CN103162831B - broadband polarization spectrometer and optical measurement system - Google Patents

Abstract

A broadband polarization spectrometer provided by the present invention includes a light source, a first reflection unit, a first light collection unit, a second light collection unit, a polarizer, a first plane reflector, a first curved reflector, a second reflection unit and a detector unit. The vertically incident broadband polarization spectrometer including a reference beam provided by the present invention improves the luminous flux efficiency of the probe beam, and realizes the complete combination of the split non-detection beam as the reference beam and the probe beam at the front end of the detector. In addition, the probe beam can be focused on the sample surface without chromatic aberration, and the polarization state of the probe beam can be maintained at the same time, and the complexity of the system is lower than that of the prior art. The present invention also provides an optical measurement system comprising the above-mentioned spectrometer.

Description

Broadband polarization spectrometer and optical measurement system

technical field

The invention relates to the field of optics, in particular to a broadband polarization spectrometer and an optical measurement system.

Background technique

With the rapid development of the semiconductor industry, the use of optical measurement technology to quickly and accurately detect the thickness, material properties and three-dimensional shape of the periodic structure of the semiconductor film is the key link to control the production process and improve productivity. It is mainly used in integrated circuits and flat panels. Displays, hard disks, solar cells and other industries that contain thin film structures. Films made of different materials and structures have different reflectivities for incident light of different polarization states at different wavelengths, and their reflection spectra are unique. Today's advanced thin film and three-dimensional structure measurement equipment, such as ellipsometer and Optical Critical Dimension (OCD), require as wide a spectrum measurement capability as possible to increase measurement accuracy, usually 190nm to 1000nm. When the structural parameters of the film are known, the reflectance spectrum of the film can be calculated through a mathematical model. When there are unknown structural parameters, such as film thickness, film optical constants, surface three-dimensional structure, etc., the unknown structural parameters can be obtained through regression analysis, fitting measurement and simulated calculation spectra.

Generally speaking, there are usually two methods for optical measurement of semiconductor thin films, absolute reflectance measurement method and ellipsometry method. As stated in Chinese patent application 201110032744.8, when using the absolute reflectance measurement method to measure, it is necessary to use a standard sample for measurement first, and record the measurement result of the standard sample as a reference value, then measure the sample to be tested, and the The measurement result is compared with the reference value measured by the standard sample, so as to obtain the relative true value of the sample to be tested. Due to the light source itself, its spectral intensity may change (drift) during actual measurement. In theory, it is generally assumed that the spectral intensity of the light source is exactly the same when measuring the standard sample and the sample to be tested, but in practice, since the sample to be tested and the standard sample cannot be measured at the same time, changes in the spectral intensity of the light source will affect the measurement results.

In view of the above reasons, those skilled in the art propose to use a reference beam to calibrate the light source fluctuation. The light emitted by the light source is divided into two beams, one beam is used as the probe light to record the optical information of the sample, and the other beam is used as the reference beam. Through the measurement of the reference beam, the spectra of the light source when measuring the reference sample and the sample to be tested can be recorded respectively. Intensity, thereby correcting the spectral intensity change of the light source during the measurement process and improving the measurement accuracy.

Measuring equipment is generally divided into optical systems with normal incidence relative to the sample surface and optical systems with oblique incidence relative to the sample surface. Due to its more compact structure, the vertical incidence optical system can usually be integrated with other process equipment to realize the integration of production and measurement and real-time monitoring. In the prior art, there are mainly the following two methods for realizing the normal incidence spectrometer calibrated by the reference beam:

(1) As shown in Figure 1, the divergent light emitted by the light source 101 passes through the lens 102 and then enters the beam splitter 103 in parallel, the light transmitted through the beam splitter 103 is used as the detection beam, and the light reflected by the beam splitter 103 is used as the reference beam . The probe beam is converged by the lens 104 and then focused to the surface of the sample 105. The reflected light on the surface of the sample 105 is reflected by the lens 104 and then vertically enters the beam splitter 103. The probe beam reflected by the beam splitter 103 is converged by the lens 107 and is incident on the detector. 108. Obtain the reflectance spectrum of the sample surface; the reference beam is vertically incident on the plane reflector 106, and after being reflected by the plane reflector 106, it is vertically incident on the beam splitter 103, and the reference beam transmitted through the beam splitter 103 is also converged by the lens 107 and incident on the detector The device 108 obtains a reference spectrum including the spectral characteristics of the light source (for example, see US Patent No. 7067818B2, No. 7189973B2 and No. 7271394B2, US Patent Application Publication No. 2005/0002037A1). In such spectrometers, the control aperture can be used to select the beam to be measured. This method has the following advantages: it can calibrate the fluctuation of the light source, but due to the use of the beam splitter, this spectrometer also has the following problems: ① The luminous flux is low, and the light beam needs to be transmitted and reflected by the light source through the same beam splitter once during the entire measurement process , into the detector. Assuming that the transmittance and reflectivity of the beam splitter are 50% each, the maximum luminous flux ratio that can be achieved between the probe beam and the reference beam is 25%; Complexity and cost are high.

(2) Insert a plane reflector in the optical path, so that part of the light emitted by the light source is incident on the plane reflector, and the other part passes through the edge of the plane reflector. The light beam reflected by the plane mirror is vertically incident on the sample surface as the probe light, and the beam passing through the edge of the plane mirror is used as the reference beam. The probe beam and the reference beam enter two different spectrometers for simultaneous measurement (for example, see Patent No.5747813 and No.6374967B1). This method has the following advantages: the detection beam and the reference beam can be measured simultaneously during the measurement process, and the spectrum and intensity changes of the light source are accurately corrected; the loss of light intensity in the system is small and the utilization rate is high. However, due to the use of two different spectrometers, the photoelectric conversion efficiency is not the same, the wavelength distribution and resolution are also different, it is not easy to calibrate the system, but will reduce the measurement accuracy. On the other hand, the optical path structure of this scheme is relatively It is complicated and difficult to adjust, and two spectrometers will increase the volume of the equipment and increase the cost of the equipment.

When inspecting a wafer with a typical size of 150 mm, 200 mm or 300 mm, the wafer surface may be uneven due to, for example, film layer stress on the wafer. Therefore, when inspecting the entire wafer, in order to achieve high-precision measurement and ensure rapid measurement of semiconductor production line yield, automatic focusing on each measurement point is one of the key technologies. Moreover, it is known to those skilled in the art that it is advantageous to focus the broadband probe beam into a relatively small-sized spot on the sample surface, because the small-sized spot can measure microstructure patterns, and the broadband probe beam can improve measurement accuracy. In this case, when a lens is used for focusing, there will be the following problems: the lens usually has chromatic aberration, and such chromatic aberration will cause different focusing positions of light of different wavelengths, increase errors, and reduce measurement accuracy; moreover, it is difficult to find Lens material with good transmission over the entire broadband wavelength range. In view of the above reasons, those skilled in the art have proposed such a method, using curved mirrors, such as elliptical mirrors, toroidal mirrors (toroidal mirrors), off-axis parabolic mirrors, etc. to focus the broadband probe beam to on the sample surface (see, eg, US Patent Nos. 5608526 and 7505133B1, US Patent Application Publication No. 2007/0247624A1 and Chinese Patent Application Publication No. 101467306A). This approach has the advantage that the mirror does not produce chromatic aberration over the entire broadband wavelength range, and the mirror can have high reflectivity over a wide wavelength range. However, the polarization state of the beam changes after reflection from a single mirror. Here we take an aluminum reflector as an example. The reflectances Rs and Rp for S and P polarized light are shown in Figure 2a for two incident angles. The upper two curves are the reflectivity Rs of S-polarized light, and the lower two curves are the reflectivity Rp of P-polarized light. The solid line corresponds to an angle of incidence of 45 degrees and the dashed line corresponds to an angle of incidence of 50 degrees. It can be seen that the reflectivity of S or P polarized light is not equal, and changes with different incident angles. The phase difference between the reflected S and P polarized light is shown in Fig. 2b, the solid line corresponds to an incident angle of 45 degrees, and the dashed line corresponds to an incident angle of 50 degrees. It can be seen that the phase difference between the reflected S and P polarized light changes, and changes with the incident angle, and is related to the wavelength. In short, when the broadband beam is reflected by the mirror, since the polarization states S and P with orthogonal polarization directions have different reflectivity and phase changes, the polarization state of the beam changes, making it difficult to control the polarization change of the beam (for example, See US Patent Nos. 6829049B1 and 6667805).

