SU1695145A1 - Ellipsometer - Google Patents

Abstract

The invention relates to optical polarization devices and can be used for the rapid non-destructive determination of physical parameters (film thickness, porosity degree, refractive index and absorption spectra, birefringence, surface roughness and quality, chemical composition, concentration of solutions, etc.) of solid and liquid materials in various fields of science and technology. The ellipsometer contains a source of monochromatic radiation 1, located in series along the beam beam forming theme 2, beam separation element, modulator and beam combining element installed with simultaneous rotation of the sample holder 9 analyzer 10 and an inaming-recording system containing a receiver 11 and a block of amplification, processing and displaying information 12 to improve measurement accuracy and increase the signal-to-noise ratio , the elements of separation and unification of polarized beams B by the proposed ellipsometer are unified into one element, made in the form of an isosceles prism 3 of double-bore This material has an axis located in the perpendicular direction of the initial propagation and passing through the linga of the trans- planes of the entrance and video of one side face of the idpar prism, lelyo or perpendicular to the base of the prism, and at the output of the prism along the common and unusual beams symmetrically with respect to the indicated plane, spherical or parabascopic mirrors 4–7 are installed; in this case, the modulator 8 is installed between the mirrors. A small loss of cure is realized in the ellipsometer; polarization of switched beams, a wide spectral range of 5 з п ф-лы, 5 Il (L S e

Description

The invention relates to optical polarization devices and can be used dp express non-destructive determination of physical parameters (thickness of the layers of the refractive index and absorption birefringence spectra, concentration concentrations, etc.) of solid and liquid materials

Known Polarization -. : an installation in which a source of collimated monochromatic radiation is successively arranged along the beam, a polarized radiation separation element containing a thin beam-splitting plate (SP) with a translucent coating and an aluminized mirror at an angle of 45 ° to the radiation incident on them, two polarizers with orthogonal azimuths of the direction of polarization, installed in two spatially separated parallel rays, the modulator and the element combining polarized rays into one, containing aluminum mirror and SP, a table for mounting the sample and a PMT with a recording system. The modulator disk interrupts the spatially separated rays, polarized in the plane of radiation incident on the sample and perpendicular to it, at different frequencies. The signal from the photomultiplier is fed to selective amplifiers tuned to these frequencies, at the output of which the signal ratio is recorded, determining the ellipsometric parameter

Tp (tg if), where Vree Rs are the modules of the reflection coefficients of the radiation intensity, polarized in the plane of incidence of radiation on the sample and perpendicular to it, respectively.

The disadvantages of this setup include the possibility of determining only one ellipsometric parameter ty, a low signal-to-noise ratio due to large losses in the separation and unification units of rays, since the reflection and transmission coefficients of semitransparent SPs are 30% and the total transmission of beams with orthogonal polarization is less than 9%. The residual birefringence in the SP material and the interference of multiply reflected rays in a thin SP leads to temperature changes in the degree of polarization of the rays and instability, which reduces the measurement accuracy of gr.

In the ellipsometer, after the source of monochromatic radiation along the beam, there is a system for forming a weakly converging or collimated radiation beam, a two-hole diaphragm, an obturator, two polarizers, each of which is optically connected with a corresponding hole in the diaphragm, are rotatably mounted around the direction of the incident radiation, an analyzer, and a receiving-recording system on them. Each of the polarizers is made in the form of sequentially installed and optically coupled two parallel mirrors with a metallic coating and two reflective plates of semiconductor material mounted at the Brewster angle to the incident radiation. The division of the beams into the upper and lower beams by the diaphragm, their successive interruption and the formation of two linearly polarized beams with different polarization azimuths followed by recording the azimuths of the analyzer AI and Ar, at which the signals on the photo receiver corresponding to the switched beams are equal, provide accuracy

determine the ellipsometric parameters andA for reproducibility of 0.05-0.1 ° in the spectral wavelength range from 0.5 to 2.5 µm.

The disadvantages of an ellipsometer include

a decrease in the signal-to-noise ratio due to radiation losses during diaphragm and reflection from semiconductor wafers. Insufficiently high degree of polarization in the visible near-IR region

the spectrum leads to a decrease in the absolute accuracy of measurements of the ellipsometric parameters ip and d. The ellipsometer does not overlap the important spectral region from 0.25 to 0.5 μm.

