JP3433844B2 - Filter device for exposure apparatus and projection exposure apparatus - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a filter for removing contaminants in the air used in an exposure apparatus, and is particularly suitable for use in an exposure apparatus having a light source that emits light in the ultraviolet range. is there.

[0002]

2. Description of the Related Art Conventionally, an apparatus for illuminating an object to be illuminated with light from a discharge lamp has been used for various purposes in various fields. Above all, a semiconductor element such as LSI or a liquid crystal display element is subjected to a photolithography process. In the reduction projection type exposure apparatus (stepper, aligner, etc.) used in the production of the light of a specific wavelength of the light output from the ultra-high pressure mercury lamp (Hg lamp, Xe-Hg lamp, etc.) ( An apparatus for illuminating a reticle on which a transfer pattern is formed with an i-line having a wavelength of 365 nm, a g-line having a wavelength of 436 nm, etc. is used.

In such a projection exposure apparatus, research and development for transferring a finer pattern on a photosensitive substrate with high resolution have been earnestly pursued. Generally, when the numerical aperture of the projection optical system of the projection exposure apparatus is NA and the wavelength of the exposure light is λ, the resolution R and the depth of focus DO of the projection exposure apparatus.
F can be expressed as: R = k ₁ · λ / NA (1) DOF = k ₂ · λ / NA ² (2) However, in the above equation, k ₁ and k ₂ are coefficients determined by the process. According to the above equation, pattern miniaturization can be achieved by either of the following two methods. Enlargement of numerical aperture NA of projection optical system Shortening of wavelength of exposure light (exposure wavelength) λ

Regarding the numerical aperture of the projection optical system of these two methods, in recent years, a projection optical system having a large numerical aperture of 0.5 to 0.6 has been realized, which improves the resolution. There is. However, if the numerical aperture NA of the projection optical system is simply increased, the depth of focus D
There is an inconvenience that the OF becomes small in inverse proportion to the square of the numerical aperture NA. Generally, in an actual semiconductor process, it is necessary to expose a circuit pattern on a wafer having a step in the previous step, and it is also necessary to absorb a flatness error of the wafer itself. It is necessary to secure a sufficiently large value.

On the other hand, in the method of shortening the exposure wavelength λ, the depth of focus DO is
F changes in proportion to the wavelength λ of the exposure light. Therefore, it is advantageous to improve the resolution by shortening the exposure wavelength λ in terms of ensuring the depth of focus. From such a background, as the exposure light in the projection exposure apparatus, from the bright line called g-line (wavelength 436 nm) of the conventionally used mercury lamp, the i-line (wavelength 365 nm) of the same mercury lamp is now used.
It is becoming mainstream to use the bright line called.

FIG. 4 shows an example of an illumination optical device using a conventional mercury lamp for a projection exposure apparatus as a light source. In FIG. 4, the light emitting point of the mercury lamp 1 is on a first focal point F1 in the elliptical mirror 2. It is located in. An opening for passing the electrode portion of the mercury lamp 1 is formed at the end of the ellipsoidal mirror 2, and, for example, aluminum or various multilayer dielectric materials are vapor-deposited on the inner surface of the ellipsoidal mirror 2, and the inner surface serves as a reflecting surface. To work. The light L emitted from the mercury lamp 1 is reflected by the inner surface of the elliptic mirror 2 and travels to the mirror 3 for bending the optical path. Aluminum or various multilayer dielectric materials are also vapor-deposited on the reflecting surface of the mirror 3, and the light reflected by the mirror 3 is focused on the second focal point F2 of the elliptic mirror 2, and the light source is placed on the second focal point F2. An image is formed.

The divergent light from the light source image is converted into a substantially parallel light flux by the input lens 4 and is incident on the narrow band bandpass filter 5. Illumination light having a wavelength selected by the bandpass filter 5 enters a fly-eye lens 6 serving as an optical integrator, and a large number of secondary light sources are formed on the focal plane on the rear side (reticle side) of the fly-eye lens 6. The divergent light from these many secondary sources is
After being reflected by the mirror 7 for bending the optical path, it is condensed by the condenser lens 8 to illuminate the pattern forming surface of the reticle 9 as the irradiated surface in a superimposed manner. The reflecting surface of the mirror 7 is also vapor-deposited with aluminum or various multilayer dielectric materials.

The entire optical system is compactly assembled by mirrors 3 and 7 for bending the optical path.
Further, the inner surface of the elliptic mirror 2 as the condenser mirror, the reflecting surface of the mirror 3 and the reflecting surface of the mirror 7 are designed so that the maximum reflectance is obtained at the wavelength of the exposure light. An ultrahigh pressure mercury lamp is used as the mercury lamp 1 in FIG. 4, and the emission spectrum distribution of this ultrahigh pressure mercury lamp is shown in FIG.
Shown in. Further, FIG. 6 (A) shows the wavelength dependence of the reflectance of an aluminum reflecting mirror in which aluminum is deposited on the reflecting surface.
FIG. 6B shows the wavelength dependence of the reflectance of a conventional representative dielectric multilayer film reflecting mirror in which a multilayer dielectric film is deposited on the reflecting surface. Further, FIG. 7 shows the wavelength dependence of the transmittance of the bandpass filter 5 in FIG. 4 when the exposure light is the i-line (wavelength 365 nm). With such a configuration, the i-line illumination light is selected, the pattern of the reticle 9 is illuminated by the selected illumination light with a uniform illuminance distribution, and the projection optical system in which the image of the pattern of the reticle 9 is omitted is shown. An image was formed on the photosensitive substrate via the.

