This application generally relates to optical measurement method, and more particularly relates to the measuring method and apparatus utilizing ellipsometric technique on semiconductor devices and wafer surface and other materials.
Ellipsometric technique is a kind of powerful multi-functional light division technology, used to obtain the characteristics information of target surface by detecting electromagnetic waves. The characteristics information may includes reflectivity, thickness, refractive index, extinction parameters, polarization, surface microstructure, particles, defects and roughness of the target surface or the thin film surface, and so on. The ellipsometric technique is widely used in the field from basic research to industrial applications, from semiconductor physics, microelectronics and other aspects of biology, because of its high sensitivity and non-destructive and non-contact and other characteristics.
Existing ellipsometry measurement technology works as follows, the light source illuminates light through the first polarizing plate (often called the polarizer) to generate a polarized light, then this polarized light illuminates on the target surface. This polarized light changed its polarized state after goes through the target surface. For example, after its going through the second polarizer (often called the analyzer), then entered the detecting device. By analyzing the light intensity, phase and polarization states reflected by the target samples, ellipsometry technology may obtain the characteristic information detected by electromagnetic waves on a target surface which even has a shorter thickness than the detecting light wavelength (small to a single atomic layer or smaller). In general, ellipsometry is a technique based on light mirror reflection, in which, the incident angle is equal to the reflection azimuthal angle, and the incident light path and the reflected light path are in the same plane (called the incident plane). In the text to follow, the parallel and perpendicular electric components between this incident plane and the polarized light by the polarizer are defined as p and s components of this light, respectively. Certainly, the polarization state of this polarized light from the target surface can be changed by various methods, including reflection, transmission, diffraction and so on. In this article, without loss of generality, the main conditions for reflection are introduced. In ellipsometry measurement based on existing technology, each measurement can only get a set of experimental data. Therefore, in the measurement process, rotating ellipsometry method is generally used. That is, the polarizer is driven by the motor to rotate to change polarizer azimuthal angle. Similarly, the second polarizer is also driven by the motor to rotate, to make the analyzer azimuthal angle also changed. Based on all these polarizer azimuthal angles, a set of data could be obtained, and then based on these data, the characteristics of the target surface can be determined. Therefore, since a set of data is needed, existing measurement methods have some shortcomings, such as long data measuring time, complex method and expensive hardware and so on.
In order to solve the problems such as, long time data measuring, complex measuring method and expensive measuring hardware caused by rotating the polarizer. In this invention, the ellipsometric polarizes the light from the light source in a certain azimuthal angle, and then the polarized light was directed onto the target surface, after reflected from the target surface, the light characteristics could be translated with the related relationship information. Finally, based on the obtained light intensity information corresponding to the target surface, the characteristics information of the target surface could be determined by the electromagnetic waves. When conducting the measurement, there always are deviations exist in the polarized azimuthal angle, incident angle and the analyzed azimuthal angle. To make this deviation no longer happen, this invention provides a reference surface for calibration, after this calibration, the measurement for the target surface's characteristics begins.
The first aspect of this present invention provides a measuring method to detect the surface characteristics information based on ellipsometry by electromagnetic wave, which includes the following steps: i, polarize the light in a fixed polarizer's azimuthal angle to generate the polarized light; ii. illuminate the polarized light on to the target surface to get the characteristics light from reflection, transmission or diffraction; iv. analyze the characteristics light or lights by the fixed analyzer azimuthal angle, to obtain their p component and s component v. detect the light intensity for p component and s component to get the corresponding light intensity information of the target surface and the phase relationship information between. Besides, based on this corresponding light intensity information of the target surface, the characteristics information of the target surface could be obtained by the certain method.
Selectively, after polarize the incident light and illuminate it onto the target surface in a fixed azimuthal angle, then use a light divider to separate the obtained characteristics light into multiple beams, and then, detect and analyze each light beam in the fixed azimuthal angle to find the phase relationship information. In the text to follow, the measurement method uses a light divider, defined as the light division ellipsometric measurement. By the light division ellipsometric measurement, more data could be obtained this time, this is based on the second respect of this invention as below.
The second aspect of this present invention provides a measuring method to detect the surface characteristics information based on ellipsometry by electromagnetic wave, which includes the following steps: i, polarize the light in a fixed polarizer's azimuthal angle to get the polarized light, ii. the polarized light be illuminated on to the target surface to get the characteristics light after the reflectivity, and/or transmission, and/or diffracted; iii. these characteristics lights be divided to get several (fixed number) corresponding several p component and s component, iv. analyze the characteristics light or lights by the several (fixed number) analyzer azimuthal angle, to obtain their p component and s component v. detect the light intensity for p component and s component to get the corresponding light intensity information of the target surface and the phase relationship information between. Besides, based on this corresponding light intensity information of the target surface, the characteristics information of the target surface could be obtained by certain method.
