JP7633137B2 - Ellipsometer and semiconductor device inspection device - Google Patents

Description

The present invention relates to an ellipsometer and an inspection device for semiconductor devices.

Since Aspnes et al. made automatic measurement possible in 1975, ellipsometry has significantly reduced measurement time and improved accuracy, and spectroscopic ellipsometry, which measures using multiple wavelengths, has also been put to practical use. Since then, it has been widely used in semiconductor manufacturing processes, taking advantage of its ability to measure dimensions such as film thickness and optical constants such as refractive index with high precision in non-destructive measurements of thin films and fine structures. Even today, it is used as an OCD (Optical Critical Dimension) measuring device to measure the dimensions of fine structures with line widths of circuit patterns on wafers of 10 nm or less, complementing length-measuring SEMs (Scanning Electron-beam Microscopes) and AFMs (Atomic Force Microscopes).

Over the past decade, semiconductor circuit structures have become more three-dimensional and complex, such as FinFET (Fin Field-Effect Transistor) in logic semiconductors and 3D-NAND in memory. Many OCDs use spectroscopic ellipsometry as their measurement principle, and in order to determine the dimensions of the semiconductor circuit structure being measured and the optical constants of the constituent materials, a model is created, and the dimensions and optical constants of the measurement target are treated as floating parameters, and the model is fitted to the measurement results to obtain a solution. Therefore, as the structure of the target becomes more complex, the number of floating parameters increases. For example, current OCD measurements of FinFETs require the use of around 20-30 floating parameters. In ellipsometry, two values, Ψ and Δ, are generally obtained as the ellipsometry coefficients, which are the measurement results, but both the ellipsometry coefficients Ψ and Δ are wavelength dependent. For this reason, in the case of spectroscopic ellipsometry, the ellipsometry coefficients Ψ and Δ can be written as Ψ(λ) and Δ(λ).

In order to find the solution for the dimension, it is necessary at a minimum to measure and obtain a number of ellipsometry coefficients Ψ and Δ that is greater than the number of floating parameters in order to fit the model. However, when there are a large number of floating parameters, a problem occurs in which the fitting converges to a combination of floating parameters that is different from the actual dimension. This is a problem called coupling, and to avoid this, it is effective to measure and fit ellipsometry coefficients Ψ and Δ that have different dependencies on the floating parameters. Therefore, it is desirable to perform ellipsometry measurements at different angles of incidence and orientations of incidence in addition to wavelengths, and to use ellipsometry coefficients Ψ and Δ that have more different dependencies on the floating parameters for fitting the model.

In terms of measurement sensitivity, the Brewster angle, at which the reflectance of P-polarized light is zero for the illumination light on a semiconductor wafer, provides the highest measurement sensitivity. For this reason, ellipsometry measurements are often performed using the Brewster angle as the angle of incidence. This angle of incidence corresponds to approximately 60° to 75° in semiconductor circuit structures. When performing measurements with such oblique incidence, spectroscopic ellipsometry measurements are often performed using an oblique incidence optical system specialized for a single angle of incidence and orientation. However, to avoid the coupling problem mentioned above, there is a growing demand for optical systems that can accommodate multiple angles of incidence and orientations of incidence.

To meet this demand, if the optical system or semiconductor wafer were moved for each measurement, the measurement time would be extremely long. Therefore, an objective lens with a large NA, such as the Brewster angle mentioned above, is used to simultaneously introduce measurement light from a wide range of incident angles and orientations. Then, the reflected light from the semiconductor wafer is guided onto the exit pupil of the objective lens to perform ellipsometry measurement. This combination of the pupil image measurement optical system and ellipsometry measurement is ideal. However, even in this case, in a configuration that combines a rotating compensator and a rotating analyzer, which are common ellipsometry measurement methods, the measurement time for each wavelength is more than one second due to the limitations of the pupil image transfer speed. Furthermore, if spectroscopic ellipsometry measurement is performed using illumination light with more than 100 wavelengths, the measurement time becomes unrealistic for an OCD measurement device in the semiconductor manufacturing process.

JP 2021-085698 A U.S. Pat. No. 5,596,411 U.S. Pat. No. 7,616,323 U.S. Pat. No. 5,953,137 U.S. Pat. No. 6,856,384 Japanese Unexamined Patent Publication No. 197706/1986 JP 2010-002846 A JP 2000-331927 A JP 2003-279854 A

Ellipsometers used in OCD measurement devices in semiconductor manufacturing processes typically require a measurement time of one to several seconds to measure one point. Due to the time allocated for measurement, usually only a few to several dozen points on the wafer can be measured, and partial changes in film thickness or line width on the wafer may be overlooked, resulting in a decrease in yield. The main reasons for this include the need for a large number of measurement points synchronized with modulation by rotating compensators and phase modulation elements, the need to measure the amount of light split into each wavelength by a dispersive element such as a diffraction grating with a high S/N ratio in the case of spectroscopic measurement, and the need to measure while switching between several types of polarization states in the illumination light in the case of Mueller matrix ellipsometry.

On the other hand, in order to increase the number of measurement points within a wafer and shorten the time required for ellipsometry measurements, it is necessary to increase the speed of moving parts such as rotating compensators. However, stability and heat generation are bottlenecks, making it difficult to improve the throughput of ellipsometry coefficient measurements for OCD measurements, etc.

Figure 1 is a block diagram illustrating an ellipsometer according to a comparative example. As shown in Figure 1, the comparative example ellipsometer 101 described in Patent Document 1 was devised to solve the above-mentioned problems, and uses a Nomarski prism 131 to cause two orthogonal polarized components to interfere on an image detector 142. An analyzer 143 then finds Ψ and Δ from the amplitude and phase of the generated interference fringes. This method is expected to realize high throughput in the semiconductor manufacturing process, and to put into practical use OCD devices that are highly stable due to the lack of moving parts.

The illumination optical system 110 includes a light source 111, a spectroscope 112, a fiber 113, an illumination lens 114, a polarizer 115, a beam splitter 116, and an objective lens 117, and illuminates the sample 50 with illumination light L1. The light collecting optical system 120 includes an objective lens 117, a beam splitter 116, and relay lenses 121 and 122, and collects reflected light R1 reflected by the sample 50. The polarizing optical element 130 includes a Nomarski prism 131, and the light receiving optical system 140 includes an analyzer 141, an image detector 142, and an analysis device 143.

This method can obtain Ψ and Δ from a single image without obtaining the Stokes parameters that are common in ellipsometry measurements. This makes the measurement efficient, and by combining it with a high-speed spectrometer, spectroscopic ellipsometry is also possible. By applying this measurement principle to pupil image measurement using a large NA objective lens 117, it is possible to simultaneously obtain ellipsometry information for multiple angles of incidence and orientations of incidence, achieving ideal performance as a measurement device for semiconductor manufacturing.

Figure 2 is a diagram illustrating interference fringes on an image detector 142 according to a comparative example. In Figure 2, a graph is shown in which the horizontal axis represents the position of the image detector 142 and the vertical axis represents the intensity of light forming the interference fringes, corresponding to the interference fringes. As shown in Figure 2, the ellipsometer 101 of the comparative example forms vertical interference fringes on the image detector 142, for example, by interfering with X-polarized light and Y-polarized light. On the vertical line passing through the center of the image detector 142, the difference in optical path length between X-polarized light and Y-polarized light is zero. If the wavelength width is wide, the contrast of the interference fringes on both sides decreases.

As described above, because partially coherent light is used as the illumination light L1, if the number of interference fringes formed on the pupil is increased, the contrast of the interference fringes will decrease in the left and right regions where the optical path length difference between the two interfering polarization components is large, as shown in Figure 2. The phenomenon in which the contrast of the interference fringes decreases in regions where the optical path length difference is large means that there is an upper limit to the number of usable interference fringes.

Figure 3 is a diagram illustrating interference fringes when the wavelength width is the same for long-, medium-, and short-wavelength illumination light in a comparative example. As shown in Figure 3, when the wavelength width is the same, the number of interference fringes is greater for short wavelengths, so the contrast of the interference fringes is lower for shorter wavelengths. To avoid this problem, it is necessary to reduce the polarization separation angle of the Nomarski prism 131 or narrow the wavelength width when measuring at short wavelengths. However, in the former case, the fringe spacing becomes wider, especially on the long-wavelength side, and the measurement resolution on the pupil decreases. In the latter case, a side effect occurs in which the amount of light at short wavelengths is reduced.

The present invention has been made to solve these problems, and provides an ellipsometer and semiconductor device inspection device that can improve the throughput of measuring the ellipsometry coefficients Ψ and Δ by suppressing the decrease in resolution caused by the widening of the fringe spacing and the decrease in light quantity caused by narrowing the wavelength width of the spectrometer, even when measurements are made using a combination of a light source with a wide wavelength range and a spectrometer.

Linearly polarized light is passed through the objective lens to simultaneously illuminate the measurement sample from multiple angles and orientations of incidence, and the light reflected by the measurement sample is made incident on the objective lens, passes through a relay lens and a polarizing optical element, and is then received by an image detector positioned at the pupil conjugate position of the objective lens.

The polarization analysis section is composed of at least two uniaxial birefringent crystals (e.g., magnesium fluoride) and a linear polarizer. The uniaxial birefringent crystals are combined so that the crystal optical axes are perpendicular to each other and to the optical axis, and the light that is reflected by the sample is separated into two polarized components in the radial direction centered on the optical axis.

These uniaxial birefringent crystals are spherical, with at least one side being convex or concave, and have no refractive power when two are combined. The two polarized light components separated by the crystal through a linear polarizer (analyzer) located just before the image detector interfere with each other on the image detector, forming concentric interference fringes. The contrast and phase information of the resulting interference fringe image are processed to determine the ellipsometry measurement results Ψ and Δ.

The ellipsometer of one embodiment includes a polarizing optical element unit that separates two polarized components with mutually orthogonal polarization directions of the reflected light, which is generated when illumination light containing at least one of linearly polarized light, circularly polarized light, and elliptically polarized light is reflected off the measurement surface of the sample, into a radial direction centered on the optical axis of the optical system of the reflected light; an analyzer unit that transmits components of the two polarized components with directions different from the polarization directions, causing interference to form concentric interference fringes; an image detector that detects the interference fringes; and an analyzer that calculates the ellipsometry coefficients Ψ and Δ from the detected interference fringes.

In one embodiment, the ellipsometer includes a polarizing optical element that separates the reflected light, which is illumination light including linear polarization and is reflected by the measurement surface of the sample, into two linearly polarized components with mutually orthogonal polarization directions, an analyzer that transmits the two linearly polarized components in directions different from the polarization directions, causing them to interfere with each other and form interference fringes, an image detector that detects the interference fringes, and an analyzer that calculates the ellipsometry coefficients Ψ and Δ from the detected interference fringes, and the fringe pitch of the interference fringes has the same value for multiple wavelengths of the illumination light.

Furthermore, in one embodiment, a semiconductor device inspection device includes the above-mentioned ellipsometer.

The present invention provides an ellipsometer and semiconductor device inspection device that can improve the throughput of measuring the ellipsometry coefficients Ψ and Δ.

FIG. 1 is a configuration diagram illustrating an ellipsometer according to a comparative example. 11 is a diagram illustrating interference fringes on an image detector according to a comparative example. 11 is a diagram illustrating interference fringes when long-wavelength, medium-wavelength, and short-wavelength illumination light have the same wavelength width according to a comparative example. 1 is a configuration diagram illustrating an ellipsometer according to a first embodiment. 4 is a diagram illustrating linearly polarized light transmitted through an analyzer unit in the ellipsometer according to the first embodiment. FIG. 4 is a diagram illustrating an example of interference fringes of reflected light that has interfered on an image detector in the ellipsometer according to the first embodiment. FIG. 5A to 5C are diagrams illustrating a method of analyzing interference fringes of reflected light that has interfered on the image detector according to the first embodiment. 5A to 5C are diagrams illustrating a method of analyzing interference fringes of reflected light that has interfered on the image detector according to the first embodiment. 5A to 5C are diagrams illustrating a method of analyzing interference fringes of reflected light that has interfered on the image detector according to the first embodiment. 5 is a diagram illustrating interference fringes of reflected light that has interfered on an image detector in the ellipsometers according to the first embodiment and the comparative example. FIG. 1 is a perspective view illustrating an example of the arrangement of an illumination optical system, a light collecting optical system, and a pupil plane according to a first embodiment. 4 is a diagram illustrating an example of the relationship between the position on the pupil plane and the angle and direction of incidence of light on a sample according to the first embodiment. FIG. FIG. 13 is a configuration diagram illustrating an inspection device for a semiconductor device according to a modified example of the first embodiment. FIG. 11 is a configuration diagram illustrating an ellipsometer according to a second embodiment. 11 is a diagram illustrating interference fringes of reflected light, radial polarization, and azimuth polarization that have interfered on an image detector in the ellipsometer according to the second embodiment. FIG. 11A and 11B are diagrams illustrating a polarizing optical element section according to a second embodiment. FIG. 11 is a configuration diagram illustrating an ellipsometer according to a third embodiment. 13 is a diagram illustrating interference fringes of reflected light, radial polarization, and azimuth polarization that have interfered on an image detector in the ellipsometer according to the third embodiment. FIG. 13A and 13B are diagrams illustrating a polarizing optical element portion according to a third embodiment. 13 is a diagram illustrating a polarizing optical element portion according to a first modified example of the third embodiment. FIG. 13 is a diagram illustrating a polarizing optical element portion according to Modification 2 of Embodiment 3. FIG. 13 is a diagram illustrating a polarizing optical element portion according to Modification 3 of Embodiment 3. FIG. 13 is a diagram illustrating a polarizing optical element portion according to a fourth modified example of the third embodiment. FIG. FIG. 11 is a configuration diagram illustrating an ellipsometer according to a fourth embodiment. 13 is a diagram illustrating an example of interference fringes of reflected light, X-polarized light, and Y-polarized light that have interfered on an image detector in the ellipsometer according to the fourth embodiment. FIG. 1 is a graph illustrating the wavelength dependence of birefringence in each material of a birefringent crystal, where the horizontal axis indicates wavelength and the vertical axis indicates birefringence. 11A and 11B are diagrams illustrating a polarizing optical element according to a fourth embodiment.

For clarity of explanation, the following description and drawings have been omitted and simplified as appropriate. In addition, the same elements in each drawing are given the same reference numerals, and duplicate explanations have been omitted as necessary.

(Embodiment 1)
An ellipsometer according to the embodiment 1 will be described. Fig. 4 is a configuration diagram illustrating the ellipsometer according to the embodiment 1. As shown in Fig. 4, the ellipsometer 1 includes an illumination optical system 10, a focusing optical system 20, a polarizing optical element unit 30, and a light receiving optical system 40. The ellipsometer 1 receives reflected light R1 that is illumination light L1 reflected by a sample 50, and obtains ellipsometry coefficients Ψ and Δ.

The illumination optical system 10 illuminates the sample 50 with illumination light L1 that includes linearly polarized light. The illumination optical system 10 includes a light source 11, a spectrometer 12, a fiber 13, an illumination lens 14, a polarizer unit 15, a beam splitter 16, and an objective lens 17.

The light source 11 generates illumination light L1. The illumination light L1 generated by the light source 11 may include light of a wide band of wavelengths. The illumination light L1 is, for example, white light. Note that the illumination light L1 generated by the light source 11 is not limited to white light, and may be monochromatic light having a specific wavelength or light having a specific wavelength range. The illumination light L1 generated by the light source 11 is incident on the spectroscope 12.

The spectroscope 12 extracts and emits light of a specific wavelength range from the incident illumination light L1. For example, the spectroscope 12 emits light with a central wavelength of 400 nm and a wavelength range of 10 nm. The illumination light L1 emitted from the spectroscope 12 enters the fiber 13.

The fiber 13 is a cable-like light-guiding member having one end and the other end. The illumination light L1 that is incident on one end of the fiber 13 is emitted from the other end of the fiber 13. The illumination light L1 that is emitted from the other end of the fiber 13 is incident on the illumination lens 14.

The illumination lens 14 is, for example, a convex lens. The illumination lens 14 changes the angular distribution of the incident illumination light L1 and irradiates the illumination light L1 onto the polarizer unit 15. For example, the illumination lens 14 converts the illumination light L1 emitted from the other end of the fiber 13 into parallel light. The parallel illumination light L1 is then incident on the polarizer unit 15.

The polarizer unit 15 receives the illumination light L1 generated by the light source 11. The polarizer unit 15 includes, for example, a linear polarizer H1 that generates linearly polarized light. Thus, the polarizer unit 15 transmits the illumination light L1 that includes linearly polarized light in one direction. For example, the polarizer unit 15 outputs the linearly polarized illumination light L1, whose polarization direction is tilted 45° with respect to the paper surface, to the beam splitter 16.

The beam splitter 16 reflects a portion of the incident illumination light L1 and transmits a portion of it. The beam splitter 16 reflects a portion of the incident illumination light L1 toward the objective lens 17. The illumination light L1 reflected by the beam splitter 16 is incident on the objective lens 17.

The objective lens 17 illuminates the sample 50 with illumination light L1 that includes linearly polarized light. The objective lens 17 focuses the illumination light L1 reflected by the beam splitter 16 into a point shape to illuminate the sample 50. The objective lens 17 has an NA. The objective lens 17 illuminates the sample 50 from all angles and all directions within the NA. It is desirable for the NA of the objective lens to be equal to or greater than a value that includes the Brewster's angle with respect to the sample 50 (e.g., 0.95).

The objective lens 17 transmits the illumination light L1 and also transmits the reflected light R1 that is reflected from the measurement surface of the sample 50. In the ellipsometer 1 of this embodiment, the optical axis C of the illumination light L1 that is incident on the sample 50 and the optical axis C of the reflected light R1 that is reflected from the sample 50 are perpendicular to the measurement surface of the sample 50.

Here, for ease of explanation of the ellipsometer 1, an XYZ orthogonal coordinate system is introduced. The Z-axis direction is defined as the optical axis C. The two directions that are perpendicular to the Z-axis direction and perpendicular to each other are defined as the X-axis direction and the Y-axis direction.

The illumination light L1 that illuminates the sample 50 contains linearly polarized light in one direction. Such illumination light L1 containing linearly polarized light in one direction is incident on the measurement surface of the sample 50 while being focused. Therefore, if the illumination light L1 is fully polarized and linearly polarized, and the optical axis C is perpendicular to the measurement surface of the sample 50, the illumination light L1 will have both P-polarized and S-polarized portions depending on the direction of incidence on the measurement surface. The P-polarized portion of the illumination light L1 is reflected as P-polarized light. The S-polarized portion of the illumination light L1 is reflected as S-polarized light.

The focusing optical system 20 focuses the reflected light R1 of the illumination light L1 reflected by the sample 50. The focusing optical system 20 includes an objective lens 17, a beam splitter 16, and relay lenses 21 and 22. The objective lens 17 is a component of the illumination optical system 10 and also a component of the focusing optical system 20. The reflected light R1 reflected by the sample 50 passes through a pupil position 23 of the objective lens 17. The pupil position 23 is re-imaged on the image detector 42 by the relay lenses 21 and 22. In this way, the focusing optical system 20 images the exit pupil of the objective lens 17 on the image detector 42. The objective lens 17 transmits the reflected light R1 of the illumination light L1 reflected by the sample 50, and causes it to enter the beam splitter 16.

The beam splitter 16 transmits a portion of the incident reflected light R1. For example, the reflected light R1 that transmits through the beam splitter 16 is incident on the relay lens 21.

The relay lens 21 focuses the reflected light R1 that has passed through the beam splitter 16, forms an image, and then causes it to enter the relay lens 22. The relay lens 22 transmits the incident reflected light R1 and causes it to enter the polarizing optical element unit 30.

The polarizing optical element unit 30 separates two linearly polarized components having mutually orthogonal polarization directions from the reflected light R1, which is the result of the illumination light L1 containing linear polarization being reflected by the measurement surface of the sample 50, into a radial direction centered on the optical axis C of the focusing optical system 20 of the reflected light R1. The polarizing optical element unit 30 includes, for example, a Wollaston lens. The polarizing optical element unit 30 may include a plurality of Wollaston lenses W10 and Wollaston lenses W20. The Wollaston lens W10 includes two birefringent crystals W11 and W12. The Wollaston lens W20 includes two birefringent crystals W21 and W22.

The birefringent crystal W11 includes a uniaxial birefringent crystal. The birefringent crystal W11 has a plane entrance surface and a concave spherical exit surface. The birefringent crystal W11 has a crystal optical axis in the X-axis direction. The birefringent crystal W12 includes a uniaxial birefringent crystal. The birefringent crystal W12 has a convex spherical entrance surface and a plane exit surface. The birefringent crystal W12 has a crystal optical axis in the Y-axis direction. Thus, the Wollaston lens W10 includes two uniaxial birefringent crystals having spherical shapes that fit together. The crystal optical axes of the uniaxial birefringent crystals of the Wollaston lens W10 are perpendicular to the optical axis C of the focusing optical system 20 and perpendicular to each other.

The birefringent crystal W21 includes a uniaxial birefringent crystal. The birefringent crystal W21 has a planar entrance surface and a concave spherical exit surface. The birefringent crystal W21 has a crystal optical axis in the Y-axis direction. The birefringent crystal W22 includes a uniaxial birefringent crystal. The birefringent crystal W22 has a convex spherical entrance surface and a planar exit surface. The birefringent crystal W22 has a crystal optical axis in the X-axis direction. Thus, the Wollaston lens W20 includes two uniaxial birefringent crystals having spherical shapes that fit together. The crystal optical axes of the uniaxial birefringent crystals of the Wollaston lens W20 are perpendicular to the optical axis C of the focusing optical system 20 and perpendicular to each other.

Each of the birefringent crystals W11, W12, W21, and W22 may contain any of the following materials: quartz, magnesium fluoride, sapphire, calcite, and αBBO. In the polarizing optical element section 30, the Wollaston lens W10 and the Wollaston lens W20 have a refractive power of 0. Therefore, the polarizing optical element section 30 does not need to focus or diffuse the reflected light R1.

In the Wollaston lens W10, the crystal optical axis of the birefringent crystal W11 and the crystal optical axis of the birefringent crystal W12 are arranged so that they are perpendicular to each other and orthogonal to the optical axis C. For example, the crystal optical axis of the birefringent crystal W11 is in the X-axis direction, and the crystal optical axis of the birefringent crystal W12 is in the Y-axis direction. As a result, the two mutually perpendicular polarized lights (X-polarized light and Y-polarized light in FIG. 1) are radially separated from the optical axis C and travel in different directions. For example, the X-polarized light is separated outward, farther from the optical axis C than the Y-polarized light. The Y-polarized light is separated inward, closer to the optical axis C than the X-polarized light.

The polarizing optical element section 30 further includes another pair of Wollaston lenses W20. In the Wollaston lens W20, the crystal optical axis of the birefringent crystal W21 and the crystal optical axis of the birefringent crystal W22 are arranged so as to be perpendicular to each other and perpendicular to the optical axis C. However, in the Wollaston lens W20, the crystal optical axis of each birefringent crystal is opposite to the crystal optical axis of each birefringent crystal of the Wollaston lens W10. For example, the crystal optical axis of the birefringent crystal W21 is in the Y-axis direction, and the crystal optical axis of the birefringent crystal W22 is in the X-axis direction. As a result, the two pairs of Wollaston lenses W10 and W20 are designed in terms of their arrangement and curvature radius so that they separate the respective polarized lights in opposite directions and the two polarized lights are again at the same point on the image detector 42. Therefore, the two polarized components of the reflected light R1 transmitted through the polarizing optical element section 30 and the analyzer section 41 are detected at the same point on the image detector 42.

The light receiving optical system 40 receives the reflected light R1 and calculates the ellipsometry coefficients Ψ and Δ. The light receiving optical system 40 has an analyzer section 41, an image detector 42, and an analyzer 43. The analyzer section 41 includes, for example, a linear polarizer H2 that transmits a linearly polarized component in a predetermined direction.

5 is a diagram illustrating linearly polarized light passing through the analyzer unit 41 in the ellipsometer 1 according to the first embodiment. As shown in FIG. 5, the analyzer unit 41 transmits linearly polarized light components in the X-direction polarization direction and the Y-direction polarization direction separated by the polarizing optical element unit 30 and inclined by 45 degrees. Thus, the analyzer unit 41 transmits the polarized light component inclined by 45 degrees from the X-direction among linearly polarized light having the X-direction polarization direction. The analyzer unit 41 also transmits the polarized light component inclined by 45 degrees from the Y-direction among linearly polarized light having the Y-direction polarization direction. In this way, the analyzer unit 41, such as the linear polarizer H2, is also arranged between the Wollaston lens W20 and the image detector 42. The transmission axis direction of this linear polarizer H2 is arranged in an orientation that is midway (45 degrees) between the optical axes of the two uniaxial birefringent crystals that form the Wollaston lenses W10 and W20. Therefore, the analyzer unit 41 transmits the components of the two mutually orthogonal polarized light segments with directions different from each other, causing them to interfere with each other and form concentric interference fringes. In other words, the two polarized light segments are transmitted through the analyzer unit 41 and are emitted as polarized light components polarized in the same direction (a direction tilted by 45 degrees), causing them to interfere with each other.

The image detector 42 receives the incident reflected light R1. The image detector 42 is disposed at a pupil conjugate position 24 that is conjugate with the pupil position 23 of the objective lens 17. The reflected light R1 contains polarization components in the same direction of two linearly polarized lights that are orthogonal to each other. Therefore, the reflected light R1 interferes on the image detector 42. As a result, concentric interference fringes are formed on the image detector 42. The image detector 42 detects the interference fringes of each polarization component that has passed through the analyzer section 41.

Figure 6 is a diagram illustrating interference fringes of reflected light that has interfered on the image detector 42 in the ellipsometer 1 according to the first embodiment. As shown in Figure 6, the interference fringes on the image detector 42 are concentric. In this embodiment, such concentric interference fringes are called Wollaston lens interference fringes. Wollaston lens interference fringes are formed by the interference of X-polarized light and Y-polarized light. In the Wollaston lens interference fringes, the fringe spacing is large near the center, and the fringe spacing is small at the edge of the pupil corresponding to the light incident at the Brewster angle. The Wollaston lens interference fringes have a circular shape in the peripheral region of the image detector 42, with positions where the optical path length difference between X-polarized light and Y-polarized light is 0.

7 to 9 are diagrams illustrating a method for analyzing the interference fringes of the reflected light R1 that has interfered on the image detector 42 according to the first embodiment. As shown in FIG. 7, the analysis device 43 acquires an image detected on the image detector 42. The analysis device 43 then converts the acquired image including the point-symmetric concentric interference fringes into polar coordinates so that the center of the concentric circle becomes the origin of the coordinate system. Next, the analysis device 43 expresses the image after the coordinate conversion in two-dimensional coordinates with the radius (radius vector) and azimuth (deflection angle) of the polar coordinate system as axes. The analysis device 43 performs a two-dimensional Fourier transform on the plane of the two-dimensional coordinates, and trims the converted frequency space image into DC and AC components.

Next, as shown in FIG. 8, the analysis device 43 performs an inverse Fourier transform on the DC component after shifting the peak position to the coordinate origin. Also, as shown in FIG. 9, the analysis device 43 performs an inverse Fourier transform on the AC component after shifting the peak position to the coordinate origin. The analysis device 43 further obtains an AC component (amplitude) which is an amplitude component and an AC component (phase) which is a phase component from the AC component. The obtained AC component (phase) directly corresponds to Δ of the ellipsometry measurement. Ψ of the ellipsometry measurement is expressed by the following formula (1). Here, DC indicates the DC component, and AC amplitude indicates the AC component (amplitude). In this way, the analysis device 43 calculates the ellipsometry coefficients Ψ and Δ from the detected interference fringes based on the result of the inverse Fourier transform.

Ψ=tan ^-1 [DC/AC amplitude ±√{(DC/AC amplitude) ² -1}] (1)

Next, the effect of this embodiment will be described. FIG. 10 is a diagram illustrating interference fringes of reflected light R1 that has interfered on the image detector 42 in the ellipsometers 1 and 101 according to the first embodiment and the comparative example. As shown in FIG. 10, the ellipsometers 1 and 101 that use interference fringes use light that is partially coherent in time and space for measurement. Therefore, in places where the optical path length difference between the two interfering polarized lights becomes large, the contrast of the interference fringes is significantly reduced. Therefore, there is an upper limit to the number of interference fringes formed in the pupil that depends on the wavelength width and the field of view.

As shown in FIG. 10, in order to compare the concentric interference fringes formed by the ellipsometer 1 of this embodiment with the interference fringes formed by the ellipsometer 101 of the comparative example, for example, the number of fringes is the same for both, about five. In this case, the interference fringes of the comparative example are equally spaced within the pupil. In contrast, the interference fringes of this embodiment are sparse near the center and dense near the periphery. Also, the width containing the same five interference fringes is the pupil diameter in the comparative example, whereas it is the pupil radius in this embodiment. Therefore, in the ellipsometer 1 of this embodiment, the interference fringes can be arranged twice as densely as in the ellipsometer 101 of the comparative example. In particular, the interference fringes can be formed several to ten times denser near the periphery of the pupil.

Here, we consider the importance of the pupil in ellipsometry measurement. FIG. 11 is a perspective view illustrating the arrangement of the illumination optical system 10, the focusing optical system 20, and the pupil plane according to the first embodiment. It is a diagram illustrating the relationship between the position on the pupil plane and the angle and direction of incidence of light on the sample 50. Light that is perpendicular to the sample 50 reaches the vicinity of the center of the pupil. On the other hand, light that is incident on the sample 50 at a large angle of incidence reaches the vicinity of the periphery of the sample 50. In ellipsometry measurement, the latter oblique incidence light is more important. In particular, ellipsometry measurement using illumination light L1 with an incidence angle of 60° to 75°, near the Brewster angle, is the most sensitive. In addition, illumination light L1 with such an incidence angle has a large incidence angle dependency and incidence direction dependency, so it can be measured with high resolution.

Figure 12 is a diagram illustrating the relationship between the position on the pupil plane and the angle and direction of incidence of light on the sample 50 in the first embodiment. NA = 0.95, which is the upper limit of the NA of a typical objective lens 17, corresponds to an angle of incidence of 72°, which almost includes the Brewster angle. NA = 0.7 corresponds to an angle of incidence of 45°. The ellipsometer 1 of this embodiment can form dense interference fringes near the more important periphery on the pupil. Therefore, it has a very significant advantage, for example, in high-precision measurements in semiconductor manufacturing.

In this way, the ellipsometer 1 of this embodiment has a unique measurement principle in which Ψ and Δ are calculated from concentric interference fringes formed by two different polarized lights. In particular, when performing spectroscopic ellipsometry measurement by combining a broadband light source with a spectrometer, it is possible to improve the resolution on the pupil and increase the amount of light, which is expected to shorten the measurement time. This makes it possible to improve measurement accuracy and throughput.

(Modification)
Next, a modified example of the first embodiment will be described. This modified example is an example in which the above-mentioned ellipsometer 1 is applied to an inspection device for a semiconductor device. FIG. 13 is a configuration diagram illustrating an inspection device for a semiconductor device according to a modified example of the first embodiment. As shown in FIG. 13, the inspection device 1A includes an ellipsometer 1, a base 80, an isolator 81, an optical surface plate 82, a stage 83, a wafer holder 84, a frame 85, an environmental chamber 86, a temperature controller unit 87, and an automatic substrate transport device 88. The ellipsometer 1 includes a light source 11, a spectroscope 12, an illumination optical system 10, a focusing optical system 20, a polarizing optical element unit 30, and a light receiving optical system 40.

The base 80 is a base. The isolator 81 provided on the base 80 isolates vibrations from the floor surface. The optical surface plate 82 is provided on the isolator 81. The stage 83 and the frame 85 are arranged on the optical surface plate 82. The stage 83 is placed on the sample 50, such as a Si wafer attached to a wafer holder 84. The frame 85 fixes the optical system. The environmental chamber 86 isolates some components of the ellipsometer 1, such as the base 80, the isolator 81, the optical surface plate 82, the stage 83, the wafer holder 84, the frame 85, etc., from the external environment. The temperature controller unit 87 maintains the inside of the environmental chamber 86 at a predetermined temperature. The automatic substrate transport device 88 transports the sample 50.

The analysis device 43 of the ellipsometer 1 captures images from an image detector 42 such as a camera and processes the ellipsometry measurement. The analysis device 43 may include a control unit 91 such as a computer, a grabber board 92 that captures and processes images from the image detector 42, a stage controller 93 that controls a stage 83 on which the sample 50 is mounted, and a light source/spectroscope controller 94 that controls the light source 11 and the spectrometer 12.

The semiconductor device inspection device 1A of this modified example is equipped with the ellipsometer 1 described above. This improves the measurement accuracy in semiconductor device inspection and increases throughput. For example, in CD measurement and overlay evaluation in the semiconductor manufacturing process, measurements were previously only taken at a few points within a single wafer, but it is now possible to perform distribution evaluation within shots, chips, and memory cells on a wafer. This contributes to improved yield and productivity in the manufacture of semiconductor devices, and also leads to reduced costs for semiconductor devices. Other configurations and effects are included in the description of embodiment 1.

(Embodiment 2)
Next, an ellipsometer according to the second embodiment will be described. The ellipsometer of this embodiment uses illumination light L1 including circularly polarized light. FIG. 14 is a configuration diagram illustrating an ellipsometer according to the second embodiment. As shown in FIG. 14, the ellipsometer 2 of this embodiment has a polarizer unit 18 instead of the polarizer unit 15. The polarizer unit 18 includes a circular polarizer that generates circularly polarized light that rotates left or right around the optical axis of the illumination light L1. The circular polarizer includes a linear polarizer H1 and a λ/4 wave plate H3.

In addition, the ellipsometer 2 of this embodiment has an analyzer section 44 instead of the analyzer section 41. The analyzer section 44 includes a circular polarizer that transmits a circularly polarized light component that rotates either left or right. The circular polarizer includes a linear polarizer H2 and a λ/4 wave plate H4.

Furthermore, the ellipsometer 2 of this embodiment has a polarizing optical element section 31 instead of the polarizing optical element section 30. The polarizing optical element section 31 includes a birefringent crystal F1. The birefringent crystal F1 is arranged in a position in the focusing optical system 20 where the reflected light R1 is focused or diffused. For example, the birefringent crystal F1 is arranged in a position where the reflected light R1 is diffused between the relay lens 21 and the relay lens 22. The birefringent crystal F1 has a parallel plate shape. Therefore, the birefringent crystal F1 has two parallel plate surfaces. The birefringent crystal F1 is a uniaxial birefringent crystal. The crystal optical axis of the birefringent crystal F1 is arranged in a direction perpendicular to the plate surface. The birefringent crystal F1 is arranged so that the plate surface is perpendicular to the optical axis. Therefore, the crystal optical axis of the birefringent crystal F1 is parallel to the optical axis C of the focusing optical system 20. The birefringent crystal F1 includes, for example, αBBO as a material. The birefringent crystal F1 may include any of quartz, magnesium fluoride, sapphire, and calcite.

In the ellipsometer 2 of this embodiment, the illumination light L1 converted to parallel light by the illumination lens 14 passes through the linear polarizer H1 and contains linearly polarized light. The illumination light L1 containing linearly polarized light then passes through the λ/4 wavelength plate H3 to become illumination light L1 containing circularly polarized light. Note that the illumination light L1 may also contain elliptically polarized light.

The illumination light L1 passes through the objective lens 17 via the beam splitter 16. This allows the sample 50 to be illuminated from all angles and all directions within the NA of the objective lens 17. In this case, it is desirable for the NA of the objective lens 17 to be equal to or greater than a value that includes the Brewster angle of the sample 50.

The reflected light R1 reflected by the sample 50 contains radially polarized light and azimuthally polarized light. The reflected light R1 passes through the pupil position 23 of the objective lens 17. The pupil conjugate position 24, which is conjugate with the pupil position 23, is imaged on the image detector 42 by the relay lenses 21 and 22.

Figure 15 is a diagram illustrating the interference fringes, radial polarization, and azimuth polarization of the reflected light R1 that interfere on the image detector 42 in the ellipsometer 2 according to the second embodiment. As shown in Figure 15, consider the action of the radial polarization and azimuth polarization contained in the reflected light R1 when they pass through the birefringent crystal F1. First, the azimuth polarization is always perpendicular to the crystal optical axis of the birefringent crystal F1. Therefore, the azimuth polarization is ordinary light. On the other hand, the polarization direction of the radial polarization is inclined with respect to the optical axis C except for the optical axis C. Therefore, it includes extraordinary light in which the crystal optical axis of the birefringent crystal F1 is parallel to the electric field.

Figure 16 is a diagram illustrating a polarizing optical element unit 31 according to embodiment 2. In Figure 16, the linear polarizer H2 and λ/4 wave plate H4 of the analyzer unit 44 are omitted in order to explain the propagation state of radial polarization and azimuth polarization. As shown in Figure 16, in the polarizing optical element unit 31 of this embodiment, the radial polarization and azimuth polarization transmitted through the birefringent crystal F1 have relatively different propagation directions and phases due to the difference in refractive index. That is, the radial polarization containing extraordinary light spreads out in a direction away from the optical axis C more than the azimuth polarization containing ordinary light. On the other hand, after transmitting through the birefringent crystal F1, the radial polarization and azimuth polarization shift parallel to each other and proceed.

After passing through the relay lens 22, the radial polarization and the azimuth polarization return to the same point on the image detector 42 (on the pupil). It is important to note that the phase difference between the two polarizations is also delayed by different amounts. On the optical axis C, the radial polarization and the azimuth polarization have the same phase, but the phase differs as they approach the periphery of the image detector 42 (on the pupil).

Between the image detector 42 and the relay lens 22, an analyzer section 44 including a λ/4 wave plate H4 and a linear polarizer H2 is disposed. The analyzer section 44 transmits only the common polarization components of the radial polarized light and the azimuth polarized light. As a result, the radial polarized light and the azimuth polarized light transmitted through the analyzer section 44 become interferable and form concentric interference fringes on the image detector 42. Here, the radial polarized light corresponds to the P polarized light on the sample 50 such as a semiconductor wafer, and the azimuth polarized light corresponds to the S polarized light on the sample 50. Therefore, the ellipsometer 2 of this embodiment is configured to enable ellipsometric measurement of the P polarized light and the S polarized light in all directions, and can perform highly sensitive measurement independent of the structure of the measurement target due to its symmetry. Other configurations and effects are included in the description of the first embodiment and the modified examples.

(Embodiment 3)
Next, an ellipsometer according to embodiment 3 will be described. In the above-mentioned embodiment 2, the position where the optical path length difference between radial polarized light and azimuth polarized light is zero is on the optical axis C of the focusing optical system 20. In contrast, in this embodiment, the position where the optical path length difference between radial polarized light and azimuth polarized light is zero is arranged more peripherally on the pupil. This allows the number of interference fringes to be densely measured.

Figure 17 is a configuration diagram illustrating an ellipsometer according to the third embodiment. As shown in Figure 17, the ellipsometer 3 of this embodiment has a polarizing optical element section 32 instead of the polarizing optical element section 31. The polarizing optical element section 32 includes a birefringent crystal F2 in addition to the birefringent crystal F1. The birefringent crystal F2 is arranged between the relay lens 22 and the analyzer section 44. The birefringent crystal F2 is a plate having an incident surface and an exit surface, and the incident surface has a concave conical shape and the exit surface has a convex conical shape. The birefringent crystal F2 has a double-sided axicon lens shape, so to speak. The birefringent crystal F2 includes a uniaxial birefringent crystal. The crystal optical axis of the birefringent crystal F2 is parallel to the optical axis C of the focusing optical system 20.

The birefringent crystal F2 has a birefringence opposite to that of the birefringent crystal F1. Specifically, the birefringent crystal F1 and the birefringent crystal F2 have mutually opposite birefringence, that is, the refractive index of ordinary light n ₀ (hereinafter referred to as n ₀ )<the refractive index of extraordinary light n _e (hereinafter referred to as n _e ), and n ₀ >n _e . For example, if the birefringent crystal F1 is a negative crystal (n ₀ >n _e ), the birefringent crystal F2 is a positive crystal (n ₀ <n _e ). If the birefringent crystal F1 is a positive crystal (n ₀ <n _e ), the birefringent crystal F2 is a negative crystal (n ₀ >n _e ).

Figure 18 is a diagram illustrating the interference fringes of the reflected light, radial polarization, and azimuth polarization that have interfered on the image detector 42 in the ellipsometer 3 according to the third embodiment. Figure 19 is a diagram illustrating the polarizing optical element unit 32 according to the third embodiment. As shown in Figures 18 and 19, by adding a birefringent crystal F2 to the birefringent crystal F1, the position of the optical path length difference of 0 between the radial polarization and the azimuth polarization on the image detector 42 can be moved to a position outside the optical axis C. Note that in Figure 19, the linear polarizer H2 and the λ/4 wave plate H4 of the analyzer unit 44 are omitted in order to explain the propagation state of the radial polarization and the azimuth polarization.

(Modifications 1 to 4)
Next, modified examples 1 to 4 of embodiment 3 will be described. Figures 20 to 23 are diagrams illustrating polarizing optical element units 33 to 36 according to modified examples 1 to 4 of embodiment 3. As shown in Figures 20 to 23, even if the polarizing optical element units 33 to 36 have a configuration different from that of the polarizing optical element unit 32, concentric interference fringes of radial polarized light and azimuth polarized light can be formed on the image detector 42 by the same principle as that of the polarizing optical element unit 32.

20, the polarizing optical element section 33 of the modified example 1 includes a birefringent crystal F1 and a birefringent crystal F3. Specifically, the polarizing optical element section 33 of the modified example 1 has the configuration of the polarizing optical element section 32 of the embodiment 3, in which the birefringent crystal F2 is omitted and the birefringent crystal F3 is disposed between the birefringent crystal F1 and the relay lens 22. Thus, the birefringent crystal F3 is disposed in an area where the reflected light R1 of the focusing optical system 20 is focused or diffused. The birefringent crystal F3 includes a uniaxial birefringent crystal having a parallel plate shape. The birefringent crystal F3 has two parallel plate surfaces. The crystal optical axis of the birefringent crystal F3 is perpendicular to the plate surface. The birefringent crystal F3 is disposed so that the plate surface is perpendicular to the optical axis C of the focusing optical system 20. The crystal optical axis of the birefringent crystal F3 is parallel to the optical axis C.

The birefringent crystal F3 has a birefringence opposite to that of the birefringent crystal F1. Specifically, the birefringent crystal F1 and the birefringent crystal F3 have mutually opposite birefringences of n ₀ <n _e and n ₀ >n _e . For example, if the birefringent crystal F1 is a negative crystal (n ₀ >n _e ), the birefringent crystal F3 is a positive crystal (n ₀ <n _e ). If the birefringent crystal F1 is a positive crystal (n ₀ <n _e ), the birefringent crystal F3 is a negative crystal (n ₀ >n _e ).

21, the polarizing optical element unit 34 of the modified example 2 includes a birefringent crystal F2 having a double-sided axicon lens shape and a birefringent crystal F4 having a meniscus lens shape. Specifically, the polarizing optical element unit 34 of the modified example 2 includes a birefringent crystal F4 arranged between the relay lens 21 and the relay lens 22 instead of the birefringent crystal F1 in the configuration of the polarizing optical element unit 32 of the embodiment 3. Thus, the birefringent crystal F4 is arranged in the area where the reflected light R1 of the focusing optical system 20 is focused or diffused. The birefringent crystal F4 has an entrance surface and an exit surface. The birefringent crystal F4 has a concave spherical entrance surface and a convex spherical exit surface. The birefringent crystal F4 includes a uniaxial birefringent crystal. The crystal optical axis of the birefringent crystal F4 is arranged to be parallel to the optical axis C of the focusing optical system 20.

The birefringent crystal F4 has a birefringence opposite to that of the birefringent crystal F2. Specifically, the birefringent crystal F2 and the birefringent crystal F4 have mutually opposite birefringences of n ₀ <n _e and n ₀ >n _e . For example, if the birefringent crystal F2 is a negative crystal (n ₀ >n _e ), the birefringent crystal F4 is a positive crystal (n ₀ <n _e ). If the birefringent crystal F2 is a positive crystal (n ₀ <n _e ), the birefringent crystal F4 is a negative crystal (n ₀ >n _e ).

As shown in FIG. 22, the polarizing optical element section 35 of the modified example 3 includes a birefringent crystal F1, a birefringent crystal F5, and axicon lenses 25 and 26. Specifically, the polarizing optical element section 35 of the modified example 3 has a configuration in which a parallel plate-shaped birefringent crystal F5 is sandwiched between concave and convex axicon lenses 25 and 26 instead of the birefringent crystal F2 having a double-sided axicon lens shape in the configuration of the embodiment 3. The birefringent crystal F5 includes a uniaxial birefringent crystal having a parallel plate shape. The birefringent crystal F5 has two parallel plate surfaces. The crystal optical axis of the birefringent crystal F5 is perpendicular to the plate surfaces. The birefringent crystal F5 is arranged so that the plate surfaces are perpendicular to the optical axis C of the focusing optical system 20. The crystal optical axis of the birefringent crystal F5 is parallel to the optical axis C.

The axicon lens 25 is disposed between the relay lens 22 and the birefringent crystal F5. Therefore, the axicon lens 25 is disposed on the incident surface side of the birefringent crystal F5. The axicon lens 25 includes, for example, glass as a material. The axicon lens 25 has a concave conical incident surface and a flat exit surface.

The axicon lens 26 is disposed between the birefringent crystal F5 and the analyzer section 44. Therefore, the axicon lens 26 is disposed on the exit surface side of the birefringent crystal F5. The axicon lens 26 includes, for example, glass as a material. The axicon lens 26 has a flat entrance surface and a convex conical exit surface.

The birefringent crystal F5 has a birefringence opposite to that of the birefringent crystal F1. Specifically, the birefringent crystal F1 and the birefringent crystal F5 have mutually opposite birefringences of n ₀ <n _e and n ₀ >n _e . For example, if the birefringent crystal F1 is a negative crystal (n ₀ >n _e ), the birefringent crystal F5 is a positive crystal (n ₀ <n _e ). If the birefringent crystal F1 is a positive crystal (n ₀ <n _e ), the birefringent crystal F5 is a negative crystal (n ₀ >n _e ).

23, the polarizing optical element section 36 of the modified example 4 includes a birefringent crystal F2, a birefringent crystal F6, and a birefringent crystal F7. Specifically, the polarizing optical element section 36 of the modified example 4 includes a birefringent crystal F6 and a birefringent crystal F7 instead of the birefringent crystal F1 in the configuration of the embodiment 3. Therefore, the birefringent crystal F6 and the birefringent crystal F7 are arranged in the area where the reflected light R1 of the focusing optical system 20 is focused or diffused. The birefringent crystal F6 and the birefringent crystal F7 include a uniaxial birefringent crystal having a parallel plate shape. The birefringent crystal F6 and the birefringent crystal F7 each have two parallel plate surfaces. The crystal optical axis of the birefringent crystal F6 and the crystal optical axis of the birefringent crystal F7 are parallel to the plate surfaces. The birefringent crystal F6 and the birefringent crystal F7 are arranged so that the plate surfaces are perpendicular to the optical axis C of the focusing optical system 20. The crystal optical axis of the birefringent crystal F6 and the crystal optical axis of the birefringent crystal F7 are perpendicular to the optical axis C of the focusing optical system 20 and are mutually perpendicular.

The birefringent crystal F6 and the birefringent crystal F7 have birefringence opposite to that of the birefringent crystal F2. Specifically, the birefringent crystal F6 and the birefringent crystal F7 and the birefringent crystal F2 have birefringence opposite to each other, n ₀ <n _e and n ₀ >n _e . For example, if the birefringent crystal F6 and the birefringent crystal F7 are negative crystals (n ₀ >n _e ), the birefringent crystal F2 is a positive crystal (n ₀ <n _e ). If the birefringent crystal F6 and the birefringent crystal F7 are positive crystals (n ₀ <n _e ), the birefringent crystal F2 is a negative crystal (n ₀ >n _e ). Thus, in the third embodiment and the first to fourth modifications, the polarizing optical element units 32 to 36 include birefringent crystals having birefringence opposite to each other, n ₀ <n _e and n ₀ >n _e . Other configurations and effects are included in the descriptions of the first and second embodiments and the modified examples.

(Embodiment 4)
Next, an ellipsometer according to the fourth embodiment will be described. FIG. 24 is a configuration diagram illustrating an ellipsometer according to the fourth embodiment. As shown in FIG. 24, in the ellipsometer 4, the polarizer section 15 includes a linear polarizer H1, and the analyzer section 41 includes a linear polarizer H2. The polarizing optical element section 37 includes Nomarski prisms 131 and 132. The polarizing optical element section 37 includes a Nomarski prism 131 including a uniaxial birefringent crystal having positive birefringence where n ₀ <n _e and a uniaxial birefringent crystal having negative birefringence where n ₀ >n _e , and a Nomarski prism 132 including a uniaxial birefringent crystal having positive birefringence and a uniaxial birefringent crystal having negative birefringence. Thus, the polarization optical element 37 separates the reflected light R1, which is the illumination light L1 including linear polarization and reflected by the measurement surface of the sample 50, into two linearly polarized components with mutually orthogonal linear polarization directions. The analyzer 41 transmits the two linearly polarized components in directions different from the polarization directions, causing them to interfere with each other and form interference fringes. The image detector 42 detects the interference fringes. The analyzer 43 calculates the ellipsometry coefficients Ψ and Δ from the detected interference fringes.

25 is a diagram illustrating the interference fringes of the reflected light R1 that interferes on the image detector 42, and the X-polarized light and the Y-polarized light in the ellipsometer 4 according to the fourth embodiment. FIG. 25 also shows the interference fringes of the reflected light R1 that interferes on the image detector 142 in the ellipsometer 101 of the comparative example. Unlike the first to third embodiments, the ellipsometer 4 of this embodiment forms the interference fringes of the X-polarized light and the Y-polarized light in a one-dimensional vertical stripe on the pupil. However, by using two or more types of crystals with different birefringence, the separation angle of the two polarized lights is changed for each wavelength. More specifically, the Nomarski prisms 131 and 132 are arranged so that the separation angle of the two polarized lights just before the image detector 42 is larger on the long wavelength side and smaller on the short wavelength side. Therefore, in this embodiment, by including the achromatic Nomarski prism 132, it is possible to suppress the decrease in contrast at both ends in the X-axis direction compared to the comparative example. In this way, the change in the interval between the interference fringes due to the difference in wavelength is reduced. Therefore, even when using a broadband light source for measurements, it is possible to narrow the spacing between interference fringes and widen the wavelength width compared to previous methods.

Fig. 26 is a graph illustrating the wavelength dependency of birefringence in each material of birefringent crystal, where the horizontal axis indicates wavelength and the vertical axis indicates birefringence. In Fig. 26, the value of (n ₀ -n _e )/{(n ₀ +n _e )/2} is normalized to 1 when the wavelength is 400 nm, and the wavelength dependency of the polarization separation angle is shown when a Nomarski prism is made of each material of αBBO, quartz, magnesium fluoride (MgF ₂ ), sapphire (Al ₂ O ₃ ), and calcite (Calcit). As shown in Fig. 26, magnesium fluoride has a relatively small wavelength dependency of the polarization separation angle, and calcite has a large wavelength dependency of the polarization separation angle. The fringe pitch of the interference fringes is the same value for multiple wavelengths of illumination light.

Figure 27 is a diagram illustrating the polarizing optical element section 37 according to embodiment 4. As shown in Figure 27, if the Nomarski prism 132 is made of calcite and the Nomarski prism 131 is made of magnesium fluoride, the polarization separation angle just before the image detector 42 will be large on the long wavelength side and small on the short wavelength side. By designing in this way, it is possible to reduce the wavelength dependency of the spacing between interference fringes. Other configurations and effects are included in the descriptions of embodiments 1 to 3 and the modified examples.

The present invention is not limited to the above-described embodiments, and can be modified as appropriate without departing from the spirit of the present invention. For example, the configurations of embodiments 1 to 4 and each of the modified examples can be combined with each other. In addition, semiconductor device inspection devices equipped with the ellipsometers of embodiments 1 to 4 and each of the modified examples are also included in the technical concept of the present embodiments.

Reference Signs List 1, 2, 3, 4 Ellipsometer 1A Inspection device 10 Illumination optical system 11 Light source 12 Spectroscope 13 Fiber 14 Illumination lens 15, 18 Polarizer section 16 Beam splitter 17 Objective lens 20 Condensing optical system 21, 22 Relay lens 23 Pupil position 24 Pupil conjugate position 25, 26 Axicon lens 30, 31, 32, 33, 34, 35, 36, 37 Polarizing optical element section 40 Light receiving optical system 41, 44 Analyzer section 42 Image detector 43 Analysis device 50 Sample 80 Base 81 Isolator 82 Optical surface plate 83 Stage 84 Wafer holder 85 Frame 86 Environmental chamber 87 Temperature controller unit 88 Automatic substrate transport device 91 Control section 92 Grabber board 93 Stage controller 94 Light source/spectroscope controller 101 Ellipsometer 110 Illumination optical system 111 Light source 112 Spectrometer 113 Fiber 114 Illumination lens 115 Polarizer 116 Beam splitter 117 Objective lens 120 Light collecting optical system 121, 122 Relay lens 130 Polarizing optical elements 131, 132 Nomarski prism 140 Light receiving optical system 141 Analyzer 142 Image detector 143 Analysis device C Optical axis H1, H2 Linear polarizer H3, H4 λ/4 wave plates F1, F2 Birefringent crystal L1 Illumination light R1 Reflected light W10, W20 Wollaston lens

Claims (20)

a polarizing optical element unit that separates two polarized components having mutually orthogonal polarization directions of reflected light, which is obtained by reflecting illumination light including at least one of linearly polarized light, circularly polarized light, and elliptically polarized light off a measurement surface of a sample, into radial directions centered on an optical axis of an optical system of the reflected light;
an analyzer unit that transmits components of the two polarized components in directions different from the polarization directions of the two polarized components, thereby causing interference and forming concentric interference fringes;
an image detector for detecting the interference fringes;
an analyzer for calculating ellipsometry coefficients Ψ and Δ from the detected interference fringes;
An ellipsometer equipped with A light source that generates the illumination light;
a polarizer unit to which the illumination light generated by the light source is incident and which transmits the illumination light including at least one of the linearly polarized light, the circularly polarized light, and the elliptically polarized light;
an objective lens that transmits the illumination light and transmits the reflected light that is reflected by the measurement surface of the sample from the illumination light;
a focusing optical system for imaging an exit pupil of the objective lens onto the image detector;
10. The ellipsometer of claim 1 further comprising: The polarizer unit includes a linear polarizer that generates linearly polarized light,
The analyzer unit includes a linear polarizer that transmits a linearly polarized component in a predetermined direction.
3. The ellipsometer of claim 2. the polarizer unit includes a circular polarizer that generates circularly polarized light that rotates left or right around an optical axis of the illumination light,
The analyzer unit includes a circular polarizer that transmits a circularly polarized component that rotates either left or right.
3. The ellipsometer of claim 2. Each of the circular polarizers includes a linear polarizer and a λ/4 wave plate.
5. The ellipsometer of claim 4. the polarizing optical element portion includes a Wollaston lens,
The Wollaston lens includes two uniaxial birefringent crystals having interdigitated spherical shapes;
The crystal optical axis of each of the uniaxial birefringent crystals is perpendicular to the optical axis of the optical system and perpendicular to each other.
4. The ellipsometer of claim 3. The polarizing optical element portion further includes another Wollaston lens,
the two polarized components of the reflected light transmitted through the polarizing optical element unit and the analyzer unit are detected at the same point on the image detector;
7. The ellipsometer of claim 6. the polarization optical element unit includes a first birefringent crystal having a parallel plate shape arranged at a position in the optical system where the reflected light is condensed or diffused,
the first birefringent crystal is a uniaxial birefringent crystal,
The crystal optical axis of the first birefringent crystal is parallel to the optical axis of the optical system.
6. An ellipsometer according to claim 4 or 5. The polarizing optical element unit is
a first birefringent crystal having a parallel plate shape arranged at a position in the optical system where the reflected light is focused or diffused;
a second birefringent crystal having a concave conical entrance surface and a convex conical exit surface;
Including,
the first birefringent crystal and the second birefringent crystal are uniaxial birefringent crystals,
a crystal optical axis of the first birefringent crystal and a crystal optical axis of the second birefringent crystal are parallel to an optical axis of the optical system;
6. An ellipsometer according to claim 4 or 5. The polarizing optical element unit is
a first birefringent crystal having a parallel plate shape arranged at a position in the optical system where the reflected light is focused or diffused;
a third birefringent crystal having a parallel plate shape arranged at a position in the optical system where the reflected light is focused or diffused;
Including,
the first birefringent crystal and the third birefringent crystal are uniaxial birefringent crystals,
a crystal optical axis of the first birefringent crystal and a crystal optical axis of the third birefringent crystal are parallel to an optical axis of the optical system;
6. An ellipsometer according to claim 4 or 5. The polarizing optical element unit is
a fourth birefringent crystal that is disposed in the optical system at a position where the reflected light is condensed or diffused, the fourth birefringent crystal having an incident surface that is a concave spherical shape and an exit surface that is a convex spherical shape;
a second birefringent crystal having a concave conical entrance surface and a convex conical exit surface;
Including,
the second birefringent crystal and the fourth birefringent crystal are uniaxial birefringent crystals,
a crystal optical axis of the second birefringent crystal and a crystal optical axis of the fourth birefringent crystal are parallel to an optical axis of the optical system;
6. An ellipsometer according to claim 4 or 5. The polarizing optical element unit is
a first birefringent crystal having a parallel plate shape arranged at a position in the optical system where the reflected light is focused or diffused;
a fifth birefringent crystal having the parallel plate shape;
a first glass substrate disposed on the incident surface side of the fifth birefringent crystal, the first glass substrate having an incident surface in a concave conical shape and an exit surface in a planar shape;
a second glass substrate disposed on the exit surface side of the fifth birefringent crystal, the second glass substrate having an entrance surface in a planar shape and an exit surface in a convex cone shape;
Including,
the first birefringent crystal and the fifth birefringent crystal are uniaxial birefringent crystals,
The crystal optical axis of the first birefringent crystal and the crystal optical axis of the fifth birefringent crystal are parallel to the optical axis of the optical system.
6. An ellipsometer according to claim 4 or 5. The polarizing optical element unit is
a sixth birefringent crystal having a parallel plate shape and arranged at a position in the optical system where the reflected light is focused or diffused;
a seventh birefringent crystal having a parallel plate shape and arranged at a position in the optical system where the reflected light is focused or diffused;
a second birefringent crystal having a concave conical entrance surface and a convex conical exit surface;
Including,
the second birefringent crystal, the sixth birefringent crystal, and the seventh birefringent crystal are uniaxial birefringent crystals;
a crystal optical axis of the second birefringent crystal is parallel to an optical axis of the optical system;
The crystal optical axis of the sixth birefringent crystal and the crystal optical axis of the seventh birefringent crystal are perpendicular to the optical axis of the optical system and perpendicular to each other.
6. An ellipsometer according to claim 4 or 5. The polarization optical element unit includes the birefringent crystal having mutually opposite birefringence, that is, a refractive index n ₀ of ordinary light is smaller than a refractive index n _e of extraordinary light, and a refractive index n ₀ of ordinary light is larger than a refractive index n _e of extraordinary light.
The ellipsometer according to any one of claims 9 to 13. a polarization optical element unit that separates the reflected light, which is illumination light including linear polarization and is reflected by a measurement surface of a sample, into two linearly polarized components having polarization directions perpendicular to each other;
an analyzer unit that transmits the two linearly polarized components in directions different from the polarization directions, causes them to interfere with each other, and forms interference fringes;
an image detector for detecting the interference fringes;
an analyzer for calculating ellipsometry coefficients Ψ and Δ from the detected interference fringes;
Equipped with
the fringe pitch of the interference fringes is the same for a plurality of wavelengths of the illumination light;
Ellipsometer. The polarizing optical element unit is
A first Nomarski prism including a uniaxial birefringent crystal having positive birefringence, where the refractive index of ordinary light _{is n0} < the refractive index of extraordinary light, _n2 , and a uniaxial birefringent crystal having negative birefringence, where the refractive index of ordinary light is _n0 > the refractive index of extraordinary light, n2,
a second Nomarski prism including the uniaxial birefringent crystal having the positive birefringence and the uniaxial birefringent crystal having the negative birefringence;
Including,
16. The ellipsometer of claim 15. the first Nomarski prism and the second Nomarski prism are arranged such that a separation angle of the two polarized lights immediately before the image detector is large on the long wavelength side and small on the short wavelength side;
17. The ellipsometer of claim 16. The birefringent crystal includes any one of quartz, magnesium fluoride, sapphire, calcite, and αBBO;
The ellipsometer according to any one of claims 6 to 14 and 16 to 17. The analysis device includes:
The image including the concentric interference fringes is transformed into polar coordinates so that the center of the concentric circle becomes the coordinate origin;
A two-dimensional Fourier transform is performed on a two-dimensional coordinate plane having the radius and the deviation angle of the polar coordinates as axes;
In the frequency space image after the two-dimensional Fourier transform, a DC component and an AC component are separated and trimmed;
Ψ and Δ are calculated based on the results of the inverse Fourier transform of the trimmed area.
The ellipsometer according to any one of claims 1 to 14. An ellipsometer comprising the ellipsometer according to any one of claims 1 to 19.
Inspection equipment for semiconductor devices.

Images