The closest in technical essence and the achieved result to the invention is an ellipsometer containing a source of monochromatic radiation arranged in series along the way.

a beam of a polarized beam separation element, a modulator, a beam combining element and an analyzer mounted rotatably, and a receiving-recording system. The elements of separation and combining of rays contain BaF2 beam-splitting plates and polished doped silicon plates, arranged parallel to each other at the Brewster angle to the incident radiation.

The most accurate measurement variant on an ellipsometer measurement with discrete (binary) modulation of the polarization state when the

two beams with different azimuths of linear polarization and register the azimuths of the analyzer Ai and A2, at which the signals on the photodetector corresponding to the switched beams are equal.

By the values of AI and A2, the ellipsometric parameters 1p D are determined. The accuracy of the determination of p and D from reproducibility for measurements with discrete (binary) modulation of the polarization state reaches ± 0.03 ° and ± 0.05 °, respectively. The main disadvantage of an ellipsometer is the large loss of intensity, which leads to a decrease in measurement accuracy and signal-to-noise ratio. Calculations of the reflection of polarized radiation with oblique incidence using the Fresnel formulas show that the intensity of a beam m, polarized perpendicular to the plane of the drawing, decreases by 60% when reflected from each of the 8aF2 transparent plates and 30% when reflected from silicon wafers. The intensity of beam A after four reflections is 8% of the intensity of the perpendicular component of the beam incident on the first separating plate. The loss of intensity of beam B after passing through the beam-splitting plates is about 40% for the component polarized in the drawing plane and about 85% for components that are perpendicular to the plane of the drawing. The ratio of the intensities of the components, and, therefore, the azimuth of the linear polarization of the beam B, is determined by the value of the refractive index of the transparent plates, which varies with the radiation wavelength and temperature. Changes in the azimuth of linear polarization with temperature fluctuations lead to a decrease in measurement accuracy and a signal-to-noise ratio. Large losses of beam intensity A (12 times) lead to a decrease in the measurement accuracy of 1p and Au and the signal-to-noise ratio, especially during spectral ellipsometric measurements. A change in the ratio of the intensities of the beams A and B and the azimuth of the linear polarization of the beam B to the wavelength leads to the need for complex calibrations, which also reduces the absolute accuracy of ellipsometric spectral measurements. The degree of polarization of the A and B beams is small in the spectral region of 0.25–0.5 µm, which also reduces the measurement accuracy of tr and A. The effects of depolarization and residual birefringence of transparent SPs lead to a decrease in the degree of beam polarization of the beam B and temperature instabilities, which also reduces the absolute measurement accuracy 1 / and A.

The aim of the invention is to improve the measurement accuracy of the ellipsometric parameters I and A and to increase the signal-to-noise ratio.

The goal is achieved by the fact that in an ellipsometer containing a source of monochromatic radiation, the element

separation of polarized beams, module and beam combining element, installed & furnace with simultaneous rotation, analyzer and receiver5 recording system, separation and combining elements of polarized beam combined into a single element made in the form of an equipotent prism of birefringent material The optical axis of which is located in a plane perpendicular to the direction of propagation of the initial beam and passes through the line crosses the r planes INPUT AND OUTPUT bi- f, BWX

The 5 faces of the prism are parallel or perpendicular to the base of the prism, and the trip ofu.::- a nosuchi and an extraordinary path; spherical or parabolic mirrors mounted ip to the initial beam direction, respectively, are installed symmetrically with respect to the indicated the modulator is mounted between the mirrors.

5 In the proposed clamper, in contrast to the known, to separate and combine polarized beams instead of the beam splitting plates and polished plates of doped silicon, a prism made of two refractive material and a system of spherical or parabolic mirrors arranged above are used, thanks to this radiation losses are reduced several times, a higher degree of polarization is achieved | | And switched beams, that in aggregate :; allows to increase the accuracy of measurements and the signal / noise ratio.

