and each group may be placed on a different base. However, as will be described later, in order to increase the productivity of the exposure process, groups of LED elements with different emission peak wavelengths may be configured as different array light sources, and the light from each light source may be combined using a beam combiner such as a dichroic mirror.

[0026] The light emitted from multiple LED elements 22 has an emission angle of about 60 to 70 degrees in half-angle, which is a very large angular distribution considering that the numerical aperture NA of a typical projection optical system for flat panel displays is around 0.1 (equivalent to about 5.7 degrees). Therefore, in order to capture the radiated light beam from multiple LED elements 22 without loss in the downstream optical system, a light-collimating unit 23 can be placed directly above each of the multiple LED elements 22 to collimate the radiated light beam. The light-collimating unit 23 is equipped with a collimating lens corresponding to the LED element. The intersections of the dotted lines in Figure 3(b) represent the optical axes of each collimating lens included in the light-collimating unit 23.

[0027] In UV-LED elements, only about 30-50% of the power input can be used as desired light, with the remainder being converted into heat. Therefore, a cooler 24 is provided to dissipate the heat generated from the LED element and suppress temperature changes of the LED element. The cooler 24 is, for example, a liquid-cooled heat sink. A refrigerant pump (not shown) circulates a refrigerant within the cooler, dissipating the heat transferred from the LED via the base 21 and cooling the LED element. Parameters related to cooling (cooling conditions) include the flow rate of the refrigerant and the temperature of the refrigerant. Increasing the flow rate of the refrigerant and lowering the temperature of the refrigerant improves cooling performance and allows for the cooling of more heat from the LED element. However, the temperature of the refrigerant is generally set to be close to the ambient temperature (room temperature) of the light source. Since the room temperature of a cleanroom for semiconductor and FPD manufacturing is generally around 22-24°C, temperatures within this temperature range are considered the ambient temperature or room temperature.

[0028] Referring to Figure 4, the configuration of the illumination optical system 1 in this embodiment will be described. The difference from the illumination optical system 1' in Figure 2 is that multiple LED array light sources are used. The other configurations are the same as those of the illumination optical system 1' in Figure 2. In Figure 4, the mask 2 is illuminated by combined light from the first light source 11a and the second light source 11b. The first light source 11a is a first LED array light source that includes multiple first LED elements (first solid-state light-emitting elements) having a first wavelength characteristic I1. The second light source 11b is a second LED array light source that includes multiple second LED elements (second solid-state light-emitting elements) having a second wavelength characteristic I2. The first LED elements having the first wavelength characteristic I1 are, for example, LED elements having a peak wavelength of 365 nm, and the second LED elements having the second wavelength characteristic I2 are, for example, LED elements having a peak wavelength of 405 nm.

[0029] In the configuration shown in Figure 4, light with different wavelength characteristics emitted from the first light source 11a and the second light source 11b is combined in the beam combiner 17 and guided to the condenser lens 12. The beam combiner 17 can be, for example, a dichroic mirror. In one example, this dichroic mirror may be a flat glass plate on which an optical thin film has been formed that transmits light around 365 nm and reflects light around 405 nm. With this configuration, a larger amount of light can be guided to the mask 2, which is the irradiated surface of the illumination optical system 1, compared to the illumination optical system 1' described in Figure 2.

[0030] Etendue is a physical quantity used when considering light capture in an optical system. Etendue is expressed as the product of the cross-sectional area of ​​light and the solid angle in one cross-section of the optical system. In the design of an exposure apparatus, etendue E in mask 2 is used. in This is determined as the device specification. Etendue E in (Assuming no loss of light) is invariant within the optical system, and E in Light with an etendue exceeding E cannot be captured by the optical system. In other words, if the etendue of the light source exceeds E S If we consider E S ≤E in The following relationship holds. Adding LED elements in the XY plane direction as shown in Figure 3(b) results in E S This corresponds to an increase in E in Anything exceeding this limit can only result in loss, and beyond this limit, it becomes impossible to improve the illuminance of the irradiated surface. However, in principle, etendue does not increase in wavelength synthesis. Therefore, by using wavelength synthesis, it becomes possible to deliver a larger amount of light to mask 2.

[0031] Fig. 5 shows an example of the spectrum of an LED element with a peak wavelength of 365 nm, the spectrum of an LED element with a peak wavelength of 405 nm, and the transmittance characteristics of a dichroic mirror. In order to minimize the loss of light quantity associated with wavelength synthesis and conduct light efficiently, it is desirable to separate the two spectra in the wavelength direction so that both spectra are surely separated into the passing region and the blocking region in the transmittance characteristics of the dichroic mirror. When the angle of incidence (AOI) of the light beam on the dichroic mirror changes, the transmittance characteristics also change in the wavelength direction as shown in Fig. 5. Therefore, when there is a spread in the angle of incidence distribution on the dichroic mirror, it is desirable to separate the spectra of the light sources by that much.

[0032] Here, let the wavelength width of the transition region indicating the steepness of the blocking characteristics in the transmittance characteristics of the dichroic mirror be Δw d and the fluctuation width of the cut-off wavelength of the transmittance characteristics due to the change in the angle of incidence of the light beam on the dichroic mirror be Δw AOI Also, assume that the center wavelength of the second wavelength characteristic is on the longer wavelength side than the center wavelength of the first wavelength characteristic. Among the wavelengths at which the intensity in the first wavelength characteristic becomes half of the peak intensity, let the wavelength on the longer wavelength side be λ 1d and among the wavelengths at which the intensity in the second wavelength characteristic becomes half of the peak intensity, let the wavelength on the shorter wavelength side be λ 2d Further, let the amount of change in the wavelength characteristics caused by the temperature change of the LED element described later be Δλ t According to the study by the inventors, the conditions for suppressing the loss of light quantity associated with wavelength synthesis are as follows. Δw d +Δw AOI +Δλ t ≦λ 2d -λ 1d ···(1)

[0033] By satisfying Equation (1), the loss of light quantity associated with wavelength synthesis can be suppressed to, for example, 10% or less.