When the conventional illumination optical device as described above is operated in a state where it is open to the outside, the optical members from the mercury lamp 1 to the bandpass filter 5 (in the case of FIG. 4, the elliptical mirror 2, the mirror 3 and the input). A fog phenomenon occurs on the surface of the lens 4 and the incident surface of the bandpass filter 5, and the reflectance or the transmittance gradually decreases, which causes a problem that the illumination efficiency decreases. It was found that this fogging phenomenon was caused by the deposition of a fogging substance on the optical member. And
As a result of analysis by an ion chromatography method, it was found that most of the cloudy substances were ammonium sulfate ((NH ₄ ) ₂ SO ₄ ).

As a result of analysis by ESCA, SEM (scanning electron microscope), etc., there are cases where silicon oxide (SiOx) adheres to the surfaces of the lens and the mirror. Adhesion of silicon oxide causes scattering if the surface is not smooth. Even if the surface is smooth, if silicon oxide adheres to the antireflection film or the high reflection film, the antireflection film or the reflection film will be out of the proper condition, and the transmittance or the reflectance will decrease, As a result, the illuminance was lowered. These cloudy substances include ammonium ions (NH ₄ ⁺ ) and sulfate ions (SO ₄ ^2- ) or organic silanols, which are present due to some factors, and far ultraviolet light (UV light) of optical members.
It is considered that they were attached to the irradiated part by photochemical reaction.

Are these ions or compound molecules generated from the surface of the black alumite (BAm) material, which is often used as a holding member and a shielding member in the illumination optical system, or is it originally present in the air? , And has been considered to have been ionized by irradiation with UV light. In the case of a black alumite material, a diazo dye is used, and since sulfuric acid is used in the black alumite treatment step, it can be a source of generation of ammonium groups and sulfuric acid groups which are constituents of ammonium sulfate. UV light was irradiated to the black alumite material in a nitrogen (N ₂ ) atmosphere, an oxygen (O ₂ ) atmosphere, and a normal air atmosphere containing a large amount of water vapor. , The most ammonium and sulfate groups were generated in the air containing water vapor, and then these ions were generated in the oxygen (O ₂ ) atmosphere. That is, it was confirmed that the contamination is likely to be promoted by the presence of water in the air.

When oxygen (O ₂ ) is irradiated with UV light, it becomes ozone with high reactivity, which promotes ion generation from the wall surface of the holding member and ionization of gas in the air. To be However, the source of silicon oxide could not be found in the illumination optical device. As a result of examining the installation environment dependency of the degree of dirt of the device, it was found that many ionic substances such as ammonium group, sulfate group and nitrate group exist in the clean room, and H
Many organic silanes such as MDS (hexamethyldisiloxane) and trimethylsilanol were detected. HMDS is a material often used as a surface treatment material when coating a photosensitive material on a wafer, and trimethylsilanol is HM.
It is a substance formed by hydrolysis of DS. As a result of the investigation, it was found that there is a very good correlation between the amount of the fogging substance of the optical member and the amount of the impurities contained in the environment. As a result of detailed examination of the degree of contamination of the optical member and the situation, it was found that the source of the cloudy substance is not in the device but in the environment in which the device is installed.

Further, in recent years, the degree of integration of semiconductor integrated circuits has further increased, and it has become necessary to form the minimum line width of the circuit in units of submicrons. As one means for coping with such miniaturization of circuits, it is conceivable to use exposure light with a short wavelength in a projection exposure apparatus for producing a semiconductor integrated circuit. Currently, KrF excimer laser light with a wavelength of 248 nm or titanium is used. -Sapphire laser harmonics,
4th harmonic of YAG laser with wavelength 266nm, wavelength 21
Attention has been paid to a 5th harmonic of 3 nm, an ArF excimer laser beam having a wavelength of 213 nm, or the like.

Conventionally, a semiconductor pattern is formed by an exposure apparatus using a light source having a long wavelength such as g-line and i-line, and a photoresist (photosensitive resin) composed of a novolac resin called a novolac resin and a photosensitive material is formed on a wafer. And apply
It has been done by the already stable process of exposing and developing it. However, if the wavelength of the light source is, for example, Kr
When the wavelength is shorter than that of the F excimer laser (248 nm), the light absorption of the resin is increased, and therefore, this novolac photoresist cannot form a well-shaped pattern. Therefore, a new type of photoresist called chemical amplification type has appeared for exposure equipment that uses a short-wavelength light source such as an excimer laser, and its good characteristics such as pattern formation characteristics and resolution have attracted attention and become the mainstream at present. Is coming. Chemically amplified photoresist is
Generally, it is composed of components such as a resin, a photosensitive acid generator, a dissolution accelerator or a crosslinking agent, and an acid is generated from the acid generator upon exposure, and the acid serves as a catalyst during baking (PEB) after exposure. The reaction of a dissolution accelerator or a crosslinking agent is promoted to form a pattern by development. The one using the dissolution accelerator forms a positive type pattern, and the one using the crosslinking agent forms a negative type pattern.

Although chemically amplified photoresists used as photoresists for exposure apparatuses using ultraviolet light are excellent in terms of resolution, it is difficult to control the catalytic action of acid generated during exposure during PEB. However, there is a problem of lack of stability. Particularly in the case of a positive photoresist, if a basic gas such as ammonia or amine is present in the air between the exposure and PEB, the generated acid reacts and escapes into the air, making the photoresist surface insoluble. There is also a phenomenon that layers are formed. When the insolubilized layer is formed, the formed pattern becomes a T-shape having an "overhang" on the upper part (T-top phenomenon), which greatly hinders the subsequent steps such as etching. Since these gases are often present in the air of the clean room, it has been a factor that hinders the use of the chemically amplified photoresist. In other words, it has been difficult to use the chemically amplified photoresist due to a substance such as ammonia that causes a cloudy substance on the surface of the optical member.