The ability of the spectrometer to control the polarization limits the application range of the spectrometer. For example, the optical critical dimension device OCD, which is widely used in the process control of integrated circuit production lines, measures the critical dimension (CD) and Three-dimensional morphology and film thickness and optical constants of multilayer materials. The spectrometer to achieve critical scale measurement requires its focusing system to control the polarization state of the beam during the focusing and optical signal acquisition process, so that the sample can be accurately measured.

Contents of the invention

The technical problem to be solved by the present invention is to provide a broadband polarization spectrometer and an optical measurement system that can realize no chromatic aberration, maintain polarization characteristics, have high luminous flux efficiency and are easy to implement.

According to one aspect of the present invention, there is provided a broadband polarization spectrometer comprising a light source, a first reflecting unit, a first light concentrating unit, a second light concentrating unit, a polarizer, a first plane reflector, a first curved reflector, a second reflection unit, detection unit, wherein:

The first reflection unit is used to divide the light beam emitted by the light source into two parts, a detection beam and a reference beam, and inject the two parts of the beam into the first light concentrating unit and the second light concentrating unit respectively;

The second reflective unit is used to simultaneously or separately receive the detection beam reflected from the sample and sequentially pass through the first curved reflector, the first plane reflector, the polarizer, the first condensing unit, and passing through the reference beam of the second condensing unit, and incident the received beam to the detection unit;

The first light concentrating unit is used to receive the detection beam, make the beam into a parallel beam and then enter the polarizer;

The polarizer is disposed between the first light concentrating unit and the first plane reflector, for allowing the parallel light beam to pass through and enter the plane reflector;

The first plane reflector is used to receive the parallel beam and reflect the parallel beam to the first curved reflector;

The first curved reflector is used to turn the parallel beam into a converging beam, and reflect the converging beam to vertically focus on the sample;

The second concentrating unit is used to receive the reference beam and make it incident to the second reflecting unit;

The detecting unit is used for detecting the light beam reflected by the second reflecting unit.

According to another aspect of the present invention, there is provided an optical measurement system including the normal incidence broadband spectrometer.

The broadband polarization spectrometer including the reference beam provided by the present invention realizes the complete combination of the split beams, improves the light flux efficiency of the probe beam, and also enables the probe beam to be focused on the sample surface without chromatic aberration while maintaining the polarization of the beam state, and the complexity of the system is lower than that of the prior art.

Description of drawings

FIG. 1 is a schematic diagram of light splitting and light combining realized by a light splitter in the prior art.

Figure 2a is a schematic diagram of the reflectivity Rs and Rp of S and P polarized light under two incident angles.

Fig. 2b is a schematic diagram illustrating the phase difference between reflected S and P polarized light.

3a to 3c are schematic diagrams for explaining the basic principle of maintaining the polarization characteristics of polarized light by two plane mirrors or one plane mirror and one off-axis parabolic mirror.

Fig. 4 is a schematic diagram of the calculated point distribution in the incident light cross-section.

Fig. 5 is a schematic diagram of the position where the maximum value of the quadratic curve can be obtained by curve fitting from the measured values of the first three different positions (ie, positions A, B and C).

6 is an optical schematic diagram showing a Rochon prism polarizer (Rochon Polarizer), in which, RP represents a Rochon prism polarizer, A represents an aperture, and S represents a sample.

7a and 7b are schematic diagrams of splitting light through two non-coplanar plane mirrors in the present invention.

Fig. 8 is a schematic diagram of combining light through two non-coplanar plane mirrors in the present invention.

Fig. 9a and Fig. 9b are the cross-sectional shape of the light beam and the image formed by the light beam obtained through simulation.

Fig. 10 is a structural diagram of a periodic shallow trench in single crystal silicon.

Fig. 11 is the absolute reflectance spectrum of periodic shallow trenches TE and TM in monocrystalline silicon in the absolute reflectance measurement method.

Fig. 12a is a schematic diagram showing a normal incidence broadband polarization spectrometer according to the first embodiment of the present invention.

Fig. 12b is an optical path diagram of focusing and imaging the sample surface and the probe beam by using the beam splitter and the pattern recognition system in the present invention.

Figures 13a and 13b are patterns observed by the pattern recognition system.

Fig. 14 is a schematic diagram showing a normal-incidence broadband polarization spectrometer according to a second embodiment of the present invention.

Fig. 15 is a schematic diagram showing a normal-incidence broadband polarization spectrometer according to a third embodiment of the present invention.

Detailed ways

Aiming at the problems existing in the prior art, the present invention proposes a vertical-incidence broadband polarization spectrometer that can achieve no chromatic aberration, maintain polarization characteristics, have high luminous flux efficiency and is easy to implement and is calibrated with a reference beam. The maximum luminous flux ratio of the probe beam and the reference beam in the spectrometer can reach 50%, and the vertical incidence broadband polarization spectrometer of the present invention only includes a spectrometer, therefore, the spectrometer measurement accuracy proposed by the present invention is higher, while the complexity and equipment The cost is lower than that of the prior art. In addition, the normal incidence band spectrometer contains at least one polarizer, so that it can accurately measure the three-dimensional topography and material optical constants of anisotropic or inhomogeneous samples, such as thin films containing periodic structures.

The basic principle of maintaining the polarization properties of polarized light by two plane mirrors or one plane mirror and one off-axis parabolic mirror is explained below with reference to FIGS. 3a, 3b and 3c.

As shown in Figure 3a, it is assumed that the S (or P) polarized beam with the incident surface of M1 as the reference is incident on the first plane mirror M1 at an incident angle of (90-θ) degrees, and is reflected by the first plane mirror M1 to the second plane mirror M2. When the incident plane of the first plane mirror M1 and the incident plane of the second plane mirror M2 are perpendicular to each other, and the inclination of M2 is such that the reflected light of M1 enters M2 at an incident angle of (90-θ) degrees, it is reflected by M1 The S (or P) polarized light with the M1 incident plane as the reference is transformed into the P (or S) polarized light with the M2 incident plane as the reference.

Now analyze the propagation of the beam and the change of the polarization state with the right-hand reference frame determined by the beam propagation direction as the +Z direction. Express the above process in a mathematical formula:

$\left\{\begin{array}{c}\mathrm{<span\; class="notranslate">\; Ex</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}\\ \mathrm{<span\; class="notranslate">\; Ey</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; p</span>}}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; a</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span>$

The polarization components E _1s and E _1p with reference to the incident plane of M1 are respectively defined as +X and +Y direction components in the right-hand reference system. After being reflected by M1,

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}\\ {\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; p</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; p</span>}}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; p</span>}}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; b</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span>$

E′ _1s , E′ _1p are the reflected light polarization components with reference to the incident surface of M1 respectively; among them, r _1s and r _1p are the polarization components of S and P light with reference to the incident surface of M1 respectively in terms of (90-θ) Angle of incidence on the reflectivity of the first plane mirror M1. and,

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; p</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\\ {\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; c</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span>$

E' _1s and E' _1p reflected by M1 are the incident polarization components -E _2p and E _2s with reference to the incident plane of M2, respectively. After being reflected by M2,

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}\\ {\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; p</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; p</span>}}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; p</span>}}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; d</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span>$

E′ _2s , E _2p are the reflected light polarization components with reference to the M2 incident surface, r _2s and r _2p are respectively the S and P light polarization components with the M2 incident surface as the reference, incident at the angle of (90-θ) The reflectivity of the second plane mirror M2.