The third aspect of this present invention provides a measuring apparatus to detect the surface characteristics information based on ellipsometry by electromagnetic wave, which including, a polarizer, to get the polarized light by the polarizing on the fixed polarized azimuthal angle, and then illuminate this polarized light onto the target surface, and then get the characteristics light after reflectivity, and/or transmission, and/or diffracted. Analyzer, to analyze the foreside characteristics light, to obtain one or several characteristics light. Detection-processing unit, to detect the light intensity of the p component and s component for one or several characteristics lights, and then to get the corresponding light intensity information of target surface and the phase relationship information between them. Besides, the calculation unit, to get the characteristics information of target surface based on the light intensity information obtained aforesaid.
The forth aspect of this present invention provides a measuring apparatus to detect the surface characteristics information based on ellipsometry by electromagnetic wave, which including a polarizer, to get the polarized light by the polarizing with a fixed polarized azimuthal angle, and then illuminate this polarized light onto the target surface, and then get the characteristics light after reflectivity, and/or transmission, and/or diffracted. Divider, is used to divide the aforesaid characteristics light after the surface reflectivity, and/or transmission, and/or diffracted light, to generate a predetermined number of analyses with the surface as described in the corresponding light division; Analyzer, is used to analyze the foreside characteristics light, to obtain one or several characteristics light. Detection-processing unit, is used to detect the light intensity of the p component and s component for one or several characteristics light, and then to get the corresponding light intensity information for target surface and the phase relationship information between them. Besides, the calculation unit, is used to get the characteristics information for target surface based on the light intensity information obtained aforesaid.
Use the ellipsometry measurement method and apparatus in this invention, a fixed polarizing azimuthal angle and analyzing azimuthal angle could be used to instead of several measurements by mechanical rotation, this costs down the hardware expense, decrease the measurements, reduce complexity of the measurement. At the meanwhile, use the method and apparatus in this invention, the calibration process in actual measurement is much simple, so that simplify the measurement process for the characteristics information obtaining process on target surface by electromagnetic waves.
By the detailed description of following words, the character, target and advantages will be shown better. In the drawings, same or similar drawings stand for same or similar steps or units or devices.
a shows the Fourier coefficients changing curve, when the incident light wavelength is 632.8 nm, the incidence azimuthal angle is 75.55 degree, and the thickness of the surface changes from 0 to 500 nm,
b shows the Fourier coefficients changing curve, when the incident light wavelength is 632.8 nm, the incidence azimuthal angle is 75.55 degree, and the thickness of the surface changes from 0 to 50 nm,
a shows the ellipsometry parameters changing curve, when the incident light wavelength is 632.8 nm, the incidence azimuthal angle is 75.55 degree, and the thickness of the surface changes from 0 to 500 nm,
b shows the ellipsometry parameters changing curve, when the incident light wavelength is 632.8 nm, the incidence azimuthal angle is 75.55 degree, and the thickness of the surface changes from 0 to 50 nm,
Based on the
In which, the light source 1 use the He—Ne laser of the output light wavelength 632.8 nm (nanometers), in addition, as known in technical staff in this field can understand, the light source 1 is not limited to this one, but also for other high-power red LED light sources. The role of small hole 2 is used for the better focus for the incident light, as known in technical staff in this field, hole 2 is not necessary for this invention. The role of hole 4 includes the avoidance of partial output from the other polarized light beam into the measurement light path, so that to avoid its impact on the measurement results, and similarly, as known in technical staff in this field, hole 4 is also not necessary for this invention. Usually, the use of small hole 2 and 4 help to improve measurement accuracy of the surface characteristics information by the electromagnetic waves
Firstly, light source 1 illuminates a light, goes through the hole 2, focused by hole 2 and then illuminates onto the polarizer 3.
In step S11, the polarizer 3 polarizes a single wavelength light in a fixed polarizer azimuthal angle P to generate a single wavelength polarized light.
Then, the polarized light goes through the hole 4, the hole 4 only allows one polarized light goes through by polarizer 3, and stop the other bunch of polarized lights in order to avoid its impact on the measurement. Thus, in step S12, the polarized light irradiated in the incident angle θ onto the target surface 5. As known of technique staff in this field, without hole 4, the light road of the polarized light from polarizer 3 may has quite difference with the light road of the polarized light from hole 4 in
The illuminated polarized light reflects to analyzer 7 by target surface, in which, the nature of this polarized light has been changed due to the target surface 5, and obtained the characteristics information of the target surface. And then, in the step 13, analyzer 7 analyzes the characteristics light from the target surface in the fixed analyze azimuthal angle A, and then divide the p component and the s component of the characteristics light and then provide them to the detection-processing device 8.