[0034] In an example of the design, the wavelength width Δw of the transition region of the transmittance characteristics of the dichroic mirror dThe wavelength is 4 nm, and the angular distribution of light incident on the dichroic mirror is 45 degrees ± 3 degrees (spread of 6 degrees). A 1-degree change in the incident angle to the dichroic mirror alters the wavelength characteristics by approximately 1 nm. Therefore, the variation in the cutoff wavelength of the transmittance characteristics due to changes in the incident angle of light on the dichroic mirror is Δw. AOI It is 6 nm. Furthermore, as will be described later, the amount of change in wavelength characteristics Δλ caused by temperature changes. t The maximum value is approximately 5 nm. Therefore, for example, λ 2d -λ 1d By selecting first and second wavelength characteristics such that the difference is 15 nm or greater, it is possible to suppress the loss of light intensity associated with wavelength synthesis and increase the productivity of the exposure equipment. In the following, regarding the radiation characteristics of LED elements, it is assumed that equation (1) is satisfied if the peak wavelengths are separated by 30 nm or more.

[0035] In the case of a mercury lamp light source, it takes several tens of minutes for the light output to stabilize after the voltage is applied, so the mercury lamp is kept lit at all times. Then, to control the illumination of mask 2 with illumination light, an illumination light shutter configured in the illumination optical system is opened and closed to control the illumination of mask 2 only when exposure is performed. On the other hand, in the case of an LED element, the time from the application of voltage to the stabilization of the light output is extremely short, on the order of nanoseconds to microseconds, so it is not necessary to keep it lit at all times. Therefore, an illumination light shutter is not necessarily required, and the illumination of mask 2 with illumination light can be controlled by controlling the timing of the power supply to the LED element. That is, the illumination of mask 2 can be controlled by powering the LED element during exposure and not powering the LED element during non-exposure. By adopting this control method of powering only at the timing necessary for exposure, power consumption can be reduced. In addition, since the time the LED element is powered can be shortened, it is also advantageous from the standpoint of the lifespan of the LED element.

[0036] However, intermittently switching the power to the LED element causes the LED element's temperature to change during the exposure operation. Figure 6 shows an example of the change in wavelength characteristics (spectrum) due to temperature changes in the LED element. For two LED elements with peak wavelengths of 365 nm and 405 nm, respectively, at a temperature of 23°C, the emission spectrum at 23°C is shown by a solid line, and the emission spectrum at 83°C is shown by a dashed line. Although it depends on the element and emission wavelength, the peak wavelength can shift to a longer wavelength side by approximately 1 nm for every 20°C increase in the LED element's temperature. Therefore, if the LED element's temperature rises from 23°C to 83°C during the exposure operation, the peak wavelength will change by 3 nm during the exposure operation.

[0037] The maximum junction temperature rating for LED elements (UV-LED elements) is typically 120-130 degrees Celsius. Exceeding this temperature can lead to degradation of the LED element's lifespan or failure. Since the temperature can vary by approximately 100 degrees Celsius from room temperature, the peak wavelength may fluctuate by about 5 nm when used within the room temperature range or the rated temperature. Parameters affecting the LED element temperature include the cooling conditions of the cooler (refrigerant flow rate and temperature) and the output power of the LED element. To operate the LED element below its rated junction temperature, the cooling conditions and the power supplied to the LED element can be adjusted accordingly. While the integration density of LED elements in an LED array light source and the thermal conductivity of the substrate also affect the LED element temperature, these are not controllable parameters during exposure equipment operation and are therefore excluded here.

[0038] In the following, the state in which the LED element temperature is near room temperature (22-24°C) is referred to as the steady state. In contrast, the state in which the LED element temperature is higher than room temperature and the wavelength characteristics have changed from the steady state is referred to as the transient state. When the LED element is repeatedly switched on and off intermittently, it is possible that the temperature will fluctuate in the transient state without returning to the steady state. The problem in the operation of the exposure apparatus is the fluctuation between the minimum and maximum temperature of the LED element during the exposure process. The state in which the temperature is at its lowest value is called the ground state, and the temperature and wavelength characteristics at this time are called the ground temperature and ground wavelength characteristics, respectively. Furthermore, the state in which the LED element temperature is at its highest value is called the maximum transition state, and the temperature and wavelength characteristics at this time are called the maximum temperature and maximum transition wavelength characteristics, respectively. In addition, the state between the ground state and the maximum transition state is called the transition state. Unless otherwise specified, when "wavelength characteristics" is referred to below, it refers to the ground wavelength characteristics. The ground state may be the same as the steady state, but it may not be the same. Quantities representing transition states may be denoted with a dash (´), and quantities representing the difference between the transition state and the ground state shall be denoted with a delta (Δ). Furthermore, quantities representing the maximum transition state shall be denoted with the subscript max.

[0039] Here, "wavelength characteristics" refers to the intensity distribution of the emission wavelength of the light source and implies "spectrum" or "emission spectrum." While wavelength characteristics are sometimes explicitly expressed as a function of wavelength λ as I(λ), they are also sometimes simply represented as I. "Peak wavelength" refers to the wavelength at which the LED element obtains maximum intensity in its emission spectrum, expressed using the subscript p as λ. p It is expressed as follows. The "centroid wavelength" refers to the wavelength obtained by weighting the wavelength according to the intensity of the emission spectrum and averaging it, and the subscript g is used as λ. g This is expressed as follows. When the wavelength characteristics of an LED element change due to temperature changes, the distribution of the wavelength characteristics before and after the change can be considered to be substantially the same shape but shifted in the wavelength direction. In other words, the amount of change in the peak wavelength caused by the temperature change of the LED element and the amount of change in the centroid wavelength can be considered to be the same. Under this assumption, the difference between the base wavelength characteristic I(λ) and the transition wavelength characteristic I'(λ) is the base peak wavelength λ pand transition peak wavelength λ p It is characterized as the difference from '. That is, Δλ with respect to the basis wavelength characteristic I(λ). p =λ p ´-λ p Given the given conditions, the transition wavelength characteristic I'(λ) is assumed to be uniquely determined. This is also true when the peak wavelength is replaced with the centroid wavelength, so the degree of change in the wavelength characteristic will be expressed as Δλ (without a subscript).