Due to the right-hand rule, the polarization direction of the S light with reference to the incident surface of M1 is the negative direction of the P light with reference to the incident surface of M2. It is stipulated that in the right-hand reference system determined by the beam propagation direction as the +Z direction, the polarization component of the S light with the M1 incident plane as the reference is always the +X axis. After the light beam is reflected by M2, the polarization direction of P light with reference to the incident surface of M2 is the positive direction of the X axis; thus, the polarization direction of S light with reference to the incident surface of M2 is the negative direction of the Y axis. have:

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; p</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\\ {\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}^{<span\; class="notranslate">\; \&prime;</span>}<span\; class="notranslate">\; =</span><span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}^{<span\; class="notranslate">\; \&prime;</span>}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; e</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span>$

E' _x , E' _y are the polarization components of the outgoing light. In the case that M1 and M2 have the same reflective material and coating structure:

$\left\{\begin{array}{c}{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; the\; s</span>}}\\ {\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 1</span>}\mathrm{<span\; class="notranslate">\; p</span>}}<span\; class="notranslate">\; =</span>{\mathrm{<span\; class="notranslate">\; r</span>}}_{\mathrm{<span\; class="notranslate">\; 2</span>}\mathrm{<span\; class="notranslate">\; p</span>}}\end{array}\right.<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span>$

Combining the above formulas are:

$<span\; class="notranslate">\; \{</span>\frac{{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}}{{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; x</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}{{\mathrm{<span\; class="notranslate">\; E.</span>}}_{\mathrm{<span\; class="notranslate">\; the\; y</span>}}^{<span\; class="notranslate">\; \&prime;</span>}}<span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; -</span><span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; g</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; .</span>$

In the above formulas (a)-(g), all variables are complex numbers. It can be seen from formula (g) that the ratio of the polarization components of the outgoing light is equal to the ratio of the polarization components of the incident light. Therefore, the polarization characteristics of the polarized light can be maintained by the above two plane mirrors.

According to the above formulas (a)-(e), those skilled in the art know that as long as the first plane mirror M1 and the second plane mirror M2 satisfy the relationship of r _2s r _1p =r _2p r _1s , the formula ( g) relationship. That is to say, if the two mirrors satisfy the relationship of r _2s r _1p =r _2p r _1s , the polarization characteristics of the polarized light can be maintained through the two mirrors. It can be seen that the system composed of two plane mirrors whose incident planes are perpendicular to each other and have the same incident angle can perfectly maintain the polarization characteristics of the incident light.

Under the assumption that one of the above two planar mirrors is replaced by an off-axis parabolic mirror with the same reflective material and coating structure, simulation calculations are carried out for the case of small numerical aperture (NA, numerical aperture). Although there will be deviations in polarization characteristics after the beam passes through the system composed of planar mirrors and off-axis parabolic mirrors, when the parallel beams are focused with a small NA, the deviation in polarization characteristics is not enough to affect the accuracy of the measurement. For stringent polarization requirements, the measurement results can be further corrected using numerical calculations.

If the first plane mirror M1 of the above two plane mirrors is replaced by an off-axis parabolic mirror OAP, taking Figure 3b as an example, the parallel light is circularly polarized before it is incident on the off-axis parabolic mirror OAP, that is, Ex =Ey, and Phase(Ex)-Phase(Ey)=90 degrees, where Ex and Ey are the amplitudes of the electric vectors of the light beam in the x and y directions respectively, and Phase(Ex) and Phase(Ey) are the The phase of the electric vector in the x and y directions. After being focused by the off-axis parabolic mirror, the half angle of the cone formed by the focused beam is 4.2 degrees (NA=0.073). The wavelength of the incident light is 210nm, and the calculated point distribution in the cross-section of the incident light is shown in Figure 4, with a total of 29 points (parts have been calibrated, for example, (0,3) to (0,0)). After numerical calculation, the intensity change and phase change of the polarization at the focal point are listed in Table 1. The polarization intensity change is defined as |Ex/Ey|-1, and the phase change is Phase(Ex)-Phase(Ey)-90. It can be seen from Table 1 that the beam symmetrical to (0,0) has a very close complementarity in polarization intensity and phase change, so when the incident light is vertically incident on the sample, it passes through the plane mirror M2, off-axis The polarization deviation on the return of the parabolic mirror OAP1 is further cancelled.

Table 1

If the second plane mirror M2 of the above two plane mirrors is replaced by an off-axis parabolic mirror OAP1, taking Figure 4c as an example, the parallel light reflected by the plane mirror M1 is focused by the off-axis parabolic mirror OAP1 , the half-angle of the cone formed by the focused beam is 6.3 degrees (NA=0.109734). When other conditions are the same as in Table 1, after numerical calculation, the intensity change and phase change of the polarization at the focal point are listed in Table 2. It can be seen from the table that there is a very close complementarity in the polarization intensity and phase change of the beam centered at (0, 0). Taking the point (-3, 0) with the largest polarization phase deviation as an example, after passing through the plane After the system composed of reflector and off-axis parabolic reflector, its phase deviation is 4.7508 degrees, but when the detection beam returns from the sample surface, the calculation point beam is reflected to point (3, 0), then the phase deviation of this point would get a fairly large (-4.3325 degrees) offset. Therefore, when the incident light is vertically incident on the sample and returns through the off-axis parabolic mirror OAP1 and the plane mirror M1, the polarization deviation will be further offset. That is, the present invention can further offset the polarization effect caused by the curved reflector as a whole.

Table 2

Therefore, adopting such a system composed of off-axis parabolic mirrors and plane mirrors can not only achieve achromatic focusing of the incident light on the sample surface, but also basically maintain the polarization characteristics of the polarized light.

The above only exemplifies the case where one of the above two plane reflectors is replaced by an off-axis parabolic reflector with the same reflective material and coating structure. Those skilled in the art should know that not only plane mirrors and off-axis parabolic mirrors, but also other curved mirrors, such as toroidal mirrors, ellipsoidal mirrors or non-quadratic mirrors, etc., any two When the above-mentioned relationship is satisfied, all kinds of reflectors can basically maintain the polarization characteristics of polarized light.

In summary, if the two mirrors have approximately the same reflective material and approximately the same coating structure and satisfy the same incident angle of the main beam and the incident planes are perpendicular to each other (within the error range allowed in this field, that is, including the incident The angles are approximately the same and the incident planes are approximately perpendicular to each other), then the polarization characteristics of any polarized light remain unchanged after passing through the two mirrors. An example of a mirror with the same reflective material and coating structure is a mirror obtained by keeping the same coating in the same vacuum chamber.

In addition, those skilled in the art can know that when the probe beam is focused on the sample surface, in order to obtain a smaller spot size and better imaging quality of the probe beam on the sample surface, the numerical aperture of the focused beam on the sample surface should be greater than or equal to that of the point light source. The numerical aperture of the beam at the point source is more favorable, and for a fixed distance, the larger the numerical aperture of the beam at the point light source, the greater the luminous flux of the beam actually collected. Therefore, the focusing method in Figure 4c can be compared with Figure 4b. The detection beam can obtain a larger numerical aperture, that is, a larger luminous flux can be obtained, and measurement accuracy can be improved.