Then in step S14, the detection-processing device 8 receives and detects the light intensity of the p component and the s component of the characteristics light, obtain one light intensity information corresponding to the target surface, and obtain the phase deviation between the p component and the s component, and then provides this light intensity information to the calculation device 9 which has electromagnetic connection with the detection-processing device 8 and not shown in
Finally, in the step S15, the calculating device 9 determines the thickness of the target surface 5 in the scheduled method based on the light intensity of the p and s component of the characteristics light. It is can be understood by technical staff in this field, there are several ways to determine the thickness information, this is not limited by this embodiment. In this embodiment, the principle of this scheduled method is as follows,
The basic formula for the output light intensity after the polarizer 3 and the analyzer 7 is,
In which, P=π/4, A=0, π/2, tan Ψ is the amplitude ratio of the p and s component of the light, Δ is the phase deviation between the p and s component of the characteristics light. The detection-processing device obtained light intensity Iout(π/4,0), Iout(π/4,π/2), can be used to calculate the Fourier coefficient
and calculate the ellipsometry parameters tan Ψ, that is:
Besides,
stands for the first-order Fourier coefficient of the output light intensity.
In this invention, the changing situation can be measured forehead. In which, the illuminated light angle θ is 75.55 degree, the light source 1 (He—Ne laser) outputs the light with wavelength 632.8 nm, when the thickness of the surface changes in the range of 0 to 500 nm, the Fourier coefficients of the curve shown in
b shows the Fourier coefficients of the curve, when the wavelength of the illuminate light is 632.8 nm, the incident angle of the polarized light is 75.55 degree, when the thickness of the surface changes in the range of 0 to 50 nm. It can be figured out that when the thickness is within 0 to 50 nm, the Fourier coefficient is the monotonic function of the surface thickness.
a shows the ellipsometry parameters of the curve, when the wavelength of the illuminate light is 632.8 nm, the incident angle of the polarized light is 75.55 degree, when the thickness of the surface changes in the range of 0 to 500 nm. It can be figured out that, when the thickness is within 0 to 50 nm, the ellipsometric parameters is the periodic function of the surface thickness.
b shows the ellipsometry parameters of the curve, when the wavelength of the illuminate light is 632.8 nm, the incident angle of the polarized light is 75.55 degree, when the thickness of the surface changes in the range of 0 to 50 nm. It can be figured out from
When process the measurement, the Fourier coefficient can be calculated based on the formula (2). When all the azimuthal angles are fixed, including the polarized azimuthal angle, analyzer azimuthal angle and the incident angle between the polarized light and the target surface, and the wavelength of the laser is known, meanwhile the refractive index of the target surface is also known, all the Fourier coefficient α, β are all known, then the control method can be used, find the most similar Fourier coefficients within the range of values from the known Fourier coefficients, then the thickness of this Fourier coefficients is the thickness of the target surface.
In theory, the Fourier coefficients α, β effects by the thickness, azimuthal angle θ, wavelength, light constants of the material target surface, the polarizer azimuthal angle P and the analyzer azimuthal angle. In which, the thickness, azimuthal angle θ, wavelength, light constants of the material target surface effect the ellipsometry parameters and the ellipsometry effects the Fourier coefficients. Use the fixed wavelength light source, since the light constants of the material is fixed, the effects on the measurement result can be deleted. Thus, the thickness of the target surface is only decided by the Fourier coefficients α, β. Based on this principle, in a single range, Fourier coefficients of different thickness and is one to one. Thus, according to the correspondence between, the thickness can be calculated.
This embodiment shows only when the light source 1 with the wavelength 632.8 nm, the incident angle of the polarized light is 75.55 degree, the function relationship between the Fourier coefficients and the ellipsometry parameters and the thickness. It is widely understood by the technical staff in this field, when use light has other wavelength and incident angle, the function relationship between the Fourier coefficients α, β and the ellipsometry parameters and the thickness will be changed, and this invention is suitable for the reality function relationship.
Since the Fourier coefficients has one to one relationship with the reflectivity, refractive index, extinction coefficient, polarization, surface microstructure, particles, defects and other surface roughness features and so on, besides the thickness of the target surface, thus, the method in this invention is not limited to the thickness of the target surface, other characteristics can be measured also, and, by the introduction of this file, the technical staff in this field can understand that this invention can also used to measure other kind of target surface and its characteristics information. Based on above, for the sake of to be simple, this file will not show further on other situations.