FIG. 1 shows the general scheme of the proposed ellipsometer; Figures 2-5 illustrate various embodiments of a separation unit for combining polarized light beams. The proposed zlipsomef contains a source of monochromatic, io5 radiation 1 (figure 1), a system for forming beam 2, an element for separating and combining polarized beams, made in the form an isosceles prism 3 of birefringent material, a mirror 4.5,

0 installed at the exit of the prism 3 along the ordinary beam B. and G.7, installed at the exit of the prism according to an extraordinary beam of light B, modulator 8 installed between mirrors 4.6 and 5.7,

5, the holder of an obrazts 9, an analyzer 10, installed with the possibility of a vrshoni around the axis of a light beam reflected from obr-nna, and a reception and recording software, containing a photoreception 1 and a unit 12 for amplifying, processing and displaying information. The prism 3 and mirrors 4,5,6,7 are installed with the possibility of simultaneous rotation around the axis of the original light beam, the prism 3 can be made of CaCO3, MgFa, CdS, etc. depending on the wavelength of the working radiation. The source of monochromatic radiation 1 can be made in the form of an incandescent lamp of a monochromator or in the form of a laser. The beam forming system 2 can be made of one or several lenses or mirrors, the optical axis of the prism 3 is located in the plane CC, perpendicular to the direction of propagation of the original radiation beam and passing through the line of intersection of the planes of the input and output side faces of the prism 3, parallel or perpendicular its foundation. Mirrors 4,5,6 and 7 are made spherical or parabolic and are located symmetrically relative to the plane CC, which is the plane of symmetry, Mirrors 4.5 (Fig. 2), mounted along the ordinary light beam are oriented so that the angle between the axis of this beam and the normal to the surface of this mirror (4 or 5) at the point of intersection of the axis with the mirror is p0C2. where tpo is the angle between the original direction of propagation of the light beam and the direction of propagation of the ordinary beam at the output of prism 3. Mirrors 6.7, mounted along the course of the unusual light beam, are oriented so that the angle between the axis of this beam and the normal to the surface of the mirror (6 or 7) at the point of intersection of the axis with the mirror, is pe / 2, where p & is the angle between the initial direction of propagation of the light beam and the direction of propagation of the unusual light beam at the output of the prism 3. Angles are determined by the ordinary and extraordinary indexes of refraction of the material of the prism 3 and the angle θ at its top from the relations

l

p0 arcsin no-sin (c) 2

sin tu arcsin (-),

 AT

 arcsin ne Sin (in-fe) -Ј,

sin arcsin (-pЈ-).

(2)

In the embodiment shown in Fig. 2, the mirrors 4,5,6 and 7 are installed at distances FI and F2 from the plane of symmetry CC equal to their focal lengths. The modulator 8 is made in the form of a shutter whose disk is located in the plane symmetry CC, the disk made two systems of holes with. the distance between the centers of the holes along the radius of the disk, equal to the distance between

The axes of the Ai 8 beams. For speed measurements, the circuit shown in Fig. 3 can be used, in which modulator 8 is made as two electro-optical modulators 13 and 14 with half-wave plates 15 and 16 installed in front of them. Electro-optical modulators work effectively when using collimated light beams. To form them, mirrors 4 and 6 are mounted on

distances x from the prism 3, equal to their focal lengths. In this case, along the initial light beam, at the output of the prism 3, the beam former 17 is installed. Fig, 4 shows a variant of the scheme that provides the most accurate measurements in a wide range of temperatures and wavelengths. In this variant, in contrast to the variant in FIG. 3, each mirror 4,5,6,7 is installed relative to the plane of symmetry

SS at distance X, determined from the expression

1.1 1 x w f

where X is the distance from prism 3 to a given mirror;

F is the focal distance of the given mirror.

The modulator 8 can be made a

the form of the obturator. When working with wide-aperture beams, with a prism 3 having a small angle in, and with long-focus mirrors, it is convenient to use the circuit shown in FIG. 5, Unlike the circuit 5 shown in FIG. 2, it contains two spherical or parabolic mirrors 18 and 19, installed at distances x from the prism 3, equal to their focal length, m

The mirrors are oriented so that the angles between the axes of the ordinary and extraordinary beams and the normals to the surface of the mirror at the points of intersection of the axes of these beams with the mirrors are respectively

p0 / 2 and (re / 2.