In this regard, conventionally, for example, a HEPA filter (High Efficiency Particulate Ai) for removing dust is used as a filter for air flowing through the interior of an illumination optical device.
r Filter) has been used conventionally. In addition, taking advantage of the fact that the decomposition of ammonium sulphate begins at about 120 ° C (edited by the Chemical Dictionary Dictionary: “The Chemical Dictionary”, Vol. 9, P6).
90, Kyoritsu Shuppan 1964), a method of avoiding the adhesion of ammonium sulfate by keeping the optical member at the temperature or higher is disclosed in Japanese Patent Application Laid-Open No. 4-128702 by the present applicant.

[0017]

[Problems to be Solved by the Invention] However, HEP
Since the A filter is a filter for removing particles, there is a disadvantage that impurities such as ammonia (ammonium ion), sulfate ion, and silicon oxide which cause the photochemical reaction as described above cannot be removed. Also,
Regarding the method of keeping the optical members above a certain temperature, the condensing mirror close to a large "heat source" such as a mercury lamp can raise the temperature relatively easily, but other optical members use a considerably large new heat source. Since it is necessary, in the case of an exposure apparatus for manufacturing a semiconductor or the like that requires strict temperature control, the heat exhausting method thereof becomes a problem.

As a method for eliminating the above inconvenience and avoiding the adhesion of ammonium sulfate ((NH ₄ ) ₂ SO ₄ ) which is a particular problem, a filter having a chemical adsorption mechanism (so-called chemical filter) is used. A method of removing ammonia or sulfate ions in the air can be considered. Further, as a method for avoiding the formation of the surface hardly soluble layer in the chemically amplified resist, a method of removing ammonia and basic gas in the air by a filter having a chemical adsorption function (chemical filter) can be considered.

However, the fogging phenomenon of the illumination system and the T generated by using the chemically amplified photoresist
It is known that the top phenomenon occurs even when the concentration of the impurity gas is extremely low, and in the case of ammonia, it is several pp.
It is known that the T-top phenomenon occurs only when b is present. Filtering such ppb-order impurities involves various technical difficulties. The main difficulty among them is the method of determining the life of the chemical filter.
The life of the chemical filter largely depends on the environment in which it is used. However, the impurity gas concentration in the ambient air in a clean room of a normal semiconductor manufacturing plant cannot be said to be constant, and varies greatly depending on the operating status of manufacturing equipment, the content of processes, the number of workers, the environment around the factory, the season, etc. To do. Therefore, even if the amount of ventilation to the chemical filter is constant, the amount of impurities passing through the filter varies.

Further, since the gas concentration at the ppb level is near the measurement limit in the ion chromatography method and the gas chromatograph method, it takes a huge amount of time to judge the filter life from the gas concentration around the exposure apparatus. There is the inconvenience of cost. Furthermore, the measurement limit of various gas sensors currently on the market is at most about 1 ppm, so it is considered from the intended use that the life of the chemical filter is diagnosed by installing these gas sensors on the downstream side of the chemical filter. Is meaningless.

In view of the above, the present invention removes impurities to an extremely low concentration level, which causes an unnecessary fogging phenomenon in an optical member in an illumination optical device or causes a T-top phenomenon when a chemically amplified resist is used. An object of the present invention is to provide a filter device for an exposure apparatus, which is provided with a means and can easily determine the life of the means.

[0022]

A filter for an exposure apparatus according to the present invention is a filter apparatus for a gas that flows in a predetermined direction around the exposure apparatus, wherein ammonia (including ammonium ion), amine compound (amine ion Filter means (42) for removing from the gas a predetermined substance selected from the group of substances consisting of (including), sulfur oxide (including sulfite ion and sulfate ion), and silicon oxide (including silicon oxide ion). An adsorbent (43) for adsorbing the predetermined substance on the gas inflow side of the filter means, and a measuring means (53, 54, 57) for measuring the amount of the substance adsorbed by the adsorbent, The determination means (58) for determining the life of the filter means (42) based on the measurement result by the measurement means is provided.

In this case, one example of the measuring means is the frequency detecting means (53, 54) for detecting the resonance frequency of the adsorbent (43) and the adsorbent (43) based on the detection result of the frequency detecting means. The calculation means (57) for obtaining the amount of the adsorbed substance. It is also desirable to install a dust removing filter (44) at the gas outlet from the filter means (42).

Further, when the exposure apparatus is installed in the chamber (50) isolated from the outside air, the filter means (42) and the adsorbent (43) are introduced into the chamber (50). May be placed in the mouth.

[0025]

According to such a filter for an exposure apparatus of the present invention, clean air in which at least a part of predetermined harmful ions and harmful organic gas is removed by the filter means (42) is introduced into the periphery of the exposure apparatus. Therefore, the white haze of the optical member is reduced, or the T-top phenomenon when using the chemically amplified resist is reduced. Also, filter means (42)
Since the adsorbent (43) for adsorbing a predetermined impurity is provided on the upstream side of the filter and the life of the filter means (42) is determined from the amount of the impurity adsorbed by the adsorbent, the concentration of the impurity gas to be detected is easy. Is high enough to be measured (for example, p
(pm order), the life of the filter means (42) can be determined with high accuracy without tedious operations.