Above shows the measurement method when the fixed polarized azimuthal angle P and the fixed analyzer azimuthal angle A are all in the ideal state, that is P=π/4, A=0, π/2, and when θ=75.55 degree. In the process of real measurement, due to the deviation in the measuring device and the light road, this condition is difficult to be satisfied. Thus, when there is deviation in polarized azimuthal angle, incident angle and the fixed analyzer azimuthal angle, the calibration action must be taken on the deviation azimuthal angle, so that to define the real fixed polarized azimuthal angle, incident angle and the fixed analyzer azimuthal angle, and based on that to obtain the real thickness of the target surface.
In this case, based on the thickness of the target surface measurement method also includes steps S20, as shown in
Below is the description of the calibration process for the polarized azimuthal angle, the analyzer azimuthal angle and the incident angle shown in the
Before the calibration, firstly the target surface 5 shall be changed to be the reference target surface 5′ used for the calibration, preferably, this reference target surface shall be SiO2 film.
Calibration process works as follows, supposing there is azimuthal angle deviation between the polarizer and the analyzer, assumed as P+δP, A+δA, where P+δP represents the deviation between the actual azimuthal angle and the current theoretical value (eg, π/4), δA represents the deviation between the actual analyzer azimuthal angle A and the current theoretical value (eg, 0 and π/2). In this case, the output light intensity after the polarizer 3 and analyzer 7 is as the following equation:
In which,
represents the theoretical Fourier coefficient, Δ is the phase deviation between the p and s components of the characteristics light.
The theoretical value of the polarized azimuthal angle is P=π/4, deviation is δP, theoretical value of the analyzer azimuthal angle is A=0, π/2, the deviation is δA, so the two light intensity are,
I
1
=I(0)=I01(1+α′cos 2δA+β′ sin 2δA)
I
2
=I(π/2)=I01(1−α′ cos 2δA−β′ sin 2δA) (4)
Based on the formula (4), the light intensity can be calculated as below,
I
01=(I1+I2)/2
After the normalization,
I
1′=1+α′ cos 2δA+β′ sin 2δA
I
2′=1−α′ cos 2δA−β′ sin 2δA (5)
So that the objective function is,
In which, t is the thickness of the surface, θ is the incident angle, δP is the deviation of the polarizer azimuthal angle of the polarizer 3, δA is the deviation of the analyzer azimuthal angle of the analyzer 7, {right arrow over (n)} is the refractive index of the reference surface, {right arrow over (k)} is the extinction coefficient for the reference surface, λ is the wavelength. On the right, I′ij is the normalized obtained light intensity, Iijtheory is the theoretical calculating light intensity. In which, i is the p and s component, j is the changing of the reference surface.
When process the calibration, suppose to use m reference surfaces with different known thickness, the wavelength of the using light is constant, so the {right arrow over (n)}, {right arrow over (k)} are also constant, X has the 2m item, after remove the linear dependencies related items from it, there are m items could be used. There are 3 unknown items including the polarized azimuthal angle δP, the incident angle θ and the analyzer azimuthal angle δA, based on the basic calculating method, all the unknown parameters could be solved only when m>3.
Known in this field, if one or more of the polarizer 3 azimuthal angle P+δP, incident angle θ, and the analyzer azimuthal angle A+δA are known, then in the calibration step, other azimuthal angle will be calibrated. Since the unknown parameters are not so many, then the required number of nonlinear equations and then the corresponding measurements on the reference surface time reduced.
Above is the description and support theory for the principle of the calibration method provide by this invention, below introduce all steps in the calibration process, referred to the
Firstly, in the step S201, polarizer 3 polarizes the light in the fixed polarized azimuthal angle P+δP to get the polarized light.
Then, in the step S202, polarized light illuminate onto the known reference surface 5′ in the incident angle θ.
Then, in the step S203, analyzer 7 analyzes the characteristic light reflected from the reference surface 5′ in a fixed analyzed azimuthal angle A+δA, and analyzes the p and s components of the characteristic light.
Then, in the step S204, detection-processing device 8 detects the light intensity of the p and s components of the characteristics light, to obtain one set of light intensity information corresponding to reference surface 5′.
In which, according to equation (6), the thickness of the reference surface 5′, the deviation of the fixed polarizer azimuthal angle δP, the incident angle θ, the fixed polarizer azimuthal angle deviation δA, and the refractive index {right arrow over (n)} of the reference surface 5′, the light extinction coefficient {right arrow over (k)} of the reference surface 5′ and, The wavelength λ is variable, With the numerical approximation principle, assume the light intensity of the detected p and s components and the sum of the square of their corresponding theoretical light intensity deviation to be the objective function, to establish a nonlinear equation for the reference surface 5′.
