JP2005257697A - Manufacturing device of semiconductor device and semiconductor device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To control a manufacturing process of a semiconductor device based on a determined dimension by including a workpiece dimension measuring process for nondestructively determining the transverse dimension of a pattern formed on a wafer in no contact at high speed by means of ellipsometry. <P>SOLUTION: A polarized state parameter is determined for a calibration pattern with a confirmed dimension formed on the wafer to form a database. The database is retrieved on the basis of the polarization parameter of a measuring sample to determine the dimension thereof. <P>COPYRIGHT: (C)2005,JPO&NCIPI

Description

  The present invention generally relates to semiconductor device manufacturing, and more particularly, to a semiconductor device manufacturing method including a workpiece dimension measuring step and a semiconductor device quality control method including the workpiece dimension measuring step.

  In a semiconductor device manufacturing line, it is required to measure a dimension of a workpiece formed on a wafer, for example, a gate pattern at high speed in a non-contact and non-destructive manner. It is also desirable to adjust the manufacturing parameters by quickly feeding back the measurement results to the manufacturing line. In particular, since the gate length dimension affects the threshold characteristics of a semiconductor device, precise control thereof is required.

Conventionally, a wafer on which a structure such as a gate pattern is formed is scanned one by one with a scanning electron microscope (SEM), and its dimensions are measured. There is also a method in which a bridge circuit is simultaneously formed on a wafer on which a gate pattern is formed, and a dimension of the gate pattern is measured by passing a current through the bridge circuit.
Blayo, N., et al., "Ultraviolet-visible ellipsometry for process controlduring the etching of submicrometer features," J. Opt. Soc. Am. A, vol.12, no.3,1995, pp.591-599.

  However, in the method using the SEM, since it is necessary to transfer the wafers flowing through the production line one by one to the vacuum chamber of the SEM, it is impossible to perform 100% inspection, and even when sampling inspection is performed, The problem that the manufacturing throughput of the semiconductor device decreases is unavoidable. In addition, when a processing dimension such as a gate length becomes about 0.1 μm as in a recent highly integrated semiconductor device, a measurement error of about 10 nm occurs due to the influence of the electron beam diameter of the SEM, and the accuracy of the measurement result And problems in terms of reproducibility.

  On the other hand, the resistance measurement method eliminates the problems of accuracy and reproducibility of measurement results that occur when using SEM, but the results cannot be obtained unless the wafer manufacturing process is completed. There arises a problem that the process cannot be feedback controlled.

On the other hand, conventionally, ellipsometry technology has been used to measure the thickness of semiconductor films and insulating films in the manufacturing process of semiconductor devices. For example, ellipsometry is also used for etching control when forming a line and space pattern in an etching process. FIGS. 22A and 22B show the configuration of an ellipsometer used in conventional ellipsometry. Among these, the device shown in FIG. 22A is a device of a type called a rotation quenching type, and the device shown in FIG. 22B is a device of a type called a quenching type.

Referring to FIG. 22A, light emitted from the light source 1 is converted into linearly polarized light having a predetermined polarization plane by a polarizer 2, and is incident on a sample 3 on which a film to be measured is formed. The light reflected by the sample 3 generally changes its polarization state into elliptically polarized light characterized by the rotation angle φ of the polarization plane and the elliptic flatness a _min / a _max = tan ψ as shown in FIG. Is detected by the detector 5. In this apparatus, the polarization state is obtained by measuring the intensity of incident light with the detector 5 while rotating the rotating analyzer 4. In the apparatus having this configuration, a quarter wavelength plate 4 a may be inserted between the rotating analyzer 4 and the sample 3.

  In the apparatus of FIG. 22B, a rotating quarter wavelength plate 4b is inserted between the rotating analyzer 4 and the sample 3, and the elliptically polarized light reflected by the sample 3 is once converted into linearly polarized light. In this state, the rotation analyzer 4 is rotated to obtain an extinction angle at which the light reaching the detector 5 is quenched.

  As described above, these ellipsometers are widely used to measure the film thickness in the manufacturing process of a semiconductor device, but the polarization state of the light beam that has passed through the structure formed on the sample 3 Is considered to contain information on the lateral dimensions of such a structure. However, conventionally, there has not been proposed a concept of measuring the dimension in the horizontal direction of a workpiece such as a line and space pattern formed on the sample 3 using an ellipsometer.

  SUMMARY OF THE INVENTION Accordingly, it is a general object of the present invention to provide a novel and useful workpiece dimension measuring method and a semiconductor device manufacturing method using such a workpiece dimension measuring method, which solve the above problems.

  A more specific object of the present invention is to provide a method for measuring a dimension of a workpiece, which uses an ellipsometry technique to measure a dimension of a structure formed on a substrate wafer with high accuracy, quickly and non-destructively, Another object of the present invention is to provide a method of manufacturing a semiconductor device using such a method for measuring a workpiece size.

The present invention solves the above problems.
As described in claim 1,
In a semiconductor device manufacturing method for manufacturing a semiconductor device based on dimensions of a structure formed on a substrate,
An incident light beam polarized at a predetermined angle with respect to the substrate surface is incident on the structure, and a polarization state of an outgoing light beam emitted from the structure on which the incident light beam is incident is measured. A semiconductor device is manufactured by setting manufacturing parameters of an etching process based on information on a dimension of the structure obtained based on an evaluation obtained from the direction parallel to the substrate surface. Or as described in claim 2,
The information on the dimensions is fed back to the etching process of the semiconductor device, and the manufacturing parameters are adjusted, or by the method for manufacturing a semiconductor device according to claim 1, or as described in claim 3.
3. The method of manufacturing a semiconductor device according to claim 1, wherein the evaluation is performed by referring to a database including a relationship between the polarization state and information on dimensions of the structure. Or as described in claim 4,
The polarization state is expressed by rotation and flatness of a polarization plane of an outgoing light beam after passing through the structure, and the measurement is performed by an ellipsometer. According to the method for manufacturing a semiconductor device according to any one of the above, or as described in claim 5,
5. The semiconductor device manufacturing method according to claim 1, wherein the incident is performed by changing an incident angle of the polarized incident light beam. 6. As described in
The measurement uses reflected light of the incident light beam, diffracted light of the incident light beam diffracted from the structure, or scattered light of the incident light beam scattered by the structure as the outgoing light beam. The semiconductor device manufacturing method according to any one of claims 1 to 5, or as described in claim 7,
The method for manufacturing a semiconductor device according to claim 1, wherein the evaluation obtains information or a width related to a cross-sectional shape of the structure based on the polarization state and a film thickness. Or as described in claim 8
The incident interrupts the incident light, and the measurement measures the polarization state of the intermittent outgoing light beam formed corresponding to the intermittent incident light beam in the irradiation state and the cutoff state. The semiconductor device manufacturing method according to any one of claims 1 to 7, or as described in claim 9,
A decomposing step of decomposing the incident light beam into a plurality of light beams; a measuring step of calculating and integrating the polarization components and intensities of the plurality of outgoing light beams emitted from the structure and corresponding to the plurality of light beams;
Obtaining information about dimensions of the structure in a direction parallel to the substrate surface based on the integrated polarization component and intensity; and
The method for manufacturing a semiconductor device according to any one of claims 1 to 8, or as described in claim 10,
It solves by the semiconductor device manufactured by the manufacturing method of the semiconductor device as described in any one of Claims 1-9.

  According to the features of the present invention, in the manufacturing process of a semiconductor device, the dimension of a pattern formed on a wafer can be measured at high speed in a non-contact and non-destructive manner, and the measurement result is fed back to the manufacturing process. It becomes possible to do.

  Further, according to the feature of the present invention, it is possible to easily obtain a pattern dimension without performing a complicated theoretical calculation by referring to a database constructed based on an experiment based on the obtained elliptical polarization parameter. Is possible.

  According to the feature of the present invention, the polarization state can be easily obtained by using an ellipsometer.

  Further, according to the feature of the present invention, it is possible to obtain not only the pattern dimension but also the angle formed by the pattern side wall with respect to the substrate surface by performing measurement while changing the incident angle of the incident light beam. .

  Further, according to the feature of the present invention, by using strong reflected light reflected from the structure, the polarization state of the reflected light beam can be stably obtained with good reproducibility.

  Further, according to the feature of the present invention, by using the diffracted light diffracted from the structure, it is possible to minimize the influence of the underlayer on the measurement result even if the pattern is relatively rough.

  Further, according to the feature of the present invention, by using the scattered light scattered from the structure, it is possible to minimize the influence of the underlayer on the measurement result even for a relatively rough pattern.

  According to the feature of the present invention, the width or cross-sectional shape of the structure can be obtained by non-destructive and non-contact.

  In addition, according to the feature of the present invention, the incident light beam is intermittently measured, and the polarization state is measured for each of the irradiation state and the blocking state, so that even when weak outgoing light such as scattered light is used, high-precision measurement is possible. Is possible.

[principle]
FIG. 1 illustrates the principle of the present invention.

  Referring to FIG. 1, in the present invention, the width W in the lateral direction of the pattern 11 is obtained by detecting the polarization state of the light beam incident on the pattern 11 formed on the substrate 10 using ellipsometry. .

The incident light beam is incident at an incident angle θ on the pattern 11, the oscillating electric field components E _p and E _s which are perpendicular to each other, to form a linear polarization of the polarization angle phi. The incident light beam passes from one side of the pattern 11 to the other side. At this time, the refractive index and the reflectance at the boundary with the substrate 10 or the boundary with the air are different for each vibration component, and the inside of the pattern. Therefore, the oscillation components E _p ′ and E _s ′ after passing through the pattern form elliptically polarized light.

  FIG. 2 shows an enlarged detail of a part of FIG.

  Referring to FIG. 2, the pattern 11 includes pattern elements 11a and 11b. Light rays a to c constituting an incident light beam in a linear polarization state are incident on the pattern element 11a, and are refracted and reflected to be elliptically polarized light. Then exit.

For example, when a monochromatic linearly polarized beam having an angle of 45 ° is incident on the pattern shown in FIG. 2 at an angle Θ, the difference in reflectance between the s-polarized component and the p-polarized component at the boundary between the surface and the ground, and in the pattern The reflected light becomes elliptically polarized light due to the phase difference that occurs when passing through. In such a case, the complex reflection coefficient ratio R _p / R _s determined by the substrate refractive index, the pattern refractive index, the pattern shape, and the incident angle can be obtained by an ellipsometer, and the complex reflection coefficient ratio R _p / R _s , The relationship between Ψ and Δ obtained by the ellipsometer is given by the following equation.

ただし、パラメータΨは図２３に説明したように、ｔａｎΨ＝ρ_ｐ ／ρ_ｓ で与えられ、またパラメータΔは位相差を表し、エリプソメトリにより測定される量である。また、パラメータΨおよびΔは、先に説明した偏光角φおよび楕円の偏平率と対応関係により結ばれている。さらに、式（１）において、総和Σはパターン１１を通過する全ての光線についてなされる。

However, as described in FIG. 23, the parameter Ψ is given by tan Ψ = ρ _p / ρ _s , and the parameter Δ represents a phase difference and is an amount measured by ellipsometry. Further, the parameters Ψ and Δ are connected by the correspondence relationship with the polarization angle φ and the ellipticity described above. Further, in equation (1), the sum Σ is made for all rays that pass through the pattern 11.

In other words, when the refractive index (dielectric constant) of the substrate 10 and the refractive index (dielectric constant) of the pattern 11 are defined and the incident angle of the incident light beam is defined, the components E _p ′ and E _s are determined by ellipsometry. The parameters Ψ and Δ are obtained by obtaining the elliptical polarization state of the outgoing light beam consisting of '.

  As can be seen from FIG. 2, when the outgoing light beam passes through the pattern 11, information on the dimensions of the pattern elements 11 a and 11 b, especially the lateral dimension W, is taken in as a change in phase. The dimension W can be estimated based on Ψ and Δ obtained from the polarization state of the beam. In one embodiment of the present invention, parameters Ψ and Δ are obtained as a database in advance for various patterns whose shapes and dimensions have been grasped by, for example, SEM, and the parameters Ψ and Δ actually obtained are stored in the database. , Δ, the dimension W of the pattern is obtained.

Hereinafter, preferred embodiments of the present invention will be described with reference to the drawings.
[First embodiment]
FIG. 3 shows a configuration of a dimension measuring apparatus for a pattern formed on a substrate using ellipsometry according to a first embodiment of the present invention.

  Referring to FIG. 3, a substrate 22 on which a pattern is formed is placed on a stage 21 and irradiated with a light beam having a wavelength of 6328 nm from a He—Ne laser 23. That is, the phase of the output light of the laser 23 is adjusted by the quarter wavelength plate 24, and further converted by the polarizer 25 into linearly polarized light having a predetermined polarization angle, for example, 45 °, and the substrate on the stage 21. 22 is irradiated at a predetermined incident angle, for example, an incident angle of 70 °. Alternatively, the output light of the laser 23 may be passed through the polarizer 25 first, and then the phase may be adjusted by the quarter wavelength plate 24. In a typical case, the light beam has a beam diameter of 50 to 150 μm on the pattern.

  The light beam that passes through the pattern on the substrate 22 and is further reflected has an elliptical polarization state, is converted into linearly polarized light by the rotating analyzer 26, and then enters the photocell 27 and is detected. At that time, detection of incident light by the photocell is executed while rotating the rotation analyzer 26.

  The photocell 27 forms an output signal corresponding to the detected intensity of incident light, and sends it to the processing device 30 in the form of a digital signal via the amplifier 28 and the A / D converter 29. The processing device 30 obtains the polarization state of the light beam incident on the photocell 27 from the supplied digital signal, and outputs the parameters Ψ and Δ therefrom.

  4A and 4B are a perspective view and a sectional view, respectively, showing an example of a resist pattern 22a formed on the substrate 22. FIG. As these patterns, an appropriate pattern on the substrate may be used, but when there is no pattern suitable for measurement, a line-and-space pattern for measurement, an appropriate place on the substrate such as a scribe line, etc. What is necessary is just to form.

Referring to FIG. 4 (A), (B) , the substrate 22 is made of a polysilicon film 22 ₂ deposited to a thickness of 153nm on the Si substrate 22 _1, the resist pattern 22a is formed on the polysilicon film 22 ₂ Is done. As the resist pattern 22a, for example, a CMS resist (trade name) for electron beam exposure manufactured by Toyo Soda Co., Ltd. can be used.

  The resist pattern 22a is a line and space pattern in which a plurality of pattern elements 22b are repeated in parallel at a predetermined pattern pitch, for example, a pitch of 0.3 μm, and is formed in a range of 1 mm × 1 mm on the substrate 22, for example. Each pattern element 22b is formed to a height of 150 nm by patterning a resist layer having a thickness of 150 nm. In such a resist pattern, even if the pattern pitch is accurately determined, the width of each pattern element 22b may change with respect to a predetermined value, which is the line-and-space pattern formed on the substrate 22. Causes the width to go crazy. For example, when such a line and space pattern constitutes the gate electrode of a MOS transistor, such a deviation leads to a change in the threshold voltage of the formed transistor.

  FIG. 5 shows the results of experimentally determining the relationship between Ψ and Δ obtained when the apparatus of FIG. 3 is applied to such a line and space pattern for various pattern widths W. However, in the experiment, the substrate 22 shown in FIGS. 4A and 4B was used as the substrate, and Ψ and Δ were obtained by changing the width W of the pattern element 22b from 100 nm to 190 nm. At that time, the value of the width W is a value confirmed by observation with an SEM. The values of Ψ and Δ were measured five times for each pattern width W.

  As can be seen from FIG. 5, one width W corresponds to one (Ψ, Δ) combination, and therefore, Ψ, Δ is measured for a given substrate and collated with the relationship of FIG. The width W of the line and space pattern formed on the substrate is required. The relationship shown in FIG. 5 is stored in the database in the processing device 30 of FIG. 3, and the processing device 30 performs such collation on the database to obtain the value of the width W.

FIG. 6 shows a plan view of the apparatus of FIG. 3. Referring to FIG. 6, a positioning member 21 </ b> A that abuts with the orientation flat 22 </ b> A of the wafer constituting the substrate 22 is formed on the stage 21. Positioning pins 21B that abut the side surfaces are formed. At this time, the line and space pattern 22a is formed on the substrate 22 so as to be orthogonal to the optical path of the light beam from the light source 23 in a state where the substrate 22 abuts the positioning members 21A and 21B. By forming the line and space pattern 22a in such an orientation, the light beam efficiently captures the information of the line and space pattern 22a in its phase. In other words, the ellipsometer of FIG. 3 has the maximum sensitivity to the width W of the line and space pattern 22a by placing the substrate 22 on the stage 21 in the orientation shown in FIG.

  The wafer orientation may be adjusted by, for example, forming a notch on the wafer and engaging the notch with a positioning pin. Furthermore, the orientation of the wafer may be adjusted using an orientation detection mechanism using LEDs.

In the configuration of FIG. 6, the wafer 22 is arranged so that the pattern elements constituting the pattern 22a are orthogonal to the optical path of the light beam. This is because each pattern element is a light beam in the plan view of FIG. You may arrange | position so that it may extend in parallel with this optical path. In this case, information on the background of the pattern 22a is obtained.

[Second Embodiment]
FIG. 7 shows a second embodiment of the present invention. However, in FIG. 7, the same reference numerals are given to the parts described above, and the description thereof is omitted.

  Referring to FIG. 7, in this embodiment, the pattern elements 22b are formed relatively sparsely on the substrate 22, so that a substantial part of the light beam emitted from the light source 23 is information on the substrate 22 itself, for example, the thickness. I will pick up the information. As a result, there is a possibility that the pattern width W obtained from the obtained Ψ and Δ is shifted from the actual one.

  In the present embodiment, the light beam emitted from the light source 23 made of a He—Ne laser is a coherent light, and the Bragg diffracted high is obtained by the line and space pattern 22a formed on the substrate 22 at a predetermined pitch. The pattern width W is obtained using the next diffracted light. In this case, a photocell is provided at the diffraction position of each diffracted light, and Ψ and Δ are obtained from the polarization state. Further, similarly to the previous embodiment, the pattern width W is obtained with reference to the database formed for each diffracted light. Further, the most probable value of the pattern width W is obtained by averaging or weighted averaging the values of W obtained from the individual diffracted lights.

  FIG. 8 uses the configuration shown in FIG. 7 to reflect the values of the polarization state parameters Ψ and Δ obtained for a resist line and space pattern with a pitch of 5 μm formed on a Si substrate with a thickness of 150 nm for detecting the polarization state. A case where light is used and a case where first-order diffracted light is used will be described. However, in FIG. 7, the incident angle of the incident light is set to 77 °, and as a result, the reflected light is generated in the direction of the reflection angle of 77 °. On the other hand, the first-order diffracted light is generated in the direction of the diffraction angle of 58 ° in the example of FIG.

As can be seen from FIG. 8, the values of the parameters Ψ and Δ are changed by changing the width W of the line and space pattern on the Si substrate in the range of 110 to 210 nm. While the value of Δ changes in the range of 10 ° to 10 ° and 30 ° to 60 °, in the first-order diffracted light, the value of Ψ changes in the range of 60 ° to 85 ° and the value of Δ changes in the range of −100 to −110 °. I understand. As can be seen from FIG. 8, when the first-order diffracted light is used, the change in the phase difference Δ is small and the width W roughly corresponds to the value of Ψ.

[Third embodiment]
FIG. 9 schematically shows a semiconductor device manufacturing line according to a third embodiment of the present invention, which incorporates the workpiece size measuring apparatus of FIG.

  Referring to FIG. 9, the semiconductor device manufacturing line includes a wafer process 101 including an exposure process and an etching process, and a process control unit 102 that controls the wafer process 101, and further processes a wafer processed in the wafer process 101. An ellipsometer 103 for inspection is provided. The ellipsometry unit 103 includes an ellipsometer having the configuration shown in FIG. 3, and irradiates the wafer to be processed in the wafer process 101 with a polarized light beam. The measured values of parameters Ψ and Δ are obtained from the polarization state of the obtained reflected light beam. Ask for.

  The measured parameters Ψ and Δ of the reflected light beam obtained by the ellipsometry unit 103 are supplied to the process control unit 102, and the process control unit 102 presets the parameters Ψ and Δ supplied from the initial condition setting unit 104. The wafer process 101 changes the exposure conditions such as exposure dose and exposure time or the etching conditions such as RF power based on the result.

  The initial condition setting unit 104 supplies shape parameters such as the film thickness and pattern width of the line and space pattern formed in the wafer process 101, or data such as the film thickness of the layer formed under the line and space pattern. Then, based on the shape parameter, the desired values of Δ and Ψ are calculated with reference to the database 104a holding the relationship between Δ and Ψ as shown in FIG. As described above, the process control unit 102 compares the combination of the parameters Δ and Ψ supplied from the initial condition setting unit 104 with the combination of the parameters Δ and Ψ obtained by the ellipsometer 103 and compares the difference between them. The wafer process 101 is controlled so as to be minimized.

  Further, in the apparatus of FIG. 9, a set of parameters Δ and Ψ constituting the database 104a is obtained by inspecting a wafer whose pattern dimension is confirmed by SEM or the like by the ellipsometry unit 103.

  In the apparatus of FIG. 9, the theoretical calculation unit 104b calculates the theoretical values of desired Δ and Ψ from the given initial parameters and supplies them to the process control unit via the initial condition setting unit 104. You may do it.

  The calculation of the theoretical value is roughly performed as follows.

A light beam is incident at an incident angle Θ ₁ in an environment having a refractive index of n _{1 on} a sample periodically formed on a substrate having a refractive index of n ₃ and including a line and space pattern having a refractive index of n ₂ . Then, as shown in FIG. 10, the light beam is repeatedly refracted and reflected in accordance with Snell's law within the sample, and is emitted from the sample at the same angle as the incident angle Θ ₁ . Therefore, the incident light is converted into m incident light beams I ₍₁₎ , I ₍₂₎ ,. . . I _(m) is resolved, and the optical path of each incident ray is obtained according to Snell's law. At that time, the incident light beam is decomposed into a p-polarized light component and an s-polarized light component, and the attenuation effect accompanying reflection and refraction is obtained for each component. More specifically, when the incident light is reflected, the intensity I ₀ of the p-polarized component and the s-polarized component of the incident light is multiplied by Fresnel's amplitude reflectances r _p and r _s for each reflection to accompany the reflection. Determine the attenuation of the light amplitude. Further, if the incident light is refracted multiplies the intensity I ₀ of the p-polarized light component and s-polarized component of the incident light, the amplitude refractive index t _p each Fresnel every time _refraction, the t _s, an optical amplitude due to refraction Determine the attenuation of. Further, when the light beam passes through an opaque medium, the attenuation effect can be _obtained by multiplying by the amplitude transmittance tk. Further to this, a phase delay δ accompanying the optical path length is added to obtain final intensities I _p (n) ′ and I _s (n) ′ of the respective polarization components. However, the amplitude reflectance _r p, _{r s,} amplitude refractive index _t p, _{t s,} amplitude transmittance _{t k} and the phase delay term δ is
r _p = tan (Θ ₁ −Θ ₂ ) / tan (Θ ₁ + Θ ₂ )
r _s = −sin (Θ ₁ −Θ ₂ ) / sin (Θ ₁ + Θ ₂ )
t _p = 2sinΘ ₂ cosΘ ₁ / sin (Θ ₁ + Θ ₂ ) cos (Θ ₁ −Θ ₂ )
_{t s = 2sinΘ 2 cosΘ 1 /} sin (Θ 1 + Θ 2)
t _k = exp (−2πkd / λ)
δ = exp (−i2πnd / λ)
Given in. Where λ is the wavelength of the incident light beam and d is the optical path length in the sample.

For example, in the example of FIG. 11, incident light beams (1) to (3) having an intensity I ₀ are reflected and refracted by a line and space pattern, and then polarized components I _p (1) ′ to I represented by the following expressions: _It is emitted as outgoing light having _p (3) ′, I _s (1) ′ to I _s (3) ′.
I _p (1) ′ = t _p1 t _p2 r _p3 t _p4 t _p5 × exp (−2πk ₂ d ₁ / λ)
exp (−i2πn ₂ d ₁ / λ) · I ₀
I _s (1) ′ = t _s1 t _s2 r _s3 t _s4 t _s5 × exp (−2πk ₂ d ₁ / λ)
exp (−i2πn ₂ d ₁ / λ) · I ₀
I _p (2) ′ = t _p1 r _p2 t _p3 × exp (−2πk ₂ d ₂ / λ)
exp (−i2πn ₂ d ₂ / λ) · I ₀
I _s (2) ′ = t _s1 r _rs2 t _s3 × exp (−2πk ₂ d ₂ / λ)
exp (−i2πn ₂ d ₂ / λ) · I ₀
I _p (3) ′ = r _p1 · I _o

I _s (3) ′ = r _s1 · I _o

Here, n ₂ and k ₂ represent the refractive index and absorption coefficient in the line and pattern, respectively, and d _{1 and} d ₂ represent the optical path lengths of the light beams (1) and (2) in the line and space, respectively. . Subscripts 1 to 5 indicate the order of the reflection point and the refraction point of each light ray counted from the incident side. See FIG.

Thus, when the intensities I (1) ′ to I (n) ′ of the respective rays are obtained, the complex reflection coefficient ratio (ρ = R _p / Rs) is
ρ = R _p / R _s = ΣI _p (n) ′ / ΣI _s (n) ′ = tan Ψexp (iΔ)
From which the parameters Ψ and Δ are
^{Ψ = tan - 1 (| ρ} |)
Δ = arg (ρ)
Is required.

  In the configuration of FIG. 9, the calculation unit 104 b performs the above calculation and supplies the obtained parameters Ψ and Δ to the process control unit via the initial condition setting unit 104.

  FIG. 12 is a flowchart showing an outline of the operation executed by the calculation unit 104b.

  Referring to FIG. 12, first, in step 1, structure conditions such as pattern film thickness, refractive index, absorptivity, underlayer film thickness, refractive index, absorptivity, and pattern pitch are input. The light beam is broken down into individual rays according to Snell's law, and an optical line is determined for each ray (i = 1-n). Further, in step 3, the intensities Ip (i) 'and Is (i)' of the polarization components p and s are determined for each light ray i.

  Next, in step 4, the integrated intensities ΣIp (n) ′ and ΣIs (n) ′ of the polarization components p and s are obtained, and the complex reflection coefficient ratio ρ is obtained in step 5 from the obtained integrated intensities. Further, in step 6, the parameters Ψ and Δ are obtained from the complex reflection coefficient ratio ρ, and in step 7, the pattern width W is obtained by referring to the database.

13, in a state where the resist pattern 22a shown in the mask in the embodiment, a polysilicon layer 22 _{and second} base is etched by RIE to remove the resist pattern 22a in FIG. 4 (A), (B) , FIG. 4 shows a relationship between parameters Ψ and Δ obtained by the measuring apparatus of FIG. However, in this case, the angle of incidence was set at 55 °, the underlying layer 22 ₁ of the polysilicon layer 22 ₂ is in the SiO ₂ instead Si. By accumulating such a relationship in the database 104a, it is possible to obtain pattern widths for various patterns.

Further, FIG. 14 shows the change of the parameters Ψ and Δ while changing the etching time using the measuring apparatus of FIG. 3 when the pitch of the mask pattern is 150 nm in the embodiment of FIGS. 4 (A) and (B). This is an example of monitoring. However, the illustrated example corresponds to the case where 10% and 30% overetching is further performed with respect to the etching end point. The polarization state is measured with the resist pattern 22a removed. In general, it is known from SEM observation that approximately 20 nm of side etching occurs when 30% overetching is performed. However, the results shown in FIG. It can be seen that side etching can be detected. In other words, the workpiece dimension measuring method using ellipsometry according to the present invention indicates that the deviation of the machining dimension can be detected with an accuracy of 10 nm or less.

[Fourth embodiment]
FIG. 15 shows the principle of the fourth embodiment of the present invention.

  Referring to FIG. 15, in this embodiment, the polarization state of the scattered light generated by the incident light beam being scattered by the pattern 22b on the substrate 22 instead of the reflected light or diffracted light of the incident light beam is obtained. The parameters Ψ and Δ are obtained by

  FIG. 16 shows a configuration for executing the scattered light measurement shown in FIG. However, in FIG. 16, the parts described above are denoted by the same reference numerals, and the description thereof is omitted.

  In the measuring apparatus of FIG. 16, the coherent light emitted from the light source 23 passes through the ¼λ phase difference plate 23 and the polarizer 25, and then enters the wafer 22 on which the pattern 22a is formed via the beam chopper 25A. . The beam chopper 25a is composed of a rotating disk having a cut formed in a part thereof, and controls on / off of coherent light incident on the wafer 22 from the light source 23.

  Further, in the configuration of FIG. 16, a lens 31 and a beam splitter 32 cooperating with the lens 31 are formed while avoiding the reflected light or diffracted light formed by the pattern 22a on the wafer 22, and the beam splitter 32 is connected to the pattern 22a. The scattered light that is generated from the light and is focused by the lens 31 is decomposed into a p component and an s component and supplied to detectors 27A and 27B, respectively. At that time, a rotation control signal is supplied as a synchronizing signal SYNC from the beam chopper 25A to the detectors 27A and 27B, and the detectors 27A and 27B respectively correspond to one cycle of the synchronizing signal SYNC. The light intensity Ip, Is is measured. According to such a configuration, the intensity of the scattered light in a state where the light beam is irradiated onto the wafer 22 is compared with the intensity of the scattered light in a state where the light beam is blocked, thereby reducing the intensity of the weak scattered light. Ip and Is can be measured with high accuracy.

  The scattered light intensities Ip and Is obtained in this way are processed in the processing device 30, and the reflection coefficient ratio tan Ψ is obtained by executing the division Ip / Is on the intensities Ip and Is. From the above, the polarization parameter Ψ is obtained.

FIG. 17 shows the relationship between the reflection coefficient ratio tan Ψ and the pattern width W in the configuration using the scattered light of FIG. However, in the example of FIG. 17, a substrate in which a polysilicon line and space pattern having a thickness of 180 nm is formed on a SiO ₂ film having a thickness of 100 nm at a pitch of 300 nm is used, and each pattern width W is 120 to 180 nm. The range is varied. In the measurement, linearly polarized light having a polarization angle of 45 ° was used as incident light in the apparatus shown in FIG.

As can be seen from FIG. 17, the dependence of the reflection coefficient ratio tan Ψ on the pattern width is observed by using scattered light. That is, the pattern width W of the line and space pattern can be obtained by obtaining the reflection coefficient ratio tan Ψ from the scattered light.

[Fifth embodiment]
By the way, in the apparatus of FIG. 3, FIG. 7, FIG. 9 or FIG. 15, the incident angle of the light beam emitted from the light source 23 is not limited to 70 °, and other angles can be used. Further, Δ and Ψ are measured by changing the incident angle in various ways, and on the basis of this, by referring to a database including the incident angle as a parameter in addition to Ψ and Δ, not only the pattern width but also the pattern It is also possible to determine the inclination angle of the side wall with respect to the substrate surface.

FIG. 18 shows a reflection of a wafer in which a polysilicon film is formed to a thickness of 182 nm on a SiO ₂ film having a thickness of 102 nm, and a resist line and space pattern having a thickness of 150 nm is formed thereon at a pitch of 500 nm. The relationship between polarization parameters Ψ and Δ obtained from light is shown.

  Referring to FIG. 18, in the state before the polysilicon film is etched while the resist pattern is formed, the parameters Ψ, W for each value of the pattern width W (W = 175 nm, W = 150 nm, W = 125 nm). Δ is obtained as shown in the upper left curve in FIG. Among these, experiments at three points indicated by white circles, white triangles, and white squares are performed at each point corresponding to the value of W, and the convergence of the obtained data is very good at each point. However, the above data is obtained by setting the incident angle of the incident light beam to 70 °.

  Next, the polysilicon film is etched using the resist line and space pattern as a mask, and polarization parameters Ψ and Δ obtained for the obtained structure are indicated by black circles, black triangles, and black squares. However, as shown in the cross-sectional SEM photograph of FIG. 19A, the black triangle is when the overetching amount is 0%, and as for the black square, the overetching amount is 30% as shown in the cross-sectional SEM photograph of FIG. 19B. Further, the black circles indicate the case where the overetching amount is 80% as shown in the cross-sectional SEM photograph of FIG. 19C.

  As can be seen from FIG. 18, the curve when the overetching amount is 0%, the curve when the overetching amount is 30%, and the curve when the overetching amount is 80% are the pattern width W. In the case of 175 nm, they are different from each other except that the difference in cross-sectional shape of the polysilicon pattern shown in FIGS. 19A to 19C is detected from the combination of the polarization parameters Ψ and Δ. It shows that you can. That is, when the structural parameters such as the pattern width W and the film thickness or refractive index are known in advance by some method, the cross-sectional shape of the pattern can be estimated from the combination of the polarization parameters Ψ and Δ obtained by ellipsometry. become.

[Example 6]
FIG. 20 shows the relationship between the polarization parameters Ψ and Δ when the thickness of the polysilicon film is changed in the same line and space pattern formed on the SiO ₂ film. However, in FIG. 20, the curve indicated by a black circle represents a case where a polysilicon line and space pattern having a thickness of 178 nm is formed on a SiO ₂ film having a thickness of 100 nm, and the curve indicated by a white circle represents polysilicon. This is a case where the thickness of the line and space pattern is 163 nm. In any case, the pitch of the line and space pattern is changed to 300 nm and the pattern width W is changed in the range of 110 nm to 200 nm.

  As can be seen from FIG. 20, the curves defined by the parameters Ψ and Δ are different between the case where the thickness of the polysilicon pattern is 178 nm and the case where the thickness is 163 nm. For this reason, the thickness of the pattern is measured first. Thus, it becomes possible to select a specific curve and obtain Ψ and Δ for the selected curve.

  FIG. 21 shows a flowchart of the pattern dimension measuring method according to the sixth embodiment of the present invention including the film thickness measuring step.

  Referring to FIG. 21, first, in step 11, the film thickness of the line and space pattern is measured, and then in step 12, a curve or characteristic curve corresponding to the measured film thickness is selected. Further, measurement by ellipsometry is performed in step 13, and the pattern width W or its cross-sectional shape is obtained by using the obtained parameters Ψ and Δ and the selected characteristic curve.

  Furthermore, the present invention is not limited to the above-described embodiments, and various modifications and changes can be made within the scope of the present invention.

It is a figure explaining the principle of this invention. It is a figure which expands and shows the principal part of FIG. It is a figure which shows the structure by 1st Example of this invention. It is a figure which shows the line and space pattern to which the apparatus of FIG. 3 is applied. It is a figure explaining the example of the database used with the apparatus of FIG. It is a figure explaining the positioning of the wafer in the apparatus of FIG. It is a figure which shows 2nd Example of this invention. FIG. 8 is a diagram illustrating an example using reflected light and first-order diffracted light in the configuration of FIG. 7. It is a figure which shows 3rd Example of this invention which applied the apparatus of FIG. 3 to control of the manufacturing process of a semiconductor device. FIG. 10 is a diagram (part 1) illustrating a principle when parameters Ψ and Δ are obtained by calculation in the apparatus of FIG. 9; FIG. 10 is a diagram (No. 2) illustrating the principle when the parameters Ψ and Δ are obtained by calculation in the apparatus of FIG. 9. 10 is a flowchart showing processing in the apparatus of FIG. 9. FIG. 4 is a diagram showing the relationship between pattern widths and parameters Ψ and Δ when the present invention is applied to a line and space pattern formed by etching in the apparatus of FIG. 3. FIG. 4 is a diagram showing a relationship between parameters Ψ and Δ as etching progresses in the apparatus of FIG. 3. It is a figure explaining the 4th example of the present invention. It is a figure which shows the specific structure of the apparatus of FIG. It is a figure which shows the example of the characteristic curve obtained by the apparatus of FIG. It is a figure explaining the change of (DELTA) -Ψ relation by the cross-sectional shape of the pattern by 5th Example of this invention. (A)-(C) are figures which show various pattern cross-sectional shapes corresponding to FIG. It is a figure explaining the change of (DELTA) -Ψ relation by the film thickness of the pattern by 6th Example of this invention. It is a flowchart which shows the pattern measurement process by 6th Example of this invention. It is a figure which shows the structure of the conventional ellipsometer. It is a figure which shows the elliptically polarized light measured by ellipsometry.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1,23 Light source 2,25 Polarizer 3 Sample 4 Rotating analyzer 4a, 24 1/4 wavelength plate 4b Rotating 1/4 wavelength plate 5, 27, 27A, 27B Detector 10, 22 Substrate 11 Pattern 11a, 11b Pattern element 21 Stage 21A Positioning Structure 21B Positioning Pin 22a Pattern 22b Pattern Element 22A Orientation Flat 25A Beam Chopper 26 Rotating Analyzer 28 Amplifier 29 A / D Converter 30 Processing Unit 31 Lens 32 Beam Splitter 101 Wafer Process Unit 102 Process Control Unit 103 Ellipsometry Part 104 initial condition setting part 104a database 104b theoretical calculation part

Claims (10)

In a semiconductor device manufacturing method for manufacturing a semiconductor device based on dimensions of a structure formed on a substrate,
An incident light beam polarized at a predetermined angle with respect to the substrate surface is incident on the structure, and a polarization state of an outgoing light beam emitted from the structure on which the incident light beam is incident is measured. A semiconductor device is manufactured by setting manufacturing parameters of an etching process based on information on a dimension of the structure obtained based on an evaluation obtained from the direction parallel to the substrate surface. Manufacturing method.   2. The method of manufacturing a semiconductor device according to claim 1, wherein the information on the dimensions is fed back to an etching process of the semiconductor device, and the manufacturing parameters are adjusted.   3. The method of manufacturing a semiconductor device according to claim 1, wherein the evaluation is performed by referring to a database including a relationship between the polarization state and information on dimensions of the structure.   The polarization state is expressed by rotation and flatness of a polarization plane of an outgoing light beam after passing through the structure, and the measurement is performed by an ellipsometer. A method for manufacturing a semiconductor device according to claim 1.   5. The method of manufacturing a semiconductor device according to claim 1, wherein the incident is performed by changing an incident angle of the polarized incident light beam. 6.   The measurement uses reflected light of the incident light beam, diffracted light of the incident light beam diffracted from the structure, or scattered light of the incident light beam scattered by the structure as the outgoing light beam. The method for manufacturing a semiconductor device according to claim 1, wherein the method is a semiconductor device manufacturing method.   The method for manufacturing a semiconductor device according to claim 1, wherein the evaluation obtains information or a width related to a cross-sectional shape of the structure based on the polarization state and a film thickness. .   The incident interrupts the incident light, and the measurement measures the polarization state of the intermittent outgoing light beam formed corresponding to the intermittent incident light beam in the irradiation state and the cutoff state. The manufacturing method of the semiconductor device as described in any one of Claims 1-7. A decomposing step of decomposing the incident light beam into a plurality of light beams; a measuring step of calculating and integrating the polarization components and intensities of the plurality of outgoing light beams emitted from the structure and corresponding to the plurality of light beams;
Obtaining information about dimensions of the structure in a direction parallel to the substrate surface based on the integrated polarization component and intensity; and
The method for manufacturing a semiconductor device according to claim 1, comprising: The semiconductor device manufactured by the manufacturing method of the semiconductor device as described in any one of Claims 1-9.

Images