[0040] Changes in the wavelength characteristics of the LED element due to temperature changes can lead to changes in the line width of the resist image (the pattern transferred onto the substrate) formed on the resist coated on plate 8 (substrate). According to the inventors' investigation, changes in line width can mainly occur due to the following three factors.

[0041] The first is the wavelength dependence of the contrast of the aerial image. As the wavelength increases, the diffraction angle increases, and the proportion of light that does not contribute to imaging increases, thus reducing the contrast of the aerial image. However, for wavelength changes of about 3 nm, for example, this has only a minor effect on the linewidth.

[0042] The second factor is the wavelength dependence of the sensitivity of the photosensitive material contained in the resist. Since sensitivity usually differs with different wavelengths, if the wavelength characteristics of the illumination light used for exposure change, the appropriate exposure amount to obtain a resist image with the desired line width will change. Conversely, if the wavelength characteristics of the illumination light change while exposure is being performed with a constant exposure amount, the line width of the resulting resist image will change.

[0043] The third reason is the standing wave effect within the resist. Depending on the film thickness and refractive index of the applied resist, and the wavelength of the illumination light, standing waves can form within the resist. If the wavelength characteristics of the illumination light change, the mode of the standing waves changes, which in turn affects the change in the resist linewidth.

[0044] Figure 7 shows the simulation results of the resist linewidth when the exposure wavelength is changed for resist film thicknesses of 1300 nm, 1400 nm, and 1500 nm. The simulation conditions are shown in Figure 16(a). The exposure wavelength was set to a single wavelength without any distribution. The mask pattern was a 1.5 μm line and space pattern, and the target linewidth was also 1.5 μm. The substrate was silicon, the numerical aperture (NA) of the projection optical system was 0.1, and the effective light source was an annular light source with σ 0.45-0.90. The exposure amount was set for each resist film thickness so that the target linewidth was achieved at 365 nm, and the exposure amount was kept constant even when the wavelength was changed. A periodic change in linewidth can be observed with respect to wavelength change, which is due to the standing wave effect described above. When the wavelength dependence of the sensitivity of the photosensitive material is not steep, the effect of standing waves is dominant as shown in Figure 7. Since the conditions for the formation of standing waves vary with film thickness, the repetition period of the increase and decrease in resist linewidth also differs when the resist film thickness is changed.

[0045] Based on the simulation results in Figure 7, Figure 8 shows the calculated linewidth change ΔCD when a wavelength change of Δλ = 3 nm occurs from the ground state. It can be seen that the sign and amount of the linewidth change when a wavelength change occurs differ depending on the exposure wavelength. For example, focusing on λ = 365 nm, ΔCD = -80 nm at a resist film thickness T = 1500 nm, while ΔCD = 14 nm at T = 1400 nm and ΔCD = 113 nm at T = 1300 nm, showing that it changes with film thickness. On the other hand, focusing on λ = 405 nm, ΔCD = 99 nm at a resist film thickness T = 1500 nm, while ΔCD = -38 nm at T = 1400 nm and ΔCD = -129 nm at T = 1300 nm. Thus, the linewidth change ΔCD changes depending on the exposure wavelength λ and the resist film thickness T.

[0046] Let ΔCD1 (first linewidth change) be the linewidth change caused by a change in the first wavelength characteristic, and ΔCD2 be the linewidth change caused by a change in the second wavelength characteristic. One feature of this embodiment is that the first and second wavelength characteristics are defined such that ΔCD1 (first linewidth change) and ΔCD2 (second linewidth change) have opposite signs. By performing exposure using the first and second wavelength characteristics such that ΔCD has opposite signs, even if a change in wavelength characteristics occurs, ΔCD cancels each other out, and the linewidth change associated with the change in wavelength characteristics can be suppressed. To explain using the example above, if a wavelength change of Δλ=3nm occurs at T=1300nm, then ΔCD=113nm at λ=365nm and ΔCD=-129nm at λ=405nm. By performing exposure by mixing these wavelengths, the linewidth changes associated with the wavelength changes at each wavelength cancel each other out, and as a result, the linewidth change is suppressed.

[0047] On the other hand, if a wavelength change of Δλ=3nm occurs at T=1300nm, then ΔCD=54nm at λ=395nm and ΔCD=113nm at λ=365nm, and ΔCD has the same sign. Thus, when the sign of the linewidth change associated with the wavelength change at each wavelength is the same, the linewidth change cannot be sufficiently suppressed when the images are mixed and exposed.

[0048] Figures 7 and 8 show simulation results for a single wavelength, whereas actual light sources have wavelength characteristics with a certain degree of chromatic spread. For example, UV-LEDs have a wavelength spread of about 10 nm in half width at half maximum. In reality, due to this chromatic spread, different standing wave modes coexist and are averaged along with the weights of the wavelength characteristics. Therefore, in actual light sources, the linewidth change is smaller than ΔCD in Figure 8, and the absolute values ​​do not match. However, if the horizontal axis of Figure 8 is viewed as the centroid wavelength in a certain wavelength characteristic, the trend of increase and decrease in linewidth change matches that of Figure 8, so it is possible to estimate the exposure wavelength using single-wavelength simulation results like those in Figure 8.

[0049] Furthermore, if the sole objective is to suppress linewidth fluctuations, it is thought that this can be achieved by using wavelengths that span more than one period of the ΔCD fluctuation shown in Figure 8. For example, in Figure 8, the increase and decrease in ΔCD around 360-390 nm has a period of 20-30 nm. Therefore, by using a mixture of LEDs with peak wavelengths of 365 nm, 375 nm, and 385 nm, a wavelength characteristic that extends over the 20-30 nm region can be obtained, and it is thought that linewidth changes in the resist image can be suppressed even when wavelength changes occur.

[0050] Thus, when distributing the exposure wavelength over the period of increase and decrease of ΔCD, it is necessary to use multiple LED elements with closely spaced peak wavelengths. However, when the peak wavelengths are close together, the efficiency of wavelength synthesis using dichroic mirrors decreases significantly. Therefore, multiple LED elements with different wavelength characteristics are mixed within a single LED array light source. When using only a single LED array light source as shown in Figure 2, this method can suppress linewidth changes due to changes in wavelength characteristics, but in this case, there is a disadvantage that sufficient light intensity cannot be obtained during exposure. To obtain sufficient light intensity for exposure, it is desirable to synthesize light from multiple LED array light sources using a beam combiner, as shown in the configuration in Figure 4, and for this to work, the wavelength characteristics of the LED elements must be spaced apart from each other.