Next, in the step S205, system bases on the relationship between foreside function number and the unknown parameters number, to decide the nonlinear equations corresponding to the reference surface 5′, whether is able to conduct all the unknown parameters in the fixed polarized azimuthal angle P+δP this incident angle θ and this fixed analyzer azimuthal angle A+δA. If this could not conduct, the reference surface 5′ shall be changed to be another reference surface 5″, and then repeat the steps S201 to S205, until all the unknown azimuthal angle parameters in the fixed polarized azimuthal angle P+δP, this incident angle θ and this fixed analyzer azimuthal angle A+δA could be conducted by several nonlinear equations corresponding to several reference surfaces.
When it is possible to conduct all the unknown azimuthal angle parameters according to one or more nonlinear equations corresponding to reference surface 5′ (and 5″ so on.), this method shall start the step 26. Based on one or more nonlinear equations corresponding to one or more reference surface, use the relationship curve between the light intensity from nonlinear optimization theory methods and surface characteristics information, conduct all the unknown parameters in the polarized azimuthal angle P+δP, this incident angle θ and could this fixed analyzer azimuthal angle A+δA.
The calculating method could be the non-linear optimization method L-M, also could be any other method that could work the non-linear equations, the detailed conducting process is depends on the current available method, not repeat. The calculating process could be finished by the calibration device 10, this calibration device could be a micro-process device, work as the programming process to finish this calibration process, or this calibration device could be a corresponding firmware, ASIC or DSP devices.
Not related to the above steps S201 to S205, similar with the steps S11 to S14, process the steps S11′ to S14′, as shown in
After determine the fixed polarized azimuthal angle P+δP, this incident angle θ and this fixed analyzer azimuthal angle A+δA, in the step S15′, Based on the light parameters of the reference surface 5, {right arrow over (n)} and {right arrow over (k)} are known, and the detected azimuthal angle parameters, put them into the equation (6), use the L-M method to result the thickness. Of course, other nonlinear functions could be used to result this thickness, the detailed resulting method is not the focus in this invention. The calculating process could be finished by the calculating device 9, this calculating device shall be a micro processing device, work as the programming process to finish this calibration process, or this calibration device could be a corresponding firmware, ASIC or DSP devices.
As the improvement for the first embodiment, between the surface and the polarizer of the light road, one or more dividers, to divide the characteristics light from the reflection of the surface, then polarize in a fixed azimuthal angle and detect the light intensity of the dividing of several lights. According to the light intensity of the p and s components of all the dividing light, obtain the characteristic information of the target surface detected by the electromagnetic wave.
In which, the light source 1 use the He—Ne laser of the output light wavelength 632.8 nm (nanometers), in addition, technical staff in this field can understand, the light source 1 is not limited to this one, but also for other high-power red LED light sources. The role of small hole 2 is used for the better focus of the incident light, as known in technical staff in this field, hole 2 is not necessary for this invention. The role of hole 4 includes the avoidance of partial output from the other polarized light beam into the measurement light path, so that to avoid its impact on the measurement results, and similarly, as known in technical staff in this field, hole 4 is also not necessary for this invention. Usually, the use of small hole 2 and 4 help to improve measurement accuracy of the surface characteristics information by the electromagnetic waves
According to the light road map, it works as below,
Firstly, light source 1 illuminates a light, goes through the hole 2, focused by hole 2 and then illuminates onto the polarizer 3.
In step S31, the polarizer 3 polarizes a single wavelength light in a fixed polarizer azimuthal angle P to generate a polarized light.
Then, the polarized light goes through the hole 4, the hole 4 only allows one polarized light goes through by polarizer 3, and stop the other bunch of polarized lights in order to avoid its impact on the measurement. Thus, in step S32, the polarized light irradiated in the incident angle θ onto the target surface 5. As known of technique staff in this field, without hole 4, the light road of the polarized light from polarizer 3 may has quite difference with the light road of the polarized light from hole 4 in
The illuminated polarized light reflects to analyzer 7 by target surface 5, in which the nature of this polarized light has been changed. In the step 33, light divider 6 divides the characteristics light from the target surface, to get the corresponding dividing light L1 and L2, and then provide them to two analyzer 7a and 7b
Then, in the step S34, the analyzer 7a analyzes the characteristics light L1 from the target surface in the fixed analyzer azimuthal angle, separate the p1 and s1 component and the p and s component of the L1, and then provide to the detection-processing device 8a, the analyzer 7b analyzes the characteristics light L2 from the target surface in the fixed analyzer azimuthal angle, separate the p2 and s2component and the p and s component of the L2, and then provide to the detection-processing device 8b.