In order to determine the life of the filter means (42), a method of first measuring the concentration of the impurity gas contained in the gas after passing through the filter can be considered. The life of the filter means (42) at issue here is
The impurity gas concentration in the gas after passing through the filter is, for example, several p
It means the time when it becomes larger than pb. Regarding the case of removing ammonia in order to suppress the T-top phenomenon, it is considered that the time when the concentration after passing through the filter means becomes 1 ppb or more is the life of the filter means. At present, the only method for measuring the ammonia concentration around 1 ppb is the ion chromatography method, but the sampling time is several hours, the cost required for one analysis is expensive, and it is a practical method for determining the filter life. Not the way. The same applies to other ions and organic gases.

In the present invention, however, the impurity gas concentration is measured on the upstream side of the filter means (42) and in the form integrated by the adsorbent (43). As compared with the case of measuring itself, even a very small amount of gas concentration can be detected without any trouble, and the life of the filter means (42) can be accurately determined.

When determining the life of the filter means (42) from the resonance frequency of the adsorbent (43), the adsorbent (43) is proportional to the amount of the predetermined impurities passing through the filter means (42). Since the resonance frequency of the filter means changes, it is possible to determine the integrated amount of impurities flowing into the filter means (42) and thus the life of the filter means (42) only by measuring the frequency.

Further, if a filter (44) for removing dust is installed at the gas outlet from the filter means (42), fine dust components such as metal powder which adversely affects the optical member are removed, and Dust that may be generated from the filter means (42) is also removed. In addition, filter means (4
2) and the adsorbent (43) are installed at the inlet (48) of the outside air to the chamber (50) isolated from the outside air,
Since harmful ions and harmful substances in the air introduced into the chamber (50) are reduced, not only contamination of optical members is prevented but also amine gas and chemically amplified positive resist coating on the wafer during the exposure operation are prevented. The reaction in the exposed area is also reduced.

[0030]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS An embodiment of a filter device for an exposure apparatus according to the present invention will be described below with reference to FIGS.
In this embodiment, the present invention is applied to an illumination optical system of a projection exposure apparatus for manufacturing a semiconductor device. In FIGS. 1 and 2, parts corresponding to those in FIG. Omit it. In addition, this embodiment also deals with the case where the i-line (wavelength 365 nm) of a mercury lamp is used as the illumination light.

FIG. 1 shows an example of an optical system of the illumination optical device of the exposure apparatus of this embodiment. In FIG. 1, the light emitted from the mercury lamp 1 is reflected by the inner surface of the elliptical mirror 2 and the optical path thereof is changed. Head towards the folding mirror 3. For the mirror 3,
A dichroic mirror that reflects ultraviolet light and transmits visible light and infrared light is used. The light reflected by the mirror 3 enters the input lens 4. The light emitted from the input lens 4 enters the bandpass filter 5, and the illumination light including the i-line selected by the bandpass filter 5 enters the fly-eye lens 6 and is formed on the exit surface of the fly-eye lens 6. Illumination light from a large number of secondary light source images is directed to a relay lens system (not shown). This relay lens system includes a fly-eye lens 6 and a condenser lens 8 shown in FIG.
And a lens system for generating a conjugate surface with the reticle 9, the lens system being disposed between and. An interference filter is used as the bandpass filter 5.

Further, in FIG. 1, the light transmitted through the mirror 3 includes a lamp monitor (not shown), a wafer global alignment (WGA) optical system, and an auto focus (A).
F) Guided to the optical system. The lamp monitor is a mercury lamp 1
, A WGA optical system is an optical system for performing a rough alignment of a wafer onto which a reticle pattern is projected, and an AF optical system is an oblique optical system for the photoresist on the wafer. An optical system for irradiating a photosensitive detection light to detect the focus position of the wafer.

In FIG. 1, the surface marked with a star, that is, the elliptical mirror 2
The inner surface, the reflecting surface of the mirror 3, the both surfaces of the input lens 4, and the incident surface of the bandpass filter 5 are the surfaces on which the fog phenomenon appears when no anti-fog measures are taken.
So far, no significant cloudy surface has appeared in the optical members subsequent to the bandpass filter 5. From this, it can be assumed that the clouding phenomenon is a photochemical reaction involving light in a wavelength range of less than 365 nm.

In this embodiment, an air curtain made of chemically clean air is provided to each optical member except the elliptic mirror 2 in the region where the fog phenomenon from the elliptic mirror 2 to the incident surface of the bandpass filter 5 is observed. Form and isolate and isolate the optical components from the chemically unclean atmosphere. As used herein, chemically clean air is a mechanism for chemically adsorbing ammonia ions (or ammonia) or sulfate ions, which are substances causing fog in the illumination system, by using an impurity removal filter generally called a chemical filter as described later. The organic silanol and the like are removed by the physical adsorption mechanism.

As a method of passing air through the chemical filter, there is a method in which a chemical filter, a HEPA filter for removing dust, a fan for blowing air, and the like are unitized when dust is a problem. In this case, in order to use the chemical filter efficiently, it is desirable to make the air flow uniform on the surface of the chemical filter. Also, if the cleanliness of particles (dust particles) in the air that has passed through the chemical filter does not matter, or if the dust generation of the primary side air, the chemical filter, and the blower fan is so low that it does not matter, The HEPA filter may be removed from the configuration. Further, in that case, if the primary side air has a sufficient positive pressure or the secondary side has a sufficient negative pressure, the blower fan may be further removed from the above configuration.