The proposed ellipsometer was manufactured in the modifications of FIG. 1,2 and 5 and was tested under laboratory conditions. In this case, a KGM 70x9 incandescent lamp and a DMR-4 monochromator or an LG-79 laser were used as the radiation source. Prism 3 was made of calcite, and the angle варь varied in various ways from 10 to 20 °. The photoreceiver 11 was equipped with a PMT 100, a FD-256 silicon photodiode, and a PbS photoresistance. The working wavelength range of the ellipsometer is 0.25-2.5 microns.

The ellipsometer operates as follows.

The monochromatic radiation from source 1 enters the beam forming system 2 (Fig. 1). The generated radiation beam is divided by the prism 3 into ordinary and extraordinary beams, which are mirrored 4 and 6 parallel to the original beam in the opposite direction. Mirrors 5 and 7 ordinary B and extraordinary B beams are directed onto a prism 3 and are combined into one beam. The modulator 8 sequentially interrupts the beams B and C, providing at the output of the prism 3 switching beams with orthogonal azimuths of linear polarization Pi and Pa directed to the sample under study. The radiation reflected from the sample passes through the analyzer 10 with the azimuth of the transmission axis A to the photodetector 11. The photodetector signal is amplified, processed and displayed in block 12, In the variants shown in Fig.2.4 and 5, the ordinary and extraordinary beams are focused in the plane of symmetry SS , which allows to significantly increase the number of holes in the obturator located in the same plane, to increase the frequency of radiation modulation f, and, therefore, to reduce the 1 / f noise and increase the signal-to-noise ratio. Note that spherical (or parabolic) mirrors invert the radiation beams that pass through the prism 3 the first and second times. Thus, the optical path is identical with two passes in a prism over the entire beam section. In the embodiment of the node for dividing and combining polarized radiation beams shown in Fig. 2, system 2 forms a collimated radiation beam, which is supplied to a prism 3, where it is divided into ordinary and extraordinary beams. Mirrors 4 and 6 focus these beams in the CC plane. The mirrors 5 and 7 form collimated beams and direct them to the prism 3, which combines these beams.

In the variant shown in FIG. 3, the original beam is focused by system 2 onto a prism 3, divided into ordinary and extraordinary beams, which are collimated by mirrors 4 and 6. Half-wave devices 15 and 16 rotate the plane of polarization of beams B and C by 90 °, In the absence of voltage on the electro-optical modulators 13 and 14, a rotation of 90 ° leads to the damping of the beams at the output of the prism 3 in the direction of the original beam. When pulses of an electric field are applied to electro-optical modulators 13 and 14, radiation pulses are generated at the output of the prism with orthogonal azimuths of linear polarization Pi and Pz, given by the azimuth of the prism unit, rigidly connected with mirrors, modulators and semi-new devices.

Fig. 4 shows a variant where the radiation beam is focused onto a prism 3. And the mirrors 4 and 6 form images of the focusing point in the plane CC. The mirrors 5.6 direct the beams on the prism 3 to the focal point, the beam shaper 17 forms a collimated or convergent radiation beam. A feature of the scheme is the constant position of the focal points in the plane of symmetry SS when the wavelength 1 of the radiation I varies and the temperature T changes, which increases the accuracy and signal-to-noise ratio in a wide interval of change of I and T

In the embodiment of the separation and convergence unit of polarized beams shown in Fig. 5, two mirrors are used to simplify the design and implement measurements with wide beams of radiation to increase the signal-to-noise ratio. The collimated radiation beams are split by the prism 3 into ordinary and extraordinary beams, the mirror 16 is directed parallel to the original beam in the opposite direction, interrupted by the obturator 8 and the mirror 19 is collimated and directed to the prism 3, combining

Beams. At 6 10 and ((e) 1 ° of puchki

after the second passage, the prisms 3 are combined with a stability of up to 0.01 ° and are almost identical.