Then in step S35, the detection-processing devices 8a and 8b receive and detect the light intensity of the p1 and s1, p2 and s2 component, obtain one light intensity information corresponding to the target surface, and obtain the phase deviation between the p1 and s1 component, and the phase deviation between the p1 and s1 component, and then provide this light intensity information to the calculation device 9 which has electromagnetic connection with the detection-processing device 8 and is not shown in
Finally, in the step S36, the calculating device 9 based on the light intensity of the p1 and s1 component and the p2 and s2 component, determines the thickness of the target surface by the preordered method. It is can be understood by technical staff in this field, there are several ways to determine the thickness information, this is not limited by this embodiment. In this embodiment, the principle of this scheduled method is as follows,
The basic formula for the output light intensity after the polarizer and the analyzer is,
When the light goes through the analyzer 7a, P=π/4, A=A1=0, π/2, tan Ψ is the amplitude ratio of the p and s component of the light, Δ is the phase deviation between the p and s component. The detection-processing device obtained light intensity I(π/4,0)out, I(π/4,π/2)out, can be used to calculate the Fourier coefficient
and can calculate the ellipsometric parameters tan Ψ, that is:
When the light goes through analyzer 7b P=π/4, A=A2=±π/4, obtain the detection-processing device collecting light intensity I(π/4,π/4)out, I(π/4,−π/4)out. Besides,
stands for the first-order Fourier coefficient. So,
In this invention, the Fourier coefficient can be calculated before head. The variation situation of the Fourier coefficient α, β for different thickness is similar with the first embodiment, when the incident angle θ is 75.55 degree, the helium-neon laser outputs the incident light of 632.8 nm wavelength, when the thickness of the surface changed from 0 to 500 nm, the variation situation of the Fourier coefficient shows as
b shows the Fourier coefficients of the curve, when the wavelength of the illuminate light is 632.8 nm, the incident angle of the polarized light is 75.55 degree, when the thickness of the surface changes in the range of 0 to 50 nm. It can be figured out that when the thickness is within 0 to 50 nm, the Fourier coefficient is the monotonic function of the surface thickness.
a shows the ellipsometry parameters of the curve, when the wavelength of the illuminate light is 632.8 nm, the incident angle of the polarized light is 75.55 degree, when the thickness of the surface changes in the range of 0 to 500 nm. It can be figured out that, when the thickness is within 0 to 50 nm, the ellipsometric parameters is the periodic function of the surface thickness.
b shows the ellipsometry parameters of the curve, when the wavelength of the illuminate light is 632.8 nm, the incident angle of the polarized light is 75.55 degree, when the thickness of the surface changes in the range of 0 to 50 nm. It is can be figured out from
When process the measurement, the Fourier coefficient can be calculated based on the formula (8). When all the azimuthal angles are fixed, including the polarized azimuthal angle, analyzer azimuthal angle and the incident angle between the polarized light and the target surface, and the wavelength of the laser is known, meanwhile the refractive index of the target surface is known, all the Fourier coefficient α, β are all known, then the control method can be used, find the most similar Fourier coefficients within the range of values from the known Fourier coefficients, then the thickness of this Fourier coefficients is the thickness of the target surface.
In theory, the Fourier coefficients α, β effects by the thickness, azimuthal angle θ, wavelength, light constants of the material target surface, the polarizer azimuthal angle P and the analyzer azimuthal angle. In which, the thickness, azimuthal angle θ, wavelength, light constants of the material target surface effects the ellipsometry parameters, and the ellipsometry effects the Fourier coefficients. Use the fixed wavelength light source, since the light constants of the material is fixed, the effects on the measurement result can be deleted. Thus, the thickness of the target surface is only decided by the Fourier coefficients α, β. Based on this principle, in a single range, one Fourier coefficients is corresponding to one thickness. Thus, according to the correspondence between them, the thickness can be calculated.
This embodiment shows only when the light source 1 with the wavelength 632.8 nm, the incident angle of the polarized light is 75.55 degree, the function relationship between the Fourier coefficients and the ellipsometry parameters and the thickness. It is widely understood by the technical staff in this field, when use light has other wavelength and incident angle, the function relationship between the Fourier coefficients α, β and the ellipsometry parameters and the thickness will be changed, and this invention is suitable for the reality function relationship.
Technical staff in this filed understand that, since the Fourier coefficients has the one to one relationship with the reflectivity, refractive index, extinction coefficient, polarization, surface microstructure, particles, defects and other surface roughness features and so on, besides the thickness of the target surface, thus, the method in this invention is not limited to the thickness of the target surface, other characteristics can be measured also, and, by the introduction of this file, the technical staff in this field can understand that this invention can also used to measure other kind of target surface and its characteristics information. Based on above, for the sake of to be simple, this file will not show further on other situations.