[0051] Even if the linewidth change ΔCD1 (hereinafter also referred to as "first linewidth change") of the resist image caused by a change in the first wavelength characteristic and the linewidth change ΔCD2 (hereinafter also referred to as "second linewidth change") of the resist image caused by a change in the second wavelength characteristic have opposite signs, if there is a difference in their absolute amounts, the linewidth changes may not be completely canceled out, and a certain degree of linewidth change may occur. Therefore, in this embodiment, the exposure amount from the first LED element (first exposure amount) and the exposure amount from the second LED element (second exposure amount) are set so that the sum of the first linewidth change and the second linewidth change is smaller than the allowable value. Specifically, the exposure amount D1 (first exposure amount) from the light with the first wavelength characteristic and the exposure amount D2 (second exposure amount) from the light with the second wavelength characteristic should be set so that ΔCD1 + ΔCD2 is small enough to cancel out the linewidth changes. Here, ΔCD1 + ΔCD2 represents the sum of the first linewidth change and the second linewidth change. One way to make the exposure amount different for the first wavelength characteristics and the second wavelength characteristics is for the control unit CNT to control the power supplied to each LED element with different wavelength characteristics, thereby creating different light outputs. That is, by controlling the power supplied to the first LED element and the second LED element, exposure of the first exposure amount by the first LED element and exposure of the second exposure amount by the second LED element can be achieved. However, this method is not limited to this, and the exposure time (the time each LED element is energized) may be changed relatively by the light with different wavelength characteristics. Specifically, multiple exposure can be considered, in which exposure is performed for a first exposure time with only the LED element with the first wavelength characteristics energized, and then exposure is performed for a second exposure time with only the LED element with the second wavelength characteristics energized. In this case, even if the light output of the first and second wavelength characteristics is the same, the integrated exposure amount can be made different by making the first exposure time and the second exposure time different.

[0052] Figure 9 is a schematic diagram showing the relationship between the change in linewidth and the change in wavelength. Here, the exposure amount D1 (first exposure amount) of the first wavelength characteristic I1 and the exposure amount D2 (second exposure amount) of the second wavelength characteristic I2 are set to D1:D2 = 1:1. Figure 9 shows the relationship between the change in linewidth ΔCD when wavelength characteristics I1 and I2 undergo an independent wavelength change of Δλ in this case. In Figure 9, the solid line shows the difference from the target linewidth that occurs when the second wavelength characteristic I2 is in the ground state and the first wavelength characteristic I1 changes by Δλ. The dashed line shows the difference from the target linewidth that occurs when the first wavelength characteristic I1 is in the ground state and the second wavelength characteristic I2 changes by Δλ. If Δλ is in the range of approximately 0 to 3 nm, ΔCD changes approximately linearly.

[0053] Here, the wavelength sensitivity representing the linewidth change per unit wavelength change in the first wavelength characteristic is denoted as (ΔCD / Δλ)1, and the wavelength sensitivity representing the linewidth change per unit wavelength change in the second wavelength characteristic is denoted as (ΔCD / Δλ)2. These correspond to the slopes of the solid and dashed lines in Figure 9, respectively. Furthermore, the maximum value of the wavelength change during exposure in the first wavelength characteristic is denoted as Δλ 1max , the maximum value of the wavelength change during exposure in the second wavelength characteristic is Δλ 2max Let's assume these quantities represent the difference between the centroid wavelength (peak wavelength) in the maximum transition state and the centroid wavelength (peak wavelength) in the ground state. The linewidth changes caused by the wavelength change in the first wavelength characteristic are (ΔCD / Δλ)1 and Δλ. 1max It is estimated as the product of (ΔCD / Δλ)² and Δλ, and the linewidth change caused by the wavelength change in the second wavelength characteristic is (ΔCD / Δλ)² and Δλ 2max It is estimated as the product of the two. When the exposure ratio is equal, if there is a difference between the linewidth change caused by the wavelength change of the first wavelength characteristic and the linewidth change caused by the wavelength change of the second wavelength characteristic, the exposure ratio should be set according to these ratios. Specifically, |(ΔCD / Δλ)²×Δλ 2max | / |(ΔCD / Δλ)1×Δλ 1max | ···(2) The ratio of D1 to D2 should be set such that the difference between the characteristic value defined by and D1 / D2 falls within an acceptable range.

[0054] Δλ 1max and Δλ 2max This depends on the power input P1 to the first LED element, the power input P2 to the second LED element, and the cooling conditions of the cooler. The higher the power input, the greater the heat generated, the more easily the temperature of the LED element rises, and the larger the maximum value of the change in wavelength characteristics. Also, if the cooling capacity of the cooler is improved by increasing the flow rate of the refrigerant, the temperature of the LED element will decrease more easily toward the temperature of the refrigerant, and the maximum value of the change in wavelength characteristics will decrease.

[0055] Here, Δλ 1max and Δλ 2max Let's consider the case where the parameters are controlled to be the same. Even if the input powers P1 and P2 are made different to make the exposure amounts D1 and D2 different, and a difference in the amount of heat generated by the LED elements occurs, it is possible to control the amount of wavelength change equally by changing the cooling conditions. In this case, equation (2) is simply, |(ΔCD / Δλ)2 / (ΔCD / Δλ)1| ···(3) It can be written as follows. Therefore, when setting cooling conditions so that the maximum transition wavelengths between different wavelength characteristics are the same, the exposure amount should be set so that the difference between equation (3) and D1 / D2 is small.