As described below, there are two ways to achieve focusing in the broadband spectrometer of the present invention.

The first method achieves focusing by observing the signal intensity variation of the collected reflected light. Compared with the in-focus state, after the position of the spectrometer slit is calibrated, the out-of-focus will cause part of the light at the periphery of the spot to be lost in the optical collection system. On the basis of preliminary focusing, the most precise focusing can be obtained by finding the maximum value of the optical signal. The mathematical method and basic steps for quickly finding the focus can be: in the vicinity of the focus, the relationship between the optical signal intensity and the defocus distance is approximated as a quadratic curve, that is, a parabola: I=-A(xx ₀ ) ² +B, Among them, I is the optical signal intensity, x ₀ is the focal point position, and A and B are coefficients. As shown in Figure 5, the position of the maximum value of the quadratic curve can be obtained by curve fitting through the measured values of the first three different positions (that is, A, B and C positions); the measured value of this position is a new value point, the curve can be fitted again; this method is iterated until |x _n+1 -x _n |<σ is theoretically satisfied, where x _n is the position of the nth focus, and x _n+1 is the nth increase In the case of the measured value of the focus position, the n+1th focus position is fitted, and σ is the accuracy of the system adjustment.

The second method is to achieve focusing by observing the imaging sharpness of the sample surface in the pattern recognition system. In the ideal focus state, after the position of the pattern recognition system is calibrated, the sample surface has the clearest image when it is in focus. When the image resolution is determined, the sharpness of the image is determined by the sharpness of the image. Sharpness indicates the contrast of the edges of an image. More precisely, sharpness is the magnitude of the derivative of brightness with respect to space. On the basis of the initial focus (ie the sample surface is identifiable in the pattern recognition system), the image sharpness can be calculated synchronously by adjusting the focus. In this way, the most precise focus can be obtained by combining the above mathematics and basic steps of finding focus quickly

As the polarizer used in the present invention, a Rochester prism polarizer RP as shown in FIG. 6 can be used. The material of the Rochester prism polarizer can be MgF ₂ , a-BBO, calcite, YVO ₄ or quartz. The Rochonian prism polarizer uses a birefringent crystal (the refractive index of the o light and the e light is different) to make the two beams of polarized light in the orthogonal direction of the incident beam pass through the interface of the Rochonian prism at a certain angle, where the o light and the incident direction Keeping it consistent, the light exits in a linearly polarized state. Different materials have different transmission spectrum ranges, _MgF2 can reach the spectral range of 130-7000nm. Since different materials have different refractive indices of o-ray and e-ray, the included angles of o-ray and e-ray in the transmitted light are also different. For example, for MgF ₂ or quartz, the angle between o-light and e-light is 1 to 2 degrees, however, for a-BBO or YVO ₄ the angle can be as high as 8 to 14 degrees. This angle also depends in part on the cutting angle θ of the Rochester prism. After the detection beam is transmitted through the polarizer, the o light is vertically incident on the sample S, and the e light is obliquely incident on the sample S at an angle α; when the reflected beam of the e light on the sample surface can enter the optical aperture range of the polarizer, The reflected beam of its e-light can also bounce off the polarizer and then into the detector, affecting the measurement. For a polarizer with a large e-light deflection angle, the reflected light of the e-light on the sample surface is not easy to re-enter the polarizer. In order to improve the measurement accuracy and avoid the influence of the reflected light of the e-light, an aperture A (as shown in Figure 6) can be set at the position above the sample surface where the o-light and e-light are separated, so as to prevent the e-light from incident on the sample surface or Its reflected light bounces back to the polarizer.

Referring to FIG. 7a , FIG. 7b and FIG. 8 , the process of realizing light splitting and light combining respectively through the first reflection unit and the second reflection unit each including two non-coplanar plane mirrors will be described below.

(1) Realize light splitting: As shown in Figure 7a, it is assumed that the divergent light beam from the point light source passes through the light source concentrating unit, such as the toroidal reflector (M6) (or achromatic lens), to form a converging light beam, and After being deflected in the incident plane, it enters the first reflection unit. The first reflection unit is composed of two non-coplanar plane mirrors M4 and M5. The plane mirror M4 contains a straight line edge, and the straight line edge is in the optical path of the above-mentioned converging light beam. Half of the converging light beam is incident on the plane reflection On the mirror M4, it is deflected in the incident plane after being reflected by the plane mirror M1 to form a converging beam B1. The other part of the converging beam passes through the straight edge of the plane mirror M1 and is incident on the plane mirror M5. After being reflected by the plane mirror M5, it is deflected in the incident plane to form a converging beam B2. The main axis direction of the plane mirror M5 is slightly inclined relative to the plane mirror M4 in the incident plane, so that the main beams of the converging beams B1 and B2 respectively reflected by the plane mirrors M4 and M5 first intersect and then separate, as shown in Figure 7a as shown; alternatively, the converging beams B1 and B2 are directly separated, as shown in Figure 7b. From then on, the light from the point light source passes through the first reflection unit, that is, the plane mirrors M4 and M5, and is divided into two beams that can be respectively used as the detection light and the reference light. Before and after splitting, the main light of the two beams of light is always in the same plane, and the straight edge of the plane mirror M5 is perpendicular to the plane.

In addition, if the emitted light beam emitted by the point light source is not converged by a light-collecting unit (such as a toroidal reflector or an achromatic lens), but directly enters the first reflective unit composed of plane reflectors M5 and M6, the above can also be achieved. Spectral effect.

(2) Combining light: As shown in Figure 8, when the probe beam reflected by the sample returns to the plane where the reference beam is located along the original path, it is a converging beam. The second reflection unit is composed of two non-coplanar plane mirrors M2 and M3. The probe beam and reference beam returned from the sample surface are respectively incident on the plane mirrors M2 and M3 that form the second reflection unit. The plane mirror M2 At least one straight edge shape is included, and the straight edge intersects with the main beam of the detection beam. After being reflected by the plane mirror M2, the detection beam is incident and focused into the spectrometer SP. The spectrometer SP is placed at the focal point of the converging probe beam. The reference beam in the same plane passes through the lens L or other light-gathering elements, such as the reflective objective lens, and becomes a converging beam, which is reflected by the plane mirror M2, deflected in the incident plane, and enters the same spectrometer SP. By rotating and/or moving the plane mirror M3 along the light direction (or in the opposite direction), the propagation direction and/or deflection position of the reference beam can be changed, so that the main beam of the reference beam coincides with the main beam of the probe light beam, and The reference and probe beams do not affect each other. The focus position of the reference beam can be adjusted by moving the condensing lens L (not shown) in the direction of the reference beam light (or in the opposite direction). That is, adjusting the plane mirror M3 and the focusing lens L can make the reference beam incident and focused into the same spectrometer SP. From then on, the detection beam and the reference beam from different directions can be incident and focused into the same spectrometer SP after being reflected by the second reflection unit.

According to the combined light process shown in Fig. 8, the simulated cross section of the probe beam and reference beam reflected by the second reflection unit is shown in Fig. detected in a spectrometer, and in the process, they do not affect each other's propagation. The image formed by focusing the probe beam and the reference beam on the spectrometer is shown in Figure 9b. In Figure 9b, the size of the focused spot formed by the probe beam and the reference beam on the spectrometer is different, because the two beams are focused Due to the different magnifications in the process. In the actual detection process, it is necessary to select an appropriate size measurement window (entrance slit) for the spectrometer, so that the reference beam can be detected as much as possible, so as to improve the signal-to-noise ratio of the reference beam spectrum and achieve the purpose of improving measurement accuracy. .

Since the plane mirror itself does not affect the convergence state of the incident light and does not produce chromatic aberration, the use of the mirror can change the propagation direction of the beam while ensuring the quality of the converged beam. The detection beam and the reference beam can be focused into the same spectrometer at the same time after passing through the above two plane mirrors. On the other hand, the plane reflector can realize the high reflectivity in the broadband spectral range, and it is very low to light intensity influence, then the design of a spectrometer in the present invention does not reduce the detection efficiency of spectrometer to probe beam and reference beam, therefore , the present invention realizes the complete combination of the beams after splitting through proper optical path design, so as to improve the efficiency of light flux, and the complexity of the system is lower than that of the prior art.

The present invention can adopt the absolute reflectance measurement method, that is, to measure the absolute reflectance of two polarization states of the sample in the orthogonal direction. To measure the absolute reflectance of a sample, do the following:

a. Measure the dark value I _d0 of the spectrometer, that is, the reading of the spectrometer when no light signal enters the spectrometer;

b. load a reference sample, for example, a bare silicon wafer, obtain the spectral value I _Si , and measure the spectral value I _R0 of the reference beam immediately before or after measuring the reference sample;

c. load and measure the sample to be measured, obtain the spectral value I, and measure the spectral value I _R of the reference beam immediately before or after measuring the sample to be measured;

d. Measure the dark value I _d of the spectrometer;

In the above steps, steps a and b only need to be performed once within a period of time, for example, within an hour, within a day, within a week or within several weeks. Steps c and d should be repeated for each measurement. If the ambient temperature does not change, or the dark value of the spectrometer does not change with time, then _Id can be replaced by _Id0 .

Thus, the reflectance of the sample is:

$\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; I</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; Si</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; )</span>\mathrm{<span\; class="notranslate">\; f</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; r</span>}<span\; class="notranslate">\; )</span>}\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; Si</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{{\mathrm{<span\; class="notranslate">\; I</span>}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}}{<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; Si</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; (</span>\frac{{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; R</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}}}{{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; R</span>}\mathrm{<span\; class="notranslate">\; 0</span>}}<span\; class="notranslate">\; -</span>{\mathrm{<span\; class="notranslate">\; I</span>}}_{\mathrm{<span\; class="notranslate">\; d</span>}\mathrm{<span\; class="notranslate">\; 0</span>}}}<span\; class="notranslate">\; )</span>}\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; Si</span>}<span\; class="notranslate">\; )</span>$

Where R(Si) is the absolute reflectance of the reference sample, R(Si) can be obtained from other measurements, or calculated from the characteristics of the reference sample, usually the reflectance of the bare silicon wafer; is the correction of the reflectance of the reference beam to the sample.

For example, in the periodic shallow trench structure, as shown in FIG. 10 , the two orthogonal polarization directions are respectively defined as a direction TM perpendicular to the linear structure and a direction TE parallel to the linear structure. When the period p is 100 nm, the line width w is 50 nm, and the groove depth t is 50 nm, the reflectivity is shown in Figure 11, where the dotted line is the reflectance in the TE polarization direction, and the solid line is the reflectance in the TM polarization direction.

After measuring the absolute reflectance of TE and TM, by comparing with the numerical simulation results and numerical regression calculation, the critical scale, three-dimensional topography, film thickness and optical constant of the periodic pattern on the sample surface can be measured. In this case, the normal incidence broadband spectrometer may also include a calculation unit, which is used to calculate the optical constant of the sample material, the thickness of the film and/or use the mathematical model calculation and curve regression fitting of the reflectance. Analyze samples for critical scale properties or three-dimensional topography of periodic structures. Rigorous Coupled-Wave Analysis (RCWA) is the commonly used calculation method for electromagnetic simulation of periodic structures today, and the regression algorithm is the Levenberg-Marquardt algorithm. In the present invention, in addition to the theoretical measurement method, the measurement process also involves the processing of changes caused by polarization sensitivity such as polarizer rotation. Such problems can be solved by numerical methods. For more specific content, please refer to US Patent No.6522406B1 and US Patent No. 6665070B1. In the present invention, the linear polarization direction of the beam passing through the polarizer is determined by the rotation angle of the polarizer, and the light source incident on the polarizer can be a beam of any polarization state. The light reflected by the sample is linearly polarized after passing through the polarizer. When the beam is incident on the detector, the reflected light of the reference sample and the reflected light of the measured sample all experience the same polarization change, so it is not required to maintain the polarization state. For optics The polarization sensitivity of the components is not required.

Hereinafter, the present invention will be described in detail by taking specific embodiments as examples.

Embodiment one

A normal incidence broadband spectrometer according to a first embodiment of the invention is shown in Fig. 12a. As shown in Figure 12a, the normal-incidence broadband spectrometer includes a broadband point light source SO, a first reflection unit (including plane mirrors M4, M5), a movable light baffle D, an aperture A, a first light-gathering unit (off-axis parabolic reflection mirror OAP2), polarizer P, the first plane mirror M1, the first off-axis parabolic mirror OAP1, the second concentrating unit (lens L), the second reflection unit (including plane mirrors M2, M3), broadband spectrum Meter SP, movable beam splitter BS (the specific location can refer to Figure 12b) and pattern recognition system IRS. The pattern recognition system IRS includes a lens L', an illumination source (not shown) and a CCD imager (not shown). The broadband point source SO can emit a divergent beam containing a broadband spectrum, typically in the deep ultraviolet to near infrared range (in the wavelength range of approximately 190nm to 1100nm). In practice, the broadband point light source SO can be a xenon lamp, a deuterium lamp, a tungsten lamp, a halogen lamp, a mercury lamp, a composite broadband light source containing a deuterium lamp and a tungsten lamp, a composite broadband light source containing a tungsten lamp and a halogen lamp, a composite broadband light source containing a mercury lamp and a xenon lamp composite broadband light sources, and composite broadband light sources containing deuterium tungsten halogen lamps. The light beams of these broadband light sources may be natural light (ie, polarization equal to zero). However, the broadband point light source can also be a natural light point light source with zero polarization degree generated by a depolarizer. Examples of broadband point light sources SO include Ocean Optics products HPX-2000, HL-2000 and DH2000, and Hamamatsu company products L11034, L8706, L9841 and L10290. The broadband spectrometer can be a charge-coupled device (CCD) or photodiode array (PDA) broadband spectrometer, for example, Ocean Optics QE65000 spectrometer or B&W TeckCypher H spectrometer. Examples of the pattern recognition system IRS include NT59-839, NT59-743 from EDMUND, FVL-5X-120D-C, FVL-6X-120D-C from SEIWA, XF-T6X-110D from Canray Optics, and the like.

The divergent light beam emitted by the broadband point light source SO is incident on the plane mirrors M3 and M4 and will be divided into two beams, one of which is the probe light and the other is the reference light. As preferably, this embodiment also includes a light source concentrating unit (curved surface reflector M6), the curved surface reflector M5 is arranged between the point light source SO and the plane reflectors M3, M4 to form a converging light beam, which is reflected by the plane The mirrors M3 and M4 are divided into two beams, one of which is the probe light and the other is the reference light. The optical paths of the two beams are introduced below: (1) The converging beam of the main light reflected by the plane mirror M4 in the horizontal plane is used as the probe light, and the aperture A is placed at the focal point of the converging beam. The probe light after passing through the diaphragm re-diverges and enters the first concentrating unit, that is, the off-axis parabolic reflector OAP2. Reflected by the mirror OAP2, a parallel beam along the horizontal direction is formed. The parallel beam is incident on the plane mirror M1 after passing through the polarizer P. The plane mirror M1 rotates the parallel beam by 90 degrees in the horizontal plane and then enters the first off-axis paraboloid Reflector OAP1, the first off-axis parabolic reflector OAP1 rotates the parallel light beam by 90° in the vertical plane, and the light beam reflected by the off-axis parabolic reflector OAP1 is a converging beam of the main light along the vertical direction. The beam is incident at normal incidence and focused on the sample surface. The reflected light on the surface of the sample passes through the first off-axis parabolic mirror OAP1, the first plane mirror M1, the polarizer P, and the off-axis parabolic mirror OAP2 in turn, and then returns to the plane where the light beam was incident on OAP2 for the first time, and forms The converging light beam is reflected by the plane mirror M2 and then vertically enters the broadband spectrometer SP. The broadband spectrometer SP will be placed at the focal point of the converging probe beam reflected by the plane mirror M2. (2) The light passing through the edge of the plane mirror M4 and reflected by the plane mirror M5 is used as the reference light. The plane reflector M5 is slightly inclined relative to M4, and the reference light reflected by the plane reflector M5 intersects with the probe light reflected by the plane reflector M4 and then separates immediately. The reference light converges to one point and becomes a divergent beam, which enters the second converging unit, that is, the focusing lens L, and then forms a converged beam, which passes through the plane mirror M3 and then vertically enters the broadband spectrometer SP.