Above shows the measurement method when the fixed polarized azimuthal angle P and the fixed analyzer azimuthal angle A1 and A2 are all in the ideal state, that is P=π/4, A1=0, π/2, A2=±π/4, and when θ=75.55 degree. In the process of real measurement, due to the deviation in the measuring device and the light road, this condition is difficult to be satisfied. Thus, when there is deviation in polarized azimuthal angle P, incident angle θ and the fixed analyzer azimuthal angle A1 and A2, Thus, when there is one or more deviation between this fixed analyzer A1 and A2, fixed polarized azimuthal angle P and the incident angle θ, the calibration action must be taken to confirm the fixed polarized azimuthal angle P and the incident angle θ, based on this fixed polarized azimuthal angle P and the incident angle θ to obtain the thickness of the target surface.
In this case, based on the thickness of the target surface measurement method also includes step S40, as shown in
Specifically, below will describe the method of calibration for the polarized azimuthal angle, analyzer azimuthal angle and the incident angle, and based on the obtained azimuthal angles and the obtained light intensity of the p and s components of the target surface, so that to process the thickness measurement.
Before the calibration, firstly change the target surface 5 in
Calibration process works as follows, suppose there is azimuthal angle deviation between the polarizer 3 and the analyzer 7a and 7b, assumed as P+δP, A1+δA1 and A2+δA2 , where δP represents the deviation between the actual azimuthal angle P and the current theoretical value (eg, π/4), δA1 and δA2 represents the deviation between the actual analyzer azimuthal angle A and the current theoretical value (eg, 0 and π/2, π/4 and −π/4). The output light intensity is as the following equation:
In which,
represents the theoretical Fourier coefficient, Δ is the phase deviation between the p and s component for the characteristic light.
The theoretical value of the polarized azimuthal angle is P=π/4, deviation is δP, theoretical value of the analyzer 7a azimuthal angle A1 is 0°/90°, the deviation is δA1, theoretical value of the analyzer 7b azimuthal angle A2 is 45°/−45°, the deviation is δA1, so the 4 light intensity are,
I
1
=I(0)=I01(1+α′ cos 2δA1+β′ sin 2δA1)
I
2
=I(π/2)=I01(1−α′ cos 2δA1−β′ sin 2δA1)
I
3
=I(π/4)=I02(1−α′ sin 2δA2+β′ cos 2δA2)
I
4
=I(−π/4)=I02(1+α′ sin 2δA2−β′ cos 2δA2) (11)
After the normalization,
So that the objective function is,
In which, {right arrow over (t)} is the thickness of the surface, 0 is the incident angle, δP is the deviation of the polarizer azimuthal angle of the polarizer 3, δA1 is the deviation of the analyzer azimuthal angle of the analyzer 7a, δA2 is the deviation of the analyzer azimuthal angle of the analyzer 7b, {right arrow over (n)} is the refractive index of the reference surface, {right arrow over (k)} is the extinction coefficient for the reference surface, is the wavelength. On the right, I′ij is the normalized obtained light intensity, Iijtheory is the theoretical calculating light intensity. In which, i is the p1 and s1, p2 a and s2component, j is the changing of the reference surface.
When process the calibration, suppose to use m reference surfaces with different known thickness, the wavelength of the using light is constant, so the {right arrow over (n)}, {right arrow over (k)} are also constant, X has the 4m item, after remove the linear dependencies related items from it, there are 2m items could be used. There are m+4 unknown items including the polarized azimuthal angle δP, δA, δA2, the incident angle θ, m unknown thickness of the reference target surfaces, based on the basic calculating method, all the unknown parameters could be solved only when 2m>m+4, that is m>4.
Known in this field, if one or more of the polarizer 3 azimuthal angle P+δP, incident angle θ, and the analyzer azimuthal angle A1+δA1 and A2+δA2 of the analyzer 7a and 7b is known, then in the calibration step, other azimuthal angle will be calibrated. Since the unknown parameters are not so many, then the required number of nonlinear equations and then the corresponding measurements on the reference surface time reduced.
Above is the description and support theory for the principle of the calibration method provide by this invention, below introduce all steps in the calibration process, referred to the
Firstly, in the step S401, polarizer 3 polarizes the light in the fixed polarized azimuthal angle P+δP to get the polarized light.
Then, in the step S402, polarized light illuminate onto the known reference surface 5′ in the incident angle θ.
Then, in the step S404, dividing device 6 divides the characteristics light reflected from the reference surface 5′, obtained 2 lights L1 and L2
Then, in the step S404, analyzer 7a and 7b analyzes the characteristic light in a fixed analyzed azimuthal angle A1+δA1 and A2+δA2, and analyzes the p1 and s1, p2 and s2 components of the characteristic light.
Then, in the step S405, detection-processing device 8a and 8b detects the light intensity of the p1and s1, p2 and s2 components of the characteristics light, to obtain one light intensity information corresponding to reference surface 5′.