[0056] (Example 1) Figure 10 shows the simulation results for Example 1. Example 1 has a base peak wavelength of 365 nm (λ 1p An LED element having a first wavelength characteristic of ) and a base peak wavelength of 405 nm (λ 2p An LED element having a second wavelength characteristic of λ is used (Figure 10(a)). The illumination light generated from these LED elements is mixed with an exposure ratio of 1:1. Comparative Example 1 has a base peak wavelength of 365 nm (λ). 1p An LED element having a first wavelength characteristic of ) and a base peak wavelength of 395 nm (λ 2pLED elements having a second wavelength characteristic are used (Figure 10(a)). The illumination light generated from these LED elements is mixed with an exposure ratio of 1:1. Other conditions are shown in Figure 16(b). For both Example 1 and Comparative Example 1, the exposure amount that results in a resist linewidth of 1.5 μm is calculated. The peak wavelengths of both the first and second wavelength characteristics are Δλ from the ground state. 1p =Δλ 2p The resist linewidth was evaluated at the same exposure dose as the ground state in the transition state wavelength characteristics, which changed by only 3 nm.

[0057] In the following, to clarify the preconditions for evaluating linewidth, we will express it as CD(Δλ1, Δλ2) using the change Δλ1 from the ground state of the first wavelength characteristic and the change Δλ2 from the ground state of the second wavelength characteristic. The linewidth in the ground state is expressed as CD(0, 0) since Δλ1 = Δλ2 = 0 nm. Also, for example, if the second wavelength characteristic is in the ground state (Δλ2 = 0) and the first wavelength characteristic is in the transition state (Δλ1 ≠ 0), it is expressed as CD(Δλ1, 0). Furthermore, when expressing the difference between the linewidth when some wavelength characteristic transition occurs and the linewidth in the ground state (CD(0, 0)), we will add Δ at the beginning.

[0058] As can be seen from Figure 10(b), in Comparative Example 1, a linewidth change of 50 nm occurs with the transition of the wavelength characteristic. This is a 3.3% difference from the target linewidth of 1.5 μm, which is a quantity that cannot be ignored in terms of linewidth control. On the other hand, in Example 1, only a linewidth change of 5 nm occurs with the transition of the wavelength characteristic. This is a 0.3% difference from the target linewidth of 1.5 μm, and it can be seen that it is suppressed to 1 / 10th compared to the Comparative Example.

[0059] Here, the second wavelength characteristic is defined as the wavelength characteristic in the ground state, and the linewidth under the condition that the first wavelength characteristic is the wavelength characteristic in the transition state where Δλ1 = 3 nm has changed from the ground state is defined as CD(Δλ1,0). Also, the linewidth under the condition that the first wavelength characteristic is defined as the wavelength characteristic in the ground state, and the second wavelength characteristic is the wavelength characteristic in the transition state where Δλ2 = 3 nm has changed from the ground state is defined as CD(0,Δλ2). These CD(Δλ1,0) and CD(0,Δλ2) are shown in Figure 10(b). In Example 1, CD(Δλ1,0) is 1540 nm, which is a positive change relative to the target linewidth, but CD(0,Δλ2) is 1469 nm, which is a negative change relative to the target linewidth. Thus, when the signs of the linewidth change in response to the change in wavelength characteristics are opposite for the first and second wavelength characteristics, even if these wavelength characteristics change simultaneously, they cancel each other out, and as a result, the change in linewidth can be suppressed.

[0060] On the other hand, in Comparative Example 1, CD(Δλ1,0) is 1539 nm, which is a positive change relative to the target linewidth, but CD(0,Δλ2) is 1510 nm, which is also a positive change relative to the target linewidth. Thus, when the sign of the linewidth change in response to the change in wavelength characteristics is the same for the first wavelength characteristic and the second wavelength characteristic, the change in linewidth is not suppressed and becomes larger when these wavelength characteristics change simultaneously. As a result, variations in linewidth occur within the same process, which can be problematic in terms of linewidth management.

[0061] Let's consider the case where the sign of the linewidth change ΔCD(Δλ1,0) associated with the change in the first wavelength characteristic and the sign of the linewidth change ΔCD(0,Δλ2) associated with the change in the second wavelength characteristic are the same. In this case, the error between the linewidth at the maximum transition wavelength characteristic and the linewidth at the base wavelength characteristic may be 3% or more relative to the target linewidth. On the other hand, let's consider the case where the signs of the linewidth change ΔCD(Δλ1,0) associated with the change in the first wavelength characteristic and the sign of the linewidth change ΔCD(0,Δλ2) associated with the change in the second wavelength characteristic are opposite. In this case, the error between the linewidth at the maximum transition wavelength characteristic and the linewidth at the base wavelength characteristic can be reduced to 3% or less relative to the target linewidth. Therefore, by applying the present invention, it is possible to suppress variations in linewidth within the same process.

[0062] (Example 2) Next, Example 2 will be described. Figure 11(a) shows the linewidth change ΔCD when the first wavelength characteristic I1(λ) and the second wavelength characteristic I2(λ) undergo an independent wavelength change of Δλ. Here, the first wavelength characteristic I1(λ) has a peak wavelength of 365 nm in the ground state, and its exposure amount is D1. The second wavelength characteristic I2(λ) has a peak wavelength of 405 nm in the ground state, and its exposure amount is D2. The ratio of exposure amounts is D1:D2=1:1. In Figure 11(a), the solid line shows the difference from the target linewidth that occurs when the second wavelength characteristic I2(λ) is in the ground state and the first wavelength characteristic I1(λ) changes by Δλ, and is expressed as ΔCD(Δλ1,0) according to the notation described above. Furthermore, the dashed line indicates the difference from the target linewidth that occurs when the first wavelength characteristic I1(λ) is in the ground state and the second wavelength characteristic I2(λ) changes by Δλ, and is expressed as ΔCD(0,Δλ2) according to the notation described above. The resist film thickness is 1425 nm, and the other conditions are as shown in Figure 16(b).