The air curtain may be formed individually for each optical member, or if the optical members are close to each other, a wide range of air curtains may be collectively formed for a plurality of optical members. FIG. 1 shows the case where the air curtain is formed for the input lens 4 alone, and the mechanism for forming the air curtain for the other optical members is not shown.

That is, in FIG. 1, air from a temperature control unit (not shown) is passed through the pipe 34 to the chemical filter 4
2 (see FIG. 2 ) is guided to the filter unit 35, and clean air from which impurities have been removed by the filter unit 35 is blown through the blower fan 45, the HEPA filter 44, and the pipe 36 to form an air curtain. 37, and clean air is blown from the blowout port 37 to the periphery of the input lens 4. The outlet 37
Are arranged at positions that do not interfere with the optical path for illumination.

That is, on the downstream side of the filter unit 35, a HEPA filter 44 for removing dust is installed. Since the HEPA filter 44 has a large pressure loss, the blower fan 45 is installed in front of it. However,
As described above, when the cleanliness of particles (dust particles) in the air that has passed through the chemical filter does not matter, or when the dust generation of the primary side air, the chemical filter, and the blower fan is so low that it does not matter. The HEPA filter 44 may be removed from the configuration.
At that time, if there is a sufficient positive pressure in the primary air,
If there is a sufficient negative pressure on the secondary side, the blower fan 45 may be removed from the above configuration.

Usually, since the HEPA filter 44 is used, the pressure loss thereof is large, so that the air on the primary side is pressure-fed to the secondary side by the blower fan 45 or the like. At this time, it is necessary to pay attention to the selection of the material of the pipe 36 and the like so that the air after passing through the filter is not mixed with impurities from the pipe or the like. Further, it is desirable that the air on the secondary side is not circulated as much as possible.

Further, in order to suppress the generation of harmful ions,
It is conceivable to change the holding material and the shielding material, which are the generation sources, to materials with less ion generation, and to flow high-purity nitrogen gas (N ₂ ) to reduce the influence of oxygen (O ₂ ). To be However, a considerable amount of nitrogen (N ₂ ) is required to flow nitrogen (N ₂ ) gas through the entire illumination optical system, which may cause suffocation. Need to be thorough.

Further, the mercury lamp 1, the elliptical mirror 2 and the mirror 3 are provided.
A duct 39 is attached so as to cover the air duct 40, and an air flow 40 for cooling the mercury lamp 1 is blown from a blowing hole 39a near the end face of the elliptic mirror 2 of the duct 39 via a sirocco fan or the like.
And the air flow 40 is forcibly exhausted from the exhaust hole 39b of the duct 39 at the bottom of the elliptic mirror 2. Further, a partition plate 41 having an opening for passing the light from the mercury lamp 1 is fixed between the elliptic mirror 2 and the mirror 3 inside the duct 39, and the partition plate 41 flows in from the opening 39c of the duct 39 facing the input lens 4. The generated air is forcibly discharged from the discharge hole 39b through the opening of the partition plate 41. In addition, the lamp monitor that passed through the mirror 3, WGA
The light for the optical system and the light for the AF optical system are extracted to the outside through the transmission window on the side surface of the duct 39.

In this case, since the elliptic mirror 2 is irradiated with heat from the mercury lamp 1 and a large amount of cooling air flow 40 exists near the elliptic mirror 2, it is not possible to form an air curtain here. Have difficulty. Therefore, the ellipsoidal mirror 2 is prevented from adhering ammonium sulfate by controlling the temperature. The temperature control for avoiding the adhesion of ammonium sulfate ((NH ₄ ) ₂ SO ₄ ) to the optical member is based on the fact that the decomposition of ammonium sulfate starts at about 120 ° C. Kai: “Chemical Dictionary” Vol.9, P690, Kyoritsu Shuppan 1964), which is a method to keep the optical member above the temperature. The elliptical mirror 2 adjacent to the mercury lamp 1, which is a large heat source, can be raised in temperature relatively easily.

FIG. 2 is an enlarged schematic view showing the internal structure of the filter unit 35 shown in FIG. 1. In FIG. 2, the chemical filter 4 is located in the middle of the cylindrical casing 51.
2 is fixed, and a gas sensor 43 for measuring an integrated flow rate of a predetermined substance in the fluid gas is arranged in the casing 51a on the upstream side of the chemical filter 42. This gas sensor 43 is a crystal unit (AT cut type) 43a.
And the adsorption film 43b applied to the surface of the vibrator 43a
Consists of. A transducer 52 is attached to one end of the crystal oscillator 43a, and the transducer 52 is composed of electrode plates sandwiching the crystal oscillator 43a and a converter (consisting of a voice coil motor or the like) for converting the vibration of the crystal oscillator 43a into an electric signal. Has been done. Then, in response to a start signal from the control unit 56 in the microprocessor (MPU) 55, the drive unit 53 gives vibration of a wide frequency to the crystal oscillator 43a via the transducer 52, and the electric signal from the transducer 52 is changed in frequency. The resonance frequency of the crystal oscillator 43a and the adsorption film 43b is obtained by analysis by the measurement unit 54, and this resonance frequency is supplied to the arithmetic unit 57 in the microprocessor 55. The calculation unit 57 subtracts the first resonance frequency from the resonance frequency to obtain the change amount ΔF and informs the determination unit 58, and the determination unit 58 determines from the change amount ΔF whether the chemical filter 42 has reached the life. . The arrow in FIG. 2 indicates the direction of gas (air) flow.

Next, the operation of the gas sensor 43 installed on the upstream side of the chemical filter 42 will be described. The surface of the crystal oscillator 43a of the gas sensor 43 has a predetermined adsorption film 4
It is covered with 3b. As the adsorption film 43b, an optimum material is selected according to the adsorption target gas.