In which, according to equation (13), the thickness of the reference surface 5′, the fixed polarizer azimuthal angle P+δP, the incident angle δ, the fixed analyzer azimuthal angle A1+δA1 and A2+δA2, and the refractive index {right arrow over (n)} of the reference surface 5′, the light extinction coefficient {right arrow over (k)} of the reference surface 5′ and the wavelength λ are variable, With the numerical approximation principle, assume the light intensity of the detected p and s components and the sum of the square of their corresponding theoretical light intensity deviation to be the objective function, to establish a nonlinear equation for the reference surface 5′.
Next, in the step S406, bases on the relationship between foreside function number and the unknown parameters number, to decide the nonlinear equations corresponding to the reference surface 5′, whether is able to conduct all the unknown parameters in the fixed polarized azimuthal angle P+δP, this incident angle δ and could this fixed analyzer azimuthal angle A+δA. If this could not conduct, the reference surface 5′ shall be changed to be another reference surface 5″, and then repeat the steps S401 to S406, until all the unknown azimuthal angle parameters in the fixed polarized azimuthal angle P+δP, this incident angle θ and could this fixed analyzer azimuthal angle A1+δA1 and A2+δA2 could be conducted by several nonlinear equations corresponding to several reference surfaces.
When it is possible to conduct all the unknown azimuthal angle parameters according to one or more nonlinear equations (13) corresponding to reference surface 5′ (and 5″ so on.), this method shall start the step S407. In which, based on one or more nonlinear equations corresponding to one or more reference surface, use the relationship curve between the light intensity from nonlinear optimization theory methods and surface characteristics information, conduct all the unknown parameters in the polarized azimuthal angle P+δP, this incident angle θ and this fixed analyzer azimuthal angle A1+δA1 and A2+δA2. The calculating method could be the non-linear optimization method L-M, also could be any other method that could work the non-linear equations, the detailed conducting process is depends on the current available method, not repeat. The calibration device 11 could finish the calculating process.
Not related to the above steps S401 to S407, similar with the steps S31 to S35, process the steps S31′ to S35′, as shown in
After determine the fixed polarized azimuthal angle P+δP, incident angle θ and this fixed analyzer azimuthal angle A1+δA, and A2+δA2 , in the step S36′, based on the light parameters of the reference surface 5, {right arrow over (n)} and {right arrow over (k)} are known, and the detected azimuthal angle parameters, put them into the equation (13), use the L-M method to result the thickness. The calculating process could be finished by the calculating device 9. Of course, other nonlinear functions could be used to result this thickness, the detailed resulting method is not the focus in this invention.
In a variable embodiment, the ellipsometry measurement device puts one or more characteristic information and the ellipsometric parameters, such as Ψ and Δ into the measurement database. Specifically, after the system calibration with above method, the obtained characteristic information of several target surfaces in same conditions and their corresponding ellipsometric parameters are reserved by their corresponding relationship. In another situation, this corresponding relationship may be conducted from the outside. Then, when divide the characteristic information of the target surface by the electromagnetic waves, the azimuthal angle of the ellipsometric measurement remain constant, use the obtained corresponding ellipsometric parameters corresponding to the current target surface, based on the corresponding relationship between the ellipsometric parameters and the characteristic information pre-reserved in the database, the characteristic information could be obtained from this corresponding relationship by the obtained ellipsometric parameters.
Above is the method, after using the target reflection change the ellipsometric state of the incident light, after the measurement for the characteristic information of the reflected characteristic light, of the characteristic information confirmation. This invention is not limited of the method of using the target reflection change the ellipsometric state of the incident light, besides, the using the target Transmittance or diffraction change the ellipsometric state of the incident light, to get the corresponding characteristic light information then get the target characteristics.
It is understood that, this invention is not limited with using the reflection only, using the transmission only, and using the diffraction only. It is be understood that this invention can be used to obtain one polarized characteristics light by reflection or transmission or diffraction, and then based on this characteristics light to obtain the characteristics information for the target surface by the electromagnetic waves. By the instruction of this invention, the calibration and measurement steps of the ellipsometric measurement method could be proved, for the sake of simplification, this instruction will not show more about this using situation of this invention.
Above is the description about the detailed processing methods of this invention. But this invention is not limited of one of above processing methods, all kinds of variations or changing could be used in the claims scope.
Number | Date | Country | Kind |
---|---|---|---|
200810042435.7 | Sep 2008 | CN | national |
Filing Document | Filing Date | Country | Kind | 371c Date |
---|---|---|---|---|
PCT/CN2009/073713 | 9/3/2009 | WO | 00 | 3/2/2011 |