[0063] As shown in Figure 11(a), ΔCD changes approximately linearly in the range of Δλ from 0 to 5 nm. Here, let (ΔCD / Δλ)1 be the linewidth change per unit wavelength change in the first wavelength characteristic, and (ΔCD / Δλ)2 be the linewidth change per unit wavelength change in the second wavelength characteristic. Then, (ΔCD / Δλ)1 = 19.4 nm and (ΔCD / Δλ)2 = -9.1 nm. Also, the maximum value of the wavelength change during exposure in the first wavelength characteristic is Δλ. 1max , the maximum value of the wavelength change during exposure in the second wavelength characteristic Δλ 2max Let's assume both are 3 nm. In that case, the linewidth change caused by the wavelength change in the first wavelength characteristic is (ΔCD / Δλ)1 and Δλ 1max It is estimated as the product of (ΔCD / Δλ)² and is 58.2 nm. Similarly, the linewidth change caused by the wavelength change in the second wavelength characteristic is (ΔCD / Δλ)² and Δλ 2max It is estimated as the product of and is -27.3 nm. Since the characteristic value of equation (2) in this embodiment is 0.47, line width changes can be suppressed by setting the exposure ratio so that D1 / D2 is close to 0.47.

[0064] Figure 11(b) shows the simulation results for the relationship between the exposure ratio D1 / D2 and the linewidth change ΔCD. The calculation conditions are the same as those used for calculating the linewidth change in Figure 11(a). For the total exposure D1+D2, the ground state resist linewidth was set to 1500 nm under each condition.

[0065] When the exposure dose for the first and second wavelength characteristics is the same, i.e., D1 / D2 = 1, the change in wavelength characteristics results in ΔCD = 31 nm, causing a linewidth error of approximately 2% relative to the target linewidth. However, ΔCD changes by changing the exposure dose ratio D1 / D2. For example, if D1 / D2 = 0.5, ΔCD = 1 nm, making it possible to suppress the linewidth change ΔCD due to the change in wavelength characteristics to a considerable extent. Thus, even when a linewidth change occurs due to a change in wavelength characteristics under the same exposure dose, it is possible to suppress the linewidth change by making the relative exposure doses different between different wavelength characteristics. Furthermore, it is desirable to set the exposure dose ratio such that the difference between the characteristic value defined by equation (2) and D1 / D2 is small.

[0066] <Second Embodiment> In the first embodiment described above, there were no particular limitations on the difference between the steady state and the ground state. Before the exposure of the first substrate in the lot, there is a period during which no exposure operation is performed, and it is fundamental that no power is supplied to the LED elements during this time. Therefore, the LED elements cool down during this time, and after a sufficient amount of time has passed, the LED elements reach the same temperature as the refrigerant. Typically, the temperature of the refrigerant is set to be the same as the ambient temperature (room temperature). That is, at the beginning of the lot, the LED elements are at a temperature close to room temperature. Considering that the exposure operation starts from this state, the steady state and the ground state usually coincide.

[0067] However, by making the steady state and the ground state different, it is possible to reduce the change in resist linewidth due to the change in wavelength characteristics during exposure. In other words, by making the lowest temperature of the LED element during exposure different from that of room temperature, the wavelength characteristics of the ground state and the steady state can be made different, and as a result, it may be possible to reduce the change in resist linewidth. One way to make the lowest temperature of the LED element during exposure different from that of room temperature is to energize the LED element with the illumination light shutter closed, and raise the temperature of the LED element from room temperature by generating heat from the LED element itself. Alternatively, the temperature of the refrigerant can be made different from that of room temperature, or the temperature of the LED element can be adjusted by providing a heater in addition to the cooler to heat it.

[0068] (Example 3) In Example 3 of the second embodiment, an LED element with a steady-state peak wavelength of 365 nm and an LED element with a steady-state peak wavelength of 395 nm are used. In Example 3, these two LED elements are operated so that their base peak wavelengths are 367 nm and 397 nm, respectively. This can be achieved by keeping each LED element at a temperature at least 40 degrees Celsius higher than room temperature. On the other hand, Comparative Example 1 uses an LED element with a steady-state peak wavelength of 365 nm and an LED element with a steady-state peak wavelength of 395 nm, operated so that their base peak wavelengths are 365 nm and 395 nm, respectively. This corresponds to the normal case where the minimum temperature during exposure of each LED element is around room temperature. The change in wavelength characteristic Δλ is set to 3 nm in both Example 3 and Comparative Example 1. An overview of the respective wavelength characteristics is shown in Figure 12(a). The calculation conditions other than the wavelength characteristics are shown in Figure 16(b).

[0069] The simulation results are shown in Figure 12(b). In Comparative Example 1, as mentioned above, both CD(Δλ1,0) and CD(0,Δλ2) are positive relative to the target linewidth, resulting in an error of 50 nm relative to the target linewidth. In Example 3, on the other hand, CD(Δλ1,0) is 1511 nm, which is positive relative to the target linewidth, but CD(0,Δλ2) is 1493 nm, which is negative relative to the target linewidth. Therefore, the changes in both cancel each other out, resulting in a linewidth change of 6 nm, which is suppressed compared to Comparative Example 1. In this way, by making the steady-state wavelength characteristics different from the ground state wavelength characteristics, it is possible to reduce the linewidth change of the resist image due to changes in wavelength characteristics during exposure.

[0070] <Third Embodiment> Next, as a third embodiment, the exposure method will be described with reference to Figures 13 to 15. Figure 13 is a flowchart of the exposure method performed by the exposure apparatus according to this embodiment. In step S1, the first wavelength characteristic I1, the second wavelength characteristic I2, and the resist film thickness are determined. In step S2, the exposure amount D1 for the first wavelength characteristic and the exposure amount D2 for the second wavelength characteristic are determined. In step S3, it is determined whether the line width error caused by the change in wavelength characteristics that occurs when exposure is performed under the conditions determined in steps S1 and S2 is within the allowable value. If the determination is YES, that is, if the line width error is within the allowable value, exposure is performed in step S4. On the other hand, if the determination is NO, that is, if the line width error exceeds the allowable value, the process returns to step S1 and the exposure conditions are reviewed again. Steps S1, S2, and S3 are also called the exposure condition determination steps. Step S4 is also called the exposure step.

[0071] Steps S1 and S2 may be performed by actual exposure or by simulation. If actual exposure is performed, for example, exposure may be performed under multiple exposure conditions, and after development, the line width of the resist image may be measured to select an exposure amount close to the target line width. However, if actual exposure is performed, if the power supplied to the LED element is high, sufficient cooling may not be possible, and the LED element may not be able to be set to the desired temperature. In this case, the power supplied to the LED element may be reduced to a level that can maintain the desired temperature, and the exposure time may be extended to compensate for the exposure amount.