When the adsorption film 43b adsorbs a gas, the adsorption film 43b undergoes a mass change, and the resonance frequency changes as shown in the following equation (3). Variation of resonance frequency dF = −2 · F ₀ ² · ΔM / (ν · S · ρ) (3) where F ₀ is the fundamental resonance frequency, and ν is bulk transverse wave velocity (3,320) propagating in the thickness direction of the crystal unit. m / s), S is the electrode area, ρ is the crystal density (2.65 × 10 ³ kg / m ³ ), and ΔM is the mass change of the adsorption film 43b.

A detailed description will be given by taking ammonia as an example of the gas to be adsorbed. In this case, a polymer film is used as the base of the adsorption film 43b that covers the surface of the crystal oscillator 43a. This polymer film is formed on the sensor surface of the gas sensor by the LB method. Here, the LB method is the Langmuir-Blodgett method, and is a method of forming a thin film composed of a layer of a single molecule or several molecules. A substance capable of immobilizing a sensor substance that reacts with a target gas and holding a reaction product in a porous polymer film is selected as a film forming material. As the polymer material, a substance capable of forming an extremely thin film is desirable, and a typical example thereof is cellulose. Examples of the sensor substance include phosphoric acid (H ₃ PO ₄ ) or potassium permanganate (KM
_n O ₄ ) etc. are selected.

Ammonia (including ammonium ions) in the environmental gas causes a neutralization reaction with phosphoric acid in the adsorption film 43b on the crystal oscillator 43a and is fixed on the adsorption film 43b. Since the neutralized ammonia does not desorb, the amount of ammonia proportional to the total amount of ammonia passing through the chemical filter 42 is equal to that of the adsorption film 43 of the gas sensor 43.
It is continuously adsorbed on the surface of b.

Now, the operation of the chemical filter 42 shown in FIG. 2 will be described. FIG. 3 is a diagram showing the ammonia concentrations before and after the chemical filter 42 (upstream side, downstream side). FIG. 3 (A) shows an experimental example in which ammonia is mainly adsorbed by physical adsorption by a chemical filter using activated carbon or the like, and FIG. 3 (B) is mainly chemical adsorption by a chemical filter using ion exchange fibers or the like. An example of an experiment for adsorbing ammonia will be shown. In FIGS. 3A and 3B, the horizontal axis represents time and the vertical axis represents ammonia concentration.

In FIG. 3, the curve f (t) is the upstream concentration, g
(T) shows the downstream concentration. The upstream concentration f (t) actually fluctuates as shown by the broken line, but for convenience, its average value is C
Calculate as ₁ . Assuming that the filter breakthrough concentration on the downstream side (downstream concentration that is a factor for determining the filter life) is C _R , the foot T _C of the perpendicular line of the intersection R between the straight line Y = C _R and the curve Y = g (t) is It is the life of the filter.

Here, when the total amount of ammonia up to the breakthrough of the filter is Q, the following equation (4) is established. Q = ∫f (t) dt = C ₁ · T _C (4) However, f (t) generally fluctuates greatly due to different environmental conditions in the clean room or seasonal factors.
It is usually difficult to determine C ₁ .

However, in this embodiment, the gas sensor 43 is installed upstream of the chemical filter 42.
Indicated by the gas sensor 43, is obtained by integrating the change dF in the resonance frequency is proportional to the total ammonia amount Q, shift amount F _R (later from an initial value of the resonance frequency indicated by the gas sensor 43 at the time of filter breakthrough, broken " The following relationship holds for "overfrequency").

F _R = ∫dF = α · Q (5) where α is a coefficient determined by the treated air volume, installation conditions, etc., and Q is an amount unique to each chemical filter, both of which can be experimentally determined. It is a large amount. Specifically, when the experimental system is different from the actual system, the upstream ammonia concentration is artificially kept almost constant by using a permeator or standard ammonia gas (however, at extremely high concentrations, Since the reaction mechanism of the chemical filter changes, it is necessary to suppress it to about several tens of ppb), while measuring the upstream and downstream concentrations of the chemical filter using an ion chromatograph, and checking the change over time of the concentration to determine the service life. T
_C is actually determined, and the total ammonia amount Q is determined by the above equation (4). The experimental system at this time must be hydrodynamically similar to the system actually used. If the experimental system is the same as the actual system, a gas sensor may be actually installed to directly measure the coefficient α and the breakthrough frequency F _R.

The values of the total amount of ammonia Q, the coefficient α, and the breakthrough frequency F _R obtained before the breakthrough of the filter are stored in the judgment unit 58 of FIG. During actual operation, the gas sensor 4
The resonance frequency of the crystal oscillator 43a of No. 3 is measured by the frequency measurement unit 54 constantly or as needed, and the measured value is transmitted to the calculation unit 57. The calculation unit 57 subtracts the first resonance frequency from the supplied resonance frequency to calculate the frequency change ΔF.
Ask for. This is equivalent because the amount of the substance adsorbed on the adsorption film 43b is indirectly obtained from the equation (3). The determination unit 58 determines that the chemical filter 42 has reached the end of its life when the variation ΔF reaches the previously stored breakthrough frequency F _R up to the time of filter breakthrough. The calculation unit 57
Then, the amount ΔM of the adsorbed substance is actually calculated from the equation (3),
The determination unit 58 may determine the life based on this amount ΔM.