Those skilled in the art can know that by adjusting and/or rotating the plane mirror M3, the reference beam can be vertically incident on the spectrometer SP; by moving the position of the focusing lens L2 along or against the incident direction of the reference beam, the The reference beam is focused into the spectrometer SP after being reflected by the plane mirror M2.

In the embodiment of the present invention, the probe beam is emitted from the broadband light source SO to before reaching the off-axis parabolic mirror OAP2, and after being reflected by the sample, leaves the off-axis parabolic mirror OAP2 and reaches the broadband spectrometer SP, and the reference beam is emitted from the broadband light source SO to the Before reaching the broadband spectrometer SP, they are all in the same vertical plane P1. The probe beam is reflected by the plane mirror M4 before reaching the off-axis parabolic mirror OAP1, and after being reflected by the sample, leaves the off-axis parabolic mirror OAP1 and reaches the plane mirror M2, and is in a plane parallel to the sample surface.

In the present invention, the light emitted by the point light source is divided into two beams of light, the probe beam and the reference beam, after being reflected by the first reflection unit (i.e., the plane reflectors M4 and M5), and the probe beam and the reference beam returned from the sample surface undergo a second reflection After the unit (namely, the plane mirrors M2 and M3), a beam of light with the cross-sectional shape of the beam shown in Figure 4a is synthesized, thereby achieving the purpose of sharing a spectrometer for the detection beam and the reference beam.

In the present invention, the movable baffle D can be moved automatically or manually to cut off the reference light and/or the detection light. And when the movable baffle D is outside the optical path of the probe light/or the reference light, it has no influence on the corresponding optical path, then after the beam switching is completed, the spectral measurement can be performed without readjusting the optical path. Therefore, the vertical incidence spectrometer of the present invention can easily and quickly switch between the reference beam and the detection beam during the measurement process.

In the present invention, the polarizer may be a film polarizer, a Glan-Thompson prism polarizer, a Rochon prism polarizer, a Glan-Taylor prism polarizer, or a Glan laser polarizer. In particular, the polarizer is preferably a Rochonian prism polarizer, and its material is preferably magnesium fluoride (MgF ₂ ). In this embodiment, the beam splitter BS and the mirrors M2 and M4 are planar reflective elements with at least one straight edge shape, such as semicircular planar reflectors or square reflectors. Those skilled in the art can know that the linear edges of the beam splitter BS, the plane mirrors M2 and M4 are parallel, and the linear edges intersect the chief light of the light beam. The straight edge is preferably acute-angled to avoid reflections of the reference beam.

In addition, the normal-incidence broadband polarization spectrometer of the present invention may further include a polarizer rotation control device for controlling the rotation of the polarizer to adjust the polarization direction of the light beam. The polarizer rotation control device can adopt various automatic rotation devices controlled by motors (manual measurement can also be realized), such as Newport Precision Rotation Stage URS150.

The sample to be measured is usually placed on an adjustable sample stage, such as an X-Y-Z-Theta or R-Theta-Z stage. In the semiconductor industry, sample sizes are typically 8-inch (200 mm) or 12-inch (300 mm) diameter wafers. In the flat panel display industry, samples often have dimensions above 1 meter. For wafers, the surface may not be flat due to, for example, thin film layer stress on the wafer. For large-scale samples, the sample surface may be distorted, or the sample platform may not be flat. Therefore, when inspecting samples, each measurement point can be refocused in order to achieve high-accuracy measurement and ensure fast measurement of semiconductor production line yield.

In the present invention, in order to prevent the light-gathering effect of the off-axis parabolic mirror OAP1 from affecting the imaging clarity of the CCD imager in the pattern recognition system IRS, the pattern recognition system IRS and the beam splitter BS are placed between the off-axis parabolic mirror OAP1 and the sample. between.

Before measuring the sample, it is necessary to use the pattern recognition system to identify and locate the measurement points on the surface of the sample. This process can be realized only by moving the beam splitter into the optical path along the direction shown by the arrow in Figure 12b. The specific operation is: move the movable beam splitter into the optical path of the detection beam (including the sample reflection beam of the detection beam), its non-reflective surface completely shields the detection beam, and its reflective surface makes the illumination beam incident on the sample surface, at this time the sample The surface pattern can be imaged in the CCD imaging system. By calculating the imaging resolution of the sample surface and focusing on the sample based on the calibrated pattern recognition system IRS, the chip sample shown in Figure 13a can be obtained in the CCD imaging system. Surface pattern, the darker square areas in the figure are the measurement points. After the measurement point identification is completed, the spectroscopic plate can be partially moved out of the optical path of the detection beam. Its non-reflective surface shields part of the detection beam, and its reflective surface can reflect the illumination beam to the sample surface. At the same time, the detection beam reflected by the sample surface and the illumination beam The light beam is reflected to the CCD imaging system, and its optical path is shown in Figure 12b. At this time, both the detection beam and the surface pattern of the sample can be imaged in the CCD imaging system, so that the light spot and the measurement point can be aligned through the movable sample platform. When the light spot is aligned with the measurement point, the image observed in the CCD imaging system is shown in Figure 13b, and the central bright spot is the image formed by the probe beam. After the above steps are completed, the measurement points can be measured. During the measurement, the spectroscopic plate BS needs to be completely moved out of the optical path of the probe beam (there is no image in the CCD imaging system at this time), so that the probe beam can freely propagate to the sample surface, thereby performing spectrum measurement.

In addition, in addition to judging the focus by observing the change of the light intensity in the spectrometer, the present invention also has another method for judging the focus, that is, to adjust the focus by observing the imaging definition of the CCD imager in the pattern recognition system IRS. The coexistence of two focusing systems improves the accuracy of the focusing of the device. In addition, the function of aligning the detection beam spot on the sample surface with the pattern on the sample surface can be realized.

According to this embodiment, those skilled in the art can know that if the plane mirror M1 and the off-axis parabolic mirror OAP1 have the same reflective material and coating structure and satisfy the conditions that the incident angles of the light beams are the same and the incident planes are perpendicular to each other, then when As the probe beam propagates in the optical path between the polarizer and the sample surface, the polarization characteristics upon reaching the sample surface remain unchanged relative to those upon leaving the polarizer, and the reflected light from the sample, usually elliptically polarized light, returns to the polarizer with the same polarization The properties also remain unchanged relative to leaving the sample. That is, as the beam travels between the polarizer and the sample surface, its polarization properties are changed only by reflection from the sample. That is, in the present invention, the focusing system and the focusing process do not affect the polarization state of the beam, and the double-beam normal-incidence polarization spectrometer can measure anisotropic thin film samples or non-uniform thin film samples according to the measurement method described above, such as measuring Critical dimension (CD), three-dimensional topography of surface periodic patterns, and film thickness and optical constants of multilayer materials.

In this embodiment, the linear polarization direction of the probe beam passing through the polarizer is determined by the rotation angle of the polarizer, and the light incident on the polarizer can be a beam of any polarization state. The light reflected by the sample is linearly polarized after passing through the polarizer. When the beam is incident on the detector, the reflected light of the reference sample and the reflected light of the measured sample all experience the same polarization change, so it is not required to maintain the polarization state. For optics The polarization sensitivity of the components is not required.

Using the vertical incidence broadband polarization spectrometer of this embodiment can not only focus the probe beam on the sample surface without chromatic aberration, but also precisely control the polarization change of the probe beam, that is, the effect of maintaining the polarization characteristics of arbitrary polarized light, and can also be achieved by The simple operation corrects the measurement error caused by the spectral change of the broadband light source in the measurement process of the absolute reflectance method. In addition, since only one spectrometer is required, the cost is not increased.

Embodiment two

FIG. 14 shows an optical path diagram of a double beam normal incidence broadband spectrometer according to the second embodiment of the present invention. The optical path components and measurement method of this embodiment are basically the same as the double-beam vertical incidence broadband polarization spectrometer of the first embodiment, but the characteristics of the optical path are slightly different. For the sake of simplicity, only a brief description of the optical path is given.

The beam emitted by the broadband point light source SO is reflected by the curved mirror M6 to form a converging beam. The first reflection unit, namely the plane mirrors M4 and M5, divides the beam into two beams, the detection beam and the reference beam. The second reflection unit, namely the plane Reflectors M2 and M3 combine the light beams and the detection beams returned from the sample surface and the reference beams into one beam and then enter the same spectrometer SP. The difference from the first embodiment is that the light beam reflected by M4 is the reference light, and passes through the edge of M4, and the light beam reflected by M5 is the detection light, and in this embodiment, the reference light and the detection light pass through both sides Mirrors do not intersect, but separate directly. The reference beam converges to a point and becomes a divergent beam, which is incident on the second converging unit (focusing lens L), and then forms a convergent beam, which is reflected by the plane mirror M3 and then vertically incident on the broadband spectrometer SP. The beam passing through the edge of M5 and reflected by M4 is the probe light. The probe light is a converging beam of the main light in the horizontal plane, and the diaphragm A is placed at the focal point of the converging beam. The probe light after passing through the diaphragm re-diverges and enters the first concentrating unit (off-axis parabolic mirror OAP2), and the divergent beam is deflected by 90 degrees in the incident plane (horizontal plane) after being reflected by the off-axis parabolic mirror OAP2. A parallel beam along the horizontal direction is formed. The parallel beam is incident on the plane mirror M1 after passing through the polarizer P. The plane mirror M1 deflects the parallel beam by 90 degrees in the horizontal plane to form a converging beam. The parabolic mirror OAP1 focuses on the sample surface after reflection, and its chief light is incident on the sample surface vertically. The reflected light on the surface of the sample passes through the first off-axis parabolic mirror OAP1, the first plane mirror M1, the polarizer P, and the off-axis parabolic mirror OAP2 in sequence, and then returns to the plane where the light beam was incident on OAP2 for the first time and forms a converging beam , the converging light beam is deflected by 90 degrees in the incident plane after being reflected by the plane mirror M2, and is vertically incident on the broadband spectrometer SP. The broadband spectrometer SP is placed at the focal point of the converging probe beam reflected by the plane mirror M2. Those skilled in the art can know that by adjusting and/or rotating the plane mirror M3, the reference beam can be vertically incident on the spectrometer SP; by moving the position of the focusing lens L2 along or against the incident direction of the reference beam, the The reference beam is focused into the spectrometer SP after being reflected by the plane mirror M3.

The same as the first embodiment, the probe beam is emitted from the broadband light source SO to before reaching the off-axis parabolic mirror OAP2, and after being reflected by the sample, leaving the second off-axis parabolic mirror OAP2 and before reaching the broadband spectrometer SP, the reference beam is from Before the broadband light source SO emits to the broadband spectrometer SP, it is in the same plane perpendicular to the sample surface. The detection beam is in a plane parallel to the sample surface after being reflected by the plane mirror M4 and before reaching the off-axis parabolic mirror OAP1, and after being reflected by the sample, leaving the off-axis parabolic mirror OAP1 and before reaching the plane mirror M2.

In this embodiment, a pattern recognition system can also be added as in the first embodiment. This embodiment can perform the same measurements as those described in the first embodiment.

Embodiment three

A two-beam normal incidence broadband spectrometer according to a third embodiment of the present invention is shown in FIG. 15 . The optical path components and measurement methods of this embodiment are basically the same as those of the first and second embodiments. For the sake of simplicity, only a brief description of the optical path of this embodiment is given below.

As shown in Figure 15, the light beam emitted by the broadband point light source SO passes through the curved mirror M6 and is divided into two beams by the plane mirrors M5 and M6, one of which is the probe light and the other is the reference light. The probe light passes through the aperture A, the off-axis parabolic mirror OAP2, the polarizer P, and then enters the plane mirror M1, the off-axis parabolic mirror OAP1, and then is vertically incident on the sample surface. The probe light reflected by the sample surface is incident on the spectrometer SP after passing through the off-axis parabolic mirror OAP1, the plane mirror M1, the polarizer P, the off-axis parabolic mirror OAP2, and the plane mirror M2; the reference light passes through the lens L, The plane mirror M3 is incident on the spectrometer SP. The difference from the first embodiment is that the incident plane of the probe beam on the plane mirror M1 and the incident plane of the off-axis parabolic mirror OAP2 are perpendicular to each other.

In addition, the probe beam is emitted from the broadband light source SO to before reaching the off-axis parabolic reflector OAP2, and after being reflected by the sample, leaves the plane reflector M1 to reach the broadband spectrometer SP, and the reference beam is emitted from the broadband light source SO to reach the broadband spectrometer SP Before, they were all in the same plane perpendicular to the sample surface. The probe beam is in a plane parallel to the sample surface when it is in the optical path between the off-axis parabolic mirror OAP2 and the off-axis parabolic mirror OAP1.

In this embodiment, a pattern recognition system can also be added as in the first embodiment. This embodiment can perform the same measurements as those described in the first embodiment.

In the above embodiment, the normal-incidence broadband polarization spectrometer also includes a diaphragm located between the polarizer and the sample, for preventing the e-light generated after passing through the polarizer from entering the sample surface, and/or Or its reflected light bounces back into the polarizer. In addition, an aperture can be set in any section of the optical path in the above embodiments, the aperture is at a position perpendicular to the main light and the center of the aperture passes through the main light, so as to adjust the actual numerical aperture of the probe light.

In addition, the normal-incidence broadband polarization spectrometer of the present invention may further include a calculation unit for calculating the optical constants of the sample material and/or for analyzing the critical dimension properties or three-dimensional morphology of the periodic microstructure of the sample material.

In the above embodiments, the first off-axis parabolic reflector OAP1 can move along the direction of the light or against the direction of the light, so that different points of the sample can be measured.

Please note that according to the teaching of this specification, those skilled in the art will understand that the vertical incidence broadband polarization spectrometer of the present invention is not limited to the specific forms disclosed in the above embodiments, as long as it is under the general concept of the present invention, it can Various modifications can be made to the broadband spectrometer of the present invention. The broadband spectrometer of the present invention can be applied to detecting the thickness of semiconductor thin film, optical mask, metal thin film, dielectric thin film, glass (or coating), laser mirror, organic thin film, optical constant and the criticality of the periodic structure that these materials form Scale and three-dimensional morphology, especially can be applied to measure all the scales of the three-dimensional structure with one-dimensional and two-dimensional periodicity in the plane formed by multilayer thin films and the optical constants of each layer of material. In addition, with the broadband spectrometer of the present invention, automatic focusing and manual focusing can also be realized.

Finally, it should be noted that the above specific embodiments are only used to illustrate the technical solutions of the present invention without limitation, although the present invention has been described in detail with reference to examples, those of ordinary skill in the art should understand that the technical solutions of the present invention can be carried out Modifications or equivalent replacements without departing from the spirit and scope of the technical solution of the present invention shall be covered by the claims of the present invention.

Claims (27)

1. A broadband polarized spectrometer, characterized in that it comprises a light source, a first reflective unit, a first light-gathering unit, a second light-gathering unit, a polarizer, a first flat reflector, a first curved reflector, a second reflector unit, detection unit, where: The first reflection unit is used to divide the light beam emitted by the light source into two parts, a detection beam and a reference beam, and inject the two parts of the beam into the first light concentrating unit and the second light concentrating unit respectively; The first light concentrating unit is used to receive the detection beam, make the beam into a parallel beam and then enter the polarizer; The polarizer is arranged between the first light concentrating unit and the first plane reflector, and is used to allow the parallel light beam to pass through and enter the first plane reflector; The first plane reflector is used to receive the parallel beam and reflect the parallel beam to the first curved reflector; The first curved reflector is used to turn the parallel light beam into a converging light beam, and focus the converging light beam vertically onto the sample after being reflected; The second reflective unit is used to simultaneously or separately receive the detection beam reflected from the sample and sequentially pass through the first curved reflector, the first plane reflector, the polarizer, the first condensing unit, and passing through the reference beam of the second condensing unit, and incident the received beam to the detection unit; The second concentrating unit is used to receive the reference beam and make it incident to the second reflecting unit; The detecting unit is used for detecting the light beam reflected by the second reflecting unit. 2. broadband polarized spectrometer according to claim 1, is characterized in that, described broadband polarized spectrometer also comprises: A light source concentrating unit, which is arranged between the light source and the first reflection unit, and is used to make the light emitted by the light source into a converging light beam. 3. broadband polarizing spectrometer according to claim 2, is characterized in that, described light source concentrating unit comprises: At least one lens and/or at least one curved mirror. 4. broadband polarized spectrometer according to claim 1, is characterized in that, described broadband polarized spectrometer also comprises: Movable beam splitter and pattern recognition system; The beam splitter and pattern recognition system are located between the first curved mirror and the sample; The pattern recognition system includes a lens, an illumination source and a CCD imager; The movable beam splitter is used to reflect the sample illumination beam provided by the pattern recognition system to the sample surface and reflect the reflected beam of the sample surface to the CCD imager; The focus is adjusted by the light intensity of the detection unit and/or by observing the sharpness of the image in the pattern recognition system. 5. according to the described broadband polarization spectrometer of claim 4, it is characterized in that: The movable beam splitter is a plane mirror with at least one straight edge, which plane mirror can be completely, partially, or not at all in the optical path of the probe beam. 6. broadband polarized spectrometer according to claim 1, is characterized in that, described broadband polarized spectrometer also comprises: A beam switching unit for switching the detection beam and the reference beam. 7. broadband polarized spectrometer according to claim 6, is characterized in that: The light beam switching unit is a movable light blocking plate arranged between the first reflecting unit and the second reflecting unit, and the light blocking plate can block the detection light beam and the reference light beam separately or simultaneously. 8. broadband polarized spectrometer according to claim 1, is characterized in that, described broadband polarized spectrometer also comprises: Adjustable sample platform for holding samples. 9. broadband polarized spectrometer according to claim 2, is characterized in that, described broadband polarized spectrometer also comprises: The diaphragm is placed in any section of the optical path of the entire optical system. 10. broadband polarized spectrometer according to claim 9, is characterized in that, described broadband polarized spectrometer also comprises: At least one aperture, the aperture is placed at the focal point of the detection beam branched by the converging beam formed by the light source concentrating unit. 11. broadband polarized spectrometer according to claim 1, is characterized in that, described broadband polarized spectrometer also comprises: At least one aperture, which is located between the polarizer and the sample, for preventing the e-ray generated after passing through the polarizer from entering the sample surface and/or its reflected light is reflected back to the polarizer . 12. broadband polarized spectrometer according to claim 1, is characterized in that, described broadband polarized spectrometer also comprises: A polarizer rotation control device for controlling the polarization direction of the polarizer. 13. broadband polarized spectrometer according to claim 1, is characterized in that, described broadband polarized spectrometer also comprises: A calculation unit for calculating the optical constants of the sample material, the thickness of the film and/or for analyzing the critical dimension properties or the three-dimensional shape of the periodic structure of the sample. 14. The broadband polarization spectrometer according to any one of claims 1-13, characterized in that: The first flat reflector and the first curved reflector have the same reflective material and coating structure, and satisfy the conditions that the incident angles of the main light are the same or approximately the same and the incident planes are perpendicular or approximately perpendicular to each other. 15. broadband polarized spectrometer according to claim 14, is characterized in that: The difference between the incident angle of the main beam of the probe beam on the first flat reflector and the incident angle on the first curved reflector is no more than 5 degrees; The error between the incident plane of the incident light beam of the first flat reflector and the incident plane of the incident light beam of the first curved reflector is less than 5 degrees. 16. The broadband polarization spectrometer according to any one of claims 1-13, characterized in that: The first reflective unit is composed of at least two plane reflectors, wherein at least one reflector has at least one straight edge and the straight edge intersects the main light of the optical path. 17. The broadband polarization spectrometer according to any one of claims 1-13, characterized in that: The first light concentrating unit is an off-axis parabolic reflector or a toroidal reflector. 18. The broadband polarization spectrometer according to any one of claims 1-13, characterized in that: The second condensing unit is a lens, a lens group, an off-axis parabolic reflector or a toroidal reflector. 19. The broadband polarization spectrometer according to any one of claims 1-13, characterized in that: The polarizer is a Rochonian prism polarizer. 20. The broadband polarization spectrometer according to any one of claims 1-13, characterized in that: The first curved reflector is an off-axis parabolic reflector or a toroidal reflector. 21. The broadband polarization spectrometer according to any one of claims 1-13, characterized in that: The inclination angle and/or spatial position of the first plane reflector is adjustable. 22. broadband polarized spectrometer according to claim 21, is characterized in that: The first plane reflector moves along the propagation direction of the chief light of the converging light beam. 23. The broadband polarization spectrometer according to any one of claims 1-13, characterized in that: The second reflection unit is composed of at least two plane mirrors, and at least one of the plane mirrors constituting the second reflection unit has at least one straight edge and the straight edge intersects the main light of the optical path. 24. The broadband polarization spectrometer according to any one of claims 1-13, characterized in that: The detection unit is a spectrometer. 25. The broadband polarization spectrometer according to any one of claims 1-13, characterized in that: The light source is a light source including multiple wavelengths. 26. The broadband polarization spectrometer according to any one of claims 1-13, characterized in that: The light source is a xenon lamp, a deuterium lamp, a tungsten lamp, a halogen lamp, a mercury lamp, a composite broadband light source comprising a deuterium lamp and a tungsten lamp, a composite broadband light source comprising a tungsten lamp and a halogen lamp, a composite broadband light source comprising a mercury lamp and a xenon lamp, or A compound broadband light source containing deuterium tungsten halogen, or the light source is a natural light point light source with zero polarization degree generated by a depolarizer. 27. An optical measurement system comprising the broadband polarization spectrometer according to any one of claims 1-13.