[0072] Refer to Figure 14 to explain the details of step S1. In step S11, the appropriate exposure amount is determined when exposure is performed at a 1:1 exposure ratio based on the first wavelength characteristic I1 and the second wavelength characteristic I2 in the ground state. Here, the first wavelength characteristic I1 and the second wavelength characteristic I2 used initially, the definition of the ground state (temperature of the LED element), and the resist film thickness are not particularly limited and can be arbitrary.

[0073] In step S12, the difference ΔCD(Δλ1,0) from the target linewidth when exposed with the exposure amount determined in S11 is evaluated based on the first wavelength characteristic of the transition state (wavelength change amount Δλ1) and the second wavelength characteristic of the ground state. In step S13, the difference ΔCD(0,Δλ2) from the target linewidth when exposed with the exposure amount determined in S11 is evaluated based on the first wavelength characteristic of the ground state and the second wavelength characteristic of the transition state (wavelength change amount Δλ2).

[0074] In step S14, it is determined whether the line width errors ΔCD(Δλ1,0) and ΔCD(0,Δλ2) evaluated in S12 and S13 have opposite signs. If the two line width errors have opposite signs, step S1 is terminated. If the two line width errors do not have opposite signs, the process proceeds to step S15. In step S15, it is considered whether the resist film thickness can be changed. If the resist film thickness can be changed, the resist film thickness is changed in step S16, and then the process returns to step S11 and is repeated.

[0075] If, in step S15, it is not possible to change the resist film thickness due to the exposure process settings, the process proceeds to step S17. In step S17, the first wavelength characteristic I1 or the second wavelength characteristic I2 is changed. Here, changing the wavelength characteristics also includes changing the ground state. That is, in step S17, the wavelength characteristics may be changed to those with different peak wavelengths, or the ground state (the lowest temperature of the LED element during the exposure process) may be changed. The change in wavelength characteristics can be performed, for example, by the method described in the second embodiment. Also, if multiple types of LED elements with different wavelength characteristics are arranged in the LED array light source, instead of changing the wavelength characteristics, an LED element with the desired wavelength characteristics may be selected.

[0076] Refer to Figure 15 to explain the details of step S2 for determining the exposure amount. In step S21, the maximum value of the wavelength change amount during exposure in the first wavelength characteristic Δλ is determined. 1max And, the maximum value of the wavelength change during exposure in the second wavelength characteristic Δλ 2max The following is determined. The initial values ​​are arbitrary, but both may be tentatively set to, for example, 3 nm. In step S22, the line width errors ΔCD(Δλ1,0), ΔCD(0,Δλ2) determined in step S1, and Δλ determined in step S21 are used. 1max , Δλ 2max The characteristic value of equation (2) is evaluated using this method.

[0077] In step S23, the exposure ratio D1 / D2 is determined such that the difference between the characteristic value evaluated in step S22 and the exposure ratio D1 / D2 is minimized. Here, it is desirable that the difference between the characteristic value and the exposure ratio D1 / D2 be 0, but this is not necessarily required; the exposure ratio D1 / D2 should be such that the error with the target line width is within the target value.

[0078] In step S24, exposure amounts D1 and D2 are determined. Specifically, the change in line width when exposure is performed while changing D1+D2 under conditions that satisfy the D1 / D2 determined in step S23 is evaluated, and D1+D2 close to the target line width is determined.

[0079] In step S25, the base temperature T1 and maximum temperature T of the first LED element are determined.1max And the base temperature T2 and maximum temperature T of the second LED element. 2max The following parameters that satisfy the conditions are determined. • Refrigerant flow rate, • Refrigerant temperature, • Input power P1, P2 of the LED element, • Exposure time. Here, the basal temperatures T1 and T2 are determined by the basal state determined in step S1, and the maximum temperature T 1max and T 2max This is Δλ determined in step S21. 1max , Δλ 2max This is determined by the base temperatures T1 and T2.

[0080] For example, let's assume the cooling conditions are standard. Next, under conditions where P1 / P2 = D1 / D2, the maximum temperature T 1max and T 2max P1 and P2 are determined as input powers that yield the maximum light output without exceeding a certain limit. Then, the exposure time is determined based on the exposure amounts D1 and D2 and P1 and P2.

[0081] In process S26, Δλ 1max , Δλ 2max The validity of Δλ is evaluated. 1max , Δλ 2max If this is appropriate, step S2 is completed. Here, if the temperature of one or both of the first and second LED elements falls below the maximum temperature, the cooling conditions may be adjusted to bring them up to the maximum temperature. Also, in step S26, Δλ 1max , Δλ 2max If it is determined that this is not appropriate, the process returns to step S21, and Δλ is determined based on the review in step S25. 1max , Δλ 2max This may be reconsidered.

[0082] According to the embodiments described above, the power input P1 to the first LED element, the power input P2 to the second LED element, and the cooling conditions of the cooler 24 can be set so that the maximum temperature of the first LED element and the second LED element during exposure is, for example, below the rated temperature. Furthermore, according to the embodiments described above, the power input P1 to the first LED element, the power input P2 to the second LED element, and the cooling conditions of the cooler 24 can be set so that the minimum temperature of the first LED element and the second LED element during exposure is different from the room temperature.

[0083] <Embodiment of Article Manufacturing Method> The article manufacturing method according to an embodiment of the present invention is suitable for manufacturing articles such as microdevices, semiconductor devices, and elements having a microstructure. The article manufacturing method of this embodiment includes the steps of forming a latent image pattern on a photosensitive agent coated on a substrate using the above-described exposure apparatus (a step of exposing the substrate) and developing the substrate on which the latent image pattern was formed in the above step. Furthermore, this manufacturing method includes other well-known steps (oxidation, film formation, vapor deposition, doping, planarization, etching, resist stripping, dicing, bonding, packaging, etc.). The article manufacturing method of this embodiment is advantageous over conventional methods in at least one of the performance, quality, productivity, and production cost of the article.

[0084] The invention is not limited to the embodiments described above, and various modifications and variations are possible without departing from the spirit and scope of the invention. Accordingly, claims are attached to disclose the scope of the invention. [Explanation of Symbols]

[0085] 1: Illumination optics, 2: Mask, 11a: First light source, 11b: Second light source, 12: Condenser lens, 13: Optical integrator, 14: Aperture diaphragm, 15: Condenser lens, 17: Beam combiner

Claims

1. An illumination optical system that illuminates a master plate with composite light from a first solid-state light-emitting element having a first wavelength characteristic and light from a second solid-state light-emitting element having a second wavelength characteristic, A projection optical system that forms a projected image on a substrate using light from the original plate, Equipped with, Based on the relationship between the film thickness of the resist coated on the substrate, the exposure wavelength, and the line width change of the pattern transferred onto the substrate, the first exposure amount, which is the amount of exposure from the first solid-state light-emitting element, and the second exposure amount, which is the amount of exposure from the second solid-state light-emitting element, are set such that the first line width change, which is the line width change of the pattern transferred onto the substrate caused by a change in the peak wavelength of the first wavelength characteristic, and the second line width change, which is the line width change of the pattern caused by a change in the peak wavelength of the second wavelength characteristic, have opposite signs, and the sum of the first line width change and the second line width change is smaller than an allowable value. An exposure apparatus characterized by the following features.

2. The first exposure amount is D １ , the second exposure amount is D ２ Toshi, D １ : D ２ When the ratio is 1:1, The wavelength sensitivity, which indicates the amount of linewidth change that occurs in response to the change in the centroid wavelength of the first wavelength characteristic, is (ΔCD / Δλ). １ The wavelength sensitivity, which indicates the amount of linewidth change that occurs in response to the change in the centroid wavelength of the second wavelength characteristic, is (ΔCD / Δλ). ２ year, The maximum value of the change in the centroid wavelength during exposure in the first wavelength characteristic is Δλ １ｍａｘ , the maximum value of the change in the centroid wavelength during exposure in the second wavelength characteristic is Δλ ２ｍａｘ In that case, |(DCD / DL) ２ ×Dl ２ｍａｘ | / |(DCD / DL) １ ×Dl １ｍａｘ | The characteristic value defined by and D １ / D ２ The first exposure amount and the second exposure amount are determined such that the difference between them is within an acceptable range. The exposure apparatus according to feature 1.

3. The system further comprises a control unit for controlling a light source unit including the first solid-state light-emitting element and the second solid-state light-emitting element, The control unit controls the power supplied to the first solid-state light-emitting element and the second solid-state light-emitting element, thereby achieving exposure of the first exposure amount by the first solid-state light-emitting element and exposure of the second exposure amount by the second solid-state light-emitting element. The exposure apparatus according to claim 1 or 2.

4. The system further comprises a control unit for controlling a light source unit including the first solid-state light-emitting element and the second solid-state light-emitting element, The control unit controls the energizing time for the first solid-state light-emitting element and the second solid-state light-emitting element, thereby achieving exposure of the first exposure amount by the first solid-state light-emitting element and exposure of the second exposure amount by the second solid-state light-emitting element. The exposure apparatus according to claim 1 or 2.

5. The control unit controls the light source unit so that the lowest temperature of the first solid-state light-emitting element and the second solid-state light-emitting element during exposure is higher than the ambient temperature of the light source unit. The exposure apparatus according to claim 3 or 4.

6. The light source unit includes a cooler for cooling the first solid-state light-emitting element and the second solid-state light-emitting element. The exposure apparatus according to any one of claims 3 to 5.

7. The exposure apparatus according to claim 6, characterized in that the power supplied to the first solid-state light-emitting element, the power supplied to the second solid-state light-emitting element, and the cooling conditions of the cooler are set so that the maximum temperature of the first solid-state light-emitting element and the second solid-state light-emitting element during exposure is below the rated temperature.

8. The exposure apparatus according to claim 6 or 7, characterized in that the power supplied to the first solid-state light-emitting element, the power supplied to the second solid-state light-emitting element, and the cooling conditions of the cooler are set so that the lowest temperature of the first solid-state light-emitting element and the second solid-state light-emitting element during exposure is different from the room temperature.

9. The exposure apparatus according to any one of claims 1 to 8, characterized in that the illumination optical system includes a beam combiner that combines light from the first solid-state light-emitting element and light from the second solid-state light-emitting element.

10. The exposure apparatus according to claim 9, characterized in that the beam combiner is a dichroic mirror.

11. Among the wavelengths at which the intensity in the first wavelength characteristic becomes half of the peak intensity, the longer wavelength is λ １ｄ λ is the wavelength on the shorter side among the wavelengths at which the intensity in the second wavelength characteristic becomes half of the peak intensity. ２ｄ In that case, λ ２ｄ -λ １ｄ The first and second wavelength characteristics are defined such that the wavelength is 15 nm or greater. The exposure apparatus according to any one of claims 1 to 10.

12. An exposure method performed by an exposure apparatus comprising: an illumination optical system that illuminates a master plate with composite light of light from a first solid-state light-emitting element having first wavelength characteristics and light from a second solid-state light-emitting element having second wavelength characteristics; and a projection optical system that forms a projected image on a substrate with light from the master plate, wherein A determination process for determining exposure conditions, The process includes an exposure step in which exposure is performed according to the determined exposure conditions, The determination step includes determining a first exposure amount, which is the amount of exposure by light from the first solid-state light-emitting element, and a second exposure amount, which is the amount of exposure by light from the second solid-state light-emitting element, based on the relationship between the film thickness of the resist coated on the substrate, the exposure wavelength, and the line width change of the pattern transferred onto the substrate, such that a first line width change, which is the line width change of the pattern transferred onto the substrate caused by a change in the peak wavelength of the first wavelength characteristic, and a second line width change, which is the line width change of the pattern caused by a change in the peak wavelength of the second wavelength characteristic, have opposite signs, and the sum of the first line width change and the second line width change is less than an allowable value. A method of exposure characterized by the following features.

13. A step of exposing a substrate using an exposure apparatus according to any one of claims 1 to 11, The process of developing the exposed substrate, A method for manufacturing an article, characterized by including a developed substrate and manufacturing an article from the developed substrate.

Images