What is important here is that the ammonia concentration in a normal clean room is about several tens of ppb, so that the desorption type gas sensor cannot detect the ammonia concentration. In this embodiment, in order to solve the problem, the adsorption film 43b of the gas sensor 43 has a property of adsorbing and fixing the passing gas. This makes it possible to calculate the passing amount of ammonia.

In this way, by installing the integrating type gas sensor 43 on the upstream side of the chemical filter 42 and measuring the total throughput, it is possible to judge the life of the chemical filter installed and used in various environments. You can The same applies to impurity gases other than ammonia. However, as the adsorption film (sensor substance) of the gas sensor, the most effective substance needs to be selected from substances capable of adsorbing and fixing the gas to be treated.

As a sensor substance for ammonia,
In addition to the above phosphoric acid, zinc chloride (ZnC1 ₂ ) or the like may be used.
Examples of the sensor substance for sulfuric acid include potassium hydroxide (KOH) and calcium hydroxide (Ca (OH) ₂ ).
Further, in FIG. 1, in addition to the case where an air curtain is formed for each individual optical member, a chemical filter is installed at the outside air inlet for controlling the outside air temperature and taking it into the chamber room where the exposure apparatus is installed. It also shows a configuration in which clean air is introduced or circulated in the chamber room.

In FIG. 1, the air taken from the outside air and adjusted to a predetermined temperature by a temperature control unit (not shown) is introduced into the blower 47 via the circulation pipe 46 and the filter unit 35A, and the pressure is raised by the blower 47. Done, H
It is blown out from the EPA filter 48 into a chamber room 50 surrounded by a chamber wall 49. Most of the clean air is returned to the temperature control unit through a predetermined circulation path in the chamber room 50, but part of the clean air is blown to the outside from the gap in the side wall of the chamber room 50. In this case, the filter unit 35A is a large capacity filter unit having the same function as the filter unit 35.

Thus, the filter units 35A and H
By introducing or circulating the clean air from which impurities have been removed by the EPA filter 48 into the chamber room 50, not only the optical members in the exposure apparatus are prevented from being contaminated, but also harmful ions and amine gas in the chamber room 50 are prevented. Since the amount of harmful gas is reduced, the reaction between the amine gas and the chemically amplified resist exposed portion on the wafer during the exposure operation is also reduced. This makes it possible to prevent the surface of the chemically amplified resist from becoming insoluble and the T-top phenomenon of the pattern.

Next, the chemical filter 42 as the impurity removing filter in FIG. 2 will be described in detail.
As a chemical filter for removing ions, there are ion exchange resins, ion exchange fibers and the like, but ion exchange fibers are suitable for gas treatment because they have a large surface area and reaction rate and are easy to mold. Ion exchange fibers are made, for example, from polypropylene fibers by radiation graft polymerization. There are two types of ion-exchange fibers, acidic cation-exchange fibers and basic anion-exchange fibers, which are used properly depending on the polarity of the target ion. In this embodiment, both filters are used, and positive ions such as ammonium ions (NH ₄ ⁺ ) and amines and basic gases are the former, and negative ions such as sulfate ions (SO ₄ ²⁻ ) or nitric oxide (NO _x ). Ions and acid gases can be adsorbed by the latter. For example, ammonium ion (NH ₄ ⁺ ) can be adsorbed by about 90% or more even at a low concentration by a neutralization reaction with a strongly acidic cation exchange fiber. Further, the negative ions are adsorbed by the neutralization reaction with the basic anion exchange fiber. These ion exchange resins are intended only for removing ionic impurities, and are not suitable for removing organic substances.

The activated carbon filter is basically effective for most impurities, but generally, the organic compound having a large molecular size, a large intermolecular force, a low solubility in water, and a polarity. The lower the value, the higher the adsorptivity. Therefore, it is effective for removing organic substances that cannot be completely removed by the ion exchange resin, for example, organic silanes such as trimethylsilanol and HMDS (hexamethyldisiloxane). When selecting an activated carbon filter, it is necessary to consider the surface area and average pore size of the activated carbon, as well as the shape and dust generation of the activated carbon.

In the present embodiment, considering the pressure loss of the apparatus, a sheet in which urethane fiber is impregnated with activated carbon, a sheet obtained by processing activated carbon fiber into a sheet shape, or a honeycomb-shaped activated carbon is preferable. Activated carbon impregnated with an acidic substance or a weak alkaline substance is also effective for removing ionic impurities. Zeolite is also effective for removing most impurities, like activated carbon, and it is necessary to select the pore size according to the size of the impurities to be removed.

Although such a filter alone can remove impurities to some extent with activated carbon or zeolite, such adsorbents are optimized by optimizing the molecular size of the impurities to determine the pore size, and therefore the ion exchange resin is used. In some cases, it may be more efficient to divide the removal of ionic substances with a filter and to remove the organic substances with an activated carbon filter or a zeolite filter to form a composite.

Although the projection type exposure apparatus is used as the exposure apparatus in the present embodiment, the filter for the exposure apparatus of the present embodiment is effectively used for the illumination optical apparatus for the projection type exposure apparatus. In addition, it is effective as a filter for an illumination optical device of a proximity type or a contact type exposure apparatus. Furthermore, it is also effective in all optical devices using ultraviolet rays.

As described above, the present invention is not limited to the above-described embodiments, and various configurations can be adopted without departing from the gist of the present invention.