The measurements of the ellipsometric parameters 1 / and D are as follows. When calibrating an ellipsometer to the position on the lumen (without a sample), by varying the azimuth of the analyzer 10, the ratio and powers of the switched beams with the determined azimuths of linear polarization (for example, PI - 30P P2 120 °) are determined by the formula

a cos2 (P2 Aoj

cos (L, Pi)

where AO is the azimuth of the analyzer corresponding to the equality of the signals on the photodetector induced by switchable beams. After calibration, the sample is set at an angle p to the incident radiation, and the analyzer unit — the photodetectors are turned at an angle of 2 p. The azimuth of the analyzer 10 is changed and the azimuth values of the analyzer AI and Aa are determined, at which the signals on the photoreceiver corresponding to the switching beams with the azimuths Pi and Pa are equal. The ellipsometric parameters and D are determined by the formulas

 / (a sin 2 Pi - sin 2 P2) tg Ai tg Aa J (a cos2 Pi -cos2 Pa)

(3)

d (a sin2 Pi -sin2 Pa) QgAl + | gA2) COStg (sin a Pa-a sin a Pi)

(four)

Investigation of the accuracy and sensitivity of measurements of Ai, Aa, V and using relations (3) and (4) allows choosing the optimal values of a, Rc and Pa for the study of these samples. The optimal values of a can be established by changing the azimuth of the additional polarizer installed after the beam forming system 2. For spectral ellipsometric measurements, the use of the additional polarizer is not helpful. Let us compare the accuracy and the signal-to-noise ratio values x on the famous ellipsometer and laboratory models of spectral and laser ellipsometers corresponding to this invention. In measurements on a known ellipsometer (prototype), the errors in determining the reproducibility are ± 0.03 ° and ± 0.05 °, respectively, and the absolute accuracy of measurement is ± 0.1 °. On the proposed zllipsometer, the measurement accuracy for reproducibility is ± 0.003 ° and ± 0.006 °, respectively, and the absolute measurement accuracy of Φ and D is D0 ± 0.05 °. The signal-to-noise ratio rises 5-7 times in comparison with the prototype due to a significant decrease in intensity losses. radiation in the elements of separation and integration of polarized beams, elimination of temperature instability due to residual birefringence in the beam-splitting plates, as well as an increase in the apertures of the working beams.

50

five

0

5 0 5 0 5 0 55

Claims (5)

1. An ellipsometer containing a source of monochromatic radiation, a system for forming radiation, a radiation separation element for ordinary and extraordinary beams, a modulator and a beam combining element installed with the possibility of simultaneous rotation, an analyzer and a receiving and recording system, differing in series on the optical axis that, in order to increase the measurement accuracy and increase the signal-to-noise ratio, the elements of separation and combining of the beams are combined into one element, made in the form an isosceles prism from a birefringent material, the optical axis of which is located in a plane perpendicular to the direction of propagation of radiation and passing through the intersection plane of the input and output side faces of the prism, at the exit of the prism along the ordinary and extraordinary beams, spherical or parabolic mirrors, with the modulator located between the mirrors. 2. The ellipsometer according to claim 1. is distinguished, and with the fact that along the ordinary and extraordinary beams, two mirrors are installed, respectively, oriented so that the angle between each axis and the normal to the surface of the mirror at the intersection point of the beam axis with the mirror is p0 / 2 for ordinary and pe / 2 for extraordinary beams, where p0 and pe are the angles between the initial radiation propagation and, respectively, the ordinary and extraordinary beams at the output of the prism. An ellipsometer in accordance with claim 2, characterized in that the mirrors are mounted at the focal lengths of the indicated symmetry symmetry. An ellipsometer in accordance with claim 2, wherein the mirrors are mounted at focal lengths from the prism. 5. The ellipsometer according to claim 1 is distinguished by the fact that each mirror is mounted relative to the plane of symmetry at a distance X, determined from the expression 1.1 1v tm + / t - g - where X is the distance or prism to the corresponding mirror, F is its focal length 6 The ellipsometer according to claim 1, that is, with the fact that along the ordinary and extraordinary beams is set by one mirror is symmetrical to the plane of symmetry at distances x from the prism equal to the focal length of the mirrors, and the mirrors are set so that the angles between the axes of the ordinary and extraordinary beams and the normals to the surface of the mirrors at the points of intersection of the axes of the beams with the mirrors are Res / 2 respectively 9 Fig.Z 7 five four / 7 % C FIG L Vo / 2 c FIG. five