[0065]

According to the projection exposure apparatus of the present invention, a predetermined ion such as ammonium ion (NH ₄ ⁺ ) or silicon oxide ion (SiO ₄ ²⁻ ), which causes white turbidity of an optical member, or its source is generated. Means to filter organic gas (chemical filter)
The source of contamination in the device is significantly reduced as it is removed by. By introducing only this filtered air into the exposure device and preventing the inflow of other air as much as possible, it is possible to prevent new inflow of harmful ions or harmful organic gas and delay the progress of fogging of the optical member. it can. Further, when a basic gas such as ammonia is removed, there is an advantage that the T-top phenomenon is suppressed when using the chemically amplified resist. Further, an adsorbent (gas sensor) for adsorbing a predetermined harmful impurity is provided on the upstream side of the filter means, and the life of the filter means is determined by measuring the amount of the impurities adhering to the adsorbent, so that the downstream of the filter. Even if the concentration of impurities in the inflowing air is low compared to the case of measuring on the side, the life of the filter means can be easily determined.

Further, by accurately determining the service life of the filter means, it is possible to highly effectively reduce fogging even in an environment where the degree of chemical cleanliness is low. Also,
Since it is not necessary to discard the filter means that can still be used in terms of safety factor, the running cost can be reduced. Further, depending on the selection of the filter means, it becomes possible to deal with other trace gases which may become a problem in the future.

Further, when the life of the filter means is judged from the resonance frequency of the adsorbent, no troublesome work such as collecting impurities adhering to the adsorbent is required, and the adhering of impurities is always or when necessary. It is possible to measure the variation amount of the frequency due to, and the lifetime can be automatically determined by the lifetime determining means based on the variation amount of the frequency. Furthermore, if a filter for removing dust is installed at the gas outlet from the filter means, fine dust components such as metal powder that adversely affect the optical member are removed, and also dust components generated from the filter means itself are removed. To be done.

Further, if the filter means of the present invention is installed at the air inlet of the chamber room of the exposure apparatus used for manufacturing semiconductors and the like, harmful ions or harmful gases such as amine gas in the chamber are reduced, so that the exposure operation is performed. At this time, the reaction between the amine gas on the wafer and the chemically amplified resist exposed portion is also reduced. As a result, it is possible to prevent the surface insolubility and the pattern T-top phenomenon in the chemically amplified positive resist.

[Brief description of drawings]

FIG. 1 is a configuration diagram showing a main part of an exposure apparatus to which an embodiment of a filter for the exposure apparatus according to the present invention is applied.

2 is a configuration diagram including a partial cross-sectional view showing an inside of a filter unit 35 and a signal processing system of FIG.

FIG. 3A is a diagram showing ammonia concentrations before and after (upstream side, downstream side) a chemical filter using activated carbon or the like,
(B) is a diagram showing the ammonia concentration before and after a chemical filter using ion-exchange fibers or the like.

FIG. 4 is a configuration diagram including an end view along a partial cross section showing a conventional illumination optical device for an exposure apparatus.

FIG. 5 is a diagram showing an emission spectrum distribution of an ultra-high pressure mercury lamp.

FIG. 6A is a diagram showing a reflectance characteristic of a conventional aluminum reflecting mirror, and FIG. 6B is a diagram showing a reflectance characteristic of a conventional representative dielectric multilayer film reflecting mirror.

FIG. 7 is a diagram showing the transmission characteristics of a conventional bandpass filter plate.

[Explanation of symbols]

1 Mercury lamp 2 elliptical mirror 3 Mirrors for bending the optical path 4 Input lens 5 bandpass filter 6 fly eye lens 35, 35A filter unit 37 Air outlet 39 duct 42 Chemical filter 43 gas sensor 43a Crystal unit 43b Adsorption membrane 44,48 HEPA filter 50 chamber room 54 Frequency measurement section 57 Operation unit 58 Judgment section

─────────────────────────────────────────────────── ─── Continuation of the front page (58) Fields surveyed (Int.Cl. ⁷ , DB name) H01L 21/027 H01L 21/02 F24F ⁷ /04-7/06

Claims (8)

(57) [Claims] 1. A filter device for a gas that flows in a predetermined direction around an exposure apparatus, wherein a predetermined substance selected from the group of substances consisting of ammonia, amine compounds, sulfur oxides, and silicon oxides is discharged from the gas. Filter means for removing, adsorbent for adsorbing the predetermined substance on the gas inflow side with respect to the filter means, measuring means for measuring the amount of the substance adsorbed by the adsorbent, and measurement result by the measuring means Determining means for determining the life of the filter means based on the above. 2. The measuring means comprises frequency detecting means for detecting a resonance frequency of the adsorbent, and computing means for obtaining the amount of the substance adsorbed on the adsorbent from the detection result of the frequency detecting means. A filter device for an exposure apparatus according to claim 1. 3. The filter device for an exposure apparatus according to claim 1, wherein a filter for removing dust is installed at an outlet of the gas from the filter means. 4. The filter means and the adsorbent are installed at an inlet of the outside air to the chamber when the exposure apparatus is installed in a chamber isolated from the outside air. 2. A filter device for an exposure apparatus according to 2 or 3. Wherein said ammonia include ammonium ions, the amine compound comprises an amine ion, said sulfur oxides comprising sulfite ions, and sulfate ions, the silicon oxide comprising silicon oxide ion The filter device for an exposure apparatus according to any one of claims 1 to 4. Wherein said adsorbent, a filter device for exposure apparatus according to any one of claims 1 5, characterized in that the gas sensor. Wherein said gas sensor, a crystal oscillator, a filter device for exposure according to claim 6, characterized in that the gas sensor and a suction film provided on the surface of the crystal oscillator. 8. A projection exposure apparatus for manufacturing a semiconductor device, comprising the filter for the exposure apparatus according to claim 1. Description: